An important goal of forensic toxicology and clinical medicine is to identify appropriate biomarkers of ethanol consumption, so as to objectively support the diagnosis of chronic excessive alcohol intake. In the common practice, screening of large population sets are commonly executed by simple and inexpensive immunometric determination of indirect biomarkers, despite their limited specificity and sensitivity. The present study proposes a multivariate model, based on several indirect biomarkers of alcohol abuse, capable of considerably enhancing their diagnostic efficiency, in terms of specificity and sensitivity. Blood samples were collected from 240 healthy non-alcohol abusers and other 183 subjects, objectively classified as non–drinkers (N=62), social drinkers (N=48) and active heavy drinkers (N=73). Five indirect biomarkers of ethanol consumption (AST, ALT, γ-GT, MCV and CDT) were determined. The original biomarkers' values, and some previously proposed mathematical combinations of them, were used to perform ROC curve analysis, which provides a diagnostic performances evaluation of the traditional approach, in terms of discrimination between healthy non-alcohol abusers and active heavy drinkers. The results from the univariate approach were compared with those obtained from a multivariate strategy based on a UNEQ class-model technique. This model was developed by splitting the 423 subjects under study into a training set (N=203) and an evaluation set (N=220). It was found that the UNEQ class-model built on γ-GT, CDT, MCV and AST can noticeably improve the screening potential of indirect biomarkers in the evaluation of alcohol misuse, with respect to the traditional univariate strategy. Moreover, this model can be easily introduced in the clinical routine work. From the present screening evaluation of healthy non-alcohol abusers, only a moderate percentage of subjects (10%20% at a fixed 10% or 5% active heavy drinker misrecognition rate) is requested to undergo more expensive and time-demanding confirmation procedures, such as the determination of ethyl glucuronate in head hair, with considerable reduction of work and expenses. The improved capability expressed by this multivariate evaluation make the reappraisal of indirect biomarkers topical, in contrast with the recent trend of considering their use void of practical significance.
Multivariate strategies for screening evaluation of chronic alcohol abuse
PIRRO, VALENTINA;Salomone A.;VINCENTI, Marco
2012-01-01
Abstract
An important goal of forensic toxicology and clinical medicine is to identify appropriate biomarkers of ethanol consumption, so as to objectively support the diagnosis of chronic excessive alcohol intake. In the common practice, screening of large population sets are commonly executed by simple and inexpensive immunometric determination of indirect biomarkers, despite their limited specificity and sensitivity. The present study proposes a multivariate model, based on several indirect biomarkers of alcohol abuse, capable of considerably enhancing their diagnostic efficiency, in terms of specificity and sensitivity. Blood samples were collected from 240 healthy non-alcohol abusers and other 183 subjects, objectively classified as non–drinkers (N=62), social drinkers (N=48) and active heavy drinkers (N=73). Five indirect biomarkers of ethanol consumption (AST, ALT, γ-GT, MCV and CDT) were determined. The original biomarkers' values, and some previously proposed mathematical combinations of them, were used to perform ROC curve analysis, which provides a diagnostic performances evaluation of the traditional approach, in terms of discrimination between healthy non-alcohol abusers and active heavy drinkers. The results from the univariate approach were compared with those obtained from a multivariate strategy based on a UNEQ class-model technique. This model was developed by splitting the 423 subjects under study into a training set (N=203) and an evaluation set (N=220). It was found that the UNEQ class-model built on γ-GT, CDT, MCV and AST can noticeably improve the screening potential of indirect biomarkers in the evaluation of alcohol misuse, with respect to the traditional univariate strategy. Moreover, this model can be easily introduced in the clinical routine work. From the present screening evaluation of healthy non-alcohol abusers, only a moderate percentage of subjects (10%20% at a fixed 10% or 5% active heavy drinker misrecognition rate) is requested to undergo more expensive and time-demanding confirmation procedures, such as the determination of ethyl glucuronate in head hair, with considerable reduction of work and expenses. The improved capability expressed by this multivariate evaluation make the reappraisal of indirect biomarkers topical, in contrast with the recent trend of considering their use void of practical significance.
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
