Detection of testosterone microdosing in healthy females.

The ready detectability of synthetic androgens by mass spectrometry (MS)-based antidoping tests has reoriented androgen doping to using testosterone (T), which must be distinguished from its endogenous counterpart making detection of exogenous T harder. We investigated urine and serum steroid and hematological profiling individually and combined to determine the optimal detection model for T administration in women. Twelve healthy females provided six paired blood and urine samples over 2 weeks prior to treatment consisting of 12.5-mg T in a topical transdermal gel applied daily for 7 days. Paired blood and urine samples were then obtained at the end of treatment and Days 1, 2, 4, 7, and 14 days later. Compliance with treatment and sampling was high, and no adverse effects were reported. T treatment significantly increased serum and urine T, serum dihydrotestosterone (DHT), urine 5α-androstane-3α,17β-diol (5α-diol) epitestosterone (E), and urine T/E ratio with a brief window of detection (2-4 days) as well as total and immature (medium and high fluorescence) reticulocytes that remained elevated over the full 14 posttreatment days. Carbon isotope ratio MS and the OFF score and Abnormal Blood Profile score (ABPS) were not discriminatory. The optimal multivariate model to identify T exposure combined serum T, urine T/E ratio with three hematological variables (% high fluorescence reticulocytes, mean corpuscular hemoglobin, and volume) with the five variables providing 93% correct classification (4% false positive, 10% false negatives). Hence, combining select serum and urine steroid MS variables with reticulocyte measures can achieve a high but imperfect detection of T administration to healthy females.