A polygenic risk score predicts future atrial fibrillation-related stroke/systemic embolism

Abstract

Abstract Background Novel strategies to identify patients at risk of stroke/systemic embolism (SSE) earlier before development of clinical atrial fibrillation (AF) might be useful. We examined whether the AF polygenic risk score (PRS) added to a clinical risk model would improve prediction of AF-related SSE in people without prevalent AF. Methods A total of 62,145 individuals (≥65 years) without prevalent AF who had additional risk factors for stroke (hypertension, diabetes, heart failure, myocardial infarction, stroke or transient ischemic attack, and peripheral artery disease) were included from the UK Biobank. The primary endpoint was AF-related SSE and the secondary endpoint was incident AF. SSE event was considered AF-related when previously undiagnosed AF was identified at the time of SSE or up to 1-year after the SSE. Results For a median of 11.5 (10.6-12.4) years, 334 AF-related SSE and 7,428 incident AF events developed. After adjusting for the clinical risk model, AF-PRS identified a significant gradient for AF-related SSE and incident AF with a 4.59-fold (HR 4.59, 95% CI 3.12-6.75, p<0.001) and 3.14-fold (HR 3.14, 95% CI 2.90-3.40, p<0.001) higher risk in the top quintile of the AF-PRS. The number needed for 5-years of AF screening at the index age of 75 years to find 1 AF-related SSE and incident AF was 91 and 10 for high AF-PRS and 500 and 22 for low AF-PRS, respectively. AF-PRS added to a clinical risk model resulted in a significant improvement in C-statistics for AF-related SSE (C-index; 0.66 to 0.72, p for difference in C-index <0.001) and incident AF (C-index; 0.65 to 0.68, p for difference in C-index <0.001). Conclusions AF screening strategy based on AF-PRS added to a clinical risk model might detect and potentially prevent more cases of AF-related SSE.