In 46 weaned piglets we surgically implanted a cannula in the jugular vein and electrodes for ECG and EMG recordings. After a 4- to 5-day recovery, piglets were hydrated, then dosed with cisplatin (5.5 mg/kg i.v.) and recorded continuously for the next 60 h. Thirteen piglets (i.e., controls) received only cisplatin. Twenty-three other piglets received, 15 min before cisplatin, an i.v. injection of granisetron (0.25, 0.5, 2 or 7 mg/kg) or ondansetron (0.5, 2 or 7 mg/kg). Ten other piglets received, in addition to cisplatin, multiple injections of granisetron (1 mg/kg) and ondansetron (3.5 mg/kg). All control piglets exhibited both acute and delayed emesis. The first vomiting occurred with a latency of 2.13 +/- 0.82 hr after cisplatin administration; emetic intensity reached a peak (5 vomits/hr) within 2 hr and then decreased rapidly. No vomiting was observed between the 16th and 18th hr. The mean number of vomits during the first 16th was 18.4 +/- 2. Delayed emesis started at the 18th hr and lasted until the 58th hr. The mean number of vomits during the whole of the delayed phase was 9.6 +/- 2.4; the highest emetic intensity (1.2 vomit/hr) occurred between the 21th and the 22th hr. Pretreatment with a 5-HT3 receptor antagonist increased significantly the latency of the first emetic event in a dose-dependent manner. However, the severity of the acute phase was reduced significantly only with granisetron at the dose of 7 mg/kg, although the severity of the delayed phase remained unchanged, irrespective of the dose of granisetron. Three about five piglets treated repeatedly with granisetron did not vomit throughout the chemotherapy course. In contrast, no complete control was observed with repetitive injections of ondansetron. Cisplatin inducing both acute and delayed vomiting in the piglet without any lethality; this animal is a suitable model in which to study the pathogenesis of delayed emesis.