Contents: Volume 101, Issue 5 (November 2005)

CARDIOVASCULAR ANESTHESIA:
c
d
e
f
g

Detlef Obal, Saskia Dettwiler, Christian Favoccia, Horst Scharbatke, Benedikt

Preckel, and Wolfgang Schlack
The Influence of Mitochondrial KATP-Channels in the Cardioprotection of
Preconditioning and Postconditioning by Sevoflurane in the Rat In Vivo
Anesth Analg 2005 101: 1252-1260.
IMPLICATIONS: Sevoflurane protects against myocardial ischemiareperfusion injury when given before (preconditioning) or after
(postconditioning) myocardial ischemia. Both effects provide additive
cardioprotection and are mediated at least in part by mitochondrial KATPchannels.
c
d
e
f
g

Guochang Hu, M. Ramez Salem, and George J. Crystal

Isoflurane Prevents Platelets from Enhancing Neutrophil-Induced
Coronary Endothelial Dysfunction
Anesth Analg 2005 101: 1261-1268.
IMPLICATIONS: Isoflurane prevented the ability of platelets to enhance the
coronary endothelial dysfunction caused by neutrophils. This effect was
associated with reductions in superoxide production and vascular adherence.
The current findings provide further evidence that the cardioprotective effects
of isoflurane involve inhibitory actions on inflammatory cells and their
interactions.

c
d
e
f
g

Michael F. Haney, Gran Johansson, Sren Hggmark, Bjrn Biber, Michael F.

Haney, Gran Johansson, Sren Hggmark, and Bjrn Biber
Myocardial Systolic Function Increases During Positive Pressure Lung
Inflation
Anesth Analg 2005 101: 1269-1274.
IMPLICATIONS: We demonstrated that positive pressure inflation of the
lungs brings about a very rapid increase in myocardial contractile function, in
a large animal model. Therefore, positive airway pressure and lung inflation
status must be standardized for serial measurements of myocardial contractile
function.

c
d
e
f
g

Leo G. Kevin, Enis Novalija, and David F. Stowe

Reactive Oxygen Species as Mediators of Cardiac Injury and Protection:
The Relevance to Anesthesia Practice
Anesth Analg 2005 101: 1275-1287.

c
d
e
f
g

Rebecca A. Schroeder and Jonathan B. Mark

Is the Valve OK or Not? Immediate Evaluation of a Replaced Aortic
Valve
Anesth Analg 2005 101: 1288-1291.
IMPLICATIONS: Although transesophageal echocardiography is a crucial
tool in evaluation newly implanted prosthetic valves, it is important to be
familiar with factors that confound echocardiographic data. Failure in this
regard may lead to unwarranted surgical intervention.

Paul C. Whiting and Nicholas J. Morgan-Hughes

Transesophageal Echocardiographic Findings in Papillary Muscle
Rupture
Anesth Analg 2005 101: 1292-1293.

PEDIATRIC ANESTHESIA:
c
d
e
f
g

Barry D. Kussman, David Wypij, James A. DiNardo, Jane Newburger, Richard

A. Jonas, Jodi Bartlett, Ellen McGrath, and Peter C. Laussen
An Evaluation of Bilateral Monitoring of Cerebral Oxygen Saturation
During Pediatric Cardiac Surgery
Anesth Analg 2005 101: 1294-1300.
IMPLICATIONS: Bi-hemispheric measurements of regional cerebral oxygen
saturation using near-infrared spectroscopy were similar in neonates and
infants undergoing biventricular repair without aortic arch reconstruction.
Further longitudinal neurological outcome studies are required to determine
whether uni- or bi-hemispheric monitoring is required in this patient
population.
c
d
e
f
g

Jin-Hee Kim, Chong-Sung Kim, Jae-Hyun Bahk, Kyung Joon Cha, Young-Sun
Park, Young-Tae Jeon, and Sung-Hee Han
The Optimal Depth of Central Venous Catheter for Infants Less Than 5
kg
Anesth Analg 2005 101: 1301-1303.

IMPLICATIONS: To create a guideline to avoid intracardiac placement of

central venous catheters in infants, we examined the distance from skin to
superior vena cava and right atrial junction by using transesophageal
echocardiography. The recommended depth of the central venous catheter
inserted through right subclavian vein is between 40-55 mm for infants less
than 5 kg.
c
d
e
f
g

Kha M. Tran, Theodore J. Ganley, Lawrence Wells, Arjunan Ganesh, Kimberly I.

Minger, and Giovanni Cucchiaro
Intraarticular Bupivacaine-Clonidine-Morphine Versus Femoral-Sciatic
Nerve Block in Pediatric Patients Undergoing Anterior Cruciate
Ligament Reconstruction
Anesth Analg 2005 101: 1304-1310.
IMPLICATIONS: Methods to control pain after knee surgery include blocking
the femoral nerve and injecting medications into the joint. Neither technique is
clearly superior. We tested the hypothesis that blocking the sciatic and femoral
nerves provides better analgesia than joint injection. Postoperatively, blocked
patients had less pain and required less medication.
c
d
e
f
g

Stephan Schubert, Gisela Stoltenburg-Didinger, Anke Wehsack, Dirk Troitzsch,

Wolfgang Boettcher, Michael Huebler, Matthias Redlin, Majid Kanaan, Michael
Meissler, Peter E. Lange, and Hashim Abdul-Khaliq
Large-Dose Pretreatment with Methylprednisolone Fails to Attenuate
Neuronal Injury After Deep Hypothermic Circulatory Arrest in a
Neonatal Piglet Model
Anesth Analg 2005 101: 1311-1318.
IMPLICATIONS: A different effect of steroid treatment could be suspected in
the neonatal age, especially in the context of ischemia or hypoxia, and the use
of steroids has to be reconsidered in critically ill neonates and infants.

AMBULATORY ANESTHESIA:
c
d
e
f
g

Brian M. Ilfeld, Thomas W. Wright, F. Kayser Enneking, Jennie A. Mace,

Jonathan J. Shuster, Eugene H. Spadoni, Terese L. Chmielewski, and Krista
Vandenborne
Total Shoulder Arthroplasty as an Outpatient Procedure Using
Ambulatory Perineural Local Anesthetic Infusion: A Pilot Feasibility
Study
Anesth Analg 2005 101: 1319-1322.

IMPLICATIONS: The results of this pilot study suggest that it is possible to

convert total shoulder arthroplasty into an outpatient procedure using
ambulatory interscalene perineural local anesthetic infusion. Additional data
are required to define the appropriate subset of patients and assess the
incidence of complications associated with this practice before its mainstream
use.
c
d
e
f
g

Tong J. Gan, Andrew Coop, Beverly K. Philip, and the Kytril Study Group
A Randomized, Double-Blind Study of Granisetron Plus Dexamethasone
Versus Ondansetron Plus Dexamethasone to Prevent Postoperative
Nausea and Vomiting in Patients Undergoing Abdominal Hysterectomy
Anesth Analg 2005 101: 1323-1329.
IMPLICATIONS: This randomized, double-blind study evaluated whether
small-dose granisetron (0.1 mg) plus dexamethasone 8 mg was as effective as
ondansetron 4 mg plus dexamethasone 8 mg for preventing vomiting during
the 0 to 2 h after tracheal extubation in patients undergoing abdominal
hysterectomy requiring general anesthesia.

c
d
e
f
g

Arash Pirat, Senay F. Tuncay, Adnan Torgay, Selim Candan, and Gulnaz Arslan
Ondansetron, Orally Disintegrating Tablets Versus Intravenous Injection
for Prevention of Intrathecal Morphine-Induced Nausea, Vomiting, and
Pruritus in Young Males
Anesth Analg 2005 101: 1330-1336.
IMPLICATIONS: The results of this study indicate that neither ODT
ondansetron 8 mg nor IV ondansetron 4 mg is more effective than placebo as
prophylaxis for IT morphine-induced postoperative nausea and vomiting in
young males. Although both regimens of ondansetron were associated with
less frequent IT morphine-induced pruritus than placebo, only the ODT form
significantly reduced the severity of pruritus and significantly decreased the
need for rescue antipruritic administration.

c
d
e
f
g

Ronald P. Olson, Alan Stone, and David Lubarsky

The Prevalence and Significance of Low Preoperative Hemoglobin in
ASA 1 or 2 Outpatient Surgery Candidates
Anesth Analg 2005 101: 1337-1340.
IMPLICATIONS: In a large population of low risk surgical patients,
asymptomatic anemia rarely occurred, and the discovery of this anemia did not
change perioperative management. Routine hemoglobin screening for anemia
is therefore not indicated in these patients.

ANESTHETIC PHARMACOLOGY:
Martin Redmond and Peter Glass
Opiate-Induced Nausea and Vomiting: What Is the Challenge?
Anesth Analg 2005 101: 1341-1342.
c
d
e
f
g

Gregory W. Roberts, Tenna B. Bekker, Helle H. Carlsen, Christine H. Moffatt,

Peter J. Slattery, and Anna F. McClure
Postoperative Nausea and Vomiting Are Strongly Influenced by
Postoperative Opioid Use in a Dose-Related Manner
Anesth Analg 2005 101: 1343-1348.
IMPLICATIONS: There is a strong relationship between the amount of
postoperative opioid used and postoperative vomiting. The accuracy of various
risk scoring approaches may have been undermined by not allowing for this
relationship. Patients likely to have larger postoperative opioid requirements
should be priority targets for opioid reduction or vomiting prevention
strategies.
c
d
e
f
g

Ching-Rong Cheng, Daniel I. Sessler, and Christian C. Apfel

Does Neostigmine Administration Produce a Clinically Important
Increase in Postoperative Nausea and Vomiting?
Anesth Analg 2005 101: 1349-1355.
IMPLICATIONS: Neostigmine does not increase postoperative vomiting.
However, data on the effect on postoperative nausea remain conflicting and
there is insufficient evidence from which to draw strong conclusions.

c
d
e
f
g

Mathias Sluga, Wolfgang Ummenhofer, Wolfgang Studer, Martin Siegemund,

and Stephan C. Marsch
Rocuronium Versus Succinylcholine for Rapid Sequence Induction of
Anesthesia and Endotracheal Intubation: A Prospective, Randomized
Trial in Emergent Cases
Anesth Analg 2005 101: 1356-1361.
IMPLICATIONS: During rapid sequence induction of anesthesia in emergent
cases, succinylcholine allowed for a more rapid endotracheal intubation
sequence and created superior intubation conditions compared with
rocuronium. However, there was no difference in the incidence of poor
intubation conditions, failed intubation attempts, and desaturations between
succinylcholine and rocuronium.

c
d
e
f
g

Heidrun Fink, Ralph Bogdanski, Peter Luppa, J. A. Jeevendra Martyn, and

Manfred Blobner
Nitric Oxide Is Not a Mediator of Inflammation-Induced Resistance to
Atracurium
Anesth Analg 2005 101: 1362-1367.
IMPLICATIONS: Increased {alpha}1-acid glycoprotein levels during
systemic inflammatory response leads to resistance towards the neuromuscular
blocking effects of atracurium. Nitric oxide may modulate {alpha}1-acid
glycoprotein synthesis. Suppression of increased nitric oxide levels, however,
did not successfully reverse the resistance to atracurium, implying a nitric
oxide-independent regulation of {alpha}1-acid glycoprotein.
c
d
e
f
g

Ryu Komatsu, Papiya Sengupta, Grigory Cherynak, Anupama Wadhwa, Daniel I.

Sessler, Jin Liu, Harrell E. Hurst, and Rainer Lenhardt
Doxapram Only Slightly Reduces the Shivering Threshold in Healthy
Volunteers
Anesth Analg 2005 101: 1368-1373.
IMPLICATIONS: The observed reduction of only 0.5{degrees}C in the
shivering threshold explains doxapram's efficacy for treatment of
postoperative shivering. However, such a reduction is unlikely to markedly
facilitate induction of therapeutic hypothermia as a sole drug.
c
d
e
f
g

Michael T. Pawlik, Ernil Hansen, Daniela Waldhauser, Christoph Selig, and

Thomas S. Kuehnel
Clonidine Premedication in Patients with Sleep Apnea Syndrome: A
Randomized, Double-Blind, Placebo-Controlled Study
Anesth Analg 2005 101: 1374-1380.
IMPLICATIONS: Patients with sleep apnea syndrome undergoing general
anesthesia usually receive no premedication drugs before surgery because of
the potential respiratory-depressing effects. Clonidine, an antihypertensive
used as premedication, stabilizes hemodynamics during anesthesia and
improves postoperative analgesia without disturbing the ventilation pattern in
patients with sleep apnea syndrome.
c
d
e
f
g

Koji Hara, Tomohiro Yamakura, Takeyoshi Sata, and R. Adron Harris

The Effects of Anesthetics and Ethanol on 2 Adrenoceptor Subtypes
Expressed with G Protein-Coupled Inwardly Rectifying Potassium
Channels in Xenopus Oocytes
Anesth Analg 2005 101: 1381-1388.

IMPLICATIONS: The {alpha}2-adrenoceptor (AR) subtypes and

GIRK1/GIRK2 channels were co-expressed in Xenopus oocytes. IV
anesthetics, ethanol, and halothane had no direct effect on any of the
{alpha}2-AR subtypes at clinically relevant concentrations. However, the
physiological actions of the {alpha}2-ARs mediated by GIRK1/GIRK2
channels may be affected by ethanol and halothane.
c
d
e
f
g

Catherine Crosby, Deborah J. Culley, Mark G. Baxter, Rustam Yukhananov, and

Gregory Crosby
Spatial Memory Performance 2 Weeks After General Anesthesia in Adult
Rats
Anesth Analg 2005 101: 1389-1392.
IMPLICATIONS: Two hours of isoflurane anesthesia, with or without nitrous
oxide, does not impair and may improve the ability of rats to acquire a novel
spatial memory task 2 wk later.

TECHNOLOGY, COMPUTING, AND

SIMULATION:
c
d
e
f
g

Mats K. E. B. Wallin, Therese Marve, and Peter K. Hakansson

Modern Wireless Telecommunication Technologies and Their
Electromagnetic Compatibility with Life-Supporting Equipment
Anesth Analg 2005 101: 1393-1400.
IMPLICATIONS: Universal Mobile Telecommunication Systems (UMTS)
(WCDMA FDD), Wireless Local Area Network (WLAN) (802.11b), and
General Packet Radio Services are three new wireless technologies that could
potentially replace hospital pager systems. Electromagnetic provocation tests,
between signals and life-supporting equipment, were performed in a clinical
setting. The study showed that signals, from UMTS and WLAN, are
electromagnetically compatible with the tested equipment.

c
d
e
f
g

Huei-Ming Yeh, Mei-Chuan Tsai, Yi-Ning Su, Rong-Ching Shen, Jeuy-Jen

Hwang, Wei-Zen Sun, and Ling-Ping Lai
Denaturing High Performance Liquid Chromatography Screening of
Ryanodine Receptor Type 1 Gene in Patients with Malignant
Hyperthermia in Taiwan and Identification of a Novel Mutation (Y522C)
Anesth Analg 2005 101: 1401-1406.
IMPLICATIONS: We screened for skeletal muscle ryanodine receptor
mutations in five Taiwanese patients with malignant hyperthermia (MH) using
denaturing high-performance liquid chromatography. We identified the
responsible mutations in all patients. This approach has the potential for the
development of better genetic tests for MH in the future.

c
d
e
f
g

Samsun Lampotang, Nikolaus Gravenstein, David A. Paulus, and Dietrich

Gravenstein
Reducing the Incidence of Surgical Fires: Supplying Nasal Cannulae with
sub-100% O2 Gas Mixtures from Anesthesia Machines
Anesth Analg 2005 101: 1407-1412.
IMPLICATIONS: In a June, 2003 Sentinel Event Alert concerning prevention
of surgical fires, the Joint Commission on Accreditation of Healthcare
Organizations recommended, as a general policy and consistent with patient
needs, supplying open delivery systems such as nasal cannulae with air or
<=30% O2. This report analyzes equipment, practice alternatives, and patient
safety considerations that address this recommendation.
c
d
e
f
g

Ramachandra R. Avula, Richard Kramer, and Charles E. Smith

Air Detection Performance of the Level 1 H-1200 Fluid and Blood
Warmer
Anesth Analg 2005 101: 1413-1416.
IMPLICATIONS: The air detector/clamp of the Level 1 H-1200 fluid and
blood warmer functioned effectively during simulated minor and massive air
embolism and would be expected to prevent lethal air embolism. The use of
ultrasonic air detection coupled with automatic shutoff is a significant safety
improvement of the Level 1 H-1200 fluid and blood warmer.

PAIN MEDICINE:
c
d
e
f
g

Myung Ha Yoon, Hong Beom Bae, and Jeong Il Choi

Antinociception of Intrathecal Adenosine Receptor Subtype Agonists in
Rat Formalin Test
Anesth Analg 2005 101: 1417-1421.
IMPLICATIONS: Intrathecal adenosine A1 and A2A receptor agonists
reduced acute pain and inflammatory hyperalgesia in the formalin test.
Further, the adenosine A1 receptor agonist was more powerful than adenosine
A2A and A3 receptor agonists for attenuating acute pain. In contrast,
intrathecal adenosine A3 receptor agonist decreased inflammatory
hyperalgesia without affecting acute pain.

c
d
e
f
g

Sang-Sik Choi, Yong-Chul Kim, Young Jin Lim, Chul-Joong Lee, Pyung-Bok
Lee, Sang-Chul Lee, Woo-Seok Sim, and Yoon-La Choi
The Neurological Safety of Epidural Gabapentin in Rats: A Light
Microscopic Examination
Anesth Analg 2005 101: 1422-1426.

IMPLICATIONS: Direct epidural injection of gabapentin in rats did not show

any neurotoxic evidence in terms of sensory-motor functions and behavior, or
by a microscopic histopathological evaluation. This study represents a first
promising step toward the trial of epidural gabapentin in a clinical setting.
c
d
e
f
g

Argyro Fassoulaki, Argyro Triga, Aikaterini Melemeni, and Constantine

Sarantopoulos
Multimodal Analgesia with Gabapentin and Local Anesthetics Prevents
Acute and Chronic Pain After Breast Surgery for Cancer
Anesth Analg 2005 101: 1427-1432.
IMPLICATIONS: Patients subjected to surgery for breast cancer were
randomly assigned to receive perioperatively gabapentin, ropivacaine in the
operating field, and EMLA around the wound or three placebos. The
multimodal analgesia reduced the analgesic requirements significantly and
attenuated the acute and chronic pain after surgery when compared with the
control group.
c
d
e
f
g

Eugene Y. Kissin, Cristina F. Freitas, and Igor Kissin

The Effects of Intraarticular Resiniferatoxin in Experimental Knee-Joint
Arthritis
Anesth Analg 2005 101: 1433-1439.
IMPLICATIONS: The intraarticular administration of a vanilloid agonist
resiniferatoxin inhibits pain behavior in knee joint arthritis. This effect is dosedependent.

c
d
e
f
g

Steven P. Cohen
Sacroiliac Joint Pain: A Comprehensive Review of Anatomy, Diagnosis,
and Treatment
Anesth Analg 2005 101: 1440-1453.

ECONOMICS, EDUCATION, AND HEALTH

SYSTEMS RESEARCH:
Zeev N. Kain
The Legend of the P Value
Anesth Analg 2005 101: 1454-1456.
c
d
e
f
g

Gregory B. Collins, Mark S. McAllister, Mark Jensen, and Timothy A. Gooden

Chemical Dependency Treatment Outcomes of Residents in
Anesthesiology: Results of a Survey
Anesth Analg 2005 101: 1457-1462.

IMPLICATIONS: Impairment among anesthesiology residents is common.

Less than half of the rehabilitated residents returning to anesthesia
successfully completed training and a significant proportion died as judged by
this nationwide survey. Improvement in overall outcome can be achieved
through redirection of these individuals into lower risk specialties.

CRITICAL CARE AND TRAUMA:

Nino Stocchetti
Brain and Sepsis: Functional Impairment, Structural Damage, and
Markers
Anesth Analg 2005 101: 1463-1464.
c
d
e
f
g

Anders Larsson, Mikls Lipcsey, Jan Sjlin, Lars-Olof Hansson, and Mats B.
Eriksson
Slight Increase of Serum S-100B During Porcine Endotoxemic Shock May
Indicate Blood-Brain Barrier Damage
Anesth Analg 2005 101: 1465-1469.
IMPLICATIONS: It is important to increase our knowledge regarding the
causes of neurologic function deterioration that frequently accompanies organ
failure in sepsis patients. S-100B is a widely used marker for brain damage
that potentially can be used to study sepsis effects on the central nervous
system.
c
d
e
f
g

Youns Assaoui, Amine Ali Zeggwagh, Acha Zekraoui, Khalid Abidi, and
Redouane Abouqal
Validation of a Behavioral Pain Scale in Critically Ill, Sedated, and
Mechanically Ventilated Patients
Anesth Analg 2005 101: 1470-1476.
IMPLICATIONS: Despite interest in pain assessment in the intensive care unit
(ICU), there is no specific pain scale for critical care patients and, particularly,
for uncommunicative patients. This study demonstrated that a behavioral scale
can be valid and reliable for measuring pain in ICU patients.
c
d
e
f
g

Jouni O. Kurola, Matti J. Turunen, Juha-Pekka Laakso, Jouko T. Gorski, Heikki

J. Paakkonen, and Tom O. Silfvast
A Comparison of the Laryngeal Tube and Bag-Valve Mask Ventilation by
Emergency Medical Technicians: A Feasibility Study in Anesthetized
Patients
Anesth Analg 2005 101: 1477-1481.

IMPLICATIONS: Airway management is of major importance in emergency

care. In this study, we found that emergency medical technician students could
use the laryngeal tube with reasonable success, and this device may provide an
alternative solution in emergency airway management.
c
d
e
f
g

Krishnan Raghavendran, Bruce A. Davidson, Jadwiga D. Helinski, Cristi J.

Marschke, Patricia Manderscheid, James A. Woytash, Robert H. Notter, and Paul R.
Knight
A Rat Model for Isolated Bilateral Lung Contusion from Blunt Chest
Trauma
Anesth Analg 2005 101: 1482-1489.
IMPLICATIONS: Lung contusion is a common problem encountered by
victims of motor vehicular accidents. We have developed a rat model for lung
contusion that does not involve any significant associated injury to the heart.
This model will be helpful in ascertaining the complex mechanisms involved
in the pathogenesis of lung contusion.
c
d
e
f
g

Ahmet Akyol, A. Can Senel, Hlya Ulusoy, Filiz Karip, and Nesrin Erciyes
Delayed Respiratory Depression After Risperidone Overdose
Anesth Analg 2005 101: 1490-1491.
IMPLICATIONS: Risperidone is an atypical antipsychotic drug used for the
treatment of schizophrenia. Although experience with risperidone overdose is
limited and the clinical presentation difficult to predict, the possibility of
delayed respiratory depression must be kept in mind.

NEUROSURGICAL ANESTHESIA:
c
d
e
f
g

Theresa Hartsell, Donlin Long, and Jeffrey R. Kirsch

The Efficacy of Postoperative Ondansetron (Zofran) Orally
Disintegrating Tablets for Preventing Nausea and Vomiting After
Acoustic Neuroma Surgery
Anesth Analg 2005 101: 1492-1496.
IMPLICATIONS: Although postoperative nausea and vomiting is common
(particularly in women) after open craniotomy for acoustic neuroma tumors,
treatment with Zofran(R) (IV and oral) significantly reduces the intensity and
frequency of this complication.

c
d
e
f
g

Leonard R. Allmond, Greg Stratmann, Sandeep M. Kunwar, and Daniel H.

Burkhardt
Epidural Blood Patch for Headache After Lumboperitoneal Shunt
Placement
Anesth Analg 2005 101: 1497-1498.

IMPLICATIONS: Epidural blood patch may be an alternative initial therapy

for some low-pressure headaches after lumboperitoneal shunt placement.
c
d
e
f
g

Ludmil Todorov, Charles E. Laurito, and David E. Schwartz

Postural Headache in the Presence of Cerebral Venous Sinus Thrombosis
(Case Report)

Anesth Analg 2005 101: 1499-1500.

IMPLICATIONS: We describe the case of a patient with peripartum cerebral
venous sinus thrombosis. It is important to consider this condition when
dealing with positional headache.

REGIONAL ANESTHESIA:
c
d
e
f
g

Mehmet Cesur, Haci A. Alici, Ali F. Erdem, Fikret Silbir, and Mustafa S. Yuksek
Administration of Local Anesthetic Through the Epidural Needle Before
Catheter Insertion Improves the Quality of Anesthesia and Reduces
Catheter-Related Complications
Anesth Analg 2005 101: 1501-1505.
IMPLICATIONS: Administration of a single-injection dose of local anesthetic
through the epidural needle as a priming solution into the epidural space
before inserting the catheter improves the quality of epidural anesthesia and
reduces the incidence of catheter-related complications.

c
d
e
f
g

Hany A. Mowafi
The Efficacy of Plethysmographic Pulse Wave Amplitude as an Indicator
for Intravascular Injection of Epinephrine-Containing Epidural Test
Dose in Anesthetized Adults
Anesth Analg 2005 101: 1506-1511.
IMPLICATIONS: Plethysmographic pulse wave amplitude is a reliable
alternative to changes in heart rate and systolic blood pressure in detecting
intravascular injection of an epinephrine-containing simulated epidural test
dose during general anesthesia.

c
d
e
f
g

Karim Asehnoune, Eric Larousse, Jean Marc Tadi, Vincent Minville, Stephane
Droupy, and Dan Benhamou
Small-Dose Bupivacaine-Sufentanil Prevents Cardiac Output
Modifications After Spinal Anesthesia
Anesth Analg 2005 101: 1512-1515.

IMPLICATIONS: Spinal injection of small-dose bupivacaine (7.5 mg) was

compared to large-dose bupivacaine (12.5 mg). The cardiac output (CO) was
depressed in the 12.5 mg group and remained stable in the 7.5 mg group.
Small-dose bupivacaine provides successful anesthesia and prevents a
decrease in CO.
c
d
e
f
g

Christos A. Iatrou, Christos K. Dragoumanis, Theodosia D. Vogiatzaki, George I.

Vretzakis, Constantinos E. Simopoulos, and Vasilios K. Dimitriou
Prophylactic Intravenous Ondansetron and Dolasetron in Intrathecal
Morphine-Induced Pruritus: A Randomized, Double-Blinded, PlaceboControlled Study
Anesth Analg 2005 101: 1516-1520.
IMPLICATIONS: Pruritus is the most common side effect of intrathecal
morphine. In this study, we evaluated the efficacy of prophylactic
administration of ondansetron and dolasetron for prevention of intrathecal
morphine-induced pruritus. Both medications offer effective antipruritic
prophylaxis, reducing the severity and incidence of pruritus after intrathecal
morphine.
c
d
e
f
g

Masataka Yokoyama, Yoshitaro Itano, Hiroshi Katayama, Hiroshi Morimatsu,

Yoshimasa Takeda, Toru Takahashi, Osamu Nagano, and Kiyoshi Morita
The Effects of Continuous Epidural Anesthesia and Analgesia on Stress
Response and Immune Function in Patients Undergoing Radical
Esophagectomy
Anesth Analg 2005 101: 1521-1527.
IMPLICATIONS: Extensive epidural block (C3-L2) was not able to suppress
leukocytosis, lymphopenia, increases in cytokines and C-reactive protein, and
changes in proportions of lymphocyte subsets after radical esophagectomy.
This suggests that tissue damage and inflammation overcome the effects of
extensive epidural block on stress response and immune function after radical
esophagectomy.
c
d
e
f
g

Ban C. H. Tsui and Corey K. C. Sze

An In Vitro Comparison of the Electrical Conducting Properties of
Multiport Versus Single-Port Epidural Catheters for the Epidural
Stimulation Test
Anesth Analg 2005 101: 1528-1530.
IMPLICATIONS: When performing the epidural stimulation test, sufficient
current can be transmitted with metal-coil reinforced catheters (single or
multi-port) but not with conventional plastic catheters.

c
d
e
f
g

Ashok Kumar and Ashim Banerjee

Continuous Maxillary and Mandibular Nerve Block for Perioperative
Pain Relief: The Excision of a Complicated Pleomorphic Adenoma
Anesth Analg 2005 101: 1531-1532.
IMPLICATIONS: Continuous maxillary and mandibular nerve blocks were a
significant and novel contribution for management of excision of pleomorphic
adenoma in an elderly patient with intractable atrial fibrillation associated with
regional wall motion abnormalities, significant reduction of ejection fraction,
hypertension, poor chest condition, and difficult airway.

GENERAL ARTICLES:
c
d
e
f
g

Gernot E. Groemer, Joseph Brimacombe, Thorsten Haas, Cristina de Negueruela,

Alexander Soucek, Michael Thomsen, and Christian Keller
The Feasibility of Laryngoscope-Guided Tracheal Intubation in
Microgravity During Parabolic Flight: A Comparison of Two Techniques
Anesth Analg 2005 101: 1533-1535.
IMPLICATIONS: Laryngoscope-guided tracheal intubation is feasible in
microgravity obtained during parabolic flight, but the success rate is
infrequent because of severe time restrictions. There were no differences in
success rate between the free-floating condition, with the head gripped
between the knees, and in the restrained condition, with the torso strapped to
the surface.
c
d
e
f
g

Jean-Pierre Estebe, Marc Gentili, Pascal Le Corre, Gilles Dollo, Franois

Chevanne, and Claude Ecoffey
Alkalinization of Intracuff Lidocaine: Efficacy and Safety
Anesth Analg 2005 101: 1536-1541.
IMPLICATIONS: Adding less bicarbonates (1.4% versus 8.4%) to lidocaine
placed in an endotracheal (ETT) can still be effective in preventing ETTrelated emergence phenomena, although it results in a slower release of
lidocaine through the ETT cuff. These results were clinically relevant despite
thyroid surgery and anesthesia without nitrous oxide.
c
d
e
f
g

Banjong Krobbuaban, Siriwan Diregpoke, Sujarit Kumkeaw, and Malin

Tanomsat
The Predictive Value of the Height Ratio and Thyromental Distance:
Four Predictive Tests for Difficult Laryngoscopy
Anesth Analg 2005 101: 1542-1545.
IMPLICATIONS: Because difficult laryngoscopy is infrequent, the incidence
of false negatives is small; nevertheless, the sensitivity of a test should be

sufficiently high to detect possible difficulties with laryngoscopy.

Measurement of the ratio of height to thyromental distance resulted in less
detection failure of difficult laryngoscopy than mouth opening, thyromental
distance, neck movement, and oropharyngeal view (modified Mallampati).
c
d
e
f
g

Barbara Kabon, Ozan Aka, Akiko Taguchi, Angelika Nagele, Ratnaraj

Jebadurai, Cem F. Arkilic, Neeru Sharma, Arundhathi Ahluwalia, Susan Galandiuk,
James Fleshman, Daniel I. Sessler, and Andrea Kurz
Supplemental Intravenous Crystalloid Administration Does Not Reduce
the Risk of Surgical Wound Infection
Anesth Analg 2005 101: 1546-1553.
IMPLICATIONS: Supplemental perioperative IV fluid administration did not
reduce the rate of wound infection. The apparent lack of benefit may have
resulted because hydration's effect on intestinal oxygenation is modest or
because the statistical power of our study was limited. Nonetheless, our results
suggest that supplemental hydration in the range tested does not impact wound
infection rate.
c
d
e
f
g

Ju-Mei Ng
Hypoxemia During One-Lung Ventilation: Jet Ventilation of the Middle
and Lower Lobes During Right Upper Lobe Sleeve Resection
Anesth Analg 2005 101: 1554-1555.
IMPLICATIONS: This case report describes a novel way of managing
hypoxemia during one-lung ventilation when continuous positive airway
pressure to the nondependent lung is not feasible.

MEETING ARTICLE:
John R. Keltner and Pamela Flood
Report of the 13th Annual Meeting of the International Society for
Anaesthetic Pharmacology
Anesth Analg 2005 101: 1556-1557.
M. Cardwell, G. Siviter, and A. Smith
Nonsteroidal Antiinflammatory Drugs and Perioperative Bleeding in
Pediatric Tonsillectomy
Anesth Analg 2005 101: 1558.
M. Zacharias, I. CS Gilmore, G. P. Herbison, P. Sivalingam, and R. J. Walker
Interventions for Protecting Renal Function in the Perioperative Period
Anesth Analg 2005 101: 1558.

LETTER TO THE EDITOR:

Leo I. Stemp
Anesthetic Depth and Long-Term Mortality
Anesth Analg 2005 101: 1559.
Neal H. Cohen
Anesthetic Depth and Long-Term Mortality
Anesth Analg 2005 101: 1559-1560.
Bhavani Shankar Kodali
Capnogram Shape in Obstructive Lung Disease
Anesth Analg 2005 101: 1560.
Baruch Krauss, Aaron Deykin, Alexander Lam, Joan J. Ryoo, Jyoti Mathad, David R.
Hampton, Paul W. Schmitt, and Jay L. Falk
Capnogram Shape in Obstructive Lung Disease
Anesth Analg 2005 101: 1560.
Ratan K. Banik
Short-Lasting Effect of Perineural Resiniferatoxin on Mechanical
Hyperalgesia
Anesth Analg 2005 101: 1560-1561.
Igor Kissin, Natasha Davison, and Edwin L. Bradley, Jr
Short-Lasting Effect of Perineural Resiniferatoxin on Mechanical
Hyperalgesia
Anesth Analg 2005 101: 1561.
Koh Mizutani and Yoshiro Toyoda
Sterilization Pouches for Perioperative Thermal Insulation
Anesth Analg 2005 101: 1561.
Stephen B. Corn
Countries, Rooms, and Plants, but Not Mammalian Cells
Anesth Analg 2005 101: 1561.
T. J. Gan, Michael P. W. Grocott, and Michael G. Mythen
Countries, Rooms, and Plants, but Not Mammalian Cells
Anesth Analg 2005 101: 1561.
Hetam Al-Takrouri and James F. Mayhew
Hyperkalemic Cardiac Arrest After Cardiopulmonary Bypass in a Child
with Duchenne Muscular Dystrophy
Anesth Analg 2005 101: 1561-1562.
Aruna Nathan, Arjunan Ganesh, Rodolfo I. Godinez, Susan C. Nicolson, and William
J. Greeley
Hyperkalemic Cardiac Arrest After Cardiopulmonary Bypass in a Child
with Duchenne Muscular Dystrophy
Anesth Analg 2005 101: 1562.
Eric L. Bloomfield
Costs Are Not the Only Thing We Should Be Concerned with in
Anesthesia
Anesth Analg 2005 101: 1562.

Martin Schuster, Andre Gottschalk, and Thomas Standl

Costs Are Not the Only Thing We Should Be Concerned with in
Anesthesia
Anesth Analg 2005 101: 1562.
Marie T. Aouad and Roula E. Hajj
Supplementation of Intrathecal Bupivacaine with Clonidine in ExPremature Neonates
Anesth Analg 2005 101: 1562-1563.
Alain Rochette, Christophe Dadure, and Xavier Capdevila
Supplementation of Intrathecal Bupivacaine with Clonidine in ExPremature Neonates
Anesth Analg 2005 101: 1563.
Ahed Zeidan, Dania Halabi, and Anis Baraka
Aerosolized Lidocaine for Relief of Extubation Laryngospasm
Anesth Analg 2005 101: 1563.
Naveen Eipe and Ashish Choudhrie
Airway Fires: Gas-Bugs Providing the Fuel?
Anesth Analg 2005 101: 1563-1564.
Mitchel B. Sosis
Anesthesiologists Must Inform Their Surgical Colleagues When There Is
a Risk of an Operating Room Fire
Anesth Analg 2005 101: 1564.
Andrew Paterson
The Case for the Role of Advanced Simulators in Trauma Management
Training Was Not Made
Anesth Analg 2005 101: 1564.
Hain Berkenstadt and Daphna Barsuk
The Case for the Role of Advanced Simulators in Trauma Management
Training Was Not Made
Anesth Analg 2005 101: 1564-1565.
Shaul Cohen, Alex Shorshtein, and Patrick Blanchfield
Subdural Hematoma Following Accidental Dural Puncture
Anesth Analg 2005 101: 1565.
Ju-Mei Ng
Thoracotomy in a Patient with a History of Local Anesthetic Allergy
Anesth Analg 2005 101: 1565-1566.
G. D. Puri and D. Nakra
ECG Artifact and BIS in Severe Brain Injury
Anesth Analg 2005 101: 1566-1567.
Tomasz Gaszynski
ECG Artifact and BIS in Severe Brain Injury
Anesth Analg 2005 101: 1567.

BOOK AND MULTIMEDIA REVIEWS:

Stanley Muravchick
On-Line Electronic Help for Anaesthesiologists.
Anesth Analg 2005 101: 1568.
James E. Duckett and Gordon H. Morewood
Fundamental Applications of Transesophageal Echocardiography: A
Society of Cardiovascular Anesthesiologists Monograph on DVD.
Anesth Analg 2005 101: 1568-1569.

ERRATA:
Erratum
Anesth Analg 2005 101: 1569.

CARDIOVASCULAR ANESTHESIA

SOCIETY

OF

CARDIOVASCULAR ANESTHESIOLOGISTS

SECTION EDITOR
KENNETH J. TUMAN

The Influence of Mitochondrial KATP-Channels in the

Cardioprotection of Preconditioning and Postconditioning by
Sevoflurane in the Rat In Vivo
Detlef Obal, MD, DEAA, Saskia Dettwiler,
Horst Scharbatke, MD, Benedikt Preckel,

cand. MD,

Christian Favoccia, MD, DEAA,

and Wolfgang Schlack, MD, PhD,

MD, DEAA,

DEAA

Klinik fur Anaesthesiologie, Universitatsklinikum Dusseldorf, Dusseldorf, Germany

Volatile anesthetics induce myocardial preconditioning

and can also protect the heart when given at the onset of
reperfusiona practice recently termed postconditioning. We investigated the role of mitochondrial KATP
(mKATP)-channels in sevoflurane-induced cardioprotection for both preconditioning and postconditioning
alone and whether there is a synergistic effect of both.
Rats were subjected to 25 min of coronary artery occlusion followed by 120 min of reperfusion. Infarct size was
determined by triphenyltetrazolium staining. The following protocols were used: 1) preconditioning (S-Pre, n
10, achieved by 2 periods of 5 min sevoflurane administration (1 MAC) followed by 10 min of washout); 2)
sevoflurane postconditioning (1 MAC of sevoflurane
given for 2 min at the beginning of reperfusion; S-Post, n
10); 3) administration before and after ischemia (S-Pre
S-Post, n 10). Protocols 13 were repeated in the
presence of 5-hydroxydecanoate (5HD), a specific
mKATP-channel-blocker (S-Pre S-Post 5HD, S-Pre

reconditioning describes a protective mechanism against myocardial ischemia that can be

initiated by short episodes of coronary occlusion
(ischemic preconditioning) or by administration of
certain drugs (pharmacological preconditioning). One
of these classes of drugs, volatile anesthetics, when
given before ischemia, induces a cardioprotective effect similar to that of ischemic preconditioning.
Whereas the end-effector of both ischemic and anesthetic preconditioning remains unknown, mitochondrial KATP (mKATP)-channels appear to play a central
role in anesthetic preconditioning (1,2).
Most research in the last years has focused on interventions like preconditioning to ameliorate irreversible
Accepted for publication June 15, 2005.
Address correspondence and reprint requests to Professor Wolfgang Schlack, DEAA, Klinik fur Anaesthesiologie, Universitatsklinikum Dusseldorf, Postfach 10 10 07, D-40001 Dusseldorf, Germany.
Address e-mail to schlack@med.uni-duesseldorf.de.
DOI: 10.1213/01.ANE.0000181336.96511.32

1252

Anesth Analg 2005;101:125260

5HD: n 10; S-Post 5HD: n 9). Nine rats served as

untreated controls (CON) or received 5HD alone (5HD,
n 10). Both S-Pre (23% 13% of the area at risk, mean
sd) and S-Post (18% 5%) reduced infarct size compared with CON (49% 11%, both P 0.05). S-Pre
S-Post resulted in a larger reduction of infarct size (12%
5%, P 0.054 versus S-Pre) compared with administration before or after ischemia alone. 5HD diminished
the protection in all three sevoflurane treated groups
(S-Pre 5HD, 35% 12%; S-Post 5HD, 44% 12%;
S-Pre S-Post 5HD, 46% 14%;) but given alone had
no effect on infarct size (41% 13%). Sevoflurane preconditioning and postconditioning protects against myocardial ischemia-reperfusion injury. The combination of
preconditioning and postconditioning provides additive
cardioprotection and is mediated, at least in part, by
mKATP-channels.
(Anesth Analg 2005;101:125260)

ischemic cell damage. However, the first minutes of

reperfusion are critical, and reperfusion paradoxically
worsens ischemic injury. Studies have demonstrated a
protective effect of brief episodes of ischemia after prolonged coronary artery occlusion that resulted in reduced infarct size. In parallel to preconditioning this
strategy was termed postconditioning (3).
Similar to the protection of volatile anesthetics given as
preconditioning stimulus, these drugs can also protect the
heart when given after ischemia during reperfusion (46).
The protection resulting in a decrease of infarct size might
be a mechanism of anesthetic postconditioning. Although both mechanisms are effective to protect the heart
against ischemia-reperfusion injury, the underlying mechanism of both strategies is still unknown. Because mKATPchannels are of such great importance in preconditioning,
we hypothesized an involvement of these channels also
during postconditioning. The present study also investigated if additive protection and a reduction in infarct size
could be achieved by combining both preconditioning and
2005 by the International Anesthesia Research Society
0003-2999/05

ANESTH ANALG
2005;101:125260

CARDIOVASCULAR ANESTHESIA
OBAL ET AL.
PERI-ISCHEMIC PROTECTION BY SEVOFLURANE

1253

Figure 1. Experimental protocol. CON control

group; 5HD 5-hydroxydecanoate; S-Pre group
preconditioned by two episodes of sevoflurane administration interspersed with 10 min of washout;
arrows mark the time of continuous 5HD administration (5 g kg1 min1); S-Post group
which received sevoflurane at the beginning of
reperfusion (anesthetic postconditioning); all rats
underwent 25 min of coronary artery occlusion
followed by 2 h of reperfusion.

postconditioning strategies. We further questioned

whether both protection mechanisms might work additively and enhance the infarct size-limiting effect of each
single intervention.

Methods
The study was performed in accordance with the regulations of the German Animal Protection Law and
was approved by the Bioethics Committee of the District of Duesseldorf.
The surgical procedures were described in detail
previously (5). In brief, -chloralose-anesthetized male
Wistar rats (mean body weight, 397 46 g) were
artificially ventilated after orotracheal intubation (rodent ventilator Rhema-Labortechnikav, Type 10 mL,
Class 34931, Germany) with a Fio2 0.4 and a positive
end-expiratory pressure of 35 cm H2O for maintaining arterial pH, Pco2, and Po2 within normal physiological limits. Body temperature was maintained at
38C by the use of a heating pad. The rats were instrumented for measurement of mean aortic pressure
(AOP, catheter in the left femoral artery), left ventricular (LV) pressure (LVP, tip manometer introduced
through the right carotid artery, Millar Instruments),
and cardiac output (CO, ultrasonic flowprobe (6S)
around the pulmonary artery; T 208; Transonic Systems Inc., Ithaca, NY). A ligature snare was looped
around a major coronary artery supplying the anterior
wall of the LV for later occlusion. Myocardial ischemia
was verified by the appearance of epicardial cyanosis
and changes in surface electrocardiogram. Successful
reperfusion was evidenced later by disappearance of
epicardial cyanosis, as described previously (6).
Hemodynamic variables were recorded after a 15min stabilization period. All rats underwent 25 min of
regional myocardial ischemia by tightening the snare

around the prepared coronary artery, followed by

120 min of reperfusion. Successful reperfusion was
evidenced by the disappearance of epicardial cyanosis. After 120 min of reperfusion, the hearts were
quickly excised and infarct size was measured by
triphenyltetrazoliumchloride (TTC) staining. Therefore, the hearts were perfused on a modified Langendorff apparatus with normal saline at 80 mm Hg perfusion pressure to washout the remaining blood. The
coronary artery was then reoccluded and 510 mL of
0.2% Evans Blue dye in 1% Dextran was infused via
the aortic root into the coronary system. This maneuver identifies the area at risk as the part of the myocardium that remains unstained. The heart was then
frozen and cut into approximately 10 transverse slices
of equal thickness (1 mm). The slices were incubated
(37C) for 15 min in buffered 1% TTC adjusted to pH
7.4 and then incubated for 2 days in 4% formaldehyde.
Viable myocardium was then identified as red stained
by TTC, whereas necrotic myocardium appears pale
gray. The area at risk and the infarcted area were
determined by planimetry using Sigma Scan Pro 5
computer software (SPSS Science Software, Chicago,
IL) and corrected for dry weight (5).
All rats were randomly assigned to one of the following groups (Fig. 1):
Control (CON, n 9): After a stabilization period of
45 min, rats were subjected to 25 min of coronary artery
occlusion followed by 120 min of reperfusion (ischemia
and reperfusion period similar in all other groups).
5-Hydroxydecanoate (5HD, n 10): IV administration of 5HD (5 g kg1 min1 in saline aqueous
solution) starting 40 min before ischemia until the
beginning of reperfusion.
Sevoflurane preconditioning (S-Pre, n 10): Rats
received sevoflurane 2.0 vol.% (corresponding to one
minimal alveolar concentration (MAC) in rats (7) for 2

1254

CARDIOVASCULAR ANESTHESIA OBAL ET AL.

PERI-ISCHEMIC PROTECTION BY SEVOFLURANE

5-min periods interspersed by 10 min of washout

10 min before coronary artery occlusion.
Sevoflurane preconditioning 5HD (S-Pre 5HD,
n 10): Rats underwent the S-Pre protocol with additional administration of 5HD starting 10 min before
the preconditioning protocol until the occlusion of the
coronary artery.
Sevoflurane during reperfusion (postconditioning) (S-Post, n 10): Rats received one MAC sevoflurane for 2 min starting at the beginning of reperfusion.
Sevoflurane during reperfusion 5HD (S-Post
5HD, n 9): Rats received 5HD starting 10 min before
ischemia until the end of sevoflurane administration
during reperfusion.
Sevoflurane during preconditioning and postconditioning (S-Pre S-Post, n 10): Both, the S-Pre and
S-Post protocol were combined and rats received one
MAC of sevoflurane before and after myocardial
ischemia.
Sevoflurane preconditioning and postconditioning
5HD (S-Pre S-Post 5HD, n 10): Rats received
sevoflurane before and after myocardial ischemia and
5HD 10 min before preconditioning until the end of
sevoflurane administration during reperfusion.
Expiratory sevoflurane concentration was measured
at the end of the endotracheal tube (Datex Capnomac
Ultima, Division of Instrumentarium Corp., Helsinki,
Finland) at a sampling rate of 200 mL/min. By using
a high inspiratory flow of 12 L/min, rapid changes of
sevoflurane concentrations could be achieved as evidenced by a very rapid decrease in expiratory sevoflurane concentration during the experimental course
(0.0 vol. % within 15 s) and by rapid changes in
hemodynamics during the first seconds of sevoflurane
administration (5,6).
LVP, its first derivative dP/dt, AOP and CO were
recorded continuously on a polygraph (Hellige 120
710 94; Freiburg, Germany) and were digitized using
an analog-to-digital converter (Data Translation, Marlboro, MA) at a sampling rate of 500 Hz and processed
later on a personal computer.
Global myocardial function was measured in terms
of LV peak systolic pressure (LVPSP). Global LV endsystole was defined as the point of minimum dP/dt
(dP/dtmin) and LV end-diastole as the beginning of
the sharp upslope of the LV dP/dt tracing. Systemic
vascular resistance (SVR) was estimated from AOP
and CO, assuming a right atrial pressure of 0 mm Hg
in the open-chest preparation.
Results are expressed as mean sd. Statistical analysis was performed by a two-way analysis of variance
for time and treatment (preconditioning versus protection against reperfusion) effects. Time effects
(changes from baseline value) during the experiments
were analyzed using Dunnetts post hoc test. If an
overall significance among groups was found, Students t-test with Bonferroni correction for multiple

ANESTH ANALG
2005;101:125260

comparisons was performed for each time point. The

same analysis was used for detection of differences in
infarct size. Changes within and between groups were
considered statistically significant with P 0.05.

Results
A total of 80 rats were used. Two rats died from
untreatable arrhythmia during ischemia and reperfusion. In the remaining 78 animals, complete data sets
were obtained.
Differences in hemodynamics were not observed
among the groups at the beginning of the experiments;
data are summarized in Tables 1 and 2 and Fig. 2.
Preconditioning with sevoflurane resulted in a reduction of global (LV) function (LVPSP, 94 18 mm Hg
and rate pressure product [RPP], 35 9 mm
Hg min1 10001 summarized for all groups which
received sevoflurane, values after second exposure to
sevoflurane; P 0.05 versus CON or 5HD). Sevoflurane also decreased SVR (189 54 mm
Hg min1 10001) and AOP (56 19 mm Hg). After
10 min of washout, recovery was similar in all groups
and hemodynamics did not differ among the groups
before coronary artery occlusion (Tables 1 and 2 and
Fig. 2).
During reperfusion, LVPSP remained nearly constant in CON (129 17 mm Hg) and 5HD (123
21 mm Hg, both at 15 min of reperfusion) groups.
Although global hemodynamics were reduced (at
least due to metabolic depression) in all groups during
initial reperfusion, sevoflurane treatment was followed by stronger reduction of global hemodynamic
variables (LVPSP, 93 13 mm Hg; RPP, 36 8 mm
Hg min1 10001; all data at 2 min of reperfusion,
all P 0.05 versus baseline); SVR (167 46 mm
Hg min L1) and AOP (52 13 mm Hg, both P
0.05) were also reduced, whereas CO remained nearly
constant (32 8 mL/min). LVPSP (94 14 mm Hg)
and AOP (57 12 mm Hg) decreased similarly in
S-Post and S-Pre S-Post at the beginning of
reperfusion.
After sevoflurane was discontinued the cardiodepressive effect in all groups was rapidly abrogated
(LVPSP, 119 19 mm Hg; RPP, 48 11 mm
Hg min1 10001; all data after 5 min of reperfusion). SVR (251 76 mm Hg min L1) and AOP (86
20 mm Hg) returned to values similar to those of the
groups that did not receive sevoflurane at beginning
of reperfusion. No further differences among the
groups were observed until the end of reperfusion.
Mean LV dry weight was 0.15 0.05 g with no
differences among groups (data for the individual
groups are given in Table 3). The ischemic-reperfused
area (area at risk) constituted 39% 16% of the LV. In
CON, infarct size was 49% 11% of the area at risk,

ANESTH ANALG
2005;101:125260

CARDIOVASCULAR ANESTHESIA
OBAL ET AL.
PERI-ISCHEMIC PROTECTION BY SEVOFLURANE

1255

Table 1. Hemodynamic Data

Coronary occlusion
Variable
HR (bpm)
CON
5HD
S-Pre
S-Pre 5HD
S-Post
S-Post 5HD
S-Pre S-Post
S-Pre S-Post 5HD
AOP (mm Hg)
CON
5HD
S-Pre
S-Pre 5HD
S-Post
S-Post 5HD
S-Pre S-Post
S-Pre S-Post 5HD
LVEDP (mm Hg)
CON
5HD
S-Pre
S-Pre 5HD
S-Post
S-Post 5HD
S-Pre S-Post
S-Pre S-Post 5HD

Baseline

S-Pre

15 min

24 min

2 min

5 min

15 min

60 min

120 min

399 52
433 48
400 21
383 37
393 33
380 28
379 30
390 33

409 35
427 48
402 22
392 33
405 39
408 37
402 21
417 31

411 27
431 50
411 32
391 36
406 29
420 27
402 19
421 38

394 27
409 50
408 26
395 24
396 25
412 34
407 37
408 28

371 25
398 50
382 47
365 48
397 23
381 35
385 31
390 33

426 29
434 42
420 51
440 49
432 36
438 25
422 21
458 46

424 32
426 58
367 33*
378 48
416 29
430 34
385 28
390 25

416 40
434 54
406 24
392 39
419 23
433 40
407 23
433 33

416 38
448 44
419 38
397 36
428 33
429 29
411 23
438 42

410 43
437 52
412 36
402 37
414 36
433 34
412 25
433 37

115 27
106 16
109 17
106 15
105 10
116 14
114 15
103 16

105 30
85 32
50 12*
53 14*
95 11
99 27
67 27*
53 16*

95 27
93 23
96 22
86 17
100 12
93 27
93 24
101 19

97 34
82 22
91 21
74 30
91 15
90 26
91 21
90 20

93 36
87 19
89 22
76 26
96 16
100 28
93 24
101 20

11 3
12 3
94
10 2
93
12 8
12 5
10 5

13 5
15 6
16 3
15 4
13 8
14 6
16 5
12 9

14 6
15 7
12 7
15 3
15 8
14 6
16 6
12 8

10 3
10 3
92
94
74
92
72
82

10 4
92
73
12 8
83
13 8
13 4
95

Reperfusion

Washout/
recovery

84 19
81 19
84 23
67 21*
59 13*
51 12*
55 12*
45 12*

81 28
75 18
88 22
74 27
86 21
81 23
88 20
88 21

81 32
84 23
88 25
75 26
93 16
102 25
87 19
88 21

16 7
16 6
15 6
14 6
11 6
12 5
14 4
95

14 8
15 7
13 7
16 7
14 9
14 7
17 7
10 5

11 6
16 6
10 6
15 5
13 8
12 7
15 8
96

80 21
80 28
77 27
72 27
92 19
83 29
87 13
83 24

67 25
68 24
72 19
68 29
85 20
76 28
76 20
68 23

13 7
16 8
12 5
14 6
12 8
11 5
13 6
97

14 3
12 4
12 7
13 6
17 7
14 7
13 5
11 7

Data are mean sd.

CON control group; S-Pre sevoflurane preconditioning; S-Post sevoflurane postconditioning; 5HD 5-hydroxydecanoate; HR heart rate; AOP
mean aortic pressure; LVEDP left ventricular end-diastolic pressure.
* P 0.05 versus CON; P 0.05 versus S-Pre; P 0.05 versus S-Post; P 0.05 versus baseline.

Table 2. Hemodynamic Data

Coronary occlusion
Variable
CO (mL/min)
CON
5HD
S-Pre
S-Pre 5HD
S-Post
S-Post 5HD
S-Pre S-Post
S-Pre S-Post 5HD
1
RPP (mm Hg min 10001)
CON
5HD
S-Pre
S-Pre 5HD
S-Post
S-Post 5HD
S-Pre S-Post
S-Pre S-Post 5HD
SVR (mm Hg min L1)
CON
5HD
S-Pre
S-Pre 5HD
S-Post
S-Post 5HD
S-Pre S-Post
S-Pre S-Post 5HD

Reperfusion

Baseline

S-Pre

Washout/
recovery

35 6
42 11
36 7
37 8
40 8
36 5
34 6
34 6

32 4
37 8
28 8
32 5
36 7
36 9
32 5
30 8

33 5
36 6
33 5
36 6
39 6
36 5
34 7
37 8

31 4
33 7
34 5
33 6
39 5
35 6
34 6
34 9

31 4
33 7
32 3
35 7
37 5
37 5
37 6
37 10

32 7
35 6
34 6
37 6
36 6
34 10
31 9
30 8

33 6
36 6
35 6
35 7
38 6
36 5
34 8
35 10

30 5
36 6
36 7
35 8
36 4
35 6
34 8
31 6

35 7
36 8
36 8
35 7
37 6
37 9
36 9
33 6

35 5
36 11
33 9
37 7
35 9
40 6
34 4
34 10

61 9
61 7
58 14
63 11
58 9
64 7
59 10
64 12

57 4
48 8
29 8*
34 6*
52 7
62 14
38 11*
37 7*

55 5
53 8
48 9
49 7
55 8
58 12
50 11
60 9

54 6
53 10
50 5
45 14
52 12
52 4
50 7
55 10

52 7
53 8
47 7
47 11
52 9
59 9
50 10
59 10

49 11
49 4
46 7
44 10
38 8*
34 10*
35 7*
35 7*

51 8
48 4
47 8
47 11
48 13
45 12
47 9
52 10

50 7
49 5
49 14
46 13
49 9
55 10
48 7
52 10

45 6
48 13
45 13
47 9
49 10
49 9
47 7
50 10

38 8
43 12
42 12
45 12
45 13
44 8
42 8
44 12

302 117
277 86
299 68
251 96
268 53
266 87
288 100
287 78

320 146
258 69
277 76
253 101
237 29
261 81
279 94
281 96

301 125
271 64
280 93
212 81
258 44
279 91
256 64
294 97

265 77
243 64
247 69
186 64
164 32
161 53*
185 45
159 50*

254 95
217 76
257 72
219 97
231 56
232 75
266 63
270 102

298 126
252 82
260 89
227 118
254 34
301 104
273 98
296 99

241 104
230 96
237 97
215 76
244 52
236 99
254 67
267 82

201 101
168 36
223 29
196 85
249 67
185 88
227 64
185 37

343 99
279 101
318 70
303 81
271 56
332 70
345 73
306 52

334 123
236 92
198 47
169 49
271 58
283 80
207 56
182 63

15 min

24 min

2 min

5 min

15 min

60 min

120 min

Data are mean sd.

CON control group; S-Pre sevoflurane preconditioning; S-Post sevoflurane postconditioning; 5HD 5-hydroxydecanoate; CO cardiac output;
RPP rate pressure product; SVR systemic vascular resistance.
* P 0.05 versus CON; P 0.05 versus S-Pre; P 0.05 versus S-Post; P 0.05 versus baseline.

1256

CARDIOVASCULAR ANESTHESIA OBAL ET AL.

PERI-ISCHEMIC PROTECTION BY SEVOFLURANE

ANESTH ANALG
2005;101:125260

Figure 2. Line plot showing the time course of left

ventricular peak systolic pressure (LVPSP) during
the experiments. Dots express mean sd.

which was not affected by administration of 5HD

alone (41% 13%) (Fig. 3). Sevoflurane preconditioning reduced infarct size to 23% 13% of the area at
risk (P 0.001 versus CON). The cardioprotective
effect of sevoflurane was greater in the group that
received sevoflurane during reperfusion alone (S-Post:
18% 5%) and further increased by administration of
sevoflurane before and after ischemia (S-Pre S-Post:
12% 5%; P 0.054 versus S-Pre). Administration of
the specific mKATP-channel blocker 5HD abolished the
protection. (S-Pre 5HD: 35% 12%; P 0.08 versus
S-Pre; S-Post 5HD: 44% 12%, P 0.05 versus
S-Post; S-Pre S-Post 5HD: 46% 14%, P 0.05
versus S-Pre S-Post).

Discussion
This study demonstrated that mKATP-channels are
also involved in sevoflurane-induced cardioprotection
during reperfusion, i.e., in anesthetic postconditioning. Thus, anesthetic preconditioning and anesthetic
postconditioning may share opening of mKATPchannels during cardioprotection. Sevoflurane postconditioning further enhanced the cardioprotection of
sevoflurane preconditioning.
In the present study we used two 5-minute periods
of sevoflurane exposure to induce anesthetic preconditioning. This was similar to a study of An et al. (8) in
isolated guinea pig hearts. From studies of ischemic
preconditioning and anesthetic preconditioning (9) it
is known that repeated transient ischemic episodes are
more effective in reducing infarct size than one single
ischemic period (10). Sevoflurane was given at only
one concentration (1 MAC) because we had demonstrated previously that a maximal cardioprotective effect can be achieved by using this concentration of

sevoflurane for 2 minutes of reperfusion (5,6). Recently published data confirmed those findings for
isoflurane (11).
Besides the anti-ischemic properties (12), volatile
anesthetics given before myocardial ischemia induce
anesthetic preconditioning, characterized by an application period separated from the subsequent ischemia
by a memory period (1315). The administration of
volatile anesthetics at the beginning of reperfusion
also protects against a lethal cell injury (5,6,16) and the
term postconditioning was recently introduced for
interventions after ischemia (17).
Initially, mKATP-channels were thought only to be
the end-effectors of preconditioning. Direct openers of
mKATP-channels (i.e., diaxozide) given before a sustained ischemia significantly prolonged the time to
ischemic contracture in isolated hearts (18) and reduced the cell damage (19). These effects could be
abolished by previous administration of the mKATPblocker 5HD without having any specific effect on
infarct size (20). Nevertheless, Pain et al. (21) demonstrated that mKATP-channels are only one part of the
signal cascade of preconditioning. They demonstrated
that release of reactive oxygen species (ROS) might be
involved in the protection mediated by ischemic preconditioning cascade, which was also shown for anesthetic preconditioning (2224). In contrast to large
amounts of ROS that might be harmful, small concentrations of ROS seem to act as a trigger of preconditioning. Different ROS species seem to have protective
properties. One possible way to initiate preconditioning might be the generation of free radicals or reactive
nitrogen species, which may, in turn, activate the intracellular protein kinase C (PKC) or tyrosine kinase
pathway, resulting in a lower threshold for mKATPchannels opening. Opening of mKATP-channels causes

ANESTH ANALG
2005;101:125260

CARDIOVASCULAR ANESTHESIA
OBAL ET AL.
PERI-ISCHEMIC PROTECTION BY SEVOFLURANE

1257

Table 3. Weights and Size of Area at Risk

Group

Body weight (g)

Left ventricle dry weight (g)

Area at risk dry weight (g)

CON
5HD
S-Pre
S-Pre 5HD
S-Post
S-Post 5HD
S-Pre S-Post
S-Pre S-Post 5HD

429 57
392 46
400 50
374 15
412 58
388 16
401 50
381 39

0.14 0.04
0.16 0.06
0.15 0.05
0.15 0.07
0.17 0.08
0.14 0.05
0.16 0.04
0.14 0.04

0.05 0.02
0.06 0.04
0.05 0.03
0.05 0.01
0.06 0.02
0.05 0.03
0.06 0.03
0.05 0.01

Body weight, left ventricular weight and area at risk size in the control group (CON) and the groups that were preconditioned by sevoflurane (S-Pre, S-Pre
S-Post) or that received sevoflurane during reperfusion (S-Post; S-Pre S-Post) without or in presence of the mitochondrial KATP-channel blocker
5-hydroxydecanoate (5HD). Data are mean sd.

Figure 3. Infarct size expressed as percent of the area at risk. Bars present mean
sd. CON control group; 5HD
5-hydroxydecanoate; S-Pre group preconditioned by 2 episodes of sevoflurane
administration interspersed by 10 min of
washout; S-Post group received
sevoflurane at the beginning of reperfusion; S-Pre S-Post group received
sevoflurane before ischemia (preconditioning) and at the beginning of reperfusion (anesthetic postconditioning). *P
0.05 versus CON or 5HD; #P 0.05 versus S-Post; P 0.05 versus S-Pre
S-Post.

a small decrease in mitochondrial membrane potential, which would slow the electron transport chain
and oxidative phosphorylation during ischemia and
reperfusion (25).
Maintenance of mitochondrial bioenergetics might
decrease generation of damaging ROS and so produce
better function and reduced cell damage after ischemia. However, there are still many questions about the
protective effect of preconditioning induced by ROS
(26,27).
Although the mKATP-channels appear to be involved in anesthetic preconditioning, so does enhanced ROS production (24). With normal cellular
adenosine triphosphate (ATP) levels volatile anesthetics did not activate the mKATP-channels as
shown in patch clamp experiments (28) but sensitized the mKATP-channels to reduced ATP levels
(29,30). Anesthetic preconditioning also modifies
the phosphorylation state of different kinases (31).
Therefore, activated and translocated PKC might be
an alternative mechanism to open mKATP-channels
(32). However, the exact mechanism leading to
mKATP-channel opening remains unclear.

It has been proposed that opening of the mKATPchannels leads to a depolarization of the mitochondrial membrane that would be favorable by reducing
the driving force for ATP generation (33). This was
speculated to preserve mitochondrial function during
ischemia and reperfusion by attenuating Ca2 loading
that would disrupt mitochondrial function by inducing Ca2 permeability transition (34). Riess et al. (35)
assessed NADH in isolated but energized hearts and
found that anesthetic exposure increased NADH (reduction) in isolated hearts, which is different from the
findings of Kohro et al. (36) who reported that an
anesthetic exposure increased flavoprotein oxidation
in de-energized isolated cardiac cells.
Others suggested that mKATP-channels opening significantly increased mitochondrial matrix volume
without changing membrane potential. Uptake of K
through the mKATP-channels would maintain mitochondrial matrix volume (25).
Recently, it was shown that short ischemic periods
interspersed with reperfusion were also cardioprotective when administered after the sustained infarct inducing ischemia, i.e., during reperfusion (3,17). Two

1258

CARDIOVASCULAR ANESTHESIA OBAL ET AL.

PERI-ISCHEMIC PROTECTION BY SEVOFLURANE

major mechanisms of cardioprotection induced by

postconditioning were discussed. First, there are passive effects that modify reperfusion injury by reduction of neutrophil accumulation (37), reduced membrane peroxygenation by impaired oxygen radical
generation and a reduced mitochondrial Ca2 load.
Prevention of sarcoplasmatic calcium oscillations during early reoxygenation by volatile anesthetics prevents hypercontracture of isolated cardiomyocytes,
thereby reducing the lethal cell injury (38).
As a more active part of postconditioning, the activation of different kinases, such as PI3-AKT kinase or
ERK1/2 kinase, was postulated to inhibit opening of the
mitochondrial permeability transition pore by phosphorylation of downstream targets such as endothelial
nitric oxide synthase (17). A recent study of Chiari et al.
(11) demonstrated that brief administration of 0.5 MAC
of isoflurane during reperfusion provided cardioprotection. They suggested that the ischemia-induced phosphorylation of PI3-AKT kinase might be enhanced in the
presence of isoflurane. However, the involvement of
PI3-AKT kinase and a possible interaction with mKATPchannels requires further investigations.
Similar to anesthetic-induced preconditioning, administration of a PKC inhibitor at the time of reperfusion partially reduced infarct size in pigs (39). Therefore, it was speculated that PKC might also be
involved in mediating the protection against reperfusion injury. However, this hypothesis also has to be
confirmed.
It is unknown whether the protection conferred by
preconditioning is caused by effects occurring during
ischemia or during initial reperfusion. An et al. (8)
demonstrated that sevoflurane preconditioning reduced Ca2 loading during ischemia, thereby attenuating myocardial cell damage and improving myocardial function during reperfusion. The protection of
volatile anesthetics appears to depend on the time of
administration. Varadarajan et al. (40) demonstrated,
in isolated guinea pig hearts, that even if sevoflurane
was given directly before ischemia the protection
against ischemia-reperfusion injury was much better
than when sevoflurane was only given during reperfusion. They speculated that the small decrease in
metabolic rate before ischemia metabolically triggers a
protective mechanism that might decrease myocardial
hypercontracture at the beginning of reperfusion.
Oxygen radicals play a dual role in protection
against ischemia and reperfusion injury. Although in
preconditioning ROS activate PKC-dependent cardioprotective signaling pathways (41), they cause irreversible cell damage during early reperfusion by direct membrane destruction and lipid peroxidation
(42,43). Administration of a mKATP-channel opener
before ischemia abolished the ROS burst at the beginning of reperfusion (44). Therefore, one might speculate that the cell injury by ROS during reperfusion was

ANESTH ANALG
2005;101:125260

reduced by abolished ROS release caused by preconditioning and by the administration of sevoflurane
during reperfusion.
Haelewyn et al. (45) demonstrated that administration
of desflurane before, during, and after ischemia resulted
in a maximal protective effect. Unfortunately, the preconditioning protocol was different from that used in the
group that received desflurane during the total procedure. Therefore, direct comparison between groups is
difficult. The authors demonstrated that the amount of
infarct size reduction was similar when desflurane was
given before or after myocardial ischemia.
Most often, 5HD is used as specific blocker of
mKATP-channels in studies performed on ischemic or
pharmacologic preconditioning. The specificity of
5HD has been questioned by Hanley et al. (46,47) who
demonstrated that 5HD is activated inside the mitochondria and further metabolized via -oxidation. Although other possible mechanisms of action of 5HD
are discussed (e.g., inhibition of mKATP-channels by
putative cytosolic 5HD or 5HD-CoA, by intermediates
of 5HD metabolism, or by inhibition of the respiratory
chain), the lack of inhibition of sevoflurane preconditioning remained unclear (47).
In the present study, we used continuous administration of 5HD to avoid possible hemodynamic or
pharmacological side effects of bolus injections. This
protocol resulted in different administration times and
total doses of 5-HD. We found no effect of 5HD on
infarct size when given for the longest time in the 5HD
group. Administration time of 5HD was shortest in
the S-Pre 5HD group and we cannot completely
exclude that this differences in the protocol might
have influenced the results of the study, i.e., that it
might explain the tendency of a smaller blockade in
the S-Pre 5HD group. However, the drug was given
for the same time before coronary artery occlusion in
the S-Post 5HD group and the effect of sevoflurane
during reperfusion was totally blocked.
The mechanism whereby opening of mKATPchannels leads to cardioprotection remains unclear.
Several possible mechanisms have been proposed.
Opening of mKATP-channels might lead to a depolarization of the mitochondrial membrane, causing dissipation of the mitochondrial potential and reducing
the force for Ca2 uptake and preservation of mitochondrial function during ischemia and reperfusion
(48). Others suggest that opening of mKATP-channels
increased mitochondrial matrix volume while membrane potential remains nearly stable (25).
Several studies have shown that apoptosis might
contribute to cardiac death, and that apoptosis induced by H2O2 can be inhibited by mKATP-channels
opening or be blocked by administration of 5HD (49).
In the present study, we used TTC staining after
2 hours of reperfusion to determine infarct size. Thus,
we cannot speculate about the mechanism of cell

ANESTH ANALG
2005;101:125260

death and the effect of sevoflurane on these

mechanisms.
In conclusion, the present study demonstrates that
mKATP-channels are not only involved in sevoflurane
preconditioning but also in postischemic cardioprotection by sevoflurane, i.e., anesthetic postconditioning.
Whether anesthetic preconditioning or anesthetic
postconditioning will become suitable strategies in the
clinical setting remains unclear. A recent study by De
Hert et al. (50) demonstrated that only administration
of sevoflurane throughout the surgical procedure reduced troponin I levels and length of stay on the
intensive care unit in patients undergoing coronary
artery surgery, whereas only prebypass (preconditioning) or administration only during reperfusion (postconditioning) were ineffective.
The work of Maximillane Kratz, cand. MD, is gratefully acknowledged.

References
1. Toller WG, Kersten JR, Pagel PS, et al. Sevoflurane reduces
myocardial infarct size and decreases the time threshold for
ischemic preconditioning in dogs. Anesthesiology 1999;91:
1437 46.
2. ORourke B. Evidence for mitochondrial K channels and their
role in cardioprotection. Circ Res 2004;94:420 32.
3. Zhao ZQ, Corvera JS, Halkos ME, et al. Inhibition of myocardial
injury by ischemic postconditioning during reperfusion: comparison with ischemic preconditioning. Am J Physiol Heart Circ
Physiol 2003;285:H579 88.
4. Schlack W, Preckel B, Stunneck D, Thamer V. Effect of halothane, enflurane, isoflurane, sevoflurane and desflurane on
myocardial reperfusion injury in the isolated rat heart. Br J
Anaesth 1998;81:9139.
5. Obal D, Preckel B, Scharbatke H, et al. One MAC of sevoflurane
provides protection against reperfusion injury in the rat heart in
vivo. Br J Anaesth 2001;87:90511.
6. Obal D, Scharbatke H, Barthel H, et al. Cardioprotection against
reperfusion injury is maximal with only two minutes of sevoflurane administration in rats. Can J Anaesth 2003;50:940 5.
7. Conzen PF, Vollmar B, Habazettl H, et al. Systemic and regional
hemodynamics of isoflurane and sevoflurane in rats. Anesth
Analg 1992;74:79 88.
8. An J, Varadarajan SG, Novalija E, Stowe DF. Ischemic and
anesthetic preconditioning reduces cytosolic [Ca2] and improves Ca2 responses in intact hearts. Am J Physiol Heart Circ
Physiol 2001;281:H1508 23.
9. Riess ML, Kevin LG, Amadou C, et al. Dual exposure to sevoflurane improves anesthetic preconditioning in intact hearts. Anesthesiology 2004;100:569 74.
10. Sandhu R, Diaz RJ, Mao GD, Wilson GJ. Ischemic
preconditioning: differences in protection and susceptibility to
blockade with single-cycle versus multicycle transient ischemia.
Circulation 1997;96:984 95.
11. Chiari PC, Bienengraeber MW, Pagel PS, et al. Isoflurane protects against myocardial infarction during early reperfusion by
activation of phosphatidylinositol-3-kinase signal transduction:
evidence for anesthetic-induced postconditioning in rabbits.
Anesthesiology 2005;102:1029.
12. Tarnow J, Markschies Hornung A, Schulte Sasse U. Isoflurane
improves the tolerance to pacing-induced myocardial ischemia.
Anesthesiology 1986;64:14756.
13. Cope DK, Impastato WK, Cohen MV, Downey JM. Volatile
anesthetics protect the ischemic rabbit myocardium from infarction. Anesthesiology 1997;86:699 709.

CARDIOVASCULAR ANESTHESIA
OBAL ET AL.
PERI-ISCHEMIC PROTECTION BY SEVOFLURANE

1259

14. Kersten JR, Schmeling TJ, Pagel PS, et al. Isoflurane mimics
ischemic preconditioning via activation of KATP channel: reduction of myocardial infarct size with an acute memory phase.
Anesthesiology 1997;87:36170.
15. Cason BA, Gamperl AK, Slocum RE, Hickey RF. Anestheticinduced preconditioning: previous administration of isoflurane
decreases myocardial infarct size in rabbits. Anesthesiology
1997;87:118290.
16. Preckel B, Schlack W, Comfe`re T, et al. Effects of enflurane,
isoflurane, sevoflurane and desflurane on reperfusion injury
after regional myocardial ischaemia in the rabbit heart in vivo.
Br J Anaesth 1998;81:90512.
17. Tsang A, Hausenloy DJ, Mocanu MM, Yellon DM.
Postconditioning: a form of modified reperfusion protects the
myocardium by activating the phosphatidylinositol 3-kinaseAkt pathway. Circ Res 2004;95:230 2.
18. Garlid KD, Paucek P, Yarov-Yarovoy V, et al. Cardioprotective
effect of diazoxide and its interaction with mitochondrial ATPsensitive K channels: possible mechanism of cardioprotection.
Circ Res 1997;81:1072 82.
19. Fryer RM, Eells JT, Hsu AK, et al. Ischemic preconditioning in
rats: role of mitochondrial KATP channel in preservation of
mitochondrial function. Am J Physiol Heart Circ Physiol 2000;
278:H30512.
20. ORourke B. Myocardial KATP channels in preconditioning. Circ
Res 2000;87:84555.
21. Pain T, Yang XM, Critz SD, et al. Opening of mitochondrial
KATP channels triggers the preconditioned state by generating
free radicals. Circ Res 2000;87:460 6.
22. Mullenheim J, Ebel D, Fradorf J, et al. Isoflurane preconditions
myocardium against infarction via release of free radicals. Anesthesiology 2002;96:934 40.
23. Novalija E, Varadarajan SG, Camara AK, et al. Anesthetic
preconditioning: triggering role of reactive oxygen and nitrogen
species in isolated hearts. Am J Physiol Heart Circ Physiol
2002;283:H44 52.
24. Kevin LG, Novalija E, Riess ML, et al. Sevoflurane exposure
generates superoxide but leads to decreased superoxide during
ischemia and reperfusion in isolated hearts. Anesth Analg 2003;
96:949 55.
25. Kowaltowski AJ, Seetharaman S, Paucek P, Garlid KD. Bioenergetic consequences of opening the ATP-sensitive K channel
of heart mitochondria. Am J Physiol Heart Circ Physiol 2001;
280:H649.
26. Stowe DF, Riess ML. Reactive oxygen species and cardiac
preconditioning: many questions remain. Cardiovasc Drugs
Ther 2004;18:8790.
27. Stowe DF, Kevin LG. Cardiac preconditioning by volatile anesthetic agents: a defining role for altered mitochondrial bioenergetics. Antioxid Redox Signal 2004;6:439 48.
28. Fujimoto K, Bosnjak ZJ, Kwok WM. Isoflurane-induced facilitation of the cardiac sarcolemmal KATP channel. Anesthesiology
2002;97:57 65.
29. Gassmayr S, Stadnicka A, Suzuki A, et al. Isoflurane sensitizes
the cardiac sarcolemmal adenosine triphosphate-sensitive potassium channel to pinacidil. Anesthesiology 2003;98:114 20.
30. Stowe DF, Heisner JS, Chung WW, Fujita S. Volatile anaesthetics
restore bradykinin and serotonin-induced coronary vasodilation after blocking nitric oxide synthase: lack of anaesthetic
effects on KATP channels and prostaglandin pathways. Eur J
Anaesthesiol 2001;18:219 30.
31. Zaugg M, Lucchinetti E, Spahn DR, et al. Volatile anesthetics
mimic cardiac preconditioning by priming the activation of
mitochondrial KATP channels via multiple signaling pathways.
Anesthesiology 2002;97:4 14.
32. Ohnuma Y, Miura T, Miki T, et al. Opening of mitochondrial
KATP channel occurs downstream of PKC-e activation in the
mechanism of preconditioning. Am J Physiol Heart Circ Physiol
2002;283:H440 7.
33. Kohro S, Hogan QH, Nakae Y, et al. Anesthetic effects on
mitochondrial ATP-sensitive K channel. Anesthesiology 2001;
95:1435 40.

1260

CARDIOVASCULAR ANESTHESIA OBAL ET AL.

PERI-ISCHEMIC PROTECTION BY SEVOFLURANE

34. Ferrari R. The role of mitochondria in ischemic heart disease.

J Cardiovasc Pharmacol 1996;28 Suppl 1:S110.
35. Riess ML, Camara AK, Chen Q, et al. Altered NADH and
improved function by anesthetic and ischemic preconditioning
in guinea pig intact hearts. Am J Physiol Heart Circ Physiol
2002;283:H53 60.
36. Sato T, ORourke B, Marban E. Modulation of mitochondrial
ATP-dependent K channels by protein kinase C. Circ Res
1998;83:110 4.
37. Heindl B, Reichle FM, Zahler S, et al. Sevoflurane and isoflurane
protect the reperfused guinea pig heart by reducing postischemic adhesion of polymorphonuclear neutrophils. Anesthesiology 1999;91:52130.
38. Siegmund B, Schlack W, Ladilov YV, et al. Halothane protects
cardiomyocytes against reoxygenation-induced hypercontracture. Circulation 1997;96:43729.
39. Inagaki K, Chen L, Ikeno F, et al. Inhibition of d-protein kinase
C protects against reperfusion injury of the ischemic heart in
vivo. Circulation 2003;108:2304 7.
40. Varadarajan SG, An J, Novalija E, Stowe DF. Sevoflurane before
or after ischemia improves contractile and metabolic function
while reducing myoplasmic Ca2 loading in intact hearts. Anesthesiology 2002;96:12533.
41. Oldenburg O, Cohen MV, Downey JM. Mitochondrial KATP
channels in preconditioning. J Mol Cell Cardiol 2003;35:569 75.
42. Zweier JL, Flaherty JT, Weisfeldt ML. Direct measurement of
free radical generation following reperfusion of ischemic myo-

ANESTH ANALG
2005;101:125260

cardium. Proc Natl Acad Sci U S A 1987;84:1404 7.

43. Becker LB. New concepts in reactive oxygen species and cardiovascular reperfusion physiology. Cardiovasc Res 2004;61:
46170.
44. Ozcan C, Bienengraeber M, Dzeja PP, Terzic A. Potassium channel openers protect cardiac mitochondria by attenuating oxidant
stress at reoxygenation. Am J Physiol Heart Circ Physiol 2002;
282:H5319.
45. Haelewyn B, Zhu L, Hanouz JL, et al. Cardioprotective effects of
desflurane: effect of timing and duration of administration in rat
myocardium. Br J Anaesth 2004;92:5527.
46. Hanley PJ, Gopalan KV, Lareau RA, et al. -oxidation of
5-hydroxydecanoate, a putative blocker of mitochondrial ATPsensitive potassium channels. J Physiol 2003;547:38793.
47. Hanley PJ, Drose S, Brandt U, et al. 5-Hydroxydecanoate is
metabolized in mitochondria and creates a rate-limiting bottleneck for -oxidation of fatty acids. J Physiol 2005;562:30718.
48. Liu Y, Sato T, Seharaseyon J, et al. Mitochondrial ATPdependent potassium channels: viable candidate effectors of
ischemic preconditioning. Ann N Y Acad Sci 1999;874:2737.
49. Akao M, Ohler A, ORourke B, Marban E. Mitochondrial ATPsensitive potassium channels inhibit apoptosis induced by oxidative stress in cardiac cells. Circ Res 2001;88:126775.
50. De Hert SG. Cardioprotective properties of sevoflurane in patients undergoing coronary surgery with cardiopulmonary bypass are related to the modalities of its administration. Anesthesiology 2004;101:299 310.

Isoflurane Prevents Platelets from Enhancing NeutrophilInduced Coronary Endothelial Dysfunction

Guochang Hu,

MD, PhD*,

M. Ramez Salem,

MD*,

and George J. Crystal,

PhD*

*Department of Anesthesiology, Advocate Illinois Masonic Medical Center, and Department of Anesthesiology, University
of Illinois College of Medicine; Department of Physiology and Biophysics, University of Illinois College of Medicine,
Chicago, Illinois

We evaluated whether platelets can enhance polymorphonuclear neutrophil-induced coronary endothelial

dysfunction, and, after observing this, whether isoflurane can prevent the effect. Neutrophils, coronary artery
segments, and platelets were obtained from 25 healthy
dogs. Coronary artery rings were exposed to neutrophils
activated with platelet-activating factor (1.0 M), and
after washing and preconstriction with U46619, were
evaluated for concentration-related responses to acetylcholine, an endothelium-dependent vasorelaxing drug.
Superoxide production by activated neutrophils was
measured spectrophotometrically. Adherence of the activated neutrophils to the endothelium of coronary segments was assessed by direct counting of neutrophils
labeled with fluorescent dye. Measurements were performed in absence and presence of isoflurane (1 minimum alveolar concentration) both with and without

vidence has emerged indicating interactions between neutrophils and platelets (1 4). Either type
of cell may release soluble factors that activate the
other type or increase its response to stimulating agents.
The primary mechanism for the platelet-neutrophil interaction appears to be a direct receptor ligand interaction involving P-selectin on the surface of the platelets
and P-selectin glycoprotein ligand-1 (PSGL-1) on the
neutrophil surface. Platelets have been shown to enhance superoxide production by neutrophils, neutrophil
adhesion to endothelial cells, and neutrophil-induced,
postischemic cardiac dysfunction in vitro (4 6) and to

Supported, in part, by funding from the Department of Anesthesiology, Advocate Illinois Masonic Medical Center, Chicago, Illinois.
Presented in part at the annual meeting of the American Society
of Anesthesiologists, Orlando, Florida, October 1216, 2002.
Accepted for publication June 10, 2005.
Address correspondence and reprint requests to George J. Crystal, PhD, Department of Anesthesiology, Advocate IL Masonic Medical Center, 836 West Wellington Avenue, Chicago, IL 606575193.
Address e-mail to gcrystal@uic.edu.
DOI: 10.1213/01.ANE.0000181340.28271.4F
2005 by the International Anesthesia Research Society
0003-2999/05

platelets. The presence of platelets enhanced the

neutrophil-induced rightward shift in the concentrationvasorelaxation response curve to acetylcholine (the concentration of acetylcholine required to elicit 50% of maximal relaxation (log M) was increased from 6.78 0.7
to 5.26 0.6), and it increased superoxide oxide production from 45.0 4.2 to 54.3 4.2 nM O2/5 106
neutrophils and adherence of activated neutrophils from
204 10 to 268 5 neutrophils/mm2. Isoflurane abolished these effects of platelets. In conclusion, platelets
enhanced the ability of neutrophils to cause coronary
endothelial dysfunction. This effect was prevented by
isoflurane. This may be attributable to an inhibitory action on superoxide production by the neutrophils leading to reduced expression of endothelial adhesion molecules and, in turn, reduced neutrophil adherence.
(Anesth Analg 2005;101:12618)

enhance neutrophil adhesion and postangioplasty vasoconstriction in the carotid artery in vivo (7). These findings all suggest an ability of platelets to enhance
neutrophil-induced endothelial dysfunction, but this has
not been confirmed experimentally.
Volatile anesthetics have been demonstrated to protect
the myocardium against ischemic-reperfusion injury (8),
but despite extensive investigation, the mechanisms underlying this effect remain unclear. Activation of protective signaling pathways within the myocytes involving
the adenosine triphosphate-sensitive potassium (KATP)
channels appears to play a prominent role (8), but various lines of evidence have suggested that inhibitory
actions on inflammatory pathways may also be involved. First, isoflurane reduced neutrophil adherence to
the endothelium and the cardiac and endothelial dysfunction caused by the activated neutrophils (9 12). Furthermore, isoflurane reduced the cytokine-induced
death of cultured endothelial and smooth muscle cells
and the release of tumor necrosis factor- (13).
Although some studies have shown that isoflurane
has no effect on platelet function (14 16), others have
suggested that it has an inhibitory effect at clinically
Anesth Analg 2005;101:12618

1261

1262

CARDIOVASCULAR ANESTHESIA HU ET AL.

ISOFLURANE PROTECTS THE ENDOTHELIUM

relevant concentrations (1720). In addition, several

reports have demonstrated that volatile anesthetics
can modulate the formation of platelet-neutrophil aggregates and expression of P-selectin on platelets but
that these effects may differ among the various anesthetics (2123). No study has assessed whether isoflurane (or any other volatile anesthetic) can affect the
ability of platelets to modulate neutrophil-induced
coronary endothelial dysfunction. Such information is
important in gaining a more complete understanding
of the mechanisms by which volatile anesthetics protect the myocardium from reperfusion injury in vivo.
The first objective of this study was to demonstrate
the ability of platelets to enhance neutrophil-induced
coronary endothelial dysfunction. After this was accomplished, we evaluated the effect of isoflurane on
this phenomenon. Mechanistic insights were obtained
by evaluating the associated changes in superoxide
production and vascular adherence of the activated
neutrophils.

Methods
The study was conducted after approval from the
institutional animal research committee at the University of Illinois at Chicago.
Experiments were performed on 25 healthy,
heartworm-free mongrel dogs of either sex (weight,
18 26 kg). The dogs were anesthetized with an IV
bolus injection of 30 mg/kg sodium pentobarbital and
mechanically ventilated (Air Shields Inc, Hatboro, PA)
with oxygen-enriched room air. The right carotid artery was cannulated to permit withdrawal of blood for
separation of neutrophils and platelets.
Neutrophils were separated as described previously
(11). Arterial blood was collected in tubes and anticoagulated with 1.6% citric acid and 2.5% sodium citrate
(pH 5.4) in 10 mL of 6% dextran solution in buffered
Hanks balanced salt solution (HBSS). The blood was
maintained at room temperature while erythrocytes
sedimented (approximately 40 min). The leukocyterich plasma layer was carefully aspirated and centrifuged at 500g at 4C for 10 min. Contaminating erythrocytes in the pellet were removed by hypotonic lysis
for 20 s with 9 mL of sterile distilled water. Subsequent addition of 3 mL of 0.6 M KCl and 15 mL of
buffered HBSS rapidly returned the cells to isotonicity.
The leukocyte-rich suspension was centrifuged at 500g
at 4C for 10 min, after which time the cells were
resuspended in 2 mL of HBSS, layered on the top of
3 mL of Ficoll-Pacque, and centrifuged again at 800g at
4C for 20 min. The resulting pellet was rinsed with
HBSS. The neutrophils were resuspended in HBSS in
preparation before experimental use. Our procedure
for neutrophil isolation yields neutrophil suspensions

ANESTH ANALG
2005;101:12618

that are 98% pure and more than 95% viable as evaluated by Wrights staining and trypan blue exclusion
(11,24).
Platelets were isolated by centrifugation as described previously (25). Briefly, arterial blood was collected in tubes and combined with citrate anticoagulant A to yield final concentrations of 9.3 mM sodium
citrate, 0.7 mM citric acid, and 14 mM dextrose. This
blood was then centrifuged at 55g at 4C for 40 min.
The platelet-rich plasma layer was carefully collected
and added into an equal volume of cold citrate anticoagulant B solution (in mM: sodium citrate 93, citric
acid 7, dextrose 105, and KCl 5, pH 6.5), and then the
mixture was centrifuged at 600g at 4C for 20 min. The
supernatant was discarded, and the resulting platelet
pellet was resuspended in a small volume of the citrate anticoagulant B. A platelet count of this suspension was obtained manually after Wrights staining
(24).
To quantify neutrophil-induced damage to the endothelium, dogs were killed with an overdose of IV
pentobarbital sodium (80 mg/kg), and the heart was
rapidly excised and immediately placed into cold
(4C), oxygenated Krebs solution with the following
composition (mM): NaCl 118, KCl 4.75, MgSO4 1.19 ,
KH2PO4 1.12 , CaCl2 2.54, NaHCO3 12.5, glucose 10.
The left anterior descending and circumflex coronary
arteries were carefully dissected, cleaned of connective and adipose tissue without disturbing the endothelium, and cut into rings 23 mm in length as described previously (11). The rings were mounted on
stainless steel hooks, placed into organ baths containing 10 mL of Krebs solution aerated with a gas mixture
of 95%O25% CO2 at 37C, and connected to isometric
force transducers (Radnoti Model TR 001). Changes in
isometric force were digitized at 3 Hz using an analogto-digital converter and analyzed using a videographic program (SPECTRUM; Triton Technology,
San Diego, CA). The rings were initially stretched to
yield a preload of 1.0 g of tension. After equilibration
for 20 min, the contraction response to KCl (30 mM)
was determined as tension was increased in 1.0-g
steps until the maximal response was observed. This
prestretch tension was used in subsequent procedures. The rings were then thoroughly washed and
the optimal contracting dose for U46619 (Upjohn,
Kalamazoo, MI), a thromboxane A2 mimetic agent,
was determined. A concentration-response curve for
U46619 was generated for each ring using concentrations over the range 2.5 to 5.0 nM. The concentration
for U46619 that caused the maximal contracting effect
was used to provide vascular tone in the acetylcholine
(ACh) and sodium nitroprusside (SNP) trials. After a
thorough washing, the rings were allowed to stabilize
at baseline tension for 20 min and then carefully removed from the chamber. The rings were then placed
in tightly capped glass test tubes containing Krebs

ANESTH ANALG
2005;101:12618

solution saturated with oxygen within a heated (37C)

shaking bath. Neutrophils (1 107 cells/mL), plateletactivating factor (PAF; 1 M) (11), and platelets (1
108 cells/mL), when appropriate, were added to the
tubes alone or together with 0.30 mM isoflurane in
Krebs solution and incubated for 25 min. This millimolar concentration for isoflurane was calculated using an isoflurane 1.0 minimum alveolar concentration
(MAC) of 1.4 vol % for the dog (26) and a buffer/gas
partition coefficient of 0.55 at 37C and 1 atm (27). A
stable level of isoflurane during all experimental protocols was confirmed by gas chromatography. The
concentrations for PAF and isoflurane in the ring studies were also used in the superoxide and adherence
studies described below. The main experimental
groups in the ring studies were as follows: 1) control;
no neutrophils, PAF, or platelets; 2) PAF and neutrophils; 3) PAF, neutrophils, and platelets; 4) PAF, neutrophils, and isoflurane; 5) PAF, neutrophils, platelets,
and isoflurane. Additional validation studies were
conducted with platelets both alone and with PAF and
with neutrophils alone.
After the incubation period with the inflammatory
cells, the rings were assessed for endothelial function.
The rings were removed from the test tubes and washed
three times with Krebs solution to remove neutrophils,
platelets, PAF, and isoflurane. They were then remounted in the chambers, and allowed to stabilize at the
appropriate pre-tension for 40 min. Indomethacin (10
M) was added to the chambers to prevent vascular
responses to prostaglandins (11,32). Once a stable contraction to the predetermined U46619 dose was obtained, endothelial function was assessed from cumulative concentration-response curves to the endotheliumdependent dilator ACh (109 to 106 M). After the rings
were again washed and allowed to stabilize, vascular
smooth muscle function was assessed from the curves to
the endothelium-independent dilator SNP (109 to 106
M). Drug concentrations are expressed as the final concentrations in the organ chamber.
Superoxide production by neutrophils in suspension
was determined by measuring the superoxide dismutase
(SOD) inhibitable reduction of ferricytochrome c to ferrocytochrome c (11). Neutrophils (5 106 cells/mL)
or/and platelets (5 107 cells/mL) were prewarmed in
a shaking incubator at 37C with 160 M cytochrome c
and 5 g/mL cytochalasin B in the absence (control) or
presence of the test agents for 5 min and then stimulated
with PAF. Half of the tubes were provided with an
excess of SOD (100 g/mL) as a control for nonspecific
activity or color generation. Five min after adding PAF,
cytochrome c reduction was measured spectrophotometrically by determining the optical density of the supernatant at 550 nm, using a Vmax kinetic microtiter
plate reader (Molecular Devices, Palo Alto, CA). Superoxide production was calculated using an extinction coefficient of 21 mM1cm1 for cytochrome c. Results are

CARDIOVASCULAR ANESTHESIA
HU ET AL.
ISOFLURANE PROTECTS THE ENDOTHELIUM

1263

expressed as nM of SOD-inhibitable O2 produced by a

suspension of 5 106 neutrophils/mL. The experimental groups were the same as those in the endothelial
dysfunction studies described above. Further validation
studies were performed to determine the ability of the
PAF-activated platelets alone to produce superoxide.
Neutrophil adherence to the endothelial surface of
the coronary artery segments was assessed using neutrophils labeled with a vital fluorescent dye, as described previously (11). Briefly, 1 mL of 4 M solution
of PKH26 dye was added to 1 mL of a neutrophil
suspension (2 107 cells/mL). After the sample was
gently mixed, it was incubated at room temperature
for 5 min; the labeling reaction was stopped by adding
2 mL plasma and incubating for 1 min. The plasmastopped sample was diluted with 4 mL of HBSS and
then centrifuged at 400g for 10 min at 4C. The resultant cell pellet was transferred to a new tube for additional duplicate washings, and the cells were resuspended. This labeling procedure yields neutrophils
possessing normal viability and function (11).
Coronary artery rings were carefully opened without damaging the endothelium and placed in the tube
containing 3 mL of HBSS with Ca2 at 37C. Labeled
neutrophils were added to each tube to achieve a final
concentration of 5 106 neutrophils/mL. The coronary vascular segments were incubated with labeled
neutrophils in the absence and presence of platelets (5
107 cells/mL) and isoflurane for 20 min. PAF was
used as an activator of the cells. After removing the
vascular segments, they were flushed gently with
HBSS. Neutrophil adherence was determined by
counting the number of neutrophils adhering to the
endothelial surface in six separate microscopic fields
under epifluorescent microscopy (490-nm excitation,
504-nm emission) (Fryer Company, Inc, Huntley, IL),
and expressed per square millimeter of endothelium.
The experimental groups were the same as those in the
endothelial dysfunction studies described above.
The following chemicals and reagents were obtained
from Sigma Chemical (St. Louis, MO): acetylcholine
chloride, SNP, Ficoll-Pacque, SOD, cytochrome c, cytochalasin B, PKH26 dye, and indomethacin. PAF and
HBSS without Mg2 and Ca2 were purchased from
Avanti Polar Lipids (Alabaster, AL) and Meditech, Inc.
(Salt Lake City, UT), respectively. All solutions were
prepared freshly on the day of the study.
Concentration-responses for vascular relaxation were
calculated as a percentage of the decrease of U46619induced isometric force. The drug concentration required to elicit 50% of the maximum relaxation response
(EC50) was calculated by linear regression analysis and
expressed as the negative logarithm of the drug concentration (log [M]). The maximal relaxation response
(Rmax, %) was also determined. An Rmax value equal to
100% indicated complete reversal of U46619-induced

1264

CARDIOVASCULAR ANESTHESIA HU ET AL.

ISOFLURANE PROTECTS THE ENDOTHELIUM

ANESTH ANALG
2005;101:12618

Figure 1. Effects of unactivated neutrophils (PMNs)

and platelet-activating factor (PAF)-activated PMNs
alone and in the presence of isoflurane (ISO) with
and without platelets (PLT) on relaxation responses
of coronary artery rings to acetylcholine (A) and
sodium nitroprusside (B). *P 0.05 versus control;
P 0.05 versus PMNPAF. Values are mean se.
Number of coronary rings for each condition presented in Table 1.

contraction. A one-way analysis of variance in combination with the Student-Newman-Keuls test were used to
evaluate treatment effects on superoxide production and
neutrophil adherence (28). Within- and between-group
differences for coronary vascular relaxation responses
were assessed using a two-way analysis of variance for
repeated measures followed by the Student-NewmanKeuls test (28). A value of P 0.05 was considered
significant throughout the study.

Results
ACh caused concentration-dependent relaxation responses in the control coronary rings (Fig. 1A). Unacti-

vated neutrophils had no effect on these responses. Activating the neutrophils with PAF reduced the relaxation
responses of the coronary rings to ACh, as reflected in a
rightward shift in the concentration-response curve, a
reduction in the maximum response, and an increase
in the EC50 (Fig. 1A and Table 1). Combining the
PAF-activated neutrophils with platelets enhanced
these effects. Isoflurane reversed the impairment to
ACh-induced coronary relaxation caused by the PAFactivated neutrophils both alone and combined with
platelets. The concentration-related vascular relaxation responses to SNP remained unchanged under all
conditions (Fig. 1B). Neither platelets alone or combined with PAF nor neutrophils alone altered the vascular relaxation responses caused by ACh (data not
shown).

ANESTH ANALG
2005;101:12618

CARDIOVASCULAR ANESTHESIA
HU ET AL.
ISOFLURANE PROTECTS THE ENDOTHELIUM

1265

Table 1. Effect of Isoflurane on Changes in Coronary Vascular Relaxation Responses Caused by Neutrophils Alone and
Combined with Platelets
Condition
Acetylcholine
Control
PMN PAF
PMN PAF PLT
PMN PAF ISO
PMN PAF PLT ISO
Sodium nitroprusside
Control
PMN PAF
PMN PAF PLT
PMN PAF ISO
PMN PAF PLT ISO

EC50(log[M])

Rmax (%)

29
25
20
17
18

7.39 0.12
6.78 0.07*
5.26 0.06*
7.26 0.12
7.31 0.09

110.1 5.4
80.2 3.5*
58.3 2.8*
108.7 3.8
110.5 4.9

29
25
20
17
18

7.35 0.06
7.37 0.08
7.36 0.09
7.45 0.03
7.32 0.07

106.7 4.2
109.5 4.7
108.2 4.5
107.9 4.6
107.3 4.5

Values are mean se.

EC50 drug concentration eliciting 50% of the maximal relaxation response; Rmax maximal relaxation response; the % value indicates the degree to which
this response reversed the baseline (pre-drug) U46619-induced contraction; PMN polymorphonuclear neutrophils; PAF platelet activating factor; PLT
platelets; ISO isoflurane.
* P 0.05 versus control; P 0.05 versus PMN PAF groups.

Figure 2 presents the effects of isoflurane on superoxide production by PAF-stimulated neutrophils alone
and in the presence of platelets. PAF stimulation of neutrophils caused a ninefold increase in superoxide production. This response was greater in the presence of
platelets. Isoflurane inhibited superoxide production by
neutrophils alone and it abolished the ability of platelets
to enhance this production. PAF-activated platelets
alone produced a negligible amount of superoxide.
Figure 3 demonstrates that PAF caused a pronounced increase neutrophil adherence to the coronary artery rings, which was enhanced by the presence of platelets. Isoflurane reduced adherence of the
activated neutrophils alone and also abolished the
platelet-induced enhancement of this effect.

Discussion
The current findings demonstrated for the first time
that platelets can enhance neutrophil-induced coronary endothelial dysfunction and that isoflurane can
prevent this effect.
There is an extensive body of evidence suggesting
that interactions between neutrophils and platelets
may be important in the pathophysiology of myocardial and coronary endothelial reperfusion injury (1 4).
Platelet-neutrophil conjugates have been demonstrated to enhance neutrophil activation and
neutrophil-induced cardiac injury (5,6), as reflected by
the enhanced increases in superoxide production, vascular adherence, and the coronary endothelial dysfunction observed in the current study. Previous studies have suggested that platelets promote neutrophil
activation via the release of several mediators, including thromboxane A2, platelet-derived growth factor,
platelet factor 4, and serotonin (3). Although the main

Figure 2. Effects of isoflurane (ISO) on superoxide production (O2)

by platelet-activating factor (PAF)-stimulated neutrophils (PMNs)
alone and in the presence of platelets (PLT). *P 0.05 versus PMN;
P 0.05 versus corresponding without ISO. Values are mean se
of at least eight observations.

mechanism for platelet-neutrophil interactions is

P-selectin on the surface of the platelets and PSGL-1
on the surface of the neutrophils, other mechanisms
have been implicated. These include 1) fibrinogen
bridging via GPIIb/IIIa on the platelet and the 2
integrin MAC-1 (CD11b/CD18) on the neutrophil, 2)
thrombospondin bridging via GPIa/IIa, GPIIb/IIIa, or
GPIV on the platelet and a specific receptor on the
neutrophil, and 3) platelet intercellular adhesion
molecule-2 (ICAM-2) binding to neutrophil lymphocyte function-associated antigen-1 (LFA-1) (29).
Platelets can adhere to the endothelium via glycoprotein 1b on the platelet surface and von Willebrand
factor in the vessel wall and this interaction can cause
release of chemotactic proteins and expression of adhesion molecules in the endothelium, thus promoting

1266

CARDIOVASCULAR ANESTHESIA HU ET AL.

ISOFLURANE PROTECTS THE ENDOTHELIUM

Figure 3. Effects of isoflurane (ISO) on platelet-activating factor

(PAF)-stimulated adherence of neutrophils to coronary vascular
endothelium in the absence and presence of platelets (PLT). *P
0.05 versus PMN; P 0.05 versus corresponding without ISO.
Values are mean se of at least 8 coronary rings.

neutrophil-endothelial interactions and endothelial injury (30). Although this mechanism may have played
a role in our adherence and vascular relaxation studies, it could not have been involved in the superoxide
studies, which did not include coronary arterial
segments.
It has been demonstrated that the platelets themselves
can produce superoxide but, as confirmed in our validation studies, this production is much smaller than that of
the neutrophils and is likely limited to intracellular signaling (31). We also demonstrated that activated platelets in the absence of neutrophils had no effect on the
relaxation responses of the coronary artery segments.
These findings, when taken together, would seem to
eliminate the possibility that the enhanced endothelial
dysfunction caused by the platelet-neutrophil combination was attributable to an independent effect of the
platelets.
Activated neutrophils produce superoxide, whose action on the coronary vasculature can result in endothelial
dysfunction. Superoxide production by the neutrophil
can be via adherence-independent or adherencedependent pathways (32). Our superoxide production
and vascular adherence findings suggest that the platelets enhanced both these pathways and that isoflurane
abolished these effects.
The mechanisms by which isoflurane inhibited the
platelet-induced enhancement of neutrophil-induced
endothelial dysfunction remain to be clarified. Several
potential mechanisms can be proposed:
1. Isoflurane had an inhibitory effect on the platelets,
which reduced interactions between the platelets and
both the neutrophils and endothelium. Previous studies
are equivocal regarding a potential role for this mechanism. Some have shown that isoflurane does not affect
platelet function (15), whereas others have suggested

ANESTH ANALG
2005;101:12618

that it has an inhibitory effect at clinically relevant concentrations (19). Moreover, studies have suggested that
isoflurane may have a facilitating, rather than inhibitory,
effect on expression of P-selectin in activated platelets
and on platelet-neutrophil binding (21). However, the
use of a larger concentration for isoflurane (2 MAC),
different agonists (adenosine diphosphate and thrombin
receptor agonist protein TRAP-6), and cells contained in
whole blood samples from a different species (humans)
make it difficult to extrapolate these latter findings to the
present study. Although isoflurane has been shown to
reduce thrombin-induced platelet adherence to the coronary vascular endothelium in isolated guinea pig hearts
(19), it has also been shown to increase expression of
glycoprotein 1b in adenosine diphosphate-activated
platelets in vitro (23). This suggests that the former finding may reflect an inhibitory effect on the endothelium
rather than on the platelets.
2. Isoflurane had an inhibitory effect on the neutrophils that rendered them unresponsive to the stimulatory action of the platelets. PAF binds to a specific receptor on the neutrophil membrane, which is the first event
in the signal transduction sequence. Ultimately, the production of superoxide by the neutrophil results from
activation and assembly of nicotinamide adenine dinucleotide phosphate (NADPH) oxidase, which is a transmembrane electron transport chain that reduces oxygen
to superoxide. The isoflurane-induced inhibition of superoxide production by neutrophils alone in the current
study could be attributable to a direct inhibitory effect on
NADPH oxidase or to an inhibitory effect at some site in
the signal transduction pathway regulating NADPH oxidase, e.g., the PAF receptor on the neutrophil membrane or the GTP-binding proteins (G proteins) that are
involved in transduction of agonist signals. This inhibitory action of isoflurane on NADPH oxidase may prevent modulation from upstream stimulatory mediators,
including those released by the platelets.
3. Isoflurane had a protective effect on the coronary
endothelium that prevented the dysfunction caused
by activated neutrophils both alone and in the presence of platelets. Free radicals released from neutrophils, e.g., superoxide, provide a stimulus for upregulation of endothelial adhesion molecules (33). The
main endothelial adhesion molecules are P-selectin,
which mediates the initial rolling and slowing of
neutrophils along the endothelial surface, and
ICAM-1, which mediates the later firm adherence of
neutrophils to the endothelial cell surface (11). In the
PAF-stimulated in vitro system used in the current
study, adherence is via ICAM-1 (32). We have demonstrated in an isolated rat heart preparation (in the
absence of platelets) that selective exposure of the
coronary vasculature to isoflurane can reduce adherence of activated neutrophils (10). Although not confirmed by direct measurements, this reduced adherence would likely have resulted in better preserved

ANESTH ANALG
2005;101:12618

endothelial function. Exposure to isoflurane may protect the coronary endothelium from the deleterious
effects of activated neutrophils both in the absence
and presence of platelets.
Our findings confirm our previous work indicating
that isoflurane has a profound inhibitory effect on neutrophil activation and functions (11). In this regard, it is
noteworthy that isoflurane was capable of attenuating,
but not abolishing, the increase in superoxide production by the activated neutrophils, whereas it could completely abolish the neutrophil-induced endothelial dysfunction and the associated increases in neutrophil
adherence. This apparent quantitative divergence can be
reconciled if the ability of activated neutrophils to interact and cause dysfunction within the endothelium was a
threshold phenomenon and isoflurane was able to reduce neutrophil activation below the critical level.
PAF is a highly active phospholipid that has been
implicated in inflammatory responses and cardiac
reperfusion injury in vivo (34). Exogenous PAF was
used as an agonist in the present study because of its
well-established ability to activate the three cell types
investigated: neutrophils, platelets, and endothelial
cells. We followed previous studies and treated the
coronary vascular rings with indomethacin to inhibit
prostaglandin synthesis (11,32). Although indomethacin has been demonstrated to increase the sensitivity
of canine coronary arteries to the constrictor effects of
U46619 (35), there is no reason to expect that this
would have influenced our results or conclusions.
A blunted response to the endothelium-dependent
vasorelaxing drug ACh could reflect impaired endothelial or vascular smooth function. SNP, a nitric oxide
donor, is an endothelium-independent vasorelaxing
drug that was used to differentiate between these
possibilities. The inability of neutrophils alone, or
combined with platelets, to attenuate the responses to
SNP implied well preserved vascular smooth function
and suggested endothelial dysfunction as an explanation for the blunted responses to ACh.
In a previous investigation (11), we performed a
control study that indicated that treatment with isoflurane alone did not alter the ACh vasorelaxation curve.
This finding eliminates the possibility that isofluranes
beneficial effect on ACh-induced coronary relaxation
in the studies involving activated neutrophils and
platelets was caused by an enhancement of baseline
endothelial function and was not related to protection
against neutrophil- and/or platelet-induced endothelial damage.
In conclusion, the present study demonstrated that
isoflurane can inhibit the ability of platelets to enhance
neutrophil-induced vascular endothelial dysfunction
in the coronary circulation. Our findings provide additional support for the notion that the cardioprotective effects of isoflurane involve inhibitory actions on
inflammatory cells and their interactions, as well as

CARDIOVASCULAR ANESTHESIA
HU ET AL.
ISOFLURANE PROTECTS THE ENDOTHELIUM

1267

KATP channel-dependent signaling pathways within

myocytes (8).
The authors thank Derrick L. Harris, BS, for technical assistance.

References
1. Li N, Hu H, Lindqvist M, et al. Platelet-leukocyte cross talk in
whole blood. Arterioscler Thromb Vasc Biol 2000;20:2702 8.
2. Dore M. Platelet-leukocyte interactions. Am Heart J 1998;135:
S146 51.
3. Siminiak T, Flores NA, Sheridan DJ. Neutrophil interactions
with endothelium and platelets: possible role in the development of cardiovascular injury. Eur Heart J 1995;16:160 70.
4. Kogaki S, Sawa Y, Matsushita T, et al. Selectin on activated
platelets enhances neutrophil endothelial adherence in myocardial reperfusion injury. Cardiovasc Res 1999;43:968 73.
5. Naum CC, Kaplan SS, Basford RE. Platelets and ATP prime
neutrophils for enhanced O2 generation at low concentrations
but inhibit O2 generation at high concentrations. J Leukoc Biol
1991;49:839.
6. Lefer AM, Campbell B, Scalia R, Lefer DJ. Synergism between
platelets and neutrophils in provoking cardiac dysfunction after
ischemia and reperfusion: role of selections. Circulation 1998;
98:1322 8.
7. Merhi Y, Provost P, Guidoin R, Latour J-G. Importance of platelets
in neutrophil adhesion and vasoconstriction after deep carotid
arterial injury by angioplasty in pigs. Arterioscler Thromb 1997;17:
118591.
8. Tanaka K, Ludwig LM, Kersten JR, et al. Mechanisms of cardioprotection by volatile anesthetics. Anesthesiology 2004;100:
70721.
9. Hu G, Vasiliauskas T, Salem MR, et al. Neutrophils pretreated
with volatile anesthetics lose ability to cause cardiac dysfunction. Anesthesiology 2003;98:712 8.
10. Hu G, Salem MR, Crystal GJ. Isoflurane and sevoflurane precondition against neutrophil-induced contractile dysfunction in
isolated rat hearts. Anesthesiology 2004;100:489 97.
11. Hu G, Vinten-Johansen J, Salem MR, et al. Isoflurane inhibits
neutrophil-endothelium interactions in the coronary circulation:
lack of role for adenosine triphosphate-sensitive potassium
channels. Anesth Analg 2002;94:849 56.
12. Mobert J, Zahler S, Becker BF, Conzen PF. Inhibition of neutrophil activation by volatile anesthetics decreases adhesion to
cultured human endothelial cells. Anesthesiology 1999;90:
1372 81.
13. de Klaver MJ, Manning L, Palmer LA, Rich GF. Isoflurane
pretreatment inhibits cytokine-induced cell death in cultured rat
smooth muscle cells and human endothelial cells. Anesthesiology 2002;97:24 32.
14. Hirakata H, Ushikubi F, Toda H, et al. Sevoflurane inhibits
human platelet aggregation and thromboxane A2 formation,
possibly by suppression of cyclooxygenase activity. Anesthesiology 1996;85:144753.
15. Hirakata H, Ushikubi F, Narumiya S, et al. The effect of inhaled
anesthetics on the platelet aggregation and the ligand-binding
affinity of the platelet thromboxane A2 receptor. Anesth Analg
1995;81:114 8.
16. Kohro S, Yamakage M. Direct inhibitory mechanisms of halothane on human platelet aggregation. Anesthesiology 1996;85:
96 106.
17. Honemann CW, Nietgen GW, Podranski T, et al. Influence of
volatile anesthetics on thromboxane A2 signaling. Anesthesiology 1998;88:440 51.
18. Blaise GA, Parent M, Laurin S, et al. Platelet-induced vasomotion of isolated canine coronary artery in the presence of halothane or isoflurane. J Cardiothorac Vasc Anesth 1994;8:175 81.

1268

CARDIOVASCULAR ANESTHESIA HU ET AL.

ISOFLURANE PROTECTS THE ENDOTHELIUM

19. Heindl B, Becker BF, Zahler S, Conzen PF. Volatile anaesthetics

reduce adhesion of blood platelets under low-flow conditions in
the coronary system of isolated guinea pig hearts. Acta Anaesthesiol Scand 1998;42:9951003.
20. Fauss BG, Meadows JC, Bruni CY, Qureshi GD. The in vitro and
in vivo effects of isoflurane and nitrous oxide on platelet aggregation. Anesth Analg 1986;65:1170 4.
21. de Rossi LW, Horn NA, Hecker KE, et al. Effect of halothane and
isoflurane on binding of ADP- and TRAP-6-activated platelets
to leukocytes in whole blood. Anesthesiology 2002;96:11724.
22. Horn NA, de Rossi L, Robitzsch T, et al. The effects of sevoflurane and desflurane in vitro platelet-leukocyte adhesion in
whole blood. Anaesthesia 2003;58:3129.
23. Frohlich D, Rothe G, Schmitz G, Hansen E. Volatile anaesthetics
induced in the expression of P-selectin and glycoprotein 1b on
the surface of platelets in vitro. Eur J Anaesthesiol 1998;15:641 8.
24. Pearson PJ, Schaff HV, Vanhoutte PM. Acute impairment of
endothelium-dependent relaxations to aggregating platelets following reperfusion injury in canine coronary arteries. Circ Res
1990;67:38593.
25. Kiechenbaum LA, Aziz M, Astiz ME, et al. Influence of rheologic changes and platelet-neutrophil interactions on cell filtration in sepsis. Am J Resp Crit Care Med 2000;161:16027.
26. Koblin DD, Eger EI II, Johnson BH, et al. Minimum alveolar
concentrations and oil/gas partition coefficients of four anesthetic isomers. Anesthesiology 1981;54:314 7.

ANESTH ANALG
2005;101:12618

27. Lerman J, Willis MM, Gregory GA, Eger EI II. Osmolarity determines the solubility of anesthetics in aqueous solutions at
37C. Anesthesiology 1983;59:554 8.
28. Zar JH. Biostatistical analysis. Englewood Cliffs, NJ: PrenticeHall, 1974.
29. Merhi Y, Provost P, Chauvet P, et al. Selectin blockade reduces
neutrophil interaction with platelets at the site of deep arterial
injury by angioplasty in pigs. Arteriocler Throm Vasc Biol 1999;
19:3727.
30. Gawaz M. Role of platelets in coronary thrombosis and reperfusion in ischemic myocardium. Cardiovasc Res 2004;61:498 511.
31. Seno T, Inoue N, Gao D, et al. Involvement of NADH/NADPH
oxidase in human platelet ROS production. Thrombosis Res
2001;103:399 409.
32. Zhao ZQ, Sato H, Williams MW, et al. Adenosine A2-receptor
activation inhibits neutrophil-mediated injury to coronary endothelium. Am J Physiol 1996;27:H1456 64.
33. Fan H, Sun B, Gu Q, et al. Oxygen radicals trigger activation of
NF- B and AP-1 and upregulation of ICAM-1 in reperfused
canine heart. Am J Physiol 2002;282:H1778 86.
34. Montrucchio G, Alloatti G, Camussi G. Role of plateletactivating factor in cardiovascular pathophysiology. Physiol
Rev 2000;80:1669 99.
35. Dusting GJ, Angus JA. Interaction of epoprostenol (PGI2) with
vasoconstrictors on diameter of large coronary arteries of the
dog. J Cardiovasc Pharmacol 1984;6:20 7.

Myocardial Systolic Function Increases During Positive

Pressure Lung Inflation
Michael F. Haney, Goran Johansson, Soren Haggmark, Bjorn Biber,
Michael F. Haney, MD, PhD, Goran Johansson, MS, Soren Haggmark, MS, and
Bjorn Biber, MD, PhD
Anesthesiology and Intensive Care Medicine, Ume University Hospital, Ume, Sweden

Lung inflation with positive airway pressure may have

rapid and dynamic effects on myocardial contractile
function. We designed this study to assess the magnitude and time to onset of myocardial function changes
during the initiation of single positive pressure lung inflation at clinically relevant inflation pressures. In 8
anesthetized 40-kg pigs, left ventricular pressures and
volumes were measured directly (conductance volumetry). A 15 cm H2O airway pressure plateau with lung
inflation (PPLI-15) was performed, and 2 single beats
from that sequence, one from resting apnea at zero airway pressure and the second from the point when the

ositive airway pressure and lung inflation can

have distinct effects on heart loading conditions
and performance that have been long recognized, although other direct and indirect effects on the
heart have been suspected (1). In patients exposed to
relatively low levels of positive airway pressure over
longer periods, increased cardiac performance and
sympathetic nervous system activation have been observed (2). Lung inflation with positive airway pressure may have dynamic effects on myocardial contractile status that have rapid onset, even during a single
breath. Our hypothesis was that increases in myocardial contractile status occur very rapidly in response
to positive airway pressure and lung inflation, and
that this could be detected at the earliest point of full
lung inflation during tidal breathing. The specific aim
was to measure myocardial contractile function
changes during onset and delivery of a single positive
pressure lung inflation at clinically relevant inflation
Supported, in part, by grants from The Swedish Medical Research
Council, The Swedish Heart-Lung Fund, and Ume University,
Sweden.
Accepted for publication June 21, 2005.
Address correspondence to Michael Haney, Anesthesiology and
Intensive Care Medicine, Ume University Hospital, 901 85 Ume,
Sweden. Address e-mail to michael.haney@anestesi.umu.se.
DOI: 10.1213/01.ANE.0000181330.57600.FF
2005 by the International Anesthesia Research Society
0003-2999/05

lungs were first maximally inflated, were selected for

analysis. Systolic function variables for zero airway
pressure and PPLI-15 were analyzed. Systolic elastance, derived from bilinear time-varying elastance
curves, increased approximately 15% during PPLI-15
from zero airway pressure. This agreed with other systolic function variables that identified an increase in left
ventricular contractile function for the lung inflation
beat. Serial measurements of myocardial function
should be conducted with constant airway pressure
and lung inflation conditions.
(Anesth Analg 2005;101:1269 74)

pressures in an in vivo large animal model with an

intact thorax.

Methods
With approval of the Ethical Research Committee of
Ume University, and in conformity with the Guide
for the Care and Use of Laboratory Animals (National
Academy of Sciences, 1996, USA), 8 Swedish Landrace
pigs (mean weight 39.9 kg) were used in this study.
They were first premedicated using IM azaperone
2 mg/kg and ketamine 12 mg/kg, and then anesthetized using IV pentobarbital 15 mg/kg and a pentobarbital infusion 1520 mg kg1 h1. No muscle relaxants were used. After a tracheotomy was
performed, animals were ventilated with tidal volumes not exceeding 10 mg/kg (Servo 900B; SiemensElema, Stockholm, Sweden) to achieve normoxia and
normocapnea (Oscar, Datex, Helsinki, Finland). Intravenous fluids were administered with the goal of
maintaining normovolemia: Ringers acetate solution
25
mg/kg
during
the
first
hour,
then
10 mL kg1 h1 throughout the study period. Body
temperature was measured and maintained between
38C and 39C using warmed IV fluids and a warming
blanket.
Anesth Analg 2005;101:126974

1269

1270

CARDIOVASCULAR ANESTHESIA HANEY ET AL.

MYOCARDIAL FUNCTION DURING INSPIRATION

A multilumen central venous catheter (Arrow International, Reading, PA) and a pulmonary artery catheter (Optimetrix; Abbott, Abbott Park, IL) were placed
via the external jugular vein system. An arterial catheter was placed with the tip in the descending aorta. A
7.5F balloon occlusion catheter (Vascular Technologies, Solna, Sweden) was placed into the inferior vena
cava (IVC). Arterial pressure, central venous pressure
(CVP), and pulmonary artery pressure were measured
using a fluid-filled catheter system and transducers
(Gabarith PMSET; Becton Dickinson, Franklin Lakes,
NJ). A 7F left ventricular (LV) pigtail combination tip
manometer and conductance catheter (CA-71083-PN;
CD Leycom, Zoetermeer, Holland) were placed
through an 8.5F introducer in the carotid artery system into the LV using fluoroscopic guidance. Best
conductance voltage signal and a stable catheter position at fluoroscopic examination were used to determine optimal conductance catheter position.
The conductance technique has been described in
depth previously (3). LV volume was measured using
the 12-electrode dual-field conductance catheter with
8 mm spacing between electrodes, and a signal
conditioning-amplifier (Leycom Sigma 5DF; Cardiodynamics, Zoetermeer, Holland). Volume calibration included establishment of a flow reference ratio to the
average of 3 thermodilution cardiac output measurements. Cardiac output was measured by thermodilution
using the thermistor-tipped pulmonary artery catheter.
Parallel conductance volume, by the hypertonic saline
injection method (4), and blood conductivity were also
measured. LV pressure and conductance data were recorded with a frequency of 250 Hz (PC Conduct; Cardiodynamics, Zoetermeer, Holland). All circulatory data
were recorded and analyzed using a digital signal acquisition and analysis software package (Acqknowledge;
Biopac Systems, Santa Barbara, CA).
The measuring sequences for the airway pressure
increase were recorded beginning at zero airway pressure (ZEEP) with the endotracheal tube disconnected.
The endotracheal tube was then connected to a
pressure-cycled ventilator (Evita 4; Drager, Lubeck,
Germany) with 15 cm H2O airway pressure plateau
for 8 s (PPLI-15), where lung inflation occurred. The
airway pressure was then released, and the measurement period ended. The animal was reconnected to
the volume-cycled ventilator set to maintain normocapnea. In a separate measurement, a preload reduction sequence was recorded during ZEEP, brought
about by a transient occlusion of the IVC by the
balloon-tipped catheter.
From each airway pressure increase sequence, 2
single heart cycles were identified for analysis: one
from the rest period at ZEEP before the positive airway pressure intervention and then a second heart
cycle during PPLI-15. The heart cycle that was analyzed for PPLI-15 was systematically selected as the

ANESTH ANALG
2005;101:1269 74

first heart beat, after the onset of PPLI-15, where CVP

had increased and reached a stable level (Fig. 1). The
time from onset of positive airway pressure 15 cm
H2O until the start of the heart cycle measured during
PPLI-15 was noted. Each heart cycle for ZEEP and
PPLI-15 was analyzed for load and systolic function
variables. End-diastole was determined either as the
point 12 or 16 ms after the R wave onset, or based on
milliseconds preceding the dP/dtmax 500 mm
Hg s1. End-systole was determined for the single
beats as maximal elastance (pressure/volume), or the
point on the pressure-volume loop that was tangent to
a line originating from the zero x-intercept, with slope
reported as Emax (5). All heart cycle times were measured and reported as heart rate (HR). Stroke work
(SW) was calculated as the integral of LV pressure and
volume during a single heart cycle. Time-varying elastance during each beat was analyzed for the ratio ()
of elastance slope during isovolumic contraction and
ejection phases (6). End-systolic elastance (Ees) was
determined as the linear regression slope for the endsystolic pressure-volume relation (ESPVR) over contiguous beats during the preload reduction sequence
(7). For each animal, an Ees slope was measured for
both the preload reduction sequence by transient IVC
balloon occlusion at ZEEP and then separately from
the preload reduction sequence that was induced by
the PPLI-15 (8).
General circulatory variables for the 2 conditions
(ZEEP and PPLI-15) were grouped for all 8 animals.
Statistical inferences for differences within the paired
measurements (ZEEP and PPLI-15) for all systolic
function variables were made using a paired Students
t-test. Significant differences were determined if the P
value was 0.05.

Results
Differences in general circulatory and systolic function
variables for paired beats at ZEEP and the PPLI-15 are
shown in Table 1. The mean time from institution of
positive airway pressure 15 cm H2O to the beat that
was analyzed for PPLI-15 was 1.5 0.1 s (sem). CVP
increased a small amount during PPLI-15, as demonstrated by a representative example in Figure 1. LV
end-systolic pressure increased slightly from ZEEP to
PPLI-15 whereas LV end-systolic volume decreased in
all animals. In other words, the LV ejected to a lower
end-systolic volume during PPLI-15, with no corresponding decrease in LV end-systolic pressure
(LVESP). There were significant increases in the
PPLI-15 measurement for both load-dependent systolic performance variables ejection fraction and SW,
as well as load-indexed systolic function variables
(SW/end-diastolic volume [EDV], Emax, and ) as
shown in Table 1. An example of representative paired

ANESTH ANALG
2005;101:1269 74

CARDIOVASCULAR ANESTHESIA
HANEY ET AL.
MYOCARDIAL FUNCTION DURING INSPIRATION

1271

Figure 1. This figure demonstrates a single

representative measurement sequence with
the single heart beats identified for zero expiratory airway pressure (ZEEP) (a), pressure
plateau with lung inflation (PPLI-15) onset
(b), and the single beat selected for measurement during PPLI-15 (c). Left ventricular pressure (LVP) and left ventricular volume (LVV)
are plotted versus time and superimposed on
each other in the upper left panel to demonstrate the temporal relation between the ZEEP
and the PPLI-15 measurement points. The simultaneous central venous pressure (CVP)
plot is shown in the lower left panel. A
pressure-volume diagram of this same sequence (right panel) illustrates the progression of pressure and volume changes related
to the institution of PPLI-15 (same events for
a, b, and c). A separate preload reduction
sequence by transient vena cava balloon occlusion is not shown here.

Table 1. Systolic Variables

Single beat parameters

HR (bpm)
CVP (mm Hg)
MPAP (mm Hg)
EDV (mL)
EDP (mm Hg)
ESV (mL)
ESP (mm Hg)
EF (%)
SW (mm Hg/mL)
SW/EDV (mm
Hg mL mL1)
Emax (mm Hg/mL)

Derived from preload

reduction sequence
Ees (mm Hg/mL)

ZEEP

PPLI-15

99 18
6.4 1.7
24 5.1
108 29
12.3 3.1
66.2 26
112 14
45 9
4253 789
42 13

98 17
8.8 1.7*
27 4.5
103 27*
15.5 4.2*
59.3 24*
114 14*
49.3 11*
4456 761*
46.1 14*

2.0 0.9
0.47 0.20

2.3 1.1*
0.54 0.23*

1.20 0.5

1.34 0.4

Values are mean sd. n 8.

ZEEP zero expiratory airway pressure; PPLI-15 15 cm H2O airway
pressure; HR heart rate; CVP central venous pressure; MPAP mean
pulmonary artery pressure; EDV end-diastolic volume; EDP enddiastolic pressure; ESV end-systolic volume; ESP end-systolic pressure;
EF ejection fraction; SW stroke work; Emax single beat end-systolic
pressure-volume relation; single beat systolic time varying elastance;
Ees end-systolic elastance.
* P 0.05, PPLI-15 value versus ZEEP using a paired Students t-test.

observations with pressure-volume loops with timevarying elastance derivation is shown in Figure 2,
along with an illustration of the derivation of the
systolic function variable in Figure 3. Ees slopes
during PPLI-15 were not significantly different from
the Ees measurements performed during ZEEP and
transient vena cava occlusion.

Discussion
These results demonstrate that increases in LV contractile function measurements occur very rapidly

during the course of single lung inflation in association with modest levels of positive pressure and inspiratory volume. This implies that myocardial function is dynamic during positive pressure breathing
within single tidal respiratory cycles. This is the first
study to observe a clear, rapid-onset myocardial function change in the pressure-volume plane with lung
inflation pressures and volumes in this relatively low
range.
When LV pressures increase during PPLI-15 (Table
1 and Fig. 1), this occurs despite a decrease in left
ventricular end-diastolic volume (LVEDV, preload). It
is not an increase in load, with unchanged myocardial
contractile function, which causes this transient
LVESP and systolic blood pressure increase. Instead,
an increase in systolic function, occurring at a lower
LVEDV (preload), produced this transient pressure
increase. These findings are in contrast to the results of
an earlier investigation (9) where a brief increase in
preload was observed as a result of positive pressure
lung inflation, which was suggested as the mechanism
for a transient increase in stroke volume and systolic
blood pressure, in the same setting. Differences in
study material, conditions, and measurement methods
for LV volumes are possible explanations for these
apparently diverging results. The conclusions of that
study: that increases in LVESP and systolic blood pressure occur because of increases in preload induced by
an inspiratory pump mechanism, probably should
not be considered a general or complete explanation
for the transient inspiratory systolic pressure increase
phenomena, in view of our results.
In the study design, there was an attempt to describe systolic function and loading from several aspects, including work (SW), the ESPVR, and time
varying elastance (). Also, a separate measurement
for Ees, which used a preload reduction over consecutive beats, was used as a corroborating variable. Another form of analysis of ESPVR (10) for changes in

1272

CARDIOVASCULAR ANESTHESIA HANEY ET AL.

MYOCARDIAL FUNCTION DURING INSPIRATION

ANESTH ANALG
2005;101:1269 74

Figure 2. This figure shows a representative example of paired selected beats for
zero expiratory airway pressure (ZEEP)
and pressure plateau with lung inflation
(PPLI-15) measurement with timevarying elastance (pressure/volume). ED
end-diastole; EJ start ejection; ES
end-systole.

Figure 3. This figure demonstrates the

derivation of the systolic function variable
from a single representative heart cycle.
The heart cycle-phase specific points are
identified. ED end-diastole; EJ start
ejection; ES end-systole. The elastance
slope for isovolemic contraction and ejection is used to calculate the ratio (6).

contractile status is to compare 2 beats from the same

ventricle, which occur in the same end-systolic pressure range, and identify the beat where the ventricle
ejects to a lower end-systolic volume as the beat with
increased myocardial contractile status. This is more
specific than Emax, given that the systolic events are
compared at the same end-systolic load (pressure), or
afterload. These ESPVR findings further support the
finding of increased myocardial contractile function
during PPLI-15.
In earlier clinical studies, increased ventricular performance, but not relation to ventricular load, has
been described in a model of increased airway and
intrathoracic pressures and experimental myocardial
depression (11). Changes in ventricular volumes in
response to positive airway pressure have also been
described, as mentioned above (8). To assess myocardial function in response to these conditions, however,

the relation of performance to load (and other conditions) must be measured accurately (12). Unless performance is indexed for the prevailing conditions, including load, then observations of change in
performance do not provide direct information about
the myocardial contractile status.
The time course of this response allows speculation
that some form of autonomic nerve system activity is
involved, although this study was not designed to
identify a neurocirculatory mechanism. Increases in
peripheral sympathetic nerve activity after initiation
of positive pressure lung inflation in humans have
been observed, although no myocardial function assessment has been performed during the immediate
ensuing period after lung inflation (13,14). One proposed mechanism is through stimulation of mechanical receptors in the lung and airway, which leads to
activation of neurocirculatory reflexes (15). Regional

ANESTH ANALG
2005;101:1269 74

blockade of lung afferent mechanoreceptor input during inflation has been shown to block increases in HR
and arterial blood pressure in anesthetized patients,
although more direct indices of myocardial function
have not been measured (16).
Earlier physiological studies in animals have demonstrated that positive airway pressure and lung inflation over seconds or minutes appear to be able to
initiate neurocirculatory reflexes that are modulated
through lung mechanoreceptors and vagal efferent
nerve signals that have been associated with bradycardia and even decreased ventricular contractile
function (1719). Other animal studies have suggested
that lower positive airway inflation pressures can produce an early increase in HR with no clear effects on
force of LV contraction (20) or a positive inotropic
response (21). This body of evidence supports the
possibility that a pulmonary afferent reflex and a
neurocirculatory effector working in the LV could be
involved, although further investigation is needed in
this area.
Extracardiac pressure presumably increases to some
extent during positive airway pressure lung inflation.
The effect of this on transmural LV pressures can be
measured in an invasive pericardial preparation, and
there is evidence to suggest that even moderate or
large continuous increases in extracardiac pressures
do not alter the relationship of ventricular performance to loading conditions as quantified by, for example, preload recruitable SW (22). When large extracardiac pressures are present, transmural ventricular
pressures need to be considered. Extracardiac pressure changes can be estimated, however, based on
observations in changes in right atrial pressures (23)
and are not large during PPLI-15. Systolic function
results, including ESPVR and , are not affected by a
small and constant difference in extracardiac pressure,
making this very unlikely as a confounding factor.
The conductance volumetry method has been criticized regarding accuracy for absolute volumes, although the method is considered very reliable for
precision or the identification of relative changes in
volumes and direction of changes. There has been
concern that lung inflation and positive pressure airway events can lead to a change in tissue volume
signal conductivity surrounding the ventricle, potentially producing errors in measured volumes. There is
evidence to suggest that no change occurs in parallel
conductance during positive airway pressure lung inflation (24).
Also, with respect to parallel conductance, there is a
theoretical concern that changes in right ventricular
configuration could potentially cause a change in parallel conductance and ventricular volume signal. If a
blood volume increase occurred in the right ventricle
during positive pressure lung inflation and this was
sensed by the conductance catheter as volume signal,

CARDIOVASCULAR ANESTHESIA
HANEY ET AL.
MYOCARDIAL FUNCTION DURING INSPIRATION

1273

then parallel conductance (and total conductancemeasured volume for the LV, but not true LV volume)
would increase. The opposite was observed, however,
with ventricular volume decreases during positive airway pressure. LV volume changes of this sort during
positive pressure ventilation have been previously described (25).
The clinical significance of these modest variations
in myocardial contractile function during the respiratory cycle may be recognition that they exist. When
performing serial measurements of contractile status,
the conditions must be standardized to the extent
possible, and this includes lung inflation as a condition that can have effects on measured myocardial
function. Therefore, analysis of myocardial contractile
function should include attempts to standardize respiratory and lung inflation events during serial measurements. It is also possible that the myocardial response in an individual subject to lung inflation can
change over time. The most certain method of eliminating potential myocardial effects of lung inflation
from serial measures is to perform measurements at
apnea with no positive airway pressure. Myocardial
function measurements during respiratory activity
can also be of interest, and longer respiratory pauses
are not always practical for individual subjects. The
data in this study indicate that serial measures of
myocardial function need to have a consistent relation
to the respiratory cycle.
In summary, a very rapid onset for increases in
myocardial contractile function during a single positive airway pressure lung inflation was observed in
this large animal model. The magnitude of the airway
pressure and lung inflation was typical of that used
for clinical support of ventilator-dependent subjects
with healthy lungs. We conclude that the rapid variation in myocardial function in relation to positive
pressure ventilation must be considered when measuring myocardial function, so that serial measures
can be collected within the same context of cardiopulmonary interactions.

References
1. Van den Berg PCM, Jansen JRC, Pinsky MR. Effect of positive
pressure on venous return in volume loaded cardiac surgical
patients. J Appl Physiol 2002;92:122331.
2. Ziegler MG, Mills PJ, Loredo JS, et al. Effect of continuous
positive airway pressure and placebo treatment on sympathetic
nervous activity in patients with obstructive sleep apnea. Chest
2001;120:88793.
3. Steendijk P, Van der Velde ET, Baan J. Left ventricular stroke
volume by single and double excitation of conductance catheter
in dogs. Am J Physiol 1993;264:H2198 207.
4. Steendijk P, Staal E, Jukema JW, Baan J. Hypertonic saline
method accurately determines parallel conductance for dualfield conductance catheter. Am J Physiol Heart Circ Physiol
2001;281:H755 63.

1274

CARDIOVASCULAR ANESTHESIA HANEY ET AL.

MYOCARDIAL FUNCTION DURING INSPIRATION

5. Suga H, Sagawa K, Shoukas AA. Load independence of the

instantaneous pressure-volume ratio of the canine left ventricle
and effects of epinephrine and heart rate on the ratio. Circ Res
1973;32:314.
6. Shoshido T. Hayashi K, Shigemi K, et al. Single-beat estimation
of end-systolic elastance using bilinearly approximated timevarying elastance curves. Circulation 2000;102:19839.
7. Kass DA, Maughan WL, Guo ZM, et al. Comparative influence
of load versus inotropic states on indexes of ventricular
contractility: experimental and theoretical analysis based on
pressure-volume relationships. Circulation 1987;76:142236.
8. Haney MF, Johansson G, Haggmark S, Biber B. Method of
preload reduction during LVPVR analysis of systolic function:
airway pressure elevation and vena cava occlusion. Anesthesiology 2002;97:436 46.
9. Vieillard-Baron A, Chergui K, Augarde R, et al. Cyclic changes
in arterial pulse during respiratory support revisited by Doppler echocardiography. Am J Respir Crit Care Med 2003;168:
671 6.
10. Baan J, Van der Velde ET, Steendijk P. Ventricular pressurevolume relations in vivo. Eur Heart J 1992;13(SupplE):2 6.
11. Pinsky MR, Summer WR, Wise RA, et al. Augmentation of
cardiac function by elevation of intrathoracic pressure. J Appl
Physiol 1983;54:950 5.
12. Kass DA, Maughan WL. From Emax to pressure-volume
relations: a broader view. Circulation 1988;77:120312.
13. Macefield VG, Wallin BG. Modulation of muscle sympathetic
activity during spontaneous and artificial ventilation and
apnoea in humans. J Auton Nerv Syst 1995;53:137 47.
14. Sellgren J, Ponten J, Wallin BG. Characteristics of muscle nerve
sympathetic activity during general anesthesia in humans. Acta
Anaesthesiol Scand 1992;36:336 45.
15. Hamaya Y, Dohi S, Takenaka-Hamaya C. Localized airway
anesthesia with lidocaine partially suppressed cardiovascular
responses to lung inflation. Anesth Analg 2000;90:84751.

ANESTH ANALG
2005;101:1269 74

16. Hamaya Y, Dohi S. Differences in cardiovascular response to

airway stimulation at different sites and blockade of the responses by lidocaine. Anesthesiology 2000;93:95103.
17. Kaufman MP, Iwamoto GA, Ashton JH, Cassidy SS. Responses
to inflation of vagal afferents with endings in the lungs of dogs.
Circ Res 1982;51:52531.
18. Cassidy SS. Stimulus-response curves of the lung inflation
cardio-depressor reflex. Resp Physiol 1984;57:259 68.
19. Ashton JH, Cassidy SS. Reflex depression of cardiovascular
function during lung inflation. J Appl Physiol 1985;58:137 45.
20. Hainsworth R. Circulatory responses from lung inflation in
anesthetized dogs. Am J Physiol 1974;226:24755.
21. Greenwood PV, Hainsworth R, Karim F, et al. Reflex inotropic
responses of the heart from lung inflation in anaesthetized dogs.
Pflugers Arch 1980;386:199 205.
22. Haney MF, Johansson G, Haggmark S, Biber B. Analysis of left
ventricular systolic function during elevated external cardiac
pressures: an examination of measured transmural left ventricular pressure during pressure-volume analysis. Acta Anaesthesiol Scand 2001;45:868 74.
23. Smiseth OA, Thompson CR, Ling H, et al. A potential clinical
method for calculating transmural left ventricular filling pressure during positive end-expiratory pressure ventilation: an
intraoperative study in humans. J Am Coll Cardiol 1996;27:
155 60.
24. Szwarc RS, Ball HA. Simultaneous LV and RV volumes by
conductance catheter: effects of lung insufflation on parallel
conductance. Am J Physiol 1998;275(2 Pt 2):H653 61.
25. Denault AY, Gorcsan J 3rd, Pinsky MR. Dynamic effects of
positive-pressure ventilation on canine left ventricular pressurevolume relations. J Appl Physiol 2001;91:298 308.

REVIEW ARTICLE

Reactive Oxygen Species as Mediators of Cardiac Injury and

Protection: The Relevance to Anesthesia Practice
Leo G. Kevin,

MD, FCARCSI,

Enis Novalija,

MD,

and David F. Stowe,

MD, PhD

Anesthesiology Research Laboratories, Departments of Anesthesiology and Physiology, Cardiovascular Research Center,
The Medical College of Wisconsin, VA Medical Center Research Service, and Department of Biomedical Engineering,
Marquette University, Milwaukee, Wisconsin

Reactive oxygen species (ROS) are central to cardiac

ischemic and reperfusion injury. They contribute to
myocardial stunning, infarction and apoptosis, and
possibly to the genesis of arrhythmias. Multiple
laboratory studies and clinical trials have evaluated
the use of scavengers of ROS to protect the heart from
the effects of ischemia and reperfusion. Generally,
studies in animal models have shown such effects.
Clinical trials have also shown protective effects of
scavengers, but whether this protection confers
meaningful clinical benefits is uncertain. Several IV
anesthetic drugs act as ROS scavengers. In contrast,
volatile anesthetics have recently been demonstrated
to generate ROS in the heart, most likely because of
inhibitory effects on cardiac mitochondria. ROS are

ree radicals and their nonradical reactants are

recognized as critical mediators of cardiac injury
during ischemia and reperfusion. They have been
implicated in reversible postischemic contractile dysfunction (myocardial stunning), cardiac cell death,
dysrhythmias, and in chronic cardiovascular diseases.
Intensive laboratory and clinical investigative effort
has focused on the capability of antioxidant therapies
to attenuate deleterious effects of free radicals in the
heart. Many investigators have also sought to demonstrate that anesthetic drugs can either prevent free
radical formation or scavenge free radicals.
The deleterious effects of free radicals have been
known for almost 50 yr. More recently, however, an
essential role for free radicals in physiologic control of
several aspects of cell function has been demonstrated.
Free radicals are, indeed, now considered as key regulatory molecules vital for life, but they cause cellular
Accepted for publication May 23, 2005.
Address correspondence and reprint requests to David Stone,
MD, PhD, Anesthesiology, 8701 Watertown Plank Road, Medical
College of Wisconsin, Milwaukee, WI 53226. Address e-mail to
dfstone@mcw.edu.
DOI: 10.1213/01.ANE.0000180999.81013.D0
2005 by the International Anesthesia Research Society
0003-2999/05

involved in the signaling cascade for cardioprotection induced by brief exposure to a volatile anesthetic
(termed anesthetic preconditioning). ROS, therefore, although injurious in large quantities, can have
a paradoxical protective effect within the heart. In
this review we provide background information on
ROS formation and elimination relevant to anesthetic
and adjuvant drugs with particular reference to the
heart. The sources of ROS, the means by which they
induce cardiac injury or activate protective signaling
pathways, the results of clinical studies evaluating
ROS scavengers, and the effects of anesthetic drugs
on ROS are each discussed.
(Anesth Analg 2005;101:127587)

and organ damage when produced in excess or when

innate antioxidant defenses are overwhelmed.
Anesthetic preconditioning (APC) is of particular
current interest to anesthesiologists. In this phenomenon, lasting protection of myocardium is elicited by
brief exposure to a volatile anesthetic. Free radicals are
now known to act as second messengers in the preconditioning cell-signaling pathway. Most remarkably, volatile anesthetic drugs have recently been
shown to enhance generation of free radicals in cardiac cells, probably by causing mild uncoupling of the
mitochondrial electron transport chain. A cardioprotective signaling pathway that is triggered by volatile
anesthetics and that utilizes free radicals as essential
mediators thus represents a paradigm shift in traditional anesthesiology teaching. First, effects of volatile
anesthetic drugs in the heart are not evanescent but
may be long lasting and, second, moderate release of
free radicals may induce a protective rather than an
injurious effect within the myocardium.
This review is intended to provide background information on free radical formation and elimination
relevant to anesthetic and adjuvant drugs with particular reference to the heart. The sources of free radicals,
the means by which they induce cardiac injury or
Anesth Analg 2005;101:127587

1275

1276

REVIEW ARTICLE KEVIN ET AL.

REACTIVE OXYGEN SPECIES AND THE HEART: RELEVANCE TO ANESTHESIA

ANESTH ANALG
2005;101:127587

activate cardiac signaling pathways, the results of trials evaluating free radical scavengers, and the effects
of anesthetic drugs on free radicals are each discussed.

Free Radicals and Reactive Oxygen

Species
A free radical is any species capable of independent
existence that contains one or more unpaired electrons
occupying an atomic orbital by itself. Radicals are
therefore formed by the loss or gain of an electron
from a nonradical. The presence of an unpaired electron usually makes the species highly reactive, although different free radical species differ greatly in
their reactivity. Redox or oxidation-reduction reactions are those reactions that involve exchange of electrons between molecular species.
The diatomic ground state oxygen molecule has
two unpaired electrons, each in a different orbital, and
indeed fulfills the definition for a free radical. If this
molecule accepts an electron the product is superoxide
radical (O2). Many of the important free radicals are
derived from oxygen. Another important radical is
nitric oxide (NO). Many of these have important intermediates, such as hydrogen peroxide (H2O2) and
peroxynitrite (ONOO), which are not free radicals
but which are highly reactive and may be responsible
for some of the biologic effects attributed to free radicals. Therefore, the term reactive oxygen species
(ROS) is used to encompass these species. Nonradical
ROS also include ozone and singlet oxygen.

Sources of ROS
Sources of free radicals are the mitochondrial electron
transport chain, the enzymes xanthine oxidase,
NADPH oxidase, lipoxygenase/cyclooxygenase and
nitric oxide synthase (NOS), and auto-oxidation of
various substances, particularly catecholamines (Fig.
1).
The tetravalent reduction of molecular O2 by the
mitochondrial electron transport chain is necessary for
generating most biologic energy. To accomplish this
there are four mitochondrial complexes (IIV) involved in energy conservation. The reduction is not
100% efficient, however, and 1% 4% of available oxygen is normally incompletely reduced and leaks from
the electron transport chain in the form of O2.
O2 is produced at complex I (NADH coenzyyme Q
reductase, also called ubiquinone oxidoreductase) and
complex III (ubiquinol cytochrome C reductase). This
process becomes greatly accelerated at supranormal
O2 tensions or after mitochondrial injury and is believed to be the primary intracardiac source of ROS

Figure 1. Formation of reactive oxygen species in cardiac myocytes.

An electron is gained by molecular O2 yielding superoxide (O2). In
myocytes, this reaction occurs primarily at complexes I and III of the
mitochondrial electron transport chain. The electrons are derived
from NADH and FADH2. O2 is converted to hydrogen peroxide
(H2O2) in the presence of superoxide dismutase, found in the mitochondrion and in the cytosol. H2O2, a non-radical, but a freely
permeable and highly reactive reactive oxygen species (ROS), is
converted to hydroxyl radical (OH) in the presence of copper or
iron (the Fenton reaction). Alternatively H2O2 reacts with O2 to
form OH in the Haber-Weiss reaction. Glutathione peroxidase catalyzes the conversion of H2O2 to nonradical water and oxygen. ETC
electron transport chain.

during ischemia and reperfusion. Cellular hypoxia decreases the activity of complex IV (cytochrome oxidase); when O2 is reintroduced, leakage of free radicals from more proximal complexes is greatly
accelerated. Although it was previously believed that
ROS formation occurred primarily or solely at reoxygenation after ischemia, it is now known that significant formation of ROS occurs during ischemia from
residual O2. This has been demonstrated in cardiomyocytes (1) and in the intact heart (2,3).
Most O2 is dismutated by manganese-superoxide
dismutase (MnSOD) in the mitochondrial matrix to
H2O2, which readily diffuses through mitochondrial
membranes. The remainder exits the mitochondria
through anion channels in the mitochondrial membrane and is then rapidly converted to H2O2 in the
cytoplasm, either spontaneously, or when catalyzed
by copper superoxide dismutase (CuSOD). H2O2 is
reduced to H2O and O2 by catalase and glutathione
peroxidase. Alternatively, H2O2 reacts with transition
metals, particularly Fe2 (the Fenton reaction), to generate hydroxyl radical (OH). These reactions are
shown in Figure 2. OH is a particularly reactive radical, such that it will react with virtually all biomolecules at diffusion limited rates. This has important
consequences, as OH will react at the site of its formation whereas more stable species, such as O2 and
H2O2, can react at more distant locations.
Xanthine oxidase is a major source of ROS after
reperfusion of ischemic tissue in several organs, although its role in the heart remains somewhat unclear.

ANESTH ANALG
2005;101:127587

REVIEW ARTICLE
KEVIN ET AL.
REACTIVE OXYGEN SPECIES AND THE HEART: RELEVANCE TO ANESTHESIA

1277

NOS normally generates the free radical NO, which

is required for a variety of physiologic processes reviewed elsewhere (8). Increasing evidence points to
the important role of the interaction between NO and
O2 to form ONOO. This reaction may act as an
endogenous scavenging system, as ONOO is considerably less reactive than OH, the formation of which is
necessarily decreased as O2 is removed. In addition,
there appear to be multiple physiologic signaling
pathways that require ONOO, including a cardioprotective pathway (9). Nonetheless, excess ONOO
can also damage cellular components (10). Furthermore, electron transfer by NOS can become uncoupled
under certain conditions, including hyperglycemia,
directly causing the generation of O2 by NOS (11).

Endogenous Antioxidant Systems

Figure 2. Reactive oxygen species (ROS) are derived from multiple

sources and are counterbalanced by enzymatic and non-enzymatic
antioxidants. Under pathologic conditions, such as ischemiareperfusion, antioxidant defenses are overwhelmed by excess ROS
and cellular injury results.

During ischemia, purine precursors are degraded to

the nucleotide derivatives hypoxanthine and xanthine
and the enzyme xanthine dehydrogenase is converted
to xanthine oxidase. Xanthine oxidase catalyzes oxidation of hypoxanthine and xanthine to uric acid while
reducing O2 to O2 and H2O2. The main source of
xanthine oxidase in human hearts remains unknown.
Endothelial cells are more likely candidates than cardiomyocytes, which appear to contain little of the
enzyme (4).
Phagocytic cells generate O2 from O2 when these
cells are activated. This is the so-called respiratory
burst, which is catalyzed by the enzyme NADPH
oxidase. ROS generated from this source play a central
role in host defense. Cardiopulmonary bypass (CPB)
is associated with significant neutrophil activation as a
result of ischemia-reperfusion injury but is also a result of a direct effect of contact of neutrophils with the
extracorporeal circuit. Neutrophils therefore provide a
major source of free radicals during and after CPB (5);
indeed, free radicals from extracardiac sources during
CPB are reported to greatly exceed those from the
heart (6). Enzymes related to NADPH oxidase have
been identified in nonphagocytic cells, including vascular smooth muscle cells (7).

There are endogenous antioxidant systems that counteract the potential for injury to cellular structures by
regulating the balance of individual ROS and their
reactants. These endogenous antioxidants are upregulated when exposure of the cell to ROS is increased;
thus the term redox homeostasis has been coined.
However, under pathologic conditions such as ischemia-reperfusion, ROS formation can rapidly overcome
antioxidant defenses and cellular injury ensues.
Major endogenous antioxidants in cardiomyocytes
include superoxide dismutase (SOD), catalase, glutathione peroxidase, glutathione, Coenzyme Q10
(ubiquinone), and vitamins C and E (Fig. 1). SOD
exists in 3 isoforms: MnSOD is within mitochondria,
CuZnSOD is in the cytoplasm, and extracellular SOD
is located extracellularly associated with glycosaminoglycans. During CPB endogenous antioxidant systems
become activated to high levels in response to increased free radicals; such is the ROS production,
however, that these rapidly become depleted so that
CPB leads to oxidant damage (12).

Injurious Effects of ROS in the Heart

Free radicals exert their biologic effects by obtaining
an electron from any molecule that may yield it, including lipids, proteins, and DNA. Lipid peroxidation
causes particularly destructive effects on cell membranes. Modification of protein structure on attack by
free radicals or their reactive products can alter function or lead to enhanced susceptibility to proteolysis.
Proteins that are misfolded or partially unfolded are
most sensitive to oxidation; thus, oxidative denaturation of dysfunctional proteins is believed to serve a
physiologic housekeeping function. At the same time,
consumption of ROS in this process aids in cellular
antioxidation (i.e., scavenging).

1278

REVIEW ARTICLE KEVIN ET AL.

REACTIVE OXYGEN SPECIES AND THE HEART: RELEVANCE TO ANESTHESIA

ROS are strongly implicated in the pathogenesis

of postischemic myocardial stunning, necrosis, programmed cell death (apoptosis), and vascular dysfunction. They may also contribute to the pathogenesis of
postischemic dysrhythmias. Myocardial stunning refers to contractile dysfunction that is reversible, as
cells remain viable, but which lasts hours to days. It is
reported to occur after CPB (13) and also after offpump cardiac surgery, albeit to a lesser extent (14).
Clinically evident heart failure may occur as a consequence of stunning. Animal models provide evidence
of a role for ROS in the pathogenesis of stunning (15).
The OH radical is suggested to be particularly responsible and several target molecules have been proposed. Protection from stunning using scavengers is
incomplete, indicating an additional contributing
mechanism.
Cell death after ischemia-reperfusion results from
necrosis and apoptosis. ROS are involved in the apoptotic pathway. Indeed, they have been shown to directly induce apoptosis after ischemia-reperfusion injury (16). Severe lipid peroxidation by ROS causes
myocardial necrosis (17).
Dysrhythmias represent an important form of morbidity during reperfusion after ischemia and in patients undergoing cardiac surgery. ROS have been
shown to have direct electrophysiologic effects that
contribute to dysrhythmias (18). An important source
of ROS leading to dysrhythmias after acute myocardial infarction may be neutrophils, as activation of
neutrophils correlated with the incidence of dysrhythmias (19). Overall, however, some doubt remains over
the exact role of ROS in dysrhythmias after ischemiareperfusion injury in humans (20).
In settings other than ischemia-reperfusion, ROS
have been implicated in the pathogenesis of cardiovascular disease. Chronic exposure of myocardium to
ROS is proposed as a hypothesis for the development
and progression of congestive cardiac failure (CCF)
(21). In support of this, cardiac failure occurs predictably when exogenous ROS are applied to the previously uninjured heart (22). Moreover, excess ROS generation (23) and inadequate antioxidant defenses (24)
have been reported in animal models of CCF. The
mechanism of the eventual myocyte contractile failure
is likely to be multifactorial; for example, sarcolemmal
Ca2 entry (25) and adrenoreceptor sensitivity (26) are
each impaired under conditions of chronic oxidative
stress. In addition, evidence links oxidative damage of
cardiac mitochondrial DNA to postischemic left ventricular remodeling and heart failure (27); this suggests that impaired mitochondrial bioenergetics secondary to chronic redox imbalance may play a pivotal
role. Human studies have shown evidence of oxidative stress in patients with CCF (28), and a relationship
between exercise tolerance and oxidative stress in
such patients has been reported (29). Nonetheless,

ANESTH ANALG
2005;101:127587

whether these findings represent a true cause-effect

relationship or merely epiphenomena in the failing
heart remains to be established. Clinical trials evaluating antioxidant therapies in CCF have failed to demonstrate clear benefit.
Increasing evidence points to a pathogenic role for
ROS in atherosclerosis. Risk factors for atherosclerosis,
including cigarette smoking, diabetes mellitus, hypertension, and hypercholesterolemia, are associated
with increased ROS levels within the coronary vasculature. Endothelial apoptosis (30), vascular smooth
muscle cell proliferation (31), and activation of matrix
metalloproteinases (32) are among several factors that
are ROS-activated and that are central to different
stages of atherogenesis. Furthermore, as described
above, the interaction between NO and O2 to form
ONOO acts to deplete NO. The vasodilating and
platelet-inhibiting effects of NO are thereby attenuated favoring vasoconstriction and thrombogenesis.
Finally, restenosis after angioplasty has been linked to
an effect of ROS. Vascular O2 production increases
after balloon angioplasty (33) and scavengers of ROS
were shown to prevent neointimal formation after
angioplasty (34).

Physiologic Effects of ROS in the Heart

The first quarter-century of ROS research focused on
damage inflicted on cellular substituents as outlined
above. Then, in the mid-1970s, ROS were shown to
activate guanylate cyclase, causing formation of the
well characterized second messenger cyclic guanosine
monophosphate (35). Gradually, as other second messenger systems were shown to be modulated by ROS,
a biologic role for ROS in cell function became accepted. Indeed, it is now known that ROS act as important mediators in signal transduction processes involved in multiple aspects of cardiac cellular function.
For several of these processes, ROS are produced
after activation of cell-surface receptors. Included
among these receptors are cytokine and G proteincoupled receptors. ROS generated in response to cytokine receptor activation are implicated in the control
of apoptotic pathways. Conversely, ROS generated in
response to activation of G protein-coupled receptors
are involved in triggering cell proliferation and hypertrophy. These findings have lead to the suggestion
that ROS are central in biologic control of cell life/
death determination. In support of this, ROS generation after activation of the growth factor receptor
TGF-1 was shown to directly regulate phase G1 of the
cell cycle (36). More recently, differentiation of embryonic stem cells into cardiomyocyte cell lines was
shown to be triggered by ROS (37), suggesting that
they have an essential role in cardiac embryogenesis.

ANESTH ANALG
2005;101:127587

REVIEW ARTICLE
KEVIN ET AL.
REACTIVE OXYGEN SPECIES AND THE HEART: RELEVANCE TO ANESTHESIA

A free radical theory of development has therefore

been proposed in which a change in redox state is a
key embryologic signal.
Two major determinants of oxygen delivery have
lately been shown to be redox regulated. The carotid
bodies control ventilatory responses to changes in arterial oxygen. The oxygen-sensing apparatus of the
carotid bodies is now believed to rely on changes in
mitochondrial ROS production in response to altered
oxygen tension in glomus type I cells (38). In the liver
and kidney, the oxygen sensor that leads to production of the hormone erythropoietin, the regulator of
red cell mass, is now known to sense changes in local
ROS levels (39). Indeed, the transcription factor that
codes for the erythropoietin gene was shown to be
unaffected by changes in oxygen tension but was
dose-dependently regulated by ROS (40). Similarly, in
the coronary vasculature, there is evidence that vascular responsiveness to changes in oxygen tension is,
in fact, mediated by superoxide (41).
The target molecules for these and other physiologic
effects of ROS are a matter of dispute, for the broad
range of redox-sensitive substituents greatly complicates an accurate appraisal. Ca2 channels, mitogenactivated protein kinase enzymes and tyrosine phosphatases (which contain a ROS-sensitive cysteine
residue in their active site) have each been shown to be
activated by ROS. Upregulation of transcription factors, including nuclear factor-B and AP-1 (42), occurs
after application of H2O2. Consequent to transcription
factor activation is altered gene expression. Indeed,
DNA microarray studies show induction of nearly
100 genes in response to oxidant stress (43). In almost all instances the details of these ROS-dependent
processes require further elucidation. Figures 3 and 4
illustrate cell signaling pathways and physiologic
consequences in which ROS have a demonstrated
role.

Antioxidant Therapies
Administration of exogenous antioxidants has been
extensively investigated as a means to attenuate myocardial ischemia-reperfusion injury and to treat or prevent chronic cardiovascular diseases. Investigations in
a variety of animal models have shown beneficial
effects of several drugs. However, clinical trials have
furnished inconsistent results. The following paragraphs summarize experience with those antioxidant
therapies that have evoked the most interest. These
include drugs that scavenge formed ROS (superoxide
dismutase [SOD], deferoxamine, vitamins C and E,
acetaminophen) and drugs that prevent or alter the
formation of ROS (deferoxamine, allopurinol).

1279

Figure 3. Cell signaling pathways in which reactive oxygen species

(ROS) have a demonstrated role. Activation of cell surface receptors
leads to the generation of ROS which have multiple intracellular
effects including modulation of ion channels and activation of several enzyme systems. Activated transcription factors cause synthesis
of proteins with wide ranging cellular effects. Mitogen-activated
protein (MAP) kinase enzymes and tyrosine phosphatase, respectively, upregulate and downregulate trancription factors. Guanylate
cyclase serves multiple functions including modulation of cell surface ion channels. CaL L-type calcium channel.

Figure 4. Reactive oxygen species (ROS)-mediated pathways utilize

different effector mechanisms to exert biologically essential effects
that appear to be particularly involved in the cell life/death cycle
(apoptosis, cell proliferation) and adaptation to stress (cell hypertrophy, cytokine production, preconditioning).

Superoxide Dismutase
SOD decreases infarct size in animal models when
given during ischemia and reperfusion (44), or in animals given a gene transfer for SOD (45). Human trials
have been disappointing, however. Although one
small clinical trial involving 23 patients reported protective effects of SOD against premature ventricular
contractions after thrombolysis (46), a larger trial involving 120 patients undergoing angioplasty in the
setting of acute infarction found no such change, nor
was there any improvement in global function, i.e.,
SOD did not attenuate stunning (47). Conflicting results between animal studies and human trials may be

1280

REVIEW ARTICLE KEVIN ET AL.

REACTIVE OXYGEN SPECIES AND THE HEART: RELEVANCE TO ANESTHESIA

the result of species differences. Perhaps more importantly, the low cellular permeability of exogenously
administered SOD likely limits potential for this therapeutic approach. Much of the production of ROS that
is theorized to be deleterious is intracellular. SOD has
a molecular weight of 32 kd and an overall negative
charge. This largely limits exogenously administered
SOD to the extracellular space. Furthermore, exogenous SOD undergoes rapid renal clearance (48).

Deferoxamine
Deferoxamine is a chelator of iron. By forming a stable
complex with ferric iron it decreases the availability of
iron for production of ROS by the Fe2 reaction. There
is also evidence that deferoxamine directly scavenges
free radicals. Like exogenously administered SOD, its
ability to penetrate cell membranes is poor (49). Nonetheless, studies in a variety of animal models have
shown protective effects of deferoxamine against ischemia-reperfusion injury with respect to contractile
function (50,51) and dysrhythmias (52). Studies in humans have shown protective effects but trials were
underpowered to demonstrate any meaningful outcome differences. For example, Ferreira et al. (53) and
Drossos and Lazou (54) included deferoxamine in cardioplegic solution for adults undergoing coronary artery bypass grafting (CABG) and valve surgery; they
found that myocardial free radical production and
lipid peroxidation were decreased but could show no
differences in functional indices.

Mannitol
Mannitol, commonly added to pump prime for CPB,
acts as a free radical scavenger, and is reported to
decrease infarction after ischemia-reperfusion injury
in some animal models (19). One small trial, underpowered to show differences in functional outcome
variables, found that mannitol decreased plasma H2O2
in patients undergoing CPB (55). A further small clinical trial found no beneficial effects of mannitol added
to cardioplegia in patients undergoing CABG (56).
Once again, this may be related to the fact that mannitol remains in the extracellular compartment and
therefore fails to penetrate to the site of ROS generation where initial damage occurs.

Allopurinol
The enzyme xanthine oxidase is an important source
of ROS during reperfusion after ischemia in several
tissues, with the possible exception of cardiac muscle,
which contains little of the enzyme. Nonetheless, inhibition of xanthine oxidase activity by allopurinol
was reported to decrease myocardial infarct size in
dogs (57). One explanation for this discrepancy is that
inhibition of extracardiac xanthine oxidase prevents

ANESTH ANALG
2005;101:127587

release of ROS from tissues remote from the heart

and thus secondarily attenuates cardiac injury. In
support of this hypothesis, Nielsen et al. (58), in a
rabbit model, demonstrated that activation of this
enzyme in hepatoenteric tissue induced significant
myocardial injury.
A literature review revealed 10 clinical trials of allopurinol in patients undergoing CABG, making it one
of the most intensively investigated antioxidants in
such patients. In eight of these, including the largest
trial which involved 169 patients (59), improved functional indices and/or decreased CK-MB release were
reported. Nonetheless, allopurinol has yet to gain acceptance in the routine management of cardiac surgery patients, likely because benefits in trials have
been largely limited to less compelling end-points
such as reduced inotrope use after termination of CPB.
Demonstration of a decrease in hospital mortality, if
there is such an effect with allopurinol, would require
a considerably larger trial than has been performed to
date.

Acetaminophen
Acetaminophen is a phenol and, like many other phenols, it has antioxidant properties. The studies of Merrill et al. (60,61) demonstrated protective effects of
acetaminophen in animal models. Acetaminophen attenuated production of OH radical in the postischemic myocardium and improved postischemic recovery in guinea pig hearts (60). Acetaminophen also
protected against exogenous H2O2, suggesting that it
might act by inhibiting the Fe2 reaction (61). There
are no human trials of acetaminophen and human
myocardial injury.

Vitamins C and E
Vitamin C (ascorbic acid) and E (-tocopherol) are
water- and lipid-soluble endogenous antioxidants, respectively. Several investigations have evaluated their
effects on myocardial ischemia-reperfusion injury
when given exogenously. Both these vitamins attenuated myocardial injury in animal models (62). Clinical
trials have shown benefits of these vitamins in patients
undergoing CABG. Dingchao et al. (63) reported that
vitamin C led to decreased cardiac enzyme release and
decreased intensive care unit and hospital stays compared with control patients. Yau et al. (64) reported
that vitamin E improved contractile function and decreased cardiac enzyme release. Sisto et al. (65) found
that a mixture of vitamins C and E and allopurinol
decreased perioperative ischemic events and cardiac
enzyme release. In contrast, Lassnigg et al. (66) found
no benefit of parenteral vitamin E in patients undergoing heart surgery, and Westhuyzen et al. (67) found
no benefit of a mixture of oral vitamins C and E in

ANESTH ANALG
2005;101:127587

REVIEW ARTICLE
KEVIN ET AL.
REACTIVE OXYGEN SPECIES AND THE HEART: RELEVANCE TO ANESTHESIA

patients undergoing CABG. The variety of preparations

used, routes of administration, and dosages makes it
difficult to draw conclusions from these disparate results. As a result, antioxidant vitamins have not become
routine in cardiac surgery. Investigation of the effects of
antioxidant vitamins on chronic cardiovascular disease
states has been more extensive and study design has
been more uniform. Placebo-controlled trials have repeatedly failed to demonstrate any useful benefit in patients with CCF or when used for secondary prevention
of ischemic heart disease (68,69).

N-acetylcysteine
N-acetylcysteine (NAC) is a glutathione precursor and
direct scavenger of several radical species. It was
shown to decrease postischemic stunning, but not infarct size, in animal hearts (70). Small studies in the
middle 1990s found that NAC decreased the neutrophil oxidative burst in patients undergoing CPB (71).
In addition, when given in combination with nitroglycerin to patients with acute myocardial infarction,
NAC decreased lipid peroxidation and improved the
cardiac index (72). Interest in this drug for therapeutic
use in patients with myocardial ischemia appears to
have waned despite these promising results.

Measures to Decrease Neutrophil-Derived ROS

During CPB
ROS produced by activated neutrophils via the respiratory burst cascade are believed to contribute to cardiac injury during and after CPB. Thus, measures that
decrease neutrophil activation during CPB are proposed to decrease ROS formation and to improve
cardiac outcome. Aprotinin, the widely used serine
protease inhibitor, decreases neutrophil activation
during bypass and has been shown to decrease ROS
production by neutrophils in vitro (73). In an in vivo
dog model, aprotinin was found to decrease postischemic contractile dysfunction (74) and in a clinical trial of
patients undergoing CPB, aprotinin was found to decrease cardiac enzyme release (75). Effects of aprotinin
are complex, however, and it is impossible at this point
to determine if cardioprotective effects result from alterations of coagulation, fibrinolysis, cytokine formation,
ROS formation, or a combination of each of these.
Leukocyte depletion by filtration was shown to decrease cardiac enzyme release and to preserve stores
of glutathione (an endogenous antioxidant) in patients
undergoing CPB (76). Two trials have shown that
heparin-coating of extracorporeal circuits decreases
evidence of ROS in blood (77,78). Corticosteroids have
been shown to protect against cardiac reperfusion injury in animal hearts (79) and were found to decrease
myocardial apoptosis in a pig model of CPB, possibly
because of a decrease in ROS (80). Multiple trials have

1281

evaluated effects of corticosteroids to improve outcome after CPB, but despite a reduction in lipid peroxidation products and the largely universal finding
of attenuated levels of inflammatory mediators, evidence does not support routine administration.

Summary: Antioxidants in Cardiac Injury

Laboratory studies have conclusively demonstrated
that excess ROS cause functionally significant injury to
cardiac tissue. In a variety of clinical settings, measurement of peroxidation production has shown clear
evidence of ROS-mediated cellular injury, and it appears reasonable to suppose that ROS are causing
important functional impairment similar to that
shown in animal studies. Unfortunately, however, the
clinical use of exogenous antioxidant substances has
generally failed to live up to their early promise.
There are a number of possible reasons for this
difference between laboratory and clinical studies.
First, there is a fundamental difficulty in successfully
delivering the drug to myocardial tissue at risk for
injury, for ischemic tissue is poorly accessible to drug
delivery; delivery of the drug solely at reperfusion
may provide suboptimal effect, as significant ROS
production is known to occur during ischemia, not
just at reperfusion. Even at reperfusion, incomplete
perfusion related to intravascular thrombosis and extravascular compression is likely to compromise drug
delivery. A second major limitation for many of these
drugs is poor cellular permeability. Scavenging of
ROS will be largely limited to the extracellular space,
but, as ROS will have exerted much cellular injury before
the excess leaks out of the cell, limited cellular protection
can be expected. Furthermore, within the reperfused
myocyte, the mitochondrial electron transport chain is
the predominant site of ROS production, and injury to
mitochondrial components may be a key determinant of
postischemic cellular function. Thus, permeability for
subcellular compartments may be a prerequisite for effective action of antioxidants in this setting.
Finally, some of the lack of congruence between
laboratory and clinical studies may relate to the fact
that most laboratory studies entail subjecting a previously normal heart to an ischemic insult. In contrast,
human studies generally occur in the setting of injury
to a heart that has already been significantly altered by
disease processes and by senescence.

Anesthetics: Interactions with Cardiac

Redox Chemistry
Propofol
Propofol has a structure similar to phenol-based scavengers such as the endogenous antioxidant vitamin E,
and its scavenging activity has been demonstrated in

1282

REVIEW ARTICLE KEVIN ET AL.

REACTIVE OXYGEN SPECIES AND THE HEART: RELEVANCE TO ANESTHESIA

animal and human models. It is effective against a

broad range of radicals, including ONOO, the product of O2 and NO (81). Intralipid, the solvent for
propofol, also has scavenging activity, but this is negligible at clinical concentrations (81).
Propofol has been shown to be protective in experimental models of injury in several organs, including
brain, liver, and heart (82,83). In intact organ models,
at least some of this protection may result from preservation of endothelium-dependent vasodilation,
which is impaired by oxidative stress. In heart models,
propofol was found to be protective against peroxidative damage and functional impairment induced by
exogenous H2O2 (84) and by ischemia-reperfusion
(85). It has been suggested that propofol-induced cardioprotection may partly result from a direct effect on
myocardial calcium influx (86), or from inhibition of
mitochondrial permeability transition (MPT) (87).
Opening of the MPT pore uncouples mitochondria
and is involved in determining pathways that lead to
apoptosis and necrosis. This latter effect may not be
independent of the radical scavenging effect, however,
as free radicals are believed to modulate the MPT (88).
Although propofol protects against ischemiareperfusion injury when given before initiation of ischemia (85), its administration on reperfusion alone
may be ineffective (89); this possibly reflects the occurrence of oxidant damage during the ischemic period (1).
Clinical studies support a protective effect of propofol against peroxidative injury. Propofol reduced levels of malondialdehyde, a marker of lipid peroxidation, after tourniquet-induced lower limb ischemiareperfusion injury (90), and propofol decreased
oxidative damage measured in platelets (91) from surgical patients. In a study by Sayin et al. (92), 24 patients were randomized to receive propofol
3 6 mg kg1 h1 or fentanyl 10 30 g kg1 h1
during CABG operations. Atrial biopsies were taken
at several time points during the bypass period, and
lipid peroxidation was measured using thiobarbutiric
acid substances. Compared with fentanyl, propofol
suppressed lipid peroxidation, but the study was underpowered to detect differences in functional outcome measures. Further studies will be required to
determine if the cellular protective effects of propofol
translate into meaningful improvements in perioperative outcome.

Barbiturates
Barbiturates have been shown to have scavenging activity in laboratory models, although to a lesser degree
than propofol. The majority of published work concerns protection against cerebral or spinal ischemiareperfusion injury. Barbiturates differ in scavenging
potency in neuronal models, as thiopental is more

ANESTH ANALG
2005;101:127587

potent than either methohexital (93) or pentobarbital

(94). Similarly, respiratory burst activity of neutrophils was suppressed more by thiopental than by
methohexital or by phenobarbital (95). ROS scavenging effects of thiopental have been used to explain
protective effects against bowel ischemia (96), but
there is scant literature concerning scavenging effects
on myocardial ischemia-reperfusion injury. No effect
of thiamylal on altering radical production was reported using an isolated heart ischemia-reperfusion
model (97). However, sodium amytal (amobarbital) is
known experimentally to reversibly block complex I
(NADH oxidoreductase) electron flow.
Clinical studies similarly show that the antioxidative effects of barbiturates are questionably significant.
No change in oxidative stress was measurable in
platelets of surgical patients after anesthesia with thiopental, in contrast to propofol, which decreased oxidative stress (91). In summary, free radical scavenging
effects of barbiturates, particularly thiopental, likely
contribute to the well-validated neuroprotective effects of these drugs, but there is insufficient evidence
to support significant scavenging effects after myocardial ischemia-reperfusion injury.

Etomidate
Etomidate had a modest effect (98) or no effect (99) on
respiratory burst activity of neutrophils at clinical concentrations. No effect of etomidate on adhesion of
neutrophils to coronary endothelium was found in
postischemic hearts (100).

Ketamine
Ketamine has a weak scavenging effect (4) and a weak
effect to suppress ROS production by neutrophils that
is probably minimal at clinical concentrations (101).
The effect is not stereoselective, suggesting that specific receptor interactions are not involved (101). Conversely, Reinke et al. (102) showed that ketamine
could generate free radicals in vivo. This was an incidental finding when rats were anesthetized with ketamine for later experiments on liver extracts; ROS
formation also occurred when ketamine was added
directly to the liver extracts. The radical product was
thought to be a relatively stable ketamine-nitroxide
radical. Interestingly, ketamine interfered with the
ability to measure other free radicals, suggesting that
this radical product was reacting with other radicals,
possibly accounting for the reported scavenging effect
of ketamine. This may have importance for investigators of ROS when laboratory animals are anesthetized
with ketamine.

ANESTH ANALG
2005;101:127587

REVIEW ARTICLE
KEVIN ET AL.
REACTIVE OXYGEN SPECIES AND THE HEART: RELEVANCE TO ANESTHESIA

Benzodiazepines
Radical scavenging activity of midazolam is significantly less than that of propofol in in vitro preparations (103) and it is likely insufficient to impact on in
vivo ROS formation. After limb-tourniquet ischemiareperfusion, a large increase in ROS production was
observed in patients sedated with midazolam whereas
no increase was seen in patients sedated with propofol
(104). Data on effects of benzodiazepines on neutrophilic respiratory burst activity are conflicting. Investigators have reported substantial effect (105), modest
effect (106), or no effect (98) of midazolam on
neutrophil-derived ROS at clinical concentrations.

Volatile Anesthetics
Volatile anesthetics have been shown to protect
against cardiac ischemia-reperfusion injury by mechanisms that are apparently independent of their effects
to cause coronary vasodilation (5) or to depress cardiac contractility (3,4). Free radical release was decreased after ischemia-reperfusion in animal hearts
exposed to halothane (107,108), isoflurane (108), or
sevoflurane (3,9). The specific mechanism of this decrease in ROS formation is unclear but it is likely to
result from initiation of an intrinsic protective effect.
Brief exposure of the heart to a volatile anesthetic
has been shown to induce a state of protection against
the effects of ischemia-reperfusion injury in animal
models (3,9,109 111). This has been termed APC because it resembles ischemic preconditioning in many
respects, including functional and structural outcome.
There is also evidence of its occurrence in humans.
Belhomme et al. (112) reported that brief pretreatment
with isoflurane decreased cardiac enzyme release in
patients undergoing CPB. In a subsequent study, improved outcome in cardiac surgery patients given volatile anesthetics, rather than propofol (113), was considered suggestive evidence of a meaningful
preconditioning effect of volatile anesthetics (114). Experimental and clinical studies on APC have been
reviewed recently (115).
APC has been shown to lead to decreased cardiac
ROS formation during and after ischemia and reperfusion (3,9); specifically, mitochondrial generation of
ROS on reperfusion after ischemia was attenuated by
APC (116). It is difficult to discern, however, based on
existing reports, if reduced ROS generation merely
accompanies cardioprotection with APC or represents
a fundamental mechanism of protection mediated by
APC. It is also difficult to determine if decreased ROS
production occurs solely as a result of APC or if volatile anesthetics have additional direct antioxidant effects when present in the heart during ischemia and
reperfusion. This latter possibility appears unlikely, as
exposure to isoflurane (117) or sevoflurane (3) directly
induces generation of O2 in the intact heart. These

1283

observations suggest that ROS generation is essential

for the initiation of APC, i.e., the free radicals are
components, possibly the initiators, of the signaling
cascade that leads to cardioprotection. This is the subject of a recent review (118).
APC can be abolished by ROS scavengers. Mullenheim at al. (111) found that ROS scavengers, given
with isoflurane to rabbits, abrogated effects of isoflurane to induce APC. Novalija et al. (9) similarly abrogated sevoflurane-induced APC effects on mechanical
and vascular function using ROS scavengers, and additionally abrogated preconditioning using the NO
scavenger L-nitroarginine methyl ester; this suggested
that NO, or possibly the NOO2 product ONOO,
is also required for APC. Kevin et al. (3) found that
myocardial ROS generation was reduced during both
ischemia and reperfusion after sevoflurane APC and
restored when a SOD mimetic was given during
sevoflurane exposure. Thus free radicals appear to
play a dual, and apparently paradoxical, role in APC
as formation of a small quantity of ROS is required to
trigger APC, while decreased free radical formation
during subsequent ischemia and reperfusion may underlie, at least in part, the functional and structural
preservation.
Although direct generation of free radicals during
(3) and after (117) anesthetic exposure was demonstrated in animal heart models, previous studies in
isolated coronary (119) and mesenteric (120) blood
vessels suggested indirectly that anesthetics might
cause ROS formation. How do volatile anesthetics
generate free radicals? Hepatotoxicity attributable to
halothane may result partly from conversion of halothane to free radicals by cytochrome P450 enzymes
(121), but there is no evidence of similar reactions
involving other volatile anesthetics in other organs.
Studies by Cohen (122) in the 1970s, demonstrated
that volatile anesthetics mildly depressed oxidative
phosphorylation. Later, Kissin et al. (123) showed that
NADH was increased in a dose-dependent manner by
each of four different anesthetics.
The above studies strongly implicate the electron
transport chain of mitochondria as the most likely
source of anesthetic-induced ROS generation, as inhibition of mitochondrial enzymes may be expected to
concurrently increase stores of reducing equivalents
and to generate free radicals at corresponding electron
transport chain complexes. Indeed, Hanley et al. (124)
demonstrated recently that halothane, isoflurane, and
sevoflurane directly inhibited complex I (NADH:
ubiquinone oxidoreductase) activity in cardiac mitochondria. Riess et al. (125), moreover, found that a
sevoflurane-induced increase in NADH level was
abolished by 5-hydroxydecanoate, a putative inhibitor
of the mitochondrial ATP-sensitive K channel; their
results indicate that the effect of volatile anesthetics on
mitochondrial enzymes might also be indirect and

1284

REVIEW ARTICLE KEVIN ET AL.

REACTIVE OXYGEN SPECIES AND THE HEART: RELEVANCE TO ANESTHESIA

mediated by altered mitochondrial ion flux. Therefore,

although volatile anesthetics do appear to generate
ROS in mitochondria, the exact mechanism of this
activity is unclear.
Volatile anesthetics alter several variables of neutrophil function, including adhesion to coronary endothelial cells and microbicidal activity. In 1911, Graham
(126) demonstrated impaired neutrophil function after
exposure to ether. Halothane (127129) and enflurane
(128) are reported to decrease ROS release from leukocytes. Isoflurane, in contrast, has been reported to
have little effect (127,129) or even to increase ROS
release by neutrophils (128). Nakagawara et al. (128)
reported inhibition of calcium mobilization in neutrophils in a dose-dependent manner with halothane and
enflurane. This inhibition correlated with effects on
O2 formation; Ca2 is believed to act as second messenger in the activation of NADPH oxidase. Protein
kinase C is also involved in this activation pathway.
Frohlich et al. (130) reported that direct protein kinase
C activation resulted in ROS formation that was unaltered by volatile anesthetics. This seems to exclude a
direct effect of anesthetics on NADPH in neutrophils
and also excludes a direct scavenging effect on formed
free radicals. Thus the current evidence suggests volatile anesthetics interfere with the neutrophil respiratory burst activation pathway by interfering with a
step proximal to protein kinase C activation, most
likely at the stage of capacitative Ca2 influx. Nonetheless, the reason for differential effects of various
volatile anesthetics is unknown.
As discussed above, ROS from neutrophils are believed to contribute to cardiac injury related to CPB,
ischemia, and reperfusion. Attenuated ROS release
from neutrophils exposed to volatile anesthetics may
contribute to cardioprotective effects of these drugs.
Interestingly, one study indicates that suppression of
neutrophil ROS formation resulting from volatile anesthetics exhibits a memory phase, i.e., is present even
after the anesthetic is discontinued; thus this may also
contribute to the cardioprotection observed with APC
(131).

Summary
During and after cardiac ischemia-reperfusion, and
during CPB, widespread ROS formation occurs in cardiac and vascular cells and in neutrophils. ROS are
central to the pathophysiology of cardiac ischemiareperfusion injury; they contribute to the genesis of
myocardial stunning, necrosis, and apoptosis, and
possibly to dysrhythmias. They have also been linked
to the development of chronic cardiovascular diseases
including CCF and atherosclerosis. Multiple animal
investigations have demonstrated attenuated cardiac
injury when scavengers of ROS were administered

ANESTH ANALG
2005;101:127587

before ischemia or in the setting of chronic cardiovascular diseases. Clinical studies have demonstrated decreased ROS-mediated tissue damage, but have failed
to conclusively demonstrate important outcome benefits of such therapies. Some IV anesthetic drugs act as
ROS scavengers. Propofol is a particularly effective
scavenger. Laboratory and clinical trials have demonstrated cardioprotective effects, but whether this property of propofol provides an advantage for patients at
risk of myocardial ischemia-reperfusion seems
doubtful.
Among the most intriguing findings in recent cardiovascular anesthesia research is the observation
that volatile anesthetics generate ROS, an effect that
paradoxically confers cardioprotection. There is
now some evidence to suggest that this property can
be effectively harnessed to the patients advantage.
Finally, one of the main lessons that recent free
radical research teaches us is that free radicals
should no longer be considered as only producing
damage. NO is essential to normal vascular function. Other ROS, like O2, H2O2, OH, or ONOO,
may induce or mediate APC and also play a part in
signaling cascades essential to normal cardiac function. Free radicals interact with each other and with
endogenous antioxidant systems. It is the balance of
all these constituents that determines their beneficial versus deleterious effects.

References
1. Vanden Hoek TL, Li C, Shao Z, et al. Significant levels of
oxidants are generated by isolated cardiomyocytes during ischemia prior to reperfusion. J Mol Cell Cardiol 1997;29:
2571 83.
2. Kevin LG, Camara AK, Riess ML, et al. Ischemic preconditioning alters real-time measure of O2 radicals in intact hearts with
ischemia and reperfusion. Am J Physiol Heart Circ Physiol
2002;283:H566 74.
3. Kevin LG, Novalija E, Riess ML, et al. Sevoflurane exposure
generates superoxide but leads to decreased superoxide during ischemia and reperfusion in isolated hearts. Anesth Analg
2003;96:949 55.
4. Rouquette M, Page S, Bryant R, et al. Xanthine oxidoreductase
is asymmetrically localised on the outer surface of human
endothelial and epithelial cells in culture. FEBS Lett 1998;426:
397 401.
5. Kawahito K, Kobayashi E, Ohmori M, et al. Enhanced responsiveness of circulatory neutrophils after cardiopulmonary
bypass: increased aggregability and superoxide producing capacity. Artif Organs 2000;24:37 42.
6. Clermont G, Vergely C, Jazayeri S, et al. Systemic free radical
activation is a major event involved in myocardial oxidative
stress related to cardiopulmonary bypass. Anesthesiology
2002;96:80 7.
7. Guzik T, Mussa S, Gastaldi D, et al. Mechanisms of increased
vascular superoxide production in human diabetes mellitus:
role of NAD(P)H oxidase and endothelial nitric oxide synthase. Circulation 2002;105:1656 62.
8. Kojda G, Kottenberg K. Regulation of basal myocardial function by NO. Cardiovasc Res 1999;41:514 23.

ANESTH ANALG
2005;101:127587

REVIEW ARTICLE
KEVIN ET AL.
REACTIVE OXYGEN SPECIES AND THE HEART: RELEVANCE TO ANESTHESIA

9. Novalija E, Varadarajan SG, Camara AKS, et al. Anesthetic

preconditioning: Triggering role of reactive oxygen and nitrogen species in isolated hearts. Am J Physiol Heart Circ Physiol
2002;283:H44 52.
10. Mungrue I, Gros R, You X, et al. Cardiomyocyte overexpression of iNOS in mice results in peroxynitrite generation, heart
block, and sudden death. J Clin Invest 2002;109:735 43.
11. Brodsky S, Gao S, Li H, Goligorsky M. Hyperglycemic switch
from mitochondrial nitric oxide to superoxide production in
endothelial cells. Am J Physiol Heart Circ Physiol 2002;283:
H2130 9.
12. McColl AJ, Keeble T, Hadjinikolaou L, et al. Plasma
antioxidants: evidence for a protective role against reactive
oxygen species following cardiac surgery. Ann Clin Biochem
1998;35:616 23.
13. Kloner RA, Przyklenk K, Kay GL. Clinical evidence for
stunned myocardium after coronary artery bypass surgery.
J Card Surg 1994;9:397 402.
14. Chang PP, Sussman MS, Conte JV, et al. Postoperative ventricular function and cardiac enzymes after on-pump versus offpump CABG surgery. Am J Cardiol 2002;89:110710.
15. Bolli R, Patel BS, Jeroudi MO, et al. Demonstration of free
radical generation in stunned myocardium of intact dogs
with the use of the spin trap alpha-phenyl N-tert-butyl nitrone.
J Clin Invest 1988;82:476 85.
16. Maulik N, Yoshida T, Das DK. Oxidative stress developed
during the reperfusion of ischemic myocardium induces apoptosis. Free Radic Biol Med 1998;24:869 75.
17. Tavazzi B, Di Pierro D, Bartolini M, et al. Lipid peroxidation,
tissue necrosis, and metabolic and mechanical recovery of
isolated reperfused rat heart as a function of increasing ischemia. Free Radic Res 1998;28:2537.
18. Nakaya H, Tohse N, Kanno M. Electrophysiological derangements induced by lipid peroxidation in cardiac tissue.
Am J Physiol 1987;253:H1089 97.
19. Dhein S, Schott M, Gottwald E, et al. The contribution of
neutrophils to reperfusion arrhythmias and a possible role for
antiadhesive pharmacological substances. Cardiovasc Res
1995;30:881 8.
20. Marczin N, El-Habashi N, Royston D. Free radicals and cardiac
arrhythmias following coronary surgery: actors of the drama
or bystanders of the spectacle. Acta Anaesthesiol Scand 2003;
47:639 42.
21. Dhalla AK, Hill MF, Singal PK. Role of oxidative stress in
transition of hypertrophy to heart failure. J Am Coll Cardiol
1996;28:506 14.
22. Blaustein AS, Schine L, Brooks WW, et al. Influence of exogenously generated oxidant species on myocardial function.
Am J Physiol 1986;250:H5959.
23. Habib FM, Springall DR, Davies GJ, et al. Tumour necrosis
factor and inducible nitric oxide synthase in dilated cardiomyophathy. Lancet 1996;347:11515.
24. Hill MF, Singal PK. Right and left myocardial antioxidant
responses during heart failure subsequent to myocardial infarction. Circulation 1997;96:2414 20.
25. Kaneko M, Lee SL, Wolf CM, Dhalla NS. Reduction of calcium
channel antagonist binding sites by oxygen free radicals in rat
heart. J Mol Cell Cardiol 1989;21:935 43.
26. Kaneko M, Chapman DC, Ganguly PK, et al. Modification of
calcium adrenergic receptors by oxygen free radicals.
Am J Physiol 1991;260:H821 6.
27. Ide T, Tsutsui H, Hayashidani S, et al. Mitochondrial DNA
damage and dysfunction associated with oxidative stress in
failing hearts after myocardial infarction. Circ Res 2001;88:
529 35.
28. McMurray J, Chopra M, Abdullah I, et al. Evidence of oxidative stress in chronic heart failure in humans. Eur Heart J
1993;14:1493 8.
29. Nishiyama Y, Ikeda H, Haramaki N, et al. Oxidative stress is
related to exercise tolerance in patients with heart failure. Am
Heart J 1998;135:11520.

1285

30. Burlacu A, Jinga V, Gafencu AV, Simionescu M. Severity of

oxidative stress generates different mechanisms of endothelial
cell death. Cell Tissue Res 2001;306:409 16.
31. Li PF, Dietz R, von Harsdorf R. Differential effect of hydrogen
peroxide and superoxide anion on apoptosis and proliferation
of vascular smooth muscle cells. Circulation 1997;96:36029.
32. Rajagopalan S, Meng XP, Ramasamy S, et al. Reactive oxygen
species produced by macrophage-derived foam cells regulate
the activity of vascular matrix metalloproteinases in vitro.
Implications for atherosclerotic plaque stability. J Clin Invest
1996;98:2579 9.
33. Shi Y, Niculescu R, Wang D, et al. Increased NAD(P)H oxidase
and reactive oxygen species in coronary arteries after balloon
injury. Arterioscler Thromb Vasc Biol 2001;21:739 45.
34. Azevedo LC, Pedro MA, Souza LC, et al. Oxidative stress as a
signaling mechanism of the vascular response to injury: the
redox hypothesis of restenosis. Cardiovasc Res 2000;47:436 45.
35. Mittal CK, Murad F. Activation of guanylate cyclase by superoxide dismutase and hydroxyl radical: a physiological regulator of guanosine 3',5'-monophosphate formation. Proc Natl
Acad Sci U S A 1977;74:4360 4.
36. Shibanuma M, Kuroki T, Nose K. Release of H2O2 and phosphorylation of 30 kilodalton proteins as early responses of cell
cycle-dependent inhibition of DNA synthesis by transforming
growth factor beta 1. Cell Growth Differ 1991;2:58391.
37. Sauer H, Rahimi G, Hescheler J, Wartenberg M. Effects of
electrical fields on cardiomyocyte differentiation of embryonic
stem cells. J Cell Biochem 1999;75:710 23.
38. Prabhakar NR. Oxygen sensing by the carotid body chemoreceptors. J Appl Physiol 2000;88:228795.
39. Neumcke I, Schneider B, Fandrey J, Pagel H. Effects of pro- and
antioxidative compounds on renal production of erythropoietin. Endocrinology 1999;140:6415.
40. Huang LE, Gu J, Schau M, Bunn HF. Regulation of hypoxiainducible factor 1alpha is mediated by an O2-dependent degredation domain via the ubiquitin-proteosome pathway. Proc
Natl Acad Sci U S A 1990;95:798792.
41. Wolin MS, Burke-Wolin TM, Mohazzab-H KM. Roles for
NAD(P)H oxidases and reactive oxygen species in vascular
sensing mechanisms. Respir Physiol 1999;115:229 38.
42. Schreck R, Rieber P, Baeuerle PA. Reactive oxygen intermediates as apparently widely used messengers in the activation of
the NF-kappa B transcription factor and HIV-1. EMBO J 1991;
10:224758.
43. Napoli C, de Nigris F, Palinski W. Multiple role of reactive
oxygen species in the arterial wall. J Cell Biochem 2001;82:
674 82.
44. Kilgore KS, Friedrichs GS, Johnson CR, et al. Protective effects
of the SOD-mimetic SC-52608 against ischemia/reperfusion
damage in the rabbit isolated heart. J Mol Cell Cardiol 1994;
26:9951006.
45. Li Q, Bolli R, Qiu Y, et al. Gene therapy with extracellular
superoxide dismutase protects conscious rabbits against myocardial infarction. Circulation 2001;103:1893 8.
46. Murohara Y, Yui Y, Hattori R, Kawai C. Effects of superoxide
dismutase on reperfusion arrhythmias and left ventricular
function in patients undergoing thrombolysis for anterior wall
acute myocardial infarction. Am J Cardiol 1991;67:7657.
47. Flaherty JT, Pitt B, Gruber JW, et al. Recombinant human
superoxide dismutase (h-SOD) fails to improve recovery of
ventricular function in patients undergoing coronary angioplasty for acute myocardial infarction. Circulation 1994;89:
198291.
48. Tsao C, Greene P, Odlind B, Brater D. Pharmacokinetics of
recombinant human superoxide dismutase in healthy volunteers. Clin Pharmacol Ther 1991;50:71320.
49. Ihnat P, Vennerstrom J, Robinson D. Synthesis and solution
properties of deferoxamine amides. J Pharm Sci 2000;89:
152536.

1286

REVIEW ARTICLE KEVIN ET AL.

REACTIVE OXYGEN SPECIES AND THE HEART: RELEVANCE TO ANESTHESIA

50. Ambrosio G, Zweier JL, Jacobus WE, et al. Improvement of

postischemic myocardial function and metabolism induced by
administration of deferoxamine at the time of reflow: the role
of iron in the pathogenesis of reperfusion injury. Circulation
1987;76:906 15.
51. Bolli R, Patel BS, Zhu WX, et al. The iron chelator desferrioxamine attenuates postischemic ventricular dysfunction.
Am J Physiol 1987;253:H1372 80.
52. Bernier M, Hearse DJ, Manning AS. Reperfusion-induced arrhythmias and oxygen-derived free radicals: studies with
anti-free radical interventions and a free radical- generating
system in the isolated perfused rat heart. Circ Res 1986;58:
331 40.
53. Ferreira R, Burgos M, Milei J, et al. Effect of supplementing
cardioplegic solution with deferoxamine on reperfused human
myocardium. J Thorac Cardiovasc Surg 1990;100:708 14.
54. Drossos G, Lazou A, Panagopoulos P, Westaby S. Deferoxamine cardioplegia reduces superoxide radical production in
human myocardium. Ann Thorac Surg 1995;59:169 72.
55. Engl M, Cavarocchi N, OBrien J et al. Influence of antioxidants
(mannitol and allopurinol) on oxygen free radical generation
during and after cardiopulmonary bypass. Circulation 1986;74:
III134 7.
56. Larsen M, Webb G, Kennington S, et al. Mannitol in cardioplegia as an oxygen free radical scavenger measured by malondialdehyde. Perfusion 2002;17:515.
57. Akizuki S, Yoshida S, Chambers DE, et al. Infarct size limitation by the xanthine oxidase inhibitor, allopurinol, in closedchest dogs with small infarcts. Cardiovasc Res 1985;19:686 92.
58. Nielsen VG, Tan S, Baird MS, et al. Xanthine oxidase mediates
myocardial injury after hepatoenteric ischemia-reperfusion.
Crit Care Med 1997;25:1044 50.
59. Johnson W, Kayser K, Brenowitz J, Saedi S. A randomized
controlled trial of allopurinol in coronary bypass surgery. Am
Heart J 1991;121:20 4.
60. Merrill G, McConnell P, Vandyke K, Powell S. Coronary and
myocardial effects of acetaminophen: protection during ischemia-reperfusion. Am J Physiol Heart Circ Physiol 2001;280:
H2631 8.
61. Merrill GF. Acetaminophen and low-flow myocardial ischemia: efficacy and antioxidant mechanisms. Am J Physiol Heart
Circ Physiol 2002;282:H13419.
62. Massey KD, Burton KP. alpha-Tocopherol attenuates myocardial membrane-related alterations resulting from ischemia and
reperfusion. Am J Physiol 1989;256:H11929.
63. Dingchao H, Zhiduan Q, Liye H, Xiaodong F. The protective
effects of high-dose ascorbic acid on myocardium against
reperfusion injury during and after cardiopulmonary bypass.
Thorac Cardiovasc Surg 1994;42:276 8.
64. Yau TM, Weisel RD, Mickle DA, et al. Vitamin E for coronary
bypass operations: a prospective, double-blind, randomized
trial. J Thorac Cardiovasc Surg 1994;108:30210.
65. Sisto T, Paajanen H, Metsa-Ketela T, et al. Pretreatment with
antioxidants and allopurinol diminishes cardiac onset events
in coronary artery bypass grafting. Ann Thorac Surg 1995;59:
1519 23.
66. Lassnigg A, Punz A, Barker R, et al. Influence of intravenous
vitamin E supplementation in cardiac surgery on oxidative
stress: a double-blinded, randomized, controlled study. Br J
Anaesth 2003;90:148 54.
67. Westhuyzen J, Cochrane AD, Tesar PJ, et al. Effect of preoperative supplementation with alpha-tocopherol and ascorbic acid
on myocardial injury in patients undergoing cardiac operations. J Thorac Cardiovasc Surg 1997;113:942 8.
68. The Heart Outcomes Prevention Evaluation Study Investigators. Vitamin E supplementation and cardiovascular events in
high-risk patients. N Engl J Med 2000;342:154 60.
69. Vivekananthan DP, Penn MS, Sapp SK, et al. Use of antioxidant vitamins for the prevention of cardiovascular diseases: a
meta-analysis of randomised trials. Lancet 2003;361:201723.

ANESTH ANALG
2005;101:127587

70. Forman MB, Puett DW, Cates CU, et al. Glutathione redox
pathway and reperfusion injury: effect of N- acetylcysteine on
infarct size and ventricular function. Circulation 1988;78:
20213.
71. Andersen LW, Thiis J, Kharazmi A, Rygg I. The role of
N-acetylcystein administration on the oxidative response of
neutrophils during cardiopulmonary bypass. Perfusion 1995;
10:21 6.
72. Arstall MA, Yang J, Stafford I, et al. N-acetylcysteine in combination with nitroglycerin and streptokinase for the treatment
of evolving acute myocardial infarction: safety and biochemical effects. Circulation 1995;92:2855 62.
73. Hallett MB, Shandall A, Young HL. Mechanism of protection
against reperfusion injury by aprotinin: roles of polymorphonuclear leucocytes and oxygen radicals. Biochem Pharmacol 1985;34:1757 61.
74. McCarthy R, Tuman KJ, OConnor C, Ivankovich AD. Aprotinin pretreatment diminishes postischemic myocardial contractile dysfunction in dogs. Anesth Analg 1999;89:1096 100.
75. Wendel HP, Heller W, Michel J, et al. Lower cardiac troponin
T levels in patients undergoing cardiopulmonary bypass and
receiving high-dose aprotinin therapy indicate reduction of
perioperative myocardial damage. J Thorac Cardiovasc Surg
1995;109:1164 72.
76. Di Salvo C, Louca LL, Pattichis K, et al. Does activated neutrophil depletion on bypass by leukocyte filtration reduce
myocardial damage? A preliminary report. J Cardiovasc Surg
1996;37:93100.
77. Bozdayi M, Borowiec J, Nilsson L, et al. Effects of heparin
coating of cardiopulmonary bypass circuits on in vitro oxygen
free radical production during coronary bypass surgery. Artif
Organs 1996;20:1008 16.
78. Borowiec JW, Venge P, Henze A, et al. Biomaterial-dependent
blood activation during simulated extracorporeal circulation: a
study of heparin coated and uncoated circuits. Thoracic Cardiovasc Surg 1997;45:295301.
79. Valen G, Kawakami T, Tahepold P, et al. Glucocorticoid pretreatment protects cardiac function and induces cardiac heat
shock protein 72. Am J Physiol Heart Circ Physiol 2000;279:
H836 43.
80. Pearl JM, Nelson DP, Schwartz SM, et al. Glucocorticoids reduce ischemia-reperfusion-induced myocardial apoptosis in
immature hearts. Ann Thorac Surg 2002;74:830 6.
81. Kahraman S, Demiryurek AT. Propofol is a peroxynitrite scavenger. Anesth Analg 1997;84:11279.
82. Young Y, Menon D, Tisavipat N, et al. Propofol neuroprotection in a rat model of ischaemia reperfusion injury. Eur J
Anaesthesiol 1997;14:320 6.
83. Navapurkar V, Skepper J, Jones J, Menon D. Propofol preserves the viability of isolated rat hepatocyte suspensions under an oxidant stress. Anesth Analg 1998;87:11527.
84. Kokita N, Hara A. Propofol attenuates hydrogen peroxideinduced mechanical and metabolic derangements in the isolated rat heart. Anesthesiology 1996;84:11727.
85. Kokita N, Hara A, Abiko Y, et al. Propofol improves functional
and metabolic recovery in ischemic reperfused isolated rat
hearts. Anesth Analg 1998;86:252 8.
86. Buljubasic N, Marijic J, Berczi V, et al. Differential effects of
etomidate, propofol, and midazolam on calcium and potassium channel currents in canine myocardial cells. Anesthesiology 1996;85:10929.
87. Sztark F, Ichas F, Ouhabi R, et al. Effects of the anaesthetic
propofol on the calcium-induced permeability transition of rat
heart mitochondria: direct pore inhibition and shift of the
gating potential. FEBS Lett 1995;368:101 4.
88. Kowaltowski A, Castilho RF, Vercesi AE. Mitochondrial permeability transition and oxidative stress. FEBS Lett 2001;495:
125.
89. Ebel D, Schlack W, Comfere T, et al. Effect of propofol on
reperfusion injury after regional ischaemia in the isolated rat
heart. Br J Anaesth 1999;83:903 8.

ANESTH ANALG
2005;101:127587

REVIEW ARTICLE
KEVIN ET AL.
REACTIVE OXYGEN SPECIES AND THE HEART: RELEVANCE TO ANESTHESIA

90. Aldemir O, Celebi H, Cevik C, Duzgun E. The effects of propofol or halothane on free radical production after tourniquet
induced ischaemia-reperfusion injury during knee arthroplasty. Acta Anaesthesiol Scand 2001;45:12215.
91. De La Cruz JP, Zanca A, Carmona JA, de la Cuesta FS. The
effect of propofol on oxidative stress in platelets from surgical
patients. Anesth Analg 1999;89:1050 5.
92. Sayin MM, Ozatamer O, Tasoz R, et al. Propofol attenuates
myocardial lipid peroxidation during coronary artery bypass
grafting surgery. Br J Anaesth 2002;89:242 6.
93. Smith DS, Rehncrona S, Westerberg E, et al. Lipid peroxidation
in brain tissue in vitro: antioxidant effects of barbiturates. Acta
Physiol Scand 1979;105:5279.
94. Cole DJ, Cross LM, Drummond JC, et al. Thiopentone and
methohexital, but not pentobarbitone, reduce early focal cerebral ischemic injury in rats. Can J Anaesth 2001;48:80714.
95. Weiss M, Buhl R, Birkhahn A, et al. Do barbiturates and their
solutions suppress FMLP-induced neutrophil chemiluminescence? Eur J Anaesthesiol 1994;11:3719.
96. Dalsing MC, Sieber P, Grosfeld JL, et al. Ischemic bowel: the
protective effect of free-radical anion scavengers. J Pediatr Surg
1983;18:360 4.
97. Tamaki F, Oguchi T, Kashimoto S, et al. Effects of propofol on
ischemia and reperfusion in the isolated rat heart compared
with thiamylal. Jpn Heart J 2001;42:193206.
98. Krumholz W, Demel C, Jung S, et al. The effects of thiopentone,
etomidate, ketamine and midazolam on several bactericidal
functions of polymorphonuclear leucocytes in vitro. Eur J Anaesthesiol 1995;12:141 6.
99. Gelb AW, Lok P. Etomidate reversibly depresses human neutrophil chemiluminescence. Anesthesiology 1987;66:60 3.
100. Szekely A, Heindl B, Zahler S, et al. Nonuniform behavior of
intravenous anesthetics on postischemic adhesion of neutrophils in the guinea pig heart. Anesth Analg 2000;90:1293300.
101. Weigand MA, Schmidt H, Zhao Q, et al. Ketamine modulates
the stimulated adhesion molecule expression on human neutrophils in vitro. Anesth Analg 2000;90:206 12.
102. Reinke LA, Kotake Y, Moore DR, Nanji AA. Free radical formation during ketamine anesthesia in rats: a cautionary note.
Free Radic Biol Med 1998;24:1002 6.
103. Tsuchiya M, Asada A, Maeda K, et al. Propofol versus midazolam regarding their antioxidant activities. Am J Respir Crit
Care Med 2001;163:26 31.
104. Cheng YJ, Wang YP, Chien CT, Chen CF. Small-dose propofol
sedation attenuates the formation of reactive oxygen species in
tourniquet-induced ischemia-reperfusion injury under spinal
anesthesia. Anesth Analg 2002;94:161720.
105. Nishina K, Akamatsu H, Mikawa K, et al. The inhibitory effects
of thiopental, midazolam, and ketamine on human neutrophil
functions. Anesth Analg 1998;86:159 65.
106. Kang MY, Tsuchiya M, Packer L, Manabe M. In vitro study on
antioxidant potential of various drugs used in the perioperative period. Acta Anaesthesiol Scand 1998;42:4 12.
107. Ikeya K, Kashimoto S, Kume M, Kumazawa T. The influence of
N-nitro-L-arginine methyl ester (L-NAME) on hydroxyl free
radical formation in the post-ischemic reperfused heart of
anesthetized rats. Masui J 2001;50:36570.
108. Nakamura T, Kashimoto S, Oguchi T, Kumazawa T. Hydroxyl
radical formation during inhalation anesthesia in the reperfused working rat heart. Can J Anaesth 1999;46:470 5.
109. Cope DK, Impastato WK, Cohen MV, Downey JM. Volatile
anesthetics protect the ischemic rabbit myocardium from infarction. Anesthesiology 1997;86:699 709.
110. Kersten JR, Schmeling TJ, Pagel PS, et al. Isoflurane mimics
ischemic preconditioning via activation of KATP channels: reduction of myocardial infarct size with an acute memory
phase. Anesthesiology 1997;87:36170.

1287

111. Mullenheim J, Ebel D, Frabetadorf J, et al. Isoflurane preconditions myocardium against infarction via release of free radicals. Anesthesiology 2002;96:934 40.
112. Belhomme D, Peynet J, Louzy M, et al. Evidence for preconditioning by isoflurane in coronary artery bypass graft surgery.
Circulation 1999;100:II340 4.
113. De Hert SG, Ten Broecke PW, Mertens E, et al. Sevoflurane but
not propofol preserves myocardial function in coronary surgery patients. Anesthesiology 2002;97:429.
114. Warltier D, Kersten JR, Pagel PS, Gross GJ. Editorial view:
anesthetic preconditioning: serendipity and science. Anesthesiology 2002;97:13.
115. De Hert S, Turani F, Mathur S, Stowe DF. Cardioprotection
with volatile anesthetic: mechanism and clinical implications.
Anesth Analg 2005;100:1584 93.
116. Novalija E, Kevin LG, Eells JT, et al. Anesthetic preconditioning improves ATP synthesis and reduces ROS formation in
mitochondria after ischemia by a redox dependent mechanism.
Anesthesiology 2003;98:1155 63.
117. Tanaka K, Weihrauch D, Kehl F, et al. Mechanism of preconditioning by isoflurane in rabbits: a direct role for reactive
oxygen species. Anesthesiology 2002;97:148590.
118. Stowe DF, Kevin LG. Cardiac preconditioning by volatile anesthetic agents: a defining role for altered mitochondrial bioenergetics. Antiox Redox Signal 2004;6:439 48.
119. Park KW, Dai HB, Lowenstein E, et al. Oxygen-derived free
radicals mediate isoflurane-induced vasoconstriction of rabbit
coronary resistance arteries. Anesth Analg 1995;80:11637.
120. Yoshida K, Okabe E. Selective impairment of endotheliumdependent relaxation by sevoflurane: oxygen free radicals participation. Anesthesiology 1992;76:440 7.
121. Lind RC, Gandolfi AJ, Hall PD. The role of oxidative biotransformation of halothane in the guinea pig model of halothaneassociated hepatotoxicity. Anesthesiology 1989;70:649 53.
122. Cohen PJ. Effect of anesthetics on mitochondrial function. Anesthesiology 1973;39:153 64.
123. Kissin I, Aultman DF, Smith LR. Effects of volatile anesthetics
on myocardial oxidation-reduction status assessed by NADH
fluorometry. Anesthesiology 1983;59:44752.
124. Hanley PJ, Ray J, Brandt U, Daut J. Halothane, isoflurane and
sevoflurane inhibit NADH:ubiquinone oxidoreductase (complex I) of cardiac mitochondria. J Physiol 2002;544:68793.
125. Riess ML, Novalija E, Camara AK, et al. Preconditioning with
sevoflurane reduces changes in nicotinamide adenine dinucleotide during ischemia-reperfusion in isolated heart. Anesthesiology 2003;98:38795.
126. Graham EV. The influence of ether and ether anesthesia on
bacteriolysis, agglutination, and phagocytosis. J Infect Dis
1911;8:14175.
127. Lieners C, Redl H, Schlag G, Hammerschmidt DE. Inhibition
by halothane, but not by isoflurane, of oxidative response to
opsonized zymosan in whole blood. Inflammation 1989;13:
62130.
128. Nakagawara M, Takeshige K, Takamatsu J, et al. Inhibition of
superoxide production and Ca2 mobilization in human neutrophils by halothane, enflurane, and isoflurane. Anesthesiology 1986;64:4 12.
129. Frohlich D, Rothe G, Schwall B, et al. Effects of volatile anaesthetics on human neutrophil oxidative response to the bacterial
peptide FMLP. Br J Anaesth 1997;78:718 23.
130. Frohlich D, Rothe G, Wittmann S, et al. Nitrous oxide impairs
the neutrophil oxidative response. Anesthesiology 1998;88:
128190.
131. Hu G, Vasiliauskas T, Salem MR, et al. Neutrophils pretreated
with volatile anesthetics lose ability to cause cardiac dysfunction. Anesthesiology 2003;98:712 8.

CASE REPORT

Is the Valve OK or Not? Immediate Evaluation of a Replaced

Aortic Valve
Rebecca A. Schroeder,

MD,

and Jonathan B. Mark,

MD

Department of Anesthesiology Durham Veterans Medical Center Duke University School of Medicine Durham, North
Carolina

Transesophageal echocardiography is a crucial tool in intraoperative evaluation of newly implanted/repaired

heart valves because suspected valvular malfunction
needs to be identified and sometimes surgically corrected. Although color Doppler is often adequate in
evaluating the expected regurgitant jets, as well as excluding pathologic paravalvular leaks, spectral Doppler techniques are the most commonly used methods
for estimating transvalvular gradients in the operating

ransesophageal echocardiography (TEE) is useful for intraoperative assessment of newly implanted prosthetic heart valves. The TEE evaluation should include documentation of leaflet
movement with 2-dimensional imaging, absence of
paravalvular regurgitation, and estimation of transvalvular pressure gradients. Abnormalities identified
in the operating room (OR) may require immediate
surgical correction.
We present a case in which dysfunction of an aortic
mechanical valve was suspected because of measurement of unusually high peak velocities on spectral
Doppler analysis. This case highlights the dangers of
relying on peak transvalvular velocities as the primary
indicator of valvular stenosis.

Case Report
A 63-yr-old, 84 kg, 179 cm male (body surface area, 2.05)
presented for aortic valve (AV) replacement for severe aortic
stenosis (AS). An echocardiogram showed severe concentric
left ventricular (LV) hypertrophy, preserved systolic LV
function, and a bicuspid, stenotic AV (area 0.9 cm2).
Accepted for publication June 9, 2005.
Address correspondence and reprint requests to Rebecca A.
Schroeder, MD, Department of Anesthesiology Durham Veterans
Medical Center Duke University School of Medicine VAMC
(112C), 508 Fulton St. Durham, NC 27705. Address electronic
mail to schro016@mc.duke.edu.
DOI: 10.1213/01.ANE.0000181339.39448.F0

1288

Anesth Analg 2005;101:128891

room. However, these methods are subject to a variety

of confounding factors, including subvalvular gradients and pressure recovery. Other methods of valve
area estimation should also be used when evaluating a
prostethic aortic valve, including the continuity equation and the left ventricular outflow tract/aortic valve
velocity ratio.
(Anesth Analg 2005;101:1288 91)

Because of extensive calcification of the aortic annulus, the

patient underwent AV replacement with a 21-mm Carbomedics bileaflet valve. The initial, pre-bypass, cardiac output
(CO) was measured via intermittent thermodilution to be 4.5
L/min. At the time of separation from cardiopulmonary
bypass, CO was 6.8 L/min while dopamine and vasopressin
were infused. A small, eccentric paravalvular leak was visualized by TEE. One leaflet of the prosthesis moved well;
the other was not visible. Continuous wave Doppler measurement of the AV showed a peak velocity of 4.8 m/s, a
peak gradient of 91 mm Hg, and a mean gradient of 42 mm
Hg. According to the manufacturers specifications, the expected mean gradient for this valve is 12.3 mm Hg (1). The
surgeon was informed, but as the patient was hemodynamically stable, the chest was closed and the patient was taken
to the intensive care unit (ICU) still on the same vasopressor
infusions. His hematocrit at that time was 28%.
After 2 h, the patient was returned to the OR because of
excessive bleeding. Reexamination of the prosthetic valve
showed high peak velocity (5.14 m/s) with a peak gradient
of 106 mm Hg and mean gradient of 45 mm Hg, a measurement confirmed 5 times by 2 operators (Fig. 1). The
second leaflet still could not be seen. A cardiologist was
consulted who concurred that there was an obstruction at
the valvular level and strongly recommended surgical intervention. By this time, however, the chest had again been
closed. An unsuccessful attempt was made to document
motion of the second leaflet with C-arm fluoroscopy. However, given the patients clinical improvement (CO, 8.2
L/min), he was taken to the ICU for observation.
The following morning, he was taken to the cardiac catheterization lab, where, under adequate fluoroscopy, both
leaflets were seen to be functioning normally. A subsequent
follow-up transthoracic echocardiogram (TTE) at 1 yr
showed a peak gradient of 25 mm Hg and a mean gradient
of 15 mm Hg.
2005 by the International Anesthesia Research Society
0003-2999/05

ANESTH ANALG
2005;101:1288 91

CASE REPORT

1289

Figure 1. A continuous-wave Doppler measurement through the prosthetic aortic valve

from the transgastric position. The peak velocity recorded is 5.3 meters/s, which is abnormally high, even for mechanical prosthetic
valves.

Discussion
Use of TEE after valvular procedures is essential in
ensuring adequate functioning of the newly implanted
or repaired valve (2 6). In our case, 2-dimensional
images of the prosthesis suggested a problem with one
leaflet, whereas Doppler examination showed an unusually high peak velocity through the new valve,
raising clinical suspicion of prosthesis malfunction.
However, fluoroscopy in the catheterization laboratory the next day showed normal motion of both
leaflets. We believe several factors must be considered
for proper interpretation of Doppler data in this confusing clinical setting.
Bernoullis Equation, as it is applied to echocardiography, is adapted to give the modified Bernoulli
Equation as follows:

Pressure Gradient 4 V22V12)

where V2 is the peak velocity through the valve, and
V1 is the subvalvular velocity; that is, through the LV
outflow tract (LVOT) (7). In general practice, V1 is
generally ignored, yielding the simplified form of the
equation:

Pressure Gradient 4 V22)

However, in cases such as ours, where V1 is significantly more than 1.0 m/s (2.6 m/s in this case), the
modified Bernoulli Equation must be used (Fig. 2).
Bird et al. (8) showed that subvalvular gradients (41
17 mm Hg) are present in patients with AS but without evidence of anatomic subvalvular obstruction.
Disregarding V1 introduces an unacceptable degree of
error in these patients.

A second point is the need to calculate valve area

using the continuity equation. With a measured LVOT
diameter of 0.9 cm, V1 of 2.6 m/s, and V2 of 5.14 m/s,
the prosthetic valve area is 1.30 cm2, a value close to
1.54 cm2, the effective area of a 21-mm Carbomedics
valve (1). Even more useful is measurement of the
peak flow ratio (LVOT/AV), which in this case (2.6/
5.14) is approximately 0.5, very near the published
normal for this size and type of valve (0.4) (9,10). Both
of these calculations confirm normal function of the
new valve.
Pressure recovery is responsible for the observed
differences between Doppler-derived gradients and
catheter-measured gradients and is probably the least
recognized factor involved. As flow traverses a narrow orifice, the flow stream contracts (vena contracta)
and subsequently expands on reaching a wider passage. Pressure exerted by the flow stream decreases
with increasing velocity, and recovers as the fluid
slows distal to the maximal obstruction. Thus, pressure measured in the LVOT will decrease at the level
of the AV and recover in the proximal ascending
aorta. Spectral Doppler measurements allow calculation of the maximal pressure gradient between the
proximal flow stream (LVOT) and the vena contracta
(valve orifice), whereas catheter measurements record
gradients between the proximal flow stream (LVOT)
and the recovered pressure beyond the vena contracta
(proximal ascending aorta).
Pressure recovery may be responsible for a significant
portion of our misleading measurements. In St. Jude
bileaflet valves, the proportion of the peak laboratorymeasured gradient (analogous to the Doppler-derived
gradient) attributed to pressure recovery was 53% for

1290

CASE REPORT

ANESTH ANALG
2005;101:1288 91

Figure 2. A high-repetition pulsed-wave

Doppler measurement with the sampling volume placed in the left ventricular outflow
tract documenting a significant degree of subvalvular obstruction. This invalidates the use
of the simplified form of Bernoullis equation.

the central orifice and 29% for the slightly larger side
orifices (11). Interestingly, a smaller proximal aorta
may be an important predictor for pressure recovery.
In a series of 23 patients with native aortic stenosis, all
of those with Doppler-catheter gradient differences
more than 20 mm Hg had aortas that measured 3 cm
in diameter (12). In our patient, the diameter of the
proximal aorta was 2.8 cm.
Several other factors may have been important in our
case. The presence of postbypass anemia and an increased CO, as well as factors such as patient/prosthetic
mismatch and technical errors, may have contributed.
The possibility of measurement or operator error must
also be considered. Measurement through the smaller
central orifice of a bileaflet prosthesis will yield a higher
peak velocity than the larger side orifices, an error decreased by multiple measurements.
Diagnostic evaluation in the OR can use a variety of
echocardiographic and other maneuvers depending
on the expertise of the anesthesiologist and the surgeon, as well as the stage of the operation. Use of
epiaortic scanning in the hands of a skilled surgeon
may settle the question definitely by imaging the leaflets more clearly. However, when using Doppler techniques, the same issues would apply as with TEE. If
the chest has been closed, or if doubt remains before
leaving the OR, a TTE could be performed, which may
be able to image the leaflets or give a more superior
Doppler signal of the LVOT given its more flexible
position on the chest. Various case reports have been
published describing similar cases in which gradients
have been measured with the use of needles placed
directly into the LV and ascending aorta. Unfortunately, these directly-measured gradients are subject
to error from similar sources as Doppler-measured

gradients (that is, pressure recovery), as well as confounding factors from impact energy interacting with
single-holed needles and acoustic shadowing from the
prosthesis itself (13). Thus, these data should be interpreted with caution.
In summary, we report a case during which an
unusually high peak velocity measured after aortic
prosthetic valve implantation, combined with an inability to visualize leaflet motion, raised serious concerns of prosthesis malfunction or stuck leaflet. In this
setting, it is imperative to consider the presence of a
subvalvular gradient and apply the modified rather
than the simplified Bernoulli Equation. Furthermore,
effective valve area should be calculated using the
continuity equation and, most importantly, by the
LVOT/AV velocity ratio to assess AV function (normal, 0.35 0.5 for range of prosthetic valves) (13). The
phenomenon of pressure recovery should be considered in evaluating prosthetic valve function when
small prostheses (19, 21 mm) are implanted, and especially when the proximal ascending aorta is small.
Other echocardiographic techniques may be helpful,
especially epiaortic scanning by a skilled surgeon, and
TTE, if the chest is already closed.

References
1. Carbomedics heart valves pressure gradient measurement. Available at: http://www.mitroflow.com/professional_surgeon_
pressure.asp. Accessed January 24, 2005.
2. Oh JK, Taliercio CP, Holmes DR Jr, et al. Prediction of the
severity of aortic stenosis by Doppler aortic valve area
determination: prospective Doppler-catheterization correlation
in 100 patients. J Am Coll Cardiol 1988;11:122734.
3. Khandheria BK, Seward JB, Tajik AJ. Transesophageal echocardiography. Mayo Clin Proc 1994;69:856 63.

ANESTH ANALG
2005;101:1288 91

4. Chaliki HP, Click RL, Abel MD. Comparison of intraoperative

transesophageal echocardiographic examinations with the operative findings: prospective review of 1918 cases. J Am Soc
Echo 1999;12:237 40.
5. Nowrangi SK, Connolly HM, Freeman WK, Click RL. Impact of
intraoperative transesophageal echocardiography among patients undergoing aortic valve replacement for aortic stenosis.
J Am Soc Echo 2001;14:863 6.
6. Ionescu AA, West RR, Proudman C, et al. Prospective study of
routine perioperative transesophageal echocardiography for
elective valve replacement: clinical impact and cost-saving implications. J Am Soc Echocardiogr 2001;14:659 67.
7. Hemodynamic assessment. In: Oh JK, Seward JB, Tajik AJ, eds.
The echo manual. Philadelphia: Lippincott Williams & Wilkins,
1999:59 72.
8. Bird JJ, Murgo JP, Pasipoularides A. Fluid dynamics of aortic
stenosis: subvalvular gradients without subvalvular obstruction. Circulation 1982;66:835 40.

CASE REPORT

1291

9. Chafizadeh ER, Zoghbi WA. Doppler echocardiographic assessment of the St. Jude Medical prosthetic valve in the aortic position
using the continuity equation. Circulation 1991;83:21323.
10. Prosthetic valves. In: Otto CM, ed. Textbook of clinical echocardiography. Philadelphia: WB Saunders, 2004:355 83.
11. Bech-Hanssen O, Caidahl K, Wallentin I, et al. Aortic prosthetic
valve design and size: relation to Doppler echocardiographic
findings and pressure recovery: an in vitro study. J Am Soc Echo
2000;13:39 50.
12. Baumgartner H, Stefenelli T, Niederberger J, et al. Overestimation of catheter gradients by Doppler ultrasound in patients
with aortic stenosis: a predictable manifestation of pressure
recovery. J Am Coll Cardiol 1999;33:1655 61.
13. Rehfeldt KH, Click RL. Prosthetic valve malfunction masked by
intraoperative pressure measurements. Anesth Analg 2002;94:
857 8.

ECHO ROUNDS

Transesophageal Echocardiographic Findings in Papillary

Muscle Rupture
Paul C. Whiting,

MD,

and Nicholas J. Morgan-Hughes,

MD

Department of Anaesthesia and Critical Care Medicine, Northern General Hospital, Sheffield, United Kingdom

39-yr-old man presented to our hospital with a

4-h history of chest pain. Despite percutaneous
coronary intervention with stenting of occluded
circumflex and right coronary arteries and placement
of an intraaortic balloon pump, he remained hypotensive and in pulmonary edema. Two transthoracic
echocardiograms performed within the first 24 h of
admission failed to show any significant structural
cardiac abnormality. Forty-eight hours after presentation he was admitted to our cardiac intensive care unit

Video TEE images for this case can be accessed at

www.anesthesia-analgesia.org. Click on Cardiovascular Anesthesia
in the Contents page and then the Data Supplement link located
below the title of the article.
Accepted for publication June 15, 2005.
Address correspondence and reprint requests to Nicholas J
Morgan-Hughes, MD, Department of Anesthesia and Critical Care
Medicine, Northern General Hospital, Herries Road, Sheffield, S5
7AU, UK. Address e-mail to n.hughes@sheffield.ac.uk.
DOI: 10.1213/01.ANE.0000181335.10838.7D

for tracheal intubation and ventilation. At this point,

we performed a transesophageal echocardiogram
(TEE), which revealed papillary muscle (PM) rupture.
He subsequently underwent successful mitral valve
replacement and was discharged home 3 wk later.
TEE showed many of the cardinal features of PM
rupture. The PM head was clearly seen prolapsing
into the left atrium in systole in the midesophageal
4-chamber view (Fig. 1) (video loop can be accessed at
www.anesthesia-analgesia.org) and the midesophageal 2-chamber view. Examination of the left ventricle
in diastole also revealed large-amplitude erratic motion of the ruptured PM head in the midesophageal
4-chamber and midesophageal 2-chamber views.
These TEE results are found in 65% and 90% of cases,
respectively (1). Therefore, although left atrial prolapse of the papillary muscle head is diagnostic, this
finding is absent in 35% of cases and careful examination of the left ventricular cavity will reduce the risk of
a false negative study.

Figure 1. Midesophageal 4-chamber view in

systole. The posteromedial papillary muscle
head (arrow) is seen in the left atrium.

1292

Anesth Analg 2005;101:12923

2005 by the International Anesthesia Research Society

0003-2999/05

ANESTH ANALG
2005;101:12923

ECHO ROUNDS

1293

Figure 2. Transesophageal transmitral continuous wave Doppler profile showing the V

wave cut-off sign (arrow) and a peak early
diastolic filling velocity of up to 1.5 m/s (dotted
line). Note also mirror-image artifact duplicating the mitral regurgitant signal in the reverse
channel.

Left ventricular function was hyperdynamic, although the inferior wall was less hyperdynamic. Twodimensional examination of the mitral valve revealed
flail P3 and A3 segments with prolapse of P2 and A2
segments consistent with rupture of the posteromedial
PM. Posteromedial rupture was confirmed at surgery.
Transmitral continuous wave Doppler revealed a
systolic signal that was complete and of equal intensity to the diastolic signal, suggesting severe regurgitation (Fig. 2). A late systolic reduction in the regurgitant velocity was also seen, giving the Doppler map
an asymmetrical shape. This is known as the V wave
cut-off sign (2) and implies regurgitation of a large
volume of blood into the left atrium, causing a late
systolic reduction in the transmitral gradient because
of a sharp increase in the left atrial pressure. A peak
early diastolic filling velocity (Emax) of up to 1.5 m/s
was seen (Fig. 2). An Emax more than 1.2 m/s is
indicative of severe mitral regurgitation and reflects
volume loading of the left atrium.

Color flow Doppler imaging revealed a broad-based

laterally directed mitral regurgitant jet (video loop).
Appreciation of the color flow map was enhanced by
using the maximum Nyquist limit available (82 cm/s).
Lower limits resulted in an excessive mosaic pattern.
This phenomenon, coupled with tachycardia (heart
rate, 140 to 160 bpm) and the late systolic reduction in
the mitral regurgitation, would reduce color flow
Doppler temporal resolution and could have contributed to the negative transthoracic findings.

References
1. Moursi MH, Bhatnagar SK, Vilacosta I et al. Transesophageal
echocardiographic assessment of papillary muscle rupture. Circulation 1996;94:10039.
2. Schiller NB, Foster E, Redberg RF. Transesophageal echocardiography in the evaluation of mitral regurgitation: the twentyfour signs of severe mitral regurgitation. Cardiol Clin 1993;11:
399 408.

PEDIATRIC ANESTHESIA

SOCIETY

FOR

PEDIATRIC ANESTHESIA

SECTION EDITOR
PETER J. DAVIS

An Evaluation of Bilateral Monitoring of Cerebral Oxygen

Saturation During Pediatric Cardiac Surgery
Barry D. Kussman, MBBCh*, David Wypij, PhD#, James A. DiNardo, MD*,
Jane Newburger, MD, MPH#, Richard A. Jonas, MD**, Jodi Bartlett, RN,
Ellen McGrath, RN, and Peter C. Laussen, MBBS*
Departments of *Anesthesiology, Perioperative, and Pain Medicine, Clinical Research Program, Cardiology and
Cardiovascular Surgery, Childrens Hospital Boston; Departments of Anesthesia, Cardiology, #Pediatrics, and
**Surgery, Harvard Medical School; and Department of Biostatistics, Harvard School of Public Health, Boston,
Massachusetts

Cerebral oximetry is a technique that enables monitoring of regional cerebral oxygenation during cardiac
surgery. In this study, we evaluated differences in
bi-hemispheric measurement of cerebral oxygen saturation using near-infrared spectroscopy in 62 infants
undergoing biventricular repair without aortic arch reconstruction. Left and right regional cerebral oxygen
saturation index (rSO2i) were recorded continuously
after the induction of anesthesia, and data were analyzed at 12 time points. Baseline rSO2i measurements

he reduction in mortality of infants with congenital heart disease (CHD) undergoing corrective
and palliative surgery has been accompanied by
the recognition that the survivors can be at risk for
long-term adverse neurologic sequelae, including impairment of fine motor performance and lifelong language and learning problems (1,2). In infants, brain
injury most often results from global hypoperfusion
because of alterations in cerebral hemodynamics and
metabolism during hypothermia and cardiopulmonary bypass (CPB), particularly low-flow (LF) CPB
and deep hypothermic circulatory arrest (DHCA) (3
9). The ability to measure and detect changes in cerebral oxygenation during pediatric cardiac surgery
may therefore allow for intervention strategies to improve late neurodevelopmental outcome. This notion

Supported, in part, by grants HL 063411 and RR 02172 from the

National Institutes of Health.
Accepted for publication May 5, 2005.
Address correspondence and reprint requests to Barry D. Kussman,
MBBCh, Department of Anesthesiology, Perioperative and Pain Medicine, Childrens Hospital, 300 Longwood Ave., Boston, MA 02115.
Address e-mail to barry.kussman@childrens.harvard.edu.
DOI: 10.1213/01.ANE.0000180205.85490.85

1294

Anesth Analg 2005;101:12941300

were left 65 13 and right 66 13 (P 0.17). Mean left

and right rSO2i measurements were similar (2 percentage points/absolute scale units) before, during,
and after cardiopulmonary bypass, irrespective of the
use of deep hypothermic circulatory arrest. Further longitudinal neurological outcome studies are required to
determine whether uni- or bi-hemispheric monitoring
is required in this patient population.
(Anesth Analg 2005;101:1294 1300)

has been presented in a recent review of multimodality neurological monitoring for congenital heart surgery (10).
Cerebral oximetry using near-infrared spectroscopy
(NIRS) is a developing technology for noninvasive
assessment of cerebral oxygenation (1114). The INVOS 5100 (Somanetics, Troy, MI) is the only cerebral
oximeter commercially available in the United States
and has been approved for use as a trend monitor of
cerebral oxygenation. The INVOS 5100 expresses cerebral oxygen saturation as the regional cerebral oxygen saturation index (rSO2i). The rSO2i is accordingly
a measure of the oxygen saturation of blood in the
gas-exchanging circulation (arterioles, capillaries, and
venules) of brain tissue. Bi-hemispheric measurement
of cerebral oxygen saturation is recommended by the
manufacturer, but correlation of the left and right
rSO2i in infants undergoing congenital cardiac surgery
without aortic arch reconstruction has not been reported. The aim of this study was to evaluate differences in bi-hemispheric measurement of cerebral oxygenation using NIRS during pediatric cardiac
surgery. The hypothesis was that the difference in left
and right cerebral oxygen saturation would be 10
percentage points/absolute scale units (%).
2005 by the International Anesthesia Research Society
0003-2999/05

ANESTH ANALG
2005;101:1294 1300

PEDIATRIC ANESTHESIA
KUSSMAN ET AL.
BILATERAL CEREBRAL OXYGENATION IN PEDIATRIC CARDIAC SURGERY

Methods
After IRB approval and parental informed consent,
cerebral oxygen saturation monitoring was performed
in neonates and infants undergoing hypothermic CPB
who had been enrolled in a prospective study to assess
neurodevelopmental outcome with differing hemodilution strategies (hematocrit 25% versus 35%). Eligibility criteria included scheduled surgery at or before
9 mo of age and repair of either D-transposition of the
great arteries (D-TGA), ventricular septal defect
(VSD), complete common atrioventricular canal defect, tetralogy of Fallot (TOF), or truncus arteriosus.
Exclusion criteria included a birth weight 2.3 kg, a
recognizable syndrome of congenital anomalies, an
extracardiac anomaly of more than minor severity,
previous cardiac surgery, or cardiovascular anomalies
requiring aortic arch reconstruction or additional open
procedures.
Anesthetic technique was not specifically controlled
but was conducted according to our institutional practice. Large-dose opioid anesthesia (fentanyl 100 g/
kg) was supplemented with midazolam and isoflurane, as tolerated, and neuromuscular blockade was
achieved with pancuronium. The head was turned to
just off the midline to prevent pressure or movement
on the endotracheal tube by the surgical team while
avoiding the possible effects of extremes of lateral
head position on cerebral blood flow and venous
drainage. Standard monitoring was used, including a
radial or femoral artery catheter for measurement of
systemic arterial blood pressure and intermittent
blood sampling, as well as tympanic, esophageal, and
rectal temperatures.
Surface cooling was initiated with low ambient
room temperature, a cooling mattress, and ice packs
applied to the head. Core cooling was begun with the
initiation of CPB. The bypass circuit was primed with
Plasmalyte-A (Baxter, Deerfield, IL) and blood to
achieve the desired hematocrit during cooling. A nonpulsatile roller pump with a membrane oxygenator (D
901 Lilliput 1 Open System; Dideco, Mirandola, Italy)
was used. Standard pump flow rates of 150
200
mL kg1 min1
for
full
flow
and
50 mL kg1 min1 for LF were used. Perfusion pressure was determined by CPB flow rate, with pressures
in the range of 30 40 mm Hg being acceptable at full
flow. When the rectal temperature reached 18C or
lower, and after at least 20 min of cooling, DHCA or
continuous LF bypass was begun. A pH-stat acid-base
management strategy was used in all patients during
hypothermia. Methylprednisolone (30 mg/kg), phentolamine (0.2 mg/kg), and furosemide (0.25 mg/kg)
were given at the initiation of CPB to all patients. At
the onset of rewarming, mannitol (0.5 g/kg) and
phentolamine (0.2 mg/kg) were given to all patients.
Patients were rewarmed for at least 30 min to a rectal

1295

temperature of 34C; hemofiltration was performed

during rewarming.
Cerebral oxygen saturation was measured with the
INVOS 5100. Cerebral oximetry relies on the penetration of the scalp, skull, and cranial contents by nearinfrared light and measures the ratio of oxyhemoglobin to total hemoglobin to derive a hemoglobin
oxygen saturation. The INVOS 5100 is a continuous
wave spectrometer and uses two wavelengths of nearinfrared light (730 and 810 nm) to measure the ratio of
oxyhemoglobin to total hemoglobin. The sensor contains a light-emitting diode (LED) and 2 detectors
located 30 and 40 mm from the LED, allowing removal
of the extracranial contribution of absorbed and scattered light by the application of a subtraction algorithm. Because the volume of blood is predominantly
venous, it reflects the balance between oxygen supply
and demand in the brain. The scale unit for rSO2i is
percent (%) but is referred to as an index because it is
not easily validated in vivo, is regional in nature, and
may not reflect events occurring at a distance from the
sensor. After the induction of anesthesia, one Pediatric
SomaSensor (Somanetics) was placed on the right and
one on the left forehead, according to the manufacturers guidelines. After an accommodation period, data
collection was begun and downloaded to a storage
disk every 10 s throughout the case for further analysis. The INVOS 5100 collects light absorption data 15
times a second, and when 50 samples have been collected, they are averaged to determine a new rSO2i
value, which is displayed on the screen (approximately every 4 s).
Intraoperative data were collected and analyzed at
the following time points: after induction, initiation of
CPB, 10 min after start of cooling, onset of LF, onset of
DHCA, resumption of LF, start of rewarming, 10 min
after start of rewarming, warm flow (35C), immediately off CPB, 60 min after CPB, and 6 h after CPB.
Longer term (1 yr) neurodevelopmental outcome and
assessment of brain injury with magnetic resonance
imaging (MRI) is continuing in the patients enrolled in
this study, and because we are still blinded to randomization, we could not analyze the rSO2i values in respect to hematocrit strategy.
Data analyses were performed with SAS Version
9.00 (SAS Institute, Inc, Cary, NC). Descriptive statistics for demographic, intraoperative, physiologic, and
rSO2i variables were based on the mean and standard
deviations, medians and ranges, or frequencies. Comparisons of left versus right rSO2i values were based
on paired t-tests and Bland-Altman analyses (15).
Physiologic and saturation data are expressed as mean
sd. The primary outcome variable was the mean
difference in all rSO2i values between cerebral hemispheres at each time point. Although the mean difference in left and right rSO2i is unknown, we chose a
mean difference of 10% (absolute scale units) because

1296

PEDIATRIC ANESTHESIA KUSSMAN ET AL.

BILATERAL CEREBRAL OXYGENATION IN PEDIATRIC CARDIAC SURGERY

more than 10 scale unit left-right rSO2i differences

have been reported as highly unusual, and decreases
(relative) of 20% from baseline have been associated
with possible neurologic symptoms (16 18). Because
this is an observational secondary analysis, a power
calculation was not performed in advance. Nevertheless, a sample size of 60 paired differences, assuming
a common standard deviation of 16 and a correlation
of 0.75 between paired left and right measurements,
provides 92% power to detect mean differences of 5%
between left and right measurements.

Results
Sixty-two infants had NIRS monitoring of cerebral
oxygenation. In two infants, there was unilateral failure of saturation measurements without replacement
of the sensor so that bilateral cerebral oxygen saturation data in 60 patients were suitable for analysis.
Demographic and intraoperative data for these 60
infants are shown in Table 1. After the induction of
deep hypothermia, 35 infants (58%) underwent continuous LF and 25 infants (42%) underwent DHCA.
Physiologic data are shown in Table 2.
Regional cerebral oxygen saturation at each time
point is shown in Table 3. The rSO2i increased during
cooling, decreased during the period of DHCA (when
occurring), and was restored to baseline levels by 6 h
after CPB. Left and right rSO2i were similar before,
during, and after CPB (mean difference, 2 absolute
percentage points/scale units at each time point).
These findings were the same regardless of whether
DHCA was used. The nadir rSO2i at the end of DHCA
(median, 26; range, 159 min) was left 64 17 and
right 65 14 (P 0.97). For those infants who had
continuous LF, the nadir rSO2i at the end of the LF
CPB period was left 81 13 and right 82 13 (P
0.67).
Graphical displays of individual subject data
showed similar trends between left and right rSO2i
measurements (results not shown). Representative
Bland-Altman plots assessing agreement between left
and right rSO2i measurements at postinduction and
10 min after cooling suggest that there is variability
within patients (Fig. 1). Of 634 paired left versus right
measurements on subjects, 77 (12%) were more than
10 percentage points apart and were evenly divided
between left and right. In 40 measurements, the left
side was 10 percentage points higher than the right
side, and in 37 measurements, the right side was 10
percentage points higher (P 0.82). These 77 measurements came from 28 of 60 (47%) patients; of these
measurements, 1 subject had 12, 1 had 7, 2 had 5, 4 had
4, 2 had 3, 8 had 2, and 10 had 1. These differences 10
scale points were distributed across many perfusion
phases, being more frequent from 10 min after start of
cooling on CPB to immediately off CPB.

ANESTH ANALG
2005;101:1294 1300

Table 1. Demographic and Intraoperative Data (n 60)

Median
(range)
Age at surgery (days)
Weight (kg)
Cardiac diagnoses n (%)
TGA/IVS
TGA/VSD
VSD
CAVC
TOF
TOF/PA
TA
Total CPB time (min)
Lowest tympanic temperature (C)
Minutes of DHCA (n 25)

51.5 (2263)
4.4 (2.67.3)
11 (18)
13 (22)
15 (25)
2 (3)
16 (27)
2 (3)
1 (2)
91 (50196)
22.5 (13.629.3)
26 (159)

TGA Transposition of Great Arteries; IVS intact ventricular septum;

VSD ventricular septal defect; CAVC complete atrioventricular canal;
TOF tetralogy of Fallot; PA pulmonary atresia; TA truncus arteriosus;
CPB cardiopulmonary bypass; DHCA deep hypothermic circulatory
arrest.

There were six patients who were more than 10

percentage points apart at baseline. For the 71 paired
left versus right measurements that were more than 10
percentage points apart after baseline, the left and
right rSO2i were compared with their respective
postinduction value. In six (8%) of the paired measurements, corresponding to five patients, either the
left or the right rSO2i had decreased more than 20%
(relative) less than baseline. The time points at which
this occurred were on CPB (1), resume LF (3), at start
rewarming (1), and 10 min after start rewarming (1).
In 27 (38%) paired measurements in 11 patients, both
the left and right rSO2i were within 20% (relative) of
baseline. Of the remaining 44 paired measurements, in
38 measurements (86%), from 19 patients, the left,
right, or both rSO2i had increased more than 20%
(relative) above baseline.
Of the 634 paired left versus right measurements on
subjects, 13 (2%) were more than 20 percentage points
apart. This included four measurements for which the
left side was 20 percentage points higher than the
right side and nine measurements for which the right
side was 20 percentage points higher (P 0.27).
These 13 measurements came from 4 of 60 (7%) subjects; of these measurements, 1 subject had 9, 1 had 2,
and 2 had 1 each.
The largest difference in any patient was 31 percentage points at 2 time pointsresume LF (L 40; r
71) and start rewarming (L 51, r 82)i.e., immediately after DHCA. This neonate, undergoing repair
of D-TGA with VSD and whose postinduction rSO2i
values were left 55% and right 79% (arterial oxygen
saturation, 95%), accounted for 12 of the 77 (16%)
paired left versus right measurements that were more
than 10 percentage points apart and 9 of the 13 (69%)

ANESTH ANALG
2005;101:1294 1300

PEDIATRIC ANESTHESIA
KUSSMAN ET AL.
BILATERAL CEREBRAL OXYGENATION IN PEDIATRIC CARDIAC SURGERY

1297

Table 2. Physiologic Data (n 60)

MAP
Hematocrit
(mm Hg)
(%)
Postinduction
On CPB
10 min after cooling
Onset of LF
Onset of DHCA (n 25)
Resume LF (n 25)
Start rewarming (SW)
10 min after SW
Warm flow (35C)
Off CPB
60 min post-CPB
6 h post-CPB

58.2 10.9
42.9 10.8
34.5 9.3
29.4 11.4
11.3 8.4
16.6 8.2
35.8 12.9
37.7 9.6
47.1 10.9
49.0 8.9
60.9 8.3
61.1 11.2

Pao2
(mm Hg)

Paco2
(mm Hg)

pH

35.1 6.5

155.3 121.2

42.2 9.9

7.43 0.08

1.4 0.9

29.0 5.3
28.9 5.5

736.7 68.3
730.5 78.2

70.5 13.0
71.5 16.6

7.25 0.06
7.24 0.07

2.0 0.7
2.1 0.8

32.6 3.5
34.5 2.8

481.4 132.8
463.4 114.4

45.1 11.1
37.0 6.3

7.37 0.08
7.44 0.05

2.7 1.2
2.8 1.3

36.9 4.4

263.9 148.2
125.8 49.2

43.7 8.8
39.6 6.2

7.42 0.07
7.44 0.06

2.8 1.6

Lactate
Temperature
(mmol/L)
(C)a
32.9 1.3
22.4 4.5
16.6 1.4
17.3 1.5
21.8 4.5
28.2 3.7
34.5 1.3
35.8 0.9
35.1 1.2

Results expressed as mean sd.

a
Temperature is tympanic. Pao2, Paco2, and pH values are uncorrected (37C) for temperature.
CPB cardiopulmonary bypass; LF low flow bypass; DHCA deep hypothermic circulatory arrest; SW start rewarming; MAP mean arterial blood
pressure.

Table 3. Regional Cerebral Oxygen Saturation Index

(rSO2i) (%) (n 60)
Perfusion Phase

Left

Right

P-value

Postinduction
On CPB
10 min after cooling
Onset of LF
Onset of DHCA (n 25)
Resume LF (n 25)
Start rewarming (SW)
10 min after SW
Warm flow (35C)
Off CPB
60 min post-CPB
6 h post-CPB

65 13
62 13
84 11
84 11
89 10
64 17
81 13
78 16
74 14
72 14
73 12
63 9

66 13
64 13
84 10
84 11
88 10
65 14
82 13
78 16
75 14
73 14
73 13
63 9

0.17
0.03
0.69
0.99
0.75
0.97
0.67
0.89
0.78
0.40
0.65
0.74

Results expressed as mean sd.

CPB cardiopulmonary bypass; LF low flow bypass; DHCA deep
hypothermic circulatory arrest; SW start rewarming.

paired measurements that were more than 20 percentage points apart. There were no postoperative neurologic complications in this patient, and a brain MRI
study, with spectroscopy, performed at 16 mo of age,
was normal. All patients survived to discharge, and
there were no clinical signs of neurologic injury.

Discussion
The main finding of this study is that rSO2i measurements from the left and right cerebral hemispheres
were similar in neonates and infants undergoing
biventricular repair without aortic arch reconstruction. This finding is in agreement with a study using
the INVOS 5100 to compare simultaneous right and
left rSO2i in 30 children under general anesthesia, in
which a mean difference of 1.3 2.8 (P 0.51) was
found (19).
Despite the availability of NIRS for several years, it
has not been widely used during congenital cardiac

Figure 1. Assessing agreement between bi-hemispheric measurement of cerebral oxygenation using near-infrared spectroscopy
(NIRS) at postinduction (panel A) and 10 min after cooling (panel
B). Scatter plots of the difference between left and right regional
cerebral oxygen saturation index (rSO2i) measurements versus the
average rSO2i measurements are presented. The solid line presents
the mean differences, and the dashed lines represent the estimated
limits of agreement, plotted as the mean differences 1.96 sd.

1298

PEDIATRIC ANESTHESIA KUSSMAN ET AL.

BILATERAL CEREBRAL OXYGENATION IN PEDIATRIC CARDIAC SURGERY

surgery. Explanations include limitations of the technology, insufficient outcome data to determine sensitivity and specificity, and cost of the instrumentation.
NIRS instruments differ in the technology and algorithm used to measure hemoglobin saturation, and
because large inter- and intraindividual differences in
saturation have been found among instruments, critical values of cerebral oxygenation determined by one
instrument cannot be applied to values measured by
another (19 22).
Defining a normative range for cerebral oxygen saturation in pediatric patients has been problematic. In
adults, mean rSO2i values of 71% 6% and 67%
10% have been reported for healthy volunteers and
adult cardiac surgical patients, respectively (16,23).
Using a frequency domain cerebral oximeter (the INVOS 5100 is a continuous wave oximeter), Kurth et al.
(22) found cerebral O2 saturation (ScO2) values of 68%
10% in healthy children. Mean rSO2i values of 78%
8% were reported for healthy children breathing air
at an altitude of 1610 m (24). Large interpatient variance of cerebral oxygen saturation has been shown for
neonates without CHD breathing room air (25). In
CHD, cerebral oxygenation has been shown to vary
with anatomy and arterial saturation (22). ScO2 values
similar to healthy subjects (68% 10%) were found in
patients with VSD, aortic coarctation, and single ventricle after Fontan operation, whereas ScO2 was significantly decreased in patients with patent ductus arteriosus (53% 8%), TOF (57% 12%), hypoplastic left
heart syndrome (46% 8%), pulmonary atresia (38%
6%), and single ventricle after aortopulmonary
shunt (50% 7%) or bidirectional Glenn operation
(43% 6%). The administration of sedative medication or general anesthesia, which typically causes a
decrease in cerebral oxygen consumption, is a confounding factor when attempting to establish normative ranges in children.
Baseline rSO2i measurements in the present study
were symmetrical. Although numerically similar to
those reported by Kurth et al. (22), rSO2i and ScO2 do
not necessarily represent the same level of cerebral
oxygenation. As demonstrated by the standard deviation, both at baseline and thereafter, there is a wide
range with interindividual variability of cerebral oxygenation. Baseline (or developing) cerebral oxygen
saturation asymmetry has been attributed to carotid or
intracranial arterial stenosis, intracranial spaceoccupying lesion, old infarction, extracranial lesions,
such as hemangioma, excessive frontal sinus fluid, or
a skull defect, subclavian steal, and interference from
an infrared-emitting device or ambient light (16). During cardiac surgery, cerebral oxygen saturation asymmetry may also occur after complications related to
aortic cannulation (26,27). Obstruction of the superior
vena cava during venous cannulation or on bypass

ANESTH ANALG
2005;101:1294 1300

can be associated with alterations in cerebral oxygen

saturation (28 30).
Although multiple factors influence neurologic outcome during cardiac surgery, mechanisms of cerebral
injury are quite different between adults and young
children. Risk factors for cerebrovascular disease in
children differ significantly from those in adults, in
whom arteriosclerosis is the leading cause (31). Fatty
streaks are present in infants and children, but they do
not significantly narrow the lumina of blood vessels
(32). Fibrous or atheromatous plaques that result in
narrowing of the lumen usually begin in the late teens
and early twenties (33). The most common identifiable
risk factor for childhood ischemic stroke is complex
CHD (31), including related factors of coagulation disorders, dehydration, and infection. Other causes of
childhood ischemic stroke include vasculitis, dissection, cancer, metabolic disorders, moyamoya disease,
sickle cell anemia, and perinatal complications. Variable developmental patterns in the Circle of Willis are
unlikely to account for large differences between left
and right cerebral oxygen saturation. An anatomical
study of 350 normal brains found a normal configuration of the circle of Willis in 52.3%, with the anterior
and posterior halves of the polygon varying in respect
to the type of anomaly (34). The most frequent anomaly (27.4%) was a reduced size of one of the component vessels. In no case was there absence of an anatomical connection between the right and left anterior
cerebral artery. Accessory vessels (duplications and
triplications) were found in 18.9%, the majority in the
anterior portion of the circle. A truly incomplete circle
was found in 0.6% with absence of the left posterior
communicating artery. No abnormalities of the middle cerebral artery were reported in this series.
It is important to stress that the INVOS 5100 is
approved by the Food and Drug Administration as a
trend monitor of cerebral oxygenation. Although studies in adults suggest that cerebral ischemia may occur
with (a) a decrease in rSO2i of more than 10 percentage
points/absolute scale units, (b) more than 20% decrease (relative) from baseline, or (c) an absolute value
of 50% (18,3537), the critical rSO2i or degree of
relative change from an individual established baseline in children that results in cerebral hypoxicischemic damage is uncertain. Cerebral oximetry is
thus a trend monitor of greatest value in situations in
which cerebral oxygen saturation could dangerously
change when changes in systemic hemodynamics and
oxygenation would not predict that change (38). The
similarity of left and right rSO2i measurements in the
present study raises the question of whether unilateral
monitoring of cerebral oxygen saturation is sufficient
for many patients undergoing infant cardiac surgery.
Cerebral oximetry (unilateral) as a component of multimodality neurophysiologic monitoring has been
shown to be potentially beneficial in pediatric cardiac

ANESTH ANALG
2005;101:1294 1300

PEDIATRIC ANESTHESIA
KUSSMAN ET AL.
BILATERAL CEREBRAL OXYGENATION IN PEDIATRIC CARDIAC SURGERY

surgery (17). In a prospective cohort study, an interventional algorithm was used during surgery to detect and
correct specific deficiencies in cerebral oxygenation or
perfusion, as measured by cerebral oximetry, transcranial Doppler sonography, and electroencephalography.
Limitations of this study included nonrandomization to
an intervention versus nonintervention strategy, no
longer-term evaluation of neurologic outcome, and no
determination of absolute level of cerebral oxygen saturation or transcranial Doppler flow, at which their needs
to be an intervention. The restricted space on the neonatal and infant forehead may only allow for unilateral
placement of the Pediatric SomaSensor during multimodality neurophysiologic monitoring. Although bilateral
monitoring is recommended during aortic arch reconstruction using regional LF perfusion (10) or in situations
where venous drainage is altered, a recent review suggested that bilateral monitoring was not required when
both carotid arteries were perfused under normal circumstances (39). In our study, six patients had baseline
rSO2i asymmetry more than 10 percentage points, but in
only one patient did this difference persist throughout
the procedure and up to 18 hours after surgery (after
which rSO2i monitoring was discontinued). Despite this,
the patient did not have postoperative neurologic complications and had a normal follow-up brain MRI study.
In patients with left-right differences more than 10%,
these were more frequent from 10 minutes after start of
cooling on CPB to off bypass, i.e., at the higher end of the
rSO2i scale where differences between sides would presumably not have the same adverse implications as at
the lower end of the scale. Comparing these asymmetric
pairs to their respective baseline values, in only 8% of
these patients did the left or right rSO2i decrease more
than 20% from baseline; in the other 92% of patients, the
change in rSO2i was to increase more than the baseline
value.
When advocating routine use of cerebral oximetry,
the significant cost of the disposable sensors may limit
widespread introduction of this form of regional monitoring. In the United States, the retail price of the
INVOS 5100B is $25,000, and the disposable Somasensors are between $110 and $140, with the final cost
being determined by the contract between the hospital
and Somanetics (information from Somanetics). Although an estimated break-even analysis may justify a
hospital expenditure for neurophysiologic monitoring
of $2,142 per case (17), many practitioners are hesitant
to incur additional intraoperative costs for monitoring
without large-scale prospective clinical trials showing
that it improves outcome. Although the potential of
NIRS for preventing or predicting neurologic injury
during pediatric cardiac surgery can only be established by large scale, prospective clinical trials, reducing the cost of cerebral oximetry by using a single
sensor, may promote more widespread use of the
technology.

1299

In conclusion, bi-hemispheric measurements of regional cerebral oxygen saturation using NIRS were
similar in neonates and infants undergoing biventricular repair without aortic arch reconstruction. Further
longitudinal neurological outcome studies are required to determine whether uni- or bi-hemispheric
monitoring is required during pediatric cardiac
surgery.
We thank Gene Walters for monitoring and data collection, Ludmila
Kyn for database and statistical programming, Donna Donati and
Donna Duva for data management, and Kathy Alexander for project
coordination.

References
1. Ferry PC. Neurologic sequelae of open-heart surgery in
children: An irritating question. Am J Dis Child 1990;144:
369 73.
2. Bellinger DC, Wypij D, Kuban KC, et al. Developmental and
neurological status of children at 4 years of age after heart
surgery with hypothermic circulatory arrest or low-flow cardiopulmonary bypass. Circulation 1999;100:526 32.
3. Nevin M, Colchester AC, Adams S, Pepper JR. Evidence for
involvement of hypocapnia and hypoperfusion in aetiology of
neurological deficit after cardiopulmonary bypass. Lancet 1987;
2:14935.
4. DeLeon S, Ilbawi M, Arcilla R, et al. Choreoathetosis after deep
hypothermia without circulatory arrest. Ann Thorac Surg 1990;
50:714 9.
5. Mujsce DJ, Towfighi J, Vannucci RC. Physiologic and neuropathologic aspects of hypothermic circulatory arrest in newborn
dogs. Pediatr Res 1990;28:354 60.
6. Greeley WJ, Bracey VA, Ungerleider RM, et al. Recovery of
cerebral metabolism and mitochondrial oxidation state is delayed after hypothermic circulatory arrest. Circulation 1991;84:
III400 6.
7. Shaw PJ. The neurologic sequelae of cardiopulmonary bypass:
the Newcastle experience. In: Smith P, Taylor K, eds. Cardiac
surgery and the brain. London: Edward Arnold, 1993:24 33.
8. Jonas RA, Newburger JW, Volpe JJ. Brain injury and pediatric
cardiac surgery. Boston: Butterworth-Heinemann, 1995.
9. Sakamoto T, Hatsuoka S, Stock UA, et al. Prediction of safe
duration of hypothermic circulatory arrest by near-infrared
spectroscopy. J Thorac Cardiovasc Surg 2001;122:339 50.
10. Andropoulos DB, Diaz LK, Fraser CD Jr, et al. Is bilateral
monitoring of cerebral oxygen saturation necessary during neonatal aortic arch reconstruction? Anesth Analg 2004;98:126772.
11. Jobsis FF. Noninvasive, infrared monitoring of cerebral and
myocardial oxygen sufficiency and circulatory parameters. Science 1977;198:1264 7.
12. Pollard V, Prough DS, DeMelo AE, et al. Validation in volunteers of a near-infrared spectroscope for monitoring brain oxygenation in vivo. Anesth Analg 1996;82:269 77.
13. Madsen PL, Secher NH. Near-infrared oximetry of the brain.
Prog Neurobiol 1999;58:541 60.
14. Fraser CD Jr, Andropoulos DB. Neurologic monitoring for special cardiopulmonary bypass techniques. Semin Thorac Cardiovasc Surg Pediatr Card Surg Annu 2004;7:12532.
15. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet
1986;1:30710.
16. Edmonds HL Jr, Ganzel BL, Austin EH 3rd. Cerebral oximetry
for cardiac and vascular surgery. Semin Cardiothorac Vasc
Anesth 2004;8:147 66.

1300

PEDIATRIC ANESTHESIA KUSSMAN ET AL.

BILATERAL CEREBRAL OXYGENATION IN PEDIATRIC CARDIAC SURGERY

17. Austin EH 3rd, Edmonds HL Jr, Auden SM, et al. Benefit of

neurophysiologic monitoring for pediatric cardiac surgery.
J Thorac Cardiovasc Surg 1997;114:70715, 17; discussion 15 6.
18. Samra SK, Dy EA, Welch K, et al. Evaluation of a cerebral
oximeter as a monitor of cerebral ischemia during carotid endarterectomy. Anesthesiology 2000;93:964 70.
19. Dullenkopf A, Frey B, Baenziger O, et al. Measurement of
cerebral oxygenation state in anaesthetized children using the
INVOS 5100 cerebral oximeter. Paediatr Anaesth 2003;13:
384 91.
20. Yoshitani K, Kawaguchi M, Tatsumi K, et al. A comparison of
the INVOS 4100 and the NIRO 300 near-infrared spectrophotometers. Anesth Analg 2002;94:586 90.
21. Thavasothy M, Broadhead M, Elwell C, et al. A comparison of
cerebral oxygenation as measured by the NIRO 300 and the
INVOS 5100 near-infrared spectrophotometers. Anaesthesia
2002;57:999 1006.
22. Kurth CD, Steven JL, Montenegro LM, et al. Cerebral oxygen
saturation before congenital heart surgery. Ann Thorac Surg
2001;72:18792.
23. Kim MB, Ward DS, Cartwright CR, et al. Estimation of jugular
venous O2 saturation from cerebral oximetry or arterial O2
saturation during isocapnic hypoxia. J Clin Monit Comput 2000;
16:1919.
24. Yaron M, Niermeyer S, Lindgren KN, et al. Physiologic response
to moderate altitude exposure among infants and young children. High Alt Med Biol 2003;4:539.
25. Menke J, Voss U, Moller G, Jorch G. Reproducibility of cerebral
near infrared spectroscopy in neonates. Biol Neonate 2003;83:
6 11.
26. Bar-Yosef S, Sanders EG, Grocott HP. Asymmetric cerebral
near-infrared oximetric measurements during cardiac surgery.
J Cardiothorac Vasc Anesth 2003;17:773 4.
27. Janelle GM, Mnookin S, Gravenstein N, et al. Unilateral cerebral
oxygen desaturation during emergent repair of a DeBakey type
1 aortic dissection: potential aversion of a major catastrophe.
Anesthesiology 2002;96:12635.

ANESTH ANALG
2005;101:1294 1300

28. Daubeney PE, Smith DC, Pilkington SN, et al. Cerebral oxygenation during paediatric cardiac surgery: identification of vulnerable periods using near infrared spectroscopy. Eur J Cardiothorac Surg 1998;13:370 7.
29. Sakamoto T, Duebener LF, Laussen PC, Jonas RA. Cerebral
ischemia caused by obstructed superior vena cava cannula is
detected by near-infrared spectroscopy. J Cardiothorac Vasc
Anesth 2004;18:293303.
30. Ing RJ, Lawson DS, Jaggers J, et al. Detection of unintentional
partial superior vena cava occlusion during a bidirectional cavopulmonary anastomosis. J Cardiothorac Vasc Anesth 2004;18:
472 4.
31. deVeber G. Cerebrovascular disease in children. In: Swaiman
KF, Ashwal S, eds. Pediatric neurology. 3rd ed. St. Louis:
Mosby, 1999:1099 124.
32. Strong JP. The natural history of atherosclerosis in childhood.
Ann N Y Acad Sci 1991;623:9 15.
33. Strong JP, McGill HC Jr. The pediatric aspects of atherosclerosis.
J Atheroscler Res 1969;9:251 65.
34. Alpers BJ, Berry RG, Paddison RM. Anatomical studies of the
circle of Willis in normal brain. AMA Arch Neurol Psychiatry
1959;81:409 18.
35. Cho H, Nemoto EM, Yonas H, et al. Cerebral monitoring by
means of oximetry and somatosensory evoked potentials during
carotid endarterectomy. J Neurosurg 1998;89:533 8.
36. Singer I, Dawn B, Edmonds HL Jr, et al. Patch migration: a
serious late complication of thoracotomy lead systems. Pacing
Clin Electrophysiol 1998;21:2616 20.
37. Singer I, Edmonds HL. Tissue oximetry for the diagnosis of
neurally mediated syncope. Pacing Clin Electrophysiol 2000;23:
2006 9.
38. Pollard V, Prough DS. Cerebral near-infrared spectroscopy: a
plea for modest expectations. Anesth Analg 1996;83:673 4.
39. Andropoulos DB, Stayer SA, Diaz LK, Ramamoorthy C. Neurological monitoring for congenital heart surgery. Anesth Analg
2004;99:136575.

The Optimal Depth of Central Venous Catheter for Infants

Less Than 5 kg
Jin-Hee Kim, MD*, Chong-Sung Kim, MD, Jae-Hyun Bahk, MD, Kyung Joon Cha,
Young-Sun Park, PhD, Young-Tae Jeon, MD, and Sung-Hee Han, MD

PhD,

*Department of Anesthesiology, Seoul National University Bundang Hospital; Department of Anesthesiology, Seoul
National University Hospital; Laboratory of Statistical Information Analysis, Hanyang University, College of Natural
Sciences, Seoul, Korea

To avoid fatal complications of central venous catheterization such as cardiac tamponade, the tip of the central
venous catheter (CVC) should be placed outside of the
cardiac chamber. To suggest a guideline for a proper
depth of CVC in infants, we measured the distance
from the skin puncture site to the junction between superior vena cava and right atrium (SVC-RA junction) by
using transesophageal echocardiography (TEE).
Fifty infants less than 5 kg undergoing surgery for
congenital heart disease were enrolled in this prospective study. After the induction of general anesthesia, CVC was inserted via the right subclavian

lacement of a central venous catheter (CVC) is

frequently needed during pediatric anesthesia
and intensive care. Its malposition not only
causes faulty central venous pressure measurement
but also causes fatal complications such as thrombosis,
arrhythmia, cardiac perforation, and cardiac tamponade (1 6). To avoid these fatal complications, the CVC
is recommended to be placed in the distal superior
vena cava (SVC), outside of the cardiac chamber (4 6).
Although malposition-related complications are more
common and more serious in infants (79), there are no
evidence-based guidelines for the proper depth of the
CVC for small infants. The subclavian vein, along with
the internal jugular vein, is one of the most commonly
used routes of the CVC for pediatric patients (9,10).
Thus, to suggest a guideline for the proper length of a
CVC inserted through the subclavian vein in infants, we
measured the distance from the skin puncture site to the
SVC-right atrium (RA) junction by using transesophageal echocardiography (TEE).
Accepted for publication May 25, 2005.
Address correspondence and reprint requests to Sung- Hee Han,
MD, Department of Anesthesiology and Pain Medicine, 300
Gumi-Dong, Bundang-Gu, SeongnamSi, Gyoenggi-Do, 463707,
Korea. Address e-mail to noninvasive@snubh.org or noninvasive@
hanmail.net.
DOI: 10.1213/01.ANE.0000180997.72988.FE
2005 by the International Anesthesia Research Society
0003-2999/05

vein. After the tip of the CVC was placed at the

SVC-RA junction using TEE guidance, the length of
the CVC inserted beneath the skin was measured.
The measured distance had a high correlation with
the patients height, weight, and age (r 0.88, 0.76,
and 0.64, respectively). In infants smaller than 5 kg,
the following guideline can avoid intraatrial placement of the CVC: a depth between 40 and 45 mm for
infants 2.0 3.0 kg in weight, 4550 mm for those 3.0
3.9 kg, and 50 55 mm for those more than 4.0 kg.
(Anesth Analg 2005;101:13013)

Methods
After obtaining hospital IRB approval and informed
parental consent, patients less than 5 kg in weight who
were scheduled for elective repair of ventricular septal
defect or atrial septal defect were enrolled in this
prospective study. Patients with left-sided SVC or any
extra-cardiac vascular abnormality were excluded. After the induction of general anesthesia, central venous
catheterization was performed using the Seldinger
technique. All the catheterizations were performed by
one of four anesthesiologist staff members with more
than 1 year of experience in pediatric anesthesia. Patients were positioned in slight head down position
with a rolled towel placed transversely under the
shoulder. An infraclavicular approach as described in
our previous report (11), was performed on the right
subclavian vein (RSCV). Each anesthesiologist was allowed to attempt to advance the CVC only twice. If
two anesthesiologists failed to advance the CVC into
the SCV-RA junction, another vein was chosen for
central catheterization and the patient was excluded
from the study. While one of the anesthesiologists
performed the CVC cannulation, another anesthesiologist observed the SVC-RA junction using TEE. After the
tip of the CVC was located at the SCV-RA junction,
defined as the superior border of crista terminalis in
bicaval view, the length of the CVC beneath the skin was
Anesth Analg 2005;101:13013

1301

1302

PEDIATRIC ANESTHESIA KIM ET AL.

OPTIMAL DEPTH OF CENTRAL VENOUS CATHETER FOR INFANTS

Table 1. Demographic Patient Characteristics

Age (mo)
Sex (M/F)
Height (cm)
Weight (kg)
Dx (ASD/VSD/AVSD)

2.1 1.3 (0.25.0)

26/24
54.3 4.9 (4263)
3.8 0.8 (2.24.9)
18/19/13

Values are mean sd (range) or number of patients.

Dx diagnosis; ASD atrail septal defect; VSD ventricular septal
defect; AVSD atrioventricular septal defect.

measured using the depth indicator on the CVC and an

aseptic paper ruler. When the tip of the CVC reached the
SCV-RA junction, the length of the CVC beneath the skin
was considered to be equal to the distance from the skin
puncture site to the SVC-RA junction.
All data are expressed as mean sd or median
(range). Plots of distance from the skin to the SVC-RA
junction versus age, weight, or height were made and
simple linear regression analysis was performed by
the least-squares method. Multiple linear regression
analysis was performed in standard and forward selection to identify independent factors affecting the
distance from the skin to the SVC-RA junction. Variables with the value of P 0.1 from correlation analysis were entered into regression analysis. Statistical
significance was taken as P 0.05. All statistical analyses were performed using the Statistical Package for
Social Sciences (SPSS Windows version 11.0; SPSS Inc.,
Chicago, IL).

Results
Sixty infants were initially enrolled. Ten patients
were excluded during the study because of failure
in subclavian vein puncture (n 6), and advancing
the tip of the CVC to the SVC-RA junction (n 4).
As a result, only 50 patients data were analyzed.
Patient demographic characteristics of the patients
are in Table 1.

ANESTH ANALG
2005;101:13013

The distance from the skin puncture site to the

SVC-RA junction (54.1 4.7; 41 62 mm) showed a
high correlation with the patients height, weight, and
age (r 0.88, 0.76, and 0.64, respectively). Fig. 1 shows
the scatter plot of the distance from the skin puncture
site to the SVC-RA junction versus age, weight, or
height. In standard forward multiple regression analysis, the distance from the skin puncture site to the
SVC-RA junction could be predicted from height and
weight as the formula: distance from skin to SVC-RA
junction (mm) 11.6 (0.70 height in cm) (1.14
weight in kg) (r2 0.81; P 0.0001).
Based on these results, we made simple and practical
guidelines (Table 2). When our guidelines were applied
to the present data, in 49 patients (98%), the CVC was
placed in the SVC above the RA. In the one other patient,
the tip of CVC was placed 1 mm inside the RA.

Discussion
Central venous catheterization can be performed
through various veins, such as the femoral vein (12),
internal jugular vein (13), and brachial vein (14). The
subclavian vein is one of the most frequently used
central venous routes in pediatric patients. The subclavian veins skin puncture site is less likely to
become infected than the other veins puncture sites
(15), and the patients are free to move their arms
and heads (10). The RSCV is preferred to the left
subclavian vein because of the higher risk of chylothorax (16) or vascular perforation (17) with the eft
subclavian vein.
The proper depth of a right subclavian catheter is
between 16 19 cm in adult patients (16,18,19). However, in pediatric patients, commonly used equations or guidelines are rare (20,21). A guideline suggested by Andropoulos et al. (20) is comparable to
ours. They suggested a depth of 4 cm for patients
less than 3 kg and 5 cm for patients between 35 kg.

Figure 1. The plots of distance versus age (a), weight (b), and height (c). a) Age (month): distance 49.2 (2.3 age); r2 0.40, P 0.1. b)
Weight (kg): distance 35.8 (4.7 wt); r2 0.58, P 0.001. c) Height (cm): distance 8.5 (0.8 ht); r2 0.77, P 0.0001. Distance
the distance from skin to superior vena cava and right atrial junction.

ANESTH ANALG
2005;101:13013

PEDIATRIC ANESTHESIA
KIM ET AL.
OPTIMAL DEPTH OF CENTRAL VENOUS CATHETER FOR INFANTS

Table 2. Suggested Depth of CVC Through RSCV

Weight
(kg)

Height
(cm)

Depth of
CVC (mm)

2.02.9
3.03.9
4.04.9

40.049.9
50.059.9
55.064.9

4045
4550
5055

CVC central venous catheter; RSCV right subclavian vein.

When their guidelines are applied to our results, in 30%

(15 of 50) of the patients, the predicted depths were
shallower than the measured depth by more than 8 mm.
A too-shallow position of the CVC can be associated
with serious complications such as phlebitis or thrombosis (1,3). The difference between their results and ours
comes from the way the study was approached. Their
guideline is from a mixture of results mostly from the
right internal jugular vein catheter and some RSCV catheters. As the distance from skin puncture site to the
SVC-RA junction is shorter for the right internal jugular
vein in comparison with the RSCV (18), the equation
derived from his data gives a shallower depth than ours.
In addition, the central venous anatomy of a small infant
is far different from that of large children or adults
(22,23), therefore applying an equation derived using the
data from a pool of patients consisting of patients over a
large age range may cause an improper result.
Our study was limited to infants 5 kg undergoing
RSCV catheter insertion, therefore indicating a practical guideline for depth of a CVC that should be used
for infants when using the RSCV. Using our guideline,
in 98% of the present patients, the CVC was placed at
a proper position (within 5 mm short of SVC-RA
junction). For the other 2%, the CVC was only 1 mm
deep into the RA, the possibility for the tip of a CVC
to contact with the wall of the cardiac chamber is not
likely at this position. Therefore, following our guideline, almost every CVC can be placed at a safe
position.
There are limitations to our study. We measured
the distance from the skin puncture site to the
SVC-RA junction, as we defined the ideal position of
a CVC as distal SVC outside RA. But the ideal
position of the tip of a CVC can be distal SVC
outside of pericardial reflection (4). To meet this
definition, the length of the pericardium that covers
the distal SVC should have been considered in our
study. However, it was not possible to achieve the
location of the pericardial reflection on TEE. Therefore, our results should be a guideline for the distance from the skin puncture site to the SVC-RA
junction or the maximum allowed depth rather
than the proper depth of the CVC.
From our study, we conclude that, the maximum
depth of the CVC allowed for the RSCV catheter to
avoid its intracardiac placement is between 40 55 mm
for infants less than 5 kg.

1303

The statistical analysis of the present work was supported by the

research fund of Hanyang University (HY-2003-T). We also thank
all members of the Laboratory of Statistical Data Analysis of Hanyang University College of Natural Sciences for their expert statistical analysis.

References
1. McDonough JJ, Altemeier WA. Subclavian venous thrombosis
secondary to indwelling cathers. Surg Gynecol Obstet 1971;133:
397 400.
2. Borja AR. Current status of infraclavicular subclavian vein catheterization. Ann Thorac Surg 1972;13:61524.
3. Johnson CL, Lazarchick J, Lynn HB. Subclavian venipuncture:
preventable complications; report of two cases. Mayo Clin Proc
1970;45:7129.
4. Collier PE, Blocker SH, Graff DM, Doyle P. Cardiac tamponade
from central venous catheters. Am J Surg 1998;176:212 4.
5. Collier PE, Goodman GB. Cardiac tamponade caused by central
venous catheter perforation of the heart: a preventable complication. J Am Coll Surg 1995;181:459 63.
6. Defalque RJ, Campbell C. Cardiac tamponade from central venous catheters. Anesthesiology 1979;50:249 52.
7. Bonventre EV, Lally KP, Chwals WJ, et al. Percutaneous insertion of subclavian venous catheters in infants and children. Surg
Gynecol Obstet 1989;169:2035.
8. Groff DB, Ahmed N. Subclavian vein catheterization in the
infant. J Pediatr Surg 1974;9:171 4.
9. Casado-Flores J, Valdivielso-Serna A, Perez-Jurado L, et al. Subclavian vein catheterization in critically ill children: analysis of
322 cannulations. Intensive Care Med 1991;17:350 4.
10. Huttel MS, Christensen P, Olesen AS. Subclavian venous catheterization in children. Acta Anaesthesiol Scand 1985;29:7335.
11. Jung CW, Bahk JH, Kim MW, et al. Head position for facilitating
the superior vena caval placement of catheters during right
subclavian approach in children. Crit Care Med 2002;30:2979.
12. Kanter RK, Zimmerman JJ, Strauss RH, Stoeckel KA. Central
venous catheter insertion by femoral vein: safety and effectiveness for the pediatric patient. Pediatrics 1986;77:8427.
13. Han SH, Kim SD, Kim CS et al. Comparison of central venous
catheterization in infants. J Int Med Res 2004;32:5639.
14. Kwun KB, Gorfine S, Berman M et al. Percutaneous catheterization of the brachial vein for central venous access. Surg Gynecol
Obstet 1984;159:287 8.
15. Hoyt DB. Internal jugular vein cannulation versus subclavian
vein cannulation: a surgeons view: the subclavian vein. J Clin
Monit 1985;1:613.
16. Miller R. Anesthesia, 5th ed. Philadelphia: Churchill Livingstone, 2000:1147 8.
17. Mukau L, Talamini MA, Sitzmann JV. Risk factors for central
venous catheter-related vascular erosions. J Parenter Enteral
Nutr 1991;15:513 6.
18. Andrews RT, Bova DA, Venbrux AC. How much guidewire is
too much? Direct measurement of the distance from subclavian
and internal jugular vein access sites to the superior vena cavaatrial junction during central venous catheter placement. Crit
Care Med 2000;28:138 42.
19. Czepizak CA, OCallaghan JM, Venus B. Evaluation of formulas
for optimal positioning of central venous catheters. Chest 1995;
107:1662 4.
20. Andropoulos DB, Bent ST, Skjonsby B, Stayer SA. The optimal
length of insertion of central venous catheters for pediatric
patients. Anesth Analg 2001;93:883 6.
21. Hayashi Y, Maruyama K, Takaki O et al. Optimal placement of
CVP catheter in paediatric cardiac patients. Can J Anaesth 1995;
42:479 82.
22. Cobb LM, Goss JC, Gilsdorf RB. Regional anatomy regarding the
placement of central venous cannulas. Ariz Med 1981;38:33 6.
23. Cobb LM, Vinocur CD, Wagner CW, Weintraub WH. The central
venous anatomy in infants. Surg Gynecol Obstet 1987;165:230 4.

Intraarticular Bupivacaine-Clonidine-Morphine Versus

Femoral-Sciatic Nerve Block in Pediatric Patients
Undergoing Anterior Cruciate Ligament Reconstruction
Kha M. Tran, MD*, Theodore J. Ganley, MD, Lawrence Wells,
Kimberly I. Minger, BSN, and Giovanni Cucchiaro, MD*

MD,

Arjunan Ganesh,

MBBS*,

Departments of *Anesthesiology and Critical Care Medicine, Orthopaedic Surgery, and Clinical Research, Childrens
Hospital of Philadelphia, Pennsylvania

We hypothesized that combined femoral-sciatic nerve

block (FSNB) offers better analgesia with fewer side effects than intraarticular infiltration (IA) in children undergoing anterior cruciate ligament (ACL) reconstruction.
Thirty-six
children
undergoing
ACL
reconstruction were randomized to FSNB or IA. FSNB
patients had FSNB with bupivacaine (0.125%)clonidine (2 g/kg), whereas IA patients received bupivacaine (0.25%)-clonidine (1 g/kg)-morphine
(5 mg). Postoperatively, analgesia was provided with
patient-controlled analgesia and rescue morphine. Patient demographics were similar. FSNB patients required less intraoperative fentanyl (50 40 g versus

nterior cruciate ligament (ACL) reconstructions

are more often performed in the adult population. Children have been treated more conservatively in the past for fear of injuring immature physes and causing growth deformity in this skeletally
immature population. Although the number of children undergoing ACL reconstruction has significantly
grown in the past decade, primarily because of increased awareness of these injuries within this
younger athletic population (13), there is a paucity of
data on postoperative pain management in this group
of patients. For this reason, most of the available data
on postoperative pain management after ACL reconstruction have been obtained in the adult population.
Reconstruction of the ACL is often associated with
significant immediate postoperative pain. Traditional
methods of analgesia include the intraarticular (IA)
administration of medications, either local anesthetics,
Accepted for publication May 5, 2005.
Address correspondence to Giovanni Cucchiaro, MD, Childrens Hospital of Philadelphia, Room, 9th Floor, 34th St. and
Civic Center Blvd., Philadelphia, PA 19104. Address e-mail to
Cucchiaro@email.chop.edu.
DOI: 10.1213/01.ANE.0000180218.54037.0B

1304

Anesth Analg 2005;101:130410

80 50 g; P 0.04). Visual analog scale score for FSNB

was smaller than IA in the recovery room (1.8 3 versus 5.4 3; P 0.0002) and during the first 24 h (1.6 1
versus 2.9 2; P 0.01)). FSNB morphine use in the
first 18 h was less (7 13 mg versus 21 21 mg; P
0.03). Fewer FSNB patients vomited (11% versus 50%; P
0.03). IA patients required morphine patientcontrolled analgesia sooner. After ACL reconstruction
in children, FSNB with bupivacaine-clonidine provides
better analgesia with fewer side effects than IA with
bupivacaine-clonidine-morphine.
(Anesth Analg 2005;101:1304 10)

opioids, or a combination of the two, and regional

anesthesia. Both techniques improve immediate postoperative pain control when compared with sham
intervention (4,5). However, when the effects of the
femoral nerve block on postoperative analgesia have
been compared with the IA infiltration IA (6 9), data
have been contradictory.
In this randomized, prospective study, we compared
the effects of a femoral and sciatic nerve block (FSNB)
versus the IA infiltration of a combination of medications on postoperative analgesia in children undergoing
ACL reconstruction. The primary aim of this study was
to determine whether a FSNB using bupivacaine and
clonidine could decrease the consumption of morphine
in the immediate postoperative period compared with
the IA administration of a combination of bupivacaineclonidine-morphine. The secondary aims were to determine whether the pain scores and the incidence of side
effects, such as vomiting, pruritus, and sedation, were
affected by the two different techniques.

Methods
With IRB approval and written informed consent, 36
children getting ACL reconstruction were enrolled into
2005 by the International Anesthesia Research Society
0003-2999/05

ANESTH ANALG
2005;101:1304 10

PEDIATRIC ANESTHESIA
TRAN ET AL.
ANALGESIA AFTER ANTERIOR CRUCIATE LIGAMENT RECONSTRUCTION IN CHILDREN

the study. The study included males and females from

all ethnic backgrounds and between 12 and 19 yr old.
Patients were randomized into one of two study
groups: FSNB or IA. The FSNB group received a combined FSNB using bupivacaine and clonidine. The IA
group received IA injection of bupivacaine-clonidinemorphine. The FSNB were performed by anesthesiologists, and surgeons performed the IA injections. Only
two anesthesiologists and two surgeons participated in
the study to assure uniform anesthetic and surgical
techniques.
The following data were collected by the nurses in
the recovery room and on the patient care units: (a)
pain score, assessed using a visual analog scale (VAS)
ranging from 0 to 10 cm with 10 cm being the worst
pain imaginable, (b) the amount of fentanyl used during the surgery, (c) the amount of morphine administered, (d) time to first demand dose of morphine
patient-controlled analgesia (PCA), (e) incidence of
pruritus, (f) number of episodes of vomiting, (g) doses of
ketorolac, diphenhydramine, and ondansetron administered, (h) sedation level using a De Kock scale (10), (i)
duration of sensory block testing sensation in the distribution areas of both the femoral and sciatic nerve, and (j)
duration of motor block using a modified Bromage score
(11). The evaluating nurses had been trained to assess
and document the outcomes studied.
General anesthesia was induced with IV propofol
(35 mg/kg). Patients randomized to receive a FSNB
were endotracheally intubated without neuromuscular blocking drugs to identify femoral and sciatic
nerve twitch responses during block placement; patients randomized to receive an IA were endotracheally intubated after the administration of vecuronium
(0.1 mg/kg). The anesthesiologist could use muscle
relaxants (vecuronium, 0.1/mg/kg) during the procedure in both groups of patients. Neuromuscular block
was reversed with atropine (20 g/kg) and neostigmine (30 70 g/kg) before endotracheal extubation.
IV fentanyl (1 g/kg) was administered before intubation in every patient. Anesthesia was maintained
with desflurane in nitrous and oxygen. Additional
fentanyl (1 g/kg) was allowed during surgery in the
case of persistent tachycardia and hypertension (defined as values 20% more than baseline) that did not
respond to increases of the inspired desflurane concentration. Every patient received ondansetron 50
g/kg (maximum, 4 mg) at the end of the operation.
In patients randomized to the FSNB group, femoral
and sciatic nerves were localized with a nerve stimulator (Stimuplex Dig RC, B. Braun Medical Inc., Bethlehem, PA) and insulated needle. After sterile prepping and draping, the femoral nerve was localized
with a 24-gauge, 1.5-in. insulated needle (Stimuplex,
B. Braun Medical), which was introduced just lateral
to the femoral artery and 2 cm inferior to the inguinal

1305

ligament. The nerve block was performed when contractions were lost at a stimulator intensity of 0.5 0.7
mA using 0.5 mL/kg of 0.125% bupivacaine with
1:200,000 epinephrine (maximum, 40 mL) and
clonidine 1 g/kg (maximum, 100 g) after negative
aspiration for blood and negative heart rate responses
to a test dose of 3 mL of the local anesthetic solution.
The sciatic nerve was approached from the lateral
thigh (12). A 21-gauge, 4-in. needle was used. Common peroneal or tibial nerve stimulation was sought
with a stimulator intensity of 0.5 0.7 mA. The block
was performed using 0.5 mL/kg of 0.125% bupivacaine with 1:200,000 epinephrine (maximum, 20 mL)
and clonidine 1 g/kg (maximum, 100 g) after negative aspiration and test dose. In those patients randomized to the IA group, surgeons performed the
injection of medication after tracheal intubation and
before prepping the leg. A thigh tourniquet was first
inflated to 300 mm Hg. Patients then received
1 mL/kg of 0.25% bupivacaine (maximum, 30 mL),
morphine 5 mg, and clonidine 1 g/kg. The tourniquet was kept inflated for 15 min after injection.
Before starting the arthroscopic portion of the surgery, a quadrupled hamstring autograft or an Achilles
tendon allograft was prepared, depending on patient
preference. A graft 12 cm in length and 9 or 10 mm in
diameter was used, and the grafts were tubularized
using suture material. The ruptured ACL was debrided with an arthroscopic curved full-radius shaver,
a straight biter, and a pituitary rongeur. A tibial ACL
guide was set at the tibial ACL footprint. A guide pin
and a 10-mm cannulated drill bit were used to create
the tibial tunnel. The femoral tunnel was then placed
at the usual 11 oclock position for the right knee and
the 1 oclock position for the left knee, leaving a small
rim of bone at the posterior wall of the femoral tunnel.
A femoral guide pin was overdrilled with a 10-mm
cannulated drill bit and drilled to 40 mm in length. A
femoral guide for the transverse femoral fixation system (Rigidfix Cross Pin System; Mitek, Westwood,
MA) was then placed such that the bioresorbable cross
pins were placed proximal to the femoral physis. Two
transverse tunnels were drilled using the femoral
guide and cannula. The arthroscopic camera was then
placed within the femoral tunnel to visualize the appropriate location of the transverse fixation points,
proximal to the level of the femoral physis. The camera was also placed within the tibial tunnel to evaluate
the bone stock and the tibial physis to ensure fixation
of the graft distal to the tibial physis. The graft was
passed through the tibial and femoral tunnels and
held in position while two bioabsorbable cross pins
were tamped lateral to medial across the grafts in the
femoral tunnel. The knee was placed in full extension
as the distal length of the graft was pulled distally.
Soft tissue screw fixation of the graft was then performed distal to the tibial physis.

1306

PEDIATRIC ANESTHESIA TRAN ET AL.

ANALGESIA AFTER ANTERIOR CRUCIATE LIGAMENT RECONSTRUCTION IN CHILDREN

Upon arrival to the recovery room, patients were

started on a morphine PCA (demand bolus, 20 g/
kg, lockout interval 8 min, and maximum dose in 1 h
100 g kg1 h1). If patients were in severe pain
on arrival to the recovery room (VAS 3 cm), a
single dose of IV ketorolac (0.5 mg/kg; maximum,
30 mg) was given. In case of persistent pain, boluses
of morphine 50 g/kg (maximum, 3 mg per dose)
were given until the VAS score was 3 cm. After
discharge to the care unit, morphine PCA was continued using the same settings for the next 18 h.
Patients with persistent pain (VAS 3 cm) initially
received IV ketorolac every 6 h (0.5 mg/kg; maximum, 30 mg/dose). When ketorolac was not effective and the VAS score was still more than 3 cm,
patients received repeated doses of IV morphine (50
g/kg; maximum, 3 mg per dose). Morphine PCA
was discontinued after 18 h, and patients were transitioned to oral oxycodone (0.1 mg/kg; maximum,
10 mg) and acetaminophen (15 mg/kg; maximum,
650 mg) every 4 h and discharged home. Patients
complaining of pruritus were given IV diphenhydramine (0.5 mg/kg; maximum, 50 mg) every 8 h as
required. Patients complaining of more than one
episode of vomiting were given ondansetron
(0.1 mg/kg; maximum, 4 mg per dose) every 8 h.
Sample size was based on the primary outcome measure involving the amount of morphine used in the 24 h
after ACL reconstruction. Presuming a 75% reduction of
use of morphine in patients receiving a FSNB, a sample
size of 15 in each group would have an 80% power to
detect a difference between the groups.
The outcomes between the groups were compared
with Students t-test or Fishers exact test, where appropriate, for continuous data or proportions. KaplanMeier survival curves were made, and the log-rank
test was used to compare the time to first PCA use.
Analysis of variance for repeated measures was performed to compare differences in pain scores between
the two treatments with time. Univariate linear regression was performed to verify the existence of a correlation between some of the variable considered in the
final analysis. Patients were not analyzed by intentionto-treat because patients whose blocks failed were
excluded from the final analysis a priori. Data are
presented as mean sd. Statistical significance was
defined as P 0.05.

Results
Thirty-six patients were enrolled in the study. Eighteen patients were randomized to each group. Two
patients from the FSNB group (11%) were excluded
from the final analysis because of failed nerve block.
One patient had no coverage in the distribution of the
lateral femoral cutaneous nerve, and the other had no

ANESTH ANALG
2005;101:1304 10

Table 1. Demographic Data of the Study Population

Age (yr)
Sex (M/F)
Weight (kg)
Race (white/AfricanAmerican/other)
Achilles tendon allograft/
hamstring autograft

FSNB
(n 16)

IA
(n 18)

P-value

15 2
3/13
62 12
11/3/2

15 2
9/9
64 12
13/2/3

NS
NS
NS
NS

13/3

17/1

NS

Demographic data of the study population. Data are presented as mean

sd.

FSNB femoral-sciatic nerve block; IA intraarticular injection.

coverage of the common peroneal nerve. Age, weight,

sex, race, and type of graft used for reconstruction
were similar between the groups (Table 1). Muscular
block was reversed at the end of the operation in 16
(78%) of the patients who received IA and in 1 (6%) of
the patients in the FSNB group. The total anesthesia
time, from the tracheal induction of anesthesia to extubation, was similar in the 2 groups (221 37 min in
the FSNB group and 205 24 min in the IA group; P
0.06).
The intraoperative fentanyl requirement was significantly less after FSNB (50 40 g) compared with IA
(80 50 g) (P 0.04). Fewer patients required rescue
doses of morphine in the recovery room after FSNB
compared with patients who received IA (one versus
six patients, respectively; P 0.061). The consumption
of morphine in the 18 h after the discharge from the
recovery room, including both PCA usage and rescue
doses, was significantly less in the FSNB group (7
13 mg) compared with IA group (21 21 mg) (P
0.03). Only three patients in the FSNB group required
morphine in the postoperative period compared with
15 of the IA patients (P 0.0001) (Table 2). There was
a significant difference in the duration of analgesia
between the two groups (P 0.004) (Fig. 1). Patients in
the FSNB group had a significantly longer time to first
analgesic request compared with the IA group (449
312 versus 116 222 min, respectively). After arrival
to the patient care units, more IA patients (10 of 18 or
55%) required nurse-administered morphine rescues
than FSNB patients (3 of 16 or 19%) (P 0.04). Administration of ketorolac (doses and timing) was similar between the two groups in the recovery room and
patient care units, although there was a trend towards
less ketorolac administration in the recovery room
(Table 2). The average use of morphine during the
course of the study was significantly larger in patients
in whom the surgeons used an Achilles tendon allograft (16 15 mg) compared with patients receiving a
hamstring tendon autograft (5 7 mg) over the course
of the study (P 0.02).
The mean pain score in the FSNB group was less
than in the IA group on arrival to the recovery room

ANESTH ANALG
2005;101:1304 10

PEDIATRIC ANESTHESIA
TRAN ET AL.
ANALGESIA AFTER ANTERIOR CRUCIATE LIGAMENT RECONSTRUCTION IN CHILDREN

1307

Table 2. Visual Analog Scale (VAS) Scores and

Consumption of Analgesic Medications After Anterior
Cruciate Ligament (ACL) Reconstruction in the Two Study
Groups

Intraoperative fentanyl
(g)
Postoperative morphine
(mg)
Morphine rescues in
PACU n (%)
Morphine rescues after
PACU n (%)
Morphine rescues during
study period n (%)
Ketorolac in PACU n (%)
Ketorolac after PACU
discharge n (%)
VAS score PACU (cm)
VAS score after PACU
discharge (cm)

FSNB
(n 16)

IA
(n 18)

50 40

80 50

0.04

7 13

21 21

0.03

1 (6)

6 (33)

0.06

3 (19)

10 (55)

0.04

3 (19)

15 (83)

0.0001

1 (6)
8 (50)

6 (33)
9 (50)

1.8 3
1.6 1

5.4 3
2.9 2

P-value

0.09
1.0
0.0002
0.01

Data are presented as mean sd and percent of patients in each group.

FSNB femoral-sciatic nerve block; IA intraarticular injection; PACU
postanesthesia care unit.

Figure 1. Comparison of survival curves for proportion of patients

using IV morphine patient-controlled analgesia (PCA). Patients getting femoral-sciatic nerve block (FSNB) with bupivacaine and
clonidine went longer before using PCA when compared with
patients getting intraarticular (IA) injection with bupivacaineclonidine-morphine. Some FSNB patients did not require any morphine PCA.

(1.8 3 versus 5.4 3; P 0.0002) and in the 18 h of

the study period (1.6 1 versus 2.9 2; P 0.01; Fig.
2). There was no correlation between pain score and
surgical technique. Analysis of variance showed significantly different VAS scores between the two
groups from the time patients reached the recovery
room to 12 h after the operation (P 0.04).
The incidence of vomiting was significantly different in the two groups. One patient in the FSNB and
two patients in the IA group vomited while in the
recovery room. These three patients had further episodes of vomiting during the study period. Two patients in the FSNB (11%) and nine patients in the IA

Figure 2. Visual analog scale (VAS) scores in the two groups during
the study period. The scores were significantly different from the
time of arrival to the recovery room (time 0) to 12 h after the
operation (analysis of variance; P 0.04). FSNB femoral-sciatic
nerve block; IA intraarticular injection.

group (50%) (P 0.03) vomited after their discharge

from the recovery room.
The incidence of pruritus and sedation were similar
in the two groups (Table 3), and no significant differences were found in the usage of ondansetron and
diphenhydramine. One patient in the IA group suffered from urinary retention. None of the FSNB patients experienced prolonged paresthesia. The duration of the sensory block was of 17 4 h in the area of
the femoral nerve distribution and 20 4 h in the area
of the sciatic nerve distribution. Fifty percent of the
patients experienced a motor block on the femoral or
sciatic nerve distribution. The duration of the motor
block was 4 6 h in the area of the femoral nerve
distribution and 6 7 h in the area of the sciatic nerve
distribution.
No correlation was found between sex or administration of neostigmine and episodes of vomiting (P
0.68). However, a significant correlation was found
between treatment (FSNB versus IA) and vomiting (P
0.014), as well as between consumption of morphine
and vomiting (P 0.001).

Discussion
Management of postoperative pain after ACL reconstruction is historically based on the IA infiltration of
different types of medications or on a femoral nerve
block. Most of the published studies on postoperative
pain control after ACL reconstruction have failed to
show any advantage of regional anesthesia over IA
infiltration of medication (6,8). However, the regional
anesthesia technique performed in these studies was
limited to a femoral nerve block. Data obtained from
retrospective clinical studies indicate that the addition
of a sciatic nerve block significantly improves postoperative pain after major knee operations, such as total
knee replacement and ACL reconstruction using hamstrings autograft (13,14). Anatomic studies have

1308

PEDIATRIC ANESTHESIA TRAN ET AL.

ANALGESIA AFTER ANTERIOR CRUCIATE LIGAMENT RECONSTRUCTION IN CHILDREN

Table 3. Incidence of Side Effects in the Two Study

Groups

Vomiting n (%)
Pruritus n (%)
Sedationa n (%)
Level 0
Level 1
Level 2
Level 3
Ondansetron no. of doses
Diphenhydramine no. of
doses

FSNB
(n 16)

IA
(n 18)

2 (11)
0
8 (50)
8
8
0
0
6
1

9 (50)
1 (5)
14 (78)
4
12
1
1
12
4

P-value
0.03
0.6
0.1

0.2
0.1

a
Sedation scale: 0 awake; 1 sleeping and arousable by verbal command; 2 sleeping and not aroused by verbal stimuli, but aroused to a
drowsy state by tactile stimulation; and 3 sleeping and not aroused to a
drowsy state by tactile stimulation.
FSNB femoral-sciatic nerve block; IA intraarticular injection.

shown that sensory fibers of the sciatic nerve innervate the posterior capsula of the knee and the hamstrings (15,16). Also, the tunneling of the graft most
likely encroaches in areas innervated by the sciatic
nerve.
Opposite conclusions have been reached by other
authors who have shown that a femoral nerve block
provides a better analgesia when compared with the
IA infiltration (17,18). However, in both these studies,
the IA infiltration group received only local anesthetic
when it had been demonstrated that a combination of
local anesthetic, morphine, and clonidine results in
better pain control than a local anesthetic alone in
patients undergoing arthroscopic knee surgery
(19 21).
Our data show that children who received a FSNB
had a significantly lower pain score in both the immediate and late postoperative course compared with
children who received IA medications. This resulted
in a significant less consumption of IV morphine and
incidence of side effects. This study does not confirm
recent data from Williams et al. (14) who, surprisingly,
failed to show any advantage of a simple femoral or
combined FSNB over IA injections after allograft ACL.
This may be explained by the intrinsic limitations of
studies based on retrospective data collections.
After an extensive review of the literature, we decided to combine in the two study groups, FSNB and
IA infiltration, sets of treatments and medications that
have been shown to be effective when administered
independently from each other. (a) We added a sciatic
to the femoral nerve block in the regional anesthesia
group. (b) We combined local anesthetic and clonidine
in the FSNB in an attempt to extend the duration of
the sensory block (2224) and improve the quality
of the sensory block (25). It should be emphasized that
the role of clonidine in peripheral nerve blocks is

ANESTH ANALG
2005;101:1304 10

controversial, with data arguing against any additional benefit from combining clonidine to bupivacaine or ropivacaine (26,27). (c) We combined three
different medications for the IA infiltration. The addition of clonidine (19) or morphine (5,28) to a local
anesthetic improves the quality of analgesia, and one
study showed a significant analgesic benefit from the
IA administration of both morphine and clonidine
(20). (d) Finally, the IA infiltration was performed
before the surgical procedure (28,29), and the tourniquet was left inflated after the IA infiltration to allow
binding of the medications to the local tissue (30).
We found a significant correlation among the analgesia technique, consumption of morphine, and the
incidence of vomiting, clearly indicating that the FSNB
provides better analgesia in children undergoing ACL
reconstruction, which results in less consumption of
narcotics and less frequent incidence of side effects.
We can not draw any conclusion on the effects of the
surgical technique (Achilles tendon allograft versus
hamstrings autograft) on postoperative pain level because of the small number of patients in the hamstrings group.
The most common complaints after orthopedic procedures are pain along with nausea and vomiting,
which are also the most common causes of costly
unplanned admissions (31,32). Our report shows that
a proper use of regional anesthesia can minimize these
postoperative symptoms.
Potential criticisms of this study are the fact that it
was not blinded, that there were more patients who
received neostigmine in the IA compared to the FSNB
group, and the potential for nerve injury when performing peripheral nerve blocks in patients under anesthesia. The study was not blinded because it is impossible to blind patients and evaluating personnel
when comparing regional anesthesia versus other
techniques that do not cause sensory or motor blocks
because of the obvious effects of the nerve block.
Moreover, the measured outcomes (consumption of
morphine, episodes of vomiting, and presence of pruritus) are objective findings that could have not been
influenced by the observers interpretation.
Neostigmine-induced increased motility of the gastrointestinal tract has been suggested to contribute to
postoperative nausea and vomiting. However, the effect of reversal drugs on postoperative vomiting is
unclear. We did not find any correlation between use
of neostigmine and postoperative vomiting. Several
studies have failed to show an increased incidence of
vomiting after administration of neostigmine (33,34).
Surprisingly, authors also found that smaller doses of
neostigmine and atropine may prevent postoperative
vomiting (35,36).
A large percentage of our patients in the FSNB
group experienced a motor block in the postoperative
period. This may have been related to the relative

ANESTH ANALG
2005;101:1304 10

PEDIATRIC ANESTHESIA
TRAN ET AL.
ANALGESIA AFTER ANTERIOR CRUCIATE LIGAMENT RECONSTRUCTION IN CHILDREN

large concentration of bupivacaine we used or the

addition of clonidine to the local anesthetic. The role
of clonidine in inducing or extending the duration of a
motor block is controversial. Studies have shown that
adding clonidine to a local anesthetic may delay the
recovery of the motor functions (37). Other authors
have failed to show any effect of clonidine on the
duration of motor blocks (38). With respect to the local
anesthetic, ropivacaine has been shown to induce less
motor block compared with bupivacaine (39,40), at
least when administered in the epidural space and at
diluted concentrations. In our experience, the presence
of a prolonged motor block is a source of great anxiety
for families and children, particularly in outpatient
settings. However, a proper education for both the
families and patients about this side effect may help
overcome their concerns, and we should take into consideration the type of local anesthetic and the concentration used when performing peripheral nerve blocks in
children to minimize the incidence of motor blocks.
The performance of regional anesthesia in children
is not as widely accepted as in adults because these
blocks must often be performed under general anesthesia. Peripheral nerve blocks are relatively safe
when performed in awake patients (41). There are
only few data in the literature on the safety of peripheral nerve blocks in children. A retrospective study
(42), conducted in 1996 by the French-Language Society of Pediatric Anesthesiologists, did not report any
significant complication from peripheral nerve blocks
performed in children. Although based on small populations, more recent studies have confirmed that peripheral nerve blocks and placement catheters can be
safely performed in anesthetized children (43,44).
None of the patients in our study experienced complications related to the nerve block itself. Larger prospective studies are required to confirm the safety of
this practice.
In summary, this prospective, randomized study
demonstrated that a FSNB with bupivacaine-clonidine,
as opposed to IA with bupivacaine-clonidine-morphine,
decreased immediate postoperative pain and resulted in
decreased opioid requirements and the incidence of side
effects in children getting ACL reconstruction. Further
studies are required to determine the effects of improved
pain control on rehabilitation and patient satisfaction.

References
1. Fehnel DJ, Johnson R. Anterior cruciate injuries in the skeletally
immature athlete: a review of treatment outcomes. Sports Med
2000;29:51 63.
2. Aichroth PM, Patel DV, Zorrilla P. The natural history and treatment
of rupture of the anterior cruciate ligament in children and
adolescents: a prospective review. J Bone Joint Surg Br 2002;84:3841.
3. Millett PJ, Willis AA, Warren RF. Associated injuries in pediatric and adolescent anterior cruciate ligament tears: does a delay
in treatment increase the risk of meniscal tear? Arthroscopy
2002;18:9559.

1309

4. Mulroy MF, Larkin KL, Batra MS, et al. Femoral nerve block
with 0.25% or 0.5% bupivacaine improves postoperative analgesia following outpatient arthroscopic anterior cruciate ligament repair. Reg Anesth Pain Med 2001;26:24 9.
5. Denti M, Randelli P, Bigoni M, et al. Pre- and postoperative intraarticular analgesia for arthroscopic surgery of the knee and
arthroscopy-assisted anterior cruciate ligament reconstruction: a
double-blind randomized, prospective study. Knee Surg Sports
Traumatol Arthrosc 1997;5:206 12.
6. Mehdi SA, Dalton DJ, Sivarajan V, Leach WJ. BTB ACL
reconstruction: femoral nerve block has no advantage over intraarticular local anaesthetic infiltration. Knee Surg Sports Traumatol
Arthrosc 2004;12:180 3.
7. Schwarz SK, Franciosi LG, Ries CR, et al. Addition of femoral
3-in-1 blockade to intra-articular ropivacaine 0.2% does not
reduce analgesic requirements following arthroscopic knee surgery. Can J Anaesth 1999;46:7417.
8. Frost S, Grossfeld S, Kirkley A, et al. The efficacy of femoral
nerve block in pain reduction for outpatient hamstring anterior
cruciate ligament reconstruction: a double-blind, prospective,
randomized trial. Arthroscopy 2000;16:243 8.
9. Edkin BS, McCarty EC, Spindler KP, Flanagan JF. Analgesia
with femoral nerve block for anterior cruciate ligament reconstruction. Clin Orthop 1999;369:289 95.
10. De Kock M, Crochet B, Morimont C, Scholtes JL. Intravenous or
epidural clonidine for intra- and postoperative analgesia. Anesthesiology 1993;79:52531.
11. Senard M, Kaba A, Jacquemin MJ, et al. Epidural levobupivacaine 0.1% or ropivacaine 0.1% combined with morphine provides comparable analgesia after abdominal surgery. Anesth
Analg 2004;98:389 94.
12. Pham-Dang C. Midfemoral block: a new lateral approach to the
sciatic nerve. Anesth Analg 1999;88:1426.
13. Ben David B, Schmalenberger K, Chelly JE. Analgesia after total
knee arthroplasty: is continuous sciatic blockade needed in addition to continuous femoral blockade? Anesth Analg 2004;98:7479.
14. Williams BA, Kentor ML, Vogt MT, et al. Femoral-sciatic nerve
blocks for complex outpatient knee surgery are associated with
less postoperative pain before same-day discharge: a review of
1,200 consecutive cases from the period 1996 1999. Anesthesiology 2003;98:1206 13.
15. Nelissen RG, Hogendoorn PC. Retain or sacrifice the posterior
cruciate ligament in total knee arthroplasty: a histopathological
study of the cruciate ligament in osteoarthritic and rheumatoid
disease. J Clin Pathol 2001;54:381 4.
16. Fenzl G, Zinnecker R. Topography of the sciatic nerves fibres in
regard of clinical use. Anat Anz 1987;163:10710.
17. Iskandar H, Benard A, Ruel-Raymond J, et al. Femoral block
provides superior analgesia compared with intra-articular ropivacaine after anterior cruciate ligament reconstruction. Reg
Anesth Pain Med 2003;28:29 32.
18. De Andres J, Bellver J, Barrera L, et al. A comparative study of
analgesia after knee surgery with intraarticular bupivacaine,
intraarticular morphine, and lumbar plexus block. Anesth
Analg 1993;77:72730.
19. Reuben SS, Connelly NR. Postoperative analgesia for outpatient
arthroscopic knee surgery with intraarticular clonidine. Anesth
Analg 1999;88:729 33.
20. Joshi W, Reuben SS, Kilaru PR, et al. Postoperative analgesia for
outpatient arthroscopic knee surgery with intraarticular
clonidine and/or morphine. Anesth Analg 2000;90:1102 6.
21. Boden BP, Fassler S, Cooper S, et al. Analgesic effect of intraarticular morphine, bupivacaine, and morphine/bupivacaine after arthroscopic knee surgery. Arthroscopy 1994;10:104 7.
22. Singelyn FJ, Gouverneur JM, Robert A. A minimum dose of
clonidine added to mepivacaine prolongs the duration of anesthesia and analgesia after axillary brachial plexus block. Anesth
Analg 1996;83:1046 50.
23. Casati A, Magistris L, Fanelli G, et al. Small-dose clonidine
prolongs postoperative analgesia after sciatic-femoral nerve
block with 0.75% ropivacaine for foot surgery. Anesth Analg
2000;91:388 92.

1310

PEDIATRIC ANESTHESIA TRAN ET AL.

ANALGESIA AFTER ANTERIOR CRUCIATE LIGAMENT RECONSTRUCTION IN CHILDREN

24. Iskandar H, Guillaume E, Dixmerias F, et al. The enhancement

of sensory blockade by clonidine selectively added to mepivacaine after midhumeral block. Anesth Analg 2001;93:7715.
25. Bernard JM, Macaire P. Dose-range effects of clonidine added to
lidocaine for brachial plexus block. Anesthesiology 1997;87:277 84.
26. Erlacher W, Schuschnig C, Orlicek F, et al. The effects of
clonidine on ropivacaine 0.75% in axillary perivascular brachial
plexus block. Acta Anaesthesiol Scand 2000;44:537.
27. Culebras X, Van Gessel E, Hoffmeyer P, Gamulin Z. Clonidine
combined with a long acting local anesthetic does not prolong
postoperative analgesia after brachial plexus block but does
induce hemodynamic changes. Anesth Analg 2001;92:199 204.
28. Tetzlaff JE, Dilger JA, Abate J, Parker RD. Preoperative intraarticular morphine and bupivacaine for pain control after outpatient arthroscopic anterior cruciate ligament reconstruction.
Reg Anesth Pain Med 1999;24:220 4.
29. Reuben SS, Sklar J, El Mansouri M. The preemptive analgesic
effect of intraarticular bupivacaine and morphine after ambulatory arthroscopic knee surgery. Anesth Analg 2001;92:923 6.
30. Whitford A, Healy M, Joshi GP, et al. The effect of tourniquet
release time on the analgesic efficacy of intraarticular morphine
after arthroscopic knee surgery. Anesth Analg 1997;84:7913.
31. Gold BS, Kitz DS, Lecky JH, Neuhaus JM. Unanticipated admission to the hospital following ambulatory surgery. JAMA 1989;
262:3008 10.
32. Carroll NV, Miederhoff PA, Cox FM, Hirsch JD. Costs incurred
by outpatient surgical centers in managing postoperative nausea and vomiting. J Clin Anesth 1994;6:364 9.
33. Hovorka J, Korttila K, Nelskyla K, et al. Reversal of neuromuscular blockade with neostigmine has no effect on the incidence
or severity of postoperative nausea and vomiting. Anesth Analg
1997;85:1359 61.
34. Joshi GP, Garg SA, Hailey A, Yu SY. The effects of antagonizing
residual neuromuscular blockade by neostigmine and glycopyrrolate on nausea and vomiting after ambulatory surgery.
Anesth Analg 1999;89:628 31.

ANESTH ANALG
2005;101:1304 10

35. Boeke AJ, de Lange JJ, van Druenen B, Langemeijer JJ. Effect of
antagonizing residual neuromuscular block by neostigmine and
atropine on postoperative vomiting. Br J Anaesth 1994;72:654 6.
36. King MJ, Milazkiewicz R, Carli F, Deacock AR. Influence of neostigmine on postoperative vomiting. Br J Anaesth 1988;61:4036.
37. Casati A, Vinciguerra F, Cappelleri G, et al. Adding clonidine to
the induction bolus and postoperative infusion during continuous femoral nerve block delays recovery of motor function
after total knee arthroplasty. Anesth Analg 2005;100:866 72.
38. Hutschala D, Mascher H, Schmetterer L, et al. Clonidine added
to bupivacaine enhances and prolongs analgesia after brachial
plexus block via a local mechanism in healthy volunteers. Eur J
Anaesthesiol 2004;21:198 204.
39. Brown DL, Carpenter RL, Thompson GE. Comparison of 0.5%
ropivacaine and 0.5% bupivacaine for epidural anesthesia in
patients undergoing lower-extremity surgery. Anesthesiology
1990;72:633 6.
40. Meister GC, DAngelo R, Owen M, et al. A comparison of
epidural analgesia with 0.125% ropivacaine with fentanyl versus 0.125% bupivacaine with fentanyl during labor. Anesth
Analg 2000;90:6327.
41. Auroy Y, Benhamou D, Bargues L, et al. Major complications of
regional anesthesia in France: the SOS Regional Anesthesia
Hotline Service. Anesthesiology 2002;97:1274 80.
42. Giaufre E, Dalens B, Gombert A. Epidemiology and morbidity
of regional anesthesia in children: a one-year prospective survey
of the French-Language Society of Pediatric Anesthesiologists.
Anesth Analg 1996;83:904 12.
43. Dadure C, Motais F, Ricard C, et al. Continuous peripheral nerve
blocks at home for treatment of recurrent complex regional pain
syndrome I in children. Anesthesiology 2005;102:38791.
44. de Jose MB, Tielens LK. Vertical infraclavicular brachial plexus
block in children: a preliminary study. Paediatr Anaesth 2004;
14:9315.

Large-Dose Pretreatment with Methylprednisolone Fails to

Attenuate Neuronal Injury After Deep Hypothermic
Circulatory Arrest in a Neonatal Piglet Model
Stephan Schubert, MD*, Gisela Stoltenburg-Didinger, MD, PhD, Anke Wehsack, MD*,
Dirk Troitzsch, MD*, Wolfgang Boettcher, ECCP, Michael Huebler, MD,
Matthias Redlin, MD, Majid Kanaan, MD*, Michael Meissler, MD, Peter E. Lange, MD,
and Hashim Abdul-Khaliq, MD, PhD*

PhD*,

Departments of *Paediatric Cardiology and Congenital Heart Disease, Anesthesiology, and Thoracic and Cardiovascular
Surgery, Deutsches Herzzentrum Berlin; Department of Neuropathology, University Clinic Benjamin Franklin, Free
University of Berlin; and Animal Experimental Laboratory, Charite, Humboldt University, Berlin, Germany

Conflicting results have been reported with regard to

the neuroprotective effects of steroid treatment with
cardiopulmonary bypass (CPB) and deep hypothermic
circulatory arrest (DHCA). We evaluated the mode and
severity of neuronal cell injury in neonatal piglets after
prolonged DHCA and the possible neuroprotective effect of systemic pretreatment (6 h before surgery)
with large-dose methylprednisolone (MP). Nineteen
neonatal piglets (age, 10 days; weight, 2.1 0.5 kg)
were randomly assigned to 2 groups: 7 animals were
pretreated with large-dose systemic MP (30 mg/kg)
24 h before surgery, and 12 animals without pharmacological pretreatment (saline) served as control groups.
All animals were connected to full-flow CPB with cooling to 15C and 120 min of DHCA. After rewarming to
38.5C with CPB, animals were weaned from CPB and
survived 6 h before they were killed, and the brain was

rain injury after deep hypothermic circulatory

arrest (DHCA) is still a matter of concern in
corrective cardiac surgery of some complex cardiovascular malformations in neonates (1,2). Understanding the pathophysiological changes, which may
be related to brain injury after such a cardiac surgical
procedure, is fundamental for the development of
pharmacological neuroprotective interventions (3).
Because of the contact of the patients blood with the

Accepted for publication May 10, 2005.

Address correspondence and reprint requests to Hashim AbdulKhaliq, MD, PhD, Department of Paediatric Cardiology and Congenital Heart Disease, Deutsches Herzzentrum Berlin, Augustenburger Platz 1, D-13353 Berlin, Germany. Address e-mail to
abdul-khaliq@dhzb.de.
DOI: 10.1213/01.ANE.0000180206.95542.76
2005 by the International Anesthesia Research Society
0003-2999/05

prepared for light and electron microscopy, immunohistochemistry, and TUNEL-staining. Quantitative histological studies were performed in hippocampus, cortex, cerebellum, and caudate nucleus. Systemic
pretreatment with large-dose MP lead to persistent hyperglycemia but no significant changes of cerebral perfusion. Necrotic and apoptotic neuronal cell death were
detected in all analyzed brain regions after 120 min of
DHCA. In comparison to the control group, large-dose
pretreatment with systemic MP lead to an increase of
necrotic neuronal cell death and induced significant
neuronal apoptosis in the dentate gyrus of the hippocampus (P 0.001). In conclusion, systemic pretreatment with large-dose MP fails to attenuate neuronal cell
injury after prolonged DHCA and induces regional
neuronal apoptosis in the dentate gyrus.
(Anesth Analg 2005;101:13118)

large nonphysiological surface of the cardiopulmonary bypass (CPB), a systemic inflammatory response
syndrome may develop, which is thought to aggravate
ischemia-reperfusion injury in the brain and other
organs (4,5). In addition to hypothermia, which plays
the major role in the cerebral protection during circulatory arrest, the administration of different medications have been routinely used before and after surgery to achieve organ protection. Methylprednisolone
(MP) is one of the frequently used protective drugs in
cardiac surgery for the prevention systemic inflammatory response syndrome (1,4 7). Conflicting results,
however, have been reported regarding the neuroprotective effects of steroids in several experimental settings (4,8,9). The precise mechanism of neuroprotective action of MP in the brain is still not clearly
understood (8). Various neuroprotective mechanisms,
such as dermolytic (10) or antioxidant effect (11,12),
Anesth Analg 2005;101:13118

1311

1312

PEDIATRIC ANESTHESIA SCHUBERT ET AL.

LARGE-DOSE PRETREATMENT WITH METHYLPREDNISOLONE

inhibition of macrophages (13) or the expression of

proinflammatory cytokine tumor necrosis factor
(8,14), reduction of the release of excitatory amino
acids (15), and improvement of cerebral blood flow
(CBF) (9), have been suggested. Several prospective
clinical studies have been performed during the last
decades for the evaluation of the protective effect of
MP using a dose of 30 mg/kg during cardiac surgery
with improvement of cardiac performance (4,16) or
pulmonary function and survival rate (17) or no clinical improvement (8).
The mode and time of drug administration, including pretreatment, type of ischemic injury, and the
evaluated variables in these studies, are heterogeneous and may not necessarily have implications for
clinical use of neuroprotective strategies in pediatric
cardiac surgery (9,1719). Thus, an evaluation of the
effect of MP pretreatment on neuronal cell changes in
the neonatal brain in association with bypass perfusion and DHCA is warranted. Because no significant
neuronal cell damage was found in our preliminary
experimental studies after 60 min of DHCA (21), prolonged duration of DHCA was chosen in our study to
induce significant neuronal cell damage to evaluate
any reduction of neuronal cell injury by MP. Similar
prolonged DHCA of 120 min has been used in numerous other experimental studies (4,22) to evaluate protective strategies (23). This study was performed to
evaluate the pattern of neuronal cell death and a possible improvement by pretreatment with large-dose
MP in a neonatal piglet model with prolonged DHCA
and subsequent reperfusion.

Methods
Nineteen neonatal piglets (age, 10 days; weight, 2.1
0.5 kg body weight [BW]) were included and randomly assigned in this study with approval of the
Institution of Animal Care for the State of Berlin (Reg.
0146/98). Seven animals were pretreated with largedose (30 mg/kg of BW) MP (methylprednisolone-21hydrogen-succinate; Urbason, Hoechst, Germany),
which was administered systemically at 24 h and
again at 4 h after surgery (each dose with 30 mg/kg).
Twelve animals without pharmacological treatment
served as the control group.
Sedation was initiated with IM ketamine 10 mg/kg
BW and midazolam 0.51 mg/kg BW and was then
continued with total IV anesthesia using fentanyl
(maintenance dose, 4 10 g kg1 h1 and midazolam 510 g kg1 h1). Muscle relaxation was performed with bolus pancuronium bromide 0.1 mg/kg
BW, with a maintenance dose of 0.05 mg/kg every
23 h. Continuous invasive arterial blood pressure
monitoring and blood gas sampling were achieved
with intravascular catheters in the femoral artery and

ANESTH ANALG
2005;101:13118

vein and also in the internal jugular vein. Prophylactic

antibiotic treatment consisted of IV second generation
cephalosporins. During rewarming and in the postoperative phase, dopamine (2 6 g kg1 min1) was
used for inotropic support in all animals to stabilize
cardiac function and to achieve a normal perfusion
pressure (24). Furosemide (12 mg/kg) was also used
to support urine production and normal urine production with 2 4 mL/kg BW per hour.
Nonpulsatile CPB was initiated with a roller pump (Stoeckert, Munich, Germany) by the use of a microporous
polypropylene membrane neonatal oxygenator (SafeMicro, Polystan, Denmark) with a 40-m arterial filter
adapted from our clinical use (Dideco/Sorin, Mirandola,
Italy), which had been selected for the size and flow requirements of the animals. The priming volume consisted
of 300 mL of fresh homologous blood collected from a
donor pig on the day of the study and an electrolyte solution to obtain a minimal target hemoglobin value of approximately 9 mg/dL. NaHCO3 8.4% was added to the
prime to achieve a normal pH value of 7.4 0.05. Temperature was kept constant at 38C 0.5C. After median
sternotomy and systemic heparinization (400 IU/kg), a 12
14F venous cannula was inserted into the right atrium via
the auricular appendage. After a small incision within the
purse string, an arterial cannula (8F) was inserted and secured with the purse-string tourniquet. The activated clotting time was kept at more than 480 s. The animals were
connected to the bypass machine and normothermic CPB
was established for 20 min at flow rate of 200
250 mL kg1 min1. Arterial carbon dioxide tension was
maintained throughout the hypothermic CPB at 3545 mm
Hg, uncorrected for temperature according to -stat blood
gas management, because of our institutional standard in
clinical use (25). With 30 min of systemic cooling, the minimal temperature of 15C was reached, bypass perfusion
was stopped, and DHCA was induced for 120 min. After
that, the animals were reperfused and rewarmed for 45 min
to achieve a normal rectal temperature for piglets of 38C
0.5C. The animals were then weaned from bypass and
monitored closely for 6 h under general anesthesia with
ventilation and stable hemodynamics. Fresh whole blood
was given to keep the hemoglobin value within the normal
range (910 g/dL) during and after CPB according to the
preoperative values and as reported for neonatal piglets
(24).
Three different fluorescent microspheres (FluoSpheres, Molecular Probes Inc., Leiden, Holland)
were used to evaluate regional brain blood perfusion
before and after DHCA. One milliliter of microspheresolution contains 1.0 106 fluorescent polystyrene
microspheres with a diameter of 14 m. The microsphere solution was mixed with 4 mL of aspirated
blood and was administered through the aortic cannula within 30 s. A reference blood sample was aspirated from the femoral artery for 2 min with continuous blood flow using an aspiration pump. The

ANESTH ANALG
2005;101:13118

PEDIATRIC ANESTHESIA
SCHUBERT ET AL.
LARGE-DOSE PRETREATMENT WITH METHYLPREDNISOLONE

1313

Table 1. Perioperative Parameters (mean sd)

Preoperative
Heart rate
Without MP
With MP-treatment
MAP
Without MP
With MP-treatment
CVP
Without MP
With MP-treatment

bpm
bpm

159 26
175 16

CPB

Reperfusion Postop 1 h

162 28
205 17

208 13
195 18

213 11
210 6

Postop 2 h

Postop 4 h Postop 6 h

197 5
210 15

229 23
223 17

216 8
223 2

mm Hg
mm Hg

63 14.9
78 15.1

62 8.2
64 8.6

85 7.0
76 8.0

74 14.0
73 22.3

62 2.4
81 20

66 6.7
68 7.8

70 2.5
63 6.9

mm Hg
mm Hg

5.4 1.3
3.8 2.2

4.8 1.4
5.5 2.2

5.0 0.4
8.8 1.4

9.0 1.3
9.0 2.0

8.0 1.3
8.0 2.5

6.5 1.0
8.5 3.3

8.0 1.4
7.5 2.5

MP methylprednisolone; CPB cardiopulmonary bypass; DHCA deep hypothermic circulatory arrest; MAD mean arterial blood pressure; Postop
postoperative.

hemodynamic variables and bypass flow rate were

kept stable and simultaneously documented. After the
end of the experiment, tissue samples were collected
from representative brain regions according to standardized protocols for determination of regional CBF
(rCBF). For homogeneous digestion, the samples were
incubated with alkaline sodium hydroxide (NaOH) in
45C for 48 h. After hydrolysis of the brain tissues, the
microspheres were separated from the tissue suspension using a diaphragm-operated vacuum pressure
pump (Millipore, Billerica, MA). The diaphragm consisted of a filter device with pore size of 10 m. The
filter devices with the retained microspheres were
then dried at 50C for 15 min. 2-ethoxyethyl-acetate
was added as solvent solution to the filter papers to
dissolve the different dyes from the microspheres.
Extension spectra of dissolved microspheres were
then quantified using a luminescence spectrometer.
Cerebral arteriovenous oxygen metabolization was
calculated as a difference between the arterial (CaO2)
and venous oxygen (CvO2) content according to the
formula: avDO2 CaO2 CvO2, where CaO2 or
CvO2 was calculated from the Po2 with the standardized and common formula, as previously reported
(27): (1.34 Hb SO2/100) 0.0031 Pao2 or Pvo2.
At the end of the 6-h postoperative period, the skull
was carefully opened, and the brain was immediately
removed, cut into coronal sections, and fixed for 48 h in
SOMOGYI-solution (28) consisting of paraformaldehyde, glutaraldehyde, and picric acid with neutral pH
values (7.4) for subsequent histological, immunohistochemical staining and electron microscopy studies. The
specimens for light microscopy were embedded in paraffin, sectioned at 6 m, and stained with hematoxylin
and eosin. For electron microscopy, the tissue was dehydrated and embedded in araldite. Semi-thin sections
were stained with toluidine blue. Apoptotic cell death
was either assessed by light microscopy, TUNEL (terminal transferase-mediated dUTP-digoxigenin nick endlabeling; Boehringer, Mannheim, Germany), or electron
microscopy.

Necrotic neuronal cell death was identified by eosinophilic cytoplasm and loss of cytoplasmatic structures and nuclear membrane integrity, according to
the neuropathological criteria (29,30). Characteristic
brain regions, such as the hippocampal regions CA4
and CA1, the cortical cingulate gyrus, the nucleus
caudatus, and cerebellum, were studied histologically.
For quantification of cell death, 10 40 magnifications and a ZEISS eyepiece graticule with 100 grid
squares were used. In every studied brain region, a
minimum of 500 cells were evaluated by a blinded
neuropathologist, and necrotic neurons were expressed as a percentage of the total counted neuronal
cell population. Apoptosis was predominantly found
in the dentate gyrus. Apoptotic neuronal cell death
was recognized and characterized by a sequence of
morphological criteria such as chromatin condensation, nuclear blebbing, and the presence of apoptotic
bodies (29,30). Normal and apoptotic cells were identified in hematoxylin and eosin stained paraffin sections, confirmed by semi-thin sections, and also
counted in the defined areas of 16 mm (2) using
computer-assisted morphometry in the dentate gyrus.
Mann-Whitney and Wilcoxon tests were used and
processed with StatView (Cary, NC) and SPSS (version 11.0; SPSS Inc., Chicago, IL) statistical software.

Results
The animals did not differ significantly in arterial
blood gases, arterial blood pressure, and hemoglobin,
rectal, and nasopharyngeal temperatures before surgery, before the induction of DHCA, and after the end
of CPB, as reported in Tables 1 and 2.
The BW before surgery was similar in all groups
(2380 150 g) and increased significantly in all groups
after the operation. Systemic pretreatment resulted in
a slight reduction of postoperative BW gain in the
MP-treated group (200 60 g versus 350 65 g; P
0.08). The brain weight at the end of the experiment
was similar in both groups without any statistical

1314

PEDIATRIC ANESTHESIA SCHUBERT ET AL.

LARGE-DOSE PRETREATMENT WITH METHYLPREDNISOLONE

ANESTH ANALG
2005;101:13118

Table 2. Blood Gas Variables (mean sd)

CPB

pH value
Without MP
With MP-treatment
Hemoglobin (g/dL)
Without MP
With MP-treatment
Po2 (mm Hg)
Without MP
With MP-treatment
Pco2 (mm Hg)
Without MP
With MP-treatment
Potassium (mmol/L)
Without MP
With MP-treatment
Sodium (mmol/L)
Without MP
With MP-treatment
Calcium (mmol/L)
Without MP
With MP-treatment

Reperfusion

Preoperative

37C

15C

7.46 0.04
7.44 0.06

7.45 0.04
7.40 0.05

7.42 0.02
7.41 0.05

8.48 0.89
8.86 0.63

9.24 1.47
9.91 0.81

9.42 1.47
9.01 0.51

188 5.6
181 6.4

225 3.6
228 4.3

398 8.6
368 9.3

31 1.5
32 2.1

38 2.3
38 2.6

3.24 0.76
3.22 0.78

37C

Postoperative
1h

2h

4h

6h

7.23 0.05
7.21 0.06

7.36 0.03
7.32 0.06

7.37 0.04
7.34 0.03

7.45 0.03
7.35 0.02

7.40 0.04
7.35 0.05

8.98 1.25
8.38 1.32

9.89 1.2
8.48 1.1

9.46 0.8
8.92 1.2

9.42 1.3
8.76 1.7

9.26 1.3
8.82 2.6

214.95 2.6
181.8 2.3

307 2.3
217 2.1

264 4.1
252 2.3

273 3.5
221 2.6

314 4.2
298 5.6

39 2.3
37 1.9

35 1.9
37 1.2

28 0.8
28 0.6

33 1.2
30 0.9

24 0.9
30 1.2

33 2.1
35 2.1

4.13 0.57
3.66 0.8

4.01 0.17
3.86 0.68

5.21 0.81
5.53 0.69

4.43 0.72
4.5 0.7

4.26 0.41
4.74 0.44

4.22 0.43
4.63 0.23

3.87 0.36
4.98 0.67

142 5
140 4

142 4
141 4

140 4
141 2

137 3
139 4

141 4
141 3

142 4
137 4

145 4
137 5

143 4
145 3

0.79 0.21
0.92 0.16

0.57 0.14
0.54 0.16

0.67 0.08
0.61 0.12

0.76 0.13
0.85 0.12

0.76 0.11
0.87 0.13

0.82 0.11
1.14 0.12

0.96 0.09
1.05 0.15

0.9 0.08
1.12 0.12

MP methylprednisolone; CPB cardiopulmonary bypass; DHCA deep hypothermic circulatory arrest.

Figure 1. Serial measurements (mean sd) of arterial blood glucose

during the experiment are showing significant hyperglycemia after
methylprednisolone (MP) pretreatment (#P 0.001). Blood glucose
levels are reaching preoperative values only in the control group
(without MP) but not in the MP-pretreated animals. CPB cardiopulmonary bypass.

difference (34 0.9 g versus 35 1.2 g; not

significant).
Blood glucose levels after reperfusion were significantly higher in the MP-treated animals compared
with the control group (P 0.001), as reported in
Figure 1. Blood glucose levels after DHCA did not
reach baseline values in the MP-treated group but

normalized in the group without MP. Insulin treatment of hyperglycemia was not used in any of the
animals to keep pharmacological intervention standardized in all groups.
The rCBF in both groups did not differ significantly
at baseline after CPB was started. Hypothermic perfusion resulted in a significant decrease of rCBF to
nearly 25% 40% of the baseline values in both groups
(Fig. 2). However, after reperfusion and rewarming,
the percentage recovery of rCBF from baseline in the
basal ganglia and parietal and frontal lobe of the brain
was slightly higher in the MP-pretreated animals
without reaching statistical significance.
Cerebral arteriovenous oxygen content difference
(AVDO2) showed a slight increase in the MP-treated
group during the early postoperative period, with no
statistical difference (P 0.36). The AVDO2 in both
groups was not different at normothermic CPB (control versus MP pretreatment 5.1 mL/dL versus
5.4 mL/dL) and was significantly reduced during
deep hypothermia (1.1 versus 1.0 mL/dL) (P 0.01).
AVDO2-values returned to preoperative values during the postoperative period (between 5 and 6 mL/dL
in both groups), without any statistical difference (P
0.48).
In this neonatal piglet model, the morphological
changes after prolonged DHCA consisted of hypoxic necrotic neuronal cell death particularly in the
hippocampus in both groups (Fig. 3). No significant
reduction of necrotic neuronal cell death was observed in the studied representative regions of the
hippocampus CA1-CA4 or the dentate gyrus, cingulate gyrus, and cerebellum after systemic large-dose

ANESTH ANALG
2005;101:13118

PEDIATRIC ANESTHESIA
SCHUBERT ET AL.
LARGE-DOSE PRETREATMENT WITH METHYLPREDNISOLONE

1315

Figure 2. Regional cerebral blood flow

(rCBF) measured by fluorescent microspheres (FluoSpheres) in milliliters per
gram per minute. Standardized measurements during the full-flow cardiopulmonary bypass (CPB) were performed in
both groups. The first measurement was
at normothermic CPB, and the second
measurement was at hypothermic bypass, just before the initiation of deep
hypothermic circulatory arrest (DHCA).
The third measurement is after reperfusion (REP) and rewarming. A slight increase of CBF could be detected in methylprednisolone (MP)-pretreated animals
during REP in the basal ganglia and cerebellum but without any significant statistical difference (Wilcoxon test; not
significant).

Figure 3. Significant increase of necrotic neuronal cells in the hippocampal region Cornu ammonis (CA) 1 and 4 after pretreatment
with methylprednisolone (MP) after prolonged deep hypothermic
circulatory arrest (DHCA; P 0.006 for CA1; #P 0.03 for CA4).
CA4 has been shown to be one of the most sensitive brain regions to
hypoxia and a possible target of pharmacological protective
intervention.

MP pretreatment. Single-cell apoptosis could be observed in all brain regions investigated in both
groups. However, in comparison to the animals
without pretreatment, apoptotic neuronal cell death
occurred predominantly in the dentate gyrus after
systemic MP-pretreatment (P 0.001) (Fig. 4). The
morphological aspects of apoptosis characterized by
marginal dark chromatin condensation were confirmed by semithin section (Fig. 5). In situ end labeling of sections of the dentate gyrus in neonatal
piglets showed significantly more TUNEL-positive
neurons in the MP-pretreated animals (Fig. 4).
Histological evaluation provided no signs of inflammation such as microglial activation or infiltration of
the brain parenchyma by hematogenous cells. Global
brain edema was manifested predominantly in the
astrocytes and perivascular space in all regions of the
brain. Light microscopy and electron microscopy
showed a marked perivascular swelling of the astrocytic
end feet, as reported in our preliminary work (21).

Discussion
In this neonatal animal model of prolonged DHCA
with a reperfusion period of six hours, we did not find

1316

PEDIATRIC ANESTHESIA SCHUBERT ET AL.

LARGE-DOSE PRETREATMENT WITH METHYLPREDNISOLONE

Figure 4. The percentage of apoptotic neurons in the dentate

gyrus of newborn piglets after pretreatment with methylprednisolone (MP) and prolonged deep hypothermic circulatory arrest (DHCA). Regional apoptotic neuronal cell death was increased predominantly in the dentate gyrus after pretreatment
with MP (P 0.01).

any significant neuroprotective effect after systemic

pretreatment with large-dose MP. To evaluate attenuation of brain injury after pretreatment with MP,
DHCA time was deliberately prolonged to 120 minutes, according to other experimental studies evaluating neuroprotective treatment (4,17,31,32).
With our preliminary experimental model, we described neuropathology after 60 minutes of DHCA
and six hours subsequent reperfusion, with major
changes on the astroglial and endothelial level, without detecting significant neuronal cell pathology
(21,33). Clinically, it is recommended to keep DHCA
times less than 60 minutes (34,35). Nevertheless,
longer periods of DHCA have been reported in the
literature, e.g., during corrective surgery of hypoplastic heart syndrome (35). It was confusing that systemic
pretreatment with MP was not associated with any
improvement in cerebral perfusion and oxygenation,
as has been reported (9). Improvement of regional
brain perfusion may indicate attenuation of brain injury after DHCA. However, changes of brain perfusion within the normal range in animals undergoing
DHCA may not necessarily indicate significant reduction of ischemic neuronal cell death (9). Astroglial cell
injury and perivascular edema has been found to precede neuronal injury after 60 minutes of DHCA (33).
Systemic pretreatment of MP did not significantly reduce the postoperative BW gain in comparison to that
of the control group, and brain weight at the end of the
experiment was similar in both groups. Although
quantitative evaluation of the water content may provide better information on the severity of brain edema,

ANESTH ANALG
2005;101:13118

histological and ultrastructural studies showed no significant obvious improvement of perivascular or neuronal edema in the pretreated group.
Significant neuronal cell death was expected after
prolongation of DHCA. The mode of neuronal cell
injury in this study was similar to that reported in other
experimental piglet studies with DHCA (36,37). But the
quantitative regional neuropathological assessment revealed no reduction of necrotic neuronal cell death in the
systemically pretreated animals. Similar to the findings
of other experimental studies, the hippocampal regions
CA4 and CA1 have been shown to be the most sensitive
brain regions to ischemia and suitable for the assessment
of pharmacological protective intervention (37,38).
However, MP induced significant apoptotic neuronal cell death, which was predominantly found in the
dentate gyrus of the hippocampus and cortex. However, the occurrence of apoptosis seems to be regional
and not global (30,36,37). Ischemia and reperfusion
have been found to induce both necrotic and apoptotic
neuronal cell death with variations according to mode
of ischemic injury, neuronal cell type, and regional
distributions (22,36,37). Apoptosis has been found
more frequently in immature brain regions, which
may explain the selective vulnerability of the dentate
gyrus to apoptotic cell death after ischemia (36,39). In
some physiological and pathophysiological settings
(40 42), steroid treatment itself has been found to
induce apoptotic cell death in different cell types, including the brain (43,44). But the precise molecular
mechanism by which MP induces apoptotic cell death
in the hippocampus is still unknown. One explanation
for the induction of apoptosis may be derived from a
study (45) of the rat hippocampus, where activation of
the glucocorticoid receptor increased the ratio of the
proapoptotic molecule Bax relative to the antiapoptotic molecule BcL-xL. In addition, hyperglycemia, induced by MP, may also have contributed to an increased number of apoptotic and necrotic cells in the
systemically pretreated animals (46 50).
Because of differences in the clinical and experimental
setting, time of MP application, and mode and duration
of ischemia, comparative conclusions concerning the
neuroprotective advantage of MP are difficult
(9,12,43,44,51). Where other studies have suggested a
protective effect, in contrast to our findings, one should
cautiously consider the conclusion of a protective effect
of MP in neonates and infants undergoing such extreme
physiological changes from DHCA (9).
The limitations of our study were that the concentration of steroids in the serum and at the cerebral
level were not measured, which may limit the interpretation of the protective effect of MP regarding the
mode of application and cerebral permeation. Also,
TUNEL-positive neurons may not necessarily always
indicate apoptotic cell death because DNA fragmentation has also been reported in neurons undergoing

ANESTH ANALG
2005;101:13118

PEDIATRIC ANESTHESIA
SCHUBERT ET AL.
LARGE-DOSE PRETREATMENT WITH METHYLPREDNISOLONE

1317

Figure 5. Semi-thin sections showing a part of the dentate gyrus taken in a systemic methylprednisolone (MP)-treated piglet with apoptotic
cell death characterized by marginal dark chromatin condensation (lower arrows) adjacent to well-preserved neurons (upper arrows). On the
right side, neuronal apoptosis with marked nuclear blebbing is surrounded by normal and preserved neuronal cells (arrowhead).

stages of necrosis (29,30,52). Untreated persistent hyperglycemia in the MP group may additionally have
worsened the cerebral pathology after DHCA, but it
was not considered to be controlled according to our
protocol. Additional studies should evaluate the independent role of hyperglycemia regarding aggravation
of necrosis and apoptosis. Additional biochemical,
molecular, and ultrastructural studies are required,
including treatment of metabolic side effects of MP
treatment.
In conclusion, a different effect of steroids could be
suspected in the neonatal age, especially in the context
of ischemia and reperfusion, and the use of steroids
has to be reconsidered in critically ill neonates and
infants suffering hypoxic-ischemic events (49). This
statement is emphasized by a recent meta-analysis
concerning postnatal steroid treatment in preterm infants because of bronchopulmonary dysplasia, in
which an altered neurodevelopmental outcome was
described (53,54).
We would like to thank Anne Gale of Deutsches Herzzentrum
Berlin for editorial assistance.

References
1. Jonas RA. Hypothermia, circulatory arrest, and the pediatric
brain. J Cardiothorac Vasc Anesth 1996;10:66 74.
2. Bellinger DC, Jonas RA, Rappaport LA, et al. Developmental
and neurologic status of children after heart surgery with hypothermic circulatory arrest or low-flow cardiopulmonary bypass [see comments]. N Engl J Med 1995;332:549 55.
3. du Plessis AJ, Johnston MV. The pursuit of effective neuroprotection during infant cardiac surgery. Semin Pediatr Neurol
1999;6:55 63.

4. Duffy JY, Nelson DP, Schwartz SM, et al. Glucocorticoids reduce

cardiac dysfunction after cardiopulmonary bypass and circulatory arrest in neonatal piglets. Pediatr Crit Care Med 2004;5:
28 34.
5. Lodge AJ, Chai PJ, Daggett CW, et al. Methylprednisolone
reduces the inflammatory response to cardiopulmonary bypass
in neonatal piglets: timing of dose is important. J Thorac Cardiovasc Surg 1999;117:51522.
6. Jonas RA. Deep hypothermic circulatory arrest: a need for caution. Ann Thorac Surg 1996;61:779 80.
7. Shum-Tim D, Tchervenkov CI, Laliberte E, et al. Timing of
steroid treatment is important for cerebral protection during
cardiopulmonary bypass and circulatory arrest: minimal protection of pump prime methylprednisolone. Eur J Cardiothorac
Surg 2003;24:12532.
8. Chaney MA. Corticosteroids and cardiopulmonary bypass: a
review of clinical investigations. Chest 2002;121:92131.
9. Langley SM, Chai PJ, Jaggers JJ, Ungerleider RM. Preoperative
high dose methylprednisolone attenuates the cerebral response
to deep hypothermic circulatory arrest. Eur J Cardiothorac Surg
2000;17:279 86.
10. Jamshidi J, Yoshimine T, Ushio Y, Hayakawa T. Effects of glucocorticoid and chemotherapy on the peritumoral edema and
astrocytic reaction in experimental brain tumor. J Neurooncol
1992;12:197204.
11. Marzatico F, Gaetani P, Buratti E, et al. Effects of high-dose
methylprednisolone on Na()-K ATPase and lipid peroxidation after experimental subarachnoid hemorrhage. Acta Neurol
Scand 1990;82:26370.
12. Hall ED. Neuroprotective actions of glucocorticoid and nonglucocorticoid steroids in acute neuronal injury. Cell Mol Neurobiol 1993;13:41532.
13. Kim JS, Chopp M, Gautam SC. High dose methylprednisolone
therapy reduces expression of JE/MCP-1 mRNA and macrophage accumulation in the ischemic rat brain. J Neurol Sci
1995;128:28 35.
14. Buttini M, Mir A, Appel K, et al. Lipopolysaccharide induces
expression of tumour necrosis factor alpha in rat brain: inhibition by methylprednisolone and by rolipram. Br J Pharmacol
1997;122:14839.
15. Liu D, McAdoo DJ. Methylprednisolone reduces excitatory
amino acid release following experimental spinal cord injury.
Brain Res 1993;609:2937.

1318

PEDIATRIC ANESTHESIA SCHUBERT ET AL.

LARGE-DOSE PRETREATMENT WITH METHYLPREDNISOLONE

16. Schwartz SM, Duffy JY, Pearl JM, et al. Glucocorticoids preserve
calpastatin and troponin I during cardiopulmonary bypass in
immature pigs. Pediatr Res 2003;54:917.
17. Shum-Tim D, Tchervenkov CI, Jamal AM, et al. Systemic steroid
pretreatment improves cerebral protection after circulatory arrest. Ann Thorac Surg 2001;72:146571; discussion 14712.
18. Carotti A, Emma F, Picca S, et al. Inflammatory response to
cardiac bypass in ewe fetuses: effects of steroid administration
or continuous hemodiafiltration. J Thorac Cardiovasc Surg 2003;
126:1839 50.
19. Hassan AH, von Rosenstiel P, Patchev VK, et al. Exacerbation of
apoptosis in the dentate gyrus of the aged rat by dexamethasone
and the protective role of corticosterone. Exp Neurol 1996;140:
4352.
20. Kin H, Ishibashi K, Nitatori T, Kawazoe K. Hippocampal neuronal death following deep hypothermic circulatory arrest in
dogs: involvement of apoptosis. Cardiovasc Surg 1999;7:558 64.
21. Abdul-Khaliq H, Schubert S, Stoltenburg-Didinger G, et al. Release patterns of astrocytic and neuronal biochemical markers in
serum during and after experimental settings of cardiac surgery.
Restor Neurol Neurosci 2003;21:14150.
22. Ye JYL, Del Bigio MR, Filgueiras CL, et al. Neuronal damage
after hypothermic circulatory arrest and retrograde cerebral
perfusion in the pig. Ann Thorac Surg 1996;61:1316 22.
23. Nagashima M, Shinoka T, Nollert G, et al. High-volume continuous hemofiltration during cardiopulmonary bypass attenuates pulmonary dysfunction in neonatal lambs after deep hypothermic circulatory arrest. Circulation 1998;98:II378 84.
24. Pond W, Houpt KA. The Biology of the pig. Ithaca, NY: Cornell
University Press, 1978.
25. Merkle F, Boettcher W, Schulz F, et al. Perfusion technique for
nonhaemic cardiopulmonary bypass prime in neonates and infants under 6 kg body weight. Perfusion 2004;19:229 37.
26. Kirkham FJ. Recognition and prevention of neurological complications in pediatric cardiac surgery. Pediatr Cardiol 1998;19:
331 45.
27. Bullock R, Stewart L, Rafferty C, Teasdale GM. Continuous
monitoring of jugular bulb oxygen saturation and the effect of
drugs acting on cerebral metabolism. Acta Neurochir Suppl
(Wien) 1993;59:113 8.
28. Somogyi P, Takagi H. A note on the use of picric acidparaformaldehyde-glutaraldehyde fixative for correlated light
and electron microscopic immunocytochemistry. Neuroscience
1982;1:1779 83.
29. Nicotera P, Leist M, Ferrando-May E. Intracellular ATP, a
switch in the decision between apoptosis and necrosis. Toxicol
Lett 1998;102103:139 42.
30. Leist M, Nicotera P. Apoptosis versus necrosis: the shape of
neuronal cell death. Results Probl Cell Differ 1998;24:10535.
31. Bokesch PM, Seirafi PA, Warner KG, et al. Differential
immediate-early gene expression in ovine brain after cardiopulmonary bypass and hypothermic circulatory arrest. Anesthesiology 1998;89:961 8.
32. Nollert G, Nagashima M, Bucerius J, et al. Oxygenation strategy
and neurologic damage after deep hypothermic circulatory arrest. I. Gaseous microemboli. J Thorac Cardiovasc Surg 1999;
117:1166 71.
33. Abdul-Khaliq H, Schubert S, Stoltenburg-Didinger G, et al. Protein S-100beta in brain and serum after deep hypothermic circulatory arrest in rabbits: relationship to perivascular astrocytic
swelling. Clin Chem Lab Med 2000;38:1169 72.
34. Bellinger DC, Wernovsky G, Rappaport LA, et al. Cognitive
development of children following early repair of transposition
of the great arteries using deep hypothermic circulatory arrest.
Pediatrics 1991;87:7017.

ANESTH ANALG
2005;101:13118

35. Gaynor JW, Mahle WT, Cohen MI, et al. Risk factors for mortality after the Norwood procedure. Eur J Cardiothorac Surg
2002;22:829.
36. Ditsworth D, Priestley MA, Loepke AW, et al. Apoptotic neuronal death following deep hypothermic circulatory arrest in
piglets. Anesthesiology 2003;98:1119 27.
37. Kurth CD, Priestley M, Golden J, et al. Regional patterns of
neuronal death after deep hypothermic circulatory arrest in
newborn pigs. J Thorac Cardiovasc Surg 1999;118:1068 77.
38. Myung RJ, Petko M, Judkins AR, et al. Regional low-flow perfusion improves neurologic outcome compared with deep hypothermic circulatory arrest in neonatal piglets. J Thorac Cardiovasc Surg 2004;127:1051 6; discussion 1056 7.
39. Yue XMH, Penrice J, Cooper C, et al. Apoptosis and necrosis in
the newborn piglet brain following transient cerebral hypoxiaischaemia. Neuropathol Appl Neurobiol 1997;23:16 25.
40. Chandra J, Gilbreath J, Freireich EJ, et al. Protease activation is
required for glucocorticoid-induced apoptosis in chronic lymphocytic leukemic lymphocytes. Blood 1997;90:3673 81.
41. Hicsonmez G, Ozbek N, Kale G, et al. Dramatic effect of highdose methylprednisolone on orbital granulocytic sarcoma. Pediatr Hematol Oncol 1996;13:18790.
42. Wyllie AH. Glucocorticoid-induced thymocyte apoptosis is associated with endogenous endonuclease activation. Nature
1980;284:555 6.
43. Ahlbom E, Gogvadze V, Chen M, et al. Prenatal exposure to
high levels of glucocorticoids increases the susceptibility of
cerebellar granule cells to oxidative stress-induced cell death.
Proc Natl Acad Sci U S A 2000;97:14726 30.
44. Reagan LP, Mc Ewens BS. Controverses surrounding
glucocorticoid-mediated cell death in the hippocampus. J Chem
Neuroanat 1997;13:149 67.
45. Almeida OF, Conde GL, Crochemore C, et al. Subtle shifts in the
ratio between pro- and antiapoptotic molecules after activation
of corticosteroid receptors decide neuronal fate. FASEB J 2000;
14:779 90.
46. Kondo F, Kondo Y, Makino H, Ogawa N. Delayed neuronal
death in hippocampal CA1 pyramidal neurons after forebrain
ischemia in hyperglycemic gerbils. Brain Res 2000;853:93 8.
47. Gisselsson L, Smith ML, Siesjo BK. Hyperglycemia and focal
brain ischemia. J Cereb Blood Flow Metab 1999;19:288 97.
48. de Courten-Myers GM, Kleinholz M, Wagner KR, Myers RE.
Fatal strokes in hyperglycemic cats. Stroke 1989;20:170715.
49. Chaney MA, Durazo-Arvizu RA, Nikolov MP, et al. Methylprednisolone does not benefit patients undergoing coronary
artery bypass grafting and early tracheal extubation. J Thorac
Cardiovasc Surg 2001;121:5619.
50. Chaney MA, Nikolov MP, Blakeman BP, Bakhos M. Attempting
to maintain normoglycemia during cardiopulmonary bypass
with insulin may initiate postoperative hypoglycemia. Anesth
Analg 1999;89:10915.
51. Bracken MB. Treatment of acute spinal cord injury with
methylprednisolone: results of a multicenter, randomized clinical trial. J Neurotrauma 1991;8:S4750; discussion S12.
52. Leist M, Nicotera P. Apoptosis, excitotoxicity, and neuropathology. Exp Cell Res 1998;239:183201.
53. Barrington K. The adverse neuro-developmental effects of postnatal steroids in the preterm infant: a systematic review of
RCTs. BMC Pediatr 2001;1:1.
54. Baud O. [Postnatal steroid treatment in preterm infants: risk/
benefit ratio]. J Gynecol Obstet Biol Reprod (Paris) 2005;34:
S118 26.

AMBULATORY ANESTHESIA

SOCIETY

FOR

AMBULATORY ANESTHESIA

SECTION EDITOR
PAUL F. WHITE

Total Shoulder Arthroplasty as an Outpatient Procedure

Using Ambulatory Perineural Local Anesthetic Infusion:
A Pilot Feasibility Study
Brian M. Ilfeld, MD, MS*, Thomas W. Wright, MD, F. Kayser Enneking, MD*,
Jennie A. Mace, MD*, Jonathan J. Shuster, PhD, Eugene H. Spadoni, PT,
Terese L. Chmielewski, PhD, PT, and Krista Vandenborne, PhD, PT
Departments of *Anesthesiology, Orthopaedics and Rehabilitation, Statistics, and Physical Therapy, University of
Florida, Gainesville, Florida

We investigated the feasibility of converting total shoulder arthroplasty (TSA) into an outpatient procedure using
ambulatory interscalene perineural ropivacaine infusion.
Of the patients of the first phase (n 8) who were required
to remain hospitalized for at least 1 postoperative night, 5
met discharge criteria in the recovery room. Of the subsequent patients of the second phase (n 6), all met discharge criteria in the recovery room after surgery, and 5
were discharged directly home. For all patients, postoperative pain was well controlled, oral opioid requirements

otal shoulder arthroplasty (TSA) results in severe

postoperative pain that traditionally requires
hospital admission to provide potent analgesia
with IV opioids for both baseline pain and to enable
effective rehabilitation (1 4). Perineural infusion or a
continuous nerve block does not always require hospitalization (5), and provides site-specific analgesia
Manufacturers donated the portable infusion pumps (Smiths
Medical, St. Paul, MN) and catheters (Arrow International, Reading,
PA) used for this investigation. Funding for this project provided by
the University of Florida Department of Anesthesiology, Arrow
International, and Smiths Medical. These two companies had no
input into any aspect of study design; data collection, analysis, and
interpretation; or manuscript preparation. Its contents are solely the
responsibility of the authors and do not necessarily represent the
official views of these entities.
Supported in part by General Clinical Research Center Grant
M01-RR00082 and National Institutes of Medicine, National Institute of General Medical Sciences (Grant K23-GM077026).
Some of the results of this investigation were presented at the
Annual Meeting of the American Society of Regional Anesthesia
and Pain Medicine, Orlando, FL, March 12, 2004; and Toronto,
Canada, April 22, 2005.
Accepted for publication April 25, 2005.
Address correspondence and reprint requests to Brian M. Ilfeld,
MD, Department of Anesthesiology, P.O. Box 100254, 1600 SW
Archer Rd., Gainesville, FL 32610-0254. Address e-mail to
bilfeld@ufl.edu.
DOI: 10.1213/01.ANE.0000180199.52383.CE
2005 by the International Anesthesia Research Society
0003-2999/05

and sleep disturbances were minimal, range-of-motion

consistently reached or exceeded the surgeons expectations, and patient satisfaction was high. These results suggest that TSA may be performed on an outpatient basis
using perineural local anesthetic infusion. Additional research is required to define the appropriate subset of patients and assess the incidence of complications associated
with this practice before its mainstream use.
(Anesth Analg 2005;101:1319 22)

with minor, if any, side effects (6,7). Portable infusion

pumps have been used to improve analgesia for outpatients after mild and moderately painful ambulatory
shoulder surgery (711). This pilot study was designed to evaluate the feasibility of converting TSA
into an ambulatory procedure using perineural interscalene local anesthetic infusion and portable infusion
pumps.

Methods
The investigation was divided into two phases: the
Hospitalized phase allowed for patient discharge as
early as the day after surgery, postoperative day
(POD) 1, whereas the Ambulatory phase allowed for
discharge home directly from the postanesthesia care
unit (PACU). The Hospitalized phase was prospectively
designated to conclude after 5 patients had met all
discharge criteria (Table 1) both in the PACU and POD
1, and subsequently completed their infusion successfully at home. Successful infusion was defined for
both phases as a patient 1) receiving acceptable analgesia as measured using a numeric rating pain scale
(NRS 4; scale of 0 10, 0 no pain, 10 worst
imaginable pain) throughout POD 7 (12); 2) avoiding
Anesth Analg 2005;101:131922

1319

1320

AMBULATORY ANESTHESIA ILFELD ET AL.

AMBULATORY TOTAL SHOULDER ARTHROPLASTY

ANESTH ANALG
2005;101:1319 22

Table 1. Discharge Criteria

Criteria
Analgesia
Opioids
Mobility
Orals
Oxygenation
Vital signs
Nausea
Estimated blood lossa
Medical issues

Details
Numeric rating pain score consistently 4
Patient required 5 mg of IV morphine (4 mg in PACU)
Able to ambulate without assistance or light-headedness
Tolerating oral liquids without nausea requiring treatment
Spo2 92% on room air at a respiratory rate 20/min
Temperature 38.5C
Heart rate preoperative baseline 10 (bpm)
Systolic blood pressure preoperative baseline 20 (mm Hg)
No nausea, or minimal nausea not requiring treatment
Estimated blood loss weight (kg) 10 mL
No medical issues requiring admission

PACU postanesthesia care unit.

a
Estimated blood loss was only a discharge criteria on postoperative day (POD) 0.

hospital readmission; and 3) achieving at least 50% of

the surgeons shoulder elevation and external rotation
goals (defined below) in the PACU and on POD 3. The
Ambulatory phase was designated to conclude after
successful completion of 5, of a maximum of 10,
patients.

Hospitalized Phase
After IRB approval, we prospectively enrolled patients scheduled for unilateral TSA. Subjects were required to 1) live within 2 h of the hospital; and 2) have
a caretaker who would remain with them during the
local anesthetic infusion and could return them to the
hospital if necessary. Exclusion criteria included any
contraindication to interscalene nerve block, any
known heart or lung disease, baseline Spo2 96%, a
history of opioid dependence or current chronic analgesic therapy, allergy to study medications, known
hepatic or renal insufficiency, peripheral neuropathy,
and morbid obesity (body mass index 40 kg/m2).
After written, informed consent, interscalene catheters (StimuCath; Arrow International, Reading, PA)
were placed using a technique described previously
(8). Forty milliliters of ropivacaine, 0.2%, with epinephrine, 100 g, was injected via the catheter with
gentle aspiration every 35 mL. For the surgical procedure, patients received a standardized general anesthetic without opioids.
Postoperatively, a perineural bolus was administered for an NRS 3 (20 mL of ropivacaine, 0.2%, with
epinephrine, 50 g), and an electronic, portable infusion pump (CADD-Legacy PCA; Smiths Medical, St.
Paul, MN) with a 600-mL reservoir of ropivacaine,
0.2%, was attached to the catheter (basal rate
7 mL/h, bolus 3 mL, lockout 60 min) (13). Patients received scheduled celecoxib, 100 mg twice
daily, and acetaminophen, 975 mg every 6 h. Rescue
opioid and route of administration were determined
by pain severity: oral oxycodone 5 mg (NRS 4), oral
oxycodone 10 mg (NRS 4 5), or IV morphine

2 4 mg (NRS 5). The patient and caretaker were

given verbal and written instructions on the use of the
pump and catheter along with physician telephone
and pager numbers.
In the PACU, both patients and their caretakers
received instruction on rehabilitation exercises by a
physical therapist. The primary indicator of functional
outcome after TSA is range-of-motion (14). For the
first 2 6 wk after surgery, patients undergo passive
elevation and external rotation up to surgeon-defined
maximums, or goals, to avoid damaging the rotator
cuff (3,15). Prospectively determined discharge criteria (Table 1) were subsequently evaluated and patients
discharged to the hospitals General Clinical Research
Center (GCRC), and then home the following morning
if they again met criteria for discharge.
Patients were telephoned beginning the night of
surgery, and each evening thereafter through the
night after catheter removal. On POD 3, patients returned to the GCRC to have their catheter site examined, local anesthetic reservoir replenished, and undergo a rehabilitation session with a physical
therapist. In the evening of POD 6, patients caretakers
removed the catheters using a pair of nonsterile
gloves, with a physician in telephone contact
throughout.

Ambulatory Phase
One change was made to the protocol after the completion of the Hospitalized phase: upon arrival in the
PACU, the perineural bolus (20 mL of ropivacaine,
0.2%, with epinephrine, 50 g) was administered for
an NRS 0 instead of an NRS 3.

Results
Hospitalized Phase
Eight patients were enrolled in this phase and had an
interscalene catheter placed successfully (Table 2).

ANESTH ANALG
2005;101:1319 22

AMBULATORY ANESTHESIA
ILFELD ET AL.
AMBULATORY TOTAL SHOULDER ARTHROPLASTY

1321

Table 2. Population Data

Age (yr)
Sex (female/male)
Height (cm)
Weight (kg)
Underlying etiology (osteoarthritis/failed previous TSA)
Surgery duration (min)
Estimated blood loss (mL)

Hospitalized Phase
(n 8)

Ambulatory Phase
(n 6)

62 12
4/4
173 9
88 18
8/0
189 85
305 (2002400)

68 8
2/4
169 9
80 19
4/2
174 44
325 (100600)

Values are reported as mean sd or median (minimummaximum) for parametric and nonparametric data, respectively.
TSA total shoulder arthroplasty.

One subject was erroneously included in the study

(exclusion criteria: asthma), and was removed from
the investigation. Five patients (63%) met all discharge
criteria both in the PACU and on POD 1, and were
discharged home on POD 1. Two patients (25%) did
not meet discharge criteria in the PACU. The first
because of an NRS 7 that required 4 mg of IV
morphine. This patient met discharge criteria the following morning and was discharged home on POD 1.
The second patient had an estimated blood loss of
2400 mL (Table 1), and was discharged home on POD
2.
For all patients, pain was well controlled with
5 mg of IV morphine (Fig. 1, Table 3). Postoperative
oral opioid requirements and sleep disturbances were
minimal (Table 3). All patients reached at least 50% of
the surgeon-defined range-of-motion goals (Table 3).
All subjects underwent successful perineural infusion
at home until their catheters were inadvertently dislodged (n 1, POD 4) or removed (n 6, POD 6).

Ambulatory Phase
Six patients were enrolled in this phase, all had an
interscalene catheter placed successfully, and all met
discharge criteria in the recovery room after surgery
(Table 2). However, one patient was admitted overnight secondary to operating room delays resulting in
surgery completion in the late evening. The remaining
5 patients were discharged directly from the recovery
room. All 6 patients underwent successful ambulatory
perineural infusion for 4 6 days. Postoperative pain
was well controlled (Fig. 1), oral opioid requirements
and sleep disturbances were minimal, range-ofmotion consistently reached at least 50% of the
surgeon-defined goals, and patient satisfaction was
high (Table 3). All subjects underwent successful perineural infusion at home until their catheters were
inadvertently dislodged (n 1, POD 4) or removed (n
5, POD 6).
There were no pump malfunctions or alarms and
caretakers for patients in both groups reported no
difficulty removing catheters at home.

Figure 1. Pre- and postoperative average (A), and worst (B) pain for
patients with an interscalene perineural ropivacaine infusion after
total shoulder arthroplasty. Pain was evaluated with a numeric
rating pain scale (NRS, 0 10, 0 no pain and 10 worst imaginable
pain). Data include both phases of the study and are expressed as
median (horizontal bar) with 25th75th (box) and 10th90th (whiskers) percentiles. For tightly clustered data (e.g., Panel A, postanesthesia care unit [PACU]), the median approximated the 10th and
25th percentile values. In these cases, the median is 0.0 and only the
75th and 90th percentiles are clearly noted.

1322

AMBULATORY ANESTHESIA ILFELD ET AL.

AMBULATORY TOTAL SHOULDER ARTHROPLASTY

ANESTH ANALG
2005;101:1319 22

Table 3. Opioid Requirements, Pain Scores During Physical Therapy, Sleep Disturbances, and Satisfaction Scores
POD
0
a

Median oral opioid tablet consumption

Percent of elevation goal achieved
Percent of external rotation goal
achieved
Average NRS during physical therapyb
Worst NRS during physical therapyb
Patients reporting difficulty sleeping (n)c
Awakenings for each patient (median)c
Satisfactiond

0
0.5
100 (65100)
80 (65100)
100 (100100) 100 (100100)

0
85 (75100)
100 (100100)

3.0

0.5 (03)
1 (08)
2
0

0
0

1
0

0
0

1
0

10 (1010)

3 (14)
6 (39)
1
0
10 (9.510.0)

2 (04)
3 (14)
0
0

Interscalene perineural ropivacaine, 0.2%, infusion provided from postoperative day (POD) 0 through the evening of PODs 4 6. All patients are included,
with the exception of pain scores and shoulder range-of-motion for POD 1, for which only patients remaining hospitalized on this day are included.
Not applicable (data not collected).
a
Oral opioid tablets: oxycodone 5 mg.
b
NRS numeric rating pain scale (0 10, 0 no pain and 10 worst imaginable pain); data presented as median (25th75th percentiles).
c
As a result of surgical pain.
d
Satisfaction 0 10 with 10 very satisfied; data presented as median (25th75th percentiles).

Discussion
For the patients of this pilot study who underwent
TSA and ambulatory perineural local anesthetic infusion, postoperative pain was well controlled with
baseline and breakthrough pain intensity below levels
previously reported for much smaller ambulatory orthopedic procedures (16). Patients also achieved
50% of the surgeon-defined maximal elevation and
external rotation without exception, and often reached
100% of this goal. This degree of comfort and shoulder
mobility was achieved with minimal oral opioid requirements and sleep disturbances, leading to a very
high rate of patient satisfaction.
Although this evidence demonstrates that TSA may
be performed in the ambulatory environment, it does
not define the appropriate subset of patients and incidence of complications associated with this practice
(e.g., local anesthetic toxicity or infection). Additional
research is required to define the appropriate subset of
patients and determine the complication incidence associated with this practice before its mainstream use.
The authors gratefully acknowledge the invaluable assistance of the
staff of both the Regional Anesthesia Block Room and General
Clinical Research Center, including Doug Theriaque, MS, for figure
compilation.

References
1. Borgeat A, Tewes E, Biasca N, Gerber C. Patient-controlled
interscalene analgesia with ropivacaine after major shoulder
surgery: PCIA vs PCA. Br J Anaesth 1998;81:6035.
2. Borgeat A, Schappi B, Biasca N, Gerber C. Patient-controlled
analgesia after major shoulder surgery: patient-controlled interscalene analgesia versus patient-controlled analgesia. Anesthesiology 1997;87:13437.

3. Brems JJ. Rehabilitation following total shoulder arthroplasty.

Clin Orthop Relat Res 1994:70 85.
4. Cameron B, Galatz L, Williams GR Jr. Factors affecting the
outcome of total shoulder arthroplasty. Am J Orthop 2001;30:
61323.
5. Ilfeld BM, Enneking FK. Continuous nerve blocks at home: a
review. Anesth Analg 2005;100:182233.
6. Klein SM, Grant SA, Greengrass RA, et al. Interscalene brachial
plexus block with a continuous catheter insertion system and a
disposable infusion pump. Anesth Analg 2000;91:1473 8.
7. Ilfeld BM, Morey TE, Wright TW, et al. Continuous interscalene
brachial plexus block for postoperative pain control at home: a
randomized, double-blinded, placebo-controlled study. Anesth
Analg 2003;96:1089 95.
8. Ilfeld BM, Morey TE, Wright TW, et al. Interscalene perineural
ropivacaine infusion: a comparison of two dosing regimens for
postoperative analgesia. Reg Anesth Pain Med 2004;29:9 16.
9. Macaire P, Gaertner E, Capdevila X. Continuous post-operative
regional analgesia at home. Minerva Anestesiol 2001;67:109 16.
10. Capdevila X, Macaire P, Aknin P, et al. Patient-controlled perineural analgesia after ambulatory orthopedic surgery: a comparison of electronic versus elastomeric pumps. Anesth Analg
2003;96:414 7.
11. Ilfeld BM, Enneking FK. A portable mechanical pump providing over four days of patient-controlled analgesia by perineural
infusion at home. Reg Anesth Pain Med 2002;27:100 4.
12. Cepeda MS, Africano JM, Polo R, et al. What decline in pain
intensity is meaningful to patients with acute pain? Pain 2003;
105:1517.
13. Ilfeld BM, Morey TE, Enneking FK. Delivery rate accuracy of
portable, bolus-capable infusion pumps used for patientcontrolled continuous regional analgesia. Reg Anesth Pain Med
2003;28:1723.
14. Cohen NP, Levine WN, Marra G, et al. Indwelling interscalene
catheter anesthesia in the surgical management of stiff shoulder:
a report of 100 consecutive cases. J Shoulder Elbow Surg 2000;
9:268 74.
15. Brown DD, Friedman RJ. Postoperative rehabilitation following
total shoulder arthroplasty. Orthop Clin North Am 1998;29:
535 47.
16. Rawal N, Hylander J, Nydahl PA, et al. Survey of postoperative
analgesia following ambulatory surgery. Acta Anaesthesiol
Scand 1997;41:101722.

A Randomized, Double-Blind Study of Granisetron Plus

Dexamethasone Versus Ondansetron Plus Dexamethasone to
Prevent Postoperative Nausea and Vomiting in Patients
Undergoing Abdominal Hysterectomy
Tong J. Gan,
Group

MD*,

Andrew Coop,

MBChB,

Beverly K. Philip,

MD,

and the Kytril Study

*Duke University Medical Center, Durham, North Carolina; Roche Laboratories Inc., Nutley, New Jersey; Brigham &
Womens Hospital, Boston, Massachusetts; See Appendix

In this randomized, double-blind study, we evaluated

whether small-dose granisetron (0.1 mg) plus dexamethasone 8 mg (GD) was as effective as ondansetron
4 mg plus dexamethasone 8 mg (OD) for preventing
vomiting during the 0 to 2 h after tracheal extubation in
patients undergoing abdominal hysterectomy requiring general anesthesia. Dexamethasone (D) was administered at induction of anesthesia, and granisetron (G)
or ondansetron (O) was given approximately 15 min
before tracheal extubation. Data on postoperative nausea and vomiting were collected at 0, 2, 6, and 24 h. For
the primary efficacy endpoint, most patients in each
group had no vomiting in the 0- to 2-h interval (82/87

hen therapeutic intervention to prevent postoperative nausea and vomiting (PONV) is

warranted, selective serotonin type 3 (5-HT3)
receptor antagonists are considered a first-line therapy
because of their efficacy and safety compared with
other drugs (1). The 5-HT3 receptor antagonists currently licensed in the United States for use in PONV
prophylaxis include dolasetron, granisetron (G), and
ondansetron (O). For patients at high risk of PONV,
use of a 5-HT3 receptor antagonist in combination
with another antiemetic drug may be justified to further reduce the likelihood of PONV (13).
Although a variety of drugs may have value in
combination therapy with 5-HT3 receptor antagonists
(2), droperidol and dexamethasone (D) have been the
most extensively investigated. However, the United
This research was sponsored by Roche Laboratories Inc., Nutley, NJ.
Accepted for publication July 13, 2005.
Address correspondence and reprint requests to Tong J. Gan, MD,
Box 3094, Department of Anesthesiology, Duke University Medical
Center, Durham, NC. Address e-mail to gan00001@mc.duke.edu.
DOI: 10.1213/01.ANE.0000180366.65267.F6
2005 by the International Anesthesia Research Society
0003-2999/05

[94%] for GD versus 86/89 [97%] for OD). Effectiveness of GD was demonstrated versus OD. Treatment groups were similar with regard to moderate or
severe nausea, complete response, rescue medication
use, and total control over 24 h. A descriptive assessment of adverse events showed that both combinations
were well tolerated with infrequent and similar incidences of adverse events. The combination of smalldose G administered just before tracheal extubation
plus D given at induction of anesthesia is an effective
alternative to OD in preventing vomiting during the
0- to 2-h interval after tracheal extubation.
(Anesth Analg 2005;101:13239)

States Food and Drug Administration (FDA)s black

box warning for droperidol concerning rare, potentially fatal cardiac adverse events has raised concerns
about its use as a first-line drug (2). In contrast, a
single bolus dose of D for PONV prophylaxis has not
been associated with serious side effects, although this
conclusion is based on limited data (4,5). With regard
to the effectiveness of D, a systematic review of trials
of D monotherapy found that the 8-mg dose showed
both early and late antiemetic efficacy (4). Combined
data from 3 trials conducted in adults, for example,
demonstrated that late vomiting occurred in 25% of 77
patients treated with D 8 mg compared with 48% of
the 77 patients who received placebo (relative benefit,
1.27; P 0.01). Given the safety and effectiveness of
both 5-HT3 receptor antagonists and D, the combination of these 2 classes of drugs may be an attractive
choice for prophylaxis in patients at high risk for
PONV (2,4).
Studies evaluating combination therapy in PONV
prophylaxis generally have used conventional doses
of the 5-HT3 receptor antagonists and D, but one study
Anesth Analg 2005;101:13239

1323

1324

AMBULATORY ANESTHESIA GAN ET AL.

5-HT3S PLUS DEXAMETHASONE FOR PONV PREVENTION

suggests that smaller doses of the 5-HT3 receptor antagonists may be effective. The FDA-labeled dose of G
recommended for PONV prophylaxis is 1 mg administered before induction of anesthesia or at the end of
anesthesia (6), but Mikawa et al. (7) have demonstrated that doses of G in excess of approximately
0.3 mg did not offer additional benefit. The results of
a pilot dose-ranging study suggested a trend of improved efficacy compared with placebo with G doses
as small as 0.1 mg when administered at the end of
surgery in preventing PONV during the 0- to 6-h
interval after abdominal hysterectomy (8). Thus, the
combination of small doses of G and conventional
doses of D may effectively prevent PONV.
Based largely on a study that found repeat doses of
O after failure of initial therapy to be ineffective (9), it
has been suggested that patients failing initial therapy
with a 5-HT3 receptor antagonist within the first 6
postoperative hours should not receive additional
5-HT3 receptor antagonist therapy (2). However, studies in patients with chemotherapy-induced nausea
and vomiting found that some of those who failed
initial therapy with O benefited from subsequent G
therapy (10,11). These results have yet to be investigated in patients with PONV.
In the present study, we evaluated the efficacy of
small-dose G in combination with 8 mg of D in preventing vomiting during the 0 to 2 h after tracheal
extubation in patients undergoing abdominal hysterectomy requiring general anesthesia. The primary efficacy analysis was a comparison of effectiveness between G 0.1 mg plus D 8 mg and O 4 mg plus D 8 mg,
using the FDA-labeled prophylaxis dose of O (12).

Methods
For this multicenter, prospective, randomized, doubleblind, parallel-group, effectiveness study, we obtained
approval from the IRB at each participating center. Patients provided written, informed consent before
participation.
Female patients 18 yr of age, with an ASA physical status of IIII, were eligible if they were scheduled
to undergo abdominal hysterectomy requiring general
anesthesia. Patients were excluded if they 1) had
known hypersensitivity or contraindication to study
medications, 2) had chronic nausea and vomiting or
experienced retching, vomiting, or moderate or severe
nausea in the 24 h before anesthesia, 3) had received
an antiemetic drug or a drug with antiemetic properties during the 24 h before anesthesia, 4) had a body
mass index 36, 5) were pregnant or breast feeding, or
6) had a condition requiring chronic opioid use.
Patients were stratified at randomization by history
of motion sickness and PONV and by current smoking
status (within the prior 30 days), and were randomly

ANESTH ANALG
2005;101:13239

assigned to receive either G 0.1 mg plus D 8 mg (GD)

or O 4 mg plus D 8 mg (OD). The randomization
lists were computer-generated, centrally determined,
and the randomization numbers were allocated sequentially in the order in which patients were enrolled. Study medications were prepared by the site
pharmacist, who was not involved in any other part of
the study, and presented to blinded investigators as
identical 2-mL filled syringes.
For eligible patients, demographic information was
collected and a physical examination was performed.
A standardized anesthesia regimen was followed. Premedication, if desired, was with IV midazolam 12 mg.
General anesthesia was induced with IV propofol (up to
2.5 mg/kg), with or without lidocaine (up to 50 mg IV),
and was maintained with isoflurane titrated between
0.6% and 2.5% and nitrous oxide 50%. Neuromuscular blockers were administered to facilitate endotracheal intubation and intraoperative muscle relaxation.
Fentanyl at doses of up to 6 g kg1 h1 IV was
administered. D was given immediately after induction of anesthesia in both treatment groups. Neuromuscular blockade was reversed with IV neostigmine
(up to 0.07 mg/kg) and glycopyrrolate (up to 0.02 mg/
kg). G or O was administered approximately 15 min
before tracheal extubation (defined as end of surgery).
Morphine (110 mg IV or IM) or fentanyl (up to 50 g
IV) was permitted as needed for the management of
postoperative pain.
The time of each vomiting episode and the time and
intensity of each nausea episode were recorded immediately before anesthesia and 2, 6, and 24 h after
tracheal extubation. An episode of vomiting was defined as either vomiting (expulsion of stomach contents) or retching (an involuntary attempt to vomit but
not productive of stomach contents). Vomiting and/or
retching episodes separated by 5 min were recorded
as a single episode. The intensity of each nausea episode was graded as mild (discomfort noticed but no
disruption of anticipated normal activity), moderate
(discomfort sufficient to reduce or affect anticipated
normal activity), or severe (inability to perform anticipated normal daily activity). Rescue medication could
be administered to any patient who experienced an
episode of moderate or severe nausea, an episode of
vomiting, or who requested rescue medication. The
initial rescue medication was G 0.1 mg administered
as a single IV push. Nausea and vomiting assessments
were made 30 min after rescue medication administration, and response was defined as improvement or
resolution of PONV symptoms. Subsequent PONV
symptoms could be treated with an alternative rescue
medication at the discretion of the investigator. If rescue medication was used, both the time of administration and type of medication were recorded. Adverse events were evaluated and recorded by the
investigator during the entire observation period.

ANESTH ANALG
2005;101:13239

AMBULATORY ANESTHESIA
GAN ET AL.
5-HT3S PLUS DEXAMETHASONE FOR PONV PREVENTION

1325

Table 1. Patient Demographic and Baseline Characteristics (Efficacy Population)

Characteristic

Granisetron 0.1 mg
dexamethasone
8 mg (n 87)

Ondansetron 4 mg
dexamethasone
8 mg (n 89)

48 12

48 10

56 (64)
13 (15)
15 (17)
3 (3)
70 14
14 (16)
32 (37)
24 (28)
24 (28)

54 (61)
22 (25)
11 (12)
2 (2)
74 14
19 (21)
37 (42)
22 (25)
23 (26)

Age (yr), mean sd

Race, n (%)
Caucasian
Black
Hispanic
Other
Weight (kg), mean sd
Smokers, n (%)
Alcohol consumers, n (%)
History of motion sickness, n (%)
History of PONV, n (%)
PONV postoperative nausea and vomiting, sd standard deviation.

The primary efficacy end-point was the proportion

of patients with no vomiting during the 0 to 2 h after
tracheal extubation. This time period was chosen because it reflects the typical time spent by a patient in a
postanesthesia care unit and because comparative
data concerning the antiemetic efficacy of OD were
available for this interval (13). Effectiveness was defined when the lower bound of the two-sided 95%
confidence interval, using the normal distribution approximation, for the proportion difference in the primary efficacy end-point (GD OD) crossed
15%. This threshold has been used in previous research (14,15) and was chosen to be clinically relevant
because of the wide variability in the rates of emesis
which is based on multiple patient and procedural
factors.
Secondary efficacy end-points included proportions
of patients in the 0 6 and 0 24 h intervals with no
vomiting, as well as proportions of patients in the 0 2,
0 6, and 0 24 h intervals after tracheal extubation: 1)
with no moderate or severe nausea, 2) with complete
response (no moderate or severe nausea and no vomiting), 3) using rescue medication, and 4) having total
control (no moderate or severe nausea, no vomiting,
and no rescue medication use). Additional secondary
efficacy variables were times to first vomiting episode,
first moderate or severe nausea episode, use of first
rescue medication within 24 h, and proportions of
patients responding to the initial dose of G as rescue
medication.
The proportion difference (GD OD) and a
2-sided 95% confidence interval (CI), using the normal
distribution approximation, were calculated and results described for all secondary efficacy end-points,
with no emphasis on comparing the lower bound of
the 2-sided 95% CI with the margin of 15%.
Time-to-event secondary efficacy variables were analyzed using survival analysis. A Cox proportional
hazard regression analysis was conducted with the
study center and stratification variables as covariates;

patients not experiencing the event were included as

censored observations, and the estimated hazard ratio
and 95% CIs were calculated to measure treatment
differences.
Adverse events were summarized by body system,
treatment group, severity (serious or nonserious), intensity (mild, moderate, severe, or life-threatening),
and relationship to study drug. The efficacy evaluable
population included all randomized patients receiving
at least 1 dose of either G or O and having 2-h nausea
and vomiting assessments. The safety population included all randomized patients receiving at least one
dose of study medication (either G, O, or D) and
having at least one postbaseline safety measurement.
Based on a previous study (13), 90% of patients in
both treatment groups were expected to have no vomiting in the 0- to 2-h interval. Recruitment of 85 evaluable patients per treatment group was planned to
provide a 90% power to exclude the possibility of a
treatment difference of 15% or more in proportions of
patients with no vomiting during the 0- to 2-h interval
for a 1-sided test at a 0.025 level of significance.

Results
In all, 210 patients were enrolled (GD, 101; OD,
109) from October 30, 2003 through April 27, 2004 at 19
centers in the United States. Of the 210 enrolled patients, 176 patients (87 GD, 89 OD) were considered evaluable for efficacy. The 34 patients excluded
from efficacy analyses either did not receive at least
1 dose of study medication (GD, n 6; OD, n
10), received D but not either G or O (GD, n 6;
OD, n 10), or had no 2-h nausea/vomiting assessment (GD, n 2); no patient was excluded for
reasons related to nausea or vomiting.
Treatment groups were generally similar with regard to baseline characteristics (Table 1). Diagnoses
for surgery and types of surgery were comparable

1326

AMBULATORY ANESTHESIA GAN ET AL.

5-HT3S PLUS DEXAMETHASONE FOR PONV PREVENTION

ANESTH ANALG
2005;101:13239

Table 2. Surgery Characteristics

Elapsed time
Induction to administration of dexamethasone (min)
Administration of dexamethasone to administration of
granisetron or ondansetron (min)
Administration of granisetron or ondansetron to time of
tracheal extubation (min)
Tracheal extubation to first oral intake postsurgery (h)
Tracheal extubation to discharge from PACU (h)
Maintenance of anesthesia (min)

Granisetron 0.1 mg
dexamethasone
8 mg (n 87)

Ondansetron 4 mg
dexamethasone
8 mg (n 89)

66
110 65

56
112 51

21 9

23 11

16 11
33
123 65

16 10
46
127 50

Data are mean sd.

PACU postanesthesia care unit; sd standard deviation.

across groups. Eight patients treated with GD and 2

treated with OD experienced mild nausea within
24 h before anesthesia. Times from administration of
D to administration of G or O and times from administration of G or O to tracheal extubation were similar
between treatment groups (Table 2).
For the primary efficacy end-point, 82 of 87 patients
(94%) in the GD group and 86 of 89 (97%) of those
receiving OD had no vomiting in the 0- to 2-h interval. The treatment difference (GD OD) was
2.4% with a 2-sided 95% CI of 8.5% to 3.8%. Because the lower bound of the 2-sided 95% CI (8.5%)
was more than the preset threshold of 15%, it was
concluded that G 0.1 mg plus D 8 mg was not inferior
to O 4 mg plus D 8 mg in preventing vomiting during
the 0- to 2-h interval after surgery.
Fewer patients had no vomiting in the 0 6 and
0 24 h intervals, but patients with no vomiting were
similar between treatments at both 0 6 and 0 24 h
(Table 3). The mean time to the first vomiting episode
was longer for patients receiving OD than GD
(19.6 versus 10.5 h, respectively; hazard ratio 1.34 [95%
CI: 0.6, 2.9]) primarily because the first vomiting episode for 2 patients in the OD group occurred 20 h
after tracheal extubation. However, the hazard ratio
and 95% CI indicated that times to first vomiting
episodes were similar between treatments.
Patients with no moderate or severe nausea episodes and having complete responses were similar
between treatments at all time intervals (Table 3).
Patients meeting these secondary efficacy outcomes
decreased over time, and maximum between treatment differences was 7% at all time intervals. Mean
times to first moderate or severe nausea episodes were
similar between treatments (hazard ratio 1.09 [95% CI:
0.7, 1.6]).
A comparable number of patients in both groups
used rescue medication (Table 3). Additionally, mean
times to first rescue medication use were also similar
(hazard ratio 1.31 [95% CI: 0.9, 2.0]). First use of rescue
medication decreased at a similar rate in both groups

in the first 10 h after tracheal extubation, and few

patients used rescue medication for the first time after
10 h (2 in the GD group, 4 receiving OD). Total
control was achieved by a similar number of patients
in each group at each time interval (Table 3).
Forty-eight patients treated with GD and 40
treated with OD received G as rescue medication.
One patient receiving OD received prochlorperazine
instead of G as first rescue medication and was excluded from these analyses. (Note that this patient was
included in overall analyses of rescue medication use
[Table 3].) At 30 min postdose, nausea and vomiting
symptoms had either resolved or improved in 38 of 48
patients (79%) who initially received GD and in 31 of
40 patients (78%) treated with OD. In all, 34 of 87
patients (39%) initially receiving GD and 28 of 89
(31%) OD-treated patients used more than one rescue medication in the 24 h after tracheal extubation.
The 194 patients receiving at least 1 dose of study
medication were included in safety analyses (GD, n
95; OD, n 99). Both regimens were well tolerated, and the incidence of adverse events was similar
between treatments (Table 4). Investigators judged
most adverse events to be mild or moderate in intensity and unrelated to study medication. No deaths
occurred during the study. Serious adverse events
were reported by two patients receiving GD (one
patient each: abdominal pain, postoperative fever)
and by four receiving OD (two patients: pneumonia;
one patient each: increased body temperature, pyrexia). All serious adverse events were judged to be
unrelated to study medication. One patient in the
GD group withdrew as a result of an adverse event
(vaginal hemorrhage) which was judged to be unrelated to study medication. No clinically meaningful
mean changes in vital signs were noted. Adverse
events occurring in 5% of patients in GD versus
OD were: pruritus (4% and 8%, respectively), headache (2% and 8%, respectively), hypertension (2% and
6%, respectively), and bradycardia (5% and 0%,
respectively).

ANESTH ANALG
2005;101:13239

AMBULATORY ANESTHESIA
GAN ET AL.
5-HT3S PLUS DEXAMETHASONE FOR PONV PREVENTION

1327

Table 3. Summary of Efficacy Results for 0 2-, 0 6-, and 0 24-h Intervals: Secondary Outcomes (Efficacy Population)

Efficacy outcome
No vomiting
06 h
024 h
No moderate or severe nausea
02 h
06 h
024 h
Complete response
02 h
06 h
024 h
Required rescue medication
02 h
06 h
024 h
Total control
02 h
06 h
024 h

Granisetron 0.1 mg
dexamethasone
8 mg (n 87)

Ondansetron 4.0 mg
dexamethasone
8 mg (n 89)

Proportion differencea
(95% CI)

76 (87)
72 (83)

83 (93)
77 (87)

5.9 (14.6, 2.8)

3.8 (14.4, 6.9)

66 (76)
53 (61)
42 (48)

67 (75)
59 (66)
45 (51)

0.6 (12.1, 13.3)

5.4 (19.6, 8.8)
2.3 (17.1, 12.5)

65 (75)
51 (59)
40 (46)

67 (75)
59 (66)
44 (49)

0.6 (13.4, 12.2)

7.7 (21.9, 6.6)
3.5 (18.2, 11.3)

21 (24)
35 (40)
48 (55)

19 (21)
27 (30)
41 (46)

2.8 (9.6, 15.2)

9.9 (4.2, 23.9)
9.1 (5.6, 23.8)

64 (74)
50 (57)
38 (44)

66 (74)
57 (64)
42 (47)

0.6 (13.6, 12.4)

6.6 (21.0, 7.8)
3.5 (18.2, 11.2)

Data are number (%) unless otherwise noted.

CI confidence interval, complete response no moderate or severe nausea and no vomiting, total control no moderate or severe nausea, no vomiting,
and no rescue medication use.
a
Difference between groups in proportions of patients experiencing the outcome and 95% CI (normal distribution approximation).

Table 4. Summary of Adverse Events (Safety Population)

Total number (%) patients with 1 adverse event

Total number adverse events
Number (%) patients with drug-related adverse event
Number (%) patients with severe adverse event
Number patient deaths
Number patients with serious adverse events
Number patients withdrawing as a result of adverse events

Discussion
In this randomized, double-blind effectiveness study,
G 0.1 mg plus D 8 mg was shown to be as effective as
the current standard prophylactic antiemetic combination, O 4 mg plus D 8 mg, in patients experiencing no
vomiting during 0 2 hours after tracheal extubation.
As expected (13), 90% of patients in each treatment
group had no vomiting during the 0- to 2-hour interval. Treatment groups also were comparable at all
measurement times with regard to secondary efficacy
outcomes, and both therapeutic regimens were well
tolerated.
The potential advantages of combination therapy
using drugs that act on different pathways in the
emetic response include improved efficacy, extended
duration of the antiemetic effect, the ability to combine
drugs with greater antinausea versus greater antiemetic effects, and the possibility of using smaller

Granisetron 0.1 mg
dexamethasone
8 mg (n 95)

Ondansetron 4.0 mg
dexamethasone
8 mg (n 99)

35 (37)
54
0
9 (9)
0
2
1

41 (41)
65
3 (3)
8 (8)
0
4
0

doses of individual drugs compared with monotherapies. A large trial of 6 PONV interventions found 26%
reductions in relative risks of nausea and vomiting for
each additional antiemetic administered (3). Optimal
use of a multimodal approach depends on the individual and combined efficacy and safety of the drugs
selected, the dosages used, and the timing of administration. Both G and O administered in combination
with D are more effective than the individual drugs
(2,4,16 20). With regard to timing, the 5-HT3 receptor
antagonists are most effective when administered at the
end of surgery whereas D seems to be most effective
when given before the induction of anesthesia (2). When
a range of doses is equally effective, the smallest dose is
recommended (2). Taking these published criteria for
optimal use together, the results of the present study
suggest that the combination of small-dose G administered at the end of surgery plus D given at induction is

1328

AMBULATORY ANESTHESIA GAN ET AL.

5-HT3S PLUS DEXAMETHASONE FOR PONV PREVENTION

an effective alternative to similarly administered OD

for patients requiring prevention of PONV.
A strength of our study was its randomized, doubleblind, parallel-group, active-control design which allowed combination regimens to be compared directly.
The design was limited, however, by the fact that combination therapies were not compared either with their
individual components or with placebo. Previous research (16 20) has provided comparisons, but it would
be informative to assess a range of therapeutic options in
a single study. The study was designed to evaluate
whether small-dose G plus D was noninferior to the
standard dose of O in combination with D in preventing
vomiting during the 0- to 2-hour period after tracheal
extubation. This study was not designed to investigate
superiority of either combination. In addition, although
the preset equivalence difference of 15% for the primary
efficacy outcome has been used previously (14,15), we
also believe this range to be reasonable and clinically
relevant because of the wide variability in the rates of
emesis based on multiple patient and procedural factors
such as differences in clinical practice, use of analgesia,
patient mobilization, resumption of oral intake, and
other factors. We also recognize that the impact on actual
clinical practice of any difference chosen depends on the
incidence, severity, duration, and cost of the outcome
factors that were unaccounted for in the present research. Although G has a longer half-life (8 hours) than
O, it is unclear how serum half-life correlates with clinical efficacy, which may be determined by receptor binding kinetics (6).
This is the first study to report the results of the use of
G as rescue medication in high-risk patients receiving
either GD or OD as PONV prophylaxis. Descriptively, 79% and 78%, respectively, of patients reported
improvement or resolution of PONV symptoms within
30 minutes, but the study was not designed as a treatment crossover and therefore did not adequately address the efficacy of readministration of either drug. The
absence of a placebo group further limits the ability to
assess whether the reduction of symptoms was the result
of the passage of time alone.
Although 80% of patients in each treatment group
had no vomiting during the 24-hour study period,
fewer than half in either group experienced complete
control. Continued research is needed to identify optimal combinations and dose levels of antiemetic
drugs for use in particular groups of high-risk patients. In conclusion, the efficacy of GD was found to
be noninferior to that of OD in preventing vomiting
during the 0- to 2-hour interval after surgery.
The authors thank Yin-Miao Chen, PhD, and Douglas Eberhardt, MS,
for statistical support; Kelly Celuch, PharmD, Catherine Chin, MS, and
Leena-Patel Shah, PharmD, for overall study management; and Jane G.
Murphy, PhD, and Linda E. Whetter, PhD, for editorial assistance.

ANESTH ANALG
2005;101:13239

Appendix 1
The following investigators comprised the Kytril
Study Group and had a substantial role in the conduct
of this study: Richard Beers, MD, State University of
New York Upstate Medical University, Syracuse, NY;
Kumar G. Belani, MD, University of Minnesota, Minneapolis, MN; Keith Candiotti, MD, University of Miami School of Medicine, Miami, FL; Jacques Chelly,
MD, University of Pittsburgh Medical Center Shadyside, Pittsburgh, PA; Patricia Dalby, MD, Magee Womens Hospital of the University of Pittsburgh Medical
Center, Pittsburgh, PA; Robert DAngelo, MD, Wake
Forest University, Winston-Salem, NC; Dennis Doblar,
MD, University of Alabama Birmingham, Birmingham, AL; Ashraf Habib, MD, Duke University,
Durham, NC; Charles Hantler, MD, Washington University Medical CenterBarnes Jewish Hospital, St.
Louis, MO; Hugh C. Hemmings, Jr., MD, Weill Medical College of Cornell University, New York, NY;
Daniel Katz, MD, California Anesthesia Associates,
Newport Beach, CA; Robert Knapp, DO, Brigham and
Womens Hospital, Boston, MA; Anthony Kovac, MD,
University of Kansas School of Medicine, Kansas City,
KS; Timothy I. Melson, MD, Helen Keller Hospital,
Sheffield, AL; Harold Minkowitz, MD, Memorial
Hermann-Memorial City Hospital, Houston, TX;
Naila Moghul, MD, Brigham and Womens Hospital,
Boston, MA; James Philip, MD, Brigham and Womens
Hospital, Boston, MA; Kenneth Rosenfeld, MD, State
University of New York at Stony Brook, Stony Brook,
NY; Lee Silk, MD, Brigham and Womens Hospital,
Boston, MA; Neil Singla, MD, Huntington Memorial
Hospital, Pasadena, CA; Richard A. Steinbrook, MD,
Beth Israel Deaconess Medical Center, Boston, MA;
Tracey Stierer, MD, Johns Hopkins University Hospital, Baltimore, MD.

References
1. ASHP therapeutic guidelines on the pharmacologic management of nausea and vomiting in adult and pediatric patients
receiving chemotherapy or radiation therapy or undergoing
surgery. Am J Health Syst Pharm 1999;56:729 64.
2. Gan TJ, Meyer T, Apfel CC, et al. Consensus guidelines for
managing postoperative nausea and vomiting. Anesth Analg
2003;97:6271.
3. Apfel CC, Korttila K, Abdalla M, et al. A factorial trial of six
interventions for the prevention of postoperative nausea and
vomiting. N Engl J Med 2004;350:244151.
4. Henzi I, Walder B, Tramer MR. Dexamethasone for the prevention of postoperative nausea and vomiting: a quantitative systematic review. Anesth Analg 2000;90:186 94.
5. Thomas R, Jones N. Prospective randomized, double-blind comparative study of dexamethasone, ondansetron, and ondansetron plus dexamethasone as prophylactic antiemetic therapy in
patients undergoing day-case gynaecological surgery. Br J Anaesth 2001;87:688 92.

ANESTH ANALG
2005;101:13239

6. Kytril, granisetron hydrochloride prescribing information

[package insert]. Nutley, NJ: Roche Laboratories; 2002.
7. Mikawa K, Takao Y, Nishina K, et al. Optimal dose of granisetron for prophylaxis against postoperative emesis after gynecological surgery. Anesth Analg 1997;85:652 6.
8. DAngelo R, Minkowitz H, Dalby P, et al. A randomized,
double-blind, dose-ranging, pilot study of intravenous granisetron in the prevention of postoperative nausea and vomiting
(PONV) in patients undergoing abdominal hysterectomy. Eur J
Anesthesiol 2005;22:774 9.
9. Kovac AL, OConnor TA, Pearman MH, et al. Efficacy of repeat
intravenous dosing of ondansetron in controlling postoperative
nausea and vomiting: a randomized, double-blind, placebocontrolled multicenter trial. J Clin Anesth 1999;11:4539.
10. Carmichael J, Keizer HJ, Cupissol D, et al. Use of granisetron in
patients refractory to previous treatment with antiemetics. Anticancer Drugs 1998;9:3815.
11. de Wit R, de Boer AC, vd Linden GH, et al. Effective cross-over
to granisetron after failure to ondansetron, a randomized double blind study in patients failing ondansetron plus dexamethasone during the first 24 hours following highly emetogenic
chemotherapy. Br J Cancer 2001;85:1099 101.
12. Zofran, ondansetron hydrochloride prescribing information
[package insert]. Research Triangle Park, NC: GlaxoSmithKline;
2002.
13. Sanchez-Ledesma MJ, Lopez-Olaondo L, Pueyo FJ, et al. A
comparison of three antiemetic combinations for the prevention
of postoperative nausea and vomiting. Anesth Analg 2002;95:
1590 5.
14. Eisenberg P, Figueroa-Vadillo J, Zamora R, et al. Improved
prevention of moderately emetogenic chemotherapy-induced
nausea and vomiting with palonosetron, a pharmacologically
novel 5-HT3 receptor antagonist: results of a phase III, singledose trial versus dolasetron. Cancer 2003;98:2473 82.

AMBULATORY ANESTHESIA
GAN ET AL.
5-HT3S PLUS DEXAMETHASONE FOR PONV PREVENTION

1329

15. Gralla R, Lichinitser M, Van Der Vegt S, et al. Palonosetron

improves prevention of chemotherapy-induced nausea and
vomiting following moderately emetogenic chemotherapy: results of a double-blind randomized phase III trial comparing
single doses of palonosetron with ondansetron. Ann Oncol
2003;14:1570 7.
16. Biswas BN, Rudra A. Comparison of granisetron and granisetron plus dexamethasone for the prevention of postoperative
nausea and vomiting after laparoscopic cholecystectomy. Acta
Anaesthesiol Scand 2003;47:79 83.
17. Coloma M, White PF, Markowitz SD, et al. Dexamethasone in
combination with dolasetron for prophylaxis in the ambulatory
setting: effect on outcome after laparoscopic cholecystectomy.
Anesthesiology 2002;96:1346 50.
18. Elhakim M, Nafie M, Mahmoud K, Atef A. Dexamethasone
8 mg in combination with ondansetron 4 mg appears to be the
optimal dose for the prevention of nausea and vomiting after
laparoscopic cholecystectomy. Can J Anaesth 2002;49:922 6.
19. Lopez-Olaondo L, Carrascosa F, Pueyo FJ, et al. Combination of
ondansetron and dexamethasone in the prophylaxis of postoperative nausea and vomiting. Br J Anaesth 1996;76:835 40.
20. McKenzie R, Tantisira B, Karambelkar DJ, et al. Comparison of
ondansetron with ondansetron plus dexamethasone in the prevention of postoperative nausea and vomiting. Anesth Analg
1994;79:961 4.

Ondansetron, Orally Disintegrating Tablets Versus

Intravenous Injection for Prevention of Intrathecal MorphineInduced Nausea, Vomiting, and Pruritus in Young Males
Arash Pirat, MD, Senay F. Tuncay,
Gulnaz Arslan, MD

MD,

Adnan Torgay,

MD,

Selim Candan,

MD,

and

Baskent University Faculty of Medicine, Department of Anesthesiology, Ankara, Turkey and Department of
Anesthesiology, Air Force Hospital, Etimesgut, Ankara, Turkey

In this study we compared the efficacy of orally disintegrating tablets (ODT) and IV ondansetron for
preventing spinal morphine-induced pruritus and
postoperative nausea and vomiting (PONV) in
healthy young male patients. Patients who received
bupivacaine with 0.20 mg morphine for spinal anesthesia were randomly assigned to the ODT group
(ODT ondansetron 8 mg, n 50), the IV group (4 mg
ondansetron IV, n 50), or the placebo group (n
50). Each individual was assessed for pruritus, postoperative nausea and vomiting, and pain at 0, 2, 6, 12,
18, and 24 h after surgery using three distinct visual
analog scales. The frequencies of postoperative nausea and vomiting and frequencies of requirement for
rescue antiemetic and antipruritic were recorded.
There were no significant differences among the
three groups with respect to incidence or severity of

ntrathecal (IT) morphine provides an intense and

long-lasting analgesic effect and is often used for
pain relief after surgery (1). Unfortunately, spinal
morphine is associated with a frequent incidence of
pruritus and emetic symptoms (2,3), which can limit
its use. Several drugs have been used to prevent or
treat these side effects (4 7). Of these various drugs,
5-hydroxytryptamin-3 (5HT3) receptor antagonists are
of special interest because they lack sedative effects,
do not reverse the analgesic effect of spinal morphine,
and are well tolerated by patients. Studies have indicated that IV ondansetron (a 5HT3 receptor antagonist) can reduce the frequency of IT morphine-induced

Accepted for publication April 27, 2005.

Address correspondence and reprint requests to Arash Pirat, MD,
niversitesi Hastanesi Anesteziyoloji Anabilim Dal 10.
Baskent U
Sok. No: 45 Bahcelievler 06490, Ankara, Turkey. Address electronic
mail to arashp@baskent-ank.edu.tr.
DOI: 10.1213/01.ANE.0000180830.12355.D9

1330

Anesth Analg 2005;101:13306

PONV or postoperative pain visual analog scale

scores. The incidences of pruritus in the ODT (56%)
and IV (66%) groups were significantly different
from that in the placebo group (86%) (P 0.02 for
both). Only the ODT group had significantly lower
mean pruritus visual analog scale scores at 0, 2, 6, and
12 h postsurgery than the placebo group (P 0.023
for all). The frequency of requirement for rescue antipruritic was significantly less in the ODT group than
the placebo group (P 0.013). Both ODT ondansetron 8 mg and IV ondansetron 4 mg are more effective
than placebo for preventing spinal morphineinduced pruritus, but neither form of this agent reduces spinal morphine-induced postoperative nausea and vomiting in this patient group.
(Anesth Analg 2005;101:1330 6)

pruritus and postoperative nausea and vomiting

(PONV) (8 10).
Orally disintegrating tablets (ODT) of ondansetron,
which disperse rapidly when placed on the tongue, have
been used to treat radiotherapy- and chemotherapyinduced nausea and vomiting (11,12). This new formulation of ondansetron is also attractive as a drug for
anesthesia practice, where the need for preoperative fasting and nil per os orders precludes the use of traditional
oral medications. The literature contains a few reports on
use of the ODT form of ondansetron (hereafter referred
to as ODT ondansetron) for PONV (13,14); however,
there are no published data on the use of ODT ondansetron for preventing IT opioid-induced pruritus and
PONV, and it is not clear whether the ODT formulation
of this drug is as effective as the IV form.
In this study, we compared the efficacy of ODT
ondansetron 8 mg, IV ondansetron 4 mg, and placebo
for preventing IT morphine-induced pruritus and
PONV in a group of young men who underwent
minor elective surgeries.
2005 by the International Anesthesia Research Society
0003-2999/05

ANESTH ANALG
2005;101:1330 6

AMBULATORY ANESTHESIA
PIRAT ET AL.
ONDANSETRON FOR INTRATHECAL MORPHINE-INDUCED PONV AND PRURITUS

Methods
The randomized, prospective, placebo-controlled
study was approved by the institutional ethics committee, and informed consent was obtained from each
participant. The subjects were 150 ASA I young men
who were scheduled for elective operations for inguinal hernia, cord hydrocele, and pilonidal sinus. The
exclusion criteria were history of motion sickness or
PONV, preoperative pruritus, treatment with opioids
or antiemetics within 48 hours of surgery, hypersensitivity to ondansetron, morphine, or bupivacaine, and
contraindication for or refusal of spinal anesthesia.
Cases in which dural puncture could not be performed or opioids were required to control intraoperative or postoperative pain were also excluded. None
of the patients was premedicated. Each was randomly
allocated to one of three groups using computergenerated random numbers.
Patients in the ODT group (n 50) received ODT
ondansetron 8 mg and 5 mL of normal saline IV.
Patients in the IV group (n 50) received IV ondansetron 4 mg in 5 mL saline and an oral placebo. Those
in the placebo group (n 50) received 5 mL normal
saline IV and an oral placebo. All substances (ondansetron or placebo) were administered 10 min before
spinal anesthesia. Unfortunately, at the time the study
was conducted, the manufacturer of ODT ondansetron stated it was unable to provide us with placebo
pills. Therefore, we used a mild peppermint candy
that disperses when placed on the tongue as the oral
placebo. Although this was not a one-to-one placebo
for ODT ondansetron, we believe it was a suitable
substitute because none of the patients had used ODT
ondansetron previously.
Patients, anesthesiologists who attended the intraoperative care, and nurse anesthesiologists who performed the postoperative evaluations were unaware
of patient group allocations. To ensure the study was
blinded, the nurse who prepared and administered
the study drugs was not involved in patient care.
Standard intraoperative monitoring with electrocardiogram (ECG), noninvasive arterial blood pressure,
and pulse oximetry was used. Before spinal anesthesia, normal saline 15 mL/kg IV was given over 30 to
60 min. Then the patient was placed in sitting position
and a 25-gauge Quincke needle was placed in the L2-3
or L3-4 interspace using a midline approach. Once
successful dural puncture was confirmed by observation of a free flow of cerebrospinal fluid, an IT injection of 0.5% hyperbaric bupivacaine (12.5 mg if 75 kg
body weight and 15.0 mg if 75 kg) with 0.2 mg of
preservative-free morphine was administered with
the bevel of the needle oriented caudally. In all cases,
two anesthesiologists who were blinded as to the

1331

study performed the spinal anesthesia and determined the level of sensory blockade by pinprick testing. For intraoperative sedation, each patient received
an IV injection of midazolam 0.05 mg/kg before spinal
anesthesia, and this was repeated as required during
the surgery. All patients received supplemental oxygen via nasal cannulae (rate 35 L/min) during the
procedure.
Intraoperative hemodynamic and respiratory complications were recorded. Hypotension was defined as
a 15% decrease in systolic blood pressure from baseline or the appropriate age-adjusted values. Bradycardia was defined as heart rate less than 40 bpm or as an
inappropriately slow heart rate despite hypovolemia.
Hypoxia was defined as an oxygen saturation value
90%. Hypotension was treated with IV boluses of
ephedrine 0.1 mg/kg and normal saline 5 mL/kg, and
the same doses were repeated if required. Bradycardia
was treated with atropine 1 mg IV.
At the end of the surgery, patients were transferred
to the postanesthesia care unit (PACU), where nurse
anesthesiologists who were unaware of the group allocations cared for them. A patient was transferred to
the surgery ward from the PACU when the following
criteria were met: hemodynamic and respiratory stability, full recovery from motor block, sensory block
no higher than the T10 level, no or minimal pain, no or
minimal nausea, no or minimal pruritus, and absence
of vomiting or surgical bleeding. Postoperative pain
was treated with an IM injection of diclofenac sodium
100 mg. This was administered if the patient asked for
an analgesic medication or his pain visual analog scale
(VAS) score was 5 (see VAS details below).
Each patient was assessed in the PACU and at 2, 6,
12, 18, and 24 h postoperatively for frequency and
severity of PONV, pruritus, and postoperative pain.
Patients who were sleeping at the assessment time
were considered to be symptom-free and were asked
if they had any complaints at the next assessment
time. Anesthesiology nurses who were blinded as to
the study performed all these evaluations. Nausea was
defined as an unpleasant feeling associated with inclination to vomit, and vomiting was defined as the
forceful ejection of gastric contents through the
mouth. Retching (defined as involuntary gastric and
esophageal movements of vomiting without expulsion
of vomitus) was also recorded as vomiting. Pruritus
was defined as an uncomfortable sensation of irritation of the skin or mucous membranes that provokes
the desire to scratch or rub the affected sites. Patients
were asked about the presence of PONV, pruritus, and
pain and, if present, their location and intensity. Three
distinct standard 10-cm VASs (0 representing no
symptom and 10 representing the worst imaginable
severity of the symptom) were used to determine the
intensity of PONV, pruritus, and pain. We also recorded the frequencies of vomiting episodes, use of

1332

AMBULATORY ANESTHESIA PIRAT ET AL.

ONDANSETRON FOR INTRATHECAL MORPHINE-INDUCED PONV AND PRURITUS

rescue antiemetic, and adverse reactions to ondansetron (headache, cardiac arrhythmias, extrapyramidal
signs) in the first 24 h after surgery. Patients were
asked about the presence and characteristics of headache. If a patient complained of palpitations, 12-lead
ECG was used to verify the arrhythmia. The rescue
medications for PONV and pruritus were droperidol
1.25 mg IV and diphenhydramine 10 mg IV, respectively. Droperidol was administered if a patient had 2
or more vomiting episodes, if the nausea VAS score
was 5, or if the patient asked for an antiemetic.
Diphenhydramine was given if the pruritus score was
5 or if the patient asked for an antipruritic.
The n values for the study groups were established
using power analysis. A preliminary survey of similar
patients who underwent similar operations and anesthesia revealed PONV and pruritus incidences of 55%
and 75%, respectively. Assuming respective PONV
and pruritus frequencies of 55% and 75%, the calculation revealed that 49 patients per group would be
adequate to provide a value of 0.2 and an value of
0.05 for detection of a 50% difference in incidences of
PONV and pruritus. All values are presented as mean
sd or number (%), as appropriate. For continuous
data, statistical analyses were performed with oneway analysis of variance. Repeated measures of analysis of variance were used to compare the groups
pruritus, nausea, and pain VAS scores and also to
determine the significance of group time (G T)
interactions. Where a significant difference was detected, the Students t-test with Bonferroni correction
was applied for multiple comparisons within the
groups. The time to the start of PONV and pruritus
were analyzed by means of Kaplan-Meier probability
curves and the log-rank test. The 2 test was used to
analyze categorical variables, as appropriate. For all
determinations, P values 0.05 were considered
significant.

Results
There were no significant differences among the three
groups with respect to age, height, weight, number of
smokers, dose of additional midazolam patients received, operative time, and types of surgery performed (Table 1). Both forms of ondansetron were
well tolerated. None of the 150 patients in the study
developed headache, cardiac dysrhythmias, or extrapyramidal signs. The level of spinal anesthesia extended to the T6 to T10 dermatomal distributions in all
cases. The incidences of intraoperative hypotension
for the ODT, IV, and placebo groups were 12%, 14%,
and 8%, respectively (P 0.05). There were no significant differences among the groups with respect to
amounts of ephedrine received (ODT: 1.8 6.8 mg,
IV: 1.5 4.5 mg and placebo: 0.9 3.5 mg; P 0.05 for

ANESTH ANALG
2005;101:1330 6

all comparisons). Bradycardia was detected in only 2

patients in the ODT group, and no episodes of hypoxemia were diagnosed in the study population.
The proportion of the placebo group that developed
pruritus in the first 24 h postsurgery (86%) was significantly larger than the corresponding proportions
in the ODT group (56%, P 0.001) and the IV group
(66%, P 0.017), but the rates in the ODT and IV
groups were not significantly different (Table 2).
Repeated-measures analysis of variance revealed significantly different mean pruritus VAS scores within
each group at the different time points studied (Table
3). The same analysis also demonstrated a significantly lower overall mean pruritus VAS score in the
ODT group than the placebo group (P 0.001) and no
significant G T interactions among the 3 groups. The
ODT group had significantly lower mean pruritus
VAS scores in the PACU and at 2, 6, and 12 h postsurgery than the placebo group (P 0.03 for all comparisons) (Table 3). However, the mean pruritus VAS
scores in the IV group were not significantly different
from those in the other groups at any of the time
points assessed. Forty percent of the patients in the
placebo group required rescue antipruritic treatment
in the first 24 h, and this proportion was significantly
larger than that in the ODT group (18%) (P 0.013)
(Table 2). However, there were no significant differences between the ODT and IV groups (18% and 34%,
respectively; P 0.055) or the placebo and IV groups
(40% and 34%, respectively; P 0.05) with respect to
requirement for this type of rescue therapy. At the end
of the study period, 44% of the patients in the ODT
group, 34% in the IV group, and only 14% in the
placebo group were free of pruritus (log-rank test: P
0.0008 for ODT versus placebo and P 0.03 for IV
versus placebo) (Fig. 1). There were no differences
among the groups with respect to time to onset of
pruritus (ODT: 122 114 min, IV: 132 203 min,
placebo: 113 145 min; P 0.05) or sites of pruritus.
Of the total 104 patients who developed pruritus in
the first 24 h postsurgery, 83% (n 85) experienced
this symptom at more than one body site. The most
common sites of pruritus in descending order were
the chest (66%, n 67), face (60%, n 61), back (46%,
n 47), abdomen (38%, n 39), and lower extremities
(30%, n 30).
There were no significant differences among the
three groups with respect to incidence of PONV, number of vomiting episodes, and use of rescue antiemetics in the first 24 h after surgery (Fig. 2, Tables 2 and 3).
The mean numbers of vomiting episodes during the
24-h study period in the ODT, IV, and placebo groups
were 0.6 1.4, 0.8 3.0, and 0.9 2.5, respectively (P
0.05). Repeated-measures analysis of variance revealed significantly different mean nausea VAS scores
within each group at the different time points studied

ANESTH ANALG
2005;101:1330 6

AMBULATORY ANESTHESIA
PIRAT ET AL.
ONDANSETRON FOR INTRATHECAL MORPHINE-INDUCED PONV AND PRURITUS

1333

Table 1. Subjects Physical Characteristics, Duration of Surgery, Number of Smokers, Dose of Additional Midazolam
Patients Received, and Types of Surgical Procedures Performed

Age (yr)
Body height (cm)
Body weight (kg)
Operative time (min)
Smokers
Intraoperative additional midazolam (mg)
Type of surgery
Inguinal hernia
Cord hydrocele
Pilonidal sinus

ODT
(n 50)

IV
(n 50)

Placebo
(n 50)

24 6
173 7
74 9
45 15
31 (62)
0.5 1.2

25 7
176 7
75 11
44 17
34 (68)
0.4 1.0

23 4
174 7
70 10
47 22
29 (58)
0.4 1.1

26 (52)
16 (32)
8 (16)

30 (60)
14 (28)
6 (12)

25 (50)
17 (34)
8 (16)

Values are mean sd or number (%).

ODT orally disintegrating tablets of ondansetron; IV intravenous ondansetron.

Table 2. Overall 24-h Frequencies for Pruritus, PONV,

Vomiting Episodes, and Use of Rescue Medications

Pruritus
Rescue antipruritic
PONV
Vomiting episodes
Rescue antiemetic

ODT
(n 50)

IV
(n 50)

Placebo
(n 50)

28 (56)*
9 (18)
22 (44)
12 (24)
8 (16)

33 (66)
17 (34)
20 (40)
6 (12)
12 (24)

43 (86)
20 (40)
25 (50)
9 (18)
11 (22)

Values are n (%).

ODT orally disintegrating tablets of ondansetron; IV intravenous
ondansetron; PONV postoperative nausea and vomiting.
* P 0.001 compared with placebo; P 0.017 compared with placebo;
P 0.013 compared with placebo.

and no significant G T interactions among the

groups (Table 3).
There were no significant differences among groups
with respect to overall mean pain VAS score for the
first 24 h or mean pain VAS scores at each time point
studied (Table 3). There were also no significant differences among the groups with respect to frequency
of need for additional analgesics (ODT: 16%; IV: 24%;
placebo: 22%; P 0.05) or amount of diclofenac sodium received (ODT: 90 117 mg; IV: 98 112 mg;
placebo: 112 137 mg; P 0.05).

Discussion
The results of this randomized, double-blind, placebocontrolled study suggest that, compared with placebo,
prophylactic ODT ondansetron 8 mg is associated
with less IT morphine-induced pruritus in young men
who undergo minor elective surgery. However, it appears that this ondansetron regimen does not prevent
IT morphine-induced PONV in this patient group. We
also found that the frequency of IT-morphine induced
pruritus was less with prophylactic use of ondansetron 4 mg IV than with placebo; however, this IV
treatment offered no advantage over placebo with

respect to preventing IT morphine-induced PONV in

this population.
IT morphine is an attractive option for postoperative pain control because it causes intense and longlasting analgesia, does not prolong motor recovery,
and does not delay ambulation after surgery (15).
However, the incidences of the IT morphineassociated side effects PONV and pruritus are reportedly as frequent as 60% to 80% (4,10). These undesirable side effects of IT morphine can be very disturbing
to the patient, may be refractory to conventional antiemetic and antipruritic treatments, and may affect
patient satisfaction and thus limit the use of this
method of postoperative pain control.
IT morphine induces nausea and vomiting by acting
on the chemoemetic trigger zone and vomiting center
in the area postrema (3). Serotonin receptor antagonists act on serotonin receptors in the chemoreceptor
trigger zone of the area postrema and have been
shown to be effective against chemotherapy-induced
nausea and vomiting and PONV (16,17). Several studies have tested the efficacy of these drugs for preventing nausea and vomiting induced by spinal opioids,
and the results have been mixed (6,18 22). The diversity of these results is probably, at least in part, attributable to the fact that emetic symptoms have complex
and multifactorial etiologies and are correlated with
factors such as hormonal changes, sex, age, pain, surgical procedure, duration of operation, previous
PONV, history of motion sickness, smoking status,
and weight (6). Other reasons for discrepancies among
the results of these studies could be the types and
doses of IT opioids administered, different regimens
of 5HT3 receptor antagonists tested, and different
scales and definitions of PONV that have been used.
Our findings suggest that neither IV ondansetron
4 mg nor ODT ondansetron 8 mg is more effective
than placebo as prophylaxis for IT morphine-induced
PONV in healthy young men who have no known risk
factors for these side effects.

1334

AMBULATORY ANESTHESIA PIRAT ET AL.

ONDANSETRON FOR INTRATHECAL MORPHINE-INDUCED PONV AND PRURITUS

ANESTH ANALG
2005;101:1330 6

Table 3. Visual Analogue Scale (VAS) Scores for Pruritus, Nausea and Pain at the Different Assessment Times After
Surgery
ODT (n 50)
a,f

PACU
2 hb,f
6 hb,g
12 hc
18 hd
24 he

IV (n 50)

Placebo (n 50)

Pruritus

Nausea

Pain

Pruritus

Nausea

Pain

Pruritus

Nausea

Pain

0.6 1.3
(05.7)
1.5 2.1i
(08.5)
1.4 1.9j
(08.6)
0.7 1.6k
(09.0)
0.6 1.5
(07.1)
0.4 1.1
(05.0)

0.6 1.8
(08.1)
0.5 1.2
(04.5)
0.6 1.4
(06.9)
0.8 1.9
(09.1)
0.4 1.4
(09.2)
0.1 0.3
(01.8)

0.7 1.7
(07.8)
1.0 1.9
(08.3)
1.1 1.5
(004.7)
1.6 2.2
(09.0)
1.5 2.2
(09.1)
1.3 1.9
(07.0)

0.8 1.5
(05.6)
2.2 2.6
(08.5)
2.0 2.3
(08.0)
1.7 2.2
(07.7)
1.1 2.0
(09.1)
0.5 1.1
(05.0)

0.7 2.0
(09.6)
0.3 1.2
(07.4)
0.6 1.4
(07.3)
0.8 1.5
(06.1)
0.5 1.3
(05.0)
0.2 0.9
(05.0)

0.6 1.5
(06.1)
0.6 1.2
(05.0)
1.1 1.8
(09.1)
1.6 1.9
(08.9)
1.7 2.0
(06.7)
1.6 1.9
(08.1)

1.5 2.0
(07.0)
2.8 2.4
(09.5)
2.8 2.6
(09.0)
1.8 2.2
(08.1)
1.3 2.1
(09.0)
1.2 2.3
(09.1)

0.6 1.9
(09.5)
0.8 1.7
(08.0)
0.8 1.7
(08.1)
0.9 2.0
(010)
0.5 1.3
(06.2)
0.2 0.8
(04.1)

0.8 2.0
(08.1)
1.1 1.8
(07.0)
1.3 1.8
(06.1)
2.0 2.4
(010)
2.1 2.5
(09.0)
1.9 2.6
(09.0)

Values are mean sd (range).

PACU postanesthesia care unit; ODT orally disintegrating tablets of ondansetron; IV intravenous ondansetron.
For mean pruritus VAS values: a P 0.0001 compared with 2- and 6-h values; b P 0.001 compared with 12-, 18-, and 24-h values; and c P 0.001 compared
with 12- and 18-h values.
For mean nausea VAS values: d P 0.007 compared with 12-h values; and e P 0.029 compared with 2-, 6-, and 12-h values.
For mean pain VAS values: f P 0.02 compared with 12-, 18-, and 24-h values; and g P 0.027 compared with 12- and 18-h values.
Compared with placebo h P 0.011, i P 0.023, j P 0.007, and kP 0.015.

Figure 1. Time to onset of pruritus in the first 24 h postsurgery,

shown as Kaplan-Meier curves. ODT: group that received orally
disintegrating tablets of ondansetron 8 mg; IV: group that received
IV ondansetron 4 mg. *P 0.0008, ODT versus placebo; P 0.03,
IV versus placebo (log-rank test).

We believe ours is the first study to have investigated

the efficacy of prophylactic ondansetron relative to IT
morphine-induced-PONV in a homogeneous group of
patients with no risk factors for such symptoms. We
excluded all the patients with known risk factors for
PONV, and all the participants were healthy young men.
Based on the results, it appears that routine prophylactic
administration of ondansetron for antiemetic purposes is

Figure 2. Time to onset of postoperative nausea and vomiting

(PONV) in the first 24 h postsurgery, shown as Kaplan-Meier
curves. ODT: group that received orally disintegrating tablets of
ondansetron 8 mg; IV: group that received IV ondansetron 4 mg.

not advisable in this patient group. All our patients were

men, and this might partially explain why ondansetron
did not reduce the incidence of PONV in this study.
Apfel et al. (23) noted that droperidol is only effective at
preventing PONV in females. The lack of effect on
PONV in our study setting might also have been related
to the short half-life of ondansetron (4 6 hours) and to
the fact that IT morphine-induced PONV can last for a
much longer period (up to 24 hours).

ANESTH ANALG
2005;101:1330 6

AMBULATORY ANESTHESIA
PIRAT ET AL.
ONDANSETRON FOR INTRATHECAL MORPHINE-INDUCED PONV AND PRURITUS

Pruritus is the most common side effect of IT opioids, with reported incidence rates of 57% to 95%
(9,10,18,22,24,25). Several theories have been proposed
to explain the mechanism of IT opioid-induced pruritus, but the exact pathogenesis remains unclear. The
histamine-release theory does not explain this side
effect because pruritus can also be induced by opioids
that do not cause histamine release, and antihistamines are not effective treatment (18,24,25). The efficacy of ondansetron for preventing and treating of this
form of pruritus indicates that serotonin type 3 receptors play a role (10,26). Serotonin receptor antagonists
probably inhibit the direct excitatory effect of opioids
on the non-nociceptive neurons in the posterior horn
of the medulla spinalis and the nucleus of the spinal
tract of the trigeminal nerve (24,26). Studies have demonstrated that ondansetron is effective at both treating
(26) and preventing (9,22,27) IT morphine-induced
pruritus. However, the efficacy of ondansetron for
prophylaxis against pruritus when IT lipophilic opioids are administered is not clear (19,24,25). A report
by Szarvas et al. (18) is the only one that has disputed
this beneficial effect of ondansetron with IT morphine
use. Compared with other studies mentioned above
(9,22,27), Szarvas et al. administered a much larger
amount of IT morphine (0.01 mg/kg, up to 0.7 mg)
and they did not compare the efficacy of ondansetron
with results in a placebo group. In our study, the onset
and distribution of pruritus were similar to results
that have been reported previously (22,26). We found
that both the ODT (8 mg) and IV (4 mg) forms of
ondansetron were associated with significantly less
frequent pruritus than placebo (rates 56%, 66%, and
86%, respectively; P 0.017 for all comparisons). We
also found that only the prophylactic ODT ondansetron was associated with statistically less rescue antipruritic requirement than placebo treatment (18% versus 40%, respectively; P 0.013). The same
relationship held true for the mean VAS pruritus
scores in the PACU and at 2, 6, and 12 hours postsurgery (P 0.023 for all comparisons). None of the
corresponding comparisons involving the IV group
(i.e., IV versus placebo or IV versus ODT) revealed
significant differences. These results clearly indicate
that, when compared with placebo treatment, prophylactic ODT ondansetron 8 mg effectively reduces the
incidence and severity of IT morphine-induced pruritus and is even more effective in this respect than IV
ondansetron 4 mg.
The ODT form of ondansetron seems to offer important advantages for anesthesia practice, as it
avoids IV injection while maintaining the desired
preoperative fasting state. Another important advantage of this formulation of the drug is that it can
be used in ambulatory surgery after patient discharge. Still, the current literature contains only two

1335

reports on the use of ODT ondansetron for preventing PONV after general anesthesia (13,14). Our
study is the first to have focused on prophylactic
use of ODT ondansetron for IT morphine-induced
pruritus and PONV. We used a 4-mg dose of IV
ondansetron because previous studies have shown
that this dose is adequate for preventing and treating emesis and pruritus after IT opioid administration (6,9,18,22,28). The dose of ODT ondansetron
was chosen based on the pharmacokinetic properties of oral ondansetron and its 60% to 67% bioavailability (29,30). However, it is not clear whether the
bioavailability of this newer formulation of the drug
is identical to that of conventional tablets. It could
be speculated that the higher antipruritic efficacy of
ODT ondansetron compared with the IV formulation in our study might be a result of the fact that
8 mg of the ODT form results in a larger absorbed
dose than 8 mg of conventional tablets or 4 mg of IV
form. Timing of drug administration might also be
important. In our study, both forms of ondansetron
were administered at the same time (10 minutes
before spinal anesthesia). It is logical to speculate
that the IV ondansetron would have peaked before
pruritus onset, whereas the ODT form probably
peaked closer to the time of pruritus onset. It is also
possible that ODT ondansetron was absorbed into
the circulation more slowly than the IV form, thus
providing a longer period of effective antipruritic
blood levels.
One important limitation of this study is that the
incidence of PONV in our placebo group was less
frequent than the expected value we used to calculate
required sample size (actual incidence 50% as opposed
to estimated 55%). This increases the risk of type II
error. Post hoc power analysis revealed that 50 patients
per group gave this trial a power of 74% (for 0.05).
Another limitation is that we did not measure patients blood ondansetron levels. Without comparing
the groups blood levels, it is impossible to make a
clear-cut statement about why ODT ondansetron provided superior pruritus prophylaxis in this study.
Third, the placebos that were used also constitute a
limitation. It would have been preferable to administer placebo tablets that tasted and looked identical to
the ODT form of ondansetron.
In conclusion, neither ODT ondansetron 8 mg nor
IV ondansetron 4 mg is more effective than placebo as
prophylaxis for IT morphine-induced PONV in young
men with no known risk factors for PONV. Compared
with placebo, both regimens of ondansetron were associated with less incidences of IT morphine-induced
pruritus in this patient group. However, only the ODT
form of ondansetron significantly reduced the severity
of pruritus and the need for rescue antipruritic
administration.

1336

AMBULATORY ANESTHESIA PIRAT ET AL.

ONDANSETRON FOR INTRATHECAL MORPHINE-INDUCED PONV AND PRURITUS

References
1. Brill S, Gurman GM, Fisher A. A history of neuraxial administration of local analgesics and opioids. Eur J Anaesthesiol 2003;
20:6829.
2. Watcha MF, White PF. Postoperative nausea and vomiting: its
etiology, treatment, and prevention. Anesthesiology 1992;77:
162 84.
3. Chaney MA. Side effects of intrathecal and epidural opiates.
Can J Anaesth 1995;42:891903.
4. Kjellberg F, Tramer MR. Pharmacological control of opioidinduced pruritus: a quantitative systematic review of randomized trials. Eur J Anaesthesiol 2001;18:346 57.
5. Pan PH, Moore CH. Intraoperative antiemetic efficacy of prophylactic ondansetron versus droperidol for cesarean section
patients under epidural anesthesia. Anesth Analg 1996;83:
982 6.
6. Pan PH, Moore CH. Comparing the efficacy of prophylactic
metoclopramide, ondansetron, and placebo in cesarean section
patients given epidural anesthesia. J Clin Anesth 2001;13:430 5.
7. Parlow JL, Costache I, Avery N, Turner K. Single-dose haloperidol for the prophylaxis of postoperative nausea and vomiting
after intrathecal morphine. Anesth Analg 2004;98:1072 6.
8. Dimitriou V, Voyagis GS. Opioid-induced pruritus: repeated vs
single dose ondansetron administration in preventing pruritus
after intrathecal morphine. Br J Anaesth 1999;83:8223.
9. Charuluxananan S, Somboonviboon W, Kyokong O, Nimcharoendee K. Ondansetron for treatment of intrathecal morphineinduced pruritus after cesarean delivery. Reg Anesth Pain Med
2000;25:5359.
10. Yeh HM, Chen LK, Lin CJ, et al. Prophylactic intravenous
ondansetron reduces the incidence of intrathecal morphineinduced pruritus in patients undergoing cesarean delivery.
Anesth Analg 2000;91:1725.
11. LeBourgeois JP, McKenna CJ, Coster B, et al. Efficacy of an
ondansetron orally disintegrating tablet: a novel oral formulation of this 5-HT(3) receptor antagonist in the treatment of
fractionated radiotherapy-induced nausea and emesis. Emesis
Study Group for the Ondansetron Orally Disintegrating Tablet
in Radiotherapy Treatment. Clin Oncol (R Coll Radiol) 1999;11:
340 7.
12. Davidson N, Rapoport B, Erikstein B, et al. Comparison of an
orally disintegrating ondansetron tablet with the conventional
ondansetron tablet for cyclophosphamide-induced emesis in
cancer patients: a multicenter, double-masked study. Ondansetron Orally Disintegrating Tablet Emesis Study Group. Clin
Ther 1999;21:492502.
13. Gan TJ, Franiak R, Reeves J. Ondansetron orally disintegrating
tablet versus placebo for the prevention of postdischarge nausea
and vomiting after ambulatory surgery. Anesth Analg 2002;94:
1199 200.
14. Thagaard KS, Steine S, Raeder J. Ondansetron disintegrating
tablets of 8 mg twice a day for 3 days did not reduce the
incidence of nausea or vomiting after laparoscopic surgery. Eur
J Anaesthesiol 2003;20:1537.

ANESTH ANALG
2005;101:1330 6

15. Carr D, Cousins M. Spinal route of analgesia, opioids and future

options. In: Bridenbaugh P, ed. Neural blockade in clinical
anesthesia and management of pain, 3rd ed. Philadelphia:
Lippincott-Raven, 1998:115 83.
16. Koeller JM, Aapro MS, Gralla RJ, et al. Antiemetic guidelines:
creating a more practical treatment approach. Support Care
Cancer 2002;10:519 22.
17. Gan TJ, Meyer T, Apfel CC, et al. Consensus guidelines for
managing postoperative nausea and vomiting. Anesth Analg
2003;97:6271.
18. Szarvas S, Chellapuri RS, Harmon DC, et al. A comparison of
dexamethasone, ondansetron, and dexamethasone plus ondansetron as prophylactic antiemetic and antipruritic therapy in
patients receiving intrathecal morphine for major orthopedic
surgery. Anesth Analg 2003;97:259 63.
19. Yazigi A, Chalhoub V, Madi-Jebara S, et al. Prophylactic ondansetron is effective in the treatment of nausea and vomiting but
not on pruritus after cesarean delivery with intrathecal
sufentanil-morphine. J Clin Anesth 2002;14:183 6.
20. Pitkanen MT, Numminen MK, Tuominen MK, Rosenberg PH.
Comparison of metoclopramide and ondansetron for the prevention of nausea and vomiting after intrathecal morphine. Eur
J Anaesthesiol 1997;14:1727.
21. Pitkanen MT, Niemi L, Tuominen MK, Rosenberg PH. Effect of
tropisetron, a 5-HT3 receptor antagonist, on analgesia and nausea after intrathecal morphine. Br J Anaesth 1993;71:681 4.
22. Charuluxananan S, Kyokong O, Somboonviboon W, et al. Nalbuphine versus ondansetron for prevention of intrathecal
morphine-induced pruritus after cesarean delivery. Anesth
Analg 2003;96:1789 93.
23. Apfel CC, Korttila K, Abdalla M, et al. A factorial trial of six
interventions for the prevention of postoperative nausea and
vomiting. N Engl J Med 2004;350:244151.
24. Gurkan Y, Toker K. Prophylactic ondansetron reduces the incidence of intrathecal fentanyl-induced pruritus. Anesth Analg
2002;95:1763 6.
25. Korhonen AM, Valanne JV, Jokela RM, et al. Ondansetron does
not prevent pruritus induced by low-dose intrathecal fentanyl.
Acta Anaesthesiol Scand 2003;47:12927.
26. Borgeat A, Stirnemann HR. Ondansetron is effective to treat
spinal or epidural morphine-induced pruritus. Anesthesiology
1999;90:432 6.
27. Arai L, Stayer S, Schwartz R, Dorsey A. The use of ondansetron
to treat pruritus associated with intrathecal morphine in two
paediatric patients. Paediatr Anaesth 1996;6:3379.
28. Sanchez-Ledesma MJ, Lopez-Olaondo L, Pueyo FJ et al. A comparison of three antiemetic combinations for the prevention of
postoperative nausea and vomiting. Anesth Analg 2002;95:
1590 5.
29. Roila F, Del Favero A. Ondansetron clinical pharmacokinetics.
Clin Pharmacokinet 1995;29:95109.
30. Rojanasthien N, Manorot M, Kumsorn B. Pharmacokinetics and
bioequivalence testing of generic ondansetron preparations in
healthy Thai male volunteers. Int J Clin Pharmacol Ther 1999;
37:548 54.

BRIEF REPORT

The Prevalence and Significance of Low Preoperative

Hemoglobin in ASA 1 or 2 Outpatient Surgery Candidates
Ronald P. Olson,

MD,

Alan Stone,

PhD,

and David Lubarsky,

MD MBA,

Departmentof Anesthesiology, Duke University Medical Center, Duke Health Technology Solutions, Duke University
Medical Center, Durham, North Carolina, Department of Anesthesiology, University of Miami, Miami, Florida

Asymptomatic anemia in healthy patients undergoing

low risk surgery is rare. In this retrospective study, we
examined the records of 9584 ASA class III patients
scheduled for elective low risk surgery who had a preoperative hemoglobin (hgb) test for the presence of anemia. Hgb 9 g/dL was detected in 75 patients (0.8%).
Perioperative management of anemia occurred in no
cases of elective surgery in this group. Transfusion of

everal studies have shown that preoperative

testing should be based on clinical examination
rather than routine protocols (17). Two recent
prospective studies have supported this general concept in elderly surgical patients and minimal-risk surgery on cataracts (8,9).
There is increasing realization that the perioperative
period is not an appropriate setting for screening and
investigating asymptomatic anemia (10 12). This is
largely because the prevalence is infrequent except in
neonates, the elderly, smokers, patients using anticoagulants, or patients with specific disease (12,13) and
the likelihood of perioperative management of anemia
is small (2,4).
We hypothesize that the prevalence of anemia of clinical interest by current practice in generally healthy outpatient surgical candidates is too infrequent to warrant
routine screening. There has been no large study of the
prevalence of anemia in this increasingly large proportion of surgical candidates, especially anemia as defined
in currently practical terms. This retrospective study assesses the prevalence of anemia defined as hemoglobin
(hgb) 9 g/dL and also examines whether the discovery
of this level of anemia in asymptomatic outpatient surgical candidates results in any change in perioperative
management.
Accepted for publication May 12, 2005.
Address correspondence and reprint requests to Ronald P. Olson, MD,
Duke University Medical Center, Box 3094, Durham, NC 277l0, Telephone
919 684 2025 Fax 919 681 8484, e-mail olson012@mc.duke.edu
DOI: 10.1213/01.ANE.0000180836.02142.E6
2005 by the International Anesthesia Research Society
0003-2999/05

red cells occurred in four other patients, all of whom

had hgb 9 g/dL. In all cases, management decisions
were based on clinical factors rather than the preoperative hemoglobin test. In healthy patients undergoing
low risk elective surgery, routine preoperative hgb testing is not indicated.
(Anesth Analg 2005;101:133740)

Methods
Between January 1, 1997 and April 31, 1999, 21,080 outpatient surgical procedures were performed at Duke
University Medical Center. This was a unique period
when the required databases could be efficiently collated. After obtaining IRB approval, the resu1ts of the
laboratory database, admission database, hospital electronic common data repository, and transfusion services
database were analyzed. There were 14,337 patients who
had preoperative hgb test results recorded in the hospital laboratory database within the 30 days before the
surgical procedure as well as complete demographic
data.
The laboratory database was analyzed to find all patients who had hgb 9 mg/dL. This threshold was
selected as one at which even asymptomatic anemia
might require treatment. The admission database was
used to determine the ASA physical classification, age,
and gender of all patients with a preoperative hgb
9 mg/dL. The prevalence of preoperative hgb levels
9 mg/dL in various subgroups of the 14,337 patients
were calculated and compared. The electronic common
data repository records of any of the above patients who
were ASA class 1 or 2 with preoperative hgb 9 mg/dL
were then examined to see if there was evidence of
further anemia investigations or treatment. If there was
insufficient information in the electronic record, the written hospital record was examined as well. Statistical
comparisons of group proportions were made using
Pearsons 2 test or Fishers exact test with small counts.
Alpha 0.05 was considered significant.
Anesth Analg 2005;101:133740

1337

1338

BRIEF REPORT

The ASA classification was based only on information that would have been available preoperatively
from the preoperative clinical assessment, paper medical record, and electronic medical record that included previous lab results and anesthetic records.
As an additional crosscheck, the transfusion services database was reviewed to see if any of these
patients received blood products. As no therapy was
apparently impacted by measuring the preoperative
hgb in this subset of patients, zero numerator statistics
give a 99% confidence interval that the chance of a hgb
test result impacting management is at most 0.048%
(14).

Results
Of the 14,337 surgical outpatients who had hgb tests,
9584 were ASA III. Seventy-five of these 9584 had
hgb 9 g/dL, giving a prevalence of 0.8% (95% confidence interval [CI], 0.6%1.0%). In the ASA III patients, 138 of 3499 patients had hgb 9 g/dL, giving a
prevalence of 3.9% (95% CI, 3.0% 4.9%). In the ASA
IV patients, 18 of 205 patients had hgb 9 g/dL,
giving a prevalence of 8.8% (95% CI, 5.3%13.5%). All
3 pairwise comparisons showed Fishers exact P
0.0032.
The prevalence of hgb 9 g/dL was slightly higher
in females (0.9%; 95% CI, 0.7%1.2%) than males
(0.5%; 95% CI, 0.3% 0.9%) (P 0.0340). When analyzed by age, the prevalence in patients aged 13 yr
was 4.6%, which was significantly more than in patients aged 13 yr, where it was 1.6%. In the 66 infants
1 yr of age, the rate was 6.1%, although this was not
significantly different from the rate of 4.4% in the 112
yr group.
Of the 75 ASA III patients with hgb 9 mg/dL, the
average preoperative hgb was 8.3 g/dL with a standard deviation of 0.7 g/dL. The lowest was 5.6 g/L.
The 4 patients with results 7 g/dL had clear indicators of potential anemia (advanced human immunodeficiency virus, sickle cell disease, or history of bleeding). It is questionable whether these individuals
should have been classified as ASA II. There was no
evidence of further investigation or perioperative
treatment of anemia in these 75 patients undergoing
elective surgery, including those with hgb 7 g/dL.
Review of the transfusion records showed that 4 of the
9584 ASA III patients (0.05%) received red blood cells
on the same day or the day after the hgb test. All had hgb
9 g/dL. All had clear pretest clinical indictors of potential anemia. Three had a history of anemia, one was
receiving heparin, and one had advanced cancer. There
was no evidence that the decision to transfuse or any
other perioperative management had occurred as a result of the preoperative hgb. Of these 9584 patients,
11 had received transfusions in the 6 mo before the

ANESTH ANALG
2005;101:133740

procedure. Review of the medical records showed that

in no case had a screening or preoperative hgb result
been part of the decision to transfuse.
One other patient, initially classified as ASA II, was
transfused postoperatively. She had sickle cell disease
and had presented to the emergency department with
increased vaginal bleeding; her hgb was 9 g/dL. She
was reclassified as an ASA IIE and removed from the
ASA III study group.

Discussion
Recent changes in the perioperative hospitalization
practices and transfusion thresholds have effectively
resulted in a reduced prevalence of anemia requiring
management, especially in the increasing proportion
of patients who are basically healthy, asymptomatic
with respect to anemia, and undergoing outpatient
surgery. We sought to assess the prevalence of anemia
in asymptomatic outpatient surgical candidates to see
if routine hgb screening is warranted.
In previous times, surgery usually entailed significant risk of blood loss and physiological insult as well
as a lengthy hospital admission. Most patients with a
hgb 10 g/dL received a transfusion. Therefore, the
prevalence of anemia that required transfusion was
frequent and routine hgb testing was appropriate.
Today, many surgical procedures involve minimal
risk. There has also been a change in transfusion
thresholds. The National Institute of Health, American
College of Physicians, and American Society of Anesthesiology consensuses state that transfusion is not
necessarily indicated in a patient with a hgb as low as
7 g/dL if that patient is normovolemic, asymptomatic,
and no further blood loss is anticipated (1518).
Screening in low-prevalence groups yields so many
false positives that a positive test is not useful information (10 12). Guidelines for anemia screening set
by authoritative organizations state that it should only
be done in high-risk infants once before the age of 9
months and in menstruating women every 510 years
(13,19). If the prevalence of anemia in asymptomatic
outpatient surgical candidates is similar to that in the
general population, screening criteria should be similar to those recommended by these authorities. Unindicated hgb screening, especially without the clinical context and follow-up of primary care, exposes
patients to the net negative effects of testing with little
benefit.
It has been shown that actual follow-up of abnormal
hgb results in the increasingly short perioperative period is quite rare (2,20). It has been stated that there is
now more risk of litigation from not following up a
test than from not ordering it (11,20 22). We agree.
In this study, we assessed the prevalence of hgb
9 mg/dL in healthy outpatient surgical candidates.

ANESTH ANALG
2005;101:133740

This threshold was chosen as a balance between increasing evidence that the traditional transfusion
threshold of hgb 10 mg/dL is no longer appropriate
in asymptomatic patients, (15,16,18) and the concern
of many clinicians to at least be monitoring hgb at
levels slightly below the traditional one.
The prevalence of hgb 9 g/dL in ASA III patients
in this study was only 0.8%. Generally, a prevalence of
1%5% is needed to make screening beneficial (2327).
As an example, the prevalence of iron deficiency anemia (hgb 13.5 g/dL) in adult males is 2%, and
screening is not recommended for this group (13).
This study of 9584 such patients showed that only
4 had transfusions, and all of these patients had been
expected based on information other than the preoperative hgb. Unexpected anemia requiring management is
so rare in these patients that routine hgb testing has no
benefit.
Routine laboratory tests cannot substitute for a clinical examination (19). But if there are any clinical
indicators of potential anemia, including high-risk
surgery or high-risk patients, then hgb testing is indicated. Clinical indicators of anemia are any feature of
the patient history or examination that might lead one
to suspect anemia. One such list is given in a recent
textbook (21).
There are shortcomings of this study. It is possible
that patients with low hgb results had surgery delayed
and so were excluded from the sample or only entered
the study after correction of anemia. This possibility
would exist even if hgb results for all 21,080 patients
were recorded. Analysis of the events leading up to
the referral to surgery would be needed. That information was not reliably available in this study.
Only 14,337 of 21,080 outpatients (68%) had hgb tests
performed within the 30 days before surgery. There is a
possibility of selection bias, which excluded some cases of
low hgb. However, all of the unmeasured patients proceeded to surgery without further testing, which would
have been unlikely if there had been anemia concerns. In
addition, analysis of the preoperative transfusions in the
studied group showed that all transfusion decisions were
based on historical information independent of screening
hgb tests. This suggests that these test results are rarely part
of management decisions.
More likely the bias is in the other direction, with the
clinicians choosing to not order the test because of low
probability of anemia or knowledge of previous outside
test results showing normal levels. Thus the population
studied likely had a more frequent incidence of abnormal hgb than the general ASA III outpatient surgical
population. This would effectively increase the prevalence of anemia in our tested population, making the
conclusions of this study even more valid in a less selected population.

BRIEF REPORT

1339

Because 90% of the study population was between

the ages of 13 and 79 years, conclusions cannot be
applied to children or the very old.
This large study supports previous ones showing that
selective testing based on the clinical examination is safe
(4,28), avoids the drawbacks of false positives (1,7,21),
and reduces unnecessary expense (1,6,7,28,29). The probability that preoperative hgb will impact on management is estimated to be 0.05%.

Conclusion
We conclude that asymptomatic hgb 9 g/dL in ASA
III outpatient surgical candidates is rare. When it does
occur, it is associated with clinical indicators. It does not
result in any change in management. Routine preoperative hgb testing in these patients is not indicated.

References
1. Allison JG, Bromley HR. Unnecessary preoperative investigations:
evaluation and cost analysis. Am Surg 1996;62:686 9.
2. Kaplan EB, Sheiner SB, Boeckmann AJ. The usefulness of preoperative laboratory screening. JAMA 1985;253:3576 81.
3. McLeane GJ. Preoperative measurement of haemoglobin concentration. Ulster Medical J 1990;59:145148.
4. Narr BJ, Warner ME, Schroeder DR, Warner MA. Outcomes of
patients with no laboratory assessments before anesthesia and
surgical procedures. Mayo Clin Proc 1997;72L505509.
5. Turnbull JM, Buck C. The value of preoperative screens investigations in otherwise healthy individuals. Arch Intern Med
1987;147:11015.
6. Velanovitch V. The value of routine preoperative laboratory
testing in predicting postoperative complications: a multivariate
analysis. Surgery 1991;109:236 43.
7. Wyatt WJ, Reed DN, Apelgren KN. Pitfalls in the role of standardized preadmission laboratory screening for ambulatory
surgery. Am Surg 1989;55:343 6.
8. Dzankic S, Pastor D, Gonzalez C, Leung JM. The prevalence and
predictive value of abnormal preoperative laboratory tests in
elderly surgical patients. Anesth Analg 2001;93:301 8.
9. Schein OD, Katz J, Bass EB. The value of routine preoperative
medical testing before cataract surgery. JAMA 2005;342:168 75.
10. Litaker D. Preoperative screening. Med Clin North Am 1999;83:
1565 81.
11. Marcello PW, Roberts PL. Routine preoperative studies.
Which studies in which patients? Surg Clin North Am 1996;76:
1123.
12. Pasternak LR. Preoperative assessment: guidelines and challenges. Acta Anesthesiologica Scandinavia 1997;S111:318 20.
13. Recommendations to prevent and control iron deficiency in the
United States. MMWR 1998;44(RR-3):129.
14. Hanley J, Lippman-Hand A. If nothing goes wrong, is everything all right? Interpreting zero numerators. JAMA 1983;249:
17435.
15. American College of Physicians. Practice strategies for elective
red blood cell transfusion. Ann Internal Med 1992;116:403 6.
16. American Society of Anesthesiologists. Practice guidelines for
blood component therapy. Anesthesiology 2005;84:732 47.
17. Hebert PC, Wells G, Blajchman MA. A multicenter, randomized,
controlled clinical trial of transfusion requirements in critical
care. N Engl J Med 1999;340:409 17.
18. National Institutes of Health. Perioperative Red Cell Transfusion:
NIH Consensus Statement Online. NIH Consensus Development
Conference. 1988.

1340

BRIEF REPORT

19. US Preventive Services Task Force. Guide to clinical preventive

services. Baltimore: Williams & Wilkins, 1996.
20. Golub R, Cantu R, Sorrento JJ, Stein HD. Efficacy of preadmission testing in ambulatory surgical patients. Am J Surg
1992;163:56571.
21. Roizen MF. Preoperative evaluation. In: Miller RD, ed. Millers
anesthesia. Philadelphia: Churchill Livingstone, 2005:92798.
22. Smetana GW, Macpherson D. The case against routine preoperative laboratory testing. Med Clin North Am 2003;87:7 40.
23. Kerlikowske K, Grady D, Barclay J. Positive predictive value of
screening mammography by age and family history of breast
cancer. JAMA 1993;70:2444 50.
24. Ladenson PW, Singer PA, Ain KB. American Thyroid Association guidelines for detection of thyroid dysfunction. Arch Internal Med 2000;160:15735.

ANESTH ANALG
2005;101:133740

25. Lieberman DA, Weiss DG, Bond JH. Use of colonoscopy to

screen asymptomatic adults for colorectal cancer. N Engl
J Med 2000;343:162 8.
26. Mandel JS, Bond JH, Church TR. Reducing mortality from colorectal cancer by screening for fecal occult blood. N Engl J Med
1993;328:136571.
27. Patz EF, Goodman PC, Bepler G. Screening for lung cancer.
N Engl J Med 2000;343:162732.
28. Korvin CC, Pearce RH, Stanley J. Admissions screening: clinical
benefits. Ann Internal Med 1975;83:197.
29. Nardella A, Pechet L, Snyder LM. Continuous improvement,
quality control, and cost containment in clinical laboratory
testing: effects of establishing and implementing guidelines for
preoperative tests. Arch Path Lab Med 1995;119:518 22.

ANESTHETIC PHARMACOLOGY

INTERNATIONAL SOCIETY

FOR

ANAESTHETIC PHARMACOLOGY

SECTION EDITOR
JAMES G. BOVILL

EDITORIAL

Opiate-Induced Nausea and Vomiting: What Is

the Challenge?
Martin Redmond,

MD

and Peter Glass,

MBChB

Department of Anesthesiology, SUNY at Stony Brook, Stony Brook, New York

ostoperative pain management is an everchanging and challenging conundrum.

The Agency for Healthcare Research and Quality

(previously known as the Agency for Health Care
Policy and Research) was created by Congress in 1989
was charged with the development of practice guidelines. In 1992, they released the publication Acute
Pain Management in Operative or Medical Procedures
and Trauma, which stated, Opioid analgesics are the
cornerstone of pharmacological postoperative pain
management, especially for more extensive surgical
procedures that cause moderate to severe pain. (1)
This was followed by the Joint Commission on Accreditation of Healthcare Organizations emphasis on
pain management and the recognition of pain as the
fifth vital sign. These aggressive efforts to enhance
postoperative pain management led to increasing opiate administration. Opioid use was further increased
by the availability of newer, simpler, and more effective drug delivery systems, including central neuraxial
administration, patient-controlled analgesia pumps,
drug patches, and slow release oral formulations. Although the practice of increasing opiate administration has resulted in more effective pain therapy and
patient satisfaction, it has also come at a price of
increased opiate side effects (2,3) including sedation,
respiratory depression, urinary retention, pruritus,
constipation, and nausea and vomiting. Certainly the
incidence and severity of sedation and respiratory
depression are dose-related phenomenon, but is the
incidence of these other adverse events dichotomous
(all or none), or are they also dose-dependent? This is
an important question because if these adverse events
occur independent of dose, then the many efforts to
Accepted for publication July 11, 2005.
Address correspondence to Peter S. Glass, MB.ChB., FFA (S.A.),
University Medical Center at Stony Brook, Department of Anesthesiology, Nicolls Rd - HSC Level 4, Rm 060, Stony Brook, NY 11794
8480. Address e-mail to Peter.Glass@stonybrook.edu.
DOI: 10.1213/01.ANE.0000181329.07214.D8
2005 by the International Anesthesia Research Society
0003-2999/05

provide opiate sparing have limited value. Unfortunately, although numerous studies have demonstrated that a variety of interventions or therapies can
result in opiate sparing, they are generally underpowered to demonstrate an improvement in outcomes (4).
Several recent articles have, at least in general, evaluated if opiate sparing reduces the incidence or severity of opiate side effects or what the authors have
termed clinically meaningful events (CMEs) (5,6).
This work resulted from the development of cyclooxygenase (COX)-2 analgesics for perioperative pain. In
the control group, the incidence of a CME and the
number of CMEs was related to the dose of opiate
analgesic administered. Interestingly, the authors also
found that below a morphine equivalent threshold of
approximately 10 mg, opiate-related symptoms did
not occur. In another study, Gan et al. (6) found after
a cholecystectomy and using the same symptom distress questionnaire that the opiate sparing produced
by the coadministration of a COX-2 similarly reduced
the incidence of CMEs proportional to the reduction in
opiate administration. The authors, however, were unable to discriminate the effect of dose on individual
adverse events produced by opiates. These studies
confirmed that opiate-related side effects, in general,
were reduced by decreasing opiate administration,
but they could not confirm if the incidence of specific
adverse events were decreased by reduced opiate administration. More recently, Marret et al. (7) performed a meta-analysis of studies investigating the
opiate-sparing effects of nonsteroidal antiinflammatory drugs (NSAIDs). By combining these studies,
they demonstrate that the approximately 30% opiatesparing effect that these drugs provided also resulted
in a decreased relative risk of nausea and vomiting.
More importantly, their analysis showed there was a
dose-response relationship between the amount of
opiate administered and the incidence of both nausea
and vomiting. They found a linear relationship, with
each reduction of 10 mg of morphine decreasing the
incidence of nausea by 9%. Being a retrospective metaanalysis, these data required verification by a prospective study.
Anesth Analg 2005;101:13412

1341

1342

EDITORIAL

In this issue of Anesthesia & Analgesia, Roberts et al.

(8) have confirmed, in a prospective study, that the
incidence of nausea and vomiting are both increased
in a dose-dependent manner by the amount of opiate
administered. This study included epidural, IV, and
oral opiate administration for patients after either orthopedic or abdominal surgery, thereby covering a
wide spectrum of opiate requirements. They found an
exponential relationship between opiate dose and
nausea, with each halving of the opiate dose reducing
the incidence of vomiting by 6% (and a little more for
nausea). Although there are several design problems
in the study by Roberts et al. (which are well discussed
by the authors) and although the percentage reduction
in postoperative nausea and vomiting for each milligram of opiate is not identical between the two studies, they together provide credence to the notion that
opiate-induced nausea and vomiting is dosedependent and that measures to provide opiate sparing are likely to reduce their incidence. One caveat of
the Roberts et al. study is that pain itself may induce
nausea and vomiting (9). This makes it difficult to
determine if pain or the opiate is causing the nausea
and vomiting. The meta-analysis by Marret et al. (7)
would support that this is opiate-induced, as patients
in the various studies analyzed generally had equal
pain relief whether this was provided primarily by the
opiate or the NSAID plus reduced opiate.
In many instances, the use of large-dose opiates is
unavoidable. In these incidences, administering prophylactic antiemetics is clearly indicated. The choices
for first-line antiemetics are multiple and controversial. With the knowledge that reducing opiate requirements decreases nausea and vomiting, it makes sense
to choose an antiemetic that is effective in opiateinduced nausea and vomiting and also enhances analgesic efficacy. Dexamethasone (10), small-dose
promethazine (11), and droperidol (12) have demonstrated, at least in some models of pain, both opiatesparing activity and efficacy in opiate-induced nausea
and vomiting. All of these antiemetics are relatively
inexpensive, effective, and can be used alone or in
combination. When used at appropriate doses they
also have limited side effects.
The articles by Roberts et al. and Marret et al. have
confirmed the notion that opiate-induced nausea and

ANESTH ANALG
2005;101:13412

vomiting is dose dependent. Based on this information

we should consider moving away from the concept
that opiate analgesics are the cornerstone of moderateto-severe pain management but are rather simply one
component of multimodel pain therapy. Unfortunately, all therapies used for pain management are
fraught with adverse events. The challenge for the
clinician today is to minimize the opiate component of
their pain management (thereby reducing the incidence of side effects) while still insuring optimal pain
relief.

References
1. Research AfHCPa. Acute Pain Management Guideline Panel:
Acute Pain Management: Operative or Medical Procedures and
Trauma Clinical Practice Guideline Rockville, MD: US Department of Health and Human Services, Public Health Service,
1992.
2. Wheeler M, Oderda GM, Ashburn MA, Lipman AG. Adverse
events associated with postoperative opioid analgesia: a systematic review. J Pain 2002;3:159 80.
3. Smetzer JL, Cohen MR. Pain scales dont weigh every risk. J
Pain Palliat Care Pharmacother 2003;17:6770.
4. Romsing J, Moiniche S, Mathiesen O, Dahl JB. Reduction of
opioid-related adverse events using opioid-sparing analgesia
with COX-2 inhibitors lacks documentation: a systematic review. Acta Anaesthesiol Scand 2005;49:133 42.
5. Zhao SZ, Chung F, Hanna DB, et al. Dose-response relationship
between opioid use and adverse effects after ambulatory surgery. J Pain Symptom Manage 2004;28:35 46.
6. Gan TJ, Joshi GP, Zhao SZ, et al. Presurgical intravenous parecoxib sodium and follow-up oral valdecoxib for pain management after laparoscopic cholecystectomy surgery reduces opioid
requirements and opioid-related adverse effects. Acta Anaesthesiol Scand 2004;48:1194 207.
7. Marret E, Kurdi O, Zufferey P, Bonnet F. Effects of nonsteroidal
antiinflammatory drugs on patient-controlled analgesia morphine side effects: meta-analysis of randomized controlled trials. Anesthesiology 2005;102:1249 60.
8. Roberts GW, Becker TB, Carlsen HH, et al. Postoperative nausea
and vomiting is strongly influenced by postoperative opioid use
in a dose related manner. Anesth Analg 2005;101:1343 8.
9. Andersen R, Krohg K. Pain as a major cause of postoperative
nausea. Can Anaesth Soc J 1976;23:366 9.
10. Baxendale BR, Vater M, Lavery KM. Dexamethasone reduces
pain and swelling following extraction of third molar teeth.
Anaesthesia 1993;48:961 4.
11. Chia YY, Lo Y, Liu K, et al. The effect of promethazine on
postoperative pain: a comparison of preoperative, postoperative, and placebo administration in patients following total abdominal hysterectomy. Acta Anaesthesiol Scand 2004;48:62530.
12. Lo Y, Chia YY, Liu K, Ko NH. Morphine sparing with droperidol in patient-controlled analgesia. J Clin Anesth 2005;17:2715.

Postoperative Nausea and Vomiting Are Strongly Influenced

by Postoperative Opioid Use in a Dose-Related Manner
Gregory W. Roberts, B Pharm, FSHP, BCPS*, Tenna B. Bekker, MSc Pharm,
Helle H. Carlsen, MSc Pharm, Christine H. Moffatt, MBBS, FANZCA,
Peter J. Slattery, MBBS, FANZCA, FFPMANZCA, and Anna F. McClure, B Pharm,

BCPS*

*Pharmacy Department and Department of Anesthesia, Intensive Care, and Pain Medicine, Repatriation General
Hospital, Daw Park, Australia; and Royal Danish School of Pharmacy, Copenhagen, Denmark

We prospectively examined the incidence of postoperative nausea and vomiting (PONV) in a group of 193
elderly surgical inpatients receiving no postoperative
antiemetic prophylaxis. Risk factors for PONV and
detailed data on postoperative opioid use were recorded. The overall postoperative vomiting (POV) rate
was 23.8%, whereas postoperative nausea (PON) was
51.3%. Opioid use (P 0.025), and female gender
(P 0.038) were identified as significantly influencing
POV in this relatively small population. There was a
strong logarithmic dose-response relationship between

ostoperative vomiting (POV) is a major source of

distress for patients undergoing anesthesia for
surgery and may delay discharge from, or cause
readmission to, hospital. Much work has gone into
clarifying and weighting those factors responsible for
POV, but the diversity of patient demographics and
types of surgery make it difficult to predict with certainty those patients most likely to be at risk. Medical
literature indicates a diverse array of postoperative
nausea (PON) and POV rates. In populations that
include a wide range of operative procedures, the
incidence ranges from 22% to 52% (1,2). The incidence
seems to increase in patients receiving patientcontrolled analgesia (PCA), where rates of 68%100%
for PON and POV have been recorded (3,4).
Despite the large amount of research performed in
this area, it is only recently that expert consensus
regarding the selection of patients for prophylaxis of
POV has been published (5). Prevention of PON and

Accepted for publication May 5, 2005.

Address correspondence and reprint requests to Greg Roberts,
Pharmacy Department, Repatriation General Hospital, Daws
Rd., Daw Park SA 5041, Australia. Address e-mail to
greg.roberts@rgh.sa.gov.au.
DOI: 10.1213/01.ANE.0000180204.64588.EC
2005 by the International Anesthesia Research Society
0003-2999/05

postoperative opioid dose and POV (r2 0.98, P

0.01), as well as PON (r2 0.98, P 0.01). Use of patientcontrolled analgesia or epidural analgesia was a marker
for large-dose opioid use (P 0.001) and was associated
with POV in the 24-h postoperative period of 41% and
31% respectively, compared with 11% for other patients
(P 0.001). Future studies defining risk factors for POV
should treat postoperative opioid use as a continuous
variable, rather than treat it as a dichotomous variable.
(Anesth Analg 2005;101:13438)

POV is important, not only in relation to the quality of

care afforded to each patient, but also in monetary
terms.
A number of studies have identified risk factors for
POV in various surgical populations (6 9). These
studies have often used regression analysis to construct risk scoring equations that incorporate various
risk factors, although not all groups have identified
the same risk factors as being significant contributors
to POV. This may reflect the different populations that
have been studied. The commonly identified risk factors that have consistently contributed to POV are:
female gender, nonsmoking status, history of POV or
motion sickness, extended duration of anesthesia,
postoperative opioid use, and age.
As the number of risk factors increases, so does the
chance of POV. It is possible that in some instances the
type of anesthesia and surgical procedure may also
contribute.
The above-mentioned studies have not specifically
examined the dose-response relationship between opioids and vomiting. Most clinicians would intuitively
assume that a patient receiving a large dose of postoperative opioid is more likely to vomit than a similar
patient who receives less.
We examined the effect of these known risk factors
on our patient group, with a focus on the relationship
Anesth Analg 2005;101:13438

1343

1344

ANESTHETIC PHARMACOLOGY ROBERTS ET AL.

OPIOIDS AND POSTOPERATIVE VOMITING

Table 1. Opioid Dose Equivalency10,11

Drug
Alfentanil
Codeine
Fentanyl
Morphine
Oxycodone
Pethidine
Tramadol

Oral dose

Parenteral dose

20 mg

10 g
13 mg
10 g (10 g epidural)
1 mg ( 10 g spinal)

2 mg
20 mg

7.5 mg
10 mg

between vomiting and opioid use in the 48 h postoperatively. Although well recognized, the relationship
between opioids and POV, and the degree of influence
it bears, has not been well explored. In particular,
many of the equations developed through logistic regression predicting risk of POV have only included
opioid use as a dichotomous variable, with no acknowledgment of any dose-response relationship.

Methods
The study was approved by the local Research and
Ethics Committee, and written, informed consent was
obtained from all subjects. All patients receiving surgical procedures requiring anesthesia (excluding local
anesthesia) with an expected length of stay of 2 days
who did not receive perioperative antiemetic prophylaxis were eligible. The approach to analgesia for any
given patient was at the discretion of the anesthesiologist. Those patients not using epidural analgesia or
PCA were given pain relief with a combination of IV
and oral medication, on an as required basis. Patients already receiving drugs with antiemetic properties, including corticosteroids, were excluded.
An episode of POV was defined as vomiting or
retching over any 2-min period. The severity or duration of nausea was not recorded, only if it was present
or not, as determined by the patient. Patients who
vomited were automatically included as having experienced nausea at that point. Patients routinely received postoperative rescue antiemetics if they vomited, or experienced 10 min of debilitating nausea. In
the first instance, they received 10 mg of IV metoclopramide, followed 10 min later by 4 mg of IV ondansetron if the nausea and vomiting were still not
controlled.
Nausea and vomiting episodes were recorded 0.5,
1, 2, 4, 8, 12, 24, and 48 h postoperatively. Opioid
doses, both intra- and postoperative, were recorded
for the 0- to 24-h and 24- to 48-h periods postoperatively. All opioid doses, regardless of route of administration or type of opioid, were converted to the
equianalgesic dose of IV morphine for comparative
purposes, using the values in Table 1 (10,11). These
values were predetermined on current available literature and the clinical expertise of the participating

ANESTH ANALG
2005;101:13438

anesthesiologists. Fentanyl was used for all epidurals and was considered equipotent via epidural
or IV route. One milligram of IV morphine was
considered to be equianalgesic with 10 g of spinal
morphine.
Parametric data were compared using a Students
t-test. Kaplan-Meier plots were used to examine the
incidence of POV over time. Cox regression analysis
was used to examine variables influencing POV. The
significance level was set at 5%.

Results
Data were collected on 193 patients (see Table 2 for
patient characteristics and Table 3 for perioperative
characteristics). In the first 24 h postoperatively, 23.8%
of patients experienced POV and a further 27.5% experienced PON with no associated vomiting (Table 4).
In the 24- to 48-h postoperative period, 6.5% of patients vomited and a further 23.2% experienced nausea only. One patient (0.5%) vomited for the first time
during this period and 5 patients (2.6%) experienced
PON for the first time.
Cox regression analysis included gender, history of
POV or motion sickness, smoking, duration of anesthesia, age, and opioid dose, and revealed only opioid
use (P 0.025) and female gender (P 0.038) as
factors influencing POV. The influence of history of
POV or motion sickness, smoking, duration of anesthesia, or age were not significant in this relatively
small group.
Use of PCA or epidural analgesia were markers for
large-dose opioid use in the first 24 h (91.5 and 83.2 mg
of morphine or equivalent for PCA and epidural analgesia, respectively, versus 17.5 mg for non-users, P
0.001). This was associated with more frequent POV
and PON (Table 4). The majority of POV had occurred
by 12 h in the PCA group whereas the majority of POV
for epidural patients was seen in the 12- to 24-h period
(Fig. 1). Patients not using PCA or epidural analgesia
experienced less POV and PON (P 0.001 for both).
Seven patients in this group required no opioid analgesia and did not experience PON or POV. Those
patients who experienced POV and PON in the 24- to
48-h period postoperatively had significantly larger
opioid use during this period than those who did not
(P 0.01 for both).
Patients were divided into quartiles according to
opioid dose to further examine the relationship between opioid dose and POV in the first 24 h postoperatively. The time course of POV for the 4 morphine dose quartiles is shown in Figure 2 (P 0.05).
There was a strong logarithmic dose-response relationship with POV (r2 0.98, P 0.01), as well as

ANESTH ANALG
2005;101:13438

ANESTHETIC PHARMACOLOGY
ROBERTS ET AL.
OPIOIDS AND POSTOPERATIVE VOMITING

1345

Table 2. Patient Characteristics

Type of analgesia

na

Age
(yr)

Weight
(kg)

Female
(%)

History of
POV (%)

Smokers
(%)

All patients
PCA
Epidural
Other

193
64
26
101

74 (2291)
72 (2291)
71 (6089)
76 (2689)

80 (40128)
82 (47128)
78 (54101)
79 (40121)

26.9
34.4
7.7
27.7

22.8
23.4
19.2
22.8

14.5
15.6
3.8
16.8

Age and weight data are median (range).

POV Postoperative vomiting.
a
Two patients had both epidural and patient-controlled analgesia (PCA) postoperatively and were excluded from the PCA or Epidural Analgesia columns.

Table 3. Perioperative Characteristics

Type of
analgesia

Duration
(h)

Spinal
morphinea (%)

Propofol
(%)

Propofol (mg)
(mean sd)

General
anesthetic (%)

Spinal
anesthetic (%)

Orthopedic
surgery (%)

All patients
PCA
Epidural
Other

2.3 1.3
2.3 1.8
3.2 1.5
2.1 1.3

14 (7.3)
10 (15.6)
0
4 (4.0)

124 (64.2)
42 (65.6)
17 (65.4)
65 (64.4)

148 125
135 50
209 284
143 91

115 (59.6)
43 (78.1)
13 (50.0)
58 (57.4)

84 (43.5)
25 (39.1)
14 (53.8)
44 (43.6)

81 (42.0)
46 (71.9)
16 (61.5)
17 (16.8)

Some patients received more than one type of anesthetic.

PCA patient-controlled analgesia.
a
One hundred micrograms of spinal morphine given during anesthesia.

Table 4. Postoperative Opioid Use and Vomiting/Nausea Rates

Type of
analgesia
All
PCA
Epidural
Other

0- to 24-h period

24- to 48-h period

Morphine dose (mg)

Vomiting (%)

Nausea (%)

Morphine dose (mg)

Vomiting (%)

Nausea (%)

51.9
91.5
83.2
17.5

23.8
40.6
30.8
10.9

51.3
71.9
69.2
32.7

26.4
39.9
67.7
2.7

6.5
6.7
19.2
1.5

29.7
41.7
34.6
14.9

PCA patient-controlled analgesia.

Figure 1. Patient-controlled analgesia (PCA) (, n

64), epidural analgesia (, n 26), neither (F, n
101). P 0.0001.

PON (r2 0.98, P 0.01, Fig. 3). The relationship

between POV and morphine dose in the 0- to 24-h
postoperative period was described by:
% vomiting 19.9 log (morphine dose in mg) 5.5

When patients receiving opioids via the spinal or

epidural route were removed from analysis, this relationship remained largely intact, although the doseresponse relationship with POV in this subgroup was
better suited to a linear relationship (r2 0.99, P

1346

ANESTHETIC PHARMACOLOGY ROBERTS ET AL.

OPIOIDS AND POSTOPERATIVE VOMITING

ANESTH ANALG
2005;101:13438

Figure 2. Time to first vomit for various morphine dose quartiles (P 0.05).

epidural analgesia with a previous history of POV

(35%, PON rate 71%).

Discussion

Figure 3. Log morphine dose (mg) in the immediate 24 h postoperatively versus vomiting () and nausea (F). Vomiting: r2 0.98,
P 0.008. Nausea: r2 0.98, P 0.010.

0.01 for linear, r2 0.82, P 0.09 for logarithmic, n

145). PON remained best correlated to a logarithmic
relationship (r2 0.99, P 0.01 for logarithmic versus
r2 0.88, P 0.07 for linear).
For given types of surgery there was marked variability in opioid use. In the first 24 h postoperatively,
knee arthroplasties had a median morphine dose of
70 mg (range 7134, n 33) whereas hip arthroplasties
had a median morphine dose of 47 mg (range 4 296, n
23).
The highest at-risk groups for POV were women
receiving PCA or epidural analgesia with or without a
history of POV (70% and 54%, respectively, PON rates
100% for both groups) and then men receiving PCA or

These data indicate a strong dose-response relationship between opioid use and POV. The relationship
was best described by a logarithmic association more
so than linear. A logarithmic association is biologically
consistent with receptor saturation and attainment of
a response plateau.
The relationship between POV and morphine dose
in the 0- to 24-hour postoperative period was described by:
% vomiting 19.9 log (morphine dose in mg) 5.5

This means that for each halving of the opioid dose

in this 24-hour period, there was an absolute reduction
of 6.0% in POV, assuming all other factors remained
constant. Regardless of whether the dose is halved
from 200 to 100 mg, 100 to 50 mg, or 50 to 25 mg, the
absolute reduction in POV was constant at 6.0%. The
relative reduction in these groups, however, was
14.9%, 17.5%, and 21.2%, respectively, giving greater
relative returns for dose reductions in those patients in
the smaller dosing range. If a patient had not already
experienced POV or PON in the first 24 hours postoperatively, there was only a very small likelihood
they would experience it after 24 hours.

ANESTH ANALG
2005;101:13438

PCA or epidural analgesia use was targeted toward

patients undergoing surgical procedures with known
large postoperative analgesic requirements, such as
hip and knee arthroplasties. Not surprisingly, these
types of analgesia were markers for large-dose opioid
use and hence also predictive for POV. There was
much variation in opioid use, however, even within
various surgical subgroups receiving PCA or epidural
analgesia. Accurate prediction of postoperative opioid
use, perhaps as part of a risk scoring assessment to
determine prophylactic antiemetic use, would be very
difficult.
The postoperative opioid dose seems to have a large
part in determining the likelihood of POV or PON.
The strength of the dose-response relationship with
POV has been largely unrecognized, and this may
have led to the inability of risk scoring equations to
accurately predict POV (12). Given the apparent
strength of the opioid-POV relationship, the inclusion
of opioid use as a dichotomous variable rather than a
continuous variable may undermine the ability of
these risk scoring equations to predict POV. Unfortunately, the ability to predict a patients opioid dose is
poor, even for a given type of surgery, as seen with the
large dose range in both hip and knee arthroplasties in
this study. This source of variability may have to be
something clinicians have to accept when attempting
to assess a patients POV risk. Exploration of opioidsparing strategies for the postoperative period seems
appropriate, and patients likely to have large postoperative opioid requirements should be targeted for
prophylactic antiemetic strategies.
It is important to note that although these data
indicate a strong relationship between opioid use and
POV, they do not represent definitive proof of this
relationship, the population studied in this series being relatively small. The possibility exists that postoperative opioid use may be a marker for a further
variable that is the true cause of increased POV, such
as duration of surgery, or a specific type of surgery.
Consistent with the belief that postoperative opioid
dose is a primary driver of POV, however, Buvanendran et al. (13) found a reduction in POV from 26% to
6% in 2 identical groups undergoing knee arthroplasty
when postoperative opioid was reduced by the perioperative administration of rofecoxib. It remains to be
seen if other associated variables, such as the type of
surgery, may be a marker for opioid use and hence
POV. This needs further investigation in populations
large enough to determine these possibilities.
The opioid dose was the most dominant influence in
determining POV in this patient group, followed by
female gender. The small population studied was
likely to have lacked sufficient power to determine
other known associated risk factors for POV. The addition of various known risk factors resulted in increased rates of POV, consistent with the findings of

ANESTHETIC PHARMACOLOGY
ROBERTS ET AL.
OPIOIDS AND POSTOPERATIVE VOMITING

1347

others (14). The two highest at-risk groups for POV

were women receiving PCA or epidural analgesia
with or without a history of POV (70% and 54%,
respectively). Men receiving PCA or epidural analgesia with a history of POV also experienced a frequent
rate of POV (35%).
There were two practical problems we faced in obtaining a history of POV. First, some patients had not
previously experienced anesthesia, and, second, the
occasional elderly patient described POV after the use
of ether as an anesthetic hardly a fair comparison
with the anesthetics used today. This decreased the
value of a history of POV as a reliable predictor. We
also experienced difficulty in ascertaining a clear history of motion sickness in our elderly patients.
Propofol has been shown to possess dose-related
antiemetic activity (15,16). This patient group received
a small dose of propofol as part of the induction
process, with just five patients receiving continuing
propofol as a continuing part of the anesthetic procedure. The effect of propofol in this patient group remains unknown, because the numbers were too small
for meaningful analysis. Given the small doses used,
however, the effect on POV was unlikely to be
significant.
There is inevitably some degree of subjectivity in
determining equianalgesic potency within the opioid
group and across various routes of administration.
The equivalency tables used for this study were determined before any patient data were collected and
are based on currently available literature. The duration and potency of spinal or epidural opioids is
greatly enhanced compared with the IV route, and
exact comparisons of analgesic potency are difficult.
When patients receiving spinal or epidural analgesia
were excluded, however, the dose-response relationship remained intact. Although conversions to equivalent doses of IV morphine were made on the basis of
analgesic potency, this may not be reflective of their
differing emetogenic potencies.
The incidence of POV and PON found in this study
is likely to be an underestimate of the true incidence
for two reasons. First, patients who received antiemetic prophylaxis were excluded from study. For
ethical reasons, we allowed the normal practice of
empirical prophylaxis, as judged by the anesthesiologist, to continue, and hence were unable to study an
entire surgical population uninfluenced by perioperative antiemetic prophylaxis. These excluded patients
(n 72) contained a number of higher-risk patients
who would likely have increased the rate of POV and
PON had they been eligible for inclusion. Compared
with our study group, this group had a significantly
larger proportion of orthopedic patients, women, patients with a history of POV, and fewer smokers.
Exclusion of these patients would make identification
of the higher-risk groups less likely. Second, patients

1348

ANESTHETIC PHARMACOLOGY ROBERTS ET AL.

OPIOIDS AND POSTOPERATIVE VOMITING

who experienced PON, but not yet POV, were subsequently given rescue antiemetic therapy, which may
have prevented them from progressing to POV. This
may have generated a falsely small incidence of POV,
without affecting the incidence of PON.
It is possible that accounting for the dose-response
relationship between postoperative opioids and POV
may lead to further refinement in risk scoring algorithms. The ability to predict a patients postoperative
opioid requirement is likely to be poor, however, and
this may continue to hamper accurate prediction of
which patients should be targeted for preoperative
antiemetics.

References
1. Koivuranta M, Laara E, Snare L, Alahuhta S. A survey of postoperative nausea and vomiting. Anesthesia 1997;52:4439.
2. Cohen M, Duncan P, DeBoer D, Tweed W. The postoperative
interview: assessing risk factors for nausea and vomiting.
Anesth Analg 1994;78:716.
3. Woodhouse A, Mather L. The effect of duration of dose delivery
with patient controlled analgesia on the incidence of nausea and
vomiting after hysterectomy. Br J Clin Pharmacol 1998;45:57 62.
4. Bree S, West M, Taylor P, Kestin I. Combining propofol with
morphine in patient-controlled analgesia to prevent postoperative nausea and vomiting. Br J Anaesth 1998;80:152 4.
5. Gan TJ, Meyer T, Apfel CC, et al. Consensus guidelines for
managing postoperative nausea and vomiting. Anesth Analg
2003;97:6271.

ANESTH ANALG
2005;101:13438

6. Apfel CC, Greim CA, Haubitz I, et al. A risk score to predict the
probability of postoperative vomiting in adults. Acta Anaesthesiol Scand 1998;42:495501.
7. Apfel CC, Laara E, Koivuranta M, et al. A simplified risk score
for predicting postoperative nausea and vomiting. Anesthesiology 1999;91:693700.
8. Palazzo M, Evans R. Logistic regression analysis of fixed patient
factors for postoperative sickness: a model for risk assessment.
Br J Anaesth 1993;70:135 40.
9. Sinclair D, Chung F, Mezei G. Can postoperative nausea and
vomiting be predicted? Anesthesiology 1999;91:109 18.
10. Cherny I. Opioid analgesics: comparative features and prescribing guidelines. Drugs 1996;51:71337.
11. Metz C, Gobel L, Gruber M, Hoerauf KH. Pharmacokinetics of
human cerebral opioid extraction: a comparative study on
sulfentanil, fentanyl, and alfentanil in a patient after severe head
injury. Anesthesiology 2000;92:1559 67.
12. van den Bosch JE, Kalkman CJ, Vergouwe Y, et al. Assessing the
applicability of scoring systems for predicting postoperative
nausea and vomiting. Anaesthesia 2005;60:3231.
13. Buvanendran A, Kroin JS, Tuman KJ, et al. Effects of perioperative administration of a selective cyclooxygenase 2 inhibitor on
pain management and recovery of function after knee replacement. JAMA 2003;290:2411 8.
14. Apfel CC, Kortilla K, Abdalla M, et al. A factorial design of six
interventions for the prevention of postoperative nausea and
vomiting. N Engl J Med 2004;350:244151.
15. Sneyd JR, Carr A, Byrom WD, Bilski AJ. A meta-analysis of
nausea and vomiting following maintenance of anesthesia with
propofol or inhalational agents. Eur J Anaesthesiol 1998;15:
433 45.
16. Hammas B, Thorn SE, Wattwil M. Superior prolonged antiemetic prophylaxis with a four drug multi-modal regimen: comparison with propofol or placebo. Acta Anaesthesiol Scand 2002;
46:2327.

Does Neostigmine Administration Produce a Clinically

Important Increase in Postoperative Nausea and Vomiting?
Ching-Rong Cheng,

MD,

Daniel I. Sessler,

MD,

and Christian C. Apfel,

MD

Outcomes Research Institute and Department of Anesthesiology and Perioperative Medicine, University of Louisville,
Louisville, Kentucky

Neostigmine is used to antagonize neuromuscular

blocker-induced residual neuromuscular paralysis. Despite the findings of a previous meta-analysis, the effect
of neostigmine on postoperative nausea and vomiting
remains unresolved. We reevaluated the effect of
neostigmine on postoperative nausea and vomiting
while considering the different anticholinergics as potentially confounding factors. We performed a systematic literature search using MEDLINE, Embase, Cochrane library, reference listings, and hand searching
with no language restriction through December 2004
and identified 10 clinical, randomized, controlled trials
evaluating neostigmines effect on postoperative nausea and vomiting. Data on nausea or vomiting from 933
patients were extracted for the early (0 6 h), delayed
(6 24 h), and overall (0 24 h) postoperative periods

ausea and vomiting remain among the most common perioperative complications, occurring in approximately 30% of postoperative patients (13).
Although the origin is generally believed to be multifactorial, there is increasing evidence that patient-specific
risk factors play a major role (4). However, drugs specific
to anesthesia, including volatile anesthetics, nitrous oxide, and postoperative opioids, are at least as important
and, in contrast to the patient-specific risk factors, under
the control of anesthesiologists (5,6).
Neostigmine is often used to antagonize residual neuromuscular block. Because anticholinesterases such as
neostigmine have cholinergic effects on the gastrointestinal tract (increased motility and gastric acid secretion)
and on the heart (bradycardia, cardiac arrest), they are
Supported, in part, by National Institutes of Health Grant GM
061655 (Bethesda, MD), the Gheens Foundation (Louisville, KY), the
Joseph Drown Foundation (Los Angeles, CA), and the Commonwealth of Kentucky Research Challenge Trust Fund (Louisville,
KY).
Accepted for publication May 12, 2005.
Address correspondence and reprint requests to Christian Apfel,
MD, 501 E. Broadway, Suite 210, Louisville, KY, 40202. Address
electronic mail to apfel@ponv.org.
DOI: 10.1213/01.ANE.0000180992.76743.C9
2005 by the International Anesthesia Research Society
0003-2999/05

and analyzed with RevMan 4.2 (Cochrane Collaboration, Oxford, UK) and multiple logistic regression analysis. The combination of neostigmine with either atropine or glycopyrrolate did not significantly increase the
incidence of overall (0 24 h) vomiting (relative risk,
0.91; 95% confidence interval, 0.70 1.18; P 0.48) or
nausea (relative risk, 1.24; 95% confidence interval,
0.98 1.59; P 0.08). Multiple logistic regression analysis indicated that there was not a significant increase in
the risk of vomiting with large compared with small
doses of neostigmine. Contrasting a previous analysis,
we conclude that there is insufficient evidence to conclude that neostigmine increases the risk of postoperative nausea and vomiting.
(Anesth Analg 2005;101:1349 55)

coadministered with anticholinergics such as atropine or

glycopyrrolate (7). Atropine is a tertiary amine and can
cross the blood-brain barrier to cause central effects. In
contrast, glycopyrrolate is a quaternary amine that does
not easily cross the blood-brain barrier and thus has no
important central effects (8).
Intrathecal neostigmine has been shown to cause
severe nausea and vomiting in a dose-dependent manner, probably via action on the brainstem (9). The
effect of IV neostigmine on postoperative nausea and
vomiting (PONV) remains controversial. Tramer and
Fuchs-Buder (10) concluded in their meta-analysis
that neostigmine in doses 2.5 mg increases the incidence of PONV. However, a subsequent study (11),
not included in their systemic review (10), was unable
to confirm an emetogenic effect. Furthermore, the incidence of PONV appears to be reduced when neostigmine is combined with atropine as opposed to glycopyrrolate (12).
If neostigmine were truly emetogenic, it would be
reasonable to reconsider its routine use in antagonizing neuromuscular paralysis, especially for patients at
increased risk for PONV. Our goal was thus to determine whether neostigmine administration produces a
clinically relevant increase in the risk of PONV and to
Anesth Analg 2005;101:134955

1349

1350

ANESTHETIC PHARMACOLOGY
NEOSTIGMINE AND PONV.

CHENG ET AL.

determine the extent to which risk depends on the

coadministration of the anticholinergics.

Methods
We searched for all published full reports of randomized, controlled trials that compared patients given
neuromuscular blocking antagonists (intervention)
with those allowed to recover spontaneously from
neuromuscular block (control group, i.e., placebo or
no added anticholinesterase). Included trials were required to have dichotomous outcomes (presence or
absence) for postoperative nausea (PON), postoperative vomiting (POV), or adverse events.
We searched systematically for relevant reports without any language restrictions in MEDLINE (1966 2004),
EMBASE (1980 2004), and Cochrane Library (2004, Issue 4); the date of our last electronic search was December 8, 2004. We used the free text terms nausea, vomiting, emesis, neostigmine, prostigmine, edrophonium,
antagonism, and neuromuscular block in any combination for the search. A manual scan was performed
through the reference lists of all studies in the search
results until no further relevant references could be
identified.
Authors of the original publication were contacted
if analyses end-points were insufficiently reported.
We sought to obtain separate data for PON and POV
and for the early, delayed, and overall study period of
24 h (13). Two authors (RC and CCA) independently
read each retrieved report to assess the adequacy of
randomization and blinding. The 5-point Oxford score
was used to assess the quality of the study design (14)
and differences were resolved by consensus.
We obtained information on patients, anesthetics,
type and dose of anticholinergics, type and dose of
anticholinesterases, and intervention-related adverse
effects from each included report. Dichotomous data
on harm and efficacy were extracted from the published reports. Extracted outcome data were early
PON, early POV (0 6 h postoperative cumulative incidence), delayed PON, delayed POV (6 24 h postoperative cumulative incidence), overall postoperative
PON, and overall POV (0 24 h postoperative cumulative incidence), as well as data on clinically diagnosed adverse events.
Data extracted from the relevant studies were entered into RevMan 4.2 (Review Manager, Cochrane
Collaboration, UK) and analyzed. The relative risk
(RR) with the corresponding 95% confidence intervals
(CI) was calculated for each study. The results were
pooled together using the Mantel-Haenszel method
for combining trials. The individual effect sizes were
weighted according to the reciprocal of their variance.
A random effect model was used and heterogeneity
was determined under the assumption (null hypothesis) that there were no differences in treatment effect

ANESTH ANALG
2005;101:1349 55

between trials. Results are presented as RR (95% CI).

The comparisons of neostigmine versus control were
also divided into subgroups based on whether atropine or glycopyrrolate was given.
Multiple logistic regression analysis was used to investigate the relationship of the dose of neostigmine and
POV. The analysis was corrected by center, with the
largest center being the reference group (SPSS version
12.0 for windows, SPSS Inc., Chicago, IL). An absolute
increase of 10% or a relative increase of 25%, i.e., a
quarter of the incidence, was considered to be clinically
important.

Results
We found 15 reports that met our criteria. Five of these
were subsequently excluded: two did not have control
groups (15,16), another did not report PONV results
(17), the data of one other study were insufficient to be
considered in the analysis, despite getting more information from the authors (18), and in the last excluded
study, only the combination of edrophonium and atropine was compared with placebo (19). In one trial,
although only data on PONV were published, the
authors generously provided the detailed data so that
their study could be included in the analysis (11). We
therefore performed meta-analyses on 10 comparisons
of neostigmine versus an inactive control with 933
study patients (Table 1) (11,20 28). All trials were
randomized, except one that had a pseudo-randomization; we only included those patients who received
standard anesthesia (n 41) or no reversal (n 40)
after cholecystectomy from this study (28).
Early (0 6 h) PON was reported as an outcome in 6
trials (Table 2) (11,20 23,25): 5 with glycopyrrolate
(11,2123,25) and 1 with atropine (20). The RR of suffering PON in this early period was 1.24 (0.86 1.80; P
0.25). Early POV (0 6 h) was reported in 8 studies.
Patients in six of them received glycopyrrolate (11,21
23,25,27); patients in the other two received atropine
(20,26). The RR for POV in the early postoperative
period was 1.05 (0.721.55;P 0.79).
Delayed (6 24 h) PON as an outcome was included
in 4 studies with a RR of 1.09 (0.76 1.57; P 0.64)
(11,21,23,25). All were with neostigmine combined
with glycopyrrolate versus control. Delayed POV was
an outcome in 4 studies; all were with glycopyrrolate.
The RR was 1.01 (0.58 1.78; P 0.96) (11,21,23,25).
Overall (0 24 h) PON was reported in 6 studies
with a RR of 1.24 (0.98 1.59; P 0.08) (Fig. 1)
(11,20,2225). Overall POV (0 24 h) was reported in 8
studies with coadministration of atropine or glycopyrrolate with a RR of 0.91 (0.70 1.18; P 0.48) (Fig. 2)
(11,20,2225,27,28). Thus, neostigmine was not associated with a significant increase in PON or POV in any
of the above-mentioned analyses.

ANESTH ANALG
2005;101:1349 55

ANESTHETIC PHARMACOLOGY
CHENG ET AL.
NEOSTIGMINE AND PONV.

1351

Table 1. Characteristics of Analyzed Studies

Score
(random,
blind,
dropouts)

Study ID

Interventions

Outcomes
PON, POV in day care; first
and second day nausea,
vomiting
PON, POV in PACU; 24 h
PON, POV

Notes

Boeke et al., 1994 (20)

(1, 0, 1)

79 adults

Vecuronium; neostigmine 1.5 mg and

atropine vs. no treatment

Early: day care stage

Ding et al., 1994 (21)

(1, 0, 1)

69 women

Hovorka et al., 1997 (22)

(1, 2, 1)

160 women

Janhunen and Tammisto,

1972 (28)

(0 to 1, 0, 0) 81 cholecystectomy
patients

Joshi et al., 1999 (23)

(2, 2, 0)

100 adults

King et al., 1988 (24)

(1, 0, 0)

38 adults

Lovstad et al., 2001 (25)

(2, 2, 1)

90 women

Nelskyla et al., 1998 (12)

(2, 2, 1)

100 women

Walsh et al., 1988 (26)

(2, 0, 1)

120 children

Watcha et al., 1995 (27)

(1, 2, 0)

113 children

Succinylcholine mivacurium no
Early: PACU stay
reversal vs. mivacurium
mivacurium no reversal vs.
mivacurium mivacurium
neostigmine 2.5 mg and
glycopyyrolate 0.5 mg
Mivacurium; Neostigmine 2.0 mg and PON and POV in 01 h; 2
Early: 03 h
glycopyrrolate vs. placebo
3 h; 39 h; 915 h; 1521 h;
2127 h total (027 h)
Tubocurarine; Neostigmine 2.0 mg POV (024 h)
Only included standard
atropine 1.0 mg vs. No reversal
vs. no reversal
patients
Mivacurium vs. rocuronium; Residual PACU, Phase II, 24 h PON
Early: PACU and Phase
block reversal with neostigmine
and POV
II stage
2.5 mg glycopyrrolate 0.5 mg
Tubocurarine; Neostigmine 2.5 mg vs. PON, POV in 24 h
No treatment
Mivacurium; Neostigmine 2.5 mg
06 h; 624 h; 024 h PON,
Pretreated with
glycopyrrolate 0.5 mg vs. placebo
POV; Satisfaction score
Ondansetron
Mivacurium; Neostigmine 2.0 mg
PON, POV in PACU, ward, Only PONV, no PON,
glycopyrrolate 0.4 mg vs. No
on the way home, at
POV data Recount
treatment
home, 24 h
data after got
original data
Early POV
Pancuronium; Neostigmine (60 g/
kg) atropine (20 g/kg) vs. No
treatment
Early POV, POV in 24 h
Mivacurium; Placebo vs.
Edrophonium (1 mg/kg)
atropine (10 g/kg) vs.
Neostigmine (70 g/kg)
Glycopyrrolate (10 g/kg)

PACU postanesthesia care unit; PON postoperative nausea; POV postoperative vomiting.

Table 2. Early and Delayed Postoperative Nausea and Vomiting with Neostigmine Versus Control; Results of the Metaanalyses
Outcome
Early nausea (06 h)
Early vomiting (06 h)
Delayed nausea (624 h)
Delayed vomiting (624 h)

Anticholinergics
Atropine and Glycopyrrolate
Atropine
Glycopyrrolate
Atropine and Glycopyrrolate
Atropine
Glycopyrrolate
Glycopyrrolate
Glycopyrrolate

Number
of studies

Number of
participants

Relative risk
(95% CI)

6
1
5
8
2
6
4
4

584
79
505
768
199
568
337
337

1.24 (0.861.80)
0.67 (0.361.26)
1.39 (0.971.99)
1.05 (0.721.55)
0.75 (0.521.08)
1.35 (0.882.06)
1.09 (0.761.57)
1.01 (0.581.78)

CI confidence interval.

We performed multiple logistic regression analysis

of overall POV on 9 studies with 800 patients (Table 3)
(11,12,20,2225,27,28). The average dose of neostigmine when given with glycopyrrolate was 3.02 mg/
70 kg; the average dose when given with atropine was
2.59 mg/70 kg. We used the coefficients in Table 3 to
calculate the odds ratio for a combination of interventions by odds ratio e(dose*) where e is the
natural logarithm (2.71), dose is the neostigmine
dose in mg, is the coefficient for the neostigmine
(ln 1.32 0.278), and is the coefficient for the
concomitant anticholinergic drug, 0.73 (ln 0.482) for
glycopyrrolate or 1.14 (ln 0.32) for atropine. Thus,

patients who received an average of 3.02 mg neostigmine and glycopyrrolate had an odds ratio for developing POV of 1.11 ( e(3.02*0.278 0.73)) whereas those
receiving an average of 2.59 mg neostigmine with
atropine had an odds ratio of 0.66 ( e(2.59*0.278 1.14)).
For comparison, the effect of the center, i.e., where the
study was conducted, had odds ratios ranging from
0.12 to 5.24. Thus, logistic regression analysis suggested that neostigmine does not significantly increase
overall POV.
Two studies described inadequate muscle strength
in their control groups: One patient was excluded
from a 40-patient control group because of failure to

1352

ANESTHETIC PHARMACOLOGY
NEOSTIGMINE AND PONV.

CHENG ET AL.

ANESTH ANALG
2005;101:1349 55

Figure 1. Overall postoperative nausea (0 24 h).

Figure 2. Overall postoperative vomiting (0 24 h).

regain muscle strength (20), and 2 of 50 control

patients were excluded from the efficacy analysis of
another study because they needed muscle reversal

for muscle weakness (19). There were no other reports of other side effects in patients given neostigmine or in the control patients.

ANESTH ANALG
2005;101:1349 55

ANESTHETIC PHARMACOLOGY
CHENG ET AL.
NEOSTIGMINE AND PONV.

1353

Table 3. Multiple Logistic Regression Analysis of Overall Postoperative Vomiting

Binary Outcome of 24-h Postoperative
Vomiting
Neostigmine dose (per mg)
Anticholinergic
Glycopyrrolate
Atropine
Centera
Janhunen 1972 (28)
King 1988 (24)
Boeke 1994 (20)
Watcha 1995 (27)
Nelskyla 1998 (12)
Joshi 1999 (23)
Lovstad 2001 (25)
Chhibber 1999 (11)
Constant

(se)
0.277 (0.175)
0.729 (0.507)
1.140 (0.464)
1.657 (0.339)
0.207 (0.427)
0.136 (0.330)
0.132 (0.383)
0.968 (0.321)
0.945 (0.308)
2.094 (0.448)
1.042 (0.486)
0.478 (0.199)

P value

Odds ratio (e) (95% CI)

0.110
0.030
0.150
0.010

0.000
0.630
0.680
0.730
0.003
0.002
0.000
0.030
0.020

1.32 (0.941.86)
1.00
0.48 (0.181.30)
0.32 (0.130.79)
1.00
5.24 (2.7010.19)
0.81 (0.351.88)
0.87 (0.461.67)
1.14 (0.542.42)
0.38 (0.200.71)
0.39 (0.210.71)
0.12 (0.050.30)
2.84 (1.097.35)
0.62

CI confidence interval.
a
Hovorka et al. (22) was the largest study and was used as the reference for the centers.

Discussion
We found insufficient evidence that the use of neostigmine, accompanied by either atropine or glycopyrrolate, increases the RR of early, delayed, or overall PON
or POV. This finding contrasts with a previous metaanalysis in which the authors concluded that neostigmine in doses of 2.5 mg or larger increases PONV (10).
Several differences might account for this discrepancy.
For example, a limitation of the previous metaanalysis (10) is its treatment of a study by Janhunen
and Tammisto (28) in which five different modes of
general anesthesia in patients undergoing cholecystectomy or vein stripping were compared. In Janhunen
and Tammistos pseudo-randomized study, two
groups of patients (I and II) were given meperidine
and reversed with neostigmine and atropine; however, the second group also received halothane (group
II). Tramer and Fuchs-Buder (10) combined the data
from groups I and II and compared them with those of
patients in another group (IV) who received meperidine but not neostigmine or halothane. Considering
that volatile anesthetics are a major cause of POV (5),
by including the data from patients of group II who
received halothane, Tramer and Fuchs-Buder (10)
might have some introduced bias. Moreover, because
the trial was not randomized, the proportion of patients undergoing cholecystectomy or vein stripping
was dissimilar, which may have introduced further
bias. We avoided this problem by limiting our comparison to cholecystectomy patients from Janhunen
and Tammistos group I (meperidine and neostigmine) with those of group IV (meperidine without
neostigmine). Thus, in spite of including 2 additional
studies in our meta-analysis, it incorporates only 933
patients as opposed to 1134 patients in the analysis by
Tramer and Fuchs-Buder (10).

Tramer and Fuchs-Buder also identified a dosedependent relationship between neostigmine and
PONV, which we were unable to confirm. A closer
look at their Figure 2 reveals that the label of the Y-axis
should probably be risk reduction rather than
number-needed-to-treat, as published. But even then,
the 1 at the top of the dotted line should be 0 and
values for the 1.5-mg neostigmine were less than with
no antagonism. This would represent a negative effect
at small dose and would be inconsistent with a classical pharmacological dose-response relationship. Furthermore, as dose cannot be considered as a covariate
in RevMan, we subjected the data to logistic regression analysis, which showed that the dose of neostigmine did not exert a statistically significant effect on
the rate of PONV. Furthermore, center effects (i.e.,
where the study was performed) were an order of
magnitude larger than the dose dependence, suggesting that the nonsignificant effect of neostigmine dose
is considerably smaller than other influences.
A further limitation of the previous meta-analysis is
that it did not include atropine or glycopyrrolate as
potential confounders; instead, the authors argued
that the choice of the anticholinergic partner drug
does not affect PONV. Our logistic regression analysis
revealed that atropine was associated with a statistically significant decreased risk for POV whereas glycopyrrolate was not. This result is supported by the
study from Chhibber et al. (11) in which atropine was
associated with significantly less POV when compared directly with glycopyrrolate. Thus, it could be
hypothesized that atropine is a better antiemetic than
glycopyrrolate because of its central anticholinergic
effects. As patients given atropine received about
0.5 mg less neostigmine than those given glycopyrrolate, only a multivariate analysis could correct for
those confounders. Ignoring the different antiemetic

1354

ANESTHETIC PHARMACOLOGY
NEOSTIGMINE AND PONV.

CHENG ET AL.

effects of the anticholinergic partner drugs may thus

have contributed to the appearance of a dose-response
relationship for neostigmine in the previous metaanalysis. However, because there is only one true
head-to-head comparison of atropine versus glycopyrrolate (12), these results should be interpreted with
some caution.
It is possible that the rate of PONV is dependent on
the ratio between neostigmine and the coadministered
anticholinergic drug. However, the ratio of neostigmine to glycopyrrolate was 5:1 in almost all of the
studies included in our meta-analysis, and we only
had 2 studies in which atropine was the anticholinergic. This consistency in the ratio of neostigmine to
glycopyrrolate precluded introducing this factor into
the multiple regression analysis (colinearity problem).
There were only three cases of residual muscle
weakness noted in the control groups. Among the 10
studies we used for efficacy analysis, only 1 of 933
patients was reported to have residual muscle relaxation requiring treatment (0.1%). Although this infrequent incidence of residual paralysis suggests that
antagonism of neuromuscular block may not be necessary, numerous complications have been reported
when patients are inadequately antagonized (29).
Search strategies of systematic reviews are designed
to locate all relevant studies pertinent to the question.
To achieve the highest level of evidence, metaanalyses provide an objective approach to quantify the
effects of all data available from trials. It is, therefore,
conceivable that the point estimates from metaanalyses have more external validity than single studies. However, the best approach to weighing the relative impact of studies remains in dispute. For
example, we used the standard inversed standard error to weight the studies, although this approach can
lend too much weight to small studies (30,31). Point
estimates regarding the effectiveness should, therefore, be interpreted with some caution. Another problem of meta-analyses is the heterogeneity of reported
end-points (e.g., PON and POV) and when they are
measured (early, delayed, or overall period). A consequence is that, as in our analysis, point estimates for
different outcomes are not necessarily derived from
the same trials. Thus, although analysis of the early
postoperative period suggests that neostigmine might
increase early PON but not early POV, such interpretations should be viewed with a high degree of
skepticism.
Our meta-analysis failed to demonstrate that
neostigmine leads to a clinically important increase in
the risk of PONV. This might be a result of the limited
number of patients available for analysis. For example,
the incidences of overall PON for the neostigmine and
the control groups were 41% and 33%, respectively, so
that the average group size of 280 patients only provided a 43% power to detect this absolute difference of

ANESTH ANALG
2005;101:1349 55

8% (which would be of limited clinical importance in

any case). Thus, only a large, well-designed, randomized controlled trial can fully resolve this issue (32).
Such a trial should also address whether there is a
dose-response relationship between neostigmine and
PONV and the effect of atropine versus glycopyrrolate, possibly by a using a factorial design, which has
been proven to be a powerful tool (6,33,34). Defining a
clinically important decrease in PONV of approximately 25% for the omission of neostigmine would
require at least 372 patients if an incidence of 60%
were to be studied, 524 patients for an incidence of
50%, and 744 for 40% (13).
In conclusion, our meta-analysis suggested that
neostigmine does not increase the risk of POV in the
early, delayed, or overall postoperative period and
that there is insufficient evidence to conclude that
neostigmine leads to a clinically important increase in
the risk of PON. Thus, concerns about the effect of
neostigmine on PONV should probably not influence
the clinicians decision to antagonize neuromuscular
block.
We thank Nancy Alsip (Outcomes Research Institute, University of
Louisville) for her valuable help in reviewing and editing this
manuscript. We are grateful to Kaisa Nelskyla and Kari Korttila
(Helsinki, Finland) for providing the detailed results of their study,
and to Peter Kranke (Wuerzburg, Germany) for his constructive
comments.

References
1. Watcha MF, White PF. Postoperative nausea and vomiting: its
etiology, treatment, and prevention. Anesthesiology 1992;77:
162 84.
2. Kovac AL. Prevention and treatment of postoperative nausea
and vomiting. Drugs 2000;59:213 43.
3. Gan TJ, Meyer T, Apfel CC, et al. Consensus guidelines for the
management of postoperative nausea and vomiting. Anesth
Analg 2003;97:6271.
4. Apfel CC, Kranke P, Eberhart LH, et al. Comparison of predictive models for postoperative nausea and vomiting. Br J Anaesth 2002;88:234 40.
5. Apfel CC, Kranke P, Katz MH, et al. Volatile anaesthetics may
be the main cause of early but not delayed postoperative
vomiting: a randomized controlled trial of factorial design. Br J
Anaesth 2002;88:659 68.
6. Apfel CC, Korttila K, Abdalla M, et al. A factorial trial of six
interventions for the prevention of postoperative nausea and
vomiting. N Engl J Med 2004;350:244151.
7. Turner DA, Smith G. Evaluation of the combined effects of
atropine and neostigmine on the lower oesophageal sphincter.
Br J Anaesth 1985;57:956 9.
8. Proakis AG, Harris GB. Comparative penetration of glycopyrrolate and atropine across the blood-brain and placental barriers
in anesthetized dogs. Anesthesiology 1978;48:339 44.
9. Hood DD, Eisenach JC, Tuttle R. Phase I safety assessment of
intrathecal neostigmine methylsulfate in humans. Anesthesiology 1995;82:331 43.
10. Tramer MR, Fuchs-Buder T. Omitting antagonism of neuromuscular block: effect on postoperative nausea and vomiting and
risk of residual paralysis: a systematic review. Br J Anaesth
1999;82:379 86.

ANESTH ANALG
2005;101:1349 55

11. Chhibber AK, Lustik SJ, Thakur R et al. Effects of anticholinergics on postoperative vomiting, recovery, and hospital stay in
children undergoing tonsillectomy with or without adenoidectomy. Anesthesiology 1999;90:697700.
12. Nelskyla K, Yli-Hankala A, Soikkeli A, Korttila K. Neostigmine
with glycopyrrolate does not increase the incidence or severity
of postoperative nausea and vomiting in outpatients undergoing gynaecological laparoscopy. Br J Anaesth 1998;81:757 60.
13. Apfel CC, Roewer N, Korttila K. How to study postoperative
nausea and vomiting. Acta Anaesthesiol Scand 2002;46:921 8.
14. Jadad AR, Moore RA, Carroll D, et al. Assessing the quality of
reports of randomized clinical trials: is blinding necessary? Control Clin Trials 1996;17:112.
15. Huang CH, Wang MJ, Susetio L, et al. Comparison of the
combined effects of atropine and neostigmine with atropine and
edrophonium on the occurrence of postoperative nausea and
vomiting. Ma Zui Xue Za Zhi 1993;31:113 6.
16. Naguib M, Abdulatif M, al-Ghamdi A. Dose-response relationships for edrophonium and neostigmine antagonism of rocuronium bromide (ORG 9426)-induced neuromuscular blockade.
Anesthesiology 1993;79:739 45.
17. Devcic A, Munshi CA, Gandhi SK, Kampine JP. Antagonism of
mivacurium neuromuscular block: neostigmine versus edrophonium. Anesth Analg 1995;81:10059.
18. McCourt KC, Mirakhur RK, Kerr CM. Dosage of neostigmine
for reversal of rocuronium block from two levels of spontaneous
recovery. Anaesthesia 1999;54:6515.
19. Bevan DR, Kahwaji R, Ansermino JM, et al. Residual block after
mivacurium with or without edrophonium reversal in adults
and children. Anesthesiology 1996;84:3627.
20. Boeke AJ, de Lange JJ, van Druenen B, Langemeijer JJ. Effect of
antagonizing residual neuromuscular block by neostigmine and
atropine on postoperative vomiting. Br J Anaesth 1994;72:654 6.
21. Ding Y, Fredman B, White PF. Use of mivacurium during laparoscopic surgery: effect of reversal drugs on postoperative recovery. Anesth Analg 1994;78:450 4.
22. Hovorka J, Korttila K, Nelskyla K, et al. Reversal of neuromuscular blockade with neostigmine has no effect on the incidence
or severity of postoperative nausea and vomiting. Anesth Analg
1997;85:1359 61.
23. Joshi GP, Garg SA, Hailey A, Yu SY. The effects of antagonizing
residual neuromuscular blockade by neostigmine and glycopyrrolate on nausea and vomiting after ambulatory surgery. Anesth
Analg 1999;89:628 31.

ANESTHETIC PHARMACOLOGY
CHENG ET AL.
NEOSTIGMINE AND PONV.

1355

24. King MJ, Milazkiewicz R, Carli F, Deacock AR. Influence of

neostigmine on postoperative vomiting. Br J Anaesth 1988;61:
403 6.
25. Lovstad RZ, Thagaard KS, Berner NS, Raeder JC. Neostigmine 50
g kg1 with glycopyrrolate increases postoperative nausea in
women after laparoscopic gynaecological surgery. Acta Anaesthesiol Scand 2001;45:495500.
26. Walsh C, Smith CE, Ryan B, et al. Postoperative vomiting following strabismus surgery in paediatric outpatients: spontaneous versus controlled ventilation. Can J Anaesth 1988;35:315.
27. Watcha MF, Safavi FZ, McCulloch DA, et al. Effect of antagonism of mivacurium-induced neuromuscular block on postoperative emesis in children. Anesth Analg 1995;80:7137.
28. Janhunen L, Tammisto T. Postoperative vomiting after different
modes of general anaesthesia. Ann Chir Gynaecol Fenn 1972;
61:1529.
29. Berg H, Roed J, Viby-Mogensen J, et al. Residual neuromuscular
block is a risk factor for postoperative pulmonary complications: a
prospective, randomised, and blinded study of postoperative pulmonary complications after atracurium, vecuronium and pancuronium. Acta Anaesthesiol Scand 1997;41:1095103.
30. Fuchs-Buder T, Mencke T. Use of reversal agents in day care
procedures (with special reference to postoperative nausea and
vomiting). Eur J Anaesthesiol Suppl 2001;23:539.
31. Sterne JA, Gavaghan D, Egger M. Publication and related bias in
meta-analysis: power of statistical tests and prevalence in the
literature. J Clin Epidemiol 2000;53:1119 29.
32. Cappelleri JC, Ioannidis JP, Schmid CH, et al. Large trials vs
meta-analysis of smaller trials: how do their results compare?
JAMA 1996;276:1332 8.
33. Apfel CC, Korttila K, Abdalla M, et al. An international multicenter protocol to assess the single and combined benefits of
antiemetic interventions in a controlled clinical trial of a
222222 factorial design (IMPACT). Control Clin Trials
2003;24:736 51.
34. McAlister FA, Straus SE, Sackett DL, Altman DG. Analysis and
reporting of factorial trials: a systematic review. JAMA 2003;289:
254553.

Rocuronium Versus Succinylcholine for Rapid Sequence

Induction of Anesthesia and Endotracheal Intubation:
A Prospective, Randomized Trial in Emergent Cases
Mathias Sluga, MD, Wolfgang Ummenhofer, MD, Wolfgang Studer,
Martin Siegemund, MD, and Stephan C. Marsch, MD, DPhil

MD,

Department of Anesthesia, Krankenhaus Thusis, Switzerland

When anesthesia is induced with propofol in elective

cases, endotracheal intubation conditions are not different between succinylcholine and rocuronium approximately 60 s after the injection of the neuromuscular relaxant. In the present study, we investigated whether,
in emergent cases, endotracheal intubation conditions
obtained at the actual moment of intubation under succinylcholine differ from those obtained 60 s after the
injection of rocuronium. One-hundred-eighty adult patients requiring rapid sequence induction of anesthesia
for emergent surgery received propofol (1.5 mg/kg)
and either rocuronium (0.6 mg/kg; endotracheal intubation 60 s after injection) or succinylcholine (1 mg/kg;
endotracheal intubation as soon as possible). The time
from beginning of the induction until completion of the

rapid sequence induction of anesthesia and endotracheal intubation are indicated in emergency situations in the presence of a full stomach or other conditions with an increased risk of
aspiration. Traditionally, succinylcholine has been the
neuromuscular blocking drug of choice for rapid sequence induction of anesthesia. However, as a result
of its depolarizing effect, succinylcholine can have
serious side effects and is contraindicated in many
conditions. Rocuronium has the most rapid onset of
the currently available nondepolarizing neuromuscular blocking drugs. Therefore, many studies have investigated whether rocuronium may be a suitable alternative to succinylcholine. A meta-analysis of the
Cochrane collaboration concluded that when propofol

Accepted for publication April 21, 2005.

Address correspondence to Stephan Marsch, MD, DPhil, Medizinische Intensivstation, Kantonsspital, 4031 Basel, Switzerland. Address e-mail to smarsch@uhbs.ch.
DOI: 10.1213/01.ANE.0000180196.58567.FE

1356

Anesth Analg 2005;101:135661

intubation was shorter after the administration of succinylcholine than after rocuronium (median time 95 s
versus 130 s; P 0.0001). Endotracheal intubation conditions, rated with a 9-point scale, were better after succinylcholine administration than after rocuronium (8.6
1.1 versus 8.0 1.5; P 0.001). There was no significant difference in patients with poor intubation conditions (7 versus 12) or in patients with failed first intubation attempt (4 versus 5) between the groups. We
conclude that during rapid sequence induction of anesthesia in emergent cases, succinylcholine allows for a
more rapid endotracheal intubation sequence and creates superior intubation conditions compared with
rocuronium.
(Anesth Analg 2005;101:1356 61)

is used to rapidly induce anesthesia, endotracheal intubation conditions are not statistically different between succinylcholine and rocuronium (1). Before applying this evidence in daily practice, some important
limitations of the Cochrane Review have to be recognized: (a) most of the patients receiving propofol were
elective cases; (b) only a small number of emergent
cases actually underwent a rapid sequence induction
of anesthesia and endotracheal intubation with propofol and rocuronium; and (c) in most studies included
in the meta-analysis, tracheas were intubated approximately 60 s after the injection of the neuromuscular
blocking drug, yet clinical practice may allow intubation sooner than 60 s after the injection of succinylcholine. It is currently not known whether endotracheal
intubation conditions obtained at the actual moment
of intubation under succinylcholine differ from those
obtained 60 s after the injection of rocuronium.
Accordingly, the aim of the present study was to
compare rocuronium with the current practice of the
use of succinylcholine (i.e., endotracheal intubation as
soon as possible) in patients requiring rapid sequence
induction of anesthesia and endotracheal intubation
for emergent surgery. The hypotheses to be tested
2005 by the International Anesthesia Research Society
0003-2999/05

ANESTH ANALG
2005;101:1356 61

ANESTHETIC PHARMACOLOGY
SLUGA ET AL.
ROCURONIUM VERSUS SUCCINYLCHOLINE FOR RAPID SEQUENCE INDUCTION OF ANESTHESIA

were that (a) succinylcholine would allow for an earlier completion of the endotracheal intubation sequence and (b) succinylcholine would create superior
intubation conditions at the actual time of intubation.

Methods
The study took place in the Hospital of Thusis, a rural
Level III center. All adult (age, 18 yr) patients undergoing emergent surgery under general anesthesia
were eligible. Indications for emergent surgery were
mainly trauma (the hospital is located in a tourist
region with skiing accidents in winter and climbing
accidents in summer) and laparotomies. Exclusion criteria were hyperkalemia, neurologic disorders, burns,
familial history of malignant hyperthermia, cesarean
delivery, complications during birth before delivery,
known or anticipated difficult endotracheal intubation
warranting awake fiberoptic intubation, contraindication against the use of propofol (e.g., shock) and allergy to rocuronium. The study was approved by the
regional Ethics Committee, and written informed consent was obtained during the preoperative visit. The
primary outcomes of the study were the duration of
the endotracheal intubation sequence and intubation
conditions. Using a 9-point grading system for intubation conditions (Table 1), a difference of at least 1.0
points was considered to be of clinical relevance. A
power analysis revealed that 85 patients were required for each study group to detect that difference
with a power of 0.9 and a two-sided of 0.05. To
account for protocol violations related to an emergent
procedure, we planned to enroll 90 patients per group.
Patients were randomly allocated (sealed envelopes) to receive either 0.6 mg/kg of rocuronium (Esmeron, Organon, Switzerland) or 1.0 mg/kg of succinylcholine (Lystenon, Nycomed, Switzerland) as
the neuromuscular blocking drug. No premedication
was administered. Upon arrival in the operating
room, a 18-gauge cannula was inserted in a forearm
vein. Routine monitoring was used. End-tidal carbon
dioxide was measured using the side-stream method
(Cardiocap, Datex, Finland). Electrodes of a nerve
stimulator (Healthcare NS 272; Fisher & Paykel, New
Zealand) were placed over the left ulnar nerve.
One of three experienced staff anesthesiologists
(MS, WU, or SM), assisted by a registered anesthetic
nurse and a scrub nurse, was present throughout the
whole procedure, guided the injection of drugs, and
performed the endotracheal intubation. The staff anesthesiologist was not blinded to the neuromuscular
blocking drug used, and the management of difficulties and complications, if any, was left to his discretion. To minimize bias, intubations were performed by
three different anesthesiologists who had no personal
preference for one of the two neuromuscular blocking
drugs.

1357

The endotracheal intubation sequence was defined

as time interval between the injection of propofol and
the first appearance of end-tidal carbon dioxide on the
screen of the monitor. After 3 min of the administration of oxygen, cricoid pressure was applied, and anesthesia was induced with fentanyl 2 g/kg and
propofol 1.5 mg/kg. The neuromuscular blocking
drug was injected as soon as the eyelid reflex had
disappeared, and the nerve stimulator was switched
to the single-twitch mode (rate, one twitch per second). Laryngoscopy was started either after the cessation of fasciculations in the lower extremities (2), if
any, the cessation of a visible motor response to continuous single-twitch nerve stimulation, or after 50 s
(anticipated time of intubation 60 s after the injection
of the neuromuscular blocking drug), whichever was
earlier. Endotracheal intubations were performed using a Macintosh size 3 blade and a tracheal tube
(Mallinckrodt Hi-Contour, Mallinckrodt, Ireland) with
an internal diameter of 7.5 cm in women and of 8.5 cm
in men. The timing of events was performed by the
anesthetic nurse.
Intubation conditions are usually evaluated using
the following factors: (a) ease of laryngoscopy, (b)
position and movement of the vocal cords, and (c)
response to intubation of the airway and the limbs (3).
However, previous studies differ in that either a numerical (1) or a qualitative (4) score was derived from
these factors. To allow for a comparison with both
types of scoring systems previously used, we provide
both a numerical and a qualitative rating. Both ratings
are based on a scoring system proposed for good
clinical research practice in studies of neuromuscular
blocking drugs (3). The intubating anesthesiologist
rated the ease of laryngoscopy, the movement and
position of the vocal cords, and the reaction to intubation, as demonstrated in Table 1.
Desaturation was defined as either a saturation
90% or a decrease in saturation of 5% occurring at
any time between the start of the induction sequence
and 3 min after the completion of the intubation.
Data, presented as mean sd unless otherwise
stated, were analyzed using SPSS 12.0 for Windows, a
commercially available statistical software (SPSS, Chicago, IL). Two-way analysis of variance, unpaired Students t-test, Mann-Whitney test, Fishers exact test,
and the logrank test were applied, as appropriate.
General linear modeling was used to assess differences among the 3 intubating anesthesiologists with
regard to scoring of the intubation conditions. A P
0.05 was considered to represent statistical
significance.

Results
During the study period ending with the completion
of the protocol in the 180th patient, 234 consecutive

1358

ANESTHETIC PHARMACOLOGY SLUGA ET AL.

ROCURONIUM VERSUS SUCCINYLCHOLINE FOR RAPID SEQUENCE INDUCTION OF ANESTHESIA

ANESTH ANALG
2005;101:1356 61

Table 1. Scoring System for Endotracheal Intubation Conditions

Score 3
Laryngoscopy
Jaw relaxation
Resistance to blade
Vocal cords
Position
Movement
Intubation response
Limb movement
Coughing

Score 2

Score 1

Relaxed
None

Acceptable relaxation
Slight resistance

Poor relaxation
Active resistance

Abducted
None

Intermediate
Moving

Closed
Closing

None
None

Slight
Diaphragmatic

Vigorous
Severe coughing or bucking

The factors laryngoscopy, vocal cords, and response to intubation are individually rated with a score from 1 to 3. The assignment of a score for each of the
three factors is based on the lower rating of two variables, e.g., the combination of the variables no limb movement and no coughing results in a score of
3 for the factor response to intubation, whereas the combination of the variables no limb movement and severe coughing results in a score of 1. The numerical
intubation score was obtained by summing the scores assigned to the factors laryngoscopy, vocal cords, and response to intubation. The maximum score is thus
9, whereas the minimum score is 3. The qualitative intubation scores was defined as follows:
excellent intubation conditions: all 3 factors were rated with a score of 3.
good intubation conditions: all 3 factors were rated either with a score of 3 or 2.
poor intubation conditions: the presence of one factor rated with a score of 1.
Excellent and good intubation conditions are considered clinically acceptable, whereas poor intubation conditions are considered clinically not acceptable (3).

patients underwent emergency surgery under general

anesthesia. Five had to be excluded because of predefined exclusion criteria (2 cesarean delivery, 2 hemorrhagic shock, and 1 hyperkalemia), 16 refused to
participate, and the enrollment of 33 was missed,
mainly because of high workload. One-hundredeighty patients were randomized, received the allocated treatment, and were included in the analysis
(Table 2).
The median time interval between the beginning of
the administration of propofol and the disappearance
of the eyelid reflex was 30 s (interquartile range, 18.5 s)
in the succinylcholine group and 26 s (interquartile
range, 20 s) in the rocuronium group (P 1.0). Figure
1 depicts the time interval from injection of the neuromuscular blocking drug to the cessation of a visible
motor response to continuous single-twitch nerve
stimulation of the ulnar nerve. This time interval was
significantly shorter (P 0.0001) in the succinylcholine group (median time, 40 s) compared with the
rocuronium group (median time, 70 s). Figure 2 depicts the time interval between the beginning of the
administration of propofol and the first appearance of
end-tidal carbon dioxide after endotracheal intubation, which was significantly shorter (P 0.0001) in
the succinylcholine group (median time, 95 s) compared with the rocuronium group (median time,
130 s).
Scores for endotracheal intubation conditions were
significantly higher in the succinylcholine group than
in the rocuronium group (8.6 1.1 versus 8.0 1.5; P
0.001). This difference resulted almost exclusively
from a difference in the subscore rating the response
to intubation (2.8 0.5 versus 2.3 1.0; P 0.0001),
whereas there was no difference in the subscores for
laryngoscopy (2.9 0.3 versus 2.9 0.3; P 0.91) and
vocal cords (2.9 0.4 versus 2.8 0.6; P 0.23).
Figure 3 depicts the scores for intubating conditions.

Table 2. Patient Demographics

Age (yr)
Sex (m/f)
ASA Status (I/II/III/IV)
Height (cm)
Weight (kg)

Succinylcholine
(n 90)

Rocuronium
(n 90)

43 18
39/51
14/50/24/2
170 9
70 14

49 21
36/54
14/48/28/0
167 9
68 14

Mean sd. There was no statistically significant difference between the

groups.

Note that compared with the rocuronium group, there

were significantly more excellent intubation conditions in the succinylcholine group (Fig. 3). However,
there was no difference in patients with poor intubation conditions between the groups (7 versus 12; P
0.33). General linear modeling showed (a) no significant difference among the 3 intubating anesthesiologists with regard to the rating of the intubation conditions (F2,168 0.21; P 0.81), (b) no significant
interaction of the 2 between-subject factors intubating
anesthesiologist and neuromuscular blocking drug
(F2,168 1.47; P 0.23), and (c) no significant interaction of the between-subject factor intubating anesthesiologist and the within-subject factor subscores of
intubation conditions (F4,336 0.87; P 0.48). This
indicates that there was no systematic difference in
scoring among the 3 intubating anesthesiologists.
Eighty-six of 90 patients in the succinylcholine
group and 85 of 90 patients in the rocuronium group
were intubated during the first attempt (P 1.0). All
remaining nine patients were successfully endotracheally intubated in the second attempt. The reasons for
the four failures of the first intubation attempt in the
succinylcholine group were one poor intubation condition (numerical score 3), one esophageal intubation

ANESTH ANALG
2005;101:1356 61

ANESTHETIC PHARMACOLOGY
SLUGA ET AL.
ROCURONIUM VERSUS SUCCINYLCHOLINE FOR RAPID SEQUENCE INDUCTION OF ANESTHESIA

Figure 1. Kaplan-Meyer curve of the probability of the disappearance of a visible motor response to a continuous single-twitch
stimulation of the ulnar nerve after injection of succinylcholine or
rocuronium. Time 0 denotes the injection of the neuromuscular
blocking drug. Curves differ significantly (P 0.0001; logrank
test).

1359

Figure 3. Endotracheal intubation conditions during rapid sequence

induction of anesthesia and endotracheal intubation with succinylcholine or rocuronium as the neuromuscular blocking drug. The
scoring system is explained in Table 1. *P 0.05 between the 2
neuromuscular blocking drugs (Fishers exact test).

patients (one in each group) with desaturation,

whereas 8 of 14 desaturations were associated with an
excellent intubation score. Four of 14 desaturations (2
in each group) occurred in patients with a second
intubation attempt. Compared with the patients without desaturation, the time interval from the beginning
of the administration of propofol and the completion
of the intubation was longer in patients with desaturation (134 9 s versus 116 3 s; P 0.047).

Discussion
Figure 2. Kaplan-Meyer curve of the probability of the completion
of the endotracheal intubation sequence including succinylcholine
or rocuronium as the neuromuscular blocking drug. Time 0 denotes
the beginning of the injection of the induction drug propofol. The
endotracheal intubation sequence was defined to be completed
upon the first appearance of end-tidal carbon dioxide after intubation. Curves differ significantly (P 0.0001; logrank test).

(intubation score excellent), and two difficult anatomy (intubation scores excellent and good, respectively) that could be mastered in the second attempt
by mounting the tube on a stylet. The reasons for the
five failures of the first intubation attempt in the rocuronium group were one poor intubation condition
(numerical score 4), two esophageal intubations (intubation scores excellent and good, respectively), and
two difficult anatomy (intubation score excellent in
both cases) that could be mastered in the second attempt by mounting the tube on a stylet. Thus, poor
intubation conditions were observed in only two of
the nine patients (one in each group) not intubated in
the first attempt.
A desaturation occurred in 5 of 90 patients in the
succinylcholine group and in 9 of 90 patients of the
rocuronium group (P 0.40). Poor endotracheal intubation conditions were observed in only 2 of the 14

In the present study, we compared rocuronium with

the current practice of the use of succinylcholine (i.e.,
endotracheal intubation as soon as possible) in patients requiring rapid sequence induction of anesthesia and endotracheal intubation for emergent surgery.
When succinylcholine was used as the neuromuscular
blocking drug for rapid sequence induction of anesthesia, the median intubation sequence was 35 s
shorter than when rocuronium was used. Succinylcholine created excellent intubation conditions more often
than rocuronium, and there was a statistically significant difference of 0.5 points on a 9-point grading scale
of intubation conditions in favor of succinylcholine.
However, as far as clinically acceptable intubating
conditions and failed intubation attempts are concerned, the two relaxants were not statistically
different.
Analyzing the available evidence up to the year
2000, a Cochrane Review concluded that for rapid
sequence induction of anesthesia, succinylcholine created superior endotracheal intubation conditions to
rocuronium when comparing excellent intubation
conditions. Using the less stringent clinically acceptable intubation conditions, the two drugs were not
statistically different (1). Moreover, based on a subgroup analysis, the Cochrane Review concluded that

1360

ANESTHETIC PHARMACOLOGY SLUGA ET AL.

ROCURONIUM VERSUS SUCCINYLCHOLINE FOR RAPID SEQUENCE INDUCTION OF ANESTHESIA

intubation conditions did not statistically differ between the administration of succinylcholine and rocuronium when propofol was used as the drug to induce
anesthesia (1). Several potential limitations of these
conclusions are noteworthy. Only 24 of the 1606 patients included in the Cochrane Review were emergent cases that actually underwent a true rapid sequence induction of anesthesia and endotracheal
intubation with both propofol and rocuronium. All 24
patients were part of a single study and received
1 mg/kg of rocuronium (4). Moreover, only 47 of the
640 patients included in the subgroup using propofol
as the induction drug (4 12) were emergency cases
undergoing a true rapid sequence induction of anesthesia and endotracheal intubation. From the remaining 593 elective cases, approximately 50% (n 290)
did not undergo a true rapid sequence induction of
anesthesia. Most previous studies comparing succinylcholine and rocuronium assessed endotracheal intubation conditions approximately 60 seconds after the
injection of the neuromuscular blocking drug (1).
Whereas this is an appropriate time interval for rocuronium, a delay between injection of succinylcholine
and start of laryngoscopy of 50 seconds or more does
not reflect current practice; most, if not all, anesthesiologists choosing succinylcholine for rapid sequence
induction of anesthesia and endotracheal intubation
take advantage of its rapid onset of action and start
laryngoscopy as quickly as possible, i.e., after the cessation of fasciculations. Indeed, 50 seconds after the
injection of the neuromuscular blocking drug, i.e., the
time of the beginning of the laryngoscopy in previous
studies, the intubation sequence was already completed in more than one-third of the patients in the
succinylcholine group of the present study.
Based on current evidence, the induction of anesthesia sequence of the present study was chosen to
achieve the best possible endotracheal intubation conditions for the rocuronium group. Propofol was used
as the induction drug because this anesthetic seems to
be superior to all other drugs with regard to intubating conditions after rocuronium injection (1). Fentanyl
(2 g/kg) was added to the induction sequence because opioids, in doses equivalent to alfentanil 20
g/kg, were found to significantly improve intubating conditions after rocuronium administration (13).
Intubation was attempted 60 seconds after the injection of rocuronium because this seemed to be the
earliest moment when acceptable intubation conditions can be reliably achieved (1). Rocuronium was
used in a dose of 0.6 mg/kg because there seemed to
be no benefit of larger rocuronium doses on intubation
conditions when propofol was used as the induction
drug (1). Previous work demonstrated a dosedependent effect of rocuronium on both onset and
duration of neuromuscular block (14). Thus, there is
the possibility that larger doses of rocuronium would

ANESTH ANALG
2005;101:1356 61

allow for an earlier intubation. However, all studies

comparing intubation conditions after different doses
of rocuronium did so after a predefined time interval
(usually 60 seconds after the injection of rocuronium).
Thus, it is unknown whether the earlier onset of neuromuscular block associated with doses larger than
0.6 mg/kg of rocuronium would translate into a clinical advantage, i.e., the possibility for an earlier intubation with at least the same intubation conditions
that are achievable at 60 seconds.
An important limitation of our study is the lack of a
double-blind design. Concealing the effects of drugs
that have visible effects such as fasciculations is inherently difficult. Moreover, because the two neuromuscular blocking drugs studied differ in onset time,
awareness of the time of the injection of the drug
results in unblinding. Thus, a perfect double-blind
design implies that the intubating anesthesiologist is
not able to see or overhear the patient and the team
performing the induction sequence and is immediately available to intubate the patients trachea after
the cessation of fasciculations. A rapid sequence induction of anesthesia is a high-risk procedure requiring the full attention of an appropriately trained anesthesiologist. Because, in our settings, the
simultaneous achievement of perfect blinding and optimal patient safety was not feasible, we opted for a
single-blind study design. The statistical analysis of
the effects and interactions of neuromuscular blocking
drugs and the intubation scores revealed a homogenous rating with no systematic differences among the
anesthesiologists performing the intubations.
The power of the present study was too small to
allow reliable conclusions on the incidence of complications. These issues should be addressed in large
multicenter trials. Interestingly, in the present study,
only a minority of failed first endotracheal intubation
attempts and desaturations were associated with a
low score of intubation conditions. If confirmed in
further trials, these findings may lead to a modification of the scoring system presently used.
What are the practical implications of our findings?
Choosing rocuronium instead of succinylcholine for
rapid sequence induction of anesthesia prolongs the
time of unprotected airway, i.e., the time interval from
beginning of the induction until completion of endotracheal intubation, from a median time of 95 seconds
to a median time of 130 seconds. The additional risk of
aspiration and desaturation resulting from a prolongation of the intubation sequence by a median time of
35 seconds is unknown, but it is most likely very small
in most patients. However, patients with an especially
high risk for aspiration or a desaturation may benefit
from a more rapid intubation. Choosing rocuronium
instead of succinylcholine for rapid sequence induction of anesthesia results in less optimal intubating
conditions. However, the difference between the two

ANESTH ANALG
2005;101:1356 61

ANESTHETIC PHARMACOLOGY
SLUGA ET AL.
ROCURONIUM VERSUS SUCCINYLCHOLINE FOR RAPID SEQUENCE INDUCTION OF ANESTHESIA

relaxants is small and mainly results from lower ratings in the subscore addressing the reaction to intubation, i.e., coughing or bucking. Because the reaction
to intubation occurs after the placement of the tube,
the relevance for patients safety is marginal. Until
more data on complications are available, we suggest
that anesthesiologists select the best treatment for
their patients undergoing a rapid sequence induction
of anesthesia on an individual basis by balancing intubation conditions and duration of the intubation
sequence against potential side effects.
Compared with the subgroup of patients receiving
propofol included in the recent Cochrane Review on
rapid sequence induction (1), in the present study, we
observed significantly more poor intubation conditions after both neuromuscular blocking drugs (19 of
180 versus 27 of 640 poor intubation conditions; P
0.007). Because most patients included in the Cochrane Review were elective cases not undergoing a
true rapid sequence induction of anesthesia, this difference is most likely explained by differences in the
patient population and the procedure. Although elective cases are valuable models for investigating the
effects of neuromuscular blocking drugs, findings obtained in this setting may thus not be necessarily
extrapolated to emergency situations.
In conclusion, in the context of a rapid sequence
induction of anesthesia with propofol and fentanyl in
emergent cases, succinylcholine allowed for a more
rapid endotracheal intubation sequence and created
superior intubation conditions than rocuronium. Presently, practitioners have to balance the quality of intubation conditions and the duration of the intubation
sequence against the potential for side effects. Largescale trials are required to address important safety
issues such as failed intubation attempts and desaturations associated with the use of succinylcholine or
rocuronium.

1361

References
1. Perry J, Lee J, Wells G. Rocuronium versus succinylcholine for
rapid sequence induction intubation. In: The Cochrane Library.
Issue 3. Hoboken, NJ: John Wiley & Sons, Ltd., 2004.
2. Ummenhofer WC, Kindler C, Tschaler G, et al. Propofol reduces
succinylcholine induced increase of masseter muscle tone. Can
J Anaesth 1998;45:41723.
3. Viby-Mogensen J, Englbaek J, Eriksson LI, et al. Good clinical
research practice (GCRP) in pharmacodynamic studies of neuromuscular blocking agents. Acta Anaesthesiol Scand 1996;40:
59 74.
4. Andrews JI, Kumar N, Van Den Brom RHG, et al. A large simple
randomized trial of rocuronium versus succinylcholine in rapidsequence induction of anaesthesia along with propofol. Acta
Anaesthesiol Scand 1999;43:4 8.
5. Abdulatif M, Al Ghamdi A, El Sanabary M. Rocuronium priming of atracurium-induced neuromuscular blockade: the use of
short priming intervals. J Clin Anesth 1996;8:376 81.
6. Chiu CL, Jaais F, Wang CY. Effect of rocuronium compared with
succinylcholine on intraocular pressure during rapid sequence
induction of anaesthesia. Br J Anaesth 1999;82:757 60.
7. Latorre F, Stanek A, Gervais HW, Kleemann PP. Intubation
requirements after rocuronium and succinylcholine. Anasthesiol Intensivmed Notfallmed Schmerzther 1996;31:470 3.
8. Le Corre F, Plaud B, Benhamou E, Debaene B. Visual estimation
of onset time at the orbicularis oculi after five muscle relaxants:
application to clinical monitoring of tracheal intubation. Anesth
Analg 1999;89:130510.
9. Naguib M, Samarkandi AH, Ammar A, Turkistani A. Comparison of suxamethonium and different combinations of rocuronium and mivacurium for rapid tracheal intubation in children.
Br J Anaesth 1997;79:450 5.
10. Puhringer FK, Khuenl-Brady KS, Koller J, Mitterschiffthaler G.
Evaluation of the endotracheal intubating conditions of rocuronium (ORG 9426) and succinylcholine in outpatient surgery.
Anesth Analg 1992;75:37 40.
11. Vinik HR. Intraocular pressure changes during rapid sequence
induction and intubation: a comparison of rocuronium, atracurium, and succinylcholine. J Clin Anesth 1999;11:95100.
12. Stoddart PA, Mather SJ. Onset of neuromuscular blockade and
intubating conditions one minute after the administration of
rocuronium in children. Paediatr Anaesth 1998;8:37 40.
13. Sparr HJ, Giesinger S, Ulmer H, et al. Influence of induction
technique on intubating conditions after rocuronium in adults:
comparison with rapid-sequence induction using thiopentone
and suxamethonium. Br J Anaesth 1996;77:339 42.
14. Wright PM, Caldwell JE, Miller RD. Onset and duration of
rocuronium and succinylcholine at the adductor pollicis and
laryngeal adductor muscles in anesthetized humans. Anesthesiology 1994;81:1110 5.

Nitric Oxide Is Not a Mediator of Inflammation-Induced

Resistance to Atracurium
Heidrun Fink, MD*, Ralph Bogdanski, MD*, Peter Luppa,
J. A. Jeevendra Martyn, MD, and Manfred Blobner, MD*

MD#,

*Klinik fur Anaesthesiologie der Technischen Universitat Munchen and #Institut fur Klinische Chemie und
Pathobiochemie der Technischen Universitat Munchen, Klinikum rechts der Isar, Munich, Germany; and Department of
Anesthesiology and Critical Care, Harvard Medical School and Anesthesia Services, Massachusetts General Hospital, and
Shriners Hospital for Children, Boston, MA

Resistance to atracurium as a result of increased drug

binding to 1-acid glycoprotein is associated with increased inducible nitric oxide synthase activity and increased nitric oxide levels in plasma. We investigated if
the inhibition of inducible nitric oxide synthase and
suppression of nitric oxide can reverse the resistance to
atracurium. As a model of 1-acid glycoprotein and nitric oxide increase, 84 male Sprague-Dawley rats received an IV injection of either 60 mg/kg Corynebacterium parvum (CP) or saline (control). The 2 groups (CP/
Control) were further divided into subgroups,
receiving the selective inducible nitric oxide synthase
inhibitor, N-Iminolysine, via drinking water at different concentrations. On day 4 post-CP injection, the

itric oxide (NO) is a key signal transducing molecule of many aberrant responses observed during systemic inflammation, including hemodynamic instability (1), decreased organ function (2),
decreased xenobiosis (3), and reduced drug action
(4,5). Previously we had shown that action of muscle
relaxants is also impaired at several levels by NOrelated effects (4,5). The plasma clearance of vecuronium, for example, depends on hepatic cytochrome
p450 activity, which is inhibited by NO but can be

Supported, in part, by a personal stipend from the Deutscher

Akademischer Austauschdienst (DAAD), Bonn, Germany to Heidrun Fink, a grant to Manfred Blobner from GlaxoWellcome, Germany, and by grants # GM 31569 22, GM 55082 08, and GM
611411 06 from the National Institutes of Health and from Shriners
Hospital for Children to J.A. Jeevendra Martyn.
Accepted for publication May 3, 2005.
Address correspondence to Heidrun Fink, MD, Klinik fur Anaesthesiologie der Technischen Universitat Munchen*, Klinikum rechts
der Isar, Ismaninger Str. 22, 81675 Munich, Germany. Address
electronic mail to h.fink@lrz.tum.de.
DOI: 10.1213/01.ANE.0000180832.62367.CC

1362

Anesth Analg 2005;101:13627

pharmacodynamics of atracurium were determined.

Plasma concentrations of nitric oxide, atracurium, and
1-acid glycoprotein were measured and acetylcholine
receptor numbers were quantified. In the CP groups,
N-Iminolysine suppressed nitric oxide levels in a dosedependent manner. Resistance to atracurium persisted.
1-acid glycoprotein serum levels remained increased
in all CP groups with no differences in acetylcholine
receptor expression. Our results suggest that the mechanism leading to increased expression of 1-acid glycoprotein and consecutive increased protein binding of
atracurium is not mediated by inducible nitric oxide
synthase induction and nitric oxide expression.
(Anesth Analg 2005;101:13627)

restored if NO synthesis is suppressed by the nonspecific NO synthase inhibitor NG-monomethyl-Larginine (5). Normalizing the hepatic clearance of vecuronium, however, revealed an additional effect of
sepsis or NO, evidenced as a resistance towards its
neuromuscular blocking effect. Studies with atracurium showed that the observed resistance to its
action as muscle relaxant was attributable to an
inflammation-associated increase in 1-acid glycoprotein (1-AGP), to which atracurium binds in serum (4).
The present study, therefore, addresses the question
of whether suppression of NO can reverse the acute
phase response to inflammation at a blood chemical
level, i.e., the increase of 1-AGP levels and at a functional level, i.e., by neutralizing the resistance to atracurium. Therefore, with the specific inducible nitric
oxide synthase (iNOS) inhibitor N-Iminolysine (NIL),
dose-response studies were performed on both
targets.

Methods
After governmental approval of the study (Regierung
von Oberbayern, AZ 211253170/97), 84 male
2005 by the International Anesthesia Research Society
0003-2999/05

ANESTH ANALG
2005;101:13627

Sprague-Dawley rats (Charles River GmbH, Kisslegg,

Germany) weighing 250 300 g were allowed to accommodate to the standard conditions of our animal
facility with free access to chow and water for 14 days.
Rats were divided into 2 groups: rats of the Corynebacterium parvum (CP) group (n 42) received an IV
injection of 60 mg/kg of a whole cell preparation of
heat-killed CP (Roche, Penzberg, Germany) in 0.5 mL
saline; rats of the control group (n 42) received the
solvent. The injection of heat-killed CP induces a granulomatous liver inflammation and systemic inflammatory response syndrome evidenced as increased NO
and 1-AGP levels peaking at the fourth day postinjection (4). To examine the role of NO in the altered
biochemical and pharmaceutical changes, the two
groups of rats (CP/control) were further divided into
six subgroups and received six different doses of the
selective iNOS inhibitor NIL via drinking water (2.2,
0.22, 0.022, 0.0022, 0.00022, and 0 mol/L of NIL) in
drip-free drinking bottles. NIL treatment was begun
1 day before the injection and was continued until the
fourth day after bacterial injection. Body weight and
water intake were recorded daily. On the fourth day
post-CP injection, effects of atracurium on neuromuscular function were evaluated. After these measurements the animals were killed by exsanguination. The
blood was centrifuged (3500 rpm, 4C, 10 min) and the
serum was stored for later analysis of nitrite/nitrate
(NO2/NO3) as well as 1-AGP levels. The gastrocnemius muscle was excised to determine the expression
of acetylcholine receptors.
Anesthesia was induced by inhalation of sevoflurane in a glass cylinder. After loss of consciousness,
the rats were endotracheally intubated and mechanically ventilated with oxygen in nitrous oxide (ratio
1:2). Anesthesia was maintained with 4% 6% sevoflurane. After cannulation of the left external jugular
vein, anesthesia was switched to a continuous infusion of propofol (20 40 mg kg1 h1) and fentanyl
(4 g kg1 h1). The left carotid artery was cannulated to measure mean arterial blood pressure and
perform blood gas analyses. After the start of IV anesthesia, all animals were allowed to stabilize over a
period of 60 min to eliminate effects of sevoflurane on
neuromuscular transmission. At all times the administration of anesthesia was adjusted according to cardiovascular signs of adequate anesthesia.
Throughout the experiment, arterial carbon dioxide
partial pressure (Paco2) was maintained between 36
and 44 mm Hg. Whenever necessary, base excess was
corrected with 1 mM sodium bicarbonate to values
between 2 2 mM. Arterial oxygen partial pressure
(Pao2), heart rate, and mean arterial blood pressure
were continuously monitored to ensure stable oxygenation and hemodynamic conditions throughout the
experiment. Rectal temperature was controlled between 36.8C and 37.2C with a warming blanket and

ANESTHETIC PHARMACOLOGY
FINK ET AL.
NITRIC OXIDE AND ATRACURIUM RESISTANCE

1363

heating lamp. Rats were excluded from the experiment if they were hemodynamically unstable (mean
arterial blood pressure 80 mm Hg) at the beginning
of the experiments or if their blood gas status throughout the experiment was not within the defined predetermined ranges (Pao2 100 mm Hg; pH, 7.36 7.44;
Paco2: 36 44 mm Hg; base excess 2 2 mM).
Neuromuscular function was monitored by evoked
mechanomyography (Myograph, Biometer, Copenhagen, Denmark). The sciatic nerve of the left leg was
exposed at its exit from the lumbosacral plexus and
stimulated using the train-of-four pattern (2 Hz for 2 s
every 12 s). The knee was pinned and firmly fixed. A
force transducer was connected to the Achilles tendon
and the contraction of the gastrocnemius muscle was
measured. Supramaximal stimulus and control twitch
height (T0) were established. The mechanomyographic response was stabilized over a period of at
least 10 min before determination of the individual
dose-response relation of atracurium (ED50) in each rat
using the cumulative method described previously
(4). Bolus doses of atracurium were given IV in increments between 0.2 and 0.8 mg/kg until the first twitch
of the train-of-four (T1) was below 5% of baseline
value. Each incremental dose was given only when the
previous dose had produced maximal effect, as indicated by 3 equal (2%) consecutive T1 twitches. After
the last dose of atracurium, twitch response was allowed to recover to baseline values. The recovery
interval was calculated as time at (T1/T0 75%)
time at (T1/T0 25%). After complete recovery of T1,
a continuous infusion of atracurium was started, and
the infusion rate adjusted to achieve a constant T1/T0
of 50%. After 10 min of stable T1/T0 50% at a fixed
infusion rate, steady-state conditions were assumed.
The required infusion rate was documented (infusion
rate at 50% neuromuscular block) and 1 mL of heparinized blood was withdrawn to determine total
plasma concentrations of atracurium. The blood was
immediately transferred to Eppendorff tubes containing 20 L 1M H2SO4 (to avoid degradation of atracurium) and centrifuged (3500 rpm, 10 min, 4C). The
supernatant was collected and 0.2 mL portions were
aliquoted into Eppendorff tubes containing 0.8 mL 15
mM H2SO4. The samples were immediately frozen at
70C. After blood sampling, both gastrocnemius
muscles were dissected from the surrounding structures and rapidly frozen in isopentane pre-cooled in
liquid nitrogen and stored at 70C. After this, animals were killed by exsanguination. The collected
blood was centrifuged (3500 rpm, 10 min, 4C) and
the gained serum immediately stored at 70C for
determination of 1-AGP and NO2/NO3-plasma
concentrations.
Activity of iNOS and production of NO was assessed by measuring its stable metabolites, NO2 and

1364

ANESTHETIC PHARMACOLOGY FINK ET AL.

NITRIC OXIDE AND ATRACURIUM RESISTANCE

NO3 (NO2/NO3), in plasma. Samples were deproteinized with 0.5 M NaOH and 10% ZnSO4. NO3 was then
converted to NO2 using HPLC on a cadmium column.
NO2 concentrations were determined spectrophotometrically at 540 nm using a method based on the
Griess reaction (6). Methemoglobin (MetHb) levels as
a result of NO binding to hemoglobin were measured
oximetrically as part of the blood gas analyses (Rapidlab 860, Bayer Diagnostics, Munich, Germany).
1-AGP was measured using a competitive chemiluminescence immunoassay (7). Signaling was induced by applying horseradish peroxidase-conjugated
streptavidin and quantifying the enzyme activity using an enhanced chemiluminometric method (Amerlite System, Ortho Clinical Diagnostics, Neckargemund, Germany). Polyclonal antibodies against rat
1-AGP were raised in rabbits using pure commercially available rat 1-AGP (Sigma, Deisenhofen,
Germany).
Atracurium plasma concentrations were determined by HPLC as previously described (4). In brief,
the method has a between-day imprecision of 3.7%
(coefficient of variation) at a mean value of 9.6 g/mL
atracurium. The recovery of spiked plasma samples
varied between 96% and 98.7%. The functional sensitivity of the method, being a measure of the lower
limit of quantitation, was found to be 0.8 g/mL.
Linearity of the HPLC determination could be demonstrated up to 120 g/mL. The plasma clearance of
atracurium during steady-state conditions was calculated as clearance infusion rate/plasma level.
The muscle was homogenized, the protein extracted
and the amount of membrane acetylcholine receptors
quantified by the 125I- -bungarotoxin binding assay
(8). The protein concentration of the muscle extract
was assayed using the Bio-Rad DC protein assay kit
(Bio-Rad Laboratories, Hercules, CA), and the content
of acetylcholine receptors was calculated and expressed in fmol/mg protein.
Data are described as mean values and 95% confidence intervals. Variables were compared between
groups using factorial analyses of variance with the 2
independent factors infection (CP versus control) and
NIL dose (0, 0.00022, 0.0022, 0.022, 0.22 mol/L). An
effect of NIL treatment on a respective variable was
considered if NIL or NIL infection proved to be
significant (P 0.05) and was followed post hoc by
factorial analyses of variance in the control groups
and, independently, in the CP groups to evaluate dose
dependency. If infection proved to be significant,
groups with the same NIL dose were compared by
unpaired Students t-tests post hoc. The dose-response
of atracurium was evaluated by linear regressions of
the degree of the blockade in logit scale and the respective cumulative dose of atracurium on log scale in
each animal. Slopes and intercepts of the so calculated
individual regressions served as determinates for the

ANESTH ANALG
2005;101:13627

dose-response and were subjected to statistical analysis. The doses of atracurium for a 50% neuromuscular
block (ED50) were interpolated using the means of
slope and intercept of each group and their 95% confidence borders.

Results
There was a 50% mortality rate in rats receiving 2.2
mM NIL regardless of CP or control group, making it
impossible to include data from these animals in the
statistical analysis. Rats of the 2.2 mM NIL groups that
were still alive on day four were not included in the
neuromuscular function tests because of hemodynamic instability and were killed by exsanguination.
Therefore, only some blood chemistry could be
performed.
There were no differences in water intake between
the different NIL groups of the CP or the control rats
at any day. However, from the third day postinjection
on, there was a significant difference in water intake
between CP (34 [3236], 33 [3234], 32 [30 34], 25
[24 26], 26 [24 27] mL/day from 1 day before NIL
administration till day 4) and control rats (36 [34 37],
35 [34 36], 34 [3236], 34 [3236]*, 34 [3235]* mL/day
from 1 day before NIL administration till day 4, *P
0.05 compared with CP rats). Injection of CP resulted
in an increase in NO2/NO3 and MetHb levels in
plasma. Treatment with NIL reduced NO2/NO3 and
MetHb plasma concentrations in the experimental
group in a dose-dependent fashion. NO2/NO3 levels
in CP rats receiving 2.2 mM NIL had returned to
values of control rats. NIL had no effect on NO2/NO3
and MetHb plasma concentrations in control rats. After CP injection 1-AGP levels were increased. Administration of NIL did not have any effect on the increased 1-AGP levels of the CP group. There was no
difference in expression of nicotinic acetylcholine receptors in the gastrocnemic muscle between groups
(Table 1).
The slopes of the dose-response curves did not differ between groups. The ED50 of all CP groups were
significantly larger compared with control and placebo rats (Table 2). Likewise the atracurium plasma
concentrations to maintain a 50% neuromuscular
block were also increased in the CP groups. Consistent
with these findings, the recovery index was shortened
in all groups with a larger requirement for atracurium
(Table 2). Treatment with 0.22 mM NIL shifted the
atracurium dose-response curve leftwards in both the
CP and the control group (Table 2).

Discussion
In this study, the inflammation-induced increase in
1-AGP levels could not be reversed with the iNOS

ANESTH ANALG
2005;101:13627

1365

ANESTHETIC PHARMACOLOGY
FINK ET AL.
NITRIC OXIDE AND ATRACURIUM RESISTANCE

Table 1. Nitrite/Nitrate, MetHemoglobin, 1-acid glycoprotein Levels, and Acetylcholine Receptor Concentration in
Corynebacterium Parvum and Control Animals During Treatment With Various Concentrations of N-Iminolysine
NIL-concentration
0.0 mM
NO2/NO3 (mol/L)
Ctrl
29 (2335)
CP
1372 (7831961)*
MetHb (mg/dL)
Ctrl
0.5 (0.30.7)
CP
7.5 (3.911.1)*
1-AGP (mg/mL)
Ctrl
0.6 (0.40.8)
CP
7.4 (6.18.6)*
AChR (fmol/mg)
Ctrl
29 (2334)
CP
33 (3036)

NIL
infection

NIL
effect

31 (2835)
342 (147537)*

P 0.05

NS
P 0.05

0.9 (0.61.2)
2.3 (1.82.8)*

1.3 (1.01.6)
1.5 (0.62.3)

P 0.05

NS
P 0.05

0.4 (0.30.5)
6.2 (5.27.2)*

0.4 (0.30.5)
6.6 (6.36.9)*

0.5 (0.30.6)
6.7 (5.38.0)*

NS

NA
NA

27 (2035)
25 (1931)

24 (2029)
21 (2527)

24 (2326)
27 (2232)

NS

NA
NA

0.00022 mM

0.0022 mM

0.022 mM

0.22 mM

29 (1443)
1107 (3071907)*

29 (1444)
741 (4261056)*

36 (3240)
391 (117666)*

0.6 (0.21.0)
6.2 (2.59.9)*

0.7 (0.31.1)
2.7 (1.93.5)*

0.4 (0.30.5)
6.1 (5.17.2)*
23 (2126)
30 (2425)

Data are presented as mean 95% confidence interval. NO2/NO3 levels of 7 surviving rats receiving 2.2 mM NIL were 30 (26 33) mol/L in the Control group
and 35 (31;38) in the CP group.
NO2/NO3 nitrite/nitrate; MetHb MetHemoglobin; 1-AGP 1-acid glycoprotein; AChR acetylcholine receptor; Ctrl control; CP corynebacterium parvum; NIL N-Iminolysine.
* P 0.05 compared with ctrl; NS not significant; NA not applicable because NIL and NIL infection were NS once NIL was significant its effect was
posthoc compared between all groups CP or between all groups Ctrl.

Table 2. Atracurium Pharmacodynamics

NIL-Concentration
0.0 mM

0.00022 mM

0.0022 mM

Atracurium ED50 (mg/kg)

Ctrl
0.90 (0.820.99)
0.92 (0.721.11)
0.84 (0.780.91)
CP
1.19 (1.081.32)* 1.25 (1.121.38)* 1.17 (1.031.30)*
Recovery Interval (s)
Ctrl
103 (94116)
115 (91138)
107 (95119)
CP
82 (7993)*
70 (6179)*
84 (7198)*
Infusion rate for constant 50% neuromuscular block (g kg min)
Ctrl
145 (124165)
137 (111164)
126 (121132)
CP
234 (186282)*
197 (166228)*
200 (159241)*
Atracurium plasma concentrations (g/mL)
Ctrl
4.5 (4.04.9)
4.4 (4.24.7)
4.4 (4.24.7)
CP
7.1 (6.18.0)*
8.3 (7.49.1)*
8.2 (7.39.0)*
Clearance (mL/min)
Ctrl
34 (2641)
31 (2438)
29 (2631)
CP
33 (3036)
24 (2127)
25 (2128)

NIL
infection

NIL
effect

0.70 (0.600.81)
0.98 (0.821.12)*

NS

NA
NA

112 (89135)
87 (7896)*

127 (108146)
79 (6692)*

NS

NA
NA

112 (94131)
206 (182231)*

126 (103148)
183 (150221)*

NS

NA
NA

4.2 (3.64.8)
8.9 (6.31.5)*

4.0 (3.44.7)
8.3 (7.29.4)*

NS

NA
NA

27 (2431)
25 (2228)

32 (2637)
24 (2028)

NS

NA
NA

0.022 mM

0.22 mM

0.78 (0.670.90)
1.19 (1.031.36)*

Data are presented as mean (95% confidence interval).

* P 0.05 compared with control; NS not significant; NA not applicable since N-Iminolysine (NIL) and NIL infection were NS; once NIL was significant
its effect was posthoc compared between all groups corynebacterium parvum (CP) or among all groups Ctrl. Ctrl control; NIL N-Iminolysine; ED50
effective dose to achieve 50% neuromuscular block).

inhibitor NIL despite dose-dependent attenuation of

the release of NO. However, in nonlethal doses NIL
was not able to suppress increased NO levels in animals with systemic inflammation to the levels of control animals. The increase in 1-AGP resulted in resistance to the neuromuscular blocking actions of
atracurium because of increased binding of atracurium to 1-AGP.
Resistance to nondepolarizing neuromuscular
blocking drugs can be seen in patients with systemic
inflammation or sepsis. In a previous study in rats, we
demonstrated that this resistance was not a result of
altered target-organ, i.e., acetylcholine receptor, sensitivity but was attributed to an increase in drug binding to 1-AGP (4). The present study confirms these

results. Atracurium resistance was evidenced as increased effective doses, increased plasma concentration requirement to achieve a constant neuromuscular
block, and decreased recovery times from complete
neuromuscular block. The increased 1-AGP levels
persisted in all CP groups receiving NIL, despite attenuation of inflammation, evidenced as decreasing NO
levels. Accordingly, the resistance to atracurium persisted in the all CP groups, demonstrated by a rightward
shift of the dose-response curves compared with control
animals. Interestingly, however, both CP and control
groups receiving 0.22 mM NIL had a shift of their doseresponse curve to the left again. Practically, the rightward shift (i.e., resistance to atracurium) of the doseresponse curve induced by CP could be counteracted by

1366

ANESTHETIC PHARMACOLOGY FINK ET AL.

NITRIC OXIDE AND ATRACURIUM RESISTANCE

NIL. Because, however, 1-AGP levels and atracurium

plasma concentrations remained increased, a direct inhibitory effect of NIL on neuromuscular transmission is
very likely. NIL probably displays the same direct open
or closed acetylcholine receptor channel block as the
other NOS inhibitors L-NG-monomethyl arginine and
L-nitroargine methylester (9).
The injection of CP induces a granulomatous liver
inflammation with excessive production of NO. This
endogenous NO reduces hepatic cytochrome p450 activity that can be restored by inhibiting iNOS activity
(10). Atracurium, however, is a quaternary ammonium compound degraded primarily via Hoffmann
elimination and is therefore independent of hepatic
metabolism (11). Uniform distribution and degradation of atracurium were attempted by maintaining
hemodynamic variables, acid-base balance, and temperature constant, resulting in comparable plasma
clearance among groups. We are therefore confident
that the observed differences in the effective doses and
plasma concentrations of atracurium are unrelated to
altered metabolism of atracurium.
NIL is a specific iNOS inhibitor. However, in animals
receiving the larger doses of NIL, an effect on endothelial
NO synthase (eNOS) and neural NO synthase (nNOS)
activity cannot be excluded. A decrease in muscle perfusion on the microcirculatory level through an eNOSinhibition mediated vasoconstriction might have affected our results. Furthermore, NO decreases
submaximal force by modulating excitation-contraction
coupling and therefore promotes relaxation of the skeletal muscle through the cyclic guanosine monophosphate (cGMP) pathway (12). Moreover, NO takes part in
regulating acetylcholinesterase expression in the synaptic cleft and ion channel properties by activating cGMP
(13). Finally, NO donors inhibit the evoked release of
acetylcholine reducing the end-plate potential (14).
Overall, it is impossible to speculate to what extent the
various eNOS or nNOS effects on synaptic processes
might have been suppressed. The NO levels remained
increased NIL groups up to 0.22 M. All these effects of
increased NO levels in our model, however, would have
made the muscle more susceptible rather than resistant
to muscle relaxants, i.e., leading to a leftward shift of the
dose-response curve.
Conflicting results have been published regarding
possible links between iNOS induction and production of 1-AGP. Cultured and activated liver cells increasingly produce 1-AGP during conditions of iNOS
pathway (15). In contrast, iNOS expression in hepatocytes in vitro after lipopolysaccharide stimulation is
not part of the acute-phase response and is differentially regulated from 1-AGP (16). In our in vivo experiments the specific iNOS inhibitor NIL was able to
dose-dependently suppress NO production without

ANESTH ANALG
2005;101:13627

affecting increased 1-acid glycoprotein levels. Although we did not measure iNOS activity, the approximately normalized NO metabolites NO2/NO3 prove
the efficacy of the drug administration. NO2/NO3 levels in the 2.2 M CP group were restored to normal
again. Because of the frequent lethality, however, the
rats were excluded from analysis. Reasons for the
deaths in these rats were not examined, although it
can be speculated that they were related to hypertensive vasoconstriction, especially in the pulmonary circuit. Nevertheless, our results therefore support the
notion of independent regulation of iNOS activity and
1-acid glycoprotein expression.
NO has a high affinity to hemoglobin, resulting in
the generation of MetHb. The effective suppression of
increased NO levels was also reflected in a normalization of pathologically increased MetHb levels. In the
CP groups receiving placebo or small-dose NIL treatment, however, MetHb levels of more than 5% posed
the possibility of oxidative stress. Normal Pao2 and
pH ranges and the absence of any increased lactate
levels, however, make the presence of hypoxia and an
effect on our results unlikely.
In summary, we conclude that targeting increased NO
levels to reverse altered pharmacodynamics of drugs
binding to 1-AGP, e.g., atracurium, does not seem effective. The only practically feasible way to deal with
increased plasma protein binding is to administer more
drug, in our case, more atracurium.

References
1. Henderson JL, Statman R, Cunningham JN, et al. The effects
of nitric oxide inhibition on regional hemodynamics during
hyperdynamic endotoxemia. Arch Surg 1994;129:1271 4.
2. Iskit AB, Guc O. Effects of endothelin and nitric oxide on organ
injury, mesenteric ischemia, and survival in experimental models of septic shock. Acta Pharmacol Sin 2003;24:9537.
3. Veihelmann A, Brill T, Blobner M, et al. Inhibition of nitric oxide
synthases improves detoxication in inflammatory liver dysfunction in vivo. Am J Physiol 1997;273:G530 6.
4. Fink H, Luppa P, Mayer B, et al. Systemic inflammation leads to
resistance to atracurium without increasing membrane expression of acetylcholine receptors. Anesthesiology 2003;98:82 8.
5. Blobner M, Kochs E, Fink H, et al. Pharmacokinetics and pharmacodynamics of vecuronium in rats with systemic inflammatory response syndrome: treatment with N(G)-monomethylL-arginine. Anesthesiology 1999;91:999 1005.
6. Archer S. Measurement of nitric oxide in biological models.
FASEB J 1993;7:349 60.
7. Metzger J, Blobner M, Luppa PB. Sensitive chemiluminescence
immunoassay for the determination of rat serum alpha1-acid
glycoprotein. Clin Chem Lab Med 2001;39:514 8.
8. Ibebunjo C, Martyn JA. Fiber atrophy, but not changes in acetylcholine receptor expression, contributes to the muscle dysfunction after immobilization. Crit Care Med 1999;27:275 85.
9. Scheller M, Blobner M, Von Loewenich C, et al. The NO synthase inhibitors L-Name and L-NMMA, but not L-arginine,
block the mammalian nicotinic acetylcholine receptor channel.
Toxicol Lett 1998;100 101:109 13.
10. Blobner M, Stadler J, Brill T, et al. Pharmacodynamics of atracurium and vecuronium in rats with inflammatory liver dysfunction - treatment with N-monomethyl-L-arginine. Anesthesiology 1996;85(Suppl):A811.

ANESTH ANALG
2005;101:13627

11. Merrett RA, Thompson CW, Webb FW. In vitro degradation of

atracurium in human plasma. Br J Anaesth 1983;55:61 6.
12. Kobzik L, Reid MB, Bredt DS, Stamler JS. Nitric oxide in skeletal
muscle. Nature 1994;372:546 8.
13. Ribera J, Marsal J, Casanovas A, et al. Nitric oxide synthase in
rat neuromuscular junctions and in nerve terminals of Torpedo
electric organ: its role as regulator of acetylcholine release.
J Neurosci Res 1998;51:90 102.
14. Kusner LL, Kaminski HJ. Nitric oxide synthase is concentrated
at the skeletal muscle endplate. Brain Res 1996;730:238 42.

ANESTHETIC PHARMACOLOGY
FINK ET AL.
NITRIC OXIDE AND ATRACURIUM RESISTANCE

1367

15. Itoh Y, Okanoue T, Enjo F, et al. Regulation of hepatocyte

albumin and alpha 1-acid glycoprotein secretion by monokines,
dexamethasone, and nitric oxide synthase pathway: significance
of activated liver nonparenchymal cells. Dig Dis Sci 1994;39:
851 60.
16. Geller DA, Freeswick PD, Nguyen D, et al. Differential induction of nitric oxide synthase in hepatocytes during endotoxemia
and the acute-phase response. Arch Surg 1994;129:16571.

Doxapram Only Slightly Reduces the Shivering Threshold in

Healthy Volunteers
Ryu Komatsu, MD*, Papiya Sengupta, MD*, Grigory Cherynak, MD,
Anupama Wadhwa, MD*, Daniel I. Sessler, MD*, Jin Liu, MD, Harrell E. Hurst,
Rainer Lenhardt, MD*

PhD,

and

*Outcomes Research Institute, University of Louisville; and Departments of Anesthesiology & Perioperative Medicine,
and Pharmacology and Toxicology, University of Louisville, Louisville, Kentucky

We determined the effects of doxapram on the major

autonomic thermoregulatory responses in humans.
Nine healthy volunteers were studied on 2 days: control and doxapram (IV infusion to a plasma concentration of 2.4 0.8, 2.5 0.9, and 2.6 1.1 g/mL at
the sweating, vasoconstriction, and shivering thresholds, respectively). Each day, skin and core temperatures were increased to provoke sweating, then reduced to elicit peripheral vasoconstriction and
shivering. We determined the sweating, vasoconstriction, and shivering thresholds with compensation for changes in skin temperature. Data were analyzed with paired t-tests and presented as mean sd;
P 0.05 was considered statistically significant.

verwhelming evidence in animals indicates that

even mild hypothermia provides substantial protection against cerebral (1) and myocardial ischemia (2). Mild hypothermia has been shown to improve
outcome after cardiac arrest in humans (3). Many of
these studies target core temperatures between 33 and
34C.
Induction of therapeutic hypothermia in patients having acute myocardial infarction or stroke may be compromised because tiny reductions in core temperature
trigger aggressive thermoregulatory defenses (4). The
Supported by National Institutes of Health Grant GM 061655
(Bethesda, MD), the Gheens Foundation (Louisville, KY), the Joseph
Drown Foundation (Los Angeles, CA), and the Commonwealth of
Kentucky Research Challenge Trust Fund (Louisville, KY).
Mallinckrodt Anesthesiology Products, Inc. (St. Louis, MO) donated
the thermocouples we used.
GCs current address is Department of Anesthesiology, Oklahoma University Health Sciences Center.
Accepted for publication April 25, 2005.
Address correspondence and reprint requests to Daniel I. Sessler,
MD, Outcomes Research Institute, 501 East Broadway, Suite 210,
Louisville, KY 40202. Address e-mail to sessler@louisville.edu. On
the worldwide web: www.or.org.
DOI: 10.1213/01.ANE.0000180198.13467.DF

1368

Anesth Analg 2005;101:136873

Doxapram did not change the sweating (control:

37.5 0.4C, doxapram: 37.3 0.4C; P 0.290) or
the vasoconstriction threshold (36.8 0.7C versus
36.4 0.5C; P 0.110). However, it significantly
reduced the shivering threshold from 36.2 0.5C to
35.7 0.7C (P 0.012). No sedation or symptoms of
panic were observed on either study day. The observed reduction in the shivering threshold explains
the drugs efficacy for treatment of postoperative
shivering; however, a reduction of only 0.5C is unlikely to markedly facilitate induction of therapeutic
hypothermia as a sole drug.
(Anesth Analg 2005;101:1368 73)

major autonomic thermoregulatory responses in humans are sweating, vasoconstriction, and shivering.
Each response is characterized by a threshold, which is
defined as the triggering core temperature.
Dopamine is among the most important thermoregulatory neurotransmitters, and it is well established
that increasing preoptic dopamine concentrations provokes hypothermia in mammals (5). Although doxapram stimulates release of dopamine from carotid
bodies in rats (6), it has central effects (7) that are
probably, at least in part, similarly mediated. As might
thus be expected, doxapram is a useful treatment for
postanesthetic shivering (8). Doxapram produces a
dose-dependent and substantial reduction in the shivering threshold in rabbits (9). The magnitude of this
inhibition, if similar in humans, would be clinically
important. Equally important is whether doxapram
reduces the shivering threshold from its normal value
near 35.5C (10) to approximately 34C, which may
provide protection against cerebral or myocardial ischemia. We thus tested the hypotheses that doxapram
comparably reduces the sweating, vasoconstriction,
and shivering thresholdsand that the reduction is
clinically important (i.e., approximately 1C).
2005 by the International Anesthesia Research Society
0003-2999/05

ANESTH ANALG
2005;101:1368 73

Methods
With approval of the Human Studies Committee at
the University of Louisville and written informed consent, we studied nine healthy volunteers (five men
and four women). None was obese, taking medication,
or had a history of thyroid disease, dysautonomia, or
Raynauds syndrome.
The volunteers fasted 8 h before each study day.
They dressed minimally and rested supine on a standard operating room table. Ambient temperature was
maintained near 21C. Each volunteer was studied on
two randomly assigned days: 1) controlno drug,
and 2) doxapram hydrochloride (A. H. Robins, Inc.) at
a target plasma concentration of 4 g/mL.
Doxapram was infused using a modification of the
simplified infusion regimen of Clements et al. (11),
which was stated to produce a constant plasma concentration of 2 g/mL. In this case with an objective of 4
g/mL, the infusion rates were doubled under the
assumption of linear pharmacokinetics. Specifically,
doxapram was infused at a rate of 6 mg min1 70 kg1
for the first 15 min; at a rate of 4 mg min1 70 kg1 for
the next 15 min; at a rate of 3 mg min1 70 kg1 for the
next 30 min; at a rate of 2 mg min1 70 kg1 for
the next hour, and subsequently, at a rate of
1 mg min1 70 kg1. The infusion was started 15 min
before starting thermal manipulation to allow establishment of steady-state plasma drug concentration; the infusion was continued until shivering was detected.
Core temperature was recorded from the tympanic
membrane using Mon-a-therm thermocouples
(Mallinckrodt Anesthesiology Products, Inc., St. Louis,
MO). Mean skin-surface temperature and cutaneous
heat transfer were calculated from measurements at 15
area-weighted sites. Temperatures were recorded at
1-min intervals from thermocouples connected to calibrated Iso-Thermex thermometers having an accuracy of 0.1C and a precision of 0.01C (Columbus
Instruments, Corp., Columbus, OH).
Sweating was continuously quantified on the left
upper chest using a ventilated capsule (4). We considered a sweating rate 40 g m2 h1 for at least
5 min to be significant (12). Absolute right middle
fingertip bloodflow was quantified using venousocclusion volume plethysmography at 5-min intervals
(13). The vasoconstriction threshold was determined
post hoc by an observer blinded to treatment and core
temperature.
As in previous similar studies (12), we used systemic oxygen consumption to quantify shivering. A
DeltaTrac metabolic monitor (SensorMedics Corp.,
Yorba Linda, CA) was used in canopy mode. Initiation
of shivering threshold was determined post hoc by an
observer blinded to treatment and core temperature.
End-tidal Pco2 was sampled from a catheter inserted

ANESTHETIC PHARMACOLOGY
KOMATSU ET AL.
DOXAPRAM AND THE SHIVERING THRESHOLD

1369

into one nostril; gas removed from the catheter was

returned to the canopy of the metabolic monitor.
Heart rate, end-tidal Pco2, and oxyhemoglobin saturation (Spo2) were measured continuously using
pulse oximetry, and arterial blood pressure was determined oscillometrically at 5-min intervals at the left
ankle.
Sedation was evaluated by using the responsiveness
component of the Observers Assessment of
Alertness/Sedation (OAA/S) score (Table 1) (14) at
several times: before starting drug administration, before thermal manipulation started, and at each threshold. An investigator blinded to core temperature and
treatment evaluated sedation. Blood for doxapram
analysis was sampled from the central catheter before
drug administration (blank sample) and at each
threshold. Blood was centrifuged, and the plasma was
removed and stored at 40C for up to 2 mo. The
plasma samples were then transferred to another
freezer and stored for up to 1 wk until analysis. Doxapram and its metabolites are stable when stored
20C.
Doxapram concentrations were determined in triplicate by gas chromatography/mass spectrometry using the technique of selected ion monitoring (GC/MSSIM). Reference doxapram hydrochloride was
obtained from USP (U. S. Pharmacopeia, Rockville,
MD). The method used solid phase extraction columns
(Strata-X; Phenomenex, Torrance, CA) to extract doxapram from plasma along with internal standard (diazepam) to compensate the drug loss during extraction and GC/MS processes. Doxapram and diazepam
were separated by gas chromatography according to
their different GC retention times (7.0, 5.4 min, respectively) in a capillary column (DB5; Agilent, Palo Alto,
CA) that was temperature programmed from 150 to
290C at 32/min. The drugs were then detected by
MS using SIM of characteristic ions (doxapram m/z
100, 378; diazepam m/z 256, 283) after electron ionization. Quantification used standard responses measured by ion peak area ratios versus amounts of
analyte/internal standard. Responses were linear
across the doxapram plasma concentration range of
0.255 g/mL. Limits of detection and quantification
were 0.25 g/mL. Sample replicates were assayed
on different occasions. The intra-assay coefficient of
variation (CV) was 8.8%, whereas inter-assay CV was
21%.
Nausea and vomiting incidence and severity were
recorded by blinded observers before commencement
of drug administration, before thermal manipulation
started, and at each threshold during each study. Nausea and vomiting incidence and severity were reported by the volunteer using a 4-point scale: 1
none, 2 mild, 3 moderate, or 4 severe.
Because doxapram is a potent drug to cause panic
(15), we scored the severity of panic attack symptoms

1370

ANESTHETIC PHARMACOLOGY KOMATSU ET AL.

DOXAPRAM AND THE SHIVERING THRESHOLD

ANESTH ANALG
2005;101:1368 73

Table 2. Panic Symptoms

Table 1. Responsive Component of the Observers

Assessment of Alertness/Sedation Scale
Score

Responsiveness

Responds readily to name spoken in

normal tone
Lethargic response to name spoken in
normal tone
Responds only after name is spoken loudly
and/or repeatedly
Responds only after mild prodding or
shaking
Does not respond to mild prodding or
shaking

4
3
2
1

Symptom

1
2

Nausea or abdominal distress

Paresthesia (numbness or tingling
sensations)
Feeling of choking
Palpitations, pounding heart, or
accelerated heart rate
Chest pain or discomfort
Derealization (feelings of unreality) or
depersonalization (being detached
from oneself)
Feeling dizzy, unsteady, light-headed,
or faint

3
4
5
6
7

by presence or absence of seven selected items of

Diagnostic and Statistical Manual of Mental Disorders
(DSM)-IV-TR diagnostic criteria for panic attack (Table 2) (16). These were assessed at the same times as
the OAA/S score. The total number of the symptoms
present at each time was used for statistical analysis.
The cutaneous contribution to the thermoregulatory
responsessweating (17), vasoconstriction, and shiveringis linear (18). We thus used measured skin and
core temperatures in degrees Celsius at each threshold
to calculate the core-temperature threshold that would
have been observed had skin been maintained at a
single designated temperature. For this purpose, we
used Equation 1, which corrects core temperature for
cutaneous temperature, providing response thresholds that would have been observed at a designated
core temperature:

Tcore(calculated) Tcore

Score

Tskin Tskin(designated)
(1)

We have previously described the derivation and

validation of this equation (12). We used a of 0.1 for
sweating (17) and a of 0.2 for vasoconstriction and
shivering (18). The designated skin temperature was
set at 34C, a typical intraoperative value.
We determined that a sample size of 9 would detect
a 0.6C difference in shivering thresholds between
control and drug days with 89% power. Based on a
similar study conducted with the same design in our
laboratory with another drug (unpublished observation), we assumed that the difference in shivering
thresholds on the 2 study days would have a standard
deviation of 0.495 and correlation of 0.55.
Oxygen consumption and calculated respiratory
quotient (RQ) were averaged during baseline (before
commencement of drug infusion), during the cooling
phase before shivering, and during shivering, respectively. The averaged values were then compared with
two-way ANOVA (two factors; drug, period) and post
hoc Bonferroni/Dunn tests.

After confirming that sedation, nausea severity, and

panic symptoms were similar at baseline on each
study day, the difference in the scores at each measurement time (baseline, prewarming, sweating threshold, vasoconstriction threshold, and shivering threshold) between the drug day and the control day within
each volunteer were compared with Kruskal-Wallis
tests. The incidence of nausea and combined nausea
and vomiting was compared by 2 analyses. For the
combined nausea and vomiting analysis, any score for
nausea or vomiting exceeding none was considered
positive.
Ambient temperature, humidity, mean arterial
blood pressure, heart rate, Spo2, and end-tidal Pco2 on
each study day were averaged within each volunteer
across the warming and cooling periods; the resulting
values were then averaged among volunteers. Results
for each study day were compared using paired
t-tests. Plasma concentrations of doxapram at each
threshold were compared using one-way ANOVA. All
results are presented as means sd or range, as
appropriate; P 0.05 was considered statistically
significant.

Results
Five of the participants were men; four were women.
They were 26 5 yr old, weighed 72 13 kg, and
were 175 10 cm tall. Of the possible confounding
factors that might influence thermoregulatory thresholds, ambient temperature, humidity, and heart rate
were similar on each of the study days during the
warming and cooling periods. In addition, end-tidal
Pco2 was significantly reduced, whereas Spo2 and
mean arterial blood pressure were both increased significantly by doxapram during the warming and cooling periods (Table 3).
Oxygen consumption was not significantly influenced by doxapram (P 0.293), and it was similar
between baseline and cooling phases. But it was significantly increased during shivering compared with

ANESTH ANALG
2005;101:1368 73

ANESTHETIC PHARMACOLOGY
KOMATSU ET AL.
DOXAPRAM AND THE SHIVERING THRESHOLD

1371

Table 3. Potential Confounding Factors

Mean arterial blood pressure (mm Hg)

Heart rate (bpm)
End-tidal Pco2 (mm Hg)
Spo2 (%)
Ambient temperature (C)
Relative humidity (%)

Control day

Doxapram day

P value

89 10
71 7
38 2
97 1
23.0 1.4
28.1 11.4

102 6
76 7
36 2
98 1
23.4 2.2
21.8 9.4

0.005
0.065
0.023
0.001
0.400
0.270

Values were recorded at 5-min intervals and averaged over the sweating-to-shivering period. Data are presented as mean sd; results were compared with
paired t-tests.

both baseline and the cooling phase (P 0.001). RQ

was not significantly influenced by doxapram (P
0.195), nor by the periods of the study (P 0.458)
(Table 4).
None of the study participants was sedated at baseline on either study day (OAA/S score of 5 for all
volunteers), and no sedation was observed during the
study. None of the volunteers was nauseated at the
start of the study (nausea severity score of 1) on either
study day. Additionally, as with sedation, no significant nausea was observed during the study. None of
the volunteers experienced panic at baseline on either
study day (panic symptom score total of 0). Panic
symptoms during doxapram infusion were not significantly different from those of the control day (P
0.06).
On the doxapram day, plasma concentrations of
doxapram were similar at the sweating, vasoconstriction, and shivering thresholds with plasma concentrations of 2.4 0.8 [1.4 3.5], 2.5 0.9 [1.4 3.6], and 2.6
1.1 [1.3 4.8] g/mL; (mean sd [range]) for each
threshold, respectively (P 0.930). The sweating
thresholds were similar on the doxapram (37.3
0.4C) and control (37.5 0.4C) days (P 0.290;
Table 5). The vasoconstriction thresholds were also
similar on the doxapram (36.4 0.5C; Table 5) and on
the control (36.8 0.7C) day (P 0.110). In contrast,
doxapram reduced the shivering threshold by 0.5
0.4C from 36.2 0.5C on the control day to 35.7
0.7C on the doxapram day (P 0.012; Fig. 1).

Discussion
Although doxapram significantly reduced the shivering threshold, the reduction was only 0.5C. In comparison, clonidine, at a dose of 75 g, decreases the
shivering threshold to the same extent as doxapram at
the dose used in our study (19) and tramadol, at a dose
of 250 g, decreases the shivering threshold by 0.9C,
which is slightly more than the change we obtained
with doxapram (20). Both of those drugs have proven
effective treatments for postoperative shivering. However, a reduction in the shivering threshold of this
extent is insufficient to facilitate induction of therapeutic hypothermia. We thus conclude that doxapram
as a sole drug will not serve for this purpose.

Table 4. Oxygen Consumption and Respiratory Quotient

on the Control and Doxapram Study Days
Control day Doxapram day
Oxygen consumption
(mL/min)
Baseline period
Cooling period
Shivering period
Respiratory quotient
Baseline period
Cooling period
Shivering period

217 43
226 58
315 89

210 16
250 53
351 51

0.84 0.05
0.84 0.08
0.82 0.05

0.80 0.05
0.82 0.04
0.79 0.03

Values were recorded at 1-min intervals and averaged in each period. Data
are presented as mean sd. There were no statistically significant differences
between the study days for any period.

Doxapram increased Spo2 (98% 1% versus 97%

1%) and reduced end-tidal Pco2 (36 2 versus 38
2 mm Hg), presumably as a consequence of doxaprams well known ability to improve tidal volume
and augment respiratory rate (21). The increased
mean arterial blood pressure on the doxapram day
(102 6 versus 89 10 mm Hg) doubtless resulted
from the established pressor effect of doxapram (21).
Despite being statistically significant, these small respiratory and hemodynamic effects, which were not
accompanied by increases in metabolic rates, were of
little clinical consequence and unlikely to cause adverse effects even in victims of stroke or myocardial
infarction.
A target plasma concentration of 4 g/mL was
chosen, because it seems to be the maximal therapeutic concentration that does not evoke serious side effects. A slightly larger plasma concentration of 5
g/mL increases the risk of side effects (22). Our
measured plasma concentration was within the range
(1.53.0 g/mL) that effectively increases the minute
ventilation (23).
Although we used a published infusion protocol,
the average plasma concentration at each threshold
was considerably less than the targeted concentration
of 4 g/mL in all but one patient. We are unable to
determine in the present study whether the differences between target and actual plasma levels resulted
from differences in drug disposition, clearance, or
nonlinear pharmacokinetic effects caused by dosage

1372

ANESTHETIC PHARMACOLOGY KOMATSU ET AL.

DOXAPRAM AND THE SHIVERING THRESHOLD

ANESTH ANALG
2005;101:1368 73

Table 5. Mean Skin Temperatures, Core Temperatures, and Calculated Thermoregulatory Thresholds

Sweating
Mean skin (C)
Core (C)
Threshold (C)
Doxapram concentration (g/mL)
Vasoconstriction
Mean skin (C)
Core (C)
Threshold (C)
Doxapram concentration (g/mL)
Shivering
Mean skin (C)
Core (C)
Threshold (C)
Doxapram concentration (g/mL)

Control day

Doxapram day

36.6 0.3
37.2 0.3
37.5 0.4

36.4 1.0
37.0 0.2
37.3 0.4
2.4 0.8 [1.393.49]

32.9 2.4
37.1 0.3
36.8 0.7

32.2 1.5
36.8 0.3
36.4 0.5
2.5 0.9 [1.363.61]

30.0 1.5
37.2 0.3
36.2 0.5

29.2 1.3
36.9 0.4
35.7 0.7
2.6 1.1 [1.344.82]

Thresholds were calculated based on a designated mean skin temperature of 34C. Results are presented as mean sd and range for plasma doxapram
concentrations. Only the thresholds were statistically compared, and only the shivering threshold differed significantly between the 2 treatment groups (P
0.012).

We thank Gilbert Haugh, MS, and Nancy Alsip, PhD, for their
statistical and editorial contributions, respectively (both from the
University of Louisville).

References

Figure 1. Shivering thresholds in nine healthy volunteers. The open

circles show the shivering threshold for each volunteer on the
control and doxapram days; the filled squares are the group means
(sd). The shivering threshold was 0.5C greater on the control day
than on the doxapram day, P 0.012.

scaling. Assuming a steady-state plasma concentration

was achieved before shivering occurred, the total
plasma clearance calculated from our data was 12.9
5 [range 5.8 21.5] mL kg1 min1, a value larger
than that reported by Clements et al. (11). However,
Jamali et al. (24) found that 4 of 17 neonates had much
larger clearance rates of doxapram (19.0
29.2 mL kg1 min1) than the others. They suggested that the disposition kinetics of doxapram has a
binominal distribution. If any of our participants were
rapid metabolizers of doxapram, their increased clearance rates could have profoundly reduced the actual
mean plasma concentration of the group.
In summary, doxapram at a measured plasma concentration of approximately 2.5 g/mL reduced the
shivering threshold significantly, but only by 0.5C.
This reduction explains the drugs efficacy for treatment of postoperative shivering. However, a reduction of only 0.5C is unlikely to markedly facilitate
induction of therapeutic hypothermia as a sole drug.

1. Popovic R, Liniger R, Bickler PE. Anesthetics and mild hypothermia similarly prevent hippocampal neuron death in an in
vitro model of cerebral ischemia. Anesthesiology 2000;92:
13439.
2. Dae MW, Gao DW, Sessler DI, et al. Effect of endovascular
cooling on myocardial temperature, infarct size, and cardiac
output in human-sized pigs. Am J Physiol Heart Circ Physiol
2002;282:H1584 91.
3. Bernard SA, Gray TW, Buist MD, et al. Treatment of comatose
survivors of out-of-hospital cardiac arrest with induced hypothermia. N Engl J Med 2002;346:557 63.
4. Lopez M, Sessler DI, Walter K, et al. Rate and gender dependence of the sweating, vasoconstriction, and shivering thresholds in humans. Anesthesiology 1994;80:780 8.
5. Ruwe WD, Myers RD. Dopamine in the hypothalamus of the
cat: pharmacological characterization and push-pull perfusion
analysis of sites mediating hypothermia. Pharmacol Biochem
Behav 1978;9:65 80.
6. Anderson-Beck R, Wilson L, Brazier S, et al. Doxapram stimulates dopamine release from the intact rat carotid body in vitro.
Neurosci Lett 1995;187:25 8.
7. Yanagida H, Ashizawa N, Wakushima Y, Yamamura H. Effects
of anesthetics on ponto-geniculo-occipital waves from the oculomotor nucleus of the cat. Anesthesiology 1975;42:574 83.
8. Singh P, Dimitriou V, Mahajan RP, Crossley AW. Double-blind
comparison between doxapram and pethidine in the treatment
of postanaesthetic shivering. Br J Anaesth 1993;71:685 8.
9. Okuyama K, Matsukawa T, Ozaki M, et al. Doxapram produces
a dose-dependent reduction in the shivering threshold in rabbits. Anesth Analg 2003;97:759 62.
10. Ikeda T, Kim J-S, Sessler DI, et al. Isoflurane alters shivering
patterns and reduces maximum shivering intensity. Anesthesiology 1998;88:866 73.
11. Clements JA, Robson RH, Prescott LF. The disposition of intravenous doxapram in man. Eur J Clin Pharmacol 1979;16:411 6.
12. Matsukawa T, Kurz A, Sessler DI, et al. Propofol linearly reduces the vasoconstriction and shivering thresholds. Anesthesiology 1995;82:1169 80.

ANESTH ANALG
2005;101:1368 73

13. Rubinstein EH, Sessler DI. Skin-surface temperature gradients

correlate with fingertip blood flow in humans. Anesthesiology
1990;73:5415.
14. Chernik DA, Gillings D, Laine H, et al. Validity and reliability of
the Observers Assessment of Alertness/Sedation Scale: study
with intravenous midazolam. J Clin Psychopharmacol 1990;10:
244 51.
15. Abelson JL, Nesse RM, Weg JG, Curtis GC. Respiratory psychophysiology and anxiety: cognitive intervention in the doxapram
model of panic. Psychosom Med 1996;58:30213.
16. American Psychiatric Association. DSM-IV: diagnostic and statistical manual of mental disorders. 4th ed. Washington, DC:
American Psychiatric Association; 2000.
17. Nadel ER, Horvath SM, Dawson CA, Tucker A. Sensitivity to
central and peripheral thermal stimulation in man. J Appl
Physiol 1970;29:6039.
18. Cheng C, Matsukawa T, Sessler DI, et al. Increasing mean skin
temperature linearly reduces the core-temperature thresholds
for vasoconstriction and shivering in humans. Anesthesiology
1995;82:1160 8.

ANESTHETIC PHARMACOLOGY
KOMATSU ET AL.
DOXAPRAM AND THE SHIVERING THRESHOLD

1373

19. Delaunay L, Bonnet F, Liu N, et al. Clonidine comparably decreases the thermoregulatory thresholds for vasoconstriction
and shivering in humans. Anesthesiology 1993;79:470 4.
20. De Witte JL, Kim J-S, Sessler DI, et al. Tramadol reduces the
sweating, vasoconstriction, and shivering thresholds. Anesth
Analg 1998;87:1739.
21. Winnie AP. Chemical respirogenesis: a comparative study. Acta
Anaesthesiol Scand Suppl 1973;51:132.
22. Hayakawa F, Hakamada S, Kuno K, et al. Doxapram in the
treatment of idiopathic apnea of prematurity: desirable dosage
and serum concentrations. J Pediatr 1986;109:138 40.
23. Calverley PM, Robson RH, Wraith PK, et al. The ventilatory
effects of doxapram in normal man. Clin Sci (Lond) 1983;65:
659.
24. Jamali F, Barrington KJ, Finer NN, et al. Doxapram dosage
regimen in apnea of prematurity based on pharmacokinetic
data. Dev Pharmacol Ther 1988;11:2537.

Clonidine Premedication in Patients with Sleep Apnea

Syndrome: A Randomized, Double-Blind,
Placebo-Controlled Study
Michael T. Pawlik, MD, DEAA*, Ernil Hansen, MD, PhD*, Daniela Waldhauser*,
Christoph Selig, MD, and Thomas S. Kuehnel, MD
Departments of *Anesthesiology and Otorhinolaryngology, Universitatsklinik Regensburg; and Department of
Anesthesiology, Universitatsklinik Ulm, Germany

Patients with sleep apnea often present with cardiac

diseases and breathing difficulties, with a high risk of
postoperative respiratory depression. We conducted a
randomized, double-blind, prospective study in 30 adult
patients with obstructive sleep apnea, undergoing elective ear-nose-throat surgery. The patients were randomly assigned to receive placebo or clonidine (2 g/kg
oral) the night before and the next morning 2 h before
surgery. Spo2, heart rate, mean arterial blood pressure,
snoring, and oronasal airflow were monitored for 36 h. A
standard anesthesia was used consisting of propofol and
remifentanil. Anesthetic drug consumption, postoperative analgesics, and pain score were recorded. In the
clonidine group, mean arterial blood pressures were significantly lower during induction, operation, and emergence from anesthesia. Both propofol dose required for
induction (190 32.2 mg) and anesthesia (6.3

bstructive sleep apnea syndrome (OSAS), which

is associated with obesity, hypoxemia, hypercapnia, and snoring, is characterized by upper
airway obstruction. Its prevalence is reported to be
1% 4% in the adult population with a familial predisposition (1). Although first described in patients with
an increased body mass index, OSAS also occurs in
patients with a normal body mass index.
Hypertension, arrhythmia, and coronary artery disease are frequently associated with OSAS and influence
the clinical course of the disease (2). Because of the
frequent incidence of an obstructed airway, a high risk of
respiratory depression, and the incidence of coexisting

This study was supported by the Departments of Anesthesiology

and Otorhinolaryngology, Universitat Regensburg, Germany.
Accepted for publication April 20, 2005.
Address correspondence to Michael T. Pawlik, MD, DEAA, Department of Anesthesiology, University of Regensburg, Franz-JosefStrauss-Allee 6, D-93046 Regensburg, Germany. Address e-mail to
michael.pawlik@klinik.uni-regensburg.de.
DOI: 10.1213/01.ANE.0000180194.30741.40

1374

Anesth Analg 2005;101:137480

1.3 mg kg1 h1) during surgery were significantly

reduced in the clonidine group compared with the placebo group (induction 218 32.4, anesthesia 7.70 1.5;
P 0.05). Piritramide consumption (7.4 5.1 versus
14.2 8.5 mg; P 0.05) and analgesia scores were
significantly reduced in the clonidine group. Apnea and
desaturation index were not different between the
groups, whereas the minimal postoperative oxygen saturation on the day of surgery was significantly lower in
the placebo than in the clonidine group (76.7% 8.0%
versus 82.4% 5.8%; P 0.05). We conclude that oral
clonidine premedication stabilizes hemodynamic variables during induction, maintenance, and emergence
from anesthesia and reduces the amount of intraoperative anesthetics and postoperative opioids without deterioration of ventilation.
(Anesth Analg 2005;101:1374 80)

cardiopulmonary disorders, OSAS patients are at increased risk during general anesthesia (3). Analgesics
and sedatives aggravate OSAS by decreasing pharyngeal tone, and depressing ventilatory responses to hypoxia and hypercapnia (4). For these reasons, drugs that
cause sedation and respiratory depression should be
avoided for premedication (5). Therefore, current opinion does not recommend premedication for sleep apnea
patients, although oral premedication is common practice in general anesthesia (6). The 2-agonist clonidine
possesses rapid eye movement (REM)-suppressant activity and improves the level of nocturnal hypoxemia in
patients with OSAS (7). Clonidine has been used for
some years in supplementing general anesthesia. It is
beneficial when used for preoperative medication before
general anesthesia by enhancing the effects of anesthetics
(8). Several studies have shown reduced changes in heart
rate and mean arterial blood pressure (MAP) in patients
receiving premedication with clonidine without having
adverse effects on respiratory rate (9). We hypothesize
that oral administration of 2 g/kg clonidine given the
2005 by the International Anesthesia Research Society
0003-2999/05

ANESTH ANALG
2005;101:1374 80

night before and 2 h before surgery reduces changes in

heart rate, MAP, and the requirement for perioperative
anesthetics and analgesics without adverse effects on the
apnea/hypopnea index (AHI) of OSAS patients.

Methods
After obtaining local ethics committee approval and written informed consent, we studied 30 ASA physical status II
and III patients with diagnosed OSAS, who were undergoing elective septoplasty alone or in combination with uvulopalatopharyngoplasty (UPPP) (Table 1). Patients with a
history of myocardial infarction within the last 6 mo, resting room air oxygen saturation 90%, or taking clonidine
to treat hypertension were excluded. None of the patients
had a preoperatively diagnosed myocardial ischemia. The
protocol started with an initial screening of medical history
and a clinical examination. No patient showed underlying
CO2 retention in the preoperative blood gas sample. Patients who successfully completed the screening were randomized, using a software program (Excel 2000; Microsoft
Inc., Redmond, WA), to receive either oral clonidine drops
(2 g/kg) or placebo (preparation: University of Regensburg pharmacy) the evening before and 2 h before induction of anesthesia. Patients received their usual antihypertensive medications the night before and the morning of
surgery. The sample size was calculated to detect a 50%
increase of the AHI in the study group with 0.05 and
0.95.
All patients underwent polysomnography (Eden
Trace; Edentec Monitoring System, Eden Prairie, MN)
from 08:00 pm to 08:00 am on the preoperative day,
and on the day of surgery until 08:00 am on the first
postoperative morning. Respiratory events were assessed by measuring oronasal airflow using a dual
thermistor. Chest wall movements were measured using
an inductive respiratory plethysmograph, oxyhemoglobin saturation was analyzed by finger oximetry in the
Edentec recorder, sound was recorded using a microphone placed at the level of the larynx, body position
was monitored through a sensor positioned on the chest,
and cardiac rhythm was monitored by electrocardiogram (ECG). The sleep variables used in this study were
AHI, desaturation index, and minimum oxygen saturation (MSAT). The AHI index was calculated as the number of apnea/hypopnea episodes/hour of sleep. An apnea episode was defined as a cessation of airflow lasting
10 s, whereas a hypopnea episode was defined as a
50% reduction in combined oral and nasal airflow
lasting 10 s. A desaturation event was defined as a
saturation 90% and a duration 10 s; the index was
calculated as the number of desaturation episodes/hour.
MSAT was defined as the lowest arterial oxygen saturation level detected during the observation period. An
oxygen desaturation 90% was considered abnormal.
To minimize the possibility of misinterpretation, signal

ANESTHETIC PHARMACOLOGY
PAWLIK ET AL.
CLONIDINE PREMEDICATION AND SLEEP APNEA

1375

strength and data were monitored off-line by a physician

trained in sleep medicine (TSK).
On arrival in the operating room, monitoring consisted of three-lead ECG, pulse oximetry, capnography,
heart rate, MAP (9000C; Siemens, Erlangen, Germany).
Bispectral index (BIS; Aspect Medical Systems Inc.,
Natick, MA) electrodes were applied and BIS A 2000
monitoring commenced. MAP and heart rate were recorded at intervals of 1 min during anesthetic induction
until tracheal intubation and after cessation of anesthetics at emergence from anesthesia. During surgery and
recovery, measurement was performed in 5-min intervals and hourly intervals on the ward for 8 h after
surgery on the first postoperative day. All patients had
an induction technique commencing with a bolus of
remifentanil 1 g/kg, followed by propofol 1.5 mg/kg
until loss of the eyelash reflex and a BIS target value of
40. Increased BIS values (40) were treated by repeated
bolus of 0.5 mg/kg propofol. Muscle relaxation and
tracheal intubation followed 2.5 min after IV mivacurium 0.2 mg/kg. Anesthesia was maintained during surgery with remifentanil infusion immediately commenced at 0.2 g kg1 min1 and increased to 0.5
g kg1 min1 to keep heart rate between 60 80 bpm.
A decrease of heart rate to 40 bpm was treated by
injection of 0.5 mg of atropine. Further propofol dosage
adjustments were standardized by protocol, according to
adverse hemodynamic responses and BIS measurements. In general, this consisted of 1. a bolus of
propofol and increased propofol rates if signs of light
anesthesia (systolic blood pressure 15% compared
with values before induction, or BIS 45) were present,
or 2. reduced propofol infusion rates if signs of deep
anesthesia (systolic blood pressure 15% compared
with values before induction or BIS 35) were present.
At the time of surgical dressing, all anesthetic infusions were stopped, and time until opening of eyes on
command was recorded. The total amount of intraoperative propofol and remifentanil was recorded. Tracheal extubation occurred when the patient was fully
awake and cooperative, with a respiratory rate between 10 and 20 breaths/min, and had satisfactory
pulse oximetry.
Postoperative analgesia was assessed by using a
visual analog scale (VAS) score. Standard analgesia
consisted of rectal diclofenac 1.5 mg/kg immediately
after induction of anesthesia. Piritramide 0.1 mg/kg
was administered IV at the patients request by nurses
or if the VAS score was 3. Patient pain scores were
obtained after tracheal extubation, at 10-min intervals
in the first 2 h after surgery, and then every hour for
8 h after completion of the procedure. Patients received oxygen during the first 2 h after tracheal extubation, if oxygen saturation was constantly 92%.
Because of nasal tamponades nCPAP (nasal continuous positive airway pressure), patients were not able
to use their nCPAP device postoperatively.

1376

ANESTHETIC PHARMACOLOGY PAWLIK ET AL.

CLONIDINE PREMEDICATION AND SLEEP APNEA

ANESTH ANALG
2005;101:1374 80

Table 1. Patient Demographic Characteristics of Each Group

n
Age, yr
Body mass index
Sex (male/female), n
ASA physical status (n)
Apnea index preoperative
Antihypertensive medication, n
Mean arterial blood pressure, mm Hg
Nasal CPAP therapy, n
Anesthesia time, min
Septoplasty alone
Septoplasty and UPPP
Septoplasty and tonsillectomy
UPPP and tonsillectomy

Placebo

Clonidine

15
53.8 7.8
34.3 6.1
15/0
II (4), III (11)
37.3/h 21.8
8/15
107 14.7
12/15
102 30
12
8
6
4

15
48.9 5.2
33.7 5.4
13/2
II (3), III (12)
45.2/h 24.0
10/15
105 12.1
10/15
114 36
14
10
3
3

Data are presented as absolute number, median and range, or mean sd as appropriate. There were no statistical differences between the two groups.
UPPP uvulopalatopharyngoplasty, CPAP continuous positive airway pressure.

Statistical analysis was performed using the SPSS

12.0 software package (SPSS Inc., Chicago, IL). The
Students t-test was used to compare normally distributed continuous variables between the two groups,
and the nonparametric Mann-Whitney U-test was
used for nonparametric or not normally distributed
data. Hemodynamic values were analyzed by comparing corresponding time points during induction and
emergence of anesthesia between the groups. Hemodynamic data during surgery, recovery room, and on
the postoperative ward, as well as BIS values and VAS
pain scores were added, averaged, and analyzed using
Students t-test. A paired t-test with Bonferroni adjustment for multiple comparisons was used for possible
changes of AHI, desaturation index, and MSAT within
the groups. Data are presented as means and their
standard deviations (sd). Statistical significance was
considered at P 0.05.

Results
Thirty patients were enrolled in the study, 15 patients
in each group. One patient of the placebo group had to
be removed from the hemodynamic analysis because
of persistent hypotension caused by an inadvertent
overdose of angiotensin-converting enzyme (ACE)inhibitors on the ward. Oronasal flow sensor recording was impaired in six patients because of minor
postoperative bleeding and sweating, so that AHI
could not be analyzed. Evaluation of the data was
performed the following day by a single investigator
trained in the field of somnology to ensure consistency. The investigator was not aware of the subject
randomization. The two groups were well balanced
for all demographic and baseline data (Table 1).
On the morning of surgery, the MAP before induction
of anesthesia was significantly lower in the clonidine
group than in the placebo group, whereas the heart rates

showed no difference between groups (Table 2). Only

during the first 3 min of induction were MAP values
significantly lower in the clonidine group, whereas there
was no difference in heart rate between the groups (Fig.
1A and B). MAP and heart rates were significantly lower
during the entire surgery in the clonidine-premedicated
patients than in placebo (P 0.001, t-test; Table 2).
During emergence from anesthesia, MAPs, but not heart
rates, were significantly lower in the clonidine group
(Fig. 2A and B). After emergence, differences in MAP
were significant (P 0.001). Additionally, we found
significant differences of MAP and heart rates (P
0.001) in the recovery room; significant differences in
heart rate and MAP were also observed between the
groups on the ward (Table 2).
The induction dose of propofol required to reach a
target BIS value of 40 was significantly smaller in
clonidine-premedicated patients (Table 3). Patients receiving clonidine required a significantly smaller average dosage of propofol during the entire surgery than
the placebo group (Table 3). The consumption of
remifentanil showed no significant differences between
the groups (Table 3). There was no difference in the
administered basic analgesic regimen of diclofenac (118
53.8 mg placebo versus 110 68.6 mg clonidine) and
the consumption of piritramide in the first postoperative
hour in the recovery room, but clonidine-premedicated
patients required significantly less piritramide in the 24
postoperative hours than placebo patients (Table 3).
Time until eye opening on command after cessation of
total IV anesthesia (TIVA) was significantly shorter in
the clonidine group (Table 3). The overall quality of
analgesia was excellent in both groups throughout the
observation according to psychometric evaluation (VAS
3). Mean VAS was significantly lower in the clonidine
group after tracheal extubation in the recovery room and
significantly lower in the clonidine group than in the
placebo group after 24 h (P 0.001; Table 3).

ANESTH ANALG
2005;101:1374 80

ANESTHETIC PHARMACOLOGY
PAWLIK ET AL.
CLONIDINE PREMEDICATION AND SLEEP APNEA

1377

Table 2. Perioperative Hemodynamic Variables and Requirement of Additional Antihypertensive Medication

Heart rate before induction (bpm)

Mean arterial blood pressure before induction (mm Hg)
Heart rate, during operation (bpm)
Mean arterial blood pressure during operation (mm Hg)
Heart rate, 9 h after recovery room (bpm)
Mean arterial blood pressure, 9 h after recovery room (mm Hg)
No. of additional antihypertensive drug requirements during observation
period/total patients

Placebo

Clonidine

74 16.6
115 18.5
62 13
83 14
83 13
101 16
8/15

67 13.9
92 11.3
57 11
72 12
73 11*
92 11
0/15

Data are given as mean value sd.

* P 0.05.
P 0.001.

Figure 1. Effect of clonidine premedication on mean arterial blood

pressure (A) and heart rate (B) during induction. *P 0.05, **P 0.001.

There were no airway difficulties during anesthesia in

any of the 30 patients. Hypertensive events that required
treatment postoperatively with additional antihypertensive medication occurred in 8 of 15 (53.3%) patients only
in the placebo group (Table 2). The majority of hypertensive events occurred in the first 6 h after surgery on

Figure 2. Mean arterial blood pressure (A) and heart rate (B) during
emergence from anesthesia. *P 0.05, **P 0.001.

1378

ANESTHETIC PHARMACOLOGY PAWLIK ET AL.

CLONIDINE PREMEDICATION AND SLEEP APNEA

ANESTH ANALG
2005;101:1374 80

Table 3. Intraoperative Anesthetic Drug Consumption, Postoperative Analgesic Consumption, Time Until Opening of
Eyes After Cessation of TIVA, Pain Score, and Bispectral Index

Propofol induction dose (mg)

Total propofol consumption (mg kg1 h1)
Remifentanil induction (g)
Total remifentanil consumption (g kg1 h1)
Mivacurium induction (mg)
Total mivacurium (mg)
Diclofenac consumption/24 h
Time until opening of eyes after cessation of TIVA
Piritramide consumption in the recovery room (mg/kg)
Total piritramide consumption in 24 h (mg/kg body weight)
Average pain score, during first hour in recovery room (010 VAS)
Average pain score, first 8 h after operation (010 VAS)
Mean intraoperative BIS

Placebo

Clonidine

218 32.4
10.6 8.1
77.3 4.6
12.2 3.3
20.7 4.0
24.3 6.7
118 53.8
12.3 4.0
10.6 8.0
14.2 8.5
2.8 1.8
1.8 1.3
39 5.0

190 32.2*
7.43 5.1*
77.0 7.0
10.2 2.6
19.7 4.2
22.9 7.6
110 68.6
9.1 3.1*
7.4 5.1
7.4 5.1*
2.1 1.4*
1.1 1.0**
39 4.9

Data are given as mean value sd.

TIVA total IV anesthesia, VAS visual analog scale, BIS bispectral index.
* P 0.05.

the ward. One patient in the placebo group had angina

pectoris in the recovery room and needed aspirin, heparin, and nitroglycerin. Two patients in the clonidine
group required atropine because of bradycardia
(40 bpm) during the operation; none of the patients in
the placebo group required atropine. No adverse respiratory events were observed. The baseline AHI and oxygen desaturation index were comparable in both
groups indicating that all patients had comparably severe OSAS. There were no significant differences between the groups at any time and no significant differences within the groups at different time points (Fig. 3A).
Desaturation indices showed no significant differences
between the groups at all time points. There was a significant decrease on the preoperative night, the day of
surgery, and the postoperative night compared with
baseline values within both groups (Fig. 3B). The recorded MSAT of placebo and clonidine patients showed
no changes from baseline within the groups. However,
on the day of surgery, MSAT was significantly less in
placebo-treated patients than in those receiving
clonidine. (Fig. 3C)

Discussion
OSAS has been recognized as an important disease
with special interest for the anesthesiologist because of
complex airway management and associated cardiovascular diseases. Anesthetic drugs diminish the regular action of the upper airways and compromise
ventilatory response to hypoxemia and hypercarbia
(10). It has been recommended that the amount of
central depressant drugs such as anesthetics and opioids should be minimized to avoid perioperative respiratory complications in the OSAS patient (11).
Other studies have reported a close relationship between OSAS and concomitant cardiovascular diseases

such as hypertension, coronary diseases, and stroke

(12). It is common practice to avoid administering
premedication drugs such as benzodiazepines to
OSAS patients, although premedication is desirable in
patients undergoing anesthesia (13).
In the present study, we found a perioperative incidence of 53% of hypertensive patients in the control
group. One of the sequelae of OSAS is a constantly
increased sympathetic tone as a consequence of repetitive hypoxia and arousal during sleep (14). Oral
clonidine has been shown to reduce sympathetic hyperactivity and suppress the increase in sympathetic activity
after tracheal intubation (15). We found a significant
reduction in heart rate and MAP after clonidine premedication during anesthesia and 24 hours postoperatively.
Importantly, none of our patients used his or her nCPAP, although there are reports that the use of nCPAP
reduces postoperative complications such as cardiac
events (16). Even if desirable, there may be situations in
which patients cannot use their nCPAP because of nasal
packing after having undergone nose and throat
surgery.
There was a significant reduction in both the amount
of propofol required during induction (15%) and total
propofol consumption (15%) in our patients given
clonidine, which is in accordance with other reports on
non-OSAS patients (17). In OSAS patients, short-acting
anesthetic drugs such as propofol and remifentanil are
considered advantageous, and, from a pharmacokinetic
point of view, they are useful to achieve a rapid awakening after surgery. In fact, all patients in our study were
fully awake shortly after cessation of the anesthetics and
could easily be tracheally extubated. It has been reported
that clonidine is associated with a prolonged recovery
from propofol anesthesia (18). However, we were surprised to find a significantly faster awakening after the
use of clonidine premedication in the OSAS patient,

ANESTH ANALG
2005;101:1374 80

ANESTHETIC PHARMACOLOGY
PAWLIK ET AL.
CLONIDINE PREMEDICATION AND SLEEP APNEA

1379

probably because of significantly less propofol consumption throughout the operation. Studies on the analgesic
potency of clonidine show inconsistent results, ranging
from significant opioid-sparing effects to minor analgesic effects (19,20). Importantly, we found a significant
postoperative opioid-sparing effect with a better quality
of analgesia, when clonidine was given preoperatively.
Non-opioids have been shown to ensure safe pain relief
in OSAS patients in the majority of cases, but sufficient
pain control in some cases may be achieved only with
opioids, especially after oropharyngeal surgery (21).
There are several case reports of deleterious outcomes
in OSAS patients who received opioids in the postoperative course and who were not monitored adequately
(22), thus we considered the reduction of opioid administration as an important step in both patient comfort and
safety. Although pain scores were statistically different between the groups, the finding is not of clinical significance,
because the low pain score values in both groups indicated
an excellent pain reduction in all study patients. The most
interesting finding is that clonidine-premedicated patients
required about half the dose of opioids compared with
placebo patients. One possible explanation of reduced opioid requirement and increased effect of opioids in OSAS
patients could be upregulation of -opioid receptors in the
brainstem caused by continuous hypoxemia as shown by
Moss and Laferriere (23) in a hypoxic animal model. There
is evidence of an interaction between clonidine and receptors, which is probably responsible for the significant
opioid-sparing effect we found in our study (24). However,
despite a decreased requirement for opioids, we found no
improvement of respiratory function, which could have
been for two reasons. First, the severity of the underlying
OSAS probably has a greater role than the reduction of
respiratory-depressing opioids. Second, the effects of
clonidine on patients with OSAS in a perioperative setting
are still unknown. There may be potentially depressing
effects of clonidine itself which are not antagonized by a
decreased requirement of opioids, on the ventilatory pattern of those patients. Roberge et al. (25) reported one case
of severe respiratory acidosis, hypotension, and associated
central nervous system depression in an OSAS patient taking continuous clonidine medication.
Another important aspect of OSAS is that
hypopnea/apnea is mainly associated with REM sleep
stages, which facilitates muscle relaxation and leads to
upper airway obstruction in the OSAS patient. Issa (7)
had already shown in 1992 that clonidine is able to
suppress REM sleep and thus improve nocturnal hypoxemia. Preoperative anxiety over the anticipated surgery
leads to interruption of normal sleep cycles in the OSAS

Figure 3. Time course of (A) apnea/hypopnea index (AHI), (B)

desaturation index, and (C) minimal oxygen saturation (MSAT) in
patients treated perioperatively with clonidine or placebo. Desaturation index showed a significant decrease in the preoperative night,

the day of surgery, and the postoperative night compared with

baseline values within groups (*P 0.05, **P 0.01, paired t-test
with Bonferroni adjustment for multiple comparisons). All measured variables showed no difference between both groups with the
exception of MSAT on the day of surgery (P 0.05).

1380

ANESTHETIC PHARMACOLOGY PAWLIK ET AL.

CLONIDINE PREMEDICATION AND SLEEP APNEA

patient, which is aggravated by the administration of

clonidine premedication. This could be a possible explanation for our finding of a marked decrease in AHI and
desaturation index in both groups on the morning before
operation. We speculate that the return of the AHI to
baseline values on the day of surgery in both groups was
related to postoperative pain medication. All AHI of the
first postoperative night showed neither a significant
difference between the clonidine and placebo group nor
differences within the groups compared with baseline
values. Interestingly, minimum desaturation was comparable in both groups except for the day of surgery.
This could be explained by the decreased amount of
propofol and piritramide required in the clonidinepremedicated patients. In our study, protocol polysomnography follow-up was restricted to the postoperative
night. Thus, further studies are needed to examine
whether profound REM rebounds occur after clonidine
premedication during the first week after anesthesia.
There are some limitations to this study. First, the study
was technically difficult to perform. The signal quality of
oronasal flow sensors was intermittently poor because of
patient movement and saliva or blood running over the
oronasal sensor. All analyzed recorded events were manually reevaluated by a specialist, which led to a marked
reduction of originally recorded apnea/hypopnea events.
Second, there was a high interindividual scatter in both
groups representing mild to severe OSAS with an AHI
range of 10/hour to 84/hour. However, our results do
show that patients receiving clonidine premedication had
fewer changes in their perioperative hemodynamic patterns than controls. We demonstrated that 2 g/kg oral
clonidine was able to reduce the postoperative piritramide
consumption significantly without deterioration of the respiratory pattern. We therefore conclude that premedication with 2 g/kg oral clonidine is helpful in reducing
perioperative changes in heart rate, MAP, and the amount
of intraoperative propofol. We also found quicker recovery
at the end of anesthesia and improved postoperative pain
management without adverse effects on respiratory pattern
of the OSAS patient.
The authors thank David Tracey, PhD (Department of Anatomy,
University of New South Wales, Sydney, Australia), for his help in
preparing the manuscript.

References
1. Stradling JR, Crosby JH. Predictors and prevalence of obstructive sleep apnoea and snoring in 1001 middle aged men. Thorax
1991;46:8590.
2. Milleron O, Pilliere R, Foucher A, et al. Benefits of obstructive
sleep apnoea treatment in coronary artery disease: a long-term
follow-up study. Eur Heart J 2004;25:728 34.
3. Ostermeier AM, Hofmann-Kiefer K, Schwender D. [Induction of
anesthesia for a patient with sleep apnea syndrome.] Anaesthesist 2000;49:31720.

ANESTH ANALG
2005;101:1374 80

4. Catley DM, Thornton C, Jordan C, et al. Pronounced, episodic

oxygen desaturation in the postoperative period: its association
with ventilatory pattern and analgesic regimen. Anesthesiology
1985;63:20 8.
5. Parikh SN, Stuchin SA, Maca C, et al. Sleep apnea syndrome in
patients undergoing total joint arthroplasty. J Arthroplasty 2002;
17:635 42.
6. Connolly LA. Anesthetic management of obstructive sleep apnea patients. J Clin Anesth 1991;3:4619.
7. Issa FG. Effect of clonidine in obstructive sleep apnea. Am Rev
Respir Dis 1992;145:4359.
8. Goyagi T, Tanaka M, Nishikawa T. Oral clonidine premedication reduces induction dose and prolongs awakening time from
propofol-nitrous oxide anesthesia. Can J Anaesth 1999;46:894 6.
9. Hall JE, Uhrich TD, Ebert TJ. Sedative, analgesic and cognitive
effects of clonidine infusions in humans. Br J Anaesth 2001;86:511.
10. Sollevi A, Lindahl SG. Hypoxic and hypercapnic ventilatory
responses during isoflurane sedation and anaesthesia in
women. Acta Anaesthesiol Scand 1995;39:931 8.
11. Ip MS, Lam B, Tang LC, et al. A community study of sleepdisordered breathing in middle-aged Chinese women in Hong
Kong: prevalence and gender differences. Chest 2004;125:12734.
12. Mooe T, Franklin KA, Wiklund U, et al. Sleep-disordered
breathing and myocardial ischemia in patients with coronary
artery disease. Chest 2000;117:1597 602.
13. Montravers P, Dureuil B, Desmonts JM. Effects of i.v. midazolam on upper airway resistance. Br J Anaesth 1992;68:2731.
14. Fletcher EC, Bao G, Li R. Renin activity and blood pressure in
response to chronic episodic hypoxia. Hypertension 1999;34:
309 14.
15. Yotsui T. Clonidine premedication prevents sympathetic hyperactivity but does not prevent hypothalamo-pituitary-adrenocortical
responses in patients undergoing laparoscopic cholecystectomy.
J Anesth 2001;15:78 82.
16. Gupta RM, Parvizi J, Hanssen AD, Gay PC. Postoperative complications in patients with obstructive sleep apnea syndrome
undergoing hip or knee replacement: a case-control study.
Mayo Clin Proc 2001;76:897905.
17. Higuchi H, Adachi Y, Arimura S, et al. Oral clonidine premedication reduces the EC50 of propofol concentration for laryngeal
mask airway insertion in male patients. Acta Anaesthesiol
Scand 2002;46:3727.
18. Higuchi H, Adachi Y, Arimura S, et al. Oral clonidine premedication reduces the awakening concentration of propofol.
Anesth Analg 2002;94:609 14.
19. Dimou P, Paraskeva A, Papilas K, Fassoulaki A. Transdermal
clonidine: does it affect pain after abdominal hysterectomy?
Acta Anaesthesiol Belg 2003;54:22732.
20. Nishina K, Mikawa K, Shiga M, et al. Diclofenac and flurbiprofen with or without clonidine for postoperative analgesia in
children undergoing elective ophthalmological surgery. Paediatr Anaesth 2000;10:64551.
21. Virtaniemi J, Kokki H, Nikanne E, Aho M. Ketoprofen and
fentanyl for pain after uvulopalatopharyngoplasty and tonsillectomy. Laryngoscope 1999;109:1950 4.
22. Ostermeier AM, Roizen MF, Hautkappe M, et al. Three sudden
postoperative respiratory arrests associated with epidural opioids in patients with sleep apnea. Anesth Analg 1997;85:452 60.
23. Moss IR, Laferriere A. Central neuropeptide systems and respiratory control during development. Respir Physiol Neurobiol
2002;131:1527.
24. Baker AK, Hoffmann VL, Meert TF. Interactions of NMDA
antagonists and an alpha 2 agonist with mu, delta and kappa
opioids in an acute nociception assay. Acta Anaesthesiol Belg
2002;53:20312.
25. Roberge RJ, Kimball ET, Rossi J, Warren J. Clonidine and sleep
apnea syndrome interaction: antagonism with yohimbine.
J Emerg Med 1998;16:72730.

The Effects of Anesthetics and Ethanol on 2 Adrenoceptor

Subtypes Expressed with G Protein-Coupled Inwardly
Rectifying Potassium Channels in Xenopus Oocytes
Koji Hara, MD, PhD*, Tomohiro Yamakura,
and R. Adron Harris, PhD*

MD, PhD,

Takeyoshi Sata,

MD, PhD,

*Waggoner Center for Alcohol and Addiction Research and Institute for Cellular and Molecular Biology, University of
Texas at Austin; Department of Anesthesiology, University of Occupational and Environmental Health, School of
Medicine, Kitakyushu; and Division of Anesthesiology, Niigata University Graduate School of Medical and Dental
Sciences, Japan

A wide range of physiological effects are mediated by

2-adrenoceptors (ARs) through their association with
G protein-coupled inwardly rectifying potassium
(GIRK) channels. Although 2-ARs are divided into
three subtypes (2AC), a pharmacological distinction
among the subtypes is difficult to establish because of
the lack of a selective agonist and antagonist; therefore,
little is known about the effects of anesthetics on the
2-AR subtypes. We expressed each subtype together
with GIRK1/GIRK2 subunits in Xenopus oocytes and
observed 2-AR-mediated GIRK1/GIRK2 currents to
test the effects of ethanol, halothane, and several IV anesthetics at clinical concentrations. UK 14,304, a selective 2-AR agonist, evoked GIRK1/GIRK2 currents in

wide range of physiological effects are mediated

by 2-adrenoceptors (ARs), for example, cardiovascular, sedative, antinociceptive/anestheticsparing, and hypothermic effects (1,2). The 2-ARs belong to the superfamily of -GTP binding protein (G
protein) coupled receptors and have been divided into
three subtypes (2A, 2B, and a2C) on the basis of
pharmacological and molecular cloning evidence (1,2).
The three subtypes of 2-ARs are highly homologous,
possessing a 50% 60% identity of amino acid sequences among the subtypes. They have similar pharmacological properties and signal transduction pathways mediated through the pertussis toxin-sensitive
Supported, in part, by Grants-in-Aid for Research from the Ministry of Education, Science and Culture of Japan, No. 15790842.
Accepted for publication April 20, 2005.
Address correspondence and reprint requests to Koji Hara, MD,
PhD, Department of Anesthesiology, University of Occupational
and Environmental Health, School of Medicine, 1-1, Iseigaoka, Yahatanishiku, Kitakyushu, 807-8555, Japan. Address e-mail to
kojihara@med.uoeh-u.ac.jp.
DOI: 10.1213/01.ANE.0000180190.05238.D5
2005 by the International Anesthesia Research Society
0003-2999/05

every subtype. None of the IV anesthetics, which included pentobarbital, propofol, ketamine, and alphaxalone, influenced UK 14,304-evoked potassium currents
in any of the receptor subtypes. Ethanol enhanced the
UK 14,304-evoked potassium currents, whereas halothane inhibited the currents. However, these effects
were not significantly different from those on the
baseline-GIRK1/GIRK2 current, suggesting that neither ethanol nor halothane acts directly on the 2-AR
subtypes. Although none of the drugs examined had
any effect on the 2-ARs, the physiological actions of
the 2-ARs mediated by the GIRK1/GIRK2 channels
may be affected by ethanol and halothane.
(Anesth Analg 2005;101:13818)

inhibitory G protein Gi/o. A previous study using

genetically engineered mice with dysfunctional
2A-AR has established that the 2A-subtype mediates
most of the physiological effects of 2-AR, which include sedative, analgesic, and anesthetic-sparing effects (3). The activation of the 2-ARs decreases neuronal excitation by opening G protein-coupled
inwardly rectifying potassium (GIRK) channels, by
inhibition of voltage-gated calcium channels, and by
suppression of adenylate cyclase. The 2-AR agonists
clonidine and dexmedetomidine are used in the treatment of patients with chronic pain and as adjuncts to
general anesthesia. Despite the importance of 2-AR
agonist use for anesthesia, there are few reports studying the effects of anesthetics on the 2-AR subtypes.
GIRK channels are opened by pertussis toxinsensitive G proteins in a membrane-delimited manner in vivo and are activated by various G proteincoupled receptors, including M2-muscarinic, 2adrenergic, D2-dopaminergic, -, -, and -opioid,
and -aminobutyric acid type B (GABAB) receptors
Anesth Analg 2005;101:13818

1381

1382

ANESTHETIC PHARMACOLOGY HARA ET AL.

ALPHA2 ADRENOCEPTOR SUBTYPES AND ANESTHETICS

(4). In noradrenergic neurons, such as locus coeruleus and preganglionic sympathetic neurons, the
GIRK channels underlie inhibitory postsynaptic potentials produced in response to the activation of the
2-ARs (57). Our study (8) demonstrated that the
antinociceptive effect of clonidine was reduced in
GIRK2 knockout mice. Although a precise distribution of the GIRK subtypes has not been determined,
GIRK1/GIRK2 channels are highly abundant in the
central nervous system (CNS) (9,10). Another study
demonstrated that GIRK1/GIRK2 channels expressed in Xenopus laevis oocytes can elicit a potassium current (11). The physiological effects of the
2-ARs are mediated, at least in part, by the activation of GIRK channels, and thus expressing GIRK
channels together with 2-ARs is regarded as a good
tool for studying the effects of anesthetics on 2-ARs
to mimic in vivo physiological effects.
Activation of Gq protein-coupled receptors, such as
M1 muscarinic receptor, results in activation of phospholipase C, Ca2 release from intracellular stores,
and generation of Ca2-dependent Cl currents in the
oocytes. However, Gi/o protein-coupled receptors,
such as the 2-ARs, are unsuitable for electrophysiological analysis because agonist stimulations for the
receptors do not generate measurable current. Furthermore, separating the 2-AR subtypes pharmacologically has been difficult because of the lack of a
selective agonist and antagonist. Consequently, we are
unaware of any published studies that investigated
the effect of anesthetics on the 2-AR subtypes.
In the current study, we demonstrated that each
subtype of 2-ARs could be coupled with GIRK channels in vitro. Using this system, we studied through
electrophysiology the effects of various IV anesthetics,
ethanol, and the volatile anesthetic halothane on the
2-ARs coupled with GIRK channels under the same
condition. We chose the channels composed of GIRK1
and GIRK2 subunits (GIRK1/GIRK2) for studying the
effects of the drugs on 2-AR subtypes using a twoelectrode, voltage-clamp system.

Methods
Xenopus laevis female frogs were purchased from Xenopus Express (Homosassa, FL) and Seac Yoshitomi
(Fukuoka, Japan). The three-aminobenzoic acid ethyl ester was obtained from Sigma (St. Louis, MO). XL-1 Blue
cells and the mCAPRNA Capping kit were obtained
from Stratagene (La Jolla, CA), and the QIAFilter Plasmid Maxi kit was obtained from Qiagen (Valencia, CA).
Halothane was obtained from Halocarbon Laboratories
(River Edge, NJ); 2,6-diisopropylphenol (propofol) was
obtained from Aldrich Chemical Co. (Milwaukee, WI);
alphaxalone was obtained from Research Biochemicals
International (Natick, MA). Ethanol was obtained from

ANESTH ANALG
2005;101:13818

Aaper Alcohol and Chemical (Shelbyville, KY); the pentobarbital sodium, ketamine hydrochloride, and other
reagents were purchased from Sigma. Propofol and alphaxalone were first dissolved in dimethyl sulfoxide and
then diluted in a high potassium (hK) solution. The
largest final concentration of dimethyl sulfoxide (0.01%)
did not affect the potassium current.
GIRK1 and GIRK2 subunit complimentary (c)DNAs
were subcloned into the pBluescript II KS vector,
which was kindly provided by Dr. Michel Ladunski
(Institut de Pharmacologie Moleculaire et Cellulaire,
Valbonne, France). Human 2-AR subtypes 2A, 2B,
and 2C cDNAs in a pBC expression vector were kindly
given to us by Dr. Robert J. Lefkowitz (Department of
Medicine, Duke University Medical Center, Durham,
NC). NcoI-SalI fragments coding the 2A- and 2B-ARs
were ligated and subcloned into pGEM-T Easy Vector.
For the 2C-AR, the authors introduced an NcoI site
upstream and an NdeI site downstream of the coding
region and subcloned the fragment into the NcoI and
NdeI sites of pBluescript II KS. XL-1 Blue cells were
transformed with the cDNAs, and amplified plasmid
was purified with a QIAFilter Plasmid Maxi kit. The
2A- and 2B-AR cDNAs were linearized with SalI,
and 2C-ARs were linearized with NdeI. The cRNAs
were prepared using an mCAPRNA Capping kit.
The use of animals and experiments was approved
by the Animal Care and Use Committees of University
of Texas and the Ethics Committee of Animal Care
and Experimentation by University of Occupational
and Environmental Health in Japan. The surgical procedure was performed on frogs after they were anesthetized in water containing 3-aminobenzoic acid
ethyl ester (240 mg/200 mL of water). The isolation of
Xenopus oocytes was performed, as described previously (12). Isolated oocytes were placed in modified
Barths saline (MBS) containing 88 mM of NaCl, 1 mM
of KCl, 2.4 mM of NaHCO3, 0.82 mM of MgSO4, 0.91
mM of CaCl2, 0.33 mM of Ca(NO3)2, and 10 mM of
HEPES buffer adjusted to a pH value of 7.5. The
oocytes were injected with cRNA (30 nL/oocyte) encoding a combination of GIRK channels and 2-AR
subtype in a 1:1 ratio and were individually placed in
Corning cell wells (Corning Glass Works, Corning,
NY) containing incubation medium (sterile MBS supplemented with 10 mg/L of streptomycin, 100,000
U/L of penicillin, 50 mg/L of gentamicin, 90 mg/L of
theophylline, and 220 mg/L of pyruvate). The injected
oocytes were incubated at 15C19C for 23 days and
were then used for electrophysiological recording (12).
Oocytes co-expressing GIRK1/GIRK2 channels and
each of the 2-AR subtypes were placed in a rectangular chamber (100-L volume) and perfused
(2 mL/min) with MBS. The effects of IV anesthetics
including pentobarbital, propofol, ketamine, and alphaxalone were studied using a two-electrode
voltage-clamp system, as reported previously (11).

ANESTH ANALG
2005;101:13818

The animal poles of oocytes were impaled with 2 glass

electrodes (0.510 M) filled with 3 M of KCl, and the
oocytes were voltage-clamped at 80 mV using a
Warner Instruments model OC-752B oocyte clamp
(Hamden, CT). The oocytes were initially bathed in
MBS and then changed to a hK solution containing 2
mM of NaCl, 96 mM of KCl, 1 mM of MgCl2, 1.5 mM
of CaCl2, and 5 mM of HEPES buffer adjusted to a pH
value of 7.5. Because the baseline current obtained in
physiological buffer (MBS) was small, the hK solution
was used to reverse the driving force of the channel
and to provide a large inward current, as previously
reported (13). After stable responses were established,
UK 14,304, a selective full-agonist of 2-AR, was applied in the hK solution for 15 s before returning to the
hK solution. All drugs dissolved in hK solution were
preapplied for 1 min before being co-applied with UK
14,304. The effects of the drugs were expressed as the
fraction of control responses, which were calculated
by averaging the control responses before and after
the application of the drugs.
The GIRK1/GIRK2 channels are mainly distributed
in the CNS (9,10). Our previous work (11) has shown
that the GIRK1/GIRK2 channels are not affected by IV
anesthetics. Accordingly, we co-expressed the GIRK1/
GIRK2 channels with the 2-ARs to study the effects of
the drugs. For studying the effect of ethanol, solutions
of 50, 100, and 200 mM of ethanol were initially assessed for effects on the baseline GIRK currents. Next,
ethanol was tested on UK 14,304-evoked GIRK currents. For testing the volatile anesthetic halothane, 250,
500, and 1000 M (1, 2, and 4 mean alveolar anesthetic
concentration) of halothane solutions were initially
assessed for their effects on baseline GIRK currents.
The concentrations of halothane in the chamber were
quantified by gas chromatography. Approximately
the half-maximal effective concentration (EC50) of UK
14,304 in each subtype was applied to test the effects
of the anesthetics and ethanol. Data are represented as
mean sem. Statistical analysis was performed by
one-way analysis of variance for multiple comparisons and by unpaired t-test for comparisons between
two groups. Differences were considered statistically
significant at P 0.05. The values of the EC50 and the
Hill coefficient were calculated by nonlinear regression using GraphPad Prism software version 3.0
(GraphPad Inc., San Diego, CA). All experiments were
performed at room temperature (23C).

Results
The application of UK 14,304 to the oocytes coexpressing the GIRK1/GIRK2 with each 2-AR subtype gave consistent 2-AR subtype-mediated potassium currents (Fig. 1). Initial studies using longer
application times of UK 14,304 indicated that 15 s

ANESTHETIC PHARMACOLOGY
HARA ET AL.
ALPHA2 ADRENOCEPTOR SUBTYPES AND ANESTHETICS

1383

Figure 1. Representative tracing of G protein-coupled inwardly

rectifying potassium (GIRK) current and UK 14,304-evoked GIRK
current in the oocyte expressing the 2A-adrenoceptor (AR) and
GIRK1/GIRK2 channel. The oocyte was bathed in modified Barths
saline (MBS) and then changed to a high potassium (hK) solution.
After a stable baseline current was established, UK 14,304 was
repeatedly applied for 15 s with consistent responses being
observed.

yielded a consistent and mostly maximal effect. The

oocytes were exposed to various concentrations of UK
14,304 (2A, 0.01 nM1 M; 2B, 0.1 nM10 M; 2C, 1
nM100 M) to obtain the concentration-response relationship. UK 14,304 increased 2-AR-mediated potassium current in a concentration-dependent manner
(Fig. 2). The EC50 values of UK 14,304 for 2A, 2B, and
2C were 3.4 0.3 nM, 89 3 nM, and 212 6 nM,
respectively, and the Hill coefficients were 1.0 0.1
for all the subtypes.
To obtain a control response, UK 14,304 was applied
repeatedly until a consistent response was observed.
After the control currents were obtained, concentrations of 2- and 4-times the EC50 of each IV anesthetic,
including propofol, ketamine, pentobarbital, and alphaxalone, were applied in a hK solution for 1 min
before coadministration of the anesthetic and UK
14,304 for 15 s. None of the anesthetics affected the
baseline potassium current, which is consistent with
our previous results (11). Similarly, the anesthetics
tested did not significantly affect UK 14,304-evoked
potassium currents in the oocytes co-expressing the
GIRK1/GIRK2 with each of the 2-AR subtypes (Fig.
3).
As previously demonstrated (16), ethanol enhanced
the baseline GIRK1/GIRK2 currents per se in a
concentration-dependent manner. UK 14,304-evoked
potassium currents were also enhanced by ethanol to
a degree essentially similar to that of baseline currents
(Fig. 4). Conversely, the volatile anesthetic halothane
inhibited the baseline currents at 1, 2, and 4 mean
alveolar anesthetic concentration (Fig. 5). Halothane
also inhibited UK 14,304-evoked currents, but the degree of the inhibition was not significantly different
from that of the baseline currents. There were no

1384

ANESTHETIC PHARMACOLOGY HARA ET AL.

ALPHA2 ADRENOCEPTOR SUBTYPES AND ANESTHETICS

Figure 2. The UK 14,304 concentration-response relationship in

oocytes expressing each of the 2-adrenoceptor (AR) subtypes and
G protein-coupled inwardly rectifying potassium (GIRK)1/GIRK2
channels. UK 14,304-evoked potassium currents were observed for
all the subunits in a concentration-dependent manner. Data are
represented as mean sem.

significant differences among 2-AR subtypes in these

experiments.

Discussion
We demonstrated in vitro coupling of 2-ARs and
GIRK channels in this study. The physiological effects
of 2-ARs are partly mediated through GIRK channels

ANESTH ANALG
2005;101:13818

in the CNS. Therefore, the assay system used in this

study is a good model for mimicking in vivo physiological conditions to study the effects of anesthetics on
2-ARs. Additionally, we are unaware of any published reports that compare the effects of anesthetics
on 2-AR subtypes. There are no selective agonists
and antagonists for the 2-AR subtypes, making the
pharmacological separation of the subtypes difficult.
The physiological effects of the 2-ARs are associated with the production of some anesthetic qualities,
such as sedation and analgesia. Our previous work (8)
demonstrated that an analgesic effect of clonidine is
mediated by the GIRK2 subtype, which suggests that
anesthetic actions of 2-ARs are at least partially mediated by GIRK channels. Another study showed that
the GIRK1/GIRK2 channels per se are insensitive to IV
anesthetics at clinical concentrations (11). We used the
GIRK1/GIRK2 subunits distributed predominantly in
the CNS to test anesthetics and ethanol and found that
the IV anesthetics had no effect on any 2-AR subtypes
coupled with the GIRK1/GIRK2 channels. These results indicate that the IV anesthetics tested are unlikely to act directly on any of the 2-AR subtypes.
Next, we tested the effects of ethanol on UK 14,304evoked GIRK1/GIRK2 currents. As shown by the previous study (16), ethanol enhanced baseline GIRK1/
GIRK2 currents in a concentration-dependent manner.
Similarly, ethanol increased UK 14,304-evoked currents to a degree similar to baseline currents. Last, we
tested a volatile anesthetic halothane on the GIRK1/
GIRK2-coupled 2-AR subtypes. Halothane inhibited
not only the baseline GIRK1/GIRK2 current, but also
the UK 14,304-evoked currents in a concentrationdependent manner. The degree of the inhibition was
not significantly different between the baseline currents and the UK 14,304-evoked currents.
Gq protein-coupled receptors, which include muscarinic
acetylcholine, 5-hydroxytryptamine, and metabotropic glutamate receptors, were shown to be affected by anesthetics
and ethanol (1720). Gq protein-coupled receptors may be
plausible targets of the drugs. The results of this study
suggest that, at clinical concentrations, neither the anesthetics used in this study nor ethanol have a direct effect on the
molecules of the 2-AR subtypes or on the transport of the
G protein subunit to the GIRK channels expressed in the
oocytes. It might be speculated that Gi/o protein-coupled
receptors are relatively insensitive to anesthetics and ethanol, as compared with Gq protein-coupled receptors. However, it is presumed that the physiological effects of the
endogenous 2-AR agonists, that is, catecholamines acting
through the GIRK channels, may be affected by ethanol
and halothane. Ethanol may augment endogenous agonistevoked GIRK1/GIRK2 current, which may partially explain the antinociceptive effect of ethanol. We (8) demonstrated that the antinociceptive effects of ethanol are
reduced in GIRK2-null mutant mice. Furthermore, Mao

ANESTH ANALG
2005;101:13818

ANESTHETIC PHARMACOLOGY
HARA ET AL.
ALPHA2 ADRENOCEPTOR SUBTYPES AND ANESTHETICS

1385

Figure 3. The effects of IV anesthetics on UK 14,304-evoked potassium currents. (A) UK 14,304 in the high potassium (hK) solution was
applied to oocytes expressing the 2A-adrenoceptor (AR) and G protein-coupled inwardly rectifying potassium (GIRK)1/GIRK2 channel for
15 s. After returning to the baseline current, IV anesthetics were preapplied for 1 min before being co-applied with UK 14,304 for 15 s. (B)
Effects of IV anesthetics on the UK 14,304-evoked GIRK1/GIRK2 currents. Anesthetics were applied at concentrations corresponding to 2and 4-times the half-maximal effective concentration (EC50) of the anesthetic, i.e., pentobarbital (100 and 200 M) (14), propofol (2 and 4 M)
(14), ketamine (20 and 40 M) (15), and alphaxalone (10 and 20 M) (15). None of the anesthetics that were tested showed any effect at clinical
concentrations. Data are represented as mean sem.

and Abdel-Rahman (21) demonstrated in rats that ethanol

synergistically increased the analgesic effect of clonidine.
The absence of any effect of volatile anesthetics on
2-ARs was shown in another assay system by Pentyala
et al. (22), who demonstrated that volatile anesthetics,
including halothane, did not influence 2-adrenergic signaling in isolated platelets. Halothane may inhibit endogenous 2-AR agonist-evoked GIRK1/GIRK2 current
in vivo. GIRK channels are unlikely to contribute to the
anesthetic-sparing effect of 2-AR agonist. Currently, the
clinical roles of the inhibitory effect of halothane on
GIRK1/GIRK2 channels are not clear; however, they
might be associated with some side effects, such as
tachycardia, or an excitement during anesthesia.

The results of this study also imply that the physiological effects of other receptors coupled with the
GIRK1/GIRK2 channels, such as -opioid, cannabinoid,
and GABAB receptors, may be modulated by ethanol
and halothane in the CNS. Antinociceptive effects associated with these receptors have been eliminated in
GIRK2 knockout or mutant mice (8,23), which indicates
that the GIRK2 subunits are involved in antinociceptive
effects through various neuronal systems.
Other GIRK subunits were not addressed in the
present study. Specifically, GIRK1/GIRK4 subunits
dominantly exist and couple with M2 receptors in the
heart (24). Our previous study (11) demonstrated that
GIRK1/GIRK4 subunits are insensitive to anesthetic,

1386

ANESTHETIC PHARMACOLOGY HARA ET AL.

ALPHA2 ADRENOCEPTOR SUBTYPES AND ANESTHETICS

ANESTH ANALG
2005;101:13818

Figure 4. The effect of ethanol on UK 14,304-evoked potassium currents. (A) Representative tracing of UK 14,304-evoked potassium currents
in the oocytes expressing the 2A-adrenoceptor (AR) and G protein-coupled inwardly rectifying potassium (GIRK)1/GIRK2 channel. Ethanol
(100 mM) enhanced the baseline potassium currents and UK 14,304-evoked potassium currents. (B) Ethanol (50, 100, and 200 mM) augmented
the UK 14,304-evoked potassium currents and baseline potassium currents in a concentration-dependent manner. The percent potentiation
was calculated from the following equations: percent potentiation for baseline currents 100(b/a 1), and percent enhancement for UK
14,304-evoked currents 100(d/c 1). Data are represented as mean sem. These values were not statistically different from each other.
N.S. indicates no significance.

and further experiments testing the effects of anesthetics on GIRK1/GIRK4 channels coupled with M2 receptors would be beneficial to understand the drugs
actions on cardiac function in detail.
The present study showed that all 2-AR subtypes can
be coupled with GIRK channels in vitro in a Xenopus oocyte
expression system. The 2-AR, a member of the G proteincoupled receptor family, was found to be an unlikely target
of both ethanol and the anesthetics used in this study. Some

clinically significant aspects of ethanol and volatile anesthetics may arise through the modulation of the effects of
2-ARs via GIRK channels.

The authors thank Dr. Michel Ladunski for kindly providing cDNAs of GIRK1 subunit and GIRK2 subunit and Dr. Robert J.
Lefkowitz for the generous gift of cDNAs of 2-AR subtypes.

ANESTH ANALG
2005;101:13818

ANESTHETIC PHARMACOLOGY
HARA ET AL.
ALPHA2 ADRENOCEPTOR SUBTYPES AND ANESTHETICS

1387

Figure 5. Effect of halothane on UK 14,304-evoked potassium currents. (A) Representative tracing of UK 14,304-evoked potassium currents
in the oocytes expressing the 2A-adrenoceptor (AR) and G protein-coupled inwardly rectifying potassium (GIRK)1/GIRK2 channel.
Halothane (500 M) inhibits the baseline potassium currents and UK 14,304-evoked potassium currents. (B) Halothane at 250 M, 500 M,
and 1 mM, corresponding to 1, 2, and 4 mean alveolar anesthetic concentration, respectively (14), suppressed the UK 14,304-evoked
potassium currents and baseline potassium currents in a concentration-dependent manner. The percent change for the UK 14,304-evoked
currents, which was calculated from 100(1 d/c), was essentially similar to that for the baseline currents calculated from 100(1 b/a). Data
are represented as mean sem. N.S. indicates no significance.

References
1. MacDonald E, Kobilka BK, Scheinin M. Gene targeting-homing
in on 2-adrenoceptor-subtype function. Trends Pharmacol Sci
1997;18:2119.
2. Kable JW, Murrin LC, Bylund DB. In vivo gene modification
elucidates subtype-specific functions of 2-adrenergic receptors.
J Pharmacol Exp Ther 2000;293:17.
3. Lakhlani PP, MacMillan LB, Guo TZ, et al. Substitution of a
mutant alpha2a-adrenergic receptor via hit and run gene
targeting reveals the role of this subtype in sedative, analgesic,
and anesthetic-sparing responses in vivo. Proc Natl Acad Sci
USA 1997;94:9950 5.
4. North RA. Drug receptors and the inhibition of nerve cells.
Br J Pharmacol 1989;98:1328.
5. Williams JT, Henderson G, North RA. Characterization of alpha
2-adrenoceptors which increase potassium conductance in rat
locus coeruleus neurones. Neuroscience 1985;14:95101.
6. Yoshimura M, Polosa C, Nishi S. Slow IPSP and the
noradrenaline-induced inhibition of the cat sympathetic preganglionic neuron in vitro. Brain Res 1987;419:383 6.

7. Aghajanian GK, Wang YY. Pertussis toxin blocks the outward

currents evoked by opiate and 2-agonists in locus coeruleus
neurons. Brain Res 1986;371:390 4.
8. Blednov YA, Stoffel M, Alva H, Harris RA. A pervasive mechanism for analgesia: activation of GIRK2 channels. Proc Natl
Acad Sci USA 2003;100:277 82.
9. Lesage F, Guillemare E, Fink M, et al. Molecular properties of
neuronal G-protein-activated inwardly rectifying K channels.
J Biol Chem 1995;270:28660 7.
10. Karschin C, Dissmann E, Stuhmer W, Karschin A. IRK(13) and
GIRK(1 4) inwardly rectifying K channel mRNAs are differentially expressed in the adult rat brain. J Neurosci 1996;16:
3559 70.
11. Yamakura T, Lewohl JM, Harris RA. Differential effects of general anesthetics on G protein-coupled inwardly rectifying and
other potassium channels. Anesthesiology 2001;95:144 53.
12. Dildy-Mayfield JE, Harris RA. Comparison of ethanol sensitivity of rat brain kainate, DL-alpha-amino-3-hydroxy-5-methyl-4isoxalone proprionic acid and N-methyl-D-aspartate receptors
expressed in Xenopus oocytes. J Pharmacol Exp Ther 1992;262:
48794.

1388

ANESTHETIC PHARMACOLOGY HARA ET AL.

ALPHA2 ADRENOCEPTOR SUBTYPES AND ANESTHETICS

13. Kovoor A, Henry DJ, Chavkin C. Agonist-induced desensitization of the opioid receptor-coupled potassium channel
(GIRK1). J Biol Chem 1995;270:589 95.
14. Franks NP, Lieb WR. Molecular and cellular mechanisms of
general anaesthesia. Nature 1994;367:60714.
15. Krasowski MD, Harrison NL. General anaesthetic actions on
ligand-gated ion channels. Cell Mol Life Sci 1999;55:1278 303.
16. Lewohl JM, Wilson WR, Mayfield RD, et al. G-protein-coupled
inwardly rectifying potassium channels are targets of alcohol
action. Nat Neurosci 1999;2:1084 90.
17. Durieux ME. Halothane inhibits signaling through m1 muscarinic receptors expressed in Xenopus oocytes. Anesthesiology
1995;82:174 82.
18. Durieux ME. Inhibition by ketamine of muscarinic acetylcholine
receptor function. Anesth Analg 1995;81:57 62.
19. Minami K, Minami M, Harris RA. Inhibition of 5-hydroxytryptamine
type 2A receptor-induced currents by n-alcohols and anesthetics.
J Pharmacol Exp Ther 1997;281:113643.

ANESTH ANALG
2005;101:13818

20. Minami K, Gereau RW 4th, Minami M, et al. Effects of ethanol

and anesthetics on type 1 and 5 metabotropic glutamate receptors expressed in Xenopus laevis oocytes. Mol Pharmacol 1998;
53:148 56.
21. Mao L, Abdel-Rahman AA. Synergistic behavioral interaction
between ethanol and clonidine in rats: role of alpha-2 adrenoceptors. J Pharmacol Exp Ther 1996;279:4439.
22. Pentyala S, Moller D, Chowdhury A, et al. Effects of inhalational
anesthetics on alpha2-adrenergic signaling in isolated platelets.
Toxicol Lett 1998;23;100 1, 11520.
23. Ikeda K, Kobayashi T, Kumanishi T, et al. Involvement of
G-protein-activated inwardly rectifying K (GIRK) channels in
opioid-induced analgesia. Neurosci Res 2000;38:113 6.
24. Krapivinsky G, Gordon EA, Wickman K, et al. The G-proteingated atrial K channel IKACh is a heteromultimer of two
inwardly rectifying K()-channel proteins. Nature 1995;374:
135 41.

Spatial Memory Performance 2 Weeks After General

Anesthesia in Adult Rats
Catherine Crosby*, Deborah J. Culley, MD, Mark G. Baxter,
Rustam Yukhananov, MD, PhD, and Gregory Crosby, MD

PhD,

*Department of Psychology, Harvard University, Cambridge, Massachusetts; Department of Experimental Psychology,

Oxford, UK; Department of Anesthesia, Harvard Medical School, Brigham & Womens Hospital, Boston, Massachusetts

We have previously demonstrated that general anesthesia with 1.2% isoflurane-70% nitrous oxide impairs acquisition of a radial arm maze task in both young and aged
rats when testing begins 2 days after anesthesia and in
aged rats when testing begins 2 wk later. We designed this
study to examine whether postanesthesia learning impairment is persistent in young rats. Six-month-old rats
were randomized to anesthesia for 2 h with 1.2%
isoflurane-70% nitrous oxide, 1.8% isoflurane, or a control
group that received 30% oxygen (n 10 per group). Rats
recovered for 2 wk and were then tested daily on a radial
arm maze for 14 days. There were no differences between

linical studies indicate that anesthesia and surgery are associated with early cognitive impairment in both young and aged patients (1,2).
Recovery seems to be age-dependant, however, as
only the aged suffer cognitive impairment lasting 3
mo or more (1,2). In the laboratory, we have demonstrated that a 2-h anesthetic with 1.2% isoflurane
(ISO)-70% nitrous oxide (N2O) attenuates performance improvement on a previously learned task in
aged rats but improves performance in young adult
rats (3). We found subsequently that acquisition of
new spatial memory is impaired in both young and
aged rats for 214 days after ISO N2O anesthesia
and that this learning/memory impairment persists
up to 28 days in aged rats (4,5). We have not determined whether postanesthetic learning impairment
persists in adult rats but, given the greater plasticity of
Supported, in part, by an American Geriatrics Society Jahnigen
Award and a Harvard / Hartford Foundation New Investigator
Award to D. Culley and by National Institutes of Health R01
AG20253 to G. Crosby.
Accepted for publication May 3, 2005.
Address correspondence and reprint requests to Gregory Crosby,
MD, Department of Anesthesiology, Brigham & Womens Hospital,
75 Francis St. Boston, MA 02115. Address electronic mail to
gcrosby@zeus.bwh.harvard.edu.
DOI: 10.1213/01.ANE.0000180835.72669.AD
2005 by the International Anesthesia Research Society
0003-2999/05

the controls and anesthesia groups in number of correct

choices to first error or time to complete the maze. There
was no main effect of group in terms of total number of
errors (P 0.05) but the group by day interaction was
significant (P 0.05), reflecting improved performance in
the 1.2% isoflurane-70% nitrous oxide group relative to
controls during the later days of testing (P 0.005).
Hence, in adult rats, previous general anesthesia is not
associated with impaired learning 2 wk later. In fact, previous 1.2% isoflurane-70% nitrous oxide improves maze
performance 2 wk later.
(Anesth Analg 2005;101:1389 92)

the young brain and the fact that postoperative cognitive impairment is short-lived in young and middle
aged patients, we speculated that postanesthetic spatial memory impairment resolves sooner in adult rats.
To test this hypothesis, we evaluated acquisition of
spatial memory in adult rats beginning 2 wk after
general anesthesia.

Methods
This study was approved by the Standing Committee
on the Use of Animals in Research and Teaching,
Harvard University Faculty of Arts and Sciences.
Thirty 6-mo old male Fischer 344 rats were acquired
from the National Institute of Health Aged rat colony
at Harlan and housed individually in a climate-and
humidity-controlled room on a 12-h light:dark cycle
with continuous access to food and water. After a
1-wk acclimation period, rats were randomly assigned
(n 10 per group) to receive 1.2% ISO-70% N2O-30%
oxygen (ISO N2O), 1.8% ISO-30% oxygen (ISO), or
30% oxygen alone (control). Anesthesia was induced
by placing rats in a chamber flushed with 3% ISO and
100% oxygen and intubating the trachea with a 14gauge catheter. Rats were then mechanically ventilated with the appropriate anesthetic for 2 h with a
2-mL tidal volume delivered at a rate of 45 breaths/
Anesth Analg 2005;101:138992

1389

1390

ANESTHETIC PHARMACOLOGY CROSBY ET AL.

SPATIAL MEMORY WEEKS AFTER ANESTHESIA IN RATS

min, which pilot studies demonstrate maintains Paco2

at 41.4 0.3 cm H2O (mean sem) in 6-mo-old male
Fischer 344 rats. Rectal temperature was controlled to
37C 0.5C. Arterial oxygen saturation and mean
arterial blood pressure (MAP) were measured noninvasively using a pulse oximeter and a rat tail cuff
during anesthesia. After 2 h, the anesthetics were discontinued and 100% oxygen was delivered. The rate of
ventilation was reduced until spontaneous ventilation
resumed and the trachea was extubated when the rat
was responsive. Control rats were placed in a box
flushed with 30% oxygen for 2 h and were not tracheally intubated. All animals were recovered for 30 min
in a box flushed with 40% oxygen and then placed in
their home cage.
Testing of cognitive function was performed in a
12-arm radial maze (RAM) as previously described
(35). This RAM tests spatial working memory, assesses the integrity of the frontal cortex, entorhinal
cortex and hippocampus (6,7) and can detect subtle
differences in learning caused by sedatives and anesthetics (35,8).
To ensure motivated performance, rats were foodrestricted to 85% of free-feeding body weight starting
11 days after anesthesia but had free access to water in
the home cage. Rats were adapted to the maze for
10 min/day during days 1113 after anesthesia, during which the maze was randomly scattered with food
rewards and the rat was allowed to freely explore the
maze. Formal testing began 14 days after anesthesia
and consisted of a daily 15-min session in which the
rat was placed on the central platform of the maze and
all arms were baited. The rat was allowed to choose
arms in any order until all 12 arms were visited or
15 min elapsed. A correct choice was defined as one in
which the rat entered and proceeded more than 80%
down a baited arm not previously explored. An error
was scored when the rat entered and proceeded more
than 80% down an arm it has previously visited or
failed to enter the arm in 15 min. Time to complete the
maze, number of correct choices before first error, and
error rate were recorded.
Measures of performance from the RAM (time to
complete the maze, number of correct choices to first
error, and error rate) were analyzed with repeatedmeasures analysis of variance, with anesthesia group
as a between-subject factor and day of testing as the
within-subject factor. Statistical analysis of MAP and
oxygen saturation was performed using one-way
analysis of variance followed by Dunnetts test for
multiple comparisons.

Results
Anesthesia and mechanical ventilation were physiologically well tolerated in both groups of anesthetized

ANESTH ANALG
2005;101:1389 92

rats. MAP was similar in rats anesthetized with ISO

and those anesthetized with ISO N2O and remained
within the physiologic range (104 2 mm Hg versus
106 1 mm Hg, respectively; P 0.05). In addition,
there were no differences in oxygen saturation among
rats anesthetized with ISO and ISO N2O (98.1%
0.3% versus 98.5% 0.3%, respectively; P 0.05).
In two of the three measures of performance there
were no significant differences between the controls
and the two anesthesia groups. However, there was a
significant effect of test day for each measure, indicating that learning took place across the 14 days of
testing. For time to complete the maze (Fig. 1), the
main effect of day was significant (P 0.0005) but the
main effect of group was not (P 0.05) nor were there
any group by day interactions (P 0.05). For number
of correct choices to first error (Fig. 2), the main effect
of day was significant (P 0.0005) but the effect of
group was not (P 0.05), nor were there any group by
day interactions (P 0.05). In terms of total number of
errors (Fig. 3), there was a significant effect of day (P
0.0005) and although there was no main effect of
group (P 0.05) the group by day interaction was
significant (P 0.05). This interaction reflects a difference in performance across days between the control
group and the ISO N2O group. The origin of this
interaction is superior performance in rats anesthetized with ISO N2O during later days of testing.
When we compared control and ISO N2O groups
only on the first week (days 17) of testing, there was
no significant difference (P 0.05) but a group difference was present during the second week (P 0.05).
Hence, general anesthesia with ISO N2O does not
impair and may enhance acquisition of a spatial memory task administered 2 wk after general anesthesia in
6-mo old Fischer 344 rats.

Discussion
This primary finding of this study is that there is no
impairment of spatial memory acquisition 2 or more
weeks after general anesthesia with 1.2% ISO/70%
N2O or 1.8% ISO in adult rats. In fact, anesthesia with
ISO N2O, but not a MAC equivalent concentration
of ISO, enhanced performance on the spatial memory
task. Taken together with our previous studies showing impairment in adult rats on the same task 2 days
after ISO N2O anesthesia and in aged rats 2 weeks
after anesthesia (4,5), these data support the idea that
the persistent memory/learning effects of these anesthetics are age-dependent. Thus, spatial memory performance returns to normal sooner in adult rats than
aged rats after general anesthesia with ISO N2O.
This is not the first study to demonstrate improved
memory performance in rodents after anesthesia. In a
previous study, in which adult rats partially learned

ANESTH ANALG
2005;101:1389 92

Figure 1. Average time to complete the maze. Rats (n 10 per

group) were anesthetized with 1.2% isoflurane (ISO)-70% nitrous
oxide (N2O)-30% oxygen (ISO N2O), 1.8% ISO-30% oxygen (ISO),
or 30% oxygen alone (control) and tested for 14 days beginning 2 wk
after anesthesia on a 12-arm radial maze. The main effect of testing
day was significant (P 0.0005) but the main effect of group was
not (P 0.05) nor were there any group by day interactions (P
0.05). Data were analyzed with repeated-measures analysis of variance, with anesthesia group as a between subject factor and day of
testing as the within-subject factor. Data points represent the mean
sem.

the RAM task before receiving anesthesia, we found

that ISO N2O decreased their subsequent error rate
(3). Similarly, the volatile anesthetics halothane, enflurane, and isoflurane have been shown to enhance
memory in young adult mice (9). In those studies, over
4 consecutive days mice received 1 hour of anesthesia
beginning immediately after RAM training, with the
result that error rate was reduced 60%70% (10). The
mechanisms responsible for this improved performance are not known but it is interesting that memory
consolidation is enhanced when natural sleep occurs
soon after a learning task (11). Such improved memory performance after general anesthesia has not been
reported in human surgical patients, however, perhaps because the memory improvement is subtle and
difficult to detect in the ill postsurgical patient. Nonetheless, there is evidence from volunteers that visual
memory for images of high emotional valence is enhanced at subanesthetic concentrations of thiopental
and, possibly, propofol, but the clinical relevance of
these results is unclear (12). It appears, therefore, that
general anesthetics may have differential effects on
different types of learning and may facilitate learning
under some circumstances.
One interesting feature of the improved memory
task performance in our study is that it occurred only
in the group that received N2O. This is unlikely to be
explained by differences in systemic physiology or
depth of anesthesia, as MAP and oxygen saturation
were similar and within the physiologic range in both
anesthesia groups and the delivered MAC concentrations were equivalent. N2O and ISO have different
receptor actions such that N2O is primarily an
N-methyl-d-aspartate (NMDA) receptor antagonist

ANESTHETIC PHARMACOLOGY
CROSBY ET AL.
SPATIAL MEMORY WEEKS AFTER ANESTHESIA IN RATS

1391

Figure 2. Average number of correct choices before first error. Rats

(n 10 per group) were anesthetized with 1.2% isoflurane (ISO)70% nitrous oxide (N2O)-30% oxygen (ISO N2O), 1.8% ISO-30%
oxygen (ISO), or 30% oxygen alone (control) and tested for 14 days
beginning 2 wk after anesthesia on a 12-arm radial maze. The main
effect of testing day was significant (P 0.0005) but the main effect
of group was not (P 0.05) and nor were there any group by day
interactions (P 0.05). Data were analyzed with repeated-measures
analysis of variance, with anesthesia group as a between subject
factor and day of testing as the within-subject factor. Data points
represent the mean sem.

Figure 3. Average number of errors. Rats (n 10 per group) were

anesthetized with 1.2% isoflurane (ISO)-70% nitrous oxide (N2O)30% oxygen (ISO N2O), 1.8% ISO-30% oxygen (ISO), or 30%
oxygen alone (control) and tested for 14 days beginning 2 wk after
anesthesia on a 12-arm radial maze. There was a main effect of
testing day (P 0.0005) and although there was no main effect of
group (P 0.05), the group by day interaction was significant and
reflected superior performance in rats anesthetized with ISO N2O
during the later days of testing. Data were analyzed with repeatedmeasures analysis of variance, with anesthesia group as a between
subject factor and day of testing as the within-subject factor. Data
points represent the mean sem.

with weak agonist activity at gamma-amino butyric

acid (GABA) and cholinergic receptors, whereas ISO is
primarily a positive GABA receptor modulator with
weak NMDA receptor antagonist properties. This
leads to the reasonable assumption that there was
greater NMDA receptor antagonism and less GABA
modulation in the ISO N2O group than in the ISO
group. Memory improvement in the former group is
therefore paradoxical because activation of NMDAglutamate receptors is required for long-term memory

1392

ANESTHETIC PHARMACOLOGY CROSBY ET AL.

SPATIAL MEMORY WEEKS AFTER ANESTHESIA IN RATS

formation (13,14). However, the effect of NMDA receptor antagonists on the capacity of the adult brain to
learn days or weeks later have not been studied extensively and there is evidence for both impaired acquisition (15) as well as better retention of certain
memory tasks (16). In addition, in cultured hippocampal slices, prolonged NMDA receptor blockade results
in axonal sprouting and more miniature excitatory
postsynaptic potentials (17), which are morphological
and neurophysiological features of memorydependent synaptic plasticity. Thus, there is limited
behavioral and cellular evidence for a potential facilitating effect of certain NMDA antagonists on memorydependent synaptic plasticity, but whether these explain the enhanced learning we observed on a
hippocampus-dependent spatial memory task after
ISO N2O anesthesia is unknown.
This study is limited in a number of ways. First,
although the anesthetized animals were tracheally intubated and mechanically ventilated, the controls
were not. It is unlikely that the improvement in maze
performance in the ISO N2O anesthetized rats was
attributable to tracheal intubation and mechanical
ventilation, as the ISO group was treated identically
but showed no behavioral improvement. Second, the
RAM task depends on hunger-induced motivation
and the ability to see extra-maze visual cues. However, all rats included in the study learned the maze,
had similar weights for the 11 days after anesthesia
before RAM testing, and consumed food pellets when
returned to their cage, indicating that they were able
to see visual cues and were motivated by hunger.
Finally, the RAM tests spatial memory, and it is possible that other cognitive domains may be more or less
sensitive to anesthesia-induced changes.
It is difficult to understand how general anesthesia
ablates memory during the time it is administered,
impairs it 48 hours later irrespective of age, and yet
subsequently enhances long-term memory performance in young animals and impairs it in the old.
However, memory itself is a complicated and poorly
understood phenomenon and our understanding of
how general anesthetics influence it is also poor. Anesthetics have known effects on many of neurochemical events associated with memory-dependent synaptic plasticity (18), but it is impossible to speculate how
they may influence memory acquisition over the
longer term.
In summary, our results show that in contrast to the
persistent impairment observed in aged rats (5), MAC
equivalent dosages of ISO N2O and ISO alone do
not impair spatial memory acquisition 2 weeks later in
adult rats. In fact, previous anesthesia with ISO N2O
actually improves RAM performance 2 weeks later.
Together with previous work, these results indicate
that postanesthetic memory impairment is agedependent such that adult rats recover sooner than

ANESTH ANALG
2005;101:1389 92

aged rats. Because improved task performance 2

weeks after anesthesia can not be explained by the
pharmacokinetics of the drugs involved, these data
suggest that persistent anesthesia-induced changes occur in neural structures and/or biochemical cascades
mediating memory. Such lasting cognitive changes
provide a basis for examining further the long-term
neurobiological effects of anesthesia on learning and
memory.

References
1. Moller JT, Cluitmans P, Rasmussen LS, et al. Long-term postoperative cognitive dysfunction in the elderly ISPOCD1 study.
Lancet 1998;351:857 61.
2. Johnson T, Monk T, Rasmussen LS, et al. Postoperative cognitive dysfunction in middle-aged patients. Anesthesiology 2002;
96:13517.
3. Culley DJ, Yukhananov RY, Baxter MG, Crosby G. The memory
effects of general anesthesia persist for weeks in young and
aged rats. Anesth Analg 2003;96:1004 9.
4. Culley DJ, Baxter MG, Yukhananov RY, Crosby G. Long-term
impairment of acquisition of a spatial memory task following
isoflurane-nitrous oxide anesthesia in rats. Anesthesiology 2004;
100:309 14.
5. Culley DJ, Baxter MG, Crosby CA, et al. Impaired acquisition of
spatial memory two-weeks after isoflurane and isofluranenitrous oxide anesthesia in aged rats. Anesth Analg 2004;99:
13937.
6. Decker MW, Gallagher M. Scopolamine-disruption of radial
arm maze performance: modification by noradrenergic depletion. Brain Res 1987;417:59 69.
7. Baxter MG, Holland PC, Gallagher M. Disruption of decrements
in conditioned stimulus processing by selective removal of hippocampal cholinergic input. J Neurosci 1997;17:5230 6.
8. Borde N, Jaffard R, Beracochea D. Effects of chronic alcohol
consumption or Diazepam administration on item recognition
and temporal ordering in a spatial working memory task in
mice. Eur J Neurosci 1998;10:2380 7.
9. Komatsu H, Nogaya J, Anabuki D, et al. Memory facilitation by
posttraining exposure to halothane, enflurane, and isoflurane in
ddN mice. Anesth Analg 1993;76:609 12.
10. Komatsu H, Nogaya J, Kuratani N, et al. Repetitive post-training
exposure to enflurane modifies spatial memory in mice. Anesthesiology 1998;89:1184 90.
11. Walker MP, Stickgold R. Sleep-dependent learning and memory
consolidation. Neuron 2004;44:12133.
12. Pryor KO, Veselis RA, Reinsel RA, Feshchenko VA. Enhanced
visual memory effect for negative versus positive emotional
content is potentiated at sub-anaesthetic concentrations of thiopental. Br J Anaesth 2004;93:348 55.
13. McGaugh JL. Memory: a century of consolidation. Science 2000;
287:248 51.
14. McGaugh JL, Izquierdo I. The contribution of pharmacology to
research on the mechanisms of memory formation. Trends Pharmacol Sci 2000;21:208 10.
15. Lukoyanov NV, Paula-Barbosa MM. A single high dose of dizocilpine produces long-lasting impairment of the water maze
performance in adult rats. Neurosci Lett 2000;285:139 42.
16. Mondadori C, Weiskrantz L, Buerki H, et al. NMDA receptor
antagonists can enhance or impair learning performance in
animals. Exp Brain Res 1989;75:449 56.
17. McKinney RA, Luthi A, Bandtlow CE, et al. Selective glutamate
receptor antagonists can induce or prevent axonal sprouting in
rat hippocampal slice cultures. Proc Natl Acad Sci U S A 1999;
96:11631 6.
18. Franks NP, Lieb WR. Molecular and cellular mechanisms of
general anaesthesia. Nature 1994;367:60714.

TECHNOLOGY, COMPUTING,

AND

SIMULATION

SOCIETY

FOR

TECHNOLOGY

IN

ANESTHESIA

SECTION EDITOR
STEVEN J. BARKER

Modern Wireless Telecommunication Technologies and

Their Electromagnetic Compatibility with
Life-Supporting Equipment
Mats K. E. B. Wallin,

MD, MSc*,

Therese Marve,

MSc,

and Peter K. Hakansson,

MSc

*Department of Anesthesiology and Intensive Care, Karolinska Hospital, and Med. Lab. Sci. and Tech,
Karolinska Institute, Stockholm, Sweden

Hospitals rely on pagers and ordinary telephones to

reach staff members in emergency situations. New telecommunication technologies such as General Packet
Radio Service (GPRS), the third generation mobile
phone system Universal Mobile Telecommunications
System (UMTS), and Wireless Local Area Network
(WLAN) might be able to replace hospital pagers if they
are electromagnetically compatible with medical devices. In this study, we sought to determine if GPRS,
UMTS (Wideband Code Division Multiple AccessFrequency Division Duplex [WCDMA FDD]), and
WLAN (IEEE 802.11b) transmitted signals interfere
with life-supporting equipment in the intensive care
and operating room environment. According to United
States standard, ANSI C63.18 1997, laboratory tests

well-functioning communication system is essential in an emergency hospital. The first commercial pagers were used in St. Thomas Hospital in London, England, in the mid 1950s (1). Today,
most hospitals still rely on alphanumeric pagers and
ordinary telephones to reach staff members during an
emergency situation.
In general, daily use of mobile phones and other wireless devices had increased rapidly. In hospitals, however, restrictions on the use of wireless telecommunication products are common after reports in the late 1980s
and early 1990s of malfunctions in medical devices (MD)
because of electromagnetic interference from various
electronic equipment and cellular phones (2). The risk of
interference between mobile phones and MDs mainly

Supported, in part, from Ericsson AB and TeliaSonera Sverige.

Accepted for publication April 29, 2005.
Address correspondence and reprint requests to Mats Wallin,
MD, MSc, Department of Anesthesiology and Intensive Care, Karolinska Hospital, S-171 76 Stockholm, Sweden. Address e-mail to
mats.wallin@labtek.ki.se.
DOI: 10.1213/01.ANE.0000180216.83554.00
2005 by the International Anesthesia Research Society
0003-2999/05

were performed on 76 medical devices. In addition,

clinical tests during 11 operations and 100 h of intensive
care were performed. UMTS and WLAN signals caused
little interference. Devices using these technologies can
be used safely in critical care areas and during operations, but direct contact between medical devices and
wireless communication devices ought to be avoided.
In the case of GPRS, at a distance of 50 cm, it caused an
older infusion pump to alarm and stop infusing; the
pump had to be reset. Also, 10 cases of interference with
device displays occurred. GPRS can be used safely at a
distance of 1 m. Terminals/cellular phones using these
technologies should be allowed without restriction in
public areas because the risk of interference is minimal.
(Anesth Analg 2005;101:13931400)

depends on transmission power, distance to the transmitter, and immunity (construction) of the MD, but signal characteristics, such as frequency and modulation
(pulsing), are also important (3).
Today, there are many different telecommunication systems for mobile phonesWide Area Networks (WANs)
around the world. The predominant second-generation
digital telecommunications system used for mobile phones
is Global System for Mobile communication (GSM). It is
used in the United States, Europe, parts of Asia, and other
countries. GSM uses a variant of the Time Division Multiple Access (TDMA) standard. These WAN technologies are
also suitable for telecommunication in larger local area
networks such as factories, offices, and hospitals.
Several studies have been performed testing medical equipment with signals from analog and GSM
signals (4 6). Interference tests have also been performed with mobile phones based on other telecommunication technologies (6,7) such as Code Division
Multiple Access (CDMA) and other TDMA standards,
mostly used only in the United States. For a more
accurate and detailed description of all these different
telecommunication technologies for mobile phones,
we suggest the appendix in the article in Health Physics
Anesth Analg 2005;101:13931400

1393

1394

TECHNOLOGY, COMPUTING, AND SIMULATION WALLIN ET AL.

TELECOMMUNICATION AND LIFE-SUPPORTING EQUIPMENT

by Morrissey (8). The safety distances recommended

in some articles vary; however, electromagnetic interference generally occurs only when the cellular
phones are in proximity to the MDs (3).
New technologies keep entering the market. Wireless Local Area Networks (WLANs), primarily designed for data communication and often used in laptops and some handheld pocket personal computers,
are now partly implemented in a few hospitals, although only a few studies concerning electromagnetic
compatibility have been published (9,10). In addition,
terminals with phone function using WLAN (802.11.b)
technology have recently been released on the market.
No compatibility studies on MDs have been published
in English-language journals for General Packet Radio
Service (GPRS)the data communication protocol for
GSMwhich can be implemented in the existing GSM
network, and the third-generation mobile phones system in Europe Universal Mobile Telecommunications
System (UMTS) using the Wideband Code Division
Multiple Access (WCDMA) technology with capability to send videophone calls and data.
If these new wireless communication systems can
operate in an environment with several MDs without
any electromagnetic interference, it will be possible to
replace the pager systems and to introduce new mobile services in hospitals.
The aim of this study was to investigate whether
electromagnetic interference with the MDs in an intensive care unit (ICU) and operating room (OR) could
be noticed when transmitting GPRS, WCDMA, and
WLAN signals.

Methods
This study was performed in the OR and ICU at the
Karolinska Hospital in Stockholm, Sweden. The study
protocol was approved by our local institutional ethics
committee. Provocative laboratory tests were made on
76 medical apparatuses (Table 1). In addition, clinical
tests were performed during 11 operations and almost
100 h in the ICU. About one-third of the equipment
was tested in the clinical setting; the ones tested depended on the type of operation and intensive care
given in the hospital during the weeks of the tests.
The test procedure was based on the American standard ANSI C63.18 1997 (11), which provides health
care organizations with a reproducible test method for
evaluating the electromagnetic immunity of existing
MDs to different radio-frequency (RF) transmitters.
The recommended practice applies to most MDs and
all portable RF transmitters with output power levels
up to 8 W. The standard describes how to evaluate the
performance of a MD when being exposed to RF
fields. The deviations from normal performance, i.e.,
alarms, error messages, or distortions of displayed
waveforms, are divided into 20 different categories.

ANESTH ANALG
2005;101:13931400

The signals used in our tests were WCDMA-Frequency

Division Duplex, GSM/GPRS (1800 MHz) transmitting at 4
time slots, and WLAN (IEEE 802.11b). The signal characteristics and output power levels can be found in Table 2.
The WCDMA and GPRS signals were simulated
using base band and signal generators (AMIQ and
SMHU 58; Rohde & Schwarz, GmbH, Munich, Germany) connected to a dipole antenna (738 454, 1710
2170 MHz, gain 2 dBi; Kathrein, Rosenheim, Germany) via a RF amplifier (LS Elektronik AB, Spanga,
Sweden) and a 5-m-long cable (Fig. 2). Using this
setup, it was easier to control the signal modulation,
frequency, and output power compared with using
commercial products. Besides, at the time for the
study, no commercial phones were available using the
specified characteristics. Before the tests, the output
power levels for the GPRS and WCDMA signals were
verified using a Rohde & Schwarz power meter.
The WLAN signals were generated by commercial
Personal Computer Memory Card International Association (PCMCIA) 802.11b network adapters (Symbol
Spectrum 24, Holtsville, NY). During the WLAN
tests, a large file was sent, via a peer-to-peer connection, between two laptop computers. All signals were
verified with a spectrum analyzer (Tektronix 7601,
Bracknell, UK) during all laboratory measurements.
Moreover, before and after every laboratory measurement session, the background electric field
strength was measured using a broadband electromagnetic field analyzer (EMR-20; Wandel & Goltermann, Research Triangle Park, NC) to ensure that no
strong fields from other sources were present.

Laboratory Tests
The laboratory tests gave an opportunity to test the
interference of the MDs in their normal environment
without exposing patients to potential danger. Only
after ensuring that none of the signals interfered with
the tested MDs in a critical way, clinical tests during
surgery and intensive care were then performed. The
76 different MDs (Table 1) included in the tests were
all equipment used in the central ICU, the dialyzing
unit, and ORs at the Karolinska Hospital. The MD
under test was in operating mode and centered in an
empty room, either on its own stand or, when possible, on a nonconducting table.
The transmitting antenna or laptop computer was
moved in a circle around the MD. The initial radius
of the circle was 2 m. If no interference appeared,
the radius was decreased and the test repeated. The
same procedure was performed with smaller radii
until the transmitting antenna was as close as possible to the MD, which exposed the MD to the
highest electric field strength levels because the

ANESTH ANALG
2005;101:13931400

TECHNOLOGY, COMPUTING, AND SIMULATION

WALLIN ET AL.
TELECOMMUNICATION AND LIFE-SUPPORTING EQUIPMENT

1395

Table 1. Medical Devices Tested for Interference with UMTS (WCDMA, FDD), WLAN (802.11b), and GPRS in the Study
Type of equipment
Anesthesia machine

Blood warmer
Defibrillator
Dialyzer
Electrosurgery
Perfusion system
Infusion pump

Laparoscopic equipment
Light source
Operation stand
Monitor

Producer/name
Siemens SC 900 XL (KION)
Datex ADU
Drager Julian plus
Dameca Dream
Dameca MCN 590 and 10590
Anmedic JB1
Biegler BW 585
HP CodeMaster WL
Gambro AK100
Hospal Prisma CFM
Fresenius 4008 E
Valleylab Force FX
Valleylab Force 10
Stockert S3 (ECMO)
Stockert CAPS 28-60-00 (ECMO)
Stockert CAPS
Alaris P6002
Baxter Flo-Gard 6201
Baxter Colleague
Sandoz Compat nr 199235
IVAC 591
Braun
IMED 960
IVAC 7101 EX
Braun Infusomat fmS
IVAC 531
IVAC 571
Sabratek 3030 Plus
Storz Thermoflator
Storz Xenon 300
Luxtec 9100 xenon
Transferis ALM
Datex GM 13 Ultima
Agilent M3046A
Datex AS/3 (Nokia CRT screen)
Datex AS/3 (Hitachi CRT
screen)
Datex Cardiocap 5
Datex S/5
HP 78361A (EKG)

Purchasing year
Manufactured
Manufactured
2001
2001
1993
1998
2000
1995
Manufactured
Manufactured
Manufactured
2000
1993
1994
1998
1993
1994
1994
Manufactured

Comments

1999
1998

X
X
X
X

1995
1994
1994

X
X
X

X
2001

1991

Alarmed and stopped,

GPRS 50 cm

2001
1982
1998
2001

1994
2000
2001
2001
2000

X
X
X

1990
Manufactured 1997
1998
Interference on screen,
GPRS 30 cm
1993
2001
Manufactured 2001

HP M1094A/Model 68S

1991

Datex PM8 Cardiocap II

Hellige Servomed SMS 151

1991

Siemens Sirecust 732 (EKG)

1993

Colin Press Mate 8800 (NIBP)

Marquette Solar 8000 (m. CRT)

1996
1999

Medair Life Sense

Protocol Propaq 104

2001
1998

Protocol Propaq 106 EL

Datex Light (F-LMP141-00)
Datex Satlite Trans (SpO2)

1993
1998
1993

Clinically
tested

X
X
X

Distortion of presented
ECG curve, GPRS 10 cm
Interference on screen,
GPRS 5 cm
X
Distortion of presented
ECG curve, GPRS 15 cm
Distortion of presented
ECG curve, GPRS 10 cm
Interference on screen,
GPRS 5 cm
Distortion of presented
ECG curve, GPRS 10 cm
X

1396

TECHNOLOGY, COMPUTING, AND SIMULATION WALLIN ET AL.

TELECOMMUNICATION AND LIFE-SUPPORTING EQUIPMENT

ANESTH ANALG
2005;101:13931400

Table 1. Continued
Type of equipment

Patient heater
X-ray
Telemetry
Tourniquet
TV-monitor

Ultrasonic doppler

Ultrasound
Ultrasound equipment
Vacuum suction apparatus
Ventilator

Video equipment
Incubator

Producer/name

Purchasing year

Datex TuffSat (SpO2)

Ohmeda 3800 (SpO2)
Mallinckrodt Warm touch 5800
ZIEHM Exposcop CB7-D
Swemac Biplanar 400E
GE Light speed
GE Apex pro
Zimmer ATS 2000
Panasonic M2080
Sony PVM-1443MD
Sony PVM-2043MD
Sony PVM-1453MD
Parks Flow detector 811-b
Huntleigh MiniDOPPLEX D900

2001
Manufactured 2001
1998

Aloka SSD-1400
Aloka Echo Camera SSD 680
USSC Autosonics Generator
Egnell Surgical Suction Pump
Drager Evita 4
Siemens Servo Ventilator 300
Respironics BIPAP Vision ST
Siemens Servo 900 C
Olympus CLV-U40, light source
Storz Tricam 20 222 020
Storz telecam SL PAL 202120 20
Drager Babyterm 8010

Comments

Clinically
tested

X
Manufactured 1998
Manufactured 1999
1998
1996
1991
1992
1997
1990
1997

X
X
Interference noise,
GPRS/UMTS/WLAN
100/1/5 cm,
Interference noise,
GPRS/UMTS/WLAN
400/25/50 cm, not
CE-marked
Interference noise,
GPRS 10 cm

1998
1992
2001
1994
2001
1988
1998
2001
1997
1998

X
X
X
X
X
X

UMTS Universal Mobile Telecommunications System; WCDMA FDD Wideband Code Division Multiple Access, Frequency Division Duplex; GPRS
General Packet Radio Service; WLAN Wireless Local Area Network.

electromagnetic field is strongest close to the antenna and then rapidly decreases with distance. We
also ensured that the transmitting antenna or
WLAN adapter was as close as possible to sockets
and cable ports because the MDs were expected to
be especially sensitive to electromagnetic fields at
these points. The flowchart in Figure 1 describes the
test procedure. If interference occurred, the transmitting source was moved outwards until the interference stopped. To be considered as an established
interference effect from the test signal, the distance
to the antenna and the type of interference were to
be the same in three consecutive measurements.

Clinical Tests in the Operation Unit

For tests during surgery, we chose operations where
as many MDs as possible were used. The tested signals were alternated and transmitted in periods of
1520 min during the entire operation, sometimes up
to 4 h. During surgery, it was not possible to move the

transmitting antenna in circles around the MDs as in the

laboratory tests. Instead, the transmitting antenna, or
laptop computer, was moved around in the OR to cover
as many positions as possible in distances as close as
possible to the MDs.

Clinical Tests in the ICU

The tests in the ICU were conducted similarly to the tests
during surgery. The transmitting antenna and the laptop
computer were placed at various locations in the ICU
during the course of a day. The tested signals were
alternated. In every room, there were between one and
three patients, and all patients had several different MDs
connected to them. In a normal situation, 3 surveillance
monitors, 10 infusion pumps, 2 ventilators, and 1 dialysis
machine could be operating simultaneously in each ICU
room. Apart from this equipment, diagnostic devices
such as ultrasound and radiograph were used during
the course of a day.

ANESTH ANALG
2005;101:13931400

TECHNOLOGY, COMPUTING, AND SIMULATION

WALLIN ET AL.
TELECOMMUNICATION AND LIFE-SUPPORTING EQUIPMENT

1397

Figure 1. Flowchart of the test procedure.

Results
Laboratory Tests
The laboratory tests showed that 65 of the 76 MDs
(85%) were immune to all signals tested, and only
2 devices showed interference effects caused by the
WCDMA and WLAN sources.
For WCDMA and WLAN, interference noise was
only detected from the loudspeakers of the two handheld diagnostic ultrasonic Dopplers tested (Table 1).
For WCDMA, the distances between the transmitting
source and the Dopplers were 1 and 25 cm, respectively, and for WLAN, the distances were 5 and 50 cm,

respectively. These interferences were not considered

critical because the Dopplers are diagnostic equipment, and the noise cannot be misinterpreted.
For GPRS, one of the affected MDs malfunctioned
in a critical way. At a distance of 50 cm to the
antenna, an older infusion pump alarmed, stopped,
and had to be reset. The other 10 cases of interference (Table 1) were display distortions and interference noise. The noted interference deviations were
interference pattern on three cathode ray tube
screens, distortion of electrocardiogram curves on
four older patient monitors, and characteristic interference noise from the three ultrasonic devices. The

1398

TECHNOLOGY, COMPUTING, AND SIMULATION WALLIN ET AL.

TELECOMMUNICATION AND LIFE-SUPPORTING EQUIPMENT

ANESTH ANALG
2005;101:13931400

Table 2. Technical Characteristics of the Signals in the Study

Technology

UMTS
(WCDMA FDD)

Frequency
Output power

1952.6 MHz
250 mW

Transmission

5-MHz
bandwidth

Other information

No power control
simulated

WLAN
(IEEE 801.11b)

GPRS
1802.1 MHz
1 W maximum
500-mW average
Pulse repetition frequency 217 Hz
200-kHz bandwidth
Transmission on four time slot in a row
No power control simulated

2.42.5 GHz
100 mW
Direct sequence spread spectrum
modulation
Maximum transmission rate 11 Mbit/s
Ad-hoc network

UMTS Universal Mobile Telecommunications System; WCDMA FDD Wideband Code Division Multiple Access, Frequency Division Duplex; GPRS
General Packet Radio Service; WLAN Wireless Local Area Network.

Figure 2. The test set up.

distances between the GPRS transmitting antenna

and the affected MDs varied from 5 to 50 cm except
for the ultrasonic Dopplers, where the interference
distances were as long as 1 and 4 m.

Clinical Tests
In the clinical tests, no new types of interference were
noticed. During surgery, the only interference noticed
was distortion of the curves on a monitor when the
GPRS signal transmitting antenna was very close.
During intensive care, there was no interference registered for any signal.

Discussion
The most important finding in our study is the minimal interference seen from WCDMA and WLAN
(standard 802.11b) signals in these sensitive hospital
environments (OR and ICU). Only two diagnostic ultrasonic Dopplers were affected, although we used
signals with the highest possible average power levels
to create worst-case scenarios. The WCDMA signal
transmitted 250 mW, which mostly is intended for
data traffic and is twice the power level of most WCDMA terminals. The GPRS signal in our tests used the
maximal number of time slots (4), which give an average power level of 0.5 W. Most GPRS phones transmit at 1 or 2 time slots, which gives lower average

output power levels. Complementary tests were for

made on 4 MDs, using GPRS with only 1 active time
slot (pulse frequency, 217 Hz; average power, 0.125
W). The distances when interference occurred decreased between 30% and 80% compared with the
4-time slot signal.
Both UMTS and GSM/GPRS have a feature so the
mobile terminals can reduce their output power to the
minimum the level required to reach the closest radio
base station. In an office building with a well designed
in-building network and with a distributed multiantenna system, the GSM mobile terminals average
output power levels were reduced to approximately
4% (15 mW) of the maximum level (12). This power
reduction depends on the network coverage and signal conditions that are affected by motion, but because
the mobility inside a hospital is limited, this power
fluctuation was not considered in this study. Similar
reductions can be expected for GPRS and WCDMA
signals. Consequently, the risk for electromagnetic interference in a real network will be less than reported
in this study.
In 1993, the International Electrotechnical Commission (IEC) adopted an electromagnetic compatibility
(EMC) standard for MDs, IEC 6060112, (13). It was
specified for RFs up to 1 GHz, and not until September
2001 was a new revision published that extended to
cover frequencies up to 2.5 GHz (14). Hence, many of
the electronic MDs used in hospitals today were manufactured before 2001 and have therefore not been
tested for the higher frequencies used by, for instance,
GSM/GPRS 1800 MHz, WLAN, and WCDMA. All
MDs tested in this study were purchased or manufactured before 2001, and the old infusion pump, which
was the only MD that malfunctioned critically, was
manufactured even before the first publication of the
IEC standard. The current EMC standard for MDs
specifies a 10 V/m immunity level for life-supporting
equipment and a 3 vol/m immunity level for nonlifesupporting equipment. The electric field strength decreases exponentially with distance. Figure 3 presents

ANESTH ANALG
2005;101:13931400

TECHNOLOGY, COMPUTING, AND SIMULATION

WALLIN ET AL.
TELECOMMUNICATION AND LIFE-SUPPORTING EQUIPMENT

Figure 3. The calculated (unbroken line) and measured electric field

strength levels (points) at different distances from the dipole antenna used (1952.3 MHz; 250 mW). The immunity levels for medical
devices (MDs; 10 and 3 vol/m) are the 2 horizontal lines. The dotted
line is an example from a well designed in-building network where
the maximal output power of the mobile terminals is reduced to 15
mW. The field strength levels were measured in the electromagnetic
field research laboratory at Ericsson in Stockholm, Sweden. WCDMA Wideband Code Division Multiple Access.

the calculated (unbroken line) and measured electric

field strength levels (points) at different distances
from the dipole antenna used (1952.3 MHz; 250 mW).
The immunity levels for MDs are the two horizontal
lines. As can be seen, the field strength is less than
both immunity levels at distances more than 1 m,
which means that as long as the MDs are in compliance with the IEC standard, there should be no risk of
interference. The dotted line in Figure 3 is shown as an
example from an in-building network where the terminals maximal output power is reduced to 15 mW.
At this power level, the critical distance for lifesupporting equipment is 20 cm.
Regarding the curves in Figure 3, it was surprising
that an ultrasound Doppler from 1997 gave interference noises at a distance as far as 4 m from the antenna. However, this device was not CE-marked,
which means that it need not have been in compliance
with the IEC standard.
Otherwise, all life-supporting equipment tested was
electromagnetic compatible with WCDMA and
WLAN, even with signal sources in proximity. Consequently, both theoretical data and the present study
indicate that WCDMA and WLAN have the capability
to replace old-fashioned pager systems. However, this
study cannot conclude that interference between the
tested signals and MDs can never occur. Thus, we recommend that all hospitals conduct tests according to the
American ANSI C63.18 1997 standard (11) before implementing new telecommunication technologies.
The fact that well-functioning communication systems are of highest importance for a hospital was
revealed in an official Australian survey of hospital

1399

admissions in the mid 1990s (15). Communication

problems, in a context more inclusive than just technical communication devices, were the most common
single reason responsible for (11%) preventable disabilities and deaths and were nearly twice as common
as those caused by inadequate medical skill (6%). The
numbers of deaths caused by communication problems were approximately 1500 per year. Even if modern telecommunication technologies could contribute
to decreasing this figure by only a small percent, many
lives can be saved. We have not found any published
reports about patients who have died because of an
incidence caused by electromagnetic interference in an
ICU or OR. It is thus time to put the risk for electromagnetic interference into a proper perspective and
instead use the many possibilities offered by new telecommunication technologies. New services could be
introduced in hospitals, e.g., constant access to telephones, smart alarm systems, radiograph pictures directly to the doctor via cellular phones, and video
consultations with paramedics at an accident. According to one report (16), there could also be substantial
economic advantages. In this report it is suggested
that the American health system could save $30 billion
per year with improved telecommunications. These
figures may be optimistic, but money could be saved
if a modern platform for mobile communication were
implemented in emergency hospitals. However, the
telecommunication companies have to develop terminals designed for hospital applications containing just
those functionalities important for hospital use.
Compared with WLAN, UMTS (WCDMA) has the
advantage of being designed primarily for both
speech and text communication. To receive alarms as
text-message and then be able to call back immediately with the same device will be an important improvement (17). It has a larger data transmission capacity, causes less interference than GSM/GPRS, and
is, thereby, suitable to be the technology that replaces
the hospital pager systems.
We conclude that these new wireless technologies
can be used in most sensitive parts of hospital environments. WCDMA and WLAN (802.11b, 2.4 GHz)
cause minimal interference and can be used in ICUs
and ORs. Direct contact between MDs and terminals
ought to be avoided. For GPRS (1800 MHz), it is
recommended that the hospital staff keep a distance of
1 m between the terminals and the MDs. However,
medical staff must always be aware of the fact that
interference can occur and should know how to recognize it. Hospitals are recommended to have a management protocol determining how to test their MDs
regarding EMC and, if necessary, replace old equipment with low immunity. In public areas, such as
cafeterias, corridors, and waiting areas, UMTS (WCDMA), WLAN, and GPRS (1800 MHz) mobile

1400

TECHNOLOGY, COMPUTING, AND SIMULATION WALLIN ET AL.

TELECOMMUNICATION AND LIFE-SUPPORTING EQUIPMENT

phones/terminals should be allowed without restrictions because the risk for interference is minimal.
This study was conducted with technical and financial support from
Ericsson AB and TeliaSonera Sverige.

References
1. Available at: http://www.wirelessmessaging.org/media/
info2.0.html. Accessed September 16, 2005.
2. Silberberg JL. Performance degradation of electronic medical
devices due to electromagnetic interference. Compliance Engineering 1993;10:1 8.
3. Klein AA, Djaiani GN. Mobile phones in the hospitals: past,
present and future. Anaesthesia 2003;58:3537.
4. Hietanen M, Sibakov V, Hallfors S, von Nadelstahl P. Safe use of
mobile phones in hospitals. Health Phys 2000;79:S77 84.
5. Barbaro V, Bartolini P, Baenassi M, et al. Electromagnetic interference by GSM cellular phones and UHF radios with intensive
care and operating room ventilators. Biomed Instrum Technol
2000;34:3619.
6. Morrisey JJ, Swicord M, Balanzo Q. Characterization of electromagnetic interference of medical devices due to cell phones.
Health Phys 2002;82:4551.
7. Baba I, Furuhata H, Kano T, et al. Experimental study of electromagnetic interference from cellular phones with electronic
medical equipment. J Clin Engl 1998;23:12234.
8. Morrissey JJ. Mobile phones in the hospital: improved mobile
communication and mitigation of EMI concerns can lead to an
overall benefit to healthcare. Health Phys 2004;87:82 8.

ANESTH ANALG
2005;101:13931400

9. Rice WP. 2.4 GHz WLAN EMI in medical devices. J Clin Engl
2000;25:260 4.
10. Tan KS, Hinberg I. Effects of wireless local area network (LAN)
system, a telemetry system, and electrosurgical devices on medical devices in a hospital environment. Biomed Instrum Technol
2000;34:115 8.
11. Institute of Electrical and Electronics Engineers, Inc. American
national standard recommended practice for on-site ad hoc test
method for estimating radiated electromagnetic immunity of
medical devices to specific radio-frequency transmitters. (standard C63.18). Piscataway, NJ: IEEE, 1997.
12. Persson T, Tornevik C, Larsson LE, Loven J. GSM mobile phone
output power distribution by network analysis of all calls in
some urban, rural and in-office networks, complemented by test
phone measurements. Bioelectromagnetics (BEMS) Abstract
book (p 182) from the Twenty-Fourth Annual Meeting, Quebec
City, Quebec, Canada, June 2327, 2002.
13. IEC 60612, collateral standard: electromagnetic compatibility
requirements and tests. 1st ed. Geneva: International Electrotechnical
Commission, 1993.
14. IEC 60612, collateral standard: electromagnetic compatibility
requirements and tests. 2nd ed. Geneva: International Electrotechnical
Commission, 2001.
15. Zinn C. 14,000 preventable deaths in Australian hospitals. BMJ
1995;310:1487.
16. Little AD. Telecommunications: Can it help solve Americas
health care problem? Cambridge, MA: Arthur D. Little, 1992.
17. Halkes M. A response to Mobile phones in the hospital: past,
present and future, Klein AA, Djaiani GN, Anaesthesia 2003;
58:3537. Anaesthesia 2003;58:1147.

Denaturing High Performance Liquid Chromatography

Screening of Ryanodine Receptor Type 1 Gene in Patients
with Malignant Hyperthermia in Taiwan and Identification of
a Novel Mutation (Y522C)
Huei-Ming Yeh, MD*, Mei-Chuan Tsai, BS, Yi-Ning Su, MD, Rong-Ching Shen,
Jeuy-Jen Hwang, MD, Wei-Zen Sun, MD*, and Ling-Ping Lai, MD, PhD
Departments of *Anesthesiology, Medical Genetics, and Internal Medicine, National Taiwan University Hospital; and
Institute of Pharmacology, National Taiwan University, Taipei, Taiwan

We performed the present study to identify the mutation in patients in Taiwan with malignant hyperthermia
(MH). We also test the hypothesis that a denaturing
high-performance liquid chromatography (DHPLC)
protocol can be used for mutation detection in these patients. We identified five Taiwanese patients with typical clinical presentations of MH after general anesthesia. We also enrolled 50 healthy volunteers. Polymerase
chain reaction was used to amplify the ryanodine receptor (RYR1) gene mutation hot spots and DHPLC
techniques were used to screen for mutations. Upon detection of a heterozygous elution pattern in DHPLC
analysis, DNA sequencing reaction was performed to

he present study was performed to identify the

mutation in patients in Taiwan with malignant
hyperthermia (MH). We also tested the hypothesis that a denaturing high-performance liquid chromatography (DHPLC) protocol can be used for mutation detection in these patients.
MH is a life-threatening skeletal muscle disorder
that is triggered by depolarizing muscle relaxant
and/or inhaled anesthetics in susceptible individuals
(1,2). It is an autosomal-dominantly inherited, genetically heterogeneous myopathy (3). Its main symptoms
include tachycardia, supraventricular or ventricular
arrhythmia, tachypnea, muscle rigidity, hypercarbia,
hyperthermia, rhabdomyolysis, and renal failure. The
This work was supported in part by Grant 93S74 from the National Taiwan University Hospital, Taipei, Taiwan.
Accepted for publication April 28, 2005.
Address correspondence and reprint requests to Dr. Ling-Ping
Lai, No. 1, Jen-Ai Rd., Section 1, Institute of Pharmacology, National
Taiwan University, Taipei, Taiwan, 100. Address e-mail to
lai@ha.mc.ntu.edu.tw.
DOI: 10.1213/01.ANE.0000180214.74580.39
2005 by the International Anesthesia Research Society
0003-2999/05

identify the nucleotide variations. We identified a RYR1

mutation in all 5 patients with MH. There were 4 different mutations in the 5 patients: Tyr522Cys, Arg552Trp,
Val2168Met, and Thr2206Arg. Among the 5 patients, 2
unrelated patients had the same Thr2206Arg mutation.
Three of the mutations had been reported before, but
the Tyr522Cys mutation was novel. None of the MHrelated mutations were found in the control group. In
conclusion, we identified RYR1 mutations in 5 Taiwanese patients with MH using a DHPLC-based approach.
A DHPLC-based genetic test may be developed as a
noninvasive and convenient test for MH.
(Anesth Analg 2005;101:14016)

pathology of MH is an increase in the concentration of

cytosolic calcium which is released from the sarcoplasmic reticulum via the ryanodine receptor type 1
(RYR1) (4). The increase of cytosolic calcium stimulates contracture and metabolic responses that finally
create acid/base and electrolyte imbalances, causing
cell damage and death.
In about 50% of the MH families, this disease is
linked to the MHS-1 locus, which encodes the human
skeletal muscle RYR1 (5,6). More than 40 RYR1 mutations have been described to associate with MH (7).
Most RYR1 mutations appear to be clustered in 23
mutation hot regions, the N-terminal amino acid residues 35 614 region, the central amino acid residues
21622458 region, and the C-terminal amino acid residues 4000 4898 region (less frequent).
In vitro contracture test (IVCT) is a common way to
detect MH-susceptible patients (8,9). However, IVCT
is invasive, laborious and should be performed in
specialized laboratories. With the advent of modern
molecular biology techniques, screening for mutations
in candidate genes is a possible alternative to identify
Anesth Analg 2005;101:14016

1401

1402

TECHNOLOGY, COMPUTING, AND SIMULATION

DENATUTING HPLC SCREENING FOR RYR1

YEH ET AL.

MH-susceptible patients. Mutation detection in the

RYR1 gene can be preformed using direct DNA sequencing techniques. However, the RYR1 gene is one
of the largest genes in the human genome. Direct
DNA sequencing is laborious. In the present study, we
designed a DHPLC protocol to screen the DNA samples for sequence variations before sequencing. This
method is rapid, inexpensive, and is high throughput.
We also described the detailed protocol for DHPLC
screening of the RYR1 mutation hot regions. We also
reported the mutations, including a novel one, in five
Taiwanese patients with MH.

ANESTH ANALG
2005;101:14016

Table 1. Base Sequence of Primer Pairs and Conditions of

Polymerase Chain Reactions
Exon

Primer

Ta (C)

CAGAGAGTTGGTGGCAGTGA
GGAAATGGGAAATTGGGAGT
TTTCCCCTGCTGACAGTGAT
TCAGGCTGCTCTGATTTTCC
GCACTCTGCAGTCCCTCAG
GCTGGAGTACAGTGGCATGA
ATGACGGGGGACAGAAAAAT
GAGGCCAAGTGTATGGATGC
GAAACCAAAGGGGCTGATTT
GCCACGGATTAGAGAAGCTG
GCCGAGTCCTGGAAAGAGAT
GTACAGGACCTCCAGGATGC
TCCACTTTCACCACCTCCTC
GGGGTGGGAGATTAAGGAAA
TCCACATTGTTCTGGTCCAA
GGTGAGCTAGCCATCGAGAG
GGTGGTGCTAGGCACAGAGT
CGGCAGAAATAGCAGAGGAA
CGCGAGCAGTGAGTCTCC3
TTGTGTCCCCAACATTGCTA
GAAAGAGGCCTGCTCTACCC
CACGAAGGATGCTGACATCT

50

9
11
12
14
15

Methods

17

We included 5 patients with MH and 50 volunteers.

All the patients were identified after general anesthesia. All of them had typical clinical presentations of
MH including high fever, marked increase of muscle
enzymes, renal failure, and hyperkalemia. Dantrolene
was used in all patients. One of the patients died
despite dantrolene therapy and supportive treatment.
None of the patients underwent IVCT because the
diagnosis was well established. This study was approved by the IRB. Peripheral blood samples were
collected after informed consent. Genomic DNA was
extracted from leukocytes using a modified proteinase
K method (Qiagen, Valencia, CA).
Primer pairs were designed to amplify the mutation
hot regions of the RYR1 gene. Table 1 shows the exons
selected, the primer sequences, and the PCR annealing
temperature. PCR was performed in thin-walled PCR
tubes in a total volume of 25 L containing 100 ng of
genomic DNA, 0.12 M of each primer, 100 M deoxynucleotide triphosphates, 0.5 U of AmpliTaq Gold
enzyme (PE Applied Biosystems, Foster City, CA),
and 2.5 L of GeneAmp 10 buffer II (10 mM trisHCl, pH 8.3, 50 mM KCl), in 2 mM MgCl2 as
provided by the manufacturer. Amplification was performed in a multiblock system thermocycler (ThermoHybaid, Ashford, UK). PCR amplifications were performed with an initial denaturation step at 95C for
10 min, followed by 35 cycles consisting of denaturation at 94C for 30 s, annealing for 30 s, and elongation at 72C for 1 min. The annealing temperatures are
shown in Table 1. A final extension step was applied
at 72C for 10 min after 35 cycles.
The DHPLC system used for detecting heteroduplexes was a Transgenomic Wave Nucleic Acid Fragment Analysis System (Transgenomic Inc., San Jose,
CA). DHPLC was performed on automated HPLC
instrumentation equipped with a DNASep column
(Transgenomic Inc. San Jose, CA). DHPLC-grade acetonitrile (9017-03; JT Baker, Phillipsburg, NJ) and triethylammonium acetate (TEAA) (Transgenomic Inc.,
Crewe, UK) were used for the mobile phases, which

39
40
45
46

50
56
50
50
50
50
50
50
50
50

Ta annealing temperature.

comprised 0.05% acetonitrile in 0.1 M TEAA (eluent

A) and 25% acetonitrile in 0.1 M TEAA (eluent B). For
heteroduplex detection, crude PCR products, subjected to an additional 3-min, 95C denaturing step,
followed by gradual reannealing from 95 to 65C
over a period of 30 min before analysis, were eluted at
a flow rate of 0.9 mL/min. The start- and end-points of
the gradient by mixing eluents A and B and the temperature required for successful resolution of heteroduplex molecules were adjusted by using an algorithm provided by the WAVEmaker system control
software, version 4.1.42 (Transgenomic Inc., San Jose,
CA). Eight microliters of PCR product was injected for
analysis in each running. Individual analytical gradient conditions for DHPLC running are described in
Table 2 and are expressed as a percentage of eluent B.
The flow rate was 0.9 mL/min, and the ultraviolet
detector was set to 260 nm. Heterozygous profiles
were identified by visual inspection of the chromatograms on the basis of the appearance of additional
earlier eluting peaks. Corresponding homozygous
profiles showed only one peak.
Amplicons were purified by solid-phase extraction
and were bidirectionally sequenced with the PE Biosystems Taq DyeDeoxy terminator cycle sequencing
kit (PE Biosystems) according to the manufacturers
instructions. Sequencing reactions were separated on
a PE Biosystems 373A/3100 sequencer. The sequencing results were compared with RYR1 mRNA (GenBank accession number NM000540). The adenosine of

ANESTH ANALG
2005;101:14016

TECHNOLOGY, COMPUTING, AND SIMULATION

YEH ET AL.
DENATUTING HPLC SCREENING FOR RYR1

1403

Table 2. The Conditions for Denaturing High-Performance Liquid Chromatography

First condition

Second condition

Exon

Size (bp)

Temperature (C)

B%

6
9
11
12
14
15
17
39
40
45
46

372
367
348
351
372
328
388
488
368
338
364

62.5
62.5
62
62
61
61
60
63
63
64
63.5

5766
5665
5766
5665
5665
5463
5766
5665
5665
5665
5665

Temperature (C)

B%

63.5
64

5564
5463

63.5
61
64.5
64.5

5160
5665
5463
5564

B% eluent B percentage.

the start codon (ATG) was numbered as the first nucleotide when expressing genetic variations.

Results
PCR amplification for all selected exons was successfully performed as revealed by agarose gel electrophoresis. In DHPLC analysis, 2 melting domains were
predicted from the base sequence of exons 11, 12, 15,
17, 39, and 40. Two different conditions were applied to these exons to increase the sensitivity (Table
2). We identified 7 different heteroduplex patterns
in amplicons for exons 6, 12, 14, 15, 17, 39, and 40,
respectively. Figure 1 shows the DHPLC pictures
for these heteroduplex patterns. Further DNA sequencing reaction identified the nucleotide variations in these heteroduplexes.
The clinical information of the study patients is
shown in Table 3. All patients received succinylcholine and volatile anesthetics for general anesthesia. All
of them had fulminant clinical presentations of MH.
We identified the mutation in all 5 patients with MH
(Tables 3 and 4). There were 4 different mutations in 5
patients. All the mutations were mis-sense mutations
and none were identified in the control group. These
mutations were Tyr522Cys (A1695G in exon 14),
Arg552Trp (C1784T in exon 15), Val2168Met (G6632A
in exon 39), and Thr2206Arg (C6747T in exon 40). All
the mutated amino acids were in sequences highly conserved among several species including nonmammal
animals (Fig. 2). A comprehensive analysis of the evolutionary conserved sites in RYR1 is available at the following Web site: http://www-nbrf.georgetown.edu/
cgi-bin/pirwww/nbrfget?uidFA1808&dbA. Among
the 4 mutations, Tyr552Cys was novel and has not yet
been reported. Two unrelated patients had the same
Thr2206Arg mutation.

Figure 1. Heteroduplex elution patterns identified by denaturing

high-performance liquid chromatography (DHPLC) in the present
study. Heteroduplex patterns are characterized by the appearance
of early eluting peaks, whereas the homoduplex profile shows only
one peak. DNA sequencing results are shown below the DHPLC
heteroduplex patterns.

There were 3 heteroduplex patterns identified in the

control group (Table 4). Further DNA sequencing reaction identified the nucleotide variations. These variations were an intronic single nucleotide polymorphism (SNP) in intron5, a synonymous SNP in exon 12
and a nonsynonymous SNP in exon 17. The nonsynonymous SNP has been reported and is not related to
MH. None of the mutations identified in MH patients
were identified in the normal controls. There was no
heteroduplex pattern detected in exon 15, in which a
common polymorphism (G1668A) had been reported
by many other researchers. Further sequencing reaction showed that all the control patients were homozygous for 1668A.

Discussion
The RYR1 gene is
genome. It locates
prises 106 exons
15,000 base pairs

one of the largest genes in human

on chromosome 19q13.1, and comencoding an mRNA longer than
(1315). The RYR1 protein has a

1404

TECHNOLOGY, COMPUTING, AND SIMULATION

DENATUTING HPLC SCREENING FOR RYR1

YEH ET AL.

ANESTH ANALG
2005;101:14016

Table 3. RYR1 Mutations and Clinical Information of the Study Patients

Patient no.

Age (yr)

Sex

Operation

Mutation

First reported

1
2
3
4
5

32
13
65
35
28

F
M
M
M
F

Thyroidectomy
Herniorrhaphy
Vocal nodule excision
Thyroidectomy
Breast augmentation

Thr2206Arg
Arg552Trp
Val2168Met
Thr2206Arg
Tyr522Cys

Brandt et al. (10)

Keating et al. (11)
Manning et al. (12)
Brandt et al. (10)
This report

Table 4. Mutations and Variations of the RYR1 Genes Identified in the Present Study
Incidence
Exon

Nucleotide change

Disease-causing mutations
14
A1695G
15
C1784T
39
G6632A
40
C6747T
Polymorphisms not related to disease
6
G(55577)T
12
A1227G
17
G1964A

Amino acid change

Patients

Control

Tyr522Cys
Arg552Trp
Val2168Met
Thr2206Arg

1/5
1/5
1/5
2/5

0/50
0/50
0/50
0/50

Leu409Leu
Ala612Thr

0/5
0/5
0/5

2/50
1/50
1/50

Figure 2. Comparison of the amino acids of ryanodine receptor type 1 among vertebrate species. All the mutations in the present study occur
at amino acids, which are highly conserved among species.

subunit size of 565 kDa and forms a homotetrameric

structure that acts as a calcium release channel in the
skeletal muscle. Its N-terminal region forms a large
foot structure as a connector that bridges the gap
between the sarcoplasmic reticulum and the T-tubule.
Its C-terminal region forms the transmembrane channel. For mutation detection, a base-by-base screening
method is laborious. There have been 40 mutations
reported. Fortunately, these mutations clustered in
several mutation hot regions. We therefore designed
to screen the hot regions first. To further enhance the
efficacy of screening, DHPLC technique was applied.
We identified the mutations in all five patients with
MH. The causal relationship between the base variations (mutations) and MH was supported by the following findings. First, the base variations all result in
a change of amino acid, which is highly conserved
among many species (Fig. 2). Second, the base variations were not identified in 50 control subjects (100
alleles). Therefore, the variations are unlikely simple
SNPs. Third, the same mutation has been reported in
the literature regarding MH. This is true for three of
the mutations in the present study.

We identified a RYR1 mutation in all 5 patients.

Therefore, the incidence of RYR1 mutation was 100%
in the present study. In reports in Western countries,
the incidence of RYR1 mutation in patients with MH
was approximately 50%70% (16). Tammaro et al. (7)
also used DHPLC techniques for screening for RYR1
mutation in 52 Italian families. They reported that the
incidence of detecting RYR1 mutation was 33% (11 of
33) in MH-susceptible families. In the present study,
we reported a much more frequent incidence. There
are several possible explanations. First, there might be
ethnic differences, and Asian patients with MH have a
more frequent incidence of RYR1 mutation. If the
Asian patients had the same incidence, the chance of
detecting RYR1 mutation in all 5 patients would be
very small (1/25 1/32). Second, all the patients in
the present study had fulminant clinical MH. The
incidence of RYR1 mutation might be different between MH-susceptible subjects and patients with documented fulminant clinical disease. Third, different
methods for mutation detection were used in different
studies. We used DHPLC screening for mutation hot
regions. In the DHPLC protocol, 2 conditions were

ANESTH ANALG
2005;101:14016

used for 6 exons because 2 melting domains were

predicted.
A novel mutation was identified in the present
study. The index case was a 21-year-old woman who
underwent a breast augmentation surgery. Succinylcholine was used for induction of general anesthesia
with isoflurane. The Tyr522Cys mutation in the RYR1
gene was not found in the normal population. This
amino acid is highly conserved across species including nonmammal animals. A different mutation at the
same amino acid (Tyr522Ser) had been reported by
Quane et al. (17). Therefore, we believe that Tyr522Cys
is a disease-causing mutation. This mutation can be
added to the list of mutation screening, especially for
Asian patients.
There are screening methods other than DHPLC for
detecting genetic variations. Single-strand conformation polymorphism analysis was the most frequently
used method for mutation screening before DHPLC
was available. However, it is time consuming, labor
intensive, and has a sensitivity between 70%90% (18
20). In the present study, we identified the mutation
for all patients with MH. A sensitivity of 100% has
been demonstrated in previous reports when DHPLC
was used for mutation screening for many genes such
as BRCA1, BRCA2, Notch3, MepC2, TSC1, and the
entire mitochondria genome (2125). Furthermore, the
DHPLC elution patterns are specific for their underlying SNP or mutation. This further increases the
power of this screening method in differentiating between common SNPs and mutations.
IVCT is often used for the diagnosis of MH susceptibility (15). However, it is invasive, costly, and should
be performed in experienced laboratories. It has a
sensitivity and specificity of 99% and 94%, respectively. Since the advent of modern molecular biology
techniques and the discovery of responsible mutations
for MH, noninvasive genetic tests are developing and
guidelines for molecular genetic test have been published (26). However, the test is preliminary and has
the following limitations. First, this test examines mutations within a list of known mutations. Mutations
not on the list might have been missed by the test.
Second, as suggested by the North America Working
Group Meeting on Malignant Hyperthermia Genetic
Testing in 2004, genetic testing is indicated for highrisk individuals such as probands with MH and their
family members or subjects with positive IVCT (27).
Genetic testing is not suitable for a general survey of
patients before undergoing general anesthesia. Furthermore, if the result of a genetic test is negative,
IVCT will still be performed. The DHPLC screening
strategy has the potential of overcoming these problems. The DHPLC method can be used for the detection of novel mutations. It also has the potential for
use in the general population because it is noninvasive, inexpensive, rapid, and is high throughput. With

TECHNOLOGY, COMPUTING, AND SIMULATION

YEH ET AL.
DENATUTING HPLC SCREENING FOR RYR1

1405

the discovery of more MH-responsible mutations and

genes, these mutations and genes can be added to the
list of DHPLC screening. If all or most of the responsible genes in a population are elucidated, a DHPLCbased genetic test may be developed as a noninvasive
and convenient test for MH. Because the DHPLC test
is a high-throughput and inexpensive procedure, it is
possible to obtain peripheral blood samples and perform PCR and DHPLC screening before operation for
all patients who need general anesthesia.
In conclusion, we applied the DHPLC mutation
screening technique in patients with MH. We identified the responsible mutation in all five patients with
MH and reported a novel disease-causing mutation.
The DHPLC screening approach can be used for the
development of a more powerful genetic test for MH.

References
1. Denborough M, Lovell R. Anaesthetic deaths in a family. Lancet
1960;2:45.
2. Nelson TE, Flewellen EH. The malignant hyperthermia syndrome. N Engl J Med 1983;309:416 8.
3. Britt BA, Locher WG, Kalow W. Hereditary aspects of malignant
hyperthermia. Can Anaesth Soc J 1969;16:89 98.
4. Mickelson JR, Louis CF. Malignant hyperthermia: excitationcontraction coupling, Ca2 release channel, and cell Ca2 regulation defects. Physiol Rev 1996;76:53792.
5. McCarthy TV, Healy JM, Heffron JJ, et al. Localization of the
malignant hyperthermia susceptibility locus to human chromosome 19q12-13.2. Nature 1990;343:562 4.
6. MacLennan DH, Duff C, Zorzato F, et al. Ryanodine receptor
gene is a candidate for predisposition to malignant hyperthermia. Nature 1990;343:559 61.
7. Tammaro A, Bracco A, Cozzolino S, et al. Scanning for mutations of the ryanodine receptor (RYR1) gene by denaturing
HPLC: detection of three novel malignant hyperthermia alleles.
Clin Chem 2003;49:761 8.
8. Rosenberg H, Reed S. In vitro contracture tests for susceptibility
to malignant hyperthermia. Anesth Analg 1983;62:41520.
9. Ording H, Brancadoro V, Cozzolino S, et al. In vitro contracture
test for diagnosis of malignant hyperthermia following the protocol of the European MH Group: results of testing patients
surviving fulminant MH and unrelated low-risk subjects. The
European Malignant Hyperthermia Group. Acta Anaesthesiol
Scand 1997;41:955 66.
10. Brandt A, Schleithoff L, Jurkat-Rott K, et al. Screening of the
ryanodine receptor gene in 105 malignant hyperthermia
families: novel mutations and concordance with the in-vitro
contracture test. Hum Mol Genet 1999;8:2055 62.
11. Keating KE, Giblin L, Lynch PJ, et al. Detection of a novel
mutation in the ryanodine receptor gene in an Irish malignant
hyperthermia pedigree: correlation of the IVCT response with
the affected and unaffected haplotypes. J Med Genet 1997;34:
291 6.
12. Manning BM, Quane KA, Ording H, et al. Identification of novel
mutations in the ryanodine receptor gene (RYR1) in malignant
hyperthermia: genotype-phenotype correlation. Am J Hum
Genet 1998;62:599 609.
13. Takeshima H, Nishimura S, Matsumoto T, et al. Primary structure and expression from complementary DNA of skeletal muscle ryanodine receptor. Nature 1989;339:439 45.
14. Zorzato F, Fujii J, Otsu K, et al. Molecular cloning of cDNA
encoding human and rabbit forms of the Ca2 release channel
(ryanodine receptor) of skeletal muscle sarcoplasmic reticulum.
J Biol Chem 1990;265:2244 56.

1406

TECHNOLOGY, COMPUTING, AND SIMULATION

DENATUTING HPLC SCREENING FOR RYR1

YEH ET AL.

15. Phillips MS, Fujii J, Khanna VK, et al. The structural organization of the human skeletal muscle ryanodine receptor (RYR1)
gene. Genomics 1996;34:24 41.
16. Rueffert H, Olthoff D, Deutrich C, et al. Mutation screening in
the ryanodine receptor 1 gene (RYR1) in patients susceptible to
malignant hyperthermia who show definite IVCT results: identification of three novel mutations. Acta Anaesthesiol Scand
2002;46:692 8.
17. Quane KA, Keating KE, Healy JMS, et al. Mutation screening of
the RYR1 gene in malignant hyperthermia: detection of a novel
Tyr to Ser mutation in a pedigree with associated central cores.
Genomics 1994;23:236 9.
18. Orita M, Iwahana H, Kanazawa H, et al. Detection of polymorphisms of human DNA by gel electrophoresis as single-strand
conformation polymorphisms. Proc Natl Acad Sci USA 1989;86:
2766 70.
19. Ravnik-Glavac M, Glavac D, Dean M. Sensitivity of singlestrand conformation polymorphism and heteroduplex method
for mutation detection in the cystic fibrosis gene. Hum Mol
Genet 1994;3:8017.
20. Xiao W, Oefner PJ. Denaturing high-performance liquid
chromatography: a review. Hum Mutat 2001;17:439 74.
21. Wagner T, Stoppa-Lyonnet D, Fleischmann E, et al. Denaturing
high-performance liquid chromatography detects reliably
BRCA1 and BRCA2 mutations. Genomics 1999;62:369 76.

ANESTH ANALG
2005;101:14016

22. Escary JL, Cecillon M, Maciazek J, et al. Evaluation of DHPLC

analysis in mutational scanning of notch3, a gene with a high
G-C content. Hum Mutat 2000;16:518 26.
23. Buyse IM, Fang P, Hoon KT, et al. Diagnostic testing for Rett
syndrome by DHPLC and direct sequencing analysis of the
MECP2 gene: identification of several novel mutations and
polymorphisms. Am J Hum Genet 2000;67:1428 36.
24. Benit P, Kara-Mostefa A, Berthelon M, et al. Mutation analysis
of the hamartin gene using denaturing high performance liquid
chromatography. Hum Mutat 2000;6:41721.
25. van Den Bosch BJ, de Coo RF, Scholte HR, et al. Mutation
analysis of the entire mitochondrial genome using denaturing
high performance liquid chromatograph. Nucleic Acids Res
2000;28:E89.
26. Urwyler A, Deufel T, McCarthy T, West S. Guidelines for molecular genetic detection of susceptibility to malignant hyperthermia. Br J Anaesth 2001;86:2837.
27. Sei Y, Sambuughin N, Muldoon S. Malignant hyperthermia
genetic testing in North America Working Group Meeting. Anesthesiology 2004;100:464 5.

MEDICAL INTELLIGENCE

Reducing the Incidence of Surgical Fires: Supplying Nasal

Cannulae with sub-100% O2 Gas Mixtures from
Anesthesia Machines
Samsun Lampotang, PhD*, Nikolaus Gravenstein,
Dietrich Gravenstein, MD*

MD*,

David A. Paulus,

MD*,

and

Departments of *Anesthesiology and Neurosurgery, University of Florida College of Medicine; and Departments of
Mechanical and Aerospace Engineering and Electrical and Computer Engineering, University of Florida College of
Engineering, Gainesville, Florida

In June 2003, the Joint Commission on Accreditation of

Healthcare Organizations (JCAHO) recommended:
As a general policy, use air or Fio2 at 30% for open
delivery (consistent with patient needs) to prevent
surgical fires. One way to interpret JCAHOs recommendation is that 100% O2 should not be indiscriminately used, and anesthesia providers should have the
ability, consistent with patient needs and their clinical
judgment, to deliver sub-100% O2 with nasal cannulae.
An auxiliary O2 flowmeter has a barbed outlet connector that offers a convenient means to connect a cannula
to an anesthesia machine and is routinely used for open

urgical fires are largely preventable incidents

that continue to occur (1). Reasons for fires may
include, among others, lack of individual and
team training, equipment design, human factors, poor
communication among care team members, and culture. Reduction of the risk of surgical fires is also one
of the Joint Commission on Accreditation of Healthcare Organizations (JCAHO) 2005 Ambulatory Care
National Patient Safety Goals: Educate staff, including operating licensed independent practitioners and
anesthesia providers, on how to control heat sources
and manage fuels, and establish guidelines to minimize O2 concentration under drapes.1
An oxidizer, a fuel, and an ignition source, the
so-called fire triangle, must all be present for a
1
JCAHO 2005 patient safety goals for ambulatory and office-based
surgery (promulgated in July, 2005): http://www.jcaho.org/
accreditedorganizations/patientsafety/05npsg/05_npsg_amb.
htm.
Accepted for publication April 29, 2005.
Address correspondence to S. Lampotang, PhD, Department of
Anesthesiology, PO Box 100254, Gainesville, FL 32610-0254. Address e-mail to slampotang@anest.ufl.edu.

DOI: 10.1213/01.ANE.0000180215.50589.02
2005 by the International Anesthesia Research Society
0003-2999/05

delivery of 100% O2. The auxiliary O2 flowmeter provides only 100% O2 and thus does not allow titration of
the O2 concentration to patient needs and may increase
the risk of surgical fires. This report clarifies the JCAHO
recommendation and describes different means of addressing it that are based primarily on using the anesthesia machine to blend a sub-100% O2 gas mixture and
delivering it via a nasal cannula. The options presented
depend on the model and manufacturer of the anesthesia machine and allow delivery via nasal cannula of O2
concentrations that range from 21% to 100%.
(Anesth Analg 2005;101:140712)

fire to occur. For example, plastic, cloth, or paper

drapes, and alcohol-based preparation solutions can
serve as fuel (2) and cautery, lasers, or endoscopes
as ignition sources (3). The auxiliary ball-in-tube O2
flowmeter mounted on many North American anesthesia machines delivers pure O2 (Table 1 and
Table 2). It is often used to supply nasal cannulae
with pure O2 during procedures performed under
sedation or monitored anesthesia care and thus may
provide the oxidizer of the fire triangle, especially in
situations in which the supplemental O2 is not
quickly dissipated.
In a June, 2003 Sentinel Event Alert, the JCAHO
recommended that health care organizations help prevent surgical fires by establishing guidelines for minimizing O2 concentration under surgical drapes.2 The
Sentinel Event Alert quoted the ECRI (a nonprofit
health services research agency): As a general policy,
use air or Fio2 at 30% for open delivery, and added
(consistent with patient needs) to the end of the
2
Joint Commission on Accreditation of Healthcare Organizations Sentinel Event Alert. Preventing surgical fires. Issue 29, June 24, 2003. http://
www.jcaho.org/aboutus/newsletters/sentineleventalert/
sea_29.htm.

Anesth Analg 2005;101:140712

1407

1408

MEDICAL INTELLIGENCE LAMPOTANG ET AL.

REDUCING THE INCIDENCE OF SURGICAL FIRES

ANESTH ANALG
2005;101:140712

Table 1. Common Gas Outlet Characteristics in Datex-Ohmeda/GE Anesthesia Machines on the North American Market

Anesthesia
machine
model

Auxiliary
ball-in-tube
O2 flowmeter
delivers
100% O2

Modulus I
Yes
Modulus II
Yes
Modulus CD
Yes
Modulus
Yes
CD-CV
Excel
Yes
Aestiva
Yes
ADU
Yes
Aespire
Yes
Avance
Yes (non-color
specific)

Available
with air
flowmeter?

FGF hose
readily
disconnected/
reconnected
from CGO by
clinician?

CGO accepts a
15-mm OD
connector, e.g.,
a 5.0-mm
endotracheal
tube connector?

Optional
Optional
Optional
Optional

Yes
Yes
Yes
Yes

Yes
Yes
Yes
Yes

Optional
Optional
Optional
Optional
Standard

Yes
No
Yes
No
No

Yes
N/A
Yes
N/A
Inspiratory port
(configurable as
CGO) accepts
15-mm OD
connector

Manufacturer
recommends
against
connection
of nasal
cannula
to CGO?

Auxiliary
CGO
available?

Auxiliary CGO
accepts a 15-mm
OD connector,
e.g., an
endotracheal
tube connector?

Manufacturer
recommends
against
connection of
nasal cannula
to auxiliary
CGO?

No
No
No
No

N/A
N/A
N/A
N/A

N/A
N/A
N/A
N/A

No
Optional
Yes
Yes
Optional
(separate
outlet)

N/A
Yes
Yes
Yes
Yes

N/A
No
No
No
No

No
No
No
No
No
N/A
No
N/A
N/A

Data courtesy of Datex-Ohmeda.

FGF fresh gas flow; CGO common gas outlet; OD outside diameter.

Table 2. Common Gas Outlet Characteristics in Drager Anesthesia Machines on the North American Market

Anesthesia
machine
model

Manufacturer
recommends
against
connection of
nasal cannula to
CGO?

Auxiliary
ball-in-tube
O2 flowmeter
delivers
100% O2?

Available
with air
flowmeter?

FGF hose readily

disconnected/
reconnected from
CGO by clinician?

CGO accepts a 15-mm

OD connector, e.g., a
5.0-mm endotracheal
tube connector?

Yes (optional)
Yes (optional)
Yes (optional)
Yes

Optional
Optional
Optional
Standard

Yes
Yes
Yes
Yes

Yes
Yes
Yes
Yes

Undetermined
Undetermined
Undetermined
Undetermined

No
No
No
No

Yes
No

Yes
N/A

Undetermined
N/A

No
No

Standard, N/A;
optional CGO,
undetermined
N/A
N/A
Undetermined

No

Narkomed 2B
Narkomed 3
Narkomed 4
Narkomed
Mobile
Narkomed GS
Narkomed
6000 series
Fabius GS

Yes (optional)
Yes

Optional
Standard

Yes (optional)

Standard

Fabius Tiro
Julian
Narkomed MRI

Yes (optional)
Yes
Yes

Standard
Standard
Standard

Standard, no;
optional CGO, yes
No
No
Yes

Standard, N/A;
optional CGO, yes
N/A
N/A
Yes

Auxiliary
CGO
available?

No
No
No

Data courtesy of Drager Medical Inc.

FGF fresh gas flow; CGO common gas outlet; OD outside diameter.

ECRI guideline3 (bold added by authors for

emphasis).
The current practice of routinely supplying nasal
cannulae with 100% O2, which does not comply with
the JCAHO/ECRI recommendations, is reinforced by
the ease of connecting nasal cannulae to the barbed
outlet of the auxiliary ball-in-tube O2 flowmeter. Large
O2 concentrations and/or flow rates of O2-enriched
gas delivered by cannula may create a localized, O2rich environment that promotes surgical fires (Fig. 1,
bottom) when combined with an ignition source and
combustible material. ECRI reports that there are
about 100 surgical fires annually, resulting in up to 20
serious injuries and 1 or 2 deaths (4).
3

ECRI: Only You Can Prevent Surgical Fires Poster, July, 2004
at http://www.mdsr.ecri.org/static/surgical_fire_poster.pdf.

The JCAHO/ECRI recommendations are especially

relevant if drapes prevent dissipation of pure O2 and
cautery (electro- or thermal) and lasers are used in or
near the O2-rich region, usually around the head or
neck. Furthermore, the users guide for electrosurgery
units explicitly warns against O2 accumulation or pooling under surgical drapes or within the area where
electrosurgery is performed.4
Addressing the JCAHO/ECRI recommendations
poses challenges in terms of altering clinical practice,
ensuring safe clinical care for patients requiring supplemental O2, equipment, and added cost. The variety
of designs and configurations of anesthesia machines
(Tables 1 and 2) and individual clinical practices and
4

Valleylab, Force FX-C electrosurgical generator users guide.

1999.

ANESTH ANALG
2005;101:140712

MEDICAL INTELLIGENCE
LAMPOTANG ET AL.
REDUCING THE INCIDENCE OF SURGICAL FIRES

1409

Figure 1. The top figure depicts the rule of thumb that 2 L/min pure
O2 delivered via nasal cannula results in an Fio2 of about 28%. The
excess O2 can dissipate into ambient air and is diluted. In the bottom
schematic, the patient is draped for eye surgery. Excess O2 may
become trapped below the drape where an enriched O2 environment increases the risk of a surgical fire. Even though the patients
effective inspired O2 approximates only 28%, because 100% O2 exits
the nasal cannula, there is a region of indeterminate size and location where 100% O2 is present.

preferences compound these challenges. This report is

based partly on the experience gained by our academic department in examining various alternatives
that address the recent JCAHO/ECRI recommendations. This report is not about denying supplemental
O2 to a patient where there is a need. It is about
providing equipment and practice recommendations
that help prevent surgical fires during routine
procedures.

Analysis
Figure 2 describes various alternatives for addressing
the JCAHO/ECRI recommendations of generally using a fraction of delivered O2 (FDO2) 30%, depending on the individual anesthesia machine configuration and the clinical logistics.

Anesthesia Machine with Air Flowmeter

Anesthesia machines with an air flowmeter can blend
a sub-100% O2 mixture in air and deliver it to the
common gas outlet (CGO) and, if present, also the
auxiliary CGO (ACGO). A nasal cannula can be mated
via a 5-mm endotracheal tube connector to a CGO, a
Y-piece (via the CGO) or, if present, an ACGO. A gas
sampling connector can be inserted in series with the

Figure 2. An algorithm to select means to deliver sub-100% O2 to

open delivery systems such as nasal cannulae and face masks.

endotracheal tube connector (5 mm) to allow analysis of the gas composition actually delivered to the
cannula (Fig. 3). To obtain an FDO2 30%, use O2-toair ratios 1:7, e.g.: (0.5 L/min O2:3.5 L/min air) or (1
L/min O2:7 L/min air). In newer machines with an
ACGO, the blended gas mixture issuing from the bank
of flowmeters can be instantly rerouted to either the
CGO or the ACGO, by simply flipping a selector lever.
To make use of this feature, anesthesia providers must
first be made aware of the presence and capability of
the ACGO and reminded of the need to flip the selector lever to redirect the fresh gas flow (FGF) as
intended.
A simple approach, which does not require a nasal
cannula and works with all anesthesia machines with
air flowmeters, uses a corrugated hose breathing circuit to deliver an FDO2 30% gas mixture (e.g., 1
L/min O2 and 7 L/min air) directly from a Y-piece
placed close to the patients nose. High FGFs have the
advantages of washing away exhaled CO2 from under
the drapes, preventing CO2 rebreathing, and producing a breeze on the face that reduces the claustrophobic sense of asphyxiation. If an O2 analyzer is available, we recommend that the gas sampling connector
be left on the Y-piece so that FDO2 can be continuously
monitored.

1410

MEDICAL INTELLIGENCE LAMPOTANG ET AL.

REDUCING THE INCIDENCE OF SURGICAL FIRES

Figure 3. A nasal cannula mated to a common gas outlet via a 5-mm

endotracheal tube connector and a gas sampling connector with a
gas sampling line attached.

A variation is to connect a nasal cannula to the

Y-piece of a standard circle breathing circuit with a
5-mm endotracheal tube connector and supply it with
an FDO2 30% gas blend from the flowmeter bank (5).
The selector knob must be set to Bag and the adjustable pressure limiting (APL) valve fully closed. The
airway pressure gauge needle will indicate a constant
pressure in the 20- to 25-cm H2O range, depending on
the machine design. The reservoir bag will fill and can
act as a visual indicator of the patency of the delivery
circuit. If the selector knob is set to Ventilator, the
threshold (nominally 3 cm H2O) on the ventilator
pressure relief valve in an upright bellows ventilator
will allow most of the FGF to exit to the scavenging
system rather than flow to the nasal cannula. The
same would be true of an open or partially open APL
valve with the selector knob set to Bag. Importantly,
if a connector (such as the standard elbow connector)
incorporating a gas sampling port is left interposed
between the Y-piece and cannula, monitoring of the
delivered gas composition becomes possible and practical. This provides convenient verification at all times
that FDO2 30% (a high O2 alarm set at 30% will warn
when FDO2 is 30%). FDO2 monitoring will also detect
disconnection of a nasal cannula from its source if the
low FDO2 alarm is set to 22% because FDO2 will decrease to 21% (ambient air O2) during a disconnection.
Gas analysis will also confirm that the gas issuing
from the O2 supply is actually O2 and warn against
gas mix-ups. Slow O2 analyzers that are part of the
CO2 absorber in some designs monitor FDO2 whenever the breathing circuit or Y-piece is used to channel
gas to the patient but not when the CGO or ACGO is
used to supply nasal cannulae.

ANESTH ANALG
2005;101:140712

We tested the nasal cannula-at-Y-piece-method with

5 anesthesia machine designs available in our laboratory: Modulus I and II, Aestiva, Avance (GE Healthcare, Madison, WI), and the Fabius GS (Drager Medical, Telford, PA). We set the total FGF at 4 L/min, the
threshold above which patient discomfort might occur. All machines were able to deliver 30% O2 in air,
with some adjustment of the flowmeters from the
nominal 1:7 O2/air ratio required. The machines differed in whether sustained airway pressure alarms
were generated and could be eliminated by increasing
the high pressure (Pmax) limit (Table 3). For the Modulus II, Aestiva, and Avance, the high limit for sustained pressure is half the high limit for Pmax. Flows
4 L/min or an O2 flush through a nasal cannula
connected to the Y-piece generated high pressures and
alarms. Depressurization of the circuit by temporarily
opening the APL valve removed the alarm condition.
In those machines in which the FGF hose can be
readily disconnected from the CGO (e.g., Modulus,
Excel, and Narkomed product lines), a nasal cannula
can be supplied from the CGO. Some anesthesia providers, out of concerns for safety, insist on always
having the circle system fully assembled and ready for
use in the event complications arise and object to tying
up the CGO or Y-piece to supply a nasal cannula, even
when the circle system is not in use. In that case,
supplying the nasal cannula from an ACGO (if available) meets the requirements for an intact circle system available at the flip of a selector lever.
When the O2 flush button is pressed, pure O2 is
delivered to the CGO or ACGO at approximately 60
L/min, bypassing the flowmeter bank and overriding
the flowmeter, i.e., FDO2 settings. When a nasal cannula is connected to and supplied from the CGO,
ACGO, or Y-piece, clinicians must be aware that an
ill-advised O2 flush in the presence of an ignition
source could initiate a surgical fire. Conversely, the O2
flush function provides a convenient means, when
clinically indicated, of temporarily and quickly increasing Fio2 during periods when it is safe to do so
(e.g., when there are no obvious ignition sources
present or anticipated in the next minute).

Anesthesia Machine Without Air Flowmeter

In O2/N2O machines, including those in dental offices, N2O should not be used as the balance gas when
supplying nasal cannulae with FDO2 30% because (a)
as a result of its nature, it may provide unintended
analgesia or loss of consciousness, (b) it will contaminate the room because scavenging in this application
is not sufficiently reliable, and, most importantly, (c)
N2O, similar to O2, is an oxidizer and supports combustion. Heliox can be used in machines with a heliox
flowmeter. Generally, heliox is a 30% O2/70% helium
mixture but users should note that heliox may be more

ANESTH ANALG
2005;101:140712

MEDICAL INTELLIGENCE
LAMPOTANG ET AL.
REDUCING THE INCIDENCE OF SURGICAL FIRES

1411

Table 3. Sustained Pressure Alarms and Deactivation with Different Anesthesia Machines
Sustained pressure at
4 L/min flow through
nasal cannula at Y-piece
(cm H2O)

Sustained pressure alarm

triggered at default Pmax
setting of 40 cm H2O?

Modulus I/7000
Modulus II/7800

20
20

No
No (yes at FGFs 4 L/min)

Aestiva

25

Yes

Avance

21

Yes

Fabius GS

21

Yes

Anesthesia machine design

Sustained pressure
alarm deactivated by
adjusting pressure
limits?
N/A
Yes (set Pmax twice
sustained pressure)
Yes (set Pmax twice
sustained pressure)
Yes (set Pmax twice
sustained pressure)
No

FGF fresh gas flow.

expensive and that heliox is supplied in different

mixes such as 25% and 40% O2.

Source of Compressed Medical Air Available

In anesthesia machines without an air flowmeter, an
option is an add-on air/O2 gas blender (disadvantages: added expense, plumbing, mounting hardware,
and clutter).

Source of Compressed Medical Air Unavailable

An option is to use 100% O2 from the auxiliary O2
flowmeter and discontinue O2 supplementation for at
least 60 s before and during cautery use, or empirically
scavenge in the region of the head and neck.

Discussion
This technical report considers primarily the equipment aspects of addressing the JCAHO/ECRI recommendations and only touches briefly on the oxygenation consequences of FDO2 30% by nasal cannula.
The JCAHO/ECRI recommendations have not been
accepted by all clinicians. Some may have missed the
qualifying phrases: as a general policy and consistent with patient needs. Others question the merit of
FDO2 30% against simply breathing room air as a
general policy, especially at low flows that would
result in ambient air entrainment and dilution to an
effective Fio2 30%. Our interpretation is that it is not
the intent of the JCAHO Sentinel Event alert to prevent using an FDO2 30% in patients in whom there is
a need, but rather to avoid routinely using a higher
FDO2 as has been customary practice. In patients who
need a higher Fio2, simply making the surgeon aware
and limiting, or even avoiding, cautery use and gas
scavenging in the area of the drapes are all strategies
to reduce the likelihood of a surgical fire. We wrote
this report while keeping in mind that clinicians
should be able to set the FDO2 at any value between
21% and 100%, if clinically indicated or if subsequent

clinical and flammability studies suggest a different

FDO2 threshold. At the time of writing, we are not
aware of any clinical study of the JCAHO/ECRI recommendations. Although JCAHO does not apply the
2005 goals to the accreditation process for nonambulatory care facilities, such as acute care hospitals, it
does not remove the need for these facilities and their
practitioners to be proactive in surgical fire
prevention.
The JCAHO recommendations in the 2003 Sentinel
Event Alert used the term, Fio2, the abbreviation for
inspired fraction of O2. In open systems, Fio2 is unknown because (a) O2 is usually not monitored with
open systems such as nasal cannulae, and (b) patients
can always entrain ambient air in an open system so
that the actual Fio2 is typically less than the delivered
O2 fraction. A heuristic rule states that Fio2 increases
by 4% from a baseline of 20% O2 in ambient air for
every liter of pure O2 flowing through a nasal cannula
(6). That is, 1 L/min of 100% O2 flowing out of the
cannula after being diluted by entrained ambient air
results in an approximate inspired Fio2 of 0.24, 2
L/min O2 Fio2 0.28, 3 L/min O2 Fio2 0.32 (see Fig.
1, top). Thus, the JCAHO recommendations could be
misconstrued as recommending pure O2 flows 2.5
L/min as a general policy. When contacted, ECRI
confirmed that it meant the delivered fraction of O2
and, thus, it is more appropriate to use FDO2 rather
than Fio2.5 The 2005 National Patient Safety goals do
not mention the Fio2 details that were contained in the
2003 JCAHO Sentinel Event. However, the thrust of
this report is very much on point related to the goals
wherein the reports recommendations help meet the
2005 goal to establish guidelines to minimize O2 concentration under drapes.
The term O2 cannula is used interchangeably with
nasal cannula in medicine in general, subconsciously reinforcing the practice of using 100% O2 with
5

Personal communication by telephone, Al de Richemond, ECRI,

March 19, 2004.

1412

MEDICAL INTELLIGENCE LAMPOTANG ET AL.

REDUCING THE INCIDENCE OF SURGICAL FIRES

nasal cannulae. Words often carry both an explicit and

implicit meaning that can subtly influence our actions,
and terminology is an intrinsic part of medical culture.
Patient safety efforts are beginning to focus on promoting a culture of safety, and simple actions, such as
encouraging the use of nasal cannula and discouraging the use of O2 cannula are examples of the
small steps we can begin to take to improve patient
safety.
The JCAHO recommendations specify Fio2 but
are not forthcoming on the flow rate of O2. Anecdotal
evidence seems to indicate that, in undraped patients,
no surgical fires occur at flow rates of 2 L/min O2.
Rate of inflow of O2 may be as important as O2 concentration because it may be that it is the total amount
of O2 available that matters. Further investigation of
the effect of total flow of O2 on surgical fires is needed.
When using the air and O2 flowmeters to supply nasal
cannulae, FDO2 can be titrated to patient need according to Spo2 while maintaining FGF at 4 L/min to
minimize patient discomfort or, when needed, high
flow rates of pure O2 can also be delivered.
The added flow resistance from the length and narrow bore of nasal cannulae will generate back pressure as indicated by the dipping of the O2 and air
flowmeter bobbins when a nasal cannula is connected
to the CGO, Y-piece, or ACGOa phenomenon not
observed when connecting a FGF hose to a common
gas outlet. This observation raised justifiable concerns
that the added flow resistance of nasal cannulae may
affect the accuracy of the anesthesia machine flowmeters. When contacted, Datex-Ohmeda indicated that it
does not have recommendations against connecting a
nasal cannula to a CGO or ACGO (Table 1), and
Drager Medical Inc. stated that it does not make any
recommendation regarding the use of the CGO for the
supply of a nasal cannula. We conducted experiments
indicating that back pressure generated by a nasal
cannula connected to the CGO or ACGO affects the O2
and air flowmeters similarly, resulting in no net effect
in the Modulus I, II, and Aestiva anesthesia machines
(7).
Gas composition is not always monitored during
supplemental O2 provided via open delivery systems,
as was the case during the January, 2002 fatalities in
New Haven where N2O was inadvertently substituted
for O2 (8). Connecting a standard nasal cannula to an
anesthesia machine with the gas sampling connector
left in provides a quick and convenient way to monitor delivered gas composition, a capability unavailable

ANESTH ANALG
2005;101:140712

until now with O2 supplementation via standard nasal

cannulae (not the newer type that allows gas
monitoring).
Other approaches to reducing surgical fires that
address the two other legs (fuel, ignition) of the fire
triangle have been recommended by JCAHO.2 In general, removal of the fuel component (drapes, nonflammable lubricants, soaking with saline, different drape
materials, etc.) is the responsibility of the prep nurses,
whereas the ignition source (cautery, scopes, etc.) is
usually under the control of the surgical staff. Anesthesia providers are generally recognized to be in
control of the third leg of the fire triangle, the oxidizer
component. This report, written for anesthesia providers, focuses on actions under their direct control. Redundancy is an established principle in designing a
safety system. A multidisciplinary team effort will
result in a triply redundant safety system that reduces
the occurrence of surgical fires.
We thank J. S. Gravenstein, MD, for his insightful comments; Robert
Tham, Gina Petry; and Tony Bean (Datex-Ohmeda/GE) and Robert
Clark (Drager Medical Inc.) for supplying data on their respective
anesthesia machine configurations; Esperanza Olivo and Anita Yeager for editorial assistance; and Kendra Kuck and David Lizdas for
the graphics. We also thank Al de Richemond of ECRI for the
personal communication that helped clarify the ECRI recommendation that is the subject of this report.

References
1. Lowry RK, Noone RB. Fires and burns during plastic surgery.
Ann Plast Surg 2001;46:72 6.
2. Barker SJ, Polson JS. Fire in the operating room: a case report and
laboratory study. Anesth Analg 2001;93:960 5.
3. Wolf GL, Sidebotham GW, Lazard JLP, Charchaflieh JG. Laser
ignition of surgical drape materials in air, 50% oxygen, and 95%
oxygen. Anesthesiology 2004;100:116771.
4. ECRI. A clinicians guide to surgical fires: how they occur, how to
prevent them, how to put them out [guidance article]. Health
Devices 2003;32:524.
5. Lampotang S, Good ML. The anesthesia machine, anesthesia
ventilator, breathing circuit and scavenging system. In: Kirby RR,
Gravenstein N, Lobato EB, Gravenstein JS, eds. Clinical anesthesia practice. 2nd ed. Philadelphia: WB Saunders; 2002:277302.
6. Branson RD. Gas delivery systems: regulators, flowmeters, and
therapy devices. In: Branson RD, Hess DR, Chatburn RL, eds.
Respiratory care equipment. 2nd ed. Philadelphia: Lippincott
Williams & Wilkins; 1998:55 85.
7. Lampotang S, Paulus DA, Gravenstein N. FDO2 accuracy when
supplying nasal cannulae from common gas outlets [abstract].
Anesthesiology 2004;101:A565.
8. Morell RC. Gas delivery mistakes continue to kill. APSF Newsletter, Spring 2002. http://www.apsf.org/resource_center/
newsletter/2002/spring/11gasdelivery.htm.

BRIEF REPORT

Air Detection Performance of the Level 1 H-1200 Fluid and

Blood Warmer
Ramachandra R. Avula, Richard Kramer, and Charles E. Smith
Department of Anesthesiology, MetroHealth Medical Center, Case Western Reserve University, Cleveland, Ohio

We evaluated the Level 1 H-1200 fluid warmer during simulated conditions of minor to massive air embolism. The fluids we tested were crystalloid and diluted red cells (estimated hematocrit 50%) during gravity and pressure driven
flow. The air volumes tested ranged from 160 mL for crystalloid and 30150 mL for red cells. No air was observed
distal to the air detector and clamp during all test conditions.

ccidental administration of air in IV fluid bags

during the use of pressurized infusion devices
can be fatal. In a Letter to the Editor to the
Anesthesia Patient Safety Foundation (1), mention is
made of 3 cases of massive air embolism from pressure infusers that resulted in settlements of $385,000 to
$1,600,000. We are aware of five cases of iatrogenic air
embolism during pressurized infusion. Adhikary and
Massey (2) describe a case of massive venous air embolism during coronary artery surgery that occurred
while infusing crystalloid under pressure via a Level 1
rapid infuser. Air embolism was diagnosed using
transesophageal echocardiography, and the patient
was successfully resuscitated without neurologic sequelae. The frequency of fatal air embolism after pressurized administration of salvaged perioperative
blood was 4 of 127,586 blood recovery procedures, or
1:31,300 (3).
The purpose of this study was to evaluate the air
handling characteristics of a rapid infusion fluid
warming device equipped with an ultrasonic air detector and automatic line clamp (Level 1 H-1200 Fluid
Warmer; Smiths Medical, Rockland, MA) designed to
protect against accidental lethal air embolism.

Accepted for publication May 3, 2005.

Address correspondence to Charles Smith, MD, Department of
Anesthesiology, MetroHealth Medical Center, 2500 MetroHealth
Dr., MetroHealth Medical Center, Case Western Reserve University,
Cleveland, OH 44109. Address e-mail to csmith@metrohealth.org.
DOI: 10.1213/01.ANE.0000180207.30194.EE
2005 by the International Anesthesia Research Society
0003-2999/05

The device consistently alarmed and automatically shut off

flow. Air was easily purged through the gas vent-filter assembly during simulated air embolism with crystalloid but
not with red cell infusion. The use of ultrasonic air detection
coupled with automatic shutoff is a significant safety improvement of the Level 1 H-1200 fluid and blood warmer.
(Anesth Analg 2005;101:14136)

The air detector clamp monitors for presence of air

in the disposable gas vent-filter assembly by continually passing an ultrasound signal through the fluidfilled gas vent-filter assembly. Because a bolus of air
displaces the fluid in the gas vent-filter assembly, the
ultrasound signal is broken, and the clamp automatically closes off the patient line. Audible and visual
alarms are activated to alert the anesthesiologist that
the fluid flow has stopped because of the presence of
air. Existing Level 1 rapid infusers (H-1000/1025 and
H-275/H-525) can be upgraded with the integrated air
detector/clamp device.

Methods
Tests were performed in the operating rooms of
MetroHealth Medical Center. The temperature was set
at 20C. To test the amount of air in crystalloid bags,
1-L sterile infusion bags of lactated Ringers (LR) solution and 0.9% saline (crystalloid; Baxter Healthcare
Corp, Deerfield, IL) were tested immediately after
opening the bag. An 18-gauge needle attached to a
syringe was inserted into the connection port of each
inverted bag. The syringe was aspirated until fluid
was observed to enter the needle hub.
To determine the air-handling characteristics, D-100
IV fluid administration sets were used per manufacturer operating instructions (Level 1 H-1200 operators
manual). Both drip chambers were primed with LR or
saline according to test sequence. Priming volume was
65 mL. The patient line was connected to a 5-in.
T-connector (Baxter Interlink; total volume
Anesth Analg 2005;101:14136

1413

1414

BRIEF REPORT

Figure 1. Lower part of the Level 1 H-1200 fluid warmer. A threeway stopcock (A) was inserted proximal to the gas vent-filter assembly and air detector (B) for air handling tests in Experiment 2.
Note the distal clamps (black arrows) after the automatic shutoff
clamp (3).

0.84 mL), which was submerged in a liquid-filled beaker. The heat exchange temperature was 41C. Observations were by at least two individuals in all cases.
Three testing conditions were performed:
1. Crystalloid. A three-way stopcock was inserted
proximal to the gas vent and air detector/clamp (Fig.
1). Aliquots of air (1, 2, 5, 10, 15, 20, 25, 30, 35, 40, 45,
50, 55, and 60 mL) were rapidly injected into the
stopcock. The stopcock was returned to its original
position to allow flow to resume. Visual inspection for
air bubbles distal to the air detector/clamp and in the
liquid-filled beaker was performed by two observers.
Injection of 1 mL of air into the patient line distal to the
gas vent-air detector/clamp was readily visible in the
liquid-filled beaker with submerged T-connector.
Tests were repeated twice for each condition and fluid
for gravity (height 6 ft) and pressurized (constant
300 mm Hg) flow after re-priming the system and
venting of air.
2. Crystalloid. Aliquots of air (10, 15, 20, 25, 30, 35, 40,
45, 50, 55, and 60 mL) were rapidly injected into the
hub at the Y-connector just above the drip chamber
with the ratchet clamp above the Y-connector in the
closed position (Fig. 2). The clamp was then opened to
allow gravity or pressure flow. Testing conditions
were as in Experiment 1.

ANESTH ANALG
2005;101:14136

Figure 2. Upper part of the Level 1 H-1200 fluid warmer. Note the
hub (white arrow) at the Y-connector just above the drip chamber,
which was used for injecting air in Experiment 2. The ratchet clamps
(black arrows) are in the open position.

3. Red Blood Cells (RBCs). RBCs were obtained from

an outdated blood bank supply and kept refrigerated
until use. One drip chamber of the D-100 disposable
was spiked with 0.9% saline and the other with blood.
Saline 100 mL was added to each unit to yield an
estimated hematocrit of 50%. Disposable sets were
primed with saline. Aliquots of air (30, 60, 90, 120, and
150 mL) were added to the RBCs. The patient line was
attached to a cell salvage waste system using a
T-connector. Pressurized flow was performed.

Results
Fifty bags of each solution were tested. LR contained
(mean sd) 53 4 mL of air (range, 43 61 mL), and
saline contained 53 5 mL (range, 41 65 mL) of air.
A total of 200 tests were performed for crystalloid
solutions. No air was observed in the T-connector or in
the liquid-filled beaker. Automatic shutoff did not
occur with injected volumes 25 mL during gravity or
pressure flow. With larger volumes, the air detector
consistently alarmed and automatically shut off flow.
After closing the ratchet clamps under the IV bag and
the ratchet and roller clamps distal to the air detector/

ANESTH ANALG
2005;101:14136

clamp as well as insuring that a full bag of crystalloid

was spiked and primed in the drip chamber, air was
easily purged through the gas vent-filter assembly by
releasing the ratchet clamp under the IV bag and
pressing the unclamp button twice. Opening of the
distal clamps reestablished flow.
Nine tests were performed for RBC infusion. No air
was detected in the T-connector. The air detector consistently alarmed and automatically shut off flow during each test. It was impossible to purge air through
the gas vent-filter using pressurized or gravity flow of
saline as with crystalloid solutions.

Discussion
Approximately 200 mL of air can enter the circulation
in as little as four seconds during pressurized infusion
of air containing solutions (average, 49.3 mL/s; range,
43 61 mL/s) (3). Despite repeated warnings to de-air
IV bags before infusion, fully prime disposable sets,
and avoid pressurization of cell-saver blood, there is
still a hazard of accidental air embolism because of
human error. One-liter bags of crystalloid used at our
institution contain sufficient amounts of air that, if
delivered under pressure, can cause air embolism and
death. Similarly, 1-L bags of crystalloid from other
manufacturers (Abbott, Chicago, IL) have been shown
to contain 65 10 mL of air, with a range of 45 85 mL
(2). Without the air detector/clamp, the gas vent-filter
assembly of older Level 1 rapid infusers acts merely to
release gas bubbles produced by the warming of the
liquid. The amount of air that could reach the patient
would then depend on factors such as drip-chamber
volume, air volume, gas vent and filter performance,
temperature, and flow.
The use of ultrasonic air detection coupled with an
automatic shutoff is a significant safety improvement
of the Level 1 H-1200 warmer because human vigilance alone is not sufficient to detect lethal quantities
of air at rapid flow. The air detector/clamp functioned
effectively during all test conditions and would be
expected to prevent lethal air embolism similar to the
Belmont FMS 2000 (Belmont Instrument Corp, Billerica, MA) (4). However, the steps required to reprime the system with the Level 1 H1200 after air
detection and shutoff were very different from those
of the Belmont. To restore flow after air detection with
the Belmont FMS 2000, the operator follows a series of
command prompts on the screen, which result in the
semiocclusive roller head pump pushing a bolus of
fluid to clear the air into the recirculate line and back
up into the reservoir chamber (5). With the Level 1
H-1200, the following steps are required per manufacturer operators manual to reestablish fluid flow after
the patient line clamps off because of air in the gas
vent filter assembly:

BRIEF REPORT

1415

1. Close the clamps under the IV bag, the roller

clamp, and the clamp below the gas vent-filter
assembly.
2. Remove pressure from the pressure chamber.
3. Inspect the IV tubing for air.
4. Locate the air source, and correct this condition.
5. Spike an air-free bag of 0.9% saline.
6. Insert the IV bag in the pressure chamber.
7. Pressurize the chamber.
8. Prime the IV drip chamber, and open the clamps
above the gas vent-filter assembly.
9. If fluid flows freely through the IV tubing, air in
the IV line is vented out through the gas vent-filter
assembly.
10. Press the unclamp button once to enter the unclamp mode.
11. Press the unclamp button again to enter automatic operation mode.
12. If no warning signals are activated (e.g., no air
detected, system primed and ready to go), open all
clamps to reestablish fluid flow.
If fluid does not flow freely through the IV tubing in
Step 9, the gas vent-filter assembly needs to be replaced. Steps 1 8 are then repeated, followed by Steps
10 12. In the present study, with blood-containing
solutions, it was not possible to purge the D-100 disposable set of air until the gas vent-filter assembly was
removed from the unit. This was likely because of
poor air venting performance of the gas-permeable
membrane of the gas vent-filter assembly during red
cell infusion, possibly caused by proteins plugging the
filter. Instead of replacing the gas vent-filter assembly,
as recommended by the manufacturer, the gas ventfilter assembly was disconnected from the fluid
warmer after ensuring that all clamps were closed. A
new bag of 0.9% saline was spiked, and the disposable
set was cleared of air by opening up all the clamps and
allowing fluid (and air) to flow into a closed cell-saver
waste system. Once the entire system was primed
with saline and free of air, the gas vent-filter disposable was reinstalled. It is not known how the system
would perform with respect to fresh frozen plasma or
platelets. In our opinion, the advantages of automatic
air detection far outweigh the disadvantage of having
to de-air once an air embolus situation has been
avoided (e.g., near-miss). Worst case scenario is that
the operator would need to replace the gas vent-filter
assembly (recommended by the manufacturer) or disconnect the gas vent-filter disposable and patient line
from the air detection/clamp system (making sure the
distal clamps were shut off) and running saline
through the system to purge the air into a waste
container. In conclusion, use of automatic air detection
coupled to a shutoff may reduce the complication of
lethal air embolism during rapid fluid infusion.

1416

BRIEF REPORT

The authors thank Jeanne Javor, MT (ASCP), SBB, Blood Bank

Supervisor, MetroHealth Medical Center, for providing outdated
red cells and Jason Feuerman, Chester Scholar, MetroHealth Medical Center, for assistance with determining how much air was in IV
crystalloid bags. Smiths Medical ASD Inc., Rockland, MA, provided
loan of the fluid warmer and donation of disposables.

ANESTH ANALG
2005;101:14136

References
1. Comunale ME. IV fluid warmers create air embolus danger.
Anesthesia Patient Safety Foundation Newsletter 2000;15:412.
2. Adhikary GS, Massey SR. Massive air embolism: a case report.
J Clin Anesth 1998;10:70 2.
3. Linden JV, Kaplan HS, Murphy MT. Fatal air embolism due to
perioperative blood recovery. Anesth Analg 1997;84:422 6.
4. Smith CE, Kabbara A, Kramer RP, Gill I. A new IV fluid and
blood warming system to prevent air embolism and compartment syndrome. Trauma Care 2001;11:78 82.
5. Comunale ME. A laboratory evaluation of the Level 1 rapid
infuser (H1025) and the Belmont Instrument fluid management
system (FMS 2000) for rapid transfusion. Anesth Analg 2003;97:
1064 9.

PAIN MEDICINE
SECTION EDITOR
CHRISTOPH STEIN

Antinociception of Intrathecal Adenosine Receptor Subtype

Agonists in Rat Formalin Test
Myung Ha Yoon,

MD,

Hong Beom Bae,

MD,

and Jeong Il Choi,

MD

Department of Anesthesiology and Pain Medicine, Chonnam National University, Medical School, Gwangju, Korea

Adenosine has shown antinociceptive action via spinal

adenosine receptors. There are four types of adenosine
receptors: A1, A2A, A2B, and A3. We characterized the
nature of types of adenosine receptors for the control of
nociception at the spinal level. For nociception, formalin solution (5%, 50 L) was injected into the hindpaw
of male Sprague-Dawley rats. The effects of intrathecal
adenosine A1 (CPA), A2A (DPMA), and A3 (IB-MECA)
receptor agonists were examined. CPA and IB-MECA
produced limited or no effect on the early phase response of the formalin test, respectively, but the two

n endogenous purine compound adenosine

functions as an extracellular signaling molecule
within the central and peripheral nervous systems (1,2). Adenosine is released locally at tissue sites
in response to adverse events such as trauma and
ischemia and interacts with specific receptors. Four
types of adenosine receptors have been identified and
cloned as A1, A2A, A2B, and A3 (1,3). Experimental
data have elucidated the role of adenosine in the modulation of nociceptive transmission at the spinal level
(4 7). Although many pharmacologic studies (8 13)
have investigated the effect of adenosine receptor agonists for the nociceptive state, the role of types of
adenosine receptors at the spinal level have not been
definitely established.
The purpose of the present study was to understand
the physiological relevance of subtypes of adenosine
receptors in the control of nociception at the spinal
level. Thus, we examined the effects of several selective adenosine receptor agonists given intrathecally in
the formalin test, which shows an early phase of acute
Supported, in part, by Research Institute of Medical Science,
Chonnam National University.
Accepted for publication May 25, 2005.
Address correspondence and reprint requests to Myung Ha Yoon,
MD, Professor in Department of Anesthesiology and Pain Medicine,
Chonnam National University, Medical School, 8 Hakdong,
Dongku, Gwangju 501757, Korea. Address e-mail to mhyoon@
chonnam.ac.kr.
DOI: 10.1213/01.ANE.0000180994.10087.6F
2005 by the International Anesthesia Research Society
0003-2999/05

drugs depressed the late phase response. DPMA suppressed both phase responses. CPA was the most potent drug among the three in the late phase. These results suggest that spinal adenosine A1 and A2A
receptors may be involved in the modulation of the
early and the late phase responses of the formalin test,
whereas adenosine A3 receptor may be involved in the
regulation of the late phase response.
(Anesth Analg 2005;101:141721)

nociceptive response followed by a late phase response being related to more complex inflammatory
reactions. Further, we sought to determine the antinociceptive potency of these agonists under the same
nociceptive conditions.

Methods
All animal protocols were reviewed and approved by
The Institutional Animal Care Committee of the Research Institute of Medical Science at Chonnam National University. Adult Male Sprague-Dawley rats
(250 300 g) were housed in groups of 4 in standard
clear plastic cages and maintained in a temperaturecontrolled room (20C 1C) on a 12-h night/day
cycle. Food and water were provided at ad libitum.
Rats were implanted with chronic intrathecal catheters under enflurane anesthesia according to a
method described elsewhere (14). A midline incision
was made over the atlantooccipital junction. Each
polyethylene-10 catheter extended from the cisterna to
the rostral edge of the lumbar enlargement and was
externalized through the anterior part of the scalp. The
outer end of the catheter was plugged with a steel wire
and the skin was closed with 3 0 silk sutures. Only
rats with normal motor function were used; the others
were killed by volatile anesthetic overdose. After recovery from anesthesia, animals were individually
housed in cages and monitored for at least 4 5 days
before experiments.
Anesth Analg 2005;101:141721

1417

1418

PAIN MEDICINE YOON ET AL.

ADENOSINE RECEPTOR SUBTYPES AND ANTINOCICEPTION

ANESTH ANALG
2005;101:141721

Drugs used in this study were as follows: 2-chloroN6-cyclopentyladenosine (CPA, Research Biochemical
Internationals [RBI], USA), DPMA (RBI) and IBMECA (Tocris Cookson Ltd., UK). All the drugs were
dissolved in dimethylsulfoxide and intrathecally administered using a hand-driven, gear-operated syringe in a volume of 10 L solution followed by an
additional 10 L of saline to flush the catheter.
The formalin test was used to measure pain state.
Formalin 50 L 5% solution was injected subcutaneously into the plantar aspect of the hindpaw using a
30-gauge needle. Formalin injection produces the specific behavior of flinching/shaking of the affected
paw. This formalin-induced behavior was regarded as
a pain response and observed for 60 min. The number
of flinching/shaking response was counted for 1-min
periods at 1 to 2 min and 5 to 6 min and at intervals of
5 min from 10 to 60 min. The flinching response is
typically observed in two phases after formalin injection. Thus, the early and late phases of the formalin
test were defined as the period of time immediately
after injection of formalin until 10 min or 10 to 60 min
after formalin injection, respectively. Rats were killed
by volatile anesthetic overdose at the end of the formalin test.
Rats were placed in a restraining cylinder 4 5 days
after surgery for the study. After a habituation period
of 20 min, rats were assigned to one of the drug
treatment groups. Dimethylsulfoxide was used as a
control (n 6). Ninety-eight rats were used and each
group comprised 6 8 rats. Rats received only one
dose of drug. The formalin test was performed only
once in each rat.
After intrathecal administration of adenosine agonists, motor function was assessed by placingstepping and righting reflexes (n 15). The former
was assessed by placing the rat horizontally with its
back on the table, which normally causes an immediate coordinated twisting of the body to an upright
position. The latter was evaluated by drawing the
dorsum of either hindpaw of the rat across the edge of
the table, which normally causes the rat to try to put
the paw ahead into a position to walk. Motor function
was measured at 5, 10, 20, 30, 40, 50, and 60 min after
intrathecal administration of adenosine agonists at
maximum doses used in this study.
For evaluation of the time course and dose-response
of the effect of adenosine receptors agonists, A1 agonist (CPA, 0.3, 0.8, 2.7 nmol), A2A agonist (DPMA, 5.8,
19.2, 57.6, 191.8 nmol), and A3 agonist (IB-MECA, 19.6,
58.8, 196, 587.9 nmol) were intrathecally administered
10 min before the formalin injection. On the other
hand, intrathecal CPA resulted in motor disturbance
at 8.1 nmol; therefore we used 2.7 nmol of CPA as a
maximal dose. Each ED50 value (effective dose producing a 50% reduction of control formalin response)
of the three drugs was separately determined.

Data are expressed as mean sem. The time response data are presented as the number of flinches.
The dose-response data are presented as the sum of
the number of flinches in each phase. To calculate the
ED50 values of each drug, the number of flinches was
converted to percentage of control as follows: % of
control (Sum of flinching number with drug in
phase 1 [2])/(Sum of flinching number in control
phase 1 [2]) 100.

% of control

Sum of flinching number

with drug in phase 1 2
Sum of flinching number
in control phase 1 2

100

To compare the potency, each ED50 and 95% confidence intervals (CI) in 2 phases were estimated according to Tallarida and Murray (15). Dose response
data were analyzed by one-way analysis of variance
with Scheffe testing for post hoc. The level of statistical
significance was set at P 0.05.

Results
Subcutaneous injection of formalin into the plantar
region of the hindpaw resulted in a biphasic flinching
response in the injected paw. The time course effects
of intrathecal CPA, DPMA, and IB-MECA are shown
in Figure 1. Intrathecal CPA produced a limited (approximately 54% of control) suppression of the early
phase response of the formalin test, while it produced
a dose-dependent suppression of the late phase response (Fig. 2). Intrathecal DPMA dose-dependently
blocked the flinching response during the early and
the late phases of the formalin test (Fig. 2). Intrathecal
IB-MECA reduced the late phase response without
affecting the early phase response (Fig. 2).
The calculated ED50 values with 95% CI of CPA,
DPMA, and IB-MECA in the early or late phase are
shown in Table 1. The rank order of potency (defined
by ED50 in nmol) of the late phase in the formalin test
was as follows: CPA DPMA IB-MECA.
Placing-stepping and righting reflexes were normal
after intrathecal delivery of CPA, DPMA, and IBMECA at maximum doses and no other motor dysfunction was observed.

Discussion
The behavioral pain response to formalin injection is
characterized by two phases corresponding to basically different processes, with the early short-lasting
phase (phase 1) of acute pain and the late phase of
prolonged pain (phase 2). The phase 1 response results

ANESTH ANALG
2005;101:141721

Figure 1. Time course curves of intrathecal CPA (A), DPMA (B), and
IB-MECA (C) for the flinching response in the formalin test. Drugs
were administered 10 min before formalin (F) injection. Data are
presented as the number of flinches. Each point on the graph
represents the mean sem of 6 8 rats. *P 0.05 versus control.

essentially from the direct stimulation of nociceptors

of the primary afferent. The phase 1 response is therefore considered to correspond to the high level of
activity in the primary afferent. On the other hand, the
phase 2 response seems to originate from the continuous low level of small afferent input. The afferent
input generated by formalin is believed to be from
release of glutamate and substance P, which initiate a
cascade through N-methyl-d-aspartate and neurokinin1 receptors. The resulting cascade leads to a state of
facilitation that appears to be more than anticipated,
considering the diminished level of afferent input (16).
In the present study, intrathecal CPA (adenosine A1
receptor agonist) had a limited effect on the early
phase response of the formalin test but decreased the
late phase response. Intrathecal DPMA (adenosine
A2A receptor agonist) attenuated formalin-induced
pain behavior during both phases. Intrathecal IBMECA (adenosine A3 receptor agonist) suppressed the
flinching response during the late phase but not during the early phase. These observations suggest that
adenosine A1 and A2A receptors may be minimally or

PAIN MEDICINE
YOON ET AL.
ADENOSINE RECEPTOR SUBTYPES AND ANTINOCICEPTION

1419

Figure 2. Dose response curves of intrathecal CPA, DPMA, and

IB-MECA for the flinching response during phase 1 (A) and phase
2 (B) in the formalin test. Data are presented as the percentage of
control. CPA produced a limited reduction of the flinching response
during phase 1 in the formalin test, but it dose-dependently decreased the flinching response during phase 2. DPMA produced a
dose-dependent inhibition of flinching response in both phases in
the formalin test. IB-MECA did not reduce the flinching response
during phase 1 in the formalin test, but it dose-dependently decreased the flinching response during phase 2. Each point on the
graph represents the mean sem of 6 8 rats. C control. *P 0.05,
P 0.01, P 0.001 versus control.

Table 1. ED50 Values (nmol) with 95% Confidence

Intervals of Intrathecal Drugs
ED50 (95% CI)

CPA
DPMA
IB-MECA

Early Phase

Late Phase

4.7 (0.827.4)
19.5 (10.635.6)

1.2 (0.72)
55.6 (27114.3)
207.1 (100427.9)

ED50 effective dose needed to show a 50% inhibition of control formalin

flinching response; CI confidence interval; CPA 2-chloro-N6-cyclopentyladenosine; DPMA N-[2-(3,5-dimethoxyphenyl)-ethyl]adenosine; IB-MECA
N6-(3-iodobenzyl)ADO-5N-methyluronamide.

actively involved in the modulation of acute nociception in the spinal cord, respectively. In contrast, the
spinal adenosine A3 receptor may not contribute to the
control of acute nociception. On the other hand, spinal
adenosine A1, A2A and A3 receptors may play a critical
role in the modulation of the inflammatory process
occurring during formalin pain.
Adenosine may play an important role in the modulation of nociceptive inputs through adenosine A1
and A2 receptors identified in the dorsal horn of the
spinal cord (7,17). It has been reported that intrathecal

1420

PAIN MEDICINE YOON ET AL.

ADENOSINE RECEPTOR SUBTYPES AND ANTINOCICEPTION

ANESTH ANALG
2005;101:141721

adenosine A1 receptor agonists attenuated not only

the inflammatory hyperalgesia but also acute nociception (8,9,10,12). Moreover, adenosine A1 receptor agonist inhibited excitatory transmission in the spinal
cord (18). Thus, the role of spinal adenosine A1 receptor in this study is in accordance with results from
previous studies. Meanwhile, intrathecal adenosine
A2A receptor agonist failed to modify the frequency of
flinching/lifting during the entire formalin test, but it
induced a limited amount of suppression of the late
phase licking/biting responses without affecting early
phase responses (8). Moreover, spinal adenosine A2A
receptor agonist produced a modest antinociception in
the inflammatory thermal hyperalgesia model (10).
However, intraperitoneal adenosine A2A receptor agonist counteracted the flinching response induced by
the formalin test only during the early phase but not
the late phase (11). Although the activation of adenosine
A2A receptor exerts different actions, it is not clear if the
different effects observed could be, at least in part, ascribed to the difference in drugs, the activation of the
receptor, the route of drugs, or the different types of
tested stimuli. Adenosine receptor activation in the spinal cord is proposed to produce an antinociception by at
least two distinct mechanisms: presynaptic inhibition of
excitatory neurotransmitter release with subsequent reduction of substance P concentration in cerebrospinal
fluid (9) and postsynaptic inhibition of the effects of
excitatory neurotransmitters (19). The above findings
jointly suggest the adenosine A2A receptor contributing
to antinociceptive action in the spinal cord.
Unfortunately, we did not evaluate the role of adenosine A2B receptor for the modulation of nociception
because there are no available adenosine A2B-selective
agonists.
Subcutaneous administration of adenosine A3 receptor agonist produced nociceptive behavior (13).
However, stimulation of spinal adenosine A3 receptor
possibly exerted an inhibitory influence on the release
of pain-related neuropeptide in the spinal cord (20),
which may have contributed to the antinociceptive
role for spinal adenosine A3 receptor in the control of
nociception. Moreover, spinal adenosine A3 receptor
agonist reduced the late phase response of the formalin test in the present study. This is the first report of
the antinociceptive role of spinal adenosine A3 receptor in formalin-induced inflammatory hyperalgesia.
In the current study, intrathecal CPA showed a
minimal effect (54% of control) in phase 1 and IBMECA also had no effect in the same condition. Furthermore, intrathecal CPA caused motor dysfunction
at 8.1 nmol. Therefore, we could not compare the
potency of the above three drugs during phase 1 in the
formalin test. However, maximum of % of control of
intrathecal CPA, DPMA, and IB-MECA was similar in
phase 2, which made it possible to compare their
potency. Thus, the rank order of potency in phase 2

was CPA DPMA IB-MECA at the spinal level.

These observations suggest that spinal adenosine A1
receptor may be more involved than adenosine A2A
and A3 receptors in the modulation of the inflammatory hyperalgesia. Therefore, agonists for adenosine
A1 receptor other than adenosine A2A and A3 receptors may be effective in the management of the inflammatory hyperalgesia in the spinal cord. One interesting finding was the relative effectiveness of DPMA on
phase 1 and 2 responses of the formalin rest. A threefold larger ED50 for phase 2 was observed compared
with that for phase 1 with DPMA. These data suggest
that adenosine A2A receptor seems to be much more
effective on early pain than on inflammatory hyperalgesia, which in turn supports the theory that agonists
for adenosine A1 receptor may be useful in the treatment of acute pain in the spinal cord.
In conclusion, adenosine A1 and A2A receptors but
not adenosine A3 receptor exhibit an antinociceptive
profile in acute nociception in the spinal cord. However, all adenosine A1, A2A and A3 receptors are involved in the control of a formalin-induced inflammatory hyperalgesia.

References
1. Ralevic V, Burnstock G. Receptors for purines and pyrimidines.
Pharmacol Rev 1998;50:41392.
2. Williams M, Jarvis MF. Purinergic and pyrimidinergic receptors
as potential drug targets. Biochem Pharmacol 2000;59:1173 85.
3. Klotz KN. Adenosine receptors and their ligands. Naunyn
Schmiedebergs Arch Pharmacol 2000;362:38291.
4. Sawynok J, Sweeney MI. The role of purines in nociception.
Neuroscience 1989;32:557 69.
5. Sawynok J. Adenosine receptor activation and nociception. Eur
J Pharmacol 1998;347:111.
6. Furst S. Transmitters involved in antinociception in the spinal
cord. Brain Res Bull 1999;48:129 41.
7. Sawynok J. Purines in pain management. Curr Opin CPNS
Invest Drugs 1999;1:2738.
8. Poon A, Sawynok J. Antinociception by adenosine analogs and
an adenosine kinase inhibitor: dependence on formalin concentration. Eur J Pharmacol 1995;286:177 84.
9. Sjolund KF, Sollevi A, Segerdahl M, Lundeberg T. Intrathecal
adenosine analog administration reduces substance P in cerebrospinal fluid along with behavioral effects that suggest antinociception in rats. Anesth Analg 1997;85:62732.
10. Poon A, Sawynok J. Antinociception by adenosine analogs and
inhibitors of adenosine metabolism in an inflammatory thermal
hyperalgesia model in the rat. Pain 1998;74:235 45.
11. Borghi V, Przewlocka B, Labuz D, et al. Formalin-induced pain
and mu-opioid receptor density in brain and spinal cord are
modulated by A1 and A2a adenosine agonists in mice. Brain Res
2002;956:339 48.
12. Boyle DL, Moore J, Yang L, et al. Spinal adenosine receptor
activation inhibits inflammation and joint destruction in rat
adjuvant-induced arthritis. Arthritis Rheum 2002;46:3076 82.
13. Sawynok J, Zarrindast MR, Reid AR, Doak GJ. Adenosine A3
receptor activation produces nociceptive behaviour and edema
by release of histamine and 5-hydroxytryptamine. Eur J Pharmacol 1997;333:17.
14. Yaksh TL, Rudy TA. Chronic catheterization of the spinal subarachnoid space. Physiol Behav 1976;17:1031 6.
15. Tallarida RJ, Murray RB. Manual of pharmacologic calculations
with computer programs. New York: Springer-Verlag, 1987.

1421

ANESTH ANALG
2005;101:141721

PAIN MEDICINE
YOON ET AL.
ADENOSINE RECEPTOR SUBTYPES AND ANTINOCICEPTION

16. Yaksh TL. Preclinical models of nociception. In: Yaksh TL,

Lynch C III, Zapol WM, et al., eds. Anesthesia: biologic foundations. Philadelphia: Lippincott-Raven, 1997:685718.
17. Choca JI, Green RD, Proudfit HK. Adenosine A1 and A2 receptors of the substantia gelatinosa are located predominantly on
intrinsic neurons: an autoradiography study. J Pharmacol Exp
Ther 1988;247:757 64.
18. Lao LJ, Kumamoto E, Luo C, et al. Adenosine inhibits excitatory
transmission to substantia gelatinosa neurons of the adult rat

spinal cord through the activation of presynaptic A(1) adenosine receptor. Pain 2001;94:31524.
19. DeLander GE, Wahl JJ. Behavior induced by putative nociceptive neurotransmitters in inhibited by adenosine or adenosine
analogs coadministered intrathecally. J Pharmacol Exp Ther
1988;246:56570.
20. Mauborgne A, Polienor H, Hamon M, et al. Adenosine receptormediated control of in vitro release of pain-related neuropeptides from the rat spinal cord. Eur J Pharmacol 2002;441:4755.

The Neurological Safety of Epidural Gabapentin in Rats:

A Light Microscopic Examination
Sang-Sik Choi, MD*, Yong-Chul Kim,
Pyung-Bok Lee, MD*, Sang-Chul Lee,

MD*,
MD*,

Young Jin Lim, MD*, Chul-Joong Lee, MD*,

Woo-Seok Sim, MD, and Yoon-La Choi, MD

*Department of Anesthesiology and Pain Medicine, Seoul National University College of Medicine; and Departments of
Anesthesiology and Pain Medicine and Diagnostic Pathology, SungKyunKwan University College of Medicine, Seoul,
Korea

Gabapentin acts primarily on the central nervous system. Therefore, we hypothesized that the direct epidural administration of gabapentin could have various advantages over its oral administration with
respect to required dose, side effects, and efficacy.
However, before administering gabapentin into the
epidural space in a clinical setting, its neurotoxicity
must be examined in animals. Thus, we evaluated
neurotoxicity of epidural gabapentin by observing
behavioral and sensory-motor changes, and by histopathological examinations of spinal cords and dorsal root ganglia in the rat. Twenty-seven rats were
randomly divided into 3 groups, which were administered 0.3 mL (30 mg) of epidural gabapentin (group

he effects of spinoaxial or oral gabapentin on

neuropathic pain have been well established by
animal (1 4) or clinical studies (57). The mechanism of gabapentin is unclear, but may involve binding to the spinal 2 calcium channel subunit (8). Previous studies had shown that the mRNAs for the 2
subunits are expressed at high levels in sensory neurons of dorsal root ganglion and in the spinal cord
dorsal horn (9,10). In addition, binding studies have
shown that the 2-1 and 2-2 subunits bind gabapentin with high affinities (8).
In a clinical setting, the main purpose of spinoaxial
administration of drugs is to reduce the dosage, maximize efficacy, and minimize side effects when the
action site of the drug involved is the central nervous
system. In addition, pain clinicians prefer the epidural

Accepted for publication April 25, 2005.

Address correspondence and reprint requests to Chul-Joong Lee,
MD, Department of Anesthesiology and Pain Medicine, Seoul
National University, College of Medicine, 28 Yeongeon-dong,
Chongno-ku, Seoul, 110-744 Korea. Address e-mail to may97lee@
yahoo.com.kr.
DOI: 10.1213/01.ANE.0000180197.32577.9B

1422

Anesth Analg 2005;101:14226

G, n 9) and the same volume of epidural alcohol

(group A, n 9) or normal saline (group N, n 9). No
rats in groups G and N showed sensory-motor dysfunction, behavioral change, or histopathological abnormalities over a 3-wk observation period, whereas
all rats in group A showed abnormalities. We conclude that the direct epidural injection of gabapentin
in rats did not show any neurotoxic evidence in terms
of sensory-motor functions and behavior, or by a microscopic histopathological evaluation. This study
represents a first promising step toward the trial of
epidural gabapentin in a clinical setting.
(Anesth Analg 2005;101:14226)

administration of analgesics to intrathecal administration because it is less invasive and is likely to reduce
the risk of neurotoxicity, especially when the administration is conducted over an extended period (11).
These observations led us to hypothesize that the spinoaxial administration of gabapentin could have
many advantages over its oral administration in a
clinical setting.
After the introduction of spinoaxial catheterization
at the end of 19th century, various drugs were injected
clinically or experimentally via the spinoaxial route.
Moreover, some of these drugs were either neurotoxic
or potentially neurotoxic. In addition, preservatives
added to drugs may be neurotoxic when injected spinoaxially (12). Therefore, if a drug is to be administered intrathecally or epidurally in a clinical setting, its
safety must be determined initially by preliminary
animal studies, because any toxic effect could have
serious consequences (13). However, no animal studies have been performed on the possible neurotoxicity
of spinoaxially injected gabapentin despite its excellent analgesic effect, and thus we undertook the
present study to evaluate the neurotoxicity of epidurally injected gabapentin in a rat model.
2005 by the International Anesthesia Research Society
0003-2999/05

ANESTH ANALG
2005;101:14226

PAIN MEDICINE
CHOI ET AL.
THE SAFETY OF EPIDURAL GABAPENTIN

1423

Table 1. Changes in Body Weight After Epidural Drug Injection

Weight

Days after
injection

Group N

Group G

Group A

P value

Pre (n 9)
2nd day (n 9)
7th day (n 6)
21st day (n 3)

271 30
291 24
330 16
385 4

284 18
292 19
308 19
351 27

291 35
285 34
287 34*
301 7*

0.334
0.906
0.028
0.004

Values are mean sd.

Group G: 0.3 mL (30 mg) of epidural gabapentin, and groups A and N: the same volume of epidural alcohol and normal saline, respectively. Pre before
epidural injection of test drugs.
P 0.05 versus Pre.

Methods
The experimental protocol used was reviewed and
approved by our institutional Animal Care and Use
Committee. All rats had free access to food and water
and were individually housed under a 12-h light/dark
cycle for 1 wk.
Anesthesia was induced by placing a rat in a closed
box containing 4% enflurane in oxygen (3 L/min) with
spontaneous ventilation. After loss of consciousness,
anesthesia was maintained with 2%3% enflurane via
a loose-fitting mask. After applying a sterile dressing,
epidural catheterization was performed, as we described previously (14,15) with some modification.
Briefly, a 3-cm midline skin incision was made at the
T13L1 intervertebral space. Using a fine microscissors, a small hole was made at the center of the ligament flavum and a PE-10 catheter (Natsume, Japan)
was inserted and gently advanced about 3 cm caudally. The catheter tip was placed in the space between L4 and L5. Cases were excluded if blood or
cerebrospinal fluid was aspirated. A drop of cyanoacrylate (Aron-Alpha, Toagosei, Japan) was applied at
the epidural catheter entry site. To confirm correct
catheter positioning, we injected 0.15 mL of 2% lidocaine through the catheter after complete recovery
from anesthesia, and we defined a correct epidural
catheter placement as one that showed paralysis of the
hindlimbs while the forelimbs retained normal motor
power. If the test solution was accidentally injected
intrathecally or IV, sudden respiratory arrest with or
without cardiac arrest was observed; such cases were
excluded from the study. The fascia and skin were
sutured and antibiotic ointment was applied. After
confirming correct epidural catheter placement, we
examined gait, spinal deformity, and behavioral abnormalities for 3 days. If the rats showed abnormal
findings during the 3-day observation period, they
were excluded from this study.
Twenty-seven male Sprague-Dawley rats, weighing
250 350 g, were successfully prepared for this study,
and these rats were divided equally into 3 groups.
Under general anesthesia, 30 mg (0.3 mL, 100 mg of
gabapentin dissolved in 1 mL of distilled water) of
preservative-free gabapentin (Sigma-Aldrich, St.

Table 2. Evaluation of Pinch-Toe Test After Epidural

Drug Injection
Days after injection Group N Group G
Pinch-toe test
2nd day (n 9)
7th day (n 6)
21st day (n 3)
Motor deficita
2nd day (n 9)
7th day (n 6)
21st day (n 3)

Group A

8 (89)*
6 (100)*
3 (100)*

8 (89)*
6 (100)*
3 (100)*

Values are expressed as number (%) of abnormal rats.

Group G: 0.3 mL (30 mg) of epidural gabapentin, and groups A and N: the
same volume of epidural alcohol and normal saline, respectively.
a
Grade 1 normal gait with no evidence of motor paresis; grade 2
normal gait with slight hindpaw deformity, such as plantar flexion of toes;
grade 3 slight gait disturbance with motor weakness and/or an inverted
hindpaw; and grade 4 prominent limping gait with a dropped hindpaw.
The rats with a motor disturbance of grade 2 or above were considered to
have a motor deficit (17).
* P 0.05 versus groups N and G.

Louis, MO) was injected via an epidural catheter in

group G (n 9). In groups A (n 9) and N (n 9), the
same volume of 40% alcohol and 0.9% normal saline
were injected via an epidural catheter, respectively.
After recovery from anesthesia, rats were individually
housed under a 12-h light/dark cycle.
In the pilot study, we evaluated the extent of
spreading of contrast medium by fluoroscopy, and
found that the total spreading spinal segments of
0.3 mL of contrast medium was 10 or 11, which was
enough to affect the entire spinal cord segment
cropped for histopathological examination.
The epidural gabapentin dosage was determined in
this study by observing the development of behavioral
changes, e.g., motor deficit, crying, or aggressive behavior. Epidural gabapentin was started from 0.3 mg
(0.3 mL, 1 mg of gabapentin was dissolved in 1 mL of
distilled water), which was the maximal amount injected intrathecally in previous studies (2 4). We did
not find any motor deficit of grade 2 or more or any
behavioral changes at the dose of 30 mg (0.3 mL,
100 mg of gabapentin dissolved in 1 mL of distilled
water), which was 100 times the previously reported
maximal intrathecal dose (2 4).

1424

PAIN MEDICINE CHOI ET AL.

THE SAFETY OF EPIDURAL GABAPENTIN

ANESTH ANALG
2005;101:14226

Table 3. Neuropathological Findings of Spinal Cords and Nerves Under Light Microscopic Examination After Epidural
Drug Injection
Group N

Dural hypertrophy
Synechia
Local neuritis
Meningeal inflammation
Local myelopathy
Myelin loss
Peripheral neuropathy
Local Infarction

Group G

Group A

2nd
(n 3)

7th
(n 3)

21st
(n 3)

2nd
(n 3)

7th
(n 3)

21st
(n 3)

2nd
(n 3)

7th
(n 3)

21st
(n 3)

2
2

2
2
2
2

2
3*

2
2
3*
3*

3*
3*

Values are expressed as number of positive animals.

Group G: 0.3 mL (30 mg) of epidural gabapentin, and groups A and N: the same volume of epidural alcohol and normal saline, respectively. 2nd, 7th, and
21st: days after epidural injection of test drugs.
* P 0.05 versus corresponding data of groups N and G.

Acute toxicity was evaluated 2 days after injection

and chronic toxicity 7 and 21 days after injection. One
examiner unaware of the study details observed motor and sensory deficits. Pinch-toe testing and motorfunction evaluation were started 2 days after drug
injection to exclude the possibility of the systemic
effect of gabapentin. Pinch-toe testing was used to
evaluate motor and sensory deficits (16). Motor function was assessed using a previously devised scoring
system with some modification (17). Grades were defined as follows: grade 1 normal gait with no evidence of motor paresis; grade 2 normal gait with
slight hindpaw deformity, such as plantar flexion of
toes; grade 3 slight gait disturbance with motor
weakness and/or an inverted hindpaw; and grade 4
a prominent limping gait with a dropped hindpaw.
The rats with a motor disturbance of grade 2 or above
were considered to a have motor deficit.
The spinal cord was cropped in 3 rats of each group
2 days after injection to evaluate acute toxicity. To
evaluate chronic toxicity, the spinal cords of the remaining 6 rats/group were cropped on the 7th (3
rats/group) and on the 21st days (3 rats/group) after
injection. Rats were killed under general anesthesia (as
described above), by transcardial perfusion with 4%
paraformaldehyde in 0.1 M phosphate buffer. Approximately 1-cm lengths of spinal cord, caudal and
rostral to the catheter tip, were obtained. For light
microscopy, spinal cords were fixed with 10% neutral
formalin solution. Using standard tissue slide preparation methods, three blocks were made per sample.
The slides were prepared from 4- to 5-m sections
stained with hematoxylin and eosin. To evaluate myelin loss, Luxol fast blue (a specific myelin stain) was
added. Microscopic tissue analysis results (15,18,19)
were classified in 7 ways, namely as, dural hypertrophy, adhesion of meninges to surrounding tissues,
local neuritis, meningeal inflammation, local myelopathy, myelin loss, and peripheral neuropathy. One

pathologist blinded to this study analyzed histological

change.
Intergroup comparisons of body weights were analyzed using Kruskal-Wallis one-way ANOVA on
ranks and this was followed by Dunns method. Motor and sensory deficit and histopathological spinal
cord data were analyzed using Fishers exact test with
a Bonferroni correction. A P value 0.05 was considered significant.

Results
All rats in groups N and G showed normal behavior
throughout the study period, whereas all rats in group
A showed both reduced activity and appetite. Rats in
group A showed significantly low body weights 7 and
21 days after injection (P 0.05, Table 1). All rats in
groups N and G responded normally to pinch-toe
testing and had a normal gait at each observation
point. However, all rats in group A, except one rat on
day 2 after injection, showed an abnormal response to
pinch-toe testing. These animals also showed hindpaw deformity, and a gait disturbance of grade 2 or
more (P 0.05, Table 2).
No histological lesions on hematoxylin and eosin
and Luxol fast blue stains were observed in groups N
or G at any time. However, in group A, 2 of the 3
spinal cords obtained 2 days after injection showed
myelin loss and peripheral neuropathy and all of
those obtained 7 and 21 days after injection showed
various neuropathies (P 0.05, Table 3 and Figs. 1 and
2).

Discussion
We administered 30 mg of preservative-free gabapentin into the epidural space. In humans, this 30 mg

ANESTH ANALG
2005;101:14226

PAIN MEDICINE
CHOI ET AL.
THE SAFETY OF EPIDURAL GABAPENTIN

1425

Figure 1. The light micrographic findings of spinal cords or nerves on 2 days

(A, D, and G), 7 days (B, E, and H), and
21 days (C, F, and I) after epidural
injection of alcohol (AC), gabapentin
(DF), and normal saline (GI). Hematoxylin and eosin stain (200). In the
epidural alcohol group, peripheral
neuropathy and myelin loss in the spinal nerve (A) and lymphocytic infiltration in the spinal cords (B) are seen.
Lymphocytic infiltration, dural hypertrophy, synechia, and meningeal inflammation in the spinal cords (C) are
seen. In epidural gabapentin or normal
saline injection groups, no inflammation or neuropathy are seen.

Figure 2. The light micrographic findings of the spinal cords or

nerves after epidural injection of alcohol (A and B), gabapentin (C
and D), and normal saline (E and F). Luxol fast blue stain. The
right-side figures (200) are the quadrangle of the adjacent left-side
figures (40), respectively. In the epidural alcohol group, pale
myelins with the presence of many vacuoles of spinal nerve are seen
(B). In the epidural gabapentin or normal saline injection groups, no
abnormal morphologies of myelin are seen.

(approximately 100 mg/kg) is equivalent to an epidural dose of 6000 mg. Moreover, the required dose
for the epidural route is about 1/30 of that required
for oral administration (20). Thus, the 30 mg of epidural gabapentin administered may be equivalent to
an oral administration of about 180,000 mg for a human adult. Because the maximal recommended oral
dose for gabapentin is 3600 mg in human adults, and
because the dura mater of small animals has greater
diffusibility than that of the human (21), we decided
that a 30-mg dose of epidural gabapentin was sufficient for neurotoxicity evaluation purposes in the rat.
In the present study, all rats in group A showed
reduced activity and appetite, poor weight gain, and
acute and chronic neurotoxicity findings on histopathological examination, all of which might be considered as a sequela of alcohol-induced neurotoxicity
(2224). However, all rats in group G, similar to group
N, showed no motor or sensory deficits and no abnormal histopathological findings at any time.
The symptoms of neuronal damage might be
present even in the absence of histological change (25).
Thus, neurotoxicity should also be assessed by examining sensory, motor, and behavioral changes. In the
present study, no findings of neuronal damage, such
as persistent motor function changes or sensory losses,
were observed in rats administered gabapentin
epidurally.
In this study, spinal cord neurotoxicity was determined by light microscopy. The use of electron microscope with morphometric methods for analyzing cell
loss might have provided more information about the

1426

PAIN MEDICINE CHOI ET AL.

THE SAFETY OF EPIDURAL GABAPENTIN

safety of epidural gabapentin. We are considering a

further study including electron microscopy.
In conclusion, gabapentin administered epidurally
to rats did not cause sensory, motor, behavioral, or
histopathological changes that might suggest neurotoxicity, which implies that gabapentin could be directly administered epidurally in a clinical setting.
However, similar studies on different animals are required to obtain reliable safety data on the epidural
application of gabapentin before use in a clinical
setting.

References
1. Shimoyama N, Shimoyama M, Davis AM, et al. Spinal gabapentin is antinociceptive in the rat formalin test. Neurosci Lett
1997;222:657.
2. Jun JH, Yaksh TL. The effect of intrathecal gabapentin and
3-isobutyl gamma-aminobutyric acid on the hyperalgesia observed after thermal injury in the rat. Anesth Analg 1998;86:
348 54.
3. Yoon MH, Yaksh TL. The effect of intrathecal gabapentin on
pain behavior and hemodynamics in the formalin test in the rat.
Anesth Analg 1999;89:434 9.
4. Cho HS, Kim MH, Choi DH, et al. The effect of intrathecal
gabapentin on mechanical and thermal hyperalgesia in neuropathic rats induced by spinal nerve ligation. J Korean Med Sci
2002;17:2259.
5. Caraceni A, Zecca E, Bonezzi C, et al. Gabapentin for neuropathic cancer pain: a randomized controlled trial from the Gabapentin Cancer Pain Study Group. J Clin Oncol 2004;22:2909 17.
6. Levendoglu F, Ogun CO, Ozerbil O, et al. Gabapentin is a first
line drug for the treatment of neuropathic pain in spinal cord
injury. Spine 2004;29:74351.
7. Bennett MI, Simpson KH. Gabapentin in the treatment of neuropathic pain. Palliat Med 2004;18:511.
8. Marais E, Klugbauer N, Hofmann F. Calcium channel 2
subunits: structure and gabapentin binding. Mol Pharmacol
2001;59:1243 8.
9. Yusaf SP, Goodman J, Pinnock RD, et al. Expression of voltagegated calcium channel subunits in rat dorsal root ganglion
neurons. Neurosci Lett 2001;311:137 41.
10. Luo ZD, Calcutt NA, Higuera ES, et al. Injury type-specific
calcium channel 2-1 subunit up-regulation in rat neuropathic
pain models correlates with antiallodynic effects of gabapentin.
J Pharmacol Exp Ther 2002;303:1199 205.

ANESTH ANALG
2005;101:14226

11. Eimerl D, Papir-Kricheli D. Epidural capsaicin produces prolonged segmental analgesia in the rat. Exp Neurol 1987;97:
169 78.
12. Malinovsky JM, Lepage JY, Cozian A, et al. Is ketamine or its
preservative responsible for neurotoxicity in the rabbit? Anesthesiology 1993;78:109 15.
13. Yaksh TL, Collins JG. Studies in animals should precede human
use of spinally administered drugs. Anesthesiology 1989;70:
4 6.
14. Kim YC, Lim YJ, Lee SC. Spreading pattern of epidurallyadministered contrast media in rabbits. Acta Anaesthesiol
Scand 1998;42:10925.
15. Lim YJ, Sim WS, Kim YC, et al. The neurotoxicity of epidural
hyaluronic acid in rabbits: a light and electron microscopic
examination. Anesth Analg 2003;97:1716 20.
16. Bajrovic F, Sketelj J. Extent of nociceptive dermatomes in adult
rats is not primarily maintained by axonal competition. Exp
Neurol 1998;150:11521.
17. Kawakami M, Weinstein JN, Spratt KF, et al. Experimental
lumbar radiculopathy: immunohistochemical and quantitative
demonstrations of pain induced by lumbar nerve root irritation
of the rat. Spine 1994;19:1780 94.
18. Madsen JB, Jensen FM, Faber T, Bille-Hansen V. Chronic catheterization of the epidural space in rabbits: a model for behavioural and histopathological studies. Examination of meptazinol
neurotoxicity. Acta Anaesthesiol Scand 1993;37:30713.
19. Gurun MS, Leinbach R, Moore L, et al. Studies on the safety of
glucose and paraben containing neostigmine for intrathecal administration. Anesth Analg 1997;85:31723.
20. Rocco AG, Chan V, Iacobo C. Algorithm for the treatment of
pain in advanced cancer. Hosp J 1989;5:93103.
21. Hogan QH, Stadnicka A, Stekiel TA, et al. Mechanism of mesenteric venodilation after epidural lidocaine in rabbits. Anesthesiology 1994;81:939 45.
22. Singler RC. Alcohol neurolysis of sciatic and femoral nerves.
Anesth Analg 1981;60:5323.
23. Porges P, Zdrahal F. Intrathecal alcohol neurolysis of the lower
sacral roots in inoperable rectal cancer. Anaesthetist 1985;34:
6279.
24. Pelissier J, Viel E, Enjalbert M, et al. Chemical neurolysis using
alcohol (alcoholization) in the treatment of spasticity in the
hemiplegic. Cah Anesthesiol 1993;41:139 43.
25. Svensson BA, Alari L, Post C. Repeated intrathecal injections of
dezocine produce antinociception without evidence for neurotoxicity in the rat: a study of morphometric evaluation of spinal
cord histology. Anesth Analg 1992;75:3929.

Multimodal Analgesia with Gabapentin and Local Anesthetics

Prevents Acute and Chronic Pain After Breast Surgery
for Cancer
Argyro Fassoulaki, MD, PhD, DEAA*, Argyro Triga,
and Constantine Sarantopoulos, MD, PhD, DEAA

MD,

Aikaterini Melemeni,

MD*,

*Department of Anesthesiology, Aretaieion Hospital, Medical School, University of Athens; Department of

Anesthesiology, St. Savas Hospital, Athens, Greece; and Department of Anesthesiology, Medical College of Wisconsin,
Milwaukee

We evaluated the effect of multimodal analgesia on

acute and chronic pain after breast surgery for cancer.
Fifty patients scheduled for breast cancer surgery were
blindly randomized to receive gabapentin, eutectic
mixture of local anesthetics cream, and ropivacaine in
the wound or three placebos. Pain (visual analog scale)
and analgesics were recorded in the postanesthesia care
unit (PACU) 3, 6, and 9 h and 8 days after surgery. Three
and 6 mo later, patients were assessed for chronic pain.
The treatment group consumed less paracetamol in the
PACU (469 versus 991 mg; P 0.002) and less Lonalgal (1.0 versus 4.4 tablets; P 0.003) than the controls,
exhibited lower visual analog scale scores at rest in the
PACU (P 0.001) and on postoperative Days 1, 3, and 5

(P 0.040, P 0.015, and P 0.045, respectively), and

after movement in the PACU (P 0.001) and on postoperative Days 2, 4, and 8 (P 0.028, P 0.007, and P
0.032, respectively). Three and 6 mo after surgery, 18 of
22 (82%) and 12 of 21 (57%) of the controls reported
chronic pain versus 10 of 22 (45%) and 6 of 20 (30%) in
the treatment group (P 0.028 and P 0.424, respectively); 5 of 22 and 4 of 21 of the controls required analgesics versus 0 of 22 and 0 of 20 of those treated (P
0.048 and P 0.107, respectively). Multimodal analgesia reduced acute and chronic pain after breast surgery
for cancer.
(Anesth Analg 2005;101:142732)

evere trauma or surgery may be a cause of chronic

pain. Scars and injury of soft tissues and of sensory
afferents, which innervate the surgical area, may all
contribute to chronic pain. Breast surgery for cancer, thus
modified radical mastectomy or lumpectomy plus axillary
lymph node dissection, is associated with chronic pain
development referred to the area of mastectomy, the axilla,
and the proximal medial arm (1).
This chronic pain is persistent and burning or pricking
in nature. The incidence of chronic postmastectomy pain
reported by several studies varies significantly. In a retrospective cohort study, 43% of the patients suffered
from chronic postmastectomy pain, but the incidence in
young women was 65% (2). Chronic postmastectomy
pain may impair womens professional lives or other
activities. This is a major problem considering the number of women operated on for breast cancer.

Acute and chronic pain management after breast

cancer surgery has been investigated previously (3 6).
Three of the four studies also examined the impact of
analgesic regimens on chronic pain development. Although local anesthetics may have some beneficial
effect, regimens that provide an efficient postoperative
analgesia and reduce analgesic consumption for acute
pain relief do not prevent chronic pain development
to a satisfactory degree.
We hypothesized that multimodal analgesia may be
effective in preventing both acute and chronic postmastectomy pain. We investigated the effect of multimodal
analgesia, comprising of local anesthetics applied on the
surgical area and gabapentin administered orally, for
acute pain and analgesic requirements in patients operated on for breast cancer. We also investigated the impact of these drugs on chronic pain development and
sensation abnormalities.

Accepted for publication April 26, 2005.

Address correspondence and reprint requests to A. Fassoulaki,
MD, PhD, DEAA, 57-59 Raftopoulou St., 11744 Athens, Greece.
Address e-mail to afassou1@otenet.gr or fassoula@aretaieio.uoa.gr.

Methods

DOI: 10.1213/01.ANE.0000180200.11626.8E
2005 by the International Anesthesia Research Society
0003-2999/05

Approval from the Hospital Ethics Committee and

patients written informed consent were obtained.
Anesth Analg 2005;101:142732

1427

1428

PAIN MEDICINE FASSOULAKI ET AL.

ANALGESIA AND POSTMASTECTOMY PAIN

Fifty women, aged between 32 and 59 yr old and ASA

physical status I or II, admitted to St. Savas Hospital
for breast cancer surgery were enrolled in the study.
Eligibility criteria were age 59 yr and body weight
not exceeding the 20% of the ideal. Patients with
chronic pain of any cause, consuming analgesics, antidepressants, calcium channel blockers, antiepileptics, sedatives, or hypnotics during the last month or
radiotherapy or chemotherapy before surgery and
those with inoperable breast cancer were not recruited
for the study. Operations were performed by two
surgeons highly experienced in breast surgery. All
patients were assessed the day before surgery. They
were informed that some treatment might be helpful
in attenuating postoperative discomfort. The visual
analog scale (VAS) scoring pain was explained.
Each patient was randomly assigned to a control
group (A) or to a treatment group (B). Fifty envelopes,
25 containing odd and 25 containing even numbers,
obtained from a computer-generated table, were prepared and sealed. According to the odd or even number obtained after opening a sealed envelope, drawn
but not read by a nurse, patients were randomized to
Group A (control treated with placebo capsules, placebo
cream, and normal saline) or to Group B (treatment with
gabapentin, eutectic mixture of local anesthetics [EMLA]
cream, and ropivacaine solution), respectively.
The study was conducted in a double-blind manner.
Placebo capsules were identical in appearance with
the gabapentin capsules. The same number of capsules was packaged in group-specific bottles and
coded as bottle A and bottle B for the control and
treatment groups, respectively. A white odorless
cream was the control treatment corresponding to the
EMLA cream. Similarly, cream for each group was
kept in boxes labeled as A and B for the control and
treatment groups, respectively.
An independent anesthesiologist, who did not participate in the study or data collection, read the number contained in the envelope and made group assignments. She also prepared for the surgeons the
envelopes, bottles with capsules, boxes with creams,
and the solution of normal saline or ropivacaine in
identical syringes. Except for the independent anesthesiologist, no other physician or nursing staff member was aware of the interventions administered to
each patient.
To avoid motor impairment and possibly excessive
numbness interfering with the blinding of the study,
brachial plexus and intercostal nerves were not infiltrated with the local anesthetic but irrigated instead.
Regarding EMLA cream and possible interference
with blinding, EMLA or placebo was applied in the
morning after pain assessment, and patients were instructed to remove it 3 4 h later. By the next morning
when pain was assessed again, the effect had worn off.

ANESTH ANALG
2005;101:142732

After randomization to Group A or B, patients were

treated as follows: Group A received placebo capsules
identical to capsules of gabapentin, every 6 h, starting
the evening before surgery (18:00) and continued until
the eighth postoperative day. Twenty grams of placebo cream was applied in the wound area, starting
the day of surgery until the third postoperative day.
Intraoperatively, the brachial plexus in the axilla was
irrigated with 10 mL of normal saline and irrigation of
the third, fourth, and fifth intercostal spaces was performed with a total of 3 mL of the same solution.
Patients of Group B received 400 mg of gabapentin
with the same time points and of the same duration as
the control group. Also, 20 g of EMLA cream (2.5% of
lidocaine and 2.5% of prilocaine) was applied from the
day of surgery until the third postoperative day. As
previously described, 5 g of EMLA was applied over
the sternal area, 5 g close to the wound, and 10 g
around the incision in the axilla (4). Intraoperatively,
irrigation of the brachial plexus block in the axilla with
10 mL of 0.75% ropivacaine and irrigation of the third,
fourth, and fifth intercostal spaces with 3 mL of the
same solution was performed by the surgeon.
Premedication was omitted. Intraoperatively, noninvasive monitoring was used (electrocardiogram,
heart rate, arterial blood pressure, Spo2, and capnography). All patients received 10 mg of metoclopramide
and 0.25 mg of droperidol and were administered
oxygen for 3 4 min. Anesthesia was induced with
thiopental 2 mg/kg and propofol 2 mg/kg. Intubation
of the trachea was facilitated with rocuronium
0.6 mg/kg, and anesthesia was maintained with 2% of
sevoflurane end-tidal concentration and 70% nitrous
oxide in oxygen. At the end of surgery, neuromuscular block was antagonized with neostigmine 2.5 mg
and atropine 1.2 mg, the trachea was extubated, and
patients were transferred to the postanesthesia care
unit (PACU). The operation performed was either a
modified radical mastectomy or lumpectomy plus axillary lymph node dissection.
All patients remained in the PACU for at least 1 h.
Those who complained of pain in the PACU received
1.2 g of paracetamol IM. In the ward, Lonalgal tablets
(Boehringer Ingelheim, Italy) were given, with each
tablet containing 500 mg of paracetamol and 30 mg of
codeine. Acute postoperative pain at rest and after
movement was assessed using the VAS by an anesthesiologist blinded to group assignment. The scale consists of a 100-mm horizontal line with 0 representing
no pain and 100 represented the worst unbearable
pain. Patients were asked to draw a line vertical to the
horizontal line showing the intensity of their pain.
Movement consisted of the operated side arm abduction by 90 degrees. Pain was recorded at time 0 (arrival
to the PACU), 3, 6, and 9 h after surgery and each
morning from the first to the eighth postoperative day.

ANESTH ANALG
2005;101:142732

Before discharge from the hospital, all patients were

instructed to note the pain they might have at home
and analgesic needs. We defined chronic pain as the
presence of pain 3 mo after surgery, independent of its
intensity or analgesic requirements. Three and 6 mo
after surgery, patients were interviewed by a blinded
anesthesiologist as to whether they received postoperative chemotherapy or radiotherapy and if they experienced pain and or abnormal sensations in the
chest, axilla, or the arm of the operated side. The
presence of pain (yes versus no pain), the pain intensity (no pain 0, mild pain 1, moderate pain 2,
or severe pain 3), and the analgesic requirements at
home, if any, were recorded. The methodology regarding anesthetic technique, postoperative analgesia
protocol, and acute and chronic pain assessment were
similar to those in our previous studies (2 4).
Initial sample size estimation showed that approximately 24 patients should be included in each group to
ensure a power of 0.80 for detecting a 50% difference
in analgesic consumption (reduction of Lonalgal tablets use) between the two groups during the first 8
postoperative days and a reduction in the incidence of
chronic pain by 50%. The standard deviation of numbers of Lonalgal tablets consumed during the first 8
postoperative days, estimated from initial pilot observations, was approximately 3.5. The incidence of
chronic pain, estimated also from initial pilot observations and previous studies (2 4), was approximately
70% in the control group. The error was assumed to
be 0.05.
Levenes test was performed for equality of variances. Demographics, duration of surgery, and time
until the first analgesic requirement were compared
between the groups with two-tailed unpaired Students t-test. Paracetamol consumed in the PACU, the
number of Lonalgal tablets consumed by each group
of patients during the 8 days after surgery, and the
VAS scores at rest and after movement were compared between the groups using the Mann-Whitney
U-test. The number of patients who had modified
radical mastectomy, received chemotherapy, radiotherapy, analgesics in PACU, developed chronic pain
in the chest, axilla, or arm, absent or decreased sensation, and developed chronic pain overall and the intensity of chronic pain were compared between the
groups with Pearson 2 test with continuity correction.
The number of patients in each group who required
analgesics for chronic pain 3 and 6 mo after surgery
was compared with Fishers exact test. The intensity of
chronic pain between the two groups 3 and 6 mo after
surgery was compared with 2 test. The effect of the
type of surgery, i.e., modified radical mastectomy versus lumpectomy plus axillary dissection on chronic
pain development, was compared within each group
with Fishers exact test. The SPSS 11 software for

PAIN MEDICINE
FASSOULAKI ET AL.
ANALGESIA AND POSTMASTECTOMY PAIN

1429

Macintosh was used for statistical analysis (SPSS Base

10.0 for Macintosh; SPSS Inc, Chicago, IL).

Results
Figure 1 illustrates the flow diagram of the study.
Patient recruitment began March 15, 2001, and was
completed on January 20, 2004. Exit assessments
were to be completed by July 2004 when all patients
had a 6-mo follow-up. Two of the patients in the
control group developed local inflammation and
thrombosis in the axilla in the early postoperative
period and received nonsteroidal antiinflammatory
drugs (NSAIDS). A third patient in the control group
developed depression within the first month after surgery and was treated with tricyclic antidepressants.
These side effects are not related to the interventions
applied.
Demographics, duration of surgery, the number of
patients who had modified radical mastectomy versus
lumpectomy plus axillary dissection, and the number
of patients who received chemotherapy or radiotherapy did not differ between the two groups (Table 1).
The results for acute pain assessment are shown in
Table 1. The number and percent of patients in each
group who required analgesia in PACU, the time to
first analgesic requirement, paracetamol requirements
in the PACU, and Lonalgal tablet requirements for
the first 8 postoperative days in each group are shown
in Table 1 (Fig. 2). The VAS scores at rest and after
movement in each group are shown in Figures 3 and
4, respectively.
Chronic pain assessment 3 and 6 months after surgery in patients treated with analgesics versus the
controls are shown in Table 2. The incidence of total
chronic pain 3 mo after surgery was equally distributed between the controls who underwent lumpectomy (82%) or mastectomy (80%) and those in the
treatment group who underwent lumpectomy (46%)
or mastectomy (44%).

Discussion
Our results demonstrate that the analgesic drugs administered per protocol of the study significantly reduced the analgesic consumption after surgery and
the development of chronic pain three months after
breast surgery for cancer. The difference in chronic
pain between the groups became less evident six
months after surgery.
Previous studies assessing acute pain after breast
surgery for cancer have shown a beneficial effect of
regional block, of local application of local anesthetics
(3,4), and of drugs, such as mexiletine, stabilizing the
neural membrane (5). Brachial plexus block with local

1430

PAIN MEDICINE FASSOULAKI ET AL.

ANALGESIA AND POSTMASTECTOMY PAIN

ANESTH ANALG
2005;101:142732

Figure 1. The flow of patients studied. EMLA eutectic mixture of local anesthetics; NSAIDS nonsteroid antiinflammatory drugs; 3 M
3 mo.

anesthetic during surgery provides satisfactory analgesia, but arm movement is temporarily impaired,
which is undesirable, particularly for fast-track interventions (3). Local application of EMLA cream, a mixture of lidocaine and prilocaine, effectively reduced
postoperative analgesic requirements associated with
breast surgery for cancer and the incidence of longterm pain (4).
Most recently, gabapentin has been shown to reduce
analgesic requirements for acute postoperative pain
(6,7). Gabapentin was introduced as an antiepileptic
and is extensively used in the treatment of neuropathic pain (8 11). The drug at clinically relevant concentrations reduces the membrane voltage-gated calcium currents in dorsal root ganglia neurons of
neuropathic rats and, to a lesser degree, in nonneuropathic rats (12). It may produce analgesia by decreasing neurotransmitter release by sensory neurons, a
calcium-dependent process (12).

Side effects reported for gabapentin were somnolence, confusion, dizziness, and ataxia. However, the
drug was given for 8 weeks to doses titrated up to
3600 mg/d, unless severe adverse effects were developed (10). We did not observe intolerable side effects
for the daily dose and the duration of treatment, as
determined by the protocol of the study. Some sedation that our patients exhibited, particularly in the
beginning of treatment, is desirable before and immediately after surgery. This analgesic treatment was not
associated with adverse effects and may be appropriate not only for those patients who suffer, but also
prophylactically to all patients who are potential candidates for developing chronic pain.
The effective treatment of acute pain usually is not
associated with prevention of chronic pain (4 6). We
found that mexiletine 600 mg/d or gabapentin
1200 mg/d reduced postoperative analgesic requirements but did not significantly affect the development

ANESTH ANALG
2005;101:142732

PAIN MEDICINE
FASSOULAKI ET AL.
ANALGESIA AND POSTMASTECTOMY PAIN

1431

Table 1. Demographics and Characteristics of the 50 Patients Recruited for the Studya
Age (yr) (n 25)
Body weight (kg) (n 25)
Height (cm) (n 25)
Duration of surgery (min)
Patients who had MRM vs. lumpectomy (n 22)
Patients who had chemotherapy (n 22)
Patients who had radiotherapy (n 22)
Patients requiring analgesia in PACU
Time to first analgesia in PACU (min)
Paracetamol IM (mg) in PACU
Lonalgal tablets during the first 8 postoperative days

Control group

Treatment group

P-value

48 8.1
63 5.9
163 4.1
77 17.3*
5/22 (23%)
19/22 (86%)
8/22 (36%)
19/23 (83%)*
23 9
991 465**
4.43 4.95

49 8.4
64 8.0
164 5.9
87 13.9*
9/22 (41%)
18/22 (82%)
6/22 (27%)
9/23 (39%)*
28 18
469 599**
1.00 1.40

0.682
0.513
0.741
0.021
0.332
1.000
0.0747
0.007
0.303
0.003
0.016

a
Number and (%) of patients who required analgesia in postanesthesia care unit (PACU) time to first analgesia, IM paracetamol requirements (mg) in PACU,
and Lonalgal tablets consumed during the first 8 days after surgery.
Values are mean sd.
MRM modified radical mastectomy; IM intramuscular. Comparisons are between the treatment and control groups.
*, **, and all statistically significant difference.

Figure 2. The daily consumption of Lonalgal tablets in the control

and treatment group.

Figure 3. The visual analog scale (VAS) scores at rest in the treatment and in the control groups 0, 3, 6, and 9 h after surgery and
from the first to the eighth postoperative day.

of chronic pain (6). In this study, gabapentin 1600 mg/d

and local anesthetic to the surgical area, in the forms of
solution and cream, proved as effective as previous analgesic regimens to prevent acute postoperative pain but

Figure 4. The visual analog scale (VAS) scores after movement in

the treatment and in the control groups 0, 3, 6, and 9 h after surgery
and from the first to the eighth postoperative day.

superior in preventing chronic pain after breast surgery

for cancer. Unlike opioids, neither local anesthetics nor
gabapentin are associated with nausea and vomiting,
nor do they enhance bleeding, like NSAIDs. In addition,
local application of EMLA is simple, noninvasive, and
produces minimal side effects. Oral gabapentin and application of EMLA are preferable over parenteral analgesia and can be continued after hospital discharge. This
multimodal analgesic technique is considered particularly advantageous in the ambulatory setting.
Prevention of chronic pain development is difficult and
seems to require prolonged and probably multimodal
treatment. Also, although several studies assess acute postoperative pain and analgesic techniques and drugs, the
issue of chronic pain development has been largely ignored. Chronic pain may be resistant to opioids, and
when these are used, adverse outcome can be more
serious than in short-term use for acute pain (13).
Local anesthetics suppressing the afferent nociceptive traffic and inflammatory reaction may prevent the
development of chronic pain. Peripheral neural block,
by nerve sheath local anesthetic infusion, reduces the

1432

PAIN MEDICINE FASSOULAKI ET AL.

ANALGESIA AND POSTMASTECTOMY PAIN

ANESTH ANALG
2005;101:142732

Table 2. Patients with Chest, Axillary, Upper Arm, and Overall Chronic Pain, Absent or Decreased Sensation, and
Patients who Required Analgesics at Home 3 and 6 mo After Surgery
3 mo
No. of patients
Chest pain
Axilla pain
Arm pain
Chronic pain (total)
Absent or decreased sensation
No. of patients who needed analgesics

6 mo

Control
n (%)

Treatment
n (%)

P-value

Control
n (%)

Treatment
n (%)

P-value

7/22 (32)
10/22 (45)
13/22 (59)
18/22 (82)
17/22 (77)
5/22 (23)

7/22 (32)
3/22 (14)
5/22 (23)
10/22 (45)
16/22 (73)
0/22 (0)

1.00
0.045
0.038
0.028
1.00
0.048

5/21 (24)
6/21 (29)
7/21 (33)
12/21 (57)
17/21 (81)
4/21 (19)

3/20 (15)
3/20 (15)
3/20 (15)
6/20 (30)
13/20 (65)
0/20 (0)

0.697
0.454
0.277
0.151
0.424
0.107

incidence of chronic pain up to 12 months after lower

limb amputation (14). Intraoperative and postoperative epidural block with mepivacaine reduces longterm postthoracotomy pain (15). Mechanisms of prevention of chronic pain after injury by gabapentin are
not clear. With regard to neurogenic processes, in the
rat, gabapentin suppresses the ectopic discharge activity from injured peripheral nerves (16) and blocks the
injury-induced neuronal hyperactivity (17). Excitatory
transmission in the spinal cord was inhibited by gabapentin only in preparations from neuropathic animals
showing hyperalgesia and not from normal animals
(18). Local application of EMLA cream had a beneficial
effect on the incidence of chronic pain (4). However,
our results demonstrate that the multimodal analgesic
regimen of local anesthetics and gabapentin is the
most effective in preventing chronic pain development after breast surgery for cancer.
The incidence of chronic pain six months after surgery was 30% in the treatment group and 57% in the
control group. Although multimodal analgesia seems
to prevent chronic pain, even six months after surgery,
this difference was not statistically significant. Our
study was not powered for chronic pain assessment
six months after surgery, but a larger clinical trial may
show a significant difference in chronic pain between
treatment and control groups at this time point.
In conclusion, multimodal analgesia with local anesthetics and gabapentin reduced analgesic consumption
after breast surgery for cancer without side effects. The
incidence of chronic pain development, the intensity of
chronic pain, and the long-term analgesic requirements
associated with breast surgery were also less three
months after surgery. More studies are required to investigate chronic pain for longer periods of time because
for some women, chronic pain may persist in duration
and intensity beyond the six-month period.
We thank Dr. Anteia Paraskeva, who acted as the independent
anesthesiologist, helping with randomization and blinding of the
study.

References
1. Wallace MS, Wallace AM, Lee J, Dokke MK. Pain after breast
surgery: a survey of 282 women. Pain 1996;66:195205.
2. Smith WCS, Bourne D, Squair J, et al. A retrospective cohort
study of post mastectomy pain syndrome. Pain 1999;83:915.
3. Fassoulaki A. Brachial plexus block for pain relief after modified
radical mastectomy. Anesth Analg 1982;61:986 7.
4. Fassoulaki A, Sarantopoulos C, Melemeni A, Hogan Q. EMLA
reduces acute and chronic pain after surgery for breast cancer.
Reg Anesth Pain Med 2000;25:350 5.
5. Fassoulaki A, Sarantopoulos C, Melemeni A, Hogan Q. Regional
block and mexiletine: the effect on pain after cancer breast
surgery. Reg Anesth Pain Med 2001;26:223 8.
6. Fassoulaki A, Patris K, Sarantopoulos C, Hogan Q. The analgesic effect of gabapentin and mexiletin after breast surgery for
cancer. Anesth Analg 2002;95:98591.
7. Dirks J, Fredensborg BB, Christensen D, et al. A randomized
study of the effects of single-dose gabapentin versus placebo on
postoperative pain and morphine consumption after mastectomy. Anesthesiology 2000;97:560 4.
8. Mao J, Chen LL. Gabapentin in pain management. Anesth
Analg 2000;91:680 7.
9. Cheng J, Pan H, Eisenach JC. Antiallodynic effect of intrathecal
gabapentin and its interaction with clonidine in a rat model of
postoperative pain. Anesthesiology 2000;92:1126 31.
10. Rowbotham M, Harden N, Stacey B, et al. Gabapentin for the
treatment of postherpetic neuralgia: a randomized controlled
trial. JAMA 1998;280:1837 42.
11. Serpell MG, Neuropathic Pain Study Group. Gabapentin in
neuropathic pain syndromes: a randomised, double-blind,
placebo-controlled trial. Pain 2000;99:557 66.
12. Sarantopoulos C, McCallum B, Kwok WM, Hogan Q. Gabapentin decreases membrane calcium currents in injured as well as in
control mammalian primary afferent neurons. Reg Anesth Pain
Med 2002;27:4757.
13. Ballantyne JC, Mao J. Opioid therapy for chronic pain. N Engl
J Med 2003;349:194353.
14. Fisher A, Meller Y. Continuous postoperative regional analgesia
by nerve sheath block for amputation surgery: a pilot study.
Anesth Analg 1991;72:300 3.
15. Obata H, Saito S, Fujita N, et al. Epidural block with mepivacaine before surgery reduces long-term post-thoracotomy pain.
Can J Anaesth 1999;46:112732.
16. Pan HL, Eisenach JC, Chen SR. Gabapentin suppresses ectopic
nerve discharges and reverses allodynia in neuropathic rats.
J Pharmacol Exp Ther 1999;288:1026 30.
17. Kanai A, Sarantopoulos C, McCallum JB, Hogan Q. Painful
neuropathy alters the effect of gabapentin on sensory neuron
excitability in rats. Acta Anaesthesiol Scand 2004;48:50712.
18. Patel MK, Gonzalez M, Bramwell S, et al. Gabapentin inhibits
excitatory synaptic transmission in the hyperalgesic spinal cord.
Br J Pharmacol 2000;130:1731 4.

The Effects of Intraarticular Resiniferatoxin in Experimental

Knee-Joint Arthritis
Eugene Y. Kissin,

MD*,

Cristina F. Freitas,

BA,

and Igor Kissin,

MD, PhD

*Arthritis Center, Boston University School of Medicine, Boston, Massachusetts and Department of Anesthesiology,
Perioperative and Pain Medicine, Brigham and Womens Hospital, Harvard Medical School, Boston, Massachusetts

In this study we sought to determine whether an intraarticular administration of a vanilloid agonist resiniferatoxin (RTX) produces an analgesic effect in experimental arthritis. Knee joint inflammation was induced
in rats by intraarticular carrageenan (2%, 30 L). Pain
score and left/right hind leg weight distribution ratio
were used to assess pain behavior. Changes in knee dimensions were evaluated by measuring external circumference and intraarticular area (ultrasound scanning). The intraarticular administration of RTX
(0.0003% or 0.003%, 30 L) provided a significant analgesic effect. Twenty-four hours after RTX administration, the pain score was reduced from 15.1 4.7 to 6.9

anilloids bind to the transient receptor potential

type V1 (TRPV1) channels, nonselective cation
ionophores that play an important role in integration of afferent noxious signals generated by
inflammatory mediators (1). Vanilloid agonists (capsaicin and its analogs) can produce selective and
long-lasting blockade of afferent C-fibers and some
A-fibers (2). The cloning of vanilloid receptors (3,4)
heralded the rapid advances in vanilloid pharmacology and has begun to play an important role in the
development of new compounds with analgesic properties. This new development put a new emphasis on
the results of previous studies with capsaicin. Resiniferatoxin (RTX) is an ultrapotent capsaicin analog with
a unique spectrum of activities (2). Its initial excitatory
effect relative to its inactivating effect is far less pronounced compared with capsaicin (2). Several studies
(57) convincingly demonstrated that capsaicin selectively suppresses the conduction of impulses in the
C-fibers and some A-fibers.

Supported by National Institutes of Health grant GM065834.

Accepted for publication May 25, 2005.
Address correspondence and reprint requests to Eugene Kissin,
MD, Boston University School of Medicine, 715 Albany Street, E-5,
Boston, MA 02118 2526. Address e-mail to ekissin@arthritis.bu.edu.
DOI: 10.1213/01.ANE.0000180998.29890.B0
2005 by the International Anesthesia Research Society
0003-2999/05

4.4 (P 0.01) with 0.0003% and was abolished (P

0.0001) with 0.003%. The improvement in weight distribution ratio lasted for several days after the RTX administration. Reduction in knee circumference demonstrated that intraarticular RTX suppressed the
carrageenan-induced edema by at least one third. Ultrasound scanning revealed no RTX-induced decrease of
the intraarticular area. The experiments demonstrated
that intraarticular RTX inhibits pain behavior in kneejoint arthritis and that this effect is dose-dependent.
These results suggest a new direction for peripheral
analgesia.
(Anesth Analg 2005;101:14339)

Using the ability of vanilloid agonists to selectively

suppress the conduction of noxious impulses, we
demonstrated that RTX-induced blockade of the sciatic and saphenous nerves completely prevents the
development of hyperalgesia caused by intraplantar
injection of carrageenan (8). Otsuki et al. (9), using the
less potent capsaicin, showed that its application to
the sciatic and femoral nerves reduces pain behavior
induced by the injection of monosodium urate crystals
into a knee joint.
The main aim of this study was to determine
whether localized intraarticular administration of RTX
produces an analgesic effect in experimental arthritis.
Changes in pain behavior were assessed with indices
based on two basic signs in acute knee inflammation:
limp (10,9) and decrease in weight load on the affected
leg (9,11).

Methods
Male Sprague-Dawley rats weighing 225275 g were
used for the experiments. The rats were housed with a
12-h light/dark cycle and provided with food and
water ad libitum. The protocol for this study was approved by the Institutional Panel on Laboratory Animal Care.
Anesth Analg 2005;101:14339

1433

1434

PAIN MEDICINE KISSIN ET AL.

INTRAARTICULAR VANILLOID IN ARTHRITIS

Joint inflammation was induced by carrageenan

(Sigma Chemical, St. Louis, MO) injected percutaneously using a 30-gauge needle through the infrapatellar ligament into the left knee joint cavity (2%, 30 L)
with the animal under brief halothane (2%) anesthesia.
The changes in pain behavior were determined during
a period of approximately 10 days after the carrageenan injection. Pain score and weight-bearing index
were used to assess pain behavior. Pain score determination was based on walking pattern. Each rat was
placed on an open bench. That enabled the rat to walk
freely. The scoring was similar to that used by Otsuki
et al. (9). A score of 0 was assigned if the rat walked
without a limp; a score of 1 if the rat used the injected
paw with a limp; a score of 2 if the rat walked on three
legs. If an animal walked both with a limp and on 3
legs the highest score was assigned. To initiate ambulation, gentle backward tugging on the tail was provided. The walking patterns were observed every
5 min for 1 h. The sum of the 12 scores (0 24) obtained
during the 1-h session was recorded as a cumulative
pain score. Weight-bearing index was determined by
measuring weight distribution between left (inflamed)
and right (contralateral control) hindlegs. An Incapacitance Analgesia Meter (Stoelting CO, Wood Dale, IL)
was used for this aim (11). The rat was placed into the
plastic box of the device where it was kept sitting
during a 2-min period. The integrated paw pressure
during this period was displayed separately for the
right and left leg. The ratio between the pressures of
the right and left leg was calculated as left/right hind
leg weight distribution ratio. The measurement was
repeated with a 1-h interval in between. The average
value of these two measurements was used as an
index of joint discomfort.
In addition to the behavioral indices, two measurements reflecting changes of the knee dimensions
were done. With the animal under light halothane
anesthesia, the circumference of the knee was determined with a flexible tape measure (12). Ultrasound
(US) scanning was used to assess intraarticular
changes of the knee. Real-time B-mode US was performed using a Titan (SonoSite Inc., Bothel, WA) US
system with a 10-MHz probe (axial resolution of
0.76 mm at depth of 30 mm). US examination was
performed under halothane (2%) anesthesia, with the
knee shaved and flexed to 90. The area between the
patella and tibia was scanned in a midline longitudinal axis. A hypoechoic or anechoic area encompassed
by the patellar ligament, the femur, and the tibia was
measured by fitting elliptical calipers between these
structures. A representative image is shown in Figure
1. The average of three measurements of the left (inflamed) and right (contralateral control) knee was
used for the assessment of intraarticular changes.

ANESTH ANALG
2005;101:14339

Figure 1. The longitudinal view of the left knee after the intraarticular injection of carrageenan. A sketch (upper part) and representative B-mode image of the knee joint. F femur; T tibia; P
patella; L patellar ligament; ellipse the measured area. In the
sonogram, the knee is flexed at 90.

RTX (Sigma Chemical) 30 L was injected into the

knee joint cavity percutaneously through the intrapatellar ligament with the animal under brief halothane
(2%) anesthesia. Two concentrations of RTX were
used, 0.0003% (0.09 g) and 0.003% (0.9 g), on the
basis of our previous experiments (8). Before an experiment, RTX (initially dissolved in dimethyl sulfoxide to a concentration of 1 g/L and stored at 80C
under nitrogen) was diluted in 0.9% saline with 0.3%
Tween 80 (to avoid precipitation). RTX was injected
into the knee joint with India ink to confirm the intraarticular site of the injection. Each animal was dissected at the end of the experiment to confirm the
localization of the injection. To prevent the initial excitatory effect of RTX, intraarticular bupivacaine
(0.5%, 10 L) was administered (providing local anesthetic effect for approximately 1 h) 10 min before RTX.
After bupivacaine-RTX administration, the animals
exhibited no behavioral signs associated with pain
(flinching or licking of the paw).
Five groups of rats (n 8 per group) were used. In
two groups that received carrageenan-bupivacaineRTX (CBR), RTX (0.0003% or 0.003%) was used against
the background of the carrageenan-induced joint inflammation. After baseline measurements of the variables, carrageenan was administered in these two
groups and variables were measured 3 h and 23 h
later. Then bupivacaine was injected followed by RTX
and the variables were measured 3 and 24 h and 2, 3,

ANESTH ANALG
2005;101:14339

6, and 8 days after the RTX injection. Three other

groups of animals were used as controls. The
carrageenan-bupivacaine-vehicle (CBV) group received the injection of carrageenan and the injection of
bupivacaine 24 h later followed by an injection of the
vehicle for RTX (dimethyl sulfoxide in saline with
Tween 80). The measurements of variables were the
same as in the CBR groups. The animals in the
bupivacaine-RTX (BR) group did not receive an injection of carrageenan before both of the agents were
administered. In the SV group, saline was used instead of bupivacaine and vehicle was used instead of
RTX. Before each of the measurements each rat was
placed on a bench and carefully observed for a minute
or two. Notes were made regarding the animals gait,
the posture of the affected hindpaw, and the condition
of the knee.
The rats were assigned to the groups randomly
(blocked randomization) and those responsible for the
measurements of the effects were blinded as to the
type of treatment. Each animal was housed individually and identified only with a sequential number
assigned before the beginning of the experiment. Because the control solution and the RTX solution were
contained in identical vials and the solutions were
both clear, they were indistinguishable to the
examiner.
Data were analyzed with a two-way (group and
time) analysis of variance, with time treated as a
repeated-measures factor. Comparisons among
groups at each time were performed with one-way
analysis of variance. Multiple comparisons among
means were made with Fishers protected leastsquares difference test. The results were declared significant if the P value was 0.05. Data are reported as
the mean sd.

Results
The comparison of two groups of animals (CBV and
CBR) illustrating the magnitude and time course of
changes in pain score and weight distribution ratio
after the injections of carrageenan and RTX is presented in Figure 2. In the CBV group, 3 h after the
vehicle injection the pain score was 20.5 2.6 (P
0.0001). It lasted for several days, and on day 2 (3 days
after carrageenan injection), the pain score was 3.8
5.4 (P 0.05). Changes in the weight distribution ratio
lasted much longer, approximately 1 wk.
In the CBR group (reflecting the effect of 0.0003%
RTX), the effects of RTX on both pain score and weight
distribution ratio continued for a long period after its
intraarticular injection. The difference between the
CBV and CBR groups with pain score was significant
at 3 h (decrease by 68%, P 0.0001), and 24 h (by 54%,
P 0.01). Relative to the maximum pain score (seen at

PAIN MEDICINE
KISSIN ET AL.
INTRAARTICULAR VANILLOID IN ARTHRITIS

1435

Figure 2. The magnitude and time course of changes in pain score

and weight distribution ratio after injections of carrageenan and
resiniferatoxin. Pain score a cumulative score based on walking
patterns observed during an 1-h period. Weight distribution ratio
a ratio of weight distribution between left (inflamed) and right
(contralateral control) hindlegs of the sitting rat. Each column represents the mean sd for the carrageenan-bupivacaineresiniferatoxin (CBR) group and the carrageenan-bupivacainevehicle (CBV) group (n 8 per group). Carrageenan (C: 2%, 30
L), bupivacaine (B: 0.5%, 10 L), resiniferatoxin (R: 0.0003%, 30
L), or vehicle (V: 30 L) were injected into the left knee joint
cavity. Time h (hours) or d (days) before () or after the injection
of resiniferatoxin. Statistical significance: aP 0.0001 versus CBV at
3 h, bP 0.01 versus CBV at 24 h, cP 0.02 versus CBV at 3 h,
d
P 0.02 versus CBV at 24 h, eP 0.05 versus CBV at 3 d.

21h; Fig. 2) in the same group (CBR) the decrease in

pain score was also significant at 3 and 24 h (P
0.0001 for both). For weight distribution ratio, the
difference between the CBV and CBR was statistically
significant at 3 h (P 0.002), 24 h (P 0.02), and
3 days (P 0.05).
The RTX-induced changes in pain behavior were
dose-dependent (Fig. 3). Three hours after the RTX
injection, the pain score was reduced from 20.5 2.6
to 6.5 4.3 (P 0.0001) with a concentration of
0.0003% and was abolished (P 0.0001) with a concentration of 0.003%. Twenty-four hours after RTX, the
pain score was reduced from 15.1 4.7 to 6.9 4.4
(P 0.01) with 0.0003% and was abolished (P
0.0001) with 0.003%. The difference between pain
scores in the 2 RTX groups was statistically significant
at 3 h (P 0.005) and 24 h (P 0.01). In contrast to the
pain score, the weight distribution ratio at 24 h was
not reversed by 0.003% RTX. Although the reversal at
24 h had a tendency to be greater with 0.003% (0.72
0.19) than with 0.0003% (0.56 0.23), the difference
was not statistically significant.
RTX clearly decreased the knee swelling induced by
carrageenan (Fig. 4). At 24 h the increases in the circumference from baseline was 8.2 1.9 mm in the
CBV group and 5.0 2.3 mm in the CBR group (P
0.02 for the difference between groups). At 3 d the
difference between the groups continued to be statistically significant (P 0.01).

1436

PAIN MEDICINE KISSIN ET AL.

INTRAARTICULAR VANILLOID IN ARTHRITIS

Figure 3. The dose-dependence of the effects of resiniferatoxin

(RTX) on pain score and weight distribution ratio 3 h (left side) and
24 h (right side) after its injection into the knee joint cavity. Each
point represents the mean sd for 3 groups (n 8 per group) with
0% (vehicle only), 0.0003%, and 0.003% of RTX injected into the left
knee joint cavity in a volume of 30 L 24h after the induction of
carrageenan-induced inflammation. Statistical significance: aP
0.0001 versus 0% RTX, bP 0.005 versus 0.0003% RTX, cP 0.002
versus 0% RTX, dP 0.001 versus 0% RTX, eP 0.0005 versus 0%
RTX, fP 0.0001 versus 0% RTX, gP 0.01 versus 0.0003% RTX,
h
P 0.02 versus 0% RTX, iP 0.0005 versus 0% RTX.

Figure 4. The effect of resiniferatoxin on the carrageenan-induced

increase in the knee circumference. CBR carrageenan-bupivacaine-resiniferatoxin group; CBV carrageenan-bupivacainevehicle group (n 8 per group). Carrageenan (C: 2%, 30 L),
bupivacaine (B: 0.5%, 10 L), resiniferatoxin (R: 0.0003%, 30 L)
or vehicle (V: 30 L) was injected into the knee-joint cavity. Boxes
display differences in the increases () of the knee circumference
(mean sd) from baseline level in the CBR group (upper line)
versus the CBV group (lower line) at 24 h (A) and 3 d (B).

The US comparison of the left (inflamed) and right

(contralateral control) knee demonstrated a profound
effect of carrageenan. Three hours after carrageenan
injection the area of the knee joint chosen for the
comparison was 6.0 1.1 mm2 on the inflamed side
and 3.7 0.5 mm2 on the contralateral side (P
0.001). Three days later (CBV group), the difference

ANESTH ANALG
2005;101:14339

was almost the same: 6.9 2.7 mm2 versus 3.3

0.7 mm2 (P 0.001). In contrast to the knee circumference (Fig. 4), there was no RTX-induced decrease of
the hypoechoic/anechoic area on day 3, when limping
had completely disappeared; the values in the CBV
and CBR groups were identical.
In the BR and SV groups (control), no pain behavior
was evident. Rats of both groups walked without a
limp; therefore a score of 0 was recorded at all time
points after the intraarticular injection. Injection of
bupivacaine-RTX without the background of
carrageenan-induced inflammation (BR group) did
not cause any significant changes in the weight distribution ratio. The index values were 1.09 0.28 (baseline), 1.15 0.25 (3 h after the injection), 1.01 0.21
(24 h), 1.02 0.09 (2 d), and 1.10 0.11 (3 d).
Rat behavior did not indicate any sedative effect of
RTX both at 3 and 24 h after the agent injection.

Discussion
Our results demonstrated that a single intraarticular
injection of RTX produces an analgesic effect in
carrageenan-induced arthritis. With the limping index
the effect lasts more than 24 hours; as long as this
model provided the possibility to observe the effect.
This effect is dose-dependent and can be induced by
relatively small doses of RTX, starting from 0.09 g
(0.0003%, 30 L). This dose is approximately 1000
times smaller than the dose producing analgesia after
systemic administration of RTX in rats (13). It is of
interest that when intraarticular and systemic equianalgesic doses of morphine were compared (using an
inflamed knee model in rats) (14) the intraarticular
dose was 10 times smaller than the systemic dose. This
may indicate that, compared with morphine, RTX has
a much wider margin of safety regarding acute systemic toxicity.
The experiments demonstrate a difference in the
time course of RTX effects between pain score and
weight distribution ratio. The changes induced by carrageenan injections disappeared by the third day
(from 24 hours to 2 days, Fig. 2) with pain score; at
the same time, they lasted for 8 9 days with weight
distribution ratio. As a result, the RTX effect could be
observed for a longer period of time when the weight
distribution ratio was used. This difference is difficult
to explain. One of the possible explanations of the
discrepancy between the tests is that limping score
selectively reflects intensive pain on movement,
whereas weight distribution ratio (with the rat in the
box of the device) may be able to reveal even mild
pain in a sitting position. Another explanation is that
the weight distribution test may reflect not only pain
but also a nonpainful joint discomfort that could last
much longer than pain.

ANESTH ANALG
2005;101:14339

It is of interest to compare the effects of RTX in

arthritis and those in a rat model of postoperative pain
(15). Equal concentrations of RTX (0.0003%) used in
both models produce analgesia of similar duration.
The incisional pain study also indicated that RTX does
not suppress motor activity.
Localized analgesic effects of RTX have previously
been demonstrated in carrageenan-induced foot models of hyperalgesia with percutaneous injection of the
agent to the sciatic and saphenous nerves (8) or its
direct administration into the plantar tissue (16). Neubert et al. (16) reported that a peripheral injection of
RTX produced the local analgesic effect by reversible
inactivation of the nerve terminal endings. Otsuki et
al. (9), using a monosodium urate model of knee joint
arthritis, demonstrated the analgesic effect of capsaicin administered systemically to the neonatal rat or
locally to the sciatic and femoral nerves. Although the
authors did not find complete analgesia with capsaicin, they did observe a significant effect. Topical
capsaicin (cream) has been used to relieve pain of
osteoarthritis, and several reviews confirmed its effectiveness using meta-analysis (17,18). However, the relatively weak effectiveness of capsaicin cream, the need
for multiple applications, and skin irritation limits its
use.
Changes in the knee circumference in our experiments clearly indicated that intraarticular RTX suppressed the carrageenan-induced edema by at least
one third (Fig. 4). In a previous study (8), pretreatment
with RTX injected to the sciatic and saphenous nerves
reduced carrageenan-induced edema of the foot to a
similar degree. Lam and Ferrell (19), using the Evans
blue content method, demonstrated that knee inflammation induced by intraarticular carrageenan was reduced by 44% in the knees of rats that had previously
been injected with capsaicin. They also observed that
chronic joint denervation produced a 37% reduction in
the inflammation. These observations indicate that at
least one component of the antiinflammatory effect of
vanilloid agonists is to counteract neurogenic inflammation. This confirms an earlier observation that local
application of capsaicin to the sciatic nerve prevents
neurogenic inflammation in the lateral part of the
dorsal skin of the rats paw (20).
Our US measurements of the knee-joint cavity demonstrated that US makes it possible to measure very
small areas (3 6 mm2) in the knee cavity. This is
comparable with the reported ability of US to measure
a knee bursa with a short axis of 2 mm (21). We
found that carrageenan injection significantly increases the hypoechoic/anechoic area representing
the joint cavity, probably reflecting changes from joint
effusion. This increase did not regress in either the
control (CBV) or the RTX-treated (CBR) group even
4 days after carrageenan administration, when the
knee circumference had almost completely recovered.

PAIN MEDICINE
KISSIN ET AL.
INTRAARTICULAR VANILLOID IN ARTHRITIS

1437

This may be explained by the slow process of knee

exudate resorption as compared with the recovery of
knee inflammation. The absence of any difference between the CBR and CBV groups indicates that RTX
does not have any significant effect on the changes
measured by US, which probably results from the
absence of the effect on exudate resorption. The knee
exudate may be a factor responsible for the discrepancy between the difference in the duration of the
changes in pain score (shorter) and weight distribution ratio (longer) after the injection of carrageenan
(Fig. 2). If this discrepancy indicates that the weightbearing test reflects not only pain but also a nonpainful joint discomfort, the joint effusion might be responsible for this discomfort.
TRVP1 is activated not only by vanilloid ligands but
also by noxious heat and low pH and thus can be
viewed as a molecular integrator of noxious stimuli in
peripheral terminals of primary sensory neurons (22).
An increase in TRPV1 expression in peripheral nociceptors is critical for the maintenance of inflammatory
hyperalgesia (23). TRPV1 exhibits a time- and Ca2dependent desensitization, a long-lasting refractory
state during which the receptor does not respond to
vanilloids or other stimuli (24). Several studies have
demonstrated that perineural administration of capsaicin can selectively block the conduction of impulses
in the C-fibers and A-fibers at the site of capsaicin
application (57). For example, Chung et al. (6) found
that capsaicin applied to the sciatic nerve completely
eliminated responses to noxious heat stimuli and decreased responses to noxious mechanical stimuli. At
the same time, they observed that responses to innocuous mechanical stimuli were increased. Xin et al. (25)
reported that capsaicin markedly attenuated voltagegated Na currents in dorsal root ganglion (DRG)
cells. They suggested that this effect may account for
the transient blockade of the nerve conduction seen
with capsaicin and may explain its analgesic
properties.
The molecular mechanism of TRPV1 receptor desensitization has remained elusive. In the last few
years, involvement of calmodulin (CaM) and CaM
kinase II (binding of Ca2/CaM complex to the
TRPV1 and phosphorylation by CaM kinase II, respectively) in the phenomenon of TRPV1 receptor regulation has been shown (26,27). These results indicate
that the influx of Ca2 through TRPV1 may feed back
on the channels, inhibiting their gating. This type of
feedback inhibition could play a role in the desensitization of the TRVP1 receptor produced by capsaicin
(27). However, the molecular mechanism of TRPV1
receptor desensitization and the mechanism of transient blockade of nerve conduction seen with administration of vanilloid agonists might be quite different.
It is well established that vanilloid agonists can kill
adult sensory neurons in culture and that this effect is

1438

PAIN MEDICINE KISSIN ET AL.

INTRAARTICULAR VANILLOID IN ARTHRITIS

most likely mediated by calcium influx (28). There is

evidence that the increase in intracellular calcium produced by vanilloids can result in levels of calcium in
sensory neurons in vivo that are high enough to cause
irreversible neuronal damage. A significant loss of
DRG neurons induced by systemic capsaicin has been
reported (29); however, the dose of the agent was
extremely large: 100 mg/kg subcutaneously. Systemic
RTX administered in a single dose of 150 g/kg to 500
g/kg caused a loss of C-fibers sensory functions that
lasted more than 4 weeks (30). The long duration of
this effect was the basis for the suggestion that the
local treatment of peripheral nerves with vanilloid
agonists results in a permanent impairment of the
function of vanilloid-sensitive afferent nerve fibers.
Whether it is a result of the degeneration of the fibers
or a long-lasting but reversible loss of their functions
is a matter of controversy.
Initial electron microscopic studies did not find any
axonal degeneration with perineural (31) or topical
(32) application of capsaicin even at large (up to 1.5%)
concentrations. In more recent studies of the neurotoxic effects of capsaicin applied to the skin (33,34),
administered perineurally (35), or into the urinary
bladder (36), the authors visualized nerve fibers by
immunohistochemical methods based on the determination of substance P (SP), calcitonin gene-related
peptide (CGRP), and/or the pan-neuronal marker
protein gene product (PGP 9.5). These studies demonstrated a profound loss of nerve-fiber staining in the
epidermis (or urinary bladder wall), leading some of
these authors to conclude that capsaicin induced axonal degeneration. However, none of these studies
used electron microscopy to confirm this conclusion.
When Avelino and Cruz (37) re-examined the problem
of capsaicin-induced axonal degeneration in the rat
bladder with the parallel use of immunohistochemical
and electron microscopy methods, they found that
both capsaicin and RTX caused desensitization and a
profound reduction in SP and CGRP immunoreactive
fibers without causing significant nerve changes demonstrated by electron microscopy. The authors concluded that vanilloids applied intravesically (urinary
bladder) in full desensitizing concentrations exert
their effect on the bladder (including depletion of SP
and CGRP) lasting for 8 to 12 weeks without producing nerve-fiber degeneration. They suggested that the
PGP immunoreactivity that was used in several capsaicin studies as an index of axon destruction could be
lost because vanilloids cause an axonal transport
blockade that slows the arrival of PGP to peripheral
axons, a process that occurs without nerve degeneration.
The controversy regarding nerve fiber degeneration
versus long-lasting loss of function without degeneration is possibly relevant to the large concentrations of
vanilloid agonists, which are far in excess of the concentrations that produce analgesia lasting only several

ANESTH ANALG
2005;101:14339

days. Even if RTX is used at very large concentrations

that cause selective degeneration of nerve endings
(C-fibers and A-fibers) of a joint, nerve function will
be able to eventually recover because of the process of
fiber regeneration. This process should be faster and
more complete after intraarticular RTX administration
than after RTX application to the sciatic nerve because
of the involvement of relatively short nerve terminals
with intraarticular RTX. Karai et al. (38) suggested the
nociceptive neuronal or nerve terminal deletion as an
effective and broadly applicable strategy for pain
management. Such an approach is an acceptable alternative to a reversible interruption of the nerve fiber
functions without their destruction that is achievable
at smaller concentrations of vanilloid agonists.
In conclusion, our experiments demonstrated that intraarticular RTX inhibits pain behavior in carrageenaninduced knee-joint arthritis and that this effect is
dose-dependent.
The authors thank Edwin L. Bradley, Jr., Professor of Biostatistics at
University of Alabama at Birmingham, for statistical analysis.

References
1. Caterina MJ, Julius D. The vanilloid receptor: a molecular gateway to the pain pathology. Annu Rev Neurosci 2001;24:487517.
2. Szallasi A, Blumberg PM. Vanilloid (capsaicin) receptors and
mechanisms. Pharmacol Rev 1999;51:159 212.
3. Caterina MJ, Schumacher MA, Tominaga M, et al. The capsaicin
receptor: a heat-activated ion channel on the pain pathway.
Nature 1997;389:816 24.
4. Hayes P, Meadows HJ, Gunthrope MJ, et al. Cloning and functional expression of a human orthologue of rat vanilloid
receptor-1. Pain 2000;88:20515.
5. Petsche U, Fleischer E, Lembeck F, Handwerker HO. The effect
of capsaicin application to a peripheral nerve on impulse conduction in functionally identified afferent nerve fibres. Brain
Res 1983;265:233 40.
6. Chung JM, Lee KH, Hori Y, Willis WD. Effects of capsaicin
applied to a peripheral nerve on the responses of primate spinothalamic tract cells. Brain Res 1985;329:2738.
7. Pini A, Lynn B. C-fibre function during the 6 weeks following
brief application of capsaicin to a cutaneous nerve in the rat. Eur
J Neurosci 1990;3:274 84.
8. Kissin I, Bright CA, Bradley EL. Selective and long-lasting neural blockade with resiniferatoxin prevents inflammatory pain
and hypersensitivity. Anesth Analg 2002;94:1253 8.
9. Otsuki T, Nakahama H, Niizuma H, Suzuki J. Evaluation of the
analgesic effects of capsaicin using a new rat model for tonic
pain. Brain Res 1986;365:235 40.
10. Okuda K, Nakahama H, Miyakawa H, Shima K. Arthritis induced in cat by sodium urate: a possible animal model for tonic
pain. Pain 1984;18:28797.
11. Bove SE, Calcaterra SL, Brooker RM, et al. Weight bearing as a
measure of disease progression and efficacy of antiinflammatory compounds in a model of monosodium
iodoacetate-induced osteoarthritis. Osteoarth Cart 2003;11:
82130.
12. Yu CY, Koo ST, Kim CH, et al. Two variables that can be used
as pain indices in experimental animal models of arthritis. Neurosci Meth 2002;115:10713.

ANESTH ANALG
2005;101:14339

13. Xu XJ, Farkas-Szallasi T, Lundberg JM, et al. Effects of the

capsaicin analogue resiniferatoxin on spinal nociceptive mechanisms in the rat: behavioral, electrophysiological, and in situ
hybridization studies. Brain Res 1997;752:52 60.
14. Nagasaka H, Awad H, Yaksh TL. Peripheral and spinal actions
of opioids in the blockade of the autonomic response evoked by
compression of the inflamed knee joint. Anesthesiology 1996;85:
808 16.
15. Kissin I, Davison N, Bradley EL Jr. Perineural resiniferatoxin
prevents hyperalgesia in a rat model of postoperative pain.
Anesth Analg 2005;100:774 80.
16. Neubert JK, Karai L, Jun JH, et al. Peripherally induced resiniferatoxin analgesia. Pain 2003;104:219 28.
17. Zhang WY, Li Wan Po A. The effectiveness of topically applied
capsaicin: a meta-analysis. Eur Clin Pharmacol 1994;46:51722.
18. McQuay HJ, Moore RA. An evidence based resource for pain
relief. Oxford: Oxford University Press, 1998.
19. Lam FY, Ferrell WR. Inhibition of carrageenan induced inflammation in the rat knee joint by substance P antagonist. Ann
Rheum Dis 1989;48:928 32.
20. Jancso G, Kiraly E, Jancso-Gabor A. Direct evidence for an
axonal site of action of capsaicin. Naunyn-Schiedeberg Arch
Pharmacol 1980;313:91 4.
21. Kane D, Balint PV, Sturrock RD. Ultrasonography is superior to
clinical examination in the detection and localization of knee joint
effusion in rheumatoid arthritis. J Rheumatol 2003;30:966 71.
22. DiMarzo V, Blumberg PM, Szallasi A. Endovanilloid signaling
in pain. Curr Opin Neurobiol 2002;12:3729.
23. Ji RR, Samad TA, Jin SX, et al. p38-MAPK activation by NGF in
primary sensory neurons after inflammation increases TRPV1
levels and maintains heat hyperalgesia. Neuron 2002;36:57 68.
24. Julius D, Basbaum AI. Molecular mechanisms of nociception.
Nature 2001;413:20310.
25. Xin S, Wachtel RE, Gebhart GF. Capsaicin sensitivity and
voltage-gated sodium currents in colon sensory neurons from
rat dorsal root ganglia. Am J Physiol 1999;277;G1180 8.
26. Rosenbaum T, Gordon-Shaag A, Munari M, Gordon SE. Ca2/
calmodulin modulates TRPV1 activation by capsaicin. J Gen
Physiol 2004;123:53 62.
27. Jung J, Shin J, Lee SY, et al. Phosphorylation of vanilloid receptor 1 by Ca2/calmodulin-dependent kinase II regulates its
vanilloid binding. J Biol Chem 2004;279:7048 54.

PAIN MEDICINE
KISSIN ET AL.
INTRAARTICULAR VANILLOID IN ARTHRITIS

1439

28. Chard PS, Bleakman D, Saviage JR, Miller RJ. Capsaicininduced neurotoxicity in cultured dorsal root ganglion neurons:
Involvement of calcium-activated proteases. Neuroscience 1995;
65:1099 108.
29. Jancso G, Kiraly E, Joo F, et al. Selective degeneration by capsaicin of a subpopulation of primary sensory neurons in the
adult rat. Neurosci Lett 1985;59:209 14.
30. Szallasi A, Blumberg PM. Vanilloid receptor loss in rat sensory
neurons associated with long term desensitization to resiniferatoxin. Neurosci Lett 1992;136:51 4.
31. Ainsworth A, Hall P, Wall PD, et al. Effects of capsaicin applied
locally to adult peripheral nerve. II Anatomy and enzyme and
peptide chemistry of peripheral nerve and spinal cord. Pain
1981;11:379 88.
32. Szolcsanyi J, Jancso-Gabor A, Joo F, et al. Functional and fine
structural characteristics of the sensory neuron blocking effect
of capsaicin. Naunyn-Schiedeberg Arch Pharmacol 1975;287:
157 69.
33. Reilly DM, Ferdinando D, Johnston C, et al. The epidermal
nerve fibre network: characterization of nerve fibers in human
skin by confocal microscopy and assessment of racial variations.
Br J Derm 1997;137:16370.
34. Simone DA, Nolano M, Johnson T, et al. Ontradermal injection
of capsaicin in humans produces degeneration and subsequent
reinnervation of epidermal nerve fibers: correlation with sensory function. J Neurosci 1998;18:894759.
35. Dux M, Sann H, Schemann M, Jancso G. Changes in fibre
populations of the rat hairy skin following selective chemodenervation by capsaicin. Cell Tissue Res 1999;296:4717.
36. Dasgupta P, Chandiramani VA, Beckett A, et al. The effect of
intravesical capsaicin on the suburothelial innervation in patients with detrusor hyper-reflexia. BJU International 2000;85:
238 45.
37. Avelino A, Cruz F. Peptide immunoreactivity and ultrastructure of rat urinary bladder nerve fibers after topical desensitization by capsaicin or resiniferatoxin. Auto Neurosci 2000;86:
37 46.
38. Karai L, Brown DC, Mannes AJ, et al. Deletion of vanilloid
receptor 1-expressing primary afferent neurons for pain control.
J Clin Invest 2004;113:1344 52.

REVIEW ARTICLE

Sacroiliac Joint Pain: A Comprehensive Review of Anatomy,

Diagnosis, and Treatment
Steven P. Cohen,

MD

Pain Management Divisions, Departments of Anesthesiology and Critical Care Medicine, Johns Hopkins Medical
Institutions, Baltimore, MD and Walter Reed Army Medical Center, Washington, DC

Sacroiliac (SI) joint pain is a challenging condition affecting 15% to 25% of patients with axial low back pain,
for which there is no standard long-term treatment. Recent studies have demonstrated that historical and
physical examination findings and radiological imaging are insufficient to diagnose SI joint pain. The most
commonly used method to diagnose the SI joint as a

Anatomy
The sacroiliac (SI) joint is the largest axial joint in the
body, with an average surface area of 17.5 cm2 (1).
There is wide variability in the adult SI joint, encompassing size, shape, and surface contour. Large disparities may even exist within the same individual
(2,3). The SI joint is most often characterized as a large,
auricular-shaped, diarthrodial synovial joint. In reality, only the anterior third of the interface between the
sacrum and ilium is a true synovial joint; the rest of the
junction is comprised of an intricate set of ligamentous
connections. Because of an absent or rudimentary posterior capsule, the SI ligamentous structure is more
extensive dorsally, functioning as a connecting band
between the sacrum and ilia (4). The main function of
this ligamentous system is to limit motion in all planes
of movement. In women the ligaments are weaker,
allowing the mobility necessary for parturition (Figs. 1
and 2).
The SI joint is also supported by a network of muscles
that help to deliver regional muscular forces to the pelvic
bones. Some of these muscles, such as the gluteus maximus, piriformis and biceps femoris, are functionally connected to SI joint ligaments, so their actions can affect
joint mobility. The potential for vertical shearing is
present in approximately 30% of SI joints, owing to the
Accepted for publication April 27, 2005.
Address correspondence and reprint requests to Steven P. Cohen,
MD, Johns Hopkins Hospital Pain Management Center 550 North
Broadway, Suite 301 Baltimore, MD 21205. Address electronic mail
to scohen40@jhmi.edu.
DOI: 10.1213/01.ANE.0000180831.60169.EA

1440

Anesth Analg 2005;101:144053

pain generator is with small-volume local anesthetic

blocks, although the validity of this practice remains
unproven. In the present review I provide a comprehensive review of the anatomy, function, and mechanisms of injury of the SI joint, along with a systematic
assessment of its diagnosis and treatment.
(Anesth Analg 2005;101:1440 53)

more acute angulation of the short, horizontal articular

component (5).
Age-related changes in the SI joint begin in puberty
and continue throughout life. During adolescence, the
iliac surface becomes rougher, duller, and coated in
some areas with fibrous plaques. These senescent
changes accelerate during the third and fourth decades
of life and are manifested by surface irregularities, crevice formation, fibrillation and the clumping of chondrocytes. Degenerative changes on the sacral side generally
lag 10 20 yr behind those affecting the iliac surface. In
the sixth decade, motion at the joint may become markedly restricted as the capsule becomes increasingly collagenous and fibrous ankylosis occurs. By the eighth
decade of life, erosions and plaque formation are inevitable and ubiquitous (4).

Innervation
The innervation of the SI joint remains a subject of much
debate. The lateral branches of the L4-S3 dorsal rami are
cited by some experts as composing the major innervation to the posterior SI joint (1). Other investigators claim
that L3 and S4 contribute to the posterior nerve supply
(6,7). The innervation of the anterior joint is similarly
ambiguous. Early 20th century German literature asserts
the anterior SI joint is supplied by the obturator nerve,
superior gluteal nerve and the lumbosacral trunk (8).
More recent literature suggests the anterior joint is innervated by L2-S2 (1), L4-S2 (9), and the L5-S2 ventral
rami (10). Some authors have even suggested that the
anterior SI joint is devoid of nervous tissue (7,11). In a
study testing the ability of L5 dorsal ramus and S1-4
2005 by the International Anesthesia Research Society
0003-2999/05

ANESTH ANALG
2005;101:1440 53

REVIEW ARTICLE
COHEN
SACROILIAC JOINT PAIN

1441

Figure 1. Posterior view of the articulations and associated ligaments of the

sacroiliac joint and surrounding structures. Drawing by Jee Hyun Kim.

Figure 2. Anterior view of the articulations and associated ligaments of the

sacroiliac joint and surrounding structures. Drawing by Jee Hyun Kim.

lateral branch blocks to protect the SI joint from an

experimental stimulus, 6 of 10 subjects retained the ability to perceive ligamentous probing (12).
A neurophysiologic study conducted in cats identified 29 mechanosensitive afferent units, 26 of which
were found in the joint capsule and 3 in adjacent
muscles (13). Twenty-eight of these units were classified as nociceptive and 1 as proprioceptive. Among
these 29 receptive fields, 16 were located in the proximal third of the posterior SI joint and 11 in the middle

third. The average mechanical threshold of an SI joint

nociceptive unit was 70 g, as compared to the 6 g mean
mechanical threshold for lumbar facet joint nociceptive units and the 241 g threshold for units residing in
the anterior lumbar disk (14 16). This indicates that
the pain sensitivity of the SI joints may be lower than
that of the lumbar facet joints but higher than the
anterior portions of lumbar discs. As all animals underwent posterior midline incisions, somatosensory
units in the anterior SI joint were not stimulated.

1442

REVIEW ARTICLE COHEN

SACROILIAC JOINT PAIN

ANESTH ANALG
2005;101:1440 53

Function and Biomechanics

Prevalence

The SI joints are designed primarily for stability. Their

functions include the transmission and dissipation of
truncal loads to the lower extremities, limiting x-axis
rotation, and facilitating parturition. Compared to the
lumbar spine, the SI joints can withstand a medially
directed force 6 times greater but only half the torsion
and 1/20th of the axial compression load (17). These
last 2 motions may preferentially strain and injure the
weaker anterior joint capsule (18).
There have been numerous attempts to discern
the biomechanics of the SI joint. These motion studies can be summarized as follows: the SI joint rotates
about all 3 axes, although the movements are very
small and difficult to measure (19,20). Miller et al.
(21) studied the load-displacement behavior of single and paired SI joints in 8 elderly cadavers. Various static test loads were applied in the superior,
lateral, anterior and posterior directions, and rotations about all 3 axes were measured. These tests
were conducted with one and both ilia fixed. The
authors found that with 1 leg immobile, movements
in all planes ranged from between 2 to 7.8 times
more than that measured with both legs fixed. In a
series of cadaveric studies, Vleeming et al. (22,23)
found that the total range of motion during flexion
and extension at the SI joint rarely exceeded 2 degrees, with 4 degrees being the upper limit during
sagittal rotation. In another cadaver study, Brunner
et al. (24) found that the main motion in male specimens
tended to be translation, whereas in female specimens it
was rotational. The maximum range of motion in this
study was 1.2 degrees in men and 2.8 degrees in women.
In a study by Egund et al. (25) examining SI joint movements in 4 volunteers using radiographic stereophotogrammetry, the authors found the maximal rotations
and translations to be 2.0 degrees and 2.0 mm, respectively. A larger study (n 24) by Jacob and Kissling (26)
conducted in healthy young and middle-aged men and
women found similarly small limits on rotation (1.7 degrees) and translation (0.7 mm). However, one individual with a history of SI joint pain exhibited more than
6 degrees of rotation about the y-axis. Finally, Sturesson
et al. (27) measured multiple SI joint movements in 25
patients diagnosed with SI joint pain. Movements in all
planes were found to be small, with translations never
exceeding 1.6 mm, and rotations being limited to 3 degrees. No differences were found between symptomatic
and asymptomatic joints, leading the authors to conclude that 3-dimensional motion analysis was not useful
for identifying painful SI joints in most patients. Possible
exceptions to the finding that hypermobility is not a
typical cause of SI joint pain include traumatic instability, multiparity, muscular atrophy, and lower motor
neuron disease (28).

Although it is widely acknowledged that dysfunctional SI joints may cause low back pain (LBP), the
prevalence of this condition has not been well studied.
Prevalence studies are further compromised by the
fact that most have used either physical examination
findings and/or radiological imaging techniques to
make the diagnosis of SI joint pain. The largest of
these is a retrospective study by Bernard and
Kirkaldy-Willis (29), who found a 22.5% prevalence
rate in 1293 adult patients presenting with LBP.
Diagnoses in this series were based predominantly
on physical examination.
Schwarzer et al. (30) conducted a prevalence study
involving 43 consecutive patients with chronic LBP
principally below L5-S1 using fluoroscopically guided
SI joint injections. Fifty-seven other patients with LBP
were excluded on the basis of more rostral symptoms.
Three criteria were used to diagnose SI joint pain:
analgesic response to local anesthetic (LA), abnormalities on post-arthrography computed tomography
(CT) scanning, and concordant pain provocation during joint distension. Using significant pain relief after
LA injection as the sole criterion for diagnosis, the
prevalence of SI joint pain in the 43 subjects was
determined to be 30% (95% confidence interval [CI],
16% 44%), with 4 patients obtaining complete pain
relief. Using analgesic response combined with a ventral capsular tear (the most common radiologic finding) as the criteria, the prevalence decreased to 21%
(95% CI, 9%33%). Only 7 patients satisfied all 3 diagnostic criteria, for a lower limit prevalence rate of
16% (95% CI, 5%27%). The presence of groin pain
was the only referral pattern found to distinguish
patients with SI joint pain from those with LBP of
non-SI joint origin.
Maigne et al. (31) conducted a prevalence study in
54 patients with unilateral LBP using a series of blocks
done with different LA based on International Spinal
Injection Society guidelines (32). Nineteen patients
had a positive response (75% pain relief) to the
lidocaine screening block. Among these patients, 10
(18.5%) responded with 2 h pain relief after the
confirmatory block with bupivacaine and were considered to have true SI joint pain (95% CI, 9%29%).
Based on these studies, the prevalence of SI joint pain
in carefully screened LBP patients appears to be in the
15%25% range.

Mechanism of Injury
The mechanism of SI joint injury has previously been described as a combination of axial loading and abrupt rotation (18). On an anatomic level, pathologic changes affecting many different SI joint structures can lead to
nociception. These include capsular or synovial disruption,

ANESTH ANALG
2005;101:1440 53

REVIEW ARTICLE
COHEN
SACROILIAC JOINT PAIN

1443

Table 1. Sacroiliac Joint Involvement and Other Characteristics of the Adult Spondylarthropathies
Clinical
characteristic

Ankylosing
spondylitis

Reactive arthritis
(Reiters syndrome)

Clinical
Sacroiliitis
Symmetry of SI
Joint
Involvement
Onset
Peripheral Joint
Involvement
Sex Ratio

Almost universal

20%30%

20%30%

15%25%

Symmetric or
alternating

Mostly asymmetric

Mostly asymmetric

Mostly asymmetric

35 years
20%30%

Young to middle-aged
Approximately 90%

Young to middle-aged
Almost universal

Young to middle-aged
15%30%

M:F 3:1

M:F 5:1

Males females

Males females

Psoriatic arthritis

Enteropathic arthritis

SI sacroiliac.

capsular and ligamentous tension, hypomobility or hypermobility, extraneous compression or shearing forces, abnormal joint mechanics, microfractures or macrofractures,
chondromalacia, soft tissue injury, and inflammation.
Mechanistically, there are numerous reported etiologies for
SI joint pain. To simplify matters, these causes can be divided into intraarticular and extra-articular sources. Arthritis and infection are two examples of intraarticular causes of
SI joint pain. Extra-articular sources are the more common
of the two and include enthesopathy, fractures, ligamentous injury, and myofascial pain. Clinical studies have demonstrated significant pain relief after both intraarticular and
periarticular SI joint injections (3336).
In addition to etiologic sources, there are numerous
factors that can predispose a person to gradually develop SI joint pain. Risk factors that operate by increasing the stress borne by the SI joints include true
and apparent leg length discrepancy (37), gait abnormalities (38), prolonged vigorous exercise (39), scoliosis (40), and spinal fusion to the sacrum (41). Whereas
increased SI joint uptake using scintigraphy has been
demonstrated after lumbar spine fusion (42), at least
one study examining the long-term effects of spinal
fusion on SI joint function concluded that neither biomechanical nor anatomical changes were more common in fusion patients than in those who underwent
decompression procedures (43). Lumbar spine surgery has also been purported to trigger SI joint pain
for reasons unrelated to increased force transmission.
These factors include SI ligament weakening and/or
surgical violation of the joint cavity during iliac graft
bone harvest (44) and postsurgical hypermobility (45).
Pregnancy predisposes women to SI joint pain via
the combination of increased weight gain, exaggerated
lordotic posture, the mechanical trauma of parturition,
and hormone-induced ligamental laxity (46,47). The
laxity associated with pregnancy is attributable to increased levels of estrogen and relaxin, and it predisposes parturients to sprains of the SI joint ligaments.
SI subluxation has also been reported to cause back
pain in pregnancy (48).
Inflammation of one or both SI joints is considered to be an early and prominent symptom in all

seronegative and HLA-B27-associated spondylarthropathies (49). Although the precise etiology of

spondylarthropathy remains unknown in most patients, the strong association with HLA-B27 supports the view that these conditions are attributable
to a genetically determined immune response to
environmental factors in susceptible individuals. In
a subset of patients with Reiters syndrome/reactive
arthritis, the disease is clearly induced by infection
(50) (Table 1).
The specific etiologies that can result in SI joint pain
are widespread and protean. Potential causes range
from rare events such as pyogenic infection (51) and
malignancy (52), to more mundane occurrences such
as bracing ones legs in a motor vehicle accident (53),
falls (53), athletic injuries (54), prolonged lifting and
bending (55), and torsional strain (55). In a retrospective study by Chou et al. (56) assessing the inciting
events in 54 patients with injection-confirmed SI joint
pain, the authors found trauma was the cause in 44%
of patients, 35% were idiopathic, and 21% were attributed to the cumulative effects of repeated stress. In the
24 patients who cited trauma as the source of their
pain, the most common events were motor vehicle
accidents (n 13), falls onto the buttock (n 6), and
childbirth (n 3).

Diagnosis and Presentation

History and Physical Examination
One of the most challenging aspects of treating SI joint
pain is the complexity of diagnosis. Literally dozens of
physical examination tests have been advocated as
diagnostic aids in patients with presumed SI joint pain
(57). Many involve distraction of the SI joints, with 2 of
the most common ones being Patricks test and Gaenslens test. Despite the plethora of diagnostic tests, clinical studies have for the most part demonstrated that
neither medical history nor physical examination findings are consistently capable of identifying dysfunctional
SI joints as pain generators (30, 58, 59) (Table 2). In
addition, Dreyfuss et al. (60) found 20% of asymptomatic

1444

REVIEW ARTICLE COHEN

SACROILIAC JOINT PAIN

ANESTH ANALG
2005;101:1440 53

Table 2. Studies Assessing Accuracy of History and Physical Examination in the Diagnosis of Injection-Confirmed
Sacroiliac (SI) Joint Pain
Author, year

Number and type of

patients

Diagnostic
standard

Results

Comments

Schwarzer et al., 1995

(30)

Cross-sectional,
analytic study

43 pts with chronic

axial LBP principally
below L5S1

Used 75% pain

relief to a single
block as dx
criteria.

No PE test was of
predictive value in
predicting subsequent
response to block. Only
groin pain found to be
more common in pts
with () dx block.

Dreyfuss et al., 1996

(59)

Prospective study
assessing value
of hx and 12 PE
tests in
diagnosing SI
joint pain

85 pts with axial LBP

principally below L5

Used 90% pain

relief to a single
SI joint block as
criteria for dx.

45 pts had a () block, 40

a () block. No
historical or PE finding
predicted response to
block.

Maigne et al., 1996

(31)

Prospective study
assessing the
prevalence of
SI joint pain
using double
blocks and the
accuracy of
pain
provocation
tests to dx the
disorder.

67 pts with chronic

unilateral LBP
without extension
below the knee.

Criteria for dx was

75% pain
relief after
lidocaine
screening block
followed by
similar pain
relief lasting
2 hours after
confirmatory
bupivacaine
block.

Broadhurst and Bond,

1998 (105)

Double-blind
study
determining the
value of
Patricks
(FABER) test,
posterior shear
test and
resisted
abduction in
the dx of SI jt
pain.
Prospective
cohort study
assessing
predictive
value of
provocative
tests in dx SI
joint pain.
Prospective
validity study
to identify
association
between PE &
facet,
discogenic, and
SI jt pain.

40 pts with LBP and

pain reproduction
using the 3 PE tests.

All pts had either

saline or 1%
lidocaine
injected into 1 SI
jt. Dx based on
70% pain
reduction when
provocative test
performed after
SI jt block.

Of 54 pts completing the

study, 19 experienced
good pain relief with
the screening block, and
10 of these (18.5%) recd
2 hours pain relief
after the confirmatory
block. There was no
association between any
clinical variable or pain
provocation test and a
() response to both
blocks.
Patricks test found to
have 77% sensitivity and
100% specificity;
posterior shear test had
80% sensitivity and
100% specificity; resisted
abduction had 87%
sensitivity and 100%
specificity.

30% of pts dxed with SI

jt pain based on dx
block, 21% based on
response to block and
ventral capsular tear
on CT scanning, and
16% based on
response to block,
capsular tear and
pain provocation.
Response to previous
therapy not indicative
of SI joint pain. Pts
with negative
response slightly
more likely to have
pain above L5.
7 pts excluded b/c SI jt
could not be entered,
3 b/c of sciatic palsy
after screening block
and 3 b/c they
remained pain-free
after screening block.

50 pts without
spondylarthropathy
who had ()
response to 3 dx PE
tests.

Used 80% pain

relief to a single
block as criteria
for dx.

SI jt blocks were () in 30
pts for a positive
predictive value of 60%.

55 of 102 pts with

chronic axial LBP
underwent SI joint
blocks.

Used not only

pain relief (
80%) to a single
block as dx
criteria, but also
concordant pain
provocation.

22 of 57 injected joints had

() response to SI jt
block. Pts with SI jt pain
rarely had midline LBP
or pain above L5.
Positive correlation
noted between SI jt pain
and increased pain
when rising from sitting,
unilateral pain and 3
positive pain
provocation tests.

Slipman et al., 1998

(58)

Young et al., 2003

(106)

Study type

No pt had 70% pain

reduction after saline
injection.

2 of 3 () dx tests had
to be Patricks test
and sacral sulcus
tenderness. Steroid
added to dx block,
with average
symptom reduction
being 30.5%.
5 provocation tests used
to examine the SI
joint. Clinical
evaluation done by
physical therapists.
Also sought to
identify clinical
determinates of
discogenic LBP and
lumbar facet pain.

Dx diagnosis, diagnostic; Hx history; Jt joint; PE physical examination; LBP low back pain; pts patients; CT computed tomography; b/c
because.

adults had positive findings on 3 commonly performed

SI joint provocation tests.
The reliability of provocative SI joint maneuvers
and alignment/mobility tests has also been questioned, with most studies conducted by chiropractors
and physical therapists. Whereas some of these studies
have found moderate to high inter-examiner reliability
(61 63), most have not (64 68). Generally, reproducibility has been found to be greater for provocative tests
than for mobility and alignment assessments. In the
Dreyfuss et al. study (59) conducted in 85 patients with

injection-confirmed SI joint pain, there was moderate

agreement (kappa 0.6) between chiropractors and
medical doctors with regard to provocative maneuvers
of painful joints. Even when agreement was perfect, the
maneuvers were still found to lack diagnostic utility.

Radiological Studies
Results of studies examining radiologic findings in
patients with SI joint pain have been similarly disappointing. In studies by Maigne et al. (69) and Slipman

ANESTH ANALG
2005;101:1440 53

et al. (70), the investigators found sensitivities of 46%

and 13%, respectively, for the use of radionuclide bone
scanning in the identification of SI joint pain. Despite
the high specificites in these studies (89.5% for Maigne
et al. and 100% for Slipman et al.), the low sensitivies
indicate bone scanning is a poor screening test for SI
joint pain. Poor correlation with diagnostic injections
and symptoms have also been found for CT and radiographic stereophotogrammetry (71,27). In a retrospective analysis by Elgafy et al. (71), CT imaging was
found to be 57.5% sensitive and 69% specific in diagnosing SI joint pain.

Pain Referral Patterns

There have been several attempts to identify pain
referral patterns from SI joints. In one of the earliest
studies conducted in 10 asymptomatic volunteers,
Fortin et al. (72) performed provocative SI joint
injections using contrast and lidocaine. Sensory
changes were localized to the ipsilateral medial buttock inferior to the posterior superior iliac spine in 6
of the 10 subjects. In 2 subjects, the area of hyperesthesia extended to the superior aspect of the
greater trochanter. The last 2 subjects experienced
sensory changes radiating into the upper thigh. In a
follow-up study, independent examiners selected 16
individuals among 54 with chronic LBP whose pain
diagrams most closely resembled the pain referral
patterns obtained in the first study (73). These 16
patients proceeded to undergo provocative SI joint
injections with contrast and LA. All 16 experienced
concordant pain during the injection, with 14 obtaining pain relief after deposition of LA. Ten patients reported 50% pain reduction. Six of the 16
patients had ventral capsular tears revealed during
arthrography. After the SI joint injections, provocative discography and lumbar facet joint injections
were performed in 9 patients each. In none of the
patients was either test positive.
Slipman et al. (74) conducted a retrospective study
to determine the pain referral patterns in 50 patients
with injection-confirmed SI joint pain. In contrast to
the findings by Fortin et al. (72) and Schwarzer et al.
(30), the authors found the most common referral
patterns for SI joint pain to be radiation into the buttock (94%), lower lumbar region (72%), lower extremity (50%), groin area (14%), upper lumbar region (6%),
and abdomen (2%). Twenty-eight percent of patients
experienced pain radiating below their knee, with 12%
reporting foot pain. Based on the existing data, the
most consistent factor for identifying patients with SI
joint pain is unilateral pain (unless both joints are
affected) localized predominantly below the L5 spinous process (30,59,7274).

REVIEW ARTICLE
COHEN
SACROILIAC JOINT PAIN

1445

Diagnostic Blocks
It is often assumed that an analgesic response to a
properly performed diagnostic block is the most reliable method to diagnose SI joint pain. Although this
may seem to be self-evident, the validity of intraarticular SI joint blocks remains unproven. There are many
factors that can impact on the sensitivity and specificity of diagnostic blocks. These include the placebo
effect, convergence and referred pain, neuroplasticity
and central sensitization, expectation bias, unintentional sympathetic blockade, systemic absorption of
LA, and psychosocial issues (75). In addition, SI joint
block can be one of the most challenging spinal injection procedures. Extravasation of LA to surrounding
pain-generating structures such as muscles, ligaments,
and lumbosacral nerve roots can lead to false-positive
blocks. Conversely, failure to obtain adequate LA
spread to the anterior and cephalad portions of the SI
joint can result in false-negative blocks. In a classic
study by North et al. (76) examining the specificity
and sensitivity of a battery of lumbosacral LA blocks
in 33 patients with a chief complaint of sciatica, the
authors found the specificity of all blocks to be exceedingly low. SI joint blocks were not performed in this
study.
In a pilot study by Fortin et al. (72) mapping SI joint
referral patterns in asymptomatic volunteers, extravasation of contrast (mean 1.6 mL injected) occurred in 9 of 10
subjects during SI joint injection, with half having at least
moderate spread outside the joint. After the injection of
LA, 40% of subjects noted lower extremity numbness,
indicating inadvertent anesthetization of the lumbosacral nerve roots. In the Maigne et al. (31) study, 3 of the
initial 67 patients were excluded because of sciatic
palsy after the screening block and another 7 were
excluded because penetration of the SI joint was impossible. Other investigators have reported much less frequent (5%) failure rates with fluoroscopically guided
SI joint injections (30,59,77). Technical difficulties may be
more frequently encountered in elderly patients and
those with spondylarthropathies, in whom degenerative
changes are more pronounced. If persistent difficulties
entering the SI joint are encountered using fluoroscopy,
the 3-dimensional imaging capabilities of CT may facilitate entry into the joint (34, 78) (Fig. 3).
Regardless of the imaging modality used to confirm
intraarticular injection, SI joint injections should never be
performed blindly. Rosenberg et al. (79) performed a
double-blind study in 37 patients (39 joints) to determine
the accuracy of clinically guided SI joint injections using
CT imaging as the standard. The authors found that
intraarticular injection was accomplished in only 22% of
patients, whereas sacral foraminal spread occurred 44%
of the time. In 3 patients, no contrast was seen on CT
scanning, indicating probable vascular uptake. In 24% of
injections, contrast extended into the epidural space.

1446

REVIEW ARTICLE COHEN

SACROILIAC JOINT PAIN

ANESTH ANALG
2005;101:1440 53

that have been advocated for this disorder. Generally,

these treatments can be divided into 2 categories:
those directed at correcting the underlining pathology
and those aimed at alleviating symptoms. For both of
these categories, the evidence supporting any one
therapy is limited by the lack of controlled outcome
studies.

Psychosocial Issues

Figure 3. Anteroposterior fluoroscopic image demonstrating a rightsided sacroiliac joint block with minimal extra-articular extravasation of contrast.

As for facet blocks, some experts have advocated using a series of SI joint blocks to reduce the incidence of
false-positives. In a prospective study involving 67 patients with unilateral LBP, SI joint-compatible referral
patterns and joint tenderness, Maigne et al. (31) sought
to determine the prevalence of SI joint pain using a series
of blocks with 2 different LA. In the 54 patients who
completed the study, 19 obtained 75% pain relief with
the lidocaine screening block. After the bupivacaine confirmatory block, only 10 of the 19 patients achieved
75% pain relief lasting 2 or more hours, for a prevalence rate of 18.5%. The false-positive rate of 17% in this
study is less than that previously reported for lumbar
facet blocks (80). Yet without a diagnostic gold standard, there is no way of determining how many true
positives were false positives and how many false positives were actually true positives. In clinical practice
confirmatory SI joint blocks are almost never performed
because a) the block itself is considered to be definitive
treatment; b) double-blocks are not cost-effective (81);
and c) the negative consequences of obtaining a false
false-positive block (misdiagnosing true SI joint pain)
outweigh the ramifications of overdiagnosing the condition. In summary, there is no infallible, universally accepted method for diagnosing pain originating in the SI
joint(s).

Treatment
The treatment of SI joint pain is widely acknowledged
to be one of the most challenging problems confronting pain physicians. Evidence supporting this statement can be seen by the plethora of different therapies

Recent studies have provided incontrovertible evidence that psychopathology and other psychosocial
factors can influence both the development of chronic
pain conditions and the response to treatment. In a
study by Polatin et al. (82) conducted in 200 chronic
LBP patients, the authors found that 77% met lifetime
criteria and 59% demonstrated current symptoms for
at least one psychiatric diagnosis, with the most common being depression, substance abuse, and anxiety
disorders. Notably, more than 50% of those with depression and more than 90% of patients with substance abuse or an anxiety disorder experienced
symptoms before the onset of LBP. Most, but not all,
studies have shown untreated psychopathology to
negatively affect LBP treatment outcomes (83).
In addition to psychiatric illness, social factors have
been demonstrated to impact the prognosis of LBP.
These include return-to-work issues, secondary gain,
catastrophizing, poor role models, codependency behavior, inadequate coping mechanisms, and attitudes,
beliefs, and expectations (84). To optimize outcomes,
the identification and treatment of concomitant psychosocial issues is of paramount importance. This is
best accomplished via a multidisciplinary approach.

Conservative Management
The non-interventional management of SI joint pain
should ideally address the underlining pathology. In
patients with true or apparent leg length discrepancy,
this might include the use of shoe inserts to more
equitably distribute the load borne by the SI joints.
Because leg length discrepancies are frequently found
in asymptomatic individuals (37) and many patients
already compensate for their lower extremity length
difference by altering their gait or posture, most experts recommend starting out cautiously with inserts
that correct only half the incongruity. For SI joint pain
resulting from altered gait mechanics and spine malalignment, physical therapy and osteopathic or chiropractic manipulation have been reported to reduce
pain and improve mobility (85,86). However, there are
no prospective, controlled studies supporting these
modalities.
Nonsurgical stabilization programs have been advocated for SI joint pain. These range from the application of pelvic belts that reduce the sagittal rotation
of incompetent SI joints in pregnant women (87,88) to

ANESTH ANALG
2005;101:1440 53

REVIEW ARTICLE
COHEN
SACROILIAC JOINT PAIN

1447

Table 3. Evidence Supporting Various Pharmacologic Therapies in Spondylarthropathies

Drug class/name

Conditions studied

Evidence for efficacy

Antibiotics

Re spA

Moderate evidence when a triggering organism such as

Yersinia or Chlamydia is suspected. No evidence
otherwise.
Strong evidence for infliximab and etanercept in AS,
weak evidence for other agents in AS. Weak evidence
for Ps spA and undifferentiated spA.
Most studied in Ps spA and AS, whereby effects on
peripheral arthritis may be for axial symptoms. For
axial symptoms evidence is very weak. Mixed evidence
for Re spA and juvenile onset spA. Weak evidence for
Ps spA.
Moderate evidence
Mixed evidence
Mixed evidence for AS. Weak evidence Ps spA. No
evidence for other anti-malarial drugs.
Moderate evidence
Only placebo-controlled trial showed reduction in pain
for amitriptyline.
Strong evidence of Ps spA. Moderate evidence only for
peripheral symptoms in AS. Efficacy may be better for
IV than oral treatment.
Strong evidence

Anti-tumor necrosis
factor-alpha

AS, Ps spA,
Undifferentiated spA

Sulfasalazine

All types

Cyclosporine
D-Penicillinamine
Quinine and its
derivatives
Oral corticosteroids
Tricyclic antidepressants

Ps spA
AS
AS, Ps spA,
seronegative spA
AS
AS

Methotrexate

AS and Ps spA

Nonsteroidal antiinflammatory drugs

Gold salts
Levamisole
Leflunomide

All, but mostly AS and

Ps spA
Ps spA
Seronegative spA
Ps spA, AS

Azathioprine
Bisphosphonates

Re spA, Ps spA,
seronegative spA
Mostly AS

Bromocriptine

All types

Heat-killed
Mycobacterium vaccae
Interleukin 1 antagonists
Monoclonal antibodies

Ps spA

Moderate evidence. Effect on axial symptoms may be

for peripheral arthritis.
No evidence for re spA. Weak evidence for PsA and
other seronegative spA.
No evidence

AS
Ps spA

Weak evidence
Weak evidence

Moderate-strong evidence
Moderate-strong evidence
Moderate-strong evidence for Ps spA. In AS, effects on
peripheral arthritis for axial symptoms.
Moderate evidence

Strong evidence one or more placebo-controlled trials coupled with evidence from other studies; moderate evidence 1 placebo-controlled study with
moderate support from comparative or open-label studies, or strong support from comparative or open-label studies; weak evidence some support from
open-label or comparative studies, or mixed evidence from placebo-controlled studies; no evidence the evidence against outweighs the evidence supporting
efficacy; AS ankylosing spondylitis; SpA spondylarthropathy; Re spA reactive spondylarthropathy/Reiters syndrome; Ps spA Psoriatic spondylarthropathy.

exercise-induced pelvic stabilization programs (89). In

a study by Mooney et al. (89), the authors found that
5 women with injection-confirmed SI joint pain had
electromyographic-documented hyperactivity of the
ipsilateral gluteus muscles and contralateral latissimus muscle compared with 15 asymptomatic control
patients. After a 2-1/2 month exercise program, all 5
patients achieved a significant reduction in pain and a
return of myoelectric activity to normal patterns.
Ankylosing spondylitis (AS), the most well known
and studied spondylarthropathy, is an inflammatory
rheumatic disease characterized by spine and SI joint
involvement that manifests as spondylitis and sacroiliitis. There are numerous studies assessing the efficacy of various pharmacologic treatment modalities in
AS and other spondylarthropathies, but the conclusions that can be drawn from these studies are limited

by several factors. These include the protean nature of

rheumatic involvement, the systemic manifestations accompanying the disorders, and the fact that although the
outcome measures generally include changes in pain
complaints and mobility, most do not specifically address SI joint pain. Table 3 lists the evidence supporting
various pharmacotherapies for AS and other
spondylarthropathies.

Intraarticular Injections
Intraarticular injections with steroid and LA often
serve the dual function of being therapeutic and aiding in diagnosis. To summarize these studies, most
but not all investigators have found radiologically
guided SI joint injections to provide good to excellent
pain relief lasting from 6 mo to 1 yr (Table 4). Along

1448

REVIEW ARTICLE COHEN

SACROILIAC JOINT PAIN

ANESTH ANALG
2005;101:1440 53

Table 4. Clinical Studies Evaluating Corticosteroid Injections for Sacroiliac (SI) Joint Pain
Study type

Number and type of

patients

Maugars et al., 1992 (107)

Prospective
observational

24 patients w/seronegative
spondylarthropathy

42 corticosteroid
injections without
LA. Eighteen pts
underwent bilateral
injections, 6
unilateral.

Bollow et al., 1996 (78)

Prospective
observational

66 patients with
spondylarthropathy

103 corticosteroid
injections without
LA. Used an average
of 10 ml of
superficial LA per
joint before during
procedure.

Braun et al., 1996 (34)

Prospective,
observational

30 patients with
spondylarthropathy

54 corticosteroid
injections without
LA.

Clear improvement in
pain and MRI
demonstrated
inflammation in 83%
of patients, lasting 8.9
/5 months.

Maugars et al., 1996 (33)

Placebo-controlled
double-blind

10 patients with
spondylarthropathy,
13 joints. Pts with
degenerative SI
joints and complete
ankylosis excluded.

13 total joints injected.

6 were injected with
corticosteroid
without LA and 7
with normal saline. 6
of 7 placebo pts were
re-injected with
steroid at 1 month.

Luuk-kainen et al., 1999

(35)

Randomized,
controlled study

20 pts with seronegative

spondylarthropathy
recd steroid and LA;
10 pts recd saline
and LA. All pts had
unilateral blocks.

Dussault et al., 2000 (77)

Retrospective chart
review

24 pts with pain

overlying the SI jt (n
11), LBP (n 10)
or hip pain (n 3),
31 joints. Data
evaluated on 28
joints.

All pts underwent

unilateral,
periarticular
injections. 10 recd
corticosteroid
without LA; 10 recd
normal saline with
LA.
Total of 31 joints
injected with
corticosteroid and
LA. 4 subjects
underwent bilateral
injections, 3 repeat
injections.

5 steroid joints had

good or very good
pain relief at 1 month
vs. 1 in placebo
group. Overall, 12/14
SI joints had good or
very good results at 1
month, 8/13 at 3
months and 7/12 at 6
months.
At 2-month follow-up,
VAS pain scores
decreased
significantly in the
steroid but not saline
group.

Gunaydin et al., 2000 (108)

Prospective
observational

9 pts with
spondylarthropathy.

16 joints injected with

corticosteroid
without LA.

Hanly et al., 2000 (109)

Prospective, singleblind (evaluators)

19 pts with
spondylarthropathy.
13 had radiologic
evidence of
sacroiliitis and 6 had
normal imaging
studies.

All pts underwent

bilateral injections
with corticosteroid
without LA.

Author, year

Treatment

Primary outcome

Comments

67% of joints
experienced 80%
pain relief, 19% 50%
80% improvement
and 14% had 50%
pain relief. Mean
duration of
improvement 8.4 /
4.2 months.
92.5% of patients had
significant
improvement of pain
after a mean of 1.7
wks. Mean duration
of pain relief 10 /
5 months.

Dx made by PE and
radiologic studies.
Fluoroscopy used to
guide injections.
Good pain relief was
correlated with
shorter duration of
symptoms.

Pain decreased by
80% in 7 joints, by
50%70% in 11 joints
and 50% in 10
joints. More than 50%
relief was obtained in
55% of joints with
normal radiographs,
in 62% of joints with
degenerative joint
disease, and in the
only pt with
ankylosing
spondylitis.
7 of 9 pts reported
improvement (mean
decrease in VAS
scores 49%, mean
duration of pain
relief 10.8 / 5.6
months).
Both groups
experienced
significant pain relief
1 month after
injections, with no
difference between
groups. 6 months
postinjection, there
was no difference in
pain or stiffness
compared to baseline
in either group.

Dx made by PE. CT
used to guide
injections. No
difference in
Schobers sign or
range of motion
before and after
treatment. ESR and
CRP decreased after
rx.
Dx made by PE and
contrast
enhancement on
dynamic MRI. CT
used to guide
injections. No
difference in
Schobers sign
before and after rx.
Both ESR and CRP
decreased after rx.
Dx made by PE and
radiologic studies.
Fluoroscopy used to
guide injections. One
pt developed
radicular pain that
lasted 3 weeks.

Injections were
periarticular, not
intra-articular. Dx
made by PE and
radiologic studies.
Fluoroscopy used to
guide injections.
Injections done by
radiologists with
fluoroscopic
guidance. In 1 pt the
joint could not be
penetrated and 2 pts
developed lower
extremity weakness.

Dx by PE and
radiographic studies.
MRI used to guide
injections. CRP but
not ESR decreased
after rx.
CT used to guide
injections. Pts
randomized by
radiologic imaging,
not PE. Spinal
mobility was not
changed in either
group over the
study period. Main
fault of this trial is
that radiological
studies have been
shown to be
insensitive as a
screening tool for SI
joint pain.
Continued

ANESTH ANALG
2005;101:1440 53

REVIEW ARTICLE
COHEN
SACROILIAC JOINT PAIN

1449

Table 4. Continued
Author, year

Study type

Number and type of

patients

Treatment

Primary outcome

Pereira et al., 2000

(110)

Prospective
observational

12 patients with
spondylarthropathy (3 with
ankylosing spondylitis) and
buttock pain

24 injections with
corticosteroid without
LA. 9 pts had bilateral
injections.

Clinical improvement noted

in 10 patients, with a mean
pain-free period of 9.6
months.

Pereira et al., 2000

(111)

Prospective
observational

10 patients with sacroiliitis

(9 bilateral)

21 injections with
corticosteroid without
LA. 9 pts had bilateral
injections.

Good to excellent pain relief

in 8 of 10 pts lasting a mean
of 13.5 months. The 2 nonresponders suffered from
fibromyalgia and reactive
depression.

Ojala et al., 2001

(112)

Prospective
observational

20 patients with low back

pain

Total number of
injections not noted.
Used corticosteroid
with LA.

60% of patients had

significant short-term pain
reduction after injections.

Karabacakoglu et al., Prospective

2002 (113)
observational

17 patients with ankylosing

spondylitis.

5 pts underwent
bilateral injections.
Used corticosteroid
without LA.

15 of 17 patients reported
good relief 1 month after
injection, with 2 reporting
fair relief.

Luuk-kainen et al.,
2002 (36)

Randomized,
controlled study

24 pts without
spondylarthropathy.

At 1-month follow-up, VAS

pain scores decreased
significantly more in the
steroid group than in the
saline group.

Katz et al., 2003 (41)

Retrospective chart 34 pts with low back pain

review
after lumbar spinal
fusion to the sacrum.

All pts underwent

unilateral,
periarticular
injections. 13 pts
recd corticosteroid
and LA, with 11 pts
receiving normal
saline and LA.
Total number of
injections not noted.
Used corticosteroid
with LA.

Fischer et al., 2003

(114)

Randomized,
controlled

89 children with juvenile

spondyarthropathy. 56
were responders to
NSAIDs (control group)
and 33 were nonresponders (treatment
group).

Comments
MRI used to guide
injections. In 1 pt
adequate needle position
was not obtained due to
software failure. Three
months after injection
there was a significant
decrease in marrow
edema in 2 pts, a marked
decrease in 5 pts and a
moderate decrease in 3
pts.
MRI used to guide
injections. Subchondral
marrow edema resolved
on follow-up MRI
minimally in 3 pts,
partially in 3 pts and
completely in 3 pts.
Dx made by history and
PE. All pts had normal
imaging studies. MRI
used to guide SI joint
injections.
CT used to determine
needle puncture point
and angle of
intervention.
Fluoroscopy used to
guide injections.
Injections were
periarticular, not
intraarticular. Dx made
by PE. No pt had
radiologic evidence of
sacroiliitis. Fluoroscopy
used to guide
injections.
Dx considered based on
history and PE.
Fluoroscopy used to
guide injections. 10 pts
excluded because they
had multiple injections
at the same visit.

59% (n 20) of pts had

75% pain relief 15
45 minutes after injections
and were thus diagnosed
with SI joint pain. 11 of
the 20 experienced 75%
relief lasting 2 wks,
while 6 had moderate pain
relief.
Treatment group recd
87.5% of children who recd Dx made clinically and
corticosteroid without
injections reported
by MRI evidence of
LA injections plus
significant decrease in
sacroiliitis. CT used to
NSAIDs (27 bilateral
their pain complaints over
guide injections. Oneinjections). The
the 20-month follow-up
third of patients who
control group was
period (mean VAS pain
recd injections
continued on NSAIDs
score decreased from 6.9
demonstrated
without injections.
to 1.8). The control group
continued joint
showed similar
destruction despite
improvement in pain
absence of subjective
scores, with no difference
complaints.
between groups.

Dx diagnosis; CT computed tomography; Rx treatment; PE physical examination; ESR erythrocyte sedimentation rate; CRP C-reactive protein;
LA local anesthetic; pts patients; MRI magnetic resonance imaging; VAS visual analog scale; LBP low back pain; NSAIDs nonsteroidal
antiinflammatory drugs.

with a multitude of studies demonstrating prolonged

pain relief after intraarticular SI joint steroid injections, double-blind studies have shown a beneficial
effect for periarticular corticosteroid treatment as well
(35,36).

Radiofrequency Denervation Procedures

Several investigators have performed radiofrequency
(RF) denervation procedures in an attempt to provide
prolonged pain relief to patients suffering SI joint
pain. The techniques used have ranged from denervating the nerves supplying the SI joint (90 92) to cre-

ating lesions in the joint itself (93), with one study

using a combination of the two (94). The success rates
of studies targeting the nerve supply are higher than
those focusing on the joint itself, with approximately
two thirds of patients reporting significant pain relief.
The major drawback to percutaneous RF denervation
procedures is that they should not be expected to
alleviate pain emanating from the ventral SI joint. In
the study by Schwarzer et al. (30), ventral capsular
pathology was shown to account for 69% of all CT
pathology in the 13 patients with a positive response
to diagnostic SI joint blocks. Complicating matters

1450

REVIEW ARTICLE COHEN

SACROILIAC JOINT PAIN

ANESTH ANALG
2005;101:1440 53

Table 5. Clinical Studies Evaluating Radiofrequency Procedures in the Treatment of Sacroiliac Joint Pain
Treatment

Primary outcome

Comments

Ferrante et al., 2001 (93)

Author, year

Retrospective study

Study type

33 pts, 50 joints

Number of patients

Multiple, 90C, 90
second lesions made
at 1 cm intervals
as high in the
postero-inferior joint
as possible.

Gevargez et al., 2002 (94)

Prospective
observational
study

38 patients, including
13 who underwent
bilateral treatment

Three 90C, 90 second

lesions in the
posterior
interosseous SI
ligaments and 1
lesion of the
L5 dorsal ramus.

Did not specify % pain

relief required
during diagnostic SI
joint injections for
inclusion. Only
postero-inferior joint
denervated.
Did not specify % pain
relief required
during diagnostic SI
joint injections for
inclusion.

Cohen and Abdi, 2003 (90)

Retrospective study

18 patients

80C, 90 second
lesions of the L4
and L5 dorsal rami
and S13 lateral
branches.

Yin et al., 2003 (92)

Retrospective study

14 patients, including
4 who underwent
previous spine
surgery

Buijs et al., 2004 (91)

Prospective
observational
study

38 patients, 43 joints

80C, 60 second
lesions of the
L5 dorsal ramus
sensory branch and
S1S3 dorsal rami
lateral branches
depending on
stimulation results.
All pts had L5 and
S1 branches
lesioned. 11 pts had
a lateral branch at
S2 and 6 at S3 that
were lesioned.
80C 60 second lesions
of the S13 dorsal
rami in all pts and
L4L5 dorsal rami in
about half the pts.

36.4% of pts obtained

50% pain relief 6months postprocedure. Average
duration of pain
relief was 12.0 1.2
months.
Three months after
treatment, 34.2%
were pain-free, 31.6%
reported a substantial
decrease in pain,
18.4% obtained a
slight decrease in
pain and 7.9%
reported no pain
reduction.
13 of 18 pts with SI
joint pain obtained
50% pain relief with
L4 and L5 dorsal
rami and S13 lateral
branch blocks, with
2 deriving long-term
relief. 8 of 9 pts who
underwent RF
denervation obtained
50% pain relief 9months postprocedure.
64% of patients
obtained 50%
consistent pain relief
at 6 months, with
36% obtaining
complete relief. 5
patients reported
50% pain relief, and 2
reported no relief
whatsoever.

At 12-week follow-up,
34.9% of procedures
(26.3% of pts)
resulted in complete
pain relief and
another 32.6% (34.2%
of pts) reported
50% pain relief.

Inclusion criteria was

50% pain relief
with SI joint blocks.
In 6 pts, empirical
lesions were made
at the S3 lateral
branch because of
failure to obtain
concordant
stimulation.

Inclusion criteria was

70% pain relief
after 2 separate SI
joint deep
interosseous
ligament injections.

Inclusion criteria
included 50%
pain relief with SI
joint blocks.
Outcomes of pts
receiving additional
L45 dorsal rami
denervation not
compared to pts
undergoing only S1
3 denervation.

pts patients; RF radiofrequency.

further are that the nerves lesioned during RF procedures innervate other pain-generating structures besides the SI joint, and the SI joint is likely innervated
by other nerves inaccessible for denervation (Table 5).

Surgical and Other Invasive Interventions

In 1999, Srejic et al. (95) reported 1216 mo of significant pain relief in 4 patients with SI joint pain who
received a series of 3 intraarticular injections with hyaluronic acid. Three of these patients had postsurgical SI
joint pain and one suffered from severe osteoarthritis of
the spine. The rationale for this treatment stems from
studies demonstrating long-term pain relief with hyaluronic acid injections in degenerative joint disease of the
knee (96). However, one meta-analysis found only scant
evidence for the use of viscosupplementation in knee
osteoarthritis, especially when lower molecular weight

hyaluronate formulations are used (97). In addition to

questionable efficacy, another factor that mitigates
against intraarticular hyaluronic acid injections for SI
pain is that degenerative joint disease accounts for only
a small percentage of cases.
Proliferative therapy (a.k.a. prolotherapy) has been
advocated as a treatment for nonspecific LBP and SI joint
pain (98). The rationale behind the use of prolotherapy
is that the ligaments and other soft tissue structures are
of primary importance in the development of LBP. Thus,
the injection of a drug promoting fibroblast hyperplasia
should theoretically increase the strength and reduce
sensitization of these structures. In a double-blind study,
Ongley et al. (99) found that LBP patients who received
6 wk of proliferant therapy had lower pain scores and
disability indices at their 6-mo follow-up than control
patients who received saline injections. Despite these

ANESTH ANALG
2005;101:1440 53

findings, the lack of specific diagnoses, the numerous

other treatment differences between groups, and the targeting of pain generators outside the SI joints limit the
relevance of this study.
In patients with SI joint pain unresponsive to more
conservative measures, several investigators have advocated surgical stabilization. Unfortunately, all published
reports on SI joint fusion have been small case series or
retrospective studies. Whereas the primary indications
for SI joint fixation are either joint instability or fractures
(100,101), successful arthrodesis has also been reported
for degenerative joint disease (102). In some patients,
successful stabilization can be done percutaneously using CT guidance (103). Regardless of the underlying
etiology, based on the existing studies the long-term
success rate for SI joint fusion appears to be in the range
of 70% (100 102). Neuroaugmentation of the third sacral
nerve root has also been reported to provide adequate
pain relief in 2 patients with severe SI joint pain unresponsive to conventional therapy (104).

Conclusions
The SI joint is a real yet underappreciated pain generator in an estimated 15% to 25% of patients with
axial LBP. Whereas historical and physical examination findings have been previously advocated as useful tools in identifying patients with SI joint pain,
more recent studies have demonstrated they have limited diagnostic value. Presently, small-volume diagnostic blocks remain the most commonly used method
for diagnosing this disorder, although their validity
remains unproven. Owing to the complexity of the
joint, the mechanisms of SI pain are numerous and
ill-defined. When a pathological condition such as leg
length discrepancy or altered gait mechanics is
present, correcting the underlying defect is the safest
and most reliable treatment option. Intraarticular and
periarticular corticosteroid injections have been
shown in most, but not all, studies to provide good to
excellent pain relief lasting up to 10 mo in patients
with and without spondylarthropathy. One promising
area in the treatment of SI joint pain is RF denervation,
although the conclusions that can be drawn are limited by the heterogeneous methods used and the lack
of controlled studies.

References
1. Bernard TN, Cassidy JD. The sacroiliac syndrome. Pathophysiology, diagnosis and management. In: Frymoyer JW, ed. The
adult spine: principles and practice. New York: Raven, 1991;
210730.
2. Dijkstra PF, Vleeming A, Stoeckart R. Complex motion tomography of the sacroiliac joint: an anatomical and roentgenological study [in German]. Rofo 1989;150:635 42.
3. Ruch WJ. Atlas of common subluxations of the human spine
and pelvis. Boca Raton, FL: CRC Press, 1997.

REVIEW ARTICLE
COHEN
SACROILIAC JOINT PAIN

1451

4. Bowen V, Cassidy JD. Macroscopic and microscopic anatomy

of the sacroiliac joint from embryonic life until the eighth
decade. Spine 1981;6:620 8.
5. Mitchell FL Jr. The muscle energy manual, Vol. 1. East Lansing,
MI: MET Press, 1995.
6. Murata Y, Takahashi K, Yamagata M, et al. Sensory innervation of the sacroiliac joint in rats. Spine 2000;16:20159.
7. Grob KR, Neuhuber WL, Kissling RO. Innervation of the sacroiliac joint in humans [in German]. Z Rheumatol 1995;54:11722 .
8. Pitkin HC, Pheasant HC. Sacrarthrogenic telalgia I: a study of
referred pain. J Bone Joint Surg 1936;18:11133.
9. Solonen KA. The sacroiliac joint in the light of anatomical,
roentgenological and clinical studies. Acta Orthop Scand 1957;
27(suppl):127.
10. Ikeda R. Innervation of the sacroiliac joint: macroscopic and
histological studies. J Nippon Med Sch 1991;58:58796.
11. Fortin JD, Kissling RO, OConnor BL, Vilensky JA. Sacroiliac
joint innervation and pain. Am J Orthop 1999;28:68 90.
12. Dreyfuss P, Park K, Bogduk N. Do L5 dorsal ramus and S1 4
lateral branch blocks protect the sacroiliac joint from an experimental pain stimulus? A randomized, double-blinded controlled trial. Presented at the International Spinal Injection
Society 8th Annual Scientific Meeting, San Francisco, CA, September 8 10, 2000.
13. Sakamoto N, Yamashita T, Takebayashi T, et al. An electrophysiologic study of mechanoreceptors in the sacroiliac joint
and adjacent tissues. Spine 2001;26:E468 71.
14. Yamashita T, Cavanaugh JM, El-Bohy AA, et al. Mechanosensitive afferent units in the lumbar facet joint. J Bone Joint Surg
(Am) 1990;72:86570.
15. Yamashita T, Minaki Y, Oota, et al. Mechanosensitive afferent
units in the lumbar intervertebral disc and adjacent muscle.
Spine 1993;18:2252 6.
16. Minaki Y, Yamashita T, Ishii S. An electrophysiological study
on the mechanoreceptors in the lumbar spine and adjacent
tissues. Neurol Orthop 1996;20:2335.
17. Dreyfuss P, Dreyer SJ, Cole A, Mayo K. Sacroiliac joint pain.
J Am Acad Orthop Surg 2004;12:255 65.
18. Dreyfuss P, Cole AJ, Pauza K. Sacroiliac joint injection techniques. Phys Med Rehabil Clin North Am 1995;6:785 813.
19. Walker JM. The sacroiliac joint: a critical review. Phys Ther
1992;72:90316.
20. White AA, Panjabi MM. Clinical biomechanics of the spine.
2nd ed. Philadelphia: JB Lippincott, 1990.
21. Miller JA, Schultz AB, Andersson GB. Load-displacement behavior of the sacroiliac joints. J Orthop Res 1987;5:92101.
22. Vleeming A, van Wingerden JP, Dijkstra PF, et al. Mobility in
the sacroiliac joints in the elderly: a kinematic and radiological
study. Clin Biomech 1992;7:170 6.
23. Vleeming A, van Wingerden JP, Snijders CJ, et al. Load application to the sacrotuberous ligament: influence on sacroiliac
joint mechanics. Clin Biomech 1989;4:204 9.
24. Brunner C, Kissling R, Jacob HA. The effects of morphology
and histopathologic findings on the mobility of the sacroiliac
joint. Spine 1991;16:11117.
25. Egund N, Olsson TH, Schmid H, Selvik G. Movements in the
sacroiliac joints demonstrated with roentgen stereophotogrammetric analysis. Acta Radiol Diagn 1978;19:833 45.
26. Jacob H, Kissling R. The mobility of the sacroiliac joints in
healthy volunteers between 20 and 50 years of age. Clin Biomech 1995;10:352 61.
27. Sturesson B, Selvik G, Uden A. Movements of the sacroiliac
joints: a roentgen stereophotogrammetric analysis. Spine 1989;
14:1625.
28. Harrison DE, Harrison DD, Troyanovich SJ. The sacroiliac
joint: a review of anatomy and biomechanics with clinical
implications. J Manipulative Physiol Ther 1997;20:60717.
29. Bernard TN, Kirkaldy-Willis WH. Recognizing specific characteristics of nonspecific low back pain. Clin Orthop 1987;217:
266 80.
30. Schwarzer AC, Aprill CN, Bogduk N. The sacroiliac joint in
chronic low back pain. Spine 1995;20:317.

1452

REVIEW ARTICLE COHEN

SACROILIAC JOINT PAIN

31. Maigne JY, Aivaliklis A, Pfefer F. Results of sacroiliac joint

double block and value of sacroiliac pain provocation tests in
54 patients with low back pain. Spine 1996;21:1889 92.
32. Bogduk N. International Spinal Injection Society guidelines for
the performance of spinal injection procedures. Part I: zygapophysial joint blocks. Clin J Pain 1997;13:285302.
33. Maugars Y, Mathis C, Berthelot JM, Charlier C, Prost A. Assessment of the efficacy of sacroiliac corticosteroid injections in
spondyloarthropathies: a double-blind study. Br J Rheumatol
1996;35:76770.
34. Braun J, Bollow M, Seyrekbasan F, et al. Computed tomography guided corticosteroid injection of the sacroiliac joint in
patients with spondyloarthropathy with sacroiliitis: clinical
outcome and follow-up by dynamic magnetic resonance imaging. J Rheumatol 1996;23:659 64.
35. Luukkainen R, Nissila M, Asikainen E, et al. Periarticular
corticosteroid treatment of the sacroiliac joint in patients with
seronegative spondyloarthropathy. Clin Exp Rheumatol 1999;
17:88 90.
36. Luukkainen R, Wennerstrand PV, Kautiainen HH, et al. Efficacy of periarticular corticosteroid treatment of the sacroiliac
joint in non-spondyloarthropathic patients with chronic low
back pain in the region of the sacroiliac joint. Clin Exp Rheumatol 2002;20:52 4.
37. Schuit D, McPoil TG, Mulesa P. Incidence of sacroiliac joint
malalignment in leg length discrepancies. J Am Podiatr Med
Assoc 1989;79:380 3.
38. Herzog W, Conway PJ. Gait analysis of sacroiliac joint patients.
J Manipulative Physiol Ther 1994;17:124 7.
39. Marymont JV, Lynch MA, Henning CE. Exercise-related stress
reaction of the sacroiliac joint: an unusual cause of low back
pain in athletes. Am J Sports Med 1986;14:320 3.
40. Schoenberger M, Hellmich K. Sacroiliac dislocation and scoliosis. Hippokrates 1964;35:476 9.
41. Katz V, Schofferman J, Reynolds J. The sacroiliac joint: a potential cause of pain after lumbar fusion to the sacrum. J Spinal
Disord Tech 2003;16:96 9.
42. Onsel C, Collier BD, Meting K, et al. Increased sacroiliac joint
uptake after lumbar fusion and/or laminectomy. Clin Nucl
Med 1992;17:2837.
43. Frymoyer JW, Howe J, Kuhlmann D. The long-term effects of
spinal fusion on the sacroiliac joint and ilium. Clin Orthop
1978;134:196 201.
44. Ebraheim NA, Elgafy H, Semaan HB. Computed tomographic
findings in patients with persistent sacroiliac pain after posterior iliac graft harvesting. Spine 2000;25:204751.
45. Frymoyer JW, Hanley E, Howe J, et al. Disc excision and spine
fusion in the management of lumbar disc disease: a minimum
ten-year follow-up. Spine 1978;3:1 6.
46. Albert H, Godskesen M, Westergaard J. Prognosis in four
syndromes of pregnancy-related pelvic pain. Acta Obstet Gynecol Scand 2001;80:50510.
47. Berg G, Hammar M, Moller-Nielsen J, et al. Low back pain
during pregnancy. Obstet Gynecol 1988;71:715.
48. Daly JM, Frame PS, Rapoza PA. Sacroiliac subluxation: a common treatable cause of low-back pain in pregnancy. Fam Pract
Res J 1991;11:149 59.
49. Bollow M, Braun J, Hamm B. Sacroiliitis: the key symptom of
spondylarthropathies. 1. The clinical aspects [in German]. Rofo
1997;166:95100.
50. Cush JJ, Lipsky PE. Reiters syndrome and reactive arthritis. In:
Koopman WJ, ed. Arthritis and allied conditions: a textbook of
rheumatology, Vol. 1, 14th ed. Philadelphia: Lippincott Williams & Wilkins, 2001:1324 44.
51. Dunn EJ, Bryan DM, Nugent JT, Robinson RA. Pyogenic infections of the sacro-iliac joint. Clin Orthop 1976;118:1137.
52. Humphrey SM, Inman RD. Metastatic adenocarcinoma mimicking unilateral sacroiliitis. J Rheumatol 1995;22:970 2.
53. Fortin JD. Sacroiliac joint dysfunction: a new perspective. J
Back Musculoskeletal Rehabil 1993;3:31 43.

ANESTH ANALG
2005;101:1440 53

54. Baquie P, Brukner P. Injuries presenting to an Australian

sports medicine center: a 12-month study. Clin J Sport Med
1997;7:28 31.
55. LeBlanc KE. Sacroiliac sprain: an overlooked cause of back
pain. Am Fam Physician 1992;46:1459 63.
56. Chou LH, Slipman CW, Bhagia SM, et al. Inciting events initiating injection-proven sacroiliac joint syndrome. Pain Med
2004;5:26 32.
57. Cohen SP, Rowlingson J, Abdi S. Low back pain. In: Warfield
CA, Bajwa ZA, eds. Principles and practice of pain medicine.
2nd ed. New York: McGraw-Hill, 2004: 273 84.
58. Slipman CW, Sterenfeld EB, Chou LH, et al. The predictive
value of provocative sacroiliac joint stress maneuvers in the
diagnosis of sacroiliac joint syndrome. Arch Phys Med Rehabil
1998;79:288 92.
59. Dreyfuss P, Michaelsen M, Pauza K, et al. The value of medical
history and physical examination in diagnosing sacroiliac joint
pain. Spine 1996;21:2594 2602.
60. Dreyfuss P, Dreyer S, Griffin J, et al. Positive sacroiliac screening tests in asymptomatic adults. Spine 1994;19:1138 43.
61. Laslett M, Williams M. The reliability of selected pain provocation tests for sacroiliac joint pathology. Spine 1994;19:12439.
62. McCombe PF, Fairbank JC, Cockersole BC, Pynsent PB. Reproducibility of physical signs in low-back pain. Spine 1989;14:
908 18.
63. Kokmeyer DJ, Van der Wurff P, Aufdemkampe G, Fickenscher
TC. The reliability of multitest regimens with sacroiliac pain
provocation tests. J Manipulative Physiol Ther 2002;25:42 8.
64. Riddle DL, Freburger JK. Evaluation of the presence of sacroiliac joint region dysfunction using a combination of tests: a
multicenter intertester reliability study. Phys Ther 2002;82:
772 81.
65. Potter NA, Rothstein JM. Intertester reliability for selected
clinical tests of the sacroiliac joint. Phys Ther 1985;65:16715.
66. Freburger JK, Riddle DL. Measurement of sacroiliac joint
dysfunction: a multicenter intertester reliability study. Phy
Ther 1999;79:1134 41.
67. Carmichael JP. Inter- and intra-tester reliability of palpation for
sacroiliac joint dysfunction. J Manipulative Physiol Ther 1987;
10:164 71.
68. Meijne W, van Neerbos K, Aufdemkampe G, Van der Wurff P.
Intraexaminer and interexaminer reliability of the Gillet test. J
Manipulative Physiol Ther 1999;22:4 9.
69. Maigne JY, Boulahdour H, Chatellier G. Value of quantitative
radionuclide bone scanning in the diagnosis of sacroiliac joint
syndrome in 32 patients with low back pain. Eur Spine J
1998;7:328 31.
70. Slipman CW, Sterenfeld EB, Chou LH, et al. The value of
radionuclide imaging in the diagnosis of sacroiliac joint syndrome. Spine 1996;21:2251 4.
71. Elgafy H, Semaan HB, Ebraheim NA, Coombs RJ. Computed
tomography findings in patients with sacroiliac pain. Clin
Orthop 2001;382:112 8.
72. Fortin JD, Dwyer AP, West S, Pier J. Sacroiliac joint: pain
referral maps upon applying a new injection/arthrography
technique. Part I: asymptomatic volunteers. Spine 1994;19:
1475 82.
73. Fortin JD, Aprill CN, Ponthieux B, Pier J. Sacroiliac joint: pain
referral maps upon applying a new injection/arthrography
technique. Part II: clinical evaluation. Spine 1994;19:14839.
74. Slipman CW, Jackson HB, Lipetz JS, et al. Sacroiliac joint pain
referral zones. Arch Phys Med Rehabil 2000;81:334 8.
75. Hogan QH, Abram SE. Neural blockade for diagnosis and
prognosis: a review. Anesthesiology 1997;86:216 41.
76. North RB, Kidd DH, Zahurak M, Piantadosi S. Specificity of
diagnostic nerve blocks: a prospective, randomized study of
sciatica due to lumbosacral disease. Pain 1996;65:77 85.
77. Dussault RG, Kaplan PA, Anderson MW. Fluoroscopy-guided
sacroiliac joint injections. Radiology 2000;214:2737.

ANESTH ANALG
2005;101:1440 53

78. Bollow M, Braun J, Taupitz M, et al. CT-guided intraarticular

corticosteroid injection into the sacroiliac joints in patients with
spondyloarthropathy: indication and follow-up with contrastenhanced MRI. J Comput Assist Tomogr 1996;20:51221.
79. Rosenberg JM, Quint DJ, de Rosayro AM. Computerized tomographic localization of clinically-guided sacroiliac joint injections. Clin J Pain 2000;16:18 21.
80. Schwarzer AC, Aprill CN, Derby R, et al. The false-positive
rate of uncontrolled diagnostic blocks of the lumbar zygapophysial joints. Pain 1994;58:195200.
81. Bogduk N, Holmes S. Controlled zygapophysial joint blocks:
the travesty of cost-effectiveness. Pain Med 2000;1:24 34.
82. Polatin PB, Kinney RK, Gatchel RJ, et al. Psychiatric illness and
chronic low-back pain. The mind and spine: which goes first?
Spine 1993;18:66 71.
83. Fayad F, Lefevre-Colau MM, Poiraudeau S, et al. Chronicity,
recurrence, and return to work in low back pain: common
prognostic factors [in French]. Ann Readapt Med Phys 2004;
47:179 89.
84. Seres JL. Evaluating the complex chronic pain patient. Neurosurg Clin N Am 2003;14:339 52.
85. Cibulka MT, Delitto A. A comparison of two different methods
to treat hip pain in runners. J Orthop Sports Phys Ther 1993;
17:172 6.
86. Osterbauer PJ, De Boer KF, Widmaier R, et al. Treatment and
biomechanical assessment of patients with chronic sacroiliac
joint syndrome. J Manipulative Physiol Ther 1993;16:8290.
87. Vleeming A, Buyruk HM, Stoeckart R, et al. An integrated
therapy for peripartum pelvic instability: a study of the biomechanical effects of pelvic belts. Am J Obstet Gynecol 1992;
166:12437.
88. Damen L, Spoor CW, Snijders CJ, Stam HJ. Does a pelvic belt
influence sacroiliac joint laxity? Clin Biomech 2002;17:495 8.
89. Mooney V, Pozos R, Vleeming A, et al. Exercise treatment for
sacroiliac pain. Orthopedics 2001;24:29 32.
90. Cohen SP, Abdi S. Lateral branch blocks as a treatment for
sacroiliac joint pain: a pilot study. Reg Anesth Pain Med 2003;
28:1139.
91. Buijs EJ, Kamphuis ET, Groen GJ. Radiofrequency treatment of
sacroiliac joint-related pain aimed at the first three sacral dorsal rami: a minimal approach. Pain Clinic 2004;16:139 46.
92. Yin W, Willard F, Carreiro J, Dreyfuss P. Sensory stimulationguided sacroiliac joint radiofrequency neurotomy: technique
based on neuroanatomy of the dorsal sacral plexus. Spine
2003;28:2419 25.
93. Ferrante FM, King LF, Roche EA, et al. Radiofrequency sacroiliac joint denervation for sacroiliac syndrome. Reg Anesth
Pain Med 2001;26:137 42.
94. Gevargez A, Groenemeyer D, Schirp S, Braun M. CT-guided
percutaneous radiofrequency denervation of the sacroiliac
joint. Eur Radiol 2002;12:1360 5.
95. Srejic U, Calvillo O, Kabakibou K. Viscosupplementation: a
new concept in the treatment of sacroiliac joint syndrome: a
preliminary report of four cases. Reg Anesth Pain Med 1999;
24:84 8.
96. Petrella RJ, DiSilvestro MD, Hildebrand C. Effects of hyaluronate sodium on pain and physical functioning in osteoarthritis
of the knee: a randomized, double-blind, placebo-controlled
trial. Arch Int Med 2002;162:292 8.

REVIEW ARTICLE
COHEN
SACROILIAC JOINT PAIN

1453

97. Lo GH, LaValley M, McAlindon T, Felson DT. Intra-articular

hyaluronic acid in treatment of knee osteoarthritis: a metaanalysis. JAMA 2003;290:311521.
98. Hauser RA. Punishing the pain: treating chronic pain with
prolotherapy. Rehab Manag 1999;12:26 30.
99. Ongley MJ, Klein RG, Dorman TA, et al. A new approach to the
treatment of chronic low back pain. Lancet 1987;2:143 6.
100. Simpson LA, Waddell JP, Leighton RK, et al. Anterior approach and stabilization of the disrupted sacroiliac joint.
J Trauma 1987;27:13329.
101. Dabezies EJ, Millet CW, Murphy CP, et al. Stabilization of
sacroiliac joint disruption with threaded compression rods.
Clin Orthop 1989;246:16571.
102. Waisbrod H, Krainick JU, Gerbershagen HU. Sacroiliac joint
arthrodesis for chronic lower back pain. Arch Orthop Trauma
Surg 1987;106:238 40.
103. Arand M, Kinzl L, Gebhard F. Computer-guidance in percutaneous screw stabilization of the iliosacral joint. Clin Orthop
2004;422:2017.
104. Calvillo O, Esses SI, Ponder C, et al. Neuroaugmentation in the
management of sacroiliac joint pain: report of two cases. Spine
1998;23:1069 72.
105. Broadhurst NA, Bond MJ. Pain provocation tests for the assessment of sacroiliac joint dysfunction. J Spinal Disord 1998;
11:3415.
106. Young S, Aprill C, Laslett M. Correlation of clinical characteristics with three sources of chronic low back pain. Spine J
2003;3:460 5.
107. Maugars Y, Mathis C, Vilon P, Prost A. Corticosteroid injection
of the sacroiliac joint in patients with seronegative spondylarthropathy. Arthritis Rheum 1992;35:564 8.
108. Gunaydin I, Pereira PL, Daikeler T, et al. Magnetic resonance
imaging guided corticosteroid injection of the sacroiliac joints
in patients with therapy resistant spondyloarthropathy: a pilot
study. J Rheumatol 2000;27:424 8.
109. Hanly JG, Mitchell M, MacMillan L, et al. Efficacy of sacroiliac
corticosteroid injections in patients with inflammatory
spondyloarthropathy: results of a 6 month controlled study.
J Rheumatol 2000;27:719 22.
110. Pereira PL, Gunaydin I, Duda SH, et al. Corticosteroid injections of the sacroiliac joint during magnetic resonance: preliminary results [in French]. J Radiol 2000;81:223 6.
111. Pereira PL, Gunaydin I, Trubenbach J, et al. Interventional MR
imaging for injection of sacroiliac joints in patients with sacroiliitis. AJR Am J Roentgenol 2000;175:265 6.
112. Ojala R, Klemola R, Karppinen J, et al. Sacro-iliac joint arthrography in low back pain: feasibility of MRI guidance. Eur J
Radiol 2001;40:236 9.
113. Karabacakoglu A, Karakose S, Ozerbil OM, Odev K.
Fluoroscopy-guided intraarticular corticosteroid injection into
the sacroiliac joints in patients with ankylosing spondylitis.
Acta Radiol 2002;43:4257.
114. Fischer T, Biedermann T, Hermann KG, et al. Sacroiliitis in
children with spondyloarthropathy: therapeutic effect of CTguided intra-articular corticosteroid injection [in German].
Rofo 2003;175:814 21.

ECONOMICS, EDUCATION,

AND

HEALTH SYSTEMS RESEARCH

SECTION EDITOR
RONALD D. MILLER

EDITORIAL

The Legend of the P Value

Zeev N. Kain,

MD, MBA

Center for the Advancement of Perioperative Health and Department of Anesthesiology & Pediatrics & Child Psychiatry,
Yale University School of Medicine, New Haven, Connecticut

lthough there is a growing body of literature

criticizing the use of mere statistical significance
as a measure of clinical impact, much of this
literature remains out of the purview of the discipline
of anesthesiology. Currently, the magical boundary of
P 0.05 is a major factor in determining whether a
manuscript will be accepted for publication or a research grant will be funded. Similarly, the Federal
Drug Administration does not currently consider the
magnitude of an advantage that a new drug shows
over placebo. As long as the difference is statistically
significant, a drug can be advertised in the United
States as effective whether clinical trials proved it to
be 10% or 200% more effective than placebo. We submit that if a treatment is to be useful to our patients, it
is not enough for treatment effects to be statistically
significant; they also need to be large enough to be
clinically meaningful.
Unfortunately, physicians often misinterpret statistically significant results as showing clinical significance as well. One should realize, however, that with
a large sample it is quite possible to have a statistically
significant result between groups despite a minimal
impact of treatment (i.e., small effect size). Also, study
outcomes with lower P values are typically misinterpreted by physicians as having stronger effects than
those with higher P values. That is, most clinicians
agree that a result with a P 0.002 has a much greater
treatment effect than a result of P 0.045. Although
this is true if the sample size is the same in both
studies, it is not true if the sample size is larger in the
study with the smaller P value. This is of particular
concern when one realizes that most pharmaceutically
funded studies have very large sample sizes and effect
sizes are typically not reported in these types of studies. In the following editorial I highlight some of issues
Supported, in part, by National Institutes of Health grants
NICHD, R01HD37007 02.
Accepted for publication June 16, 2005.
Address correspondence and reprint requests to Zeev N. Kain,
MD MBA, Department of Anesthesiology, Yale University School of
Med, 333 Cedar Street, New Haven, CT 06510. Address e-mail to
zeev.kain@yale.edu.
DOI: 10.1213/01.ANE.0000181331.59738.66

1454

Anesth Analg 2005;101:14546

related to this complex problem. Please note that a

detailed discussion of the underlying statistics involved in this topic is beyond the scope of this
editorial.
When examining the report of a clinical trial investigating a new treatment, clinicians should be interested in answering the following three basic questions:
1. Could the findings of the clinical trial be solely a
result of a chance occurrence? (i.e., statistical
significance)
2. How large is the difference between the primary
end-points of the study groups? (i.e., impact of
treatment, effect size)
3. Is the difference of primary end-points between
groups meaningful to a patient? (i.e., clinical
significance)
It was Sir Ronald A. Fisher, an extraordinarily influential British statistician, who first suggested the
use of a boundary to accept or reject a null hypothesis,
and he arbitrarily set this boundary at P 0.05; where
P stands for probability related to chance (1,2). That
is, the level of statistical significance as defined by
Fisher in 1925 and as used today refers to the probability that the difference between two groups would
have occurred solely by chance (i.e., probability of 5
in 100 is reported as P 0.05). Fishers emphasis on
significance testing and the arbitrary boundary of P
0.05 has been widely criticized over the past 80 yr.
This criticism was based on the rationale that focusing
on the P value does not take into account the size and
clinical significance of the observed effect. That is, a
small effect in a study with large sample size has the
same P value as a large effect in a study with a small
sample size. Also, P value is commonly misinterpreted when there are multiple comparisons, in which
case a traditional level of statistical significance of P
0.05 is no longer valid. Fisher himself indicated some
25 yr after his initial publication that If P is between
0.1 and 0.9 there is certainly no reason to suspect the
hypothesis tested. If it is below 0.02 it is strongly
indicated that the hypothesis fails to account for the
whole of the facts. We shall not often be astray if we
draw a conventional line at 0.05. . . (3). Indeed, this
2005 by the International Anesthesia Research Society
0003-2999/05

ANESTH ANALG
2005;101:1454 6

issue has been addressed in multiple recent review

articles and editorials in the general medical and psychological literature (4 8).
In an attempt to address some of the limitations of
the P value, the use of the confidence intervals (CI) has
been advocated by some clinicians (9). One should
realize, however, that these two definitions of statistical significance are essentially reciprocal (10). That is,
getting a P 0.05 is the same as having a 95% CI that
does not overlap zero. CIs can also, however, be used
to estimate the size of difference between groups in
addition to merely indicating the existence or absence
of statistical significance (11). This later approach,
however, is not widely used in the medical and psychological literature, and today CIs are mostly used as
surrogates for the hypothesis test rather than considering the full range of likely effect size.
The group of statistics called effect sizes designate
indices that measure the magnitude of difference between groups, controlling for variation within the
groups; effect sizes can be thought of as a standardized difference. In other words, although a P value
denotes whether the difference between two groups in
a particular study is likely to occur solely by chance,
the effect size quantifies the amount of difference between the two groups. Quantification of effect size
does not rely on sample size but instead relies on the
strength of the intervention. There are a number of
different types of effect sizes and a description of these
various types and formulae is beyond the scope of this
editorial. We refer the interested reader to review
articles that describe the various types of effect sizes
and their calculation methodology (12,13). Effect sizes
of the d type are the most commonly used in the
medical literature, as they are primarily used to compare two treatment groups. D type effect size is defined as the magnitude of difference between two
means, divided by the sd [(Mean of control group
Mean of treatment group)/sd of the control group].
Thus, the d effect size is dependent on variation
within the control group and the differences between
the control and intervention groups. Values of the d
type effect sizes range from to , where zero
denotes no effect and values less than or more than
zero are treated as absolute values when interpreting
magnitude. Conventionally, d type effect sizes that are
near 0.20 are interpreted as small, effect sizes near 0.50
are considered medium, and effect sizes in the range
of 0.80 are considered large (14). However, interpretation of the magnitude of an effect size depends on
the type of data gathered and the discipline involved.
Effect sizes of another typethe risk potency type
include likelihood ratios such as odds ratio, risk ratio,
risk difference, and relative risk reduction. Clinicians
are probably more familiar with these less abstract
statistics and it may be helpful to realize that likelihood statistics are a type of effect size.

EDITORIAL

1455

Clinicians should be cautioned to not interpret magnitude of change (effect size) as an indication of clinical significance. The clinical significance of a treatment should be based on external standards provided
by patients and clinicians. That is, a small effect size
may still be clinically significant and, likewise, a large
effect size may not be clinically significant, depending
on what is being studied. Indeed, there is a growing
recognition that traditional methods used, such as
statistical significance tests and effect sizes, should be
supplemented with methods for determining clinically significant changes. Although there is little consensus about the criteria for these efficacy standards,
the most prominent definitions of clinically significant
change include: 1) treated patients make a statistically
reliable improvement in the change scores; 2) treated
patients are empirically indistinguishable from a normal population after treatment, or 3) changes of at
least one sd. The most frequently used method for
evaluating the reliability of change scores is the
Jacobson-Truax method in combination with clinical
cutoff points (15). Using this method, change is considered reliable, or unlikely to be the product of measurement error, if the reliable change index (RCI) is
more than 1.96. That is, when the individual has a
change score more than 1.96, one can reasonably assume that the individual has improved.
Unfortunately, most of the methods above are difficult to adopt in the perioperative arena, as comparison with a normal population is not an option in most
trials, and the RCI, which controls for statistical issues
involving the assessment tool, is a somewhat complicated and controversial technique. Thus, clinical significance in the perioperative arena may be best assessed by posing a particular question such as is a
change of 8.5% reduction in intraoperative bleed clinically significant? or how many sd does this change
represent? Obviously, both of these questions have a
subjective component in them and although it is traditionally agreed that at least a 1-sd change is generally needed for clinical significance, this boundary has
no scientific underpinning. The validity of a clinical
cutoff for these last two methods can be improved by
establishing external validity (e.g., patient perspective) for the decision. For example, Flor et al. (16) have
conducted a large meta-analysis that was aimed at
evaluating the effectiveness of multidisciplinary rehabilitation for chronic pain. The investigators found
that pain among the patients who received the intervention was indeed reduced by 25%. This reduction
was certainly statistically significant and had an effect
size of 0.7. Colvin et al. (17), however, reported earlier
that patients would consider only a 50% improvement
in their pain levels as a treatment success. Thus, in
this example, a reduction of 25% in pain scores may be
statistically, but not clinically, significant. Clearly this
is a developing area that warrants further discussion.

1456

EDITORIAL

In conclusion, we suggest that reporting of perioperative medical research should continue beyond reporting results consisting primarily of descriptive and
statistically significant or nonsignificant findings. The
interpretation of findings should occur in the context
of the magnitude of change that occurred and the
clinical significance of the findings.

References
1. Fisher RA. Statistical methods for research workers, 1st ed.
Edinburgh: Oliver and Boyd, 1925. Reprinted by Oxford University Press.
2. Fisher RA. Design of experiments. 1st ed. Edinburgh: Oliver and
Boyd, 1935. Reprinted by Oxford University Press.
3. Fisher RA. Statistical methods for research workers. London:
Oliver and Boyd, 1950:80.
4. Borenstein M. Hypothesis testing and effect size estimation in
clinical trials. Ann Allergy Asthma Immunol 1997;78:511.
5. Matthey S. P 0.05: but is it clinically significant? Practical
examples for clinicians. Behav Change 1998;15:140 6.
6. Cummings P, Rivara FP. Reporting statistical information in
medical journal articles. Arch Pediatr Adolesc Med 2003;157:
321 4.

ANESTH ANALG
2005;101:1454 6

7. Greenstein G. Clinical versus statistical significance as they

relate to the efficacy of periodontal therapy. J Am Dent Assoc
2003;134:1168 70.
8. Sterne JAC, Smith GD, Cox DR. Sifting the evidence: whats
wrong with significance tests? Another comment on the role of
statistical methods. BMJ 2001;322:226 31.
9. Simon R. Confidence intervals for reporting results of clinical
trials. Ann Intern Med 1986;105:429 35.
10. Feinstein AR. P-values and confidence intervals: two sides of
the same unsatisfactory coin. J Clin Epidemiol 1998;51:355 60.
11. Gardner MG, Altman DG. Confidence intervals rather than P
values: estimation rather than hypothesis testing. BMJ 1986;292:
746 50.
12. Kirk R. Practical significance: A concept whose time has come.
Educ Psychol Meas 1996;56:746 59.
13. Snyder P, Lawson S. Evaluating results using corrected and
uncorrected effect size estimates. J Exper Educ 1993;61:334 349.
14. Cohen J. Statistical power analysis for the behavioral sciences,
2nd ed. Mahwah, New Jersey: Lawrence Erlbaum, 1988.
15. Jacobson NS, Truax P. Clinical significance: A statistical approach to defining meaningful change in psychotherapy research. J Consult Clinic Psych 1991;59:129.
16. Flor H, Fydrich T, Turk DC. Efficacy of multidisciplinary pain
treatment centers: a meta-analytic review. Clin J Pain 1992;49:
22130.
17. Colvin DF, Bettinger R, Knapp R, et al. Characteristics of patients with chronic pain. South Med J 1980;73:1020 3.

Chemical Dependency Treatment Outcomes of Residents in

Anesthesiology: Results of a Survey
Gregory B. Collins, MD, Mark S. McAllister,
Timothy A. Gooden, MD

MD,

Mark Jensen,

DO,

and

Alcohol and Drug Recovery Center, Department of Psychiatry and Psychology, Cleveland Clinic Foundation/P48,
Cleveland, Ohio

Substance abuse is a potentially lethal occupational

hazard confronting anesthesiology residents. We
present the results of a survey sent to all United States
anesthesiology training programs regarding experience with and outcomes of chemically dependent residents from 1991 to 2001. The response rate was 66%.
Eighty percent reported experience with impaired residents and 19% reported at least one pretreatment fatality. Despite this familiarity, few programs required preemployment drug testing or used substance abuse
screening tools during interviews. The majority of impaired residents attempted reentry into anesthesiology
after treatment. Only 46% of these were successful in
completion of anesthesiology residency. Eventually,

ubstance abuse disorders among physicians are

a major health and societal concern. Up to 14%
of all physicians become chemically dependent
at some point in their careers (13). Anesthesiology
appears to be the specialty with the largest percentage
of impaired physicians (4,5). Despite constituting 4%
of the United States physician population, anesthesiologists represent up to 13% of physicians treated for
chemical dependency (1,4). Anesthesiology residents
endure substantial risk, as the largest morbidity and
mortality from substance abuse in physicians occurs
within the first 5 yr after medical school (6). Substance
abuse among anesthesiology residents is now well
recognized. Talbott et al. (4) reported that while anesthesiology residents represented only 4.6% of the US
resident physician population at the time of their
study, 33.7% of all residents presenting to the Medical

Accepted for publication May 12, 2005.

Address correspondence and reprint requests to Gregory B. Collins, MD, Section Head, Alcohol and Drug Recovery Center, Department of Psychiatry and Psychology, Cleveland Clinic
Foundation/P48, 9500 Euclid Avenue, Cleveland, OH 44195. Address e-mail to colling@ccf.org.
DOI: 10.1213/01.ANE.0000180837.78169.04
2005 by the International Anesthesia Research Society
0003-2999/05

40% of residents who underwent treatment and returned to medical training entered another specialty.
The mortality rate for the remaining anesthesiology residents was 9%. Long-term outcome was reported for
93% of all treated residents. Of these, 56% were successful in some specialty of medicine at the end of the survey period. We hypothesize that specialty change afforded substantial improvement in the overall success
rate and avoided significant mortality. Redirection of
rehabilitated residents into lower-risk specialties may
allow a larger number to achieve successful medical
careers.
(Anesth Analg 2005;101:145762)

Association of Georgias Impaired Physicians Program were residents training in anesthesia.

Only a few studies have examined treatment outcomes and the feasibility of reentry into anesthesia
training by impaired residents. In general, these studies have produced conflicting results. A survey of
residency program directors in 1989 found that the
rate of successful reentry into anesthesia residency, as
defined by completion of training without relapse after rehabilitation, was only 34% for parenteral opiateabusing residents compared to 70% for non-opiate
abusers (7). Twenty-six deaths were reported, with
over half of these occurring as a consequence of relapse. Twenty-five percent of the parenteral opiate
abusers who relapsed died as a result. The authors
concluded that reentry of the parenteral opiateabusing resident into training, with the expectation
that two thirds will relapse and one in four of these
will die, represents too large a risk for all involved.
They suggested that redirection into another specialty
may be the most prudent course of action.
Less discouraging outcomes have been observed by
others. A 1999 report found no significant difference in
the rate of sustained recovery, defined as longer than
2 yr of abstinence, between impaired anesthesiologists
(41% of whom were residents) and randomly selected
Anesth Analg 2005;101:145762

1457

1458

ECONOMICS, EDUCATION, AND HEALTH SYSTEMS RESEARCH

TREATMENT OUTCOMES OF IMPAIRED RESIDENTS

impaired physician controls (81.3% and 86.1%, respectively) over a 12-yr period (8). Similar relapse rates
were noted as well (40.6% versus 44.6%, respectively).
Rates of sustained recovery and relapse for the subset
of anesthesiology residents did not differ significantly
from the parent or control groups (88.9% and 38.5%,
respectively). The authors noted that anesthesiologists
were encouraged to change specialty more frequently
than controls. In an earlier report, similar rates of
sustained abstinence were found among anesthesiologists and physicians of other specialties (69% and
73%, respectively) referred to the California Physicians Diversion Program (9). More than half of the
successfully treated anesthesiologists were parenteral
opiate abusers. Although only six residents were studied, four eventually did well and the remaining two
were still enrolled at study completion. The authors
did not consider a single relapse as a treatment failure,
noting that it often served as a stepping stone towards
long-term recovery. Both reports concluded that recovery rates for reentering anesthesiology professionals are similar to their medical counterparts.
These reports suggest improved outcomes based on
the trend towards highly structured monitoring and
long-term follow-up of recovering physicians. Outcomes for anesthesiology residents were not the primary focus of these studies. Therefore, we conducted
a nationwide survey of anesthesiology residency programs to examine current treatment outcomes, feasibility of successful reentry, and long-term career stability of these individuals.

Methods
An anonymous 20-question survey (Appendix) was
developed to obtain data regarding the extent of substance abuse in anesthesiology residents, treatment
outcome, and the existence of specific substance abuse
policies from 1991 to 2001. There were 176 anesthesiology residency training programs identified using
the American Medical Association (n 160) and
American Osteopathic Association (n 16) graduate
medical education registries as of January 2001. In
August 2001, a copy of the survey and a cover letter,
along with a self-addressed stamped envelope, were
sent to each residency program director. In September
2001, a second letter was sent to increase compliance.
All returned surveys were reviewed and scored by
hand. Data were compiled and analyzed with Origin
6.1 (OriginLab Corporation, Northampton, MA). Continuous variables were described as mean sd, and
categorical variables were described with frequencies
and percentages.

COLLINS ET AL.

ANESTH ANALG
2005;101:145762

Results
Of the 176 residency programs initially identified, 7 no
longer existed. Responses were received from 111 of
the 169 active programs yielding a response rate of
66%. The duration of program directorship ranged
from 1 yr to 30 yr, with a mean of 7.1 6.2 yr.
One-hundred-ten responding programs provided current resident roster data at the time of survey completion, yielding an active population of 3569 active
residents.
The first part of the survey elicited responses regarding the use of proactive substance abuse screening in the resident selection process as well as preemployment urine toxicology testing. Among the 111
responding programs, 18 (16%) routinely performed
substance abuse screening in the selection process and
17 (15%) required pre-employment urine toxicology
testing. Once having identified a resident as potentially impaired, 92% of responding programs initiate
evaluation and treatment. Three fourths of the responding programs have some form of on-site chemical dependency treatment and monitoring.
The second portion of the survey investigated training program experience with chemically dependent
residents. At least one fatality before intervention,
manifest as either drug overdose or suicide, was reported by 21 (19%) of the program directors. Eighty
percent of the responding programs identified at least
one resident as chemically dependent within the period from 1991 to 2001. A total of 230 cases of impairment were reported from the 111 responding programs, giving an average of 2.1 1.8 residents per
program over the 10-yr reporting period. Among the
230 residents identified, 199 received treatment within
the reporting interval and 31 were enrolled in an
active treatment program at the time of completion of
the survey. Thus, the prevalence of known active
chemical dependency in anesthesiology residents
among the reporting programs was 0.87% (31/3569).
The final portion of the survey ascertained the outcome of rehabilitation of impaired residents, as evidenced by reentry into medicine and specialty pursuit,
level of career attainment, and long-term well-being.
A flow chart demonstrating the ultimate outcome of
all treated residents is shown in Figure 1. Among the
199 residents receiving treatment, 32 (16%) left medicine entirely, whereas 167 (84%) continued to pursue
postgraduate training after rehabilitation. Of those remaining in medicine, 153 (92%) initially returned to
anesthesia whereas 14 (8%) pursued a different specialty. The distribution of initial career pursuit among
all individuals after rehabilitation is depicted in Figure
2. Of those returning to residency in anesthesia, 127
(83%) were allowed to return to their original residency program and the remaining 26 (17%) residents

ANESTH ANALG
2005;101:145762

ECONOMICS, EDUCATION, AND HEALTH SYSTEMS RESEARCH

COLLINS ET AL.
TREATMENT OUTCOMES OF IMPAIRED RESIDENTS

1459

Figure 1. Ultimate outcome of 199 impaired anesthesiology residents after chemical dependency treatment.

Figure 3. Distribution of career pursuit in the 199 chemically dependent anesthesiology residents at the end of the survey period.
Less than half of all treated residents were successful in completing
anesthesiology training and there was significant associated mortality. Those who chose a different specialty immediately after treatment (early group) were followed by a significant proportion that
left anesthesia after an initial attempt to return (late group).

Discussion

Figure 2. Distribution of career pursuit in the 199 chemically dependent anesthesiology residents immediately after rehabilitation.
The majority of residents who remained in medicine after treatment
attempted to return to anesthesiology training.

transferred to a different program. Eventually, however, 53 of the cohort of 153 reentering anesthesiology
residents also pursued a different specialty. Therefore,
67 of the 167 (40%) residents who returned to medicine after treatment ultimately trained in a different
specialty. Of the 199 chemically dependent residents
who completed treatment, at the time of survey completion only 91 (46%) successfully reentered and completed training in anesthesia. Six of the residency programs reported 9 relapse-related fatalities. The
distribution of ultimate career pursuit among all
treated individuals is shown in Figure 3.
Finally, the surveyed residency programs were
asked to identify the number of successfully treated
residents who were in stable recovery and actively
practicing in any medical specialty at the time of survey completion. Responses were received from 101
(91%) of the 111 programs completing the survey covering 185 (93%) of the original cohort of 199 treated
residents. Among this group, 104 (56%) were chemically abstinent and professionally stable in any medical specialty after completion of residency training.

We performed this study to determine if improvement

has been made in the treatment outcomes of residents
in anesthesiology compared with earlier studies (5,7).
Previous surveys of anesthesiology residency programs have documented that most programs have
experience with chemically dependent residents. Our
finding that 80% of responding programs have had at
least one impaired trainee is the largest reported percentage. Published incidences of impairment among
anesthesiologists in the US over the past 3 decades
range between 1% and 2% (1,10,11). The results of our
survey demonstrated an average of just over 2 residents per reporting program being identified as chemically dependent over the 10-year period. Despite recommendations for substance abuse policies (12) and
narcotic control systems (13) in anesthesia departments across the country, chemical impairment remains a major occupational hazard.
Although the focus of many reports has been on the
increasing recognition of substance abuse among anesthesiology professionals, no recent reports have examined the long-term impact of rehabilitation, particularly concerning successful reentry of residents into
training. The most recent report on this topic included
only 13 residents (8). Our results indicate that 92%
(153/167) of those returning to medicine after treatment for chemical dependency attempted to resume
their training in anesthesia, with 83% (127/153) of
those returning to their original residency program.
Previous reports have demonstrated that 61% to 83%
of anesthesiology professionals similarly attempt to

1460

ECONOMICS, EDUCATION, AND HEALTH SYSTEMS RESEARCH

TREATMENT OUTCOMES OF IMPAIRED RESIDENTS

reenter anesthesia (5,79). We did not attempt to determine how many of these individuals were counseled regarding redirection into another specialty. It
has been our institutional experience that most residents do return to their original program, as it offers
the path of least resistance in terms of familiarity
and avoidance of the stigma a history of substance
abuse may bring when reapplying for training positions in a different specialty.
Only 59% (91/153) of those returning to residency
in anesthesia were ultimately successful in completing
training. This equates to a failure rate of approximately 40%. We were unable to determine from our
data how many individuals relapsed after resuming
anesthesia training, but six programs reported at least
one relapse-related death. Only 42% of chemically
dependent anesthesiology residents overall were successful in completing training in a previously mentioned report (7). Moreover, 14% of all reentrants in
this report died as a result of relapse. Equally discouraging was a similar survey of anesthesiology training
programs in Australia and New Zealand, which revealed that only one of 13 impaired trainees ultimately
completed residency, although follow-up data in this
study were limited (14)
We were able to gather long-term data on 93%
(185/199) of the study residents and we found that
56% (104/185) of these who underwent rehabilitation
and returned to medicine were doing well in the practice of any specialty at the end of the survey period.
We were unable to determine what proportion of
these 104 individuals were successful in anesthesiology. However, the loss of 16% (32/199) of treated
residents who leave medicine entirely after treatment
is cause for concern. Perhaps this reflects a flawed
selection process for students entering medical training. The additional loss of 67 of the 167 remaining
trainees who left anesthesia for other specialties either
immediately or subsequent to an initial attempt at
reentry likely reflects the difficulty in returning successfully to anesthesiology training. Finally, the
substance-related deaths of 9 of the remaining 100
anesthesiology residents after treatment is a grim reminder of the risks and mortal consequences of an
ill-advised return to the specialty. It is likely that such
mortality could be decreased or avoided with redirection to lower risk specialties.
Very few anesthesiology programs were found to
perform screening for substance abuse during the resident selection process (16%) or pre-employment toxicology testing (15%) of new house staff hires. This
was surprising considering both the increased overall
awareness of physician impairment and the higher
risk of development of chemical dependency attributed to the practice of anesthesia. Of particular concern are the observations of Talbott et al. (4) and

COLLINS ET AL.

ANESTH ANALG
2005;101:145762

Gallegos et al. (15) that anesthesiology residents undergoing treatment for dependency cited drug availability as a major determinant of their choice of career.
Both suggested that preventative measures against
addiction should be implemented earlier, i.e., during
the resident selection process. According to a survey
of academic anesthesiology programs, current efforts
to reduce controlled substance abuse, namely increased mandatory education and tighter narcotic accounting procedures, appear largely ineffective (10).
Detection of the impaired resident is frequently difficult. Rarely are these individuals caught red
handed. A number of observational signs have been
suggested as suspicious for substance use among anesthesia providers (1). Unfortunately, a conspiracy of
silence often exists despite observance of these workplace symptoms. Moreover, addicted individuals are
extremely unlikely to admit to their abuse.
Although the addition of focused substance abuse
screening and toxicology testing to the resident selection process does not constitute foolproof mechanisms
for prevention, in light of the potential devastating
consequences, additional means of early detection
with the purpose of cautioning those at risk are warranted. The Association of Program Directors of Internal Medicine suggests that pre-employment drug testing provides a clear deterrent to addicted individuals
from entering programs where such screening is mandated (16). Currently, no studies have examined the
impact of pre-employment drug testing on the incidence of resident substance abuse.
Selection of high-risk individuals into anesthesiology residency programs may have contributed to the
observed persistence in the rate of substance abuse. In
a longitudinal study of a single midwestern medical
school class, validated screening tools were used to
assess the degree of substance abuse among these
students (17). Approximately 20% of the students interviewed during their fourth year of medical school
met the criteria for substance abuse or dependence in
the 12 months before the interview. The development
of prevention strategies for the anesthesiology profession should consider the use of similar risk assessment
tools during the resident selection process. Indeed,
others have cited risk factors, including a family history of chemical dependency and substance abuse
(particularly marijuana) at an early age, as significant
predictors of future substance abuse in residents
(17,18). These factors, combined with the occupational
stress and the ready availability of potent narcotics,
may produce an extremely high-risk environment for
predisposed individuals. Although the results of such
screening instruments should not be used as absolute
indications for denial or acceptance into training, we
believe that broader use of these tools may prove

ANESTH ANALG
2005;101:145762

ECONOMICS, EDUCATION, AND HEALTH SYSTEMS RESEARCH

COLLINS ET AL.
TREATMENT OUTCOMES OF IMPAIRED RESIDENTS

useful in discouraging high-risk individuals and augmenting overall substance abuse prevention in this
population.
Finally, we suggest that redirection of the recovering anesthesiology resident into other specialties of
medicine by rehabilitation providers may be an important step in the management of these individuals.
As a provider approved by the state licensing board,
our institution receives many impaired physician referrals. It is generally our practice to recommend specialty change to anesthesia residents because of the
frequent recidivism rate. As our survey demonstrates,
however, most do choose to return to anesthesia.
Based on personal accounts of those we have treated,
including that from one of the authors, the temptation
encountered on returning to the candy store often
makes continuing in anesthesia a futile endeavor despite structured monitoring and follow-up. Those who
do change specialty based on recommendation often
demonstrate greater insight into their disease. Although we do not suggest that relapse necessarily
foretells long-term failure, it has been shown, in this
report and others, that in the setting of reentry into
anesthesia, relapse is associated with significant mortality. Relapse is most common in the early period of
recovery. Therefore, it would seem imprudent to place
the newly recovering resident back into a stressful
operating room environment where potent narcotics
are readily available. It is clear that prospective studies
are needed to assess novel prevention strategies as
well as the success of impaired resident reentry in
terms of continued abstinence, completion of residency training, and long-term career stability.
Some limitations to this study deserve mention.
First, this retrospective study relies completely on the
recall of 10 years of data, a period longer than the
average tenure of the respondents. To facilitate this
and the overall compliance, our survey instrument
was designed to be brief and general. Given an average of just over 2 impaired residents per reporting
program and the often notable nature of these cases, it
is likely that if surveys were returned, they were indeed accurate. On the other hand, many details of the
individual, the rehabilitation, and the residency that
may have influenced the outcome were not solicited.
Second, to what extent the data reflect the true magnitude of the problem and the true population at risk
is a point of concern. As mentioned, detection of these
individuals is often difficult and delayed, with a significant proportion undiscovered until after training.
In this regard, the true scope of the problem may be
understated. Conversely, over-representation of the
problem may have occurred because a significant proportion of the true population at risk was not represented based on our response rate. One reason for lack
of response may be the absence of the perceived target
data on behalf of the nonrespondents.

1461

Despite nearly 30 years of increased awareness, education and monitoring of controlled substances, as
well as the treatment of affected individuals, our survey suggests that little progress has been made in
improving the prevalence, mortality, or feasibility of
successful reentry of impaired anesthesiology residents. Increased use of substance abuse screening of
prospective house staff coupled with pre-employment
drug testing may allow for identification of high-risk
individuals and, where appropriate, direction into
other fields of medicine. The effectiveness of such
measures should be the topic of future research.
Present strategies of monitoring and aftercare, intensive counseling and mandatory self-help meetings
have not lead to improved outcomes. Our data would
support a more proactive approach.

Appendix
Text of the survey that was sent to the directors of all
United States anesthesiology residency programs in
August and September 2001.
Please answer the following questions in regards to
the past TEN years (19912001):
1. How many years have you served as residency
program director?
2. What is the total number of residents in your
residency program?
3. On average, how many residents graduate per
year?
4. Are standardized tools or other forms of assessment for chemical dependency in individuals utilized
during your residency candidate interviewing
process?
5. Does your program require urine drug testing of
residency candidates prior to acceptance?
Prior to employment?
6. In the past 10 years, how many residents in your
program have been identified as chemically
dependent?
(Including alcohol, street drugs, anesthesia and
prescription medications)
7. How many of those residents identified as chemically dependent received evaluation and treatment?
8. How many residents are currently undergoing
chemical dependency treatment or monitoring?
9. How many substance-related fatalities or suicides
have there been among residents in your program
prior to intervention/treatment for substance abuse?
Following treatment?
10. Do residents in your program unconditionally
receive treatment if found to be substance abusers?
11. Do residents receive continued benefits while
undergoing treatment for chemical dependency in
terms of:
Continued salary? Y/N Time off for treatment?
Y/N Reentry assurance? Y/N

1462

ECONOMICS, EDUCATION, AND HEALTH SYSTEMS RESEARCH

TREATMENT OUTCOMES OF IMPAIRED RESIDENTS

12. Does your institution have a chemical dependency treatment and monitoring program for residents in place?
Is it on site?
13. Of those residents receiving treatment, how
many were allowed to return to your training
program?
14. Of those residents receiving treatment, how
many successfully completed your training program?
15. Of those residents receiving treatment, how
many returned to a different anesthesiology residency
program?
16. Of those residents receiving treatment, how
many left medicine completely?
17. Of those residents receiving treatment, how
many immediately left anesthesiology for another specialty of medicine?
18. Of those residents receiving treatment, how
many initially returned to anesthesiology but subsequently left for another specialty?
19. Of those former residents receiving treatment,
how many are currently doing well in practice in any
specialty of medicine?
20. How many residents were arrested? Convicted?
Jailed?

References
1. Farley WJ. Addiction and the anaesthesia resident. Can J Anaesth 1992;39:R113.
2. Farley WJ, Talbott GD. Anesthesiology and addiction. Anesth
Analg 1983;62:465 6.
3. Herrington RE. The impaired physician-recognition, diagnosis,
and treatment. Wisc Med J 1979;78:213.

COLLINS ET AL.

ANESTH ANALG
2005;101:145762

4. Talbott GD, Gallegos KV, Wilson PO, Porter TL. The Medical
Association of Georgias Impaired Physicians Program. Review
of the first 1000 physicians: analysis of specialty. JAMA 1987;
257:292730.
5. Ward CF, Ward GC, Saidman LJ. Drug abuse in anesthesia
training programs. A survey: 1970 through 1980. JAMA 1983;
250:9225.
6. Alexander BH, Checkoway H, Nagahama SI, Domino KB.
Cause-specific mortality risks of anesthesiologists. Anesthesiology 2000;93:92230.
7. Menk EJ, Baumgarten RK, Kingsley CP, et al. Success of reentry
into anesthesiology training programs by residents with a history of substance abuse. JAMA 1990;263:3060 2.
8. Paris RT, Canavan DI. Physician substance abuse impairment:
anesthesiologist vs. other specialties. J Addict Dis 1999;18:17.
9. Pelton C, Ikeda RM. The California Physicians Diversion Programs experience with recovering anesthesiologists. J Psychoactive Drugs 1991;23:42731.
10. Gravenstein JS, Kory WP, Marks RG. Drug abuse by anesthesia
personnel. Anesth Analg 1983;62:46772.
11. Booth JV, Grossman D, Moore J, et al. Substance abuse among
physicians: a survey of academic anesthesiology programs.
Anesth Analg 2002;95:1024 30.
12. Lecky JH, Aukburg SJ, Conahan TJ 3rd, et al. A departmental
policy addressing chemical substance abuse. Anesthesiology
1986;65:414 7.
13. Adler GR, Potts FE 3rd, Kirby RR, et al. Narcotics control in
anesthesia training. JAMA 1985;253:3133 6.
14. Weeks AM, Buckland MR, Morgan EB, Myles PS. Chemical
dependence in anaesthetic registrars in Australia and New Zealand. Anaesth Intensive Care 1993;21:1515.
15. Gallegos KV, Browne CH, Veit FW, Talbott GD. Addiction in
anesthesiologists: drug access and patterns of substance abuse.
QRB Qual Rev Bull 1988;14:116 22.
16. Aach RD, Girard DE, Humphrey H, et al. Alcohol and other
substance abuse and impairment among physicians in residency
training. Ann Int Med 1992;116:24554.
17. Flaherty JA, Richman JA. Substance use and addiction among
medical students, residents, and physicians. Psych Clin N Amer
1993;16:189 97.
18. Yarborough WH. Substance use disorders in physician training
programs. J Oklahoma State Med Assoc 1999;92:504 7.

CRITICAL CARE

AND

TRAUMA

SECTION EDITOR
JUKKA TAKALA

EDITORIAL

Brain and Sepsis: Functional Impairment, Structural

Damage, and Markers
Nino Stocchetti,

MD

Milan University, Neuroscience ICU, Ospedale Maggiore Policlinico IRCCS, Milano, Italy

epsis affects the brain, and the impairment of

brain function resulting from sepsis is often associated with severe infectious disease. The effects of sepsis on the brain are detectable in previously
healthy brains but are amplified in cases with concomitant brain injury, as after traumatic brain injury
or subarachnoid hemorrhage. Previous injuries, in
fact, increase brain vulnerability to the complex cascade of events summarized in the term septic
encephalopathy.
In fatal cases, various anatomical lesions have been
demonstrated, such as proliferation of astrocytes and
microglia in the cortex, cerebral infarcts, brain purpura, multiple small white matter hemorrhages, central pontine myelinolysis, or disseminated microabscesses (1). In survivors, on the contrary, reversible
changes, such as reduction of cerebral blood flow,
capillary leakage, and dysfunction of the blood-brain
barrier (BBB), are involved (2,3).
The interplay between direct effects of sepsis resulting from toxic mediators and indirect effects, such as
arterial hypotension, hyperthermia, and increased intracranial pressure, contributes to the confused picture
of the brain during sepsis. A marker of cerebral damage during sepsis could have considerable value, both
for clarifying the mechanisms involved and for quantifying the degree of brain injury.
In the study by Larsson et al. (4) in this issue of
the journal, a possible mechanism is investigated
and a marker proposed in a porcine model of endotoxemic shock. After endotoxin infusion, the plasma
concentration of S-100B, a protein thought to be a
marker of glial damage, increases. This significant
increase could derive from glial or Schwann cell
damage, accompanied by an opening of the BBB.
The study detects an S-100B increase concurrent
with the endotoxin infusion that is not replicated in

Accepted for publication June 9, 2005.

Address correspondence to Nino Stocchetti, MD, Milan University, Neuroscience ICU, Ospedale Maggiore Policlinico IRCCS, Via
F.Sforza 35, 20122 Milano, Italy. Address electronic mail to
stocchet@policlinico.mi.it.
DOI: 10.1213/01.ANE.0000181337.03655.0C
2005 by the International Anesthesia Research Society
0003-2999/05

the control group. A clear association between the

infusion, its temporal profile, and the S-100B plasma
levels is therefore demonstrated.
It would be tempting to conclude that this S-100B
increase indicates brain damage, suggesting glial destruction and BBB opening. If that is the case, a pathogenetic mechanism for brain damage resulting from
endotoxin and the role of S-100B as a marker of brain
dysfunction could be advocated. Unfortunately, the
specific setting of this experiment does not provide all
the information required for such a promising
conclusion.
S-100B has been acknowledged as a marker of brain
damage; for instance, after traumatic brain injury (5).
However, there are extra-cerebral potential sources of
S-100B. In a recent study considering cerebral and
extra-cerebral infections, one fourth of patients without cerebral infection showed an increase of serum
S-100B (6). There is the possibility that this increase
was a result of brain damage concurrent with extracerebral infections, but this was not confirmed.
In trauma, S-100B is detectable in the absence of
brain injury (7,8). The largest study (387 patients and
healthy volunteers) concluded that serum S-100B is a
marker of brain damage that correlates with injury
severity, but, as major extracranial injuries also increase S-100B levels, specificity is still a problem (9).
Unfortunately, data proving or excluding potential
extra-cerebral sources of S-100B during sepsis are not
available.
In the experiment by Larsson et al., a large variability of the baseline concentration of S-100B, before endotoxin infusion, confirms that in some animals a
large concentration of the marker is present with an
intact brain, but all the reasoning is based on the
assumption that S-100B increase reflects brain damage. That is likely, but not proven, as we do not have
data measured in the cerebrospinal fluid, in the cerebral tissue, or in the cerebral veins.
Another intriguing question that cannot be answered in this experiment is the relative role of BBB
disruption. Assuming that S-100B tissue concentration
in the brain increases as a result of endotoxin infusion,
Anesth Analg 2005;101:14634

1463

1464

EDITORIAL

the concentration detected in the plasma depends both

on this increase, which creates a gradient between
tissue and intravascular concentration, and on the reflection coefficient of the interface between tissue and
blood. If the BBB permeability is slightly increased, a
much larger tissue concentration should be suspected
for achieving the plasma concentrations detected in
this study compared with a situation in which BBB
freely allows the transit of proteins.
The measurement of an independent marker of BBB
permeability would have been extremely useful to
clarify the results of this experiment. First, the degree
of BBB disruption is important per se, as BBB opening,
leading to vasogenic brain edema, is thought to be an
important mechanism of septic encephalopathy.
Moreover, if the degree of BBB opening could be
measured, some estimate of the tissue concentration of
S-100B could be attempted.
The multifaceted properties of S-100B, which acts
both as a marker of glial damage and of BBB function,
require a separation of the two components. Direct
measurements of brain variables, both in vivo (such as
microdialysis, intracranial pressure, and sagittal sinus
sampling) and postmortem, would greatly increase
the amount of information obtainable from this relatively simple experimental model.
Brain and sepsis remains a difficult and relatively
unexplored topic, and contributions such as the paper

ANESTH ANALG
2005;101:14634

by Larsson et al. (4) are welcome. The authors clearly

admit the weaknesses of their work but also highlight
the potential of this field of research.

References
1. Jackson AC, Gilbert JJ, Young GB, Bolton CF. The encephalopathy of sepsis. Can J Neurol Sci 1985;12:3037.
2. Bowton DL, Bertels NH, Prough DS, Stump DA. Cerebral blood
flow is reduced in patients with sepsis syndrome. Crit Care Med
1989;17:399 403.
3. Voigt K, Kontush A, Stuerenburg HJ, et al. Decreased plasma and
cerebrospinal fluid ascorbate levels in patients with septic encephalopathy. Free Radic Res 2002;36:7359.
4. Larsson A, Lipcsey M, Sjolin J, et al. Slight increase of serum
S-100B during porcine endotoxemic shock may indicate bloodbrain barrier damage. Anesth Analg 2005;101:14659.
5. Pelinka LE, Kroepfl A, Leixnering M, et al. GFAP versus S100B in
serum after traumatic brain injury: relationship to brain damage
and outcome. J Neurotrauma 2004;21:1553 61.
6. Unden J, Christensson B, Bellner J, et al. Serum S100B levels in
patients with cerebral and extracerebral infectious disease. Scand
J Infect Dis 2004;36:10 3.
7. Anderson RE, Hansson LO, Nilsson O, et al. High serum S100B
levels for trauma patients without head injuries. Neurosurgery
2001;48:1255 8.
8. Unden J, Bellner J, Eneroth M, et al. Raised serum S100B levels
after acute bone fractures without cerebral injury. J Trauma 2005;
58:59 61.
9. Savola O, Pyhtinen J, Leino TK, et al. Effects of head and extracranial injuries on serum protein S100B levels in trauma patients. J Trauma 2004;56:1229 34.

Slight Increase of Serum S-100B During Porcine

Endotoxemic Shock May Indicate Blood-Brain
Barrier Damage
Anders Larsson, MD, PhD*, Miklos Lipcsey, MD, Jan Sjolin, MD,
Lars-Olof Hansson, MD, PhD*, and Mats B. Eriksson, MD, PhD

PhD*,

Departments of *Medical Sciences and Surgical Sciences, Uppsala University Hospital, Sweden

Septic shock is a condition that affects many organs,

but little is known about the effects on the central
nervous system. S-100B, an acidic low molecular
weight protein, has attracted considerable interest as
a marker for brain damage and disintegration of the
blood-brain barrier. It is released into the cerebrospinal fluid and blood from brain tissue after brain damage. We studied S-100B in a porcine model of endotoxemic shock that resembles human Gram-negative
septic shock. Ten piglets received IV endotoxin, and
plasma samples were collected before the endotoxin
infusion and each hour (1 6 h) during the endotoxin

-100B is a 20-kDa protein belonging to the S-100/

calmodulin/troponin C superfamily of calciumbinding proteins. S-100 was originally isolated
from the human brain and is considered a glial-cellspecific protein (1). Today, 20 monomers of the S-100
family have been identified based on structural and
functional similarities (2,3). Most of the S-100 proteins
exist as dimers and are expressed in a cell-specific
manner. Two of the S-100 monomers, designated
S-100A1 and S-100B (4), are highly conserved between
species and are found as homo- (BB) and heterodimers (A1B) in the cytoplasm of central nervous system (CNS) glial cells and in certain peripheral cells,
e.g., Schwann cells, melanocytes, adipocytes, and
chondrocytes (5). S-100A1B and S-100BB are also
present in malignant tissues, most notably in melanoma and, to a lesser extent, in glioma, thyroid cell
carcinoma, and renal cell carcinoma (2). Determina-

Supported, in part, by grants from the Laerdal Foundation for

Acute Medicine, Stavanger, Norway.
Accepted for publication April 20, 2005.
Address correspondence and reprint requests to Anders Larsson,
MD, PhD, Department of Medical Sciences, University Hospital S-751
85 Uppsala, Sweden. Address e-mail to anders.larsson@akademiska.se.
DOI: 10.1213/01.ANE.0000180193.29655.6A
2005 by the International Anesthesia Research Society
0003-2999/05

infusion. S-100B was measured by sandwich enzymelinked immunosorbent assay. Low levels of plasma
S-100B were detected, but there was a significant increase in S-100B during Hours 15 in comparison
with the 0 values. We determined that endotoxemia
causes a very small but significant increase in the levels of the widely used brain damage marker serum
S-100B. However, it cannot be excluded that the increase in S-100B could be caused by release from organs other than the brain.
(Anesth Analg 2005;101:14659)

tion of S-100B in serum has been shown to be clinically

useful for prognosis and treatment monitoring of patients diagnosed with malignant melanoma (6 9).
S-100B has been considered to be a marker of the
blood-brain barrier and has thus been used in the
management of patients with brain damage from
causes, such as traumatic head injury, perinatal asphyxia, cardiac arrest, cardiac surgery, and stroke (10
13). S-100B is often considered as a specific brain damage marker (astroglial and Schwann cells), but it may
also be increased because of damage to other tissues
(14). S-100B has previously been measured in pigs
with acute encephalopathy (15,16) and during circulatory arrest and cardiopulmonary resuscitation
(17,18) and has been shown to be related to coma
depth, time of anoxia, and long-term outcome after
cardiac arrest (19). Even a minor head injury, such as
repeated heading by soccer players, results in increased S-100B levels (20).
Deterioration of neurologic function accompanies
organ failure in sepsis patients (21). Patients who have
sepsis with stable or supported hemodynamics and
adequate oxygenation may manifest with altered mental status. Cerebral hypoperfusion, anoxia, and progressive edema of the brain have been suggested as
potentially causative, but the pathogenesis remains
Anesth Analg 2005;101:14659

1465

1466

CRITICAL CARE AND TRAUMA LARSSON ET AL.

S-100B AND PORCINE ENDOTOXEMIC SHOCK

unknown. The brain is the source of several inflammatory mediators that have an impact on cerebrovascular function. Experimental porcine endotoxemic
shock has been shown to cause significant increases of
intracranial pressure and increased cerebral blood volume despite arterial hypotension (22). The aim of the
present investigation was to study whether the bloodbrain marker S-100B, which can repeatedly be analyzed in serum samples, is increased in porcine endotoxemic shock and, if so, whether this is related to
changes in the systemic vascular permeability.

Methods
The study included 14 domestic breed piglets of both
sexes weighing between 24 and 30 kg (average 27 kg).
Ten piglets received an endotoxin infusion, and four
piglets served as a control group. All animals were
between 12 and 14 wk old and apparently healthy. All
piglets were handled according to the guidelines of
the Swedish National Board for Laboratory Animals
and the European Convention on Animal Care. The
experiment was approved by the Animal Ethics Committee of the University of Uppsala, Sweden.
The study inclusion criteria included patients with
no apparent preexisting diseases, Pao2 10 kPa
(75 mm Hg) in arterial blood, and a mean pulmonary
artery pressure (MPAP) 2.7 kPa (20 mm Hg) at baseline and 20 min after completed preparatory procedure (see below).
General anesthesia was induced by injecting a
mixture of 6 mg/kg of tiletamin-zolazepam (Zoletil
forte vet, Virbac Laboratories, Carros, France),
2.2 mg/kg of xylazine (Rompun Vet, Bayer, Leverkusen, Germany), and 0.04 mg/kg of atropine
(Atropin, NM Pharma, Stockholm, Sweden) IM.
Anesthesia was then maintained with sodium pentobarbital
(8
mg kg1 h1;
PentobarbitalNatrium, Apoteket, Ume, Sweden), pancuronium bromide (0.26 mg kg1 h1; Pavulon,
Organon, Oss, The Netherlands), and morphine
(0.48 mg kg1 h1, Morphine, Pharmacia, Uppsala, Sweden) dissolved in 2.5% glucose solution
given as a continuous infusion. Also, sodium chloride infusion was given, resulting in a total fluid
administration rate of 30 mL kg1 h1.
A bolus dose of 20 mg of morphine was given IV before
the performance of a tracheotomy, which was performed to
secure a free airway during the experiment. The animals
were artificially ventilated throughout the experimental
procedure (Servo 900C, Siemens-Elema, Stockholm,
Sweden). During surgical stimulation, i.e., catheter insertion, 30% oxygen in N2O was given, after which the gas
mixture was set to a fraction of inspired oxygen (Fio2)
of 0.3 in medical air during the rest of the experiment.
The ventilation after preparation was adjusted to yield

ANESTH ANALG
2005;101:14659

a Paco2 between 5.0 and 5.5 kPa. The respiratory rate

was 25 breaths/min, and the inspiratory-expiratory
ratio was 1:3. Respirator settings were then kept constant throughout the experiment. To monitor hemodynamics and to permit blood sampling, a cervical
artery was catheterized, and a central venous line and
a 7F Swan-Ganz catheter (equipped with thermistor)
were inserted through the internal jugular vein into
the superior caval vein and the pulmonary artery,
respectively. A minor vesicotomy was performed, and
a urinary catheter was introduced into the bladder. A
heating pad (Operatherm 200W, KanMed, Bromma,
Sweden) was used to keep the animals at a constant
body temperature. After preparation, there was a 20min stabilization period before baseline values were
registered and baseline blood samples were taken.
Hypodynamic endotoxemic shock was induced in
10 of the piglets with a 3-h continuous infusion of
Escherichia coli endotoxin (O111:B4; Sigma Chemicals,
St. Louis, MO), where at least doubling of the initial
MPAP during the first hour was take as a sign of
endotoxemic shock. This model has previously been
described by us (23) and others (24).
S-100B was measured by a human S-100 sandwich
enzyme-linked immunosorbent assay (ELISA; S100
EIA, Prod. No. 708 10; CanAg Diagnostics AB, Gothenburg, Sweden). The antibodies in the ELISA are
raised against bovine S-100B and detect S-100A1B and
S-100BB. S-100B is highly conserved between species,
with only a single amino acid difference between human and bovine S-100B. The ELISA can thus be used
to detect S-100 from different species. Hemoglobin
was analyzed with an ABL 300 (Radiometer, Brnhj,
Denmark) as a marker for increased vascular permeability. A commercial sandwich ELISA was used for
the detection of interleukin-6 (Quantikine porcine
IL-6, P6000; R&D Systems, Minneapolis, MN). The
lower detection limit was 10 pg/mL, and the assay
had an intraassay coefficient of variation (CV) of 5%
and total CV of 10%.
For statistical analysis, the relative changes for all
values were calculated setting baseline values as zero
indexes. Differences between the animals at different
times were calculated by an analysis of variance test.
These calculations were based on the individual values for each pig at each time of measurement. Spearman rank correlation coefficients (R) and P values
were used for comparison between S-100B and hemoglobin values. A P value 0.05 was considered
significant.

Results
The animals were comparable in physiological baseline variables. All pigs given endotoxin responded to
this challenge with a marked increase in MPAP and a

ANESTH ANALG
2005;101:14659

1467

CRITICAL CARE AND TRAUMA

LARSSON ET AL.
S-100B AND PORCINE ENDOTOXEMIC SHOCK

Table 1. Physiologic Variables Expressed as Absolute Values During Endotoxin Infusion (mean sd). Interleukin-6 is
expressed as geometric mean sd.
Time (h)
pH value
Heart rate (bpm)
Mean arterial blood pressure
(mm Hg)
Mean pulmonary arterial
pressure (mm Hg)
Pulmonary static compliance
(mL/cm H2O)
White blood cells (106)
IL-6 (pg/mL)

7.5 0.0
118 13
106 20

7.4 0.1
114 23
100 28

7.3 0.1
124 42
90 30

7.3 0.0
158 41
75 20

7.3 0.0
147 36
80 22

7.3 0.0
138 20
87 27

7.3 0.0
143 19
84 35

19 2.6

34 5.3

34 4.4

38 4.9

35 3.6

32 5.9

29 6.3

16 1.8

13 2.9

12 2.8

10 2.1

10 1.7

11 1.6

11 1.5

12.9 4.2
10 1.0

6.1 2.0
105 3.7

3.7 2.0
3851 1.6

3.0 1.0
5889 1.5

3.8 1.6
3735 2.3

4.5 2.6
1558 2.7

6.3 2.5
680 3.1

continuous deterioration in circulatory and respiratory variables (Table 1). One pig died after the first
hour of the experiment.
At baseline (0 h), all animals had low S-100B values.
S-100B increased during the first 3 4 h of the experiment and decreased afterward to return to baseline
values (Figs. 1 and 2). Both relative and absolute
S-100B values were increased from 1 h to 5 h in comparison with 0 h (P 0.05). There was no significant
difference between 0-h and 6-h values. No increase in
S-100B values was detected in any of the animals in
the control group (results not shown).
Hemoglobin increased from baseline to 3 h as a sign
of hemoconcentration and increased vascular permeability. From 4 h to 6 h, there was a decrease in
hemoglobin concentration (Fig. 3). The hemoglobin
values at the end of the experiment (6 h) were not
significantly different from the baseline values. The
change in hemoglobin over time was similar to the
changes observed for S-100B, except that the amplitude of the changes was less pronounced for hemoglobin. There was also a very strong Spearman rank
correlation between the relative hemoglobin and
S-100B values (RS 0.57; P 0.0001).

Figure 1. S-100B levels in percentage of baseline value (0 value) in

individual animals. The results are presented as median, quartiles,
and range.

Discussion
Human septic illness is frequently accompanied by
abnormal cerebral manifestations, where the severity
of brain dysfunction correlates with fatal outcome
(25). In various porcine models of endotoxemia, expression of inflammatory mediators has been detected
in the brain (26), and cerebrovascular dysfunction is
associated with maldistribution of the cerebral blood
flow (21). The exact mechanism by which encephalopathy secondary to sepsis ensues is not known, but
amino acid derangements, blood-brain barrier disruption, abnormal neurotransmitters, and direct CNS effects may all contribute to this effect on the CNS (27).
Nonlipophilic substances, and especially charged proteins, such as S-100B, have a poor ability to penetrate
the tight junctions of the blood-brain barrier. When

Figure 2. S-100B levels in individual animals. The results are presented as median, quartiles, and range.

this barrier is disrupted, its penetrability is highly

increased. Interestingly, not only patients suffering
from infections that primarily affect the brain, but also
those with severe, primarily extracerebral infectious
diseases, had increased levels of S-100B, although the
levels were less frequently increased and less often

1468

CRITICAL CARE AND TRAUMA LARSSON ET AL.

S-100B AND PORCINE ENDOTOXEMIC SHOCK

Figure 3. Hemoglobin levels in percentage of baseline value (0

value) in individual animals. The results are presented as median,
quartiles, and range.

expressed among the patients with extracerebral disease (28).

There is no specific test available for porcine S-100B,
but S-100B is highly conserved between species, with
only a single amino acid substitution. The test used in
the study uses antibodies directed against bovine
S-100B and can be used to detect S-100B from different
mammalian species. It has previously been shown that
porcine S-100B can be detected by assays intended for
human use (15,16). Because serum S-100B is a marker
for brain and blood-brain barrier damage, serum samples may be an interesting alternative to cerebrospinal
samples in the assessment of cerebral injury in endotoxemic shock. If similar results are found in patients
with septic shock, analysis of serum S-100B may offer
an option to monitor organ dysfunction of the CNS in
this patient group. However, this requires further
investigation.
In this model of porcine endotoxemia leading to
infrequent mortality but frequent morbidity, a small
but significant increase in S-100B was found. No increase in S-100B was detected in the control group,
indicating that the S-100B increase was caused by the
endotoxemia and not to the surgical procedures or the
anesthesia. This is in agreement with previous reports
that showed that anesthesia did not cause an increase
in S-100B (15). Furthermore, our anesthetic procedure
is based on continuous infusion of pentobarbital,
which is frequently used to reduce increased intracranial pressure (29). Change in plasma hemoglobin concentration was used as a marker for increased vascular
permeability. We found a strong correlation between
the relative hemoglobin and S-100B values. The correlation is in agreement with a general increase in vascular permeability. At 3 h, the mean base excess was
3 mmol/L, excluding the possibility that severe ischemic damage was responsible for the observed increase in S-100B. Thus, we propose that the endotoxin

ANESTH ANALG
2005;101:14659

effect on glial cellular integrity and blood-brain barrier

disruption was the main cause of this increase, especially because the kinetic of the S-100B increase was
similar to the increase in hemoglobin that was used as
a marker for increased permeability.
In conclusion, the increase of S-100B found in this
study indicates that there is minor blood-brain barrier
damage during porcine endotoxemic shock and that
this increase is correlated to the degree of systemic
vascular permeability. However, we cannot exclude
the possibility that other cells could be the source of
the S-100B increase. S-100B could have been a potential marker for studies of CNS effects during septicemia, but further studies are required, e.g., to elucidate
the effects of cerebral blood flow on S-100B release
from the brain to the blood circulation.
The S-100B assay was generously provided by CanAg Diagnostics
AB, Gothenburg, Sweden. The technical assistance of Anders Nordgren is greatly appreciated.

References
1. Moore BW. A soluble protein characteristic of the nervous system. Biochem Biophys Res Commun 1965;19:739 44.
2. Zimmer DB, Cornwall EH, Landar A, Song W. The S100 protein
family: history, function, and expression. Brain Res Bull 1995;
37:41729.
3. Heizmann CW, Fritz G, Schafer BW. S100 proteins: structure,
functions and pathology. Front Biosci 2002;7:1356 68.
4. Schafer BW, Wicki R, Englelkamp D, et al. Isolation of a YAC
clone covering a cluster of nine S100 genes on human chromosome 1q21: rationale for a new nomenclature of the S100
calcium-binding protein family. Genomics 1995;25:638 43.
5. Takahashi K, Isobe T, Ohtsuki Y, et al. Immunohistochemical
study on the distribution of alpha and beta subunits of S-100
protein in human neoplasm and normal tissues. Virchows Arch
B Cell Pathol Incl Mol Pathol 1984;45:38596.
6. Banfalvi T, Gilde K, Gergye M, et al. Use of serum 5-S-CD and
S-100B protein levels to monitor the clinical course of malignant
melanoma. Eur J Cancer 2003;39:164 9.
7. Mrtenson ED, Hansson LO, Nilsson B, et al. Serum S-100b
protein as a prognostic marker in malignant cutaneous melanoma. J Clin Oncol 2001;19:824 31.
8. Hauschild A, Englel G, Brenner W, et al. S100B protein detection
in serum is a significant prognostic factor in metastatic melanoma. Oncology 1999;56:338 44.
9. Wunderlich MT, Ebert AD, Kratz T, et al. Early neurobehavioral
outcome after stroke is related to release of neurobiochemical
markers of brain damage. Stroke 1999;30:1190 5.
10. Martens P, Raabe A, Johnsson P. Serum S-100 and neuronspecific enolase for prediction of regaining consciousness after
global cerebral ischemia. Stroke 1998;29:2363 6.
11. Rosen H, Rosengren L, Herlitz J, Blomstrand C. Increased serum
levels of the S-100 protein are associated with hypoxic brain
damage after cardiac arrest. Stroke 1998;29:4737.
12. Ingebrigtsen T, Romner B, Marup-Jensen S, et al. The clinical
value of serum S-100 protein measurements in minor head
injury: a Scandinavian multicentre study. Brain Inj 2000;14:
104755.
13. Michetti F, Gazzolo D. S100B protein in biological fluids: a tool
for perinatal medicine. Clin Chem 2002;48:2097104.
14. Anderson RE, Hansson LO, Nilsson O, et al. High serum S100B
levels for trauma patients without head injuries. Neurosurgery
2001;48:1255 8.

ANESTH ANALG
2005;101:14659

15. Ytrebo LM, Nedredal GI, Korvald C, et al. Renal elimination of

protein S-100beta in pigs with acute encephalopathy. Scand
J Clin Lab Invest 2001;61:21725.
16. Ytrebo LM, Ingebrigtsen T, Nedredal GI, et al. Protein S-100beta:
a biochemical marker for increased intracranial pressure in pigs
with acute hepatic failure. Scand J Gastroenterol 2000;35:546 51.
17. Pokela M, Anttila V, Rimpilainen J, et al. Serum S-100beta
protein predicts brain injury after hypothermic circulatory arrest in pigs. Scand Cardiovasc J 2000;34:570 4.
18. Krieter H, Denz C, Janke C, et al. Hypertonic-hyperoncotic
solutions reduce the release of cardiac troponin I and s-100 after
successful cardiopulmonary resuscitation in pigs. Anesth Analg
2002;95:1031 6.
19. Rosen H, Sunnerhagen KS, Herlitz J, et al. Serum levels of the
brain-derived proteins S-100 and NSE predict long-term outcome after cardiac arrest. Resuscitation 2001;49:18391.
20. Mussack T, Dvorak J, Graf-Baumann T, Jochum M. Serum S-100B
protein levels in young amateur soccer players after controlled
heading and normal exercise. Eur J Med Res 2003;8:457 64.
21. Burrell R. Human responses to bacterial endotoxin. Circ Shock
1994;43:13753.
22. Hariri RJ, Ghajar JB, Bahramian K, et al. Alterations in intracranial pressure and cerebral blood volume in endotoxemia. Surg
Gynecol Obstet 1993;176:155 66.

CRITICAL CARE AND TRAUMA

LARSSON ET AL.
S-100B AND PORCINE ENDOTOXEMIC SHOCK

1469

23. Mutschler D, Eriksson M, Wikstrom B, et al. Microdialysisevaluated myocardial COX-mediated inflammation and early
circulatory depression in porcine endotoxemia. Crit Care Med
2003;31:1780 5.
24. Santak B, Radermacher P, Adler J, et al. Effect of increased
cardiac output on liver blood flow, oxygen exchange and metabolic rate during long-term endotoxin-induced shock in pigs.
Br J Pharmacol 1998;124:1689 97.
25. Young GB, Bolton CF, Austin TW, et al. The encephalopathy
associated with septic illness. Clin Invest Med 1990;13:297304.
26. Vellucci SV, Parrott RF. Expression of mRNAs for vasopressin,
oxytocin and corticotrophin releasing hormone in the hypothalamus, and of cyclooxygenases-1 and -2 in the cerebral vasculature, of endotoxin-challenged pigs. Neuropeptides 1998;32:
439 46.
27. Green R, Scott LK, Minagar A, Conrad S. Sepsis associated
encephalopathy (SAE): a review. Front Biosci 2004;9:1637 41.
28. Unden J, Christensson B, Bellner J, et al. Serum S100B levels in
patients with cerebral and extracerebral infectious disease.
Scand J Infect Dis 2004;36:10 3.
29. Censullo JL, Sebastian S. Pentobarbital sodium coma for refractory intracranial hypertension. J Neurosci Nurs 2003;35:
252 62.

Validation of a Behavioral Pain Scale in Critically Ill,

Sedated, and Mechanically Ventilated Patients
Youne`s Assaoui, MD*, Amine Ali Zeggwagh, MD, PhD*, Acha Zekraoui,
Khalid Abidi, MD*, and Redouane Abouqal, MD, PhD*

MD*,

*Service de Reanimation Medicale et de Toxicologie Clinique, Hopital Ibn Sina; and Laboratoire de Biostatistiques, de
Recherche Clinique et Epidemiologique, Faculte de Medecine et de Pharmacie, Rabat, Morocco

Assessing pain in critically ill patients, particularly in

nonverbal patients, is a great challenge. In this study,
we validated a behavioral pain scale (BPS) in critically
ill, sedated, and mechanically ventilated patients. The
BPS score was the sum of 3 subscales that have a range
score of 1 4: facial expression, upper limb movements,
and compliance with mechanical ventilation. Two assessors observed and scored pain simultaneously with
the BPS at rest and during painful procedures. The psychometric properties of the BPS that were studied were
reliability, validity, and responsiveness. We achieved
360 observations in 30 patients. The BPS was internally
reliable (Cronbach 0.72). The intraclass correlation

ssessment and management of pain in critically

ill patients have recently received increased attention (13). Scientific advances in understanding pain mechanisms, multidimensional methods of
pain assessment, and analgesic pharmacology have
improved pain management practices. However, pain
assessment for critically ill patients, especially for nonverbal patients, continues to present a challenge for
clinicians and researchers. Critically ill patients are
unable to communicate effectively for several reasons,
including tracheal intubation, reduced level of consciousness, restraints, sedation, and administration of
paralyzing drugs (4 6).
Pain experts agree that a patients self-report of pain
intensity is the most valid measure (4). Unfortunately,
most of the existing scales are designed for use with
patients who can respond verbally to assessment commands. Consequently, pain management in nonverbal
patients, such as elderly patients with cognitive impairment, is often guided by less precise and wholly
Accepted for publication April 6, 2005.
Address correspondence and reprint requests to Youne`s Aissaoui,
MD, Service de Reanimation Medicale et Toxicologie Clinique, BP
1005, Hopital Ibn Sina, 10001 Rabat, Morocco. Address e-mail to
shadowyounes@hotmail.com.
DOI: 10.1213/01.ANE.0000182331.68722.FF

1470

Anesth Analg 2005;101:14706

coefficient to evaluate inter-rater reliability was high

(0.95). Validity was demonstrated by the change in BPS
scores, which were significantly higher during painful
procedures, with averages of 3.9 1.1 at rest and 6.8
1.9 during procedures (P 0.001), and by the principal
components factor analysis, which revealed a large
first-factor accounting for 65% of the variance in pain
expression. The BPS exhibited excellent responsiveness, with an effect size ranging from 2.2 to 3.4. This
study demonstrated that the BPS can be valid and reliable for measuring pain in noncommunicative intensive care unit patients.
(Anesth Analg 2005;101:1470 6)

untested methods (7). Other methods, such as observational pain tools, must be used in a lieu of patients
self-reports of pain (8). The limited amount of data
suggests that certain observable behaviors may be
valid indicators of pain (9,10). Pain behaviors can be
markers of the existence, intensity, and causes of pain.
Indeed, observing pain behaviors is a common
method of assessing pain, especially when patients are
unable to verbalize.
Nevertheless, no pain scale comprising behavioral
indicators has been validated in the intensive care unit
(ICU), except the one developed by Payen et al. (11).
The latter consisted of a behavioral pain scale (BPS),
which was used to assess pain in patients who had
undergone thoracic or abdominal surgery or who had
been admitted for management of multiple trauma.
However, its psychometric properties were insufficiently studied, and it has never been validated in a
medical ICU. In addition, validation of any pain tool
requires repeated tests of reliability, validity, and responsiveness across samples, settings, and observers.
Therefore, the purpose of this prospective study,
which sampled from a population of critically ill patients who were sedated and mechanically ventilated,
was to validate Payen et al.s (11) behavioral scale as a
measure of pain using psychometric methods.
2005 by the International Anesthesia Research Society
0003-2999/05

ANESTH ANALG
2005;101:1470 6

Methods
The study was performed over a 6-mo period in a 12-bed
ICU of the university teaching hospital Ibn Sina, Rabat,
Morocco. The hospital ethical committee approved the
study protocol, and because this observational study did
not require any deviation from routine medical practice,
informed consent was not required.
We included patients who were older than 16 yr,
mechanically ventilated, sedated, and unconscious.
Inclusion criteria were chosen because they precluded
the use of an auto evaluation pain scale. Patients who
were quadriplegic, receiving neuromuscular blocking
medications, or had a peripheral neuropathy were
excluded. Exclusion criteria were primarily selected to
not include patients whose diseases or medications
might compromise expression of the pain behaviors.
To assess pain intensity, we used the BPS described
by Payen et al. (11). The BPS is based on a sum of three
subscales: facial expression, upper limb movements,
and compliance with mechanical ventilation (Table 1).
Each subscale is scored from 1 (no response) to 4 (full
response). Therefore, possible BPS scores range from 3
(no pain) to 12 (maximum pain).
In addition to the BPS scores, mean arterial blood
pressure and heart rate were also collected, which
were measured by multimodal monitors. These two
hemodynamic variables were collected because previous studies had shown that increased heart rate and
increased arterial blood pressure are the most frequent
physiological indicators of pain noted by observing
nurses (9).
The patients sedation levels were assessed using
the Ramsay scale (12). The Ramsay scale rates sedation
level on a scale from 1 to 6, with higher levels indicating greater degrees of sedation. This instrument
proved satisfactorily reliable and valid (13).
Sample characteristics were also recorded, including
age, sex, Acute Physiology and Chronic Health Evaluation (APACHE) II score (14), and diagnosis categories.
APACHE II score was calculated for the first 24 h.
For each patient, the BPS scores and the two physiological variables were collected three times a day by
the various teams of nurses (morning team, afternoon
team, and night team). Each team comprised four
nurses and one nurses aide. Assessments were made
by two evaluators to measure the inter-rater agreement. The two assessors were the nurse and the physician in charge of the patient. They made their assessments simultaneously but without any communication
between them. The assessors were not randomized, for
reasons of convenience.
Evaluation of the BPS and the physiological variables was made at rest and during painful procedures
to appreciate the BPS responsiveness. The two painful
procedures chosen were tracheal suction and peripheral venous cannulation. They were selected because

CRITICAL CARE AND TRAUMA

AISSAOUI ET AL.
PAIN ASSESSMENT IN CRITICALLY ILL PATIENTS

1471

their painful characters had been demonstrated in several previous studies (1517) and because they were
part of the routine care that was normally planned for
the patients. No additional interventions or procedures were performed on the patients for the benefit
of the study.
The assessments were done in the first 48 h after
admission to the ICU. However, for patients who were
not being ventilated at the time of their admission but
who were ventilated later during their stay, the assessments were made in the first 48 h after mechanical
ventilation began.
Twelve physicians and 16 nurses participated in the
study. Before the beginning of the study, a training session was provided to teach assessors how to complete
BPS, followed by a probation period (15 days), during
which the BPS was tested on some patients (n 4).
Quantitative variables were expressed as mean sd,
and significance for all statistical tests was set at P 0.05.
The sample size required for validation of the BPS
was established using the precision of a coefficient,
such as Cronbach or Intraclass Correlation Coefficient (ICC) (18). Thus, with a precision of Cronbach
of 0.90 0.05 as an objective, and for a scale of 3
subscales, it was required to include 2530 patients in
the study.
The validation of an instrument measuring a subjective variable (like pain) requires a comparison with
a gold standard. Nevertheless, no pain scale has
been validated in critically ill patients who were unable to communicate effectively because of the presence of artificial airways or underlying pathologies.
Consequently, we had to validate the BPS with indirect arguments, which consisted of checking the psychometric properties of reliability, validity, and
responsiveness.
Reliability refers to the lack of measurement error in a
scale and includes internal consistency and inter-rater
reliability. Internal consistency is an indication of how
the items within a scale are interrelated. Cronbach is
one method of assessing internal consistency (19). A high
Cronbach value reflects high internal consistency. Generally, a value larger than 0.7 is regarded as satisfactory.
Inter-rater reliability (or inter-rater agreement) is the
ability of a new instrument to obtain similar measures
with different assessors. It was assessed using the intraclass correlation coefficient (ICC) (20). Theoretically, the
ICC can range from 0 (no agreement) to 1.0 (perfect
agreement). Generally, a value larger than 0.8 is regarded as satisfactory (20). The ICC was calculated for
the BPS and for each subscale of the BPS separately. A
95% confidence interval (CI) for the coefficient was
derived.
Validity is the degree to which an instrument measures what it claims to measure (21). Validity was
established in three ways: construct validity, change in

1472

CRITICAL CARE AND TRAUMA AISSAOUI ET AL.

PAIN ASSESSMENT IN CRITICALLY ILL PATIENTS

ANESTH ANALG
2005;101:1470 6

Table 1. The Behavioral Pain Scale (11)

Item
Facial expression

Upper limb movements

Compliance with mechanical ventilation

Description

Score

Relaxed
Partially tightened (e.g., brow lowering)
Fully tightened (e.g., eyelid closing)
Grimacing
No movement
Partially bent
Fully bent with finger flexion
Permanently retracted
Tolerating movement
Coughing but tolerating ventilation for the most of time
Fighting ventilator
Unable to control ventilation

1
2
3
4
1
2
3
4
1
2
3
4

BPS scores during pain, and factor structure of the

BPS.
Construct validity is the extent to which scores on a
scale correlate with scores of other measures in predicted ways (21). We hypothesized that a significant
correlation would be found between the BPS scores
and the two physiological variables that were supposed to measure the same concept (pain). We also
tested the correlation between the BPS and the Ramsay scale. Spearman nonparametric coefficients were
used.
Change in BPS scores was assessed by comparing
the BPS scores at rest and after painful procedures. We
hypothesized that if the BPS really measures pain, the
BPS scores should be much higher during painful
procedures than while the patient is at rest. Wilcoxon
paired tests (nonparametric) were used.
Furthermore, the factor structure of the BPS was
extracted by performing exploratory principal components factor analysis. This is a statistical procedure
that enables the underlying dimensions of a scale to be
determined (21).
Responsiveness refers to an instruments ability to
detect important changes over time in the concept
being measured, even if those changes are small (22).
The magnitude of this property was assessed by the
effect size. This coefficient is calculated by dividing
the difference between the mean BPS scores at rest and
during painful procedures by the sd of the mean
scores at rest. The effect size is considered small if it is
less than 0.2, moderate if it is near 0.5, and large if it is
more than 0.8 (22).

Results
The various teams assessed 38 patients. However, the
assessments of 8 patients could not be included for 3
major reasons: (a) the patient died before the end of
the assessments (n 2), (b) the presence of exclusion
criteria (administration of neuromuscular blockade) (n

Table 2. Principal Patient Characteristics

Age (y)
Sex: men/women (n)
APACHE II score
Diagnostic categories (n)

39 19*
18/12
17 7.8*
Nontraumatic coma (11)
Acute intoxication (7)
Respiratory failure (5)
Sepsis (5)
Status epilepticus (2)

* Values expressed as mean sd.

APACHE Acute Physiology and Chronic Health Evaluation.

3), and (c) an incomplete or incorrect collection of

data (n 3).
Thirty patients were included. The principal patient
characteristics are presented in Table 2. Each patient
was assessed three times a day (morning, afternoon,
and night), by two observers (a physician and a
nurse), and at two different times (at rest and during
painful procedures). Thus, the various teams achieved
360 observations (30 patients 2 observers 2 different times 3 times per day). Realization of a complete assessment usually required 3 4 min.
All patients were sedated with midazolam in continuous infusion except one patient who received thiopental (status epilepticus). The mean amount of midazolam administered was 5.6 2.5 mg/h. The Ramsay
scale had an average value of 3.9 1.6. For analgesia,
the drug frequently used was morphine, also in continuous perfusion. The mean amount of morphine
administered was 3 0.7 mg/h.
Change in physiological variables is shown in Table
3. There was a significant increase in both hemodynamic variables during painful procedures. The amplitude of this increase was 10.7% for heart rate and
2.6% for mean arterial blood pressure.
Cronbach values indicated that the BPS had good
internal consistency (Cronbach 0.72). ICC to evaluate the inter-rater agreement were high for all subscales of the BPS. For facial expression, ICC was 0.91
(95% CI, 0.88 0.93). For upper limb movements, ICC
was 0.90 (95% CI, 0.87 0.92). For compliance with

ANESTH ANALG
2005;101:1470 6

CRITICAL CARE AND TRAUMA

AISSAOUI ET AL.
PAIN ASSESSMENT IN CRITICALLY ILL PATIENTS

Table 3. Physiological Variables at Rest and During

Painful Procedures
Rest
Heart rate (bpm)
Mean arterial blood
pressure (mm Hg)

103 22
77 26

Painful
procedures

P-value

114 23
79 27

0.001
0.042

Values expressed as mean sd.

mechanical ventilation, ICC was 0.89 (95% CI, 0.85

0.92). ICC for the total score of the BPS was 0.95 (95%
CI, 0.94 0.97). These values showed excellent interrater agreement. We also compared the BPS scores
obtained by the three teams of caregivers. There was
no significant difference (Table 4).
No significant correlation was found between the
BPS scores and the physiological variables for variability. The correlation coefficients were r 0.16 (P
0.13) for heart rate and r 0.02 (P 0.84) for mean
arterial blood pressure. When the correlation between
the BPS scores and Ramsay scale was investigated, as
expected, a significant negative correlation emerged (r
0.432; P 0.001). The higher the sedation level,
the lower the BPS scores (Fig. 1).
BPS scores obtained at rest and during painful procedures appear in Table 5. The scores were significantly greater during painful procedures than at rest
and did not differ between the two categories of painful procedures (tracheal suction and peripheral venous cannulation). Moreover, all subscale scores were
significantly higher during painful procedures.
Using exploratory principal components factor analysis, we found a large first factor, which accounted for
65% of the variance in pain expression, with strong correlation of the subscales with this factor, including coefficients of 0.90 for facial expression, 0.85 for upper limb
movements, and 0.64 for compliance with mechanical
ventilation. Table 6 shows the correlation matrix between the subscales of the BPS. The 3 subscales were
significantly correlated (all P 0.001), with a high correlation between facial expression and upper limb movements (r 0.70) and moderate correlations between
compliance with mechanical ventilation and the 2 other
subscales (r 0.40 with facial expression and r 0.29
with upper limb movements).
The effect size for responsiveness was large for the
three subscale scores and for the total BPS scores
(Table 5). These results showed an excellent responsiveness and, consequently, the excellent ability of the
BPS to quantify change in clinical status and detect
painful procedures.

Discussion
This validation study showed that the BPS had good
psychometric properties when used with critically ill

1473

patients. In particular, the BPS showed a high interrater reliability (ICC 0.95) and a satisfactory internal
consistency (Cronbach 0.72). Validity of the BPS
was demonstrated by a significant increase in BPS
scores during painful procedures and by principal
components factor analysis that identified a large first
factor, which accounted for 65% of the variance in
pain expression. Furthermore, the BPS exhibited an
excellent responsiveness, suggesting that this is a
powerful tool to detect the impact of painful stimulation in ICU patients.
Each of our patients was assessed by three teams of
nurses to remove a possible bias caused by assessments being made by the same caregivers. Results
showed that there was no significant difference
among the evaluations made by the three teams.
At rest, theoretically, the BPS scores should be equal
to 3, indicating the absence of pain. However, the
mean BPS scores, which were near 4, suggest the
possibility of preexisting background pain before any
procedure was performed. Indeed, our patients, like
all ICU patients, are subjected to a multitude of painful constraints, including various tubes (nasogastric
and endotracheal), central and arterial lines, wrist restraints, etc. Another explanation could be that the
amount of analgesic infusion was insufficient. This
fact highlights the need for an instrument that can be
used to titrate and adapt analgesia in critical care.
Pain is a stressor that produces a sympathetic stimulation (tachycardia, change in arterial blood pressure,
diaphoresis, and change in pupillary size) (4,23).
These physiological variations can help to detect pain
among patients with impaired mental status
(4,8,23,24). Puntillo et al. (9), in a study of patients
having difficulties with verbal communication (mechanically ventilated or having been tracheally extubated less than four hours), showed that the most
frequently noted physiological indicators of pain were
increased heart rate and increased arterial blood pressure. In our study, heart rate and arterial blood pressure increased significantly during painful procedures, with the increase for heart rate measuring
approximately 10%. These results coincide with the
observations of clinicians who generally associate pain
with a variation of from 10% to 20% in physiological
variables (25). However, it is agreed that these physiological indicators lack specificity in the ICU and can
be influenced by many medications (vasopressors,
adrenergic blockers, antiarrhythmics, sedative drugs,
etc.) and pathological conditions (sepsis states, shock,
hypoxia, and fear) (4). Moreover, no significant correlation was found among the BPS scores and the two
physiological variables in our study. Unfortunately,
the absence of an objective measure of pain in ICU
patients limited the testing of construct validity. The
study of Payen et al. (11) had the same results, and no
published study with a sufficient level of scientific

1474

CRITICAL CARE AND TRAUMA AISSAOUI ET AL.

PAIN ASSESSMENT IN CRITICALLY ILL PATIENTS

ANESTH ANALG
2005;101:1470 6

Table 4. Behavioral Pain Scale Scores as Assessed by Three Nursing Teams

Rest
Painful procedures

Morning team

Afternoon team

Night team

P-value*

3.8 1.2
6.6 1.7

3.7 0.9
6.8 1.7

3.9 1.2
6.6 2.2

0.44
0.46

Values expressed as mean sd.

* Friedman test.

Figure 1. Correlations between the behavioral pain scale (BPS) and

the Ramsay scale.

evidence has found a correlation among these physiological variables and pain (9).
However, the correlation between the BPS and Ramsay scale was negative and significant. The logical direction of the association is the higher the sedation level, the
lower the ability to express painful behaviors.
In the present study, the BPS yielded a Cronbach
of 0.72, thus fulfilling Nunnally and Bernsteins (26)
criterion for satisfactory internal consistency. The
inter-rater reliability of the BPS was found to be excellent (ICC 0.95). This indicates that the BPS produces consistent scores from different assessors. Reliability is an essential property when caregivers are
numerous, as in the ICU.
The BPS total and subscale scores were significantly
higher during the procedures (Table 5). This change in
BPS scores testifies to the instruments capacity to
detect and discriminate pain and provides the evidence that the BPS is a valid measure of pain. It is also
important that all of the subscales changed, indicating
that they all have the same ability to discriminate pain.
Principal factor analysis revealed that a large first
factor was dominant and that the three subscales were
strongly related to this factor, which means each of the

BPS subscales contributed to the overall pain assessment rating. The largest contributor was facial expression (r 0.90), followed by upper limb movements (r
0.85), and then compliance with mechanical ventilation (r 0.64). Furthermore, the positive significant
correlation found among the three subscales demonstrates that they evaluate the same concept, which, in
this case, was pain intensity.
This analysis has shown that behavioral indicators
can be a valid and reliable measure of pain. Few
studies have evaluated pain behaviors in the ICU
(9,10,25). The most recent one (10) identified specific
procedural pain behaviors such as grimacing, rigidity,
wincing, shutting of eyes, verbalization, and clenching
of fists. But in that study, the patients were awake and
could measure their pain with a numeric rating scale.
In fact, facial expression, which contributed most to
the pain rating in our study, is a sign found in various
works measuring both acute and chronic pain
(25,27,28). Prkachin (27) has suggested that four facial
actions carry the bulk of facial information about pain:
lowering the brow, tightening and closing of the eyelids, wrinkling of the nose, and raising the upper lip.
He has also provided evidence of the existence of a
universal facial language of pain. The facial scales,
which are especially useful for measuring pain in infants and children, highlight the value of this type of
signal (4,23,29). Pediatric scales also rely on upper
limb movements as a measure of pain (23,29). In our
study, upper limb movements contributed as much as
facial expression to the pain rating. Compliance with
mechanical ventilation, adapted from the Comfort
scale (11), had a moderate but effective contribution to
pain assessment. The reason could be that this subscale might be affected by some factors unrelated to
pain, such as hypoxemia, bronchospasm, and mucous
plugging, which can lead to coughing and some fighting of the ventilator.
In addition to these psychometric properties, the
BPS showed good feasibility, in as much as the average time of assessment was only four minutes. The
short time required will make the BPS suitable for
everyday clinical use.
This study has two limitations. First, one aspect of
the validation process has not been addressed, namely
the criterion validity (validity of the BPS in comparison with another validated pain scale). We could have
compared the BPS to subjective rating of the level pain
by an independent rater (a nurse) on a visual analog

ANESTH ANALG
2005;101:1470 6

CRITICAL CARE AND TRAUMA

AISSAOUI ET AL.
PAIN ASSESSMENT IN CRITICALLY ILL PATIENTS

1475

Table 5. Behavioral Pain Scale (BPS) Total Scores and BPS Subscale Scores at Rest and During Painful Procedures, with
the Effect Size

BPS subscales
Facial expression
Morning team
Afternoon team
Night team
Upper limb movements
Morning team
Afternoon team
Night team
Compliance with mechanical ventilation
Morning team
Afternoon team
Night team
BPS total
Morning team
Afternoon team
Night team

Rest

Painful
procedure

P*-value

Effect size

1.2 0.6
1.1 0.25
1.2 0.3

2.6 1
2.8 1.1
2.7 1.2

0.0001
0.0001
0.0001

2.3
6.8
5

1.1 0.2
1 0.2
1.2 0.5

2 0.7
1.9 0.8
1.9 0.9

0.0001
0.0001
0.0001

4.5
4.5
1.4

1.5 0.6
1.6 0.6
1.5 0.5

2 0.9
2.1 0.9
2 0.9

0.046
0.005
0.006

0.8
0.8
1

3.8 1.2
3.7 0.9
3.9 1.2

6.6 1.7
6.8 1.7
6.6 2.2

0.0001
0.0001
0.0001

2.3
3.4
2.2

* Wilcoxon paired test.

Table 6. Correlation Matrix Among the Items of the Behavioral Pain Scale

Facial expression
Movements of upper limbs
Compliance with mechanical ventilation

Facial
expression

Movements of
upper limbs

Compliance with
mechanical ventilation

1
0.70
0.41

1
0.29

Values shown represent Spearman nonparametric correlation coefficients; all correlations were statistically significant at P 0.001.

scale (VAS). However, apart from the BPS, no other

validated instrument has been developed to measure
the level of pain in mechanically ventilated ICU patients, and the VAS has never been validated in such
patients. In addition, a number of studies have found
that from 35% to 55% of nurses under-rate patient pain
when using the VAS (4). This precludes any analysis
of criterion validity in which the new instrument
would be compared to a reference instrument.
The second limitation of our study is that the sample of critical care patients observed was small. Future
studies will have to include more patients.
We conclude that the present study provides evidence that the BPS has good psychometric properties.
This instrument might prove useful to measure pain in
uncommunicative critically ill patients and to evaluate
the effectiveness of analgesic treatment and adapt it.
Further studies are required to determine whether the
use of this scale can really improve management of
pain in the critical care setting.
The authors gratefully acknowledge all the nurses and physicians
who participated in this study, Dounia Benzarouel for her assistance
with data collection, and Youne`s Lahrech and Khalil Zakari for their
help during the writing of this manuscript.

References
1. Carroll KC, Atkins PJ, Herold GR, et al. Pain assessment and
management in critically ill postoperative and trauma patients.
Am J Crit Care 1999;8:10517.
2. Puntillo KA. Pain assessment and management in the critically
ill: wizardry or science? Am J Crit Care 2003;12:310 6.
3. Edrek MA, Pronovost J. Improving assessment and treatment of
pain in the critically ill. Int J Qual Health Care 2004;16:59 64.
4. Hamill-Ruth RJ, Marohn ML. Evaluation of pain in the critically
ill patient. Crit Care Clin 1999;15:3554.
5. Puntillo KA. The phenomenon of pain and critical care nursing.
Heart Lung 1988;17:26270.
6. Shannon K, Bucknall T. Pain assessment in critical care: what have
we learnt from research. Intensive Crit Care Nurs 2003;19:154 62.
7. Taylor LJ, Herr K. Pain intensity assessment: a comparison of
selected pain intensity scales for use in cognitively intact and
cognitively impaired African American older adults. Pain
Manag Nurs 2003;4:8795.
8. Puntillo KA, Stannard D, Miakowski C, et al. Use of pain
assessment and intervention notation (P.A.I.N) tool in critical
care nursing: nurses evaluations. Heart Lung 2002;31:30313.
9. Puntillo KA, Miakowski C, Kehrle K, et al. Relationship between
behavioral and physiological indicators of pain, critical care
patients self reports of pain, and opioid administration. Crit
Care Med 1997;25:1159 66.
10. Puntillo KA, Morris AB, Thompson CL, et al. Pain behaviors
observed during six common procedures: results from Thunder
Project II. Crit Care Med 2004;32:4217.
11. Payen JF, Bru O, Bosson JL, et al. Assessing pain in critically ill
sedated patients by using a behavioral pain scale. Crit Care Med
2001;29:2258 63.

1476

CRITICAL CARE AND TRAUMA AISSAOUI ET AL.

PAIN ASSESSMENT IN CRITICALLY ILL PATIENTS

12. Ramsay MAE, Savege TM, Simpson BRJ, Goodwin R. Controlled sedation with alphaxolone-alphadolone. Br Med J 1974;
2:656 9.
13. Riker RR, Picard JT, Fraser GL. Prospective evaluation of the
Sedation-Agitation Scale for adult critically ill patients. Crit
Care Med 1999;27:13259.
14. Knaus W, Draper EA, Wagner DP, Zimmerman JE. APACHE II:
a severity of disease classification system. Crit Care Med 1985;
13:818 29.
15. Puntillo KA. Dimensions of procedural pain and its analgesic
management in critically ill surgical patients. Am J Crit Care
1994;3:116 22.
16. Puntillo KA, White C, Morris AB, et al. Patients perceptions
and responses to procedural pain: results from Thunder Project
II. Am J Crit Care 2001;10:238 51.
17. Vaghadia H, al-Ahdal OH, Nevin K. EMLA patch for venous
cannulation in adult surgical outpatients. Can J Anaesth 1997;
44:798 802.
18. Feldt LS. The approximate sampling distribution of KuderRichardson reliability coefficient twenty. Psychometrika 1965;
30:357370.
19. Cronbach LJ. Coefficient alpha and the internal structure of
tests. Psychometrika 1951;16:297334.

ANESTH ANALG
2005;101:1470 6

20. Shrout PE, Fleiss JL. Intraclass correlation: uses in assess rater
reliability. Psychol Bull 1979;86:420 8.
21. Kline P. A psychometrics primer. London: Free Association
Books, 2000.
22. Wright JG, Young NL. A comparison of different indices of
responsiveness. J Clin Epidemiol 1997;50:239 47.
23. Franck LS, Greenberg CS, Stevens B. Pain assessment in infants
and children. Pediatr Clin North Am 2000;47:487512.
24. Leisifer D. Monitoring pain control and charting. Crit Care Clin
1990;6:28394.
25. Terai T, Yukioka H, Asada A. Pain evaluation in the intensive care
unit: observer-reported faces scale compared with self-reported
visual analog scale. Reg Anesth Pain Med 1998;23:14751.
26. Nunnally JC, Bernstein IH. Psychometric theory. 3rd ed. New
York: McGraw-Hill, 1994:83113.
27. Prkachin KM. The consistency of facial expression of pain: a
comparison across modalities. Pain 1992;51:297306.
28. LeResche L, Dworkin SF. Facial expressions of pain and emotions in chronic TMD patients. Pain 1988;35:71 8.
29. Mathew PJ, Mathew JL. Assessment and management of pain in
infants. Postgrad Med J 2003;79:438 43.

A Comparison of the Laryngeal Tube and Bag-Valve Mask

Ventilation by Emergency Medical Technicians: A Feasibility
Study in Anesthetized Patients
Jouni O. Kurola, MD*, Matti J. Turunen, MD*, Juha-Pekka Laakso, RN,
Jouko T. Gorski, RN, MSc, Heikki J. Paakkonen, RN, MSc, and Tom O. Silfvast,

MD

*Department of Anaesthesia and Intensive Care, Kuopio University Hospital; Emergency Services College, Kuopio; and
Department of Anaesthesia and Intensive Care Medicine, Helsinki University Central Hospital, Finland

Airway management is of major importance in emergency care. The basic technique for all health care providers is bag-valve mask (BVM) ventilation, which requires skill and may be difficult to perform.
Endotracheal intubation, which is the advanced
method for securing the airway, is a demanding technique that has been shown to be associated with infrequent success, even when used by experienced paramedical personnel. Therefore, alternative airway
devices have been sought. The use of the laryngeal tube
(LT) by experienced anesthesia personnel had been
studied in anesthetized patients and manikins in emergency medical training. We decided to evaluate the
ability of inexperienced firefighter-emergency medical

uccessful and rapid airway management to ensure adequate oxygenation and ventilation is of
major importance in emergency care. Whereas
bag-valve mask (BVM) ventilation is the initial
method of choice for all health care providers, the
technique frequently requires considerable skill and
has been reported to be difficult (1 6). Endotracheal
intubation (ETI) is, on the other hand, considered to be
the gold standard of advanced airway management.
However, training of ETI and especially maintenance
of the skills learned is difficult, and poor success rates
have been reported, even by experienced paramedical
personnel (7,8). Therefore, alternative methods that
may be used by less experienced care providers, such
as fire fighters, have been investigated (9). These may
be of value, especially in situations where BVM is
found difficult.

Accepted for publication April 6, 2005.

Address correspondence and reprint requests to Jouni Kurola,
MD, Department of Anesthesia and Intensive Care, Kuopio University Hospital, PO Box 1777, FIN-70210, Kuopio, Finland. Address
e-mail to jouni.kurola@kuh.fi.
DOI: 10.1213/01.ANE.0000182330.54814.70
2005 by the International Anesthesia Research Society
0003-2999/05

technician students (fire-EMT) to insert the LT or perform BVM in anesthetized patients. Thirty fire-EMTs
randomly inserted the LT (n 15) and performed 1 min
of ventilation or used the BVM (n 15). We found that
all students successfully (100%) inserted the LT. Those
who inserted the LT on the first attempt (73%) required
48.2 14.7 s for the insertion. Both the LT and BVM
provided adequate oxygenation and ventilation. In this
study, we found that inexperienced fire-EMT students
inserted LT and performed 1-min ventilation with a
reasonable success rate and insertion time in anesthetized patients.
(Anesth Analg 2005;101:147781)

One of these alternative methods, the Laryngeal

tube (LT; VBM Medizintechnik GmbH, Sulz, Germany), is an oropharyngeal airway with two cuff balloons that are inflated with a single syringe. The device has been successfully tested and used in
simulated cardiac arrest in manikins and also in anesthetized patients (9 15). We evaluated the use of the
LT by trained but clinically inexperienced emergency
medical technician (EMT) students and compared its
use with BVM in anesthetized patients.

Methods
The firefighter EMT training provided at the national
Emergency Services College in Kuopio, Finland, is
basic training for prehospital emergency medical care.
Students usually have no previous health care education. The airway management curriculum for fireEMTs at the college consists of 20 h of didactic lessons
and simulations. It includes BVM using a one-person
technique to assist ventilation and the use of the LT
and ETI to secure the airway in a cardiac arrest
situation.
Anesth Analg 2005;101:147781

1477

1478

CRITICAL CARE AND TRAUMA KUROLA ET AL.

LARYNGEAL TUBE IN EMERGENCY AIRWAY MANAGEMENT

For our study, 30 EMT students in the second semester of their education were asked to participate in
a test comparing the use of the LT and BVM in anesthetized patients. During their first semester, 6 mo
earlier, the students had participated in a study on the
use of the LT (9). Thus, they received a 2-h refresher
course on the use of the LT on a manikin.
After the study protocol had been approved by the
IRB of the hospital, written informed consent was
obtained from 30 ASA physical status I-II women who
were day-surgery outpatients. Consent was obtained
during normal preoperative evaluation by the attending anesthesiologist. All patients were scheduled for a
gynecological procedure. The patients were monitored with 3-lead electrocardiogram, pulse oximetry
(Spo2), noninvasive arterial blood pressure (NIBP),
and end-tidal CO2 (ETco2). Ventilation volumes and
airway pressures were measured with a side-stream
spirometry device (Datex-Ohmeda CS 3; Datex Corp.,
Helsinki, Finland) placed between the LT or the mask
and the bag reservoir. The patients were premedicated
with 2 g of paracetamol orally, and after the anesthesiologist had cannulated a peripheral vein, anesthesia
was induced with fentanyl 2 g/kg and propofol
23 mg/kg. Thereafter, anesthesia was maintained
with sevoflurane.
The students were assigned either to insert the LT or
to use BVM according to a previously created randomization list, but they were unaware of which technique
they would use until the beginning of the test. Monitoring devices were connected, and all airway equipment was put in place on a table next to the patient
before the study protocol was begun. Irrespective of
the preassigned method, the anesthesiologist induced
anesthesia while the student observed the induction.
When the anesthesiologist, based on clinical variables
(heart rate [HR], NIBP, and deepness of sleep), judged
the level of anesthesia to be adequate, he stopped
ventilating the patient, and those students allocated to
use BVM (n 15) were told to place the mask on the
patients face and start ventilation. The students ventilated the patients for 1 min, during which period
respiratory mechanics were measured. After the
minute-long test, the surgical procedure was performed under anesthesia as scheduled.
In the LT group (n 15), the anesthesiologist ventilated the patient after the induction of anesthesia.
The test was started when the anesthesiologist asked
the student to insert the LT. A maximum of three
insertion attempts were allowed, and if the third attempt was unsuccessful, the test was stopped, and the
anesthesiologist managed the airway. After each
failed or prolonged attempt, the student performed
10 s of BVM ventilation to normalize Spo2 before the
next attempt. The time interval to place the LT was
measured from the first attempt when the student was
opening the patients mouth until the first recorded

ANESTH ANALG
2005;101:147781

Table 1. Patient Demographics (mean sd)

LT
Age (yr)
Height (cm)
Weight (kg)
Body surface area (m2)
Body mass index
(kg/m2)

46.1 13.5
159.1 4.2
67.7 11.9
1.7 0.2
26.7 4.3

BVM

P-value

41.7 15.2
NS
164.1 4.8 0.05
72.0 16.4
NS
1.8 0.2
NS
26.8 6.6
NS

LT laryngeal tube; BVM bag-valve mask (P NS).

ventilation as measured with spirometry. After successful placement of the LT, the student ventilated the
patient for 1 min, during which period respiratory
values were measured with spirometry. After 1 min,
the test was ended, and the anesthesiologist thereafter
maintained anesthesia. An LT of size 4 was used in all
patients.
NIBP and HR were measured before and after the
induction of anesthesia and after insertion of the LT
and ventilation or 1 min after BVM ventilation. Spo2
was measured continuously, as was ETco2, as soon as
ventilation was started. In both groups, the students
were asked to ventilate patients with a frequency of
their own ventilatory rate, using approximately half of
the tidal volume of the bag (Laerdal Inc., Stavanger,
Norway). An Fio2 of 1.0 was used during the test.
Results were analyzed using the Windows SPSS
software, version 10.0 (SPSS Inc., Chicago, IL). Continuous data are presented as mean with standard
deviation if not stated otherwise. We used a nonparametric t-test to test differences between groups.

Results
The patients demographics are presented in Table 1.
All students required to insert the LT did so successfully. A mean number of 1.4 0.7 attempts was required to place the LT correctly. Eleven students (73%)
successfully placed the LT on the first attempt. They
required 48.2 14.7 s from opening the patients
mouth until the first measurable ventilation was recorded. For the whole group, the average time required for insertion, from opening the mouth until the
initiation of ventilation, was 80.0 58.0 s. Expired
minute volume was 6.2 2.2 L in the LT group and 6.7
2.4 L in the BVM group (P not significant [ns]).
The corresponding expired tidal volumes were 464
135 mL and 495 151 mL, respectively (P ns; Fig. 1).
Peak airway pressure was 23.5 9.7 m Hg and
minute-ventilation airway leak was 2.4 1.4 L in the
LT group compared with 19.2 8.6 mm Hg and 3.2
1.7 L in the BVM group (P ns; Fig. 2). During
ventilation, ETco2 was 4.8 0.5 kPa in the LT group
and 4.2 1.0 kPa in the BVM group (P ns). There
was no difference in Spo2 or HR between the groups

ANESTH ANALG
2005;101:147781

CRITICAL CARE AND TRAUMA

KUROLA ET AL.
LARYNGEAL TUBE IN EMERGENCY AIRWAY MANAGEMENT

1479

Figure 1. Ventilation variables (medians, quartiles, and min/max [excluding one outlier] (P not significant). LT laryngeal tube; BVM
bag-valve mask.

Figure 2. Peak airway pressures and airway leaks in different groups (medians, quartiles, and min/max) (P not significant). LT laryngeal
tube; BVM bag-valve mask.

1480

CRITICAL CARE AND TRAUMA KUROLA ET AL.

LARYNGEAL TUBE IN EMERGENCY AIRWAY MANAGEMENT

ANESTH ANALG
2005;101:147781

Table 2. Oxygenation During Study Phases (mean sd)

(P ns)
Spo2 (%)

LT

BVM

Baseline
Lowest during study period
After ventilation

99 1
91 8
99 1

99 1
92 8
99 1

LT laryngeal tube; BVM bag-valve mask (P ns).

(Table 2). Mean arterial blood pressure after 1 min of

ventilation was 104 18 mm Hg in the LT group and
78 11 mm Hg in the BVM group (P 0.0001). Other
arterial blood pressure variables did not differ (Fig. 3).

Discussion
In this study, we found that clinically inexperienced
EMTs who had undergone manikin training successfully inserted the LT in anesthetized patients. Also,
they were able to maintain adequate oxygenation and
ventilation using both the LT and BVM.
The optimal airway device should be easy to use
and have a frequent first-attempt success rate. Multiple attempts prolong the apneic periods, and although
these devices are inserted without instrumentation,
there is a risk that repeated attempts will make the
patient prone to complications, as shown with orotracheal intubation (7).
In this setting, the EMTs placed the LT with a success rate of 100%. The first-attempt success rate of 73%
was somewhat less than that reported previously in
the hands of anesthesiologists with previous clinical
LT experience, who documented a first-attempt success rate of 90% (15). The time required for insertion in
that study was 35.1 15.9 s, compared with our
finding of 48.2 14.7 s. It is notable that our patients
were well oxygenated, fasting, and relatively healthy.
In an emergency situation such as a cardiac arrest,
patients are probably less well oxygenated, and one
may argue whether this delay is acceptable under
those circumstances. However, tracheal intubation
also requires time, and BVM is often difficult, as well
(1 8).
Comparing the observed success rates for inexperienced fire-EMTs using the LT with those achieved by
inexperienced postanesthesia care unit nurses after
manikin training using oropharyngeal airway devices,
such as the ProSeal (PLMA) or the Classic (LMA)
laryngeal mask airway, we noted that the overall success with the LT (100%) seemed more frequent than
that with the PMLA (88%) or the LMA (90%), whereas
the first-attempt success rate was less with the LT
(73%) compared with the LMA (85%) and the PLMA
(83%). The corresponding times required for insertion
(39 13 s for the LMA and 43 19 s for the PMLA)
were comparable to those recorded with the LT in this
study (16).

Figure 3. Mean arterial blood pressures (median, range, and min/

max). Baseline mean arterial blood pressure (P not significant
[ns]), mean arterial blood pressure after anesthesia induction (one
data point missing in BVM group; P ns), and mean arterial blood
pressure after ventilation (two data points missing in LT group; P
0.0001). LT laryngeal tube; BVM bag-valve mask.

1481

ANESTH ANALG
2005;101:147781

CRITICAL CARE AND TRAUMA

KUROLA ET AL.
LARYNGEAL TUBE IN EMERGENCY AIRWAY MANAGEMENT

With the LT, oxygenation, ventilation, minute volumes, and tidal volumes have been found comparable
to those achieved with ETI (9). Comparing the LT with
the BVM, we found that inexperienced EMTs could
provide adequate oxygenation and ventilation with
comparable lung volumes and airway pressures in
anesthetized patients using both methods. The LT
seems to cause a slight but significant increase in mean
arterial blood pressure as a response to insertion.
The obvious limitation of this study is its elective
design with stable patients in a controlled environment. These findings cannot be directly extrapolated
into emergency care but imply that the LT may provide a good alternative to BVM in basic life support
and in cases where BVM proves difficult. The students
were not blinded to the success of their ventilation,
which was continuously shown on a monitor. This
may have contributed to the observed performance of
BVM.
We conclude that inexperienced emergency personnel can successfully use the LT in anesthetized patients only after manikin training, and the LT may
provide an alternative to BVM. Its use in emergency
situations should be evaluated.

4. Elling R, Politis J. An evaluation of emergency medical technicians ability to use manual ventilation devices. Ann Emerg
Med 1983;12:765 8.
5. Martin PD, Cyna AM, Hunter WA, et al. Training nursing staff
in airway management for resuscitation: a clinical comparison
of the facemask and laryngeal mask. Anaesthesia 1993;48:337.
6. Rumball CJ, MacDonald D. The PTL, Combitube, laryngeal
mask and oral airway: a randomized prehospital comparative
study of ventilatory device effectiveness and cost-effectiveness
in 470 cases of cardiorespiratory arrest. Prehosp Emerg Care
1997;1:110.
7. Mort TC. Emergency tracheal intubation: complications associated with repeated laryngoscopic attempts. Anesth Analg 2004;
99:60713.
8. Bradley JS, Billows GL, Olinger ML, et al. Prehospital oral
endotracheal intubation by rural basic emergency medical technicians. Ann Emerg Med 1998;32:26 32.
9. Kurola J, Harve H, Kettunen T, et al. Airway management in
cardiac arrest: comparison of the laryngeal tube, tracheal intubation and bag-valve mask ventilation in emergency medical
training. Resuscitation 2004;61:149 53.
10. Dorges V, Ocker H, Wenzel V, Schmucker P. The laryngeal tube:
a new simple airway device. Anesth Analg 2000;90:1220 2.
11. Dorges V, Wenzel V, Schuman T, et al. Intubating laryngeal
mask airway, laryngeal tube, 1100 ml self-inflating bag: alternatives for basic life support? Resuscitation 2001;51:18591.
12. Genzwuerker HV, Finteis T, Slabschi D, et al. Assessment of the
use of the laryngeal tube for cardiopulmonary resuscitation in a
manikin. Resuscitation 2001;51:291 6.
13. Dorges V, Wenzel V, Neubert E, Schmucker P. Emergency airway management by intensive care unit nurses with the intubating laryngeal mask airway and the laryngeal tube. Crit Care
2000;4:369 76.
14. Ocker H, Wenzel V, Schmucker P, et al. A comparison of the
laryngeal tube with the laryngeal mask airway during routine
surgical procedures. Anesth Analg 2002;95:1094 7.
15. Wrobel M, Grundmann U, Wilhelm W, et al. Laryngeal tube
versus laryngeal mask airway in anaesthetised non-paralysed
patients: a comparison of handling and postoperative morbidity. Anaesthesist 2004;53:702 8.
16. Coulson A, Brimacombe J, Keller C, et al. A comparison of the
ProSeal and classic laryngeal mask airways for airway management by inexperienced personnel after manikin-only training.
Anaesth Intensive Care 2003;31:286 9.

References
1. Cummins RO, Austin D, Graves JR, et al. Ventilation skills of
emergency medical technicians: a teaching challenge for emergency medicine. Ann Emerg Med 1986;15:118792.
2. Clayton TJ, Pittmann JA, Gabbott DA. A comparison of two
techniques for manual ventilation of the lungs by nonanaesthetists: the bag-valve-facemask and the cuffed orotracheal airway (COPA). Anaesthesia 2000;56:756 9.
3. Walsh K, Cummins F, Keogh J, et al. Effectiveness of mask
ventilation performed by hospital doctors in an Irish tertiary
referral teaching hospital. Ir Med J 2000;93:557.

A Rat Model for Isolated Bilateral Lung Contusion from

Blunt Chest Trauma
Krishnan Raghavendran, MD*, Bruce A. Davidson, BS, Jadwiga D. Helinski,
Cristi J. Marschke, Patricia Manderscheid, PhD, James A. Woytash, MD,
Robert H. Notter, MD, PhD, and Paul R. Knight, MD, PhD

BS,

Departments of *Surgery, Anesthesiology, and Pathology, State University of New York (SUNY) at Buffalo, Buffalo; and
Department of Pediatrics, University of Rochester, Rochester, New York

Lung contusion affects 17%25% of adult blunt trauma

patients, and is the leading cause of death from blunt
thoracic injury. A small animal model for isolated bilateral lung contusion has not been developed. We induced lung contusion in anesthetized rats by dropping
a 0.3-kg weight onto a precordial protective shield to
direct the impact force away from the heart and toward
the lungs. Lung injury was characterized as a function
of chest impact energy (1.8 2.7 J) by measurements of
arterial oxygenation, bronchoalveolar lavage (BAL) albumin and cytology, pressure-volume mechanics, and
histopathology. Histology confirmed bilateral lung
contusion without substantial cardiac muscle trauma.

ung contusion is a common problem affecting the

care of trauma patients in the intensive care unit.
It is the most frequently diagnosed intrathoracic
injury resulting from blunt trauma, affecting 17%25%
of adult patients with this diagnosis (1,2). Lung contusion is also an independent risk factor for the development of pneumonia and severe clinical acute lung
injury (ALI) and the acute respiratory distress syndrome (ARDS) (2). Clinical-pathological correlations
in patients with lung contusion often show a lack of
correlation between the extent of contused lung and
the severity of hypoxemia (1). Patients with extensive
contusions visible on conventional chest radiography

This study was supported in part by National Institutes of Health

Grants HL-48889 (PRK, BAD), HL69763 (PRK, BAD), and HL56176
(RHN).
Accepted for publication April 26, 2005.
Address correspondence and reprint requests to Krishnan
Raghavendran, MD, Department of Surgery (Division of Trauma
and Critical Care), SUNY at Buffalo, David K. Miller Building, Suite
316, 462 Grider St., Suite 316, Buffalo, NY 14215. Address e-mail to
kraghave@ecmc.edu.
DOI: 10.1213/01.ANE.0000180201.25746.1F

1482

Anesth Analg 2005;101:14829

Rats receiving 2.7 J of chest impact energy had 33% mortality that exceeded prospectively defined limits for
sublethal injury. Hypoxemia in rats with maximal sublethal injury (2.45 J) met criteria for acute lung injury at
24 h, improving by 48 h. BAL albumin levels were
highest at 24 h, and remained elevated along with increased BAL leukocytes and decreased lung volumes at
48 h. We concluded that an impact energy of 2.45 J induces isolated, bilateral lung contusion and provides a
useful model for future mechanistic pathophysiological assessments.
(Anesth Analg 2005;101:14829)

may not be hypoxemic, whereas those with lung contusions small enough to be detectable only by computed axial tomography (CAT) scan can exhibit significant transient or prolonged oxygenation deficits
(3). A better understanding of the pathophysiological
mechanisms and cellular processes occurring in isolated lung contusion injury could significantly improve the diagnosis, prognosis, and treatment of this
important condition.
Studies investigating the pathophysiology and
cellular/molecular mechanisms of lung contusion require the use of meaningful, reproducible animal models. Various models for lung contusion have been developed, particularly in large animals such as canines,
swine, and monkeys (4 13). Although these models
have important specific applications, they also have limitations that can include frequent mortality, open chest
trauma induction, the presence of penetrating as well as
blunt trauma, and/or the existence of substantial concomitant cardiac trauma. Another disadvantage of large
animal models of lung contusion for mechanistic investigations is the lack of molecular probes and other celland mediator-specific reagents, which are much more
widely available for small animals such as mice and rats.
Studies in small-animal species also can generally be
2005 by the International Anesthesia Research Society
0003-2999/05

ANESTH ANALG
2005;101:14829

CRITICAL CARE AND TRAUMA

RAGHAVENDRAN ET AL.
A RAT MODEL FOR LUNG CONTUSION

1483

done with greater ease and less cost compared with

large-animal studies. There is currently no reliable and
reproducible small-animal model for isolated bilateral
lung contusion. Herein, we report a rat model of lung
contusion injury caused by external blunt thoracic trauma; we examined its dependence on chest impact energy
and time based on arterial oxygenation, pressurevolume (P-V) mechanics, and levels of albumin and cells
in bronchoalveolar lavage (BAL). In addition, we performed histopathological assessments of lung and cardiac tissue. Results indicate that this closed-chest smallanimal model will be helpful in future mechanistic
studies on the pathophysiology and therapy of lung
contusion.

Methods
All procedures on rats were approved by the Institutional Animal Care and Use Committee (IACUC) at
the State University of New York at Buffalo, and complied fully with state, federal, and National Institutes
of Health regulations. A modified version of a previously described model of bilateral lung contusion was
used (14). Lung contusion was induced in halothaneanesthetized adult male Long-Evans rats (250 300 g
body weight; Harlan Sprague-Dawley, Indianapolis,
IN) by dropping a hollow aluminum cylindrical
weight (0.3 kg) through a vertical stainless steel tube
onto a Lexon platform resting on the chest. The platform was suspended on Teflon guides to minimize
friction and facilitate energy transfer to the anesthetized animal. A key feature of the model was a precordial protective shield (Plexiglas), which was attached to the undersurface of the Lexon platform and
directly contacted the chest (Fig. 1). This shield protected the heart from contusion, and directed the impact energy to the lateral aspects of the chest. The
xiphoid process was clearly marked on each animal,
and the shield was reproducibly placed entirely on the
chest without intrusion onto the neck or the abdomen.
The impact energy E (in Joules, J) of the falling weight
was calculated from Equation 1:

E mgh

(1)

where m is the mass of the aluminum weight (in

kilograms), g is gravitational acceleration (9.8 m/s2),
and h is the height of weight above the Lexon platform
(in meters). Calculations assumed that all the potential
energy of the weight was transferred to the animal,
neglecting frictional dissipation. The heights for the
hollow cylindrical weight above the chest were calculated to generate external chest impact energies of 1.8,
2.0, 2.2, 2.45, and 2.7 J.
Arterial oxygenation (arterial partial pressure of oxygen, Pao2) was measured at 48 h in all groups, and
additionally at 8 min, 4, and 24 h for animals injured

Figure 1. Schematic of rat model of isolated bilateral pulmonary

contusion. A hollow cylindrical weight held in place with a pin at a
defined height was dropped through a steel tube onto a Lexon
platform with an underlying precordial shield contacting the chest
of an anesthetized rat. Teflon guides minimize fractional dissipation
and maximize energy transfer to the animal. The precordial shield
allowed the induction of bilateral pulmonary contusion but spared
the mediastinum and associated structures. The severity of lung
contusion was dependent on the impact force of the falling weight
in Joules. See text for details.

with 2.45 J of impact energy. At a given time, subgroups of animals were allowed to breathe 98% O2 for
5 min (fraction of inspired oxygen [Fio2] 0.98).
Halothane anesthesia was then induced, and a midline
incision made through the peritoneum. Blood (0.5 mL)
was collected from the descending aorta in a heparinized syringe, followed by analysis with an ABL5 blood
gas analyzer (Radiometer America, Westlake, OH).
Pulmonary P-V mechanics were measured immediately after blood samples were collected at 48 h postcontusion. A 14-gauge steel cannula was inserted into
the trachea of a halothane-anesthetized rat through a
2-cm ventral midline neck incision, and was secured
with a silk suture. The animal was then exsanguinated
by transection of the abdominal inferior vena cava. Air
was injected into the lungs at a rate of 25 mL/min by
a syringe pump connected to the tracheal cannula.
Inflation pressure was monitored continuously by an
in-line pressure transducer connected to an Apple
PowerBook G4 (Apple Computer, Cupertino, CA)

1484

CRITICAL CARE AND TRAUMA RAGHAVENDRAN ET AL.

A RAT MODEL FOR LUNG CONTUSION

equipped with a National Instruments data acquisition board (Austin, TX) and software written by the
laboratory in Lab VIEW 6.0 (National Instruments).
When pressure reached 40 cm H2O, the syringe pump
was reversed and deflation pressures were monitored.
Volumes were calculated based on the rate of injection
or withdrawal, and were normalized to kilogram
body weight.
After measurements of mechanics, a ventral midline
incision was made through the sternum and the lung
vasculature was flushed by injecting 20 mL of Hanks
balanced salt solution into the beating right ventricle.
BAL was then performed by alternating injection and
collection of 5 10 mL of 37C normal saline through
the tracheal cannula. Recovered BAL fluid was centrifuged at 1500g at 4C for 3 min to pellet cells, and the
supernatant analyzed for albumin content (described
below). The cell pellet was resuspended in 4 mL of
phosphate-buffered normal saline 0.1% sodium
azide, and total numbers of erythrocytes (RBCs) and
leukocytes (WBCs) were determined using a Multisizer 3 Coulter Counter (Beckman Coulter, Fullerton,
CA). Differential cell counts were performed on cytocentrifuge preparations (Cytospin 3; Shandon Southern Instruments, Sewickley, PA) stained with DiffQuik (Baxter, Detroit, MI).
Albumin concentrations in BAL fluid were determined by ELISA with a polyclonal rabbit anti-mouse
albumin antibody (provided by Dr. Daniel Remick,
University of Michigan, Ann Arbor, MI), a horseradish
peroxidase-labeled goat anti-rabbit immunoglobulin
G (BD Biosciences Pharmingen, San Diego, CA), and
rat albumin (Sigma, St. Louis, MO) as a standard.
Histopathological evaluations were blinded and
were performed by an experienced laboratory pathologist. Pulmonary and cardiac tissue were obtained
from rats over a 48-h period after initial chest contusion with maximal sublethal impact energy of 2.45 J.
For histological assessments of pulmonary tissue, the
lungs were gradually inflated with 1% formalin, and
tissue sections were stained with hematoxylin and
eosin (H&E). Histological assessments of cardiac tissue also used H&E staining, and included evaluations
of all four chambers of the heart in multiple sections.
Data are expressed as mean sem (standard error
of the mean) for n animals per group. One-way
ANOVA with Dunns post hoc test was used for intergroup pairwise comparisons. Values for correlations
such as r value and P value (Fishers r-z test) were
ascertained with the Statview software program.

Results
Three of the 39 injured rats failed to survive (Table 1).
Two of 6 rats in the 2.7-J chest impact energy group

ANESTH ANALG
2005;101:14829

Table 1. Mortality Rates in Different Blunt Chest Energy

Impact Groups
Impact energya
(Joules)

No. of
animals

Mortality rate (%)

1.80
2.00
2.20
2.45
2.70

9
9
9
6
6

0/9 (0 )
0/9 (0 )
1/9 (11)
0/6 (0 )
2/6 (33)

a
Impact energy was calculated from Equation (1) in Methods. Based on
criteria defined in consultation with the Institutional Animal Care and Use
Committee at State University of New York Buffalo, 2.45 J was chosen as the
maximal sublethal impact energy. See text for details.

died as a result of major intrathoracic bleeding from

lung lacerations, as determined at necropsy. In addition, 1 of 9 rats in the 2.2-J group died of intrathoracic
bleeding from a lacerated liver without injury to the
heart. All three rats that failed to survive died without
waking from anesthesia immediately after the initial
blunt trauma insult. Rats that received external blunt
chest impact energies of 2.45, 2, and 1.8 J all survived
(Table 1). Mortality in the 2.7-J impact energy group (2
of 6 animals) exceeded a 15% limit that was chosen
prospectively in consultation with the IACUC to define sublethal lung contusion injury. Thus, additional
experiments used 2.45 J as the maximal sublethal chest
impact energy condition.
All surviving animals demonstrated no evidence of
intrathoracic or intraabdominal bleeding or visceral
injury, and radiography and physical evaluation did
not reveal any associated rib fractures. In addition,
injured animals exhibited normal movement, activity,
and behavior in room air throughout the 48-h period
after lung contusion. More detailed gross observations
were made in rats injured with a chest impact energy
of 2.45 J. Gross examination at death indicated that, at
least superficially, both lungs were affected by contusion in a relatively uniform pattern in injured animals.
Areas of contusion were acutely hemorrhagic (e.g., at
4 h). By 24 h postinjury, the contused lungs had an
altered gross appearance, with prominent purplish
surface discolorations. By 48 h, areas of contused lung
appeared to be brownish in color. In contrast to the
lungs, there was no evidence of significant cardiac
trauma on gross examination.
Histological sections were examined to assess the
severity of pulmonary and cardiac tissue injury in rats
receiving the maximal sublethal chest impact energy
of 2.45 J. At 4 h and 12 h post-contusion, pulmonary
tissue showed diffuse areas of intraalveolar hemorrhage with disruption of alveoli (Fig. 2A). Perihilar
tissue was also involved, with affected areas of parenchyma in many cases extending to the visceral surface
of the pleura. Similar patterns of acute injury were

ANESTH ANALG
2005;101:14829

CRITICAL CARE AND TRAUMA

RAGHAVENDRAN ET AL.
A RAT MODEL FOR LUNG CONTUSION

1485

Figure 2. Histopathology of evolving tissue injury in rats with isolated lung contusion. A, Hematoxylin-eosin (H&E)-stained section of lung
tissue at 4 h (magnification 10). Disruption of alveoli is apparent, with blood-filled air spaces; areas of injury extend to the visceral pleural
surface. B, H&E-stained lung tissue section at 24 h (10). Neutrophils are apparent in the parenchyma and alveoli, along with evidence of
alveolar and interstitial edema. C, H&E-stained left lung tissue section at 48 h (10). Neutrophils are apparent in the air spaces, along with
some thickening of the alveolar lining. D and E, Representative H&E-stained sections of epicardial surfaces of right ventricle (D) and left
ventricular (E) tissue at 12 h after lung contusion (10). No disruption of the cardiac muscle is seen. F, H&E of the left ventricular tissue (20)
showing presence of congestion. No significant disruption of cardiac muscle is apparent. Similar histological results showing no significant
cardiac muscle disruption in the left and right atria and ventricles were found at earlier times in the 48-h period after initial lung contusion
(data not shown). See text for details.

apparent bilaterally, although quantitative morphometrics were not done. At 24 h post-contusion, a predominantly neutrophilic pattern of leukocyte infiltration was apparent in the alveolar spaces, and
significant atelectasis was also observed (Fig. 2B). At
48 h, thickening of the alveolar lining with a continuing leukocytic infiltration was apparent along with
intraalveolar edema (Fig. 2C). Tissue examination of
all four chambers of the heart did not reveal any
substantial disruption of cardiac muscle at any time
(Fig. 2DF at 12 and 24 h post-contusion).

The ratio of Pao2/Fio2 was determined at 48 h

post-contusion in rats with sublethal lung contusion
injury (1.8 2.45 J). Before blood sampling from the
abdominal aorta, animals breathed 98% oxygen to fix
Fio2 at 0.98. This brief period of high Fio2 breathing
highlighted differences in hypoxemia between experimental groups, and also facilitated subsequent absorption atelectasis to standardize P-V compliance
measurements. Pao2/Fio2 ratios at 48 h in rats given
chest impact energies of 1.8 2.45 J were equivalent to
uninjured controls (e.g., Table 2 for 2.45 J). However,

1486

CRITICAL CARE AND TRAUMA RAGHAVENDRAN ET AL.

A RAT MODEL FOR LUNG CONTUSION

ANESTH ANALG
2005;101:14829

Table 2. Evolution of Acute Pulmonary Injury over 24 h in Rats with Lung Contusion from a Chest Impact Energy of
2.45 Joules
Time after lung
contusion

Pao2/Fio2 ratio
(mm Hg)

Total RBC number

in BAL

Total neutrophil
number in BAL

Mean albumin
concentration in BAL
(g/mL)

Control (n 6)
8 min (n 89)
4 h (n 89)
24 h (n 9)
48 h (n 9)

432 15
109.8 14*
169.5 25.6*
282.5 34.1*
393 33.9

1.9 0.5 106

1.92 0.29 108*
2.71 0.02 108*
1.96 0.12 108*
5.91 1.70 107*

1.0 105
6.9 2.8 105*
3.7 0.77 106*
2.06 0.33 107*
1.38 0.26 107

22.7 3.0
427.8 67.6*
664.9 84.0*
529.0 80.4*
217.3 52.4*

Pao2 was measured in blood samples obtained after animals breathed 98% oxygen (Fio2 0.98) for 5 min at the times shown. Bronchoalveolar lavage (BAL)
samples obtained at these times were analyzed for numbers of erythrocytes (RBCs), neutrophils, and albumin (Methods). Values are mean sem (except for mean
albumin concentrations) for (n) animals per group as shown.
* P 0.05 versus control; P 0.05 versus animals at 4 h; P 0.05 versus animals at 8 min.

Pao2/Fio2 ratios at 8 min, 4 h, and 24 h post-contusion

in rats receiving the 2.45-J level of impact indicated
severe acute arterial hypoxemia, with Pao2/Fio2 ratios
of 109 14 mm Hg at 1 h and 169 25 mm Hg at 4 h
after contusion (Table 2). Significant arterial hypoxemia was also evident at 24 h post-contusion in rats
injured with 2.45 J of chest impact energy (Table 2).
Albumin levels in BAL were measured to assess
permeability injury to the alveolocapillary membrane,
as in previous studies of acute inflammatory lung
injury (1519). At 48 h post-contusion, injured rats
showed significant energy-dependent increases in
BAL albumin (r value 0.62 and P value 0.001 by
Fishers r-z test) (Fig. 3). Albumin levels in BAL at 48 h
post-contusion were 211 53 g/mL for rats receiving 2.45 J of chest impact energy, and 146 34 g/mL
for those receiving 2.2 J, compared with 22.7 3
g/mL in controls (Fig. 3). Additional time-course
studies showed that albumin levels in BAL from rats
injured with 2.45 J of chest impact energy were increased to an even greater extent in the first 24 h
post-contusion (Table 2). Early increases in BAL albumin levels were consistent with the severe arterial
hypoxemia observed in these animals at 8 min and 4 h
post-contusion (Table 2).
The cellular response in BAL varied with both chest
impact energy and time. At 48 h post-contusion, the
total numbers of leukocytes (WBCs) in BAL increased
as chest impact energy increased (Fig. 4). Animals
injured with 2.45 or 2.2 J had the most total WBCs per
milliliter in BAL fluid (5.18 3.89 107 and 3.44
2.43 107, respectively, compared with 9.19 1.13
106 in noninjured controls; Fig. 4). Additional studies
in animals injured with 2.45 J of chest impact energy
showed that the numbers of neutrophils in BAL were
significantly increased at 24 h post-contusion (Table
2). Animals injured with 2.45 J of impact energy also
had increased numbers of RBCs in BAL compared
with uninjured controls at 8 min, 4 h, and 24 h postcontusion, but with no significant differences between
values at these time points (Table 2).

Figure 3. Albumin levels in bronchoalveolar lavage (BAL) fluid at

48 h after isolated lung contusion injury. Bar diagrams show albumin levels in BAL fluid (g/mL, mean sem) from rats at 48 h after
lung contusion from blunt chest trauma of 1.8 2.45 J. Animal numbers were n 9 at each impact energy level except 2.45 J (n 6).

Quasistatic P-V inflation/deflation measurements

were done at 48 h post-contusion in rats given different levels of chest impact energy (1.8 2.45 J) (Fig. 5).
Both inflation and deflation lung volumes at fixed
pressures more than approximately 8 cm H2O were
decreased in injured animals relative to controls. Decreases in lung volumes were most pronounced in rats
injured with the two highest sublethal chest impact
energies investigated (2.2 and 2.45 J) (Fig. 5).

Discussion
This study has defined a reproducible, sublethal
model of isolated bilateral lung contusion in rats induced by focused external blunt chest trauma. Rats
receiving 1.8 2.7 J of external chest impact energy
were examined. To develop a model that would be
most applicable for future mechanistic studies on the
evolution and pathophysiology of blunt traumainduced lung contusion, an arbitrary upper limit of

ANESTH ANALG
2005;101:14829

Figure 4. Leukocyte numbers in bronchoalveolar lavage (BAL) fluid

at 48 h after lung contusion. Data are total numbers of leukocytes
(WBCs) per milliliter of BAL fluid at 48 h post-contusion for rats
injured with chest impact energies of 1.8, 2.0, 2.2, and 2.45 J. Values
are mean sem for n 9 except in the 2.45-J group (n 6).

15% mortality was prospectively chosen for maximal

sublethal injury. This mortality limit was low to guarantee substantial numbers of survivors for lung injury
assessments, whereas still allowing for some animal
deaths in response to a significant blunt chest trauma.
Rats injured with 2.7 J of chest impact energy had
mortality exceeding this limit (Table 1), whereas rats
receiving 2.45 and 2.2 J of impact energy had the most
severe sublethal injury. Injured rats had no rib fractures on chest radiographs, and bilateral lung contusion was present on gross or histological examination
without substantial disruption of cardiac muscle tissue (Fig. 2).
Isolated lung contusion injury was found to be dependent on time in addition to impact energy. Rats
injured with 2.45 J of external chest impact energy had
significantly increased levels of albumin and inflammatory leukocytes in BAL at 48 hours post-contusion
(Figs. 3 and 4), as well as decreased inflation/deflation
lung volumes at fixed pressure (Fig. 5). Hypoxemia
had significantly improved in these injured animals at
48 hours, but time course studies for rats receiving
2.45 J of impact energy showed substantial acute reductions in Pao2/Fio2 ratios that met criteria for clinical ALI and/or ARDS at 8 minutes, 4 hours, and
24 hours (Table 2). Albumin levels in BAL from rats
injured with 2.45 J of chest impact energy were significantly increased over controls at 8 minutes, 4 hours,
24 hours, and 48 hours post-contusion, with peak increases occurring at the earlier times (Table 2). Rats
receiving 2.45 J of chest impact energy also had an
acute inflammatory response, with increased numbers
of neutrophils at 24 hours post-contusion (Table 2).

CRITICAL CARE AND TRAUMA

RAGHAVENDRAN ET AL.
A RAT MODEL FOR LUNG CONTUSION

1487

Figure 5. Quasistatic pressure-volume (P-V) curves measured at

48 h after lung contusion. P-V inflation and deflation curves at 48 h
post-contusion are shown for rats injured with external chest impact
energies of 1.8, 2.0, 2.2, and 2.45 J. Pressure values are in cm H2O
and volumes are in milliliter per kilogram body weight. Points
along each curve represent the mean of values for 9 animals at each
energy level except 2.45 J (n 6).

This cellular response became more lymphocytic by

48 hours, consistent with a transition from an acute
injury to a subacute or reparative stage. Histological
analysis also confirmed the presence of neutrophils at
24 hours (Fig. 2).
Blunt chest trauma is a common problem in the care
of critically ill trauma patients (20), and thoracic
trauma accounts for 20%25% of adult deaths caused
by trauma (21). Lung contusion is the most frequently
diagnosed intrathoracic injury resulting from blunt
trauma (1,2), and is an important risk factor for the
development of other conditions such as pneumonia
and ALI/ARDS. Various large-animal models for lung
contusion have been developed, including studies in
canines, swine, and monkeys (4 13) (Table 3). In addition, a rat model of combined lung and heart contusion is available (22), and a blast-induced lung contusion model in mice has also been reported (23). The
present rat model has the advantages of being specific
in terms of lung contusion, and also uses an impact
induction method that is highly relevant for clinical
blunt thoracic trauma (e.g., as occurs in motor vehicle
accidents). As described in Table 3, many of the features and responses found in our model are consistent
with lung contusion injury in humans and in largeanimal models.
One important feature of the current rat model is
that it involves isolated (or at least relatively isolated)

1488

CRITICAL CARE AND TRAUMA RAGHAVENDRAN ET AL.

A RAT MODEL FOR LUNG CONTUSION

ANESTH ANALG
2005;101:14829

Table 3. Selected Features of Lung Contusion Found in Animal Models and Human Patients
Model

Mode of injury

Pulmonary
histopathology

BAL (quantitative
and qualitative)

Dogs (Border et al.,

1968) (4)

Weight dropped on
shielded chest

Atelectasis and air

trapping

Not reported

Mongrel dogs
(Nichols et al.,
1968) (7)

Sliding steel bar with

plate

Focal atelectasis
followed by air
trapping and
consolidation

Not reported

Swine (Moomey et al., Captive gun bolt with

Alveolar hemorrhage
1998, and Davis et
mechanical continuing
and atelectasis
al., 2000) (12,13)
ventilation
Rat model (Wang et
al., 2003) (22)
Mice (Knoferl et al.,
2003) (23)

Rat model (present

study)

Human studies
(retrospective)
(2,6,24,25)

Not reported

Swinging pendulum
Alveolar bleeding and Not reported
(combined cardiac
atelectasis
lung contusion)force
1.76 J
Blast wave from
Alveolar disruption
Not reported
compressed air
with hemorrhage

Hypoxemia
Severe hypoxemia;
changes over time
not reported
Peak at 24 h
improving by
48 h; changes
with time not
reported
Immediate and
persists for 8 h;
further
progression not
reported
Not reported

BAL albumin

Pressure-volume
compliance

Not reported

Reduced

Not reported

Reduced

10- to 20-fold
increase in BAL
protein

Not reported

Not reported

Not reported

Early hypoxia that Not reported

peaks at 6 h, and
correlates with
the force of the
blast wave
Weight dropped onto a Alveolar disruption
Significant numbers Early hypoxia that Peaks at 48 h, and
platform with a
with hemorrhage;
of WBCs in BAL,
improves by 48 h
correlates with
protective shield to
significant
which correlate
and correlates
impact energy
prevent cardiac injury
neutrophil
with energy impact with impact
(2.2, 2.45 J)
infiltration by 24 h,
energy
with minimal
associated cardiac
injury
Most common cause is
Alveolar disruption,
Not reported
Hypoxia that
Not studied
motor vehicular
hemorrhage, and
improves over
accidents
progression in some
23 days or can
to acute respiratory
progress to
distress syndrome
respiratory failure
over a period of
days

Not reported

Reduced, and
correlates with
impact energy;
LIP (lower
inflection point)
is present
Reduced, with LIP
present;
improves with
PEEP and
recruitment
maneuvers

BAL Bronchoalveolar lavage, WBC leukocytes, PEEP positive end-expiratory pressure.

lung contusion injury without substantial cardiac or

abdominal trauma. Wang et al. (22) reported an animal model of thoracic trauma without protection of
the heart in which mortality was primarily associated
with blunt myocardial injury (Table 3). In the present
study, cardiac trauma was minimized by the use of a
special platform and precordial shield that directed
the energy of external blunt chest impact toward the
lateral aspects of the chest and away from the heart.
This approach is based on the work of Stockman and
Roscher (14) in 1977, which described blunt trauma
injury in rats caused by a falling weight in a related
system. The current model differs from this previous
model in the materials and specific design of the impact delivery apparatus and how it was positioned on
the chest (e.g., positioning was relative to the marked
xiphoid process to enhance reproducibility) (Table 3).
Cardiac output and other variables of cardiac dysfunction were not measured in the present study, because
these are not specific to blunt cardiac trauma. The
hypoxemia and pulmonary hemodynamic changes
(specifically acute reactive pulmonary hypertension)
associated with lung injury have been shown in many
animal studies to seriously affect cardiac function
(13,26 28). In addition, in the present study, we examined a much more extensive range of physiological,
histopathological, and BAL assessments of lung contusion injury, and also investigated its progression

over time and its severity as a function of external

chest impact energy (Tables 1 and 2 and Figs. 25).
Many current animal models of chest contusion are
affected by how the injury is induced (Table 3). For
example, one common method of inducing lung injury
in such models involves firing an empty handgun into
the ipsilateral lung after opening the thoracic cavity
(7,11). This mechanism of injury induction in an
opened chest is very dissimilar to that found in reallife situations, and may significantly alter the associated details of injury and inflammation. Other available models of chest contusion do involve closed
thoracic trauma, but include continuing mechanical
ventilation at the time of injury that can influence the
specifics of pulmonary pathology (12,13). In our experiments, rats were not ventilated before, during, or
after lung contusion.
In addition to incorporating a clinically relevant
method of injury induction, future studies of lung
injury in the rat model can also be expected to benefit
from the greater availability of specific molecular
probes for investigating the inflammatory response in
rodents (mice, rats) compared with many large-animal
species.
Our primary emphasis in this study was on defining
the technical variables and basic characteristics of the
rat model of isolated lung contusion rather than on

ANESTH ANALG
2005;101:14829

detailed mechanistic investigations of pathophysiology. However, the results display several features consistent with other forms of acute inflammatory pulmonary injury in animals. Energy-dependent increases in
the levels of albumin in BAL are consistent with increasing levels of acute injury to the integrity of the
alveolocapillary membrane (Table 2, Fig. 4). Injury to
the alveolar epithelium as well as the capillary endothelium is indicated, because albumin had to be
present in the alveolar lumen to be accessible to lavage. The fact that albumin levels in BAL were greatest early in the course of injury at 1 and 4 hours
post-contusion, and had begun to recover by 48 hours
(Table 2), shows that substantial tearing or gross disruption of the alveolocapillary membrane did not occur in injured animals. The transient nature of the
arterial hypoxemia observed in injured rats (Table 2,
Fig. 3) also indicates that isolated lung contusion from
blunt chest trauma is at least partly reversible.
The results in rats reported here support the paradigm that a severe but isolated lung contusion could
lead to transient injury with a relatively rapid recovery. Our laboratory has previously reported that transient acute pulmonary injury in small animals also
occurs in response to other insults such as the aspiration of acid (1518). In contrast, we and others have
demonstrated that a second insult to the lung (twohit hypothesis) exacerbates inflammatory pulmonary
injury and allows it to become progressive (29). Many
patients with blunt chest trauma experience a transient clinical course with rapid recovery, whereas others exhibit disease that is severe and progressive
(2,6,24,25). The latter clinical picture of profound, persistent hypoxemia may reflect lung contusion injury
that is exacerbated by secondary events such as gastric
aspiration that are also associated with the development of ALI/ARDS. In addition to having utility in
future mechanistic studies on isolated lung contusion,
the rat model reported herein could also be adapted to
help investigate specific interactions between contusion injury and additional insults such as gastric
aspiration.

References
1. Miller PR, Croce MA, Bee TK, et al. ARDS after pulmonary
contusion: accurate measurement of contusion volume identifies high-risk patients. J Trauma 2001;51:223 8.
2. Cohn SM. Pulmonary contusion: review of the clinical entity.
J Trauma 1997;42:9739.
3. Schild HH, Strunk H, Weber W, et al. Pulmonary contusion: CT
vs. plain radiograms. J Comput Assist Tomogr 1989;13:41720.
4. Border JR, Hopkinson BR, Schenk WG Jr. Mechanisms of pulmonary trauma: an experimental study. J Trauma 1968;8:47 62.
5. Geller E, Khaw BA, Strauss HW, et al. Technetium-fibrinogen
lung scanning in canine lung contusion. J Trauma 1984;24:
611 8.

CRITICAL CARE AND TRAUMA

RAGHAVENDRAN ET AL.
A RAT MODEL FOR LUNG CONTUSION

1489

6. Fulton RL, Peter ET. The progressive nature of pulmonary

contusion. Surgery 1970;67:499 506.
7. Nichols RT, Pearce HJ, Greenfield LJ. Effects of experimental
pulmonary contusion on respiratory exchange and lung mechanics. Arch Surg 1968;96:72330.
8. Obertacke U, Neudeck F, Majetschak M, et al. Local and systemic reactions after lung contusion: an experimental study in
the pig. Shock 1998;10:712.
9. Malkusch W, Hellinger A, Konerding M, et al. Morphometry of
experimental lung contusion: an improved quantitative method.
Anal Cell Pathol 1995;8:279 86.
10. Erickson DR, Shinozaki T, Beekman E, Davis JH. Relationship of
arterial blood gases and pulmonary radiographs to the degree
of pulmonary damage in experimental pulmonary contusion.
J Trauma 1971;11:689 94.
11. Hellinger A, Konerding MA, Malkusch W, et al. Does lung
contusion affect both the traumatized and the noninjured lung
parenchyma? A morphological and morphometric study in the
pig. J Trauma 1995;39:7129.
12. Davis KA, Fabian TC, Ragsdale DN, et al. Endogenous adenosine and secondary injury after chest trauma. J Trauma 2000;49:
892 8.
13. Moomey CB Jr, Fabian TC, Croce MA, et al. Cardiopulmonary
function after pulmonary contusion and partial liquid ventilation. J Trauma 1998;45:28390.
14. Stockmann U, Roscher R. [Pulmonary contusion after blunt
chest trauma: an experimental model (authors translation)].
Thoraxchir Vask Chir 1977;25:2113.
15. Davidson BA, Knight PR, Helinski JD, et al. The role of tumor
necrosis factor-alpha in the pathogenesis of aspiration pneumonitis in rats. Anesthesiology 1999;91:486 99.
16. Knight PR, Druskovich G, Tait AR, Johnson KJ. The role of
neutrophils, oxidants, and proteases in the pathogenesis of acid
pulmonary injury. Anesthesiology 1992;77:772 8.
17. Knight PR, Rutter T, Tait AR, et al. Pathogenesis of gastric
particulate lung injury: a comparison and interaction with acidic
pneumonitis. Anesth Analg 1993;77:754 60.
18. Kennedy TP, Johnson KJ, Kunkel RG, et al. Acute acid aspiration lung injury in the rat: biphasic pathogenesis. Anesth Analg
1989;69:8792.
19. Nader-Djalal N, Knight PR, Davidson BA, Johnson K. Hyperoxia exacerbates microvascular lung injury following acid aspiration. Chest 1997;112:160714.
20. Shorr RM, Crittenden M, Indeck M, et al. Blunt thoracic trauma:
analysis of 515 patients. Ann Surg 1987;206:200 5.
21. LoCicero J 3rd, Mattox KL. Epidemiology of chest trauma. Surg
Clin North Am 1989;69:159.
22. Wang ND, Stevens MH, Doty DB, Hammond EH. Blunt chest
trauma: an experimental model for heart and lung contusion.
J Trauma 2003;54:744 8.
23. Knoferl MW, Liener UC, Seitz DH, et al. Cardiopulmonary,
histological, and inflammatory alterations after lung contusion
in a novel mouse model of blunt chest trauma. Shock 2003;19:
519 25.
24. Kollmorgen DR, Murray KA, Sullivan JJ, et al. Predictors of
mortality in pulmonary contusion. Am J Surg 1994;168:659 63.
25. Schreiter D, Reske A, Scheibner L, et al. [The open lung concept:
clinical application in severe thoracic trauma]. Chirurg 2002;73:
3539.
26. Schuster DP, Perez JE, Trulock EP, et al. Cardiac dysfunction
during acute lung injury induced by oleic acid in dogs. Am Rev
Respir Dis 1986;133:519 25.
27. Steenbergen C, Deleeuw G, Rich T, Williamson JR. Effects of
acidosis and ischemia on contractility and intracellular pH of rat
heart. Circ Res 1977;41:849 58.
28. Teplinsky K, OToole M, Olman M, et al. Effect of lactic acidosis
on canine hemodynamics and left ventricular function.
Am J Physiol 1990;258:H11939.
29. Knight PR, Davidson BA, Nader ND, et al. Progressive, severe
lung injury secondary to the interaction of insults in gastric
aspiration. Exp Lung Res 2004;30:53557.

CASE REPORT

Delayed Respiratory Depression After Risperidone Overdose

Ahmet Akyol, MD, A. Can Senel,
and Nesrin Erciyes, MD

MD,

Hulya Ulusoy,

MD,

Filiz Karip,

MD,

Department of Anesthesiology and Critical Care, Karadeniz Technical University, Trabzon, Turkey

Risperidone is an atypical antipsychotic drug used for

the treatment of schizophrenia. Both positive and negative symptoms are prominent with its use. Metabolism
occurs mainly in the liver, where risperidone is
changed by CUP2D6 to an active metabolite,
9-hydroxyrisperidone. The half-lives of risperidone
and its metabolite are 3 and 7 h, respectively. Genetic

ntipsychotic drugs, also called major tranquilizers and neuroleptics, are used to treat schizophrenia and other types of psychoses. Toxicity
of these drugs may result from suicidal overdose or
from adverse reactions after therapeutic administration. Risperidone overdose is rare but life-threatening.
All antipsychotics produce extrapyramidal side effects
(EPS), such as acute dystonic reactions, which can
cause respiratory difficulty by pharyngeal and laryngeal muscle spasm. The introduction of drugs that
produce minimal EPS and of atypical antipsychotics
(e.g., risperidone) has allowed separation of antipsychotic from neuroleptic effects and prevents the interchange of terms. The serotonin-dopamine antagonist
concept contends that antipsychotics that are more
potent at 5-hydroxytryptamine 2A receptors than at
D2 receptors (e.g., risperidone) have a low EPS liability (1).

Case Report
A 26-yr-old woman who had been treated with risperidone
for 3 mo was found unconscious on the street. The patient
was admitted to a state hospital where it was determined
that she had ingested 30 mg of risperidone (a suicide attempt after an argument with her husband) rather than her
usual amount of 6 mg/d. Her medical history included
long-term schizophrenia.
Accepted for publication April 22, 2005.
Address correspondence and reprint requests to Ahmet Akyol,
MD, Department of Anesthesiology and Critical Care, Karadeniz
Technical University, Faculty of Medicine, 61080 Trabzon, Turkey.
Address e-mail to ahmetmelike@yahoo.com.
DOI: 10.1213/01.ANE.0000180195.57760.E9

1490

Anesth Analg 2005;101:14901

polymorphism is seen in the 6% 8% of white patients

who are considered poor metabolizers. In poor metabolizers, the half-life extends to 20 30 h. We present an
unusual case of unanticipated delayed respiratory depression after risperidone overdose.
(Anesth Analg 2005;101:1490 1)

On Day 3 (after 52 h), respiratory distress developed, and

the patient was transferred to our university hospital. On
admission, she was hypotensive, with an arterial blood pressure (BP) of 80/50 mm Hg, and she had a Glasgow Coma
Scale (GCS) score of 8 of 15. A chest radiograph showed no
abnormality. Serum electrolytes and her electrocardiogram
were normal (heart rate, 50 bpm). The slow heart rate and BP
responded initially to treatment that included IV fluids and
a single dose of 0.5 mg of atropine.
Two hours after arrival in the intensive care unit (ICU), the
patients respiratory drive began failing, and she started
gasping. The blood gas analysis at this time, under O2 6
L/min with face mask, showed a pH value of 7.39, Po2 of
51 mm Hg, Pco2 of 54 mm Hg, and HCO3 of 24.3 mmol/L,
with a respiratory rate of 5 breaths/min. After 10 min, the
patient had a respiratory arrest and was endotracheally
intubated and ventilated. Sedation was maintained by
propofol infusion. A computed tomography scan of the head
was performed and showed no abnormalities. There were
no other causes of respiratory depression, and she was not
receiving any other respiratory depressant drugs. The patient remained hemodynamically stable, and laboratory tests
were normal.
The patient started to make respiratory efforts on Day 5.
Her reported GCS was 13. After 6 h, she managed to breathe
spontaneously with a continuous positive airway pressure
mode. She was tracheally extubated on Day 6 and discharged from the ICU on Day 8.

Discussion
Toxicity of antipsychotic drugs results from unintentional or intentional overdose or from adverse reactions after therapeutic administration. EPS are a result
of basal ganglia dopamine D2 receptor blockade and
consist of four main drug-induced syndromes. These
can be divided into reversible syndromes, which occur
within days to weeks of the onset of antipsychotic
2005 by the International Anesthesia Research Society
0003-2999/05

ANESTH ANALG
2005;101:1490 1

therapy (acute dystonia, parkinsonism, and akathisia)

and a potentially irreversible syndrome that occurs
after months to years of therapy (tardive dyskinesia).
Acute dystonic reactions (dyskinesias) consist of intermittent spasmodic or sustained involuntary contractions of muscles in the face, neck, trunk, and occasionally the extremities. Resulting clinical
manifestations include trismus, tongue protrusion,
torticollis, opisthotonos, and respiratory difficulty.
Pharyngeal and laryngeal muscle spasm may produce
respiratory distress and asphyxia (2,3).
Reported side effects of risperidone are lethargy,
dystonia, hypotension, tachycardia, dysrhythmia, impaired concentration, and abnormal temperature regulation (4 7). The recommended dose is 2 mg on Day
1, 4 mg on Day 2, and 6 mg on Day 3; however, in the
elderly (or in patients with hepatic or renal failure),
the dose is 0.5 mg up to 2 mg on Day 3. This patient
was under treatment with risperidone for three
months, and her usual medication was 6 mg/d.
Delayed respiratory failure has not been a prominent feature in overdose cases. Rassam and Srinivasa
(8) reported a 72-year-old patient with respiratory depression after accidental risperidone overdose, and
Acri and Henretig (9) reported cases of impaired respiration when risperidone was given concomitantly
with other respiratory depressant drugs resulting in
no additive side effects.
In this case, we evaluated whether acute dystonic
reaction contributed to the clinical course of our patient, who demonstrated lethargy, tongue protrusion,

CASE REPORT

1491

and pharyngeal-laryngeal muscle spasm, all of which

were related to respiratory dysfunction.
Experience with risperidone overdose is limited and
the clinical presentation difficult to predict after an
overdose. Therefore, patients who are being treated
for risperidone overdose should be monitored for hypotension, sedation, and respiratory depression. It
should be noted that respiratory depression may be
seen as late as the third day after risperidone
overdose.

References
1. Huttunen M. The evolution of the serotonin-dopamine antagonist concept. J Clin Psychopharmacol 1995;15:4 8.
2. Rupniak NMJ, Jenner P, Marsden CD. Acute dystonia induced by
neuroleptic drugs. Psychopharmacology (Berl) 1986;88:4035.
3. Sweet C. Drug-induced dystonia. Am J Psychiatry 1975;132:
5327.
4. Himstreet JE, Daya M. Hypotension and orthostasis following a
risperidone overdose. Ann Pharmacother (United States) 1998;
32:26770.
5. Brown K, Levy H, Brenner C, et al. Overdose of risperidone. Ann
Emerg Med 1993;22:1908 10.
6. Vecchio FL, Hamilton RJ, Hoffman RJ. Risperidone overdose.
Am J Emerg Med 1996;14:95 6.
7. British Medical Association and the Royal Pharmaceutical Society of Great Britain. Risperidone. British National Formulary
2001;41:1829.
8. Rassam S, Srinivasa R. Respiratory depression after accidental
risperidone overdose. Am J Emerg Med 2002;20:570 4.
9. Acri AA, Henretig FM. Effects of risperidone in overdose.
Am J Emerg Med 1998;16:498 501.

NEUROSURGICAL ANESTHESIA
SECTION EDITOR
DAVID S. WARNER

The Efficacy of Postoperative Ondansetron (Zofran) Orally

Disintegrating Tablets for Preventing Nausea and Vomiting
After Acoustic Neuroma Surgery
Theresa Hartsell,

MD, PhD*,

Donlin Long,

MD, PhD,

and Jeffrey R. Kirsch,

MD*

*Departments of Anesthesiology and Critical Care Medicine and Neurosurgery, The Johns Hopkins Medical Institutions,
Baltimore, Maryland

Postoperative nausea and vomiting is a frequent complication of craniotomy. We evaluated the ability of intraoperative IV ondansetron followed by postoperative
ondansetron in an orally disintegrating tablet formulation to reduce the frequency and severity of postoperative nausea and vomiting in a prospective, randomized,
placebo-controlled double-blind trial of 60 patients undergoing acoustic neuroma resection. Each patient received intraoperative ondansetron (4 mg IV) or placebo
30 min before case end. Postoperatively, patients received ondansetron in an orally disintegrating tablet
formulation (8 mg BID) or placebo twice a day for up to
72 h. Metoclopramide was available as rescue therapy
for both groups. Severity of nausea (as measured on a
10-cm visual scale), number of emetic episodes, and

ostoperative nausea and vomiting (PONV) is a

common complication after craniotomy. Postcraniotomy nausea occurs in approximately half of
all patients, with 39 percent of patients experiencing one or
more episodes of emesis (1). Of all craniotomy patients,
those undergoing resection of an acoustic neuroma would
seem to experience PONV most frequently, possibly because of tumor and/or surgical involvement of the vestibular nerve.
Fabling et al. (2) evaluated the efficacy of preemptive ondansetron in adult patients undergoing infratentorial craniotomy. They observed that treatment with ondansetron before emergence from
anesthesia prevented early PONV but did not have an
effect on delayed PONV (after 12 h). In the current study
Supported, in part, by GlaxoSmithKline, Research Triangle Park,
North Carolina.
Accepted for publication May 31, 2005.
Address correspondence and reprint requests to Jeffrey R. Kirsch,
MD, Professor and Chair, Department of Anesthesiology and Perioperative Medicine, Oregon Health and Science University, 3181
S.W. Sam Jackson Park Rd. UHS-2, Portland, OR 97239. Address
e-mail to kirschje@ohsu.edu.
DOI: 10.1213/01.ANE.0000181007.01219.38

1492

Anesth Analg 2005;101:14926

requirement for rescue therapy were recorded. In the

immediate postoperative period, nausea severity was
less in patients treated with ondansetron than placebo
(3.3 4.1 versus 7.3 4.2; P 0.001) and fewer patients
experienced vomiting (3 of 28 versus 11 of 32; 2 P
0.01). More patients required some form of rescue treatment in the placebo group on the first postoperative
day (26 of 32 versus 16 of 28; 2 P 0.01). We conclude
that after acoustic neuroma surgery IV ondansetron
treatment prevents immediate postoperative nausea
and vomiting. Postoperative treatment with ondansetron in an orally disintegrating tablet formulation was
associated with less frequent rescue therapy as compared with placebo on the first postoperative day.
(Anesth Analg 2005;101:14926)

we evaluate the therapeutic efficacy of intraoperative ondansetron combined with postoperative, preemptive treatment with an orally disintegrating tablet (ODT) formulation of ondansetron (Zofran ODT, GlaxoSmithKline,
Research Triangle Park, NC; henceforth ondansetron ODT)
in the prevention of delayed PONV after infratentorial craniotomy in patients with a diagnosis of acoustic neuroma.
Ondansetron ODT was used because it was previously
demonstrated to be well tolerated by patients and to have
similar efficacy as oral ondansetron for cyclophosphamideinduced emesis in cancer patients (3). Ondansetron ODT
has also been demonstrated to significantly reduce the incidence of postdischarge nausea and vomiting in ASA
physical status III patients undergoing outpatient gynecological laparoscopy (4). This formulation may be particularly helpful in patients who have difficulty swallowing
postoperatively because of trauma to lower cranial nerves
during surgery.
This study was designed to test the hypothesis that
ondansetron (IV followed by ODT) would reduce both
the frequency and severity of PONV in a population of
adult patients undergoing craniotomy for acoustic
neuroma resection.
2005 by the International Anesthesia Research Society
0003-2999/05

ANESTH ANALG
2005;101:14926

NEUROSURGICAL ANESTHESIA
HARTSELL ET AL.
PONV AFTER ZOFRAN ODT

1493

Table 1. Standardized Anesthetic for Acoustic Neuroma Resection

Event
Induction

Placement of Mayfield Head Holder

Maintenance

Closure
Randomization (blinded) (30 min before
extubation)
Tracheal extubation

Drug

Standardized dose

Midazolam
Sodium Thiopental
Fentanyl
Rocuronium
Sodium Thiopental
Lidocaine
Fentanyl
Rocuronium or
Pancuronium
Isoflurane
Fentanyl
Isoflurane
Study medication

14 mg IV
35 mg/kg IV
710 g/kg IV
0.61.2 mg/kg IV
Up to 3 mg/kg IV
Up to 1 mg/kg IV
12 g/kg/h IV infusion
Titrated to maintain 23 twitches for
train-of-four monitoring
0.4%1.5% end-tidal in oxygen
Discontinued
Reduced/discontinued
As provided by pharmacy, IV bolus
injection
10 g/kg (maximum 1 mg)
70 g/kg (maximum 5 mg)

Glycopyrrolate
Neostigmine
100% oxygen

Methods
After obtaining IRB approval, we conducted a randomized, prospective, placebo-controlled, doubleblind trial in ASA IIII patients 18 80 yr of age undergoing acoustic neuroma resection at The Johns
Hopkins Hospital. After obtaining informed consent,
60 patients were divided into 2 groups (ondansetron
IV plus ondansetron ODT, n 28; placebo IV plus
placebo ODT, n 32). Exclusion criteria included any
patients experiencing emesis in the 24 h preceding
surgery, taking antiemetic medications during the immediate preoperative period, or with a history of intolerance to serotonin receptor antagonists. Patients
requiring postoperative mechanical ventilation were
also not included because they would not be able to
express their degree of PONV and would likely require additional sedation.
All ASA IIII patients between the ages of 18 and 80
yr scheduled for acoustic neuroma resection at our
institution between October 2000 and December 2002
were approached for informed consent during the
initial anesthesia evaluation on the day of surgery. The
patients age, gender, ASA classification, medical history, and current state of health were recorded with
emphasis on the presence of systemic disease, recent
emesis and/or use of antiemetic medications, and
drug sensitivities. The first day of the last menstrual
cycle was recorded for female subjects. Tumor size
was recorded from either preoperative or postoperative sources (e.g., pathology reports, imaging studies).
All study participants received routine perioperative
monitoring for acoustic neuroma resection, including
continuous electrocardiogram, pulse oximetry, temperature, noninvasive blood pressure, intraarterial blood pressure via radial or femoral arterial line, somatosensory
evoked potentials, brainstem auditory evoked potentials

(BAERs), and electromyography of the facial nerve. A standardized isoflurane/opioid anesthetic was provided (Table
1), representing routine anesthetic management for this
procedure in our institution. All patients in the study received both intraoperative (standard dose of 8 mg, IV at the
time of incision) and postoperative dexamethasone (as requested by the surgeon). At the completion of surgery the
subjects trachea was extubated if deemed appropriate by
the attending anesthesiologist. Patients requiring continued
mechanical ventilation for any reason were excluded from
further participation to avoid the confounding effects of the
endotracheal tube on the gag response and because of the
possibility of limited patient communication.
At the time of randomization the research pharmacist was provided with the patients gender and
whether a change in eighth cranial nerve function (as
determined by BAERs at completion of tumor resection) was detected, as both of these factors appeared to
independently alter the incidence of PONV in our
preliminary observations. Postoperatively, patients received antiemetic medication on a regularly scheduled
basis, either orally disintegrating ondansetron (ondansetron ODT 8 mg BID) or placebo for up to 72 h. The
first dose of ondansetron or placebo was administered
between 4 h and 8 h after tracheal extubation. Metoclopramide (10 mg IV prn every 8 h) was provided as
rescue therapy to any patient as needed for PONV
events.
For each patient, the number of emetic episodes and
requirement for rescue antiemetic treatment was determined from nursing notes and medication administration records. Severity of nausea was recorded using a 10-cm visual analog scale (VAS) twice daily
during an interview with the patient. No nausea
was considered a VAS score of 0, whereas the occurrence of any emesis during the preceding 12-h period
was considered a VAS score of 10. Patient satisfaction

1494

NEUROSURGICAL ANESTHESIA
PONV AFTER ZOFRAN ODT

HARTSELL ET AL.

with control of PONV, as well as the presence of any

side effects potentially related to antiemetic therapy,
was also recorded using a VAS and by discussion with
the physician and nursing staff. Decision to withdraw
from the study could be made by an individual patient
or primary care team at any time during the postoperative period. All patients stayed in the study long
enough to obtain data on the day of surgery.
The study safety committee reviewed the accumulated data after enrollment of 30 patients to determine
appropriateness of continuing the study to the
planned completion of 60 patients. The safety committee was instructed to stop the study if there appeared
to be an overwhelming benefit of drug treatment or if
drug treatment was found to have unexpected toxic
effects. The investigators were unaware of the findings
of the safety committee but were advised to complete
the planned study of 60 patients.
Values are expressed as mean sem. Comparisons
between groups for parametric data were accomplished with Students t-test. Expected frequencies for
patients receiving ondansetron were compared with
the observed frequencies in patients treated with placebo using 2 tests. P values of less than or equal to
0.05 were considered statistically significant.

Results
A total of 133 patients with a preoperative diagnosis of
acoustic neuroma presented to our operating rooms
during the period of patient recruitment. Sixty patients participated in the study. Other patients either
refused to participate in the study (n 20), presented
to the operating room on a day no investigator was
available to obtain informed consent (n 7), had
medical conditions that were in our exclusion criteria
(n 28, of which 10 were excluded because they were
not tracheally extubated at the end of the case) or
presented at a time when the IRB approval had expired and was being re-reviewed (n 18). Of the 60
patients who entered the study, there were no differences between groups in demographic data (Table 2).
The most common presenting complaint in patients in
both groups was hearing loss, followed by tinnitus
and vertigo.
In the immediate postoperative period, the VAS
score for nausea was lower (P 0.001) in patients
treated with intraoperative IV ondansetron (3.3 4.1)
as compared with patients treated with intraoperative
IV placebo (7.3 4.2). In the placebo group, 11 of 32
patients experienced immediate postoperative vomiting. However, only 3 of 28 patients in the ondansetron
group experienced immediate postoperative vomiting
(2 P 0.01). In the placebo group, 11 of 32 patients
asked to be removed from the study before completing the 3 days of evaluation (6 on the morning of

ANESTH ANALG
2005;101:14926

Table 2. Demographic Data for Patients Treated with

Placebo (IV Placebo and Placebo ODT) or Ondansetron (IV
Ondansetron and Ondansetron ODT)

Age (yr)
Female/male
Time in OR (h)
BAER lost during surgery
(number of pts)
Tumor size (cm)

Placebo

Ondansetron

50 12
21/11
7.6 1.5
20 of 32

51 10
17/11
8.2 1.9
12 of 27

2.1 0.8

1.8 0.9

There were no differences between groups for any of the demographic

variables.
OR operating room; BAER brainstem auditory responses; pts
patients; ODT orally disintegrating tablet.

postoperative day 1). In the ondansetron group, 7 of

28 patients asked to be removed from the study before
completing the 3 days of evaluation (none on the
morning of postoperative day 1 but 6 on the afternoon
of postoperative day 1). Female gender was associated
with worse intensity of immediate postoperative nausea (higher VAS) in patients treated with intraoperative placebo (P 0.001) but not in patients treated
with intraoperative IV ondansetron. Loss of vestibular
nerve function did not predict severity of immediate
postoperative nausea in patients treated with placebo
or ondansetron.
After 24 h of recovery, emesis was rare and not
different between groups (1 patient in the ondansetron
group; 2 patients in the placebo group). However,
more patients still required some form of rescue treatment in the placebo group (26 of 32 in the placebo
group; 16 of 28 in the ondansetron group; 2 P 0.01)
on the first postoperative day. On the second postoperative day, rescue treatment was required in 5 of 21
remaining patients in the placebo group and in 2 of 22
remaining ondansetron-treated patients.
We did not observe any complications associated
with IV or ondansetron ODT treatment. Two patients
withdrew from the study because they did not like the
taste of the ondansetron ODT. A third patient developed a transient superficial mucosal ulcer under her
tongue but completed the study.

Discussion
The main finding of the study is that IV ondansetron
treatment, administered 30 minutes before the end of
surgery, prevents immediate PONV in patients undergoing acoustic neuroma resection. Postoperative ondansetron ODT treatment in patients treated with intraoperative IV dexamethasone plus ondansetron was
associated with less frequent rescue therapy on the
first postoperative day as compared with patients receiving IV intraoperative placebo plus dexamethasone
and postoperative placebo-ODT. In most patients,

ANESTH ANALG
2005;101:14926

PONV after resection of acoustic neuroma is selflimiting and does not benefit from preventive treatment with ondansetron after the second postoperative
day.
Before initiating the current study, therapy for
PONV in our hospital involved the use of intraoperative dexamethasone prophylaxis and postoperative
metoclopramide or prochlorperazine as needed in response to emesis or patient complaint of nausea (i.e.,
rescue therapy). Serotonin type 3 (5-hydroxytryptamine; 5-HT3) receptor antagonists such as ondansetron (Zofran) have proven effective in reducing early
PONV in patients undergoing middle ear surgery
(5,6). An initial study of ondansetron in pediatric craniotomy patients failed to reach statistical significance
(7) when evaluating total number of emetic events, but
this early study was not designed to reveal differences
between ondansetron and placebo groups in either
severity of nausea or in requirement for rescue
therapy.
In a retrospective analysis the frequency of PONV
was increased after infratentorial as compared with
supratentorial craniotomy (1). However, this was not
substantiated in a prospective analysis that compared
postoperative pain and nausea in patients having either craniotomy or spine surgery (8). In a retrospective
analysis initial PONV (first four postoperative hours)
was more frequent in patients undergoing craniotomy
under general anesthesia as compared with patients
who have craniotomy in the awake state (9). The increased frequency of PONV in patients after acoustic
neuroma surgery probably relates to disruption of the
vestibular system during surgery (10). Although
movement often provokes PONV after acoustic neuroma surgery, lack of movement may delay recovery
and prevent adaptation to vestibular disruption. Our
study suggests that aggressive use of antiemetic strategies, including ondansetron ODT, may allow patients
the opportunity to participate in vestibular adaptation
exercises earlier.
In a prospective, double-blind, placebo-controlled
study, ondansetron 4 mg IV administered within
1 hour after the end of surgery was found to decrease
the incidence of both nausea and vomiting for the first
24 postoperative hours. However, Fabling et al. (2)
demonstrated that 8 mg IV ondansetron at skin closure did not decrease the incidence of PONV at any of
their individual measurement points but was associated with a decrease in the overall likelihood of developing PONV within the first 12 postoperative
hours. The authors of this study suggested further
evaluation of scheduled antiemetic therapy during the
first 48 hours after infratentorial surgery. We believe
that our study addresses this suggestion.
All patients in this study were treated with dexamethasone before surgical incision at the request of

NEUROSURGICAL ANESTHESIA
HARTSELL ET AL.
PONV AFTER ZOFRAN ODT

1495

the faculty surgeon with the intent to decrease postsurgical brain edema. Dexamethasone also has antiemetic properties. In particular, dexamethasone was
demonstrated to be effective for prophylaxis of PONV
after tympanomastoid surgery (11). Some authors (12)
have demonstrated similar efficacy for dexamethasone in comparison to ondansetron, but this has not
been evaluated specifically in neurosurgical patients.
Many authors believe that overall superior antiemetic
therapy is achieved by combining dexamethasone
with a 5-HT3 receptor antagonist (13). Although the
mechanism by which dexamethasone decreases the
incidence of PONV is unknown, it is likely that it
affects multiple important pathways (13). In the current study it was not possible to determine if the
effects of ondansetron ODT would have been similar
in the absence of concurrent dexamethasone
treatment.
Although ondansetron ODT was effective in decreasing
the intensity and frequency of PONV in this patient population, many patients in this group also benefited from
multimodal antiemetic therapy. We believe that this study
supports the conclusion that when one drug has efficacy in
preventing or treating PONV it should be continued, using
a dosing strategy that is consistent with its pharmacodynamic and pharmacokinetic properties. However, when a
single drug does not demonstrate adequate efficacy, combination therapy should be initiated (14,15). Repeat treatment with a drug (single modality therapy) that shows no
efficacy after the first dose is not likely to be effective after
the second dose (16). In our control group multimodal
therapy was achieved with the combination of dexamethasone and metoclopramide. Prophylactic metoclopramide
has been demonstrated in one study to be more effective
than ondansetron in preventing PONV in neurosurgical
patients (17). Our data demonstrate that ondansetron plus
dexamethasone prevents PONV better than dexamethasone alone. Although ondansetron ODT decreased the
need for rescue treatment with metoclopramide, once metoclopramide was administered nausea intensity and frequency of emesis were similar between groups.
The authors thank the nurses of the Johns Hopkins Neuro-sciences
Critical Care Unit for their help in facilitating data collection and the
pharmacists of the Johns Hopkins Investigational Pharmacy for
their dedication to the success of our investigation.

References
1. Fabling JM, Gan TJ, Guy J, et al. Postoperative nausea and
vomiting: a retrospective analysis in patients undergoing elective craniotomy. J Neurosurg Anesthesiol 1997;9:308 12.
2. Fabling JM, Gan TJ, El Moalem HE, et al. A randomized, doubleblind comparison of ondansetron versus placebo for prevention
of nausea and vomiting after infratentorial craniotomy. J Neurosurg Anesthesiol 2002;14:1027.

1496

NEUROSURGICAL ANESTHESIA
PONV AFTER ZOFRAN ODT

HARTSELL ET AL.

3. Davidson N, Rapoport B, Erikstein B, et al. Comparison of an

orally disintegrating ondansetron tablet with the conventional
ondansetron tablet for cyclophosphamide-induced emesis in
cancer patients: a multicenter, double-masked study. Ondansetron Orally Disintegrating Tablet Emesis Study Group. Clin
Ther 1999;21:492502.
4. Gan TJ, Franiak R, Reeves J. Ondansetron orally disintegrating
tablet versus placebo for the prevention of postdischarge nausea
and vomiting after ambulatory surgery. Anesth Analg 2002;94:
1199 200.
5. Jellish WS, Thalji Z, Fluter E, Leonetti JP. Ondansetron versus
droperidol or placebo when given prophylactically for the prevention of postoperative nausea and vomiting in patients undergoing middle ear procedures. J Clin Anesth 1997;9:451 6.
6. Honkavaara P. Effect of ondansetron on nausea and vomiting
after middle ear surgery during general anaesthesia. Br J Anaesth 1996;76:316 8.
7. Furst SR, Sullivan LJ, Soriano SG, et al. Effects of ondansetron
on emesis in the first 24 hours after craniotomy in children.
Anesth Analg 1996;83:325 8.
8. Irefin SA, Schubert A, Bloomfield EL, et al. The effect of craniotomy location on postoperative pain and nausea. J Anesth
2003;17:22731.
9. Manninen PH, Tan TK. Postoperative nausea and vomiting after
craniotomy for tumor surgery: a comparison between awake
craniotomy and general anesthesia. J Clin Anesth 2002;14:
279 83.
10. Herdman SJ, Clendaniel RA, Mattox DE, et al. Vestibular adaptation exercises and recovery: acute stage after acoustic neuroma resection. Otolaryngol Head Neck Surg 1995;113:77 87.

ANESTH ANALG
2005;101:14926

11. Liu YH, Li MJ, Wang PC, et al. Use of dexamethasone on the
prophylaxis of nausea and vomiting after tympanomastoid surgery. Laryngoscope 2001;111:1271 4.
12. Wattwil M, Thorn SE, Lovqvist A, et al. Dexamethasone is as
effective as ondansetron for the prevention of postoperative
nausea and vomiting following breast surgery. Acta Anaesthesiol Scand 2003;47:8237.
13. Henzi I, Walder B, Tramer MR. Dexamethasone for the prevention of postoperative nausea and vomiting: a quantitative systematic review. Anesth Analg 2000;90:186 94.
14. Eberhart LH, Mauch M, Morin AM, et al. Impact of a multimodal anti-emetic prophylaxis on patient satisfaction in high-risk
patients for postoperative nausea and vomiting. Anaesthesia
2002;57:10227.
15. Gan TJ, Meyer T, Apfel CC, et al. Consensus guidelines for
managing postoperative nausea and vomiting. Anesth Analg
2003;97:6271.
16. Kovac AL, OConnor TA, Pearman MH, et al. Efficacy of repeat
intravenous dosing of ondansetron in controlling postoperative
nausea and vomiting: a randomized, double-blind, placebocontrolled multicenter trial. J Clin Anesth 1999;11:4539.
17. Pugh SC, Jones NC, Barsoum LZ. A comparison of prophylactic
ondansetron and metoclopramide administration in patients
undergoing major neurosurgical procedures. Anaesthesia 1996;
51:1162 4.

CASE REPORTS

Epidural Blood Patch for Headache After Lumboperitoneal

Shunt Placement
Leonard R. Allmond, MD*, Greg Stratmann,
and Daniel H. Burkhardt, MD*

MD, PhD*,

Sandeep M. Kunwar,

MD,

Department of *Anesthesia and Perioperative Care and Neurological Surgery, The University of California at San
Francisco

Headaches complicating lumboperitoneal (LP) shunt

placement have been attributed to shunt failure with
resultant high intracranial pressure or to overdrainage
with resultant low intracranial pressure. In this case, a
17-yr-old girl had symptoms of a low-pressure headache after LP shunt placement alleviated by an epidural

he cause of headache after lumboperitoneal (LP)

shunt placement has been attributed to shunt
failure with resultant high intracranial pressure
or to overdrainage with resultant low intracranial
pressure (13). Headaches caused by shunt failure
(high-pressure headache), typically present in the
same way as preshunt headaches, and those caused by
shunt overdrainage (low-pressure headache) resemble
a postdural puncture headache (PDPH) having a
strong postural component (1). We present a case of
low-pressure headache after placement of an LP shunt
treated with an epidural blood patch, thus suggesting
an alternative cause for headaches after LP shunt
placement.

Case Report
A 17-yr-old girl with pseudotumor cerebri had persistent
headaches and increased cerebrospinal fluid (CSF) pressure
over a 1-mo period before admission, despite placement of a
ventriculoperitoneal shunt 1 yr before. Her medical history
was significant for hypothyroidism, for which she took levothyroxine. Her physical examination and laboratory values
were unremarkable. There were no imaging studies within
the year before admission. Computed tomography scan of
the head performed 1 yr before admission was unremarkable. The patient underwent placement of an LP shunt under
Accepted for publication May 19, 2005.
Address correspondence and reprint requests to Leonard R. Allmond, MD, Department of Anesthesiology and Perioperative Care,
S-436, University of California at San Francisco, San Francisco, CA,
94143-0427. Address e-mail to l_allmond@yahoo.com.
DOI: 10.1213/01.ANE.0000181002.23135.F9
2005 by the International Anesthesia Research Society
0003-2999/05

blood patch. The success of this therapy suggests postdural puncture as a possible cause for low-pressure
headache after LP shunt placement. Epidural blood
patch may be an alternative initial therapy for some
low-pressure headaches after LP shunt placement.
(Anesth Analg 2005;101:14978)

general anesthesia at the L3-4 level using a Spetzler LP shunt

kit (Integra Neurosciences, Plainsboro, NJ). Later that day,
she complained of a postural frontal/occipital headache accompanied by nausea and vomiting. An abdominal binder
was applied, and she received Vicodin, acetaminophen,
morphine, metoclopramide, and ondansetron to treat her
symptoms. Despite these measures, the headache remained
unchanged.
Because the patients clinical presentation was indistinguishable from that of a PDPH, it was postulated that the
origin of the symptoms was CSF leakage around the LP
shunt catheter. The patient was treated conservatively with
caffeine and bed rest for 24 h without improvement in
symptoms. An epidural blood patch was performed on postoperative Day 4 at the L4-5 interspace. Under full sterile
precautions, including 1 g of cefazolin IV, the epidural space
was cannulated with an 18-gauge Tuohy needle, and 20 mL
of blood, withdrawn from an antecubital vein in a sterile
manner, was injected into the space. Shortly thereafter, the
patient was able to ambulate without any recurrence of her
headache and was discharged later that day. There were no
long-term sequelae from the blood patch.

Discussion
We present a case of low-pressure headache after LP
shunt placement successfully treated with an epidural
blood patch. The incidence of low-pressure headache
after placement of an LP shunt is 15%20%, with the
cause being attributed to shunt overdrainage (2,3).
Headache presumably results from low CSF pressure
caused by shunt overdrainage. Low-pressure headaches after LP shunt placement are usually treated by
shunt revision or replacement. In one case series, lowpressure headaches were the second most common
reason for shunt revision or replacement (2).
Anesth Analg 2005;101:14978

1497

1498

CASE REPORTS

This case suggests that PDPH is a possible cause of

low-pressure headache after LP shunt placement. The
clinical presentations of PDPH and low-pressure
headache after LP shunt placement are indistinguishable. Computed tomography myelography, retrograde radionuclide myelography, cisternography, or
thin section magnetic resonance imaging could potentially diagnose a CSF leak secondary to postdural
puncture, but these studies were not performed because we wanted to determine the therapeutic potential of an epidural blood patch (4). In this case, a
14-gauge Tuohy needle with a 1.6 mm outside diameter (OD) was used to access the subarachnoid space
with subsequent placement of a 1.5 mm OD LP shunt
catheter. The size difference between the Tuohy needle and the catheter could have allowed CSF leakage
around the catheter and PDPH. An analogous situation is the placement of intrathecal epidural catheters
after inadvertent dural puncture while performing
epidural anesthesia. Placement of catheters in the dural puncture hole decreases the incidence of, but does
not completely prevent, PDPH (5,6). Perhaps many of
the low-pressure headaches after LP shunt placement
that have been attributed to shunt overdrainage were
actually caused by CSF leakage around the catheter.
Epidural blood patches performed after wet tap
during epidural catheter placement or spinal anesthetics have a 70%98% success rate (4). Although it is
unclear to what degree a placebo effect contributes to
this success rate, it seems plausible that a blood patch
could resolve some low-pressure headaches after LP
shunt placement, given the theoretical possibility of
CSF leaking around the catheter. The cost savings of
preventing unnecessary LP shunt revisions for overdrainage could be significant, and the risk of an epidural blood patch compares favorably with the risk of
reoperation (3,7). Common complications of epidural
blood patch include back or neck pain and transient
fever. More serious, rare complications such as bleeding, infection, and arachnoiditis may also occur (7). A

ANESTH ANALG
2005;101:14978

randomized, controlled trial could determine the riskbenefit ratio of epidural blood patch for low-pressure
headache after placement of LP shunts.
Some additional aspects of the case are noteworthy.
Entrance of the epidural space at the shunt level was
avoided to prevent catheter dislodgment. Also, a
20 mL volume was chosen for the blood patch because
this volume has been reported to result in the most
frequent success (4). Finally, measures to prevent secondary infection of the LP shunt were taken, including
full sterile precautions and prophylactic antibiotics.
In summary, a case of PDPH after LP shunt placement treated successfully by an epidural blood patch
was presented. Further study is required to determine
whether epidural blood patches are a suitable alternative initial therapy for low-pressure headaches after
LP shunt placement.

References
1. Roth PA, Cohen AR. Management of hydrocephalus in infants
and children. In: Tindall GT, Cooper PR, Barrow DL, eds. The
practice of neurosurgery. Baltimore: Williams and Wilkins, 1996:
270728.
2. Eggenberger ER, Miller NR, Vitale S. Lumboperitoneal shunt for
the treatment of pseudotumor cerebri. Neurology 1996;46:
1524 30.
3. Rosenberg ML, Corbett JJ, Smith C, et al. Cerebrospinal fluid
diversion procedures in pseudotumor cerebri. Neurology 1993;
43:10712.
4. Turnbull DK, Shepherd DB. Post-dural puncture headache:
pathogenesis, prevention, and treatment. Br J Anaesth 2003;91:
718 29.
5. Ayad S, Demian Y, Narouze SN, Tetzlaff JE. Subarachnoid catheter placement after wet tap for analgesia in labor: influence on
the headache in obstetric patients. Reg Anesth Pain Med 2003;
28:5125.
6. Liu N, Montefiore A, Kermarec N, et al. Prolonged placement of
spinal catheters does not prevent postdural puncture headache.
Reg Anesth 1993;18:110 3.
7. Candido KD, Stevens RA. Post-dural puncture headache: pathophysiology, prevention and treatment. Best Pract Res Clin Anaesthesiol 2003;17:451 69.

Postural Headache in the Presence of Cerebral Venous

Sinus Thrombosis
Ludmil Todorov,

MD,

Charles E. Laurito,

MD,

and David E. Schwartz,

MD

Department of Anesthesiology, University of Illinois at Chicago

Cerebral venous sinus thrombosis (CVST) can present

with a headache similar to that after a dural puncture.
We report on a patient who developed postural headache
after epidural anesthesia for delivery. The headache became more intense during the following 6 days, and the
patient had a tonic clonic seizure. A magnetic resonance
angiogram demonstrated CSVT, and anticoagulation

erebral venous sinus thrombosis (CVST) can

present with a headache similar to that following a dural puncture. There have only been few
reports of CVST occurring concomitantly with postdural puncture headache (PDPH). We present such a
case.

Case Report
A previously healthy 23-yr-old woman presented to the
emergency room with postdural headache. A week before
her admission, she had had an uneventful vaginal delivery
under epidural anesthesia. A few hours postpartum, the
patient developed an occipital headache, which was provoked by sitting or standing and was relieved by lying
down. The patient also had occasional blurry vision and
flashes of light. The pain was partially relieved with nonsteroidal antiinflammatory drugs and Tylenol No. 3 (acetaminophen 300 mg and codeine 30 mg), and the patient was sent
home 48 h after delivery.
Over the next 4 days, the pain became more intense and
less responsive to oral analgesics. When she presented to the
emergency room, she was afebrile, normotensive, and without neurologic deficits. The anesthesiologist on call was
contacted, and the decision was made to perform an epidural blood patch. However, just before the procedure, the
patient had a tonic clonic seizure. Lorazepam and phenytoin
were administered. The patient was drowsy after the
seizure but had no neurologic deficits. A cranial, noncontrast computed tomography scan revealed no acute bleeding. A lumbar puncture was performed using a 20-gauge
Accepted for publication May 19, 2005.
Address correspondence and reprint requests to Ludmil Todorov,
MD, Department of Anesthesiology, 1740 West Taylor St., suite
3200, Chicago, IL 60612. Address e-mail to ludmil@rocketmail.com.
DOI: 10.1213/01.ANE.0000181003.37968.CB
2005 by the International Anesthesia Research Society
0003-2999/05

therapy was started, with resolution of the symptoms

over 2 wk. Any postdural-puncture headache that loses its
positional character, becomes persistent, or does not improve with a properly performed blood patch should
raise the suspicion of CVST.
(Anesth Analg 2005;101:1499 500)

needle. The cerebrospinal fluid (CSF) showed no infective

agents. Her arterial blood pressure was increased to 150/
100 mm Hg, and magnesium and antihypertensives were
started for possible postpartum eclampsia. The patient
was admitted to the intensive care unit. On the next day,
the patients headache had worsened and was not completely relieved by lying down. Because of concern for
CVST, magnetic resonance imaging (MRI) was performed,
which revealed areas of hypointensity in the superior
sagittal sinus and edema in the surrounding parenchyma
compatible with venous thrombosis. The patient was anticoagulated with heparin and transferred to our institution for neurosurgical observation. A hypercoagulation
workup revealed protein S deficiency.
Over the next 2 days, the patient continued to have postural headaches with dizziness, with no relief from IV caffeine or oral analgesics. The pain service was consulted at
this point and asked to perform an epidural blood patch.
However, as the patient was already anticoagulated and had
no neurologic deficits, the decision was made not to perform
the procedure at that time.
Over the next 2 days, the severity of the headache decreased, and it became nonpostural. Its distribution was
mostly frontal and less occipital. Coumadin therapy was
initiated, and the heparin infusion was tapered once the goal
International Normalized Ratio (INR) of 2-3 was reached.
The patient was discharged from the hospital 10 days after
her readmission. A week later, her headache completely
disappeared. She has now been headache-free and asymptomatic for more than 5 months.

Discussion
CVST is an uncommon complication of pregnancy
with an incidence of between 1:3000 (1) and 1:10,000
(2). Factors that predispose to this condition include
the hypercoagulable state of pregnancy and hereditary conditions, including factor V Leiden mutation,
Anesth Analg 2005;101:1499500

1499

1500

CASE REPORTS

deficiencies of protein C, protein S, and antithrombin

III (3). Sinus thromboses related to pregnancy usually
occur from the third trimester to four weeks postpartum (1). The main symptoms include headache, seizures, impaired consciousness, nausea, and vomiting
(3). The diagnosis can be made by MRI. Prompt systemic anticoagulation is the mainstay of medical management and leads to good outcomes in most patients
who present with intact sensorium. Patients with persistent CVST and neurologic deterioration despite optimal medical management are considered candidates
for endovascular therapy (infusion of thrombolytics,
mechanical thrombectomy, and angioplasty with
stenting) (3).
There have been very few reports of CVST occurring concomitantly with PDPH (4 9). Because CVST
can present with postural headache, the symptoms
can often be confused with those of PDPH (10). In
most cases, the patients received epidural blood
patches and only after the failure to treat the headache
were further diagnostic steps taken, which revealed
the presence of CVST. Our case is unique in that we
were faced with three possible causes for the patients
headache. The initial headache was probably caused
by an unrecognized dural puncture; however, the
worsening headache and change in pain character at
the time of readmission were likely produced by the
venous thrombosis. This is made more plausible by
the fact that these changes occurred around the time
when the patient had a seizure. It is also important to
consider the diagnostic lumbar puncture as a contributing factor to the patients headache at the time of
evaluation by the pain service, because a 20-gauge
needle can cause a significant CSF leak. Some authors
have suggested that the intracranial hypotension resulting from a dural leak might predispose to CVST,
especially in patients with hereditary prothrombotic
conditions, because the hypotension leads to venous
dilation and blood stasis (4,6). Whether a blood patch
can prevent this complication is unknown.
We were faced with the dilemma of whether to
continue the anticoagulation or stop it to perform an
epidural blood patch. A prolonged CSF leak might be
associated with persistent intracranial hypotension
that can lead to intracranial hemorrhage (10). There
have been few reports of subdural hematoma after
neuraxial blocks (11,12). However, our patient did not

ANESTH ANALG
2005;101:1499 500

have any focal neurologic deficits at the time of consultation. An epidural blood patch would have required reversing the anticoagulation in the face of
CVST, a condition associated with a 6%18% mortality
(13). These factors resulted in the decision to continue
conservative management.
In conclusion, we present the case of a patient with
postural headache after epidural anesthesia who developed CVST. CVST can mimic PDPH and should
always be considered in the differential diagnosis,
especially if the pain changes from positional to nonpositional, indicating increased intracranial pressure.
The timely institution of anticoagulant therapy might
prevent neurological deterioration. Any PDPH that
loses its positional character, becomes persistent, or
does not improve with a properly performed blood
patch should raise the suspicion of CVST.

References
1. Nazziola E, Elkind MS. Dural sinus thrombosis presenting three
months postpartum. Ann Emerg Med 2003;42:5925.
2. Bansal BC, Gupta RR, Prakash C. Stroke during pregnancy and
puerperium in young females below the age of 40 years as a
result of cerebral venous sinus thrombosis. Jpn Heart J 1980;21:
171 83.
3. Benveniste RJ, Patel AB, Post KD. Management of cerebral
venous sinus thrombosis. Neurosurg Q 2004;14:2735.
4. Wilder-Smith E, Kothbauer-Margreiter I, Lammle B, et al. Dural
puncture and activated protein C resistance: risk factors for
cerebral venous sinus thrombosis. J Neurol Neurosurg Psychiatry 1997;63:351 6.
5. Chisholm ME, Campbell DC. Postpartum postural headache
due to superior sagittal sinus thrombosis mistaken for spontaneous intracranial hypotension. Can J Anaesth 2001;48:302 4.
6. Aidi S, Chaunu MP, Biousse V, Bousser MG. Changing pattern
of headache pointing to cerebral venous thrombosis after lumbar puncture and intravenous high-dose corticosteroids. Headache 1999;39:559 64.
7. Schou J, Scherb M. Postoperative sagittal sinus thrombosis after
spinal anesthesia. Anesth Analg 1986;65:5412.
8. Benzon HT, Iqbal M, Tallman MS, et al. Superior sagittal sinus
thrombosis in a patient with postdural puncture headache. Reg
Anesth Pain Med 2003;28:64 7.
9. Borum SE, Naul LG, McLeskey CH. Postpartum dural venous
sinus thrombosis after postdural puncture headache and epidural blood patch. Anesthesiology 1997;86:48790.
10. Turnbull DK, Shepherd DB. Post-dural puncture headache:
pathogenesis, prevention and treatment. Br J Anaesth 2003;91:
718 29.
11. Acharya R, Chhabra SS, Ratra M, Sehgal AD. Cranial subdural
hematoma after spinal anesthesia. Br J Anaesth 2001;86:8935.
12. Kelsaka E, Sarihasan B, Baris S, Tur A. Subdural hematoma as a
late complication of spinal anesthesia. J Neurosurg Anesthesiol
2003;15:479.
13. Allroggen H, Abbott RJ. Cerebral venous sinus thrombosis.
Postgrad Med J 2000;76:125.

REGIONAL ANESTHESIA
SECTION EDITOR
TERESE T. HORLOCKER

Administration of Local Anesthetic Through the Epidural

Needle Before Catheter Insertion Improves the Quality of
Anesthesia and Reduces Catheter-Related Complications
Mehmet Cesur, Haci A. Alici, Ali F. Erdem, Fikret Silbir, and Mustafa S. Yuksek
Department of Anesthesiology and Reanimation, Ataturk University, Erzurum, Turkey

Epidural catheter placement offers flexibility in block

management. However, during epidural catheter insertion, complications such as paresthesia and venous and
subarachnoid cannulation may occur, and suboptimal
catheter placement can affect the quality of anesthesia.
We performed this prospective, randomized, doubleblind study to assess the effect of a single-injection dose
of local anesthetic (20 mL of 2% lidocaine) through the
epidural needle as a priming solution into the epidural
space before catheter insertion. We randomized 240 patients into 2 equal groups and measured the quality of
anesthesia and the incidence of complications. In the
needle group (n 100), catheters were inserted after
injection of a full dose of local anesthetic through the
needle. In the catheter group (n 98), the catheters
were inserted immediately after identification of the
epidural space. Local anesthetic was then injected via

ontinuous epidural anesthesia is popular for

many surgical, obstetric, and analgesic procedures because placement of a catheter offers
flexibility to extend, intensify, and maintain block.
However, during epidural catheter insertion, complications such as paresthesia and inadvertent venous
and subarachnoid cannulation may occur; these, in
turn, may lead to transient or permanent paralysis,
convulsion, and postdural-puncture headache (1).
Furthermore, suboptimal catheter placement within
the epidural space affects the spread and quality of
anesthesia (13), risking failure of the anesthetic and
requiring placement of a second epidural catheter or
the need for general anesthesia. Some studies suggest
that the incidence of complications and failures may
be reduced by injecting a priming dose of local
anesthetic or saline through the epidural needle before
Accepted for publication May 12, 2005.
Address correspondence and reprint requests to Mehmet Cesur,
Yunus Emre Mah Tomurcuk Sok, Ikizler Apt. B Blok No: 7/7, 25080
Yenisehir/Erzurum, Turkey. Address e-mail to mcesur@atauni.edu.tr.
DOI: 10.1213/01.ANE.0000181005.50958.1E
2005 by the International Anesthesia Research Society
0003-2999/05

the catheter. We noted the occurrence of paresthesia,

inability to advance the catheter, or IV or subarachnoid
catheter placement. Sensory and motor block were assessed 20 min after the injection of local anesthetic. Surgery was initiated when adequate sensory loss was confirmed. In the catheter group, the incidence of
paresthesia during catheter placement was 31.6% compared with 11% in the needle group (P 0.00038). IV
catheterization occurred in 8.2% versus 2% of patients
in the catheter and needle groups, respectively (P
0.048). More patients in the needle group had excellent
surgical conditions than the catheter group (89.6% versus 72.9; P 0.003). We conclude that giving a singleinjection dose via the epidural needle before catheter
placement improves the quality of epidural anesthesia
and reduces catheter-related complications.
(Anesth Analg 2005;101:15015)

catheter insertion (4 6). This has been disputed (7,8),

but in studies showing a lack of effect, either a small
and possibly inadequate volume of local anesthetic or
normal saline, which would dilute local anesthetic
subsequently injected, was given. The use of a large
priming dose of local anesthetic has not been studied.
The purpose of this prospective, randomized, doubleblind study was to assess the effect on anesthetic
quality and complications of single-injection of local
anesthetic through the needle as a priming solution
into the epidural space before insertion of the catheter.

Methods
After obtaining institutional ethics committee approval and informed consent, 240 ASA class I-II consecutive adult patients undergoing elective surgery
with epidural anesthesia were enrolled in this prospective, randomized, double-blind study. Patients in
whom central blocks were contraindicated and patients with spinal column disorders, including scoliosis and herniated disks, or previous spinal surgery
were excluded. In addition, obstetric patients (20 in
Anesth Analg 2005;101:15015

1501

1502

REGIONAL ANESTHESIA CESUR ET AL.

LOCAL ANESTHETIC BOLUS BEFORE EPIDURAL CATHETER INSERTION

the needle group and 22 in the catheter group) were

excluded because they differ anatomically from other
surgical patients.
Patients were assigned into two equal groups using
a computer-generated randomization list. In the needle group, the epidural catheters were advanced into
the epidural space after a single injection of a full dose
of local anesthetic (20 mL of 2% lidocaine) through the
needle. In the catheter group, the epidural catheters
were inserted without injection of any solution and
local anesthetic was then injected via the catheter. All
procedures were performed by the same experienced
anesthesiologist.
On arrival in the operating room, automatic noninvasive arterial blood pressure monitoring, electrocardiograph, and pulse oximetry were commenced, and
10 15 mL/kg of Ringers lactate solution was infused
before the procedure. Baseline demographic data and
vital signs were recorded before surgery. All patients
were premedicated with 3 mg of midazolam and
0.5 mg of atropine IM. With the patient in the left
lateral position, lumbar epidural punctures were performed at the L4-5 or L3-4 interspace using a midline
approach with 18-gauge Tuohy needles and the lossof-resistance technique with 1-2 mL of saline. In the
needle group, after identification of epidural space
and a negative aspiration test for blood or cerebrospinal fluid, 3 mL of 2% lidocaine with epinephrine 5
g/mL was injected through the needle as a test dose.
The syringe was then disconnected to observe for drip
back. The patients were also observed for any increase
in heart rate that would indicate an intravascular injection of epinephrine and were questioned about dizziness, tinnitus, a metallic taste in the mouth, or sudden warmth or numbness in the legs. If these
responses were negative after 5 min, the remainder of
the full 20 mL of local anesthetic was injected in 3 divided doses. A 20-gauge multiorifice epidural catheter
(Minipack; Portex Ltd., Kent, UK) was inserted 3 cm
into the epidural space through the cranially directed
tip of the epidural needle. After removal of the Tuohy
needle, the catheter was fixed to the skin, and the
patients were turned to the supine position. In the
catheter group, identification of the epidural space,
aspiration test, test dose, and incremental injection
were performed as before, except that no local anesthetic was injected before catheter placement.
Paresthesia during insertion of the catheter, inability
to advance the catheter, and IV and subarachnoid
cannulation were noted by the attending anesthesiologist (i.e., not a blinded independent observer). IV or
subarachnoid cannulation was detected by aspiration
of frank blood or cerebrospinal fluid through the catheter. If intravascular or subarachnoid cannulation occurred, the catheter was withdrawn 1 cm. If this did
not lead to withdrawal from the vein or subarachnoid
space, the catheter was removed. If it was not possible

ANESTH ANALG
2005;101:15015

Table 1. Patient Characteristics (mean sd)

Needle group
(n 100)

Catheter group
(n 98)

78/22
52 24
70 24
172 16
90/10

82/16
55 22
74 18
166 24
82/16

Sex (m/f)
Age (yr)
Weight (kg)
Height (cm)
ASA I/II

There were no differences between groups.

Table 2. Surgical Procedures and Durations (mean sd)

Hysterectomy
Prostatectomy
Varicocelectomy
Inguinal herniorrhaphy
Duration of surgery

Needle group
(n 100)

Catheter group
(n 98)

11
34
13
42
54 24

13
37
9
39
51 20

There were no differences between groups.

to thread the catheter, it and the needle were withdrawn together. The procedure was then repeated at a
different level; if unsuccessful again, general anesthesia was given. In the needle group, if the catheter
could not be advanced and the surgery was of short
duration, surgery was commenced under singleinjection epidural anesthesia. These patients were excluded from the analysis.
Twenty minutes after the main dose, sensory block
levels and the degree of motor block were assessed
bilaterally by a blinded independent observer. Sensory block was assessed with ice and motor block by
the Bromage scale (0 no block, 1 hip movement
block, 2 hip and knee block, and 3 complete block
in hip, knee, and ankle). Complete loss of cold sensation to T8 on both sides was regarded as sufficient for
surgery.
The term failed epidural was used for situations
in which either it was impossible to insert the catheter
or there was no sensory block after injection of the
local anesthetic (9). Unilateral block, unblocked sacral
segments, low level and unblocked segments, or a
patchy block were regarded as incomplete block
before surgery (9). If these situations were observed,
an additional 10 mL (5 mL 5 mL) of anesthetic
solution was administered in both groups. If they
persisted despite the additional dose, they were accepted as persistent incomplete block before surgery,
and general anesthesia was administered. Preoperative bilateral complete loss of cold sensation to T8 and
the absence of a patient complaining of discomfort
during surgery was defined as excellent surgical conditions. In patients complaining of discomfort, if the
additional dose had not previously been administered, it was now given. Patients complaining of discomfort despite the additional injection already given

ANESTH ANALG
2005;101:15015

REGIONAL ANESTHESIA
CESUR ET AL.
LOCAL ANESTHETIC BOLUS BEFORE EPIDURAL CATHETER INSERTION

1503

Table 3. Characteristics of Epidural Block (median [range]) or mean sd)

Lumbar puncture interspace L45/L34

Median peak dermatomal level at 20 min
Motor block at 20 min (Bromage)
Lidocaine (mg)
Ephedrin (mg)
Perioperative fluid infusion (mL)

Needle group
(n 100)

Catheter group
(n 96)a

70/30
T10 (T115)
2 (03)
450 50
24 12
976 122

67/29
T9 (T124)
2 (23)
435 46
28 10
980 136

a
Two patients in the catheter group, in whom repeated IV catheterizations occurred and underwent a general anesthesia, were not included. There were no
differences between groups.

were asked if they would like sedation. This was given

as propofol and fentanyl (bolus induction dose
0.5 mg/kg of propofol and 1 g/kg of fentanyl IV
followed by propofol 12 mL/min and 1 g/kg of
fentanyl bolus every half hour). If complaints persisted despite this sedation, they were accepted as
inadequate anesthesia, and general anesthesia was
given.
Arterial blood pressure, heart rate, and oxygen saturation were measured and recorded every 5 min for
the duration of the surgical procedure. Hypotension
(systolic blood pressure 70% of baseline), bradycardia, (heart rate 50 bpm), and desaturation (Spo2
90%) were recorded. Hypotension was treated with
IV ephedrine 515 mg and bradycardia with 0.5 mg of
IV atropine; desaturation was treated with oxygen via
a face mask. The type and duration of surgical procedures, and amount of perioperative IV fluid given
were documented.
The primary outcome was excellent surgical condition, and a 15% difference in the incidence of the
excellent surgical condition was considered to be clinically important. According to a priori power analysis,
114 patients were sufficient to provide 90% power to
detect this difference between groups, accepting a
two-tailed () error of 5%. However, obstetric patients
(20 in the needle group and 22 in the catheter group)
were excluded because they differ anatomically from
other surgical patients. A post hoc power analysis was
performed with respect to the observed difference of
89.6%72.9%, with the sample size of 96 in each group.
The power was calculated as 84% with a two-tailed ()
error of 5%. Catheter-related complications, including
paresthesia during catheterization, inability to thread
the catheter, and inadvertent intrathecal and IV catheterization were secondary outcomes.
Statistical analysis was performed by SPSS for Windows (version 10.0) statistical package (SPSS Inc., Chicago, IL). Patient characteristics were analyzed using
the t-test for independent groups. Block height was
compared using Wilcoxon rank sum test. Perioperative anesthesia quality and incidences of catheterrelated complications were analyzed using the 2 test
in a 2 2 contingency table or Fisher exact test in 2
2 contingency table. Values are presented as numbers

Table 4. Incidence of Perioperative Complications

Hypotension
Bradycardia
Nausea
Vomiting

Needle group
(n 96)a

Catheter group
(n 85)a

18 (18.8)
2 (2.1)
10 (10.4)
1 (1.0)

15 (17.7)
1 (1.2)
8 (9.4)
1 (1.2)

Values are n (%).

a
Four patients in the needle group and 11 patients in the catheter group,
who experienced discomfort and required sedation or general anesthesia,
were not considered. There were no differences between groups.

(%), mean sd, or median (range). A P value 0.05

was considered statistically significant.

Results
There were no significant differences in demographic
or surgical data, epidural block characteristics, or incidence of perioperative complications between the
groups (Table 1, Table 2, Table 3, and Table 4). There
were no failed or incomplete blocks. The incidence of
catheter-related complications is shown in Table 5.
During catheter placement, the incidences of paresthesia and IV catheterization were more frequent in the
catheter group: 31 (31.6%) versus 11 (11%) (P
0.00038) and 8 (8.2%) versus 2 (2%) (P 0.048),
respectively.
Significantly more patients required catheter removal because of IV or subarachnoid cannulation or
inability to advance the catheter in the catheter group
(13 [13.3%] versus 4 [4%]; P 0.02). In 13 patients in
the catheter group, catheter insertion was attempted
through another space. In two of these, IV placement
was again detected, and general anesthesia was
instituted.
Anesthesia quality is shown in Table 6. Excellent
surgical conditions were more frequently encountered
in the needle group (86 [89.6%] versus 70 [72.9%]; P
0.003). The catheters were reinjected during the surgery as required. There was no difference between the
groups in the number of patients who required reinjection through the catheter as demonstrated in Table
6. Despite the additional injections, 4 (4.2%) patients in

1504

REGIONAL ANESTHESIA CESUR ET AL.

LOCAL ANESTHETIC BOLUS BEFORE EPIDURAL CATHETER INSERTION

ANESTH ANALG
2005;101:15015

Table 5. Incidence of Preoperative Catheter-Related Complications

Paresthesia during catheter insertion

Intrathecal catheterization
Intravenous catheterization
Inability to advance the catheter

Needle group
(n 100)

Catheter group
(n 98)

P-value

11 (11)
1 (1)
2 (2)
1 (1)

31 (31.6)
3 (3.1)
8 (8.2)a
2 (2.0)

0.00038
NS
0.048
NS

Values are n (%).

NS not significant (P 0.05).
a
Repeated IV catheterizations were observed in two patients in the catheter group, and each were analyzed as a single event.

Table 6. Anesthesia Quality During Operation

Needle group
(n 96)a

Catheter group
(n 96)b

P-value

86 (89.6)
4 (4.2)
15 (15.6)
4 (4.2)

70 (72.9)
14 (14.6)
18 (18.8)
11 (11.5)

0.003
0.013
NS
NS

Excellent surgical conditions

Discomfort, but intervention not necessary
Required reinjection through catheter
Inadequate anesthesia despite reinjection

Values are n (%).

NS Not significant (P 0.05).
a
4 patients in the needle group, in whom a catheter could not be placed into the epidural space, and b 2 patients in the catheter group, in whom repeated IV
catheterizations occurred and underwent to general anesthesia, were not included in the analysis.

the needle group and 11 (11.5%) in the catheter group

complained of discomfort. Therefore, they were accepted as having inadequate anesthesia and were either sedated or underwent general anesthesia.

Discussion
We have demonstrated improved surgical conditions
with the administration of a single-injection dose
through an epidural needle before epidural catheter
placement. Also, the single-injection administration
followed by catheter insertion was associated with
fewer paresthesias during insertion and fewer IV catheterizations. In addition, fewer anesthetic interventions were required.
Paresthesia during epidural catheter insertion has
been reported in up to 60% of parturients (10), and the
frequency of venous and subarachnoid cannulation
has been reported between 0.2% and 11% and between
0.26% (11) and 0.6% (12), respectively. Paresthesia
may be associated with transient or permanent neurological injury (13) and may be unpleasant for the
patient. Unnoticed venous and subarachnoid cannulation may lead to convulsions, total spinal anesthesia,
or postdural puncture headache.
Expansion of the epidural space by priming it with
local anesthetic before advancement of the catheter
may reduce the likelihood of both paresthesia and
inadvertent venous or subarachnoid cannulation
(14,15). Rolbin et al. (7) and Scott and Beilby (8) reported no advantage in injecting fluid into the epidural space before catheter insertion, but they administered much smaller volumes of fluid (3 and 5 mL,

respectively) for priming. However, Mannion et al. (4),

Tseng et al. (5), and Gadalla et al. (6) all noted a
significant reduction in the incidence of extradural
vein cannulation by routinely injecting 10 mL of saline
priming fluid into the epidural space before catheter
insertion. Saline, however, dilutes the local anesthetic
injected; in this study, we therefore administered a
single-injection dose of local anesthetic (20 mL) as a
priming solution.
Despite a correct technique, some segments may
remain unblocked because of inadequate spread of
local anesthetic within the epidural space. This may be
related to variations in epidural anatomy (16), although a transforaminal or anterior catheter positioning is a more likely explanation (2,7,17). Suboptimal
positioning of the epidural catheter is common. Using
radiography, Sanchez et al. (18) and Bridenbaugh et al.
(19) showed that the intended catheter placement was
often not achieved. Lim et al. (20) found that the
catheter tip could be advanced without coiling for 4
cm or less in only 13% of cases. Hogan (21) found that
lateral catheter deviation is a more common cause of
asymmetric block than anatomic barriers to the spread
of the local anesthetic solution. When epidural anesthesia is incomplete, additional injections or catheter
manipulation may provide reliable surgical anesthesia, suggesting suboptimal positioning of the catheter.
Both the type of catheter (22) and its optimal depth
of insertion (3,23) have been questioned. We used a
multi-port epidural catheter inserted only 3 cm in the
epidural space; these catheters give better anesthesia
and require less manipulation than uniport ones (22),
and insertion to no more than 3 4 cm into the epidural
space minimizes complications and the incidence of

ANESTH ANALG
2005;101:15015

REGIONAL ANESTHESIA
CESUR ET AL.
LOCAL ANESTHETIC BOLUS BEFORE EPIDURAL CATHETER INSERTION

inadequate anesthesia, even in obstetric patients (17).

Our single-injection dose via epidural needle before
catheter placement led to fewer cases of catheter replacement and inadequate anesthesia. Although no
imaging method was used in this study, in view of the
possibility of the suboptimal positioning and the malfunction of the epidural catheter, as mentioned above,
injection through the epidural needle resulted in a
more even distribution of local anesthetic solution and
a more adequate anesthetic action.
Thus, this single-injection administration before catheter insertion offers the advantages of a single injection
technique plus the flexibility of epidural catheterization.
The requirement of relatively large volumes of local
anesthetic as priming solution in the single-injection/
catheter technique may be a disadvantage, and the direct
catheter technique is preferable if it is essential to restrict
dose and level block in special patients.
In summary, we report that the administration of
local anesthetics through the epidural needle before
epidural catheter placement improves the quality of
epidural anesthesia and decreases the risk of catheterrelated complications.

References
1. Cousins JC, Veering BT. Epidural neural blockade. In: Cousins
MJ, Bridenbaugh PO, eds. Neural blockade in clinic anesthesia
and management of pain. 3rd ed. Philadelphia: LippincottRaven, 1998:243321.
2. Doughty A. A precise method of cannulating the lumbar epidural space. Anaesthesia 1974;29:635.
3. Asato F, Goto F. Radiographic findings of unilateral epidural
block. Anesth Analg 1996;83:519 22.
4. Mannion D, Walker R, Clayton K. Extradural vein puncture: an
avoidable complication. Anaesthesia 1991;46:5857.
5. Tseng CH, Li AH, Kuo-Sheng H, et al. Prior epidural injection of
10 ml normal saline reduces the incidence of inadvertent venous
puncture in epidural catheterization. Acta Anaesthesiol Sin
1995;33:2730.
6. Gadalla F, Lee SH, Choi KC, et al. Injecting saline through the
epidural needle decreases the IV epidural catheter placement
rate during combined spinal-epidural labour analgesia. Can J
Anaesth 2003;50:3825.

1505

7. Rolbin SH, Halpern SH, Braude BM. Fluid through the epidural
needle does not reduce complications of epidural catheter insertion. Can J Anaesth 1990;37:337 40.
8. Scott DA, Beilby DS. Epidural catheter insertion: the effect of
saline prior to threading in non-obstetric patients. Anaesth Intensive Care 1993;21:284 7.
9. Portnoy D, Vadhera R. Mechanisms and management of an
incomplete epidural block for cesarean section. Anesthesiol Clin
North America 2003;21:39 57.
10. Sarna MC, Smith I, James JM. Paraesthesia with lumbar epidural catheters: a comparison of air and saline in a loss-ofresistance technique. Anaesthesia 1990;45:10779.
11. Richardson MG, Lee AC, Wissler RN. High spinal anesthesia
after epidural test dose administration in five obstetric patients.
Reg Anesth 1996;21:119 23.
12. Carr MF, Hehre FW. Inadvertent lumbar puncture. Anesth
Analg 1962;41:349 53.
13. Gerancher JC, Liu SS. Complications of neuraxial (spinal/
epidural/caudal) anesthesia. In: Benumof JL, Saidmen LJ, eds.
Anesthesia & perioperative complications. 2nd ed. Philadelphia:
Mosby, 1999:50 65.
14. Brown DL. Spinal, epidural and caudal anesthesia. In: Miller
RD, ed. Anesthesia. 5th ed. Philadelphia: Churchill Livingstone,
2000:1491519.
15. Verniquet AJW. Vessel puncture with epidural catheters. Anaesthesia 1980;35:660 2.
16. Usubiaga JE, Dos Reis A, Usubiaga LE. Epidural misplacement
of catheters and mechanisms of unilateral blockade. Anesthesiology 1970;32:158 61.
17. Beilin Y, Bernstein HH, Zucker-Pinchoff B. The optimal distance
that a multiorifice epidural catheter should be threaded into the
epidural space. Anesth Analg 1995;81:301 4.
18. Sanchez R, Acuna L, Rocha F. An analysis of the radiological
visualization of the catheters placed in the epidural space. Br J
Anaesth 1967;39:4859.
19. Bridenbaugh LD, Moore DC, Bagdi P, Bridenbaugh PO. The
position of plastic tubing in continuous-block techniques: an
x-ray study of 552 patients. Anesthesiology 1967;29:10479.
20. Lim YJ, Bahk JH, Ahn WS, Lee SC. Coiling of lumbar epidural
catheters. Acta Anaesthesiol Scand 2002;46:603 6.
21. Hogan Q. Epidural catheter tip position and distribution of
injectate evaluated by computed tomography. Anesthesiology
1999;90:964 70.
22. Michael S, Richmond MN, Birks RJS. A comparison between
open-end (single hole) and closed-end (three lateral holes) epidural catheters. Anaesthesia 1989;44:578 80.
23. DAngelo R, Berkebile BL, Gerancher JC. Prospective examination of epidural catheter insertion. Anesthesiology 1996;84:
88 93.

The Efficacy of Plethysmographic Pulse Wave Amplitude as

an Indicator for Intravascular Injection of EpinephrineContaining Epidural Test Dose in Anesthetized Adults
Hany A. Mowafi, MB, Bch, MSc, MD
Department of Anesthesia, Faculty of Medicine, King Faisal University, Saudi Arabia

In this study, I evaluated the efficacy of plethysmographic pulse wave amplitude (PPWA) in detecting intravascular injection of a simulated epidural test dose
containing 15 g of epinephrine in adults during either
sevoflurane or isoflurane inhaled anesthesia and compared its reliability to the classical heart rate (HR; positive if 10 bpm) and systolic blood pressure (SBP; positive if 15 mm Hg) criteria. Eighty patients were
randomized to receive either 1 mean alveolar anesthetic
concentration of sevoflurane or 1 mean alveolar anesthetic concentration of isoflurane (n 40 for each anesthesia group). Patients in each anesthesia group were
further randomized to receive either 3 mL of 1.5% lidocaine containing 15 g of epinephrine IV or 3 mL of
saline IV (n 20 each). HR, SBP, and PPWA were monitored for 5 min after injection. Injection of the test dose

ombined epidural and general anesthesia is a popular technique for providing anesthesia and postoperative analgesia to patients undergoing surgery (1,2). To avoid the potentially life-threatening
cardiovascular and central nervous toxicity associated
with intravascular injection of large amounts of local
anesthetic solutions, an epidural test dose containing 15
g of epinephrine is used (35). However, the hemodynamic criteria of positive intravascular injection of the
epinephrine-containing test dose may be unreliable during anesthesia (6 8). This is mainly because of the decreased heart rate (HR) response to epinephrine in anesthetized patients and the need for invasive monitoring to
detect the peak increases in systolic blood pressure (SBP).
Although a previous study demonstrated that digital skin blood flow, as measured by a laser Doppler
flowmeter, is a reliable marker for an intravascular
Accepted for publication May 12, 2005.
Address correspondence and reprint requests to Hany A. Mowafi,
Department of Anesthesia, King Fahd University Hospital, PO Box
40081, Al-Khobar 31952, Saudi Arabia. Address e-mail to hany_
mowafi@hotmail.com.
DOI: 10.1213/01.ANE.0000181004.72325.D6

1506

Anesth Analg 2005;101:150611

resulted in peak PPWA decrease by 61% 17% and

58% 15% at 61 12 s and 63 13 s in the sevoflurane
and isoflurane groups, respectively. Positive PPWA criterion, as determined from peak increases during saline
administration, was a decrease in PPWA 10%. Using
this value, the sensitivity, specificity, positive predictive, and negative predictive values of PPWA were
100% in both anesthetic groups. On the contrary, sensitivities of 85% and 95% were obtained based on HR criterion in the sevoflurane and isoflurane patients, respectively, and a sensitivity of 90% was obtained in
both anesthesia groups on the basis of SBP criterion. In
conclusion, PPWA is a reliable alternative to conventional hemodynamic criteria for detection of an intravascular injection of epidural test dose.
(Anesth Analg 2005;101:1506 11)

test-dose injection during anesthesia (9), the cost and

availability of laser Doppler flow meters may limit the
routine use of this diagnostic tool. However, plethysmographic pulse wave monitoring is readily available
for routine use by anesthesiologists. Its amplitude indicates blood volume changes in the fingertips (10)
and, thus, may be used for detection of the intravascular test-dose injection. Additionally, because inhaled anesthetics differ in their depressant actions and
hemodynamic responses to catecholamines, the efficacy of the epinephrine-containing test dose may vary
under the influence of different anesthetics (11).
The aim of this study was to investigate whether
changes in plethysmographic pulse wave amplitude
(PPWA) can be used as a new criterion for detecting
intravascular injection of an epinephrine-containing
test dose in adults under stable sevoflurane or isoflurane anesthesia and to compare its reliability to the
classical hemodynamic criteria.

Methods
After local research committee approval and informed
patient consent, 80 ASA physical status I or II patients,
2005 by the International Anesthesia Research Society
0003-2999/05

ANESTH ANALG
2005;101:1506 11

REGIONAL ANESTHESIA
MOWAFI
PLETHYSMOGRAPHIC PULSE WAVE AND EPIDURAL TEST DOSE

1507

Figure 1. Example of the data collected by Datex-Ohmeda S/5 collect software for one case during sevoflurane anesthesia after an injection
of the simulated test dose. The waveforms are for electrocardiogram (ECG) and plethysmographic (pleth) pulse wave. The trends are for the
heart rate (HR), noninvasive systolic blood pressure (SBP), end-tidal (Et) sevoflurane, and plethysmographic pulse wave amplitude (PPWA),
which is presented using integer arithmetic so that the reading of 100 is 1% of the PPWA.

aged 18 50 yr old, scheduled to undergo general anesthesia for elective surgery, were included in the
study. Exclusion criteria included a history of smoking, diabetes mellitus, cardiovascular diseases, or use
of medications affecting the cardiovascular system.
Patients were randomly allocated using an online research randomizer (http://www.randomizer.org)
into 2 equal groups (40 patients each) to receive either
1 mean alveolar anesthetic concentration (MAC) of
sevoflurane or 1 MAC of isoflurane.
Patients were premedicated with 10 mg of diazepam orally 90 min before surgery. Electrocardiographic HR, noninvasive oscillometric SBP, and
PPWA were measured using an S/5 anesthesia monitor (Datex-Ohmeda, Helsinki, Finland), and the data
were collected using Datex-Ohmeda S/5 collect software every 10 s. The S/5 collect reads data through the
monitor serial port connected to a notebook personal
computer. The oximeter probe used to monitor the
plethysmographic pulse wave was attached to the
middle fingertip of the hand contralateral to the site of
SBP monitoring and was wrapped in a towel to minimize heat loss and contamination with ambient light.
The plethysmographic pulse wave displayed on the
monitor is derived from the infrared signal of the
sensor. The display scale or the amplitude factor is
automatically detected and set to optimum when the
probe is first connected to the patient. From that moment, the gain is fixed, and any changes in the amplitude reflect the pulsatile blood flow in the finger. The
PPWA is measured in percentage from the light that
passed the tissue at a certain heart rhythm. The formula used is:

PPWA I/Imax * 100%

where (Imax) is the maximum intensity of light transmitted to the receiving diode after absorption by tissues and nonpulsatile blood in the finger and (I) is
the difference between maximum and minimum
transmitted light caused by the additional light absorbed by the pulsatile blood.
Anesthesia was induced with fentanyl 2 g/kg and
propofol 2.5 mg/kg IV. Tracheal intubation was facilitated with vecuronium 0.1 mg/kg IV. Anesthesia was
maintained with a stable 1 MAC end-tidal concentration of sevoflurane or isoflurane according to the assignment groups in addition to 60% nitrous oxide in
oxygen. The lungs of patients were mechanically ventilated, and minute volume was set to maintain endtidal CO2 at 4 4.7 kPa. End-tidal gas concentrations
were measured with S/5 compact airway module
M-CAiOVX attached to the S/5 anesthesia monitor
(Datex-Ohmeda). Fluid administration was standardized to 10 mL kg1 h1 of Ringers lactate solution,
and the ambient temperature was maintained at
25C26C.
When hemodynamic variables, PPWA, and endtidal concentrations were stable for 5 min and at least
10 min had elapsed after the anesthetic induction, each
group of patients was further randomized to receive
either 3 mL of isotonic saline (n 20) or 3 mL of 1.5%
lidocaine containing 15 g of epinephrine IV (n 20)
as a simulated test dose via a peripheral IV catheter for
3 s flushed with 10 mL of saline. After the injection,
blood pressure cycling was set to every minute for
5 min. S/5 collect software (Datex-Ohmeda) was used
to collect HR, SBP, PPWA, Spo2, and end-tidal concentrations every 10 s (Fig. 1). Collected data were
later analyzed at 20-s intervals for HR and PPWA and
at 1-min intervals for SBP. In addition, maximal HR,
SBP, and PPWA responses were noted. Anesthesia

1508

REGIONAL ANESTHESIA MOWAFI

PLETHYSMOGRAPHIC PULSE WAVE AND EPIDURAL TEST DOSE

ANESTH ANALG
2005;101:1506 11

Table 1. Patient Characteristics

Sevoflurane group
Saline (n 20)
Age (yr)
Height (cm)
Weight (kg)
Sex (men/women)
Preinduction
SBP (mm Hg)
HR (bpm)
PPWA (I/Imax 100)
Preinjection
SBP (mm Hg)
HR (bpm)
PPWA (I/Imax 100)

Test dose
(n 20)

Isoflurane group
Saline (n 20)

Test dose
(n 20)

37 10
165 12
71 19
14/6

33 8
163 8
72 15
13/7

33 7
163 11
77 18
12/8

32 9
162 7
77 20
11/9

127 14
85 13
1.4 0.7

124 9
86 12
1.6 0.9

126 13
84 13
1.6 0.8

125 11
85 14
1.7 0.7

105 15*
72 9*
4.7 1.1*

101 11*
70 14*
4.8 1.2*

103 10*
75 12*
4.4 1.0*

103 11*
76 16*
5.5 2.1*

Data are mean sd or numbers.

SBP systolic blood pressure; HR heart rate; PPWA plethysmographic pulse wave amplitude; I change in light intensity received by the monitor;
Imax maximum light intensity received by the monitor.
* P 0.05 versus preinduction values.

was conducted and data were collected by the attending anesthesiologist who was blinded to the injected
test dose. All measurements were made with the patient in the supine position before surgery.
Power analysis was based on a pilot study of 10
patients (5 in each anesthesia group). More than 15
patients were required in each group during sevoflurane anesthesia, and more than 17 patients were required in each group during isoflurane anesthesia to
detect a maximum PPWA difference of 25% from the
preinjection values with a type I error of 0.05 and type
II error of 0.20. Positive HR and SBP responses to the
IV test dose were prospectively defined from previous
reports (6) as a HR increase of 10 bpm and a SBP
increase of 15 mm Hg within 2 min of administration. Ninety-five percent confidence intervals applicable to 99% of the general population (12) were calculated for PPWA after the injection of saline in each
anesthetic group. PPWA increases more than 95% tolerance limits were defined as positive criteria for detection of an intravascular injection of the test dose.
Sensitivity (true positives/[true positives false negatives]), specificity (true negatives/[true negatives
false positives]), positive predictive values (true
positives/[true positives false positives]), and negative predictive values (true negatives/[true negatives
false negatives]) were determined for HR, SBP, and
PPWA variables.
Data were tested for normal distribution using the
Kolmogorov-Smirnov test. Differences between
groups in demographic data and baseline values of
hemodynamic variables, and PPWA were analyzed
using one-way analysis of variance or 2 test as appropriate. For comparison of different observations
within and between the groups, data were first analyzed by repeated-measures analysis of variance, and
differences were then calculated by post hoc testing

(Newman-Keuls test). Fisher exact test was used to

compare sensitivities between groups. Analysis was
performed using Statistica software version 6.0 for
windows (Statsoft, Inc., Tulsa, OK). Data were presented as mean sd in the text and Table 1 and as
mean 95% confidence intervals in the figures.

Results
There were no significant differences between groups
with respect to age, weight, height, and sex distribution. There were also no significant differences in the
preinduction HR, SBP, and PPWA (Table 1). After the
induction of anesthesia and achievement of a steady
anesthetic concentration, SBP and HR decreased,
whereas PPWA increased, significantly from preinduction values. There were, however, no significant
differences between groups regarding preinjection
(baseline) data.
IV injection of the test dose produced significant
increases in HR (Fig. 2) and SBP (Fig. 3) in both
groups. Maximal increases in HR in the sevoflurane
and isoflurane groups were 14 6 bpm and 18
8 bpm at 47 14 s and 48 16 s after test-dose
injections, respectively. Maximal increases in SBP in
sevoflurane and isoflurane groups were 19 8 mm
Hg and 20 7 mm Hg at 84 30 s and 93 31 s after
test-dose injections, respectively. As shown in Figure
4, there were significant decreases in PPWA from the
preinjection value between 40 and 200 s in the sevoflurane group and between 20 and 120 s in the isoflurane
group. The average largest percent decreases in the
PPWA were 61% 17% and 58% 15% at 61 12 s
and 63 13 s after test-dose injections in the sevoflurane and isoflurane groups, respectively.
After the injection of saline, the peak percent decrease in PPWA was 2% 2% (mean sd) in both

ANESTH ANALG
2005;101:1506 11

Figure 2. Changes in heart rate (HR) after injection of the test dose
containing 15 g of epinephrine or isotonic saline during sevoflurane and isoflurane anesthesia (n 20 for each group). Vertical bars
denote 0.95 confidence intervals. *Significant difference versus preinjection values (time 0).

REGIONAL ANESTHESIA
MOWAFI
PLETHYSMOGRAPHIC PULSE WAVE AND EPIDURAL TEST DOSE

1509

Figure 4. Changes in plethysmographic pulse wave amplitude

(PPWA), expressed as percent from the preinjection value, after
injection of the test dose containing 15 g of epinephrine or isotonic
saline during sevoflurane and isoflurane anesthesia (n 20 for each
group). Vertical bars denote 0.95 confidence intervals. *Significant
difference versus preinjection values (time 0).

Table 2. Sensitivity, Specificity, Positive, and Negative

Predictive Values Based on Hemodynamic and
Plethysmographic Pulse Wave Amplitude Criteria
Sevoflurane Isoflurane
group
group

Figure 3. Changes in systolic blood pressure (SBP) after injection of

the test dose containing 15 g of epinephrine or isotonic saline
during sevoflurane and isoflurane anesthesia (n 20 for each
group). Vertical bars denote 0.95 confidence intervals. *Significant
difference versus preinjection values (time 0).

groups. Using this value, the 95% confidence interval

for 99% of the population was calculated to be from
5% to 9%, and thus, PPWA criterion for identification of intravascular injection of the test dose should
be a decrease of PPWA 10% from the preinjection
value. The sensitivity, specificity, positive predictive
value, and negative predictive value were all 100%
based on PPWA criterion. In contrast, neither the HR
nor the indirectly measured SBP changes were totally
reliable in detection of intravascular injection of test
dose (Table 2). The two-tailed Fisher exact test resulted in P values of 0.23 and 1.0 for the comparison of
the sensitivities between the PPWA and HR criteria in
the sevoflurane and isoflurane groups, respectively

By HR 10 bpm increase
Sensitivity
Specificity
Positive predictive value
Negative predictive value
By SPB 15 mm Hg
increase
Sensitivity
Specificity
Positive predictive value
Negative predictive value
By PPWA 10% increase
Sensitivity
Specificity
Positive predictive value
Negative predictive value

85 (17/20) 95 (19/20)
100 (20/20) 100 (20/20)
100 (17/17) 100 (19/19)
87 (20/23) 95 (20/21)
90 (18/20) 90 (18/20)
100 (20/20) 100 (20/20)
100 (18/18) 100 (19/19)
91 (20/22) 91 (20/22)
100 (20/20)
100 (20/20)
100 (20/20)
100 (20/20)

100 (20/20)
100 (20/20)
100 (20/20)
100 (20/20)

Data are % (n).

SBP systolic blood pressure; HR heart rate; PPWA plethysmographic pulse wave amplitude.

and a P value of 0.48 for the comparison of the sensitivities of PPWA and SBP criteria in both anesthetic
groups. Additionally, there were no statistically significant differences between the two anesthetics in
sensitivities based on all the criteria measured.

Discussion
One of the major findings in the present study was
that, using the previously recognized criteria for positive response during anesthesia, neither the HR nor

1510

REGIONAL ANESTHESIA MOWAFI

PLETHYSMOGRAPHIC PULSE WAVE AND EPIDURAL TEST DOSE

the indirectly measured SBP were 100% reliable in

detecting an intravascular injection of an epinephrinecontaining test dose during stable 1 MAC of sevoflurane or isoflurane anesthesia. The study also introduced PPWA as a new criterion for detecting
intravascular injection of the test dose and demonstrated 100% efficacy based on this criterion regardless
of the inhaled anesthetic used.
Failure to demonstrate 100% efficacy on the basis of
the conventional hemodynamic criteria indicates a
limitation of the HR and noninvasive SBP responses as
markers for detecting an intravascular test-dose injection. This is in agreement with previous reports (9,11),
which found the HR response not reliable in detecting
intravascular injection during sevoflurane anesthesia.
A previous study (13) showed that a HR increase
10 bpm was associated with a 100% sensitivity and
specificity under end-tidal isoflurane concentration
1% and that the HR responses to test-dose injection
were depressed in a concentration-dependent manner.
The reduced HR response to test-dose injection during
isoflurane anesthesia in the present study may have
been due to the larger concentration of isoflurane administered (1.2%). As with previous reports (9), the
use of intermittent and noninvasive measurement of
SBP decreased its reliability to detect intravascular
injection because the short-lived SBP increases may
easily be missed.
Photoelectric plethysmography, a technique first
described decades ago (14) and almost always routinely available on the operating room monitors (15),
is not fully used in clinical practice (16). Studies comparing PPWA and laser Doppler measurements, thermography, or forearm flow measurements have demonstrated correlation in the dynamic changes between
PPWA and digital blood flow (17,18).
Changes in finger PPWA correspond to changes in
the blood volume pulsations and depend on the distensibility of the vascular wall and the intravascular
pulse pressure. Usually, the effect of autonomic impulses upon distensibility is so strong that it predominates over the opposite effect of pulse pressure. Decreases in PPWA connected with pain and other
stressful stimuli are caused by vasoconstriction of the
finger arterial bed rather than changes in the pulse
pressure (19). As with endogenous catecholamines,
drugs cause a predictable reduction in PPWA (20). In
the present study, injection of the epinephrinecontaining test dose resulted in significant decreases
in PPWA from 20 to 140 seconds in the isoflurane
group and from 40 to 200 seconds in the sevoflurane
group. The difference in the time the PPWA was decreased may be attributed to a more potent inhibitory
effect of isoflurane on the response to epinephrine
(21). The strong maximal PPWA reduction after IV
injection of 15 g of epinephrine (61% and 58% in the
sevoflurane and isoflurane groups, respectively) may

ANESTH ANALG
2005;101:1506 11

indicate that a smaller dose of epinephrine will be

enough to produce a positive response (10% PPWA
decrease).
Several measures were taken in the present study to
eliminate sources of error with finger plethysmographic monitoring (10). Patients who smoked, those
with vascular disease, or taking medication that affects vascular resistance were excluded. Anesthetic
technique, drugs, and monitoring site were standardized. Patients were maintained normothermic, and their
hands were covered with a dark blanket to avoid regional hypothermia and contamination with ambient
light. All measurements were made with the patient in
the supine position with the transducer at the level of the
heart. Importantly, to avoid individual variations in absorption, scattering and reflection of the emitted light,
only within-subject changes in the PPWA, were reported
as every patient served as his or her control.
Our study must be interpreted with some caution.
First, our results apply only to healthy young adults
undergoing general anesthesia with either sevoflurane
or isoflurane. Other patient populations, such as the
elderly, those with vascular disease, those receiving
vasoactive drugs, or those receiving different anesthetics may have altered plethysmographic pulse
wave responses and thus require separate determination of the response criteria and the reliability profile.
Second, our study design was based on the injection of
the full test dose of IV epinephrine. In actual clinical
practice, however, only a fractional dose may be inadvertently injected IV. Thus, a dose-response study is
required to determine the minimum effective dose of
epinephrine required to elicit 100% reliability based
on the PPWA criterion. Third, although this study
showed that conventional hemodynamic criteria were
imperfect markers of test dose injection, and that
PPWA was a reliable indicator, it failed to demonstrate that the PPWA criterion was more effective than
the HR or the noninvasive SBP criteria. This is because
of the smaller number of patients assigned in our
study (type II error). Retrospective power analysis of
our data revealed that at least 48 and 150 patients
would be required to provide a power 0.8 (P 0.5)
for detecting a statistically significant difference in
sensitivities between the PPWA and HR criteria in the
sevoflurane and isoflurane groups, respectively. Similarly, 74 patients would be required for detecting
statistically significant differences in sensitivities between PPWA and noninvasive SBP criteria in both
anesthetic groups. Finally, the ability of the attending
anesthesiologist to detect a 10% decrease in the PPWA
displayed on the oscilloscope was not tested in the
present study because this should be addressed in a
blinded, prospective manner. However, the addition
of numerical values indicating the PPWA, termed the
Perfusion Index (22), to new pulse oximeters will enhance the clinical applicability of the PPWA criterion.

ANESTH ANALG
2005;101:1506 11

In conclusion, unlike HR or noninvasive SBP,

PPWA was found to be a reliable indicator of intravascular injection of an epinephrine-containing epidural test dose in anesthetized adults. Further studies
are warranted to determine whether this novel
method is applicable to other patient populations, under different anesthetic techniques, or with smaller
doses of epinephrine.

References
1. Dunet F, Pfister Ch, Deghmani M, et al. Clinical results of
combined epidural and general anesthesia procedure in radical
prostatectomy management. Can J Urol 2004;11:2200 4.
2. Dahl JB, Daugaard JJ, Rasmussen B, et al. Immediate and prolonged effects of pre- versus postoperative epidural analgesia
with bupivacaine and morphine on pain at rest and during
mobilisation after total knee arthroplasty. Acta Anaesthesiol
Scand 1994;38:557 61.
3. Lee PK, Kim JM. Lumbar epidural blocks: a case report of a
life-threatening complication. Arch Phys Med Rehabil 2000;81:
158790.
4. Mulroy MF, Norris MC, Liu SS. Safety steps for epidural injection of local anesthetics: review of literature and recommendations. Anesth Analg 1997;85:1346 56.
5. Guinard JP, Mulroy MF, Carpenter RL, Knopes KD. Test doses:
optimal epinephrine content with and without beta-adrenergic
blockade. Anesthesiology 1990;73:386 92.
6. Takahashi S, Tanaka M. Reduced efficacy of simulated epidural
test dose in sevoflurane-anesthetized adults. Can J Anaesth
1999;46:433 8.
7. Tanaka M, Yamamoto S, Ashimura H, et al. Efficacy of an
epidural test dose in adult patients anesthetized with isoflurane:
lidocaine containing 15 micrograms epinephrine reliably increase arterial blood pressure, but not heart rate. Anesth Analg
1995;80:310 4.
8. Liu SS, Carpenter RL. Hemodynamic responses to intravascular
injection of epinephrine-containing epidural test doses in adults
during general anesthesia. Anesthesiology 1996;84:817.

REGIONAL ANESTHESIA
MOWAFI
PLETHYSMOGRAPHIC PULSE WAVE AND EPIDURAL TEST DOSE

1511

9. Mowafi HA. Digital skin blood flow as an indicator for intravascular injection of epinephrine-containing simulated epidural
test dose in sevoflurane-anesthetized adults. Anesth Analg.
2005;101:584 8.
10. Blanc FB, Haig M, Troli M, Sauve B. Computerized photoplethysmography of the finger. Can J Anaesth 1993;40:271 8.
11. Tanaka M, Toshiaki N. A comparative study of hemodynamic
and T-wave criteria for detecting intravascular injection of the
test dose (epinephrine) in sevoflurane anesthetized adults.
Anesth Analg 1999;89:32 6.
12. Glantz SA. Confidence Intervals for the entire population. In:
Glantz SA, ed. Primer of biostatistics. Third ed. New York:
Mcgraw-Hill, 1992;2127.
13. Tanaka M, Takahashi S, Kondo T, Matsumiya N. Efficacy of
simulated epidural test dose in adult patients anesthetized with
isoflurane: a dose response study. Anesth Analg 1995;81:98792.
14. Hertzman AB. The blood supply of various skin areas as estimated by the photoelectric plethysmograph. Am J Physiol 1938;
124:328 40.
15. Murray WB, Foster PA. The peripheral pulse wave: information
overlooked. J Clin Monit 1996;12:36577.
16. Eichhorn JH, Cooper JB, Cullen DJ, et al. Standards for patient
monitoring during anesthesia at Harvard Medical School.
JAMA 1986;256:101720.
17. Marszalek A. The use of selected methods in assessing peripheral circulation of blood. Med Pr 2000;51:299 309.
18. Alam TA, Seifalian AM, Baker D. A review of methods currently
used for assessment of in vivo endothelial function. Eur J Vasc
Endovasc Surg 2005;29:269 76.
19. Dorlas JC, Nijboer JA. Photo-electric plethysmography as a
monitoring device in anesthesia. Br J Anaesth 1985;57:524 30.
20. Hoffman BB, Lefkowitz RJ. Chatecholamines and sympathomimetic drugs. In: Gilman AG, Hardman JG, Limbird LE, et al,
eds. Goodman and Gilmans the pharmacological basis of therapeutics. New York: Pergamon Press, 1996:199 284.
21. Akata T, Kodama K, Takahashi S. Volatile anaesthetic actions on
norepinephrine-induced contraction of small splanchnic resistance arteries. Can J Anaesth 1995;42:1040 50.
22. Hager H, Reddy D, Kurz A. Perfusion index: a valuable tool to
assess changes in peripheral perfusion caused by sevoflurane.
Anesthesiology 2003;99:A593.

Small-Dose Bupivacaine-Sufentanil Prevents Cardiac Output

Modifications After Spinal Anesthesia
Karim Asehnoune, MD*, Eric Larousse, MD*, Jean Marc Tadie,
Stephane Droupy, MD, and Dan Benhamou, MD*

MD*,

Vincent Minville,

MD*,

*Service dAnesthesie-Reanimation et Unite Propre de Recherche de lEnseignement Superieur-Equipe dAccueil

(UPRES-EA 3540) and Service dUrologie, Centre Hospitalo-Universitaire de Bicetre, AP-HP, Le Kremlin Bicetre, France

Spinal injection of small-dose (SD) bupivacaine decreases the likelihood of hypotension compared with
large-dose (LD) bupivacaine. We assumed that a SD of
bupivacaine could also prevent the decrease in cardiac
output (CO). Patients undergoing elective urologic,
lower abdominal, or lower limb surgery under spinal
anesthesia were included in this prospective randomized study. Spinal injection consisted of 5 g of sufentanil and either SD (7.5 mg of hyperbaric bupivacaine
with glucosemonohydrate80 mg/mL; n 19 patients)
or LD (12.5 mg of hyperbaric bupivacaine with glucosemonohydrate80 mg/mL; n 19 patients). CO

ypotension is a common complication during

spinal anesthesia (SA) (1). This complication is
more hazardous in elderly patients because they
may have decreased physiological reserve and compromised blood supply to various vital organs (2).
Many different techniques, such as IV crystalloid and
vasopressor administration, have been used to prevent this complication (2,3). However, rapid infusion
of large amounts of IV fluid may be detrimental to
patients with cardiac dysfunction (3,4). Moreover,
ephedrine and vasopressors may lead to serious cardiac side effects (hypertension or tachycardia) (5).
Ben-David et al. (6) demonstrated that a small dose
(SD; 7.5 mg) of bupivacaine with fentanyl was suitable
for patients undergoing ambulatory surgery. The usefulness of SD bupivacaine (10 mg) was demonstrated for motor blockade and time to recovery (7).
Dose-response data on clinical anesthetic characteristics for spinal bupivacaine indicate that SD can be
used for short duration surgery (6,8). SDs of bupivacaine (10 mg) decreased the likelihood of prolonged
Accepted for publication May 12, 2005.
Address correspondence and reprint requests to Karim Asehnoune, MD, Service dAnesthesie-Reanimation, Hopital de
Bicetre, 94275 Le Kremlin Bicetre, France. Address e-mail to
asehnounekarim@hotmail.com.
DOI: 10.1213/01.ANE.0000180996.91358.CC

1512

Anesth Analg 2005;101:15125

(impedance cardiography), arterial blood pressure, and

heart rate) were measured at 1 min before performance
of spinal block and 2, 10, and 30 min after the intrathecal
injection. Sensory level was also assessed at 30 min. CO
was higher in the SD group as compared with the LD
group from 2 min to 30 min after spinal anesthesia.
Moreover, CO increased at 2 min in the SD group and
decreased at 10 and 30 min in the LD group compared
with baseline value. In conclusion, SD bupivacaine provides successful anesthesia and gives better CO stability than LD.
(Anesth Analg 2005;101:15125)

detrusor blockade, inability to void, and prolonged

time to hospital discharge. Furthermore, Ben-David et
al. (6) reported that mean arterial blood pressure
(MAP) remained unchanged in 50 patients undergoing SD bupivacaine SA for ambulatory surgery. This
finding is of particular interest in patients at a high
risk for developing spinal-induced hypotension. Kamenik and Paver-Erzen (9), using the noninvasive impedance cardiography method, reported a decreased
cardiac output (CO) after large-dose (LD; 15 mg) bupivacaine. However, the impact of SD bupivacaine on
CO is not known. We hypothesized that a reduced
dose of bupivacaine with sufentanil could provide
successful anesthesia and stability of CO. We therefore
conducted this study to compare CO changes after
spinal injection of LD (12.5 mg) or SD (7.5 mg)
bupivacaine.

Methods
After approval by our Human Studies Committee,
written informed consent was obtained from 38 ASA
physical status I and II patients undergoing elective
urologic, lower abdominal, or lower limb surgery under SA. Patients with any contraindication to regional
anesthesia (patients refusal, hemostasis abnormalities, or local infection), dementia, allergic reaction to
2005 by the International Anesthesia Research Society
0003-2999/05

ANESTH ANALG
2005;101:15125

local anesthetics, anemia (hemoglobin 10 g/dL), and

pregnant women were excluded from the study. The
patients were fasted overnight, and oral fluid intake
was allowed for up 6 h before the operation. Premedication consisted of oral hydroxyzine 100 mg 1 h before
surgery. No IV fluid was infused before entering the
study. The study was randomized and unblinded regarding the anesthesia solution. The patients received
either a SD SA (SD group) or a LD SA (LD group). The
SD group received 7.5 mg of hyperbaric bupivacaine
with glucosemonohydrate80 mg/mL (Marcaine 0.5%;
Astra, Sodertalje, Sweden) and 5 g of sufentanil; the LD
group received 12.5 mg of hyperbaric bupivacaine with
glucosemonohydrate80 mg/mL and 5 g of sufentanil.
After arrival in the operating room, an IV catheter was
placed in the patients forearm. Subarachnoid puncture
was performed using a midline approach with a 25gauge Sprotte needle at the L4-5 interspace with the
patient in the sitting position. The Sprotte needle aperture was oriented in a cephalad direction throughout the
intrathecal injection. The injection lasted 30 s in both
groups. The patients were placed supine within 10 s of
the injection. CO was measured with the impedance
cardiography method (ICM) (NCCOM3; BoMed Medical Manufacturing, Irvine, CA). Impedance cardiography electrodes were placed as recommended by the
manufacturer just after the IV line was inserted. Slow
mode data recording then started. Heart rate (HR) was
continuously recorded, and noninvasive MAP, systolic
blood pressure (SBP), and diastolic blood pressure (DBP)
were measured by an automated blood pressure cuff
every 2.5 min with an AS3 monitor (Datex-Ohmeda
SAS, Champagne du Mont dOr, France). Four sets of
data were recorded. Each set included HR, SBP, MAP,
DBP, and CO. The first set of measurements (baseline
values) took place after the insertion of the IV line and
before the subarachnoid puncture (TO). The second set
was performed 2 min after the end of the local anesthetic
injection, and the third and fourth set were performed
10 min and 30 min after the end of the injection, respectively. A blinded observer assessed the dermatome level
of sensory blockade bilaterally with an ice-cold alcoholimmersed sponge 30 min after the injection of the local
anesthetic (T3).
Hypotension, defined as a decrease in SBP to
100 mm Hg or 70% of baseline, was treated with IV
boluses of ephedrine 6 mg repeated every 3 min.
Bradycardia, defined as HR 55 bpm, was treated
with atropine 1 mg.
Kamenik and Paver-Erzen (9) measured CO (ICM)
before and 30 min after SA (LD bupivacaine 15 mg)
and noted a 25% decrease 30 min after SA (in the
group of patients receiving no IV fluid therapy). Our
patient population was older compared with Kamenik
and Paver-Erzens patients (9); thus, we hypothesized a
30% decrease in CO in the LD group and no changes in
the SD group. A power calculation for a 30% difference

REGIONAL ANESTHESIA
ASEHNOUNE ET AL.
SMALL-DOSE BUPIVACAINE AND CARDIAC OUTPUT

1513

in CO with a probability level of 0.05 and power of 0.80

(1-) yielded a sample size of 16 patients. Accordingly,
19 patients were enrolled in both groups.
To compare demographic and surgical data, as well
as patient characteristics, between groups, a 2 test or
a Students t-test was used. Hemodynamic data between groups and changes from baseline were compared using variance analysis for repeated measurements, followed by paired Students t-test. P 0.05
was considered statistically significant. Data are presented as mean sd unless stated otherwise.

Results
Thirty-six patients completed the study. There were
no refusals and no dropouts. The ICM device failed to
monitor CO for one patient in each group. All procedures were successfully performed under SA. The two
groups were similar with respect to age, sex, ASA
physical status, weight, and height (Table 1). The median block level was two segments higher in the LD
group as compared with the SD group (Table 1; P
0.05). No statistically significant differences were
found between the groups with respect to baseline
values of CO, SBP, MAP, DBP, and HR. Two patients
in each group were treated for hypotension. In the LD
group, one patient received 6 mg and the other received 12 mg of ephedrine. In the SD group, 2 patients
received 6 mg of ephedrine. In the LD group, one
patient received 1 mg of atropine. All complications
were successfully treated, as described above. The
exclusion of the treated patients did not modify the
results.
CO was significantly higher in the SD group as
compared with LD group 2, 10, and 30 min after SA. In
the SD group, CO significantly increased 2 min after
SA (P 0.0002) and returned to baseline values (T0)
after 10 and 30 min. In the LD group, CO significantly
decreased 10 (P 0.001) and 30 min (P 0.0002) after
SA (Fig. 1).
HR and DBP were significantly lower in the LD group
as compared with the SD group (at 10 and 30 min and at
30 min, respectively). SBP, MAP, and DBP significantly
decreased after SA in the SD group as compared with T0,
whereas HR remained unchanged. SBP, MAP, DBP, and
HR significantly decreased in the LD group as compared
with T0 (Table 2).

Discussion
These results demonstrate that the spinal injection of
SD (7.5 mg) bupivacaine plus 5 g of sufentanil provides successful anesthesia and gives better CO stability than LD (12.5 mg) bupivacaine plus 5 g of sufentanil. The hemodynamic stability in the SD group is
also reflected in the significantly higher values in HR

1514

REGIONAL ANESTHESIA ASEHNOUNE ET AL.

SMALL-DOSE BUPIVACAINE AND CARDIAC OUTPUT

ANESTH ANALG
2005;101:15125

Table 1. Demographic Data, and Sensory Block Level in Two Groups of Patients
Demographic data

SD group (n 19)

LD group (n 19)

P-value

Sex (M/F)
ASA (I/II)
Age (yr)
Weight (kg)
Height (cm)
Sensory level #

13/6
8/11
64 16
74 12
168 8
T8 (T412)

13/6
9/10
60 20
73 18
171 9
T6 (T48)

NS
NS
NS
NS
NS
0.006

Data are expressed as ratio or mean sd, except #, which is the median (range).
LD large dose; SD small dose; NS nonsignificant.

Figure 1. Cardiac output variations in small dose (SD) and large

dose (LD) groups. Cardiac output (CO) 2, 10, and 30 min after spinal
anesthesia (SA) is expressed as percent changes. The baseline value
of CO (T0: before SA) represents the value 100%. **P 0.01 compared with T0 in the same group; #P 0.05 versus the LD group;
##P 0.01 versus the LD group.

and DBP observed after SD compared with LD

bupivacaine.
In this study, a noninvasive ICM for CO measurements was used. Although the absolute value of
CO measurement is controversial, its value for monitoring the trends of CO changes has been accepted by
most researchers (9,10). Comparisons with thermodilution (TD) and direct Fick methods reported that ICM
measures CO accurately in different conditions (in the
supine position, during head-up tilt, and after the
induction of anesthesia) (11,12). Moreover, this
method can be used in circumstances where there are
immediate and large changes in sympathetic tone.
Indeed, Sageman (13) demonstrated that accurate and
reliable measurements of CO are possible in a lower
body negative pressure model.
An effective approach to minimizing cardiac side
effects after SA is to use SDs of local anesthetics.
However, whereas the use of a single shot of SD local
anesthetic for SA may limit hypotension, it may not
provide acceptable anesthesia. Some authors, therefore, added a lipophilic opioid for anesthetic synergy
(7,14).
HR and DBP were significantly lower in the LD
group compared with the SD group, and only two
patients in each group developed hypotension. This
infrequent incidence of hypotension in the SD group is
in accordance with previously published data (15,16).
A reason for the infrequent incidence of hypotension

in the LD group is the low limit set in our study

(17,18). Indeed, if we had operationally defined hypotension as a decrease in systemic arterial pressure
80% of baseline, seven patients (39%) in the LD
group and five patients (28%) in the SD group would
have developed hypotension.
In the present study, CO was significantly lower in
the LD group compared with the SD group. In the
supine position, SA leads to blood pooling in the
denervated lower extremities with a reflex vasoconstriction in the nonblocked upper extremities (19).
When this vasoconstriction of the nonblocked areas
remains effective, the CO is unchanged (19,20). Our
results indicate that SD bupivacaine-sufentanil gives
better CO stability than the LD, probably by maintaining this vasoconstriction reflex. Indeed, the sensory
block level was significantly lower in the SD group
compared with the LD group, although sympathetic
level is very variable and may not correspond to sensory level.
It has been reported that 4 mg of bupivacainefentanyl SA caused dramatically less hypotension
than 10 mg of bupivacaine and nearly eliminated the
need for vasopressors in the elderly (7). In this study,
we showed that SD bupivacaine preserves the CO,
even with 7.5 mg. This point is critical because the
dose of 4 mg used by Ben-David et al. (7) in elderly
patients might not be sufficient for surgeries in which
the need for an intense motor block would require
larger doses of local anesthetic.
This preliminary study has several limitations. First,
a small number of patients were studied. However,
hemodynamic changes were consistently obtained,
suggesting that our results can be generalized to the
whole population of patients studied under similar
conditions. Second, our results were obtained in a
patient population 40 years of age and may not be
extrapolated to other conditions such as younger patients, different operations, and varying levels of hydration. A further limitation is that the magnitude of
changes might be altered by a different approach (TD
or direct Fick method). Crucially, the measurement of
CO might not be accurate when the patient is moving
because of interference caused by changes in electrode
position. Despite all this, individual patients plots of

ANESTH ANALG
2005;101:15125

REGIONAL ANESTHESIA
ASEHNOUNE ET AL.
SMALL-DOSE BUPIVACAINE AND CARDIAC OUTPUT

1515

Table 2. Hemodynamic Changes Immediately Before Spinal Anesthesia, and 2 min, 10 min, 30 min After Spinal
Anesthesia in Two Groups of Patients
Hemodynamic variables
Small-dose group
HR (bpm)
SBP (mm Hg)
MAP (mm Hg)
DBP (mm Hg)
Large-dose group
HR (bpm)
SBP (mm Hg)
MAP (mm Hg)
DBP (mm Hg)

Before SA

2 min after SA

10 min after SA

30 min after SA

73 16
150 32
105 19
83 16

77 11
133 31*
92 23*
76 19

74 1#
126 25**
89 16**
71 16**

71 14#
132 26**
96 16*
75 15#

68 12
147 18
98 17
78 12

67 7
130 19*
89 16
74 19

62 10*
117 16**
87 18*
65 13**

60 10**
124 22**
91 21
64 14**

SA spinal anesthesia; HR heart rate; SBP systolic blood pressure; MAP mean arterial blood pressure; DBP diastolic blood pressure.
* P 0.05; ** P 0.01 compared with before SA value; # P 0.05 versus large-dose group.

hemodynamic changes reassured us of the validity of

our results. In addition, a noninvasive technique
would be highly acceptable to patients and more
likely to be used routinely.
In conclusion, despite limitations of ICM monitoring of CO, our study shows that a decrease in CO
observed with LD bupivacaine (12.5 mg) is not associated with SD bupivacaine (7.5 mg), and the use of SD
bupivacaine may prevent cardiovascular side effects.

References
1. Tarkkila PJ, Kaukinen S. Complications during spinal
anesthesia: a prospective study. Reg Anesth 1991;16:101 6.
2. Lim HH, Ho KM, Choi WY, et al. The use of intravenous
atropine after a saline infusion in the prevention of spinal
anesthesia-induced hypotension in elderly patients. Anesth
Analg 2000;91:1203 6.
3. McCrae AF, Wildsmith JAW. Prevention and treatment of hypotension during central neural block. Br J Anaesth 1993;70:672 80.
4. Buggy D, Higgins P, Moran C, et al. Prevention of spinal
anesthesia-induced hypotension in the elderly: comparison between preanesthetic administration of crystalloids, colloids and
no prehydration. Anesth Analg 1997;84:106 10.
5. Critchley LAH, Stuart JC, Conway F, Short TG. Hypotension
during subarachnoid anaesthesia: haemodynamic effects of
ephedrine. Br J Anaesth 1995;74:373 8.
6. Ben-David B, Levin H, Salomon E, et al. Spinal bupivacaine in
ambulatory surgery: the effect of saline dilution. Anesth Analg
1996;83:716 20.
7. Ben-David B, Frankel R, Arzumonov T, et al. Minidose
bupivacaine-fentanyl spinal anesthesia for surgical repair of hip
fracture in the aged. Anesthesiology 2000;92:6 10.
8. Liu SS, Ware PD, Allen HW, et al. Dose-response characteristics
of spinal bupivacaine in volunteers: clinical implications for
ambulatory anesthesia. Anesthesiology 1996;85:729 36.

9. Kamenik M, Paver-Erzen V. The effects of lactated ringers

solution infusion on cardiac output changes after spinal anesthesia. Anesth Analg 2001;92:710 4.
10. Critchley LAH. Electrical bioimpedance for non-invasive cardiac measurement. Crit Care Med 1998;26:1460 1.
11. Jarvela K, Honkonen SE, Jarvela T, et al. The comparison of
hypertonic saline (7.5%) and normal saline (0.9%) for initial
fluid administration before spinal anesthesia. Anesth Analg
2000;91:14615.
12. Koobi T, Kaukinen S, Turjanmaa VM, Uusitalo AJ. Whole-body
impedance cardiography in the measurement of cardiac output.
Crit Care Med 1997;25:779 85.
13. Sageman WS. Reliability and precision of a new thoracic electrical bioimpedance monitor in a lower body negative pressure
model. Crit Care Med 1999;27:1986 90.
14. Tejwani GA, Rattan AK, McDonald JS. Role of spinal opioid
receptor in the antinociceptive interactions between intrathecal
morphine and bupivacaine. Anesth Analg 1992;74:726 34.
15. Hartmann B, Junger A, Klasen J, et al. The incidence and risk
factors for hypotension after spinal anesthesia induction: an
analysis with automated data collection. Anesth Analg 2002;94:
15219.
16. Vercauteren MP, Coppejans HC, Hoffmann VL, et al. Smalldose hyperbaric versus plain bupivacaine during spinal anesthesia for cesarean section. Anesth Analg 1998;86:989 93.
17. Liu SS, McDonald SB. Current issues in spinal anesthesia. Anesthesiology 2001;94:888 906.
18. Butterworth J. Physiology of spinal anesthesia: what are the
implications for management? Reg Anesth Pain Med 1998;23:
370 3.
19. Arndt JO, Hock A, Stanton-Hicks M, Stuhmeier KD. Peridural
anesthesia and the distribution of blood in supine humans.
Anesthesiology 1985;63:616 23.
20. Larousse E, Asehnoune K, Dartayet B, et al. The hemodynamic
effects of pediatric caudal anesthesia assessed by esophageal
Doppler. Anesth Analg 2002;94:1165 8.

Prophylactic Intravenous Ondansetron and Dolasetron in

Intrathecal Morphine-Induced Pruritus: A Randomized,
Double-Blinded, Placebo-Controlled Study
Christos A. Iatrou, MD, PhD, Christos K. Dragoumanis, MD,
Theodosia D. Vogiatzaki, MD, PhD, George I. Vretzakis MD, PhD,
Constantinos E. Simopoulos, MD, PhD, and Vasilios K. Dimitriou,

MD, PhD

Departments of Anesthesia and Surgery, Democritus University of Thrace, Alexandroupolis, Greece, and the Department
of Anesthesia, G. Gennimatas Hospital, Athens, Greece

Pruritus is the most common side effect of intrathecal

morphine for postoperative pain relief. Activation of
central 5-hydroxytryptamine subtype 3 (5-HT3) receptors is one of its possible mechanisms. The role of 5-HT3
antagonists in the prevention of pruritus has not been
clearly established. In a prospective, randomized,
double-blind, placebo-controlled study, we evaluated
the efficacy of prophylactic administration of ondansetron and dolasetron for the prevention of intrathecal
morphine-induced pruritus. The patients were randomized into 3 groups to receive either 4 mg ondansetron IV (group O, n 35), 12.5 mg dolasetron IV (group
D, n 35) or 5 mL placebo (group P, n 35) 30 min
before administration of spinal anesthesia with 10 to
17.5 mg of 0.5% hyperbaric bupivacaine and 0.25 mg of
morphine for urologic, orthopedic, or vascular surgery.

ntrathecal morphine improves postoperative analgesia, but it is accompanied by a frequent incidence

of postoperative nausea and vomiting (PONV) and
pruritus (1). Pruritus is the most common side effect of
intrathecal morphine, with a reported incidence of
62% to 94% (2 4). It is unpleasant for patients and
difficult to treat, and its prevention remains a challenge for anesthesiologists (5,6). Although the exact
mechanism is unclear, it seems that intrathecal morphines interaction with central 5-hydroxytryptamine
subtype 3 (5-HT3) receptors (6) plays some role in the
genesis of pruritus. As a result, 5-HT3 receptor antagonists could be effective in its control. Ondansetron
has been used for this purpose with conflicting results
Accepted for publication June 9, 2005.
Address correspondence and reprint requests to Christos Dragoumanis MD, University Hospital of Alexandroupolis, Level 2,
Room: 54.10, Dragana, Alexandroupolis 68100, Greece. Address
e-mail to christosdr@panafonet.gr.
DOI: 10.1213/01.ANE.0000181338.35454.6A

1516

Anesth Analg 2005;101:151620

Patients were evaluated for incidence and severity of

pruritus at arrival to the postanesthesia care unit and at
2, 4, 8, and 24 h postoperatively. The incidence and severity of pruritus was significantly less frequent in the
ondansetron and dolasetron groups compared with
placebo (34%, 20%, and 66% respectively, P 0.01). Patients who received 5-HT3 antagonist reported significantly less total severity of pruritus compared with placebo during the first 8 h and the severe pruritus was
observed only in patients within P group (P group: 4 of
35; 11%, O or D group: 0 of 35; 0%, P 0.05). We conclude that the prophylactic use of ondansetron and dolasetron helps to reduce the incidence and severity of
intrathecal morphine-induced pruritus.
(Anesth Analg 2005;101:1516 20)

(3,4,7,8). Dolasetron, another 5-HT3 receptor antagonist, is usually used to control nausea and vomiting
associated with chemotherapy and PONV (9,10).
However, its antipruritic activity has not been evaluated. Therefore, we conducted a prospective, randomized, double-blind, placebo-controlled study to determine the effectiveness of dolasetron and ondansetron
for the prevention of pruritus after spinal anesthesia
performed with bupivacaine and morphine in patients
undergoing elective vascular, orthopedic, or urologic
surgery.

Methods
After institutional ethics committee approval and
written informed consent, we studied 119 adult ASA
physical status IIII patients undergoing elective urologic, vascular, or orthopedic surgery under spinal
anesthesia. Exclusion criteria included patient refusal
to participate in the study, contraindication for spinal
2005 by the International Anesthesia Research Society
0003-2999/05

ANESTH ANALG
2005;101:1516 20

REGIONAL ANESTHESIA
IATROU ET AL.
ONDANSETRON AND DOLASETRON FOR INTRATHECAL MORPHINE-INDUCED PRURITUS

anesthesia, prolonged QT interval, electrolyte disturbances, age older than 75 yr, morbid obesity, sleep
apnea syndrome, coexisting skin disorder, and systemic diseases with pruritus. Patients with a history of
allergy associated with any drug included in the study
were also excluded. The study was designed in a
prospective, randomized, double-blind, and placebocontrolled fashion. A computer randomization program was used to allocate the patients into 3 groups:
12.5 mg dolasetron IV (group D), 4 mg ondansetron IV
(group O), or 5 mL of normal saline (group P) 30 min
before administration of spinal anesthesia. Ondansetron and dolasetron were diluted in normal saline to a
volume of 5 mL. Patients and anesthesiologists performing the spinal and collecting the postoperative
data were blinded as to the study drugs.
Every patient was prehydrated with Ringers lactate
solution 510 mL/kg. Spinal anesthesia was performed at the L2-3 or L3-4 interspace with a 25-gauge
Quincke-type needle using 10 to 17.5 mg of 0.5% hyperbaric bupivacaine plus 0.25 mg of preservative-free
morphine. No sedation was used before the procedure. Midazolam, in 0.5 mg IV increments, was used
for intraoperative sedation at the discretion of the
anesthesiologist.
The patients were followed for 24 h postoperatively.
Sedation level was evaluated with a 3-degree scale (1
wide awake; 2 drowsy; 3 responds to stimulation). Postoperative wound pain at rest was assessed
with a 10-cm visual analogue scale (VAS). Rescue
treatment for postoperative pain was provided with
meperidine patient-controlled analgesia (PCA), 1 mg/
mL, 10 mg bolus, 20-min lockout, and 4-h limit of
50 mg. Patients were instructed to use the PCA if their
pain became worse than mild.
Pruritus was evaluated at arrival in the postanesthesia care unit and at 2, 4, 8, and 24 h postoperatively
by 2 blinded investigators (senior residents in anesthesiology). Pruritus was defined as the sensation that
provokes the desire to scratch. The patients were questioned about the presence, location, and degree of
pruritus. The degree of pruritus was classified as 0
no pruritus, 1 mild pruritus, 2 moderate pruritus,
and 3 severe pruritus. Severe pruritus was treated
with 3 mg nalbuphine IV (11). Patients were also
asked about the presence of PONV. Patients who reported vomiting or intense nausea received 10 mg IV
metoclopramide.
Statistical analysis was performed using the Statistical Package for the Social Sciences (SPSS for Windows, version 11, SPSS, Inc., Chicago, IL). We considered a 50% reduction in the incidence of pruritus to be
clinically important. Power analysis was performed to
determine the sample size with a probability for a type
II error of 0.2 and type I error of 0.05. To detect a 50%
reduction in the incidence of pruritus, using the results of a pilot study, in which pruritus was present in

1517

13 (65%) of 20 patients who received placebo, a sample

size of 35 patients in each group was estimated to be
required. Dropouts of the study were compensated for
by admission of a new patient into the study protocol
in a blind and randomized manner any time a patient
was excluded. Statistical analysis was performed with
2 test for nominal data, Kruskal-Wallis test followed
by Mann-Whitney U-test for ordinal data, and analysis of variance for continuous data. Confidence interval (CI) at 95% was calculated for the incidences of
pruritus and PONV. The level of significance was set
at P 0.05. For pairwise intergroup comparisons we
used the Bonferroni correction for level (e.g., P
0.017 for pairwise comparisons among three groups).

Results
From December 2003 to March 2004, we enrolled 119
patients in our study, 14 of whom were excluded for
the following reasons: inadequate anesthesia-received
general anesthesia (n 3), reoperation (n 2), incomplete collection of postoperative data (n 5), and
protocol violations (n 4). Therefore, 105 patients
completed the trial, 35 in each group. The demographic and surgical data of patients who completed
the study are listed in Table 1.
The overall incidence of pruritus differed significantly among the groups (P 0.01). Pairwise comparisons showed that this significant difference was attributable to a less frequent incidence of pruritus in
groups D and O compared with group P. The incidence of pruritus between the O and D groups was
not significantly different (P 0.18) (Table 2). In the
first 8 h postoperatively the severity of pruritus
among groups was significantly different and was
significantly less in the O and D groups compared
with placebo in the intergroup comparisons (Table 3).
The overall allocation of the pruritus in the trigeminal,
cervical, thoracic, and lumbar dermatomes is shown in
Figure 1. Patient gender did not have a significant
influence on the incidence of pruritus (Table 2). Severe
pruritus, requiring rescue treatment, was observed
only within group P (P group: 4 of 35, 11%; O and D
groups: 0 of 35, 0%, P 0.05 compared with the
Kruskal-Wallis test).
The incidence of PONV was significantly different
between the two treatment groups, but it did not reach
the adjusted significance level in the pairwise comparisons (Table 2), and no significant difference among
any groups occurred in the number of rescue treatments for PONV.
There were no significant differences in the sum of
pain VAS scores during the observations or the mean
total meperidine consumption among groups as compared with one-way analysis of variance (Table 2).
The number of patients who received midazolam
and the total dose of midazolam administered to them

1518

REGIONAL ANESTHESIA IATROU ET AL.

ONDANSETRON AND DOLASETRON FOR INTRATHECAL MORPHINE-INDUCED PRURITUS

ANESTH ANALG
2005;101:1516 20

Table 1. Characteristics of Patients, Intraoperative, and Surgical Data

Variable

Placebo (n 35)

Dolasetron (n 35)

Ondansetron (n 35)

Age (yr)
Sex (M/F)
Body height (cm)
Body weight (kg)
ASA I/II/III
Duration of surgery (min)
Patients received sedation*
Total dose of midazolam (mg)
Type of surgery
Varicose vein surgery
Leg surgery
Transurethral urologic procedures

57 14
20/15
171 5
78 9
25/7/3
63 12
17 (49%)
1.1 0.5

61 11
18/17
170 4
76 10
28/5/2
59 16
11 (31%)
1.0 0.5

62 7
19/16
169 5
77 9
26/8/2
56 16
14 (40%)
0.8 0.3

8 (23%)
9 (26%)
18 (51%)

10 (29%)
12 (34%)
13 (37%)

7 (20%)
11 (31%)
17 (49%)

Values are mean sd or n (%).

* P 0.34 when compared with chi-square test; P 0.17 when total dose of midazolam administered to patients who received sedation intraoperatively
compared with Kruskal-Wallis test.

Table 2. Incidence of the Pruritus and Postoperative Nausea and Vomiting, Sum of Pain Visual Analogue Scale Scores
for Observations in Postanesthesia Care Unit, 2, 4, 8, and 24 h, Rescue Meperidine Delivered by Patient-Controlled
Analgesia for 24 h

Incidence of pruritus*
95% CI
Males
Females
Incidence of PONV
95% CI
Sum of pain VAS scores (cm)
Rescue meperidine (mg)

Placebo (n 35)

Dolasetron (n 35)

Ondansetron (n 35)

23/35 (66%)
51% to 81%
14/20 (70%)
9/15 (60%)
15/35 (43%)
27% to 59%
5.5 2.5
17.1 13.8

7/35 (20%)
8% to 32%
3/18 (18%)
4/17 (29%)
6/35 (17%)
5% to 29%
4.9 2.3
13.4 13.9

12/35 (34%)
18% to 50%
5/19 (26%)
7/16 (44%)
8/35 (23%)
9% to 37%
5.7 2.5
16.3 12.1

Values are number of patients (%) or mean sd.

PONV postoperative nausea and vomiting; VAS visual analog scale; PACU postanesthesia care unit; CI confidence interval.
* P 0.01; P 0.05 when compared with chi-square test; P 0.001; P 0.01 when compared to placebo with chi-square test.

intraoperatively did not differ significantly among

groups (Table 1). Patients were mainly wide awake
during the five data collection times, and the level of
sedation was similar among groups.

Discussion
The mechanism of intrathecal morphine-induced pruritus is complex. Although itch-specific neuronal pathways are different from pain pathways, they are close
in interaction. Continuing activity of the painprocessing system tonically suppresses activity in the
spinal itch-processing neurons. Thus, inhibition of
pain can unmask pruritus (e.g., intrathecal morphineinduced pruritus) and pruritus can be inhibited by
pain (e.g., antipruritic effect of scratching) (12). The
serotoninergic system seems to play the role of modulator, providing a balance between nociception and
antinociception in the network of pain processing neurons (13,14). Furthermore, morphine is able to activate
5-HT3 receptors by a mechanism independent of opioid receptors (15). The dorsal horn area of the spinal
cord and the spinal tract of the trigeminal nerve in the

medulla are abundant in 5-HT3 receptors. Direct stimulation of 5-HT3 receptors by morphine appears to be one
of the mechanisms of intrathecal morphine-induced pruritus; thus, their occupation by a 5-HT3-receptor antagonist may prevent this pruritus. The results of our study
support this hypothesis, demonstrating that patients
who received preemptive 5-HT3-receptor antagonists reported significantly less pruritus and less severity during
the first 8 h postoperatively compared with patients who
received placebo. The frequency of pruritus was reduced
by 48% and 70% for ondansetron and dolasetron, respectively, compared with placebo. Both drugs thus offer
effective prophylaxis against intrathecal morphineinduced pruritus, at least for the first 8 h postoperatively.
Five clinical studies have previously evaluated the
efficacy of 5-HT3 receptor antagonists as prophylaxis
for intrathecal morphine-induced pruritus; all involved ondansetron (3,4,7,8,16). In three of these studies ondansetron in a dose 4 8 mg IV was more effective than placebo in the prevention of pruritus
associated with spinal anesthesia using bupivacaine
and 0.15 0.3 mg morphine (3,4,7). In the two other
studies, ondansetron was ineffective in preventing

ANESTH ANALG
2005;101:1516 20

REGIONAL ANESTHESIA
IATROU ET AL.
ONDANSETRON AND DOLASETRON FOR INTRATHECAL MORPHINE-INDUCED PRURITUS

1519

Table 3. Pruritus Score, Assessed at Arrival in Postanesthesia Care Unit, at 2, 4, 8, and 24 h Postoperatively

Placebo
No pruritus
Mild pruritus
Moderate pruritus
Severe pruritus
Dolasetron
No pruritus
Mild pruritus
Moderate pruritus
Ondansetron
No pruritus
Mild pruritus
Pairwise comparisons

In PACU*

2 h*

4 h*

8 h*

24 h

12 (34%)
15 (43%)
6 (17%)
2 (6%)

12 (34%)
13 (37%)
9 (26%)
1 (3%)

11 (31%)
14 (40%)
9 (26%)
1 (3%)

11 (31%)
16 (46%)
8 (23%)

31 (89%)
4 (11%)

29 (83%)
2 (6%)
4 (11%)

28 (80%)
4 (11%)
3 (9%)

27 (77%)
8 (23%)

32 (91%)
2 (6%)
1 (3%)

35 (100%)

27 (77%)
5 (14%)
3 (9%)
P-O P-D D-O

24 (68%)
8 (23%)
3 (9%)
P-O P-D D-O

23 (66%)
10 (28%)
2 (6%)
P-O P-D D-O

25 (71%)
9 (26%)
1 (3%)
P-O P-D D-O

33 (94%)
2 (6%)
P-O P-D D-O

Values are number of patients (%).

Pairwise comparisons between groups P-O placebo versus ondansetron, P-D placebo versus dolasetron, and D-O dolasetron versus ondansetron.
PACU postanesthesia care unit.
* P 0.001 when compared with Kruskal-Wallis test; P 0.001; P 0.01 using Mann-Whitney U-test in pairwise comparisons.

Figure 1. Allocation of pruritus in dermatomes.

pruritus (8,16). Szarvas et al. (8) reported a high frequency (73%) of intrathecal morphine-induced pruritus
in a population similar to ours. However, the dose of
intrathecally-administered morphine (0.57 0.13 mg)
was about twofold larger than the dose of morphine
(0.25 mg) that we administered. This suggests that the
antipruritic efficacy of ondansetron depends on the dose
of intrathecally administered morphine. Yazigi et al. (16)
reported the failure of 8 mg ondansetron to prevent
pruritus after intrathecal injection of 25 g sufentanil
and 0.1 mg morphine but speculated that this related to
the lipophilic properties of sufentanil and delayed administration of ondansetron. In our study, the percentage of patients who reported pruritus in the placebo
group (66%) was similar to that of previous investigations (6,7).
Although both 5-HT3 antagonists in our study reduced the incidence and severity of pruritus, this side
effect still occurred in 12 (34%) and 7 (20%) of the 35
patients in each group who received ondansetron and
dolasetron respectively, showing the complexity of the
pathogenesis of pruritus. Pruritus in these patients
may be prevented with a larger dose of these 5-HT3
antagonists, prostaglandin synthesis inhibition (17),

blocking opioid receptors (4,18), or a combination of

therapies.
The prophylaxis of intrathecal morphine-induced
pruritus with ondansetron or dolasetron increases the
cost of perioperative care by 11.26 Euros (14.6 US$)/
patient or 4.04 Euros (5.2 US$)/patient respectively.
Rescue treatments were administered to only 4 patients (11%) with severe pruritus within the placebo
group, arguing against the routine use of prophylaxis
for pruritus because of unnecessary drug administration and increased perioperative care cost. However,
decreasing the overall frequency of intrathecal
morphine-induced pruritus with preemptive ondansetron or dolasetron administration may improve the
quality of postoperative care and might economize
medical and nurse working time spent even for patients who complained of mild or moderate pruritus
without requiring antipruritic treatment. Furthermore, pruritus induced by intrathecal morphine is
occasionally difficult to treat and there are no known
risk factors, except pregnancy (5), to identify patients
who are at high risk to develop pruritus.
In this study there was no statistically significant
difference (P 0.12) in the severity of pruritus among

1520

REGIONAL ANESTHESIA IATROU ET AL.

ONDANSETRON AND DOLASETRON FOR INTRATHECAL MORPHINE-INDUCED PRURITUS

groups at the 24th postoperative hour. However,

power analysis for this comparison showed that its
power was only 43%. Consequently, a type II error
cannot be excluded. We attributed the low statistical
power of the comparison of severity of pruritus at the
24th postoperative hour to an overall decrease of severity of pruritus because of elimination of morphine
from the central nervous system. A larger sample size
(approximately 250 patients) would be required to
achieve adequate statistical power. Nonetheless, this
larger sample size could offer sufficient power to compare the antipruritic potency of ondansetron and dolasetron. Our study was performed in nonpregnant
patients. However, pruritus is more frequent in the
obstetric population because of increased cephalic
spread of spinally administered drugs and an interaction of estrogens with opioid receptors (5). Thus, our
findings only apply to the nonpregnant population.
PONV are common side effects of postoperative
analgesia after intrathecal morphine. Our results show
a reduction in the incidence of PONV in patients who
received 5-HT3 antagonists as compared with placebo.
However, our study was not designed to evaluate this
phenomenon. A study oriented to PONV with appropriate size and exclusion criteria could potentially
more clearly define the impact of ondansetron and
dolasetron on PONV in the setting of intrathecally
administered morphine (19).
We conclude that preemptive administration of
both ondansetron and dolasetron to patients receiving
intrathecal morphine for postoperative analgesia provides a significant reduction of intrathecal morphineinduced pruritus without affecting pain relief. The use
of prophylaxis involves is expensive. Thus, identifying
high-risk populations would maximize its benefits.

References
1. Gwirtz KH, Young JV, Byers RS, et al. The safety and efficacy of
intrathecal opioid analgesia for acute postoperative pain: seven
years experience with 5969 surgical patients at Indiana University Hospital. Anesth Analg 1999;88:599 604.
2. Charuluxananan S, Kyokong O, Somboonviboon W, et al. Nalbuphine versus propofol for treatment of intrathecal morphineinduced pruritus after cesarean delivery. Anesth Analg 2001;93:
1625.

ANESTH ANALG
2005;101:1516 20

3. Yeh HM, Chen LK, Lin CJ, et al. Prophylactic intravenous

ondansetron reduces the incidence of intrathecal morphineinduced pruritus in patients undergoing cesarean delivery.
Anesth Analg 2000;91:1725.
4. Charuluxananan S, Kyokong O, Somboonviboon W, et al. Nalbuphine versus ondansetron for prevention of intrathecal
morphine-induced pruritus after cesarean delivery. Anesth
Analg 2003;96:1789 93.
5. Beilin Y, Bernstein HH, Zucker-Pinchoff B, et al. Subhypnotic
doses of propofol do not relieve pruritus induced by intrathecal
morphine after cesarean section. Anesth Analg 1998;86:310 3.
6. Szarvas S, Harmon D, Murphy D. Neuraxial. opioid-induced
pruritus: a review. J Clin Anesth 2003;15:234 9.
7. Dimitriou V, Voyagis GS. Opioid-induced pruritus: repeated
versus single dose ondansetron administration in preventing
pruritus after intrathecal morphine. Br J Anaesth 1999;83:8223.
8. Szarvas S, Chellapuri RS, Harmon DC, et al. A comparison of
dexamethasone, ondansetron, and dexamethasone plus ondansetron as prophylactic antiemetic and antipruritic therapy in
patients receiving intrathecal morphine for major orthopedic
surgery. Anesth Analg 2003;97:259 63.
9. Korttila K, Clergue F, Leeser J, et al. Intravenous dolasetron and
ondansetron in prevention of postoperative nausea and
vomiting: a multicenter, double-blind, placebo-controlled study.
Acta Anaesthesiol Scand 1997;41:914 22.
10. Piper SN, Suttner SW, Rohm KD, et al. Dolasetron, but not
metoclopramide prevents nausea and vomiting in patients undergoing laparoscopic cholecystectomy. Can J Anaesth 2002;49:
1021 8.
11. Somrat C, Oranuch K, Ketchada U, et al. Optimal dose of
nalbuphine for treatment of intrathecal-morphine-induced pruritus after caesarean section. J Obstet Gynaecol Res 1999;25:
209 13.
12. Ikoma A, Rukwied R, Stander S, et al. Neurophysiology of
pruritus: interaction of itch and pain. Arch Dermatol 2003;139:
1475 8.
13. Zeitz KP, Guy N, Malmberg AB, et al. The 5-HT3 subtype of
serotonin receptor contributes to nociceptive processing via a
novel subset of myelinated and unmyelinated nociceptors.
J Neurosci 2002;22:1010 9.
14. Arcioni R, della Rocca M, Romano S, et al. Ondansetron inhibits
the analgesic effects of tramadol: a possible 5-HT(3) spinal receptor involvement in acute pain in humans. Anesth Analg
2002;94:15537.
15. Fan P. Nonopioid mechanism of morphine modulation of the
activation of 5-hydroxytryptamine type 3 receptors. Mol Pharmacol 1995;47:4915.
16. Yazigi A, Chalhoub V, Madi-Jebara S, et al. Prophylactic ondansetron is effective in the treatment of nausea and vomiting but
not on pruritus after cesarean delivery with intrathecal
sufentanil-morphine. J Clin Anesth 2002;14:183 6.
17. Colbert S, OHanlon DM, Galvin S, et al. The effect of rectal
diclofenac on pruritus in patients receiving intrathecal morphine. Anaesthesia 1999;54:948 52.
18. Cepeda MS, Alvarez H, Morales O, Carr DB. Addition of ultralow dose naloxone to postoperative morphine PCA: unchanged analgesia and opioid requirement but decreased incidence of opioid side effects. Pain 2004;107:41 6.
19. Apfel CC, Roewer N, Korttila K. How to study postoperative
nausea and vomiting. Acta Anaesthesiol Scand 2002;46:921928.

The Effects of Continuous Epidural Anesthesia and Analgesia

on Stress Response and Immune Function in Patients
Undergoing Radical Esophagectomy
Masataka Yokoyama, MD, Yoshitaro Itano, PhD, Hiroshi Katayama, MD,
Hiroshi Morimatsu, MD, Yoshimasa Takeda, MD, Toru Takahashi, MD,
Osamu Nagano, MD, and Kiyoshi Morita, MD
Department of Anesthesiology and Resuscitology Okayama University Medical School 251, Shikata-cho, Okayama City,
Okayama 700 8558, Japan

We investigated whether perioperative extensive epidural block (C3-L) affects postoperative immune response in patients undergoing radical esophagectomy.
Patients undergoing radical esophagectomy were randomly assigned to either general anesthesia with continuous epidural infusion via 2 epidural catheters that
was continued for postoperative analgesia (group E, n
15) or intraoperative general anesthesia and postoperative IV morphine analgesia (group G, n 15).
Plasma levels of stress hormones, cytokines, C-reactive
protein (CRP), leukocyte counts, and distribution of
lymphocyte subsets were assessed before and after surgery and on postoperative days (PODs) 1 and 3. In comparison with group E, significant increases in plasma
epinephrine level at the end of surgery (P 0.05) and
norepinephrine level at the end of surgery (P 0.01)
and on POD1 (P 0.01) and POD3 (P 0.01) and significant decrease in cluster of differentiation (CD4/
CD8 ratio) at the end of surgery (P 0.05) were

fferent neural blockade by epidural anesthesia

can decrease the postoperative neuroendocrine
stress response (1). A neuroendocrine response
blunted by epidural anesthesia also could affect postoperative immune function because the immune and nervous systems bidirectionally communicate and influence
each other (2). It has been argued that nociception and
proinflammatory cytokines play a mutual up-regulatory
role (3). Therefore, pain management may influence the
immune response in the postoperative period. It has

Accepted for publication June 15, 2005.

Address correspondence to and reprint requests to Masataka
Yokoyama, MD, Department of Anesthesiology and Resuscitology,
Okayama University Medical School, 251, Shikata-cho, Okayama
City, Okayama 700 8558, Japan. Address e-mail to masayoko
@cc.okayama-u.ac.jp.
DOI: 10.1213/01.ANE.0000184287.15086.1E
2005 by the International Anesthesia Research Society
0003-2999/05

observed in group G. However, there were no significant differences in other variables between groups. In
both groups, plasma cortisol, adrenocorticotropic hormone, interleukin (IL)-1, IL-6, IL-10, and CRP levels
were increased after surgery (each group P 0.01) and
IL-1, IL-6, IL-10, and CRP were still increased on
POD1 and POD3 (each change, each group P 0.01).
Leukocyte counts were increased on POD1 (each group
P 0.05) and POD3 (each group P 0.01). The proportion of lymphocytes decreased from the end of surgery
to POD3 (each group P 0.01). The proportion of B cells
was increased on POD1 (each group P 0.01); that of
natural killer cells was decreased at POD1 and POD3
(each group P 0.01). We conclude that tissue damage
and inflammation apparently overcome the effects of
extensive epidural block on stress response and immune function in radical esophagectomy.
(Anesth Analg 2005;101:15217)

been reported that the alterations in lymphocyte subsets

and the increase in white cell counts induced by surgery
and general anesthesia can be prevented by epidural
analgesia (4 6). Furthermore, Beilin et al. (7,8) reported
that patients treated with patient-controlled epidural analgesia (PCEA) exhibited attenuated proinflammatory
cytokine response in the postoperative period. These
reports resulted from procedures below the umbilicus.
In upper abdominal or major surgery, however, the effect of epidural anesthesia and analgesia on attenuation
of the stress response and preservation of immune function is controversial (9,10). Major surgery induces a series of inflammatory responses, and the stress response
is often exaggerated after major surgery to the detriment
of the patient (11,12). Postoperative immune suppression
can predispose patients with cancer to the development
of postoperative infection (13) or facilitate postoperative
tumor growth and metastasis (14,15).
Anesth Analg 2005;101:15217

1521

1522

REGIONAL ANESTHESIA YOKOYAMA ET AL.

IMMUNE RESPONSE AFTER RADICAL ESOPHAGECTOMY

Abdominothoracic esophagectomy with multiple

lymph node dissection (radical esophagectomy) is a
major surgical procedure with frequent morbidity and
mortality (16). Theoretically, blunting the perioperative stress response and preserving immune function
are critical issues for patients undergoing this procedure. Several studies compared postoperative epidural analgesia with postoperative IV analgesia and
reported reductions in time spent in the intensive care
unit (ICU) and fewer respiratory complications for
patients receiving epidural analgesia (17,18). However, the effects of epidural anesthesia and analgesia
on the stress response and immune function during
and after radical esophagectomy have not been
studied.
The present study was conducted to determine
whether epidural anesthesia and analgesia affect the
stress response and immune function in patients undergoing radical esophagectomy. In previous studies
of procedures above the umbilicus, it appears that the
incomplete effect of epidural anesthesia on the stress
response was largely attributable to incomplete neural
blockade at the surgical site (19). It is reported that a
block of the sensory fibers included in the phrenic
nerves is required for a complete block of the endocrine stress response to upper abdominal surgery (20).
Thus, two epidural catheters for extensive nerve
blockade (C3-L) were used in this study.

Methods
Institutional and ethics committee approval was obtained for this study, and all participants gave written
informed consent. The study group comprised 30 patients (ASA physical status III) with diagnosed
esophageal cancer and scheduled for radical esophagectomy. Because the cough reflex is suppressed for
several days after radical esophagectomy, all patients
were mechanically ventilated at least until the morning of postoperative day (POD) 3. The surgical procedure included thoracoabdominal esophagectomy,
multiple lymph node dissection, and a cervical incision for the anastomosis. Patients were excluded from
the study if they had recently taken corticosteroids or
nonsteroidal antiinflammatory medications. The participants were randomly assigned to one of two methods of anesthesia and postoperative pain control. Fifteen patients received general anesthesia with
continuous epidural infusion (CEI) through two epidural catheters during surgery; CEI was continued for
postoperative pain control (group E). Fifteen patients
received general anesthesia during surgery with postoperative analgesia accomplished by continuous IV
morphine (group G).
On the day of surgery, patients in both groups were
premedicated IM with 50 mg hydroxyzine and 0.5 mg

ANESTH ANALG
2005;101:15217

atropine. An IV cannula was placed for fluid administration, and a radial arterial cannula was placed for
continuous arterial blood pressure monitoring and
blood sampling. In group E, 2 epidural catheters were
inserted at the T3-4 and T10-11 interspaces. The analgesic area was checked by pinprick after 20 min of
epidural injection of 8 mL 1.5% lidocaine via each
epidural catheter and was confirmed to extend at least
from the C3 to L2 dermatomes in all patients. Then,
CEI of 1% lidocaine with 4 g/mL fentanyl was administered at 6 mL/h during surgery. In both groups,
general anesthesia was induced IV with 4 mg/kg thiopental and 50 100 g fentanyl. Intubation of the tracheal was facilitated with 0.1 mg/kg vecuronium.
Ventilation was controlled mechanically to maintain
the partial pressure of expiratory carbon dioxide between 30 and 35 mm Hg, as measured by capnography. General anesthesia was maintained with 50%
nitrous oxide in oxygen and isoflurane. Intermittent
boluses of 50 g fentanyl were given as necessary.
Vecuronium was used to maintain muscular blockade.
After the induction of general anesthesia, a doublelumen catheter was inserted via the right internal
jugular vein for continuous central venous pressure
monitoring and fluid administration. Heart rate (by
electrocardiography), CO2 production (by capnography), arterial blood pressure, central venous pressure,
blood oxygen saturation (by pulse oximetry), and
bladder temperature were monitored. Lactated Ringers solution was infused at 20 mL kg1 h1 during
the induction of anesthesia and to maintain preoperative central venous pressure. Hypotension (arterial
blood pressure 90 mm Hg) was corrected by increasing the rate of IV fluid infusion and by administration
of ephedrine. Blood was replaced as necessary; the
hemoglobin level was maintained at more than 8.0
g/dL with transfusion of packed red blood cells, and
the albumin level was maintained at more than 3.0
g/dL with administration of 5% albumin or freshfrozen plasma. Bladder temperature was maintained
between 36C and 37C with the use of a warming
device (Bair Hugger; Augustine Medical Inc., Minneapolis, MN). Arterial blood gases, serum electrolytes
(Na, K, Cl, Ca2), and blood glucose levels were
analyzed (Stat Profile M; Noba Medical, Waltham,
MA) at least every 1 h, and these values were maintained within normal ranges by appropriate
treatments.
After surgery, patients in both groups were transferred to the ICU and mechanically ventilated. While
on the ventilator, patients were lightly sedated (Ramsey scale 2; cooperative and tranquil) (21) by continuous infusion of propofol.
CEI of 0.2% ropivacaine with 4 g/mL fentanyl was
administered at 4 mL/h via each epidural catheter
after surgery for postoperative pain control. If a patient complained of pain, an epidural injection of 5 mL

ANESTH ANALG
2005;101:15217

0.2% ropivacaine was administrated as a supplemental analgesic. In group G, patients received an initial
loading dose of morphine (4 5 mg) for pain relief on
arrival at the ICU and a continuous infusion of 1 mg/h
morphine was started. If a patient complained of pain,
a 2.5 mg IV bolus of morphine was administered as a
supplemental analgesic. Pain intensity was assessed at
the end of surgery and every 12 h after surgery with a
box scale (0 10, where 0 is no pain and 10 is the worst
pain ever). The independent observer who was not
involved in this study assessed pain analgesic area by
pinprick and pain intensity.
Arterial blood samples (10 mL) were collected before and at the end of surgery and on POD1 and
POD3. Blood samples were treated with EDTA, and
2 mL of blood was used for the analysis of lymphocyte
subsets and blood cell counts. The rest of the sample
was centrifuged, and the plasma was frozen at 80C
until further analysis. The plasma concentrations of
epinephrine, norepinephrine, cortisol, and adrenocorticotropic hormone (ACTH) were measured, the number of leukocyte was counted, the distribution of lymphocyte subsets was analyzed, and the plasma levels
of interleukin (IL)-1, IL-6, IL-10, tumor necrosis factor (TNF)-, and C-reactive protein (CRP) were measured before and at the end of surgery and on POD1
and POD3.
Lymphocyte subsets were analyzed by flow cytometry (EPICS ELITE; Coulter, Miami, FL) with
fluorescence-labeled antibodies specific to the cell
markers (Coulter). A 0.1-mL blood sample was incubated for 30 min with monoclonal antibodies at 4C in
the dark. The samples were processed with a Q-prep
Immunology Station (Coulter), which lyses the erythrocytes in a semiautomatic fashion, stabilizes the leukocytes, and fixes the cells. The percentage of lymphocytes relative to the total leukocyte count was
determined through differential gating after triplecolor staining. The following antibodies to lymphocyte antigens were used and cell types determined:
cluster of differentiation (CD)3-CD19 (B cells),
CD3CD19- (T cells), CD3CD4 (inducer and
helper T cells), CD3CD8 (suppressor and cytotoxic
T cells), and CD3-CD16CD56 (natural killer [NK]
cells). This method of lymphocyte subset analysis is
accurate, with a high degree of specificity and precision; the coefficient of variation we obtained was
2.5%. The total leukocyte number and the percentage of total lymphocytes were measured with a cell
counter (MAXM-Retic; Coulter).
The plasma concentrations of epinephrine and norepinephrine were measured by high-performance liquid chromatography with electrochemical detection
according to the method described by Weicker et al.
(22). The sensitivity limit of this method was 1 pg/mL
for each catecholamine. Commercially available radioimmunoassay kits were used to measure the plasma

REGIONAL ANESTHESIA
YOKOYAMA ET AL.
IMMUNE RESPONSE AFTER RADICAL ESOPHAGECTOMY

1523

concentrations of ACTH (Allegro HS-ACTH; Nichols

Institute, San Juan Capistrano, CA) and cortisol
(Amerlex Cortisol RIA Kit; Amersham, Arlington
Heights, IL). The sensitivity limits were 1 pg/mL for
ACTH and 0.1 g/dL for cortisol. The plasma concentrations of IL-1, IL-6, IL-10, and TNF- were measured by enzyme-linked immunosorbent assay kits,
and the detection levels were as follows: 0.125 pg/mL
for IL-1, 0.15 pg/mL for IL-6, 0.5 pg/mL for IL-10,
and 0.35 pg/mL for TNF-.
Sample size was determined based on our preliminary study of patients undergoing esophageal surgery. Power calculation had indicated that 15 patients
would be required per group to detect a difference of
25% in proportion of lymphocyte, white cell count,
and plasma levels of epinephrine, cortisol, and IL-1
with a power of 80%. We felt the smallest clinically
significant difference in pain score and in CRP would
be 2 and 2.0 mg/dL, respectively. Two groups of 15
patients were also sufficient to detect both these differences with a power of 80%. Because the values of
ACTH, IL-6, IL-10, and TNF-varied so widely, more
than 100 patients would be required to detect a difference of 25% in these values with a power of 80%. We
therefore showed these differences with an error of
0.05 and an insufficient power. Data are expressed as
mean sd. Students t-test (unpaired) was used to
analyze between-group differences in patients characteristics (age, weight, height), duration of surgery,
amounts of fluid and blood infused, and amount of
blood lost during surgery. Pain scores and plasma
levels of epinephrine, norepinephrine, cortisol, ACTH,
IL-1, IL-6, IL-10, TNF-, and CRP, and the leukocyte
count, and the percentage of lymphocyte subsets were
analyzed by the Kruskal-Wallis test followed by
Dunns method within and between groups. Differences were considered statistically significant at P
0.05.

Results
Patient characteristics, duration of surgery, amounts
of fluid and blood infused, and amount of blood lost
during surgery were similar between the groups (Table 1). Pain scores in group G were significantly higher
than in group E at the end of surgery (P 0.01), but
there was no significant different between groups
thereafter (Fig. 1). The epidural-blocked area at the
end of surgery was not assessed exactly because of
surgical dressing and slight sedation, but no pain at
the cervical, thoracic and abdominal regions was affirmed in all patients of group E. Doses of fentanyl
administered in group E and the dose of morphine
administered in group G for the first 24 h after surgery
were 797 47 g and 35.8 3.4 mg, respectively. The
mean dose of propofol administered for the first 24 h

1524

REGIONAL ANESTHESIA YOKOYAMA ET AL.

IMMUNE RESPONSE AFTER RADICAL ESOPHAGECTOMY

ANESTH ANALG
2005;101:15217

Table 1. Patient Characteristics and Data

Group
E
G

Age
(yr)

Sex
(M/F)

Weight
(kg)

Height
(cm)

Duration of
surgery
(min)

Amount of
blood lost
(mL)

60 8
62 9

13/2
12/3

61 9
60 7

162 10
161 8

616 69
648 66

886 308
927 273

Amount of
Amount of fluid
blood infused
infused
(mL)
(mL)
333 309
360 295

4363 612
4418 791

Values are mean sd, n 15. Group E general anesthesia with epidural anesthesia and epidural analgesia; group G: general anesthesia and IV
patient-controlled analgesia. There were no significant differences between groups.

Figure 1. Pain scores. Pain score in group G is significantly higher

than that in group E at the end of surgery (P 0.01), and this value
is significantly higher than those at other time points (P 0.01).
Values are mean sd, n 15. End-Op end of surgery; POD
postoperative day. **P 0.01 versus group E; P 0.01 versus
other points.

after surgery was similar between groups (1.0

0.1 mg kg1 h1 in group E and 1.0
0.0 mg kg1 h1 in group G). All variables were
similar between groups before surgery.
The plasma concentration of epinephrine increased
significantly in group G (P 0.05) to a level significantly more than that in group E at the end of surgery
(P 0.05; Fig. 2A). The plasma concentration of norepinephrine increased significantly in group G (P
0.01) at the end of surgery and on POD1 and POD3 to
a level significantly more than that in group E (P
0.01; Fig. 2B). The CD4/CD8 ratio decreased significantly at the end of surgery in group G (P 0.05)
to a value significantly less than that in group E (P
0.05; Fig. 5D). However, there were no significant
differences in any other values between the groups
(Figs. 25). In both groups, the following postoperative values differed significantly in comparison to the
preoperative values. The plasma levels of cortisol and
ACTH were increased at the end of surgery (each
group P 0.01; Figs. 2C and D), and IL-1, IL-6, IL-10,
and the CRP levels were increased at the end of surgery and on POD1 and POD3 (each group: P 0.01;
Figs. 3A, B, C and Fig. 4C). The leukocyte count was
increased on POD1 (each group P 0.05) and POD3
(each group P 0.01) (Fig. 4A), whereas the proportion of lymphocytes decreased from the end of surgery
to POD3 (each group P 0.01; Fig. 4B). Among the
lymphocyte subsets, the proportion of B cells was
increased on POD1 (each group P 0.01; Fig. 5A), and
that of NK cells decreased on POD1 and POD3 (each
group P 0.01; Fig. 5C).

Figure 2. Plasma concentrations of epinephrine (A), norepinephrine

(B), and cortisol (C), and adrenocorticotropic hormone (ACTH) (D).
Values are mean sd, n 15. End-Op end of surgery; POD
postoperative day. *P 0.05 versus group E; **P 0.01 versus
group E; P 0.05 versus preoperative; P 0.01 versus
preoperative.

Discussion
Our findings indicate that extensive epidural block by
two epidural catheters was not able to suppress leukocytosis, lymphopenia, increases in cytokines and
CRP, or changes in proportions of lymphocyte subsets
after radical esophagectomy. Thus, it appears that tissue damage and inflammation overcome the effects of
extensive epidural block on stress responses and immune functions in patients undergoing radical
esophagectomy.
There is indirect evidence that increased plasma
norepinephrine changes reflect increased sympathetic
activity during stress (23). The increase in the plasma
level of catecholamines was suppressed and postoperative pain was not observed in group E at the end of
surgery. Thus, CEI via two epidural catheters was able
to suppress activation of the sympathetic and somatic
nervous systems during surgery. However, CEI did
not suppress the increase in plasma levels of cortisol
and ACTH. These results indicate that the stress response was not completely suppressed during surgery
by epidural anesthesia, despite blockade of the nervous system.

ANESTH ANALG
2005;101:15217

Figure 3. Plasma concentrations of interleukin (IL)-1 (A), IL-6 (B),

IL-10 (C), and tumor necrosis factor (TNF)- (D). Values are mean
sd, n 15. End-Op end of surgery; POD postoperative day.
P 0.01 versus preoperative.

Figure 4. Total leukocyte count (A), percentage of lymphocytes (B),

and C-reactive protein (CRP) (C). Values are mean sd, n 15.
End-Op end of surgery; POD postoperative day. P 0.05
versus preoperative; P 0.01 versus preoperative.

It is reportedly difficult to prevent the surgical stress

response in procedures above the umbilicus (9,10).
There are a few likely causes of the incomplete effect
of epidural anesthesia on the stress response to procedures above the umbilicus (19). First, the doses of
local anesthetic used may not be sufficient to produce
complete neural blockade at the surgical site (20). Second, cytokines directly activate the stress response
and are released in larger quantities after surgery
above the umbilicus (24). Segawa et al. (20) reported
that a block of the sensory fibers included in the
phrenic nerves (C35 spinal nerves) is required for a

REGIONAL ANESTHESIA
YOKOYAMA ET AL.
IMMUNE RESPONSE AFTER RADICAL ESOPHAGECTOMY

1525

Figure 5. Distribution of lymphocyte subsets of B cells (A), T cells

(B), natural killer (NK) cells (C), and CD4/CD8 ratio (D). Values
are mean sd, n 15. End-Op end of surgery; POD postoperative day. *P 0.05 versus group E; P 0.05 versus preoperative; P 0.01 versus preoperative.

complete block of the endocrine stress response to

upper abdominal surgery. We used two epidural catheters to provide extensive neural blockade (C3-L2)
because most patients in our preliminary study could
not develop analgesia of C3-L2 by the one catheter
method described by Segawa et al. (20). Although
administration of 1.5% lidocaine provided extensive
neural blockade before surgery, we could not confirm
the blocked area during continuous infusion of 1%
lidocaine with fentanyl. Although the analgesic area
was assessed at the end of surgery, it was difficult to
check the blocked area exactly by pinprick because of
the surgical dressing and incomplete emergence from
anesthesia. The doses may not have been adequate to
provide complete neural blockade at the surgical site
during surgery. However, suppression of catecholamines and absence of pain were achieved.
Taken together, our findings indicate that the direct
effects of cytokines were the likely cause of the surgical stress response after radical esophagectomy.
The postoperative neuroendocrine response to
stress is now believed to be mediated by cytokines
such as IL-1, IL-6, and TNF-, as well as by some of
the more classic stress hormones including cortisol
and the catecholamines (24 26). IL-6 is the primary
stimulus for acute responses, and plasma levels of IL-6
are reportedly related to the severity of surgical
trauma (27). Silverman et al. (28) demonstrated recently that IL-6 can directly stimulate corticosteroid
secretion from the adrenal cortex, and there is evidence suggesting that it can also stimulate ACTH
secretion from the pituitary gland. We found that
plasma levels of IL-6 increased significantly after surgery in both of our groups. Therefore, the increase in

1526

REGIONAL ANESTHESIA YOKOYAMA ET AL.

IMMUNE RESPONSE AFTER RADICAL ESOPHAGECTOMY

stress hormones during surgery appears to be mediated by cytokines that are not suppressed by epidural
blockade. IL-10, which has strong antiinflammatory
and immunoinhibitory activities, was also not affected
by epidural analgesia, and the peak level of IL-10 was
similar to that of IL-6.
CRP was increased significantly in both groups after
surgery; the peak level was observed on POD3. The
consistently high level of CRP after surgery indicates
that surgical inflammation continues for several days
after radical esophagectomy. Volk et al. (5) also reported that PCEA, in comparison with patientcontrolled analgesia (PCA), had no influence on altered levels of circulating cytokines (IL-6, IL-8, IL-10)
or indicators of the stress response (CRP and cortisol)
after major spine surgery.
In our study, leukocytosis and lymphopenia were not
prevented by epidural anesthesia and analgesia. Similar
changes in lymphocyte subsets were observed in both
groups, with the exception of the CD4/CD8 ratio at the
end of surgery. The absence of a significant decrease in
the CD4/CD8 ratio during surgery might have been
attributable to epidural anesthesia. However, the ratio
returned to the presurgical value even in the group
without epidural analgesia, and the values remained at
presurgical levels on POD1 and POD3 in both groups.
The proportion of NK cells decreased in both groups
during and after surgery. Although Volk et al. (5) reported that postoperative epidural anesthesia preserves
lymphocytes after major spine surgery, our findings suggest that CEI only affected immune function slightly
after radical esophagectomy.
There are several limitations to the current study.
First, a small participant size might have affected the
results. We could not provide a sufficient power with
regard to IL-6, IL-10, and TNF-. However, it was
difficult to recruit more than 100 patients scheduled
for radical esophagectomy for short-term study in our
hospital. We, thus, showed these differences with an
error of 0.05 and an insufficient power. Second, it was
impossible to blind this study because one group had
two epidural catheters. However, the independent observer who was not involved in this study collected
pain scores by a box scale. This protocol likely did not
cause a bias between groups. Third, sedation during
intubation/ventilation might have affected results.
However, patients were slightly sedated (Ramsey
scale 2) and were able to give pain scores easily with
a box scale. The dose of propofol administered was
similar in both groups and the dose was small
(1.0 mg kg1 h1), so it was unlikely to cause
changes in pain scores and other valuables.
The present findings indicate that combined
regional/general anesthesia with epidural anesthesia
for blockade of afferent neural impulses does not attenuate stress-induced cytokine production during

ANESTH ANALG
2005;101:15217

and after radical esophagectomy. The use of two separate epidural catheters with independent risks for
complications may remove most clinical relevance.
However, this does not mean that epidural anesthesia
and analgesia by a single catheter are useless for postoperative care. Tsui et al. (17) observed fewer cardiovascular complications, less frequent morbidity and
mortality, and a shorter hospital stay for patients undergoing esophageal surgery with epidural analgesia
compared with patients with PCA. Smedstad et al.
(18) compared postoperative epidural analgesia with
postoperative IV analgesia and reported a reduction in
the time spent in the ICU and in the total time spent in
the hospital for patients treated epidurally. A multimodal approach for patients undergoing abdominothoracic esophagectomy including thoracic epidural
analgesia, early tracheal extubation, and forced mobilization could reduce the ICU stay (16).
In conclusion, our data demonstrated that extensive
epidural block by two catheters was not able to preserve immune function and suppress acute inflammatory responses after radical esophagectomy. We recommend epidural analgesia by a single catheter to
improve recovery after radical esophagectomy as a
multimodal approach.

References
1. Kehlet H. Modification of responses to surgery by neural
blockade: clinical implications. In: Cousins MJ, Bridenbaugh
PO, eds. Neural blockade in clinical anesthesia and management of pain. Philadelphia, JB Lippincott, 1988:14590.
2. Basedovsky HO, Del Rey A. Immune-neuro-endocrine
interactions: facts and hypotheses. Endocr Rev 1996;17:64 102.
3. Watkins LR, Maier SF, Goehler LE. Immune activation: the role
of pro-inflammatory cytokines in inflammation, illness responses, and pathological pain states. Pain 1995;63:289 302.
4. Tonnesen E, Wahlgreen C. Influence of extradural and general
anaesthesia on natural killer cell activity and lymphocyte subpopulations in patients undergoing hysterectomy. Br J Anaesth
1988;60:500 7.
5. Volk T, Schenk M, Voigt K, et al. Postoperative epidural anesthesia preserves lymphocyte, but not monocyte, immune function after major spine surgery. Anesth Analg 2004;98:1086 92.
6. Moor CM, Desborough JP, Powell H, et al. Effects of extradural
anaesthesia in interleukin-6 and acute phase response to surgery. Br J Anaesth 1994;72:2729.
7. Beilin B, Shavit Y, Trabekin E, et al. The effects of postoperative
pain management on immune response to surgery. Anesth
Analg 2003;97:8337.
8. Beilin B, Bessler H, Mayburd E, et al. Effects of preemptive
analgesia on pain and cytokine production in the postoperative
period. Anesthesiology 2003;98:1515.
9. Tsuji H, Asoh T, Takeuchi Y, Shirasaka C. Attenuation of adrenocortical response to upper abdominal surgery with epidural
blockade. Br J Surg 1983;70:122 4.
10. Hjortso NC, Christensen NJ, Andersen T, Kehlet H. Effects of
the extradural administration of local anaesthetic agents and
morphine on the urinary excretion of cortisol, catecholamines
and nitrogen following abdominal surgery. Br J Anaesth 1985;
57:400 6.
11. Kehlet H. The surgical stress response: should it be prevented?
Can J Surg 1991;34:5657.

ANESTH ANALG
2005;101:15217

12. Bone RC. Toward an epidemiology and natural history of SIRS

(systemic inflammatory response syndrome). JAMA 1992;268:
34525.
13. Meakins JL. Surgeons, surgery, and immunomodulation. Arch
Surg 1991;126:494 8.
14. Pollock RE, Lotzova E, Stanford SD. Mechanism of surgical
stress impairment of human perioperative natural killer cell
cytotoxicity. Arch Surg 1991;126:338 42.
15. Eggermont AM, Steller EP, Sugarbaker PH. Laparotomy enhances intraperitoneal tumor growth and abrogates the antitumor effects of interleukin-2 and lymphokine-activated killer
cells. Surgery 1987;102:71 8.
16. Brodner G, Pogatzki E, Van Aken H, et al. A multimodal approach to control postoperative pathophysiology and rehabilitation in patients undergoing abdominothoracic esophagectomy. Anesth Analg 1998;86:228 34.
17. Tsui SL, Law S, Fok M, et al. Postoperative analgesia reduces
mortality and morbidity after esophagectomy. Am J Surg 1997;
173:472 8.
18. Smedstad KG, Beattie WS, Blair WS, Buckley DN. Postoperative
pain relief and hospital stay after total esophagectomy. Clin J
Pain 1992;8:149 53.
19. Liu S, Carpenter RL, Neal JM. Epidural anesthesia and analgesia. Their role in postoperative outcome. Anesthesiology 1995;
82:1474 506.
20. Segawa H, Mori K, Kasai K, et al. The role of the phrenic nerves
in stress response in upper abdominal surgery. Anesth Analg
1996;82:121524.

REGIONAL ANESTHESIA
YOKOYAMA ET AL.
IMMUNE RESPONSE AFTER RADICAL ESOPHAGECTOMY

1527

21. Ramsey MAE, Savage TM, Simpson BRJ, Goodwin R. Controlled sedation with alfaxolone-alphadolone. Brit Med J 1974;
2:656 9.
22. Weicker H, Feraudi M, Hagele H, Pluto R. Electrochemical
detection of catecholamines in urine and plasma after separation with HPLC. Clin Chim Acta 1984;141:1725.
23. Bravo EL, Tarazi RC. Plasma catecholamines in clinical
investigation: a useful index or a meaningless number? J Lab
Clin Med 1982;100:155 60.
24. Naito Y, Tamai S, Shingu K, et al. Responses of plasma adrenocorticotropic hormone, cortisol, and cytokines during and after
upper abdominal surgery. Anesthesiology 1992;77:426 31.
25. Sapolsky R, Rivier C, Yamamoto G, et al. Interleukin-1 stimulates the secretion of hypothalamic corticotropin-releasing factor. Science 1987;238:522 4.
26. Baumann H, Prowse KR, Marinkovic S, et al. Stimulation of
hepatic acute phase response by cytokines and glucocorticoids.
Ann NY Acad Sci 1989;557:280 95.
27. Cruickshank AM, Fraser WD, Burns HJ, et al. Response of
serum interleukin-6 in patients undergoing elective surgery of
varying severity. Clin Sci 1990;79:1615.
28. Silverman MN, Miller AH, Biron CA, Pearce BD. Characterization of an interleukin-6 and adrenocorticotropin-dependent,
immune-to-adrenal pathway during viral infection. Endocrinology 2004;145:3580 9.

BRIEF REPORT

An In Vitro Comparison of the Electrical Conducting

Properties of Multiport Versus Single-Port Epidural
Catheters for the Epidural Stimulation Test
Ban C. H. Tsui,

MD, MSc, FRCP*

and Corey K. C. Sze

*Department of Anesthesiology and Pain Medicine, University of Alberta, Edmonton, Alberta, Canada; Faculty of
Engineering, University of Alberta, Edmonton, Alberta, Canada

Effective conduction of electricity through a catheter is

essential for the success of the epidural stimulation test.
In this in vitro study we examined the electrical conductivity of single and multiport epidural catheters (with
and without embedded metal elements) after being
primed with normal saline. Seven different types of 19gauge catheters (n 5), either single-port or multiport
catheters, with or without embedded metal elements,
were studied. The proximal end of each epidural catheter was connected to the cathode of a nerve stimulator
via an electrode adapter. The catheter, primed with normal saline, was placed at the bottom of a syringe filled
with 5 different volumes of saline (1, 2, 3, 4 and 5 mL)
and attached to an electrode adapter. The voltage of the
peripheral nerve stimulator was measured using an oscilloscope. The electrical resistance between the proximal and distal end of the catheter was calculated using
Ohms Law. In catheters without metal elements the

he use of epidural electrical current (110mA) to

confirm epidural catheter location is a useful
technique in pediatric patients (13). All previous studies on the epidural stimulation test used
single-port metal-coil reinforced epidural catheters
(Flexitip, Arrow International, Reading, PA) (13).
However, multiport catheters with similar metal elements are now available. Because there is clinical evidence suggesting that multiport catheters provide a
more even spread of local anesthetic within the epidural space (4,5), it is important to examine the suitability of these multiport catheters for transmitting
Supported, in part, by a Clinical Investigatorship Award, Alberta
Heritage Foundation for Medical Research, Alberta, Canada.
Accepted for publication May 31, 2005.
Address correspondence and reprint requests to Ban C.H. Tsui,
MSC, MD, FRCP(C), Department of Anesthesiology and Pain Medicine, 8 120 Clinical Sciences Building, University of Alberta, Edmonton, Alberta, Canada T6G 2G3. Address electronic mail to
btsui@ualberta.ca.
DOI: 10.1213/01.ANE.0000181006.36917.3E

1528

Anesth Analg 2005;101:152830

electrical resistances were too high to be measured. In

catheters that had metal elements, the mean electrical
resistances of the same catheter design (single-port or
multiport) were similar. However, the electrical resistances of the multiport metal reinforced epidural catheters were significantly lower (P 0.05) than the singleport metal coil reinforced epidural catheters. The
volume of saline in the syringe had no impact on the
measured electrical resistances. This study suggests
that multiport metal reinforced epidural catheters have
low electrical resistances and, thus, are a reasonable alternative to single-port catheters for transmitting sufficient current for performing the epidural stimulation
test. On the other hand, epidural catheters without
metal elements (single-port or multiport) are not suitable for performing the stimulation test.
(Anesth Analg 2005;101:1528 30)

sufficient current when performing the epidural stimulation test. This in vitro study examines if multiport
epidural catheters, with or without metal elements,
can transmit sufficient current (10 mA) for the epidural stimulation test.

Methods
Five samples of 7 different 19-gauge epidural catheters
were studied: 1) single-port nylon catheters from BectonDickinson (Deseret Medical, Sandy, UT); 2) multiport
nylon catheters from Becton-Dickinson; 3) single-port
metal reinforced catheters from Arrow International; 4)
single-port metal reinforced catheters from B. Braun (Bethlehem, PA); 5) multiport metal reinforced catheters
from B. Braun; 6) single-port metal reinforced catheters
from Portex (Sims, Markham, Ontario, Canada); and 7)
multiport metal reinforced catheters from Portex. The
proximal end of the epidural catheter was connected to
the cathode of a nerve stimulator via an electrode
adapter (Johans ECG Adapter, Arrow International Inc.).
2005 by the International Anesthesia Research Society
0003-2999/05

ANESTH ANALG
2005;101:1528 30

BRIEF REPORT

1529

Table 1. Mean Calculated Electrical Resistances of

19-Gauge Epidural Catheters

Figure 1. Experimental setup.

To try to mimic situations in which a catheter tip would

be located at varying distances from the nerve, the syringe was filled with 5 different volumes of normal
saline (1, 2, 3, 4, and 5 mL) and the catheter tip was
placed at the bottom of the syringe (Fig. 1). We ensured
that all ports were immersed in the fluid. The syringe
was in turn attached to an electrode adapter. The intensity of current delivered by the nerve stimulator was
calibrated before the experiments. The stimulator was set
at an intensity of current of 10 mA, a pulse width of 0.2
ms, and a frequency of 1 Hz. The current output of the
peripheral nerve stimulator was measured using a
factory-calibrated oscilloscope (Tektronix Inc., Beaverton, OR). The electrical resistance was calculated using
Ohms Law (r V/I), where V was the measured voltage (volts) and I was the current output from the nerve
stimulator (MaxiStim Model ST5; Life-Tech, Inc., Stafford, TX). Data are presented as mean sd. A one-way
analysis of variance test, followed by a post hoc TurkeyKramer multiple comparisons test was performed. Statistical significance was determined as P 0.05.

Results
The electrical resistances of the various single-port
and multiport epidural catheters are summarized in
Table 1. A decrease in voltage could not be measured
in the nylon catheters from Becton-Dickinson, probably because of their high electrical resistances. For all
other catheters with metal elements, the electrical resistances of these catheters remained constant (i.e., no
change), despite changing the volume of saline in the
syringe from 1 to 5 mL. The electrical resistances of the
multiport catheters from B. Braun were significantly
(P 0.05) less than single port catheters from Arrow,
B. Braun, and Portex. Multiport catheters from Portex
also demonstrated the same statistically significant
result (P 0.05). However, there was no statistically

Brand of catheter

Metal
coil

Single-port
(Kohm)

Multiport
(Kohm)

Becton-Dickinson
Arrow

No
Yes

N/A

Mean sd
B. Braun

Yes

Mean sd
Portex

Yes

N/A
16.6
17.1
17.6
17.4
16.8
17.1 0.4
16.0
16.2
16.9
16.3
17.2
16.5 0.5
17.3
17.7
17.5
17.0
17.2
17.3 0.3

Mean sd

13.2
12.5
13.1
13.2
11.9
12.8 0.6*
13.9
13.1
12.8
13.8
13.8
13.5 0.5

N/A: The electrical resistance was too high to be calculated with a 10-mA
current. The electrical resistance of all metal-coiled catheters remained constant despite varying volumes of saline (1 to 5 mL).
* A one-way ANOVA, followed by a post-hoc test Turkey-Kramer multiple comparisons test was performed. The electrical resistances of the multiport catheters from B. Braun were significantly (p 0.05) less than single-port
catheters from Arrow, B. Braun and Portex.
The electrical resistances of the multiport catheters from Portex were
significantly (P 0.05) less than single-port catheters from Arrow, B. Braun
and Portex. However, there was no statistically significant difference among
the different single-port catheters or among the different multiport catheters.

significant difference among the different single-port

catheters or among the different multiport catheters.

Discussion
This is the first in vitro study to examine the electrical
conductivity of single and multiport epidural catheters (with and without embedded metal elements)
after they had been primed with normal saline. Multiport metal reinforced catheters are a reasonable alternative to single-port 19-gauge metal reinforced catheters for transmitting sufficient current (10 mA) for
performing the epidural stimulation test. On the other
hand, regular nylon epidural catheters without metal
elements (single-port or multiport) are not suitable for
performing the stimulation test.
In the original technique described by Tsui et al.
(6,7), the stimulating current (0.2 ms; 1 Hz) was conducted through the injected normal saline into the
epidural space via an electrically conducting metal
coil embedded in the catheter. Although saline can
conduct electric current, the high impedance of this
fluid through the lengthy, narrow epidural catheter
hinders the flow of current. Any air lock within the
lumen will also deter current flow. As all portable

1530

BRIEF REPORT

nerve stimulators have limited voltage ranges, low

electrical resistance catheters are essential to allow
sufficient current (10 mA) to be conducted from the
source to the tip for performing the epidural stimulation test. This was evident in this experiment where
the high electrical resistance regular nylon epidural
catheters (single-port or multiport) could not transmit
adequate current (Table 1).
Our results demonstrated that the volume of saline
in the syringe had no impact on the measured electrical resistance between the two conducting points. This
may be because the narrow catheter opening accounted for most of the electrical resistance and was
therefore not significantly affected by small volume
changes in the highly conductive normal saline. In
contrast, the additional holes in the new multiport
metal-coil reinforced catheters allowed several contact
metal points for the conducting saline and, thus, significantly decreased the electrical resistance of these
catheters compared with single-port catheters. Therefore, as regards electrical conduction, new multiport
metal reinforced catheters are a reasonable alternative
to 19-gauge metal reinforced catheters for transmitting

ANESTH ANALG
2005;101:1528 30

sufficient current to perform the epidural stimulation

test. Conventional plastic catheters are not suitable for
performing the epidural stimulation test.

References
1. Tsui BC, Seal R, Koller J, et al. Thoracic epidural analgesia via the
caudal approach in pediatric patients undergoing fundoplication
using nerve stimulation guidance. Anesth Analg 2001;93:11525.
2. Tsui BC, Finucane B. Verifying accurate placement of an epidural
catheter tip using electrical stimulation. Anesth Analg 2002;94:
1670 1.
3. Tsui BC, Wagner A, Cave D, Kearney R. Thoracic and lumbar
epidural analgesia via the caudal approach using electrical stimulation guidance in pediatric patients: a review of 289 patients.
Anesthesiology 2004;100:6839.
4. DAngelo R, Foss ML, Livesay CH. A comparison of multiport
and uniport epidural catheters in laboring patients. Anesth
Analg 1997;84:1276 9.
5. Dickson MA, Moores C, McClure JH. Comparison of single,
end-holed and multi-orifice extradural catheters when used for
continuous infusion of local anaesthetic during labour. Br J Anaesth 1997;79:297300.
6. Tsui BC, Finucane B. Epidural stimulator catheter. Tech Reg
Anesth Pain Manage 2002;6:150 4.
7. Tsui BC, Gupta S, Finucane B. Confirmation of epidural catheter
placement using nerve stimulation. Can J Anaesth 1998;45:640 4.

CASE REPORT

Continuous Maxillary and Mandibular Nerve Block for

Perioperative Pain Relief: The Excision of a Complicated
Pleomorphic Adenoma
Ashok Kumar,

MD

and Ashim Banerjee,

MD

Department of Anaesthesiology and Critical Care, UCMS and GTB Hospital, Delhi, India

A 68-yr-old hypertensive patient with ischemic heart

disease and intractable atrial fibrillation with stable
hemodynamics and poor chest compliance underwent pleomorphic adenoma of the parotid gland using regional anesthesia with continuous propofol sedation. Continuous maxillary and mandibular nerve

axillary and mandibular nerve blocks are often

performed for diagnostic and therapeutic purposes. We report the application of continuous
maxillary and mandibular nerve blocks for intraoperative and postoperative analgesia in a patient scheduled for the excision of pleomorphic adenoma of the
parotid gland.

Case Report
A 68-yr-old 56-kg male patient with ischemic heart disease,
intractable atrial fibrillation, and controlled hypertension
with a pleomorphic adenoma of the parotid gland was
scheduled for total parotidectomy. The patients difficult
airway was Mallampatti grade III. The preoperative electrocardiogram (EEG) showed ST elevation in V1V4 leads and
left ventricular hypertrophy. Echocardiography revealed a
markedly reduced left ventricular ejection fraction of 36%
with significant regional wall motion abnormality in the
anterior wall.
A regional anesthetic technique was selected because of
the patients pulmonary and cardiac conditions, as well as
his difficult airway.
The patient was familiarized with the visual analog scale
pain scoring system technique (0 10 cm scale) and written
informed consent for blocks was obtained.
Under aseptic technique, after local infiltration with lidocaine
1% at the midpoint of the zygomatic arch, a 16-gauge IV
cannula was inserted at the midpoint of the lower margin of
Accepted for publication June 14, 2005.
Address correspondence and reprint requests to Ashok Kumar, MD,
Flat # 5, Parivar Apartments Plot # 30, I. P Extension Patparganj, Delhi
110092 India. Address e-mail to prof_ashok@rediffmail.com.
DOI: 10.1213/01.ANE.0000181332.74791.FC
2005 by the International Anesthesia Research Society
0003-2999/05

blocks were performed and excision was performed

with complete intraoperative and postoperative pain
relief without compromising the patients hemodynamic variables.
(Anesth Analg 2005;101:15312)

the zygomatic arch and advanced perpendicularly until it contacted the lateral pterygoid plate. For right maxillary nerve
block, the cannula was then withdrawn slightly and advanced
cephaloanteriorly 1 cm to enter the pterygopalatine fossa (Figure 1). The stylet was removed and an 18-gauge epidural
catheter (Portex) was advanced 0.5 cm past the cannula tip. The
cannula was removed, and the catheter was anchored and a
filter was attached.
After negative aspiration, a 2-mL test dose of lidocaine 2%
with epinephrine (1 in 200,000) was injected. As there was
no evidence of intravascular injection, 8 mL of 0.25% bupivacaine was subsequently administered through the catheter.
To provide a right mandibular nerve block, a 16-gauge IV
cannula was inserted to contact the lateral pterygoid plate.
The depth of the pterygoid plate was noted, and the cannula
was withdrawn and redirected slightly posteriorly (Fig. 2) to
a position just behind the posterior border of the lateral
pterygoid plate.
The cannula was advanced by a distance of 5 mm further,
the stylet was removed, and an 18-gauge catheter was inserted 1.0 cm past the cannula tip and anchored. Injection of
bupivacaine 0.25% 6 mL was performed through the catheter after a negative test dose.
The surgical field was evaluated for satisfactory anesthesia
by pinprick. The patient was sedated with propofol 70 mg IV
bolus followed by an infusion at a rate of 50
g kg1 min1. Surgery was allowed to proceed.
The patients vital signs remained normal throughout surgery. The pattern of atrial fibrillation persisted but with no
acute changes in the ECG. The patient was sedated but
responding to verbal commands. The propofol infusion was
stopped at the end of the surgery, the patient awoke completely pain-free.
In the evening, the visual analog scale score was 1. A
top-up dose of 4 mL 0.25% bupivacaine was administered
through each of the catheters at 8 pm (approximately 5.5 h
after the initial bolus) when the visual analog scale score was
3.5 and repeated every 12 h for the next 72 h. The visual

Anesth Analg 2005;101:15312

1531

1532

CASE REPORT

Figure 1. Continuous maxillary nerve block. i) Cannula being inserted perpendicularly at midpoint of the lower margin of zygomatic arch and advanced to contact the lateral pterygoid plate. ii)
Subsequent cephaloanterior catheter placement 0.5 cm into the
pterygopalatine fossa.

analog scale score noted throughout the 72-h postoperative

period ranged between 1.5 and 3.0. The patient was fed by
Ryles tube. Jaw opening was adequate and pain-free.
The two catheters were removed 72 h postsurgery approximately 4 h after the last injection with no evidence of any
sensory or motor deficit or any local infection in the region.

Discussion
In this patient, continuous maxillary and mandibular
nerve blocks, using 0.25% bupivacaine injected via
18-gauge epidural catheters, provided adequate intraoperative and postoperative analgesia.
Continuous mandibular nerve blocks have been
performed for postoperative analgesia after surgical
repair of a fractured mandible (1) and for management
of trigeminal neuralgia (2). Our patient is the first to
receive continuous maxillary and mandibular nerve
blocks for intraoperative anesthesia and postoperative
analgesia.
In the past few years, continuous nerve blocks have
enjoyed a significant surge of interest, especially for
major shoulder procedures (continuous interscalene),
for major surgeries of the upper extremity below the
shoulder (continuous infraclavicular and axillary),
and for major lower extremity surgery (continuous
sciatic, femoral, and lumbar plexus blocks) (3). These
techniques provide superior analgesia and encourage
early mobilization and restoration of limb function.
When performing a maxillary nerve block, it must
be considered that the pterygopalatine fossa is extremely vascular (1) as a result of pterygoid plexus of

ANESTH ANALG
2005;101:15312

Figure 2. Continuous mandibular nerve block. i) Cannula inserted

perpendicularly to contact the lateral pterygoid plate. ii) Cannula
redirected posteriorly in a horizontal plane and advanced approximately 5 mm to final position just behind the posterior border of
the lateral pterygoid plate. The catheter was advanced 1.0 cm past
the cannula tip.

veins and that there is a possibility of intravascular

injection or hematoma formation. We prefer the lateral
extraoral approach for performance of mandibular
block, as (1) it is more convenient to place a catheter by
this approach.
Finally, proper caution must be exercised to prevent
hematoma formation, infection, kinking, or obstruction of the catheter. It was important in our exercise to
keep the catheter as close to the target nerve as possible to ensure a proper anesthetic effect. As in our
patient there was a possibility that the catheter tip may
have been displaced by chewing, jaw opening, and
eating (2), proper position of the tip was confirmed by
regular assessment of pain relief after the top-up dose.
In conclusion, continuous maxillary and mandibular nerve block is an excellent approach for obtaining
adequate intraoperative and postoperative analgesia.
However, a randomized trial is needed to determine
the potential applications of this technique.

References
1. Singh B, Bhardwaj V. Continuous mandibular nerve block for
pain relief: a report of two cases. Can J Anaesth 2002;49:9513.
2. Umino M, Kohase H, Ideguchi S. Sakurai N. Long term pain
control in trigeminal neuralgia with local anaesthetics using an
indwelling catheter in mandibular nerve. Clin J Pain 2002;18:
196 9.
3. Morris GF, Lang SA. Continuous parasacral sciatic nerve block:
two case reports. Reg Anesth 1997;22:469 72.

GENERAL ARTICLES

The Feasibility of Laryngoscope-Guided Tracheal Intubation

in Microgravity During Parabolic Flight: A Comparison of
Two Techniques
Gernot E. Groemer*, Joseph Brimacombe, Thorsten Haas, Cristina de Negueruela,
Alexander Soucek, Michael Thomsen, and Christian Keller
*Institute of Astrophysics and Department of Anaesthesia and Intensive Care Medicine, Leopold-Franzens University,
Innsbruck, Austria; Department of Anaesthesia and Intensive Care, Cairns Base Hospital, Australia; Space Mission
European Space Agency, European Space Research and Technology Centre, Noordwijk, The Netherlands; European
Space Agency, European Space Research Institute, Frascati, Italy; and Technical University of Denmark, Denmark

We determined the feasibility of laryngoscope-guided

tracheal intubation (LG-TI) in microgravity obtained
during parabolic flight and tested the hypothesis that
LG-TI is similarly successful in the free-floating condition, with the patients head gripped between the anesthesiologists knees, as in the restrained condition, with
the torso strapped to the surface. Three personnel with
no experience in airway management or microgravity
participated in the study. LG-TI of a sophisticated fullsize manikin was attempted on seven occasions in each
condition by each investigator after ground-based
training. The parabolic flights, which took place in an
Airbus 300 over the Atlantic Ocean, provided 23 s of
microgravity. During this time, the investigator opened
a box with airway equipment, performed LG-TI, and
attached and held onto a self-inflating bag. The efficacy

here is a potential need for airway management in microgravity because astronauts are at
an increased risk of cardiopulmonary arrest,
and surgery may be required during extended spaceflight,
such as a mission to Mars (1,2). In 1978, LeJeune (3) hypothesized that laryngoscope-guided tracheal intubation
(LG-TI) would be difficult in microgravity because the anterior force exerted by the laryngoscope causes the head
and neck to move out of the field of view, and the hand not
holding the laryngoscope cannot synchronously stabilize
the head and neck and insert the tracheal tube. In 2000, our
group tested this hypothesis using a deep pool to simulate

Accepted for publication May 23, 2005.

Address correspondence and reprint requests to J. Brimacombe,
Department of Anesthesia and Intensive Care, Cairns Base Hospital,
The Esplanade, Cairns 4870, Australia. Address e-mail to
jbrimaco@bigpond.net.au.
DOI: 10.1213/01.ANE.0000181001.25777.53
2005 by the International Anesthesia Research Society
0003-2999/05

of ventilation was assessed during level flight by

squeezing the bag and noting whether the manikin sensors indicated a tidal volume 300 mL. There were no
differences in ventilation success (41% versus 33%) or
time to successful insertion (both 18 s) between the freefloating and the restrained conditions. More than 90%
of failures were caused by the inability to insert the tracheal tube within 23 s. There were no differences in performance among investigators. We conclude that LG-TI
is feasible in microgravity obtained during parabolic
flight, but the success rate is infrequent because of severe time restrictions. There were no differences in success rate between the free-floating condition, with the
head gripped between the knees, and the restrained
condition, with the torso strapped to the surface.
(Anesth Analg 2005;101:15335)

microgravity and found that the success rate for anesthesiologists in the free-floating condition was 15%, increasing
to 92% if the manikin was strapped to a surface using
restraints (4). A major limitation of restraints is the time
taken to apply them during a cardiac arrest. Potentially
faster options include the use an extraglottic airway device,
a self-retaining laryngoscope, or to grip the patients head
between the anesthesiologists knees. Also, there are no
data about the feasibility of airway management in true as
opposed to simulated microgravity. In the following manikin study, we determined the feasibility of LG-TI in microgravity during parabolic flight and tested the hypothesis
that LG-TI is similarly successful in the free-floating condition, with the head gripped between the knees, and in the
restrained condition, with the torso strapped to a surface.

Methods
Research approval was obtained from the University
of Innsbruck, the European Space Agency, and the
Anesth Analg 2005;101:15335

1533

1534

GROEMER ET AL.
INTUBATION DURING PARABOLIC FLIGHT

ANESTH ANALG
2005;101:15335

Table 1. Ventilation Success and Time to Successful

Insertion with the Manikin Floating Free with its Head
Between the Knees (Free-floating condition) and Attached
to the Floor with Restraints (Restrained condition)
Free-floating Restrained
condition
condition
n 22
n 21
Ventilation success
Time to successful insertion (s)
Cause of failure
Failure to insert
Malposition

9 (41)
18 3

7 (33)
18 3

11 (50)
2 (9)

15 (67)
0 (0)

Data are numbers (%) or mean sd.

Failed insertions excluded from time data. Restraints applied before
microgravity.
No significant differences between groups.

Figure 1. Laryngoscope-guided tracheal intubation with the manikin floating free with the head between the knees.

French Ministry of Defense. One woman and 2 men

(age, 26 29 yr; weight, 62 80 kg; height, 174 185 cm)
without airway management or SCUBA diving experience participated in the study. Training was in five
stages. Stage 1 involved 8 h of theory about airway
anatomy and physiology. Stage 2 involved 100 supervised LG-TIs of a standard advanced life support
(ALS) manikin (Laerdal International A/S, Kopenhagen, Denmark). Stage 3 involved two supervised LGTIs of anesthetized patients. Stage 4 involved a certified course in SCUBA diving. Stage 5 involved 10
supervised LG-TIs of a full-sized manikin submerged
in a freshwater pool using neutrally buoyant equipment and personnel: five in the free-floating condition,
in which the manikin was held for intubation by gripping its head between the knees (Fig. 1), and five in the
restrained condition, in which the manikin was held
for intubation by a Velcro restraint attached to the
inner surface of the aircraft similar to that attached to
the inner surface of the International Space Station and
Space Shuttle.
All investigators were given 0.4 mg of oral scopolamine to prevent motion sickness. The manikin was a
customized, full-body, 50-kg ALS manikin that contained thoracic wall sensors that measured the efficacy
of ventilation. The parabolic flights took place within
1 wk of completion of the training program. The parabolic flights were performed on three consecutive
days in June 2004 on an Airbus 300 flown over the
North Atlantic Ocean. The parabolic flight cycle lasted
approximately 3 min and comprised (a) rapidly pulling up at 1.8 g from 6100 to 8500 m at an angle of
47 degrees, (b) reducing the throttle so that the plane
was in free fall for 23 s, (c) pulling out when the
descent angle was 42 degrees, and (d) flying level for
1.5 min before beginning the cycle again.

The airway management equipment, which consisted of a size 7.5-mm tracheal tube (Mallinckrodt
Medical, Lo-Contour, Athlone, Ireland), a 20-mL syringe, a self-inflating bag, and a size 3 Macintosh
laryngoscope, was in a sealed box adjacent to the
manikin. All investigators attempted LG-TI on seven
occasions in each of the 2 conditions in random order.
During level flight, the investigators positioned themselves above the head of the manikin. The attempt
began at the end of the pull-up phase when microgravity commenced. Once inserted, the cuff was inflated with air 10 mL and the self-inflating bag attached. The tracheal tube and self-inflating bag were
then held in place during the pull-out phase to prevent
dislodgement. The efficacy of ventilation was assessed
by a trained observer during level flight by squeezing
the bag and noting whether the manikin sensors indicated a tidal volume 300 mL. All insertions were
recorded using a video camera placed in a fixed position. The cause of failed insertion and the time taken to
insert were assessed on the ground by analyzing video
recordings taken from a fixed position during the
parabolic flight sequence. Insertion time was from
when the investigator opened the sealed box to attachment of the self-inflating bag.
Statistical analysis was with paired t-test for the
time to successful insertion and 2 test for the ventilation success rate. Unless otherwise stated, data are
presented as mean sd. Significance was taken as P
0.05.

Results
There were no differences in ventilation success or
time to successful insertion between the free-floating
condition and the restrained condition (Table 1). More
than 90% of failures were caused by an inability to
insert the tracheal tube within 23 s. There were no
differences in performance among investigators.

ANESTH ANALG
2005;101:15335

Discussion
The success rate for LG-TI was similarly infrequent for
inexperienced personnel in the free-floating condition,
with the manikins head gripped between the knees,
and in the restrained condition, with the manikin
strapped to a surface. The percentage rate for failed
attempts is not surprising because the average time
taken to perform LG-TI in simulated microgravity in a
deep pool is approximately 35 seconds (4). We suspect
that success rates would have been improved if microgravity had lasted a minute. We suggest that gripping the patients head between the knees should be
adopted by astronauts attempting LG-TI during cardiopulmonary resuscitation in microgravity and that
this be included in their training program.
Alternative strategies for single-handed tracheal intubation in the free-floating condition include the use
of a self-retaining laryngoscope blade or the use of an
airway intubator, such as the intubating laryngeal
mask airway (4). Neither of these techniques has been
tested in true or simulated microgravity. An alternative strategy is to use an extraglottic airway device,
and our group found that the classic and intubating
laryngeal masks and cuffed oropharyngeal airway
had more frequent success than LG-TI in simulated
microgravity (4). Perhaps LG-TI is too difficult for the
infrequent user. Trips to Mars may take as many as
three years.
Our study has two limitations. First, we did not
compare LG-TI in the free-floating condition with or
without the knee grip. However, our previous study
showed that the free-floating condition without the
knee grip had much less success than the restrained
condition (4). Second, in the restrained condition, the
straps were applied before microgravity commenced,

GROEMER ET AL.
INTUBATION DURING PARABOLIC FLIGHT

1535

making the time comparisons biased against the freefloating condition. We attached the straps before microgravity because of the limited time frame. Data
from our previous study in simulated microgravity
indicated that strap application takes 510 seconds,
making it likely that the free-floating position would
be quicker than the restrained condition (4).
Finally, our study highlights the time difficulties
of assessing airway management techniques during
parabolic flight. Nonetheless, given the limitations of
simulated microgravity environments (such as the
deep pool and neutrally buoyant equipment) and the
staggering expense of flying higher parabolas (such as
could be achieved in SpaceShipOne (5)), we feel further studies are worthwhile.
We conclude that LG-TI is feasible in microgravity
obtained during parabolic flight, but success is infrequent because of severe time restrictions. There were
no differences in success rate between the free-floating
condition, with the manikins head gripped between
the knees, and in the restrained condition, with the
torso strapped to a surface (Fig. 1).

References
1. Kirkpatrick AW, Campbell MR, Novinkov OL, et al. Blunt
trauma and operative care in microgravity: a review of microgravity physiology and surgical investigations with implications
for critical care and operative treatment in space. J Am Coll Surg
1997;184:44153.
2. Campbell MR, Billica RD, Jennings R, Johnston S. Laparoscopic
surgery in weightlessness. Surg Endosc 1996;10:1117.
3. LeJeune FE. Laryngeal problems in space travel. Aviat Space
Environ Med 1978;49:13479.
4. Keller C, Brimacombe J, Giampalmo M, et al. Airway management during spaceflight: a comparison of four airway devices in
simulated microgravity. Anesthesiology 2000;92:1237 41.
5. McKee M. Pioneering private space flight. New Sci 2004;2479:18.

Alkalinization of Intracuff Lidocaine: Efficacy and Safety

Jean-Pierre Estebe, MD, PhD*, Marc Gentili, MD, PhD*, Pascal Le Corre, PharmD, PhD,
Gilles Dollo, PharmD, PhD, Francois Chevanne, BSc, and Claude Ecoffey, MD*
*Service dAnesthesie Reanimation Chirurgicale 2; Laboratoire de Biopharmacie, UPRES EA 3892, Universite Rennes 1,
Polyclinique Saint Vincent, Rennes, France

When alkalinized lidocaine instead of air is used to fill

the endotracheal tube (ETT) cuff, coughing, and bucking are decreased during extubation when ventilation is
controlled with N2O. However, sodium bicarbonate
(NaHCO3) used to transform lidocaine hydrochloride
(L-HCl) to lidocaine base induces a pH increase that
could be irritating for mucosa in the case of cuff rupture.
Therefore, we determined, in a randomized controlled
study with controlled patient ventilation without N2O,
whether the smallest concentrations of NaHCO3 (1.4%
versus 8.4%) reduced diffusion (in vitro evaluation) and
other secondary clinical benefits. After pH determination of different solutions (2 mL of 2% L-HCl and 2 to
6 mL of 8.4%, or 1.4% NaHCO3), an in vitro lidocaine
diffusion through the ETT cuffs was evaluated (2 mL of
2% L-HCl and 3 mL of 8.4% or 1.4% NaHCO3). Then,
adult patients scheduled for total thyroidectomy surgery were consecutively enrolled (n 20 for each
group). The ETT cuff was filled with air (group air) or
with alkalinized lidocaine (2 mL of 2% L-HCl) using
8.4% (group large dose) or 1.4% (group small dose) of

n endotracheal tube (ETT) induces emergence

phenomena. Controlling the pressure of the cuff
and inflation of the cuff by filling it with nitrous
oxide (N2O) (1) with close monitoring of N2O concentration (2) were proposed to improve ETT cuff tolerance. Deflation of the cuff (3) and filling the cuff with
saline (4) or with lidocaine (L-HCl) have been also
recommended (59). However, a very small amount
(1%) of L-HCl diffuses through the ETT cuff (10).
Hence, large L-HCl doses (200 to 500 mg) have been
used (7), which may be dangerous if the cuff ruptures
as a result of L-HCl vascular absorption. The addition
of 8.4% of sodium bicarbonate (NaHCO3) (i.e., alkalinization) to small doses of L-HCl (40 mg) increased

Accepted for publication May 12, 2005.

Address correspondence and reprint requests to Jean-Pierre Estebe, MD, PhD, Service dAnesthesie Reanimation Chirurgicale 2,
Hopital Hotel Dieu: 2 rue de lHotel Dieu, 35000, Rennes, France.
Address e-mail to jean-pierre.estebe@chu-rennes.fr.
DOI: 10.1213/01.ANE.0000180995.24211.89

1536

Anesth Analg 2005;101:153641

NaHCO3. After tracheal extubation, sore throat was

evaluated by visual analog scale as the main end-point
of the study. Hoarseness, bucking, dysphonia, dysphagia, cough, restlessness, and postoperative nausea and
vomiting were also evaluated. There was a slight tendency toward a slower release when a small concentration of NaHCO3 was used (i.e., 1.4%). Compared with
group air, the alkalinized-lidocaine groups had a significant reduction in sore throat during the 24-h postoperative period (P 0.0001). The difference was not significant between the two alkalinized lidocaine groups.
This increase in ETT tolerance was confirmed by the
analysis of secondary end-points. No laryngospasm,
rupture of ETT cuff, or depression of the swallowing
reflex were recorded. A decrease in sore throat during
the postoperative period was recorded when the cuff
was inflated with a small dose of alkalinized lidocaine
(i.e., 40 mg of L-HCl and 1.4% of NaHCO3) rather than
with air when ventilation was controlled without N2O.
(Anesth Analg 2005;101:1536 41)

diffusion of L-HCl through the ETT cuff (11,12). However, the clinical relevance of alkalinized L-HCl was
never evaluated during anesthesia without N2O.
Although L-HCl is used clinically as spray or jelly, its
pH, reported around 5, could be irritating for tracheal
mucosa during clinical use or in case of rupture of a cuff
filled with L-HCl (13). NaHCO3 is necessary to transform L-HCl in lidocaine-base to increase diffusion
through the ETT cuff (i.e., 65% of diffusion for 6 h for
hydrophobic neutral form versus 1% of diffusion for
charged L-HCl) (10). On the other hand, the pH of some
NaHCO3 and L-HCl mixtures, in the high range of human physiology, could be irritative if a cuff ruptures
(14). The aim of this study was to determine a mixture
with the most efficient diffusion of L-HCl together with
the most physiological pH in case of cuff rupture.
To evaluate the consequences of using different concentrations and volumes of NaHCO3, we first performed an in vitro evaluation. To evaluate the local
anesthetic effect in vivo of using alkalinized L-HCl, we
conducted a double-blind randomized on three parallel trial groups (i.e., control group with air and two
2005 by the International Anesthesia Research Society
0003-2999/05

ANESTH ANALG
2005;101:1536 41

groups with two concentrations of NaHCO3 added to

the same dose of lidocaine) in patients scheduled for
thyroidectomy, ventilated without N2O, to evaluate
ETT-induced emergence phenomena.

Methods
A 2% L-HCl solution was used (Xylocaine 2%, AstraZeneca, Rueil Malmaison, France). After injection of
2 mL of 2% L-HCl (40 mg) into the ETT cuff (Sheridan,
Hudson Respiratory Care, Temecula, CA; polyvinyl
chloride (PVC) cuff), a supplementary volume of 3 mL
was added at 2 concentrations: 8.4%, or 1.4% of
NaHCO3 (B. Braun Medical S.A., Boulogne, France).
Release of lidocaine from ETT cuffs was performed
using a Distek dissolution test system model 5100A
(Distek INC, North Brunswick, NJ). It consisted of 4
independent cylindrical flasks with spherical bottoms
each containing 900-mL release medium (i.e., simulated intestinal fluid: monobasic potassium phosphate
6.8 g/L) consisting of pH 7.4 phosphate buffer thermostated at 37C and a rotating paddle apparatus
operating at 100 rpm. To check if variable concentrations of NaHCO3 in the cuff could modify L-HCl
release, 4 sets of ETT were immersed in cylindrical
flasks, and one of the following 2 solutions was placed
inside the cuff (2 cuffs for each concentration). The
L-HCl concentration was measured continuously at
205 nm every 15 min during a 24-h period using an
Uvikon spectrophotometer model 922 (Kontron Instruments, St. Quentin en Yvelines, France). Each ETT
was tested only once. Intracuff pressure before or after
immersion was not recorded. pH determination of
different solutions was performed with the same dose
of L-HCl (2 mL of 2%) and various volumes of
NaHCO3 (2 to 6 mL of 8.4%, or 1.4% of NaHCO3).
Institutional Ethics Committee approval and written informed patient consent were obtained. Adult
patients scheduled for total thyroidectomy surgery
(ASA physical status III) were consecutively enrolled. Patients were excluded from the study if they
had an anticipated difficult tracheal intubation, had
risk factors for postoperative aspiration of gastric contents, or had respiratory disease or recent respiratory
tract infection. Patients were randomized into one of
three groups: ETT cuff was filled with air (group air)
or with alkalinized L-HCl using 8.4% (group large
dose) or 1.4% (group small dose) of NaHCO3. The ETT
was lubricated with sterile water.
Oral alprazolam (0.5 mg) was administered 23 h
before the surgery. The anesthetic care team performed the standard anesthesia. After establishing IV
access and routine monitors, propofol 2.5 mg/kg,
sufentanil 0.35 g kg1 h1, and atracurium
0.6 mg/kg were used for anesthesia. Tracheal intubation was performed using tracheal tube (Murphy,

ESTEBE ET AL.
ALKALINIZATION OF INTRACUFF LIDOCAINE

1537

Allegiance, Malaysia, low volume, high pressure; PVC

cuff) 6.57.0 mm inner diameter for women and 7.0
7.5 mm inner diameter for men) by the anesthesiology
team. Lubrication of the ETT was performed with
sterile water. ETT cuffs (low volume and high pressure) were inflated according to the randomized protocol by the experimenter. ETT cuffs were inflated at
the minimal occlusive volume (i.e., no leakage was
detected under controlled ventilation). In the control
group, the cuff was initially slowly inflated with air.
For alkalinized-L-HCl-filled cuff groups, 2 mL of
L-HCl 2% (Xylocaine, AstraZeneca, Paris, France)
was initially injected into a cuff, and then a supplementary volume of 8.4% or 1.4% NaHCO3 was added.
Cuff pressure was recorded with initial pressure less
than 30 cm H2O (Mallinckrodt, Seelscherf 1, Germany). If an air leak was recorded during the surgery
1 mL of NaHCO3 (8.4% or 1.4%) was added for liquid
groups or 1 mL of air was added in the control group.
The anesthesiology team, unaware of the experimental
protocol, delivered anesthesia. Ventilation was controlled to maintain an end-tidal CO2 of 4.55.5 kPa,
and no gastric tube was inserted. Patients were in the
supine position. Any member of the surgical or anesthesia team applied topical anesthesia or vasoconstrictor to the larynx when indicated. Maintenance anesthesia included: air/O2 (50%/50%), sevoflurane (1 to
1.2%), and sufentanil 0.35 g kg1 h1 until surgical
closure and dressing (time T0). Patients were then
administered 100% oxygen and placed in the recovery
room.
When all of the tracheal extubation criteria were met
(return of neuromuscular function confirmed using
train-of-four peripheral nerve stimulation, ability to
follow verbal commands, regular spontaneous ventilation), tracheal extubation was performed just after
suctioning at the discretion of the physician in charge
of the patient. Time of spontaneous ventilation time
(time between emergence of spontaneous breathing
and extubation) was recorded. The gas and liquid
volumes withdrawn from ETT cuffs at extubation
were recorded. Cough and restlessness were checked
before extubation, excluding cough on suctioning and
extubation times. A blinded nurse evaluated the sore
throat in the recovery room with a visual analog scale
(VAS, 0 10 cm) after extubation (at 30 min and 1, 2, 3,
6 and 24 h). Other complaints of throat discomfort,
such as hoarseness, bucking, dysphonia, and dysphagia, were systematically evaluated as present or absent. Impairment of swallowing reflex was evaluated
by oral intake of a glass of water 15 min after extubation. Hemodynamic variables and postoperative nausea and vomiting were also recorded for 2 to 4 h as
routine control in the recovery room.
Sample-size calculation was based on our previous
studies (10 12); our primary efficacy variable was the
incidence of sore throat as measured by VAS. We

1538

ESTEBE ET AL.
ALKALINIZATION OF INTRACUFF LIDOCAINE

ANESTH ANALG
2005;101:1536 41

Figure 1. Percentage of lidocaine release in vitro as a function of time, from

the cuff filled with 40 mg of lidocaine
hydrochloride (2 mL 2%) and additional 3 mL of sodium bicarbonate at
8.4% and 1.4% concentrations (n 2
for each concentration).

postulated that if the alkalinized L-HCl had only a

volume effect avoiding the overinflation phenomena,
no effect could be recorded if ventilation was controlled without N2O (null hypothesis) and the results
would be equivalent to the air group data. We estimated that using alkalinized L-HCl would decrease
the rate of sore throat by 25%30% as evaluated by
VAS compared with the air group (alternative hypothesis) (11,12). Based on these estimates, we calculated a
sample size that would permit a type I error of 5%
with a type II error of 5% and power of 95%.
Enrollment of 20 patients in each group was required.
Patients were withdrawn from the study when the
trachea was not intubated on the first attempt. Results
are presented as mean sd. Data were analyzed using
the analysis of variance followed by unpaired Students t-test with Bonferroni correction for parametric
data. Kruskal-Wallis and Mann-Whitney U-tests were
used for nonparametric data. Statistical significance
was defined as P 0.05. Patient randomization was
performed using a computerized list and the same
investigator performed the filling protocol of ETT but
was excluded from all other periods (i.e., anesthesia,
intubation, and extubation time).

Results
Concerning the effect of various concentrations of
NaHCO3, there was a slight tendency for a slower
release when the smaller concentration (1.4%) of
NaHCO3 was used compared with the larger concentration NaHCO3 (8.4%) (Fig. 1). At 3 h, 15% of L-HCl
was released throughout the ETT cuff when 1.4% of
NaHCO3 was added, versus 25% when 8.4% of
NaHCO3 was used. At 6 h 36% of L-HCl was released
throughout the ETT cuff when 1.4% of NaHCO3 was

added versus 45% when 8.4% of NaHCO3 was used.

In vitro pH determination is summarized in Table 1.
Sixty patients participated in the study and none was
excluded (i.e., all patients were tracheally intubated at
first attempt). There was no statistically significant difference among groups regarding surgery and anesthesia
characteristics (Table 2). There were no problems with
endotracheal intubation or cuff inflation. During controlled ventilation, no air leak was recorded. Compared
with the control group there was a slight but not significant difference in the initial volume of the solution
injected into the cuff (3.9 0.6 mL, 4.3 0.5 mL, 4.1
0.4 mL for groups air, large dose, and small dose, respectively; P 0.08). There was no significant difference
among groups in ETT cuff pressure recording. The volume withdrawn at extubation time was not significantly
different among groups (3.4 0.8 mL, 3.5 0.5 mL, and
3.4 0.5 mL for groups air, large dose, and small dose,
respectively).
Compared with group air, group large dose and
group small dose had significant reductions in sore
throat (mean end-point for efficacy) during the 24-h
postoperative period (P 0.0001)(Fig. 2). The difference was not significant between the two alkalinized
L-HCl groups. This increase in ETT tolerance was
confirmed by the analysis of secondary and safety
end-points (Table 3). There was a significant prolongation of spontaneous ventilation time and the time to
tracheal extubation. The good tolerance of the ETT
was associated with less cough and restlessness before
suctioning and extubation. No difference was recorded in cough reflex at extubation time. Neither
laryngospasm, nor depression of the swallowing reflex, was recorded. Based on postoperative nausea and
vomiting and hoarseness the control group (air-filled
cuff) displayed less tolerance (Table 3). There was no

ANESTH ANALG
2005;101:1536 41

ESTEBE ET AL.
ALKALINIZATION OF INTRACUFF LIDOCAINE

1539

Table 1. In Vitro pH Evaluation of Solutions with 2 mL of 2% Lidocaine and Various Concentrations and Various
Volumes of Sodium Bicarbonate
1.4% NaHCO3

pH

8.4% NaHCO3

2 mL

3 mL

4 mL

6 mL

2 mL

3 mL

4 mL

6 mL

7.44

7.47

7.59

7.63

7.95

7.99

8.07

8.09

Table 2. Demographic Patient Data for the Air Group, Lidocaine Alkalinized Groups with 8.4% or 1.4% of Sodium
Bicarbonate

Group size

Air
(n 20)

8.4% NaHCO3 2% lidocaine

2 mL
(n 20)

1.4% NaHCO3 2% lidocaine

2 mL
(n 20)

Age (yr)
Weight (kg)
Height (cm)
Female/male
Smoking (%)
Surgery (min)
Sufentanil (g/kg/h)

47 13
62 7
165 6
17/3
40
58 13
0.34 0.03

48 15
64 8
166 9
15/5
35
56 8
0.36 0.02

48 12
65 10
167 9
15/5
35
57 6
0.35 0.04

Figure 2. Visual analog scale (VAS, 0 100 mm) scores of sore throat
during the first postoperative days for air group (group air; black
symbol), lidocaine alkalinized groups with 8.4% (group large dose;
gray symbol) or 1.4% (group small dose: open symbol) of sodium
bicarbonate (n 20 for each group). *P 0.0001 group air versus
liquid groups (large or small dose).

difference between liquid groups. There was a trend

of reduced hypertension and tachycardia in the control group and the alkalinized L-HCl groups but it was
not significant (Table 3).

Discussion
This is the first study in which the ETT cuff filled with
alkalinized L-HCl was evaluated in anesthetized patients with controlled ventilation without N2O. Our
results showed a significant improvement of the ETTinduced emergence phenomena from general anesthesia when alkalinized L-HCl was used instead of air to

fill the ETT cuff. VAS scores for sore throat were
similar before surgery (Fig. 2). Although thyroid surgery was responsible for pain in the cervical area,
patients clearly reported a decrease of VAS scores for
sore throat in the two alkalinized L-HCl groups.
The incidence of coughing and sore throat on emergence from general anesthesia in the presence of ETT
has been estimated to range from 38% to 96% (8,15). In
our control group, coughing was reported in 70% of
patients and sore throat was evaluated at 30 15 mm
using the VAS. These results were in agreement with
previous studies (8,10 12,15). Our data confirmed the
lack of increased cuff pressure and cuff volume after
air inflation without N2O (4,16). It has been reported
that the overinflation occurring during general anesthesia was attributable to an increase in temperature
and, most importantly, because of more rapid NO2
diffusion into the cuff than out from the cuff (1,13,16).
This overinflation of the ETT cuff has been associated
with damage to the pharyngeal mucosa and recurrent
laryngeal nerve palsy (17). The lack of hyperpressure
is probably one advantage of liquid filling of ETT cuffs
(18,19). However, despite the absence of overinflation
in our control group, filling the cuff with alkalinized
L-HCl allowed a significant improvement of ETT cuff
tolerance. The effect on thyroid surgical pain could not
be excluded. However, surgical pain (i.e., pressure
threshold) was not specifically evaluated.
It has been reported that L-HCl injected alone had a
slow diffusion rate across the ETT cuff (1% of release
during the 6-hour period) (11). For a clinical effect,
large doses of L-HCl (200 to 500 mg) were believed to
be required (59). In addition to the potential adverse
effect of these large doses in case of rupture, there was
no real advantage compared to saline (4). The use of

1540

ESTEBE ET AL.
ALKALINIZATION OF INTRACUFF LIDOCAINE

ANESTH ANALG
2005;101:1536 41

Table 3. Secondary End-points of Endotracheal Tube-induced Emergence Phenomena: Time of Spontaneous Ventilation
Before Extubation (T0 Isoflurane and Propofol Were Stopped), Cough Effort and Restlessness Before Extubation,
PONV, Dysphonia, and Hoarseness After Extubation. Hemodynamic Data 5 Minutes After Extubation: Systolic and
Diastolic Arterial Blood Pressure and Cardiac Frequency for Air Group, Lidocaine Alkalinized Groups with 8.4% or 1.4%
of Sodium Bicarbonate

Spontaneous ventilation time (min)

Cough effort (%)
Restlessness (%)
PONV (%)
Dysphonia (%)
Hoarseness (%)
AP systolic (mmHg)
AP diastolic (mmHg)
Cardiac frequency (beats/min)

Air
(n 20)

8.4% NaHCO3 2% lidocaine

2 mL
(n 20)

1.4% NaHCO3 2% lidocaine

2 mL
(n 20)

4.5 2
70
30
85
5
75
128 39
80 10
94 18

12 3
5
0
30
5
40
119 15
74 11
89 20

11 4*
5
5
40
5
20
123 19
73 11
85 17

Values are mean sd or %.

PONV postoperative nausea and vomiting; AP arterial blood pressure.
* P 0.0001.

alkalinized local anesthetics into the ETT cuff offers

the advantages of minimal stress response to smooth
tracheal extubation and cough-free emergence. We
previously reported that alkalinization of L-HCl allowed the diffusion of 65% of the neutral base form of
L-HCl through the hydrophobic structure of the PVC
cuff within a 6-hour period and showed that the use of
a small dose (40 mg) of alkalinized L-HCl markedly
improved ETT tolerance during the first postoperative
day (11,12). It appears that only the hydrophobic neutral form of L-HCl was able to diffuse across a membrane, while for charged alkalinized L-HCl only a
permeation phenomenon occurred. Following the
Henderson-Hasselbach equation (i.e., the ratio between ionized and nonionized species being a function of both the pK of the substance and the pH of the
dissolving medium) the addition of NaHCO3 to alkalinized L-HCl alkalinizes the L-HCl solution. This provides the corresponding hydrophobic base and allows
the diffusion of this uncharged form through the hydrophobic PVC wall of the cuff more readily than the
alkalinized L-HCl and allows for the best release profile observed with the lidocaine base (10). In line with
this concept of alkalinization, we have previously reported that the amount of L-HCl diffusing across the
ETT cuff in the presence of NaHCO3 was proportional
to the dose of L-HCl applied (20 40 mg) (10). Our in
vitro and in vivo studies showed no cuff rupture or
obstruction (10 12).
No significant difference was reported between the
use of 8.4% and 1.4% NaHCO3 to alkalinized L-HCl
for VAS scores for sore throat (mean total clinical
volume 4 0.5 mL; 2 mL 2% of lidocaine and 2
0.5 mL of NaHCO3). Those clinical data were confirmed by our in vitro results showing a similar release
profile of L-HCl from the cuff. There was no significant difference in the release of L-HCl with a different

brand of ETT high-pressure-low-volume PVC cuff

(unpublished data). The slight difference of L-HCl
release observed as a function of NaHCO3 concentration in our in vitro results seems to have no clinical
effect. This slight delay of release with the small concentration of NaHCO3 could be useful for long duration surgery. Based on pH in vitro evaluation, we
report an increase of pH with the NaHCO3 concentration (i.e., at the same volume) and an increase of pH
with NaHCO3 volume (Table 1) without beneficial
clinical effect. We have previously reported, in a pharmacokinetic study, that diffusion with 8.4% NaHCO3
gives a very small maximal plasma concentration of
L-HCl (Cmax 0.08 g/mL) (10 12), which is smaller
than when L-HCl was used topically (0.431.5 g/mL)
(20) or IV (23 g/mL) (21,22). We have also shown, in
vitro, that variation in volumes of 8.4% of NaHCO3 (1
to 7 mL) injected into the cuff had no effect on the
diffusion of 40 mg L-HCl (11); in this study we have
shown that the concentration of NaHCO3 had no clinical effect on the L-HCl diffusion.
If alkalinization of L-HCl improved cuff tolerance
during the evaluation period, the local anesthetic effect did not depress the swallowing reflex, indicating
palsy of the vocal cords, consistent with our earlier
report (11,12). The decrease in cough before tracheal
extubation should not be attributed to a depression of
cough reflex during the suction-extubation act. It
probably results from an increase in ETT tolerance
attributable to a local effect rather than a systemic
effect because it was reported at intubation time with
a high level of plasma L-HCl after IV administration
(2 mg/kg given lidocaine levels 3 g/mL) (22). The
efficacy of IV or topical administration of L-HCl appears to be short-acting. To reduce sore throat at extubation time requires local application of L-HCl (with

ANESTH ANALG
2005;101:1536 41

a specific device) (20), IV administration before extubation (15), or large doses of local application of L-HCl
at intubation time for short-duration surgery
(90 min) (23). Prolongation of the time of spontaneous ventilation must be seen as an improvement of
ETT tolerance rather than as an adverse effect. Because
of the experimental anesthetic protocol, similar for
each group (i.e., maintenance of anesthesia until dressing), a prolongation of the time of spontaneous ventilation was observed. However, differences in recovery
room stay were not observed overall. In our clinical
practice, the increase of ETT tolerance allows for earlier reduction of anesthesia and spontaneous ventilation at the end of surgery. Quiet tracheal extubation
without sore throat was easily obtained in the recovery room and allowed a decrease in adverse effects, as
previously observed (i.e., postoperative nausea and
vomiting) (10 12).
As in our previous studies with N2O (10 12), this
study performed in a clinical setting without N2O for
controlled ventilation confirmed that alkalinized
L-HCl (i.e., base L-HCl) injected into the cuff, instead
of air, was clinically effective and safer in reducing
postoperative sore throat. Using a solution close to the
physiological pH and a small dose of L-HCl (40 mg)
reduces the risks of local anesthetic vascular absorption and mucosal irritation in case of ETT rupture,
although ETT rupture has never been reported. Conversely, some cases of cuff rupture have been reported
when L-HCl was used as lubricant or for local anesthesia (24). Hence, the current findings support the
use of a 1.4% NaHCO3 concentration to refill the cuff
of the ETT.
We conclude that there is a decrease in sore throat
during the postoperative period when the cuff is inflated with a small dose of alkalinized L-HCl (i.e.,
small dose of L-HCl:40 mg and small dose of
NaHCO3:1.4%) rather than with air when ventilation
is controlled without N2O. This technique is also applicable for the indirect effects of tracheal extubation,
including restlessness, hoarseness, and dysphonia.
Such a drug delivery system should be considered in
clinical practice to improve a patients tolerance of
anesthesia (with and without N2O) and intensive care
and, most importantly, in the case of cardiovascular
disease, intracranial or intraocular hyperpressure, or
hyperreactive pulmonary disease.

References
1. Tu HN, Saidi N, Lieutaud T, et al. Nitrous oxide increases
endotracheal cuff pressure and the incidence of tracheal lesions
in anesthetized patients. Anesth Analg 1999;89:18790.
2. Karasawa F, Ohshima T, Takamatsu I, et al. The effect on
intracuff pressure of various nitrous oxide concentrations used
for inflating an endotracheal tube cuff. Anesth Analg 2000;91:
708 13.

ESTEBE ET AL.
ALKALINIZATION OF INTRACUFF LIDOCAINE

1541

3. Karasawa F, Matsuoka N, Kodama M, et al. Repeated deflation

of gas-barrier cuff to stabilize cuff pressure during nitrous oxide
anesthesia. Anesth Analg 2002;95:243 8.
4. Bennett MH, Isert PR, Cumming RG. Postoperative sore throat
and hoarseness following tracheal intubation using air or saline
to inflate the cuff: a randomized controlled trial. Anaesth Intensive Care 2000;28:408 13.
5. Sconso JM, Moscicki JC, Difazio CA. In vitro diffusion of lidocaine across endotracheal tube cuffs. Reg Anesth 1990;15:37 40.
6. Navarro RM, Baughman VL. Lidocaine in the endotracheal tube
cuff reduces postoperative sore throat. J Clin Anesth 1997;9:
394 7.
7. Hirota W, Kobayashi W, Igarashi K, et al. Lidocaine added to a
tracheostomy tube cuff reduces tube discomfort. Can J Anaesth
2000;47:412 4.
8. Fagan C, Frizelle HP, Laffey J, et al. The effects of intracuff
lidocaine on endotracheal-tube-induced emergence phenomena
after general anesthesia. Anesth Analg 2000;91:2015.
9. Altintas F, Bozkurt P, Kaya G, et al. Lidocaine 10% in the
endotracheal tube cuff: blood concentrations, haemodynamic
and clinical effects. Eur J Anaesthesiol 2000;17:436 42.
10. Dollo G, Estebe JP, Le Corre P, et al. Endotracheal tube cuffs
filled with lidocaine as a drug delivery system: in vitro and in
vivo investigations. Eur J Pharm Sci 2001;13:319 23.
11. Estebe JP, Dollo G, Le Corre P, et al. Improvement of effect of
intracuff lidocaine on endotracheal-tube-induced emergence
phenomena by alkalinization. Anesth Analg 2002;94:22730.
12. Estebe JP, Delahaye S, Le Corre P, et al. Alkalinization of intracuff lidocaine and use of gel lubrication protect against tracheal
tube- induced emergence phenomena. Br J Anaesth 2004;92:
361 6.
13. Soltani HA, Aghadavoudi O. The effect of different lidocaine
application methods on postoperative cough and sore throat.
J Clin Anesth 2002;14:15 8.
14. Coe AJ, Jain RK, Estebe JP. Alkalinization of intra-cuff lidocaine
and use of gel lubrication protect against tube-induced emergence phenomena. Br J Anaesth 2004;93:477 8.
15. Gonzales RM, Bjerke RJ, Drobycki T, et al. Prevention of endotracheal tube-induced coughing during emergence from general
anesthesia. Anesth Analg 1994;79:7925.
16. Dullenkopf A, Gerber AC, Wiess M. Nitrous oxide diffusion
into tracheal tube cuffs: comparison of five different tracheal
tube cuffs. Acta Anaesthesiol Scand 2004;48:1180 4.
17. Lev R, Rosen P. Prophylactic lidocaine use preintubation: a
review. J Emerg Med 1994;12:499 506.
18. Combes X, Schauvliege F, Peyrouset O, et al. Intracuff pressure
and tracheal morbidity: influence of filling with saline during
nitrous oxide anesthesia. Anesthesiology 2001;95:1120 4.
19. Ahmad NL, Norsidah AM. Change in endotracheal tube cuff
pressure during nitrous oxide anaesthesia: a comparison between air and distilled water cuff inflation. Anaesth Intensive
Care 2001;29:510 4.
20. Diachun CD, Tunink BP, Brock-Utrie JG. Suppression of cough
during emergence from general anesthesia: laryngotracheal lidocaine through a modified endotracheal tube. J Clin Anesth
2001;13:44757.
21. Nishino T, Hiraga K, Sugimori K. Effects of IV lidocaine on
airway reflexes elicited by irritation of the tracheal mucosa in
humans anesthetized with enflurane. Br J Anaesth 1999;64:
6827.
22. Yukioka H, Yoshimoto N, Nishimura K, et al., Intravenous
lidocaine as a suppressant of coughing during tracheal intubation. Anesth Analg 1985;64:1189 92.
23. Minogue SC, Ralph J, Lampa MJ. Laryngotracheal topicalization
with lidocaine before intubation decreases the incidence of
coughing on emergence from general anesthesia. Anesth Analg
2004;99:12537.
24. Walmsley AJ, Burville LM, Davis TP. Cuff failure in polyvinyl
chloride tracheal tubes sprayed with lignocaine. Anaesthesia
1988;43:399 401.

The Predictive Value of the Height Ratio and Thyromental

Distance: Four Predictive Tests for Difficult Laryngoscopy
Banjong Krobbuaban, MD, Siriwan Diregpoke,
and Malin Tanomsat, BN

BN,

Sujarit Kumkeaw,

BN,

Department of Anesthesiology, Chaiyaphum Hospital, Thailand

Preoperative evaluation of anatomical landmarks and

clinical factors help identify potentially difficult laryngoscopies; however, predictive reliability is unclear.
Because the ratio of height to thyromental distance
(RHTMD) has a demonstrably better predictive value
than the thyromental distance (TMD), we evaluated the
predictive value and odds ratios of RHTMD versus
mouth opening, TMD, neck movement, and oropharyngeal view (modified Mallampati). We collected data
on 550 consecutive patients scheduled for electivesurgery general anesthesia requiring endotracheal intubation and then assessed all five factors before surgery. An experienced anesthesiologist, not apprised of
the recorded preoperative airway assessment, performed the laryngoscopy and grading (as per Cormack

ailure in managing the airway is the most significant cause of morbidity and mortality in anesthetized patients (1). Difficult laryngoscopy (defined by poor glottic visualization) is synonymous
with difficult intubation during surgery in most patients (2). Difficult intubation is reported in 1.5%13%
of patients (312). Preoperative evaluation is important for the risk of difficult airway management, but
which anatomical landmarks and clinical factors are
the best predictors is debated (3,5,1316).
Several studies describe prediction schemes using a
single risk factor or a multifactorial index (3,10,12,17).
One test for difficult laryngoscopy is the thyromental
distance (TMD), which varies with patient size (7).
However, several studies question whether the TMD
is either sensitive or specific enough to be used as the
only predictor of difficult laryngoscopy (7,11,15,19).
Although Schmitt et al. (18) found that the ratio of
height to TMD [RHTMD Height (cm)/TMD (cm)]
Accepted for publication May 23, 2005.
Address correspondence and reprint requests to Banjong Krobbuaban, MD, Department of Anesthesiology, Chaiyaphum Hospital, Bannakan Road, Muang Chaiyaphum, Thailand. Address e-mail
to Albkb@diamond.mahidol.ac.th.
DOI: 10.1213/01.ANE.0000181000.43971.1E

1542

Anesth Analg 2005;101:15425

and Lehanes classification). Difficult laryngoscopy

(Grade 3 or 4) occurred in 69 patients (12.5%). RHTMD
had a higher sensitivity, positive predictive value, and
fewer false negatives than the other variables tested. In
the multivariate analysis, three criteria were found independent for difficult laryngoscopy (neck movement
80 degrees; Mallampati Class 3 or 4, and RHTMD
23.5). The odds ratio (95% confidence interval) of the
RHTMD, Mallampati class, and neck movement were
6.72 (3.29 13.72), 2.96 (1.635.35), and 2.73 (1.14 6.51),
respectively. The odds ratio for RHTMD was the largest
and thus may prove a useful screening test for difficult
laryngoscopy.
(Anesth Analg 2005;101:15425)

had a better predictive value than the TMD, no published study has quantified its sensitivity, specificity,
and positive predictive value (PPV) versus other bedside tests for assessing a patients airway for difficult
laryngoscopy. We, therefore, conducted a prospective,
blind study of the predictive value of the RHTMD
versus four other methods of airway assessment for
difficult laryngoscopy.

Methods
The protocol was approved by the Ethics Committee
of our hospital before commencing this study, and all
patients gave written, informed consent. We then
studied 550 consecutive ASA physical status III adult
patients scheduled to receive general anesthesia requiring endotracheal intubation for elective orthopedic, urologic, abdominal, and gynecologic surgery.
Patients younger than 18 yr of age, with obvious
malformations of the airway, edentulous, or requiring
a rapid sequence induction or awake intubation were
excluded from the study to avoid the introduction of a
variable that might independently affect predictability
of difficult laryngoscopy. Preoperative airway assessment was performed for all patients by the same anesthesiologist to avoid interobserver variability. The
2005 by the International Anesthesia Research Society
0003-2999/05

ANESTH ANALG
2005;101:15425

KROBBUABAN ET AL.
PREDICTING DIFFICULT LARYNGOSCOPY

Table 1. Demographic Data of all Patients and

Distribution of Laryngoscopic View (mean sd)
Category

Value

Men n (%)
Women n (%)
Age (yr)
Weight (kg)
Height (cm)
ASA class n (%)
I
II
TMD (cm)
RHTMD
Interincisor gap (cm)
Mallampati class
n (%)
1
2
3
4
Laryngoscopic view
n (%)
1
2
3
4

261 (48)
289 (52)
45 15
58 10
160 8
398 (72)
152 (28)
7.1 0.8
22.8 2.8
4.0 0.6
126 (23)
183 (33)
168 (31)
73 (13)
362 (66)
119 (22)
62 (11)
7 (1)

TMD thyromental distance; RHTMD ratio of height to TMD.

tests used to predict difficult laryngoscopy were measurement of mouth opening, TMD, maximum range of
head and neck movement, and assessment of the oropharyngeal view. The standard examination method
was used for each test.
Mouth opening was assessed by measuring the interincisor gap. Each patient was asked to open his or
her mouth as wide as possible, and the distance between the upper and lower incisors at the midline was
measured (20). TMD was measured from the bony
point of the mentum while the head was fully extended and the mouth closed (9).
The maximum range of head and neck movement
was assessed using the method described by Wilson et
al. (12). The patient was asked to extend his or her
head and neck fully while a pencil was placed vertically on the forehead. Then, while the pencil was held
firmly in position, the head and neck were flexed. The
range of head and neck movement was classified as 1
80 degrees or 2 80 degrees.
The oropharyngeal view was assessed using a modified Mallampati classification (21). While seated, each
patient was asked to open his or her mouth maximally
and to protrude the tongue without phonation (22).
The view was classed as (a) good visualization of the
soft palate, fauces, uvula, and tonsillar pillars; (b) pillars obscured by the base of the tongue but the soft
palate, fauces, and uvula visible; (c) soft palate and
base of the uvula visible; and (d) soft palate not visible
(21).

1543

We also assessed height, RHTMD, body weight, and

body mass index. Each patient was routinely monitored during the entire procedure by electrocardiography, pulse oximetry (for measurement of oxygen
saturation), and a noninvasive arterial blood pressure
monitor.
After the administration of oxygen, all patients were
anesthetized using standard drugs including midazolam 0.03 mg/kg, fentanyl 12 g/kg, and propofol 2.5
mg/kg and then paralyzed using neuromuscular
blocking drugs to facilitate orotracheal intubation. The
patients lungs were ventilated by mask with 100%
oxygen. Laryngoscopy was performed after the loss of
the fourth twitch in the train-of-four. The head of the
patient was placed in the sniffing position, and laryngoscopy was performed, with a Macintosh Number 3 laryngoscope blade, by a nurse anesthesiologist
with 7 yr experience.
Glottic visualization was assessed using a modified
Cormack and Lehane (23) classification without external laryngeal manipulation. This classification involved four grades of glottic visualization: Grade 1
complete of the vocal cords; Grade II the inferior
portion of the glottis; Grade III only the epiglottis;
and Grade IV a nonvisualized epiglottis. External
laryngeal pressure was permitted after evaluation of
the insertion of the endotracheal tube. Difficult laryngoscopy in this study was set at Cormack and Lehane
Grades 3 and 4. After evaluation, endotracheal intubation was performed after standard anesthetic
management.
The preoperative assessment data and the laryngoscope findings were used to evaluate the predictive
value of each test for difficult laryngoscopy. The sensitivity, specificity, PPV, and negative predictive value
(NPV) of each test were calculated. In addition, receiver operating characteristic curves were used to
identify the optimal predictive cutoff points for
RHTMD measurement, interincisor gap, and TMD.

Results
A total of 550 patients were included. Laryngoscopies
were possible for all of the patients. There were no
failed tracheal intubations. Demographic data and the
distribution of the Cormack and Lehane (23) grades
are presented with the mean for the interincisor gap,
TMD, and RHTMD (Table 1). The receiver operating
characteristic curves of the interincisor gap, TMD, and
RHTMD are presented in Figure 1.
The optimal cutoff point for the RHTMD, TMD, and
interincisor gap for predicting difficult laryngoscopy
was 23.5 (sensitivity, 77%; specificity, 66%), 6.5 cm
(sensitivity, 52%; specificity, 71%), and 3.5 cm (sensitivity, 39%; specificity, 69%), respectively. An interincisor gap 3.5 cm, a TMD 6.5 cm, neck movement

1544

KROBBUABAN ET AL.
PREDICTING DIFFICULT LARYNGOSCOPY

ANESTH ANALG
2005;101:15425

Figure 1. Receiver operating characteristic curves for the ratio of

height to thyromental distance (RHTMD), thyromental distance
(TMD), and interincisor gap (IIG) in the prediction of difficult
laryngoscopy.

80 degrees, and a Mallampati Class 3 or 4 were the

factors selected predictive of difficult laryngoscopy.
Difficult laryngoscopy (Cormack and Lehane Grade 3
or 4) was reported in 69 patients (12.5%; Table 2).
RHTMD had a higher sensitivity, PPV, and fewer
false negatives than the other factors. The multivariate
analysis odds ratios (95% confidence interval) of the
RHTMD, Mallampati class, and neck movement were
6.72 (3.29 13.72), 2.96 (1.635.35), and 2.73 (1.14 6.51),
respectively. The interincisor gap 3.5 cm and TMD
6.5 cm were not recognized as independent variables for difficult laryngoscopy.

Discussion
In our study, the incidence of difficult laryngoscopy
was frequent (12.5%) compared with other studies

(35). Our results might therefore favor a low NPV

and a high PPV. RHTMD allows a higher PPV (24%)
than previously reported (9%12%) (35) but, nonetheless, implies that predicted difficult laryngoscopy
is not observed in between 4 and 5 patients.
Ideally, any preoperative assessment scheme for
difficult laryngoscopy should have a high sensitivity
and specificity and produce few false positives and
negatives. The advantage of RHTMD is its higher
sensitivity than other tests, thus false-negative predictions are minimized. The consequence of a falsenegative result may be deleterious and even lifethreatening; therefore, decreasing false-negative
prediction is far more important than falsely predicting difficult laryngoscopy in unaffected patients. Because difficult laryngoscopy is infrequent, the incidence of false negatives is small. Nevertheless, a test
should be sufficiently sensitive to detect possible difficulties with laryngoscopy. Although all the tests in
this series were not highly sensitive, RHTMD measurement resulted in the least amount of detection failure for difficult laryngoscopy of the other four tests.
This is our most important finding.
Using a multivariate analysis, we found that the
tests using neck movement 80 degrees, a Mallampati
Class 3 or 4, and RHTMD 23.5 were the major factors
for predicting difficult laryngoscopy. The RHTMD
had the highest odds ratio for prediction of a difficult
laryngoscopy. Interestingly, we did not find any association between difficult laryngoscopy and the interincisor gap 3.5 cm or TMD 6.5 cm. Our study
agrees with reports from Randell (19) and Chou and
Wu (24) who suggested that although TMD has been
investigated, it has proven of little value in predicting
difficult intubation. Furthermore, Savva (11) found no
correlation between the interincisor gap and the view
on laryngoscopy.
Schmitt et al. (18) found that RHTMD 25 can be
used to predict difficult laryngoscopies for white men
and women. They suggested that it might not apply to
other races. In our study, a RHTMD 23.5 was a
determining factor for predicting a poor laryngeal
view among Thai patients. This difference is small,

Table 2. Accuracy of Risk Factors in Predicting Difficult Laryngoscopy (n)

Risk factors

TP

TN

FP

FN

Sens
(%)

Spec
(%)

PPV
(%)

NPV
(%)

Odds
ratio

95% CI

P-value

IIG 3.5 cm
TMD 6.5 cm
NM 80
MPC 3 or 4
RHTMD 23.5

22
36
9
48
53

331
345
449
288
317

150
136
32
193
164

42
33
60
21
16

39
52
13
70
77

69
71
93
60
66

15
21
22
20
24

89
91
88
93
95

0.81
0.78
2.73
2.96
6.72

0.451.45
0.411.51
1.146.51
1.635.35
3.2913.72

NS
NS
0.0236
0.0003
0.0000

IIG interincisor; TMD thyromental distance; NM neck movement; MPC Mallampati class; RHTMD ratio of height to thyromental distance; TP
true positive; TN true negative; FP false positive; FN false; Sens sensitivity; Spec specificity; PPV positive predictive value; NPV negative
predictive value; CI confidence interval; NS not significant.

ANESTH ANALG
2005;101:15425

and further investigation is required to determine the

significance of ethnicity.
This was a carefully conducted randomized design
involving a single physicians assessment of the airway and a single nurse anesthesiologists assessment
of intubation difficulty. Whereas inter-observer variability was minimized, it leaves open the possibility of
bias based on how representative these two individuals are of the population of anesthesiologists.
In conclusion, several studies have evaluated different clinical risk factors, alone or in combination, for a
useful method to predict difficult intubation. Our results suggest that RHTMD may be a useful bedside
screening test for preoperative prediction of difficult
laryngoscopy.
This research was supported by Chaiyaphum Hospital under the
Ministry of Health, Thailand. The authors thank Dr. Yongyut
Gumpupong, Director of Chaiyaphum Hospital, for his support and
guidance and Mr. Bryan Roderick Hamman at the Faculty of Medicine, Khon Kaen University, Thailand, for assistance with the
English-language presentation.

References
1. Caplan RA, Posner KL, Ward RJ, Cheney FW. Adverse respiratory events in anesthesia: a closed analysis. Anesthesiology
1990;72:828 33.
2. Benumof JL. Difficult laryngoscopy: obtaining the best view.
Can J Anaesth 1994;41:3615.
3. Arne J, Descoins P, Fusciardi J, et al. Preoperative assessment for
difficult intubation in general and ENT surgery: predictive
value of a clinical multivariate risk index. Br J Anaesth 1998;80:
140 6.
4. Langeron O, Masso E, Huraux C, et al. Prediction of difficult
mask ventilation. Anesthesiology 2000;92:1229 36.
5. Rose DK, Cohen MM. The airway: problems and predictions in
18,500 patients. Can J Anaesth 1994;41:372 83.
6. Rocke DA, Murray WB, Rout CC, et al. Relative risk analysis of
factors associated with difficult intubation in obstetric anesthesia. Anesthesiology 1992;77:6773.

KROBBUABAN ET AL.
PREDICTING DIFFICULT LARYNGOSCOPY

1545

7. Butler PJ, Dhara SS. Prediction of difficult laryngoscopy: an

assessment of thyromental distance and Mallampati predictive
tests. Anaesth Intensive Care 1992;20:139 42.
8. Frerk CM. Predicting difficult intubation. Anaesthesia 1991;46:
1005 8.
9. Lewis M, Keramati S, Benumof JL, Berry CC. What is the best
way to determine oropharyngeal classification and mandibular
space length to predict difficult laryngoscopy? Anesthesiology
1994;81:69 75.
10. Oates JDL, Macleod AD, Oates PD, et al. Comparison of two
methods for predicting difficult intubation. Br J Anaesth 1991;
66:3059.
11. Savva D. Prediction of difficult tracheal intubation. Br J Anaesth
1994;73:149 53.
12. Wilson ME, Spiegelhalter D, Robertson JA, Lesser P. Predicting
difficult intubation. Br J Anaesth 1988;61:211 6.
13. Crosby ET, Cooper RM, Douglas MJ, et al. The unanticipated
difficult airway with recommendations for management. Can J
Anaesth 1998;45:75776.
14. Turkan S, Ates Y, Cuhruk H, et al. Should we reevaluate the
variables for predicting the difficult airway in anesthesiology?
Anesth Analg 2002;94:1340 4.
15. Tse JC, Rimm EB, Hussain A. Predicting difficult endotracheal
intubation in surgical patients scheduled for general
anaesthesia: a prospective blind study. Anesth Analg 1995;81:
254 8.
16. The ASA task force on guidelines for management of the difficult airway: practice guidelines for management of the difficult
airway. Anesthesiology 1993;78:597 602.
17. EI-Ganzouri AR, McCarthy RJ, Tuman KL, et al. Preoperative
airway assessment: predictive value of a multivariate risk index.
Anesth Analg 1997;84:419 21.
18. Schmitt HJ, Kirmse M, Radespiel-Troger M. Ratio of patients
height to thyromental distance improves prediction of difficult
laryngoscopy. Anaesth Intensive Care 2002;30:7635.
19. Randell T. Prediction of difficult intubation. Acta Anaesthesiol
Scand 1996;40:1016 23.
20. Aiello G, Metcalf I. Anaesthetic implication of temporomandibular joint disease. Can J Anaesth 1992;39:610 6.
21. Samsoon GLT, Young JRB. Difficult tracheal intubation: a prospective study. Anaesthesia 1987;42:48790.
22. Mallampati SR, Gatt SP, Gugino LD, et al. A clinical sign to
predict difficult tracheal intubation: a prospective study. Can
Anaesth Soc J 1985;32:429 34.
23. Cormack RS, Lehane J. Difficult tracheal intubation in obstetrics.
Anaesthesia 1984;39:110511.
24. Chou H, Wu T. Thyromental distance: shouldnt we redefine its
role in the prediction of difficult laryngoscopy? Acta Anaesthesiol Scand 1998;42:1.

Supplemental Intravenous Crystalloid Administration Does

Not Reduce the Risk of Surgical Wound Infection
Barbara Kabon, MD*, Ozan Akca, MD, Akiko Taguchi, MD*, Angelika Nagele,
Ratnaraj Jebadurai, MD*, Cem F. Arkilic, MD*, Neeru Sharma, MD*,
Arundhathi Ahluwalia, MD*, Susan Galandiuk, MD, James Fleshman, MD,
Daniel I. Sessler, MD, and Andrea Kurz, MD#

RN*,

Departments of *Anesthesiology and Surgery, Washington University, St. Louis, Missouri; Outcomes Research Institute
and Departments of Anesthesiology & Perioperative Medicine and Surgery, University of Louisville, Louisville,
Kentucky; Department of Anesthesiology and General Intensive Care, Vienna General Hospital, University of Vienna,
Vienna, Austria; and #Department of Anesthesiology, University of Bern, Bern, Switzerland

Wound perfusion and oxygenation are important determinants of the development of postoperative wound
infections. Supplemental fluid administration significantly increases tissue oxygenation in surrogate
wounds in the subcutaneous tissue of the upper arm in
perioperative surgical patients. We tested the hypothesis that supplemental fluid administration during and
after elective colon resections decreases the incidence of
postoperative wound infections. Patients undergoing
open colon resection were randomly assigned to smallvolume (n 124, 8 mL kg1 h1) or large-volume (n
129, 16 18 mL kg1 h1) fluid management. Our
major outcomes were two distinct criteria for diagnosis
of surgical wound infections: 1) purulent exudate combined with a culture positive for pathogenic bacteria,

ound infections are common and serious

complications of surgery. For example, there
is a 9%27% risk of wound infection in patients undergoing colon surgery (1). Surgical wound
infections prolong hospitalization by 520 days (2),
Supported by the Erwin-Schroedinger Foundation (Vienna, Austria), National Institutes of Health Grants GM 61655 and DE 14879
(Bethesda, MD), the Joseph Drown Foundation (Los Angeles, CA),
the Gheens Foundation (Louisville, KY), and the Commonwealth of
Kentucky Research Challenge Trust Fund (Louisville, KY). OA is
the recipient of a Research Training Grant from the Foundation for
Anesthesia Education and Research (Rochester, MN). TycoMallinckrodt donated the thermometers and thermocouples used in
this study.
None of the authors has any financial relationship with any
company related to this research.
Accepted for publication May 5, 2005.
Address correspondence and reprint requests to Dr. Andrea
Kurz, Department of Anesthesiology, Inselspital, University of
Bern, Bern, Switzerland. Address e-mail to andrea.kurz@insel.ch.
On the worldwide web: www.or.org.
DOI: 10.1213/01.ANE.0000180217.57952.FE

1546

Anesth Analg 2005;101:154653

and 2) Center for Disease Control criteria for diagnosis

of surgical wound infections. All wound infections diagnosed using either criterion by a blinded observer in
the 15 days after surgery were considered in the analysis. Wound healing was evaluated with the ASEPSIS
scoring system. Of the patients given small fluid administration, 14 had surgical wound infections; 11 given
large fluid therapy had infections, P 0.46. ASEPSIS
wound-healing scores were similar in both groups: 7
16 (small volume) versus 8 14 (large volume), P
0.70. Our results suggest that supplemental hydration
in the range tested does not impact wound infection
rate.
(Anesth Analg 2005;101:1546 53)

increase intensive care unit (ICU) admissions (3), and

substantially increase cost (4).
Wound repair and resistance to infection both depend in part on wound tissue oxygen tension and can
potentially be improved by increasing arterial oxygen
tension, even of fully saturated blood (3,5). Consistent
with this theory, Hopf et al. (6) showed that tissue
oxygenation better predicts the risk of surgical infection than does the Center for Disease Control (CDC)
index. Oxygen is also a critical substrate for tissue
repair and wound healing because prolyl and lysyl
hydroxylases, and, therefore, collagen synthesis are
Po2 dependent (7).
Various factors influence tissue oxygenation. For
example, mild hypothermia reduces tissue oxygenation (8) and triples the risk of infection (2); conversely, supplemental perioperative oxygen halves
the risk of infection by increasing tissue oxygenation
(3). However, even supplemental oxygen fails to improve oxygenation in hypoperfused tissues (9). It is
2005 by the International Anesthesia Research Society
0003-2999/05

ANESTH ANALG
2005;101:1546 53

thus unsurprising that experimental wound hypoperfusion aggravates infections and reduces scar formation in animals (10,11).
There is considerable evidence that colloid administration titrated to maximize cardiac output reduces
complications, including those related to wound healing (1214). However, goal-directed fluid administration is technically difficult, expensive, and somewhat
invasive. Adequate vascular volume is thus usually
defined by hemodynamic stability and good urinary
output. The difficulty with this approach is that hypovolemia reduces peripheral tissue perfusion, as evidenced by decreasing tissue oxygen tension, before
reducing arterial blood pressure, increasing heart rate,
or reducing urinary output (9).1
Guided only by hemodynamic responses, clinicians
may inadequately compensate for the enormous fluid
losses associated with major surgery. Consistent with
this theory, tissue oxygenation is poor in many surgical patients (9) but can be significantly improved by
supplemental fluid administration (9,15). We, therefore, tested the hypothesis that doubling the rate of
crystalloid fluid administration during and after elective colon resection decreases the incidence of postoperative wound infection.

Methods
The study was conducted with approval from the IRBs
at Washington University in St. Louis and the University of Louisville. Written, informed consent was obtained from 256 patients aged 18 80 yr undergoing
open elective colon resection with an anticipated duration of surgery 2 h. We excluded patients having a
recent history of fever or infection, susceptibility to
malignant hyperthermia, congestive heart failure, diuretic therapy, any sort of renal failure, or a history of
pulmonary edema.
Patients fasted for at least 8 h before surgery. All
patients received standard mechanical bowel preparation the night before surgery using standard Fleets
phospha soda oral electrolyte solution. In-patients
were given IV fluid at a rate of 2 mL kg1 h1
during bowel preparation and throughout the preoperative evening. A third of the patients were admitted
to the hospital on the day of surgery; they performed
their bowel preparation at home.
Intraluminal antibiotics were not used. Cefotetan (2
g) and metronidazole (1500 mg) were given IV during
induction of anesthesia. Surgeons were encouraged to
restrict prophylactic antibiotics to 48 h. Additional
antibiotics (e.g., to treat clinically suspected infections)
were administered per judgment of the attending
surgeon.
1
Hopf HW. Subcutaneous tissue oxygen tension in wellresuscitated trauma patients. Crit Care Med, 1994;22:A60.

KABON ET AL.
SUPPLEMENTAL FLUID AND WOUND INFECTION

1547

Anesthesia was induced with IV sodium thiopental

(35 mg/kg) or propofol (2 4 mg/kg), vecuronium
(0.1 mg/kg), and fentanyl (13 g/kg). Anesthesia
was maintained with isoflurane in 40% oxygen and
60% nitrous oxide supplemented by fentanyl (12
g kg1 h1) and vecuronium. Inspiratory oxygen
concentrations were adjusted to maintain oxygen saturation 95%. The lungs were mechanically ventilated at a tidal volume of 10 mL/kg at a rate sufficient
to maintain end-tidal Pco2 near 35 mm Hg. A bladder
catheter was inserted in each patient. Intraoperative
distal esophageal temperature was maintained at 36C
using forced-air and IV fluid warming.
After induction of anesthesia and endotracheal intubation, patients were assigned to small or large perioperative hydration using computer-generated randomized codes that were kept in opaque, sealed,
sequentially numbered envelopes. Both fluid management schemes were within the range of normal practice and were complemented by replacement of blood
loss with crystalloid at a 3:1 ratio. Patients assigned to
small fluid management were given maintenance of lactated Ringers solution of 8 10 mL kg1 h1 intraoperatively and for the first postoperative hour. Those
assigned to large fluid management were given a fluid
bolus of 10 mL/kg before induction of anesthesia and
subsequently maintained with lactated Ringers solution at a rate of 16 18 mL kg1 h1 throughout surgery and the first postoperative hour. Additional fluid
was given to patients as necessary to maintain urinary
output 1 mL kg1 h1. Similarly, additional fluid
was administered when mean arterial blood pressure
decreased to 70% of preinduction value and was
unresponsive to minor adjustments in the isoflurane
concentration.
Postoperative fluid was given to both groups at a rate
of 2 mL kg1 h1 until the first postoperative morning.
Additional fluid was given to patients as necessary to
maintain urinary output 1 mL kg1 h1. Subsequently, fluid was given at the discretion of the attending surgeon. Plasma expanders (i.e., hetastarch or albumin) were not given.
Leukocyte filtered allogeneic blood was administered only as necessary to maintain the prospectively
determined target hematocrit that was based on the
patients age and cardiovascular status as previously
described (2,3).
Fascial closures were accomplished with running
double-stranded monofilament absorbable suture.
The skin was closed tightly with staples or absorbable
sutures. Throughout surgery, wound protectors were
used. Wound dressing was standardized; pressure
dressings were not used.
Postoperative pain relief was maintained with
patient-controlled morphine analgesia (2-mg bolus,
6-min lockout). Prophylactic antiemetics were not
given, as per standard of care. Postoperative nausea

1548

KABON ET AL.
SUPPLEMENTAL FLUID AND WOUND INFECTION

and vomiting were treated with ondansetron (4 mg

IV). Additional doses were given as necessary by clinicians blinded to group assignment. During recovery, patients were given supplemental oxygen via nasal prongs at a rate of 2 L/min; additional oxygen was
given as necessary to maintain oxygen saturation
95%.
After 1 h of recovery, the anesthesia record and
perioperative fluid administration records were sealed
in an envelope marked: Anesthesia Record. Do not
open unless necessary for clinical care until date
(16 days after surgery). Thus, the surgeons and investigators evaluating wound infections and healing
were unable to determine group assignment or perioperative fluid management from patient records.
Appropriate clinical characteristics of each treatment group were tabulated. Preoperative laboratory
values and historical factors likely to influence wound
healing or resistance to infection were recorded; they
included smoking history, preoperative hemoglobin
concentration, coexisting systemic diseases, and drug
therapy.
Anesthetic and hemodynamic data, as well as core
temperatures and forearm minus fingertip skintemperature gradients, were measured throughout the
treatment period in 15-min intervals.
Preoperative risk of infection was evaluated using
the CDC Study on the Efficacy of Nosocomial Infection Control (SENIC) score, where 1 point each was
assigned for 3 diagnoses, surgical duration 2 h,
abdominal site of surgery, and the presence of a contaminated or dirty-infected wound (16). Infection risk
was further quantified using the National Nosocomial
Infection Surveillance System (NNISS), in which risk
was predicted based on type of surgery, ASA physical
status rating, and surgical duration (17). Our primary
outcome measure was the incidence of postoperative
wound infection. Surgical wounds were evaluated
daily by a physician blinded to group assignment.
After discharge, the same blinded physician evaluated
patients during their 2-wk clinic visits. Patients not
returning to the clinic were contacted by phone by an
investigator blinded to group assignment and treatment. All wound infections diagnosed in the 15 days
after surgery were considered in the data analysis. We
restricted the diagnostic interval to 15 days because no
infections were detected between 16 and 30 days in
our previous study in this surgical population (2).
Our major outcome was either of two distinct criteria for diagnosis of surgical wound infections. The
first, as in our previous studies (2,3) was purulent
exudate and a positive culture. The second was the
1992 revision (18) of the CDC criteria for surgical
wounds (16) with the exception that we restricted our
diagnostic period to 15 rather than 30 days.

ANESTH ANALG
2005;101:1546 53

Wound healing and infections were numerically

scored using the ASEPSIS system (19). This is an established and validated system for quantifying surgical wound infections and evaluating wound healing.
Subcutaneous oxygen tension (PsqO2) in the upper
arm was evaluated in a subgroup of 56 patients with a
polarographic Clark-type tissue oxygen sensor (Licox,
Inc., Germany). Because the hypothesis of our outcome study was based on the results of this preliminary study, tissue oxygenation in these patients was
reported previously (15). However, randomization
and follow-up of this subgroup were properly performed within the current outcome study.
We recorded the incidence of nausea and vomiting
and antiemetic use in each group, during recovery,
and on the first postoperative day. The severity of
nausea was evaluated after 1 postoperative hour with
a 100-mm visual analog scale (0 mm no nausea;
100 mm worst possible nausea). Postoperative pain
was evaluated with a 100-mm visual analog scale at
30-min intervals during the first hour of recovery and
again after 24 h (0 mm no pain; 100 mm worst
possible pain).
Attending surgeons, who were unaware of the patients group assignment, made the decision to discharge the patient from the hospital. Discharge was
based on routine surgical considerations, including
return of bowel function, control of infections (if any),
and adequate healing of the incision.
Based on previous studies (3), we anticipated that
the incidence of surgical wound infections without
and with supplemental fluid administration would be
11% and 5%, respectively, in patients maintained at a
core temperature of 36C and with 30% oxygen administration. Our original study design thus called for
a total of 750 participants. The prospective criterion for
ending the study after 250 patients was a difference in
the incidence of surgical wound infection with a twotailed P value 0.0125. After 500 patients, the study
was to be stopped if P 0.0161. To compensate for the
2 initial analyses, a P value of 0.0356 was to be required upon completion of all 750 patients. To avoid a
Type I statistical error resulting from multiple comparisons, infection incidence was the sole criterion for
stopping the study early (20).
The number of postoperative wound infections and
all other categorical outcomes in each volume group
were compared using a 2 test with Yates correction
for continuity. Additionally, multivariable logistic regression was used to examine the relationship of surgical wound infection rate and fluid management
while controlling for potential confounding factors.
Wound-healing scores and other ordinal variables
were compared with Mann-Whitney ranked sum test.
Days of hospitalization and other continuous variables were compared with unpaired, two-tailed t-tests.

ANESTH ANALG
2005;101:1546 53

KABON ET AL.
SUPPLEMENTAL FLUID AND WOUND INFECTION

1549

Results
An initial data analysis was performed as planned
after 250 patients. However, an additional six patients
were enrolled while the analysis was being conducted.
These patients were included in the analysis, which
was thus based on 256 patients; of these, 253 completed the trial (Fig. 1).
Based on the results in the initial 253 patients, we
performed 2 sample-size calculations. The first indicated that approximately 750 patients would be required to provide an 80% power for excluding a twofold difference in infection rates in each volume
group. The second indicated that 4100 patients
would be required to have an 80% power to identify a
statistically significant difference between the treatments. The Data Safety Monitoring Board thus concluded that the hypothesis was unlikely to be confirmed even with a very large number of patients; they
also concluded that the apparent effect of supplemental fluid administration (if any) was likely to be clinically unimportant compared with other interventions.
The study was thus stopped.
Demographic and morphometric characteristics
were similar in patients assigned to small-volume (n
124) and large-volume (n 129) fluid management.
Potential confounding factors were also similar.
SENIC and NNISS scores were virtually identical in
the 2 treatment groups (Table 1); anesthetic and surgical management were also similar (Table 2).
Large-volume patients received perioperatively almost twice as much total fluid as the small-volume
patients (5.7 2 versus 3.1 1.5 L) (Table 2). Among
the 253 patients, there were 14 infections in patients
assigned to small fluid management and 11 in those
assigned to large fluid management (P 0.462).
The overall infection rate was 9.9%, which was frequent compared with the 4.2% rate predicted by the
NNISS score. ASEPSIS wound-healing scores were similar in the 2 volume groups, as were the times at which
solid food was tolerated (Table 3). Four of 13 wound
cultures failed to grow pathogenic bacteria. The overall
infection rate at the University of Louisville was 14%,
whereas it was 9% at Washington University (P 0.48).
Wound infection rates did not differ significantly among
patients admitted the morning of surgery (11.9%) and
patients who were admitted the night before surgery
and, therefore, had IV hydration during mechanical
bowel preparation (7.7%; P 0.281).
Although preoperative comorbidities did not differ,
preoperative hemoglobin concentrations were less in
patients who developed infections (11.2 1.7 versus
12.2 1.9 g/dL, P 0.01).
Including hemoglobin concentration, inpatient versus outpatient hydration, comorbidities, and other potential confounding factors in a multivariate analysis

Figure 1. Trial profile. Withdrawn includes patients who withdrew consent at any point during the follow-up period.

confirmed that volume management did not influence

infection rate.
Eleven patients were admitted to the ICU, 3 in the
small-volume group versus 8 in the large-volume
group, P 0.14 (Table 3). None of the patients died.
The duration of hospitalization was virtually identical
in the small and large fluid management groups.
However, the duration of hospitalization was 13.9
10.0 days in the infected patients, whereas it was 6.4
3.0 days in the uninfected patients (P 0.001).
About a quarter of the patients experienced postoperative nausea during the initial 24 postoperative
hours. There was no difference in the incidence of
nausea or vomiting in the 2 volume groups, either in
the first hour or during the first 24 h (Table 4).

Discussion
Infection rates did not differ significantly in the patients assigned to small (11.3%) or large (8.5%) fluid
management. This finding was surprising because our
previous study suggested that aggressive fluid management should decrease infection rate. There are,
however, at least 3 possible reasons why we were
unable to reduce infection rate even though supplemental fluid increases PsqO2 (15). The first is that
although aggressive fluid management increases subcutaneous tissue oxygenation by approximately
15 mm Hg (15) and previous work suggests that a
15-mm Hg improvement in tissue oxygenation is
likely to reduce infection risk (6), this increase is small
compared with providing supplemental inspired oxygen, which doubles tissue oxygenation from approximately 60 to 110 mm Hg (3).

1550

KABON ET AL.
SUPPLEMENTAL FLUID AND WOUND INFECTION

ANESTH ANALG
2005;101:1546 53

Table 1. Patient Characteristics

Sex (male/female)
Weight (kg)
Age (yr)
ASA status (I/II/III)
Smoker (yes/no)
Diagnosis (%)
Cancer
Inflammatory bowel disease
Other
Operative site (%)
Colon
Rectum
Hemoglobin (g/dL)
Study site (St. Louis/Louisville)
Bowel preparation (inpatient/outpatient)
SENIC score 1/2/3 (n)
NNISS score 1/2/3 (n)
Infection rate predicted by NNISS (%)

Small volume
(n 124)

Large volume
(n 129)

60/64
77 18
53 14
13/81/17a
25/99

65/64
78 17
52 15
9/86/18a
25/104

61
30
9

62
32
6

65
35
12.0 1.8
97/27
78/46
49/70/5
68/34/7a
4.2

70
30
12.5 2.1
106/23
84/45
39/86/4
64/42/5a
4.2

P value
0.750
0.689
0.859
0.642
0.876
0.684

0.394
0.424
0.431
0.814
0.248
0.528

a
Some data from these categories were missing. SENIC is the Center for Disease Control Study on the Efficacy of Nosocomial Infection Control. NNISS is the
National Nosocomial Infection Surveillance System score.

Table 2. Perioperative Management

Small volume
(n 124)
Preoperative crystalloid (L)
Intraoperative data
Fentanyl (g)
Isoflurane (%)
Mean arterial blood pressure (mm Hg)
Heart rate (bmp)
Crystalloid (L)
Urine output (mL)
Estimated blood loss (mL)
Red-cell transfusion
Patients (n)
Total units (n)
Duration of surgery (h)
Core temperature (C)
Fio2 (%)
Arterial oxygen saturation (%)
End-tidal Pco2 (mm Hg)
Skin-temperature gradient (C)
Postoperative data
Crystalloid (L)
Oxygen saturation (%)
VAS

Large volume
(n 129)

P value

0.7 0.3
196 156
0.9 0.2
81 10
80 12
2.5 1.3
310 276
333 349

199 133
0.9 0.2
81 9
80 13
3.9 1.9
490 393
322 331

0.866
0.452
0.876
0.745
0.001
0.001
0.796

11
22
2.6 1.1
35.9 0.5
45.8 14.5
98.9 1.1
32.5 2.9
0.5 1.4

9
22
2.6 1.0
35.9 0.6
43.3 10.3
99.0 1.1
32.0 2.7
0.5 1.3

0.312

0.909
0.720
0.123
0.556
0.152
0.972

0.6 0.3
98.4 3.2
39.9 26.3

1.1 0.4
97.8 3.3
39.5 28.3

0.001
0.099
0.911

VAS is visual analog pain scores. Postoperative results were obtained during the first hour of recovery. Skin-temperature gradient is calculated as forearm
temperature minus fingertip temperature; values exceeding 0C indicate vasoconstriction.

The second factor is that in our previous study we

measured tissue oxygen partial pressure in a surrogate
wound in the upper arm rather than adjacent to the
abdominal wound or in the intestines per se. Oxygenation at the arm, wound, and colon are similarly improved by supplemental inspired oxygen (2123).

However, supplemental fluid administration, specifically crystalloids, might affect tissue oxygenation differently in injured and uninjured tissue. Thus, in this
specific case, a surrogate wound on the arm might not
reflect perfusion and oxygenation of the actual
wound. It is a major limitation of our study that we

ANESTH ANALG
2005;101:1546 53

KABON ET AL.
SUPPLEMENTAL FLUID AND WOUND INFECTION

1551

Table 3. Principal Results

Infection diagnosed by pus and a positive culture (%)

Infection diagnosed by CDC criteria (%)
Superficial
Deep
Peritoneal
Any CDC infection
Total infection by either criterion (%)
ASEPSIS score
Intensive care admission (%)
First solid food (postoperative days)
Duration of hospitalization (days)

Small volume

Large volume

P value

5.7

4.7

0.720

6.5
5.7
0.0
10.5
11.3
8 14
2.4
4.2 1.9
7.3 4.0

3.9
5.4
2.3
7.0
8.5
7 16
6.2
4.4 2.7
7.0 5.4

0.354
0.939
0.247
0.322
0.462
0.698
0.140
0.448
0.701

CDC is Center for Disease Control. ASEPSIS is a wound-healing score (19). Infections as diagnosed by pus and a positive culture and by CDC criteria are not
mutually exclusive. Consequently, the total infection column is not the sum of each infection type.

Table 4. Postoperative Nausea and Vomiting

Small
Large
volume volume P value
1 h after surgery (n 253)
Nausea
Vomiting
Rescue ondansetron
024 h after surgery (n 203)a
Nausea
Vomiting
Rescue ondansetron

12.1
0.8
8.9

12.4
0.0
5.4

0.941
1.00
0.287

24.7
4.1
18.6

28.3
2.8
16.0

0.566
0.614
0.635

Data are percentage.

a
Data collected at St. Louis site only.

did not evaluate wound tissue oxygen tension. Furthermore, recent data, which were unavailable at the
time of our study, suggest that intestinal oxygenation
in swine is minimally influenced by hydration (24).
Hydration strategies, at least in the range tested, thus
seem to have less effect on intestinal oxygenation than
other proven methods of reducing infection risk such
as supplemental oxygen administration (24).
The third factor is that our current results provide
only a 37% power for detecting a factor-of-two treatment effect. It thus remains possible that supplemental hydration does reduce infection risk. Although our
results thus suggest that the effects of hydration are
small in our general surgical patient population, in
some specific high-risk patient populations, any decrease of wound infection rate might be important.
It is important to recognize that even our smallvolume group was given adequate fluid and was not
hypovolemic by any routine clinical criteria. For example, mean arterial blood pressure and heart rate
were virtually identical in the two groups. That supplemental fluid administration did not reduce infection risk is by no means an indication that hypovolemia is harmless. In fact, hypovolemia reduces healing
of colonic anastomoses (25).

Our primary result was that the risk of surgical

wound infection was similar in patients given 3.1 versus 5.7 L of crystalloid perioperatively; bowel function
and duration of hospitalization were also similar in
the 2 treatment groups. This result contrasts with
those of Brandstrup et al. (26) who found that fluid
restriction (with the goal of maintaining an unchanged
body weight) reduced the risk of complications in
patients undergoing colon resection. They also differ from Holte et al. (27) who reported that a fixed
regimen of 3000 mL versus 500 mL of crystalloid
for elective laparoscopic cholecystectomy improved
outcome.
Additionally, numerous studies report that
Doppler-directed colloid administration improves recovery characteristics and shortens hospitalization.
For example, Gan et al. (12) found that goal-directed
colloid administration reduced the risk of complications, sped return of bowel function, and shortened
hospitalization in patients undergoing major surgery
with anticipated blood loss exceeding 500 mL. Mythen
and Webb (13) reported that Doppler-directed colloid
administration improved gut mucosal perfusion (as
determined by gastric tonometry), reduced the risk of
complications, and shortened hospitalization in cardiac surgery patients. And finally, Sinclair et al. (14)
similarly reported that goal-directed colloid administration improved recovery characteristics and shortened hospitalization in patients undergoing hip fracture repair.
Taken together, available literature indicates that
goal-directed colloid administration improves outcomes, especially shortening the duration of hospitalization. Fixed-volume regimens, in contrast, are variously reported as being detrimental, neutral, or
beneficial. It thus seems likely that variability among
patients precludes selecting a fixed mL kg1 h1
fluid regimen that will optimally serve most patients,
even within a particular type of surgery. We also note
that study designs concerning types and amounts of
fluids and primary outcomes in the major studies

1552

KABON ET AL.
SUPPLEMENTAL FLUID AND WOUND INFECTION

often differed widely. Furthermore, fluid regimens

that are optimal for one organ or system may well
prove detrimental to another. Results of vascular volume studies must thus be interpreted in light of the
type of fluid, management strategy, patient population, and specific outcomes.
We (28) and others (29) have shown that supplemental perioperative fluid administration reduces the
risk of postoperative nausea and vomiting. For example, patients given an initial loading dose of 15 mL/kg
crystalloid experienced less nausea and vomiting after
gynecologic surgery than those given 2 mL/kg (28).
However, the incidence of nausea and vomiting during the first 24 hours after surgery was similar in our
current treatment groups. This is consistent with other
studies in which a benefit of supplemental fluid administration was not apparent (30,31).
Our overall infection rate was approximately twice
that predicted by the NNISS score (17). In this respect,
the current results are similar to our previous ones
(2,3). It is likely that our infection rate was relatively
frequent in part because the studies were primarily
conducted in tertiary referral hospitals that cater to
relatively sick patients who often have serious underlying conditions. We also included patients at high
risk of infection including those taking immunosuppressant medications. And, finally, it is worth noting
that the NNISS data set is limited by being based on
self-reports by surgeons, whereas wounds in our studies were evaluated daily by trained physicians using
two sets of strict diagnostic criteria. However, our
observed infection rate is still less than in most recent
studies, which report infection rates in similar patient
populations 15% (26,32). Our patient population was
diverse, but no more so than in all the other studies
mentioned. We performed a multivariant analysis to
evaluate the potential contribution of coexisting diseases and found that there was no evidence that preexisting patient factors contributed substantially to
our results or would alter our conclusions. Our infected patients remained in the hospital a full week
longer than uninfected patients, suggesting that our
diagnostic criteria were appropriate and that our infections were indeed clinically important.
Approximately 5% of our patients were admitted to
the ICU, twice as many in the large-volume group as
compared with the small volume. One 63-year-old
female patient (ASA II with no cardiac history) from
the large-volume group was admitted to the ICU because she developed pulmonary edema. Treatment
was conservative without any ventilatory support,
and she was discharged from the ICU the next day.
She had no further complications and was discharged
from hospital on day 8. Three patients, all in the
large-volume group, were admitted because of
surgical-site infection. Two of them developed deep
infection and sepsis. Other reasons for ICU admission

ANESTH ANALG
2005;101:1546 53

were myocardial infarction (large-volume group),

atrial arrhythmia (large-volume group), postoperative
psychoses (both groups) or because of surgical complications. However, ICU admissions were not our
primary outcome and thus we neither can draw any
conclusions about the association of group assignments and ICU admission nor about fluid management and the different complications leading to these
admissions.
Our patients were all anesthetized with nitrous oxide, and although there are many advantages to nitrous oxide, its use precludes simultaneous administration of supplemental oxygen. Although we have
demonstrated that supplemental oxygen halves infection risk (3), a recent smaller study by Pryor et al. (32)
concluded just the opposite. Furthermore, the administration of 80% oxygen was not standard of care at the
time when we performed this study and, in fact, is still
not as of the writing of this article. Infection rates may
have been less had we substituted supplemental oxygen for nitrous oxide. However, it seems unlikely that
doing so would alter our overall conclusion that supplemental fluid administration does not reduce infection risk.
In summary, infection rates were similar in patients
given small fluid replacement (approximately 3.100 mL)
and in those given large fluid replacement (approximately 5.700 mL). The apparent lack of benefit from
supplemental fluid may have resulted because the effect
of hydration on intestinal oxygenation is modest or because the statistical power of our study was limited.
Nonetheless, our results suggest that supplemental hydration in the range tested does not have a major impact
on wound infection risk.
We greatly appreciate the assistance of Anthony G. Doufas, MD
(Department of Anesthesiology, University of Louisville), Yunus
Muneer Shaw, MD (Department of Anesthesiology, University of
Louisville), and Nobutada Morioka, MD (Outcomes Research
Institute, University of Louisville). We also thank Gilbert Haugh,
MA, for statistical assistance, Nancy Alsip, PhD, for editorial assistance, and Diane Delong, RN, for patient recruitment and followup coordination (Outcomes Research Institute, University of
Louisville).

References
1. Haley RW, Culver DH, Morgan WM, et al. Identifying patients
at high risk of surgical wound infection: a simple multivariate
index of patient susceptibility and wound contamination. Am J
Epidemiol 1985;121:206 15.
2. Kurz A, Sessler DI, Lenhardt R. Perioperative normothermia to
reduce the incidence of surgical-wound infection and shorten
hospitalization. Study of Wound Infection and Temperature
Group. N Engl J Med 1996;334:1209 15.

ANESTH ANALG
2005;101:1546 53

3. Greif R, Akca O, Horn EP, et al. Supplemental perioperative

oxygen to reduce the incidence of surgical-wound infection.
Outcomes Research Group. N Engl J Med 2000;342:1617.
4. Bremmelgaard A, Raahave D, Beir-Holgersen R, et al.
Computer-aided surveillance of surgical infections and identification of risk factors. J Hosp Infect 1989;13:118.
5. Jonsson K, Hunt TK, Mathes SJ. Oxygen as an isolated variable
influences resistance to infection. Ann Surg 1988;208:7837.
6. Hopf HW, Hunt TK, West JM, et al. Wound tissue oxygen
tension predicts the risk of wound infection in surgical patients.
Arch Surg 1997;132:9971004; discussion 5.
7. Prockop DJ, Kivirikko KI, Tuderman L, Guzman NA. The biosynthesis of collagen and its disorders. Part 1. N Engl J Med
1979;301:1323.
8. Sheffield CW, Sessler DI, Hopf HW, et al. Centrally and locally
mediated thermoregulatory responses alter subcutaneous oxygen tension. Wound Rep Reg 1997;4:339 45.
9. Jonsson K, Jensen JA, Goodson WH 3rd, et al. Assessment of
perfusion in postoperative patients using tissue oxygen measurements. Br J Surg 1987;74:2637.
10. Miles AA, Miles EM, Burke JP. The value and duration of
defense reactions of the skin to the primary lodgement of bacteria. Br J Exp Pathol 1957;38:79.
11. Hartmann M, Jonsson K, Zederfeldt B. Importance of dehydration in anastomotic and subcutaneous wound healing: an experimental study in rats. Eur J Surg 1992;158:79 82.
12. Gan TJ, Soppitt A, Maroof M, et al. Goal-directed intraoperative
fluid administration reduces length of hospital stay after major
surgery. Anesthesiology 2002;97:820 6.
13. Mythen MG, Webb AR. Perioperative plasma volume expansion reduces the incidence of gut mucosal hypoperfusion during
cardiac surgery. Arch Surg 1995;130:4239.
14. Sinclair S, James S, Singer M. Intraoperative intravascular volume optimisation and length of hospital stay after repair of
proximal femoral fracture: randomised controlled trial. BMJ
1997;315:909 12.
15. Arkilic CF, Taguchi A, Sharma N, et al. Supplemental perioperative fluid administration increases tissue oxygen pressure.
Surgery 2003;133:49 55.
16. Haley RW, Morgan WM, Culver DH, et al. Update from the
SENIC project. Hospital infection control: recent progress and
opportunities under prospective payment. Am J Infect Control
1985;13:97108.
17. Culver DH, Horan TC, Gaynes RP, et al. Surgical wound infection rates by wound class, operative procedure, and patient risk
index. National Nosocomial Infections Surveillance System.
Am J Med 1991;91:152S7.

KABON ET AL.
SUPPLEMENTAL FLUID AND WOUND INFECTION

1553

18. Horan TC, Gaynes RP, Martone WJ, et al. CDC definitions of
nosocomial surgical site infections, 1992: a modification of CDC
definitions of surgical wound infections. Infect Control Hosp
Epidemiol 1992;13:606 8.
19. Byrne DJ, Malek MM, Davey PG, Cuschieri A. Postoperative
wound scoring. Biomed Pharmacother 1989;43:669 73.
20. Fleming TR, Harrington DP, OBrien PC. Designs for group
sequential tests. Control Clin Trials 1984;5:348 61.
21. Chang N, Goodson WH 3rd, Gottrup F, Hunt TK. Direct measurement of wound and tissue oxygen tension in postoperative
patients. Ann Surg 1983;197:470 8.
22. Ratnaraj J, Kabon B, Talcott MR, et al. Supplemental oxygen and
carbon dioxide each increase subcutaneous and intestinal intramural oxygenation. Anesth Analg 2004;99:20711.
23. Kabon B, Nagele A, Reddy D, et al. Obesity decreases perioperative tissue oxygenation. Anesthesiology 2004;100:274 80.
24. Hiltebrand LB, Hager H, Pestel G, et al. Effects of different
volume treatments on intestinal and subcutaneous tissue oxygenation during anesthesia. Anesthesiology 2004;101:A666.
25. Foster ME, Laycock JR, Silver IA, Leaper DJ. Hypovolaemia and
healing in colonic anastomoses. Br J Surg 1985;72:831 4.
26. Brandstrup B, Tonnesen H, Beier-Holgersen R, et al. Effects of
intravenous fluid restriction on postoperative complications:
comparison of two perioperative fluid regimens: a randomized
assessor-blinded multicenter trial. Ann Surg 2003;238:641 8.
27. Holte K, Klarskov B, Christensen DS, et al. Effect of fluid administration on recovery after laparoscopic cholecystectomie: a
randomized, double blind study. Eur J Anaesthesiol 2003;
20(Suppl 30):A-36.
28. Ali SZ, Taguchi A, Holtmann B, Kurz A. Effect of supplemental
pre-operative fluid on postoperative nausea and vomiting. Anaesthesia 2003;58:780 4.
29. Magner JJ, McCaul C, Carton E, et al. Effect of intraoperative
intravenous crystalloid infusion on postoperative nausea and
vomiting after gynaecological laparoscopy: comparison of 30
and 10 ml kg-1. Br J Anaesth 2004;93:3815.
30. McCaul C, Moran C, OCronin D, et al. Intravenous fluid loading with or without supplementary dextrose does not prevent
nausea, vomiting and pain after laparoscopy. Can J Anaesth
2003;50:440 4.
31. Spencer EM. Intravenous fluids in minor gynaecological surgery. Their effect on postoperative morbidity. Anaesthesia 1988;
43:1050 1.
32. Pryor KO, Fahey TJ 3rd, Lien CA, Goldstein PA. Surgical site
infection and the routine use of perioperative hyperoxia in a
general surgical population: a randomized controlled trial.
JAMA 2004;291:79 87.

CASE REPORT

Hypoxemia During One-Lung Ventilation: Jet Ventilation of

the Middle and Lower Lobes During Right Upper Lobe
Sleeve Resection
Ju-Mei Ng,

FANZCA

Department of Anaesthesia & Surgical Intensive Care, Singapore General Hospital, Singapore

A 64-yr-old man underwent right thoracotomy and upper lobectomy for lung carcinoma. Hypoxemia on onelung ventilation was being managed with continuous
positive airway pressure to the nondependent lung
when a sleeve resection had to be performed. As this
positive airway pressure would no longer be maintained with the bronchus open, an alternate method of

redicting which patients will desaturate during

one-lung ventilation (OLV) is difficult (1). Different methods of improving arterial oxygenation on OLV include intermittent inflation of the
collapsed lung with oxygen (2), administration of IV
almitrine (3) nitric oxide (4) or both (5), lung recruitment (6), and application of continuous positive airway pressure (CPAP) to the nondependent lung (7,8).
This case report describes the use of CPAP to the right
lung, which was converted to high-frequency jet ventilation (HFJV) of the middle and lower lobes during
unexpected right upper lobe sleeve resection.

Case Report
A 64-yr old man with lung carcinoma presented for right
upper lobectomy. He had a 40 pack-year smoking history
and single vessel coronary artery disease. Preoperative arterial blood gases (Fio2, 0.21) showed a Pao2 of 91 mm Hg and
Paco2 of 39 mm Hg. His forced expiratory volume in 1 s was
1.43 L and forced vital capacity was 2.58 L. An epidural
catheter was placed at T5-6 level under local anesthesia, and
induction of anesthesia was uneventful. His trachea was
intubated with a left-sided 39F double-lumen tube (DLT).
Correct position and depth of the DLT in the left main
bronchus was verified fiberoptically in both supine and left
Accepted for publication May 3, 2005.
Address correspondence and reprint requests to J. M. Ng, FANZCA, Department of Anesthesia & Surgical Intensive Care, Singapore General Hospital, Outram Road, Singapore 169608, Republic of
Singapore. Address electronic mail to gannjm@sgh.com.sg.
DOI: 10.1213/01.ANE.0000180834.13549.E4

1554

Anesth Analg 2005;101:15545

oxygenation was necessary. This report describes the

successful use of jet ventilation via an airway exchange
catheter placed in the bronchus intermedius through
the tracheal lumen of a left-sided double-lumen endobronchial tube. Oxygenation was maintained and surgical access was good during the 15-min resection.
(Anesth Analg 2005;101:1554 5)

lateral decubitus positions. The patient received a 6 mL

epidural bolus of 0.25% bupivacaine followed by an infusion
of 0.2% ropivacaine fentanyl 2 g/mL at 6 mL/h. Anesthesia was maintained with desflurane (end-tidal concentration 5.0% 6.0%) in 100% oxygen and atracurium at 20 mg/h.
A muscle-sparing thoracotomy incision was made over the
right fifth intercostal space.
Ten minutes into OLV, oxygen saturation (Sao2) decreased
from 97%98% on two-lung ventilation to 84% 85%. He was
ventilated with tidal volume 8 mL/kg and respiratory rate
of 12 breaths/min, and peak airway pressures were 28 cm
H2O. Urgent bronchoscopy showed good positioning of the
DLT and there was good air entry in the dependent lung
with no adventitious sounds on auscultation. Positive endexpiratory pressure of 5 cm H2O was applied to the dependent lung with no improvement. Increasing the positive
end-expiratory pressure to 7.5 cm H2O worsened the Sao2.
With 5 cm H2O of dependent lung positive end-expiratory
pressure and 5 cm H2O of nondependent lung CPAP, Sao2
improved to 95%96%. CPAP was decreased to 2 cm H2O
and Sao2 remained at 94%95%. Operative conditions were
satisfactory. Unfortunately, to attain tumor-free margins, a
sleeve resection was necessary, and CPAP was no longer
feasible. Discontinuation of CPAP resulted in desaturation
within 5 min. An 11F 83-cm airway exchange catheter (CCAE-11.0 83-DLT; Cook Inc, Bloomington, IN) was placed
in the bronchus intermedius. As the catheter and fiberoptic
bronchoscope were too large to be passed concurrently
through the tracheal lumen of the DLT, the bronchoscope
was first passed until its tip was at the distal bronchus
intermedius. The marking at the proximal tracheal lumen
was noted. With the bronchoscope withdrawn, the catheter
was then inserted to this same distance through the tracheal
lumen (Fig. 1). This allowed for HFJV of the middle and
lower lobes with the bronchus open. The Monsoon Universal Jet Ventilator (v2.1e; Innovmedics Pte Ltd., Singapore)
with a respiratory rate of 150 breaths/min, inspiratory time
2005 by the International Anesthesia Research Society
0003-2999/05

ANESTH ANALG
2005;101:1554 5

Figure 1. The airway exchange catheter in the bronchus intermedius. RUL right upper lobe; RML right middle lobe; RLL
right lower lobe.

30%, and driving pressure of 1 mbar was used. Sao2 immediately improved to 98%99% and surgical exposure was
adequate. Inflation of the middle and lower lobes was observed. HFJV was continued for 15 min until bronchial
closure. CPAP with 2 cm H2O was then resumed until
clamping of the supplying pulmonary artery and improvement in the Sao2.

Discussion
Although this patient had pathology in his nonventilated lung and an obstructive pattern on preoperative
pulmonary function tests, he developed hypoxemia
on OLV. His hypoxemia was corrected with application of CPAP to the nondependent lung. Unfortunately, a sleeve resection had to be performed and an
alternate means of treating hypoxemia had to be formulated, as CPAP could no longer be maintained with
the open bronchus exposed to atmospheric pressure.
HFJV provides satisfactory oxygenation and good surgical access during OLV for thoracotomy (9 11),
where it has been delivered via a DLT (9), a small
uncuffed endobronchial tube (10), and the lumen of
the bronchial blocker of a Univent tube (11). In this
patient, selective HFJV to the right middle and lower
lobes would be comparable to selective lobar collapse
of the right upper lobe (12). This not only maintained
oxygenation but also provided optimal surgical access
by collapsing the right upper lobe. It was achieved,
without manipulation of the patients DLT, by passing
a catheter of adequate length such that its tip would lie
well within the bronchus intermedius and beyond the
surgical incision in the right main bronchus. Although
the lower and middle lobes were inflated, there were
minimal respiratory movements and with the right

CASE REPORT

1555

upper lobe collapsed, there was adequate surgical exposure and good operative conditions. Expiratory
flow limitation leading to increased lung volume, especially in the context of chronic obstructive airway
disease, is an eminent danger. Care was taken to ensure proper positioning of the catheter fiberoptically
and the catheter secured to prevent distal migration
into a single lobe, which would further increase the
risk of barotrauma.
Although the use of a ventilating catheter passed
into the distal bronchus for tracheal and carinal resection has been described, this case shows an adaptation; treatment of hypoxemia on OLV by placement in
the bronchus intermedius. The indications for HFJV
are expanded beyond ventilating both lungs or a single lung to that of ventilating 2 lobes during management of hypoxemia on OLV. For those without ready
access to HFJV, an intermediate ventilation mode (e.g.,
low frequency jet ventilation with an injector) may be
a viable option in a similar situation. This case demonstrates a novel way of managing hypoxia on OLV
when the nondependent bronchus is open to the
atmosphere.

References
1. Guenoun T, Journois D, Silleran-Chassany J, et al. Prediction of
arterial oxygen tension during one-lung ventilation: analysis of
preoperative and intraoperative variables. J Cardiothorac Vasc
Anesth 2002;16:199 203.
2. Malmkvist G. Maintenance of oxygenation during one-lung
ventilation: effect of intermittent reinflation of the collapsed
lung with oxygen. Anesth Analg 1989;68:763 6.
3. Dalibon N, Moutafis M, Liu N, et al. Treatment of hypoxemia
during one-lung ventilation using intravenous almitrine.
Anesth Analg 2004;98:590 4.
4. Sticher J, Scholz S, Boning O, et al. Small-dose nitric oxide
improves oxygenation during one-lung ventilation: an experimental study. Anesth Analg 2002;95:1557 62.
5. Moutafis M, Liu N, Dalibon N, et al. The effects of inhaled nitric
oxide and its combination with intravenous almitrine on Pao2
during one-lung ventilation in patients undergoing thoracoscopic procedures. Anesth Analg 1997;85:1130 5.
6. Tusman G, Bohm SH, Sipmann FS, Maisch S. Lung recruitment
improves the efficiency of ventilation and gas exchange during
one-lung ventilation anesthesia. Anesth Analg 2004;98:1604 9.
7. Capan LM, Turndorf H, Patel C, et al. Optimization of arterial
oxygenation during one-lung anesthesia. Anesth Analg 1980;59:
84751.
8. Slinger P, Triolet W, Wilson J. Improving arterial oxygenation
during one-lung ventilation. Anesthesiology 1988;68:2915.
9. Nakatsuka M, Wetstein L, Keenan RL. Unilateral highfrequency jet ventilation during one-lung ventilation for thoracotomy. Ann Thorac Surg 1988;46:654 60.
10. El-Baz NM, Kittle CF, Faber LP, Welsher W. High-frequency
ventilation with an uncuffed endobronchial tube: a new technique for one-lung anesthesia. J Thorac Cardiovasc Surg 1982;
846:823 8.
11. Williams H, Gothard J. Jet ventilation via a Univent tube for
sleeve pneumonectomy. Eur J Anaesthesiol 2001;186:4079.
12. Campos JH. Effects on oxygenation during selective lobar versus total lung collapse with our without continuous positive
airway pressure. Anesth Analg 1997;85:583 6.

MEETING ARTICLE

Report of the 13th Annual Meeting of the International

Society for Anaesthetic Pharmacology
John R. Keltner,

MD, PhD,

and Pamela Flood,

MD

Department of Anesthesia University of California, San Francisco, San Francisco, California

he International Society of Anaesthetic Pharmacology (ISAP) held its 13th Annual Meeting on
October 22, 2004, in Las Vegas, Nevada. The
major theme this year was The Pharmacology of Pain
and Analgesia. The Meeting Chair was Steven Shafer
(Stanford University), and the Meeting Vice Chair was
Pamela Flood (Columbia University). Dr. Shafer and
Dr. Flood welcomed the participants, noting their
pleasure at the collection of basic and clinical scientists
as well as scientists from industry. It is the hope of the
organizers that the combination of backgrounds and
expertise will evoke fruitful discussion that will propel the field into the future.
The first session, moderated by Dr. Tony Gin (Chinese University of Hong Kong), included lectures on
the current understanding of pain circuitry as well as
descending modulatory systems. Dr. Timothy J. Brennan (University of Iowa) discussed the special characteristics of postoperative pain. He noted that after
surgery, pain induced by patient movement is relatively refractory to parental opiates and persists for
23 days. Dr. Brennan has pioneered a rat model for
postoperative pain that has a number of similarities to
clinical postoperative pain, including its timing and
pharmacological susceptibility. Dr. Brennan also described studies using primary afferent fibers to study
electrical activity induced by nociceptive input. Results from such studies suggest that background or
continuing activity of A and C-fibers in the region of
surgical incisions may signal pain at rest. Possible
mediators for rest pain are inflammatory cytokines
that activate the fibers or direct trauma caused to
terminals of these nerve fibers. More proximally, central sensitization from postoperative pain can also be
appreciated in dorsal horn neuron recordings. Pain on
movement and limb protection may have features attributable to mechanical allodynia.
Accepted for publication June 9, 2005
Address correspondence to John R. Keltner, MD, PhD, Department of Anesthesia University of California, San Francisco, Box
1654, 2255 Post Street, MZ Bldg N PAIN, San Francisco, CA 94143
1654. Address electronic mail to keltnerj@anesthesia.ucsf.edu.
DOI: 10.1213/01.ANE.0000181328.51477.A2

1556

Anesth Analg 2005;101:15567

Also in the first session, Dr. Donna L. Hammond

(University of Iowa) presented the current thinking on
the descending modulatory systems that affect pain
perception. Her summary included descriptions of her
work and Dr. Howard Fields work characterizing
descending pain modulation circuits in the brainstem:
the periaqueductal gray, the dorsal lateral pontine
tegmentem, and the rostral ventral medulla. Dr. Hammond made the interesting point that efferent modulatory pain pathways, similar to afferent pain pathways, are susceptible to long-term plasticity
secondary to persistent pain states.
The second session, moderated by Dr. Talmage
Egan (University of Utah), included lectures on cancer
pain and brain imaging of pain. Dr. Patrick W. Mantyh
(University of Minnesota) has developed a mouse
model of metastatic cancer spread to bones. He has
studied local mechanisms that mediate the pain
caused by metastatic tumors as well as central reorganization of spinal cord that is induced by neurochemical signaling. His work suggests there are inflammatory and neuropathic components and components
specific to the tumor itself that participate in the generation and maintenance of cancer pain. Defining
when and how these different components contribute
to cancer pain will be pivotal for the development of
new and more selective analgesics for the treatment of
cancer pain.
In the second half of the second session, Dr. Robert
C. Coghill (Wake Forest University) presented compelling images of the brain during pain perception. Dr.
Coghill showed that the degree of regional activation
in brain images of pain perception was closely correlated with individual subject visual analog scale reports of pain stimuli. This important result helps establish that metabolic brain activity meaningfully
reflects individuals experiences of pain. Dr. Coghill
also demonstrated, with an elegant series of experiments, that an individuals expectation of pain affects
his/her experience of pain, independent of the degree
of nociceptive input.
The third session, moderated by Dr. Steven Shafer,
included a presentation entitled Perspectives on the
2005 by the International Anesthesia Research Society
0003-2999/05

ANESTH ANALG
2005;101:1556 7

FDA by Dr. Bob A. Rappaport as well as discussions

of three featured abstracts. Dr. Rappaport discussed 1)
the regulatory framework that guides the decisionmaking process at the Food and Drug Administration
(FDA), 2) an overview of the regulatory role in drug
development and marketing, and 3) an overview of
the agencys draft analgesic drug development guidance document.
Dr. Satoru Sakurai presented the Zsigman Award
winning poster, Exogenous ATP Potentiates and
Aminophyline Reverses Propofol-Induced Sedative/
Hypnotic Effects as Assessed by BIS in Human Volunteers. This poster suggests that the sedative/
hypnotic effects of propofol may be mediated at least
in part via central purinergic mechanisms. Another
noted poster was titled Beta-blockers Alter Initial
Drug Distribution by Decreasing Cardiac Output and
Altering Blood Flow Distribution in Man, by Dr. Tom
C. Krejcie (Northwestern University). Finally, the
poster The Influence of Propofol Administration Rate
on the ke0 value. Preliminary Results, by Dr. Nicolaas
De Neve concludes that the ke0 of propofol depends
on the time course of drug delivery.
The final session was moderated by Dr. Pamela Flood.
Dr. James C. Eisenach (Wake Forest University) presented a review of his Top 10 Papers in Pain and
Analgesic Pharmacology This Year. Dr. Eisenachs top
paper choice, What Decline in Pain Intensity is Meaningful to Patients with Acute Pain? (Cepeda et al., Pain
2003;105:1517) describes agreement of pain scales and
the clinically meaningful effects of drugs on the scales.

MEETING REPORT
KELTNER AND FLOOD
13TH ANNUAL MEETING OF ISAP

1557

The plenary lecture, The Neurobiology of Acute

and Persistent Pain, was presented by Dr. Allan I.
Basbaum (UCSF). Dr. Basbaum described how persistent pain states, such as arthritis or complex regional
pain syndrome, defy our conventional understanding
of pain mechanisms. Recent studies have established
that clinical pain results from changes in the organization of pain transmission systems. Furthermore, it
appears that the circuits that underlie the transmission
of pain signals are dynamic. Therefore, chronic pain is
not merely prolonged acute pain but is rather a manifestation of an altered nervous system. Dr. Basbaum
discussed molecular changes in the nervous system
that have been implicated in the development of persistent pain, including changes in receptor systems
and alterations in neural networks. Finally, Dr. Basbaum argued that directing clinical pain treatments at
processes underlying the maintenance of chronic pain,
as opposed to symptoms produced by pain, is likely to
be more effective in long-term pain management.
These presentations affirmed that significant
progress has been made in our understanding of the
pharmacology of acute and chronic pain. This meeting
also confirmed that the future promises further exciting developments in pain pharmacology. We look
forward to welcoming you at our 14th Annual ISAP
Meeting next year in New Orleans, Louisiana on the
Friday before the meeting of the American Society of
Anesthesiologists, which will focus on the use of Genetics as a Research Tool.

Cochrane Corner
Nonsteroidal Antiinflammatory Drugs and
Perioperative Bleeding in Pediatric
Tonsillectomy

Interventions for Protecting Renal Function

in the Perioperative Period

Cardwell M, Siviter G, Smith A

Zacharias M, Gilmore ICS, Herbison GP, Sivalingam

P, Walker RJ

Background: Nonsteroidal antiinflammatory drugs (NSAIDs) are used

for pain relief after tonsillectomy in children. However, as they inhibit
platelet aggregation and prolong bleeding time, they could cause increased perioperative bleeding. The overall risk remains unclear.
Objectives: The primary objective of this review was to assess the
effects of NSAIDs on bleeding for pediatric tonsillectomy. There is
good evidence (Kokki 2003; Romsing 1997) to show that NSAIDs are
effective analgesics in children. It was not the purpose of our review
to question this but rather to assess the risk of bleeding when
NSAIDs are used for pain relief after pediatric tonsillectomy.
Search strategy: We searched the Cochrane Central Register of Controlled Trials (CENTRAL) (The Cochrane Library, Issue 3, 2004); MEDLINE (inception until August 2004); EMBASE (from inception until
August 2004), Current Problems (produced by the UK Medicines Control Agency); MedWatch (produced by the United States Food and
Drug Administration) and the Australian Adverse Drug Reactions
Bulletin in December 2001. The Cochrane Anesthesia Review Groups
hand-search coordinator performed hand-searching as required. We
also contacted manufacturers and researchers in the field.
Selection criteria: We included randomized controlled trials assessing NSAIDs in children up to and including 16 yr of age undergoing
elective tonsillectomy or adenotonsillectomy.
Data collection and analysis: Two authors independently assessed
trial quality and extracted the data. We contacted study authors
when necessary for additional information. We also collected information on adverse effects from the trials.
Main results: We included 13 trials involving 955 children. All included
trials compared NSAIDs with other analgesics or placebo and evaluated bleeding requiring surgical intervention. NSAIDs did not significantly alter number of perioperative bleeding events requiring surgical intervention; Peto odds ratio 1.46 (95% confidence interval, 0.49 to
4.40). Seven trials involving 471 children evaluated bleeding not requiring surgical intervention. NSAIDs did not significantly alter number of perioperative bleeding events not requiring surgical intervention; Peto odds ratio 1.23 (95% confidence interval, 0.44 to 3.43). Ten
trials involving 837 children considered postoperative nausea and
vomiting. There was less nausea and vomiting when NSAIDs were
used as part of the analgesic regime compared with when NSAIDs
were not used; odds ratio 0.40 (95% confidence interval, 0.23 to 0.72).
Reviewers conclusions: NSAIDs did not cause any increase in bleeding requiring a return to the operating room. There was significantly
less nausea and vomiting when NSAIDs were used compared to
alternative analgesics.
Citation: Cardwell M, Siviter G, Smith A. Non-steroidal antiinflammatory drugs and perioperative bleeding in paediatric tonsillectomy (Cochrane Review). In: The Cochrane Library, Issue 2,
2005. Chichester, UK: John Wiley & Sons, Ltd. Copyright Cochrane
Library; reproduced with permission.

Background: A number of methods have been used to try to protect

kidney function in patients undergoing surgery. These include the
administration of dopamine, diuretics, calcium channel blockers, angiotensin converting enzyme (ACE) inhibitors and hydration fluids.
Objectives: For this review, we selected randomized controlled trials,
which used different methods to protect renal function during the
perioperative period. In examining these trials, we considered outcomes such as renal failure and mortality, as well as changes in the
renal function tests, including urine output, creatinine clearance,
free water clearance, fractional excretion of sodium and renal
plasma flow.
Search strategy: We searched the Cochrane Central register of
Controlled Trials (CENTRAL) (The Cochrane Library, Issue 4,
2004), MEDLINE (1966 to 2004) and EMBASE (1988 to 2004) and
hand-searched six journals (British Journal of Anesthesia; Anesthesia & Analgesia; Anesthesiology; Annals of Surgery; Journal
of Thoracic and Cardiovascular Surgery and Journal of Vascular
Surgery).
Selection criteria: We selected all randomized controlled trials in an
adult population undergoing surgery in which a treatment measure
was used for the purpose of renal protection in the perioperative
period.
Data collection and analysis: We selected 37 studies for inclusion in
this review. As well as analysis of the data from all the studies, we
also performed subgroup analysis for type of interventions, types of
surgical procedures and those with preexisting renal dysfunction.
We undertook sensitivity analysis on studies with high methodological quality.
Main results: The review included data from 37 studies, comprising
a total of 1227 patients. Of these, 658 received some form of treatment and 569 acted as controls. The interventions were mostly using
different drugs such as dopamine, diuretics, calcium channel blockers, ACE inhibitors, or selected hydration fluids. The results indicated that certain interventions showed some benefits, but all the
results suffered from significant heterogeneity. Hence we can draw
no conclusions about the effectiveness of these interventions in
protecting the kidneys during surgery.
Reviewers conclusions: There is no reliable evidence from available
literature to suggest that interventions during surgery can protect
the kidneys from damage. However, there is a need for more studies
of high methodological quality. One particular area for further
studies may be on patients with preexisting renal dysfunction undergoing surgery.
Citation: Zacharias M, Gilmore ICS, Herbison GP, Sivalingam P,
Walker RJ. Interventions for protecting renal function in the perioperative period. (Cochrane Review). In: The Cochrane Library,
Issue 3, 2005. Chichester, UK: John Wiley & Sons, Ltd. Copyright
Cochrane Library, reproduced with permission

1558

Anesth Analg 2005;101:1558

2005 by the International Anesthesia Research Society

0003-2999/05

Letters to the Editor

Anesthetic Depth and Long-Term Mortality
To the Editor:
As I read Cohens editorial about anesthetic depth and long term
mortality (1), I was waiting for him to ask the obvious question:
What about regional anesthesia? Although studies have arguably
failed to show a lack of short-term mortality benefit associated with
regional anesthesia compared with general anesthesia, what about
long term outcome 1 or 2 yr after surgery?
Further, if anesthetic depth has a negative effect on long term
patient outcome, then it is logical to conclude that any anesthetic
depth is deleterious to healthits only a matter of statistical
degree. And if that is the case, then the obligatory conclusion is that
general anesthesia is deleterious to health and should be avoided if
a suitable alternative exists. Of course, everything is relative, and
regional anesthesia certainly has its own morbidity and mortality
rates (2).
As an anesthesiologist who is also an internist and critical care
physician, the priority in my own practice has long been not to
avoid intraoperative anesthetic complications but to maximize patients short-term and long-term functional status (and thereby
quality of life). This is particularly important in older patients,
especially the frail and those with organic brain disease and deteriorating mentation. As a result, regional anesthesia or more precisely, avoiding general anesthesia when possibleis the obvious
choice to me, for almost anyone.
Even so, a secondary issue is that there are commonly conflicts
and technical problems that require pushing the envelope when it
comes to using regional anesthesia. Two examples are, first, the
elderly frail patient who presents for colectomy: such an operation
is easily accomplished with regional anesthesia but often requires
some sedation that may impair the airway and/or expose the patient to the risk of aspiration. Too, such an anesthetic is often quite
a bit more laborious for the anesthetist than a simple general
anesthetic.
The second example is typified by a patient who presented to me
the other day for an urgent below-knee amputation: a frail elderly
man with mild dementia and significant lung disease that left him
with labored breathing. It was clear to me that general anesthesia
would very likely be a life-changing event for him, both from a
mental status and a respiratory standpoint. But he was also anticoagulated with a platelet inhibitor, so a spinal was seemingly contraindicated. What to do? A narrow gauge spinal, after a long
discussion of risk versus benefits with the family, did fine.
So ultimately, how far is it acceptable to push the envelope to
avoid general anesthesia? Clearly, we need more information and
more data to fully and responsibly analyze the risk-benefit equation
and be able to elucidate it to our patients and their families.
I would like to see the authors of some of the well known
comparison studies (35) go back and follow up their patients to see
if we could gain further information about long-term functional and
mortality outcomes, comparing those who had regional anesthesia
(alone) versus general anesthesia.
Leo I. Stemp, MD
Western Mass Critical Care
Mercy Medical Center
Springfield, MA
leos@cox.net

References
1. Cohen NH. Anesthetic depth is not (yet) a predictor of mortality! Anesth Analg
2005;100:13.
2. Lee LA. Complications associated with regional anesthesia. Review of ASA Closed
Claims. APSF Newsletter, Spring 2004:5 6.
3. Yeager MP, Glass DD, Neff RK, Brinck-Johnsen T. Epidural anesthesia and analgesia
in high-risk surgical patients. Anesthesiology 1987;66:729 36.

2005 by the International Anesthesia Research Society

4. Bode RH, Lewis KP, Zarich SW, et al. Cardiac outcome after peripheral vascular
surgery: comparison of general and regional anesthesia. Anesthesiology 1996;84:313.
5. OHara DA, Duff A, Berlin JA, et al. The effect of anesthetic technique on postoperative
outcomes in hip fracture repair. Anesthesiology 2000;92:94757.
DOI: 10.1213/01.ANE.0000180252.75160.22

In Response:
In his Letter to the Editor regarding the article by Monk et al. (1) and
my editorial (2), Dr. Stemp raises some additional important points
that warrant further comment. The study by Monk et al. was undertaken to assess the impact of general anesthesia on morbidity. It
did not include a more comprehensive evaluation of morbidity, nor
was it designed to compare outcomes associated with general anesthesia versus regional anesthesia. In the editorial I raised some
concerns about the study design and the legitimacy of the conclusion of the article but did not address some of the broader issues
raised by Dr. Stemp related to anesthetic technique and its impact
on long-term outcome.
Clearly, every anesthesiologist should be concerned about the
sequelae of our clinical decision-making on the patients intraoperative course, immediate postoperative care, and pain management but also on subsequent quality of life no matter what anesthetic technique is used. Over the past few years we have assessed
many of the early postoperative implications of our care, but we
have only recently begun to evaluate the long-term impact of our
intraoperative decision-making. The Monk articleand the letter
from Dr. Stemp emphasize that what we do in the operating room
has more significant and longer-term impact than many patients or
their providers have acknowledged.
Dr. Stemp has made the judgment that for the sickest patients, he
prefers to use regional anesthesia whenever possible; he suggests
that although there may be relative contraindications to regional
techniques, for many patients the risks associated with regional
anesthesia are less than the risks associated with general anesthesia.
Although in some situations, I agree with his conclusions, as a
fellow critical care provider, I do not agree that in all situations
avoiding elective airway control and mechanical ventilatory support is necessarily more risky than providing regional anesthesia
with light sedation. For many intensive care unit patients, we
provide general anesthesia (often without acknowledging that we
are doing so) to facilitate care, yet are able to wean the patients from
the support without causing serious long-term sequelae. On the
other hand, I agree with Dr. Stemp that avoiding instrumentation of
the airway or hemodynamic variability whenever possible is probably desirable.
The experience of Dr. Stemp and his thoughtful comments
emphasize two important points that have been only partially
addressed in the debate about the Monk article. First, we need a
great deal more data on the indications for and risks associated
with both regional and general anesthesia. Is Dr. Stemps approach appropriatepushing the envelope of regional anesthesia
because of the more serious complications associated with general anesthesia? We are aware of many of the short-term risks of
regional anesthesia but do not have a great deal of data on longterm outcomes.
Second, we need to be much more concerned about outcomes
associated with all aspects of our anesthetic management not only in
the early postoperative period but also as our decisions impact
wound healing, rapidity of rehabilitation, and quality of life. For
example, we need further studies to document the most appropriate
hemoglobin for the surgical patient, the impact of anesthesia on
immune status, and the impact of intensive glucose control in
patients undergoing procedures other than cardiac surgery.
Once again the manuscript by Monk et al. has initiated a very
valuable debate about the impact of our clinical decisions on patient
Anesth Analg 2005;101:155967

1559

1560

LETTERS TO THE EDITOR

outcome. Although we do not yet have straightforward answers to

the question about the relationship between anesthetic depth and
mortality, nor do we know which anesthetic technique is best under
most circumstances, this discussion has been valuable in stimulating further investigation into the long-term outcomes associated
with our clinical management decisions. If, after further investigation the cumulative deep hypnotic time proves not to have any
predictive value, the debate it has generated is still valuable to us
and our patients.
Neal H. Cohen, MD, MPH, MS
UCSF School of Medicine
University of California San Francisco
San Francisco, CA
cohenn@medsch.ucsf.edu

References
1. Monk TG, Saini V, Weldon BC, Sigl JC. Anesthetic management and one-year mortality
after noncardiac surgery. Anesth Analg 2005;100:4 10.
2. Cohen NH. Anesthetic depth is not (yet) a predictor of mortality! Anesth Analg
2005;100:13.

Capnogram Shape in Obstructive Lung

Disease

ANESTH ANALG
2005;101:1559 67

Finally, demarcation to locate angle b may not be possible in time

capnograms obtained in some patients with chronic obstructive
lung disease. These time capnograms are quite distinct from those
shown by the authors. There is prolongation of phase II that merges
almost in a linear fashion with up-sloping phase III. Under these
circumstances, however, a volume capnogram will have a distinct
demarcation between phases II and III.
Bhavani Shankar Kodali, MD
Department of Anesthesiology
Brigham and Womens Hospital
Harvard Medical School
Boston, MA
bhavani@capnography.com

References
1. Krauss B, Deykin A, Lam A, et al. Capnogram shape in obstructive lung disease.
Anesth Analg 2005;100:884 8.
2. Bhavani Shankar K, Philips JH. Defining segments and phases of a time capnogram.
Anesth Analg 2000;91:9737.
3. Kumar AY, Bhavani Shankar K, Moseley HS, Delph Y. Inspiratory valve malfunction
in a circle system: pitfalls in capnography. Can J Anaesth 1992;39:9979.
4. Kodali BS. www.capnography.com: an animated website. Anesth Analg. 2001;93:1364.
5. Carbon dioxide. In: Lumb AB, ed. Nunns respiratory physiology, 5th ed. London:
Butterworth Heinemann, 2004:222 48.
6. Bhavani Shankar K, Moseley H, Kumar Y, Delph Y. Capnometry and anesthesia. A
Review Article. Can J Anaesth 1992;39:61732.
7. Bhavani Shankar K, Kumar AY, Moseley H, Hallsworth R. Terminology and the
current limitations of time capnography: a brief review. J Clin Monit 1995;11:175 82.

To the Editor:

DOI: 10.1213/01.ANE.0000180367.29387.9F

The study by Krauss et al. (1) raises some important concerns that
must be considered in the interpretation of the results and
conclusion.
First, the underlying physiological principles of time capnography (carbon dioxide versus time) as employed by Krauss et al.
are old and have inherent fallacies that could lead to erroneous
results (2 4). There is no way the authors could determine inspiratory time and expiratory time accurately from time capnograms without simultaneously measuring and superimposing
respiratory flow rates over the capnograms (2). The side-stream
technology of carbon dioxide (CO2) measurement, as described
by the authors, is limited in this regard. The authors assumption
that the onset of the up-stroke and the onset of down-stroke
denote the beginning of expiration and inspiration, respectively,
is not always true. For the same reason, the use of the old
terminology of ABCDE is not physiologically correct and has
been replaced by phase 0 (inspiratory phase) and the three phases
of expiration: Phase I (dead space gases), Phase II (mixing of dead
space gases with alveolar gases), and Phase III (alveolar gases)
(27). The beginning of phase I (expiration) and the beginning of
phase 0 (inspiration) cannot be delineated in a time capnogram
without superimposing respiratory flow rates over time capnograms. Phase I actually begins before the up-stroke, as CO2
concentration in dead space gases is zero, assuming that there is
no rebreathing. Similarly, expiration may end before the actual
down-stroke on the capnogram, the remainder being expiratory
pause. Therefore, incorporation of the inspiratory and expiratory
times as determined by the authors into their algorithm for
analysis could result in methodological errors.
Second, volume capnography is a better technique for studying
the hypothesis in question because the slope of phase III (alveolar
plateau) in the volume capnogram is a better reflection of the
ventilation perfusion status of the lung than the corresponding
slope in a time capnogram (7). For example, the incidence of positive slope phase III, as seen in volume capnograms, is 100%, which
is not reflected in a time capnogram (the phase III appears flat). This
is a result of exponentially decreasing expiratory flow that is completed in the very proximal part of alveolar plateau (phase III). A
smaller volume of expired gases (approximately the final 15%)
occupies half of the time available for expiration, so that a similar
change in Pco2 is distributed over a greater length of time in the
time capnogram than in the volume capnogram (7). Moreover,
the expiration may be complete well before the down-stroke,
and the reminder of phase III could be expiratory pause (depending
on the respiratory rate) until the onset of the next inspiration.

In Response:
We appreciate Dr. Kodalis comments on some of the theoretical
advantages of volumetric capnography as they relate to capnogram
interpretation. It is unclear, as it has not been studied, how volumetric capnography would perform in relation to time-based capnography in discriminating between different disease states. However, regardless of the method of capnogram interpretation, our
analytic techniques clearly discriminated among obstructive, restrictive, and normal capnograms. Further, we were able to demonstrate that the magnitude of these differences increased with
disease severity and that the observed changes in capnogram shape
correlated to changes in spirometry. These discriminations of capnogram morphology are not dependent on locating the precise start
of inhalation or exhalation or on defining the precise intervals
associated with capnogram phases. Rather, they characterize the
overall curvature of the capnogram.
Baruch Krauss, MD, EdM
Aaron Deykin, MD
Alexander Lam, MD
Joan J. Ryoo, MD
Jyoti Mathad, MD
David R. Hampton, PhD
Paul W. Schmitt, PhD
Jay L. Falk, MD
Division of Emergency Medicine
Childrens Hospital and Harvard Medical School
Boston, MA
baruch.krauss@mac.com

Short-Lasting Effect of Perineural

Resiniferatoxin on Mechanical Hyperalgesia
To the Editor:
I have read with great interest the report by Kissin et al. (1).
Intrathecal administration of small-dose (100 ng) resiniferatoxin
does not influence responses to mechanical stimuli as evaluated by
von Frey hair stimulation (2). Responses to von Frey hair stimulation are considered as light touch response. The current article did
not disagree with the above finding but opined that greater concentrations of resiniferatoxin prevented responses to noxious mechanical stimuli. It is very interesting that the duration of the effect of
resiniferatoxin on noxious mechanosensitivity is much shorter
(48 h) than that of noxious heat sensitivity (3 wk).

ANESTH ANALG
2005;101:1559 67

It is difficult to understand why the authors did not use any

vehicle control in showing dose dependency of resiniferatoxin in
Figures 1 and 2. Because they dissolved resiniferatoxin in dimethyl
sulfoxide and Tween 80, it may be necessary to know if these
substances have any effect on animal behaviors. The first sentence in
the Results section was Resiniferatoxin-induced hyperalgesia was
dose dependent. However, the authors, in fact, discussed
resiniferatoxin-induced hypoalgesia but not hyperalgesia in the text
and in Figures 1 and 2, although this could be a printing error. In the
fourth paragraph of Results, the authors stated, Perineural resiniferatoxin prevented mechanical incision-induced hyperalgesia. The
correct sentence may be Perineural resiniferatoxin prevented
incision-induced mechanical hyperalgesia.
Ratan K. Banik, MBBS, PhD
Department of Anesthesia
University of Iowa Hospitals and Clinics
Iowa City, IA
ratan-banik@uiowa.edu

LETTERS TO THE EDITOR

1561

readily available in an operating room, easy to use, transparent and

visible, thin and light, inexpensive, and easy to discard after use.
Koh Mizutani,

MD

Department of Anesthesia
Fuchu Hospital
Izumi, Japan
k_mizutani@fuchu-hospital.co.jp

Yoshiro Toyoda,

MD

Department of Anesthesia
Osaka Kohseinenkin Hospital
Osaka, Japan

References
1. Sessler DI, McGuire J, Sessler AM. Perioperative thermal insulation. Anesthesiology
1991;74:8759.
2. Maglinger PE, Sessler DI, Lenhardt R. Cutaneous heat loss with three surgical drapes,
one impervious to moisture. Anesth Analg 2005;100:738 42.
DOI: 10.1213/01.ANE.0000180245.28215.99

References
1. Kissin I, Davison N, Bradley EL Jr. Perineural resiniferatoxin prevents hyperalgesia in
a rat model of postoperative pain. Anesth Analg 2005;100:774 80.
2. Karai L, Brown DC, Mannes AJ, et al. Deletion of vanilloid receptor 1-expressing
primary afferent neurons for pain control. J Clin Invest 2004;113:1344 52.
DOI: 10.1213/01.ANE.0000180247.66058.BE

Countries, Rooms, and Plants, but Not

Mammalian Cells

In Response:

To the Editor:

We thank Dr. Banik for noticing the typographical error. As far as a

possible effect of the vehicle is concerned, the control experiments in
this regard can be seen in Figures 3 and 4 of the paper (see the
Sal-Veh groups). There was no effect observed in these groups.
Igor Kissin, MD, PhD
Natasha Davison, BS
Edwin L. Bradley, Jr, PhD

In the article Perioperative Fluid Management and Clinical Outcomes in Adults (1), the authors make references to the cell wall.
Mammalian cells do not possess a cell wall. As such, I suspect the
authors should have stated cell membrane.
Stephen B. Corn, MD

Department of Anesthesiology, Perioperative and Pain Medicine

Brigham and Womens Hospital
Harvard Medical School
Boston, MA

Brigham and Womens Hospital

Harvard Medical School
Boston, MA
corn@zeus.bwh.harvard.edu

Reference
1. Grocott MPW, Mythen MG, Gan TJ. Perioperative fluid management and clinical
outcomes in adults. Anesth Analg 2005;100:1093 6.
DOI: 10.1213/01.ANE.0000180244.99261.40

Sterilization Pouches for Perioperative

Thermal Insulation

In Response:

To the Editor:

Department of Anesthesiology
Duke University Medical Center
Durham, NC
gan00001@mc.duke.edu

Any single layer of passive insulators reduces cutaneous heat loss

(1). Surgical drapes are commonly used for thermal insulation (2).
We wrap the head and untreated body areas in open sterilization
pouches in place of surgical drapes (Fig. 1). This novel method
sufficiently maintains core and peripheral temperature over the
long term. Among their many advantages, sterilization pouches are

We thank the author for identifying this common error. It is good to

know that the rest was perfect!
T. J. Gan, MD

Michael P. W. Grocott, BSc, MRCP,

Michael G. Mythen, MD, FRCA

FRCA

Centre for Anaesthesia

University College London
London, UK

Hyperkalemic Cardiac Arrest After

Cardiopulmonary Bypass in a Child with
Duchenne Muscular Dystrophy
To the Editor:

Figure 1. Thermal insulation using open sterilization pouches.

We read with interest the article by Nathan et al. (1) concerning a

hyperkalemic cardiac arrest in a child with unsuspected Duchenne
muscular dystrophy. As the authors state in their discussion, children with undiagnosed Duchenne muscular dystrophy are at an
increased risk when exposed to depolarizing muscle relaxants; however, the unexpected rhabdomyolysis associated with general anesthesia drugs is still a problem and no specific disease appears to be
related to this phenomenon (2 4). Because of this, we disagree with
the authors recommendation of routine preoperative screening of
all patients for elevated serum creatine kinase. Although an elevated serum creatine kinase is highly sensitive for myopathies, it

1562

LETTERS TO THE EDITOR

lacks specificity (4). Serum creatinine kinase may be elevated after

trauma and heavy exercise. Universal screening can result in frequent and unnecessary delays in the operating room schedule and
also lead to the Ulysses Syndrome of mental and physical disorders which follow the discovery of a false positive result (5).
Careful and thorough preanesthetic evaluation is essential in
helping to identify undiagnosed myopathies, but as with the patient
in this report and others (6), this method is not foolproof.
We agree with the authors that screening in the neonatal period
may be helpful but will need a formal cost analysis before this can
be accomplished.
Hetam Al-Takrouri, MD
Arkansas Childrens Hospital
University of Arkansas Medical Center
Little Rock, AR

James F. Mayhew,

MD

Batson Childrens Hospital

University of Mississippi Medical Center
Jackson, MS
jmayhew@anesthesia.umsmed.edu

References
1. Nathan A, Ganes A, Godinez RI, et al. Hyperkalemic cardiac arrest after cardiopulmonary bypass in a child with unsuspected Duchenne muscular dystrophy. Anesth
Analg 2005;100:672 4.
2. Medina KA, Mayhew JF. Generalized muscle rigidity and hypercarbia with halothane
and isoflurane. Anesth Analg 1998;86:297 8.
3. Maccario M, Fumagalli C, Dottori V, et al. The association between rhabdomyolysis
and acute renal failure in patients undergoing cardiopulmonary by pass. J Cardiovas
Surg 1996;37:1539.
4. Pedrozzi NE, Ranelii GP, Tomasetti R, et al. Rhabdomyolysis and anesthesia: a report
of two cases and review of the literature. Pediatr Neurol 1996;15:254 7.
5. Mercer R. The Ulysses syndrome. Can Med Assoc J 1972;106:1223.
6. Al-Takrouri H, Martin TW, Mayhew JF. Hyperkalemic cardiac arrest following succinylcholine administration: the use of extracorporeal membrane oxygenation in an
emergency situation. J Clin Anesth 2004;16:449 51.

ANESTH ANALG
2005;101:1559 67

per unit of output should be part of the bigger picture. The specialty
of anesthesiology is now involved in perioperative medicine, and
anesthesiologists participate more often in the preoperative and
postoperative care of patients. Costs per unit of quality-adjusted life
years should be analyzed to obtain the most accurate answer to
whether less expensive or more expensive costs are justified.
In the United States alone, the health care economy is more than
$1 trillion (2). Technology contributes to 40% of health care inflation,
yet that fact seems to be ignored. The current solution to the problem seems to be either decreasing reimbursement or denying health
care services. Schuster et al. (1) stated that the cost advantage of
spinal anesthesia over general anesthesia tends to decrease as the
duration of the procedure lengthens. That difference may be important. However, that cost needs to be evaluated in terms of the final
follow-up outcomes of patients. This approach is basic technology
assessment. Knowledge of whether a treatment, drug, or procedure
is less than $50,000 per quality-adjusted life years saved will allow
physicians to determine whether the best financial and medical
result is being obtained in the United States or elsewhere in the
world. My compliments to the authors. My hope is that all physicians do long-term follow-up studies to determine whether the
techniques we render to our patients make a true difference in
long-term outcome.
Eric L. Bloomfield, MD, MS, FCCM
Departments of Anesthesiology and Critical Care Medicine
Mayo Clinic
Jacksonville, FL
bloomfield.eric@mayo.edu

References
1. Schuster M, Gottschalk A, Berger J, Standl T. A retrospective comparison of costs for
regional and general anesthesia techniques. Anesth Analg 2005;100:786 94.
2. Bloomfield EL. The impact of economics on changing medical technology with reference to critical care medicine in the United States. Anesth Analg. 2003;96:418 25.

DOI: 10.1213/01.ANE.0000180242.35595.32

DOI: 10.1213/01.ANE.0000180368.65153.8E

In Response:

In Response:

We appreciate the comments of Drs. Al-Takrouri and Mayhew. We

agree with them that universal screening for elevated creatine
kinase during the preoperative evaluation is unnecessary. We had
been very careful in preparing our manuscript (1) and had only
recommended estimating creatine kinase in children with a history
of unexplained motor delay. We also agree with them on the need
for a cost analysis for neonatal screening.
Aruna Nathan, MD

We thank Dr. Bloomfield for his comments on our paper. Indeed,

the economic evaluation of a medical procedure is complex. Our
aim was to estimate cost differences between different anesthesia
techniques and analyze the cost drivers behind them. To calculate
the long-term economic impact of medical treatments and procedures is a very important, but very tricky, task. To calculate the
cost of quality-adjusted life years for specific anesthesia procedures, as Dr. Bloomfield suggests, might be very difficult given
the complex interaction of surgical and anesthesiological complications and the low morbidity and mortality achieved by modern
anesthesia science. We absolutely agree that we will need much
more solid data both on outcomes and on costs before we are able
to realize such a long-term assessment of anesthesia techniques.
Martin Schuster, MD, MA
Andre Gottschalk, MD
Thomas Standl, MD

Department of Anesthesia and Critical Care Medicine

The Childrens Hospital of Philadelphia

Arjunan Ganesh, MBBS

Rodolfo I. Godinez, MD,
Susan C. Nicolson, MD
William J. Greeley, MD

PhD

Department of Anesthesia and Critical Care Medicine

The Childrens Hospital of Philadelphia
Department of Anesthesia
University of Pennsylvania School of Medicine
Philadelphia, Pennsylvania
ganesha@email.chop.edu

Department of Anesthesiology
University Hospital Hamburg-Eppendorf
Hamburg, Germany
m.schuster@uke.uni-hamburg.de

Reference
1. Nathan A, Ganesh A, Godinez RI, et al. Hyperkalemic cardiac arrest after cardiopulmonary bypass in a child with unsuspected Duchenne muscular dystrophy. Anesth
Analg 2005;100:672 4.

Costs Are Not the Only Thing We Should

Be Concerned with in Anesthesia
To the Editor:
The study by Schuster et al. (1) comparing the costs of three different regional anesthesia techniques with general anesthesia according to duration of surgical procedure, although a retrospective
comparison, was an admirable evaluation.
Although cost does need to be considered, sights should be set on
a more opportunistic analysis. Costs are only one component. Cost

Supplementation of Intrathecal Bupivacaine

with Clonidine in Ex-Premature Neonates
To the Editor:
We read with interest the report of Rochette et al. (1) about
neonates 60 wk postconceptual age (PCA) with former prematurity in 50% of the cases, scheduled for inguinal herniotomy
who received spinal bupivacaine supplemented with various
doses of clonidine. Supplementation with 1 g/kg intrathecal
clonidine doubled the duration to full motor recovery as compared with a control group. This dose of clonidine was not
associated with significant hemodynamic or respiratory alterations in the early postoperative period.

ANESTH ANALG
2005;101:1559 67

We performed spinal anesthesia with 1 mg/kg bupivacaine and 1

g/kg clonidine in two ex-premature and anemic neonates undergoing bilateral inguinal herniotomy (postconceptual ages, 49 and 39
weeks). The duration of the block until full motor recovery was 120
and 90 min, respectively. Both neonates were sedated during surgery, but respiratory rate was maintained at 30 40 breaths/min
during surgery and in the postanesthesia care unit. One episode of
intraoperative bradycardia (85 bpm) occurred in one neonate and
was treated with 0.1 mg atropine. Postoperatively, 24-h apnea monitoring did not reveal any respiratory depression.
Extending the duration of a single-shot spinal anesthesia with
intrathecal clonidine is attractive because it avoids the use of supplemental anesthesia in high-risk neonates. However, despite the
safe use of this technique in Rochettes study and in our patients,
monitoring of apnea postoperatively is required because concerns
have been raised regarding the potential for respiratory depression
after caudal clonidine in neonates (2 4). A 24-h investigation on a
large number of patients should be conducted to confirm the absence of long-term effects of spinal clonidine in newborns.
Marie T. Aouad, MD
Roula E. Hajj, MD
Department of Anesthesiology
American University of Beirut-Medical Center
Beirut, Lebanon
mm01@aub.edu.lb

References
1. Rochette A, Raux O, Troncin R et al. Clonidine prolongs spinal anesthesia in newborns:
a prospective dose-ranging study. Anesth Analg 2004;98:56 9.
2. Breschan C, Krumpholz R, Likar R et al. Can a dose of 2 g kg1 caudal clonidine
cause respiratory depression in neonates? Paediatr Anaesth 1999;9:813.
3. Bouchut JC, Dubois R, Godart J. Clonidine in preterm-infant caudal anesthesia may be
responsible for postoperative apnea. Reg Anesth Pain Med 2001;26:835.
4. Fellmann C, Gerber AC, Weiss M. Apnoea in a former preterm infant after caudal
bupivacaine with clonidine for inguinal herniorrhaphy. Paediatr Anaesth 2002;12:637 40.
DOI: 10.1213/01.ANE.0000180238.73522.01

In Response:
We thank Drs. Aouad and Hajj for their interest in our work (1), as
well as for their comments. A further confirmation of the efficacy of
our approach is welcome. Since this work was initiated, we have
performed about 500 clonidine-supplemented spinal anesthesia in
neonates and this series highlights our positive opinion. However,
we fully agree with Aouad and Hajjs recommendation to monitor
closely these patients for 24 h, because i) spinal anesthesia reduces
but does not eliminate the risk of postoperative apnea (2) and ii)
clonidine may affect postoperative respiratory course, as advocated
by the three case reports our colleagues mentioned. We have specifically addressed this question in a 24-h descriptive study to be
published on more than 120 infants (3). As the results appear
encouraging, we have designed a prospective comparative study to
clarify benefits and limits of this approach.
Alain Rochette, MD
Christophe Dadure, MD
Xavier Capdevila, MD, PhD
Department of Anesthesiology and Critical Care Medicine
University Hospital Lapeyronie
Montpellier, France
a-rochette@chu-montpellier.fr

References
1. Rochette A, Raux O, Troncin R, et al. Clonidine prolongs spinal anesthesia in newborns: a prospective dose-ranging study. Anesth Analg 2004;98:56 9.
2. Shenkman Z, Hoppenstein D, Litmanowitz I, et al. Spinal anesthesia in 62 premature,
former-premature or young infants: technical aspects and pitfalls. Can J Anesth 2002;
49:2629.
3. Rochette A, Troncin R, Raux O, et al. Clonidine added to bupivacaine in neonatal
spinal anesthesia: a prospective comparison in 124 preterm and term infants. Ped
Anesth 2005; in press.

LETTERS TO THE EDITOR

1563

Aerosolized Lidocaine for Relief of

Extubation Laryngospasm
To the Editor:
Topical and/or IV lidocaine has been previously reported as effective to control laryngospasm (1). The following case report shows
that an aerosolized lidocaine can also successfully control partial
tracheal extubation laryngospasm.
A 28-yr-old woman, primigravida, was admitted for cesarean
delivery. A rapid sequence induction was performed using propofol
and succinylcholine, followed by cisatracurium. Trachea was intubated using a 7.0-mm (inner diameter) cuffed tube. After surgery,
neostigmine 0.05 mg/kg and atropine 0.02 mg/kg were injected IV
to reverse the residual muscular blockade. The trachea was extubated when the patient was fully awake. After extubation, the
patient developed partial laryngospasm and Spo2 decreased to 86%.
A mixture of 5 mL of 2% lidocaine without epinephrine with equivolume of 0.9% normal saline was nebulized by 100% oxygen. A few
minutes later, the patient showed a complete relief of laryngospasm
and oxygen saturation increased to 100%.
In obstetric anesthesia anatomical changes associated with pregnancy, as well as nasal congestion and pharyngeal edema, may
exacerbate the risk of hypoxemia after extubation laryngospasm (2).
Previous reports have shown that topical lidocaine can abolish
laryngeal chemoreflex and mechanoreflex (3). Our report shows
that an aerosolized lidocaine may also depress the laryngeal reflexes
and, hence, can be used to control partial extubation laryngospasm.
The technique may not be advisable in patients with full stomach.
Ahed Zeidan, MD
Dania Halabi, MD
Department of Anesthesiology
Sahel General Hospital
Beirut, Lebanon
doczeidan@hotmail.com

Anis Baraka,

MD, FRCA

Department of Anesthesiology
American University of Beirut Medical Center
Beirut, Lebanon

References
1. Koc C, Kocaman F, Aygenc E, et al. The use of preoperative lidocaine to prevent stridor
and laryngospasm after tonsillectomy and adenoidectomy. Otolaryngol Head Neck
Surg 1998;118:880 2.
2. Pilkington S, Carli F, Dakin MJ, et al. Increase in Mallampati score during pregnancy.
Br J Anaesth 1995;74:638 42.
3. McCulloch TM, Flint PW, Richardson MA, Bishop MJ. Lidocaine effects on the laryngeal chemoreflex, mechanoreflex, and afferent electrical stimulation reflex. Ann Otol
Rhinol Laryngol. 1992;101:5839.
DOI: 10.1213/01.ANE.0000180236.27026.9F

Airway Fires: Gas-Bugs Providing the Fuel?

To the Editor:
We read with interest the description of an airway fire and commend Singla et al. (1) for a succinct review of these adverse events.
The described components of fire (2)ignition sources, oxidizing
agents, and fuelsare well known. Reports reviewed suggest an
increased incidence of airway fires during surgery on diseased
lungs (3 6). These diseased lungs usually harbor a mixed flora.
Notable among these are the aerobes Staphylococci and Pseudomonas,
whereas Bacteroids and Peptostreptococci are the major anaerobes (7).
We hypothesize the role of gas-forming microorganisms in airway fires. Anaerobic organisms are known to play a major role in
pleuro-pulmonary infections. They are involved in infections such
as pneumonia, aspiration pneumonia, lung abscess, and empyema
(8). In one study (9) the isolation rate of anaerobic bacteria was as
high as 93.3% in patients with lung abscess and up to 22.7% with
nosocomial pneumonia. Some of these anaerobes are avid flammable gas producers (e.g., methane) (10). Anaerobic organisms produce methane from short chain organic acids by the reduction of

1564

LETTERS TO THE EDITOR

carbon dioxide and acetate by hydrogen. This methanogenesis (10)

has been demonstrated in vitro and may occur in anaerobic infections of the lung where underlying tissue destruction is occurring at
the hands of anaerobic organisms.
During surgery on a lung that has such infective foci, the electrocautery tip may come in contact with methane or other flammable
gases produced by the anaerobes. In the presence of high concentrations of oxygen (increased Fio2) or increased pressures
(continuous positive airway pressure or positive end-expiratory
pressure), the situation may truly become explosive! We speculate that the amount of flammable gas produced is small and
therefore the effects are limited. Nonetheless, these may become
significant if another flammable substance (a dry gauze or polyvinyl chloride tracheal tube) is present, and serious fires may
result. These fires are less frequent in surgery on other sites of
documented anaerobic infections, probably because of the absence of high concentrations of oxygen seen in the airway. Our
hypothesis, though anecdotal, is supported by the frequent observation of sparks when the large bowel is opened with electrocautery. The cause has been attributed to flammable gastrointestinal gases (11), again produced by anaerobes.
This hypothesis may therefore explain the increased reports of
these fires in surgeries (tracheotomy or tracheoesophageal fistulae
repair) on diseased lungs. We advise caution during surgery on the
airways when the underlying lung is diseased and may have such
gas-forming microorganisms. We believe that simple monitoring
devices that detect traces of inflammable gases (such as those used
in industry) may provide conclusive evidence of the role of methane
and other flammable gases in airway fires.
Naveen Eipe, MD
Department of Anaesthesia
neipe@yahoo.com

Ashish Choudhrie,

MS

Department of Surgery
Padhar Hospital
Padhar, India

Dr. Campagna does not wish to respond.

References
1. Singla AK, Campagna JA, Wright CD, Sandberg WS. Surgical field fire during a repair
of bronchoesophageal fistula. Anesth Analg 2005;100:1062 4.
2. A clinicians guide to surgical fires: how they occur, how to prevent them, how to put
them out. Health Devices 2003;32:524.
3. Wu CC, Shen CH, Ho WM. Endotracheal tube fire induced by electrocautery during
tracheostomy: a case report. Acta Anaesthesiol Sin 2002;40:209 13.
4. Rogers ML, Nickalls RW, Brackenbury ET, Salama FD, et al. Airway fire during
tracheostomy: prevention strategies for surgeons and anesthetists. Ann R Coll Surg
Engl 2001;83:376 80.
5. Lai A, Ng KP. Fire during thoracic surgery. Anaesth Intensive Care 2001;29:3013.
6. Baur DA, Butler RC. Electrocautery-ignited endotracheal tube fire: case report.
Br J Oral Maxillofac Surg 1999;37:1423.
7. Kedzia A, Kwapisz E, Wierzbowska M. Incidence of anaerobic bacteria in respiratory
tract infections [in Polish]. Pneumonol Alergol Pol 2003;71:68 73.
8. Verma P. Laboratory diagnosis of anaerobic pleuropulmonary infections. Semin Respir
Infect 2000;15:114 8.
9. Konishi M, Mori K, Yoshimoto E, Takahashi K, et al. Clinical evaluation of anaerobic
infections in patients with bronchopulmonary infections diagnosed by transtracheal
aspiration [in Japanese]. Kansenshogaku Zasshi 1999;73:675 80.
10. Bal AS, Dhagat NN. Upflow anaerobic sludge blanket reactor: a review. Indian J
Environ Health 2001;43:1 82.
11. Macdonald AG. A brief historical review of non-anaesthetic causes of fires and explosions in the operating room. Br J Anaesth 1994;73:84756.
DOI: 10.1213/01.ANE.0000180235.87151.E6

Anesthesiologists Must Inform Their

Surgical Colleagues When There Is a Risk
of an Operating Room Fire
To the Editor:
I read with interest the report by Singla et al. (1) on the occurrence
of a fire during the surgical repair of a bronchoesophageal fistula.
The authors reported that the fire occurred when cautery was

ANESTH ANALG
2005;101:1559 67

employed in an atmosphere rich in oxygen while slightly bloody

surgical sponges were present.
The authors stated, This event draws attention to the need for all
members of the patient care team in an operating room to understand
the contribution of their actions to the development of a surgical fire.
I agree. However, they do not explicitly state that the anesthesia care
team must inform surgical and nursing personnel about the presence
of a high concentration of oxygen in the surgical field and the need to
employ only moistened gauze sponges to minimize the risk of a fire
when using cautery. Obviously, anesthesiologists are the ones most
likely to be aware of the presence of a high concentration of oxygen
because they are administering it. Moreover, they must be cognizant of
the risk factors for the occurrence of a fire and be proactive in informing other member of the operating room team about these dangers to
minimize the risks to the patient.
Mitchel B. Sosis, MD, PhD
Department of Anesthesiology
Holy Redeemer Hospital and Medical Center
Meadowbrook, PA
mitchelsosis@hotmail.com

Dr. Campagna does not wish to respond.

Reference
1. Singla AK, Campagna JA, Wright CD, et al. Surgical field fire during repair of a
bronchoesophageal fistula. Anesth Analg 2005;100:1062 4.
DOI: 10.1213/01.ANE.0000180234.78115.2E

The Case for the Role of Advanced

Simulators in Trauma Management
Training Was Not Made
To the Editor:
Despite the claims by Barsuk et al. (1) their study does not add
usefully to the current data on simulation in medical training. Their
methodology only allowed them to show the expected decline in
knowledge over time after training and that teaching of deficient
knowledge before testing will result in an improvement. Benefit for
simulation could have been tested had they used four groups. The
pre-intervention group would remain the same, highlighting the
current gaps in knowledge. The intervention groups would consist
of one group of fresh Advanced Trauma Life Support graduates,
one group taught the deficiencies in an Advanced Trauma Life
Support style with no simulator, and one group taught the same
information with a simulator. The beneficial effects of advanced
simulation techniques could then be shown only if the fourth group
was significantly better than the other two intervention groups.
Retesting at 1 yr would allow examination of any long-term benefits
of advanced simulation training.
The article referred to as demonstrating construct validity for the
use of simulators in trauma training (2) is a pilot study examining
the use of a simulator as an alternative assessment tool. Similarly,
the value of advanced simulation in promoting trauma management is not shown by Marshall et al. (3) Their article uses a simulator for pre- and post-Advanced Trauma Life Support testing of
candidates and as such does not use the simulator as a training tool.
Perhaps the only tentative conclusion would be that the use of a
simulator increases, rightly or wrongly, subjects confidence in their
own abilities in a simulated trauma scenario.
The pre-intervention group appears to have had one third of the
simulator familiarization time allocated to the intervention group
before testing. It is likely that doctors used to performing in a
situation involving a simulator will perform better when tested, so
for valid comparison it is necessary to ensure that all examined
groups have similar previous experience with medical simulators.
With changes to the way doctors are trained and improvements in
the fidelity with which simulators recreate the real clinical situation,
it is likely that they will become an increasingly important learning
tool. With the inherent high costs (3) we require more research so

ANESTH ANALG
2005;101:1559 67

that their role in training may be defined, optimized, and the

financial outlay justified.
Andrew Paterson, MBChB, BSc
Department of Anaesthesia
Cheriton House
James Cook University Hospital
Middlesbrough, UK
a.d.paterson@doctors.org.uk

References
1. Barsuk D, Ziv A, Lin G, et al. Using advanced simulation for recognition and correction
of gaps in airway and breathing management skills in prehospital trauma care. Anesth
Analg 2005;100:8039.
2. Gordon JA, Tancredi D, Binder W, et al. Assessing global performance in emergency
medicine using high fidelity patient simulator: a pilot study. Acad Emerg Med 2003;
10:472.
3. Marshal RL, Smith JS, Gorman P, et al. Use of a human patient simulator in the
development of resident trauma management skills. J Trauma 2001;51:1721.

LETTERS TO THE EDITOR

without midline shift. The subdural hematoma was selfcontained, without increased intrathecal pressure and was not
accompanied by somnolence, vomiting, or focal neurologic abnormalities. Therefore, it did not require surgical evacuation. The
patient recovered uneventfully and was discharged home 6 days
later.
We suspect that a decrease in intrathecal pressure from an excessive leak of cerebrospinal fluid caused stress on the cerebral vessels
of this 46-yr-old patient with a history of cigarette smoking and
subsequently precipitated the intracranial bleed.
Shaul Cohen, MD
Alex Shorshtein, MD
Patrick Blanchfield, MD

DOI: 10.1213/01.ANE.0000180233.58171.68

Department of Anesthesia
UMDNJ-RWJMS
New Brunswick, NJ
cohensh@umdnj.edu

In Response:

DOI: 10.1213/01.ANE.0000180261.70703.5A

We appreciate Dr. Patersons comments and gratefully accept the

opportunity to reply. We believe our study demonstrated the benefits of using Advanced Simulation to assess deficiencies in the
management of trauma patients, as a tool to facilitate changes to a
Trauma Management training curriculum, followed by an improvement in performance. To the best of our knowledge, previous studies, using Advanced Simulation modalities, have only recommended curriculum changes but have not implemented them.
In addition, our study demonstrated the benefits of using a relatively
simple skill-based simulator (TraumaMan, Laerdal, Stavanger, Norway) to advance skill transference to an Advanced Simulation Modality (SimMan, Laerdal, Stavanger, Norway). These findings are not
trivial in view of the limited data published demonstrating the beneficial effects of simulation-based training.
We acknowledged, in our manuscript, the design limitation mentioned by Dr. Paterson. We believe our findings are the first step to
a larger longitudinal study in which we would be delighted to
cooperate with any other interested Simulation Center.
Hain Berkenstadt, MD
Daphna Barsuk, MD
Department of General Surgery C
The Israel Center for Medical Simulation
Sheba Medical Center
Tel Hashomer, Israel
berken@netvision.net.il

Subdural Hematoma Following Accidental

Dural Puncture
To the Editor:
We would like to report a case of subdural hematoma that
occurred after accidental dural puncture. A 46-yr-old (88 kg, 168
cm) primipara at 38 wk intrauterine pregnancy was scheduled for
elective cesarean section for breech presentation. The patient had
a history of cigarette smoking one pack per day for 20 yr. A
loss-of-resistance to air technique was attempted in sitting position at L3-4 interspace. A sudden excessive leak of cerebrospinal
fluid was noticed upon advancing the epidural needle 6.5 cm
from the skin. A decision was made to insert the epidural catheter 3 cm intrathecally and use it for spinal anesthesia. The glass
syringe followed by the stylet could not stop the constant leak of
cerebrospinal fluid, and an excessive amount of cerebrospinal
fluid leaked until the catheter was inserted. A total of 10 mg of
hyperbaric bupivacaine was injected via the catheter and provided a satisfactory bilateral T3-S5 spinal block for an uneventful
cesarean section. Two hours from the accidental dural puncture,
in the postanesthesia care unit, the patient complained of severe
non-positional headache located in the left forehead. The epidural catheter was removed and patient underwent a computed
tomographic scan, which revealed a left subdural hematoma

1565

Thoracotomy in a Patient with a History of

Local Anesthetic Allergy
To the Editor:
A 42-yr-old woman presented for thoracotomy and left upper lobectomy for lung carcinoma. She had a shellfish allergy (nonspecific
pruritus) and an alleged allergy to Xylocaine but no family history
of local anesthetic-induced adverse reactions. Two months earlier,
she developed local erythema, pruritus, and wheal immediately
after subcutaneous injection of Xylocaine, presumably for insertion
of an IV cannula before bronchoscopy. There was no generalized
rash, wheezing, or hypotension. This same reaction occurred after a
second injection performed a few minutes later at a different site.
She was labeled as allergic to Xylocaine and bronchoscopy was
abandoned. She subsequently had a computed tomographic-guided
aspiration biopsy of her left upper lobe nodule and a chest tube
placed for post-procedural pneumothorax although she could not
recall if any local anesthetic was used for these procedures. All this
had occurred in a hospital out of the country.
After discussion with the patient and surgeon regarding the
benefits of thoracic epidural anesthesia for thoracotomy, the likelihood of a true local anesthetic allergy and the risks associated with
a local anesthetic challenge, a decision was made to preoperatively
challenge her with ropivacaine. In the operating room with full
resuscitation facilities available, an IV line was established and
standard monitors were placed. Subcutaneous injections of
preservative-free 0.2% ropivacaine were administered beginning
with 0.5 mL and followed by 1.0-mL and 2.0- mL doses. The patient
was observed for 15 min after each injection for clinical signs of
allergic reaction. None were observed and an epidural catheter was
placed at T4-5 level with IV fentanyl 150 g and midazolam 3 mg.
Intraoperative analgesia was provided with a 6 mL bolus of ropivacaine 0.2% fentanyl 2 g/mL followed by an infusion of the
same mixture at 6 mL/h. Perioperative anesthetic care and surgery
proceeded uneventfully and the patient received excellent postoperative analgesia.
A true immunologic reaction to local anesthetic is rare. In most
instances local anesthetic allergy can be excluded from history and
the safety of local anesthetic verified by progressive challenge (13).
The exact preparation of the Xylocaine used previously could not be
ascertained and therefore her reaction could have been a result of
the preservative, antioxidant, or metabolites. Skin tests and subcutaneous challenge have been reported to be safe (4). The testing was
limited to one drug, ropivacaine without a preservative. Epidural
opioids alone could have been utilized for perioperative analgesia,
but the full benefits of epidural analgesia in terms of perioperative
respiratory complications and dynamic pain relief are better
achieved with a local anesthetic plus opioid combination.

1566

LETTERS TO THE EDITOR

ANESTH ANALG
2005;101:1559 67

The safety and utility of provocative challenge testing has been

previously reported. However, consultation with an allergy specialist for formal testing should be considered if the patient had a
previous anaphylactic reaction to consider other potential responsible agents.
Ju-Mei Ng, FANZCA
Department of Anaesthesia & Surgical Intensive Care
Singapore General Hospital
Singapore
gannjm@sgh.com.sg

References
1. Fisher MM, Bowey CJ. Alleged allergy to local anaesthetics. Anaesth Intensive Care
1997;25:611 4.
2. Schatz M. Skin testing and incremental challenge in the evaluation of adverse reactions
to local anesthetics. J Allergy Clin Immunol 1984;74:606 16.
3. Eggleston ST, Lush LW. Understanding allergic reactions to local anesthetics. Ann
Pharmacother 1996;30:8517.
4. Troise C, Voltolini S, Minale P, et al. Management of patients at risk for adverse
reactions to local anesthetics: analysis of 386 cases. J Investig Allergol Clin Immunol
1998;8:1725.
DOI: 10.1213/01.ANE.0000180260.93604.7B

Figure 2.

ECG Artifact and BIS in Severe Brain Injury

To the Editor:
Myles and Cairo (1) reported an unexpected increase in bispectral
index values in a patient with ischemic brain injury, which was
confirmed by cerebral angiography and attributed to electrocardiogram artifact. Gaszynski (2) refuted these findings and attributed
them to insular activity of the brain. We observed similar findings
to those of Myles and Cairo in a patient with a ruptured intracranial
aneurysm in whom all the tests of brain death were positive. We
have been using bispectral index as an adjuvant to diagnose brain
death routinely in our intensive care unit. In the present case, we
observed bispectral index values fluctuating between 0 and 50 with
suppression ratio varying from 100% to 25% without any significant
electromyographic activity on electromyographic strength indicator
(Figs. 13). A careful analysis of the bispectral index monitor revealed that it was intermittently selecting electrocardiogram waveform and was falsely reading it as electroencephalogram (Fig. 4).
Normally electrocardiogram and electromyographic artifacts are
filtered by the bispectral index monitor, but in the present case it
was actually selecting electrocardiogram tracings and they were
corresponding to patients actual electrocardiogram tracings. Therefore, we recommend that electroencephalogram tracings should be

Figure 1.

Figure 3.

Figure 4.

ANESTH ANALG
2005;101:1559 67

LETTERS TO THE EDITOR

1567

Figure 1. Bispectral index record in 35-yr-old male potential organ donor. Brain death was caused by hypoxia during cardiac arrest and
delayed cardiopulmonary resuscitation at the accident scene.
carefully observed and the numerical values displayed on bispectral
index monitor should not be relied on exclusively.
G. D. Puri, MD, PhD
D. Nakra, MD
Department of Anaesthesia and Intensive Care
Post Graduate Institute of Medical and Research
Chandigarh, India
gdpuri007@hotmail.com

Dr. Myles does not wish to respond.

References
1. Myles PS, Cario S. Artifact in the Bispectral Index in a patient with severe ischemic
brain injury. Anesth Analg 2004;98:706 7.
2. Gaszynski T. BIS in brain injury. Anesth Analg 2005;100:293 4.
DOI: 10.1213/01.ANE.0000180375.64387.27

In Response:
Puri described a possible explanation for bispectral index values
increasing in a patient with brain death. However, in some cases of

potential organ donors, changes in electrocardiogram cannot be

connected with bispectral index level increase and therefore other
possibilities should be considered. Brain death is defined as brainstem death, which does not mean death of all structures of the
central nervous system (new modified definition). The example of
another difficult case is presented in Figure 1. This record was not
accompanied by significant changes in electrocardiogram, heart
work, or other possible artifacts.
Tomasz Gaszynski, MD, PhD
Department of Anaesthesiology and Intensive Therapy
Medical University of Lodz
Barlicki University Hospital
Lodz, Poland
tomgaszyn@poczta.onet.pl

SECTION EDITOR
NORIG ELLISON

Book and Multimedia Reviews

On-Line Electronic Help for Anaesthesiologists.
Perel A, ed. Brussels, Belgium: European Society of
Anaesthesiologists, 2004. Available through Philips Medical
Systems.
On-Line Electronic Help for Anaesthesiologists (OLEH) is a compact
disc that installs a digital handbook of basic information relevant to
anesthesiologists on most PCs that run a current Windows operating system. OLEH consists of many HyperText Markup Language
(HTML) files linked together to implement an electronic book that
must be viewed using a Web browser, Microsoft Internet Explorer
5.5 or higher. The stated intent of this product, developed by a
peer-review process of the European Society of Anaesthesiologists,
is to provide concise, authoritative, searchable, and interlinked
point of care information to anesthesiologists in the operating
room (OR). For those who work with the Philips IntelliVue family of
patient monitors, the professional version of the OLEH program,
available from Philips Medical Systems, appears in a window on the
OR monitor that can be opened while keeping most of the monitors
values for vital signs still visible on the display.
Installation on a PC is quick and automatic. The home page is
nicely designed with navigation and help buttons on the right and
major topic categories (Drugs, Fluids and Electrolytes, Regional
Anesthesia, Difficult Airway, etc.) on the left. There is a general
index and a drug list index, and most related topics are linked
together in an easily understandable way, i.e., clicking on dantrolene in the drug index takes you directly to the summary of
treatment for malignant hyperthermia in the topic list. Most descriptions are concise and correct but many are self-evident ( e.g., for
treatment of hypoxia, administer oxygen 100% immediately) for
anyone with even basic anesthesia skills and experience. All of
OLEHs hyperlinks are simply shortcuts to the other included
HTML files and they do not link to Web sites even if an internet
connection is present.
This product is the obvious replacement for the well-worn paper
pocket handbooks carried by so many anesthesiologists in training.
Nevertheless, it shares their deficiencies which include unavoidable
oversimplification, limited scope, and, with time, outdated information. OLEH is also less likely than a paperback book to be instantly
available in a clinical emergency because it may not be in a handy
back pocket, although installation on a Philips monitor, if those are
used for OR patient monitoring, will minimize this limitation. Nevertheless, OLEH will disappoint anyone expecting a quantum step
forward in clinically valuable information systems, which this reviewer believes should be capable of full-color graphic and video
representations and voice recognition and synthesis, as well as
maintaining and refreshing a continuing connection to the Internet
to insure that all information is as up-to-date as possible.
OLEH may be helpful for anesthesiologists who have a few quiet
minutes with their laptops in the OR (if theyre not Apple users!) to
browse topics of interest. This reviewer certainly would not object to
having OLEH installed and available on his OR patient monitors
(which happen to be compatible Philips products), but he does not
believe it will fulfill the editors assertion that OLEH can improve
the anaesthesiologists performance in all the situations they meet.
Only education and experience, not just text-based information, can
accomplish that task.
Stanley Muravchick,

MD, PhD

Professor of Anesthesia
University of Pennsylvania School of Medicine
Philadelphia, PA

1568

Anesth Analg 2005;101:156870

Fundamental Applications of Transesophageal

Echocardiography: A Society of Cardiovascular
Anesthesiologists Monograph on DVD.
Reeves ST, Shanewise JS, eds. Baltimore, MD: Lippincott Williams
& Wilkins, 2005. ISBN 1-58255-761-6. $99.00.
Each year, the Society of Cardiovascular Anesthesiologists produces
a multi-authored monograph focusing on a specific area of interest
to their membership. This year the monograph has been produced
for the first time as a DVD. The topic, transesophageal echocardiography, is well suited to this new format and the editors have
obviously labored to take advantage of the many unique aspects of
a computer-based publication.
The monograph is compatible with both Windows (2000, XP
Professional) and Macintosh (OS X) operating systems with a Pentium II 550 or faster processor or a 550 MHz or faster processor,
respectively. Both systems require at least 128 MB RAM and 1 MB
hard disk space, as well as a DVD-ROM drive. For this commentary,
the monograph was reviewed on appropriately configured Macintosh and Windows systems.
The editors are Drs. Scott Reeves and Jack Shanewise, both of
whom are well-known educators in the field of cardiac anesthesiology. The monograph is composed of 17 chapters and there are 34
contributors representing the fields of cardiac anesthesiology, cardiothoracic surgery, and cardiology. As the contributors represent
17 different institutions, a broad sampling of current clinical approaches is presented. There are some inconsistencies in style from
chapter to chapter, as well as some repetition of basic material, but
overall the editors have produced a cohesive and easily readable
package of information.
Once the DVD is loaded into the computer, the table of contents
can be accessed, and the user can select one of the 17 chapters. Each
chapter is self-contained; it is not necessary to review each chapter
in sequence. The chapters begin with a text presentation, and relevant figures, tables, and video loops are then hyperlinked to the
text. Once a figure is reviewed, return to the text is rapid.
Most chapters begin with a brief review of the topic, emphasizing published data regarding the natural history of the condition under discussion, the results of various treatment modalities, and then follow with a discussion of the role of TEE in
diagnosis and surgical decision-making. Some chapters cover
broad topics such as Indications and Outcomes or Hemodynamics. However, in keeping with the editors stated goals,
focus on specific disease states encountered in clinical practice or
tackle specific controversial issues in echocardiography (e.g.,
mechanisms and treatment of ischemic mitral regurgitation,
moderate aortic stenosis in patients undergoing coronary artery
bypass). Many of the chapters make effective use of case presentations to illustrate specific clinical problems. While much of the
monograph is devoted to discussion of problems commonly encountered by most cardiac anesthesiologists, there is also coverage of less common problems, such as congenital heart disease in
adults, ventricular assist devices, and employment of TEE during
treatment of atrial or ventricular septal defects with transvenous
occlusive devices.
As with previous SCA monographs, this work is not meant to be
a comprehensive step-wise discussion of the topic selected. Thus, it
would not be an appropriate choice for the novice wishing an
introduction to clinical echocardiography. However, any clinician
with at least a working knowledge of echocardiography will find
much worthwhile information in this multidisciplinary text.
2005 by the International Anesthesia Research Society

ANESTH ANALG
2005;101:1568 70

BOOK AND MULTIMEDIA REVIEWS

Most of the shortcomings of the monograph are technical in

nature. Many of the video clips are slow to load (up to 2 minutes
on the Windows system). Exacerbating this problem is the fact
that text and graphics windows cannot be viewed simultaneously, leaving the reader staring at a blank screen waiting for
the video. Allowing readers to continue reading the text while
loading video images in a separate window would be a significant improvement.
The text formatting is extremely basic and no distinction (font,
size, color) is made between major and minor headings. In addition,
the reader must flip from one page to the next rather than scroll
as is common in software applications today. The net effect of these
two drawbacks is a disjointed reading experience, and it is occasionally difficult to follow the narrative flow from one section of text
to the next within a given chapter.

1569

Finally, the resolution of the tables is poor overall, and many with
small print are close to being illegible. This is a peculiar problem for
a work produced as a DVD, where memory (and thus image resolution) was presumably in abundance.
Despite these minor flaws, this years monograph is a worthwhile addition to the library of all echocardiographers who practice in the perioperative environment. It deserves to be read,
discussed, and debated and will hopefully serve as an example
for future publications on perioperative echocardiography.
James E. Duckett, MD
Gordon H. Morewood, MD, FASE
Department of Anesthesiology
Bryn Mawr Hospital
Bryn Mawr, PA
ducks6@aol.com

Erratum
In the review of Australias First Anesthesthetic Department: 75 Years at the RPA (Anesth Analg 2005;101:
930 1), the philanthropist whose financial support for basic research and education created the first
academic anesthesia department in Australia and for whom the Chair was named was incorrectly identified. Lord Nuffield provided this financial support. The surgeon who was largely responsible for obtaining
Lord Nuffields support was John Lowenthal.