Abstract 13957: Peripheral Metabolite Profiles Predict Cardiomyopathy in a Cohort of Cardiac Catheterization Patients
Background: The mechanisms underlying systolic heart failure remain incompletely characterized. Molecular technologies such as metabolomics have been successfully applied to identify novel biomarkers and mechanisms of cardiovascular disease risk. We hypothesized that similar metabolomic profiles would identify novel biomarkers cardiomyopathy and elucidate biological mechanisms.
Methods: The study population consisted of individuals identified from the Duke CATHGEN biorepository. Cardiomyopathy cases were defined as having an ejection fraction [EF]<45% and were further stratified into ischemic cardiomyopathy (CADindex≥32,N=380) and nonischemic cardiomyopathy (CADindex≤23,N=95). Controls were defined as having no CAD (CADindex≤23) and EF≥45% (N=1291). Principal components analysis (PCA) was used to reduce the large number of correlated metabolites into uncorrelated factors. Association with cardiomyopathy was tested using linear regression models in a basic (race, age and sex) and a multivariable (hypertension, dyslipidemia, smoking, diabetes, and body mass index) adjusted model. Results are reported as p-values following Bonferroni correction for multiple comparisons at the level of metabolite factors.
Results: PCA identified 14 metabolomic factors, grouped in biologically plausible clusters. Four factors were strongly associated with cardiomyopathy in both basic and multivariable analyses (p≤4E-4). These included factors composed of ketone related metabolites (basic and multivariable p=0.0002), urea cycle amino acids (basic p=3E-3, multivariable p=4E-4), short chain acylcarnitines (basic p=5E-7, multivariable p=7E-6), and C22 acylcarnitine (basic p=2E-5, multivariable p=0.7E-6). While the short chain acylcarnitines were only associated with ischemic cardiomyopathy, the other metabolite factors were associated with both ischemic and non-ischemic cardiomyopathy.
Conclusions: Using targeted quantitative metabolomics, we have identified metabolites reporting on components of mitochondrial function and the urea cycle to be associated with the presence of cardiomyopathy. These observations provide insight into potential disease pathways and serve as potential biomarkers of systolic dysfunction.
- © 2013 by American Heart Association, Inc.