OBJECTIVES: To identify novel biomarkers able to discriminate between alcohol-dependent, non-AD alcohol drinkers and controls using metabolomics.

METHOD: Urine samples were collected from 30 alcohol-dependent persons who did not yet start AD treatment, 54 social drinkers and 60 controls, who were then analysed using 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 to develop the discriminatory model. The reproducibility was done using intraclass correlation coefficient (ICC).

RESULTS: The OPLS-DA revealed significant discrimination between AD and other groups with sensitivity 86.21%, specificity 97.25% and accuracy 94.93%. Six biomarkers were significantly associated with AD in the multivariate logistic regression model. These biomarkers were cis-aconitic acid, citric acid, alanine, lactic acid, 1,2-propanediol and 2-hydroxyisovaleric acid. The reproducibility of all biomarkers was excellent (0.81-1.0).

METHODS: A cross sectional study was conducted on three groups: individuals with alcohol use disorders (n=30), social drinkers (n=54) and alcohol-naive controls (n=60). 1H NMR-based metabolomics was used to obtain the metabolic profiles of plasma samples. Data were processed by multivariate principal component analysis (PCA) and orthogonal partial least squares-discriminant analysis (OPLS-DA) followed by univariate and multivariate logistic regressions to produce the best fit-model for discrimination between groups.

RESULTS: The OPLS-DA model was able to distinguish between the AUD group and the other groups with high sensitivity, specificity and accuracy of 64.29%, 98.17% and 91.24% respectively. The logistic regression model identified two biomarkers in plasma (propionic acid and acetic acid) as being significantly associated with alcohol use disorders. The reproducibility of all biomarkers was excellent (0.81-1.0).

METHOD: In this work, resting-state EEG-derived features were utilized as input data to the proposed feature selection and classification method. The aim was to perform automatic classification of AUD patients and healthy controls. The validation of the proposed method involved real-EEG data acquired from 30 AUD patients and 30 age-matched healthy controls. The resting-state EEG-derived features such as synchronization likelihood (SL) were computed involving 19 scalp locations resulted into 513 features. Furthermore, the features were rank-ordered to select the most discriminant features involving a rank-based feature selection method according to a criterion, i.e., receiver operating characteristics (ROC). Consequently, a reduced set of most discriminant features was identified and utilized further during classification of AUD patients and healthy controls. In this study, three different classification models such as Support Vector Machine (SVM), Naïve Bayesian (NB), and Logistic Regression (LR) were used.

RESULTS: The study resulted into SVM classification accuracy=98%, sensitivity=99.9%, specificity=95%, and f-measure=0.97; LR classification accuracy=91.7%, sensitivity=86.66%, specificity=96.6%, and f-measure=0.90; NB classification accuracy=93.6%, sensitivity=100%, specificity=87.9%, and f-measure=0.95.