The use of remifentanil during neurosurgery has similar effects on cerebral blood flow and carbon dioxide reactivity when compared with other opioids [1-3]. The possibility to quickly adjust the concentration of this opioid at the effecter site is appealing and allows rapid response to changes in nociceptive stimuli. Though postoperative pain is usually not severe in craniotomy, the use of remifentanil imposes the need for other opioids for postoperative pain relief . This could possibly reduce the postoperative advantages of this short-acting opioid. One of the advantages mentioned is the possibility to detect postoperative bleeding or other complications earlier because of the rapid recovery. We investigated the postoperative effect of a small dose of piritramide given before the end of the operation in patients treated with a remifentanil-based protocol, as compared with a fentanyl-based protocol without additive piritramide during operation. Fentanyl has a long history of predictable influences on intracranial pressure and haemodynamics in neurosurgery. Remifentanil is reported to have the same safety profile [5,6].
The institutional review board of the VU University Medical Centre approved the protocol for this single centre prospective randomized double-blinded study. Based on a difference of 2 cm (standard deviation (SD): 2 cm) on the primary effect measure (visual analogue pain score), using standard power analysis (significance level of 5%, power of 80%), the appropriate sample size was established at 16 patients per arm of the study. We included elective patients with an ASA classification I or II, a Glasgow Coma Score of 15, an expected duration of surgery of more than 1 h and aged from 18 to 70 yr old. Patients were excluded if they had contraindications for the study medication. After informed consent they were randomized to the two study groups by a sealed envelope process. All patients and observers (nurse or resident anaesthesiologist) were blinded until leaving the recovery room at the end of the study.
Every patient received midazolam 7.5 mg orally 1 h before surgery. An unblinded anaesthesiologist performed the anaesthesia. For induction we used thiopental 4-6 mg kg−1 and rocuronium 0.6 mg kg−1. Maintenance was achieved using isoflurane 0.6-1 MAC in a 2: 1 nitrous oxide: oxygen mixture and rocuronium by pump to maintain one twitch of a train-of-four (TOF-Watch®, Organon, The Netherlands). Administration of rocuronium was stopped 30 min before the end of surgery. The remifentanil-piritramide group (Group A) received anaesthesia using remifentanil by continuous infusion. The infusion rate at induction was 50 μg kg−1 h−1 and was lowered after intubation to 12.5 μg kg−1 h−1. If there was an autonomic response during surgery, such as a change in heart rate, hypertension (>20% deviation from preoperative value), movement or sweating, the rate was adjusted by steps of 5 μg kg−1 h−1. No bolus was allowed. In the case of hypotension (>20% deviation from preoperative value) the pump rate was lowered by the same steps. After closure of the dura mater (30-45 min before the end of surgery), this group received a low dose of piritramide (0.1 mg kg−1) intravenously (i.v.). Remifentanil was stopped after the removal of the Mayfield head holder. The fentanyl group (Group B), received a bolus of 3 μg kg−1 fentanyl at induction followed by a continuous infusion of 3 μg kg−1 h−1. The rate adjustments were based on the same criteria as described above, by changing the infusion rate by 1 μg kg−1 h−1. A bolus of 50 μg was given when the effect was too slow according to the anaesthesiologist. The pump was stopped after closure of the dura mater. No other opioids were administered. Isoflurane and nitrous oxide were replaced by a high flow of oxygen after removal of the Mayfield head holder as in Group A. All patients were extubated when they were able to swallow and cough, and spontaneous respiration was adequate (as shown by normalization of end-tidal carbon dioxide concentration). Both groups received 1 g of paracetamol rectally postoperatively. When the patient requested more pain medication, or as indicated by the blinded observer, 2.5 mg piritramide was given i.v.
A VAS for pain, the Glasgow Coma Score, a modified Aldrete Score and piritramide consumption were evaluated every half an hour up to 120 min postoperatively by the blinded observer. Using the same time frame, a sample of arterial blood was analysed for carbon dioxide tension (PaCO2) (Rapid lab 865®, Chiron Diagnostics, The Netherlands). All observers were instructed how to perform the scores in a uniform manner at the start of the study.
Data were analysed using SPSS 9.0 for windows. Comparisons for baseline differences and treatment effects were analysed using t-test for independent samples for parametric data, and χ2-test (with continuity correction) and the U-test for non-parametric data. Univariate analyses of variance were performed to control for possible confounding effects. For all outcome measures the two-sided significance level was set at 5%.
Throughout the year 2002, 42 patients for elective craniotomy were enrolled in this study. Twenty-one patients were included in each group. After randomization we excluded two patients, one in Group A with a pituitary tumour who received a transsphenoidal approach and one patient in Group B with a supratentorial mass, who had pulmonary emphysema on chest X-ray which had not been diagnosed before and which was a contraindication to the use of nitrous oxide. Patient characteristics (Table 1) were comparable in both groups except for age. Age was therefore incorporated in the analysis of vari-ance model as a confounding variable for all analyses. Further analysis revealed that older patients (50 yr or more) experience significantly more pain at all measurement points than younger patients (P = 0.003-0.007). Age was incorporated as a confounding variable and used as an interaction term in the analysis of variance model for pain scores. The indications for surgery are listed in Table 2. No early or late postoperative complication was reported. Time from removing the headpins to extubation did not differ significantly between groups (extubation time Group A: 10.03 ± 4.12 min, extubation time Group B: 11.45 ± 5.20 min, P = 0.343), although in the fentanyl group two patients received 2.5 mg of i.v. nalorphine before extubation to increase the respiratory rate. The postoperative results are listed in Table 3 and show a higher concentration of PaCO2 in Group B, which was close to significant at 30 min (P = 0.053) and 90 min (P = 0.060). No significant difference was found for pain intensity between the groups. However, 13 patients in the remifentanil/piritramide group received one or more extra doses of 2.5 mg piritramide as opposed to seven patients in the fentanyl group (P = 0.113).
The present study did not find a more rapid recovery in the remifentanil-piritramide protocol group. Both groups were comparable on the modified Aldrete score. The Glasgow Coma Score showed a non-significant tendency towards lower scores in the fentanyl group. We administered a smaller dose of piritramide than usually recommended in patients receiving remifentanil at the end of the procedure. The total extra dose of piritramide in the remifentanil-piritramide group required postoperatively for pain management might explain this lack of difference in recovery. We showed however, that fentanyl can be used by continuous infusion for a long period of time with acceptable time to extubation when stopping the infusion at a suitable time despite the long context sensitive half life. Lack of differences in time to extubation was reported earlier [6,7]. Talke and colleagues  proved this to be independent of the use of isoflurane. Both Balakrishnan and colleagues , and Guy and colleagues  reported mean times to extubation of 5 min after remifentanil using comparable protocols whereas we found 10 min in our study. This can possibly be explained by the piritramide used or the fact that both of our study groups received the same partial pressures of isoflurane in nitrous oxide. The MAC of isoflurane combined with remifentanil is believed to be lower compared to the combination of isoflurane and fentanyl [5,9]. Our findings, however, are in accordance with the findings of Sneyd and colleagues , who used remifentanil combined with propofol and nitrous oxide. They reported a mean extubation time of 11 min.
To our knowledge postoperative PaCO2 has not been assessed in this type of study. As expected, PaCO2 was higher when using fentanyl, but the difference compared to remifentanil was not significant. Also the levels found were too low to have clinical relevance in our opinion.
We have confirmed earlier reports that patients need more and earlier additional opioids postoperatively for pain management after remifentanil compared with fentanyl [4,6,11]. The dose of piritramide we used during operation could not reduce this difference. However, the difference found in pain intensity between both groups, could be explained by an inadequate dose of piritramide or the influence of age in this study. We chose this small dose because in our experience a higher dose prolonges the time to extubation.
Although randomization was performed in this study, a sampling artefact in age difference between groups did occur. If age was left out of the analysis of variance, significant differences would be found for pain score at 0 and 90 min in favour of the fentanyl group, and for PaCO2, with higher levels at 30 and 90 min in the fentanyl group.
In conclusion, we have shown that 0.1 mg kg−1 of piritramide given before the end of an elective craniotomy using a remifentanil-based technique has no postoperative advantage compared with the use of a continuous infusion of fentanyl as a sole-opioid regime. The remifentanil-piritramide group experienced more pain, resulting in higher postoperative piritramide consumption with a tendency to a higher Glasgow Coma Score while recovery was comparable. Higher arterial carbon dioxide concentrations measured in the fentanyl group were acceptable and did not increase during the recovery period.
We would like to thank all the people working in the recovery room, for their effort, enthusiasm and patience in the scoring of our patients.
1. Warner DS, Hindman BJ, Todd MM, et al
. Intracranial pressure and hemodynamic effects of remifentanil
versus alfentanil in patients undergoing supratentorial craniotomy
. Anesth Analg
2. Ostapkovich ND, Baker KZ, Fogarty-Mack P, et al
. Cerebral blood flow and reactivity is similar during remifentanil
O and fentanyl
O anesthesia. Anesthesiology
3. Baker KZ, Ostapkovich ND, Sisty MB, et al
. Intact cerebral blood flow reactivity during remifentanil
/nitrous oxide anesthesia. J Neurosurg Anesth
4. Verchère E, Grenier B, Mesli A, et al
. Postoperative pain
management after supratentorial craniotomy
. J Neurosurg Anesth
5. Warner DS. Experience with remifentanil
in neurosurgical patients. Anesth Analg
6. Guy J, Hindman BJ, Baker KZ, et al
. Comparison of remifentanil
in patients undergoing craniotomy
for space-occupying lesions. Anesthesiology
7. Coles JP, Leary TS, Monteiro JN, et al
. Propofol anesthesia for craniotomy
: a double-blind comparison of remifentanil
, alfentanil and fentanyl
. J Neurosurg Anesth
8. Talke P, Caldwell JE, Brown R, et al
. A comparison of three anesthetic techniques in undergoing craniotomy
for supratentorial intracranial surgery. Anesth Analg
9. Balakrishnan G, Raudzens P, Samra SK, et al
. A comparison of remifentanil
in patients undergoing surgery for intracranial mass lesions. Anesth Analg
10. Sneyd JR, Whaley A, Dimpel HL, et al
. An open randomized comparison of alfentanil, remifentanil
and alfentanil followed by remifentanyl in anaesthesia for craniotomy
. Br J Anaesth
11. Gerlach K, Uhlig T, Hüppe M, et al
-propofol versus sufentanil-propofol anaesthesia for supratentorial craniotomy
: a randomized trial. Eur J Anaesth