Effectiveness and safety of direct oral anticoagulants among patients with non-valvular atrial fibrillation and liver disease: A multinational cohort study

Abstract

Background and aimsThe effects of direct oral anticoagulants (DOACs) in patients with non-valvular atrial fibrillation (NVAF) and liver disease remain poorly understood. Our multinational cohort study assessed the effectiveness and safety of DOACs in this high-risk population. MethodsWe assembled two population-based cohorts in United Kingdom and in Québec of NVAF patients with liver disease initiating DOACs or vitamin K antagonists (VKAs) between 2011 and 2020. Using an as-treated exposure definition, we compared DOACs to VKAs and apixaban to rivaroxaban. After inverse probability of treatment weighting, Cox proportional hazards models estimated site-specific hazard ratios (HRs) and 95 % confidence intervals (CIs) of ischemic stroke and major bleeding. Site-specific estimates were pooled using random-effects models. Analyses were repeated among NVAF patients with cirrhosis. ResultsThere were 11,881 NVAF patients with liver disease (2683 with cirrhosis). Among those, 8815 initiated DOACs (4414 apixaban, 2497 rivaroxaban) and 3696 VKAs. The HRs (95 % CIs) for DOACs compared to VKAs were 1.01 (0.76–1.34) for ischemic stroke and 0.87 (0.77–0.99) for major bleeding. Results were consistent among patients with cirrhosis. The HRs (95 % CIs) for apixaban compared to rivaroxaban were 0.85 (0.60–1.20) for ischemic stroke and 0.80 (0.68–0.95) for major bleeding. This decreased bleeding risk was not observed among patients with cirrhosis (HR, 1.01; 95 % CI 0.72–1.43). ConclusionsAmong NVAF patients with liver disease, DOACs were as effective and slightly safer than VKAs, and apixaban was as effective but safer than rivaroxaban. The safety benefit with apixaban was not present among patients with cirrhosis.