Background: Since prolongation of survival by chemotherapy has been questioned, palliation balanced with an acceptable quality of life is the primary aim in treating patients with metastatic breast cancer. Patients and Methods: 224 patients from 11 centers were randomized to treatment with 40 mg/m2 of Adriamycin or with 40 mg/m2 of epirubicin or with 12 mg/m2 of mitoxantrone each in combination with 600 mg/m2 of cyclophosphamide every 3 weeks. A special monitoring system including follow-up until death guaranteed valid information on response criteria, survival and quality of life. Results: Treatment outcome was not statistically different between the three groups in terms of best response rate, response duration, time to progression or survival. There were, however, statistically significant differences between the three treatment groups in terms of toxicity and quality of life. Most important, treatment outcome was influenced by the following negative prognostic factors: disease-free interval less than 18 months; metastases at more than one organ site; performance status according to WHO > 1; prior adjuvant chemotherapy; age less than 40 years. Conclusions: This meticulously monitored prospectively randomized study shows that prognostic factors are more important than the chosen treatments for the outcome in metastatic breast cancer patients.