In order to better explore the toxicity and the activity of high dose epirubicin (120 mg/m2, 3 weeks) we analyzed a population of 127 metastatic breast cancer patients, treated in a randomized clinical trial conducted to evaluate the cardioprotective effect of dexrazoxane against epirubicin induced cardiotoxicity. All the patients had a diagnosis of metastatic breast cancer, an ECOG performance status < or = 2 and normal hematologic, renal, hepatic and cardiac function. No prior adjuvant chemotherapy including anthracycline was allowed. Epirubicin was given at the dose of 120 mg/m2 i.v. bolus every 3 weeks. One hundred twenty five patients were evaluable for toxicity and response. Seventeen patients (11%) had a complete response and 47 patients (37%) a partial response, for an overall response rate of 48%. The median progression free and overall survivals were 8.3 months and 18.3 months, respectively. Grade 3 and 4 leukopenia were observed in 8% and 7% of the patients, respectively. The most frequent nonhematological grade 3 toxicities were alopecia (87%), nausea and vomiting (16%), and mucositis (8%). Cardiotoxicity, defined as occurrence of congestive heart failure, decrease in resting left ventricular ejection fraction (L-VEF) to < or = 45%, or 20 EF units decrease from baseline L-VEF, was observed in 19% of the patients, after a median cumulative dose of epirubicin of 720 mg/m2 (range 120-1440). This study confirms in a large series of patients the activity of high dose epirubicin; however, the high incidence of cardiotoxicity requires a careful evaluation of cardiac risk factors before treatment.