In a multicentre trial, 78 patients with a variety of malignancies, who had experienced insufficient control of emesis (greater than or equal to 3 episodes within 24 hours) while receiving standard antiemetics during previous chemotherapy, were randomly assigned to receive tropisetron 5mg once daily for 5 days or conventional antiemetic drugs. No attempt was made to standardise the conventional antiemetic treatment, which was given according to the usual practice of the participating institutions. Emesis was evaluated by counting emetic episodes and nausea by asking the patients to record on a diary chart the duration and severity of the nausea. Emesis was much better controlled with tropisetron than with standard drugs, complete control during the first 24 hours being achieved in 42% and 8% of patients, respectively, (p less than 0.001). Nausea was of significantly shorter duration (6.9 vs 10.3 hours; p less than 0.01) and was less severe (p less than 0.005) in the tropisetron group. The patients' overall assessment of treatment outcome was markedly better for tropisetron than for the standard antiemetic therapy. The superior efficacy of tropisetron was especially marked during the first 24 hours. For delayed nausea, no significant difference between treatments was seen. No serious adverse effects were observed.