Risk of admission to hospital for serious infection after initiating tofacitinib versus biologic DMARDs in patients with rheumatoid arthritis: a multidatabase cohort study.

ABSTRACT

BACKGROUND: The risk of serious infection is a major concern when prescribing a disease-modifying antirheumatic drug (DMARD) for rheumatoid arthritis. Little evidence exists with regard to serious infection risk associated with tofacitinib, a targeted synthetic (ts)DMARD, compared with biologic DMARDs (bDMARDs) among patients with rheumatoid arthritis. We compared the risk of serious infection in patients with rheumatoid arthritis initiating tofacitinib versus one of the seven bDMARDs. METHODS: In this multidatabase cohort study, we identified eight mutually exclusive groups of patients with rheumatoid arthritis initiating tofacitinib or one of the seven bDMARDs using US public (Medicare 20012-15) and private (Optum Clinformatics 2012-18 and IBM MarketScan 2012-17) health insurance programmes. Eligible patients were aged 18 years or older, and had one inpatient visit or two or more outpatient visits (7-365 days apart) for rheumatoid arthritis using International Classification of Diseases 9th and 10th revision codes. The primary outcome was a composite endpoint of admission to hospital for serious infection including bacterial, viral, or opportunistic infection based on the inpatient principal diagnosis code. Secondary outcomes were individual types of serious infections and herpes zoster. We adjusted for more than 60 potential confounders through propensity score-based inverse probability treatment weighting in each database. Database-specific adjusted hazard ratios (aHRs) were combined using a fixed-effects meta-analysis. FINDINGS: We identified 130 718 patients with rheumatoid arthritis across the three databases. During 100 790 person-years of follow-up, 3140 serious infections occurred (crude incidence rate per 100 person-years 3·12, 95% CI 3·01-3·23). The aHR for serious infection associated with tofacitinib was 1·41 (95% CI 1·15-1·73) versus etanercept, 1·20 (0·97-1·49) versus abatacept, 1·23 (0·94-1·62) versus golimumab, and 1·17 (0·89-1·53) versus tocilizumab. The serious infection risk with tofacitinib was similar to adalimumab (1·06, 0·87-1·30) and certolizumab (1·02, 0·80-1·29), and was lower than infliximab (0·81, 0·65-1·00). Tofacitinib was associated with a 2-fold higher risk of herpes zoster versus all bDMARDs. INTERPRETATION: This study found potential differences between tofacitinib and several bDMARDs in the risk of admission to hospital for serious infection, as well as herpes zoster, in patients with rheumatoid arthritis. These results contribute to the evolving understanding of the overall risk-benefit profile of tofacitinib. FUNDING: Division of Pharmacoepidemiology and Pharmacoeconomics, Brigham and Women's Hospital, Harvard Medical School, Harvard University, MA, USA.