Propofol/remifentanil vs sevoflurane/remifentanil for long lasting surgical procedures: a randomised controlled trial


Dr Jan Höcker


We compared the haemodynamics, emergence and recovery characteristics of total intravenous anaesthesia using propofol/remifentanil with sevoflurane/remifentanil anaesthesia, under bispectral index guidance, in 103 patients undergoing surgical procedures lasting > 3.5 h. Time to tracheal extubation was significantly shorter in the propofol group than in the sevoflurane group (mean (SD) 8.3 (3.5) min vs 10.8 (4.6) min, respectively; p = 0.0024), but further recovery was comparable in both groups. There were no significant differences in haemodynamic parameters, intensity of pain or postoperative nausea and vomiting. During and after anaesthesia of comparable depth for long lasting surgical procedures, both propofol/remifentanil and sevoflurane/remifentanil enable haemodynamic stability and fast emergence. The shorter time to extubation in the propofol group does not offer a relevant clinical advantage.

Rapid emergence from anaesthesia and postoperative recovery of cognitive function even after long lasting surgical procedures, as well as haemodynamic stability, are important requirements of modern anaesthetics. Generally, both propofol and sevoflurane meet these criteria [1]. Comparisons of emergence and recovery after propofol and sevoflurane anaesthesia have produced differing results dependent on patients' status, surgical procedures and the analgesics used [2, 3]. For total intravenous anaesthesia with propofol and remifentanil, mean (SD) times to tracheal extubation of between 3.8 (1.9) min and 11.0 (3.7) min [4,5] have been reported. Extubation times after sevoflurane anaesthesia have also greatly differed, ranging from 10.0 (3.2) min to 18.0 (17.0) min [6, 7]. One reason for this variability and a common problem of pharmacological studies is the non-comparable depth of anaesthesia between study groups. Many authors have adjusted applied dosages of anaesthetics to the mean dosages recommended by the manufacturer, and have estimated depth of anaesthesia only by clinical criteria such as blood pressure, heart rate and similar parameters [2]. In contrast, application of electroencephalogram-based monitoring of depth of anaesthesia allows adjustment of the delivery of anaesthetics according to a defined target.

It has been shown that prolonged anaesthesia results in prolonged emergence and postoperative recovery [8]. Therefore, especially during long lasting surgical procedures, advantages and recovery profiles of different modern anaesthetics may become more apparent. This study was designed to compare emergence and recovery characteristics, patient satisfaction and analgesic demand after long lasting surgical procedures (> 3.5 h) using propofol/remifentanil vs sevoflurane/remifentanil during bispectral index (BIS) controlled anaesthesia.


After approval by the institutional review board and written informed consent, 103 patients of ASA physical status 1–2, aged 18–65 years, and scheduled for elective abdominal and urological surgical procedures estimated to last > 3.5 h, were enrolled in the study. Exclusion criteria were pregnancy, abuse of drugs or alcohol, hyperlipoproteinaemia or routine use of sedative drugs. The patients were randomly assigned by opening a sealed envelope to receive propofol by target controlled infusion or sevoflurane.

All patients received 20 mg chlorazepate orally the evening before surgery, and 7.5 mg midazolam orally 45 min before surgery. Standard monitoring was used, including relaxometry and BIS (BIS XP, Aspect Medical Systems, Natick, MA). Five minutes before induction of anaesthesia, psychomotor and cognitive function were assessed using the Trieger Dot Test and Digit Symbol Substitution Test [9, 10]. For the Trieger Dot Test, patients have to connect printed dots on a paper sheet within 1 min; missed dots and the distance line–dot (mm) are noted [11]. For the Digit Symbol Substitution Test, patients were asked to replace digits with appropriate symbols located in a legend at the top of the page within 1 min [11]. After adequate pre-oxygenation, anaesthesia was induced with 0.3−1 etomidate and 0.5 μ−1.min−1 remifentanil infusion. Tracheal intubation was facilitated with 0.6−1 rocuronium. After tracheal intubation, the lungs were ventilated with 40% oxygen in air with a fresh gas flow of 1 l.min−1 until the end of surgery, adjusted to achieve an end-expiratory Pco2 of 4.7–5.3 kPa. The remifentanil infusion was initially adjusted to 0.25 μ−1.min−1. In the propofol group, propofol-target controlled infusion (Diprifusor®, Zeneca Pharmaceuticals, Macclesfield, Cheshire, UK) was started to reach a blood concentration of 2 μ−1: in the sevoflurane group, sevoflurane was initially titrated to an end-expiratory concentration of 0.5 MAC. Any increase or decrease in heart rate or mean arterial blood pressure > 10% was followed by adjustment of the remifentanil infusion by ± 0.05 μ−1.min−1. Propofol and sevoflurane were titrated continuously to maintain BIS between 40 and 50 throughout the surgical procedure until the final suture was finished. In case of BIS < 40 or > 50, either the propofol infusion was adjusted by ± 0.5 μ−1 blood concentration or the end-expiratory sevoflurane concentration adjusted by 0.25 MAC. Additional rocuronium (10 mg) was given as required to keep the train-of-four stimulation < 3 throughout surgery. Intra-operatively, all patients were covered with a convective air warming system (Warmtouch®, Mallinckrodt Medical, Hazelwood, MO). Vital signs, nasopharyngeal temperature, ventilation parameters and concentration of anaesthetics were recorded manually every minute for 5 min after induction of anaesthesia, after skin incision and after the final suture until tracheal extubation. During the surgical procedure, recording was performed every 5 min. Values of BIS were recorded continuously.

Fifteen minutes before the end of surgery all patients received 7.5 mg piritramide intravenously for analgesia. Once the final suture was finished, delivery of anaesthetics was stopped and the fresh gas flow increased to 6 l.min−1 until extubation. Criteria for extubation were as follows: sufficient spontaneous breathing (end-expiratory Pco2 < 6 kPa and Spo2 > 90%); temperature > 36 °C; systolic blood pressure and heart rate within 20% of baseline; train-of-four ratio > 0.7; and an awake and co-operative patient. Time from discontinuation of anaesthesia to extubation (‘extubation time’), vital signs and BIS values were recorded. After extubation the patient was transferred to the postanaesthesia care unit, where standard monitoring was recorded every 10 min, along with the Observer Assessment of Alertness and Sedation (OAAS) Scale [12], Post-anaesthesia Discharge Scoring System (PADSS) [13], modified Aldrete score [14] and Ramsay score [15]. The Trieger Dot and the Digit Symbol Substitution Tests were performed every 20 min. Postoperative pain (visual analogue scale; VAS), analgesic demand, postoperative nausea and vomiting (PONV) and shivering were also recorded. Postoperative pain (VAS > 4) was treated with 3.75 mg piritramide, and shivering was treated with 75 μg clonidine, intravenously. Patients were judged ready for discharge from the postanaesthesia care unit with a score of ≥ 3 in OAAS, ≥ 9 in PADSS and modified Aldrete score, a VAS < 4 and Spo2 > 95% with oxygen 2 l.min−1 or > 92% without oxygen. Postoperative recovery (from admission to the postanaesthesia care unit until discharge) was assessed by a single investigator blinded to the allocation of groups. On the next day, patients were interviewed using a standardised questionnaire about the occurrence of nausea, vomiting, shivering, strongest pain and the effect of pain therapy (using a four-point scale consisting of none, mild, moderate or severe.) Patients were also asked to rate their preference for a similar anaesthetic procedure in the future using a four-point scale.

A sample size of 49 patients was calculated based on previous studies of time to extubation following propofol and sevoflurane as the primary endpoint [2, 3, 11, 16], with a difference of 5 min, 90% power, and an α error of 0.05. To compensate for dropouts, we enrolled 55 patients in each group. Statistical analysis performed using GraphPad Prism v.4.03 for Windows (GraphPad Software, San Diego, CA). A Kolmogorov-Smirnov test was used to test for Gaussian distribution. Haemodynamic data and normalised BIS values between groups were tested by two-way anova factoring for group and time. One-way anova for repeated measures followed by Dunnett's post-test correction and the Friedman test with Dunn's post-test correction were performed to detect changes over time within the same group. Values between groups were compared using the unpaired Student's t-test or Mann–Whitney U-test. Proportions were compared with Fisher's exact test or Chi-squared test, as appropriate. A p value of <  0.05 was considered statistically significant.


The two study groups were comparable with respect to patients' characteristics, type and duration of surgery and duration of anaesthesia (Table 1). Baseline blood pressure and heart rate and intra-operative BIS values, body temperature and arterial oxygen saturation were also similar. Five minutes after induction of anaesthesia, systolic blood pressure in the propofol group was significantly lower than in the sevoflurane group, subsequently, systolic blood pressure was approximately 10 mmHg lower in the sevoflurane group till the end of surgery (Fig. 1a). Heart rate was similar between groups, as was arterial oxygen saturation. Both the total dose of remifentanil (mean (SD) 7.3 (2.9) mg in the propofol group and 6.9 (2.5) mg in the sevoflurane group) and the remifentanil infusion rates at defined times during anaesthesia did not differ significantly (Fig. 1b). Calculated average propofol target plasma concentration throughout anaesthesia was 2.55 (0.85) μ−1 and average end-expiratory sevoflurane concentration was 1.21 (0.28) vol% (Fig. 1c). Values for BIS were similar between groups during surgery (Fig. 1d). Extubation time was significantly shorter in the propofol group than in the sevoflurane group (8.3 (3.5) min vs 10.8 (4.6) min, respectively; p = 0.0024). At extubation, BIS values in the propofol group were significantly lower than in the sevoflurane group but in the further recovery period these differences disappeared (Fig. 1d). Average calculated propofol plasma concentration at extubation was 1.19 (0.52) μ−1; the average sevoflurane concentration was 0.17 (0.16) vol%.

Table 1.   Patients' characteristics, duration and type of surgery, and duration of anaesthesia in patients undergoing prolonged anaesthesia with either propofol/remifentanil or sevoflurane/remifentanil. Values are mean (SD) or number.
 Propofol(n = 51)Sevoflurane(n = 52)
  • *

    Radical prostatectomy or nephrectomy.

Age; years 53 (12) 53 (13)
Gender; F : M 24 : 27 25 : 27
Weight; kg 76 (15) 79 (13)
Body mass index; kg.m−2 25 (4) 26 (4)
ASA physical status; 1 : 2  7 : 44 10 : 42
Duration of surgery; min220 (58)234 (70)
Type of surgery; abdominal: urological* 37 : 14 34 : 18
Duration of anaesthesia; min281 (60)295 (67)
Figure 1.

 Changes in haemodynamic parameters, anaesthetic requirements and bispectral index (BIS) in patients undergoing prolonged anaesthesia with either propofol/remifentanil (grey; n = 51) or sevoflurane/remifentanil (black; n = 52). Values are mean (SD). a) Heart rate (▪) and blood pressure (bsl00066). p = 0.004 at induction + 5 min. b) Remifentanil infusion rates. c) Target propofol plasma concentration and sevoflurane concentration. d) BIS values. p = 0.007 at extubation.

There were no significant differences in physiological parameters in the postanaesthesia care unit. Recovery of cognitive function was comparable in both groups, although the cognitive function scores had not reached baseline values 80 min after the end of anaesthesia (Table 2). No significant differences were found in the incidence of shivering (propofol group 21/51 (41%) and sevoflurane group 18/52 (35%)) or PONV (propofol group 10/51 (20%) and sevoflurane group 13/52 (25%)), or in the intensity of pain (Table 3). The median (interquartile range [range]) amount of piritramide required was 11.25 (7.5–15 [3.75–30]) mg in the propofol group and 12.19 (7.5–15 [3.75–26.25]) mg in the sevoflurane group, and clonidine administration was 75 (0–150 [0–150]) μg in both groups. Patients were judged ready for discharge after 82 (34) min in the propofol group and after 79 (34) min in the sevoflurane group, although actual discharge was later than this (172 (56) min and 182 (76) min, respectively), primarily because of deficits in the organisation of the patients' transport.

Table 2.   Recovery of cognitive function in patients undergoing prolonged anaesthesia with either propofol/remifentanil (n = 51) or sevoflurane/remifentanil (n = 52). Values are median (IQR [range]).
 Pre-operative baseline40 min after extubation80 min after extubation
  • *

    See text for details.

Trieger Dot Test*42 (33–50 [22–55])42 (34–48 [20–55])26 (15–33 [7–52])30 (18–38 [11–54])34 (26–40 [13–55])34 (28–40 [12–55])
Digit Symbol Substitution Test*25 (18–34 [14–40])25 (18–33 [12–42])8 (5–12 [2–22])8 (5–11 [1–27])13 (6–17 [3–30])14 (7–19 [4–35])
Table 3.   Pain intensity (visual analogue scale) in patients undergoing prolonged anaesthesia with either propofol/remifentanil (n = 51) or sevoflurane/remifentanil (n = 52). Values are mean (SD).
 Propofol(n = 51)Sevoflurane(n = 52)
10 min after extubation6.3 (3.3)5.7 (3.5)
30 min after extubation5.0 (2.6)4.7 (2.6)
60 min after extubation3.4 (2.0)3.3 (2.2)

The 24-h follow-up interview revealed no difference in the incidence of nausea, vomiting, shivering, intensity of pain or patient satisfaction with the anaesthetic procedure.


The main findings of our study are as follows: first, both propofol and sevoflurane supplemented with remifentanil are suitable for anaesthesia for long lasting surgical procedures, each allowing comparable haemodynamic stability and fast postoperative recovery. Second, controlling and adjusting the depth of anaesthesia using BIS monitoring resulted in cost effective consumption of anaesthetics and analgesics in both groups. Third, although there was a slightly shorter time to extubation in the propofol group, this did not result in faster postoperative recovery.

Previous studies comparing propofol and sevoflurane in children and adolescents after intermediate anaesthetic procedures (∼ 90 min) have also demonstrated reduced time to extubation for propofol (11.8 (4.2) vs 15.0 (5.6) (min)), which did not result into earlier discharge from the postanaesthesia care unit[2]. Without monitoring the depth of anaesthesia, the amounts of propofol and sevoflurane administered – albeit in younger patients – were higher than in our study, whereas the remifentanil infusion rates were comparable. In another study during EEG-monitored depth of anaesthesia, shorter extubation times for propofol/remifentanil were observed compared to sevoflurane/remifentanil (8.1 (2.4) vs 11.4 (3.5) min, respectively) (Beger FA, Grouven U, Lüllwitz E, Hausdörfer J, Schultz A. Aufwachzeiten bei EEG-gesteuerten Narkosen: Propofol vs. Sevofluran Dt. Anasthesie Kongress, München, 2002). Postoperative recovery was not reported.

In our study, monitoring depth of anaesthesia was primarily intended to allow for an equivalent dosing of anaesthetics in both groups. The predefined BIS target of 40–50 during anaesthesia was accomplished, with values lying between 42 and 49 throughout. Interestingly, our use of propofol and sevoflurane was at the lower end of the clinical dosage range, whereas our use of remifentanil was similar to that commonly used clinically. In previous studies without depth of anaesthesia monitoring propofol, infusion rates for maintenance of anaesthesia were between 53 (8) and 135 (6) μ−1.min−1, whereas end-expiratory sevoflurane concentrations were between 1.2 (1.0) vol% (with nitrous oxide) and 1.45 (0.34) vol% (with air) [11, 16]. This may explain why extubation times in our study were shorter than in previous studies of short or intermediate term anaesthesia without depth of anaesthesia monitoring. A reduction of anaesthetic drug administration by using BIS or comparable EEG monitoring has been reported previously, as well as faster discharge from the recovery unit after EEG controlled anaesthesia [17]. The application of an EEG-based monitoring, especially for prolonged anaesthesia to increase safety and for economical reasons seems to be justified.

As a result of a posthoc subgroup analysis of patients < 60 years vs patients > 60 years, age seemed to have an effect on extubation time. This could be observed only in the propofol group but not in the sevoflurane group. Further investigations with appropriate study design would be necessary to verify this assumption.

For analysis of postoperative recovery, a variety of scoring systems are available, among which the Ramsay score and OAAS are most widely used and accepted. Other scoring systems also include vital parameters, intensity of pain and state of activity and the patients' general state (e.g. modified Aldrete score and PADSS). Scoring systems are relatively objective and their application is easy. Nevertheless, our study can be criticised because the scoring systems we used may not be able to detect minimal changes in the patients' state during the recovery period. Recovery of cognitive function was analysed by tests classified as the most sensitive for residual cortical disorders caused by anaesthetics [11]. Patients in both groups demonstrated significant delay in recovery of cognitive function, even 1.5 h after the end of anaesthesia, without any difference between the groups. In agreement with our results, Biedler et al.[18] were able to detect impairment of cognitive function even 4 h after anaesthesia lasting < 1 h. It has been previously demonstrated that cognitive function is more rapidly restored after sevoflurane or desflurane anaesthesia than after propofol, but in those studies propofol was given with nitrous oxide, which might explain the different results [19]. Our data suggest that duration of anaesthesia, even with modern drugs such as propofol and sevoflurane, has an effect on the recovery of sophisticated cognitive functions that is often not realised in clinical practice. An important factor in the performance and interpretation of the applied tests is that the individual motivation of the patient to participate actively has a strong effect on the results obtained.

We found a similar incidence of PONV and postoperative shivering between groups. Though propofol has intrinsic anti-emetic properties [20], previous studies also found no significant differences in the incidence of PONV and shivering after long-term anaesthesia with sevoflurane/desflurane vs propofol [11, 21, 22]. Potentially, the administration of piritramide might have blunted any beneficial effect of propofol. The incidence of PONV is influenced by many factors, for example female sex, young age, obesity and co-morbidities (e.g. diabetes), and is associated with specific surgical procedures [23]. The patients we studied did not differ in these factors. Furthermore, the incidence of PONV was low and our study consequently underpowered in this regard. Intra-operative warming resulted in a similar and acceptable body temperature in both groups at the end of surgery. This may have contributed to the similar incidence of shivering in both groups.

In our study, the ability to discharge patients from the recovery unit was defined by different score parameters. This was achieved comparably fast in both groups. Interestingly, actual discharge took considerably longer; this may be explained, at least in part, by deficiencies in the organisation of patient transport to the wards. Furthermore, immediate discharge was not always required, especially when actual occupancy of the unit was low.