Abstract
Objective: Metabolomic profiling has been used to identify novel biomarkers and mechanisms of cardiovascular disease risk. We will review seminal studies in this arena.
Background: Cardiovascular disease (CVD) is the greatest public health burden in the United States and globally. Clinical models for CVD risk are incomplete. Novel molecular technologies such as metabolomic profiling hold great promise in improving such models. In our own work, we have used targeted, quantitative metabolomic profiling to identify independent, incremental biomarkers and mechanisms for CVD risk.
Methods: We have used mass-spectrometry based, quantitative, targeted metabolomic profiling of 63 metabolites in peripheral, fasting plasma collected from a large cardiovascular cohort of patients referred for cardiac catheterization at Duke University (CATHGEN, N=3900). Principal components analysis (PCA) was used for multidimensional data reduction. PCA-derived factors were compared between coronary artery disease (CAD) cases and non-CAD controls using linear regression. Association between PCA-derived factors and incident CVD events were tested using Cox proportional hazards time-to-event analysis. Incremental predictive capabilities were tested using reclassification analyses.
Results: We found that a novel biosignature composed of short chain dicarboxylacylcarnitine (SCDA) metabolites is independently and incrementally predictive of incident CVD events. Specifically, the SCDA factor predicted incident death in multivariable models (HR 1.17 [1.05–1.31], P = .005). In reclassification analyses, 27% of intermediate-risk patients were correctly reclassified (net reclassification improvement 8.8%). We have shown these same metabolites to be heritable in families burdened with early onset CVD. Preliminary results of a genomewide association study (GWAS) of SCDA levels have identified several genetic variants reporting on a shared biological pathway potentially mediating CVD event risk.
Conclusions: We (and others) have successfully used metabolomic profiling in cardiometabolic diseases for both “form” (identification of novel biomarkers for improved risk prediction) and “function” (identification of novel mechanisms of disease).