Abstract

Background: Heart failure (HF) is a life-threatening aging-related syndrome with growing impact on the global population. It is estimated that >8 million adults in the US will be living with HF by 2030. HF with preserved EF (HFpEF), defined by an ejection fraction ≥50%, frequently has delayed diagnosis and limited treatment options. HFpEF progression can be affected by both gene and environment via different molecular pathways, which can be assessed by epigenetic factors measured by DNA methylation (DNAm). Methods: The DNAm levels of peripheral blood cells at enrollment into the Million Veteran Program (MVP) were profiled using the Illumina’s EPIC array. After excluding samples with low quality or discordant sex, and merging with the phenotypic data, the multi-ancestry (~70%, 24%, 5.4%, and 0.6% are European, African, Hispanic, and Asian Americans, respectively) epigenetic association study in the MVP included 1658 with incident HFpEF and 33926 controls without any record of HF. Incident HFpEF was defined as the presence of a diagnostic code for HF after enrollment into MVP and a measured left ventricle ejection fraction ≥50% within 90 days of first appearance of HF diagnosis code. For each DNAm site, the association with HFpEF was examined in a multiple regression model adjusted for age, sex, race/ethnicity, calculated cell type proportions, and potential batch effects. Epigenome-wide significance (EWS) threshold (p<6.6×10 -8 ) corrected for multiple testing was used to identify DNAm sites associated with incident HFpEF. Results: We identified 106 EWS DNAm sites associated with incident HFpEF across 22 chromosomes, including the most significant DNAm in chromosomes 9 ( AQP3 , p=2.50×10 -27 ), 17 ( SOCS3 , p=3.39×10 -26 ), 19 ( SBNO2 , p=2.64×10 -25 ), 6 ( FGD2 , p =1.49×10 -24 ), and 18 ( NDUFV2 , p=1.57×10 -24 ). Most of the identified DNAm (92 out of 106, 86.8%) showed negative association with incident HFpEF. After additional adjustment for body mass index and diabetes status, the associations of 54 and 47 DNAm sites remained EWS with incident HFpEF. DNAm age (years) predicts incident HFpEF (OR 1.045, 95% CI 1.039-1.052, p=6.52×10 -14 ) independently from chronological age. Conclusion: The significant epigenetic associations with incident HFpEF suggest the crucial role of epigenetic regulation in the pathogenesis of HFpEF. These findings may lead to novel targets for treatment and prevention which could eventually reduce the burden of HFpEF among the growing aging population.