We conducted a prospective, randomized, open, single-center, parallel group study comparing the anti-emetic efficacy and toxicity of granisetron with that of ondansetron in patients receiving moderately emetogenic chemotherapy. From December 1994 to May 1995, patients who were to receive moderately emetogenic chemotherapy for the first time or who had not received chemotherapy (80 to 100 mg/m2 of cisplatin or 40 mg/m2 of doxorubicin) within 4 weeks previously were enrolled in this study. The following anti-emetic regimens were used: 3 mg of granisetron were given intravenously before chemotherapy for a single dose; 8 mg of ondansetron were given intravenously before chemotherapy and then every 8 hours for a total of 3 doses, plus 8 mg of an oral maintenance dose every 12 hours for 5 consecutive days. We evaluated 97 patients (48 received granisetron and 49 received ondansetron). In the first 24 hours after chemotherapy, complete and major responses were achieved in 76.6% of the patients receiving granisetron and in 72.9% of patients receiving ondansetron (p = 0.9033). Additionally, there was no difference in the control of delayed nausea and vomiting between the two groups (51.1% versus 54.2%, p = 0.9200), and there were no significant adverse effects or toxicities. We have concluded that a single dose of granisetron is as effective in prophylaxis of emesis induced by moderately emetogenic chemotherapy as a triple dose of ondansetron plus oral maintenance.