Prevalent-New User Design for Biosimilar Safety Using Saudi Real-World Data

Incidents of infections associated with biosimilar biological disease-modifying antirheumatic drugs (bDMARDs) have been communicated to the Saudi Food and Drug Authority (SFDA). This study aimed to evaluate the infection risk related to biosimilar bDMARDs and assess the applicability of the prevalent new-user design within a Saudi cohort.

Using the Saudi Real-World Evidence Network (SRWEN), patients aged ≥ 14 years who initiated or switched to biosimilar versus reference bDMARDs between 2016 and 2025 were identified. A prevalent new-user design that accounted for prior exposure to reference bDMARDs was implemented. Time-conditional propensity score (TCPS) matching (1:1 and 1:3) balanced baseline covariates. The study outcome was the risk of any infection, identified via ICD-10 codes in the main analysis, in addition to laboratory results and antimicrobial prescriptions in the sensitivity analyses. Hazard ratios (HRs) were estimated using Cox proportional hazards models with robust sandwich estimators.

Among 5721 patients contributing 59 856 treatment episodes, 2178 were biosimilar initiators and 2909 were switchers (from the reference product). Adalimumab accounted for 69.5% of use. Over a median 1.6-year follow-up, infection incidence rates were comparable between reference and biosimilar groups (456.7 vs. 422.0 per 1000 person-years). Adjusted HRs showed no significant difference in infection risk (HR 0.89; 95% CI 0.57-1.39, p-value: 0.61), with consistent findings across sensitivity and disease-matched analyses where the adjusted HR was 0.87 (95% CI 0.51-1.50, p-value: 0.62).

In this real-world Saudi cohort, no statistically significant difference in infection risk was observed between biosimilar and reference bDMARD. Larger studies and further evaluation of the prevalent new-user design are needed to enhance post-marketing biosimilar surveillance in the region.