BACKGROUND AND OBJECTIVES: Urinary steroid profiling is used in doping controls to detect testosterone abuse. A testosterone over epitestosterone (T/E) ratio exceeding 4.0 is considered as suspicious of testosterone administration, irrespectively of individual heterogeneous factors such as the athlete's ethnicity. A deletion polymorphism in the UGT2B17 gene was demonstrated to account for a significant part of the interindividual variability in the T/E between Caucasians and Asians. Here, the variability of urinary steroid profiles was examined in a widely heterogeneous cohort of professional soccer players. Method: The steroid profile of 57 Africans, 32 Asians, 50 Caucasians and 32 Hispanics was determined by gas chromatography-mass spectrometry. RESULTS: Significant differences have been observed between all ethnic groups. After estimation of the prevalence of the UGT2B17 deletion/deletion genotype (African: 22%; Asian: 81%; Caucasian: 10%; Hispanic: 7%), ethnic-specific thresholds were developed for a specificity of 99% for the T/E (African: 5.6; Asian: 3.8; Caucasian: 5.7; Hispanic: 5.8). Finally, another polymorphism could be hypothesised in Asians based on specific concentration ratio of 5alpha-/5beta-androstane-3alpha,17beta-diol in urine. CONCLUSION: These results demonstrate that a unique and non-specific threshold to evidence testosterone misuse is not fit for purpose. An athlete's endocrinological passport consisting of a longitudinal follow-up together with the ethnicity and/or the genotype would strongly enhance the detection of testosterone abuse. Finally, additional genotyping studies should be undertaken to determine whether the remaining unexplained disparities have an environmental or a genetic origin.