Effectiveness and Safety of Oral Anticoagulants among patients with Atrial Fibrillation and Advanced Kidney Disease

Background: The net clinical benefit of oral anticoagulants (OAC) among patients with atrial fibrillation (AF) and advanced kidney disease is unclear.

Objectives: To compare the effects of use vs. non-use of specific OACs among patients with AF and chronic kidney diseases (CKD) stage 4-5, stratified by dialysis status.

Methods: We used US Medicare claims data (aged ≥65 years, 2013-2020) and Optum Clinformatics® Data Mart Database (aged ≥18 years, 2013-2023) to identify patients with non-valvular AF, a CHA2DS2-VASc score ≥2 (male) or ≥3 (female), and CKD stage 4-5 and those with end-stage kidney disease (ESKD) on dialysis. We established three new user vs. non-user comparative cohorts for apixaban, rivaroxaban, warfarin, respectively. For each new user (no preceding use in the preceding year), we identify non-users using incidence density sampling method. Time zero was the dispensing date of OAC for users and the sampling date for the non-users. We conducted stratified analysis by 1) ESKD on dialysis; 2) CKD stage 4 or 5 not on dialysis. The outcome was a composite of hospitalization with ischemic stroke, major bleeding, and all-cause mortality. We used 1:1 propensity score (PS) matching to balance 88 baseline covariates. We used Cox proportional hazard models to estimate hazard ratios (HR) and 95% confidence intervals (CI).

Results: Pooling from the two databases for the dialysis cohort, our dialysis cohort included 6,111 PS-matched pairs for apixaban, 222 PS-matched pairs for rivaroxaban, and 6,111 PS-matched pairs for warfarin. Compared to non-users of OAC, we found a reduced risk of the composite outcome associated with apixaban (HR [95% CI], 0.68 [0.47- 0.97]) but not rivaroxaban (HR, 0.90 [0.62, 1.31]) or warfarin users (HR 0.77 [0.54, 1.10]). The CKD stage 4 and 5 non-dialysis cohort included 6,803 PS-matched pairs for apixaban, 1,105 pairs for rivaroxaban, and 4,158 pairs for warfarin. Compared to non-users of OAC, we found OAC use was associated with a reduced risk of the composite outcome for apixaban users (HR [95% CI], 0.66 [0.49, 0.89]), and warfarin users (HR 0.84 [0.72, 0.98]), but not rivaroxaban (HR, 0.86 [0.73, 1.03]).

Conclusions: Based on two US national cohorts, we found that use of apixaban, but not warfarin or rivaroxaban, was associated with a net clinical benefit among patients with AF and ESKD on dialysis. Among those with CKD stages 4 to 5 not on dialysis, use of OAC was associated with a net clinical benefit for apixaban and warfarin, but not rivaroxaban.