We analyzed the effects of the 5-HT3 receptor antagonist granisetron on both acute and delayed phases of cisplatin-induced emesis in the conscious piglet. Animals that received a high dose of cisplatin (5.5 mg/kg i.v.) were observed continuously for 60 h. Seventeen piglets were treated with cisplatin only and acted as controls. In experimental animals, granisetron (administered before cisplatin) was administered either as a single initial injection (7 mg/kg), alone or in combination with dexamethasone (40 mg), or as multiple injections (1 mg/kg) given every 5 h during the first 30 h of the experiment (cumulative dose: 7 mg/kg). Two other groups of piglets were treated with dexamethasone (40 mg) alone or with multiple injections of ondansetron (7 injections at 3.5 mg/kg), respectively. The latency to the first emetic episode was significantly increased in all groups that received a 5-HT3 receptor antagonist, whatever the agent and the protocol of administration. Piglets treated solely with dexamethasone exhibited a latency similar to that of controls. The total number of emetic events during the 60 h was significantly reduced only in the group of piglets treated repeatedly with granisetron and in the group that received an initial dose (7 mg/kg) of granisetron in combination with dexamethasone. We observed that 3 out of 8 piglets treated repeatedly with granisetron did not vomit throughout the experiment. These results demonstrate that granisetron, when administered repeatedly, is efficacious against delayed emesis. They also suggest that serotonin may be involved in the production of the delayed phase of cisplatin-induced emesis.