Ondansetron and droperidol are commonly prescribed as single drugs for the prevention of postoperative nausea and vomiting (PONV) (1). However, a significant portion of patients continue to suffer from PONV despite the use of an adequate dose of either drug (2,3). A logical approach is to combine the two drugs because they act on the different receptors of the vomiting pathway. However, the treatment response may be modified with drug interactions. Thus, therapeutic and adverse effects are exaggerated with synergistic drug combinations. Conversely, larger doses are required to produce the same effect if the drugs are antagonistic to each other. On the basis of a stepwise logistic regression model, a multicenter trial (4) reported an additive interaction between ondansetron and droperidol, suggesting that the drugs act independently.
Alternatively, the interaction profile may be characterized by comparing the observed response of a drug combination with that predicted from additivity. In this regard, an additive drug interaction can be determined by multiplying responses of individual treatments. Therefore, synergism is defined if the observed response is significantly better than predicted and vice versa for antagonism. The primary purpose of the present study was to confirm the pharmacologic interaction between ondansetron and droperidol using this simple nonparametric approach. In this study, we compared the antiemetic effects of ondansetron or droperidol alone and when combined in moderate-to-high risk patients undergoing gynecologic laparoscopic surgery. We also determined the side effect profiles, particularly the electrocardiographic (ECG) changes, after administration of this drug combination. As a secondary objective, we performed a meta-analysis of published clinical trials to characterize the overall interaction between the two drugs.
The study was approved by the Clinical Research Ethics Committee. Written informed consent was obtained from all patients. Patients were eligible for the study if they were ASA physical status 1 or 2, aged between 18 and 45 yr, and scheduled for laparoscopic gynecologic surgery. Patients were excluded if they had preexisting nausea or vomiting, or had received opioids or drugs with known antiemetic properties in the 24 h before surgery. Pregnant patients and patients with a history of esophageal reflux, opioid or alcohol abuse were also excluded. All patients were interviewed during the preoperative assessment. A history of motion sickness and PONV was noted. The day of the current menstrual cycle was also recorded. Preoperative full blood counts, renal and liver function tests, and 12-lead ECG were performed and were repeated 24 h after surgery. No preanesthetic medication was prescribed, and the patients were fasted for at least 6 h before surgery.
In the operating room, routine ASA monitoring was applied. Patients were randomly assigned to one of the four treatment groups according to a computer-generated random number concealed in an opaque envelope. Patients in the control group received saline 10 mL, and those allocated to antiemetic groups received ondansetron 4 mg, droperidol 1.25 mg, or a combination of ondansetron 4 mg and droperidol 1.25 mg. Study drugs, up to 10 mL with normal saline, were prepared by an anesthesiologist independent of the study and were injected IV over 30 s. Arterial blood pressure and heart rate were recorded noninvasively immediately before and frequently after the start of drug administration. Immediate reactions including pain on injection, signs of allergy, extrapyramidal symptoms, dizziness, headache, chest or abdominal discomfort were specifically noted.
Five minutes after study drug injection, an ECG (lead II) tracing was obtained in all patients. The QT and RR intervals were measured and averaged over five consecutive cardiac cycles. The heart rate adjusted QT (QTc) interval was calculated according to the formula of Bazett (5).
After ECG recordings, anesthesia was then induced with fentanyl 2 μg/kg and propofol 2–3 mg/kg. Atracurium 0.5 mg/kg was administrated to facilitate tracheal intubation. Anesthesia was maintained with nitrous oxide 70% and isoflurane 0.5–1.0% in oxygen. The lungs were mechanically ventilated and the end-tidal carbon dioxide concentration was maintained between 5.0–5.5%. At the end of surgery, anesthesia was discontinued and residual neuromuscular blockade was antagonized with neostigmine 40 μg/kg and atropine 20 μg/kg in all patients. The trachea was extubated when the patient became fully awake. Anesthetic time was defined from the start of induction to the time when the anesthetic, including nitrous oxide, was discontinued. The subsequent period until the patient responded to verbal command was recorded as the recovery time.
After surgery, all patients were monitored in the postanesthesia care unit for 1 h, after which the patient returned to the ward. Analgesia was initially provided with IV morphine 1 mg and was repeated every 5 min until the patient was comfortable. In the ward, patients received IM morphine 0.15 mg/kg every 4 h or 1–2 tablets of dologesic (containing paracetamol 325 mg and dextropropoxyphene 32.5 mg) every 6 h as needed. The incidence of nausea and emetic episode (retching or vomiting), as well as the severity of nausea, pain, and sedation were recorded at 0, 1, 2, 4, 6, 24, and 48 h after surgery. Patient interviews were conducted in a standardized fashion by trained nurses who were blinded to the study drug. Retching or vomiting separated by at least 1 min was considered as separate emetic episodes. Severity of nausea and pain were graded using a 10-cm visual analog scale printed on a slide rule bar (Astra USA, Westborough, MA). Sedation was scored as 1 = alert, 2 = asleep, alert after arousal, 3 = asleep, drowsy after arousal, 4 = asleep, difficult to rouse, 5 = unarousable. IM prochlorperazine 12.5 mg was given as rescue antiemetic if there were two or more emetic episodes, and nausea persisted for more than 10 min or upon patient’s request to relieve nausea or vomiting. In a subgroup of 160 patients (the first 40 patients in each group), an additional 12-lead ECG was also obtained 2–3 h after surgery. Because of logistic reasons, we did not record an ECG from more patients. We noted that the changes of QTc intervals were consistently resolved by 2–3 h after surgery, and we therefore decided to stop further ECG recordings after the first 160 patients.
The primary end-point of the study was a complete response to treatment. This was defined as no nausea or emetic episodes or receiving rescue antiemetic during the entire 48 h observation. Adverse events were also recorded.
The interaction between ondansetron and droperidol was evaluated by comparing the predicted and observed efficacy of the drug combination. Assuming the two drugs act independently, and hence produce an additive effect, the predicted rate of complete response after combining ondansetron and droperidol is given as
This expression can be rewritten as
The drug combination is defined as synergistic if the observed complete response rate is more than what was expected. The reverse is considered as antagonistic. The 95% confidence intervals (CI) of the difference between the observed and predicted complete response rates were also calculated by the bootstrap resampling method.
An estimate of the sample size was based on the reported response rate after administration of ondansetron, droperidol, and their combination (4,6–10). Assuming the variation in the difference between the observed and predicted response rates is 5%, we calculated that 92 patients per group would have a 90% power at 5% significance level to determine the mode of interaction. We therefore planned to recruit 100 women in each group to allow for patient dropout and missing data. Patients successfully randomized to one of the treatment groups comprised the intention-to-treat population for all primary and safety analyses. The proportion of patients with complete response, the primary end-point, was compared among groups using χ2 test. Intergroup differences were compared between groups using Fisher’s exact test. The time to the first nausea or emetic episode was calculated using the Kaplan–Meier analysis and was compared among groups using log-rank test. The effects of potential variables that may influence the incidence of PONV were analyzed by Cox regression. Changes in QTc intervals were compared among groups using factorial analysis of variance with repeated measures. This was followed by a post hoc Tukey test for intergroup comparison.
We searched the databases of Medline (from 1966), Embase (from 1982), and the Cochrane Library without language restriction. Different search strategies were used with the free text terms “droperidol,” “ondansetron,” “nausea,” “vomiting,” or “emesis” “randomized,” “surgery,” “surgical,” or “postoperative,” and combinations of these words. We included only full reports of randomized, placebo-controlled trials that tested the antiemetic efficacy of combining droperidol and ondansetron with either drug alone. Trials that did not have a placebo group were excluded, because the relative antiemetic efficacy could not be determined. We took the effect size of each included study as the difference between the observed and predicted rate of complete response described above. A pooled estimate was calculated by combining the effect sizes according to the DerSimonian and Laird’s random effects method (11).
The study recruited patients between September 2002 and June 2004. Among 518 eligible patients, 400 were enrolled into the study. Six patients were withdrawn after randomization because of cancellation of surgery (three in the control group; two in the droperidol group and one in the ondansetron and droperidol group). None of these patients received the assigned study drug. Therefore, the final dataset consisted of 394 patients (Fig. 1). One patient in the control group mistakenly received metoclopramide before the end of surgery. This patient was included in the allocated group for all analyses. She continued to suffer from PONV.
At baseline, patient characteristics and operative details were similar among groups (Table 1). Overall, 285 patients (72%) reported three or more of the common risk factors for PONV, namely female gender, history of PONV or motion sickness, nonsmoker or use of postoperative opioids. The incidence of risk factors was evenly distributed among groups.
During the entire study period, 68 of 97 patients (70.1%) in the control group had one or more episodes of PONV. These were reported most often in the postanesthetic care unit. The incidence decreased gradually over time (Fig. 2). PONV was rare after the first postoperative day. The median time to the first episode of PONV in the control group was 4.7 h (95% CI 3.7–5.7) (Fig. 3). The incidence of PONV after a single dose of ondansetron 4 mg or droperidol 1.25 mg was 28.0% and 29.6%, representing a relative reduction of PONV by 60% and 58%, respectively. The respective times to first PONV episode after ondansetron and droperidol administration were also delayed, 9.0 h (95% CI 8.0–9.9) and 8.9 h (95% CI 7.9–9.1). The combination of ondansetron and droperidol further reduced the incidence of PONV to 12.1% (i.e., 78% relative reduction). The observed rate of complete response after combination antiemetic therapy, 87.9% (95% CI 79.8–93.6) was similar to that predicted by our model, 88.2% (95% CI 82.1–92.9), Fisher’s exact test, P = 0.94.
Table 2 shows the analgesic requirement after surgery. There was no difference in pain scores among groups. Morphine consumption and the number of dologesic tablets received throughout the study period were also similar among groups.
Multivariate analysis using the Cox regression model suggested that nonsmoker, history of motion sickness, history of PONV increased the likelihood of current PONV. The odds ratios were 2.07 (95% CI 1.38–3.10), 1.82 (95% CI 1.27–2.61) and 2.38 (95% CI 1.61–3.51), respectively. The need for of postoperative morphine was low and did not affect the incidence of PONV, odds ratio 1.14 (95% CI 0.67–1.65), P = 0.38.
There was no adverse event during the injection of study drug. Arterial blood pressure and heart rate did not change after drug administration. Sedation scores were similar among groups, and no patient had a score more than 2 over the entire study period.
The effects of study drugs on QTc interval are summarized in Table 3. Baseline QTc intervals were similar among groups. At 5 min after injection of the treatment medications, there was an increase in QTc interval by 2.7–3.5% (P < 0.001). The mean (±sd) increase in QTc interval after the combination therapy, 12.5 ± 36.8 ms, was similar to that after ondansetron, 9.9 ± 34.7 ms (P > 0.99), or droperidol alone, 11.3 ± 24.3 ms (P > 0.99). Three hours after surgery, the changes in QTc interval had resolved. In the postoperative interviews, there was no report of palpitation or cardiac dysrrhythmia.
We considered six trials with data from 5378 patients for the meta-analysis (4,6–10). Two studies (n = 400) involved children aged between 1 and 14 yr (7,10). Figure 4 shows the pooled analysis of the interaction effect between ondansetron and droperidol for preventing early (0–6 h) and overall PONV. There was no difference between the expected and observed incidence of complete response. The data suggested that the addition of ondansetron to droperidol produced an additive effect. A separate analysis of adult and pediatric data did not demonstrate differences in the interaction effect.
Our study shows that both ondansetron and droperidol are equally effective in preventing PONV after gynecologic laparoscopic procedures. Each of the drugs reduced the incidence of PONV by about 60% (from 68% to 29%). The incidence was further reduced by another 60% when both drugs were given together. Based on our model of interaction, the combination therapy produced an additive effect.
A number of studies have shown that the combination of ondansetron and droperidol provided better prophylaxis against early or overall PONV when compared with either drug alone in adult and pediatric populations (6–10). However, until recently, none of the studies evaluated the pharmacologic interaction between drug combinations. In a large randomized controlled trial, Apfel et al. (4) evaluated the efficacy of ondansetron, droperidol, dexamethasone, remifentanil, avoidance of nitrous oxide, propofol infusion, and their combinations for the prevention of PONV. In 5199 patients at high risk of PONV, logistic regression analysis showed no interaction among treatments. These data suggested that all treatments acted independently of each other through their specific mechanisms of action, indicating an additive interaction between treatment combinations. This approach, however, assumes that the data can be fitted with linear models and that the relationships between treatments are linear (12). Also, a relatively large sample size is required. As per the model proposed by Apfel et al., 368 patients per group are needed to demonstrate a 9% difference in the incidence of PONV between individual (27%) and a combination of two antiemetic therapies (18%) (13).
We designed the present study to facilitate interaction analysis for a two-drug combination. This method does not assume linearity between treatments, and requires fewer subjects. Overall in our meta-analysis, we found an additive interaction between ondansetron and droperidol for preventing PONV. Nonetheless, we noted a small study actually demonstrated antagonism with the combination. In this study of children undergoing strabismus repair, the incidence of PONV at 2 h after surgery was more frequent in patients receiving a combination of ondansetron and droperidol compared with either drug alone (7). However, patient characteristics differed among groups. In particular, children receiving combination therapy had longer surgery and were exposed to more anesthetics. This may have accounted for the differences in PONV.
In agreement with Apfel et al.’s method (4,13), our approach for determining drug interaction is limited to a fixed dose combination. Therefore, the finding may not be extrapolated to other dose combinations. In this study, we chose to compare ondansetron 4 mg with droperidol 1.25 mg and their combination because they were commonly prescribed. An overall picture of the pharmacodynamic interactions between the two drugs will require a response surface analysis, in which the combined drug response is determined over the entire dose range of both drugs (14). Unfortunately, this will require a large study of thousands of patients.
Side effect profiles may change when antiemetic therapies are combined. The reported incidence of adverse events after ondansetron are headache, 3.3%; constipation, 4.3%, and transiently increased liver enzymes, 3.2% (15). Droperidol, on the other hand, may cause extrapyramidal symptoms (0.2–1.1%), dizziness (0.7%), and sedation (4.2%) (16). In a meta-analysis, Habib et al. (17) showed no changes in the incidence of adverse events between the combination and ondansetron alone, and there were few patients with headache and dizziness. We did not observe adverse events associated with the administration of ondansetron, droperidol, or their combination. It would appear that the combination therapy is well tolerated. However, the study was not adequately powered to determine the interaction of uncommon end-points.
The timing of study drug administration may not have been optimal in the present study. There was some advantage of administering droperidol after induction of anesthesia to avoid the potentially unpleasant dysphoric effect. Similarly, a previous study (18) suggested that ondansetron given at the end of a procedure provided better protection against PONV compared with that administered before anesthetic induction. However, giving the study drugs at different times will increase the complexity of the study requiring the use of a double-dummy design. Furthermore, it will prevent measuring the changes in QTc interval when the drugs are combined. To simplify the blinding procedure, and to avoid the confounding changes in QTc interval with anesthetics, we administered all study drugs 5 min before induction of anesthesia.
Both ondansetron and droperidol induce prolongation of the QTc interval (19,20). Consequently, there are concerns whether these drugs, given individually or when combined, will precipitate life-threatening events such as torsade de pointes. Indeed, the United States Food and Drug Administration has issued a “black box” warning on the use of droperidol (for doses >2.5 mg) because of the risk of cardiac arrhythmia with such doses (19,20). Our data confirmed previous reports that low-dose droperidol 1.25 mg or ondansetron 4 mg induced a modest and transient prolongation of the QTc interval (21,22). Interestingly, in our study, the combination of both antiemetics did not worsen the changes in QTc interval. Although it remains uncertain as to how the prolongation of QTc interval would affect the risk of perioperative cardiac arrhythmias, a number of recent publications have indicated that the risk in the general population is increasingly small (19,21,22). In this regard, our data suggest that the risk after combination therapy was not worse compared with individual antiemetics.
In summary, this study demonstrated that a combination of ondansetron 4 mg and droperidol 1.25 mg produced an additive effect for preventing PONV after laparoscopic gynecologic surgery. At these dosages, there was no demonstrable side effect. In particular, there was no clinically significant effect on QTc interval, even when ondansetron was combined with droperidol.
1. Gan TJ, Meyer T, Apfel CC, et al. Consensus guidelines for managing postoperative nausea and vomiting. Anesth Analg 2003;97:62–71.
2. Tramer MR, Reynolds DJ, 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;87:1277–89.
3. Eberhart LH, Morin AM, Bothner U, Georgieff M. Droperidol and 5-HT3-receptor antagonists, alone or in combination, for prophylaxis of postoperative nausea and vomiting. A meta-analysis of randomised controlled trials. Acta Anaesthesiol Scand 2000;44:1252–7.
4. Apfel CC, Korttila K, Abdalla M, et al. A factorial trial of six interventions for the prevention of postoperative nausea and vomiting. N Engl J Med 2004;350:2441–51.
5. Bazett HC. An analysis of the time-relation of electrocardiograms. Heart 1920;7:353–67.
6. Pueyo FJ, Carrascosa F, Lopez L, et al. Combination of ondansetron and droperidol in the prophylaxis of postoperative nausea and vomiting. Anesth Analg 1996;83:117–22.
7. Klockgether-Radke A, Neumann S, Neumann P, et al. Ondansetron, droperidol and their combination for the prevention of post-operative vomiting in children. Eur J Anaesthesiol 1997;14:362–7.
8. Bugedo G, Gonzalez J, Asenjo C, et al. Ondansetron and droperidol in the prevention of postoperative nausea and vomiting. Br J Anaesth 1999;83:813–4.
9. Wu O, Belo SE, Koutsoukos G. Additive anti-emetic efficacy of prophylactic ondansetron with droperidol in out-patient gynecological laparoscopy. Can J Anaesth 2000;47:529–36.
10. Shende D, Bharti N, Kathirvel S, Madan R. Combination of droperidol and ondansetron reduces PONV after pediatric strabismus surgery more than single drug therapy. Acta Anaesthesiol Scand 2001;45:756–60.
11. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials 1986;7:177–88.
12. Jaccard J. Interaction effects in logistic regression. Thousand Oaks, CA: Sage Publications, 2001.
13. Apfel CC, Korttila K, Abdalla M, et al. An international multicenter protocol to assess the single and combined benefits of antiemetic interventions in a controlled clinical trial of a 2X2X2X2X2X2 factorial design (IMPACT). Control Clin Trials 2003;24:736–51.
14. Jonker DM, Visser SA, van der Graaf PH, et al. Towards a mechanism-based analysis of pharmacodynamic drug–drug interactions in vivo. Pharmacol Ther 2005;106:1–18.
15. Tramèr MR, Reynolds DJM, Moore RA, McQuay HJ. Efficacy, dose-response, and safety of ondansetron in prevention of postoperative nausea and vomiting: a qualitative systematic review of randomized placebo-controlled trials. Anesthesiology 1997;87:1277–89.
16. Eberhart LH, Morin AM, Bothner U, Georgieff M. Droperidol and 5-HT3-receptor antagonists, alone or in combination, for prophylaxis of postoperative nausea and vomiting. A meta-analysis of randomised controlled trials. Acta Anaesthesiol Scand 2000;44:1252–7.
17. Habib AS, El-Moalem HE, Gan TJ. The efficacy of the 5-HT3 receptor antagonists combined with droperidol for PONV prophylaxis is similar to their combination with dexamethasone. A meta-analysis of randomized controlled trials. Can J Anaesth 2004;51:311–19.
18. Tang J, Wang B, White PF, et al. The effect of timing of ondansetron administration on its efficacy, cost-effectiveness, and cost-benefit as a prophylactic antiemetic in the ambulatory setting. Anesth Analg 1998;86:274–82.
19. White PF. Droperidol: a cost-effective antiemetic for over thirty years [Editorial]. Anesth Analg 2002;95:789–90.
20. Shafer SL. Safety of patients reason for FDA black box warning on droperidol. Anesth Analg 2004;98:551–2.
21. White PF, Song D, Abrao J, et al. Effect of low-dose droperidol on the QT interval during and after general anesthesia: a placebocontrolled study. Anesthesiology 2005;102:1101–5.
22. Charbit B, Albaladejo P, Funck-Brentano C, et al. Prolongation of QTc interval after postoperative nausea and vomiting treatment by droperidol or ondansetron. Anesthesiology 2005;102:1094–100.