The purpose of this study was to evaluate the efficacy and safety of four different doses of granisetron when administered as a single intravenous (i.v.) dose for prophylaxis of cisplatin-induced emesis in a multicenter, randomized, parallel-group, double-blind investigation. A total of 353 chemotherapy-naive patients were enrolled, stratified according to cisplatin dose (moderate dose: 50-80 mg/m2, n = 169; high dose: 81-120 mg/m2, n = 184) and randomized to one of four granisetron doses: 5, 10, 20, or 40 micrograms/kg. Control of emesis was evaluated by the percentages of patients attaining complete response (no vomiting or retching, and no rescue medication) and major response (< or = 2 episodes of vomiting or retching, and no rescue medication). Patients were assessed on an inpatient basis for 18-24 h. Safety analyses consisted of adverse events and laboratory parameter changes. Complete response rates over 24 h after chemotherapy were 23%, 48%, 48%, and 44% for granisetron doses of 5, 10, 20, and 40 micrograms/kg, respectively, in the combined patient population (P = 0.011 for linear trend); 29%, 56%, 58%, and 41%, respectively, in the moderate-dose cisplatin stratum (P = 0.278 for linear trend); and 18%, 41%, 40%, and 47%, respectively, in the high-dose cisplatin stratum (P = 0.011 for linear trend). Transient headache was the most frequently reported adverse event (19%). There was no evidence of association between increased dose and headache. A single 10-, 20- or 40-micrograms/kg dose of granisetron is comparably effective in controlling nausea and vomiting associated with moderate or high-dose cisplatin chemotherapy. Granisetron was safe and well tolerated at all doses.