Abstract

Background: Most registration trials in rheumatoid arthritis (RA) include a placebo arm in the setting of an incomplete response to disease-modifying antirheumatic treatment (DMARD-IR). A minimum duration of 6 months is required despite serious methodological and ethical shortcomings.

Objective: To study whether a 3-month placebo period is sufficient to prove efficacy.

Methods: Meta-analysis of placebo- or active control trials of biological agents in DMARD-IR RA, comparing the contrast in ACR response between experimental and control groups at 3 and 6 months.

Results: Twenty trials yielded 15 placebo and 18 active control contrasts (>10 000 patients). At 3 months active treatment showed a highly significant contrast with placebo for ACR20 and ACR50 in every instance. As all groups improved further the mean contrast at 6 months was unchanged. For ACR70 the contrast was clearly greater at 6 months owing to further improvement only in the experimental groups. In active control trials contrasts were smaller, and for ACR50 and ACR70 these decreased somewhat owing to “catch-up” responses in the control groups.

Conclusion: The placebo phase of registration trials for RA can be limited to 3 months. An accompanying viewpoint proposes that patients receiving placebo should then be switched to standard of care, allowing a more valid and comprehensive assessment, including safety.