This is the first international, multi-centre, double-blind, randomised, parallel group study to directly compare the efficacy and safety profile of a single intravenous dose of ondansetron (8 or 32 mg) with granisetron (3 mg) in the control of cisplatin-induced acute emesis. A total of 496 patients were randomised to receive one of three anti-emetic treatments prior to cisplatin chemotherapy (> or = 50 mg/m2). Of these, 165 and 162 patients received 8 and 32 mg of ondansetron, respectively, and 169 patients received 3 mg of granisetron. Complete control of emesis (0 emetic episodes) over 24 h was reported in 59% of patients in the 8-mg ondansetron group, 51% of patients in the 32-mg ondansetron group and 56% of patients in the granisetron group. Complete or major control (< or = 2 emetic episodes) was achieved in 76 and 74% of patients in the 8- and 32-mg ondansetron group, respectively, compared with 78% of patients in the granisetron group. Nausea graded none or mild 24 h after the start of cisplatin infusion was reported in 71 and 69% of patients in the 8- and 32-mg ondansetron groups, respectively, and in 73% of patients in the granisetron group. There were no significant differences between the treatment groups when global satisfaction scores were compared. Logistic regression models were fitted to assess any interaction between treatments and prognostic factors (age, gender, alcohol use, cisplatin dose or concomitant chemotherapy) on complete or major response, but there was no evidence of interaction for any factor. All three anti-emetic treatments were well tolerated and no severe or unexpected drug-related adverse events were observed with ondansetron or granisetron. Headache, the most reported drug-related adverse event for all three treatment groups, occurred in 9% of all patients. In summary, no significant difference was observed between any of the treatment groups with respect to emesis, nausea or drug-related adverse events.