We compared a relatively short regimen of monochemotherapy with epirubicin versus polychemotherapy with CMF (cyclophosphamide, methotrexate, 5-fluorouracil) as adjuvant treatment for stage I and II breast cancer patients. 348 patients with oestrogen receptor negative (ER-) node negative and ER- or ER+ node-positive with <10 nodes were accrued. CMF was given intravenously (i.v.) on days 1 and 8, every 4 weeks, for six courses; epirubicin was given weekly for 4 months. Postmenopausal patients received tamoxifen for 3 years. The primary endpoints were overall survival (OS), relapse-free survival (RFS) and event-free survival (EFS). Outcome evaluation was performed both in eligible patients and in all randomised patients according to the intention-to-treat principle. 8 randomised patients were considered ineligible. At a median follow-up of 8 years, there was no difference in OS (Hazard Ratio (HR)=1.11, 95% Confidence Interval (CI): 0.77-1.61, P=0.58), EFS (HR=1.14, 95% CI: 0.78-1.64, P=0.48), and RFS (HR=1.14, 95% CI: 0.8-1.64, P=0.48) between the two arms for all of the patients. At 8 years, the RFS percentages (+/-Standard Error (S.E.)) were 65.4% (+/-4%) in the CMF arm and 62.7% (+/-4%) in the epirubicin arm; for EFS these were 64.2% (+/-4%) for CMF and 60.8% (+/-4%) for epirubicin, respectively. A significant difference in RFS (P=0.015) was observed in patients with 4-9 positive nodes in favour of the CMF arm. Toxicity in the two arms was superimposable except for more frequent grade 3 alopecia in the epirubicin-treated patients (P=0.001). Overall, at a median follow-up of 8 years, there were no differences between the two arms in terms of OS, EFS and RFS.