60 patients with metastatic breast cancer were entered in a phase II study using folinic acid, 5-fluorouracil bolus and infusion and mitoxantrone with or without cyclophosphamide. 47 had measurable visceral metastases and 13 had exclusively bone metastases. 36 had received previous adjuvant or metastatic treatment (33/36 with anthracycline-based regimens). Overall response rate in visceral metastatic patients was 57.1% [95% confidence interval (CI) 35.4-78.8%]; 45.5% and 70% in previously and non-previously treated patients, respectively; duration of response was 9 and 13 months, respectively. 10 out 13 patients with exclusive bone metastases improved for a median time of 18 months. Median survival was 22 months for the 60 patients; 18 and 31 months for previously and non-previously treated patients, respectively. Cyclophosphamide was scheduled only in the absence of nadir grade 4 neutropenia. However, this toxicity occurred in the first 7 patients. For this reason, we chose to avoid cyclophosphamide in patients over 60 years, or with a performance status of 1-2, or who had received previous chemotherapy. Overall, cyclophosphamide was stopped due to nadir grade 4 neutropenia in 17/24 patients for whom this drug was planned. When mitoxantrone, 5-fluorouracil and folinic acid were used at the doses scheduled, the addition of cyclophosphamide appeared feasible in only about 25% of the patients. Furthermore, survival was identical for patients receiving or not receiving cyclophosphamide. Therefore, cyclophosphamide does not contribute substantially to this regimen. This study confirms the value of folinic acid, 5-fluorouracil and mitoxantrone in metastatic breast cancer.