(143) A Population-Based Assessment of the Use of Oral Anticoagulants and the Risk of Major Bleeding Among Patients with Non-Valvular Atrial Fibrillation Treated With Antiarrhythmics

Background: The comparative safety of different oral anticoagulants in patients with non-valvular atrial fibrillation (NVAF) additionally treated with antiarrhythmics (AA) is poorly understood.

Objectives: To compare the risk of major bleeding associated with direct oral anticoagulants (DOACs) versus vitamin K antagonist (VKAs) among patients with NVAF who are using AA.

Methods: Using the United Kingdom´s Clinical Practice Research Datalink linked to hospitalization data, we assembled a study cohort of patients with NVAF who were using AA and initiated treatment with DOACs or VKAs between 2011 and 2020. Study cohort entry was the date of anticoagulant initiation. Using an as treated exposure definition, we assessed the risk of major bleeding, defined as hospitalization with bleeding, associated with concomitant use of DOACs and AA compared to concomitant use of VKAs and AA. Secondary outcomes were intracranial hemorrhage (ICH), gastrointestinal bleeding (GIB), and other major bleeding. We used inverse-probability-of-treatment-weighting (IPTW) based on propensity scores (PS) with >20 potential confounders. PS were calculated separately for prevalent and new AA users, with estimates subsequently pooled across user types. Cox models estimated weighted hazard ratios (HRs) and 95% confidence intervals (CIs) for each outcome. In secondary analyses, we stratified by demographics, baseline bleeding risk, type of AA user, and the individual DOACs apixaban and rivaroxaban, re-estimating the IPTW for each stratum. In sensitivity analyses, we assessed the impact of exposure and outcome misclassification, informative censoring, and residual confounding.

Results: Our study cohort included 14,687 patients initiating DOACs (n=7,896) or VKAs (n=6,791) while prescribed an AA. After IPTW, baseline characteristics were well balanced between groups. Concomitant use of DOACs and AA was not associated with the risk of major bleeding compared to concomitant use of VKAs and AA (crude incidence rates: 3.8 vs 3.6 per 100 person-years; weighted HR, 0.90; 95% CI, 0.75-1.08). There was a numerically decreased risk of ICH (weighted HR, 0.64; 95% CI, 0.25-1.64) but no association with the risk of GIB (weighted HR, 0.96; 95% CI, 0.71-1.29) or other major bleeding (weighted HR, 0.89; 95% CI, 0.71-1.12). Secondary analyses showed no effect measure modification, except when stratifying by type of AA use (prevalent AA use: weighted HR, 0.78: 95% CI, 0.63-0.97; new AA use: weighted HR, 1.24; 95% CI, 0.89-1.75 / p for interaction: 0.03). Sensitivity analyses were consistent with the primary analysis (HRs ranging from 0.79 to 1.07).

Conclusions: Our population-based study showed that compared to VKAs, DOACs were not associated with the risk of major bleeding among patients with NVAF treated with AA.