Abstract P026: Incident Atrial Fibrillation in an African-American Cohort: The Jackson Heart Study

Abstract

Introduction: Although many risk factors for atrial fibrillation (AF) are more prevalent in African Americans (AA) than in whites, AF incidence appears to be lower. AF incidence data in AA come largely from US cohorts enrolled in the late 1980s, but the prevalence of risk factors including obesity and smoking has changed over recent decades. Until now, AF incidence information was not available from the Jackson Heart Study (JHS), a community-based cohort study of cardiovascular disease among AAs which enrolled 5,306 participants in 2000-2004. We ascertained AF using two methods, determined age- and sex-specific AF incidence rates, and studied associations of demographic, anthropometric, and cardiovascular risk factors with AF in this contemporary cohort. Methods: Participant characteristics were ascertained at the baseline study visit. Incident AF cases through 2012 were identified from (1) study ECGs and (2) hospital discharge diagnosis code surveillance. As a second method, incident AF was ascertained from ECG and hospital surveillance supplemented by (3) Medicare claims data for inpatient and outpatient care for those enrolled in fee-for-service Medicare. Age- and sex-specific AF incidence rates in JHS were compared with those of AA participants in the Cardiovascular Health Study (CHS) and Multi-Ethnic Study of Atherosclerosis (MESA). Cox proportional hazards models were used to determine the association of risk factors with incident AF. Results: Among the 4,557 JHS participants without prevalent AF at baseline and with complete data, prevalence at baseline was high for hypertension (56%, 2,546/4,557) and diabetes (22%, 985/4,557), but relatively low for current smoking (13%, 586/4,557). Among participants without prevalent AF, using all three AF ascertainment sources over an average of 8.5 years of follow-up, we identified 260 incident AF cases. Compared with AF ascertainment from ECG and hospital surveillance alone, the inclusion of Medicare claims data identified incident AF an average of 69 days earlier and led to an additional 41 prevalent and 36 incident AF cases. Age- and sex-specific AF incidence rates in JHS were broadly similar to those among AAs in CHS and MESA. In a multivariable model, current smoking, hypertension, higher BMI, and a self-reported history of myocardial infarction were all associated with incident AF (HR, 1.70, 95% CI, 1.16-2.51; HR, 1.77, 95% CI, 1.23-2.55; HR per 5 units, 1.19, 95% CI, 1.08-1.32; and HR, 2.19, 95% CI, 1.51-3.17; respectively). Conclusions: In JHS, the inclusion of Medicare claims data in AF ascertainment moved the AF diagnosis date earlier on average and increased the number of AF cases identified, showing it to be a key tool in AF ascertainment. The associations of demographic, anthropometric, and cardiovascular risk factors with AF in this contemporary AA cohort were largely similar to those observed in earlier studies of AA participants.