The aim of the study was to evaluate the role of epidoxorubicin plus paclitaxel combination (ET) vs single agent paclitaxel (T), as second-line chemotherapy treatment in advanced ovarian cancer patients in early progression within 12 months after platinum-based chemotherapy. From October 1994 up to June 1999, 234 patients from 34 Italian hospitals were randomised to receive: (A) epidoxorubicin (E) 80 mg m(-2) + paclitaxel (T) 175 mg m(-2) (3 h infusion), every 21 days for 4-6 cycles. (B) Paclitaxel 175 mg m(-2) (3 h infusion) every 21 days for 4-6 cycles. Evaluable for survival analysis were 106 and 106 patients in ET and T arm, respectively. Platinum-based monochemotherapy was the first-line treatment in 43% patients, while polichemotherapy containing anthracyclines was the preferred first-line therapy in 22% patients. The median time from the end of first-line therapy to randomisation was 3 months. Treatment was completed in 87 and 85% of T and ET arm, respectively. Haematological toxicity was significantly more common in ET group (ECOG grade 3-4 neutropenia: 37.4% in ET vs 18.2% in T arm). Neuropathies were similar in both arms (sensory: ECOG grade 2-3: 12.1% in ET vs 14.7% in T arm, motor: 6.1% in ET vs 5.3% in T arm). Objective response was achieved in 37.4% of patients in ET group and in 46.9% of patients in T arm. At a median follow-up of time of 48 months, a total of 180 patients progressed and 163 patients died. Survival analysis showed no difference between ET and T (median time to progression: 6 months for both regimens, median survival: 12 and 14 months for ET and T, respectively; hazard ratio for mortality of ET vs T: 1.17 (95% CI 0.86-1.59; P=0.33). The ET regimen does not seem to be more effective than T in refractory advanced ovarian cancer patients in early progression after platinum-based chemotherapy. Despite an acceptable response rate, the control of disease progression remains poor.