Balancing Anticoagulation Decisions in Patients on Dialysis with Atrial Fibrillation

For over 20% of 678,000 patients with ESRD in the United States, the diagnosis of atrial fibrillation (AF) presents a clinical conundrum.1 Patients receiving hemodialysis (HD) not only have a higher prevalence of AF, but also are at greater risk of heart failure, hypertension, and diabetes2—all factors that enhance stroke risk in the setting of AF.3 Additionally, this same population is at higher risk of bleeding compared with the general population.1,2,4 Many patients with ESRD and AF will have a risk profile lending to systemic anticoagulation, and warfarin has traditionally been the mainstay of therapy for this indication.2 The balance of efficacy and safety for warfarin among patients requiring HD with AF, however, is not clear. At best, warfarin offers protection from cardiovascular events when studied in large retrospective studies (hazard ratio, 0.57; 95% confidence interval, 0.37 to 0.86) without increasing bleeding.3 In other reports, including aggregate data from meta-analyses of 37,349 patients in four studies, warfarin offers no reduction in mortality, ischemic events, or stroke but instead, increases major bleeding (adjusted hazard ratio, 1.35; 95% confidence interval, 1.11 to 1.64; P=0.003).5 Because of the unique mechanism of action, warfarin has also been reported to cause the rare effect of calcific uremic arteriopathy in patients with ESRD, which possesses a short-term mortality of 50%.4 These seemingly contrasting data provide a setting to seek out safer alternatives. Although all nonvitamin K–dependent oral anticoagulants (NOACs) are renally eliminated to some extent, they may still be preferred to warfarin in some patients with ESRD because of ease of administration, comparable or improved efficacy, and lower incidence of intracranial hemorrhage and fatal bleeds.6–8 This need has led to the use of NOACs in this population, with apixaban used in up to 10% of patients on HD in one 2015 cohort.9 Unfortunately, clinical data for the use of these newer agents among patients with renal dysfunction are limited. For example, major trials providing insight into apixaban use among patients with AF excluded patients with either a serum creatinine of 2.5 mg/dl or an estimated creatinine clearance of <25 ml/min.6,7 Furthermore, the lower dose of apixaban used in the AF trial in only 5% of the participants was not solely on the basis of renal function but also was used for patients with low body weight or advanced age. Despite being excluded from clinical evidence, rivaroxaban and apixaban have both expanded their Food and Drug Administration label to include dosing recommendations for patients with ESRD and AF receiving HD. In both instances, this decision was on the basis of the results of a single pharmacokinetic (PK) study including eight patients with ESRD receiving one dose of the medication. Although an increase in Area Under the Curve (AUC) of >30% was observed in both studies,10,11 the studied dose was used for the labeling update recommendations. For apixaban, this resulted in the unadjusted dose of 5 mg twice daily for patients with ESRD and AF. Beyond the small population and sole use of PK end points, a crucial concern of these methods was the plasma concentrations not being at steady-state levels and therefore, under-representing the drug exposure.12 Fortunately, independent investigators have attempted to rectify this problem with PK studies comparing different dosing strategies at steady-state concentrations.13,14 The study conducted by Mavrakanas et al.14 evaluating apixaban in patients with ESRD in this issue of the Journal of the American Society of Nephrology was designed to determine the steady-state PK in stable patients with AF on HD receiving apixaban 2.5 mg twice daily and then 5 mg twice daily for 8 consecutive days. PK data were collected on days 1 and 8 of each dosing phase to assess accumulation with repeated dosing. A total of seven and five patients participated in the 2.5- and 5-mg dosing phases of the study, respectively. Significant increase in AUC from day 1 to 8 was seen in patients on HD receiving apixaban 2.5 mg twice daily (298.6 versus 1009.8; P<0.001), suggesting a clinically relevant increase in exposure at steady-state concentrations. Furthermore, the authors conclude that the AUC with the 2.5-mg twice daily dose was similar to that in patients requiring the reduced dose in previous trials and above the tenth percentile of that in the patients receiving the full dose. Additionally, the 5-mg twice daily dose in the patients on HD achieved an AUC above the 90th percentile of this dose in healthy individuals. Although the study provides insight into a clear problem, there were several limitations noted, including small sample size. Additionally, patients with an increased bleeding risk or treated with dual antiplatelet therapy were excluded, which limits the ability to determine bleeding propensity among patient with an elevated baseline risk.15 No data were provided addressing if patients <60 kg or >80 years old were included. Furthermore, the AUC=0–24 was estimated by doubling the measured AUC=0–12. The threshold for comparable PK parameters was AUC measurement between the tenth and 90th percentiles of the reference range, which was referred from venous thromboembolism trials as opposed to AF trials. However, this trial provides steady-state PK data in contrast to the current labeling information that could be vital to providers caring for patients with ESRD requiring anticoagulation. Similar steady-state PK results have been published indicating that a reduced dose of rivaroxaban 10 mg daily for 7 consecutive days in six patients on HD achieved comparable AUC measurements with those in healthy volunteers receiving 20 mg daily.13 Some of the most fascinating insights from both the rivaroxaban and the apixaban single-dose PK studies10,11 are the signal for drug accumulation present in these studies and the relatively smaller dose recommendations that result from studies where the drugs have reached PK steady state (10 versus 15 mg daily for rivaroxaban and 2.5 versus 5 mg twice a day for apixaban).10,13,14,16 These changes in doses are important, however, as evidenced in this report, because the 5-mg twice a day dosing resulted in drug exposure that is twice that of healthy, historical controls.14 The safety implications of these findings cannot be overstated, and caution is requested when interpreting these single-dose trials and applying this information to drug resources and package labeling. Regardless of the lack of long-term efficacy and safety data with NOACs and unclear net benefit with warfarin, systemic anticoagulation is desired in many patients in the ESRD population. Many factors must be considered when administering medications to patients with ESRD beyond the lack of renal elimination. This point has been illustrated when patients with ESRD have been shown to require a lower dose of warfarin compared with patients with normal renal function. Patients with ESRD have been shown to have increased absorption due to uremic toxins affecting the intestinal barrier and efflux pumps, increased circulating free drug due to decrease in plasma albumin, a decrease in Cytochrome P450-mediated drug metabolism also secondary to uremic toxins, and fluctuating drug concentrations with fluid removal during HD.16 Similarly, the totality of the drug’s characteristics, not solely renal elimination, needs to be considered when adjusted dosage in patients with ESRD, including dialyzability of the molecule. Additional patient characteristics must also be considered when choosing anticoagulation agents among patients requiring HD, including advanced age, fragility, weight, drug interactions, and concomitant medications that increase bleed risk. The desire to provide monitoring for degree of anticoagulation or access to a predictable reversal agent must also come into consideration when choosing an agent. Bleeding prediction scores that takes into account patients on HD and the use of NOACs should also be developed to further guide therapy. Presence of mechanical heart valves, known clots, or venous thromboembolism must also be considered. Ulimately, the decisions to anticoagulate or not and then of which agent to use will be highly patient-specific and should weigh not only the thromboembolic versus bleeding risk, but also the overall life expectancy and potential to benefit from treatment. Among patients in whom warfarin is being avoided or is undesirable, this information provides insights that reduced doses of apixaban (2.5 mg twice a day) or rivaroxaban (10 mg daily) can be potential alternatives for patients with AF requiring HD. Because patients age >80 years old or weighing <60 kg have not been clearly represented in the PK studies to date, apixaban is best avoided in patients with either of these characteristics. Two additional novel agents for stroke prevention in patients with AF, edoxaban (a factor Xa inhibitor) and dabigatran (a direct thrombin inhibitor), are also available but should be avoided in the presence of severe renal impairment. Disclosures E.N.D. has no disclosures. J.E.S. is part of the Xarelto speaker’s bureau.