We propose a new screening method for testosterone (T) doping in sport. The current method for detecting T administration is based on finding a T to epitestosterone ratio (T/E) in urine that exceeds six. The difficulties with T/E are that T administration does not always result in a T/E>6 and that a rare individual will have T/E>6 in the absence of T administration. Our previous studies reveal that carbon isotope ratio helps to determine the origin of the urinary T because the values for T and its metabolites decrease after the administration of exogenous T. In this study, we present a rapid and efficient screening sample preparation method based on three successive liquid-solid extractions, deconjugation with E. coli beta-glucuronidase after the first extraction, acetylation after the second extraction, and a final extraction of the acetates. The 13C/12C of two T metabolites (5beta-androstane-3alpha,17beta-diol and 5alpha-androstane-3alpha,17beta-diol) and one pregnanediol as endogenous reference (5beta-pregnane-3alpha,20alpha-diol) was measured by gas chromatography-combustion-isotope ratio mass spectrometry (GC-C-IRMS) on 10 ml of urine collected from 10 healthy men before and after T administration. Following T administration, the 13 C/12C of 5beta-androstane-3alpha,17beta-diol diacetate and 5alpha-androstane-3alpha,17beta-diol diacetate declined significantly from -26.2 per thousand to -30.8 per thousand and from -25.2 per thousand to -29.9 per thousand, respectively and the 13C/12C of 5beta-pregnane-3alpha,20alpha-diol diacetate was unchanged. In addition, the ratio of androstanediols to pregnanediol increased in the post-T urines.