CLINICAL INVESTIGATION
Early Postoperative Complications After
Intracranial Surgery
Comparison Between Total Intravenous and Balanced Anesthesia
Giuseppina Magni, MD, PhD,* Italia La Rosa, MD,* Simona Gimignani, MD,*
Guido Melillo, MD,w Carmela Imperiale, MD,* and Giovanni Rosa, MD*
Abstract: This prospective study was performed to compare the
incidence of complications occurring after neurosurgical proce-
dures in patients anesthetized with either sevoflurane-fentanyl or
propofol-remifentanil anesthesia. We enrolled 162 American
Society of Anesthesiologists (ASA) I to III patients (82 females
and 80 males, Glasgow 15) undergoing elective neurosurgical
procedures. Anesthesia was conducted using either propofol-
remifentanil (T group; n = 80 patients) or sevoflurane-fentanyl
(S group; n = 82 patients). All patients were monitored in the
postanesthesia care unit for 6 hours after extubation. We
analyzed and compared in both groups the incidence of high
severity complications such as respiratory events (PaO2
<90 mm Hg; PaCO2 >45 mm Hg) and neurologic events
(seizures, new motor or sensory deficit, unexpected delay of
awakening) and the incidence of low severity complications such
as hypertension (mean arterial pressure increase above 30% of
baseline), hypotension (mean arterial pressure decrease below
30% of baseline), pain, shivering, nausea, and vomiting. A total
of 162 complications occurred in 92 patients (57%) with 50
patients (31%) having had 1, 26 patients (16%) having had 2,
and 16 patients (10%) having had 3 or more events. The most
frequent complication was respiratory impairment (28%) which
was frequently reported only in the first postoperative hour. Out
of the total number of complicating events, 77 (48 %) were
found in group S, and 85 (52%) in group T (P = ns). Severe
complications were rarely reported and evenly distributed in the
2 anesthetic groups. Similarly, no difference could be demon-
strated in the composite incidence of less serious complications
between the 2 anesthetic regimens tested in this study. This study
confirms that the recovery period after neurosurgical procedures
remains a time of great potential danger to patients given the
high incidence of postoperative complicating events indepen-
dently from the anesthetic strategy.
Received for publication December 13, 2006; accepted April 12, 2007.
From the *Department of Anesthesia and Intensive Care, Policlinico
Umberto I, University of Rome ‘‘La Sapienza’’; and wIstituto
Dermopatico dell’Immacolata, IRCCS, Vascular Pathology, Rome,
Italy.
Reprints: Giuseppina Magni, MD, PhD, Department of Anaesthesia
and Intensive Care Policlinico Umberto I, Viale del Policlinico
155-00100 Rome, Italy (e-mail: gmagni@yahoo.com).
Copyright r 2007 by Lippincott Williams & Wilkins
J Neurosurg Anesthesiol  Volume 19, Number 4, October 2007
Key Words: remifentanil, propofol, neuroanesthesia, neuro-
surgery, sevolurane
(J Neurosurg Anesthesiol 2007;19:229–234)
T he incidence of postoperative complications occurring
in the recovery room is reported to be as high as 30%
for the general surgical population and even higher
(54.5%) for neurosurgical patients.1–5 The most frequent
events reported are nausea and vomiting, shivering,
respiratory and cardiovascular impairment.
In the effort to find the best anesthetic strategy,
several studies have been made comparing total intra-
venous anesthesia and inhalation techniques in terms of
incidence of postoperative nausea and vomiting, recovery
of cognitive function, cost, and patient satisfaction in
various surgical settings including neurosurgical proce-
dures.6–9 However, because volatile and intravenous
anesthetics have been tested mostly in selected patients
with supratentorial tumors,9–11 we performed a prospec-
tive study to test the hypothesis that intravenous
anesthesia may reduce the incidence of postoperative
adverse events in neurosurgical patients. The primary end
point was to evaluate the extent and gravity of early
complications occurring in patients undergoing a wide
range of neurosurgical procedures and anesthetized with
either fentanyl-sevoflurane or propofol-remifentanyl
combination.
METHOD
After approval by the institutional review board and
written informed consent, 162 patients American Society
of Anesthesiologists (ASA) physical status I to III,
scheduled for elective intracranial surgery, were enrolled
in the study. All patients were under steroid treatment.
Criteria of exclusion were preoperative decrease in the
level of consciousness, (Glasgow Coma Scale <15)
complications during surgery (unanticipated brain swel-
ling, injury to the cranial nerves), and anticipated
duration of surgery (>5 h). An isotonic crystalloid saline
solution (10 mL/kg preoperatively and 4 to 6 mL/kg/h
during surgery) was infused through a peripheral
intravenous catheter, and a second line was inserted for
229
Magni et al
drugs administration. A balance between fluids intro-
duced and lost was checked every hour. Blood was
administered if the amount of hemoglobin fell below
9 mg/dL.
In all patients anesthesia was induced with propofol
(3 mg/kg), fentanyl (3 mcg/kg), and vecuronium (0.1 mg/kg)
in O2 100%. Preoperatively, the patients were randomly
divided in 2 groups using a computer-generated rando-
mization scheme. In group S (82 patients), anesthesia was
maintained using sevoflurane with an end tidal of 1.5% to
2% to achieve 1.3 to 1.8 times minimum alveolar
anesthetic concentration. Fentanyl bolus (0.7 mcg/kg)
were added in S group when considered necessary by
the attending anesthesiologist. In group T (80 patients)
anesthesia was maintained with continuous infusion of
propofol (10 mg/kg/h for the first 10 min, then reduced
to 8 mg/kg/h for the following 10 min, and reduced to
6 mg/kg/h thereafter)12 and remifentanil (0.5 to 0.25 mcg/
kg/min reduced to 0.05 to 0.1 mcg/kg/min after dural
opening). Vecuronium was administered as neuromuscu-
lar blocking agent according to train of four (TOF)
monitoring in both groups. After intubation of the
trachea, mechanical ventilation was begun. An inspired
mixture of air and oxygen (3:1) was administered.
Ventilation was adjusted to achieve a partial pressure of
carbon dioxide in arterial blood (PaCO2) of 35 mm Hg.
At dural opening, brain relaxation was assessed by
the attending neurosurgeon by using a 4-point scale: (1)
relaxed brain; (2) mild brain swelling, acceptable; (3)
moderate brain swelling, no treatment required; and (4)
severe swelling, requiring treatment (mannitol 0.5 mg/kg).
Arterial blood pressure was measured via a radial
artery cannula connected to a Haemomed transducer
(Siemens) positioned before surgery, after sedation with
midazolam (0.03 mg/kg IV) and local infiltration of
lidocaine. Before induction mean arterial pressure
(MAP) and heart rate were recorded for 5 minutes with
the patient in resting state. Baseline mean blood pressure
was defined as the mean of the 3 lowest values recorded
1 minute immediately before induction of anesthesia.
Intraoperatively, MAP and heart rate were maintained
within predetermined limits: propofol, remifentanil,
fentanyl, and sevoflurane doses were adjusted to maintain
the mean blood pressure within a range of ± 20% of
preanesthesia level with heart rate <90 bpm. Vital signs
were recorded throughout the surgery period: MAP more
than 20% above baseline, heart rate responses more than
90 beat/min, swelling, or movement were used during
maintenance anesthesia to justify an increase in drugs
administration. In the volatile anesthetic group, addi-
tional boluses of fentanyl 0.7 mcg/kg were given if the
patient failed to respond to increases in the level of the
primary anesthetic (sevoflurane up to a minimum alveolar
anesthetic concentration of 1.8%). In the total intrave-
nous anesthesia group the remifentanyl infusion rate was
increased by increments of 0.1 mcg/kg/min when the
infusion rate of propofol was insufficient to maintain an
adequate level of anesthesia. If relative hypertension
(MAP above 30% of baseline value) persisted, despite
230
J Neurosurg Anesthesiol  Volume 19, Number 4, October 2007
achievement of maximal allowed anesthetic concentra-
tion, episodes of relative hypertension or tachycardia
(lasting more than 1 min) were treated with labetalol
(25-mg bolus). Anesthetics were decreased only in
response to a reduction of MAP of 20% of preinduction
values that was not responsive to replacement of intra-
operative fluid losses. When clinically indicated, a
vasopressor (ephedrine 5-mg IV) was administered.
Sevoflurane inhalation, propofol and remifentanil
infusion were reduced once the bone flap was secured and
stopped after skin dressing. Intraoperative normothermia
was maintained with a convective device blanket.
Infiltration of the scalp with bupivacaine 0.25% was
accomplished in all patients before incision of the skin. At
the end of surgery all patients received ketorolac 30 mg,
labetalol (100 mg, fractionated in doses of 25 mg IV
bolus), nefopam 10 mg, and ondansetron 4 mg. All
patients after extubation were monitored in the recovery
room for 6 hours and received supplemental oxygen at a
flow rate of 8 L/min (FiO2 40%) during the entire period
of observation.
Postoperative Management
Postoperative complications were defined for the
purposes of the present study as reported in Table 1.
PaO2, PaCO2, and pH were monitored during surgery
and for 6 hours after extubation. Blood samples for gas
analysis (i-stat Abbott gas analyzer) were taken every
hour during surgery, every 15 minutes after extubation
for the first hour and every hour for the following 5
hours. Emergence time was measured as the time between
drug interruption and patient’s eye opening (sponta-
neously or on verbal prompting). Extubation time was
measured as the time elapsing from anesthetic disconti-
nuation and extubation (performed when the patient
obeyed verbal commands). Moreover, the patient’s
neurologic status (evaluation of level of consciousness,
presence of new motor or sensory deficit, and presence of
brain stem reflexes) was checked every 15 minutes, the
manifestation of shivering, postoperative nausea and
TABLE 1. Definition of Complications
High severity complications
Respiratory
PaO2 <90 mm Hg
PCO2 >45 mm Hg
Reintubation
Neurologic
Seizures
New motor or sensory deficit
Unexpected delay of awakening
Low severity complications
Nausea, vomiting*
Hemodynamic
Hypertension (MAP increase above 30% of baseline)
Hypotension (MAP decrease below 30% of baseline)
Pain*
Shivering
*Pain and nausea requiring rescue medication (VAS above 50).
r 2007 Lippincott Williams & Wilkins
J Neurosurg Anesthesiol  Volume 19, Number 4, October 2007 Early Complications After Neurosurgery

vomiting, and the request for analgesic medications were

TABLE 2. Demographic Characteristics, Surgery and
monitored and recorded by an observer blinded to the Anesthesia Duration, Anesthetic Technique, and Site of
anesthetic management. In the preoperative area, patients Space-occupying Lesions in the 2 Groups
completed baseline Visual Analog Scale (VAS) grading a
Characteristics Group T (n = 80) Group S (n = 82) P
100 mm baseline VAS with 0 = no pain or nausea and
100 = worst possible pain or nausea.13 The incidence of Age (y) 52.3 ± 14.4 53.4 ± 15.4 n.s.
Height (cm) 165.3 ± 12 163.2 ± 13.1 n.s.
pain and nausea requiring rescue medication (VAS above Weight (kg) 75.3 ± 20 74.1 ± 18 n.s.
50) was recorded and treated with ketorolac 30-mg IV Sex (F/M) 29/31 27/33 n.s.
and metoclopramide 10-mg IV, respectively. Shivering ASA (I/II/III) 35/24/21 37/23/22 n.s.
was treated with nefopam 10-mg IV.14 The incidence of Surgery duration (min) 270.1 ± 53 266 ± 50 n.s.
Anesthesia duration (min) 300.8 ± 85.0 303.7 ± 68.0 n.s.
episodes of hypotension (MAP decrease below 70% of Supratentorial surgery 50 54 n.s.
the baseline value for more than 1 min), hypertension Infratentorial surgery 23 20 n.s.
(MAP increase above 130% of the baseline value for Aneurysm clipping 7 8 n.s.
more than 1 min) was recorded and treated.
For the purpose of statistical analysis, neurologic
events (seizures, new motor deficit, and unexpected lack therapy was necessary to decrease brain bulking. Blood
of awakening) and need for reintubation were defined as transfusion was provided in 18 patients (10 in T and 8 in
high severity complications. The composite incidence of S group; P: ns).
the severe complications was analyzed separately from all The mean temperature for the 2 groups on entry
the remaining events (lower severity complications). into the recovery room was 36.5 ± 0.3 for T group and
36.7 ± 0.5 for S group, respectively (P: ns). The mean
Statistical Analysis emergence time (13.3 ± 4.9 min for group S vs.
Data are expressed as mean ± standard deviation 12.8 ± 6.1 min for group T; P = 0.92) and extubation
(for continuous variables) or as percentages (for propor- time (18.4 ± 2.1 min for group S vs. 18.0 ± 2.1 min for
tions). Data were blindly recorded on specially produced group T; P = 0.80) were similar in the 2 groups.
paper record forms and then stored into a computer Complications were reported in 92/162 patients
database. The Stata 8.0 software package was used for (57%); 50 patients experienced only 1 complication
computer analysis. Differences between continuous (30%); 26 patients 2 complications (16%); and 16 patients
variables were analyzed by unpaired t test, differences more than 2 complications (10%). The composite
between proportions were analyzed with the w2 test. The incidences of serious and less serious complications are
results were considered statistically significant for P reported separately in Tables 3A and B, respectively:
values <0.05. no statistically significant difference was found between
To compute sample size for this study, a 70% T and S anesthetic techniques comparing the composite
incidence of at least one postoperative complication was incidences of less serious complicating events. Overall
hypothesized from the available literature in an unse- frequency of complicating events per category is reported
lected group of patients.5 We estimated that a clear in Table 4. Out of the total number of complicating
clinical benefit could be established in the T group if a events, 82 (50%) were found in S group, and 82 (50%) in
30% or greater decrease in the complication incidence T Group: no statistically significant difference was found
could be obtained in these patients. Therefore, with b set between T and S anesthetic techniques comparing the
to 0.2 and a to 0.05 (single sided), it was estimated that a frequency of complicating events (Table 4).
minimum of 76 patients were needed for each study Respiratory complications were present in 28% of
group. patients. After the first hour, the incidence of respiratory
complications decreased to 2.4% (Table 5). Reintubation
RESULTS was necessary in 2 patients having both PaCO2 and PaO2
From a total number of 168 eligible patients, 6 abnormal levels: in 1 case because of persistent epileptic
subjects, in whom early awakening was not considered crisis and in the other because of deterioration of
safe, were excluded. The analysis was, therefore, carried consciousness.
out on 162 patients. Supratentorial surgery (104 cases) Pain requiring rescue medication was present in
included: 100 tumors and 6 arterovenous malformations. 39/162 (24%) of patients and was treated with a second
Infratentorial surgery included 38 tumors and 5 caver- dose of ketorolac 30 mg; pain was significantly reduced in
nous angiomas. All the aneurysms were supratentorial. 85% (33/39) of patients (VAS<50).
Baseline characteristics, demographic data, includ- A significant correlation between preoperative ASA
ing age, sex, and ASA group, site of space-occupying classification and incidence of complications analyzed
lesions, duration of surgery, and anesthesia were similar was demonstrated (Table 6).
in the 2 treatment groups and are reported in Table 2. The
requirement of fentanyl administration never exceeded
the dosage of 500 mcg/patient. In 25 patients (14 for DISCUSSION
group S and 11 for group T, P: ns) intracranial volume The result of our study shows that early postopera-
was graded as severely abnormal and decompression tive complications in patients undergoing neurosurgical

r 2007 Lippincott Williams & Wilkins 231

Magni et al J Neurosurg Anesthesiol  Volume 19, Number 4, October 2007

TABLE 3. Complications
Group S Group T Total
A. Patients with high severity complications
Neurologic 3 patients 2 patients 5 patients
Reintubation 2 patients 0 2 patients
B. Patients with lower severity complications and no. complicating events
No. complicating events
1 26 patients (26 events) 22 patients (22 events) 48 patients (48 events)
2 11 patients (22 events) 13 patients (26 events) 24 patients (48 events)
>2 7 patients (29 events) 10 patients (32 events) 17 patients (61 events)
2 patients: 3 events 8 patients: 3 events 10 patients: 3 events
2 patients: 4 events 2 patients: 4 events 4 patients: 4 events
3 patients: 5 events 0 patient: 5 events 3 patients: 5 events
Total 44 patients (77 events) 45 patients (80 events) 89 patients (157 events)
P: ns (w2 for the 3  2 table).

procedures are common. No difference between the 2 dependent incidence of respiratory impairment using
anesthetic regimens could be demonstrated in terms of blood gas analysis. We found that 28% of patients had
composite incidence of low severity postoperative com- respiratory impairment, whereas in the paper by Manni-
plications. However, given the small number of severe nen and colleagues,5 respiratory complications occurred
complications observed in our series, we cannot make any only in 6.3% of patients. This difference may be due to
inference from the comparison of severe events in the 2 the different definition of respiratory impairment between
groups because of the insufficient statistical power; even Manninen and our study; Manninen defined hypoxia as
so, this is a difficulty that applies to all analyses of rare SpO2 <90% and hypoventilation as respiratory rate <8
events. whereas we used PaO2 <90 mm Hg for hypoxia and
In our study, complications were reported in 57% of PaCO2 >45 mm Hg for hypercarbia. These methodolo-
patients; 30% of patients experienced only 1 complication gical differences may account for the higher incidence of
16% of patients 2 complications, and 10% of patients respiratory complication found in our series of patients.
more than 2 complications. The incidence of complica- In most of the available literature, arterial blood gas
tions reported in our study is in agreement with analysis was much less frequently used for definition of
Manninen and colleagues5 who reported complications respiratory impairment16 compared with ventilatory
in 59% of patients with brain tumor and in 83% of frequency (<10 bpm) or oxygen saturation (<90%).
patients who underwent vascular surgery. Moreover, an When blood gas analysis is used, a partial pressure of
overall complication rate of 51% was reported by Wong carbon dioxide greater than 50 mm Hg is the most
and coworkers15 in patients undergoing neurosurgical frequently used end point.17 We acknowledge that, in this
procedures treated with propofol-remifentanil anesthesia. study, definitions for hypoxia and hypercarbia were set
Apart from previously published reports, the most to lower thresholds than those applied to the general
frequent complication encountered in our study was surgical population, but we considered neurosurgical
respiratory impairment. In our series craniotomy was patients more prone to the detrimental effects of hypoxia
associated with a very high rate of blood gas abnormal- and hypercarbia on cerebral blood flow autoregulation.18
ities in the first postoperative hour, regardless of
technique used. This study is the first evaluating time
TABLE 5. Respiratory Complications During Recovery Period
PaO2 PaCO2 PaO2 <90 and
TABLE 4. Complicating Events per Category <90 mm Hg >45 mm Hg PaCo2 >45 mm Hg Total
Complications S T P (S vs. T) Total T15 min 18 18 10 46 (28%)
T30 min 6 9 4 19 (12%)
Shivering 16/82 13/80 n.s. 29/162 (18%) T45 min 2 1 2 5 (3%)
Pain 18/82 21/80 n.s. 39/162 (24%) T1 1 1 2 4 (2.4%)
PONV 10/82 11/80 n.s. 21/162 (13%) T2 0 1 2 3 (2%)
Hypertension 9/82 13/80 n.s. 22/162 (13%) T3 0 0 3 3 (2%)
Hypotension 0/82 0/80 n.s. 0/162 (3%) T4 0 2 0 2 (1.2%)
Neurologic 3/82 2/80 n.s. 5/162 (3%) T5 1 1 0 2 (1.2%)
PaO2 <90 mm Hg 9/82 9/80 n.s. 18/162 (11%) T6 0 1 0 1 (0.6%)
PaCO2 >45 mm Hg 10/82 8/80 n.s. 18/162 (11%)
PaO2 <90 PaCO2 >45 5/82 5/80 n.s. 10/162 (6%) T15 min = 15 min after extubation; T30 min = 30 min after extubation;
Reintubation 2/82 0/80 n.s. 2/162 (1%) T45 min = 45 min after extubation; T1 = 1 h after extubation; T2 = 2 h after
extubation; T3 = 3 h after extubation; T4 = 4 h after extubation; T5 = 5 h after
PONV indicates postoperative nausea and vomiting. extubation; T6 = 6 h after extubation.

232 r 2007 Lippincott Williams & Wilkins

J Neurosurg Anesthesiol  Volume 19, Number 4, October 2007
TABLE 6. Preoperative ASA Status and Complication
Incidence
ASA I ASA II ASA III
Patients with 22 (30%) 35 (74%) 35 (81%)
complications (%)
Patients without 50 (70%) 12 (26%) 8 (19%)
complications (%)
Total 72 (44%) 47 (29%) 43 (27%)
2
P<0.01 (w for the 2  3 table).
Constant or episodic hypoxemia and hypoventilation are
common after major operations in the early and late
postoperative period.19,20 In our study respiratory com-
plications were evaluated for 6 hours after surgery: we
found that 11% of patients had hypoxemic episodes in
the postanesthesia care unit (PACU) while receiving O2
supplementation by face-mask. This is in agreement with
Russell and Graybeal21 who reported early episodes of
desaturation in 15% to 25% of postoperative patients
despite prophylactic oxygen administration. Experimen-
tal and clinical studies have demonstrated an adverse
effect of tissue hypoxia on wound healing and on
resistance to bacterial wound infections.22 Respiratory
depression may be considered one of the most important
adverse effect when considering anesthetic techniques;
although respiratory impairment was not frequently
reported after the first 3 hours and the interventions
may not be critical in nature, more accurate monitoring
and control of this parameter may be desirable in the
early and late postoperative period as well.
The second most frequent complication recorded
in our study was pain: the overall incidence of pain in
our series of patients was 24%. We did not administer
morphine or other opioids to prevent a potential
confounding factor in the evaluation of neurologic status.
We preferred the use of ketorolac 30 mg, administered by
IV route when the anesthetics were discontinued. In our
Institution, we routinely use ketorolac in craniotomy
patient although we are aware that many neurosurgeons
are concerned to use nonsteroidal anti-inflammatory
drugs in these patients. Ketorolac is considered safe and
effective for the treatment of pain after major general
surgery without increasing the level of sedation.23,24 In
our study, all patients who complained of pain during
their stay in the PACU received 30-mg supplement of
ketorolac. After the second dose, treatment with keto-
rolac provided adequate pain relief in 85% of patients.
Neurologic examination did not deteriorate during the
following 48 hours and the cerebral computed tomo-
graphic scan, performed during the 2 days after surgery,
did not show any case of postoperative bleeding in any
patient. Other studies reported higher incidence of pain
after craniotomy, varying from 40% to 84%, with the
maximum incidence occurring 12 hours after surgery.25,26
However, the studies from Quiney and colleagues25 and
De Benedittis and colleagues26 did not account for
intraoperative opioid use, nor did they compare the
r 2007 Lippincott Williams & Wilkins
Early Complications After Neurosurgery
neurosurgical patients with a control group. Differences
in the methodological approach in defining pain intensity
and the duration of the time of observation may account
Total
for discrepancies between these studies and ours. More-
92 over, the infiltration of the scalp with bupivacaine,27 used
70
in our study before incision, may have reduced post-
operative pain at least for the first few postoperative
162 hours. In support to our observation, Dunbar and
colleagues28 reported that the typical craniotomy patient
has less postoperative pain and requires less postoperative
opioids compared with other procedures. In our experience,
severe pain is not a persistent problem in craniotomy
patients because adequate pain relief was achieved in most
patients with only 2 doses of ketorolac (30 mg).
An high shivering rate (18%) was present in our
series of patients even though the patients were normo-
thermic at the arrival in the PACU. Because both core
and skin temperature contribute to thermoregulatory
control29 we assumed that the peripheral temperature
(not measured in our study) may have played a role in
determining shivering. An higher incidence of shivering
(30%) compared with our study was reported by
Wong and colleagues15 in patients receiving mostly
meperidine at the end of neurosurgical procedure.
Nefopam used in our study may have been more
protective than other drugs in shivering prevention. A
second dose of nefopam (5 mg) was effective in abolishing
shivering in all patients.
We found hypertensive episodes during the recovery
period in 13% of patients. In a previous paper by our
group hypertensive episodes were found in 29% of
patients.9 We assume that the increased dose of labetalol
given at the end of surgery (100 mg in the present study
vs. 50 mg in the previous one) may be responsible for the
lower percentage of patients having postoperative hyper-
tension in the current study.
Nausea and vomiting are the most frequently
reported complication by Manninen and colleagues5
(28% in the tumor group). Quiney and colleagues25
reported that 37% of neurosurgical patients complained
of severe nausea or vomiting, 35% of patients complained
of moderate nausea for at least 2 hours after surgery with
no statistically significant differences in the severity of
emetic symptoms when comparing patients undergoing
craniotomy at different sites. In our study, we found a
13% overall incidence of nausea and vomiting. Our result
is in agreement with that of Kathirvel et al30 who reported
an incidence of 11% of postoperative emesis in patients
receiving ondansetron 4 mg given at the time of dural
closure. Premedication with ondansetron before the end
of surgery together with the decision of not using opioids
for pain relief, may account for the lower incidence of
nausea and vomiting in our study.
Moreover, we found a positive correlation between
the ASA physical status and frequency of complications.
This finding is in agreement with previous published
paper reporting that ASA physical status was one of the
variables associated with a greater risk of developing any
PACU complications after general surgery.31 Although
233
Magni et al
the ASA status may not describe properly neurologic
patients because of the presence of tumors or because of
the presence of systemic side effects, it still remains a
reference indicator used worldwide to assess the physical
status of patients undergoing surgery and also to stratify
their risk of developing complications after neurosurgical
procedures.
We acknowledge that the power of the study was
inadequate for the comparison of severe complications in
the 2 groups of patients. For the same reason, we did not
compare separately supratentorial and infratentorial
surgery. Finally, our results are focused in the early 6
hours after surgery and we are not aware of complicating
events occurred after this period of time.
CONCLUSIONS
The recovery period after neurosurgical procedures
remains a time of great potential danger to patients.
Despite the potential advantages of early recovery after
neurosurgical procedures, postoperative complications
as respiratory, pain, hypertension, shivering, nausea,
and vomiting are still frequent. In particular, craniotomy
is associated with a very high rate of blood gas
abnormalities mainly during the first postoperative hour,
regardless of the technique used. No difference in the
composite incidence of less severe complications between
the 2 anesthetic regimens tested in this study could be
demonstrated. Because the power of the study was
adequate only for the comparison of less severe complica-
tions, we cannot draw any conclusions from the
comparison of severe events in the 2 groups of patients.
Further clinical research, aimed at analyzing the mechan-
isms contributing to excessive postoperative complication
rates, is therefore mandatory.
REFERENCES
1. Bruder N, Pellissier D, Grillot P, et al. Cerebral hyperemia during
recovery from general anesthesia in neurosurgical patients. Anesth
Analg. 2002;94:650–654.
2. Basali A, Edward JM, Kalfas I, et al. Relation between periopera-
tive hypertension and intracranial haemorrhage after craniotomy.
Anesthesiology. 2000;93:48–54.
3. Himmelseher S, Pfenninger E. Anesthetic management in neuro-
surgical patients. Curr Opin Anaesthesiol. 2001;14:483–490.
4. Bruder N, Stordeur J, Ravussin P, et al. Metabolic and hemody-
namic changes during recovery and tracheal extubation in neuro-
surgical patients: immediate versus delayed recovery. Anesth Analg.
1999;89:674–678.
5. Manninen PH, Raman SK, Boyle K, et al. Early postoperative
complications following neurosurgical procedures. Can J Anesth.
1999;46:7–14.
6. Larsen B, Seitz A, Larsen R. Recovery of cognitive function after
remifentanil-propofol anesthesia: a comparison with desflurane and
sevoflurane anesthesia. Anesth Analg. 2000;90:168–174.
7. Sneyd JR, Andrews CJH, Tsubokawa T. Comparison of propofol/
remifentanil and sevoflurane/remifentanil for maintenance of
anaesthesia for elective intracranial surgery. Br J Anaesth. 2005;94:
778–783.
8. Ozkose Z, Ercan B, Unal Y, et al. Inhalation versus total
intravenous anesthesia for lumbar disc herniation. J Neurosurg
Anesthesiol. 2001;13:296–302.
234
J Neurosurg Anesthesiol  Volume 19, Number 4, October 2007
9. 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 supra-
tentorial lesions. J Neurosurg Anesthesiol. 2005;17:134–138.
10. Talke P, Caldwell JE, Brown R, et al. A comparison of three
anesthetic techniques in patients undergoing craniotomy for
supratentorial intracranial surgery. Anesth Analg. 2002;95:430–435.
11. 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 fenta-
nyl/nitrous oxide. Anesthesiology. 1993;78:1005–1020.
12. Roberts FL, Dixon J, Lewis GTR, et al. Induction and maintenance
of propofol anaesthesia. Anaesthesia. 1998;43(suppl):14–17.
13. Thomeé R, Grimby G, Wrigth BD, et al. Rasch analysis of Visual
Analog Scale measurements before and after treatment of patello-
femoral pain syndrome in women. Scand J Rehabil Med. 1995;
27:145–151.
14. Rosa G, Pinto G, Orsi P, et al. Control of post anaesthetic shivering
with nefopam hydrochloride in mildly hypothermic patients after
neurosurgery. Acta Anaesthesiol Scand. 1995;39:90–95.
15. Wong AYC, O’Regan AM, Irwin MG. Total intravenous anaes-
thesia with propofol and remifentanil for elective neurosurgical
procedures: an audit of early postoperative complications. Eur J
Anaesthesiol. 2006;23:586–590.
16. Cashman JN, Dolin SJ. Respiratory and haemodynamic effects of
acute postoperative pain management: evidence from published
data. Br J Anaesth. 2004;93:212–223.
17. Tsui SL, Irwing MG, Wong CM, et al. An audit of the safety of an
acute pain service. Anaesthesia. 1997;52:1042–1047.
18. Brian JE Jr. Carbon dioxide and the cerebral circulation.
Anesthesiology. 1998;88:1365–1386.
19. Rose DK, Cohen MM, Wigglesworth DF, et al. Critical respiratory
events in the postanesthesia care unit. Patient, surgical, and
anesthetic factors. Anesthesiology. 1994;81:410–418.
20. Rosenberg J, Rasmussen GI, Wojdemann KR, et al. Ventilatory
pattern and associated episodic hypoxaemia in the late post-
operative period in the general surgical ward. Anaesthesia. 1999;
54:323–328.
21. Russell GB, Graybeal JM. Hypoxemic episodes of patients in a
postanesthesia care unit. Chest. 1993;104:899–903.
22. Kehlet H, Rosenberg J. Late post-operative hypoxaemia and organ
dysfunction. Eur J Anaesthesiol Suppl. 1995;10:31–34.
23. Feld JM, Laurito CE, Beckerman M, et al. Non-opioid analgesia
improves pain relief and decreases sedation after gastric bypass
surgery. Can J Anaesth. 2003;50:336–341.
24. Forrest JB, Camu F, Greer IA, et al. Ketorolac, diclofenac, and
ketoprofen are equally safe for pain relief after major surgery.
Br J Anaesth. 2002;88:227–233.
25. Quiney N, Cooper R, Stoneham M, et al. Pain after craniotomy.
A time for reappraisal? Br J Neurosurg. 1996;10:295–299.
26. De Benedittis G, Lorenzetti A, Migliore M, et al. Postoperative pain
in neurosurgery: a pilot study in brain surgery. Neurosurgery.
1996;38:466–470.
27. Bloomfield EL, Schubert A, Secic M, et al. The influence of scalp
infiltration with bupivacaine on hemodynamics and postoperative
pain in adult patients undergoing craniotomy. Anesth Analg.
1998;87:579–582.
28. Dunbar PJ, Visco E, Lam AM. Craniotomy procedures are
associated with less analgesic requirements than other surgical
procedures. Anesth Analg. 1999;88:335–340.
29. Cheng C, Matsukawa T, Sessler DI, et al. Increasing mean skin
temperature linearly reduces the core-temperature thresholds for
vasoconstriction and shivering in humans. Anesthesiology. 1995;82:
1160–1168.
30. Kathirvel S, Dash HH, Bhatia A, et al. Effect of prophylactic
ondansetron on postoperative nausea and vomiting after elective
craniotomy. J Neurosurg Anesthesiol. 2001;13:207–212.
31. Wolters U, Wolf T, Stutzer H, et al. ASA classification and
perioperative variables as predictors of postoperative outcome.
Br J Anaesth. 1996;77:217–222.
r 2007 Lippincott Williams & Wilkins