Defining the optimal biological monotherapy in rheumatoid arthritis: A systematic review and meta-analysis of randomised trials

Objectives To summarize and compare the benefits and harms of biological agents used as monotherapy for rheumatoid arthritis (RA) in order to inform decisions for patients who are intolerant to conventional DMARD therapy. Methods We searched MEDLINE, EMBASE, CENTRAL, and other sources for randomised trials that compared biological monotherapy with methotrexate, placebo, or other biological monotherapies. Primary outcomes were ACR50 and the number of patients who discontinued due to adverse events. Our network meta-analysis was based on mixed-effects logistic regression, including both direct and indirect comparisons of the treatment effects, while preserving the randomised comparisons within each trial. PROSPERO identifier: CRD42012002800. Results The analysis comprises 28 trials (8602 patients), including all nine biological agents approved for RA. Eight trials included “DMARD-naïve”, and 20 “DMARD-Inadequate responder” (DMARD-IR) patients. All agents except anakinra and infliximab were superior (p < 0.05) to placebo (i.e., no DMARD treatment) with regard to ACR50. Etanercept and rituximab were superior to anakinra (p = 0.018 and p = 0.049, respectively). Tocilizumab was superior to adalimumab (p = 0.0082), anakinra (p = 0.0083), certolizumab (p = 0.037), and golimumab (p = 0.049). No differences among etanercept, tocilizumab, and rituximab were found (p > 0.52). However, because rituximab was evaluated in just 40 patients, our confidence in the estimates is limited. When including only DMARD-IR trials, the same statistical pattern emerged; in addition etanercept and tocilizumab were superior to abatacept. At recommended doses, both etanercept and tocilizumab were superior to adalimumab and certolizumab. No statistically significant differences among biological agents were found with respect to discontinuation due to adverse events (p > 0.068). Conclusions Evidence from randomised trials suggests that most biological agents are effective as monotherapy. Although our confidence in the estimates is limited, etanercept or tocilizumab may be the optimal choice for most patients who need treatment with biological monotherapy. However, given our limited confidence in the estimates including possibility of bias, it is appropriate to strongly weight patients׳ preferences and values in the final treatment choice.