BACKGROUND Only a few studies have evaluated the long-term effects of adjuvant chemotherapy for breast cancer. Furthermore, neither the relation between the risk of second malignancies and type of adjuvant regimen utilized nor the interaction between chemotherapy and breast irradiation or age of the patients have been described in detail. METHODS A total of 2,465 patients entered into prospective studies of CMF-based adjuvant chemotherapy carried out at the Milan Cancer Institute between June 1973 and July 1990 were evaluated. The median follow-up was 12.0 years and detailed information about therapy was available for all patients. RESULTS At 15 years, the cumulative actuarial risk of second malignancies (excluding contralateral breast cancer and basal skin cancer) was 6.7% +/- 0.8% for the total series. The figures were 8.4% +/- 2.9% after local-regional treatment alone, 6.4% +/- 0.9% following CMF, and 5.1% +/- 1.0% following CMF plus Adriamycin (doxorubicin; Farmitalia-Carlo Erba, Milan, Italy). Compared to the general female population, the relative risk following CMF-based adjuvant chemotherapy was 1.29. Three patients, all of whom had received CMF-based chemotherapy, developed acute non-lymphocytic leukemia (cumulative risk 0.23% +/- 0.15%; relative risk 2.3). No differences were evident when breast irradiation was considered, but the cumulative risk of second tumors was slightly higher in women aged > or = 50 years at surgery (7.7% +/- 1.3%) than in younger patients (6.0% +/- 1.0%). CONCLUSIONS At present, there is no evidence of a significantly increased risk of second malignancies following adjuvant CMF-based chemotherapy such as the one given in this case series. A low risk of acute leukemia was associated with the cumulative total dose of cyclophosphamide administered, and breast irradiation did not enhance this risk. CONCLUSIONS Our findings suggest that there is no reason to omit alkylating agents from short-term effective adjuvant chemotherapy.