BACKGROUND B lymphocytes are implicated in the pathogenesis of multiple sclerosis. We aimed to assess efficacy and safety of two dose regimens of the humanised anti-CD20 monoclonal antibody ocrelizumab in patients with relapsing-remitting multiple sclerosis. METHODS We did a multicentre, randomised, parallel, double-blind, placebo-controlled study involving 79 centres in 20 countries. Patients aged 18-55 years with relapsing-remitting multiple sclerosis were randomly assigned (1:1:1:1) via an interactive voice response system to receive either placebo, low-dose (600 mg) or high-dose (2000 mg) ocrelizumab in two doses on days 1 and 15, or intramuscular interferon beta-1a (30 μg) once a week. The randomisation list was not disclosed to the study centres, monitors, project statisticians or to the project team at Roche. All groups were double blinded to group assignment, except the interferon beta-1a group who were rater masked. At week 24, patients in the initial placebo, 600 mg ocrelizumab, and interferon beta-1a groups received ocrelizumab 600 mg; the 2000 mg group received 1000 mg. Our primary endpoint was the total number of gadolinium-enhancing lesions (GEL) and T1-weighted MRI at weeks 12, 16, 20, and 24. Analyses were done on an intention-to-treat basis. This trial is registered with ClinicalTrials.gov, number NCT00676715. RESULTS 218 (99%) of the 220 randomised patients received at least one dose of ocrelizumab, 204 (93%) completed 24 weeks of the study and 196 (89%) completed 48 weeks. In the intention-to-treat population of 218 patients, at week 24, the number of gadolinium-enhancing lesions was 89% (95% CI 68-97; p<0·0001) lower in the 600 mg ocrelizumab group than in the placebo group, and 96% (89-99; p<0·0001) lower in the 2000 mg group. In exploratory analyses, both 600 mg and 2000 mg ocrelizumab groups were better than interferon beta-1a for GEL reduction. We noted serious adverse events in two of 54 (4%; 95% CI 3·0-4·4) patients in the placebo group, one of 55 (2%; 1·3-2·3) in the 600 mg ocrelizumab group, three of 55 (5%; 4·6-6·3) in the 2000 mg group, and two of 54 (4%; 3·0-4·4) in the interferon beta-1a group. CONCLUSIONS The similarly pronounced effects of B-cell depletion with both ocrelizumab doses on MRI and relapse-related outcomes support a role for B-cells in disease pathogenesis and warrant further assessment in large, long-term trials. BACKGROUND F Hoffmann-La Roche Ltd, Biogen Idec Inc.