Apixaban May Offer a ‘Practical Alternative’ for Dialysis Patients With Afib

Two anticoagulants showed similar safety and efficacy in peritoneal dialysis (PD) patients with newly diagnosed nonvalvular atrial fibrillation (Afib), according to a retrospective study of real-world data.

Among 660 patients, the risks of five efficacy outcomes — ischemic stroke, myocardial infarction, systemic thromboembolism, cardiovascular death, and a cardiovascular composite — were comparable between those treated with apixaban (Eliquis) and warfarin (Coumadin) in an intention-to-treat analysis.

Findings remained consistent in an as-treated analysis, Mingyue He, MD, of Baylor College of Medicine in Houston, reported at the National Kidney Foundation’s Spring Clinical Meeting.

Similarly, no significant differences were reported across five safety outcomes, which included intracranial bleeds, major bleeding, clinically important bleeding, any stroke or intracranial bleeding, and all-cause mortality.

“Given the elevated stroke risk in kidney failure and the challenges of maintaining therapeutic anticoagulation with warfarin, apixaban may represent a practical alternative where anticoagulation is pursued,” He said.

Prior data in non-dialysis populations have favored apixaban over warfarin for preventing stroke or systemic embolism while showing less bleeding and lower mortality risk. A 2019 meta-analysis of patients with late-stage chronic kidney disease also suggested apixaban was superior in reducing adverse outcomes.

Apixaban is generally easier to manage as it does not require dietary restrictions or routine monitoring of international normalized ratio (INR) levels like warfarin does. While Afib affects roughly 0.5% to 2% of the general population, the prevalence jumps to between 11% and 25% for patients with kidney failure, He pointed out.

“Long-term anticoagulation remains challenging in this population,” she explained. “Patients with kidney failure on hemodialysis face both increased risk of thrombosis and increased risk of major bleeding. Additionally, the commonly used risk scores — such as HAS-BLED score and the CHA₂DS₂-VASc score — perform suboptimally in patients with kidney failure on dialysis, making the risk-benefit assessment quite challenging.”

Clinical evidence in this area is sparse, as kidney failure patients are often excluded from pivotal randomized trials, He pointed out. Most observational studies also group hemodialysis and PD patients together, with hemodialysis patients making up the vast majority of the cohorts.

“PD-specific data remains limited, but the distinction is very important because hemodialysis patients and peritoneal dialysis patients differ in several clinically relevant ways,” she noted. He pointed to factors such as intradialytic heparin exposure and repeated vascular cannulation in hemodialysis patients, which create different bleeding risks compared with patients on PD.

Because of these differences, data from hemodialysis-dominant cohorts may not be generalizable to PD patients. Currently, no randomized controlled trials directly compare apixaban and warfarin specifically in the PD population.

This observational, population-based cohort study utilized the U.S. Renal Data System to assess Medicare beneficiaries diagnosed with Afib between January 2014 and December 2019. Using propensity-score matching, the researchers compared 330 new users of apixaban and 330 new users of warfarin.

At baseline, the average participant age was 68 year, 76% were white, and 17% were Black. The mean dialysis vintage was 4 years. The median CHA₂DS₂-VASc score was 4, indicating a high stroke risk. Additionally, 20% of the cohort was receiving P2Y12 antiplatelet therapy, a fifth had prior cerebrovascular disease, and 5% had a history of clinically important bleeding.

Anticoagulation began a median of 6 days after Afib diagnosis in both groups. While treatment duration was generally short, it was notably shorter in the apixaban group (68 days) compared with the warfarin group (96 days).

Two control outcomes — pneumonia and hip fracture — showed no difference between the groups, suggesting a lack of major residual confounding factors. However, He noted that because of the observational design, residual confounding cannot be entirely ruled out. Other limitations included the modest sample size and early treatment discontinuation.

“Confirmation of these findings in PD-specific randomized trials is still needed,” He concluded.