A multicentre study was performed at four oncology centres in Sweden, in 160 chemotherapy-naive women, primarily with ovarian or endometrial carcinomas. Abdominal surgery preceded chemotherapy in 146 (91%) women, and another 39 (24%) women had a history of radiotherapy. The chemotherapy regimens contained cisplatin (50-100 mg/m2), in combination with a variety of other agents. In Course 1, all patients received tropisetron (5 mg i.v. Day 1; 5 mg p.o. Days 2-6) and 84% of patients achieved total or partial control of vomiting; 95% of patients achieved total or partial control of nausea during the first 24 hours. Vomiting was least successfully controlled on Day 2 (73% total or partial control) and Days 2-4 for the control of nausea (81, 83, 88% total or partial control, respectively). Patients with partial response in Course 1 (39% of patients) were randomized to addition of dexamethasone or placebo in Course 2. In Course 2, tropisetron plus dexamethasone (Group B2) prevented acute vomiting in 75% of patients, compared with 40% of patients receiving tropisetron plus placebo (Group B1). Over the entire 6 days, there was no vomiting at all in 54% and 20% of B2 and B1 patients, respectively. In Course 2, acute nausea was prevented in 75% of patients (receiving B1) and in 37% of patients (receiving B2). For nausea, over the complete 6 days, the figures were 64% and 3% (p < 0.001). This indicated that patients with incomplete control of emesis in Course 1 benefited from the addition of dexamethasone, provided that it was added for each of the 6 days studied.(ABSTRACT TRUNCATED AT 250 WORDS)