Background: Epitestosterone [E] has for a long time been considered as a biologically inactive androgen. However, recently a distinct antiandrogenic activity of this naturally occurring endogenous epimer of Testosterone has been demonstrated. Especially the ratios of testosterone/epitestosterone (T/E) seem to be key as inhibition of epitestosterone on androgen activity was postulated. As in autism, a higher androgen activity was implied. We, therefore, suggested higher levels of T/E ratios of children with autism versus children with typical development. Methods: Urine probes of 22 girls with autism (BMI 18.7 ± 4.3; average age 12.3 ± 3.8 years) and a sample of 51 controls (BMI 17.0 ± 2.6; average age 11.9 ± 4 years), as well as 61 boys with autism (BMI 17.04 ± 2. average age 11.9 ± 2.5 years) and 61 control boys (BMI 17.0 ± 2.6; average age 11.1 ± 3.0 years), were analyzed with gas chromatography mass spectrometry. RESULTS: The average T/E ratio of all boys with autism was 2.5 ± 1.8 versus 2.4 ± 1.3 in boys with typical development, respectively. No significant difference between boys with autism versus boys with typical development could be detected (p = 0.977). In girls with autism, the average T/E ratio was 1.4 ± 0.9 versus 2.0 ± 1.4 in girls with typical development, whereby a significant difference could be detected (p = 0.0285). Further, polynomial analysis of the third degree were conducted, showing a dependence from age with reasonable coefficients of determination (0.075 < R2 < 0.22, all samples). Discussion: As encompassing steroid hormone analysis are expensive and work-intensive, we hoped to find an easily applicable biomarker to support diagnostics in autism. However, as a relatively small sample of only 22 girls with autism were analyzed and menstrual cycle and pubertal status were only partly controllable through the matching of BMI and age, the question arises if it was an incidental finding. Nevertheless, one suggestion might be that epitestosterone has the effect of a competitive inhibition on the androgen receptor, which would probably help to explain the higher prevalence of autism in boys as compared to girls. Presumably, as no significant difference was detected in boys, this effect might not be as relevant from a steroid hormone perspective, and other effects such as altered 17/20-hydroxylase activity as previously shown in boys and girls with autism seem to have more relevance. Analysis of larger samples, including plenty of metabolites and enzymatic cascades, as well as the role of backdoor pathway activity of androgen synthesis of girls with autism, are demanded in order to validate current findings of altered steroid hormones in autism.