A mixed treatment comparison of the short-term efficacy of biologic disease modifying anti-rheumatic drugs in established rheumatoid arthritis.

BACKGROUND: The short-term efficacy of biological disease modifying anti-rheumatic drugs (bDMARDs) for the treatment of established moderate to severe rheumatoid arthritis (RA) has been demonstrated by various randomized placebo or active treatment controlled trials. However, there is a lack of direct comparison of these agents. SCOPE: To compare the short-term efficacy of nine bDMARDs - abatacept, adalimumab, anakinra, certolizumab, etanercept, golimumab, infliximab, rituximab and tocilizumab - in patients with established RA. FINDINGS: A systematic review was conducted to obtain all available efficacy data for each included bDMARD. Medline, EMBASE, and Cochrane clinical trials were searched for trials in patients with RA. Twenty-seven trials were retrieved from a systematic literature search and included in the meta-analysis. Mixed treatment comparison (MTC) techniques were used to perform indirect comparisons. Analyses were conducted to estimate the odds ratio of an ACR20, ACR50, and ACR70 response at approximately six months if treated with a bDMARD compared with placebo or methotrexate. Between-drug comparisons were also made. The analyses were performed including recommended doses only (as per the product information). All drugs except anakinra and golimumab demonstrated a statistically significant advantage compared to control treatment for ACR20 responses. The between-drug comparisons revealed a statistically significant advantage for certolizumab compared to most bDMARDs for ACR20, ACR50 and ACR70 response and for etanercept versus adalimumab and anakinra for ACR20 and ACR50 response, as well as a statistically significant advantage for tocilizumab versus anakinra for ACR50 response. CONCLUSION: The analyses, using MTC of efficacy of nine bDMARDs suggest that treatment with anakinra is inferior to other bDMARDs and that etanercept and certolizumab may be more effective than other bDMARDs. There are some limitations of our analyses due to MTC assumptions, variations in trial design and the fact that only ACR outcomes at six months were included.