Abstract / Summary
Cardiovascular diseases (CVDs) are highly heritable, but pathogenesis at the organ and physiological level remains poorly defined. Polygenic risk scores (PRSs), which estimate individual genetic susceptibility to a disease, may allow for the identification of associated subclinical structural abnormalities. Ultimately, identifying where cardiovascular polygenic risk manifests can shape mechanistic hypotheses, guide early interventions, and motivate prevention trials for cardiac remodelling. This study investigated the cardiac features associated with elevated polygenic risk for five common CVDs [heart failure (HF), coronary artery disease (CAD), atrial fibrillation (AF), abdominal aortic aneurysm (AAA) and ischaemic stroke (IS)] using 28 imaging-derived phenotypes (IDPs) from cardiac magnetic resonance imaging of ~62,000 participants in the UK Biobank cohort. We identified 97 significant PRS-IDP associations as well as 71 IDP-incident disease associations (FDR [≤] 0.05). We further identified 32 significant putative mediators between CVD PRSs and incident disease events, revealing that across CVDs, polygenic risk manifested as distinct patterns in cardiac structures. HF implicated all cardiac chambers, particularly left ventricular and left atrial dysfunction alongside enlarged aorta. AF was characterised by biatrial enlargement and reduced ejection fractions, most prominently in the left atrium. IS and CAD predominantly involved left ventricular hypertrophy, specifically left ventricular wall thickness for IS, whereas AAA was mainly characterised by enlarged descending aorta. Overall, our findings empirically demonstrate that cardiac IDPs partially mediate polygenic risk for CVDs, uncovering a heterogeneous landscape of organ-level mechanisms across CVDs.