Evaluating the Real-World Safety of Icosapent Ethyl Versus Omega-3 Polyunsaturated Fatty Acid in Nationwide US Veterans Cohort: Examining Atrial Fibrillation and Bleeding Endpoints.

Purpose: The REDUCE-IT randomized trial demonstrated a cardiovascular benefit of icosapent ethyl (IPE) but also raised potential safety signals for atrial fibrillation (AF) and serious bleeding. We aimed to evaluate the real-world safety of IPE versus mixed omega-3 polyunsaturated fatty acid (OM-3) formulations.

Methods: This retrospective active comparator new-user cohort study compared rates of new-onset AF and major bleeding (MB) among adult new users of IPE versus OM-3 in 2020-2024 US Veterans Affairs data. Daily drug exposure was determined via prescription dispensing dates. AF and MB outcomes were identified via validated algorithms based on the International Statistical Classification of Diseases and Related Health Problems, 10th revision, clinical modification. Confounding was accounted for via nearest-neighbor pairwise propensity score (PS) matching. The PS, constructed via logistic regression, was informed by expert-identified variables meeting the disjunctive cause criterion. Cox regression was used to estimate adjusted hazard ratios (aHRs), interpretable as average treatment effects for the treated.

Results: Cohorts for analyses of AF and MB endpoints included 1927 and 2015 people, respectively, in each of the IPE and OM-3 exposure groups. The median age was 70 years, and the groups exhibited a predominance of white (80%) males (93%). The median follow-up time was 1.29 years per person. Baseline covariates were well balanced by treatment arm after PS matching. Incidence rates for AF were 7.29 versus 7.48 per 100 person-years among new users of IPE versus OM-3. The aHR for AF was 1.15 (95% confidence interval 0.82-1.63). Incidence rates for MB were 3.27 versus 3.35 per 100 person-years among new users of IPE versus OM-3. The aHR for MB was 1.22 (95% confidence interval 0.87-3.02).

Conclusions: Our measures of association were consistent with the null, but we were unable to rule out harms from IPE (vs. OM-3) more modest than a 63% increased rate of AF and threefold increased rate of MB.