Two hundred and twenty-four patients with advanced breast cancer were enrolled in a multicenter prospective randomized clinical study and received either doxorubicin (40 mg/m2), or epirubicin (40 mg/m2) or mitoxantrone (12 mg/m2) each combined with cyclophosphamide (600 mg/m2) i.v. In the patient collective the following response rates were observed: complete response 12.1%; partial response 30.6%; stable disease 40.5%; progressive disease 16.8%. A complete response was observed significantly less often in patients where more than one organ site was involved as compared to those patients with only one metastatic site. The mean time period required to reach a best response was 3.7 months. There was no significant difference between the response rates in the three arms. In comparing the observed toxicities in 1,434 treatment cycles, there was a significant difference with regard to leukocytopenia (mitoxantrone arm exhibiting more than either epirubicin and doxorubicin) although infections did not occur more frequently in the mitoxantrone arm; with regard to alopecia, mitoxantrone and epirubicin arms both exhibited less than doxorubicin. It is noteworthy that no patient who had previously received adjuvant chemotherapy achieved a complete response (p = 0.006). The overall significance of these findings can only be clearly evaluated when survival times can be measured.