Despite recent advances in control of acute emesis, delayed nausea and vomiting following cisplatin-based chemotherapy remain a significant cause of treatment-related morbidity. Ondansetron, a selective 5HT3 receptor antagonist, is effective in preventing acute emesis in the initial 24-h period following high-dose cisplatin. The efficacy and safety of ondansetron in preventing the delayed emesis syndrome during days 2-5 after cisplatin (> or = 100 mg/m2) were evaluated in a double-blind, placebo-controlled multicentre trial. 50 patients having two or fewer emetic episodes during the first 24 h were randomised to receive ondansetron (16 mg) or placebo orally three times daily beginning 24 h after cisplatin. Rates of complete control of emesis were higher in ondansetron-treated patients during each study day, 59-78%, compared with 39-50% in placebo-treated patients, but the differences were statistically superior only on the third study day (P = 0.009). 40% of patients in the ondansetron treatment arm and 33% treated with placebo had complete control of emesis during the entire 4-day study period (P = 0.648). Withdrawal from study due to nausea and vomiting occurred in 13% of ondansetron-treated patients compared with 33% in the placebo arm (P = 0.102). Control of nausea was better with ondansetron, but differences were not statistically significant. Adverse effects of oral ondansetron given in this dose schedule were minimal. These data suggest that the delayed emesis syndrome may be partially mediated through the 5HT3 receptor, but that a serotonin antagonist alone provides inadequate control. Further investigation of ondansetron-based therapy in this clinical setting is warranted.