Apixaban Dosing Patterns Versus Warfarin in Patients With Nonvalvular Atrial Fibrillation Receiving Dialysis: A Retrospective Cohort Study

Background & Objectives: Comparison of clinical outcomes across anticoagulation regimens using different apixaban dosing or warfarin is not well-defined in patients with nonvalvular atrial fibrillation (AF) who are receiving dialysis. This study compared these outcomes in a US national cohort of patients with kidney failure receiving maintenance dialysis. Study Design: Retrospective cohort study. Setting & Participants: Patients receiving dialysis represented in the US Renal Data System database 2013-2018 who had AF and were treated with apixaban or warfarin. Exposure: First prescribed treatment with apixaban dosed according to the label, apixaban dosed below the label, or warfarin. Outcome: Ischemic stroke/systemic embolism, major bleeding, and all-cause mortality. Analytical Approach: Cox proportional hazards models with inverse probability of treatment weighting. Analyses simulating an intention-to-treat (ITT) approach as well as those incorporating censoring at drug switch or discontinuation (CAS) were also implemented. Inverse probability of censoring weighting was used to account for possible informative censoring. Results: Among 17,156 individuals, there was no difference in risk of stroke/systemic embolism among the label-concordant apixaban, below-label apixaban, and warfarin treatment groups. Both label-concordant (HR, 0.67 [95% CI, 0.55-0.81]) and below-label (HR, 0.68 [95% CI, 0.55-0.84]) apixaban dosing were associated with a lower risk of major bleeding compared with warfarin in ITT analyses. Compared with label-concordant apixaban, below-label apixaban was not associated with a lower bleeding risk (HR, 1.02 [95% CI, 0.78-1.34]). In the ITT analysis of mortality, label-concordant apixaban dosing was associated with a lower risk versus warfarin (HR, 0.85 [95% CI, 0.78-0.92]) while there was no significant difference in mortality between below-label dosing of apixaban and warfarin (HR, 0.97 [95% CI, 0.89-1.05]). Overall, results were similar for the CAS analyses. Limitations: Study limited to US Medicare beneficiaries; reliance on administrative claims to ascertain outcomes of AF, stroke, and bleeding; likely residual confounding. Conclusions: Among patients with nonvalvular AF undergoing dialysis, warfarin is associated with an increased risk of bleeding compared with apixaban. The risk of bleeding with below-label apixaban was not detectably less than with label-concordant dosing. Label-concordant apixaban dosing is associated with a mortality benefit compared to warfarin. Label-concordant dosing, rather than reduced-label dosing, may offer the most favorable benefit-risk trade-off for dialysis patients with nonvalvular AF.