1. European Journal of Anaesthesiology 2006; 23: 586–590
© 2006 Copyright European Society of Anaesthesiology
doi: 10.1017/S0265021506000214
Original Article
Total intravenous anaesthesia with propofol and remifentanil
for elective neurosurgical procedures: an audit of early
postoperative complications
A. Y. C. Wong, A. M. O’Regan, M. G. Irwin
University of Hong Kong, Queen Mary Hospital, Department of Anaesthesiology, Hong Kong
Summary
Background and objectives: This was a prospective audit to assess the incidence and characteristics of early postopera-
tive complications in the recovery room in extubated patients after elective neurosurgical procedures using propo-
fol and remifentanil-based total intravenous anaesthesia. Methods: Vital signs (temperature, conscious level,
respiratory rate, oxygen saturation, pulse and blood pressure) and postoperative complications (shivering, nausea,
vomiting and cardiorespiratory) were analysed in 145 adult patients over a 1-yr period. Results: The overall shiver-
ing, postoperative nausea and vomiting and postoperative hypertension (systolic blood pressure more than 25% of
the preoperative value) incidences were 30.3%, 16.6% and 35.2%, respectively. Fifty-one percent of the patients
had at least one of the above complications. The complication rates were found to be widely different among vari-
ous types of neurosurgery. The surgical procedures were divided into five groups: supratentorial craniotomy, poste-
rior fossa craniotomy, intracranial vascular procedures, transphenoidal hypophysectomy and extracranial procedures.
Median extubation time was similar in all groups and patients were fully conscious with no hypoxia in the recovery
room. The intracranial vascular group had the highest shivering and postoperative nausea and vomiting rates
(58.8% and 29.4%, respectively). In the supratentorial craniotomy group, 46% of the patients had hypertension.
The overall complication rate (presence of any complications) was highest in the supratentfial craniotomy (55.4%),
posterior fossa craniotomy (75%) and intracranial vascular (76.5%) groups. Shivering and overall complication
rate was significantly related to the anaesthetic time (P 0.001 and 0.02, respectively). Conclusions: Despite the
potential advantages of total intravenous anaesthesia in titratability, rapid return of consciousness and reduced res-
piratory complications, making it suitable for planned extubation at the end of neurosurgery, the postoperative
complications of shivering, postoperative nausea and vomiting and hypertension were still high.
Keywords: ANAESTHESIA INTRAVENOUS; INTRAVENOUS ANAESTHETICS, propofol; ANALGESICS OPIOID,
remifentanil; NEUROSURGICAL PROCEDURES, craniotomy, supratentorial, posterior fossa, extracranial, intracranial vascular,
transphenoidal hypophysectomy; COMPLICATIONS POSTOPERATIVE, shivering, nausea, vomiting, hypertension.
Introduction These are mostly connected with nausea and vomiting,
shivering, respiratory and cardiovascular derange-
The incidence of postoperative complications in the ments. The overall incidence of early postoperative
recovery room is reported to be as high as 30% [1–4]. complications in the neurosurgical patients has been
reported to be higher [5] but the actual anaesthetic
Correspondence to: Andrew Y. C. Wong, Department of Anaesthesiology, Univer- choice was not specified. In our institution total intra-
sity of Hongkong, F2, Queen Mary Hospital, 102 Pokfulam Road, Hong Kong.
E-mail: wongyca@netvigator.com; Tel: 852 2855 5791; Fax: 852 2855 3384 venous anaesthesia (TIVA) has become the standard
Accepted for publication 16 January 2006 EJA 3579 anaesthetic technique. The purpose of this prospective
First published online 1 March 2006 single institution audit was to assess the incidence and

2. TIVA with propofol and remifentanil for elective neurosurgical procedures 587
characteristics of early postoperative complications in Statistical analysis was performed using Microsoft
adult patients following elective neurosurgical proce- Excel 2002 and SPSS Version 11 (SPSS Inc., Chicago,
dures using propofol and remifentanil-based TIVA. IL, USA). For parametric data, t-test (e.g. anaesthetic
time, extubation time, respiratory rate, blood pressure
(BP), temperature), analysis of variance (ANOVA)
Methods
(e.g. BP among different groups) and Tukey’s test for
The study was approved by our local institutional post hoc comparisons were used. For non-parametric
ethics review board. It was prospective and all elective data, 2-test with continuity correction (e.g. PONV,
adult neurosurgical patients from December 2002 to shivering) for initial and post hoc comparisons and
November 2003 who had received TIVA with propo- Kruskal–Wallis test (e.g. GCS and oxygen saturation
fol and remifentanil and immediate postoperative among different groups) were used. Logistic regression
extubation were included. Anaesthesia was induced was used to test for correlation between two paramet-
with a target-controlled infusion (TCI) of propofol ric samples. P 0.05 was considered significant.
(Diprifusor™; Astrazeneca, UK) and a remifentanil
infusion (0.3 µg kg 1 min 1). Muscle paralysis was
Results
achieved with rocuronium prior to tracheal intubation.
Anaesthesia was maintained with propofol TCI and During the study period, 148 adult elective patients
a remifentanil infusion (0.15–0.3 µg kg 1 min 1) with were scheduled for neurosurgery. They all received
an atracurium or cisatracurium infusion titrated TIVA for their neurosurgical operation. On emer-
to maintain muscle relaxation (one twitch, train of gence, one patient convulsed and two had decreased
four (Innervator; Fisher & Paykel Electronics Ltd, limb power. They were re-anaesthetized, kept intu-
Auckland, New Zealand)). Neostigmine and atropine bated, sent to the computed tomography (CT) suite
were used to reverse paralysis at the end of surgery. and excluded from the study. One hundred and fifty
Overall administration and adjustment of TIVA was patients were included in the review. Thirteen (9%)
left to the discretion of the attending anaesthesiologist had pre-existing hypertension. The overall shivering,
according to the usual clinical parameters without PONV and postoperative hypertension (systolic blood
neurophysiological depth of anaesthesia monitoring. pressure (SBP) more than 25% of the preoperative
Due to the rapid offset of remifentanil, in some value) were 30.3%, 16.6% and 35.2%, respectively.
patients a transition analgesic was administered close Fifty-one percent of the patients had at least one of the
to the end of surgery for initial postoperative pain above complications. The complication rates were
relief. Transition analgesics used were meperidine, found to be widely different among various types of
fentanyl or morphine (n 68, 42 and 9, respectively) neurosurgery. In order to further analyse the charac-
administered at the discretion of the attending anaes- teristics of the postoperative complications, the neu-
thesiologist. A standard form was completed by a rosurgical procedures were divided into five groups:
trained recovery room nurse and was used to collate supratentorial craniotomy (SC), posterior fossa crani-
patient information for the first hour or until discharge otomy (PC), transphenoidal hypophysectomy (TH),
from the recovery room, should this be sooner. Tym- intracranial vascular (IV) and extracranial (EC)
panic membrane temperature (ThermoScan, B Braun, (Table 1). There was no significant patient character-
Germany) was measured on admission. The following istic difference among the five groups. All the patients
data was recorded on recovery room admission and had normal temperature, GCS (median 14–15), respi-
every 15 min thereafter: direct arterial pressure, pulse ratory rate (median respiratory rate of up to 21 min 1)
oximeter oxygen saturation (SaO2) and respiratory rate and SaO2 (median SaO2 99–100%) in the recovery
(Viridia physiological monitor; Hewlett Packard, room. The anaesthetic time, time to extubation and
USA), Glasgow coma score (GCS and power in all change in SBP from baseline collected in the recovery
limbs), shivering (yes or no) and postoperative nausea room are shown in Table 2. PC and TH groups took
and/or vomiting (postoperative nausea and vomiting; the most and least anaesthetic time, respectively.
yes or no) every 15 min. Other information on the data Median extubation time was up to 12 min. All groups
sheet included patient characteristics data, pre-existing showed an increase in SBP from baseline. Using logis-
medical disease such as hypertension, surgical pathol- tic regression, the extubation time was found to be
ogy and operation, preoperative assessment (including significantly correlated and directly proportional with
physiological information and neurological assess- age (P 0.02) and anaesthetic time (P 0.05).
ment), anaesthetic time and extubation time (defined Throughout the recovery room stay, the number of
as the time from application of head dressing to safe patients with shivering, PONV, SBP 25% and 30%
extubation when the patient had opened his or her greater than the preoperative value in each surgical
eyes, obeyed commands and resumed adequate respi- group is shown in Table 3. The rates of PONV ranged
ration of more than 8 breaths min 1). from 7.1% in group EC to 29.4% in group IV but
© 2006 Copyright European Society of Anaesthesiology, European Journal of Anaesthesiology 23: 586–590

3. 588 A. Y. C. Wong et al
Table 1. Neurosurgical groups (number of patients).
Known
Groups n Diagnosis hypertension ASA (mode)
SC 74 Tumour (73), abscess (1) 9 2
PC 20 Tumour (16), syringomyelia (2), cyst (1), CSF leak (1) 1 2
IV 17 AVM (13), aneurysm (2), trigeminal neuralgia (2) 1 2
TH 20 Pituitary tumour (20) 0 3
EC 14 Skull defect (6), skull osteoma (1), spine tumour (5), spine AVM (2) 2 1
P 0.60 0.06
SC: supratentorial craniotomy; PC: Posterior fossa craniotomy; IV; intracranial vascular; TH: transphenoidal hypophysectomy; EC: extracranial.
Table 2. Duration of anaesthesia, time to extubation and SBP an early postoperative overall complication rate of
difference (SBPd) on arrival in the recovery room, in each surgical
group. SBPd is the change in SBP from the preoperative value in 51% which is close to the 54.4% reported by
percentage. Manninen and colleagues in 1999, in which the
anaesthetic technique was not stated [5]. Studies
Anaesthesia Extubation comparing TIVA with inhalational opioid anaesthe-
Groups time (min) time (min) SBPd (%) sia are warranted.
Patients after neurosurgery are prone to PONV [10].
SC 309.5 99.4 5–11 20.5 19.4 In a review of early postoperative complications fol-
PC 385.6 164 4–9 18.6 13.5
IV 346.2 98 5–9 14.2 13.4 lowing neurosurgical procedures, PONV was the
TH 220.6 61.4 6–12 13.5 12.8 commonest complication (38.7%), although the
EC 262.4 163.1 5–12 9.7 12.3 anaesthesia technique was not mentioned in that study
P 0.01 0.50 0.11 [5]. In Fabling and colleagues study on neurosurgical
SC: supratentorial craniotomy; PC: posterior fossa craniotomy; patients anaesthetized with fentanyl and isoflurane,
IV: intracranial vascular; TH: transphenoidal hypophysectomy; 48% and 74% of the patients in the ondansetron and
EC: extracranial. placebo groups, respectively, reported nausea during
Values are mean SD or interquartile range. SBPi is the percentage the first hour postoperatively [11]. TIVA with propofol
increase in SBP from baseline.
has been found to produce less PONV compared to
no significant difference was detected. Post hoc com- inhalational anaesthesia [12–15] in non-neurosurgical
parisons always showed significantly lower compli- patients. This may explain the lower incidence in
cation rates in TH and EC groups. With all surgical our study, but in the IV group, the rate is relatively
groups combined, shivering in the recovery room high (29.4%). Retching and vomiting in such
was more common after long anaesthesia (P 0.01) patients may cause devastating haemorrhagic compli-
and was not affected by the choice of transition anal- cations and a vigorous antiemetic regime should,
gesic. The overall complication rate (presence of any therefore, be adopted in this group. A high shivering
complications) was also found to be significantly rate was apparent in most groups, even though the
correlated to duration of anaesthesia (P 0.02). patients arriving in the recovery room were normo-
thermic (normal tympanic core temperature). The
Discussion peripheral or skin temperature was, however, not
measured in the study. Both core and skin tempera-
TIVA with propofol and remifentanil allows rapid tures contribute to thermoregulatory control [16]. The
and predictable titration of anaesthesia, with swift skin temperature was not measured and may have a
recovery of consciousness and respiration even after more important contribution than core temperature
prolonged administration. Propofol also has the addi- to postoperative shivering. Shivering was related to
tional benefits of reducing PONV, cerebral metabolic the type of surgery (P 0.02), with the highest rate
oxygen consumption, IV pressure and increasing cere- (58.5%) in the IV group. Longer anaesthetic time was
bral perfusion pressure [6–9]. TIVA with propofol and associated with a higher shivering rate (P 0.01).
remifentanil is commonly used in neuroanaesthesia This may be due to inadequate insulation of the
and has become the standard anaesthetic technique for patient’s limbs, particularly the hand with the pulse
elective neurosurgery in our institution. oximeter and intravenous lines. Frequent exposure
The main complications observed were PONV, of the limb to allow inspection of the lines is impor-
shivering and hypertension (SBP 25% greater than tant to detect loose connections and to make sure
the baseline value). Of the 145 patients, 74 had at that TIVA is being delivered properly. Although
least one of these three complications, giving rise to meperidine has been reported to be more effective
© 2006 Copyright European Society of Anaesthesiology, European Journal of Anaesthesiology 23: 586–590

4. TIVA with propofol and remifentanil for elective neurosurgical procedures 589
Table 3. Number of patients with complications during the recovery room stay (%).
Groups Shivering PONV SBPd 25% 30% 1 complication
SC 24 (32.4%) 11 (14.9%) 34 (46%) 28 (37.8%) 41 (55.4%)
PC 6 (30%) 4 (20%) 8 (14%) 7 (35%) 15 (75%)
IV 10 (58.8%) 5 (29.4%) 6 (35.3%) 3 (17.6%) 13 (76.5%)
TH 1 (5%) 3 (15%) 4 (20%) 3 (15%) 7 (35%)
EC 3 (21.4%) 1 (7.1%) 1 (7.1%) 1 (7.1%) 3 (21.4%)
P 0.02 0.53 0.05 0.08 0.01
SC: supratentorial craniotomy; PC: posterior fossa craniotomy; IV: intracranial vascular; TH: transphenoidal hypophysectomy; EC: extracranial.
SBPd is the percentage increase in SBP from baseline.
than other opioids at reducing postoperative shiver- characteristics and complications that may be associ-
ing, it was not found to be associated with a lower ated with this technique. Another limitation with our
shivering rate than fentanyl or morphine in the pres- study is that the results are from a single-centre study.
ent study [17,18]. A multi-centre study on this topic is warranted.
Postoperative hypertension is frequent after neuro- To summarize, the principal early postoperative
surgery [19]. In one study, more than 90% of patients complications found in our study were PONV, shiv-
in the recovery room had a BP exceeding the preoper- ering and hypertension. There was an overall com-
ative value by 20% [20]. A propofol and remifentanil plication rate of 51%, but a significant difference
combination facilitates rapid titration of anaesthesia among the groups. Complications are more common
and stable intraoperative haemodynamics. However, in in intracranial procedures than EC ones: highest in IV
the recovery room 46% of the patients in the SC group surgery (76.5%), and lowest in EC surgery (21.4%).
had hypertension. Hypertension has been associated Although the findings with our TIVA technique are
with an increased risk of postoperative intracra- similar to previous reports in non-neurosurgical
nial haemorrhage and cerebral hyperaemia [21,22]. patients in terms of recovery, absence of respiratory
Anaesthesiologists using a TIVA technique should complications and reduced PONV, the incidence of
consider employing antihypertensive therapy more other complications, especially hypertension, remains
liberally during emergence from anaesthesia and in high. Further research into the mechanisms con-
the recovery room, particularly in this subgroup. tributing to a high postoperative complication rate
Postoperative hypertension is significantly more com- in intracranial procedures especially IV surgery is
mon in the SC group than in the TH and EC groups warranted.
(P 0.03 and 0.01, respectively), suggesting a
relationship between intracranial surgery and post-
operative hypertension. References
Our results showed that patients undergoing IV
procedures such as PC and IV always produced higher 1. Kluger MT, Bullock MFM. Recovery room incidents: a
review of 419 reports from the anaesthetic incident monitor-
complication rates than the EC ones. The mecha-
ing study (AIMS). Anaesth Intens Care 2002; 57: 1060–1066.
nisms leading to increased complications after IV 2. Zelcer J, Wells DG. Anaesthetic-related recovery room
procedures warrant further research. Overall, long complications. Anaesth Intens Care 1987; 15: 168–174.
anaesthetic time was positively correlated to an 3. Hines R, Barash PG, Watrous G, O’Connor T. Complica-
increased incidence of postoperative complications tions occurring in the postanaesthesia care unit: a survey.
(P 0.02). This may be reflected in the IV group Anesth Analg 1992; 74: 503–509.
which had the second longest anaesthetic time, high 4. Rose DK. Recovery room problems or problems in the
shivering and hypertension rates. PACU. Can J Anesth 1996; 43: 116–122.
Due to perceived pharmacokinetic and dynamic 5. Manninen PH, Raman SK, Boyle K, El-Beheiry H. Early
advantages, TIVA with propofol and remifentanil has postoperative complications following neurosurgical pro-
become the standard anaesthetic technique in neuro- cedures. Can J Anesth 1999; 46: 7–14.
6. Petersen KD, Landsfeldt U, Cold GE et al. Intracranial
surgery in our institution. We did not include a con-
pressure and cerebral hemodynamic in patients with cere-
trol group and do not know our complication rates bral tumours: a randomised prospective study of patients
should a different anaesthetic technique (e.g. inhala- subjected to craniotomy in propofol-fentanyl, isoflurane-
tional) be used, all other factors being equal. fentanyl, or sevoflurane-fentanyl anesthesia. Anesthesiology
Consequently we have had to rely on previously 2003; 98: 329–336.
published data for comparison. We feel, however, that 7. Todd MM, Warner DS, Sokoll MD et al. A prospective, com-
this study does help to elucidate the early recovery parative trial of three anesthetics for elective supratentorial
© 2006 Copyright European Society of Anaesthesiology, European Journal of Anaesthesiology 23: 586–590

5. 590 A. Y. C. Wong et al
craniotomy. Propofol/fentanyl, isoflurane/nitrous oxide, and 15. Chittleborough MC, Osborne FA, Rudkin GE et al. Double-
fentanyl/nitrous oxide. Anesthesiology 1993; 78: 1005–1020. blind comparison of patient recovery after induction with
8. Kahveci FS, Kahveci N, Alkan T, Goren B, Korfali E, propofol or thiopentone for day-case relaxant general anes-
Ozluk K. Propofol versus isoflurane anesthesia under hypo- thesia. Anaesth Intens Care 1992; 20: 169–173.
thermic conditions: effects on intracranial pressure and local 16. Cheng C, Matsukawa T, Sessler DI et al. Increasing mean
cerebral blood flow after diffuse traumatic brain injury in skin temperature linearly reduces the core-temperature
the rat. Surg Neurol 2001; 56: 206–214. thresholds for vasoconstriction and shivering in humans.
9. Cenic A, Craen RA, Lee TY, Gelb AW. Cerebral blood vol- Anesthesiology 1995; 82: 1160–1168.
ume and blood flow responses to hyperventilation in brain 17. Pauca AL, Savage RT, Simpson S, Roy RC. Effect of pethi-
tumors during isoflurane or propofol anesthesia. Anesth dine, fentanyl and morphine on postoperative shivering in
Analg 2002; 94: 661–666. man. Acta Anaesth Scand 1984; 28: 138–143.
10. Gan TJ, Meyer T, Apfel C, Chung F et al. Consensus guide- 18. Alfonsi P, Sessler D, Du Manoir et al. The effects of meperi-
lines for managing postoperative nausea and vomiting. dine and sufentanil on shivering threshold in postoperative
Anesth Analg 2003; 97(1): 62–71. patients. Anesthesiology 1998; 89: 43–48.
11. Fabling JM, Gan TJ, El-Moalem HE, Warner DS, Borel CO. 19. Olsen KS, Pedersen CB, Madsen JB, Ravn LI, Schifter S.
A randomised, double-blind comparison of ondansetron Vasoactive modulators during and after craniotomy: relation
versus placebo for prevention of nausea and vomiting to postoperative hypertension. J Neurosurg Anesth 2002; 14:
after infratentorial craniotomy. J Neurosurg Anesth 2002; 14: 171–179.
102–107. 20. Gibson BE, Black S, Maass L, Cucchiara RF. Esmolol for
12. Gupta A, Stierer T, Zuckerman R et al. Comparison of the control of hypertension after neurologic surgery. Clin
recovery profile after ambulatory anesthesia with propofol, Pharmacol Ther 1988; 44: 650–653.
isoflurane, sevoflurane and desflurane: a systematic review. 21. Basali A, Mascha EJ, Kalfas I, Schubert A. Relation between
Anesth Analg 2004; 98(3): 632–641. perioperative hypertension and intracranial hemorrhage after
13. Hofer CK, Zollinger A, Buchi R et al. Patient well-being craniotomy. Anesthesiology 2000; 93: 48–53.
after general anaesthesia: a prospective, randomized, con- 22. Bruder N, Pellissier D, Grillot P, Gouin F. Cerebral hyper-
trolled multi-centre trial comparing intravenous and inhala- emia during recovery from general anesthesia in neurosur-
tion anaesthesia. Brit J Anaesth 2003; 91(5): 631–637. gical patients. Anesth Analg 2002; 94: 650–654.
14. Raftery S, Sherry E. Total intravenous anesthesia with
propofol and alfentanil protects against postoperative nau-
sea and vomiting. Can J Anaesth 1992; 39: 37–40.
© 2006 Copyright European Society of Anaesthesiology, European Journal of Anaesthesiology 23: 586–590