AHA Scientific Sessions 2026 Abstract · Nov 2, 2026
Multiple Biomarkers and Risk of Heart Failure in Atrial Fibrillation
By Sullivan M, Cordova Sanchez AA, Short S et al.
Introduction: Atrial fibrillation (AF) and heart failure (HF) have a bidirectional relationship. Although multiple mechanisms including structural ventricular remodeling, impaired calcium handling, and neurohormonal activation may link AF to HF, the pathophysiology of this association remains poorly defined. Biomarkers may aid in clarifying this relationship by identifying pathways that contribute to the development of HF in AF and by enabling biomarker-based risk-stratification. Methods: The REasons for Geographic And Racial Differences in Stroke (REGARDS) study enrolled 30,239 Black and White adult participants between 2003 and 2007. In this study, we included participants with baseline AF and excluded those with baseline HF or stroke (owing to the selection criteria for biomarker measurement in a prior study on stroke outcomes). Fifteen cardiovascular risk biomarkers were measured at baseline. Hazard ratios (HRs) for incident HF per standard deviation (SD) increment of each biomarker were calculated utilizing Cox models adjusted for demographics, HF risk factors, and cardiovascular medications. Results: We identified 2,400 participants with AF at baseline and no prevalent HF or stroke; the median age was 70 (IQR 62-77) years, 49% were female, and 30% were Black. With a median follow-up of 8 years, 340 participants developed HF. In the multivariable models (see Figure), N-terminal pro-B-type natriuretic peptide (HR per SD 2.34; 2.02-2.70), growth differentiation factor 15 (HR 1.55; 1.32-1.82), cystatin C (HR 1.50; 1.35-1.66), gamma-glutamyl transferase (HR 1.28; 1.15-1.43), C-reactive protein (HR 1.25; 1,12-1.41), high-density lipoprotein (HR 1.20; 1.05-1.35), interleukin-6 (HR 1.19; 1.08-1.32), and D-dimer (HR 1.13; 1.01-1.27) were associated with the development of HF in individuals with AF. Higher triglycerides (HR 0.81; 0.71-0.92) and higher estimated glomerular filtration rate (HR 0.66; 0.60-0.73) were associated with a lower risk of incident HF. There were no associations for galectin-3, factor VIII, lipoprotein (a), creatinine, low-density lipoprotein, and total cholesterol. Excluding individuals with baseline elevation in NT-proBNP, which may reflect undiagnosed heart failure, yielded similar results. Conclusions: Several biomarkers were associated with incident HF in AF. The most significant biomarkers predisposing individuals to HF related to myocardial stress, systemic inflammation, and worse renal function. NT-proBNP demonstrated the strongest association with incident HF in individuals with AF, likely reflecting myocardial stretch, elevating filling pressures, and subclinical cardiac dysfunction that contribute to the progression from AF to HF. Results suggest possible targets for prevention.