Abstract 10343: Serum Metabolomics of Human Chronic Stable Angina and Myocardial Infarction: A Probative Discriminant Function Analysis
To address the shortcomings of conventional biochemical approach for detection of myocardial infarction (MI), we applied metabolomics approach using proton nuclear magnetic resonance (1H NMR) spectroscopy as a surrogate method for fast diagnosis of coronary artery disease (CAD), and especially MI. So our aim was to discover differential biomarkers between healthy control, chronic stable angina (CSA) and MI (of various durations from 1-30 days) with a rapid and non-invassive approach. Study comprises serum samples from non-diabetic angiographically proven CAD (coronary artery disease) including CSA (n=36) and MI (n=35) and age matched healthy controls (n=25). Conventional serum troponin level was measured in all subjects. 1H NMR derived serum metabolic data were analyzed using multivariate principal component analysis (PCA) followed by orthogonal partial least-squares discriminant analysis (OPLS-DA). OPLS-DA derived model validity was confirmed using receiver operating characteristic (ROC) analysis. Serum troponin level was significantly different in MI than CSA (5.7±3.5 vs. 0.012±0.01 ng/ml, p<0.001) and MI than HC (5.7±3.5 vs. 0.011±0.01 ng/ml, p<0.001). PCA and OPLS-DA followed by jack-knife approach revealed that the combination of valine, glutamine, citrate, malonate, and cis-aconitate biomarkers were vigorously able to differentiate CAD from HC with maximum ROC (0.96), sensitivity (93%) and specificity (94%). Similar approach determined that isoleucine, valine, and cis-aconitate biomarkers were powerfully able to discriminate MI from CSA (0.96 ROC, 92% sensitivity, 93% specificity). Valine, alanine, glutamine and cis-aconitate were significantly different in CSA than HC (0.96 ROC, 93% sensitivity, 94% specificity). Glutamine, citrate and cis-aconitate were significantly different in MI than HC (0.96 ROC, 94% sensitivity, 94% specificity). 1H NMR based serum metabolic screening appears to be a robust, rapid, and non-invasive approach to differentiate not only CAD from HC but also MI from CSA. Importantly, metabolomic approach can be used to rapid differentiation of CAD which is crucial to start tailored treatment, leading to prompt and appropriate care of patients in the emergency department setting.
- © 2013 by American Heart Association, Inc.