Assessment of the Quality of Doping Substances Identification in Urine by GC/MS/MS.

Abstract

Direct control of doping in sports is based on the analysis of active substances and/or their metabolites in urine samples of the athletes by GC/MSn or LC/MSn. The World Anti-Doping Agency, WADA, defined criteria for the agreement between retention times, RT, or relative retention times, RRT, and abundance ratios, AR, of characteristic ions of the mass spectrum of the analyte in a calibrator (positive control) and the sample. Strict criteria for confirming analyte presence were defined to reduce false positive results rates, FP. However, these criteria can lead to high rates of false negative results, FN. This work presents a methodology to define statistically sound criteria for the agreement between RRT and AR that allow keeping the FN under control. This work also determined the FP of identifications. The statistical criteria were set from Monte Carlo simulations of correlated RT and ion abundances. The simulation of AR and signal noise was also used to estimate the FN and FP of identifications based on the criteria defined by WADA. The developed tools were successfully applied to the control of nine doping substances in urine samples by GC/MS/MS. The estimated FN were tested from independent experimental tests proving estimates are accurate. The criteria defined by WADA are associated with extremely low FP but, in some cases, associated with FN much larger than 50%. The statistically sound identification criteria allow a more convenient balance between FN and FP. The user-friendly spreadsheet used in this work is made available as Supporting Information.