As valid therapeutic alternatives to adriamycin, with a more favourable safety profile, epirubicin (E) and novantrone (N) were compared in combination with fluorouracil (F) and cyclophosphamide (C) in a prospective randomized clinical trial as first-line treatment for metastatic breast cancer (mbc). 158 women with mbc were randomly allocated to receive FEC or FNC regimen; the dosage in mg/m2 was as follows: 500 for C and F, 75 for E and 10 for N. All drugs were administered iv. on day 1 and recycled on day 21. In 141 evaluable patients the response rate (CR+PR) was better in the FEC (43.6%) than in the FNC regimen (30.3%) (95% C.I. of 32% to 55% versus 14% to 34%), without any statistically significant difference. Differences in response rate were significantly in favour of FEC group in previously untreated patients (57.6% versus 25%, p = .02), and in postmenopausal women (46.1% versus 23.6%, p = .01). No significant differences between the two treatment arms were observed in terms of either time to progression or duration of response and survival. The most important dose-limiting toxicity was hematological (leuko-and thrombocytopenia were significantly higher in FNC-treated patients). This difference in hematological toxicity sustained a significantly different incidence of delays in administering chemotherapy courses, which precluded the administration of comparable doses of all drugs in both groups. The incidence of complete alopecia was significantly higher in FEC-treated patients, while no clinical or instrumental evidence of CHF was observed with either regimen. Due to its more favourable therapeutic profile, the E-containing regimen seems a suitable first-line treatment for previously untreated patients with mbc, while the FNC combination should be offered to women refusing hair loss.