Secondary Logo

Journal Logo

Ambulatory Anesthesia: Research Report

A Randomized, Double-Blind Comparison of the NK1 Antagonist, Aprepitant, Versus Ondansetron for the Prevention of Postoperative Nausea and Vomiting

Section Editor(s): Glass, Peter S. A.Gan, Tong J. MD*; Apfel, Christian C. MD, PhD; Kovac, Anthony MD; Philip, Beverly K. MD§; Singla, Neil MD, PhD; Minkowitz, Harold MD; Habib, Ashraf S. MBBCh, MSc, FRCA*; Knighton, Jennifer MS#; Carides, Alexandra D. PhD#; Zhang, Hong PhD#; Horgan, Kevin J. MD#; Evans, Judith K. MD#; Lawson, Francasca C. MD#

Author Information
doi: 10.1213/01.ane.0000263277.35140.a3

Nausea and vomiting are common postoperative complications, and can occur in up to 80% of patients at high risk for postoperative nausea and vomiting (PONV) without antiemetic prophylaxis (1–4). However, currently available antiemetics, including IV 5HT3 receptor antagonists, do not provide complete protection (5), and there is still a medical need for more effective therapies to prevent PONV.

Substance P, a regulatory peptide that is the preferred endogenous ligand at neurokinin-1 (NK1) receptors, is found in the gastrointestinal tract (vagal afferents) and areas of the central nervous system thought to be involved in the vomiting reflex (including the nucleus tractus solitarii and area postrema) (6–9). Although 5HT3 receptor antagonists have questionable efficacy against centrally induced emesis, nonpeptide NK1 receptor antagonists have demonstrated activity against both peripheral and central emetic stimuli in animal models (10–14). Evidence suggesting the potential efficacy of NK1 receptor antagonists against PONV was obtained in pilot studies of two other compounds in this class that were assessed in patients undergoing major gynecologic surgery (15,16).

Aprepitant is the first NK1 receptor antagonist available for clinical use as an antiemetic (17). As part of combination therapy with other antiemetics, aprepitant is approved for use and recommended in consensus guidelines for the prevention of chemotherapy induced nausea and vomiting (CINV) (18); the clinical profile of aprepitant suggests that it may provide benefit against PONV as well. Aprepitant is a highly selective, brain-penetrating NK1 antagonist with a long half-life of 9–12 h and preclinical efficacy against opioid-induced emesis (11,12,17). The present study is the first to evaluate the efficacy and side effect profile of oral aprepitant for the prevention of PONV. An active antiemetic, IV ondansetron, was chosen as the control. Therefore, the objective of this study was to compare the clinical profile of aprepitant versus that of ondansetron. The primary efficacy hypothesis was that in the 24 h after surgery, the proportion of patients with complete response, defined as no vomiting and no use of rescue therapy, would be significantly higher among those taking aprepitant than among those taking ondansetron.


Twenty-nine centers participated in this study (protocol 090), conducted from September 26, 2003 to November 24, 2004. Appropriate IRB approval was obtained, and all patients gave written informed consent.

Eligible patients were at least 18-yr-old, were scheduled to undergo open abdominal surgery requiring an overnight hospital stay, met criteria of ASA physical status of I-III, and were scheduled to receive general anesthesia including nitrous oxide with volatile anesthetics. Patients were excluded who were pregnant or breast-feeding, undergoing surgery requiring routine placement of a nasogastric or oral-gastric tube, or receiving spinal regional or propofol-maintained anesthesia. Also excluded were those who had vomiting of any organic etiology, had vomited for any reason within 24 h of surgery, or had abnormal laboratory values as specified by the protocol (alanine aminotransferase or aspartate aminotransferase >2.5 × upper limit of normal, bilirubin >1.5 × upper limit of normal, or creatinine >1.5 × upper limit of normal). Medications metabolized by CYP3A4 and known to have a narrow therapeutic index were prohibited, and patients taking such medications who were unable to discontinue them for the duration of the study were excluded.

Patients were allocated to a treatment group within each clinical site using a computer-generated random allocation schedule stratified according to gender. To ensure blinding among staff employed by the sponsor who were involved with the study, the randomization schedule was generated by a statistician who was otherwise uninvolved with the study. On the day of surgery, patients were randomized to one of three preoperative treatments: oral aprepitant 40 mg, oral aprepitant 125 mg, or IV ondansetron 4 mg. Matching placebos were used to maintain blinding. Blinded allocation schedules and supplies of aprepitant were provided by the sponsor, and each study site designated an unblinded pharmacist to receive, store, and prepare the ondansetron and saline placebo.

Patients received aprepitant or placebo 1–3 h before anticipated induction of anesthesia, and ondansetron or placebo by push over 2–5 min before induction, according to the prescribing information for ondansetron. Anesthesia consisted of optional premedication with a benzodiazepine; induction with any anesthetic; opioids; neuromuscular blocking drugs; maintenance of anesthesia with nitrous oxide (50%–70%) with a volatile anesthetic; and the reversal of neuroblockade with neostigmine (2–5 mg) in combination with either atropine or glycopyrrolate. Patients could not receive additional prophylactic antiemetics within 24 h preoperatively, intraoperatively, or postoperatively, although patients could request rescue antiemetics for established PONV.

All intraoperative medications and the duration of anesthesia were recorded. The time of last suture/staple was recorded as T0 (hours). Efficacy was assessed at 0–48 h (at 0, 2, 6, 24, and 48 h) after surgery. Patients were monitored continuously in the postanesthesia care unit, and emetic episodes and/or use of rescue therapy were recorded throughout the hospital stay. An emetic episode was defined as one or more continuous episodes of vomiting (oral expulsion of stomach contents) or retching (an attempt to vomit that is not productive of stomach contents); distinct episodes were those occurring at least 1 min apart. Nausea was assessed at 2, 6, 24, and 48 h postoperatively, at any time the patient complained of nausea, and immediately before administration of rescue medication. Patients rated nausea on an 11-point Verbal Rating Scale (VRS), with 0 equal to “no nausea” and 10 equal to “nausea as bad as it could be.” Rescue medication was offered if the patient had more than one episode of vomiting or retching, if the patient had nausea lasting longer than 15 min, or if the patient requested it for established nausea (VRS score >0) or vomiting.

Safety was assessed by physical examinations and laboratory tests, and included 12-lead electrocardiograms performed at baseline and 24 h postoperatively; all adverse events were recorded until 14 days after surgery. Additional safety assessments included awakening time (interval between end of surgery and patient's ability to obey commands) and duration of recovery from anesthesia (postanesthesia recovery score of 8 on a 0–10 scale) (19).

Patients were discharged ≥24 h postoperatively based on clinical criteria. For patients discharged before 48 h, the study coordinator contacted the patient at the 48-h timepoint to record emetic episodes, nausea VRS, and/or rescue therapy. Adverse events occurring within 14 days of surgery were documented at a follow-up visit or call required within 3 wk after surgery.

The sponsor managed the data and performed the analyses for this study, which were reviewed by the primary authors and others. The primary efficacy end point was the proportion of patients with complete response, defined as no vomiting and no use of rescue therapy, in the 24 h after surgery, to be tested for superiority of aprepitant. Secondary end points were no vomiting 0–24 h, no rescue therapy 0–24 h, and no vomiting 0–48 h. The primary efficacy analyses were conducted on a modified intent-to-treat (MITT) population, which included all patients who received study drug, underwent surgery under general anesthesia, and had at least one posttreatment assessment. Treatment comparisons were made with logistic regression models that included terms for treatment and sites. Although gender was a stratification factor in the randomization, <10% of the patients were male; therefore, the model did not include gender as a factor. A step-down procedure was used to account for multiple tests within the primary hypothesis (two doses of aprepitant compared with ondansetron) as follows: aprepitant 125 mg and ondansetron were compared at the 0.05 significance level; if the difference in complete response rate was significant, aprepitant 40 mg was then compared with ondansetron at the 0.05 level. A similar method was used to account for the two dose-level comparisons for each secondary end point, as well as for the multiple secondary end points themselves. The single highest VRS score recorded for each patient over 0–24 h was considered the individual peak nausea score; treatment groups were compared using Wilcoxon's ranked sum test. A post hoc analysis was performed to determine percentages of patients in each treatment group with peak nausea VRS scores in the range of 0 (i.e., no nausea) to 4 (no significant nausea), based on recent research on nausea categorization scales suggesting that a VRS score of 4 is a relevant cutoff representing no nausea to mild nausea (20). Kaplan-Meier curves were generated for time to first emesis during the first 48 h and log-rank testing was used to compare treatments. The rates of incidence of emetic episodes were compared among treatment groups using a Poisson regression analysis, which included factors for use of rescue medication.

Based on a sample size of 240 evaluable patients per treatment group, the study had 90% power to detect a 15 percentage-point difference between the aprepitant 125 mg group and the ondansetron group for the primary end point, assuming a 50% response rate for ondansetron and a rate of 65% for aprepitant 125 mg.


Patient Enrollment

Disposition of the 805 patients randomized into the study is shown in Figure 1. All randomized patients who received study drug (n = 766) were included in the tolerability assessments; data were excluded from 39 patients who either did not receive any study drug (n = 34) or who, due to incorrect preparation of the vial for IV administration at the study site, received either placebo for aprepitant and placebo for ondansetron, or active aprepitant and active ondansetron (n = 5). The efficacy analyses excluded these 39 patients plus another 14 patients who either did not undergo surgery but did receive drug (n = 1), did not have posttreatment efficacy assessments (n = 2), or did not receive Vial B but received active drug in Bottle A (n = 11). An additional 19 patients, for whom significant violations of study conduct and data collection occurred at one site, were also excluded from the efficacy analyses before unblinding occurred, although safety data from these 19 patients were included in the safety assessments.

Figure 1
Figure 1:
Figure 1.

No differences were noted among groups in terms of reasons for discontinuation, and no patients discontinued due to an adverse event. The treatment groups also did not differ in terms of patient baseline characteristics, including risk factors for PONV (Table 1). Among patients (<10%) who did not undergo gynecologic surgery, procedures included prostatectomy, intestinal resection, hernia repair, bladder surgery, cholecystectomy, or nephrectomy.

Table 1
Table 1:
Baseline Characteristics by Treatment Group, for All Patients Who Received Active Study Drug


No significant interactions were observed between treatment and investigative site, age, duration of surgery, or risk factors for PONV. As shown in Figure 2, there was no significant difference in the percentage of patients with no vomiting and no rescue (complete response) over 0–24 h between aprepitant 40 mg (45%) or 125 mg (43%) and ondansetron (42%; P > 0.5 for both odds ratios of aprepitant:ondansetron). Likewise, in a post hoc sensitivity analysis accounting for the 19 patients excluded from the MITT analysis, the rates of complete response did not differ among groups (45%, 44%, and 42% for aprepitant 40 mg, 125 mg, and ondansetron, respectively; P > 0.5 for both odds ratios). Thus, the primary hypothesis that aprepitant would have been superior to ondansetron with respect to complete response was not met. Over 0–24 h, the treatments did not differ significantly in terms of no use of rescue therapy (45%, 44%, and 46% for aprepitant 40 mg, 125 mg, and ondansetron, respectively) (Fig. 3). However, larger proportions of patients in both aprepitant groups had no vomiting compared with the ondansetron group (odds ratios = 3.2 for aprepitant 40 mg versus ondansetron and 6.8 for aprepitant 125 mg versus ondansetron; P < 0.001 for both ratios) (Fig. 4). Similar results were observed for the 0–48 h interval (odds ratios = 2.7 for aprepitant 40 mg versus ondansetron and 6.9 for aprepitant 125 mg versus ondansetron; P < 0.001 for both ratios) (Fig. 4). When rates of incidence of vomiting were adjusted for use of rescue therapy in the Poisson regression analysis, the ratio of episodes of vomiting (aprepitant:ondansetron) was 1:4 for aprepitant 40 mg and 1:7 for aprepitant 125 mg.

Figure 2
Figure 2:
Figure 2.
Figure 3
Figure 3:
Figure 3.
Figure 4
Figure 4:
Figure 4.

During the first 48 h after surgery, both doses of aprepitant delayed the time to first vomiting compared with ondansetron (P < 0.001) (Fig. 5).

Figure 5
Figure 5:
Figure 5.

The distributions of peak nausea VRS scores were not different across treatment groups (median = 5.0 in all groups; lower 25% = 0.0 in all groups; upper 25% = 7.5 in the aprepitant 40 mg group, 8.0 in the aprepitant 125 mg group, and 8.0 in the ondansetron group). Fifty percent of patients in the aprepitant 40 mg group and 49% of patients in the aprepitant 125 mg group had peak scores 0–4, compared with 43% of patients in the ondansetron group. As for the primary end point, sensitivity analyses performed for each secondary end point showed that the between-treatment results were unchanged regardless of inclusion or exclusion of the 19 patients originally excluded from the MITT efficacy analyses.


A summary of adverse events is displayed in Table 2. No significant differences were observed in the incidences of serious clinical adverse events reported in the three treatment groups. One patient in the aprepitant 40 mg group with myeloproliferative disorder died 28 days postoperatively. The death was determined by the investigator not to have been related to study drug. One serious adverse event, a case of mild constipation, which prolonged the hospital stay of a patient who received aprepitant 125 mg, was considered related to study drug. Two patients (1 in the aprepitant 125 mg group and 1 in the ondansetron group) had a serious adverse event consistent with respiratory depression but neither was considered related to study drug, and there were no reported serious adverse events consistent with excessive sedation. There was no between-treatment difference in the incidence of most of the common adverse events including headache (Table 2). The treatment groups did not differ significantly in terms of awakening time, duration of recovery from anesthesia, or percentages of patients with QTc interval prolongations on electrocardiograms performed 24 h after surgery (Table 2).

Table 2
Table 2:
Summary of Adverse Events (AEs)


The NK1 antagonist, aprepitant, is currently used for the prevention of CINV, but data suggest that drugs of this class may also have efficacy in the prevention of PONV. This study compared aprepitant with ondansetron, an active treatment widely used for prevention of PONV, in patients undergoing open abdominal surgery. Aprepitant was evaluated at two doses (40 and 125 mg), selected based on earlier studies of aprepitant for prevention of CINV (22,23); however, the study was not designed formally to compare these doses. Several risk factors were represented in the patient population, which consisted mainly of women (94%); about one-third of patients had a history of PONV, about two-thirds were nonsmokers, and all but four patients received postoperative opioids. The primary hypothesis—that aprepitant would be superior to ondansetron for complete response 0–24 h after surgery—was not met.

Comparisons for each variable of the primary end point (no vomiting and no use of rescue) showed that aprepitant provided better protection against vomiting in the initial 24 h after surgery. The difference between aprepitant and ondansetron was even more pronounced when antiemetic efficacy was assessed over a longer period (0–48 h after surgery), consistent with the shorter half-life of ondansetron compared with that of aprepitant. Aprepitant also delayed the time to first vomiting compared with ondansetron. For the no-vomiting end points, efficacy may not necessarily be attributable only to study drug, as some patients who had no vomiting may have received rescue therapy. However, an additional analysis that accounted for use of rescue therapy showed that, on average, four to seven episodes of vomiting were recorded in the ondansetron group for every one episode recorded in each of the aprepitant groups. Furthermore, smaller studies have reported that ondansetron may be more effective if administered at the end of surgery rather than before induction (24), which suggests that the efficacy of ondansetron in the present study may be less than what it might have been if ondansetron had been given at the end of surgery. However, the study was not designed to assess this possibility, as the design called for ondansetron to be given before induction according to the label.

By contrast with the incidence of no vomiting, the groups were not different in terms of use of rescue therapy, the other variable of complete response. Although explicit criteria were provided regarding the administration of rescue therapy, the perception of nausea is subjective, and the frequency at which rescue therapy was offered, or the threshold for administration of rescue for nausea, may have differed among institutions or clinical staff caring for the patient. Thus, a patient with slight nausea may have received rescue (and therefore failed the complete response end point), whereas another patient who did not take rescue despite moderate or severe nausea may have achieved the complete response end point. Indeed, of 289 patients who had nausea assessed within 1 h before the administration of rescue medication, 71 patients (25%) had no significant nausea (VRS ≤4); 114 (39%) had a VRS 5–7, and 104 (36%) had a VRS ≥8. It is therefore possible that frequent use of rescue therapy, or its variable relationship with nausea, may have influenced the incidence of complete response.

The distributions of peak nausea scores did not differ significantly across groups, although the proportion of patients in the aprepitant 40 mg group with no significant nausea exceeded that in the ondansetron group by 7 percentage points. These data suggest that the antinausea effect of aprepitant may be at least similar to that of ondansetron, which has been shown to prevent nausea as well as emesis (25). Additional analyses of nausea data will be included in a separate report.

Minimizing the incidence of postoperative vomiting has a range of potential benefits, including decreasing the likelihood of serious complications, such as aspiration, and allowing the patient to be discharged and/or to resume normal activities sooner. Reducing the likelihood that a patient will vomit can also be of particular importance after gastro-esophageal surgery, neurosurgery, or procedures requiring wiring of the jaw. Pharmacoeconomic benefits may also be realized; a previous study on cost-analysis showed that the incidence of vomiting determined the rate of unexpected hospital admission and hence increased costs (26). Thus, reducing the likelihood that a patient will actually vomit can be expected to have direct and favorable clinical consequences.

The adverse event profiles across all three groups were typical of surgical patients and did not differ significantly. Gastrointestinal disorders were the most commonly reported serious adverse events. Among all adverse events, pruritus, nausea beyond 48 h, constipation, and decreased hemoglobin were the most commonly reported and incidence rates did not differ across groups. No between-treatment differences were seen for other laboratory assessments, including electrocardiogram findings at 24 h after surgery. Special attention was given to certain safety variables in order to evaluate the possibility of a drug interaction between aprepitant, which is a dose-dependent CYP3A4 inhibitor (17), and other CYP3A4-metabolized drugs commonly used in the perioperative setting, such as fentanyl or midazolam. The findings indicated that aprepitant did not appear to differ from ondansetron in terms of potential influence on the pharmacodynamics of midazolam or fentanyl.

In conclusion, aprepitant was generally well tolerated in this study, which assessed more than 700 patients undergoing general anesthesia. Although aprepitant did not show superiority for the complete response, nausea control and rescue antiemetic use, its prevention of vomiting was better than that of ondansetron for both the 24- and 48-h postoperative periods. Further studies of aprepitant are needed to demonstrate efficacy in other populations including children.


The authors thank Dr. T. Reiss for his helpful comments on the manuscript.


1. Watcha MF, White PF. Postoperative nausea and vomiting: its etiology, treatment, and prevention. Anesthesiology 1992;77: 162–84.
2. Cohen MM, Duncan PG, DeBoer DP, Tweed WA. The postoperative interview: assessing risk factors for nausea and vomiting. Anesth Analg 1994;78:7–16.
3. Apfel CC, Läärä E, Koivuranta M, et al. A simplified risk score for predicting postoperative nausea and vomiting. Anesthesiology 1999;91:693–700.
4. Gan TJ, Ginsberg B, Grant AP, Glass PS. Double-blind, randomized comparison of ondansetron and intraoperative propofol to prevent postoperative nausea and vomiting. Anesthesiology 1996;85:1036–42.
5. Apfel CC, Korttila K, Abdalla M, et al; IMPACT Investigators. A factorial trial of six interventions for the prevention of postoperative nausea and vomiting. N Engl J Med 2004;350:2441–51.
6. Leslie RA. Neuroactive substances in the dorsal vagal complex of the medulla oblongata: nucleus of the tractus solitarius, area postrema, and dorsal motor nucleus of the vagus. Neurochem Int 1985;7:191–211.
7. Grunberg SM, Hesketh PJ. Control of chemotherapy-induced emesis. N Engl J Med 1993;329:1790–6.
8. Veyrat-Follet C, Farinotti R, Palmer JL. Physiology of chemotherapy-induced emesis and antiemetic therapy. Predictive models for evaluation of new compounds. Drugs 1997;53:206–34.
9. Amin AH, Crawford TBB, Gaddum JH. The distribution of substance P and 5-hydroxy-tryptamine in the central nervous system of the dog. J Physiol 1954;126:596–618.
10. Gonsalves S, Watson J, Ashton C. Broad spectrum antiemetic effects of CP-122,721, a tachykinin NK1 receptor antagonist, in ferrets. Eur J Pharmacol 1996;305:181–5.
11. Tattersall FD, Rycroft W, Francis B, et al. Tachykinin NK1 receptor antagonists act centrally to inhibit emesis induced by the chemotherapeutic agent cisplatin in ferrets. Neuropharmacology 1996;35:1121–9.
12. Tattersall FD, Rycroft W, Hill RG, Hargreaves RJ. Enantioselective inhibition of apomorphine-induced emesis in the ferret by the neurokinin1 receptor antagonist CP-99994. Neuropharmacology 1994;33:259–60.
13. Tattersall FD, Rycroft W, Cumberbatch M, et al. The novel NK1 receptor antagonist MK-0869 (L-754,030) and its water soluble phosphoryl prodrug, L-758,298, inhibit acute and delayed cisplatin-induced emesis in ferrets. Neuropharmacology 2000;39:652–63.
14. Hesketh PJ, Van Belle S, Aapro M, et al. Differential involvement of neurotransmitters through the time course of cisplatin-induced emesis as revealed by therapy with specific receptor antagonists. Eur J Cancer 2003;39:1074–80.
15. Gesztesi Z, Scuderi PE, White PF, et al. Substance P (neurokinin-1) antagonist prevents postoperative vomiting after abdominal hysterectomy procedures. Anesthesiology 2000;93:931–7.
16. Diemunsch P, Schoeffler P, Bryssine B, et al. Antiemetic activity of the NK1 receptor antagonist GR205171 in the treatment of established postoperative nausea and vomiting after major gynaecological surgery. Br J Anaesth 1999;82:274–6.
17. EMEND® capsules (Merck) (aprepitant). USA [package insert]. 2006.
18. Gralla RJ, Roila F, Tonato M. The 2004 Perugia Antiemetic Consensus Guideline process: methods, procedures, and participants. Support Care Cancer 2005;13:77–9.
19. Aldrete JA, Kroulik D. A postanesthetic recovery score. Anesth Analg 1970;49:924–34.
20. Boogaerts JG, Vanacker E, Seidel L, et al. Assessment of postoperative nausea using a visual analogue scale. Acta Anaesthesiol Scand 2000;44:470–4.
21. Miettinen O, Nurminen M. Comparative analysis of two rates. Stat Med 1985;4:213–26.
22. Chawla SP, Grunberg SM, Gralla RJ, et al. Establishing the dose of the oral NK1 antagonist aprepitant for the prevention of chemotherapy-induced nausea and vomiting. Cancer 2003;97: 2290–300.
23. de Wit R, Hesketh PJ, Warr D, et al. The oral NK1 antagonist aprepitant for prevention of nausea and vomiting in patients receiving highly emetogenic chemotherapy: a review. Am J Cancer 2005;4:35–48.
24. 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.
25. Apfel C, Paura A, Jokela R, et al. Ondansetron comparably reduces the relative risk of nausea and vomiting. Anesth Analg 2005;S4:100.
26. Hill RP, Lubarsky DA, Phillips-Bute B, et al. Cost-effectiveness of prophylactic antiemetic therapy with ondansetron, droperidol, or placebo. Anesthesiology 2000;92:958–67.
© 2007 International Anethesia Research Society