Risks of malignancies related to tofacitinib and biological drugs in rheumatoid arthritis: Systematic review, meta-analysis, and network meta-analysis.

Abstract

To summarize and compare the risks of malignancies accompanying biologic DMARDs (b-DMARDs) and tofacitinib in rheumatoid arthritis (RA) in randomized clinical trials (RCTs) and long-term extension studies (LTEs). Articles in Medline, Embase, Cochrane Library, and the Web of Science dated from 2000 to February 2015. Selection criteria were as follows: (1) focus on RCTs or LTEs in RA; (2) treatment with b-DMARDs or tofacitinib; (3) data on malignancies; and (4) a minimum follow-up of 12 weeks. Data included publication details, study design, risk of bias, number and types of malignancies, and patient characteristics and treatments. Of 113 articles and one updated report that were meta-analyzed, overall malignancies in RCTs showed odds ratio (95% confidence intervals) of 1.01 (0.72, 1.42) for all TNF antagonists, 1.12 (0.33, 3.81) for abatacept, 0.54 (0.20, 1.50) for rituximab, 0.70 (0.20, 2.41) for tocilizumab, and 2.39 (0.50, 11.5) for tofacitinib. Network meta-analysis of overall malignancies showed odds ratio (95% predictive intervals) of 1.68 (0.48-5.92) for infliximab, 0.79 (0.44-1.40) for etanercept, 0.93 (0.43-2.03) for adalimumab, 0.87 (0.28-2.75) for certolizumab, 0.87 (0.39-1.95) for golimumab, 1.04 (0.32-3.32) for abatacept, 0.58 (0.21-1.56) for rituximab, 0.60 (0.16-2.28) for tocilizumab, and 1.15 (0.24-5.47) for tofacitinib. Marginal numerical differences in the incidence rate of solid and hematological malignancies and non-melanoma skin cancers appeared in LTEs. In RCTs, treatment of RA with b-DMARDs or tofacitinib does not increase the risk for malignancies. Generalizability of the differences in the rate of specific malignancies encountered in LTEs requires continuous pharmacovigilance of real-world patients.