Abstract P100: Heritability of Atrial Fibrillation

Background: Atrial fibrillation (AF) is a common and heritable arrhythmia associated with substantial morbidity. Previous reports have implicated over 30 genes and 16 loci in the pathogenesis of AF, but the genetic architecture of the arrhythmia has not been systematically characterized. Methods: We assessed AF heritability in 120,286 unrelated individuals of European ancestry (2,981 with AF) from the population-based UK Biobank. We ascertained AF based on self-reported disease history and/or billing codes from medical record data. To validate the phenotype definition, we examined genetic associations with AF for variants at known susceptibility loci reported in prior independent genome-wide association studies. We then estimated the proportion of AF variance explained by additive genetic variation (heritability) using a variance components method (BOLT-REML) based on ~800,000 independent high quality imputed variants with minor allele frequencies (MAF) ≥ 1%. We further evaluated the heritability of AF by sex, allele frequency, and established AF loci from association studies (+/- 500 kb of a top associated variant). All analyses were adjusted for baseline age, sex, array, and 15 principal components of ancestry. Results: The average age was higher in AF cases than in referents (62.3 years vs. 56.8 years), and more AF cases were male (69% vs. 47%). We observed expected associations between genetic variation and AF, validating our AF definition. The heritability of AF was 22.3% (95% confidence interval [CI] 15.8%-28.8%), and no substantive difference was observed between males and females. We observed that the AF variance was mainly explained by common variants with MAF ≥ 5% (20.6%, 95%CI 15.2%-25.9%) rather than low-frequency variants with MAF between 1%-<5% (0%, 95%CI 0.0%-0.06%). Only 4.7% (95%CI 3.7%-5.7%) of AF variance was explained by known AF susceptibility loci. Conclusions: Using a population-based sample of European ancestry, we observed that a substantial proportion of variance in AF risk is attributable to additive genetic variation. AF variance is predominantly accounted for by common variants. Only a small proportion of AF variance is attributable to known AF loci. Our findings suggest that further genetic discovery, with an emphasis on common variation, is warranted to understand the causal genetic basis of AF.