Apixaban for End-Stage Kidney Disease

Abstract

T he use of vitamin K antagonists (VKAs) for stroke prevention in atrial fibril-lation (AF) among patients dependent on dialysis remains controversial. In this population, VKAs are associated with higher rates of major bleeding, particularly intracranial hemorrhage, and difficulty maintaining the international normalized ratio in the therapeutic range. A recent meta-analysis of 12 observational studies totaling 17 380 patients on hemodialysis (4010 of whom received VKAs) reported a nonsignificant reduction in the risk of ischemic stroke associated with VKAs (hazard ratio [HR], 0.74; 95% CI, 0.51-1.06), an increase in total bleeding risk (HR, 1.21; 95% CI, 1.03-1.43), and a numeric near doubling of the risk for hemorrhagic stroke (4 studies; HR, 1.93; 95% CI, 0.93-3.98). 1 In addition to the lack of convincing efficacy and more concerning harm, warfarin has been implicated in the accelerated decline of renal function either through parenchymal microbleeds in the setting of excessive anticoagulation or vascular calcifications that result from its inhibition of the vitamin K cycle and γ-carboxylation of matrix Gla protein. 2,3 Alternatives to VKAs are needed for this medically complex group of patients. Randomized trials to date of apixaban versus warfarin for AF excluded patients with severe and end-stage kidney disease, calculated creatinine clearance <25 mL per minute. 4,5 The US Food and Drug Administration cautiously extended apixaban use to patients with end-stage kidney disease on hemodialysis based on limited pharmacokinetic data. 6-8 The recommended dose of apixaban was 5 mg twice daily with a reduction in dose to 2.5 mg twice daily for either ≥80 years of age or body weight ≤60 kg. The expectation or hope was that the favorable results of the ARISTOTLE trial (Apixaban for Reduction in Stroke and Other Thromboembolic Events in Atrial Fibrillation) would be translated to patients with AF dependent on dialysis, specifically less major hemorrhage, less intracerebral hemorrhage, and lower mortality compared to VKAs. Rates of major bleeding among patients enrolled in the ARISTOTLE trial were 2.13% per year in the apixa-ban group compared with 3.09% per year in the warfarin group (HR, 0.69; 95% CI, 0.60-0.80). In this issue of Circulation, Siontis et al 9 report their findings from a retrospective study of the use of apixaban versus warfarin among Medicare patients with AF identified in the US Renal Data System. Given the nonrandomized design, cohorts were matched using a prognostic score with patient factors derived from Medicare claims data. Drug exposure was determined using serial prescription refill data from Medicare Part D. Prespecified outcomes were determined from International