BACKGROUND It is well known that chemotherapy-induced nausea and vomiting not only produce undesirable physical effects but also reduce patients' willingness to continue chemotherapy. The effect of a 5HT3 receptor antagonist, granisetron, on antiemesis was studied in patients who were receiving cisplatin-based chemotherapy. METHODS Forty patients with malignant diseases, who were scheduled to receive cisplatin-based chemotherapy, were randomly assigned to receive either a single intravenous dose of granisetron (3 mg) (n = 20) or multiple doses of metoclopramide (2 mg/kg) (n = 20) as trial. The efficacy of antiemetic therapy was assessed based on the patient's subjective assessment of nausea and vomiting, and the physician's global evaluation of nausea and vomiting in the first 24 hours and during the following 6 days. RESULTS During the first 24 hours, there was a statistically significant advantage for the granisetron group in terms of prevention of both nausea and vomiting. The number of complete responders in the first 24 hours for the granisetron and metoclopramide groups was 16 (80%) and 9 (45%), respectively (p < 0.05). The control rate of anti-nausea over 6 days was 45% and 5% in the granisetron and metoclopramide groups, respectively (p < 0.001). The frequency of adverse events in the granisetron and metoclopramide groups was 10% and 45%, respectively. Analysis of laboratory data showed no significant difference between these two groups prior to and after administration of chemotherapy. CONCLUSIONS A single 5-minute infusion of 3 mg granisetron was more effective and produced a less adverse effect than multiple doses of metoclopramide in controlling chemotherapy-induced emesis, both in the first 24 hours and over a 6-day period.