Abstract Body: Objective: Hyperglycemia (pre-diabetes and diabetes, DM) is associated with heart failure (HF). We aimed to identify distinct metabolites for subclinical cardiac dysfunction (CD), a precursor of HF, in hyperglycemic vs. euglycemic groups.
Method: We used data from the ARIC study (Atherosclerosis Risk in Communities). In HF-free 2492 participants at baseline (2011-2013), 1297 were hyperglycemic (HbA1c>5.7%, fasting glucose>100 mg/dL, DM medication, or a DM diagnosis) and 1195 were euglycemic. We performed logistic regression for the association of 790 metabolites and CD, defined by echocardiographic abnormalities (LV hypertrophy, systolic or diastolic dysfunction) or elevated biomarkers (NTproBNP>125 pg/mL or HS troponin T>14 ng/L in women, >22 ng/L in men) at baseline in two glycemic groups separately. We used Cox regression to evaluate the association between CD-related metabolites (i.e., significant metabolites in the cross-sectional analyses) with HF risk. Analyses were adjusted for clinical risk factors and multiple comparisons (FDR< 5%) and replicated in the Hispanic Community Health Study/Study of Latinos (HCHS/SOL).
Results: 34 out of 790 and 16 out of 790 metabolites were associated with CD in the hyperglycemic (15% Black, 33% men) and euglycemic (22% Black, 47% men) groups, respectively. Metabolites previously identified as microvascular disease-related markers (e.g., pseudouridine, N6-carbamoylthreonyladenosine, N6-acetyllysine, N2, N5-diacetylornithine) were associated with CD in the hyperglycemic group (Fig1). Carbohydrate and cofactor-derived metabolites (e.g., gulonate, erythrocyte) were associated with CD in the euglycemic group (Fig1). 10 and 12 distinct CD-related metabolites in hyperglycemic and euglycemic groups, respectively, were also prospectively associated with HF risk (Hazard Ratios 1.2-1.9)(Fig2). 24 out of 34 and 11 out of 16 CD-related metabolites in the hyperglycemic and euglycemic groups, respectively, were available for validation in HCHS/SOL (n 1202, 34% men). The results were consistent with ARIC, where 10 and 12 distinct CD-related metabolites showed nominal significant association with incident HF in two glycemic groups, respectively.
Conclusion: Metabolites known for microvascular complications (retinopathy, kidney disease) were associated with CD among hyperglycemic participants, supporting the premise that microvascular dysfunction contributes to HF pathogenesis in people with hyperglycemia.