Background: The use of biologic agents has revolutionized the treatment of rheumatoid arthritis (RA). However, there is much uncertainty about whether any agent may be preferable.
Purpose: The aim of the study was to evaluate the comparative efficacy of biologic agents with a disease-modifying antirheumatic drug (DMARD) in RA patients without prior exposure to a DMARD, that is, DMARD naive.
Methods: MEDLINE, Cochrane, and Clinicaltrials.gov were searched from 1990 to August 2013 for randomized controlled trials comparing biologic agents in conjunction with a DMARD and DMARDs alone in DMARD (methotrexate [MTX])-naive RA patients. Information on patient characteristics, disease duration, and the American College of Rheumatology (ACR) 20/50/70/90 response rates after 52 weeks was extracted.
Results: Six randomized controlled trials totaling 9 study arms fulfilled the inclusion criteria. Data were analyzed by direct and indirect pairwise comparisons of 2 drugs against a common comparator. In the direct comparison, all 6 biologic therapies were associated with significantly higher likelihood of achieving an ACR20 compared with MTX alone (mean ORs, 1.43-2.99). For ACR50 and ACR70, all biologic agents except golimumab showed statistically significant mean ORs of 1.31 to 2.52 (ACR20) and 1.79 to 2.59 (ACR50). At ACR90, abatacept 10 mg/kg, adalimumab 40 mg, and rituximab 500 and 1000 mg were significantly better compared with MTX (mean ORs 1.92-2.89). The indirect comparison for ACR20 showed etanercept 50 mg significantly favored against adalimumab 40 mg (OR, 1.05-3.34), golimumab 50 mg (OR, 1.16-4.07), infliximab 3 mg/kg (OR, 1.21-3.61), and infliximab 6 mg/kg (OR, 1.02-3.06). At ACR50, etanercept 50 mg and rituximab 1000 mg showed significantly higher ORs compared with golimumab 100 mg at ORs 1.06 to 3.42 and ORs 1.07 to 3.42, respectively. No significant differences were observed in the biologic agents for indirect pairwise comparisons at ACR70 and ACR90.Lack of head-to-head clinical trial data directly comparing biologic agents makes indirect meta-analysis the only substitute. Safety and cost of these agents were not evaluated. Only a small number of trials could be evaluated because of the strict inclusion criteria required for an indirect meta-analysis. Unmeasured confounders could contribute to trial heterogeneity. The data on golimumab were difficult to reconcile with the other trials because of methodological differences.
Conclusions: Overall, biological agents in conjunction with a DMARD performed similarly in the settings evaluated. However, there were some statistically significant differences. Etanercept 50 mg appears superior to adalimumab 40 mg, golimumab 50 mg, and infliximab 3 and 6 mg/kg at ACR20. Rituximab 1000 mg and etanercept 50 mg appeared superior to golimumab 100 mg at ACR50 in DMARD-naive patients. No agent was superior to all others at each ACR level.