The detection of the administration of an androgen such as testosterone that could be present normally in human bodily fluids is based upon the methodical evaluation of key parameters of the urinary profile of steroids, precisely measured by GC/MS. Over the years, the markers of utilization were identified, the reference ranges of diagnostic metabolites and ratios were established in volunteers and in populations of athletes, and their stability in individual subjects was studied. The direct confirmation comes from the measurement of delta (13)C values reflecting their synthetic origin, ruling out a potential physiological anomaly. Several factors may alter the individual GC/MS steroid profile besides the administration of a testosterone-related steroid, the nonexhaustive list ranging from the microbial degradation of the specimen, the utilization of inhibitors of 5alpha-reductase or other anabolic steroids, masking agents such as probenecid, to inebriating alcohol drinking. The limitation of the testing strategy comes from the potentially elevated rate of false negatives, since only the values exceeding those of the reference populations are picked up by the GC/MS screening analyses performed by the laboratories on blind samples, excluding individual particularities and subtle doping. Since the ranges of normal values are often described from samples collected in Western countries, extrapolating data to all athletes appears inefficient. Furthermore, with short half-life and topical formulations, the alterations of the steroid profile are less pronounced and disappear rapidly. GC/C/IRMS analyses are too delicate and fastidious to be considered for screening routine samples. An approach based upon the individual athlete's steroid profiling is necessary to pick up variations that would trigger further IRMS analysis and investigations.