Abstract 4141779: Metabolomic Markers of Aortic Stenosis Development: the Atherosclerosis Risk in Communities (ARIC) Study

Background: Prior metabolomics studies in aortic stenosis (AS) have focused on patients with severe stenosis and their response to intervention. Limited data exist regarding circulating metabolites predictive of AS development. Aim: Identify circulating metabolites associated with aortic valve (AV) hemodynamics and incident AV events. Hypothesis: We hypothesized that large-scale metabolomics can identify novel circulating metabolites relevant to AS development. Methods: Among participants in the community-based ARIC cohort study who underwent protocol echocardiography and metabolomic profiling by chromatography mass spectrometry (Metabolon) at study Visit 5 (2011-2013) and were free of AV replacement, we related metabolite values to AV peak velocity (AVmax) and dimensionless index (DI) cross-sectionally and to incident AV events using multivariable linear and Cox regression models. Metabolites with ≤10% missingness (n=736 metabolites) were included and normalized using log2 transformation. Post-Visit 5 AV events were ascertained based on ICD codes for AV-related hospitalization and interventions. Regression models were adjusted for age, sex, race, center, heart rate, blood pressure, estimated glomerular filtration rate, and cardiovascular (CV) risk factors and diseases. Statistical significance was based on a False Discovery Rate (FDR) <0.05. Results: Among 5,107 participants included, mean age was 75±5 years, 57% were women, and 18% reported Black race. Mean LVEF was 65 ±7%, AVmax 1.3 ±0.4 m/s, and DI 0.76 ±0.14. After adjusting for CV risk factors and disease, 80 metabolites were cross-sectionally associated with AVmax at FDR <0.05, 28 of which were further associated with DI ( Figure A ). Over a median follow-up of 8 [Q1- Q3, 7-9] years, 245 experienced an AV-related hospitalization. Of the 28 metabolites associated with AV hemodynamics, 20 were associated with incident AV-related hospitalization ( Figure B ). Amino acids and amino sugar metabolites were associated with greater hemodynamic AV stenosis and a higher risk of AV events, while higher dicarboxylic fatty acids and other lipids values were associated with better AV hemodynamics and a lower risk of AV-related hospitalization ( Figure C ). Conclusion: Large-scale metabolomics in an older community-based cohort identifies novel circulating markers of AS risk. While amino acid metabolites associated with high risk, dicarboxylic fatty acids were associated with lower risk.