Safety of synthetic and biological DMARDs: a systematic literature review informing the 2025 update of the EULAR recommendations for the management of rheumatoid arthritis

OBJECTIVES: This study aims to perform a systematic literature review (SLR) concerning the safety of synthetic and biological disease-modifying antirheumatic drugs (DMARDs) for the 2025 update of European Alliance of Associations for Rheumatology (EULAR) recommendations for the management of rheumatoid arthritis (RA). METHODS: Medline, Embase, Cochrane CENTRAL, and Web of Science were searched for observational and randomised controlled trials with a primary endpoint of DMARD safety on the conventional synthetic (cs-), biological (b-), and targeted synthetic-DMARDs, as well as glucocorticoids, published between January 14, 2022, and January 22, 2025. Separate searches on DMARD monitoring were conducted from database inception to January 22, 2025. A comparator group was required for inclusion. All safety outcomes were included. RESULTS: A total of 3837 articles were identified, with 321 selected for full-text review; 71 articles were included. Across the evidence base, infections were the most frequently assessed outcome: 13 studies examined serious or hospitalised infections, usually as composite endpoints of bacterial, opportunistic, and herpes zoster infection, while 1 addressed nonserious infections. Serious infections were more common with bDMARDs than csDMARDs. Janus kinase inhibitors (JAKis) showed a higher herpes zoster risk than bDMARDs. Tuberculosis risk was not increased with JAKis compared with bDMARDs, but was higher with infliximab and adalimumab compared with etanercept. Fifteen studies evaluated malignancy, split evenly between analyses of any malignancy and those excluding nonmelanoma skin cancer (NMSC); 2 focused on melanoma and 2 on NMSC. Increased NMSC was noted in patients with RA using DMARDs compared with the general population, with no link to a specific DMARD. Cardiovascular and thromboembolic events were reported in 20 studies. No consistent evidence of increased major adverse cardiovascular events risk with JAKis compared with bDMARDs was identified. Venous thromboembolism risk appeared elevated with JAKis compared with bDMARDs, driven mainly by pulmonary embolism. Fourteen studies reported retention and adverse event-related withdrawals, and 8 assessed other specific adverse events. Gastrointestinal perforation and demyelinating disease were each reported in 3 studies. No eligible articles were identified in searches on DMARD monitoring. CONCLUSIONS: There has been a notable increase in studies evaluating safety outcomes, with the majority of these being observational studies focusing primarily on malignancy, thromboembolic, and cardiovascular events, with most studies pertaining to JAKi safety. A substantial proportion of studies in this SLR relied on claims databases to evaluate safety outcomes, a practice that carries important methodological limitations for safety research. Surprisingly, not many studies looked into glucocorticoid safety outcomes over the past 3 years. This SLR, along with the SLR on efficacy of DMARDs, informed the 2025 update of the EULAR recommendations for management of RA with synthetic and biological DMARDs.