| Abstrakt: |
Currently, alcohol-dependence (AD) diagnosis depends on questionnaires and some biomarkers. However, both lack specificity and sensitivity. Metabolomics using nuclear magnetic resonance spectroscopy (NMR) may provide novel technique for the diagnosis of AD. The study aims to find novel biomarkers of AD in plasma. Blood samples of 30 alcohol-dependent, 54 social drinkers and 60 controls were collected. Plasma samples were obtained and analysed by NMR. Data analysis was done using multivariate analysis including principal component analysis (PCA) and orthogonal partial least square discriminate analysis (OPLS-DA) followed by univariate and multivariate logistic regression. The OPLS-DA model revealed 39 bins with variable influence on projection (VIP) value more than 1, significantly discriminated AD from social drinkers and controls. The sensitivity, specificity and accuracy of the model were 64.3%, 98.2% and 91.2%, respectively. In the univariate logistic regression analysis, 9 peaks were significantly associated with AD with the p value ≤0.1. In the multivariate logistic regression analysis, 4 regions were significantly associated with AD with area under the receiver operating coefficient (AUROC) of 0.961. The sensitivity, specificity and accuracy of the model were 78.6%, 98.2% and 94.2%, respectively. From the four significant regions, two biomarkers (propionic and acetic acid) were identified using the B-BIOREFCODE, the Chenomx, the BMRB databases and 2D HSQC spectra. The study showed that plasma metabolomics was able to find novel biomarkers of AD. These biomarkers if validated, they will aid the precise diagnosis of AD in the future. [ABSTRACT FROM AUTHOR] |
