Rev Esp Anestesiol Reanim.

2014;61(2):64---72

Revista Española de Anestesiología

y Reanimación
www.elsevier.es/redar

ORIGINAL ARTICLE

Cerebral oxygenation in patients undergoing shoulder surgery in

beach chair position: Comparing general to regional anesthesia
and the impact on neurobehavioral outcome
J. Aguirre a,∗ , A. Borgeat a , T. Trachsel a , I. Cobo del Prado b , J. De Andrés b,c , P. Bühler a

a
Division of Anesthesiology, Balgrist University Hospital, Zurich, Switzerland
b
Anesthesia, Critical Care and Multidisciplinary Pain Management Department, Valencia University General Hospital, Valencia,
Spain
c
Anesthesia Division, Surgical Specialties Department, Valencia School of Medicine, Spain

Received 20 May 2013; accepted 5 August 2013

Available online 9 October 2013

KEYWORDS Abstract
Cerebral oxygenation; Background: Ischemic brain damage has been reported in healthy patients after beach chair
Near-infrared position for surgery due to cerebral hypoperfusion. Near-infrared spectroscopy has been
spectroscopy; described as a non-invasive, continuous method to monitor cerebral oxygen saturation. How-
Beach chair position; ever, its impact on neurobehavioral outcome comparing different anesthesia regimens has been
Shoulder surgery; poorly described.
Cerebral desaturation Methods: In this prospective, assessor-blinded study, 90 patients undergoing shoulder surgery in
event; beach chair position following general (G-group, n = 45) or regional anesthesia (R-group; n = 45)
Neurobehavioral were enrolled to assess the prevalence of cerebral desaturation events comparing anesthesia
outcome regimens and their impact on neurobehavioral and neurological outcome. Anesthesiologists
were blinded to regional cerebral oxygen saturation values. Baseline data assessed the day
before surgery included neurological and neurobehavioral tests, which were repeated the day
after surgery. The baseline data for regional cerebral oxygen saturation/bispectral index and
invasive blood pressure both at heart and auditory meatus levels were taken prior to anesthesia,
5 min after induction of anesthesia, 5 min after beach chair positioning, after skin incision and
thereafter all 20 min until discharge.
Results: Patients in the R-group showed significantly less cerebral desaturation events
(p < 0.001), drops in regional cerebral oxygen saturation values (p < 0.001), significantly better
neurobehavioral test results the day after surgery (p < 0.001) and showed a greater hemody-
namic stability in the beach chair position compared to patients in the G-group.

∗ Corresponding author.
E-mail address: jose.aguirre@balgrist.ch (J. Aguirre).

0034-9356/$ – see front matter © 2013 Sociedad Española de Anestesiología, Reanimación y Terapéutica del Dolor. Published by Elsevier España, S.L. All rights reserved.
http://dx.doi.org/10.1016/j.redar.2013.08.002
Comparing general to regional anesthesia and the impact on neurobehavioral outcome 65

Conclusions: The incidence of regional cerebral oxygen desaturations seems to influence the
neurobehavioral outcome. Regional anesthesia offers more stable cardiovascular conditions for
shoulder surgery in beach chair position influencing neurobehavioral test results at 24 h.
© 2013 Sociedad Española de Anestesiología, Reanimación y Terapéutica del Dolor. Published
by Elsevier España, S.L. All rights reserved.

PALABRAS CLAVE Oxigenación cerebral en pacientes operados del hombro en posición sentada:
Oxigenación cerebral; comparación de anestesia general y regional e impacto en la respuesta
Espectroscopia de neuroconductual
infrarrojo cercano;
Resumen
Posición sentada
Antecedentes: Se han registrado lesiones isquémicas cerebrales en pacientes sanos ocasion-
(«silla de playa»);
adas por hipoperfusión cerebral después de una intervención quirúrgica en posición sentada
Cirugía de hombro;
(«silla de playa»). La espectroscopia de infrarrojo cercano se ha descrito como un método no
Episodio de
invasivo continuo para supervisar la saturación de oxígeno en el cerebro. No obstante, apenas
desaturación
se ha descrito su impacto en el resultado neuroconductual que compara los distintos tipos de
cerebral;
anestesia.
Respuesta
Métodos: En este estudio prospectivo con enmascaramiento doble se reclutaron 90 pacientes
neuroconductual
que habían sido operados del hombro en posición sentada con anestesia general (grupo G,
n = 45) o regional (grupo R, n = 45), para evaluar la incidencia de episodios de desaturación en
función de la anestesia suministrada y el impacto que esto suponía en su respuesta neurológica
y neuroconductual. Se ocultaron a los anestesistas los valores de saturación regional de oxígeno
cerebral. En la recopilación de los datos de referencia evaluados el día antes de la operación
se incluían pruebas neurológicas y neuroconductuales, que se repitieron el día después de la
misma. Se tomaron valores de referencia de saturación regional de oxígeno cerebral/índice
biespectral y presión invasiva a nivel del corazón y el meato auditivo antes de la anestesia,
5 min antes de la anestesia, 5 min después de la colocación en posición sentada, después de la
incisión y consecutivamente cada 20 min hasta el final.
Resultados: Los pacientes en el grupo R manifestaron muchos menos episodios de desatu-
ración cerebral (p < 0,001), bajadas de los valores de saturación regional de oxígeno cerebral
(p < 0,001), resultados mucho mejores en las pruebas neuroconductuales del día posterior a la
operación (p < 0,001) y presentaron una mayor estabilidad hemodinámica en la posición sentada,
en comparación con los pacientes del grupo G.
Conclusiones: La incidencia de desaturaciones regionales de oxígeno cerebral parece influir en
la respuesta neuroconductual. La anestesia regional ofrece mayor estabilidad cardiovascular
en las cirugías del hombro en posición sentada, de manera que influye en los resultados de las
pruebas neuroconductuales después de 24 h.
© 2013 Sociedad Española de Anestesiología, Reanimación y Terapéutica del Dolor. Publicado
por Elsevier España, S.L. Todos los derechos reservados.

Background The near-infrared spectroscopy (NIRS) is a non-invasive

Beach chair position is common for arthroscopic and open
shoulder surgery facilitating shoulder joint access and opti-
mal visualization due to the anatomic upright position of
the shoulder.1 In patients with normal preoperative regional
cerebral oxygen saturation (rScO2 ) values, changes have
been shown during the beach chair position (BCP) follow-
ing general anesthesia.2,3 Rare but catastrophic case reports
have been described in this position like visual loss, nerve
injury, cerebral infarction and death.2 The etiology of these
complications remains unclear, but a correlation with the
upright position and consecutive cerebral hypoperfusion
and cerebral ischemia has been hypothesized.1 Addition-
ally, arthroscopic shoulder surgery is often performed using
controlled hypotension protocols to minimize blood loss and
possibly leading to further reduction of cerebral blood flow.3
Therefore, a more cautious permissive hypotension manage-
ment has been suggested for the beach chair position.1,6,7
and continuous method to measure rScO2 . The technique
has demonstrated to accurately recognize cerebral hypo-
perfusion leading to cerebral desaturation events (CDEs)
that would remain undetected with the conventional mon-
itoring with possible unfavorable neurological outcome.4
Recently, studies have shown that NIRS can reliably indi-
cate cerebral hypoperfusion during shoulder surgery in
BCP.3,5---11
According to literature review and database search-
ing, no previous study using NIRS for beach chair position
has compared the effects of different anesthesia regimens
assessing their impact on the number of CDEs and on
neurobehavioral outcome using a wide spread controlled
hypotension protocol.12 Moreover, we also focused on the
effects of the anesthesia regimen on hemodynamics, on the
effects of intraoperative CDEs on post anesthesia care unit
(PACU) bypassing/stay duration and on the influence of car-
diovascular risk factors on CDEs.
66
Methods
After approval by the local ethical committee (Kantonale
Ethikkommission Zürich) and written informed consent,
we included 90 ASA I-III patients scheduled for elective,
monolateral shoulder surgery in BCP in our prospec-
tive, assessor-blinded study. Exclusion criteria were age
<18 years, history of central neurological diagnosis (tran-
sient ischemic attack, stroke, bleeding, syncope, chronic
headache, cervical disk herniation, spinal cord injury,
recent vision impairment/loss, cerebral tumor or metasta-
sis, orthostatic hypotension), recent myocardial infarction
(<6 months), known relevant carotid stenosis (>40%) or
known flow disturbance of vertebral arteries, pregnancy,
allergies to any drug used for anesthesia and known neu-
robehavioral disorders. The primary endpoint of our study
was the incidence of CDEs (defined as rScO2 decrease ≥20%
from baseline or absolute value <55% for >15 s)2 compar-
ing standardized anesthesia regimens for shoulder surgery
in BCP with controlled hypotension. Secondary endpoints
were the effects of the anesthesia regimen on rScO2 , hemo-
dynamics (invasive blood pressure at heart and acustic
meatus level) and on neurological and neurobehavioral out-
comes, the effects of intraoperative CDEs on post anesthesia
care unit (PACU) bypassing/stay duration and the influence
of cardiovascular risk factors on CDEs. Group allocation
was according to our clinical standard: general anesthe-
sia (G-group) was performed whenever surgeons wanted to
perform an immediate postoperative neurological test of
the operated arm (shoulder prothesis, latisimus dorsi mus-
cle transfer, complicated osteosynthesis) or when a good
muscle relaxation was wished for surgery (Latarjet opera-
tion) and regional anesthesia (R-group) in all other cases
(rotator cuff repair, acromio-clavicular resection, diagnostic
arthroscopies, simple osteosynthesis, etc.). These standards
include that patient with impaired cardio-vascular or pul-
monary situation considered high risk for anesthesia are
treated with regional anesthesia independently of surgery
for more stable conditions.
The evening prior to surgery, a standardized neurolog-
ical control (pupil size and reaction, lateralization tests
of both extremities, Glasgow Coma Scale GCS, Minimental
Test) and neurobehavioral tests (Trail Making A, Trail Making
B) were conducted as baseline measurements by an anes-
thesiologist not involved in the anesthesia management of
the patient. Trail Making A/B are neuropsychological tests
focusing on visual attention and task switching. They con-
sist of two parts in which the patient is instructed to connect
a set of 25 dots (numbers in A, alternating numbers and
letters in B) as fast as possible. They provide information
about visual search speed, speed of processing, scanning,
mental flexibility, as well as executive functioning. Both
tests were suggested by Murkin et al. in a consensus to
assess neurobehavioral outcomes after cardiac surgery.13
The tests were performed by the patients in the pre- and
postoperative sessions using the non-operated arm. Cere-
bral oximetry values are affected by several factors as
type of anesthetic used, depth of anesthesia, FiO2 , PaCO2
and blood pressure management.14 Therefore, anesthetic
management was standardized in both groups, and the anes-
thesiologist and the anesthesia nurse in charge of each
patient were blinded for the rScO2 value during the whole
J. Aguirre et al.
procedure by covering the NIRS monitor and switching off
its alarms. Patients were premedicated using 0.1 mg/kg oral
midazolam one hour before the arrival in the induction
room. There, standard monitoring was installed (oxygen sat-
uration, ECG, non-invasive blood pressure measurement and
placement of an i.v. line on the non-operated arm). Addi-
tionally, all patients got a bispectral index (BIS) electrode on
the forehead for comparison with rScO2 value changes and a
radial artery line for invasive blood pressure measurement
(IBP) on the non-operated arm. Further, two NIRS sensors
were applied both sides on the forehead just above the
BIS electrode (INVOS® Near-infrared spectroscopy monitor-
ing, Somametics, Covidien, MA, USA). Baseline oxygenation
parameters were set 5 min after initial monitoring instal-
lation. Arterial oxygen saturation (SpO2 ), and rScO2 values
were measured in patients while breathing a 40% air/oxygen
mixture. Blood pressure was invasively recorded at heart
and acustic meatus level. Measurements were recorded at
baseline, 5 min after induction of general/regional anesthe-
sia, 5 min after beach chair positioning, at surgery start
and then all 20 min until discharge to PACU or ward and
at periods of undesirable events (deoxygenation, hypoven-
tilation states). BIS and rScO2 values were continuously
recorded by the monitor system for later evaluation. Inter-
scalene catheter using neurostimulation was placed as
described elsewhere using 40 ml ropivacaine 0.5% in the R-
group.17 The catheter was also placed in patients of the
G-group to avoid difference in the neurobehavioral tests due
to pain medication but it was not activated in the G-group
until discharge to the PACU to avoid additional hemody-
namic alterations to general anesthesia. General anesthesia
was induced with a propofol target-controlled infusion (TCI)
using 100---200 ␮g of fentanyl and rocuronium for intubation
and propofol/remifentanil TCI for maintenance keeping the
loss of consciousness (LOC) effect site concentration at its
1.5 - double value.15,16 The inspired oxygen fraction was kept
at 40% and the end-tidal CO2 between 4.6 and 5.3 kPa. The
R-group was sedated using a propofol and remifentanil TCI
guided for a clinically assessed conscious sedation using the
Observers’s Assessment of Alertness/Sedation Scale (OAA/S
1---5, for 1: deep sleep and 5: alert).17 All interventions were
performed in BCP with 65◦ from horizontal. We used a con-
trolled hypotension protocol for a systolic blood pressure at
heart level (SAPheart ) ≤ 100 mmHg. The intervention proto-
col for a SAPheart > 100 mmHg for more than 3 min included:
(1) deepening of sedation with propofol/remifentanil
according to clinical findings, (2) clonidine 75---150 ␮g i.v.,
(3) labetalol 10 mg bolus i.v. The intervention protocol
for a SAPheart < 80 mmHg for more than 3 min included:
(1) crystalloid bolus of 250 ml, (2) ephedrine i.v. 2.5---5 mg,
(3) norepinephrine i.v. 5---10 ␮g. Body temperature was
kept > 36.5 ◦ C using a Bair Hugger Temperature Management
System for air-warming and measured all 10 min at tympanic
level. After surgery and activation of regional anesthesia in
the G-group after extubation, patients of both groups were
eligible for PACU bypass if they reached an Aldrette score
≥8 with a VAS <30. These were also the criteria for dis-
charge from PACU to the ward.18 Postoperative analgesia
was performed with a patient controlled regional analge-
sia regimen with ropivacaine 0.3% for 48 h after surgery and
standard analgesia using acetaminophen and metamizol to
avoid central acting agents.
Comparing general to regional anesthesia and the impact on neurobehavioral outcome
Sample size calculation was based on previous data show-
ing that 50---80% of patients with general anesthesia in
BCP demonstrated a CDE episode of >20% decrease from
rScO2 baseline.2,5,12 We hypothesized according to literature
and unpublished previous results in a pilot study that we
would observe 50% fewer CDEs with a SD of 15% in patients
with regional anesthesia. Group sample size of 40 patients
achieves 90% power with a 2-tailed significance level (˛) of
0.01 using 2 or Fisher exact test with continuity correc-
tion. To compensate for dropouts, 45 patients per group
were included in this study. Discrete data were compared
using Fisher exact test. Ordinal data and continuous data
that were not normally distributed are presented as median
and range. These data were compared between groups using
the Mann---Whitney U test and within groups using Wilcoxon
signed rank test. Normally distributed continuous data were
tested with the Shapiro---Wilk test and presented as mean
and ±SD. The rScO2 values at different time points were
compared using an analysis of variance for repeated mea-
surements (ANOVA) including Bonferroni correction. Possible
correlations of CDEs with patient risk factors were analyzed
using Pearson chi-square and Fisher’s exact tests.
A p < 0.05 was considered to be significant. Statistical
analysis was performed using computer SigmaStat Version
19 (SPSS Science, Chicago, IL).
Results
No patient had to be excluded from the study. Patient
and surgical characteristics were are summarized in
Tables 1 and 2. In the G-group more cristalloids and fen-
tanyl were given and more patients had to be treated due to
SAPheart <80 mmHg (p = 0.011) whereas in the R-group more
patients had to be treated due to SAPheart >100 mmHg. The
mean clonidine dose used for SAPheart >100 mmHg treatment
was 100 ␮g and the mean ephedrine dose used for SAPheart
<80 mmHg treatment was 10 mg. At baseline, bilateral rScO2
values were equivalent between the groups with differences
<5% between left and right within the group. There was a
significant difference 5 min after beach chair position with
an increase of rScO2 values in the R-group (p < 0.05) com-
pared to baseline and a decrease of rScO2 values in the
G-group compared to baseline (p < 0.05) without side differ-
ence. There was a significant increase in rScO2 values 5 min
after regional and general anesthesia induction compared to
the respective baselines with a clear increase in the G-group
(p < 0.001). The rScO2 values in the R-group remained more
stable compared to baseline in comparison to the values of
the G-group from the time point of beach chair position-
ing until discharge to the PACU (p < 0.05). There was no
difference in paCO2 between both groups at the different
measured time points (p in all > 0.6). The incidence of CDEs
in the G-group was 71.1% with 51.1% showing a rScO2 value
decrease of ≥20% from baseline and 20% showing an absolute
rScO2 value <55% for more than 15 s. Six patients had a com-
bination of both definitions. The duration of the episodes
ranged from 35 s to 10 min. In the R-group there were 2.2%
of CDEs with one case showing an absolute rScO2 value <55%
for 30 s. The difference of these events between the groups
was significant (Table 2). Patients with CDEs in the G-group
showed worse results in the neurobehavioral tests compared
67
p < 0.05 vs baseline
100.0
∗ ∗
90.0
80.0
mmHG
Map heart RA
Map heart GA
70.0
Map AM RA
Map AM GA
60.0
50.0
m line
m .RA
BC
20 t
40 in
60 n
80 n
10 in
12 in
14 in
in
U
u
C
m
m
m
m
C
a.
PA
se
0
0
0
Ba
in
in
5
5
Figure 1 Mean arterial pressures (presented as mean values)
are shown at different measurement point levels (heart level
and acustic meatus level) for general anesthesia and regional
anesthesia. The * indicate the statistical significant difference
between the groups from anesthesia induction until surgery
start (p < 0.01). The dashed line indicates the times where MAP
was significant lower compared to the respective baselines in
both groups (p < 0.05). BC: beach chair; GA: general anesthesia;
MAP: mean arterial pressure; PACU: post anesthesia care unit;
RA: regional p < 0.05 vs. baseline anesthesia.
to patients without CDEs (p < 0.001) (Table 3) but the neu-
rological test were not influenced in this subgroup: Glasgow
Coma Scale scores of 15 and Minimental State Examination
scores of 30 in both subgroups and in the R-group at baseline
and after 24 h. The neurobehavioral test results of the only
patient in the R-group with a CDE were not worse compared
to patients without CDEs of both groups (p = 0.93). Accord-
ing to the continuous registration of BIS and rScO2 values
there was no correlation between BIS value and CDEs in both
groups.
The association of CDEs with patient risk factors is shown
in Table 3. There were weak correlations with coronary
artery disease (0.367, p = 0.013) and arterial hypertension
(0.414, p = 0.006) and the odds ratios were 2.0 and 0.12,
respectively. Hemodynamic data are presented in Table 2
showing a more accentuated decrease of MAP at heart
and auditory meatus level after beach chair positioning
compared to baseline in the G-group (p = 0.001). In both
groups the decrease in MAP was significant compared to
the respective baselines until discharge (p < 0.05) and sig-
nificant between the groups after anesthesia induction until
surgery (p < 0.01) (Fig. 1). Pulse rate was not different com-
pared to baseline at 5 min after beach chair position and
between the groups during surgery. No differences between
the groups or compared to the respective baselines were
noted in SpO2 levels throughout surgery. There were no
neurological symptoms excluding the blocked operated arm
in neither groups compared to baseline. After removal of
the interscalene catheter there were no related neurolog-
ical problems in any of the 90 patients. The results of the
neurobehavioral tests on the first postoperative day were
significantly worse compared to baseline only in the G-group
(p < 0.001). The R-group showed no significant alteration of
the results 24 h after surgery compared to baseline. The
BIS values did not correlate with the rScO2 values and the
subgroup with CDEs did not show any BIS alterations during
these episodes compared to the rest of patients in the same
group. In the R-group 44 patients fulfilled criteria to bypass
68 J. Aguirre et al.

Table 1 Patient characteristics.

Variables G-group (n = 45) 95% CI R-group (n = 45) 95% CI P
Sex (m/f) 24 (53.3%)/21(46.7%) --- 27 (60%)/18 (40%) --- 0.523
Age (y) 57.6 ± 14.5 53.3/61.9 48.8 ± 10.5 45.6/51.9 0.005
Weight (kg) 80.0 ± 9.8 76.0/81.9 78.8 ±11.2 75.4/82.2 0.91
Height (cm) 173.67 ± 7.1 171.5/175.8 173.0 ± 7.8 170.6/175.4 0.821
BMI (kg/m2 ) 26.1 ± 1.9 25.5/26.7 26.4 ±3.6 25.3/27.4 0.741
ASA Physical status 2 (1---3) 1.6/2.0 2 (1---3) 1.7/2.1 0.504
Preoperative hemoglobine (g/dl) 12.9 ± 1.1 12.5/13.2 13.2 ± 1.4 12.8/13.6 0.188
Blood transfusion 5 (11.1%) --- 0 --- 0.021
Coronary artery disease 7 (15.6%) --- 7 (15.6%) --- 1
Arterial hypertension 15 (33.3%) --- 14 (31.1%) --- 0.822
Diabetes mellitus 7 (15.6%) --- 7 (15.6%) --- 1
Smoker 23 (51.1%) --- 21 (46.7%) --- 0.673
COPD 7 (15.6%) --- 6 (13.3%) --- 0.764
Data expressed as number (%) or mean ± standard deviation. 95% confidence interval (CI) calculated where appropriated. G/R-group:
general/regional anesthesia group; COPD: chronic obstructive pulmonary disease.

the PACU and were discharged to the ward. In the G-group
only 2 patients fulfilled PACU bypass criteria (p < 0.001). The
group of patients of the G-group without CDEs met discharge
criteria a mean of 45 min earlier compared to those with
CDEs events (<0.05). There was no incidence of PONV in the
R-group whereas 15% of patients in the G-group had to be
treated for PONV (p = 0.003).
Discussion
Central neurological complications following shoulder
surgery in beach chair position are uncommon7 but poten-
tially devastating.4,7 Interestingly, all reported catastrophic
events were under general anesthesia. The exact mecha-
nism remains debatable but there seems to be a correlation
to a relative decrease in cerebral perfusion leading to a
decrease of oxygenation in the brain and spinal cord. Associ-
ated to patient-specific cardiovascular risk factors this drop
in blood pressure could lead to a decrease in cerebral per-
fusion with clinical consequences.19 The autoregulation of
cerebral perfusion is impaired in hypertensive patients, its
reactivity to CO2 impaired with age and additionally the
normal regulatory mechanisms to maintain cerebral per-
fusion are often blunted or absent due to vasodilatatory
effects of inhalational agents.19 The controlled hypotension
with systolic blood pressures below 100 mmHg is standard
practice and is mostly monitored non-invasively at heart
level. This practise could be an additional risk factor1 as
brachial non-invasive blood pressure monitoring specially in
controlled hypotension protocols seems to overestimate the
cerebral perfusion pressure due to the hydrostatic gradient
estimated at 0.77 mmHg/1 cm of brain elevation according
to the waterfall theory of perfusion.3 These findings are in
accordance with our results showing a clear overestimation
of MAP measured at heart level compared to acustic meatus
level. Patients in general anesthesia clearly show the most
relevant drop in blood pressure at the acustic meatus level
5 min after beach chair positioning. This drop is in accor-
dance with current literature describing a drop in cerebral
blood flow in anesthetized patients in head-up position.19
The influence of positioning under general anesthesia
using sevoflurane on rScO2 has been investigated in different
shoulder surgery studies.2,3,5,9,11,13,14 The common results
were a posture-dependent drop in blood pressure and in
rScO2 in BCP, which was not found in the lateral decubitus
position.2 Tange et al. did not see any impairment in rScO2 at
all in BCP, probably due to the fact, that MAP at heart level
was maintained between 75 and 85 mmHg using the alpha-
adrenergic receptor agonist phenylephrine.8,19 Yadeau et al.
studied the effect of regional anesthesia with propofol seda-
tion on cerebral oxygenation on 99 patients in BCP for
shoulder surgery using controlled hypotension measured at
heart level.7 They found a frequent incidence of hypotension
(99%) but CDEs were uncommon (10%) suggesting a posi-
tive effect of regional anesthesia on cerebral oxygenation.
However, we have to consider that the controlled hypoten-
sive protocol was not standardized an was used according
to an ad hoc decision. Recently, Koh et al. compared gen-
eral anesthesia using sevoflurane to regional anesthesia with
propofol sedation for 60 patients in BCP but without con-
trolled hypotension and with an implemented treatment
protocol for CDEs.10 They could show a CDE rate of 56.7%
in the asleep group compared to 0% in the regional anes-
thesia group. The results of the regional anesthesia group
are compatible with our findings. Contrary to common def-
inition, Koh et al. did not consider a total rScO2 value
<55% to be a CDE.2 However, they registered this event
not only in 23.3% of the asleep patients but also in 3.3%
of the regional anesthesia patients. Furthermore, the treat-
ment protocol probably masked further CDEs due to induced
hemodynamic improvement. Due to different anesthesia and
monitoring regimens, blood pressure management, inter-
vention protocols and different primary outcomes, previous
studies cannot be compared. Moreover, despite the blood
pressure impairment with rScO2 decrease no impact on clin-
ical outcome was measured in these studies. Recently, two
studies used inappropriate tests for neurobehavioral out-
come in small size study groups and found no correlation
after general anesthesia.5,9 These results would suggest
that impaired rScO2 values remain without clinical conse-
quences as different studies have shown CDEs of 50% - 80%
 Comparing general to regional anesthesia and the impact on neurobehavioral outcome
Table 2 Surgical/perioperative characteristics.
Variables G-group (n = 45) 95% CI R-group (n = 45) 95% CI P
Surgery 7 Rotator cuff --- 30 Rotator cuff --- ---
repair repair
20 shoulder 6 Arthroscopies
prothesis 4 Shoulder
10 Latarjet prothesis
5 Latissimus 5
dorsi transfer Osteosynthesis
3
Osteosynthesis
Surgery time (min) 84.7 ± 32.7 74.9/94.5 72.6 ± 30.5 63.4/81.7 0.119
Cristalloid infusion 1005.6 ± 341.5 902/1108 771.8 ± 363.5 662/881 <0.001
(ml)
Colloid infusion (ml) 538.9 ± 106.0 507.7/570.7 516.7 ± 63.1 497.7/535.6 0.284
PONV 8 (17.8%) --- 0 --- 0.003
Treatment for SAP he 15 (33.3%) --- 5 (11.1%) --- 0.011
<80 mmHg
Treatment for SAP he 4 (8.9%) --- 31 (68.9%) --- <0.001
>100 mmHg
CDEs with rScO2 drops 23 (51.1%) --- 0 (0%) --- <0.001
≥20% of BL
CDEs with absolute 9 (20.0%) --- 1 (2.2%) --- 0.007
rScO2 value < 55%
for >15 s
MAP he BL/5 min 91.47 ± 11.6/67.67 88.0/94.9/63.9/71.5 97.07 ± 12.1/88.47 ± 12.5 93.4/100.7/84.7/92.2 <0.001
after BC (mmHg) ±12.7
MAP am BL/5 min 91.47 ± 11.6/50.93 ± 13.9 88.0/94.9/46.7/55.1 97.07 ± 12.1/66.53 ± 14.4 93.4/100.7/62.2/70.9 0.001
after BC (mmHg)
rScO2 left BL/left 66.02 ± 3.6/64.04 ± 5.8 64.9/67.1/62.3/65.8 67.02 ± 7.1/68.33 ± 7.1 64.9/69.2/66.2/70.5 0.001
5 min after BC (%)
rScO2 right BL/right 66.82 ± 4.0/64.24 ± 6.5 65.6/68.0/62.2/66.2 66.91 ± 7.5/68.18 ± 8.2 64.67/69.2/65.7/70.7 <0.001
5 min after BC (%)
TMT A BL/24 h 30.51 ± 2.0/46.09 29.9/31.1/44.2/47.9 31.24 ± 1.2/32.09 ± 1.2 30.9/31.6/31.7/32.5 <0.001
after surgery (s) ±6.2
TMT B BL/24 h 77.31 ± 3.4/95.69 ± 5.6 76.3/78.4/94.0/97.4 79.0 ± 3.6/81.0 ± 3.3 78.0/80.1/80.0/82.0 <0.001
after surgery (s)
Data expressed as number (%) or mean ± standard deviation. 95% confidence interval (CI) calculated where appropriated. Am: acustical meatus; BC: beach chair; BL: baseline; CDE:
cerebral desaturation event; G/R-group: general/regional anesthesia group; he: heart level; MAP: mean arterial pressure; PONV: postoperative nausea and vomiting; rScO2 : regional
cerebral oxygen saturation; SAP: systolic arterial pressure; TMT A/B: trail making test A/B.

69
 70
Table 3 Comparison of patient and treatment variables between non-desaturation and desaturation patient groups.
Variables NON CDE (n = 19) 95% CI CDE (n = 26) 95% CI P
Sex (m/f) 10 (52.6%)/9 (47.4%) --- 14 (53.8%)/12 (46.2%) --- 0.936
Age (year) 56.74 ± 17.8 (48.1/65.3) --- 58.31 ±11.8 (53.6/63.1) --- 0.645
Weight (kg) 77.9 ± 7.5 74.3/81.6 79.7 ± 11.3 75.1/84.3 0.662
Height (cm) 173.2 ± 7.1 169.8/176.6 174.0 ± 7.2 171.1/176.9 0.764
BMI (kg/m2 ) 26.0 ± 1.6 25.9/26.7 26.2 ± 2.1 25.4/27.1 0.637
ASA physical Status 2 (1---3) 1.4/1.96 2 (1---3) 1.8/2.28 0.054
Preoperative hemoglobine (g/dl) 13.1 ± 1.1 12.6/13.6 12.7 ± 1.1 12.3/13.2 0.249
Blood transfusion 2 (10.5%) --- 3 (11.5%) --- 0.915
Coronary artery disease 0 (0%) --- 7 (26.9%) --- 0.014
Arterial hypertension 2 (10.5%) --- 13 (50%) --- 0.006
Diabetes mellitus 3 (15.8%) --- 4 (15.4%) --- 0.97
Smoker 7 (36.8%) --- 16 (61.5%) --- 0.102
COPD 6 (31.6%) --- 1 (3.8%) --- 0.011
Surgery time (min) 87.63 ± 29.5 73.4/101.9 82.50 ± 35.2 68.3/96.7 0.327
Treatment for SAP he <80 mmHg 8 (42.1%) --- 7 (26.9%) --- 0.286
Treatment for SAP he >100 mmHg 1 (5.3%) --- 3 (11.5%) --- 0.465
CDEs with rScO2 drops ≥20% of BL 0 (0%) --- 26 (100%) --- <0.001
CDEs with rScO2 value < 55% for >15 s 0 (0%) --- 9 (34.6%) --- 0.004
MAP he BL/5 min after BC (mmHg) 90.84 ± 11.9/72.63 ± 15.5 85.1/96.6/65.2/80.1 91.92 ± 11.5/64.04 ±8.9 87.3/96.6/ 60.4/67.6 0.049
MAP am BL/5 min after BC (mmHg) 90.84 ± 11.9/56.32 ± 17.9 85.1/96.6/47.7/64.9 91.92 ± 11.5/47.00 ± 8.6 87.3/96.6/ 43.5/50.5 0.031
rScO2 left BL/left 5 min after BC (%) 66.11 ± 4.1/67.84 ± 6.2 64.1/68.1/64.9/67.3 65.96 ± 3.3/61.27 ± 3.6 64.6/67.3/ 59.8/62.7 <0.001
rScO2 right BL/right 5 min after BC (%) 67.42 ± 3.8/68.05 ± 7.3 65.6/69.3/64.6/71.6 66.38 ± 4.1/61.46 ±4.2 64.7/68.1/59.8/63.2 0.003
TMT A BL/24 h after surgery (s) 30.00 ± 2.8/41.68 ± 5.8 26.7/31/3/38.8/44.4 30.88 ± 1.1/49.31 ± 4.3 30.4/31.4/ 46.6/51.0 <0.001
TMT B BL/24 h after surgery (s) 77.58 ± 4.4/91.89 ± 3.1 75.5/79.7/90.4/93.4 77.12 ± 2.7/98.46 ± 5.3 76.0/78.2/96.5/100.6 <0.001
Data expressed as number (%) or mean ± standard deviation. 95% confidence interval (CI) calculated where appropriated. Am: acustical meatus; BC: beach chair; BL: baseline; CDE:
cerebral desaturation event; COPD: chronic obstructive pulmonary disease; he: heart level; MAP: mean arterial pressure; rScO2 : regional cerebral oxygen saturation; SAP: systolic arterial
pressure; TMT A/B: trail making test A/B.

J. Aguirre et al.
Comparing general to regional anesthesia and the impact on neurobehavioral outcome
which is in accordance with our results for the G-group.2,5,12
Recently, Salazar et al. focused on the neurobehavioral out-
come after general anesthesia with sevoflurane and nitrous
oxide on 50 patients in BCP using a brief cognitive test
battery (RBANS) measured prior to anesthesia (baseline),
prior to discharge home after surgery and after 3 days.12
Unfortunately, no controlled hypotension protocol was used,
blood pressure measurement was done only at heart level,
only a rScO2 decrease of >20% from baseline was consid-
ered to be a CDE and an intervention protocol to treat
CDEs was implemented. Despite the observed 18% CDEs they
could not find any impairment of the neurocognitive out-
come comparing the results after 3 days with baseline.
Unfortunately, the results after surgery were excluded from
analysis apparently due to interference with anesthesia and
pain medication. These results differ from our findings. We
showed a clear impairment of the neurocognitive tests per-
formed the day after surgery in the G-group. Moreover,
patients with CDEs showed an increased impairment of test
results compared to those without CDEs. This suggests, that
general anesthesia in BCP with controlled hypotension has
an impact on the frequency of CDE which is in accordance
to other publications.2,10,12 Moreover, our results suggest
that CDE influence the neurobehavioral outcome at least of
the first 24 h after surgery.13 Additionally, we could show in
accordance with other publications that the routinely used
neurologic tests to assess negative impacts of BCP or general
anesthesia including Glasgow Coma Scale and Minimental
State Examination are not sensitive enough.5,9
Using a common controlled hypotension protocol1 keep-
ing systolic blood pressure ≤100 mmHg lead to a drop in
rScO2 form 66.2 ± 4.1 to 64.6 ± 6.6 in the G-group. Also Lee
et al.6 described a significant drop in rScO2 values during BCP
after sevoflurane/remifentanil anesthesia despite an MAP of
60---65 mmHg at the auditory meatus level. Probably, their
combination of anesthesia gases negatively influenced the
cerebral perfusion regulation.19 The significant drop in MAP
under general anesthesia seems to be correlated to beach
chair position as showed by different sutdies.9,12 However,
Murphy et al.5 showed that in beach chair position and in lat-
eral decubitus position under sevoflurane general anesthesia
there was a significant decrease of MAP after positioning
in both groups compared to baseline but without differ-
ence between the groups. However, there was a significant
decrease in rScO2 values compared to baseline only in the
beach chair (80.3% CDEs) group which was also significant
compared to the lateral decubitus group (0% CDEs). This is
in accordance with our results, as we registered a signifi-
cant decrease in MAP compared to baseline during the whole
surgery but only until surgery begin a difference between
the groups. (Fig. 1) These results challenge the hypoth-
esis, that the drop in blood pressure might be the main
factor influencing regional cerebral perfusion and there-
fore clinical outcome.5,19,20 This is highlighted by the studies
by McCulloch20 and Soeding21 who analyzed the cerebral
blood flow during BCP and described a decrease in cere-
bral artery blood flow constant with drop in mean arterial
pressure during BCP with impairment of cerebral autoregu-
lation. However, if mean arterial pressure was maintained
stable, cerebral autoregulation was not impaired21 . Despite
no significant difference between the groups in MAP during
surgery in our collective (Fig. 1), we found 71% incidence
71
of CDEs in the G-group compared to 2% in the R-group. The
incidence of cardiovascular risk factors in both groups was
similar (Table 1) but obviously the cerebral autoregulation
is less compromised under regional anesthesia.19 In fact, we
could show that in the R-group an increase of rScO2 com-
pared to baseline was measured 5 min after beach chair
position compared to a drop of rScO2 in the G-group. Unfor-
tunately, the other 2 studies using rScO2 monitoring and
regional anesthesia did not focus on rScO2 values after beach
chair positioning7,10 and further studies must confirm our
findings.
Salazar et al. assessed possible risk factors for CDEs in
CBP and they found that a BMI ≥34 kg/m2 was a signif-
icant risk factor for developing CDE.12 In a second study
on 51 patients with sevoflurane and nitrous oxide general
anesthesia in BCP they excluded surgery duration as a risk
factor for developing CDE.11 Our results show that cardio-
vascular risk factors (hypertension, coronary artery disease)
are more common in patients with CDEs under standardized
general anesthesia with controlled hypotension. However,
we could only show a weak correlation. Contrary to Salazar
et al. BMI was not a risk factor for developing CDE.11,12
Probably this is due to the fact, that our patients had a
lower BMI (26.2 ± 2.1 kg/m2 ) compared to their population
(37.2 ± 8.3 kg/m2 ). Hemodynamic instability of higher BMI
under volatile anesthesia might have lead to a higher num-
ber of hypotension events in this population. Time to meet
discharge criteria was different comparing regional to gen-
eral anesthesia. This has been described in several studies
and is attributed to the negative impact of general anesthe-
sia on the respiratory, consciousness and circulation criteria
of the Aldrete score. More interesting is the difference to
meet discharge criteria within the G-groups depending on
the incidence of CDEs which is in accordance with findings in
visceral and heart surgery studies and again shows a clinical
relevant impact of CDEs.21
There are several limitations in our study. First, patients
were not randomized and anesthesia was decided accord-
ing to above mentioned clinical standard. In fact, both
groups were not perfectly comparable if we focus on
age, severity of surgery, fluids, transfusions and hyper-
tension/hypotension therapies needed. However, general
anesthesia for the most difficult cases like prothesis and
osteosynthesis is clinical standard and we could show a
higher incidence of CDEs with clinical consequences exactly
in this population. Moreover, the treatment protocols for
blood pressure used did not include drugs known to influence
cerebral blood flow compared to other recent publications
which included phenylephrine22 in their protocol proba-
bly influencing their results.23---25 Additionally, despite more
aggressive treatment (more fluids, more transfusions, more
vasopressors) to improve blood pressure and oxygen carri-
ers in the G-group there were more patients showing CDEs
in this group showing again, that blood pressure and oxy-
gen saturation are probably not the most important factors
influencing rScO2 .
Second, the neurocognitive tests were only performed
at baseline and after 24 h. Test results after 3, 6 and
12 months would be of further interest. No prior study
has compared regional anesthesia with conscious sedation
to intravenous general anesthesia focusing on the inci-
dence of CDEs depending on anesthesia regimen and on the
72
neurobehavioral outcome of these events following a com-
monly used controlled hypotension protocol for shoulder
surgery in BCP. We could show a negative impact of gen-
eral anesthesia on CDEs and on neurobehavioral outcome
probably due to systemic and regional hypotension. More-
over, cardiovascular risk factors were associated with an
increased risk of CDEs. Regional anesthesia apparently has
a less important effect on systemic and regional perfusion
reducing the CDEs and allowing a PACU bypass after surgery.
Therefore, we believe that regional anesthesia should be
performed whenever possible at least for patients with car-
diovascular risk factors. Additionally, rScO2 monitoring with
NIRS detects decreased cerebral perfusion allowing a rapid
intervention increasing patient’s safety.
Financial support
Financial support was provided solely by departmental
sources.
Conflict of interest
The authors declare no conflict of interest.
References
1. Papadonikolakis A, Wiesler ER, Olympio MA, Poehling GG. Avoid-
ing catastrophic complications of stroke and death related
to shoulder surgery in the sitting position. Arthroscopy.
2008;24:481---2.
2. Pohl A, Cullen DJ. Cerebral ischemia during shoulder surgery
in the upright position: a case series. J Clin Anesth.
2005;17:463---9.
3. Moerman AT, De Hert SG, Jacobs TF, De Wilde LF, Wouters PF.
Cerebral oxygen desaturation during beach chair position. Eur
J Anaesthesiol. 2012;29:82---7.
4. Casati A, Spreafico E, Putzu M, Fanelli G. New technology for
noninvasive brain monitoring: continuous cerebral oximetry.
Minerva Anestesiol. 2006;72:605---25.
5. Murphy GS, Szokol JW, Marymont JH, Greenberg SB, Avram MJ,
Vender JS, et al. Cerebral oxygen desaturation events assessed
by near-infrared spectroscopy during shoulder arthroscopy in
the beach chair and lateral decubitus positions. Anesth Analg.
2010;111:496---505.
6. Lee JH, Min KT, Chun YM, Kim EJ, Choi SH. Effects of beach-
chair position and induced hypotension on cerebral oxygen
saturation in patients undergoing arthroscopic shoulder surgery.
Arthroscopy. 2011;27:889---94.
7. Yadeau JT, Liu SS, Bang H, Shaw PM, Wilfred SE, Shetty T, et al.
Cerebral oximetry desaturation during shoulder surgery per-
formed in a sitting position under regional anesthesia. Can J
Anaesth. 2011;58:986---92.
8. Tange K, Kinoshita H, Minonishi T, Hatakeyama N, Matsuda N,
Yamazaki M, et al. Cerebral oxygenation in the beach chair posi-
tion before and during general anesthesia. Minerva Anestesiol.
2010;76:485---90.
9. Jeong H, Jeong S, Lim HJ, Lee J, Yoo KY. Cerebral oxygen sat-
uration measured by near-infrared spectroscopy and jugular
venous bulb oxygen saturation during arthroscopic shoul-
der surgery in beach chair position under sevoflurane-nitrous
J. Aguirre et al.
oxide or propofol-remifentanil anesthesia. Anesthesiology.
2012;116:1047---56.
10. Koh JL, Levin SD, Chehab EL, Murphy GS. Cerebral oxygenation
in the beach chair position: a prospective study on the effect of
general anesthesia compared with regional anesthesia and seda-
tion. J Shoulder Elbow Surg. 2013, http://dx.doi.org/10.1016/
j.jse.2013.01.035, pii:S1058-2746(13)00092-X [Epub ahead of
print].
11. Salazar D, Sears BW, Andre J, Tonino P, Marra G. Cerebral
desaturation during shoulder arthroscopy: a prospective obser-
vational study. Clin Orthop Relat Res. 2013 [Epub ahead of
print].
12. Salazar D, Sears BW, Aghdasi B, Only A, Francois A, Tonino P,
et al. Cerebral desaturation events during shoulder arthroscopy
in the beach chair position: patient risk factors and neu-
rocognitive effects. J Shoulder Elbow Surg. 2013;22:1228---35,
http://dx.doi.org/10.1016/j.jse.2012.12.036 [Epub ahead of
print].
13. Murkin JM, Newman SP, Stump DA, Blumenthal JA. Statement
of consensus on assessment of neurobehavioral outcomes after
cardiac surgery. Ann Thorac Surg. 1995;59:1289---95.
14. Fassoulaki A, Kaliontzi H, Petropoulos G, Tsaroucha A. The
effect of desflurane and sevoflurane on cerebral oximetry under
steady-state conditions. Anesth Analg. 2006;102:1830---5.
15. Bruhn J, Schumacher PM, Bouillon TW. Effect compart-
ment equilibration and time-to-peak effect. Importance of a
pharmacokinetic-pharmacodynamic principle for the daily clin-
ical practice. Anaesthesist. 2005;54:1021---31.
16. Wu Q, Sun B, Wang S, Zhao L, Qi F. Estimating the plasma
effect-site equilibrium rate constant (K) of propofol by fitting
time of loss and recovery of consciousness. Biol Pharm Bull.
2013.
17. Chernik DA, Gillings D, Laine H, Hendler J, Silver JM, Davidson
AB, et al. Validity and reliability of the Observer’s Assessment of
Alertness/Sedation Scale: study with intravenous midazolam. J
Clin Psychopharmacol. 1990;10:244---51.
18. Aldrete JA. The post-anesthesia recovery score revisited. J Clin
Anesth. 1995;7:89---91.
19. Murphy GS, Szokol JW. Blood pressure management during
beach chair position shoulder surgery: what do we know. Can
J Anaesth. 2011;58:977---82.
20. McCulloch TJ, Liyanagama K, Petchell J. Relative hypoten-
sion in the beach-chair position: effects on middle cerebral
artery blood velocity. Anaesth Intensive Care. 2010;38:
486---91.
21. Soeding PF, Wang J, Hoy G, Jarman P, Phillips H, Marks P, et al.
The effect of the sitting upright or ‘beachchair’ position on
cerebral blood flow during anaesthesia for shoulder surgery.
Anaesth Intensive Care. 2011;39:440---8.
22. Soeding PF, Hoy S, Hoy G, Evans M, Royse CF. Effect of
phenylephrine on the haemodynamic state and cerebral oxy-
gen saturation during anaesthesia in the upright position. Br J
Anaesth. 2013;111:229---34.
23. Koh JL, Levin SD, Chehab EL, Murphy GS. Cerebral oxygenation
in the beach chair position: a prospective study on the effect
of general anesthesia compared with regional anesthesia and
sedation. J Shoulder Elbow Surg. 2013.
24. Salazar D, Sears B, Aghdasi B, Only A, Francois A, Tonino P, et al.
Cerebral desaturation events during shoulder arthroscopy in the
beach chair position: patient risk factors and neurocognitive
effects. J Shoulder Elbow Surg. 2013.
25. Salazar D, Sears BW, Andre J, Tonino P, Marra G. Cerebral
desaturation during shoulder arthroscopy: a prospective obser-
vational study. Clin Orthop Relat Res. 2013.