Ninety previously untreated patients with advanced epithelial ovarian cancer (International Federation of Gynecology and Obstetrics stages IIC, III, and IV) were randomized, after initial cytoreductive surgery, to receive paclitaxel (Taxol; Bristol-Myers Squibb Company, Princeton, NJ) 175 mg/m2 as a 3-hour infusion with either carboplatin at an area under the concentration-time curve of 7 (group A) or carboplatin at an area under the concentration-time curve of 7 on courses 1, 3, and 5, alternating with cisplatin 75 mg/m2 on courses 2, 4, and 6 (group B). Treatment was given every 3 weeks, up to a total of six courses. Sixty-one patients (33 and 28 patients in groups A and B, respectively) had residual disease after the initial cytoreductive surgery. Patients with measurable or evaluable disease had a high overall response (82% v 57%). A 52% and 39% complete response rate for groups A and B, respectively, with no statistically significant difference between the groups was also observed. With a median follow-up of 12 months (range, 0.33 to 24 months), 29 patients have progressed (18 and 11 in groups A and B, respectively), and 13 have died (seven and six, respectively). Median time to progression was 20.36 months (range, 0.20 to 23.54 months) for group A, whereas this has not yet been reached for group B. Median survival has not yet been reached, but there is no significant difference between the two groups (P = .6972). Treatment was generally well tolerated. Grade 3 and 4 neutropenia was 20% and 32% for groups A and B, respectively, while grade 3 and 4 thrombocytopenia was 4% and 7%, respectively, with no significant difference between the two groups. In conclusion, both combinations seem very active for the treatment of advanced epithelial ovarian cancer and are associated with acceptable toxicity.