This randomized study compared the efficacy and safety of fludarabine-mitoxantrone (FM) with mini-CHVP (cyclophosphamide, doxorubicin, vindesine, prednisone) in elderly patients with advanced, low-grade non-Hodgkin's lymphoma.End points were remission rates [overall response (OR) and complete response (CR)], failure-free survival (FFS), survival and toxicity. One hundred and fifty-five patients were randomized, 144 were evaluable for safety and 142 for response. Each treatment arm was given as six monthly cycles, followed by three bimonthly cycles. FM comprised fludarabine (20 mg/m(2) i.v.), days 1-5, plus mitoxantrone (10 mg/m(2) i.v.), day 1. CHVP cycles comprised cyclophosphamide (750 mg/m(2) i.v. infusion), doxorubicin (25 mg/m(2) i.v.) and vindesine (3 mg/m(2) i.v.) on day 1, and prednisone (50 mg/m(2)) on days 1-5.FM therapy resulted in superior remission rates (OR 81% versus 64%, CR 49% versus 17%; P = 0.0004). Median FFS for FM patients was 36 months, compared with 19 months for CHVP patients, and has not yet been reached for early CR patients at 53 months. Treatment arm was the major risk factor influencing survival. Both treatments were well tolerated, with only few infectious complications.FM was more effective than CHVP in achieving OR and CR, and favorably affected the outcome.