# Genomic Risk Prediction of Incident Atrial Fibrillation in Older Individuals Without Prior Cardiovascular Disease

## TL;DR

This study shows that a genetic risk score improves the prediction of new-onset atrial fibrillation in older adults without prior heart disease.

## Contribution

The study introduces a polygenic risk score for atrial fibrillation in initially healthy older individuals, demonstrating its predictive value.

## Key findings

- AF-PRS was associated with a 1.74-fold higher risk of incident AF per standard deviation.
- AF-PRS improved risk prediction over clinical models, increasing C-index by up to 10.5%.
- The AF-PRS showed stronger association with AF in women compared to men.

## Abstract

Polygenic risk scores (PRSs) may enhance atrial fibrillation (AF) risk prediction when added to conventional risk factors. Most AF-PRS studies, however, focus on individuals with existing cardiovascular disease, rather than initially healthy older adults followed prospectively.

The objective of the study was to evaluate the predictive performance of a recent (2025) AF-PRS for incident AF in a cohort of healthy older individuals without prior cardiovascular events.

AF-PRS was calculated in 12,906 individuals aged ≥65 years without prior cardiovascular disease or AF at enrollment into the ASPREE (Aspirin in Reducing Events in the Elderly) trial. Cox proportional hazards models assessed HRs) per SD of AF-PRS, alone and with clinical risk factors (age, sex, body mass index, hypertension, diabetes, dyslipidaemia, thyroid-stimulating hormone, smoking, and alcohol). We compared AF-PRS to clinical scores (Cohorts for Heart and Aging Research in Genomic Epidemiology [CHARGE]-AF and hypertension, age, raised body mass index, male sex, sleep apnea, smoking, and Alcohol score[HARMS2-AF]). Model performance was evaluated using Harrell’s C-index and likelihood ratio tests. Sex-stratified analysis was also conducted.

Over a median 4.5-years of follow-up, 654 incident AF cases occurred. AF-PRS was associated with incident AF (adjusted HR: 1.74 per SD; 95% CI: 1.58-1.84) (compared to CHARGE-AF [HR: 1.50] and HARMS2-AF [HR: 1.32] [all P < 0.0001]). Individuals in the highest AF-PRS quintile had 5.44-fold higher risk than those in the lowest (P < 0.0001). The AF-PRS showed stronger association in women than men (HR: 7.09 vs 4.51; interaction P = 0.007). AF-PRS improved prediction beyond clinical factors, increasing the C-index by 8.2% (0.63 → 0.71), 9.5% over CHARGE-AF (0.61 → 0.70), and 10.5% over HARMS2-AF (0.57 → 0.65).

The use of an AF-PRS improves risk prediction of incident AF above clinical risk factors in older individuals without cardiovascular disease.