6/99 issue pt.

1 5/11/99 1:35 PM Page 1042

J Neurosurg 90:1042–1052, 1999

Propofol in the treatment of moderate and severe head injury:

a randomized, prospective double-blinded pilot trial

DANIEL F. KELLY, M.D., DAVID B. GOODALE, D.D.S., PH.D., JOHN WILLIAMS, M.D.,
DANIEL L. HERR, M.D., E. THOMAS CHAPPELL, M.D., MICHAEL J. ROSNER, M.D.,
JEFF JACOBSON, M.D., MICHAEL L. LEVY, M.D., MARTIN A. CROCE, M.D.,
ALLEN H. MANIKER, M.D., GERALD J. FULDA, M.D., JAMES V. LOVETT, M.D.,
OLGA MOHAN, M.D., AND RAJ K. NARAYAN, M.D.
Division of Neurosurgery, University of California Medical Center, Los Angeles, California; University
of California–Harbor Medical Center, Torrance, California; Washington Hospital Center, Washington,
DC; Highland Hospital, University of California: Davis–East Bay, Oakland, California; University of
Alabama Hospital, Birmingham, Alabama; New Jersey Medical School and Hospital, Newark, New
Jersey; University Medical Center, Las Vegas, Nevada; University of Southern California–Los Angeles
County, Los Angeles, California; Medical Center of Delaware, Newark, Delaware; University of
Tennessee Medical Center, Memphis, Tennessee; and Temple University Hospital, Philadelphia,
Pennsylvania

Object. Sedation regimens for head-injured patients are quite variable. The short-acting sedative–anesthetic agent pro-
pofol is being increasingly used in such patients, yet little is known regarding its safety and efficacy. In this multicenter
double-blind trial, a titratable infusion of 2% propofol accompanied by low-dose morphine for analgesia was compared
with a regimen of morphine sulfate in intubated head-injured patients. In both groups, other standard measures of control-
ling intracranial pressure (ICP) were also used.
Methods. Forty-two patients from 11 centers were evaluated to assess both the safety and efficacy of propofol: 23
patients in the propofol group (mean time of propofol usage 95  87 hours) and 19 patients in the morphine group (mean
time of morphine usage 70  54 hours). There was a higher incidence of poor prognostic indicators in the propofol group
than in the morphine group: patient age older than 55 years (30.4% compared with 10.5%, p  0.05), initial Glasgow Coma
Scale scores of 3 to 5 (39.1% compared with 15.8%, p  0.05), compressed or absent cisterns on initial computerized
tomography scanning (78.3% compared with 57.9%, p  0.05), early hypotension and/or hypoxia (26.1% compared with
10.5%, p = 0.07). During treatment there was a trend toward greater use of vasopressors in the propofol group. However,
the mean daily ICP and cerebral perfusion pressure were generally similar between groups and, on therapy Day 3, ICP was
lower in the propofol group compared with the morphine group (p  0.05). Additionally, there was less use of neuro-
muscular blocking agents, benzodiazepines, pentobarbital, and cerebrospinal fluid drainage in the propofol group (p 
0.05). At 6 months postinjury, a favorable outcome (good recovery or moderate disability) was observed in 52.1% of
patients receiving propofol and in 47.4% receiving morphine; the mortality rates were 17.4% and 21.1%, respectively.
Patients who received the highest doses of propofol for the longest duration tended to have the best outcomes. There were
no significant differences between groups in terms of adverse events.
Conclusions. Despite a higher incidence of poor prognostic indicators in the propofol group, ICP therapy was less inten-
sive, ICP was lower on therapy Day 3, and long-term outcome was similar to that of the morphine group. These results
suggest that a propofol-based sedation and an ICP control regimen is a safe, acceptable, and, possibly, desirable alterna-
tive to an opiate-based sedation regimen in intubated head-injured patients.

KEY WORDS • propofol • traumatic brain injury • intracranial hypertension •

neuroprotection

in the acute phase after moderate or severe be supported by published data on the use of these agents.

P
ATIENTS
head injury often exhibit agitation and fluctuations
in ventilatory capacity. A variety of sedatives, nar-
cotics, and neuromuscular blocking agents are used to
control ventilation and prevent agitation-related intracra-
nial pressure (ICP) spikes in these patients. Sedative reg-
imens in head-injured patients, however, remain non-
standardized and largely unproven. In a recently published
set of guidelines,12 no “standards” or “guidelines” could
Only the following “option” was offered: “Sedation and
neuromuscular blockade can be useful in optimizing
transport of the head-injured patient. However, both treat-
ments interfere with the neurological examination. In the
absence of outcome-based studies, the choice of sedative
is left to the physician. Neuromuscular blockade should
be employed when sedation alone proves inadequate, and
short-acting agents should be used when possible.”12

1042 J. Neurosurg. / Volume 90 / June, 1999

6/99 issue pt. 1 5/11/99 1:35 PM Page 1043
Propofol in traumatic brain injury
The short-acting sedative–anesthetic nonanalgesic
agent propofol (Diprivan; Zeneca Pharmaceuticals, Wil-
mington, DE) is being increasingly used in patients in
the neurosurgical intensive care unit (ICU), particularly
for head-injured patients.41 However, only a few studies
have addressed its efficacy in these patients.18,43,58 Propofol
is potentially advantageous in this setting, given its wide
dose response, short elimination halflife (24–64 minutes),
and neuroprotective effects.2,6,29,33,69,72 Low infusion rates
result in light-to-moderate sedation and high rates result in
electroencephalographic (EEG) burst suppression similar
to that achieved with high-dose pentobarbital.22,51,65 Pre-
viously, propofol has been dispensed as a 1% infusion.
A new 2% formulation was produced to reduce infusion
volumes, and the additive ethylenediamine tetraacetic acid
(EDTA) was included to reduce the risk of bacterial colo-
nization of the lipid vehicle. Using this new formulation,
a prospective randomized double-blinded pilot trial was
undertaken to assess the safety and efficacy of a propo-
fol-based sedation regimen compared with a regimen of
morphine sulfate in head-injured patients requiring me-
chanical ventilation. Although the primary study end point
was determination of drug safety, the clinical end points
included: 1) control of ICP and maintenance of cerebral
perfusion pressure (CPP); 2) therapy intensity for ICP
control, CPP control, and sedation; 3) 6-month neurologi-
cal outcome; and 4) treatment-related adverse events.
Clinical Material and Methods
Ethical Considerations
This protocol was independently reviewed and ap-
proved by the institutional review board of each partici-
pating institution. The trial was conducted in compliance
with the regulations governing informed consent promul-
gated by the United States Food and Drug Administration
(Code of Federal Regulations 21, Parts 50 and 56). Writ-
ten informed consent for participation was required from
the next of kin of each patient. The trial was also designed
and monitored in accordance with the procedures of Zen-
eca Ltd., which include the key elements of good clinical
practices as required by regulatory authorities.
Patient Population
Inclusion and Exclusion Criteria. Eligible patients in-
cluded those who were aged 17 years or older, had sus-
tained a closed or penetrating traumatic brain injury
resulting in a postresuscitation Glasgow Coma Scale
(GCS)61 score of 3 to 12, and required mechanical ventila-
tion. Patients were excluded from the efficacy analysis if
trial medication therapy was not begun within 96 hours
of injury and medications were not administered for a
minimum of 12 hours. Other exclusion criteria included a
history of hypersensitivity to propofol or its constituents,
renal insufficiency, hepatic failure, a known lipid metabo-
lism disorder, spinal cord injury with paraplegia or quad-
riplegia, participation in another investigational drug trial
within 31 days of trial entry, or body weight more than
130 kg.
Group Assignment and Randomization. Patients were
stratified by postresuscitation GCS score into four cate-
J. Neurosurg. / Volume 90 / June, 1999
TABLE 1
Stepwise control of ICP
Step Therapy
base- mild hyperventilation (PaCO2 30–35 mm Hg); normothermia
line (temperature 37.4˚C)
1 ventricular CSF drainage: every 5–10 mins until ICP 20 mm Hg
2* increase sedation: increase Infusions A & B per titration sched-
ule
3 neuromuscular blockade: if patient remains agitated, is over-
breathing ventilator, bucking, or shivering (0.1 mg/kg vecuro-
nium given intravenously, repeat every 30 mins as needed)
4 mannitol: if ICP remains elevated for 5 mins, administer 25–50
g every 2 hrs as needed to maintain ICP 20 mm Hg (maximum
250 g/24 hrs & maintain serum osmolality 310)
5 high-dose pentobarbital: for intractable intracranial hypertension
may be instituted after Infusions A & B have been titrated to
max rates & ICP remains 30 mm Hg w/ CPP 70 mm Hg or
ICP 40 mm Hg w/ CPP 70 mm Hg; titrate to achieve burst
suppression on EEG
* If systolic blood pressure drops to less than 100 mm Hg or CPP to less
than 60 mm Hg, both Infusions A and B should be lowered accordingly. If
the blood pressure does not improve, both infusions should be stopped.
gories (3, 4–5, 6–8, or 9–12) and randomized in a double-
blind fashion to either the propofol or morphine sedation
regimens. All patients received three simultaneous in-
fusions, A, B, and C. Infusion A was a white emulsion
containing either propofol or Intralipid (Pharmacia Up-
john, London, United Kingdom) as placebo. Infusion B
was a clear solution composed of either morphine sulfate
or normal saline as placebo. Infusion C was a low-dose
morphine sulfate infusion for analgesia. For the propofol
group patients, Infusion A consisted of propofol (ZD08-
59#1 2% [20 mg/ml] with 0.005% EDTA) and Infusion B
was normal saline. For morphine group patients, Infusion
A was Intralipid and Infusion B was morphine. Given that
propofol lacks analgesic effects, both groups also received
an initial 2.5-mg bolus of morphine followed by Infusion
C of morphine at 1 to 3 mg/hour for at least 48 hours.
Patient Management and Dosing Regimens
All patients were admitted to the ICU after initial stabi-
lization or after craniotomy for evacuation of an intra-
cranial hematoma. Management was in concordance with
previously published guidelines for managing patients
with severe head injury12 and included an algorithm for
maintaining ICP less than 20 mm Hg and CPP higher than
70 mm Hg.12 Intracranial pressure therapy was imple-
mented in a stepwise fashion as outlined in Table 1. The
mode of ICP monitoring was left to the discretion of the
investigator, although use of a ventriculostomy was en-
couraged. The sedation schedule outlined in Table 2 was
used forall patients. As trial drug infusions began, patients
were weaned from sedatives, narcotics, and neuromuscu-
lar blocking agents. Infusions A and B were initiated
simultaneously at Step 1 and were increased as needed for
agitation, bucking on the ventilator, abnormal posturing,
or ICP levels persisting above 20 mm Hg. Trial Infusions
A and B could be increased at a maximum rate of one step
every 5 minutes. In patients without physiological evi-
dence of pain (for example, those with tachypnea, tachy-
cardia, or hypertension), administration of low-dose mor-
1043
6/99 issue pt. 1 5/11/99 1:35 PM Page 1044

D. F. Kelly, et al.

TABLE 2 Efficacy Assessment

Titration scheme* Intracranial pressure and CPP measurements were re-
Infusion A Infusion B† corded on an hourly basis as previously described.38 For
Infusion Step ml/hr (µg/kg/min) ml/hr (µg/kg/min) each patient, baseline ICP and CPP before receiving study
drugs, mean daily ICP and CPP, and total number of hours
1 initial rate 2 (10) 1 (0.25) that ICP was greater than 20 mm Hg and CPP was lower
2 4 (20) 2 (0.50)
3 6 (30) 3 (0.75) than 70 mm Hg were recorded. The frequency of use of
4 8 (40) 4 (1.00) mannitol, neuromuscular blocking agents, benzodiaze-
5 10 (50) 5 (1.25) pines, barbiturates, and vasopressors, and the daily vol-
6 12 (60) 6 (1.50) ume of CSF drainage were recorded. Neurological out-
7 14 (70) 7 (1.75) come was determined at 6 months postinjury by obtaining
8 benzodiazepines allowed 16 (80) 8 (2.00) the Glasgow Outcome Scale (GOS)31 and Disability Rat-
9 18 (90) 9 (2.25)
10 20 (100) 10 (2.50) ing Scale (DRS)52 scores. Favorable outcome was defined
11 22 (110) 11 (2.75) as a GOS score of good recovery or moderate disability or
12 23 (120) 12 (3.00) a DRS score of no, partial, or moderate disability (DRS
13 25 (130) 13 (3.25) score of 0–6).
14 27 (140) 14 (3.50)
15 maximum rate‡ 29 (150) 15 (3.75)
16 31 (160) 16 (4.00) Safety Assessment
17 33 (170) 17 (4.25)
18 35 (180) 18 (4.50) Clinical Laboratory Assessments. Multiple blood sam-
19 37 (190) 19 (4.75)
20 39 (200) 20 (5.00)
ples were drawn to measure ionized calcium, total calci-
um, total magnesium, potassium, sodium, inorganic phos-
* For the propofol group, Infusion A consisted of 2% propofol and phate, blood urea nitrogen, creatinine, triglyceride, and
Infusion B consisted of normal saline; for the morphine group, Infusion A
consisted of 10% Intralipid and Infusion B consisted of morphine sulfate.
parathyroid hormone concentration.
Dosages of all trial medications were calculated using patient’s baseline Hemodynamics. Systolic, diastolic, and mean arterial
body weight (kg). The rates shown above were calculated only as an exam- blood pressures, heart rate, central venous pressure, cardi-
ple using the weight of a 65-kg adult.
† Infusion B rates are based on a MgSO4 concentration of 1 mg/ml.
ac index, and body temperature were recorded immedi-
‡ Rates higher than this were allowed only with permission of Zeneca. ately before initiation of trial drug therapy, 4 hours after
start time, and daily thereafter until 24 hours after admin-
istration of trial drugs was discontinued.
phine (Infusion C) could be stopped after 48 hours. Neu- Adverse Events. Fatalities, life-threatening events asso-
romuscular blocking agents could be given at any time ciated with trial drug use, and any other adverse events
if the patient was not adequately responsive to trial medi- that were both serious and unexpected were reported. Pa-
cations. Benzodiazepines could be used after Step 8 had tients were also evaluated daily for signs or symptoms of
been reached. Mannitol could also be administered if ele- sepsis. Criteria included: 1) tachycardia (heart rate  90
vated ICP was not responsive to ventricular cerebrospinal beats/minute); 2) hyperthermia (temperature  38˚C); 3)
fluid (CSF) drainage and to increasing Infusions A and B. hypothermia (temperature  36˚C); 4) white blood cell
Metabolic suppressive therapy, using high-dose pentobar- elevation ( 12,000/mm3); 5) white blood cell depression
bital, could be introduced for intractable intracranial hy- ( 4000/mm3); and 6) positive blood culture(s). Patients
pertension only after patients reached maximum rates for were considered to have sepsis if two of the first five cri-
Infusions A and B (Table 2). Cerebral perfusion pressure teria and the sixth criterion were met.
was maintained with the use of intravascular volume ex-
pansion and vasopressor therapy. Statistical Considerations and Sample Size

Injury Severity Determinants

Prognostic indicators were documented, including pa-
tient age, postresuscitation GCS score, systemic hypoten-
sion (systolic blood pressure  90 mm Hg), hypoxia (ar-
terial oxygen saturation  60 mm Hg), and abnormal
pupillary responses (poorly reactive or an asymmetry of ≥
2 mm) in the prehospital setting or during the first 24
hours of hospitalization.13,37,39,67 The first two computerized
tomography (CT) scans obtained in each patient were also
analyzed for perimesencephalic cistern effacement, dif-
fuse swelling, diffuse swelling with a midline shift greater
than 4 mm, diffuse injury with punctate hemorrhages,
subarachnoid hemorrhage, evacuated mass lesions, multi-
ple contusions, and gunshot wounds.17,23,40,63 An Injury Se-
verity Scale5 score was also calculated for each patient
(Table 3).
The primary end point of the trial was determination of
drug safety. Specifically, the study sought to determine
whether ZD0859#1 2% propofol significantly altered ion-
ized calcium, serum total calcium, total magnesium, po-
tassium, sodium, inorganic phosphate, or parathyroid
hormone levels when administered for sedation of severe-
ly head injured patients. Given this objective, the power
analysis was determined on the basis of projected standard
deviations (SDs) for changes in ionized calcium. Al-
though the original protocol called for 100 patients, a low-
er number of patients was deemed sufficient because
no clinically significant changes in calcium homeostasis
were reported in an earlier Diprivan 1% trial, approved
by the Food and Drug Administration in June 1996. Both
the 1% and 2% formulations have the same concentration
of EDTA (0.005%). For these reasons, this trial was ter-
minated earlier than expected because sufficient safety

1044 J. Neurosurg. / Volume 90 / June, 1999

6/99 issue pt. 1 5/11/99 1:35 PM Page 1045

Propofol in traumatic brain injury

data had been acquired. All patients treated with trial TABLE 3
drugs, regardless of duration, were included in the safety Characteristics of 42 patients with moderate or
analyses. severe head injury*
Because this was a multicenter trial, all statistical analy-
ses and results were based on appropriate pooling of data Propofol Morphine
Characteristic (23 patients) (19 patients)
across centers. Between-group analysis of continuous
variables was assessed using Student’s t-test or analysis of time from injury to protocol entry (hrs) 34  21 38  26
variance (ANOVA) for comparison of more than two time on study medications (hrs) 95  87 70  54
groups. Multiple comparisons were not adjusted for an mean propofol dose (µg/kg/min) 55  42 0
mean morphine dose (mg/hr) 1.3  0.7 10  6.7
inflated alpha. Percentage comparisons were performed age (yrs) 39  18 33  13
using a Bernoulli process (binomial distribution). The no. of patients 55 yrs† 7 (30.4%) 2 (10.5%)
Mann–Whitney U-test was used to assess differences for median postresuscitation GCS score (range) 7 (3–9) 6 (3–10)
nonparametric factors such as median GCS score and postresuscitation GCS score 3–5† 9 (39.1%) 3 (15.8%)
median number of CT diagnoses. All variances were ex- male/female ratio 18:5 17:2
pressed as an SD. All differences for which the probabili- no. of patients w/ abnormal pupils 14 of 22 10 (52.6%)
(1 or both) (63.6%)
ty value was less than 0.05 were considered significant. no. of patients w/ abnormal pupils (both) 10 of 22 10 (52.6%)
(45.5%)
no. of patients w/ early hypotension 6 (26.1%) 2 (10.5%)
Results &/or hypoxia‡
median CT diagnoses (range) 3 (0–7) 4 (0–6)
Enrollment and Patient Exclusions absent or compressed cisterns† 18 (78.3%) 11 (57.9%)
diffuse swelling w/ or w/o shift 11 (47.8%) 11 (57.9%)
From September 1, 1995 to August 18, 1996, 70 acute- evacuated hematomas (all) 10 (43.5%) 9 (47.4%)
ly head injured adults from 11 centers were randomized; epidural 4 (17.4%) 3 (15.8%)
however, five were withdrawn before being given trial subdural 6 (26.1%) 4 (21.1%)
drugs. Of the remaining 65 patients, a total of 23 patients intracerebral 3 (13%) 5 (26.3%)
multiple contusions 10 (43.5%) 8 (42.1%)
were not included for evaluation of efficacy of trial drugs gunshot wounds 2 (8.7%) 3 (15.8%)
for the following reasons. In two patients (one from each Injury Severity Scale score 27  9 24  8
treatment group), trial drugs were not administered within
96 hours of injury and, in seven patients, trial drugs were * Values are expressed as the mean  SD.
† p  0.05.
not given for at least 12 hours (four in the propofol group ‡ p  0.07.
and three in the morphine group). In seven patients in the
morphine group varying amounts of 1% propofol were
administered before beginning trial medications. Three (39.1% compared with 15.8%, p  0.05). The GCS score
patients from this group were excluded because they re- was 8 or less in 22 patients (95.7%) in the propofol group
ceived relatively large amounts of propofol during a peri- and in 17 (89.5%) in the morphine group. Pupillary abnor-
od of at least 24 hours. The remaining four patients were malities were documented within the 1st day postinjury
included for analysis because they received small sedating with approximately equal frequency. Systemic hypoten-
amounts of 1% propofol for less than 12 hours, all less sion and/or hypoxic insults were documented through the
than 20 g/kg/minute and averaging 12.3 g/kg/minute. 1st day postinjury in six patients (26.1%) receiving pro-
Two patients, both in the propofol group, were excluded pofol and in two patients (10.5%) receiving morphine (p =
because only telephone consent was obtained. Also, one 0.07). The median number of major CT diagnoses was
patient in the propofol group was excluded because police similar in both groups. However, the propofol group had
records and autopsy findings had indicated the patient sus- a higher number of patients with compressed or absent
tained a primary anoxic injury from strangulation with a cisterns on initial CT scans (78.3% compared with 57.9%,
resultant common carotid occlusion and multiple cerebral p  0.05). The frequency of other CT diagnoses was sim-
infarctions within 2 days of injury. Of the remaining ilar between the two groups (Table 3).
50 patients (27 in the propofol regimen and 23 in the
morphine regimen), eight patients (four in each treatment Trial Drug Administration
group) were lost to follow-up review, leaving 23 patients The average times after injury that trial infusions were
in the propofol group and 19 in the morphine group with initiated in the propofol and morphine groups were 34 
a 6-month follow up. These 42 patients are the focus of 21 hours and 38  26 hours, respectively. The mean dura-
this paper. tion of therapy was 95  87 hours in the propofol group
and 70  54 hours in the morphine group (p = 0.26). The
Patient Characteristics and Predictors of Outcome by average infusion rates for the propofol group were 55 
Patient Group 42 g/kg/minute of propofol and 1.3  0.7 mg/hour of
The patients’ median ages did not differ significantly morphine. For the morphine group, combining the titrat-
but there was a higher proportion of patients older than 55 able and low-dose infusions, the mean morphine infusion
years in the propofol group (30.4%) compared with the rate was 10  6.7 mg/hour.
morphine group (10.5%) (p  0.05). The median postre-
suscitation GCS score was 7 in the propofol group and 6 Intracranial Pressure, CPP, and Therapy Intensity
in the morphine group; however, the propofol group had a The modes of ICP monitoring in the propofol group
higher proportion of patients with a GCS score of 3 to 5 included ventriculostomy in 16 patients and a parenchy-

J. Neurosurg. / Volume 90 / June, 1999 1045

6/99 issue pt. 1 5/11/99 1:35 PM Page 1046

D. F. Kelly, et al.

TABLE 4
Therapy intensity for ICP and CPP control*
Propofol Morphine
Therapy (21 patients) (19 patients) p Value

drug treatment†
mannitol 45.5% 43.6% 0.09
NMB 33.7% 47.4% 0.01
benzodiazepines 3.0% 12.8% 0.001
pentobarbital 1.0% 6.4% 0.01
vasopressors 65.3% 61.5% 0.07
total no. of treatment days 101 78
CSF drainage by therapy day (mean ml  SD)
Day 1 48  55 (16) 77  112 (15) 0.38
Day 2 77  74 (15) 191  102 (16) 0.002
Day 3 75  105 (12) 179  127 (13) 0.05
Day 4 86  108 (9) 134  80 (8) 0.31
* Number of patients with CSF drainage is shown in parentheses. Ab-
breviation: NMB = neuromuscular blocking agents.
† The frequency of daily use of a given drug is expressed as a percent-
age of the total number of treatment days for each group.

propofol group (34.4  5.9 mm Hg) and the morphine

FIG. 1. Graphs demonstrating daily ICP and CPP for the propo-
fol (PRO) and morphine (MS) groups. Values are expressed as the
mean  SD. The only significant difference (asterisk) was ob-
served for ICP on therapy Day 3 in which ICP averaged 14 mm Hg
in the propofol group and 18 mm Hg in the morphine group (p 
0.05). The numbers of patients in each group on each therapy day
are shown at the top of each graph. In one patient in the morphine
group, baseline ICP and CPP values were considered erroneous by
the site investigator and were excluded. In another patient in the
morphine group, no CPP value was recorded for therapy Day 1.
mal monitor in five patients; two patients did not under-
go monitoring. The monitoring modes in the morphine
group included ventriculostomy in 16 patients and a
parenchymal monitor in three patients. For the 21 moni-
tored patients in the propofol group and the 19 patients
in the morphine group, daily mean ICP and CPP are
displayed in Fig. 1. Values were generally similar between
the two groups, except on therapy Day 3 when ICP was
lower in the propofol group (p  0.05). Additionally,
in patients with ventriculostomies, the amount of CSF
drainage was less in the propofol group on therapy Days 2
and 3 than in the morphine group (p  0.002 and p 
0.05, respectively; Table 4). Use of neuromuscular block-
ing agents, benzodiazepines, and pentobarbital was also
less in the propofol group (p  0.05 for all comparisons).
However, there was a trend toward greater use of vaso-
pressors (p = 0.07) and mannitol (p = 0.09) in the propo-
fol group. The mean daily PaCO2 was similar between the
group (36  4.2 mm Hg) while patients were receiving
study drugs (p = 0.32).
Neurological Outcome
At 6 months postinjury, a favorable outcome, as as-
sessed using the GOS, was observed in 12 (52.2%) of the
propofol patients and in nine (47.4%) of the morphine pa-
tients; the mortality rates were 17.4% and 21.1%, respec-
tively (Table 5). If the two patients without ICP monitor-
ing in the propofol group are excluded, the favorable
outcome rate is unchanged. Similarly, 14 patients (60.9%)
in the propofol group and 11 (57.9%) in the morphine
group had favorable outcomes, as assessed using the DRS.
Adverse Events and Cause of Death
The number of patients sustaining adverse events re-
sulting in study withdrawal was five (21.7%) in the propo-
fol group and eight (42.1%) in the morphine group. How-
ever only four events (17.4%) in the propofol group and
six (31.6%) in the morphine group were considered severe
in intensity (Table 6). Uncontrollable intracranial hyper-
tension was the cause of withdrawal in three patients in
the propofol group and in five in the morphine group (p =
0.06) and was considered a key factor in three of four
deaths in both groups. One patient in the propofol group
was withdrawn after 23 hours of trial medications because
of hypotension. She was receiving a moderate dose of pro-
pofol, averaging 55 g/kg/minute. She had sustained se-
vere multiple trauma with chest injuries and experienced
a hypotensive episode before beginning the study medi-
cations. She recovered to a level of moderate disability.
Another patient in the propofol group was withdrawn
because of low CPP after 29 hours of therapy. He was
receiving a relatively low dose of propofol, averaging 28
g/kg/minute, and also recovered to a level of moderate
disability. Pneumonia was documented in five patients
(21.7%) in the propofol group and in two (10.5%) in the
morphine group (p = 0.13). Sepsis occurred in three pa-
tients (13%) in the propofol group and in one (5.3%) in the
morphine group (p = 0.2). None of the infections reported
were considered directly related to the trial medications.

1046 J. Neurosurg. / Volume 90 / June, 1999

6/99 issue pt. 1 5/11/99 1:35 PM Page 1047
Propofol in traumatic brain injury
TABLE 5
Outcomes of 42 patients with moderate or severe head injury*
Propofol Morphine
6-Mo Outcome (23 patients) (19 patients)
good recovery 3 (13%) 5 (26.3%)
moderate disability 9 (39.1%) 4 (21.1%)
severe disability 5 (21.7%) 6 (31.6%)
vegetative state 2 (8.7%) 0 (0%)
death 4 (17.4%) 4 (21.1%)
lost to follow up 4 of 27 (14.8% ) 4 of 23 (17.4%)
* Outcomes based on GOS.
Regarding the effects of Intralipid administration, only
one patient was withdrawn because of hyperlipemia (peak
triglyceride level 471 mg/dl, normal range 10–250 mg/dl).
This patient from the morphine group had been receiving
high doses of trial infusions for 6 days. In the propofol
group, seven patients (30.4%) had at least one episode of
transient elevation in serum triglycerides. These patients
were receiving propofol averaging 64  53 g/kg/minute
for 164  113 hours. They tended to be older with a mean
age of 47 years. The patient with elevated triglycerides
who received study medications for the shortest time (45
hours) and with the lowest mean dose (19 g/kg/minute),
was also the oldest (75 years). He recovered to a level of
moderate disability.
Data Analysis With Inclusion of Patients Lost to
Follow Up
When the eight patients lost to follow-up review were
included in the prognostic indicator and ICU data analy-
ses, similar differences were seen between the two treat-
ment groups for most analyses. Specifically, the propofol
group (27 patients) continued to have a higher incidence
of poor prognostic indicators than the morphine group (23
patients): patient age older than 55 years, GCS score of
3 to 5, early hypotension and/or hypoxia (p  0.05 for
all comparisons), and compressed or absent cisterns on
CT scan (p = 0.1). Regarding ICP, the difference seen on
therapy Day 3 was no longer significant. However, the
amount of CSF drainage remained less on therapy Days
2 and 3 in the propofol group (p  0.01 and p  0.05,
respectively). Use of neuromuscular blocking agents, ben-
zodiazepines, and pentobarbital was also less in the propo-
fol group (p  0.05 for all comparisons). However, use of
mannitol (p  0.05) and vasopressors (p = 0.06) was less
in the morphine group. Regarding adverse events, there
were no intergroup differences in incidence of uncontrol-
lable intracranial hypertension, systemic hypotension, low
CPP, or infections.
Propofol Subgroup Analysis—Dose Effect
Patients in the propofol group were further categorized,
retrospectively, into either high-dose or low-dose sub-
groups based on the total amount of propofol received, the
time to initiation of therapy, and infusion duration. High-
dose patients were defined as receiving a total propofol
dose of at least 100 mg/kg, beginning propofol within 48
hours of injury, and remaining on infusion for a minimum
of 24 hours; low-dose patients did not meet at least one of
J. Neurosurg. / Volume 90 / June, 1999
TABLE 6
Adverse events resulting in withdrawal from study
Propofol Morphine
Event (23 patients) (19 patients)
uncontrollable intracranial hypertension* 3 (13%)† 5 (26.3%)†
reduced CPP 1 (4.3%)†‡ 0
hypotension 1 (4.3%)‡ 0
cardiac arrhythmia 0 1 (5.3%)†
hyperlipemia 0 1 (5.3%)‡
agitation 0 1 (5.3%)†‡
* p = 0.06.
† Considered severe in intensity. In the morphine group four of five
instances of uncontrollable ICP were considered severe in intensity and
one was deemed mild.
‡ Considered related to study medications.
these criteria. The 10 patients in the high-dose subgroup
were compared with 13 patients in the low-dose subgroup
and 19 patients in the morphine group (Tables 7 and 8).
Predictors of Outcome. The propofol high-dose sub-
group was more severely injured according to the initial
GCS score and had a higher incidence of multiple contu-
sions compared with both the low-dose subgroup and the
morphine group. There was also a trend toward a higher
incidence of abnormal cisterns in the high-dose group
compared with the morphine group. The low-dose sub-
group, however, consisted of patients who were signifi-
cantly older than both the high-dose subgroup and the
morphine group and had a higher incidence of evacuated
subdural hematomas compared with the high-dose sub-
group.
Intracranial Pressure and CPP Control. There were no
significant differences across the three groups regarding
mean daily ICP and CPP values. On therapy Day 2, how-
ever, there was a trend toward a higher mean CPP in the
high-dose propofol subgroup (81  8 mm Hg) compared
with the low-dose (68  19 mm Hg) subgroup and the
morphine group (72  12 mm Hg) (p = 0.16, ANOVA).
Similarly, on therapy Day 3 there was a trend toward
lower ICP in the high-dose (14  5 mm Hg) and low-dose
(15  6 mm Hg) subgroups compared with the morphine
group (18  4 mm Hg) (p = 0.12, ANOVA).
Therapy Intensity for CPP and ICP Control. Use of CSF
drainage was less in both the high-dose and low-dose pro-
pofol subgroups on therapy Days 2 and 3 compared with
the morphine group (p  0.01 and p = 0.11, respectively).
Additionally, on therapy Day 4, there was a trend of less
CSF drainage in the high-dose subgroup compared with
the morphine group (p = 0.06). Use of mannitol and neu-
romuscular blocking agents was greater in the high-dose
subgroup and the morphine group, whereas benzodiaze-
pine and pentobarbital use was greatest in the morphine
group. In contrast, use of vasopressors was significantly
greater in the high-dose subgroup compared with both the
low-dose subgroup and the morphine group.
Neurological Outcome. A favorable neurological out-
come, as assessed using the GOS and DRS, was achieved
in 70% and 80% of patients in the high-dose propofol sub-
group, compared with 38.5% and 46.2% in the low-dose
subgroup, respectively (p  0.05 for both GOS and DRS
1047
6/99 issue pt. 1 5/11/99 1:35 PM Page 1048

D. F. Kelly, et al.

TABLE 7 tensity for ICP control was less, and there was a trend for
Propofol subgroup analysis* fewer patient withdrawals for intractable intracranial hy-
pertension in the propofol group. Furthermore, 6 months
Propofol Group postinjury neurological outcome was similar in the two
High Low Morphine groups. When the eight patients lost to follow-up review
Patient Characteristics & Outcome Dose Dose Group were included in the acute data analysis, the propofol
group still had a significantly greater incidence of poor
no. of patients 10 13 19
time from injury to protocol entry (hrs) 21  12 43  23 38  26
prognostic indicators and a lower therapy intensity, as
time on study medications (hrs) 128  63 69  97 70  54 measured by use of neuromuscular blocking agents, ben-
average propofol dose (µg/kg/min) 78  44 36  31 0 zodiazepines, pentobarbital, and CSF drainage. However,
average morphine dose (mg/hr) 1.6  1.0 1.1  0.4 10  6.7 mannitol and vasopressor use was more frequent in the
age (mean  SD in yrs)† 30  14 46  18 33  13 propofol group.
patients’ age 55 yrs† 10% 46.2% 10.5% The subgroup analysis of propofol patients demonstrat-
median GCS score (range)‡ 4.5 (3–7) 7 (3–9) 6 (3–10)
GCS score 3–5‡ 60% 23.1% 15.8%
ed that the high-dose subgroup and the morphine group
abnormal pupils (1 or both) 66.7% 58.3% 52.6% were well matched in terms of age, incidence of hypoten-
(6 of 9) sion, hypoxia, and abnormal pupils. However, the high-
median CT diagnoses (range) 3.5 (1–7) 3 (0–5) 4 (0–6) dose propofol patients were more severely injured, as
absent or compressed cisterns§ 80% 76.9% 57.9% determined by GCS scores, and tended to have more unfa-
diffuse swelling w/ or w/o shift 60% 38.5% 57.4% vorable CT findings compared with the morphine group.
evacuated hematomas (all) 30% 53.8% 47.4%
epidural 30% 7.7% 15.8%
Despite these poor prognostic factors, the high-dose pro-
subdural§ 10% 38.5% 21.1% pofol subgroup tended to have better ICP and CPP con-
intracerebral 10% 15.4% 26.3% trol and required less CSF drainage and benzodiazepine
multiple contusions§ 70% 23.1% 42.1% and pentobarbital therapy. However, vasopressor use was
favorable outcome greatest in the high-dose subgroup. Finally, the high-dose
GOS score 4 or 5 70% 38.5% 47.4% propofol subgroup had the highest favorable neurological
DRS score 0–6 80% 46.2% 57.9%
outcome rate compared with both the low-dose subgroup
* High-dose propofol criteria: minimum total dose 100 mg/kg, propofol and the morphine group.
infusion started within 48 hours of injury, infusion received for minimum
of 24 hours. The average propofol dose was significantly higher in the
high-dose subgroup when compared with the low-dose subgroup (p  Methodological Problems
0.05, Student’s t-test).
† The age effect was significant across groups according to ANOVA The original target number of 100 patients for this trial
(p  0.05). The percentage of patients older than 55 years was greater in was not reached because sufficient drug safety data were
the low-dose subgroup compared with the high-dose subgroup and the obtained with fewer patients than expected. The sample
morphine group (p  0.05). size was further reduced by patient exclusions deemed
‡ The median GCS score was lower in the high-dose subgroup compared
with the low-dose subgroup (p  0.05) and the morphine group (p = 0.07). necessary to maintain study validity. Specifically, a mini-
The percentage of patients with GCS scores 3 to 5 was greater in the high- mum of 12 hours on trial drugs and onset of therapy with-
dose subgroup (p  0.05 compared with the low-dose subgroup; p  0.002 in 96 hours of injury were thought to constitute reason-
compared with the morphine group). able, albeit not ideal, windows for inclusion. It would
§ The appearance of abnormal cisterns on CT scans was more common
in the high-dose subgroup compared with the morphine group (p = 0.1). have been preferable to have a longer minimum time on
Subdural hematomas were more common in the low-dose subgroup com- trial drugs and a narrower time window for trial entry. The
pared with the high-dose subgroup (p  0.05). Multiple contusions were issue of wide variability in total dose and duration of
more common in the high-dose subgroup compared with the low-dose sub- propofol was addressed by the high- and low-dose propo-
group (p  0.01) and the morphine group (p = 0.06).
 There was a higher incidence of favorable outcome in the high-dose
fol subgroup analysis. Another confounding problem was
subgroup compared with the low-dose subgroup (p  0.05 for both GOS that seven patients in the morphine group received vary-
and DRS scores) and the morphine group (p = 0.09 for GOS scores and ing amounts of 1% propofol before trial entry. Four of
p = 0.1 for DRS scores). these patients received such small amounts that they were
included for analysis, despite the obvious methodologi-
scores), and compared with 47.4% and 57.9% in the mor- cal problem this creates. Ideally, these four patients also
phine group, respectively (p = 0.09 for GOS and p = 0.1 would have been excluded. The potential bias of exclud-
for DRS). ing the eight patients lost to follow-up review was ad-
dressed by including them in the prognostic-indicator and
ICU data analysis. The variability in treatment regimens
Discussion across centers is also a methodological issue that is seen
frequently in such clinical studies. For example, the fact
Summary of Study Findings
that ventriculostomies were used in only 16 of 23 patients
Despite randomization in this pilot study, the propofol in the propofol group and in 16 of 19 patients in the mor-
and morphine groups were not evenly matched in prog- phine group may have created bias in the analysis of ther-
nostic indicators. Specifically, the propofol group had apy intensity for ICP and CPP control. The differences in
more patients who were older than 55 years of age, had an CSF drainage on therapy Days 2 and 3 remained signifi-
initial GCS score of 3 to 5, had early hypotension and/or cant, however, even with a relatively small number of pa-
hypoxia and compressed or absent basilar cisterns on the tients with ventriculostomies.
initial CT scan. Even with this preponderance of adverse A titratable morphine infusion was used for the “stan-
risk factors, ICP was lower on therapy Day 3, therapy in- dard therapy” group because opiate-based sedative regi-

1048 J. Neurosurg. / Volume 90 / June, 1999

6/99 issue pt. 1 5/11/99 1:35 PM Page 1049

Propofol in traumatic brain injury

mens are commonly used in intubated head-injured pa- TABLE 8

tients. However, it was somewhat surprising that the mor-
phine dose was so high, averaging 10 mg per hour, even
though the titration protocol allowed for relatively liberal
use of CSF drainage and therapy with mannitol and neu-
romuscular blocking agents. This fact may attest to the
relative ineffectiveness of opiates for achieving sedation
and ICP control. Their routine use in head-injured patients
is also of questionable wisdom, given that opiate antago-
nists have been shown in some spinal cord injury models
to be neuroprotective.71 These and other disadvantages
of an opiate-based sedation regimen are further discussed
later.
Previous Head Injury Studies Using Propofol
Relatively few studies of propofol use for posttraumat-
ic coma have been reported during the last decade. In one
of the first such studies, 10 severely head-injured patients
were given a 1% propofol infusion over 24 hours after
having been started initially on infusions of morphine (5–
10 mg/hour), midazolam (5–10 mg/hour), and vecuroni-
um (0.1 mg/kg/hour).18 Propofol infusion rates averaged
48 g/kg/minute, which is similar to the mean rate in the
present study. The ICP was unchanged but CPP took an
upward trend and was significantly higher at the end of 24
hours. In another 24-hour study, severely head injured pa-
tients were randomized to receive either a 1% propofol in-
fusion or a fentanyl infusion; both groups received pan-
curonium.43 In the 15 patients randomized to propofol
(dose range 17–50 g/kg/minute), CPP increased signifi-
cantly from 75 to 93 mm Hg. The ICP remained stable in
the low ICP subgroup and decreased in the high ICP sub-
group. In the 16 patients receiving fentanyl, ICP also
decreased in the high ICP subgroup, but CPP decreased
slightly. In a more recent study of intubated moderately
and severely head injured patients, nine patients were giv-
en a 1% propofol infusion (mean dose 55 g/kg/minute,
range 36–95 g/kg/minute) and six were sedated with in-
fusions of morphine (range 0–4 mg/hour) and midazolam
(range 0–5 mg/hour).58 Throughout a mean data collection
period lasting 40 hours, no differences were observed be-
tween groups for blood pressure, ICP, CPP, or outcome.
However, propofol was associated with a significant fall
in the arteriovenous difference in oxygen at 4 hours after
onset of propofol infusion. At 8 hours postinfusion this
difference disappeared. The findings from these studies
suggest that a propofol-based regimen is safe and as ef-
fective as a narcotic-based regimen for sedation and ICP
control. It is difficult, however, to draw further conclu-
sions from these reports.
Propofol as a Neuroprotectant
Given the significant disparity in injury severity be-
tween the two treatment groups in this study, both acute
and long-term outcome data suggest a propofol-based reg-
imen may be superior to an opiate-based regimen in head-
injured patients. Several neuroprotective actions of propo-
fol may help explain this favorable effect on neurological
recovery. These include potentiation of -aminobutyrateA
(GABAergic) inhibition28,29 and inhibition of N-methyl-
D-aspartate glutamate receptors and voltage-dependent
calcium channels.4,25,49 Propofol is also a potent antioxi-
Propofol subgroup analysis: therapy intensity for ICP
and CPP control*
Propofol Group
Morphine
Therapy High Dose Low Dose Group
drug treatment intensity
mannitol† 48.4% 37.8% 43.6%
NMB‡ 45.3% 13.5% 47.4%
benzodiazepines§ 3.1% 2.7% 12.8%
pentobarbital§ 1.6% 0 6.4%
vasopressors 84.4% 32.4% 61.5%
total no. of treatment days 64 37 78
CSF drainage by therapy day (mean ml  SD)**
Day 1 44  34 (10) 55  59 (6) 77  112 (15)
Day 2 88  83 (10) 55  54 (5) 191  102 (16)
Day 3 72  103 (9) 84  132 (3) 179  127 (13)
Day 4 57  68 (8) ID (1) 134  80 (8)
* Number of patients with CSF drainage is shown in parentheses.
Abbreviations: ID = insufficient data; NMB = neuromuscular blocking
agents.
† Mannitol use was less in the low-dose subgroup compared with the
high-dose subgroup (p = 0.06).
‡ Use of NMB was less in the low-dose subgroup compared with the
high-dose subgroup and the morphine group (p  0.01).
§ Use of benzodiazepine and pentobarbital was less in the low- and high-
dose subgroups compared with the morphine group (p  0.05, except
p = 0.06 for the high-dose subgroup compared with the morphine group for
pentobarbital).
 Use of vasopressor agents was less in the low-dose subgroup and the
morphine group compared with the high-dose subgroup (p  0.001).
** The CSF drainage was less in the high- and low-dose subgroups on
therapy Day 2 (p  0.05) and Day 3 (p = 0.11) as determined by ANOVA.
On therapy Day 4, CSF drainage was less in the high-dose subgroup com-
pared with the morphine group (p = 0.06, Student’s t-test).
dant and inhibitor of lipid peroxidation, unlike barbiturate
medications.1,26,27,32,46
Similar to barbiturate agents, etomidate, and benzodi-
azepines, propofol causes cerebral metabolic suppression
primarily by potentiating the GABAergic receptor-chan-
nel complex.28,29 Propofol-induced EEG burst suppression
results in a decrease in cerebral blood flow of 38 to 58%,
a decrease in the cerebral metabolic rate of oxygen of 22
to 43%, and a decrease in the cerebral metabolic rate of
glucose of 36 to 55%.3,14,22,30,51,65 This metabolic depressant
effect is thought to be a key factor in enhanced neuro-
logical recovery and reduced neuronal damage in experi-
mental ischemia and reperfusion models.33,69,72 The clinical
importance of other actions of propofol is unclear. Spe-
cifically, the effects on NMDA and calcium channels are
uncertain. However, given the role of excitotoxicity and
intracellular calcium influx in posttraumatic mitochon-
drial dysfunction, attenuation of this influx by propofol
may be beneficial.4,49,66 The addition of the calcium chela-
tor EDTA to the propofol formulation, albeit in relatively
small quantities, may further reduce the harmful effects of
calcium. Propofol’s potent free-radical scavenging effects
and ability to inhibit lipid peroxidation may also be rele-
vant in the clinical setting by enhancing cell membrane
viability.1,26,27,32,46
In the present study, the mean propofol dose was 55
g/kg/minute. Infusion rates typically resulting in an
EEG burst suppression range from 100 to 220 g/kg/min-
ute.30,56,64 Given that cerebral metabolic measurements

J. Neurosurg. / Volume 90 / June, 1999 1049

6/99 issue pt. 1 5/11/99 1:35 PM Page 1050
were not routinely performed in this trial, it is unknown
whether significant metabolic suppression was achieved
in these patients. Additionally, it is unclear if other neuro-
protective mechanisms, including effects on NMDA re-
ceptors, calcium channels, and lipid peroxidation are bio-
logically relevant when propofol is administered at these
levels. The favorable trends in ICP control and therapy
intensity in the propofol group and the high favorable out-
come rate in the high-dose propofol subgroup suggest that
moderately high doses, even below those needed to induce
EEG burst suppression, may enhance neurological re-
covery.
Advantages and Disadvantages of Propofol
In addition to its neuroprotective qualities, there are
several other properties that make propofol an attractive
sedative for head-injured patients. First, its short elimi-
nation halflife of less than an hour facilitates rapid ti-
tration to a desired clinical effect and permits frequent
neurological assessments.2,6,54,55 Second, the wide dose re-
sponse allows it to be used as a sedative or as a metabolic
suppressive and neuroprotective agent. In contrast, the
short-acting neuroprotectant, etomidate, is not suited for
prolonged use because of renal toxicity related to its pro-
pylene glycol vehicle35 and because it suppresses adreno-
cortical steroid production.19,44 Pentobarbital, the metabol-
ic suppressant that is traditionally used, has a halflife of 15
to 48 hours and typically requires a minimum of 48 hours
after drug stoppage before patients can be assessed neuro-
logically.6,24 Third, propofol reduces ICP as demonstrated
in previous reports43,45,53 and as suggested by this study,
although this effect has not been seen consistently.18,58 In
contrast, neither midazolam (Versed) nor the opiates mor-
phine, fentanyl, or sufentanil reliably decrease ICP in
severely head injured patients.34,50,68 In one report, both
fentanyl and sufentanil increased ICP in severely head
injured patients.57 Fourth, propofol is a potent anticonvul-
sant medication and is now often used for refractory sta-
tus epilepticus.8,15,70 Reports of propofol-induced seizure-
like activity are uncommon and the clinical significance of
such proconvulsant properties is controversial.48,60 Fifth,
transcranial Doppler studies indicate that high-dose pro-
pofol anesthesia does not impair CO2 reactivity or cerebral
pressure autoregulation in noninjured humans.42,59,62 This
property is particularly relevant in head-injured patients
who often have derangements in cerebral vasoreactivi-
ty.9,36,47 Sixth, in contrast with the benzodiazipines and opi-
ates, long-term propofol use does not result in addiction
or withdrawal phenomena.7 Increasing dose requirements,
however, may occur.10,11,20 Whether this problem is related
to tolerance or an increased rate of drug clearance remains
unclear.
Regarding hemodynamic effects, this study suggests, as
have previous reports, that propofol infusion is relatively
well tolerated compared with opiate infusions.18,43,58 In this
study, vasopressors were used for CPP maintenance on
65% of the therapy days in the propofol group compared
with 61% for the morphine group. The tendency toward
hypotension with propofol can be minimized if patients
have a normal intravascular volume before initiating pro-
pofol; the infusion begins at a rate of less than 20 g/kg/
1050
D. F. Kelly, et al.
minute and does not increase by more than 10 g/kg/min-
ute every 5 minutes.56 In the absence of vasopressors, a
propofol bolus followed by an infusion rate of greater than
50 to 100 g/kg/minute reliably causes a transient reduc-
tion in blood pressure of up to 25% of baseline values,
similar to that seen when using high-dose pentobarbital
administration.14,30,45 This rapid dosing method is not rec-
ommended for neurosurgical ICU patients, given the like-
lihood of hypotension and possible precipitous drops in
cerebral blood flow. Of note, both midazolam and the opi-
ates also decrease blood pressure and CPP when given in
bolus form to severely head injured patients.34,50
Elevated triglyceride levels from the lipid vehicle can
occur with prolonged high-dose propofol infusion and are
more likely to occur in older patients. This side effect is
shortlived once the infusion is stopped or diminished.7,55
The new 2% formulation should reduce the lipid load by
50% compared with an equipotent drug dose with the 1%
formulation. Regarding risk of infection due to the lipid
vehicle, the addition of EDTA appears to be safe from an
infection standpoint. Although there was a trend toward a
higher pneumonia rate in patients in the propofol group,
these pneumonias were not attributed to drug administra-
tion per se. Further studies confirming the safety of this
formulation from an infection standpoint are nearing com-
pletion.
Finally, regarding the issue of cost, propofol is clearly
more expensive than morphine on a per unit basis. How-
ever, as this study confirms, most intubated head-injured
patients are not adequately sedated with morphine alone.
In contrast, a propofol-based sedation regimen appears to
reduce the need for concomitant narcotics and muscle
relaxants significantly. Given the lipid vehicle used with
propofol, the daily need for lipid nutritional supplements
is also decreased. Additionally, as other studies have
shown, patients receiving propofol emerge from sedation
more rapidly compared with those receiving benzodiaze-
pines, which may, in turn, shorten mechanical ventilation
time and the number of days in the ICU.7,21,54,55 Overall,
these factors appear to make a propofol sedation regimen
cost-effective; however, further study of this issue is war-
ranted.
Conclusions
In this preliminary study, we have demonstrated the
potential advantages of a propofol-based sedation and
ICP control regimen in intubated head-injured patients.
The similar favorable outcome rate in the two treatment
groups and the lower therapy intensity for ICP control,
despite a preponderance of poor prognostic factors in the
propofol group, suggests that a propofol-based regimen
with low-dose opiate analgesia is a safe and, possibly,
desirable alternative to an opiate-based regimen in head-
injured patients. Further studies are clearly needed to
define the neuroprotective dose range of propofol in both
experimental and human head injury. A randomized phase
II trial of early high-dose propofol during the period of
acute brain swelling in severely head injured patients is
currently being planned, based on the findings of this
study and in light of the favorable results of high-dose bar-
biturate studies performed over a decade ago.16,47
J. Neurosurg. / Volume 90 / June, 1999
6/99 issue pt. 1 5/11/99 1:35 PM Page 1051
Propofol in traumatic brain injury
References
1. Aarts L, van der Hee R, Dekker I, et al: The widely used anes-
thetic agent propofol can replace (-tocopherol as an antioxi-
dant. FEBS Lett 357:83–85, 1995
2. Albanese J, Martin C, Lacarelle B, et al: Pharmacokinetics of
long-term propofol infusion used for sedation in ICU patients.
Anaesthesiology 73:214–217, 1990
3. Alkire MT, Haier RJ, Barker SJ, et al: Cerebral metabolism dur-
ing propofol anesthesia in humans studied with positron emis-
sion tomography. Anesthesiology 82:393–403, 1995
4. Amorim P, Chambers G, Cottrell J, et al: Propofol reduces neu-
ronal transmission damage and attenuates the changes in calci-
um, potassium, and sodium during hyperthermic anoxia in the
rat hippocampal slice. Anesthesiology 83:1254–1265, 1995
5. Association for the Advancement of Automotive Medicine:
The Abbreviated Injury Scale. 1990 Revision. Des Plaines,
Ill: Association for the Advancement of Automotive Medicine,
1990
6. Bailie GR, Cockshott ID, Douglas EJ, et al: Pharmacokinetics
of propofol during and after long term continuous infusion for
maintenance of sedation in ICU patients. Br J Anaesth 68:
486–491, 1992
7. Barr J: Propofol: a new drug for sedation in the intensive care
unit. Int Anesthesiol Clin 33:131–154, 1995
8. Borgeat A: Propofol: pro- or anticonvulsant? Eur J Anaesthiol
Suppl 15:17–20, 1997
9. Bouma GJ, Muizelaar JP, Bandoh K, et al: Blood pressure and
intracranial pressure-volume dynamics in severe head injury:
relationship with cerebral blood flow. J Neurosurg 77:15–19,
1992
10. Boyle WA, Shear JM, White PF, et al: Tolerance and hyper-
lipemia during long-term sedation with propofol. Anesthesiol-
ogy 73:A245, 1990 (Abstract)
11. Buckley PM: Propofol in patients needing long-term sedation in
intensive care: an assessment of the development of tolerance.
A pilot study. Intensive Care Med 23:969–974, 1997
12. Bullock R, Chesnut RM, Clifton G, et al: Guidelines for the
management of severe head injury. Brain Trauma Foundation.
Eur J Emerg Med 3:109–127, 1996
13. Chesnut RM, Marshall LF, Marshall SB: Medical management
of intracranial pressure, in Cooper PR (ed): Head Injury, ed 3.
Baltimore: Williams & Wilkins, 1993, pp 225–246
14. Dam M, Ori C, Pizzolato G, et al: The effects of propofol anes-
thesia on local cerebral glucose utilization in the rat. Anesthesi-
ology 73:499–505, 1990
15. De Riu PL, Petruzzi V, Testa C, et al: Propofol anticonvulsant
activity in experimental epileptic status. Br J Anaesth 69:
177–181, 1992
16. Eisenberg HM, Frankowski RF, Contant CF, et al: High-dose
barbiturate control of elevated intracranial pressure in patients
with severe head injury. J Neurosurg 69:15–23, 1988
17. Eisenberg HM, Gary HE Jr, Aldrich EF, et al: Initial CT find-
ings in 753 patients with severe head injury. A report from the
NIH Traumatic Coma Data Bank. J Neurosurg 73:688–698,
1990
18. Farling PA, Johnston JR, Coppel DL: Propofol infusion for
sedation of patients with head injury in intensive care. A pre-
liminary report. Anaesthesia 44:222–226, 1989
19. Fellows IW, Bastow MD, Byrne AJ, et al: Adrenocortical sup-
pression in multiply injured patients: a complication of etomi-
date treatment. Br Med J 287:1835–1837, 1983
20. Foster SJ, Buckley PM: A retrospective review of two years’
experience with propofol in one intensive care unit. J Drug
Dev 2 (Suppl 2):73–74, 1989
21. Fulton B, Sorkin EM: Propofol. An overview of its pharmacol-
ogy and a review of its clinical efficacy in intensive care seda-
tion. Drugs 50:636–657, 1995
22. Gelb AW, Wilkes P, Henderson SM, et al: Cerebral blood flow
J. Neurosurg. / Volume 90 / June, 1999
and metabolism during prolonged propofol anaesthesia. Can J
Anaesth 40:A68, 1993 (Abstract)
23. Gennarelli TA, Spielman GM, Langfitt TW, et al: Influence of
the type of intracranial lesion on outcome from severe head in-
jury. A multicenter study using a new classification system. J
Neurosurg 56:26–32, 1982
24. Gilman AG, Goodman LS, Gilman A (eds): Goodman and
Gilman’s The Pharmacological Basis of Therapeutics. New
York: Macmillan, 1980, p 350
25. Hans P, Bonhomme V, Collette J, et al: Propofol protects cul-
tured rat hippocampal neurons against N-methyl-D-aspartate
receptor-mediated glutamate toxicity. J Neurosurg Anesthesi-
ol 6:249–253, 1994
26. Hans P, Deby C, Deby-Dupont G, et al: Effect of propofol on in
vitro lipid peroxidation induced by different free radical gener-
ating systems: a comparison with vitamin E. J Neurosurg An-
esthesiol 8:154–158, 1996
27. Hans P, Deby-Dupont G, Deby C, et al: Increase in antioxidant
capacity of plasma during propofol anesthesia. J Neurosurg
Anesth 9:234–236, 1997
28. Hara M, Kai Y, Ikemoto Y: Enhancement by propofol of the -
aminobutyric acidA response in dissociated hippocampal py-
ramidal neurons of the rat. Anesthesiology 81:988–994, 1994
29. Hollrigel GS, Toth K, Soltesz I: Neuroprotection by propofol in
acute mechanical injury: role of GABAergic inhibition. J Neu-
rophysiol 76:2412–2422, 1996
30. Illievich UM, Petricek W, Schramm W, et al: Electroenceph-
alographic burst suppression by propofol infusion in humans:
hemodynamic consequences. Anesth Analg 77:155–160, 1993
31. Jennett B, Bond M: Assessment of outcome after severe brain
damage. A practical scale. Lancet 1:480–484, 1975
32. Kahraman S, Kilinc K, Dal D, et al: Propofol attenuates forma-
tion of lipid peroxides in tourniquet-induced ischaemia-reper-
fusion injury. Br J Anaesth 78:279–281, 1997
33. Kochs E, Hoffman WE, Werner C, et al: The effects of propo-
fol on brain electrical activity, neurologic outcome, and neu-
ronal damage following incomplete ischemia in rats. An-
esthesiology 76:245–252, 1992
34. Lauer KK, Connolly LA, Schmeling WT, et al: Opioid sedation
does not alter intracranial pressure in head injured patients. Can
J Anaesth 44:929–933, 1997
35. Levy ML, Aranda M, Zelman V, et al: Propylene glycol toxici-
ty following continuous etomidate infusion for the control of
refractory cerebral edema. Neurosurgery 37:363–371, 1995
36. Marion DW, Bouma GJ: The use of stable xenon-enhanced
computed tomographic studies of cerebral blood flow to define
changes in cerebral carbon dioxide vasoresponsivity caused by
a severe head injury. Neurosurgery 29:869–873, 1991
37. Marmarou A, Ward JD, Young HF, et al: Impact of ICP insta-
bility and hypotension on outcome in patients with severe head
injury. J Neurosurg 75 (Suppl):S59–S66, 1991
38. Marmarou A, Ward JD, Young HF, et al: NINDS Traumatic
Coma Data Bank: intracranial pressure monitoring methodolo-
gy. J Neurosurg 75 (Suppl):S21–S27, 1991
39. Marshall LF, Gautille T, Klauber MR, et al: The outcome of
severe closed head injury. J Neurosurg 75 (Suppl):S28–S36,
1991
40. Marshall LF, Marshall SB, Klauber MR, et al: A new classifi-
cation of head injury based on computerized tomography. J
Neurosurg 75 (Suppl):S14–S20, 1991
41. Matta B, Menon D: Severe head injury in the United Kingdom
and Ireland: a survey of practice and implications for manage-
ment. Crit Care Med 24:1743–1748, 1996
42. Matta BF, Lam AM, Strebel S, et al: Cerebral pressure autoreg-
ulation and carbon dioxide reactivity during propofol-induced
EEG suppression. Br J Anaesth 74:159–163, 1995
43. Mergaert C, Herregods L, Rolly G, et al: The effect of a 24-h
propofol or fentanyl sedation on intracranial pressure. Eur J
Anaesth 8:324–325, 1991 (Abstract)
1051
6/99 issue pt. 1 5/11/99 1:35 PM Page 1052
44. Moore RA, Allen MC, Wood PJ, et al: Peri-operative endocrine
effects of etomidate. Anaesthesia 40:124–130, 1985
45. Moss E, Price DJ: Effect of propofol on brain retraction
pressure and cerebral perfusion pressure. Br J Anaesth 65:
823–825, 1990
46. Murphy PG, Myers DS, Davies MJ, et al: The antioxidant
potential of propofol (2,6-diisopropylphenol). Br J Anaesth
68:613–618, 1992
47. Nordström CH, Messeter K, Sundbärg G, et al: Cerebral blood
flow, vasoreactivity, and oxygen consumption during barbitu-
rate therapy in severe traumatic brain lesions. J Neurosurg 68:
424–431, 1988
48. Nowack WJ, Jordan R: Propofol, seizures and generalized
paroxysmal fast activity in the EEG. Clin Electroencephalogr
25:110–114, 1994
49. Orser BA, Bertlik M, Wang LY, et al: Inhibition by propofol
(2,6 di-isopropylphenol) of the N-methyl-D-aspartate subtype
of glutamate receptor in cultured hippocampal neurones. Br J
Pharmacol 116:1761–1768, 1995
50. Papazian L, Albanese J, Thirion X, et al: Effect of bolus doses
of midazolam on intracranial pressure and cerebral perfusion
pressure in patients with severe head injury. Br J Anaesth 71:
267–271, 1993
51. Ramani R, Todd MM, Warner DS: A dose–response study of
the influence of propofol on cerebral blood flow, metabolism
and the electroencephalogram in the rabbit. J Neurosurg Anes-
thesiol 4:110–119, 1992
52. Rappaport M, Hall KM, Hopkins K, et al: Disability rating scale
for severe head trauma: coma to community. Arch Phys Med
Rehabil 63:118–123, 1982
53. Ravussin PA, Tempelhoff R, Modica PA, et al: Propofol versus
thiopental-isoflurane for neurosurgical anesthesia: comparisons
of hemodynamics, CSF pressure and recovery. J Neurosurg
Anesthesiol 3:85–95, 1991
54. Ronan KP, Gallagher TJ, George B, et al: Comparison of propo-
fol and midazolam for sedation in intensive care unit patients.
Crit Care Med 23:286–293, 1995
55. Sanchez-Izquierdo-Riera JA, Caballero-Cubedo RE, Perez-
Vela JL, et al: Propofol versus midazolam: safety and efficacy
for sedating the severe trauma patient. Anesth Analg 86:
1219–1224, 1998
56. Shafer SL: Advances in propofol pharmacokinetics and phar-
macodynamics. J Clin Anesth 5 (Suppl 1):14S–21S, 1993
57. Sperry RJ, Bailey PL, Reichman MV, et al: Fentanyl and sufen-
tanil increase intracranial pressure in head trauma patients. An-
esthesiology 77:416–420, 1992
58. Stewart L, Bullock R, Rafferty C, et al: Propofol sedation in
severe head injury fails to control high ICP, but reduces brain
metabolism. Acta Neurochir 60:544–546, 1994
59. Strebel S, Lam AM, Matta B, et al: Dynamic and static cerebral
1052
D. F. Kelly, et al.
autoregulation during isoflurane, desflurane, and propofol anes-
thesia. Anesthesiology 83 (Suppl):66–76, 1995
60. Sutherland MJ, Burt P: Propofol and seizures. Anaesth Inten-
sive Care 22:733–737, 1994
61. Teasdale G, Jennett B: Assessment of coma and impaired con-
sciousness. A practical scale. Lancet 2:81–84, 1974
62. Tiecks FP, Lam AM, Aaslid R, et al: Comparison of static and
dynamic cerebral autoregulation measurements. Stroke 26:
1014–1019, 1995
63. Toutant SM, Klauber MR, Marshall LF, et al: Absent or com-
pressed basal cisterns on first CT scan: ominous predictors of
outcome in severe head injury. J Neurosurg 61:691–694, 1984
64. Vandesteene A, Trempont V, Engelman E, et al: Effect of pro-
pofol on cerebral blood flow and metabolism in man. Anaes-
thesia 43 (Suppl):42–43, 1988
65. Van Hemelrijck J, Fitch W, Mattheussen M, et al: Effect of
propofol on cerebral circulation and autoregulation in the ba-
boon. Anesth Analg 71:49–54, 1990
66. Verweij BH, Muizelaar JP, Vinas FC, et al: Mitochondrial dys-
function after experimental and human brain injury and its pos-
sible reversal with a selective N-type calcium channel antago-
nist (SNX-111). Neurol Res 19:334–339, 1997
67. Vollmer DG, Torner JC, Jane JA, et al: Age and outcome fol-
lowing traumatic coma: why do older patients fare worse? J
Neurosurg 75 (Suppl):S37–S49, 1991
68. Weinstabl C, Mayer N, Richling B, et al: Effect of sufentanil on
intracranial pressure in neurosurgical patients. Anaesthesia
46:837–840, 1991
69. Weir DL, Goodchild CS, Graham DI: Propofol: effects on
indices of cerebral ishemia. J Neurosurg Anesthesiol 1:
284–289, 1989
70. Wood PR, Browne GPR, Pugh S: Propofol infusion for the
treatment of status epilepticus. Lancet 1:480–481, 1988 (Let-
ter)
71. Young W: Spinal cord injury pathophysiology and therapy, in
Narayan RK, Wilberger JE Jr, Povlishock JT (eds): Neuro-
trauma. McGraw-Hill: New York, 1996, pp 1075–1093
72. Young Y, Menon DK, Tisavipat N, et al: Propofol neuroprotec-
tion in a rat model of ischaemia reperfusion injury. Eur J An-
aesthesiol 14:320–329, 1997
Manuscript received October 21, 1998.
Accepted in final form February 1, 1999.
This study was supported by Zeneca Pharmaceuticals, Wilming-
ton, Delaware.
Address reprint requests to: Daniel F. Kelly, M.D., Division of
Neurosurgery, Room 18-218A NPI, Box 957039, University of
California, Los Angeles, California 90095–7039. email: dfkelly@
ucla.edu.
J. Neurosurg. / Volume 90 / June, 1999