The feasibility of combination irinotecan, carboplatin and docetaxel chemotherapy as first-line treatment for advanced epithelial ovarian carcinoma was assessed. One hundred patients were randomised to receive four 3-weekly cycles of carboplatin (area under the curve (AUC) 7) followed by four 3-weekly cycles of docetaxel 100 mg m(-2) (arm A, n=51) or docetaxel 60 mg m(-2) with irinotecan 200 mg m(-2) (arm B, n=49). Neither arm met the formal feasibility criterion of an eight-cycle treatment completion rate that was statistically greater than 60% (arm A 71% (90% confidence interval (CI) 58-81%; P=0.079; arm B 67% (90% CI 55-78%; P=0.184)). Median-dose intensities were >85% of planned dose for all agents. In arms A and B, 15.6 and 12.2% of patients, respectively, withdrew owing to treatment-related toxicity. Grade 3-4 sensory neurotoxicity was more common in arm A (1.9 vs 0%) and grade 3-4 diarrhoea was more common in arm B (0.6 vs 3.5%). Of patients with radiologically evaluable disease at baseline, 50 and 48% responded to therapy in arms A and B, respectively; at median 17.1 months' follow-up, median progression-free survival was 17.1 and 15.9 months, respectively. Although both arms just failed to meet the formal statistical feasibility criteria, the observed completion rates of around 70% were reasonable. The addition of irinotecan to first-line carboplatin and docetaxel chemotherapy was generally well tolerated although associated with increased gastrointestinal toxicity. Further exploratory studies of topoisomerase-I inhibitors in this setting may be warranted.