Prevalence of Chronic Atrial Fibrillation in Dialysis Patients

Abstract

Background: Atrial fibrillation (AF) is reported to be common in patients on maintenance dialysis, but estimates of prevalence vary substantially. To date, no Medicare claims-based approach has been employed to rigorously assess prevalence of chronic AF. Methods: A novel database was created to identify patients undergoing maintenance dialysis who were dually-eligible for Medicare and Medicaid for at least 3 months in 2004-05. A Medicare claims-based algorithm was used to generate a plausible range of chronic AF prevalences using four approaches. Poisson analysis was employed to determine the demographic, functional status, comorbidity, and other factors, as assessed by the Medical Evidence Form, associated with chronic AF. Results: Of 102,748 dually-eligible individuals for whom data was complete, 21,540 (21.0%) had at least one claim for non-perioperative AF. Raw percentages (irrespective of length of follow-up time) of individuals with chronic AF ranged from 9.8% (using the most inclusive strategy) to 4.6% (the most exclusive); intermediate approaches led to closely-clustered estimates of 8.1% and 6.4%. The intermediate approaches demonstrated chronic AF prevalence to range from 64.2 (95% confidence intervals, 62.9 - 65.5) to 50.4 (49.2 - 51.7) per 1000 patient-years. Age 60 years, male sex, Caucasian race, body mass index 30 kg/m2, and inability to ambulate were associated with chronic AF; hypertension as a comorbidity was inversely associated with AF. Occurrence of AF was roughly 10 times greater in the youngest patients when compared to similar individuals not on dialysis. Conclusions: A linked Medicare-Medicaid database, together with a claims-based diagnostic algorithm, was used to generate prevalence estimates for chronic AF in dually-eligible dialysis patients. As expected, AF is far more common than in the non-dialysis population.