Can J Anesth/J Can Anesth
DOI 10.1007/s12630-014-0118-9
EVIDENCE-BASED CLINICAL UPDATE
Comparison of propofol and volatile agents for maintenance of
anesthesia during elective craniotomy procedures: systematic
review and meta-analysis
Comparaison entre propofol et agents volatils pour le maintien de
l’anesthésie pendant les interventions de craniotomie non
urgentes : revue méthodique et méta-analyse
Jason Chui, MBChB • Ramamani Mariappan, MD
Jigesh Mehta, MD • Pirjo Manninen, MD •
Lashmi Venkatraghavan, MD
Received: 4 September 2013 / Accepted: 16 January 2014
Ó Canadian Anesthesiologists’ Society 2014
Abstract
Background Both propofol and volatile anesthetics are
commonly used for maintenance of anesthesia in patients
undergoing neurosurgical procedures. The effects of these
two classes of drugs on cerebral hemodynamics have been
compared in many clinical trials The objectives of this
review were to evaluate the cerebral hemodynamic effects,
operative conditions, recovery profiles, postoperative
complications, and neurological outcomes of propofol-
based vs volatile-based anesthesia for craniotomy.
Methods MEDLINEÒ, EMBASETM, Cochrane, and other
relevant databases were searched for randomized controlled
trials that compared propofol-maintained anesthesia with
volatile-maintained anesthesia in adult patients undergoing
elective craniotomy. The primary outcome measure was the
intraoperative brain relaxation score. Secondary outcome
This article is accompanied by an editorial. Please see Can J Anesth
2014; 61: this issue.
Author contributions Jason Chui, Ramamani Mariappan, Mehta
Jigesh, Pirjo Manninen, and Lashmi Venkatraghavan helped conduct
the study and write the manuscript.
Electronic supplementary material The online version of this
article (doi:10.1007/s12630-014-0118-9) contains supplementary
material, which is available to authorized users.
J. Chui, MBChB
R. Mariappan, MD
J. Mehta, MD

P. Manninen, MD
L. Venkatraghavan, MD (&)
Department of Anesthesia, Toronto Western Hospital, University
Health Network, University of Toronto, Toronto, ON, Canada
e-mail: lashmi.venkatraghavan@uhn.on.ca
•
measures included intraoperative cerebral hemodynamics
(intracranial pressure [ICP], cerebral perfusion pressure
[CPP]), cardiovascular changes, recovery profiles,
postoperative complications, and clinical outcomes
(neurological morbidity, mortality, quality of life). A meta-
analysis was conducted using a random effects model to
compare the outcomes of the two anesthetic techniques.
Results Fourteen studies (1,819 patients) met inclusion
criteria and were analyzed. Brain relaxation scores were
similar between the two groups after dural opening;
however, ICP was lower (weighted mean difference of
-5.2 mmHg; 95% confidence interval -6.81 to -3.6) and
CPP was higher (weighted mean difference of 16.3 mmHg;
95% confidence interval 12.2 to 20.46) in patients
receiving propofol-maintained anesthesia. Postoperative
complications and recovery profiles were similar between
the two groups, except for postoperative nausea and
vomiting being less frequent with propofol-maintained
anesthesia. There were inadequate data to perform a meta-
analysis on clinical outcome.
Conclusion Propofol-maintained and volatile-maintained
anesthesia were associated with similar brain relaxation
scores, although mean ICP values were lower and CPP
values higher with propofol-maintained anesthesia. There
are inadequate data to compare clinically significant
outcomes such as neurological morbidity or mortality.
Résumé
Contexte Le propofol et les anesthésiques volatils sont des
agents fréquemment utilisés pour maintenir l’anesthésie chez
les patients subissant des interventions neurochirurgicales.
123

Les effets de ces deux classes de médicaments sur
l’hémodynamie cérébrale ont été comparés dans de
nombreuses études cliniques. Les objectifs de cette revue
étaient d’évaluer les effets sur l’hémodynamie cérébrale, les
conditions opératoires, les profils de rétablissement, les
complications postopératoires et les pronostics
neurologiques de l’anesthésie au propofol comparativement
aux agents volatils lors de craniotomie.
Méthode Nous avons effectué une recherche dans les bases
de données MEDLINEÒ, EMBASETM, Cochrane et les autres
bases de données pertinentes pour en extraire les études
randomisées contrôlées comparant le propofol aux agents
volatils pour le maintien de l’anesthésie chez les patients
adultes subissant une craniotomie non urgente. Le critère
d’évaluation principal était le score de relaxation cérébrale
peropératoire. Les critères d’évaluation secondaires
comprenaient l’hémodynamie cérébrale peropératoire
(pression intracrânienne [PIC], pression de perfusion
cérébrale [PPC]), les changements cardiovasculaires, les
profils de rétablissement, les complications postopératoires et
les devenirs cliniques (morbidité neurologique, mortalité,
qualité de vie). Une méta-analyse a été réalisée en utilisant un
modèle à effets aléatoires afin de comparer les pronostics des
deux techniques d’anesthésie.
Résultats Quatorze études (1819 patients) respectaient les
critères d’inclusion et ont été analysées. Les scores de
relaxation cérébrale étaient semblables dans les deux
groupes après ouverture de la dure-mère; toutefois, la PIC
était plus basse (différence moyenne pondérée de -5,2
mmHg; intervalle de confiance 95 % -6,81 à -3,6) et la
PPC plus élevée (différence moyenne pondérée de 16,3
mmHg; intervalle de confiance 95 % 12,2 à 20,46) chez les
patients dont l’anesthésie était maintenue avec du propofol.
Les complications postopératoires et les profils de
rétablissement étaient semblables dans les deux groupes, à
l’exception des nausées et vomissements postopératoires,
moins fréquents lorsque l’anesthésie était maintenue au
propofol. Les données étaient insuffisantes pour faire une
méta-analyse des devenirs cliniques.
Conclusion Le propofol et les agents volatils comme
maintien de l’anesthésie sont associés à des scores de
relaxation cérébrale semblables, bien que les valeurs de
PIC moyennes aient été plus basses et les valeurs de PPC
moyennes plus élevées avec une anesthésie maintenue au
propofol. Les données sont insuffisantes pour comparer les
devenirs significatifs d’un point de vue clinique tels que la
morbidité neurologique ou la mortalité.
The well-accepted goals for the anesthetic management of
patients undergoing neurosurgical procedures include the
delivery of smooth and hemodynamically stable anesthesia
123
J. Chui et al.
with provision of good operating conditions and a smooth but
rapid emergence that allows for early neurological
assessment. The ideal anesthetic agents should maintain
cerebral perfusion pressure (CPP), preserve cerebral
autoregulation and carbon dioxide reactivity, confer
neuroprotection, and not interfere with neurophysiological
monitoring. Numerous agents, both intravenous and
inhalational, as well as different techniques have been used
and studied.1,2
Over the past three decades, multiple studies have
compared propofol-maintained and volatile-based
anesthesia in patients undergoing elective craniotomy.1–8A
Most of these studies have looked at either perioperative
cerebral hemodynamic effects or recovery profiles. Few
studies have evaluated clinically significant outcomes such
as neurological morbidity, mortality, or quality of life. In
addition, an intrinsic problem of research in neuroanesthesia
is the small sample size of many studies. Thus far, a lack of
systematic reviews have been undertaken to compare
propofol-maintained anesthesia vs volatile-maintained
anesthesia for neurosurgical procedures. The objectives of
this review were to evaluate the cerebral hemodynamic
effects, operative conditions, recovery profiles,
postoperative complications, and neurological outcomes of
propofol-based vs volatile-based anesthesia for craniotomy.
Methods
This systematic review was carried out using the
recommended guidelines provided by the Cochrane
Handbook for Systematic Reviews of InterventionsB and
reported according to the PRISMA 2009 checklist.C
Search strategy
A professional librarian searched the databases, MEDLINEÒ
(from 1946 to November 10, 2013), MEDLINE In-Process &
other Non-Indexed Citations (Ovid) (November 10, 2013),
EMBASETM (from 1974 to November 10, 2013), Cochrane
Central Register of Controlled Trials (CENTRAL) in the
Cochrane Library (November 10, 2013), and Cochrane
Database of Systematic Reviews (CDSR) in the Cochrane
Library (November 10, 2013), and the procedure was reviewed
A
Ayrian E, Zelman V. Total intravenous anesthesia: Advantages for
intracranial surgery - Commentary. Neurosurgery 2007; 61(5 Suppl.
2): ONS377.
B
Higgins JP, Green S. Cochrane Handbook for Systematic Reviews
of Interventions. Version 5.1.0., March 2011. Available from URL:
http://handbook.cochrane.org (accessed (August 2013)).
C
Preferred Reporting Items for Systematic Reviews and Meta-
Analyses; Available from URL: http://www.prisma-statement.org
(accessed (August 2013)).
Propofol versus volatile agents for elective craniotomy
by two authors (L.V., J.M.). Searches were conducted using
three different components 1) terms related to inhalational
anesthetics 2) terms related to intravenous anesthetics, and 3)
terms related to craniotomy. The following MeSH terms were
used in searching MEDLINE (Appendix 1; available as
Electronic Supplementary Material) and The Cochrane
Register of Controlled Trials: ‘‘Anesthetics, Inhalation’’,
‘‘Enflurane’’, ‘‘Ethyl Ether’’, ‘‘Halothane’’, ‘‘Isoflurane’’,
‘‘Methoxyflurane’’, ‘‘Nitrous Oxide’’, ‘‘Trichloroethylene’’,
‘‘Xenon’’, ‘‘Sevoflurane’’, ‘‘Desflurane’’, ‘‘Intravenous
anesthetics’’, ‘‘Total intravenous anaesthesia’’, ‘‘Alfentanil’’,
‘‘Chloralose’’, ‘‘Diazepam’’, ‘‘Etomidate’’, ‘‘Fentanyl’’,
‘‘Methohexital’’, ‘‘Midazolam’’, ‘‘Propanidid’’, ‘‘Propofol’’,
‘‘Sodium Oxybate’’, ‘‘Sufentanil’’, ‘‘Thiamylal’’,
‘‘Thiopental’’, ‘‘Urethane’’, ‘‘lorazepam’’, ‘‘meperidine’’,
‘‘remifentanil’’, ‘‘ketamine’’, ‘‘droperidol’’, ‘‘Craniotomy’’,
‘‘volatile induction/maintenance anaesthesia’’, ‘‘intracranial
surgery’’, ‘‘intracranial operation’’, ‘‘Supratentorial surgery’’,
‘‘Neurosurgical Procedures’’, ‘‘Neurosurgery’’, ‘‘Brain
surgery’’, ‘‘Brain resection’’, and ‘‘Neurosurgery’’.
Corresponding EMTREE terms were used in searching
EMBASE (Appendix 2; available as Electronic
Supplementary Material). Randomized controlled trials
(RCT) was not used as a search term. A search of trial
registries and a manual search of the reference lists from the
selected articles were conducted to identify additional trials.
The searches were restricted to the English language and human
studies.
Selection criteria
Two independent reviewers (J.C., R.M.) evaluated the
search results and identified the eligible studies for possible
inclusion using predefined selection criteria. Adult patients
with both supratentorial and infratentorial pathologies
undergoing elective neurosurgical procedures were
included in the review. There was no restriction on the
type of neurosurgical pathology, but patients with
traumatic brain injury were excluded. Only randomized
and quasi-randomized controlled trials that compared
propofol-maintained with volatile-maintained anesthesia
were included. Propofol-maintained anesthesia was defined
as using propofol infusion, either by manual controlled or
targeted controlled infusion (TCI), for maintenance of
anesthesia. Volatile-maintained anesthesia included
sevoflurane, desflurane, or isoflurane, with or without
nitrous oxide. The primary outcome of this systemic review
was the intraoperative brain relaxation score. Secondary
outcomes included other intraoperative cerebral
hemodynamics (intracranial pressure [ICP] and cerebral
perfusion pressure), intraoperative hemodynamic events,
recovery profiles, postoperative complications, and
neurological outcomes. Studies reporting any of the
abovementioned outcome variables were included for
evaluation. Any disagreements in selection between the
two authors were resolved by discussion or by consulting
with the senior author (L.V.).
Data extraction and quality assessment
The data were extracted by J.C. and R.M. individually and
validated by the senior author (L.V.) and by double data
entry. Details of the study population, interventions, and
outcomes were extracted using a standardized electronic
data extraction form. When necessary, authors of the
selected articles were contacted to obtain missing
information for quantitative analysis. Two reviewers
(J.C., R.M.) graded the eligible studies independently
using the Cochrane risk of bias assessment scale. The
Cochrane risk of bias assessment scale is a domain-based
evaluation tool developed by Cochrane Collaboration. It
assesses seven different domains of bias in the primary
studies: random sequence generation, adequacy of
concealment of allocation, blinding of participants and
healthcare providers, blinding of outcome assessors,
incomplete outcome data, risk of selective reporting bias,
and other sources of bias.C
Outcome measurements
The primary outcome of this review was the brain
relaxation score. The brain relaxation score is a reflection
of the degree of brain swelling as evaluated by the surgeon
after opening the dura and is usually reported on a four-
point scale (slack brain, mild, moderate, and severe brain
herniation). Secondary outcomes included other
intraoperative cerebral hemodynamics (ICP and CPP),
intraoperative hemodynamic events (hypotension,
hypertension, tachycardia, bradycardia), recovery profiles
(time to eye opening, extubation, orientation, obeying
commands, and an Aldrete score [ 9), postoperative
complications (nausea, vomiting, seizures, shivering,
agitation, pain), and clinical outcomes (postoperative
neurological morbidity, mortality, and quality of life).
There were many variations with regard to timing and the
method of ICP measurement. Only studies using either
epidural or subdural ICP measurement at the time of the
first burr hole in the craniotomy were included in this
review. Similarly, only studies where CPP was calculated
with the formula using mean arterial pressure (MAP),
(CPP = MAP-ICP), at the time of the first burr hole in the
craniotomy were included in the analysis. Outcome
variables reported in different scales or using different
measuring methods or different timing were not used for
analysis.
123

Fig. 1 Study flowchart
Statistical analysis
The meta-analysis was performed using Review Manager,
version 5.1 software (The Cochrane Collaboration, Oxford,
UK) with a random effects model. Considering the clinical
heterogeneity and the potential variations in effect size
amongst the studies, a random effects model was chosen
instead of a fixed effects model. The results are presented
as a risk ratio (RR) for dichotomous data and as mean
difference (MD) for continuous data with corresponding
95% confidence intervals. The percentage of variability
across studies attributable to heterogeneity rather than
chance was estimated using the I2 statistic. The strategies
used to deal with heterogeneity in this systemic review
include i) conducting a subgroup analysis to explore the
heterogeneity in the use of anesthetic agents (propofol
infusion techniques, different types/combinations of
volatile agents, and types of opioids used) and ii)
123
J. Chui et al.
performing the sensitivity analysis according to the
methodological quality of the primary studies.
Publication bias was assessed using Begg’s funnel plots.
Results
Characteristics of included trials
The search strategy resulted in an initial yield of 1,599
citations. Thirty-six potential articles were selected after
thorough examination of titles and abstracts, and further
examination led to the exclusion of 19 studies from
analysis, as shown in Fig. 1. Finally, 14 studies
(n = 1,819) were included in the analysis.9–22 The
characteristics of the study populations are summarized
in Table 1. The sample size of these studies varies from
40-411 patients. All the included studies were RCTs that
 Table 1 Characteristics of included studies

ID Study Country No of Sample Mean % ASA Specification Propofol-based Volatile-based Ref

Centres Size age (SD) Male of lesions
I M O I M O

1 G. Citerio et al. 2012 Italy 14 411 54.8 (13.3) 51 NS Supratentorial tumours P P: 10/8/6 mg kg-1 hr-1 R P Sevo: 0.75-1.25 MAC R/F 12
and others
2 G. Banevicius et al. 2010 Lithuania 1 135 52.2 (14.2) 33 I-III Supratentorial tumours P P: 1.5-8.5 mg
kg-1 hr-1 F P Sevo: 0.5-1.8% F 13
3 E. Lauta et al. 2010 Italy 3 314 55.6 (1.1) 43.6 I-III Supratentorial tumours T P: 10/8/6 mg
kg-1
hr-1 R P Sevo: 0.7-2% R 14
and others
Propofol versus volatile agents for elective craniotomy

4 V. Bonhomme et al. 2009 Belgium 1 61 53 (17.5) 40 I-III Not mentioned P P: 2-3.5 mg
mL-1 (TCI) R P Sevo: 1-2.5% R 15
5 G. Magni et al. 2007 Italy 1 168 52.8 (14.9) 39.5 I-III Tumour, Vascular P P: 10/8/6 mg
kg-1
hr-1 R P Sevo: 1.5-2% F 16
6 J.R. Sneyd et al. 2005 UK 1 50 55.5 (12.5) 42 NS Tumour, Vascular P P: 1-2 mg
mL-1 (TCI) R P Sevo: 1.3-1.8% R 17
7 G Magni et al. 2005 Italy 1 120 52.9 (14.9) 53.3 I-III Supratentorial tumours P P: 10/8/6 mg
kg-1
hr-1 R P Sevo: 1.5-2% R 18
and others
8 KD Petersen et al. 2003 Denmark 1 117 54.4 (11.7) 47.8 NS Supratentorial tumours P P: 10/8/6 mg
kg-1
hr-1 F P Sevo or Iso: \1.5 MAC F 19
9 P. Talke et al. 2002 USA 1 40 46.5 (14.5) NS II-IV Supratentorial tumours P P: 50-200 lg
kg-1
min-1 F T Iso 0.55% with 70% N2O F 20
and others with 70% N2O
10 W. Ittichaikulthol et al. 1997 Thailand 1 60 43.8 (12.6) 45 I-II Tumour, Vascular P P: 2-12 mg
kg-1
hr-1 - P Iso with 66% N20 - 21
with 66% N2O
11 MM Todd et al. 1993 USA 1 80 51 (3.5) 45 II-III Supratentorial tumours P P: 25-200 lg
kg-1
min-1 F T Iso with 60% N2O F 22
and others
12 P. Ravussin 1991 USA 1 60 51.5 (2.7) NS I-III Tumour, Vascular and P P: 40-300 lg
kg-1
min-1 F T Iso 0.5-1.5% with 50% N2O F 23
Others
13 LD Mishra et al. 2008 India 1 120 34.79 (14.4) 68.3 I-II Not Mentioned P P: 65-190 lg
kg-1
min-1 - T Iso 0.2-2% with 66% N2O - 24
with 66% N2O
14 Sankari Santra et al. 2009 India 1 68 51.5 (13) 55.9 I-II Supratentorial tumours P P: 6-10 mg
kg-1
hr-1 F Iso, 50% N2O NS F 25
-1 -1 -1 -1 -1
ASA = American Society of Anesthesiologists; I = induction agent; M = maintenance regimen; O = opioid; P = propofol; ‘‘P = 10/8/6 mg
kg
hr ’’= propofol infused at 10 mg
kg
hr for 10 min, reduced to 8 mg
kg
hr-1
for 10 min, then reduced to 6 mg
kg-1
hr-1 for the remainder of the procedure;, F = fentanyl; Iso = isoflurane; MAC = minimum alveolar concentration; NS = not specified; N2O = nitrous oxide; R = remifentanil;
Sevo = sevoflurane; T = thiopentone; TCI = target controlled infusion

123

Fig. 2 Risk of bias graph:
review authors’ judgements
about each risk of bias item
presented as percentages across
all included studies
were conducted in either a single centre or multiple centres.
Eight of the 14 studies involved supratentorial
craniotomies, and the remaining six studies involved both
supratentorial and infratentorial surgery. The indications
for craniotomy included brain tumour resection, aneurysm
clipping, excision of arteriovenous malformations, vessel
ligation for trigeminal neuralgia, and evacuation of
hematomas. Most of the study population consisted of
young fit neurologically intact elective neurosurgical
patients (Glasgow Coma Score of 15 and American
Society of Anesthesiologists physical status I-III) with a
mean age of 52.1 (range 18-70). The distribution of sex
was quite homogeneous across different studies.
Anesthesia management is shown in Table 1. For
propofol-maintained anesthesia, the manual infusion of
propofol at rates of 10-8-6 mg
kg-1
hr-1 was the most
commonly used technique in the primary studies. Only two
studies used TCI for the administration of propofol. With
regard to volatile-maintained anesthesia, sevoflurane alone
was used in seven studies, one study used either
sevoflurane or isoflurane alone, and six studies used
isoflurane with nitrous oxide. Desflurane was not used in
any study. Mannitol was used in only eight studies. The
majority of studies reported the titration of anesthetic
agents against blood pressure not according to depth of
anesthesia monitors. In addition, in most studies, only the
proposed dosing regimen was reported and not the actual
total doses of anesthetics used.
Methodological quality of the studies
A risk of bias graph and summaries are shown in Figs 2
and 3. In eight studies, the randomization sequences were
generated by computer-generated random numbers, and in
six studies, the methods of randomization were not
specified. An open-label design was commonly used in
many studies, and only six studies provided good
documentation of adequate allocation concealment.
Blinding of participants was documented in two studies,
123
J. Chui et al.
and in four studies, blinding of outcome assessors was
reported. Selective reporting was identified in four studies.
No study was rated ‘‘yes’’ in all domains.
Meta-analysis results
The results of the meta-analysis of all outcome variables
are summarized in Table 2. All forest plots and funnel plots
are available online (Figs 4 to 35) as Electronic
Supplementary Material.
Cerebral hemodynamics
Brain relaxation score
The majority of studies used a four-point scale (slack brain,
mild, moderate, and severe brain herniation). Some studies
used a three-point scale, whereas others reported only the
number of events of brain herniation that required
treatment with mannitol. For this meta-analysis, only the
studies with complete data for a four-point scale were
included, and the RR for each category was
calculated.9,11,16,19,22 The propofol-maintained group had
lower RR for moderate herniation (RR = 0.75; 95% CI
0.58 to 0.97; P = 0.03) than the volatile-maintained group,
but there were no differences amongst the groups in other
categories.
Intracranial pressure and cerebral perfusion pressure
In the pooled analysis, propofol-maintained anesthesia had
lower ICP values than volatile-maintained anesthesia; the
overall weighted mean difference was -5.2 mmHg (95%
CI -6.81 to -3.6; P \ 0.0001).16,19,22 Similar to the ICP
comparison, there were significant variations in the CPP
measurement in different studies. The propofol-maintained
group had higher CPP values than the volatile-maintained
group; the combined CPP weighted mean difference was
16.33 mmHg (95% CI 12.2 to 20.46; P \ 0.001).16,19
Propofol versus volatile agents for elective craniotomy
Fig. 3 Risk of bias summary: review authors’ judgements about each
risk of bias item for each included study
Intraoperative hemodynamic events
The hemodynamic stability of the two anesthetic
techniques was evaluated in most of the primary studies
and was frequently reported as the number of events of
hypertension, hypotension, bradycardia, and tachycardia.
Propofol-maintained anesthesia was associated with fewer
intraoperative hypotensive events than volatile-maintained
anesthesia (RR= 0.74; 95% CI 0.61 to 0.90; P = 0.002).
There were no significant differences in the incidence of
intraoperative hypertension, bradycardia and tachycardia
between the groups.9,11,13–15,17
Recovery profiles
An objective comparison of the two anesthetic techniques
was difficult due to the variations in the variables used to
quantify the recovery profiles of the patients. Commonly
used recovery variables included mean times to: eye
opening, tracheal extubation, obeying verbal commands,
orientation, and to achieve an Aldrete score[9. Due to the
presence of the clinical and methodological heterogeneity
in the primary studies, not all recovery profile outcomes
could be combined and quantitative analysis was not
possible. Semi-qualitative evaluation of the data, the mean
difference in time to eye opening11,14,15,18 and
extubation11,14,15,19 between the groups, ranged from -4
to 2 min and from 0.1 to 2 min, respectively. Similarly, the
mean difference in time to achieve an Aldrete score [ 9
ranged from -3.75 to 0.13 min.9,11,19 The times to obey
commands14,19 and orientation19,20 were faster in the
propofol-maintained group with the mean difference
ranging from -5.5 to 0 min and from -7.5 to -23.6
min, respectively.
Postoperative complications
Postoperative complications were commonly reported as
the incidence (i.e., number) of events that occurred per
total patients. There were no significant differences
between propofol-maintained and volatile-maintained
techniques with regard to the incidence of postoperative
pain, seizure, and agitation. Nevertheless, the incidence of
postoperative nausea and vomiting (PONV) was
significantly lower with propofol-maintained anesthesia.
There was a significant heterogeneity encountered in
combining the results (Fig 13; available as Electronic
Supplementary Material); hence, subgroup analysis was
completed, which showed that nitrous oxide is an effect
modifier of propofol-maintained anesthesia in preventing
PONV. The protective effect of propofol-maintained
anesthesia on PONV was more prominent in the
subgroup where nitrous oxide was used (OR 0.1; 95% CI
0.03 to 0.35; P \ 0.0003)17,19,21 compared with the group
where no nitrous oxide was used (OR 0.65; 95% CI 0.38 to
1.10; P = 0.14).9,11,13–15
Neurological outcomes
One primary study reported the 24-hr postoperative
neurological complications as a secondary outcome and
123
 J. Chui et al.

Table 2 Summary of the meta-analysis results of all outcome measures

Outcomes Study Result P value Heterogeneity (I2)

Brain Relaxation Score

Slack Brain 1,3, 8, 11, 14 RR: 1.09 (0.9 to 1.33) P = 0.38 11%
Mild Brain Herniation 1,3, 8, 11, 14 RR: 1.15 (0.97 to 1.37) P = 0.11 5%
Moderate Brain Herniation 1,3, 8, 11, 14 RR: 0.75 (0.58 to 0.97) P = 0.03 0%
Severe Brain Herniation 1,3, 8, 11, 14 RR: 0.7 (0.45 to 1.11) P = 0.13 0%
Cerebral Haemodynamics
ICP 8, 11, 14 MD: -5.2 (-6.81 to -3.6) P \ 0.00001 0%
CPP 8, 11 MD: 16.33 (12.2 to 20.46) P \ 0.00001 0%
Intraoperative Hemodynamic Event
Intraoperative Hypotension 1, 3, 5, 6, 7, 9 RR: 0.74 (0.61 to 0.90) P = 0.002 0%
Intraoperative Hypertension 1, 3, 5, 6, 7, 9 RR: 1.17 (0.97 to 1.4) P = 0.09 25%
Intraoperative Bradycardia 1, 3, 7, 9 RR: 1.2 (0.9 to 1.61) P = 0.21 0%
Intraoperative Tachycardia 1, 3, 7, 9 RR: 0.77 (0.36 to 1.61) P = 0.49 56%
Recovery Profiles
Time to eye opening 3, 6, 7, 10 Unable to combine - 81%
Time to extubation 3, 6, 7, 11 Unable to combine - 68%
Time to obey command 6, 11 MD: -1.67 (-6.62 to 3.29) P = 0.51 60%
Time to orientation 11, 12 Unable to combine - 98%
Time to Aldrete score [9 1, 3,11 Unable to combine - 80%
Post operative Complications
PONV with Nitrous Oxide 9, 11, 13 OR: 0.1 (0.03 to 0.35) P \ 0.0003 13%
PONV without Nitrous Oxide 1, 3, 5, 6, 7 OR: 0.65 (0.38 to 1.10) P = 0.14 42%
Pain 1, 3, 5, 6, 7, 9 RR: 0.98 (0.82 to 1.17) P = 0.84 0%
Seizure 1, 3 RR: 1.09 (0.47 to 2.52) P = 0.84 0%
Agitation 1, 11 RR: 1.7 (0.71 to 4.1) P = 0.24 3%
Shivering 1, 3, 5, 7, 13 Unable to combine - 76%
ICP = intracranial pressure; CPP = cerebral perfusion pressure; MD = mean difference (95% confidence interval); PONV = postoperative
nausea and vomiting; RR = risk ratio (95% confidence interval); OR = odd ratio (95% confidence interval)

did not find differences between the two anesthetic
techniques.20 None of the included studies investigated
other clinical outcomes in the two anesthetic techniques,
such as long-term neurological morbidity, mortality, or
quality of life; consequently, quantitative analysis could
not be performed.
Subgroup analysis
Subgroup analysis was performed for each outcome
parameter with different combinations of anesthetic
regimens (propofol infusion techniques, different types/
combination of volatile agents, and types of opioids used).
With propofol-maintained anesthesia, there are no
differences in the pooled results with either manual
controlled infusions or targeted-controlled infusions.
Similarly, in the volatile-maintained group, there are no
differences in the pooled results with use of either
isoflurane or sevoflurane as the choice of volatile agent.
There are no differences in the subgroup analysis of other
outcome parameters with the addition of nitrous oxide.
Discussion
This meta-analysis showed that brain relaxation scores
were similar when comparing propofol-maintained vs
volatile-maintained anesthesia; however, with propofol-
maintained anesthesia, the initial ICP was lower and CPP
was higher when compared with volatile-maintained
anesthesia. This ICP lowering effect is not associated
with less brain swelling or better operative conditions after
dural opening, as shown by the similar brain relaxation
score with both techniques. Nevertheless, all studies that
assessed the brain relaxation scores were limited to patients
with a Glasgow Coma Score[15 undergoing supratentorial
excision of tumours. The preoperative ICP would be
expected to be normal or only slightly elevated in this

123
Propofol versus volatile agents for elective craniotomy
group of patients. Therefore, the effect of propofol-
maintained anesthesia on brain relaxation might be
expected to be small. The intraoperative hemodynamic
events and perioperative complications were similar with
the two techniques, except for the incidence of PONV. Due
to heterogeneity, we were unable to combine the variables
to assess the recovery profiles; however, mean differences
in certain important recovery variables were small, e.g., the
magnitude of minutes, and they were also of minimal
clinical significance.
Currently, there are no consensus guidelines or
recommendations suggesting the best anesthesia
technique for neurosurgical procedures. Most clinical
trials assessed perioperative cerebral and hemodynamic
variables and recovery profiles as their major outcome
measures. The major clinically relevant outcomes, such as
neurological function, morbidity, mortality, and the quality
of life, were not evaluated in most studies comparing the
two anesthetic techniques. For purposes of applying the
results of this systemic review into practice, there were no
differences in the majority of outcome variables, and due to
lack of data on clinically significant outcomes, we are
unable to recommend one anesthetic technique over the
others in patients undergoing elective craniotomy.
This systematic review has several limitations. First, all
primary studies included only relatively healthy and
neurologically intact patients; hence, the extrapolation of
our results to all neurosurgical patients is inappropriate.
Second, in this systematic review, we have segregated the
interventions into either propofol-based or volatile-based
anesthesia. This classification is generally understood and
was commonly used in the previous narrative reviews for
discussion. Nevertheless, there could be concerns about
oversimplification and generalization of different anesthetic
regimens. The volatile-based anesthesia group, in particular,
included different combinations of volatile agents
(sevoflurane, isoflurane, with or without nitrous oxide);
however, subgroup analysis did not reveal any differences in
the pooled results with different combinations of volatile
anesthetic regimens. Third, because of the significant
clinical and methodological heterogeneity amongst the
primary studies, we were unable to combine all the
outcome variables for recovery profiles. Fourth, some other
important considerations in neuroanesthesia include
preservation of autoregulation and carbon dioxide
reactivity, the neuroprotective effect, and the interference
with neuromonitoring, but these factors were not included in
this meta-analysis due to the heterogeneity of the data.
Finally, cerebral ischemia, postoperative neurological
deficits, and perioperative mortality are the important
outcome parameters, but none of the primary studies
measured these important outcomes. Instead, CPP and ICP
were used as surrogate outcomes for cerebral ischemia.
Based on this systemic review, there are several study
design features that would be beneficial for future studies.
First, blinding of patients and the assessors (surgeon) is
practically feasible in the operating theatre, and it is
important to reduce reporting and observer bias. Second,
the importance of reporting integrity should be
emphasized, as the quality of reporting across studies is
equally important to conducting the study. The CONSORT
statement, which is endorsed by the Journal, is an
evidence-based minimum set of recommendations for
reporting RCTs.D Third, clinical outcomes such as
neurological morbidity and mortality should be important
in choosing anesthetic techniques and should be the focus
of future studies.
In conclusion, this systematic review compared the
effects of propofol-maintained vs volatile-maintained
anesthesia in patients undergoing elective craniotomy.
There were similarities between propofol-maintained and
volatile-maintained anesthesia in the majority of outcomes
measured. The initial mean ICP values were lower and CPP
values were higher with propofol-maintained anesthesia
than with volatile anesthesia, but this ICP lowering effect
was not associated with less brain swelling or better
operative conditions after dural opening, as shown by
similar brain relaxation scores with both techniques. As no
clinically significant outcome differences in neurological
morbidity or mortality have been shown, conclusive
recommendations based on the results of this systemic
review cannot be made.
Acknowledgement We acknowledge the assistance of Ms. Marina
Englesakis for conducting the systematic searches for this manuscript.
Funding None.
Conflicts of interest None declared.
References
1. Engelhard K, Werner C. Inhalational or intravenous anesthetics
for craniotomies? Pro inhalational. Curr Opin Anaesthesiol 2006;
19: 504-8.
2. Hans P, Bonhomme V. Why we still use intravenous drugs as the
basic regimen for neurosurgical anaesthesia. Curr Opin
Anaesthesiol 2006; 19: 498-503.
3. Dahyot-Fizelier C, Frasca D, Debaene B. Inhaled agents in
neuroanaesthesia for intracranial surgery: pro or con (French).
Ann Fr Anesth Reanim 2012; 31: e229-34.
4. Mizumoto N, De Oliveira EP. Total intravenous anesthesia:
advantages for intracranial surgery - Commentary. Neurosurgery
2007; 61(5 Suppl. 2): 377-8.
D
Consolidated Standards of Reporting Trials: http://www.consort-
statement.org (accessed August 2013).
123

5. Weglinski MR, Perkins WJ. Inhalational versus total intravenous
anesthesia for neurosurgery: theory guides, outcome decides. J
Neurosurg Anesthesiol 1994; 6: 290-3.
6. Archer DP. TIVA versus inhalational anesthesia for
neurosurgery. J Neurosurg Anesthesiol 1994; 6: 284.
7. Ravussin P, de Tribolet N, Wilder-Smith OH. Total intravenous
anesthesia is best for neurological surgery. J Neurosurg
Anesthesiol 1994; 6: 285-9.
8. Cole CD, Gottfried ON, Gupta DK, Couldwell WT. Total
intravenous anesthesia: advantages for intracranial surgery.
Neurosurgery 2007; 61(5 Suppl 2): 369-77.
9. Citerio G, Pesenti A, Latini R, et al. A multicentre, randomised,
open-label, controlled trial evaluating equivalence of inhalational
and intravenous anaesthesia during elective craniotomy. Eur J
Anaesthesiol 2012; 29: 371-9.
10. Banevicius G, Rugyte D, Macas A, Tamasauskas A, Stankevicius
E. The effects of sevoflurane and propofol on cerebral
hemodynamics during intracranial tumors surgery under
monitoring the depth of anesthesia. Medicina (Kaunas) 2010;
46: 743-52.
11. Lauta E, Abbinante C, Del Gaudio F, et al. Emergence times are
similar with sevoflurane and total intravenous anesthesia: results
of a multicenter RCT of patients scheduled for elective
supratentorial craniotomy. J Neurosurg Anesthesiol 2010; 22:
110-8.
12. Bonhomme V, Demoitie J, Schaub I, Hans P. Acid-base status and
hemodynamic stability during propofol and sevoflurane-based
anesthesia in patients undergoing uncomplicated intracranial
surgery. J Neurosurg Anesthesiol 2009; 21: 112-9.
13. Magni G, La Rosa I, Gimignani S, Melillo G, Imperiale C, Rosa
G. Early postoperative complications after intracranial surgery:
comparison between total intravenous and balanced anesthesia. J
Neurosurg Anesthesiol 2007; 19: 229-34.
14. Sneyd JR, Andrews CJ, Tsubokawa T. Comparison of propofol/
remifentanil and sevoflurane/remifentanil for maintenance of
123
J. Chui et al.
anaesthesia for elective intracranial surgery. Br J Anaesth 2005;
94: 778-83.
15. Magni G, Baisi F, La Rosa I, et al. No difference in emergence
time and early cognitive function between sevoflurane–fentanyl
and propofol–remifentanil in patients undergoing craniotomy for
supratentorial intracranial surgery. J Neurosurg Anesthesiol 2005;
17: 134-8.
16. Petersen KD, Landsfeldt U, Cold GE, et al. intracranial pressure
and cerebral hemodynamic in patients with cerebral tumors: a
randomized prospective study of patients subjected to craniotomy
in propofol-fentanyl, isoflurane-fentanyl, or sevoflurane-fentanyl
anesthesia. Anesthesiology 2003; 98: 329-36.
17. Talke P, Caldwell J, Brown R, Dodson B, Howley J, Richardson
CA. A comparison of three anesthetic techniques in Patients
undergoing craniotomy for supratentorial intracranial surgery.
Anesth Analg 2002; 95: 430-5.
18. Ittichaikulthol W, Pausawasdi S, Srichintai P, Sarnvivad P.
Propofol vs isoflurane for neurosurgical anesthesia in Thai
patients. J Med Assoc Thai 1997; 80: 454-60.
19. Todd MM, Warner DS, Sokoll MD, et al. A prospective,
comparative trial of three anesthetics for elective supratentorial
craniotomy. Propofol/fentanyl, isoflurane/nitrous oxide, and
fentanyl/nitrous oxide. Anesthesiology 1993; 78: 1005-20.
20. Ravussin P, Tempelhoff R, Modica PA, Bayer-Berger MM.
Propofol vs. thiopental-isoflurane for neurosurgical anesthesia:
comparison of hemodynamics, CSF pressure, and recovery. J
Neurosurg Anesthesiol 1991; 3: 85-95.
21. Mishra LD, Dube SK, Singh SN, Dubey RK. Comparison between
continuous propofol infusion & conventional balanced
anaesthesia in neurosurgical patients. J Anaesth Clin Pharmacol
2008; 24: 399-402.
22. Santra S, Das B. Subdural pressure and brain condition during
propofol vs isoflurane - nitrous oxide anaesthesia in patients
undergoing elective supratentorial tumour surgery. Indian J
Anaesth 2009; 53: 44-51.