OBJECTIVE To assess safety and efficacy of a murine anti-CD4 monoclonal antibody (Mab) in a population of patients with rheumatoid arthritis (RA) compared to treatment with placebo. METHODS Fifty-eight patients with defined RA were included in this placebo controlled, randomized, double blind, multicenter study. Of the 48 women and 10 men (mean age 54.5 years), 25 were functional class II and 31 were class III, with 9 years' disease duration; the mean of previous disease modifying antirheumatic drugs was 4; 49 were taking steroids (mean dosage 11 mg/day of prednisone). Eighty percent were rheumatoid factor positive. All were in an active state of the disease with: pain > 4 (mean at inclusion 6.6), tender joints > 4 (mean 12), swollen joint count > 3 (mean 9), morning stiffness > 45 min (mean 185), erythrocyte sedimentation rate > 30 mm (mean 59) or C-reactive protein (CRP) > 30 mg/l (mean 63). Treatment was randomized between murine anti-CD4 Mab (B-F5, Diaclone, 20 mg/day) or placebo intravenously for 10 consecutive days. Efficacy was assessed with a composite index (Paulus), with evaluation of number of patients with 20 or 50% improvement in each group. Changes in measures of single clinical or biological variables were also evaluated. RESULTS The 2 groups were comparable at inclusion. Treatment was well tolerated. Mild side effects (chills, fever, rash) were seen in both groups. Percentage of patients with global 20 or 50% response did not differ between placebo and Mab groups at Day 10 or at Day 30. Evaluation of single variables showed reduced CRP, swollen joint count, and Ritchie index in some B-F5 patients at Day 10, although in the B-F5 group as a whole only CRP was significant. CONCLUSIONS No significant improvement in RA after murine anti-CD4 Mab was observed.