We performed a double-blind, randomized, placebo-controlled trial to investigate the efficacy and safety of ondansetron in preventing vomiting after tonsillectomy with or without adenoidectomy in children. Sixty children were premedicated with 0.5 mg/kg of oral midazolam and underwent inhaled induction and maintenance of anesthesia with halothane and nitrous oxide. Intravenous morphine 0.075 mg/kg, vecuronium 0.1 mg/kg, and either ondansetron 0.15 mg/kg (maximum = 8 mg), or saline placebo were administered after intravenous catheter placement. All episodes of postoperative vomiting on the day of surgery and the following day were recorded. There were no significant differences between the two groups with regard to age, weight, sex, time to endotracheal intubation, duration of surgery, duration of anesthesia, or number of patients having adenoidectomy. More patients in the ondansetron group received dexamethasone as part of the surgeon's routine protocol (P = 0.03). Twenty-three children (77%) who received ondansetron were emesis-free as opposed to only eight (27%) in the placebo group (P < 0.001). This was especially evident in the number of children who had two or more episodes of vomiting; 17 children (57%) who received placebo had two or more episodes of vomiting as opposed to only two children (7%) who received ondansetron. The incidence of emesis on the day after the surgery did not differ between the two groups. Three children who received ondansetron and one who received placebo complained of abdominal cramping on the night of the surgery. There were no other complaints or complications. Analysis of the 51 children who did not receive dexamethasone revealed similar results.(ABSTRACT TRUNCATED AT 250 WORDS)