A Metabolomics Study of Cardiac Dysfunction in Hyperglycemia ─ Findings from the Atherosclerosis Risk in Communities Study (ARIC) and the Hispanic Community Health Study/Study of Latinos (HCHS/SOL)

Objective: Hyperglycemic states (pre-diabetes and diabetes) are associated with heart failure (HF) risk. We aimed to identify distinct metabolites for subclinical cardiac dysfunction, a precursor of HF, in hyperglycemic or euglycemic individuals. Research Design and Methods: We conducted cross-sectional and prospective analyses of 2,492 HF-free participants from the Atherosclerosis Risk in Communities (ARIC) Study visit 5, 2011-2013. 1,297 participants were hyperglycemic (Hemoglobin A1c>5.7%, fasting glucose>100 mg/dL, diabetes medication or diagnosis) and 1,195 were euglycemic. We used logistic regression for the association between 790 metabolites and cardiac dysfunction, defined by echocardiographic abnormalities (LV hypertrophy, systolic or diastolic dysfunction) or elevated NTproBNP or troponin T in two glycemic groups separately. We used Cox regression for prospective association between cardiac dysfunction-related metabolites identified in the prior step and HF risk, adjusting for clinical risk factors. Analyses were replicated in the Hispanic Community Health Study/Study of Latinos (HCHS/SOL, n=5,167). Results: Microvascular disease-related metabolites (e.g., pseudouridine, N6-carbamoylthreonyladenosine, N6-acetyllysine, N2, N5-diacetylornithine) were associated with cardiac dysfunction in hyperglycemic individuals. Carbohydrate and cofactor-derived metabolites (e.g., gulonate, erythrocyte) were associated with cardiac dysfunction in euglycemic individuals. These cardiac dysfunction-related metabolites were prospectively associated with HF risk in the two glycemic groups (follow-up 7.5 years, 137 and 94 HF cases, per SD increase hazards ratios range 1-1.9 and 1.1-2.9), respectively. HCHS/SOL results were consistent with ARIC. Conclusions: Metabolites known for microvascular complications were associated with cardiac dysfunction in hyperglycemic individuals but not among their euglycemic counterparts, supporting the premise that microvascular dysfunction contributes to HF pathogenesis in diabetes.