Seventy-six patients with advanced breast cancer were entered into the current study. They were randomized to receive either idarubicin (IDA) 45 mg/m2 orally or doxorubicin (DX) 75 mg/m2 intravenously (IV), both drugs being administered every 3 weeks. Among 37 evaluable patients who received DX treatment the overall response rate was 46%, whereas it was 21% in 34 evaluable patients treated with IDA. This difference was statistically significant. In previously untreated patients the response rate with DX was 60% compared to 29% with IDA. Patients with prior chemotherapy had 29% response rate to DX in contrast to 12% with IDA. The median time to response, the median response duration, and the median time to progression were similar in both groups. The median survival of all patients was 20 months in DX arm and 14 months in IDA arm (95% confidence limits 16.69-23.31 and 10.77-17.23, respectively; P = 0.09). Both treatments produced equivalent incidence and severity of myelotoxicity. Gastrointestinal toxicity and alopecia were significantly lower in patients receiving IDA. As for cardiotoxicity, four cases of congestive heart failure were recorded among patients treated with DX whereas no cases occurred in the IDA group. The results of this study indicate that, although DX remains the best single agent available in the treatment of breast cancer, IDA may have a role in selected patients with this disease.