For comparative risk of DM among patients with RA, the data from 2 nationwide data sources in the United States, stated that abatacept is linked with a lower risk of incident DM, than with TNF‐inhibitors, in patients with RA. Comparison of abatacept with other agents was unconvincing due to reduced event counts presented for valid treatment‐effect estimation.

This observational cohort study was performed with an objective to compare the risk of diabetes mellitus (DM) in patients suffering from rheumatoid arthritis (RA) who were treated with biologic or targeted synthetic disease‐modifying antirheumatic drugs.

The data was obtained from a US commercial (Truven MarketScan, 2005‐2016) and public insurance (Medicare, 2010‐2014) claims databases. The RA patients who did not have DM were designated into 1 out of 8 exposure groups (namely, adalimumab, abatacept, certolizumab, etanercept, golimumab, infliximab, tocilizumab, or tofacitinib). The combination of a diagnosis code and beginning of a hypo-glycemic treatment helped to examine the outcome of incident DM. Stabilized inverse probability–weighted Cox proportional hazards model accounted for 56 confounding variables and estimated hazard ratios (HRs) and 95% confidence intervals (CIs). All the analyses were managed separately in 2 databases, and estimates were combined via the inverse variance meta‐analysis.

Amongst the 50,505 RA patients from Truven and 17,251 RA patients from Medicare, the rate of incidence (95% CI) for DM were 6.8 (6.1‐7.6) and 6.6 (5.4‐7.9) per 1000 person‐years. The pooled HRs (95% CI) specified a considerably higher risk of DM among adalimumab (2.00 [1.11‐3.03]) and infliximab initiators (2.34 [1.38‐3.98]) than abatacept. Pooled HR for the etanercept against abatacept was higher but not statistically significant (1.65 [0.91‐2.98]). Due to small sample size, the effect estimates for certolizumab, golimumab, tocilizumab, and tofacitinib were inaccurate than with abatacept.

Using abatacept was linked with a lower risk of incident DM in RA patients than with adalimumab or infliximab initation.