0970: Comparison of Survival on Treatment and Severe Infection Among New Users of Biosimilar vs. Originator Biologics in Inflammatory Arthritis: Population-based Evidence from a Natural Experiment Due to a Policy Change

Diane Lacaille¹, J. Antonio Avina-Zubieta², Yufei Zheng², Na Lu² and Hui Xie³, ¹Arthritis Research Canada, University of British Columbia, Vancouver, BC, Canada, ²Arthritis Research Canada, Vancouver, BC, Canada, ³Simon Fraser University, Richmond, BC, Canada

Background/Purpose: British Columbia's health policy mandated that all new anti-TNF initiations after June 2017 use biosimilars when available, providing the context for a natural experiment. Our study objective was to compare drug survival and severe infections (as surrogate markers of effectiveness and safety) after the initiation of etanercept and infliximab for inflammatory arthritis in new users of biosimilars versus originators, using historical controls pre- and post-policy change.

Methods: Study Cohort: Using administrative health data, we identified all incident users of a new biologic (i.e., without prior prescriptions over 6 months) with rheumatoid arthritis (RA), psoriasis or psoriatic arthritis (Pso/PsA), or ankylosing spondylitis (AS). The biosimilar cohort includes users starting etanercept or infliximab between 07/01/2017 and 12/31/2019, followed until 12/31/2020 (post period). Historical controls include all incident users of etanercept/infliximab originators between 01/01/2014 and 06/30/2016, followed until 06/30/2017 (pre period). To control for temporal trend, we selected new users of adalimumab (no biosimilar available over the same time periods) as a comparison group.

Outcome: Discontinuation was defined as no prescription renewal for at least 6 months, and severe infection was defined as hospitalization with an infection diagnostic code in any position.

Statistical analyses: People were followed for up to 3 years from anti-TNF initiation until death, moving out-of-province, or the end date of the follow-up, whichever occurred first. Rates were calculated. We applied weighted Cox Proportional Hazard Models to estimate the adjusted hazard ratio (aHR) of discontinuing anti-TNFs and Quasi Poisson Models to estimate the adjusted rate ratio (aRR) of severe infections. We applied propensity overlap weights to control for potential confounders. To control for temporal trend, we employed the difference-in-difference (DID) method, comparing drug survival and severe infections among new users of biosimilar vs. originator etanercept/infliximab with new users of adalimumab post- vs. pre-policy change. The DID computes the difference between the logarithms of the aHRs/aRRs for etanercept/infliximab and for adalimumab, reported as the ratio of the two aHRs/aRRs in Table 1C.

Results: Our sample includes 827 biosimilar etanercept users (RA:576, AS:171, Pso/PsA:80) and 271 infliximab users (RA:150, AS:54,Pso/PsA:67); 1312 etanercept and 230 infliximab originator users; and 2213 adalimumab originator users post- and 1773 pre-policy change. Counts and rates of discontinuation and severe infections are reported in Table 1A. After adjusting for baseline covariates (Table 1B), and after accounting for temporal trends (Table 1C), the likelihood of discontinuation and severe infection was similar for biosimilar vs. originator etanercept and infliximab users.

Conclusion: Real-world population-based data showed that incident users of biosimilar etanercept and infliximab had similar discontinuation and severe infection rates as originators, suggesting comparable effectiveness and safety for inflammatory arthritis.


D. Lacaille: None; J. Avina-Zubieta: None; Y. Zheng: None; N. Lu: None; H. Xie: None.