Ondansetron (GR 38032F), a selective antagonist of serotonin subtype 3 receptors, is effective in the prevention of emesis associated with cisplatin as well as other chemotherapeutic agents. In this randomized, single-blind, multicenter, parallel group study, we compared the efficacy and safety of intravenous (IV) ondansetron with IV metoclopramide in the prevention of nausea and vomiting associated with high-dose (greater than or equal to 100 mg/m2) cisplatin chemotherapy. Three hundred seven patients receiving their first dose of cisplatin, either alone or in combination with other antineoplastic agents, were randomized to receive ondansetron 0.15 mg/kg IV every 4 hours for three doses or metoclopramide 2 mg/kg IV every 2 hours for three doses, then every 3 hours for three additional doses. The study prohibited the concurrent administration of other antiemetics or dexamethasone. Patients receiving ondansetron had a higher rate of complete protection from emesis (40% v 30%, P = .07), a higher complete plus major response rate (65% v 51%, P = .016), a lower rate of failure (21% v 36%, P = .007), and a lower median number of emetic episodes (one v two, P = .005) than did those receiving metoclopramide. The median time to the first emetic episode was longer on ondansetron (20.5 v 4.3 hours, P less than .001). Adverse events occurred in 48% of patients receiving ondansetron and 69% of those receiving metoclopramide (P less than .001). Akathisia and acute dystonic reactions occurred only on metoclopramide; headache (controlled with acetaminophen) was significantly more frequent with ondansetron. Ondansetron is more effective, produces fewer adverse events, and is easier to administer than metoclopramide for the prevention of emesis associated with high-dose cisplatin chemotherapy.