Over the past 50 years, developments in anaesthesia have not only introduced new agents and techniques but more importantly have shown considerable progress in the improvement of post-operative patient care. As a result of these initiatives there has recently been greater interest in improving the control of post-operative nausea and emesis(PONV) [1,2]. Female patients undergoing gynaecological procedures are generally regarded as being at particularly high risk of experiencing PONV [3,4]. The reported incidence of PONV following major gynaecological surgery ranges from 58% to 81% [5-9].
Management of PONV is primarily through the use of antiemetic drugs such as metoclopramide, droperidol, and prochlorperazine, which are used either prophylactically to prevent PONV, or as treatment for established PONV. However, the reported efficacy of these agents is variable which probably reflects differences in study methodology, assessment criteria, duration of the assessment period and study populations evaluated .
Previously published placebo-controlled studies have clearly shown ondansetron to be a more effective antiemetic than placebo in the prevention of PONV, both in day-case patients and in in-patients undergoing major surgery[6-9], and also in the treatment of established PONV in out-patients .
In this study, we have compared the efficacy and safety of a single intravenous (i.v.) dose of ondansetron 4 mg (as ondansetron hydrochloride dihydrate) with metoclopramide 10 mg and placebo in in-patients having major gynaecological surgery performed under general anaesthesia. Metoclopramide was chosen as the comparative agent as it was deemed to be the most widely used antiemetic in this patient setting in the countries that participated in this study. In addition, the study aimed to assess the efficacy of an additional dose of i.v. ondansetron 4 mg in preventing further emesis in patients who received study medication and subsequently experienced PONV within the 24 h study period.
This randomized, double-blind, parallel-group, placebo-controlled comparative study was carried out in 58 centres from the following 12 countries: Australia, Canada, Denmark, France, Germany, Iceland, Israel, the Netherlands, Norway, South Africa, Sweden and the United Kingdom.
Female in-patients, aged at least 18 years, having either major intra-abdominal gynaecological surgery or vaginal hysterectomy performed under standardized general anaesthesia and hospitalized for at least 24 h post-recovery were to be recruited into the study. Patients were to be randomized to treatment in a 4:4:1 ratio (ondansetron: metoclopramide: placebo) in a double-blind manner. The appropriate Regulatory and Ethics Committee approvals were obtained, and following a full explanation of the study procedures all patients provided written informed consent to participate in the study. The reasons for not studying patients are shown in Table 1.
Anaesthesia and antiemetics
A standardized balanced general anaesthetic technique was used for all patients, however, the choice of agents actually administered was not made too restrictive so as to reflect true clinical practice in each of the participating countries.
Premedication was with any benzodiazepine, with the exception of lorazepam which has known antiemetic properties, and excluded the use of anti-emetics or opioids. Anaesthesia was induced with a number of i.v. agents, and maintained with nitrous oxide in oxygen and supplemented with halothane, enflurane or isoflurane, as required. The use of propofol either for induction or maintenance of anaesthesia was prohibited. Analgesia at induction and during maintenance of anaesthesia was provided by opioids, when required. Neuromuscular blocking agents and reversal agents were used as clinically needed. Any type of analgesic could be given post-operatively for pain relief.
Randomized study medication of either ondansetron 4 mg, metoclopramide 10 mg or placebo (as citrate buffered normal saline) was given as a slow i.v. injection administered over at least 30 s immediately prior to induction of anaesthesia. In addition, patients who experienced PONV within the first 24 h post-recovery period and required rescue antiemetic medication were given an additional rescue dose of i.v. ondansetron 4 mg. If further rescue antiemetic medication was required after this time, or if the rescue antiemetic dose of ondansetron 4 mg was found to be ineffective, any other rescue antiemetic could be administered.
The time of each emetic episode (defined as a single non-productive retch or productive vomit, or two or more retches and/or vomits occurring within 1 min of each other) during the 0-24 h period after recovery from anaesthesia (where the recovery time was defined as the patient's first response to a spoken command) was recorded.
The duration of nausea experienced during the specified time periods 0-1 h, 1-2 h, 2-4 h, and 4-24 h after recovery from anaesthesia was recorded after each patient was asked the question 'How long did any feeling of nausea last during each assessment period'? The severity of the 'worst episode' of nausea during each assessment period was recorded using a linear numerical scale from 0 to 10, where 0 represented 'no nausea' and 10 represented 'nausea as bad as it could be'. The assessment was based on each patient's response to the question 'How would you rate your worst feeling of nausea on a scale of 0-10 since the last assessment or since recovery from anaesthesia'?
The safety and tolerability of ondansetron were assessed by monitoring adverse events throughout the study and by performing appropriate laboratory tests before study medication was administered and again post-treatment, if clinically indicated. In addition, a prestudy pulse and blood pressure measurement was made and the duration of anaesthesia and time to recovery from anaesthesia was also recorded.
The number of patients to be included in the study was based on an estimated 30% of placebo-treated patients and 40% of metoclopramide-treated patients expected to experience no emesis (i.e. have complete response) during the 0-24 h post-recovery period. A total of 388 patients per active treatment group would give 80% power to detect a 10% increase in the proportion of ondansetron-treated patients experiencing no emesis compared with metoclopramide at the 5% level of significance, a difference which would also be of clinical significance. A total of 92 patients in the placebo group would give 95% power to detect a 20% increase in the complete response rate for ondansetron compared with placebo at the 5% level of significance. Allowing for 10% unevaluable patients, overall 972 patients (ondansetron n= 432, metoclopramide n = 432, placebo n = 108) were planned to be recruited.
The efficacy analysis was performed on all patients who were randomized and received study antiemetic medication prior to or during general anaesthesia and had gynaecological surgery performed. The safety analysis was performed on all patients who were randomized and received study antiemetic medication or who only received rescue ondansetron medication.
The primary assessment of response was based on the proportion of patients experiencing no emetic episodes and who received no rescue antiemetic medication and were not withdrawn from the study due to treatment failure/lack of response during the 0-24 h period after recovery from anaesthesia. Comparisons were made between all three treatment groups using Mantel-Haenszel χ2-tests.
Secondary assessments of response were based on the complete control of nausea (i.e. no nausea, no rescue antiemetic and no withdrawal), the time to the first emetic episode, the number of emetic episodes, the duration of the worst episode of nausea and the severity of the worst episode of nausea experienced during the 0-24 h period after recovery from anaesthesia. the incidence of emesis in patients who received a rescue dose of ondansetron 4 mg within the 0-24 h post-recovery period was also assessed, but not statistically tested. Comparisons were made between all three treatment groups using Mantel-Haenszel χ2-tests for the complete control of nausea and Wilcoxon rank sum tests for the other parameters.
Logistic regression analysis was used to investigate the effect of the prognostic factors on emesis and nausea during the 24 h post-recovery period. However, no statistical testing was performed across the three treatment groups when evaluating the effects of surgery type and previous history of PONV on emesis and nausea.
All significance tests were two-sided and were performed using SAS version 6.08 for the windows operating system. All analyses were adjusted for cluster of centres using the van Elteren method and no adjustments were made for missing values. Patients who received rescue antiemetic medication were regarded as treatment failures for the purposes of the analysis from the time of administration onwards. For patients regarded as treatment failures the number of emetic episodes was assigned a value of six prior to analysis and the nausea score was assigned a value of 10 prior to analysis and the patient was considered to have suffered nausea from the time of rescue to the end of the 0-24 h post-recovery period.
A total of 1074 patients were recruited into the study and 1047 of these patients received study medication. Of these patients, 1044 were subsequently correctly randomized to treatment, with 465, 462, and 117 patients having received i.v. doses of ondansetron 4 mg, metoclopramide 10 mg and placebo, respectively.
All treatment groups were similar with respect to patient demographic and base-line characteristics and type of surgery performed (Table 2). The most common type of surgery carried out was abdominal hysterectomy (66%). No patient treatment group differences were noted for concurrent medical conditions reported and the concurrent drugs administered. In addition, a comparable proportion of patients in each of the three treatment groups received the same premedicants, anaesthetic agents, muscle relaxant and reversal agents and intra-operative and post-operative analgesics, thus confirming the homogeneity of this patient population (Table 3).
Emesis. Significantly more patients who received ondansetron experienced no emetic episodes (203/465, 44%) during the 0-24 h post-recovery period compared with patients who received metoclopramide (173/462, 37%; P = 0.049) or placebo (29/117, 25%; P<0.001) (Fig. 1). In addition, statistically significantly fewer patients in the ondansetron treatment group required rescue antiemetics or were withdrawn due to treatment failure (215/465, 46%) than patients in either the metoclopramide (247/462, 53%; P = 0.026) or placebo (79/117, 68%; P<0.001) treatment groups (Fig. 1).
Ondansetron significantly reduced the total number of emetic episodes experienced by patients compared with both metoclopramide(P = 0.034) and placebo (P<0.001). The median time to the first emetic episode or treatment failure was significantly prolonged in patients who received ondansetron (19.2 h) compared with patients who received either metoclopramide(9.4 h; P = 0.013) or placebo (4.3 h; P<0.001) (Table 4).
Emesis by surgery type. Complete control of emesis by surgery type showed that ondansetron was the most effective antiemetic in abdominal and vaginal hysterectomy patients, whereas similar levels of control were observed for ondansetron and metoclopramide in gynaecological laparotomy and other gynaecological surgery patients. Based on the placebo treatment group results, vaginal hysterectomy appeared to be the most emetogenic procedure (89%), followed by gynaecological laparotomy (78%) and abdominal hysterectomy (70%) (Table 5).
Emesis by previous history of PONV. For patients with no previous history of PONV, better complete control of emesis was achieved in all three treatment groups compared with patients with a previous history of PONV. In both comparisons more ondansetron-treated patients experienced no emetic episodes and fewer ondansetron-treated patients were rescued or withdrawn due to treatment failure than in either the metoclopramide or placebo treatment groups (Table 6).
Emesis requiring an additional rescue antiemetic dose of ondansetron. A smaller proportion of patients who received an original prophylactic dose of ondansetron received a rescue dose of ondansetron (153/465, 33%) compared with those patients who originally received a prophylactic dose of either metoclopramide (186/462, 40%) or placebo (57/117, 49%). Rescue treatment with ondansetron 4 mg resulted in a complete response in ≈37% patients. The complete control of emesis with rescue ondansetron treatment by initial treatment group was ondansetron 39%, metoclopramide 36% and placebo 33%. No statistical testing of this outcome was planned or carried out because the study was not originally sized for this comparison.
Nausea. No nausea was experienced by 32% (148/465) of patients who received ondansetron in the 0-24 h period following recovery from anaesthesia. However, significantly fewer patients who received either metoclopramide(112/462, 24%; P = 0.009) or placebo (19/117, 16%; P<0.001) experienced no nausea (Fig. 1). Based on relative odds, patients who received either metoclopramide or placebo were 1.5 times (95% Cl 1.5-4.2) more likely or 2.5 times (95% Cl 1.1-2.0) more likely, respectively, to experience post-operative nausea than patients who received ondansetron.
Nausea by surgery type. A similar pattern of results observed for emesis was seen for the complete control of nausea by surgery type. Overall, more nausea than emesis was experienced by patients across the different types of surgery (Table 5).
Nausea by previous history of PONV. As for complete control of emesis, a greater proportion of patients with no previous history of PONV experienced no nausea compared with patients with a previous history of PONV. For both patients with and without a previous history of PONV ondansetron appeared to be the most effective antiemetic (Table 6).
Nausea severity and duration. The severity score of the worst episode of nausea (median score) during the 0-24 h post-recovery period was significantly lower for ondansetron treated-patients (score = 7) than for either metoclopramide-treated patients (score = 10; P = 0.018) or placebo-treated patients (score = 10; P<0.001). The total median duration of any nausea experienced was significantly reduced with ondansetron (45 min) compared with either metoclopramide (5.1 h; P = 0.041) or placebo (18 h; P<0.001)(Table 4).
Prognostic factors for experiencing PONV
The effects of surgery type, duration of anaesthesia, previous history of PONV and administration of post-operative opioids on the incidence of PONV were investigated using logistic regression analysis.
For emesis, all factors, other than duration of anaesthesia, were found to be significant (P<0.05) influences on the incidence of emesis, with a previous history of PONV and type of surgery being highly significant (P<0.001) influences.
For nausea, a previous history of PONV and surgery type were highly significant (P<0.001) influences on the incidence of nausea, and duration of anaesthesia was significant at the 10% level. However, the administration of opioids post-operatively was not found to be an important influence.
The overall incidence of adverse events was similar across the treatment groups (ondansetron 25%, metoclopramide 21%, and placebo 22%). The incidence of drug-related adverse events was low and reported in no more than 4% of patients in any treatment group. The most common adverse events reported were dizziness, headache, pruritis and hypotension, however, there were no significant differences in the incidence of these events between the three treatments(Table 7).
Twenty-three serious adverse events were reported and the incidence was similar in each treatment group. Only one serious event of laryngospasm was assessed as being probably drug-related in a patient who received ondansetron. This was diagnosed due to an inability to intubate the patient after administration of pancuronium, fentanyl, midazolam, thiopentone and suxamethonium. The patient had not received any medication to reduce salivation before anaesthesia. Blind nasal intubation was performed after spontaneous breathing had been restored. The event resolved within 10 min but was considered lifethreatening. It was also considered to be due to excessive salivation with possible aspiration of saliva.
One patient who received ondansetron died due to a pulmonary embolism and one patient who received metoclopramide was withdrawn due severe bronchospasm. Neither of these two events were considered to be drug-related.
With the recent heightened awareness of PONV it has become clear that certain factors, including the type of surgery, are likely to influence the occurrence of these post-operative problems . Patients undergoing gynaecological surgery appear to have an inherently higher risk of experiencing PONV . Within this group of patients, the literature also suggests that a greater incidence of PONV is experienced by females undergoing major inpatient gynaecological surgery, where the incidence of PONV ranges from 58% to 81% [5-9], compared with females undergoing day-case gynaecological procedures where the incidence is slightly lower (range 40% to 77%) [4, 13-15]. This may be due to factors such as surgery being more manipulative thus stimulating vagal afferents in the gut, increased use of post-operative opioids and also increased pain .
In addition, many of the commonly used anaesthetic agents are known to have varying degrees of emetogenicity associated with their use . This adds to the complexity and multifactorial nature of PONV. The background emetogenic stimulus associated with these drugs can be controlled to some extent by the choice of agents used, however, other factors such as those relating to the patient, surgery and post-operative care including ambulation still complicate the picture.
The overall placebo incidence of nausea (84%) and emesis (75%) observed in this group of major gynaecological surgery patients was comparable with previously published data [8,9]. In addition, ondansetron was also shown to be significantly more effective than placebo in preventing PONV. This has also been demonstrated in a number of published studies following this type of surgery[6-9] and also after day-case gynaecological surgery [11,12,14-16] and inpatient major orthopaedic surgery .
In this study, both ondansetron and metoclopramide were superior to placebo in preventing PONV following major gynaecological surgery. For the overall patient population ondansetron offered a statistically greater level of complete control of both nausea and emesis than did metoclopramide.
Perhaps more important findings in this study related to prolonged median time to first emetic episode and reduced duration of nausea, where ondansetron was clearly superior to metoclopramide and placebo.
Other published studies [15, 16, 18, 19] have also shown that ondansetron is a more effective antiemetic than metoclopramide in preventing PONV. A further recently published study also compared ondansetron 4 mg i.v. with metoclopramide 10 mg i.v. for the prevention of PONV in major gynaecological patients . This was a very small (n= 50) single-centre study and the antiemetics were administered at the end of surgery. In addition, no control arm was included in the study to assess the overall incidence of PONV. Complete control of emesis over the 24 h post-operative period was observed in 46% (11/24) patients who received metoclopramide and 36% (9/25) patients who received ondansetron, however, the study was not large enough to show a statistical difference between the two antiemetics. These results were almost a complete reversal of those observed in this study (ondansetron 44%, metoclopramide 37%) where a statistical difference was observed.
In this study both ondansetron and metoclopramide were administered immediately before induction of anaesthesia. The intention was to ensure the active antiemetic was in the patient's system and able to block the appropriate receptors before most of the emetic stimuli were triggered. Published data for metoclopramide  suggests that there may be a limited benefit in clinical effect when this antiemetic is given at the end of surgery. Ondansetron is only licensed to be given at induction of anaesthesia. However, one unpublished study compared the antiemetic effect of ondansetron when given either at induction of anaesthesia or at the end of surgery and no difference in clinical effect was observed (Glaxo Wellcome data on file - POND-L03).
In summary, this study supports previously published findings that ondansetron is a well-tolerated agent and is a more effective antiemetic for preventing post-operative nausea and emesis than placebo. In addition, ondansetron appears to be a more effective antiemetic than metoclopramide with an overall improvement in complete control of post-operative nausea and emesis, prolongation of time to first emetic episodes and reduction in duration of nausea.
We wish to thank the following anaesthetists for their help in conducting this study: Dr J. Abrahamsson, Dr G. Angergard, Dr U. Bang, Professor D. Bevan, Dr E. Bjornestad, Dr K. G. Branch, Dr R. Brule, Dr I. Brunsson, Dr T. H. L. Bryson, Dr B. Bryssine, Dr G. P. M. Clark, Dr C. Clinton, Professor A. R. Coetzee, Dr G. Cooper, Professor J. M. Desmont, Dr M. Dodgson, Dr L. Eidelman, Dr E. M. C. Ernst, Dr A. Fernandes, Dr A. Greemberg, Dr F. E. A. Geisler, Dr H. Guldager, Professor J. P. Haberer, Dr M. Harmer, Dr G. R. Harrison, Dr S. Helbo Hansen, Dr O. Higersen, Dr G. Hirlekar, Dr Ø. Holgersen, Dr A. Hunting, Dr M. T. Inman, Dr K. Jacobsen, Professor M. James, Professor R. M. Jones, Dr N. J. Kanstrup-Laursen, Dr P. L. Klineberg, Dr F. M. Lahaye, Professor M. C. Laxenaire, Dr D. J. Layfield, Dr E. Levy, Dr A. Mawter, Dr C. Moote, Dr E. Nordenskjold, Dr L. T. Øye, Professor A. Perl, Dr M. Pfisterer, Dr H. Reich, Dr R. Shukla, Dr G. Sigurbjornsson, Professor G. Smith, Dr S. Søfelt, Dr J. Solf, Dr J. C. Soni, Dr D. A. Thomas, Dr A. E. Van Eaden, Dr C. Virenque, Dr M. H. Wheildon.
We also thank Miss J. McRae for performing the statistical analyses, Mr D. Taylor for the data management aspects of the study, Dr E. Russell and Dr W. Paska for assistance with this manuscript, Ms A. den Baars, Ms N. Clyti, Dr C. Crawford, Ms N. Dobie, Ms A.-M. Eriksson, Ms S. Gellerstein, Mr U. Genest, Ms F. Hancock, Dr K. Hellstern, Mrs K. Hjelle, Dr P. Hunaeus, Mr G. Martin, Ms L. McGibbon, Mrs A. Simon, Mrs I. Sjöblom, Mr B. Thorlacius for their help in coordinating the study.
1 Kapur PA. Editorial: The Big 'Little Problem'. Anesth Analg
2 Kenny GNC. Risk factors for postoperative nausea and vomiting. Anaesthesia
3 Watcha MF, White PF. Postoperative nausea and vomiting. Its etiology, treatment, and prevention. Anesthesiology
4 Lerman J. Surgical and patient factors involved in postoperative nausea and vomiting. Br J Anaesth
(Suppl. 1): 24S-32S.
5 Madej TH, Simpson KH. Comparison of the use of domperidone, droperidol and metoclopramide in the prevention of nausea and vomiting following major gynaecological surgery. Br J Anaesth
6Kenny GNC, Oates JDL, Leeser J et al.
Efficacy of orally administered ondansetron in the prevention of post-operative nausea and vomiting: a dose ranging study. Br J Anaesth
7Dupeyron JP, Conseiller C, Levarlet M et al.
The effect of oral ondansetron in the prevention of postoperative nausea and vomiting after major gynaecological surgery performed under general anaesthesia. Anaesthesia
8 Helmers JHJH, Briggs L, Abrahamsson J et al.
A single iv dose of ondansetron 8 mg prior to induction of anaesthesia reduces postoperative nausea and vomiting in gynecological patients. Can J Anaesth
9 Rust M, Cohen LA. Single oral dose ondansetron in the prevention of postoperative nausea and emesis. Anaesthesia
10 Rowbotham DJ. Current management of postoperative nausea and vomiting. Br J Anaesth
(Suppl. 1): 46S-59S.
11 Pearman MH. Single dose intravenous ondansetron in the prevention of postoperative nausea and vomiting. Anaesthesia
12 Claybon L. Single dose intravenous ondansetron for the 24-h treatment of postoperative nausea and vomiting. Anaesthesia
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
14Kovac A, McKenzie R, O'Connor T et al.
Prophylactic intravenous ondansetron in female outpatients undergoing gynaeccological surgery: a multicentre dose-comparison study. Eur J Anaesth
(Suppl. 6): 37-47.
15 Malins AF, Field JM, Nesling PM, Cooper GM. Nausea and vomiting after gynaecological laparoscopy: comparison of premedication with oral ondansetron, metoclopramide and placebo. Br J Anaesth
16 Paxton LD, McKay AC, Mirakhur RK. Prevention of nausea and vomiting after day case gynaecological laparoscopy. Anaesthesia
17Gan TJ, Collis R, Hetreed M. Double-blind comparison of ondansetron, droperidol and saline in the prevention of postoperative nausea and vomiting. Br J Anaesth
18 Raphael JH, Norton AC. Antiemetic efficacy of prophylactic ondansetron in laparoscopic surgery: randomized, double-blind comparison with metoclopramide. Br J Anaesth
19 Alon E, Himmelseher S. Ondansetron in the treatment of postoperative vomiting: a randomized double-blind comparison with droperidol and metoclopramide. Anesth Analg
20 Chen PP, Chui PT, Gin T. Comparison of ondansetron and metoclopramide for the prevention of post-operative nausea and vomiting after major gynaecological surgery. Eur J Anaesth