Secondary Logo

Journal Logo

Original Article

Comparative study of the antiemetic efficacy of ondansetron, propofol and midazolam in the early postoperative period

Unlugenc, H.; Guler, T.; Gunes, Y.; Isik, G.

Author Information
European Journal of Anaesthesiology: August 2003 - Volume 20 - Issue 8 - p 668-673


Postoperative nausea and vomiting (PONV) remain common and distressing complications of many anaesthetic and surgical procedures [1]. The incidence rate of PONV has changed little over the last 30 yr. Some of the earliest volatile anaesthetics were very emetogenic, and lack of adequate pain control undoubtedly contributed to the postoperative emetic response [2]. However, in spite of improved anaesthetic agents and surgical techniques, the incidence rate of PONV in patients undergoing general anaesthesia remains high.

Many studies have shown that propofol possesses antiemetic activity [3,4], possibly via antagonism of the dopamine D2 receptor [5,6]. Because propofol in subhypnotic doses is effective against nausea and vomiting [7,8], it is possible that it possesses direct antiemetic properties. Midazolam reduces the incidence of PONV [9,10]. Unlike propofol, this benzodiazepine has no direct antiemetic activity, but controls anxiety and accompanying anticipatory nausea [1]. Ondansetron is a potent, highly selective 5-HT3 receptor antagonist that is effective and well tolerated in the treatment of PONV [11].

Although the sedative-hypnotic effects of propofol and midazolam are known, their antiemetic effects in the treatment of PONV are not. Therefore, the aim of this study was to test the antiemetic efficacy of two different hypnotic drugs (propofol 15 mg, midazolam 1 and 2 mg) in comparison with ondansetron, a potent antiemetic drug, for the treatment of established PONV following major abdominal surgery and gynaecological surgery.


This study was approved by the Ethics Committee of the Faculty of Medicine of Cukurova University, and written informed consent from 453 patients was obtained. All patients were ASA Class I-II, aged 18-65 yr, of both gender, and were scheduled for elective gynaecological or abdominal surgery under general anaesthesia. Patients were included if nausea or vomiting occurred during the first 2 h in the Postanaesthesia Recovery Unit. Exclusion criteria were: a history of motion sickness, previous postoperative vomiting, known major organ disease, ASA physical status >II, body weight >100% over ideal, a history of alcohol or drug abuse, or receipt of an antiemetic agent within 24 h.

Standard intraoperative monitoring included pulse oximetry, automated blood pressure measurement and electrocardiogram lead II. Anaesthesia was induced in all patients by thiopental (5 mg kg−1) and maintained with sevoflurane 1.5-2% in a mixture of nitrous oxide 67% in oxygen. Neuromuscular relaxation was induced either by an intravenous (i.v.) bolus of vecuronium (0.1-0.2 mg kg−1) or cisatracurium (0.1-0.2 mg kg−1) and maintained (0.03 mg kg−1) by bolus administration at 30 min intervals. Following endotracheal intubation, ventilation of the lungs was controlled (tidal volume of 10 mL kg−1, respiratory rate of 10 breaths min−1) to produce an end-tidal carbon dioxide partial pressure in the range 4.2-5.0 kPa.

At the end of surgery, patients were extubated following antagonism of residual neuromuscular block with atropine (0.015 mg kg−1) and neostigmine (0.05 mg kg−1). The non-steroidal analgesic piroxicam (0.5 mg kg−1) was given intramuscularly for postoperative pain relief. If no pain relief was obtained, increments of fentanyl (0.5-1 μg kg−1) i.v. were given to maintain a pain score <5 on a verbal analogue scale of 0-10. No antiemetics were administered intraoperatively.

After surgery, patients were observed in the recovery room. When awake (judged by spontaneous eye opening, response to commands and the ability to verbal contact), patients were asked about any nausea or vomiting. One-hundred-and-twenty of the 453 patients (26%) experiencing nausea, having a score of 2 or more on a five-point nausea scale during the first 6 h postrecovery, were randomly divided into one of four treatment groups. Propofol 15 mg (1.5 mL) in Group P, midazolam 1 mg in Group M1, midazolam 2 mg in Group M2 and ondansetron 4 mg in Group O were given i.v. by the nursing staff according to the study protocol. If this treatment failed, the same doses were repeated 5 min later. The study medications were restricted to three doses. Nausea persisting in spite of these three doses was regarded as a treatment failure, and ondansetron 4 mg was given i.v. Nausea was assessed using a linear numerical scale of 1-5 (where 1 = none; 2 = mild, once in 15 min; 3 = moderate, two or three times in 15 min; 4 = severe, four or more times in 15 min; and 5 = worst nausea, persistent, severe nausea despite treatment with antiemetics).

Sedation was assessed by the investigator using the five-point Observer's Assessment of Alertness/Sedation (OAA/S) scale [12] (where 1 = awake/alert and 5 = deep sleep). The nausea score, sedation score, systolic and diastolic blood pressure (SBP-DBP), heart rate (HR), and peripheral oxygen saturation (SPO2) were recorded by an anaesthetist of the Pain Management Team. They were blinded to the treatment group before administration of study drugs and at 5, 10, 15, 30, 60, 120 and 360 min after the administration of study drugs. The type of surgery, duration of anaesthesia, total dose of fentanyl used and number of patients requiring a second dose were also recorded. Any side-effects (e.g. sedation-OAA/S > 2, hypotension or confusion) associated with study drugs were also recorded. Patients were transferred to the ward when the study was complete and when fully awake (as judged by response to commands and ability to verbal contact).

Sample size calculation was based on a power analysis. A 60% incidence rate of emesis was predicted following major abdominal surgery and gynaecological surgery. After treatment, the incidence rate of emesis was expected to be 20%. The sample size required was 27 subjects per study group. Patients experiencing nausea, having a score of 2 or more on a five-point nausea scale during the first 6 h postrecovery, entered the study consecutively.

Statistical analyses were performed using data from the intent-to-treat population using the statistical package SPSS, v.9.0 (SPSS, Inc., Chicago, IL, USA). The Kruskal-Wallis test was used to make comparisons between groups. The effects of time and treatment group were evaluated by repeated measurement analysis. Since analysis of variance was significant, comparisons were applied using the U- and Wilcoxon signed rank sum tests. Bonferroni's correction was applied (P < 0.05/n, where n = number of comparisons) when multiple comparisons were made. The incidence of complications between the groups was analysed by using the χ2-test. Results were calculated as mean (±SD) and median(minimum − maximum).


There was no difference between groups in age, weight or gender. The duration of anaesthesia, type of surgery, total dose of fentanyl and number of patients requiring a second dose were comparable in the groups (Table 1). There was no statistically significant difference between groups in haemodynamic variables or SPO2. Both remained within the normal range throughout the 6 h observation.

Table 1
Table 1:
Patients' characteristics, type of surgery, duration of anaesthesia, total dose of fentanyl used and number of patients requiring second dose of antiemetics.

There were no significant differences between groups in mean nausea scores before medication. Repeated analysis was performed to evaluate the effect of time and treatment; both affected nausea scores (Fig. 1). Nausea scores decreased significantly with time in all groups (P < 0.05). They were significantly lower in all groups during each study period than before the administration of study drugs (P < 0.05). There were no significant differences between groups in mean nausea scores after administration of study drugs (Table 2).

Figure 1
Figure 1:
Effect of time and treatment groups on nausea scores. ⋄: Nausea score, before treatment; ▾: nausea score, 5 min after treatment; □: nausea score, 15 min after treatment.
Table 2
Table 2:
Nausea and sedation scores.

Both propofol and midazolam 2 mg were as effective and rapid in onset as ondansetron. However, four patients (13.3%) in Group P, 13 (43.3%) in Group M1, five (16.6%) in Group M2 and one (3.3%) in Group O required a second dose of the study drug to achieve a successful outcome (Table 1). In Group M1, two patients (7%) did not respond even to the third dose of midazolam (1 mg) and were given ondansetron 4 mg i.v.

There were no significant differences between groups in mean sedation scores before the start of the medication. Repeated analysis was performed to evaluate the effect of time and treatment groups; both affected sedation scores (Fig. 2). The nausea score after 15 min and the sedation scores after 60 min are not shown as figures, since there was no statistical difference between groups. Although sedation scores were increased by propofol and midazolam, neither caused deep sedation (OAA/S > 3). After the study medication, patients remained awake and/or responded to verbal contact during the study period. Sedation scores were significantly higher in Groups P and M2 at 5 and 15 min than before the administration of study drugs (P < 0.05). Between groups, sedation scores were significantly lower in Group M1 than in Groups P and M2 at 5 and 15 min (P < 0.007 and < 0.007), and lower in Group O than Groups P and M2 at 5, 15, 30, 60 and 120 min (P < 0.007 and <0.007) (Table 2). Although sedation scores were significantly greater in Group M2 than in Group M1 (P < 0.001), the first dose of midazolam in Group M2 was significantly more successful against nausea than in Group M1 (83.4 versus 56.7%) (Table 1).

Figure 2
Figure 2:
Effect of time and treatment groups on sedation scores. ⋄: Sedation score before treatment; ▾: sedation score, 5 min after treatment; □: sedation score, 15 min after treatment; ▪: sedation score, 30 min after treatment; ●: sedation score, 60 min after treatment.

Eighty-seven patients (72.5%) complained of pain despite piroxicam (0.5 mg kg−1) intramuscularly: 20 in Group P (66%), 23 in Group M1 (76%), 21 in Group M2 (70%) and 23 in Group O (76%). These patients were given boluses of fentanyl (0.5-1.0 μg kg−1) i.v. to keep pain scores <5 on a verbal analogue scale of 0-10. There was no significant difference between groups with respect to the numbers of patients requiring fentanyl supplementation and the dosages of fentanyl given i.v. (Table 1). Two patients (7%) in Group O complained of headache after a single dose of ondansetron. No further adverse effects attributable to medications were observed.

There were no statistical differences in mean nausea scores between males and females who underwent abdominal surgery; neither were there any such differences when the females undergoing abdominal procedures were compared with the gynaecological group.


PONV varies in incidence rate from 10 to 60% depending on patient groups, the type of surgery and type of anaesthetic used. The patients in this study underwent either gynaecological or abdominal surgery, both associated with a high incidence of PONV. Our overall incidence of vomiting (28%) is comparable with that found in previous studies [13]. It is usually recommended that antiemetics be given prophylactically before surgery or chemotherapy [14]. However, in the present study the drugs tested were given first when the patients had developed symptoms of nausea and vomiting.

Although hypnotics exert their antiemetic action primarily via sedation, propofol is believed to depress directly the chemoreceptor trigger zone, vagal nuclei and other centres implicated in nausea and vomiting [15]. However, Borgeat and colleagues suggested that it lacked vagolytic properties [7]. They also suggested that propofol might have failed as an antiemetic in patients undergoing laparoscopic gynaecological procedures because of uninhibited vagal stimulation. The dose at which propofol exerts its antiemetic properties has been reported as 10-20 mg. In our study, propofol 15 mg was quite successful in the treatment of PONV after gynaecological or abdominal surgery.

Ewalenko and colleagues [14] reported that when propofol was given to patients who were sufficiently awake, a complete antiemetic response was obtained within <2 h in 94% of patients (30/32) and reached a 100% success rate subsequently. In our study, 86.7% of patients were relieved by a single dose of propofol. Furthermore, patients treated with propofol did not show any relapse into PONV during the study. Thus, the antiemetic action of propofol lasted much longer than its sedative-hypnotic action. Two clinical trials have reported increased sedation of patients when propofol was used for treatment of nausea and vomiting [16,17]. In one trial, one of 22 patients receiving propofol 40 mg had an episode of apnoea [16]. In the present study, a lower dose of propofol (15 mg) was used to avoid its sedative and/or respiratory depressant effects.

Midazolam is a short-acting benzodiazepine with possible antiemetic properties. Splinter and colleagues reported that midazolam 75 μg kg−1 reduced vomiting after tonsillectomy in children [9]. Di Florio and Goucke reported that an i.v. infusion of midazolam (1.0 mg h−1) significantly reduced persistent PONV [18]. The incidence of PONV following midazolam infusion was always significantly lower than with placebo. Using midazolam as an anaesthetic, Cole found 1% of PONV in children in the early postoperative period compared with 32% in the first 2 postoperative days [10]. In our study, i.v. administration of midazolam was quite successful in relieving PONV following gynaecological and abdominal surgery.

The exact mechanism of action of midazolam is unknown. It has been postulated to include glycine-mimetic inhibitory effects, enhancement of inhibitory effects of γ-aminobutyric acid (GABA), enhanced adenosinergic effects and inhibition of dopamine release [6]. Benzodiazepines may augment adenosine-mediated inhibition of dopamine in the chemoreceptor trigger zone, which results in an antiemetic effect [6,7]. In addition, Di Florio and Goucke hypothesized that midazolam would act synergistically with antidopaminergic drugs [18].

Although several studies have addressed the antiemetic dose-response range of propofol for treatment of PONV [7,14], the exact dose at which midazolam exhibits its antiemetic properties is unknown. In our study, midazolam 2 mg was more effective than 1 mg. The success rate for a single dose of 2 mg was 83.4% and 56.7% for 1 mg. Although the 1 mg dose failed to treat PONV in about half of the patients, the second 1 mg dose of midazolam was successful in 93% of cases.

Clinicians may hesitate to use hypnotic agents for antiemetic purposes because of the danger of sedation. Although the doses we used were lower than those recommended for sedation [19], a mild degree of sedation (OAA/S = 3) was observed in some patients. However, the antiemetic effect lasted longer than the sedation. We could not evaluate the influence of propofol and midazolam on the length of stay in the recovery room since all the patients remained electively in the recovery room for the 6 h of the study.

Comparing the two doses of midazolam, sedation scores were significantly higher in Group M2 than M1, although fewer patients in Group M2 required a second dose of midazolam. The relative merits of the higher and lower doses are unclear.

Although some studies have evaluated the antiemetic effect of midazolam and propofol on the incidence of PONV, no clinical study evaluating the correlation between the sedative and antiemetic properties of these agents has been reported [9,10]. Therefore, it is still not known if there is correlation between the sedative and antiemetic properties of these agents. Further studies are needed to reveal this relationship.

The antiemetic effect and safety of ondansetron for the treatment of PONV has been shown in many studies. In a double-blind study comparing the effect of ondansetron (8 mg) and placebo, patients receiving a single i.v. dose of ondansetron had a higher success rate (78 versus 28%) for the treatment of PONV [20]. In Polati and colleagues' study, the success rate of ondansetron (4 mg) was 93.1% [21]. These results were similar to our findings, where a single bolus of ondansetron was successful in 29/30 patients (96.7%).

It has been reported that headache and dizziness are the main adverse effects of ondansetron in the dosages used for the treatment of PONV [22,23]. In our study, although the incidence rate of headache was higher in Group O (7%) than for the other study drugs, it was not statistically significant. This may be regarded as a further argument for reducing the dose of ondansetron.

In conclusion, propofol and midazolam used in subhypnotic doses were as effective as ondansetron in treating PONV in patients undergoing abdominal or gynaecological surgery without untoward sedative or cardiovascular effects. Midazolam was more effective in the higher (2 mg) dose.


The authors gratefully acknowledge the assistance of the nursing staff and thank G. Seyda, PhD, for expert statistical advice. The study was not supported by external funds.


1. Camu F, Lauwers MH, Verbessem D. Incidence and aetiology of postoperative nausea and vomiting. Eur J Anaesthesiol 1992; 9: 25-31.
2. Andersen R, Krogh K. Pain as a major cause for postoperative nausea. Can Anaesth Soc J 1976; 23: 366-369.
3. Borgeat A, Wilder-Smith OHG, Suter PM. The non-hypnotic therapeutic applications of propofol. Anesthesiology 1994; 80: 642-656.
4. Splinter WM, Rhine EJ, Roberts DJ. Vomiting after strabismus surgery in children: ondansetron vs propofol. Can J Anaesth 1997; 44: 825-829.
5. Di Florio T. Is propofol a dopamine antagonist? Anesth Analg 1993; 77: 200-201.
6. Splinter WM, MacNeill HB, Menard EA, Rhine EJ, Roberts DJ, Gould MH. Midazolam reduces vomiting after tonsillectomy in children. Can J Anaesth 1995; 42: 201-203.
7. Borgeat A, Wilder-Smith OHG, Saiah M, Rifat K. Subhypnotic doses of propofol possess direct antiemetic properties. Anesth Analg 1992; 74: 539-541.
8. Song D, Whitten CW, White PF, Yu SY, Zarate E. Antiemetic activity of propofol after sevoflurane and desflurane anesthesia for outpatient laparoscopic cholecystectomy. Anesthesiology 1999; 90: 1789-1790.
9. Splinter W, Noel LP, Roberts D, Rhine E, Bonn G, Clarke W. Antiemetic prophylaxis for strabismus surgery. Can J Ophthalmol 1994; 29: 224-226.
10. Cole WH. Midazolam in paediatric anaesthesia. Anaesth Intensive Care 1982; 10: 36-39.
11. Blackwell CP, Harding SM. Clinical pharmacology of ondansetron. Eur J Cancer Clin Oncol 1989; 25: 21-27.
12. Chernik DA, Gillings D, Laine H, et al. Validity and reliability of the Observer's Assessment of Alertness/Sedation (OAA/S) scale: study with intravenous midazolam. J Clin Psychopharmacol 1990; 10: 244-251.
13. Madej TH, Simpson KH. Comparison of the use of domperidone, droperidol and metoclopramide in the prevention of nausea and vomiting following gynaecological surgery in day cases. Br J Anaesth 1986; 58: 884-887.
14. Ewalenko P, Janny S, Dejonckheere M, Andry G, Wyns C. Antiemetic effect of subhypnotic doses of propofol after thyroidectomy. Br J Anaesth 1996; 77: 463-467.
15. Appadu BL, Strange PG, Lambert DG. Does propofol interact with D2 dopamine receptor? Anesth Analg 1994; 79: 1191-1192.
16. Gan TJ, El-Molem H, Ray J, Glass PSA. Patient-controlled antiemesis. A randomized, double blind comparison of two doses of propofol versus placebo. Anesthesiology 1999; 90: 1564-1570.
17. Harper I, Della-Marta E, Owen H, Plummer J, Ilsley A. Lack of efficacy of propofol in the treatment of early postoperative nausea and vomiting. Anaesth Intensive Care 1998; 26: 366-370.
18. Di Florio T, Goucke CR. The effect of midazolam on persistent postoperative nausea and vomiting. Anaesth Intensive Care 1999; 27: 38-40.
19. Kay NH, Sear JW, Uppington J, Cockshott ID, Douglas EJ. Disposition of propofol in patients undergoing surgery: A comparison in men and women. Br J Anaesth 1986; 58: 1075-1079.
20. Larijani GE, Gratz I, Afshar M, Minassian S. Treatment of postoperative nausea and vomiting with ondansetron: a randomized, double-blind comparison with placebo. Anesth Analg 1992; 74: 473-474.
21. Polati E, Verlato G, Finco G, et al. Ondansetron versus metoclopramide in the treatment of postoperative nausea and vomiting. Anesth Analg 1997; 85: 395-399.
22. Tramèr MR, Reynolds DJM, Moore RA, McQuay HJ. Efficacy, dose-response, and safety of ondansetron in prevention of postoperative nausea and vomiting: a quantitative systematic review of randomized placebo-controlled trials. Anesthesiology 1997; 87: 1277-1289.
23. Tramèr MR, Reynolds DJM, Stoner NS, Moore RA, McQuay HJ. Efficacy of 5-HT3 receptor antagonists in radiotherapy-induced nausea and vomiting: a quantitative systematic review. Eur J Cancer 1998; 34: 1836-1844.

ANAESTHETICS, INTRAVENOUS, propofol; ANTIEMETICS, ondansetron; BENZODIAZEPINES, midazolam; POSTOPERATIVE COMPLICATIONS, postoperative nausea and vomiting

© 2003 European Academy of Anaesthesiology