ABSTRACT NUMBER: ESOC2026A61 COMPARATIVE EFFECTIVENESS AND SAFETY OF DIRECT ORAL ANTICOAGULANTS IN ATRIAL FIBRILLATION AFTER GASTROINTESTINAL SURGERY Authors: Jiyeon Ha, Wookjin Yang, Mi-Sook Kim, Eung-Joon Lee, Matthew Chung, Boyeon Yang, Hyemin Jang, Jeong-Min Kim, Keun-Hwa Jung, Seung-Hoon Lee Published: 2026 ABSTRACT AbstractBackground and aimsGastrointestinal (GI) surgery can alter GI anatomy and physiology, potentially affecting the effectiveness of direct oral anticoagulants (DOACs). Despite widespread use of DOACs in atrial fibrillation (AF), evidence comparing their effectiveness and safety in patients with prior GI surgery is limited. This study aimed to evaluate the comparative risk–benefit profiles of DOACs in this population.MethodsUsing a nationwide Korean claims database, this observational cohort study included patients with AF prescribed a DOAC (dabigatran, apixaban, edoxaban, or rivaroxaban) after GI surgery between 2014 and 2021. Risks of clinical outcomes including ischemic stroke or systemic embolism, major bleeding, and their composites were evaluated. Inverse probability of treatment weighting was applied to adjust for confounding.ResultsA total of 8,047 patients (mean age, 73.7 years; 34.3% male) were analyzed. Dabigatran was associated with a higher risk of ischemic stroke/systemic embolism than apixaban (hazard ratio [HR], 1.66; 95% CI, 1.01–2.73). Rivaroxaban was associated with increased risk of intracranial hemorrhage (ICH) compared with apixaban (HR, 1.74; 95% CI, 1.02–2.98) and edoxaban (HR, 2.42; 95% CI, 1.25–4.69). The risks of GI bleeding, overall major bleeding, and composite outcomes were broadly comparable across DOACs. Subgroup analyses across key covariates did not reveal any significant effect modifications.ConclusionsIn patients with AF and prior GI surgery, dabigatran was associated with greater thromboembolic risk and rivaroxaban with higher ICH risk compared with other DOACs. DOAC selection in this setting should account for drug-specific risk–benefit profiles in conjunction with individual patient’s thromboembolic and bleeding risks.Conflict of interestJiyeon Ha : Nothing to disclose. Wookjin Yang: Nothing to disclose. Mi-Sook Kim: Nothing to disclose. Eung-Joon Lee: Nothing to disclose. Matthew Chung: Nothing to disclose. Boyeon Yang: Nothing to disclose. Hyemin Jang: Nothing to disclose. Jeong-Min Kim: Nothing to disclose. Keun-Hwa Jung: Nothing to disclose. Seung-Hoon Lee : Nothing to disclose.Figure 1 - belongs to Conclusions