A randomized trial to compare adjuvant treatment with an alternating regimen with conventional chemotherapy was performed. A total of 589 node-positive patients were included and stratified according to number of positive nodes (N1-3 and N > 4) and menopausal status. Premenopausal N1-3 patients were randomized to cyclophosphamide, methotrexate and fluorouracil (CMF) or CMF/4'-epirubicin, cyclophosphamide (EC), post-menopausal N1-3 patients to fluorouracil, 4 epirubicin, cyclophosphamide (FEC) or CMF/EC and pre- and post-menopausal patients with N > or = 4 to fluorouracil, 4' epirubicin, cyclophosphamide, methotrexate, prednisone (FECMP) or CMF/EC. In premenopausal patients, CMF was superior to CMF/EC in terms of disease-free survival (DFS) (65% vs 45%, P = 0.0149) and survival (72.3% vs 50.2%, P = 0.0220) whereas, for N > or = 4 patients, differences between FECMP and CMF/EC did not achieve statistical significance (DFS 35% vs 26.2%; survival 50% vs 38.1%, P = NS). For post-menopausal patients, FEC was superior to CMF/EC in DFS (58.6% vs 36.8%, P = 0.0215) and survival (66.2% vs 46%, P = 0.0155). In post-menopausal patients with N > 4, differences favouring CMF/EC were significant in DFS (40.4% vs 22%, P = 0.0371) but not in survival (47.4% vs 32.2%, P = 0.1185). Alternating regimens did not offer better results in premenopausal and post-menopausal N1-3 patients. Results regarding post-menopausal N > 4 women require further confirmation.