We designed a randomized, placebo-controlled, multicentre trial involving 51 relapsing-remitting multiple sclerosis patients to determine the clinical efficacy of mitoxantrone treatment over 2 years. Patients were allocated either to the mitoxantrone group (27 patients receiving I.V. infusion of mitoxantrone every month for 1 year at the dosage of 8 mg/m2) or to the placebo group (24 patients, receiving I.V. infusion of saline every month for 1 year) using a centralized randomization system. Disability at entry and at 12-24 months was evaluated by four blinded neurologists trained in the application of the Kurtzke Expanded Disability Scale (EDSS). In addition, the number and clinical characteristics of the exacerbations over the 24 months were recorded by the local investigators. MRI, at 0, 12 and 24 months, was performed with a 0.2 T permanent unit. MRI data were analysed by two blinded neuroradiologists. All patients underwent a clinical evaluation. A statistically significant difference in the mean number of exacerbations was observed between the mitoxantrone group and placebo group both during the 1st and the 2nd year. Although there was no statistically significant benefit in terms of mean EDSS progression over 2 years, the proportion of patients with confirmed progression of the disease, as measured by a one point increase on the EDSS scale, was significantly reduced at the 2nd year evaluation in the mitoxantrone group. Forty-two (23 mitoxantrone, 19 placebo) patients underwent all MRI examinations during the 24-month period. We observed a trend towards a reduction in the number of new lesions on T2-weighted images in the mitoxantrone group. Our study suggests that mitoxantrone might be effective in reducing disease activity, both by decreasing the mean number of exacerbations and by slowing the clinical progression sustained by most patients after 1 year from the end of treatment.