Mixed Treatment Comparison of the Treatment Discontinuations of Biologic Disease-Modifying Antirheumatic Drugs in Adults with Rheumatoid Arthritis

BACKGROUND: Introduction of biologic disease-modifying antirheumatic drugs (DMARDs) has considerably changed treatment options for rheumatoid arthritis (RA) over the past decade. Very little information is available on comparative discontinuation rates of the biologics. OBJECTIVE: To compare treatment discontinuations for 9 biologic DMARDs in adults with RA. METHODS: We searched electronic databases through May 2012 to retrieve randomized controlled trials (RCTs) of patients with RA that compared biologic DMARDs with placebo or another biologic DMARD. The primary outcome was treatment discontinuation during the blinded phase of the trials, measured as overall withdrawals, withdrawals resulting from lack of efficacy, and withdrawals resulting from adverse events. Random-effects meta-analysis estimated the effect size for individual agents, and adjusted indirect comparisons were made between biologics using mixed treatment comparisons (MTC) meta-analysis. RESULTS: Forty-four trials were included in the analysis. In comparison with placebo, biologics were less likely to be withdrawn because of lack of efficacy (OR 0.22, 95% CI 0.17 to 0.27) and more likely to be withdrawn because of an adverse event (OR 1.41, 95% CI 1.16 to 1.70). Based on the MTC, certolizumab had the most favorable overall withdrawal profile, followed by etanercept and rituximab. Certolizumab had lower relative withdrawal rates resulting from lack of efficacy than adalimumab, anakinra, and infliximab. Anakinra had higher relative withdrawal rates resulting from lack of efficacy than most other biologics. Certolizumab and infliximab had more, while etanercept had fewer, withdrawals because of adverse events than most other drugs. CONCLUSIONS: Based on MTC using data from RCTs, differences in discontinuation rates were observed, generally favoring certolizumab, etanercept, and rituximab over other biologic DMARDs. These potential differences need to be further explored in head-to-head trials or well-conducted observational studies