Analisi statistica multivariata del profilo steroideo umano: applicazioni in ambito clinico, tossicologico e sportivo

The Technical Document TD2014EAAS was drafted by the World Anti-Doping Agency (WADA) in order to fight the spread of endogenous anabolic androgenic steroids (EAAS) misuse in several sport disciplines. In particular, adoption of the so-called Athlete Biological Passport (ABP) ¿ Steroidal Module allowed control laboratories to identify anomalous EAAS concentrations within the athletes' physiological urinary steroidal profile. Cases of exogenous administration of EAAS for doping purposes progressively increased, since a clear distinction between endogenous and exogenous origin of EAAS found in urine samples is still challenging for anti-doping laboratories. As remarked by recent WADA reports, anabolic agents (including EAAS) represent the most abundant misused class of substances in elite sports. Moreover, anabolic effects are currently obtained also with low doses of EAAS mixtures. In all these cases, the wide inter-individual variability of EAAS levels does not lead to straightforward identification of drug misuse and confuses the interpretation of individual steroid profiles.. More in details, gas chromatography combined with either single or tandem mass spectrometry (GC-MS and GC-MS/MS) is indicated by WADA as an appropriate analytical technique to detect a selected group of EAAS in urine samples of the athletes, including six steroidal biomarkers and five urinary ratios. Moreover, longitudinal measurements are performed over time, by "in competition" and "out competition" controls and collected data are elaborated with an on-purpose Bayesian approach; in particular, such approach convert the classic Population-based approach into an innovative Subject-based Approach. In this contest, we decided to provide a differentapproach for the detection of EAAS misuse, based on multivariate statistical analysis in order to exploit all the biomarkers remarked by the ABP ¿ Steroidal Module. More in details, 6 volunteers' samples have been collected and analyzed using GC-MS and then data have been processed using Principal Component Analysis (PCA) combined with multivariate Hotelling's T2-test; then our results were compared to ones obtained with WADA's approach. The other task of this work was to develop a model based to be applied in clinical diagnostic, too: GC-MS method was extended to 18 endogenous androgen steroids and multivariate models were developed (i.e. combination of PCA and Hotelling's T2-test and, in addition, PLS-DA) in order to distinguish health subjects from the ones suffering from steroidal-correlated pathologies (or, eventually, taking pharmaceutical drugs containing steroids) . Finally, clinical and sport diagnostics were combined, working out a model capable of discriminating the samples of healthy individuals from the ones related to doped subjects or people with pathological conditions related to their steroidal profile .