Ondansetron in the prophylaxis of postoperative vomiting: a meta-analysis.

STUDY OBJECTIVES To assess the efficacy of ondansetron and the incidence of headache when used as prophylaxis for postoperative vomiting. DATA SOURCES Studies were identified using MEDLINE (January 1990 to July 1997) with the key words: "vomiting," "ondansetron," and "surgery" and/or "anesthesia." STUDY SELECTION All prospectively randomized trials in which ondansetron and placebo had been administered for prevention of postoperative vomiting. DATA EXTRACTION Data on patient numbers, gender and ages, drug dosage, outcomes (incidence of vomiting and incidence of headache), time of outcome (up to 8 hrs or up to 48 hrs postoperatively), and baseline risk for vomiting according to type of surgical intervention were extracted. MAIN RESULTS 48 trials involving 12,078 patients (10,390 adults and 1688 children) met the selection criteria. No evidence was found that the drug became more effective at doses larger than 4 mg. The dose of 8 mg was not statistically demonstrated to be superior (p = 0.558), while that of 1 mg was barely effective. The meta-analysis indicated that when the incidence of vomiting is elevated (e.g., the combined average of the placebo groups of 48%), on the statistical (i.e., hypothetical) grounds of six patients being treated with 4 mg of ondansetron, one will not vomit due to the treatment and, of the rest, two patients would have vomited despite the treatment and three patients would not have vomited anyway. The overall incidence of headache was 7.05% in ondansetron groups versus 6.16% in placebo groups. CONCLUSIONS While ondansetron is an effective antiemetic with minimal adverse effects, the data obtained on the numbers needed to be treated calculation for prophylaxis of postoperative vomiting should be considered in future cost-effective strategies of postoperative management.