Left atrial function and polygenic risk scores for predicting atrial fibrillation and related stroke/systemic embolism

Abstract

Abstract Objective To explore whether the addition of left atrial (LA) function, polygenic risk score (PRS), or both to clinical risk scores improves predictions of atrial fibrillation (AF) and related stroke/systemic embolism (SSE). Methods A total of 36,919 individuals without AF who were assessed for LA emptying fraction and AF-PRS were included from the UK Biobank imgaing enhancement. The primary aims were incident AF and AF-related SSE. AF-related SSE was ascertained when previously undiagnosed AF was identified at the time of an SSE or up to 1-year after an SSE. Results Among the 36,919 participants (median 64 years; 52.4% women), 535 incident AF and 27 AF-related SSE developed over a median follow-up of 2.9 years. Both LA emptying fraction and AF-PRS were associated with the risk of incident AF (hazard ratio [HR] per standard deviation [SD] decrease in LA emptying fraction; 2.13; 95% CI, 1.99-2.27 and HR per SD increase in AF-PRS [95% CI]; 1.65 [1.52-1.79]) and AF-related SSE (HR per SD decrease in LA emptying fraction [95% CI]; 3.52 [2.85-4.35] vs. HR per SD increase in AF-PRS [95% CI]; 1.86 [1.30-2.68]). The addition of LA emptying fraction to a clinical risk score was associated with a significant improvement in C-statistics for incident AF (C-index; 0.704 to 0.788, p for difference in C-index <0.001) and AF-related SSE (C-index; 0.666 to 0.868, p<0.001) but not for AF-PRS (p for difference in C-index > 0.05). The addition of LA emptying fraction was associated with a net reclassification of 26.2% and 21.8% for incident AF and AF-related SSE whereas 11.5% and none with the addition of AF-PRS, respectively. There was limited additive predictive utility with the combination of the two markers over LA emptying fraction alone for both incident AF and AF-related SSE (p for difference in C-index > 0.05). Conclusions Adding LA function to a clinical risk score was associated with statistically significant and clinically meaningful improvements in risk predictions of AF and AF-related SSE compared with adding AF-PRS. The addition of both LA function and AF-PRS to a clinical risk score had limited additive predictive utility compared with the addition of LA function alone.