Four hundred forty-one women with operable breast cancer with histologically positive axillary nodes were randomized to receive either combination cyclophosphamide (60 mg/m2 orally everyday for 1 year); fluorouracil (300 mg/m2 intravenously [IV] weekly for 1 year); methotrexate (15 mg/m2 IV weekly for 1 year); vincristine (0.625 mg/m2 IV for 10 weeks); prednisone (30 mg/m2 orally days 1 to 14, 20 mg/m2 days 15 to 28, 10 mg/m2 days 29 to 42) (CMFVP) or single-agent melphalan (L-PAM) (5 mg/m2 orally every day for 5 days every 6 weeks for 2 years) chemotherapy after a modified or radical mastectomy between January 1975 and February 1978. Patients were stratified according to menopausal status and number of positive nodes (one to three, more than three nodes) before randomization. Seventy-eight patients were ineligible, most (56) because they were registered more than 42 days from surgery. Maximum duration of follow-up is 12 years, with a median of 9.8 years. The treatment arms were balanced with respect to age, menopausal status, and number of positive nodes. Among eligible patients, disease-free survival and survival were superior with CMFVP (P = .002, .005, respectively). At 10 years, 48% of patients treated with CMFVP remain alive and disease-free and 56% remain alive, compared with 35% alive and disease-free and 43% alive on the L-PAM arm. Disease-free survival and survival were significantly better with CMFVP compared with L-PAM only in premenopausal patients and patients with four or more positive nodes. Both regimens were well tolerated, although toxicity was more severe and more frequent with CMFVP. We conclude that after 10 years of follow-up, adjuvant combination chemotherapy with CMFVP is superior to single-agent L-PAM in patients with axillary node-positive primary breast cancer. The major advantage is in premenopausal women and in patients with more than three positive axillary nodes.