Metoclopramide is an active antiemetic against cisplatin-induced acute emesis. However, the optimal administration method (continuous infusion versus intermittent short infusion) for metoclopramide has not yet been clearly defined. We have conducted a randomized crossover study to compare the antiemetic efficacy of continuous infusion of metoclopramide with that of intermittent short infusion of metoclopramide in 54 evaluable patients. Patients were stratified according to sex and were randomized to receive either a continuous-infusion regimen (regimen A) or an intermittent-short infusion regimen (regimen B). Patients were switched to the alternate therapy in the second course. In regimen A, metoclopramide at 3 mg/kg i.v. was given before cisplatin, and then metoclopramide at 4 mg/kg was infused intravenously over 7.5 hours. In regimen B, metoclopramide at 3 mg/kg i.v. was followed by 2 mg/kg i.v. for two doses. Dexamethasone and diphenhydramine were given intravenously in both regimens. There was no significant difference between two regimens in their ability to prevent emesis. Complete protection (no episode of emesis) and major protection (< or = 2 episodes of emesis), respectively, were obtained by 67% (95% confidence interval: 53-79%) and 85% (95% confidence interval: 73-93%) of all patients given regimen A and by 59% (95% confidence interval: 45-72%) and 81% (95% confidence interval: 68-91%) of those given regimen B. The two regimens were also equally effective in controlling nausea. However, male patients showed better control of nausea and vomiting than did female patients, regardless of treatment regimen. Toxicity was mild in both regimens and was well tolerated. Our findings indicate that both continuous-infusion metoclopramide and intermittent-short infusion metoclopramide are effective in controlling cisplatin-induced acute nausea and vomiting.