The use of steroid hormones as growth promoters in cattle has been banned within the European Union since 1988 but can still be fraudulently employed in animal breeding farms for anabolic purposes If an efficient monitoring of synthetic compounds (screening and confirmation) is ensured today by many laboratories, pointing out suspicious samples from a natural steroids abuse remains a tricky challenge clue to the difficulty to set relevant threshold levels for these endogenous compounds The development of focused profiling or untargeted metabolomic approaches is then emerging in this context, with the objective to reveal potential biomarkers signing an exogenous administration of such natural steroids This study aimed to assess sample preparation procedures based on microextraction and adapt them to high throughput urinary profiling or metabolomic analyses based on gas chromatography-mass spectrometry measurement. Two techniques have been tested and optimised, namely solid phase microextraction (SPME) and microextraction by packed sorbent (MEPS). using five model steroid metabolites (16 alpha-hydroxyandrosterone, 2 alpha-hydroxytestosterone, 11-keto,5 beta-androstanedione, 6 alpha-hydroxyestradiol and 7 beta-hydroxypregnenolone). The considered performance criteria included not only the absolute response of the targeted compounds but also the robustness of the materials, and the global aspect of the diagnostic ion chromatograms obtained. After only five successive urinary extractions, a clear degradation of the SPME fiber was observed which led to discard this method as a relevant technique for profiling, whereas no degradation was observed on MEPS sorbent. Repeatability and recovery yields were calculated from urine samples fortified at 500 mu g L(-1) and extracted by MEPS. They were found respectively below 11% and above 60% for all model compounds Detection limits were in the 5-15 mu g L(-1) range depending on the compounds, and a good linearity was observed on the 10-75 mu g L(-1) range (R(2) > 0 99) This methodology was applied on urine samples collected from control versus androstenedione-treated bovines, revealing a significant concentration increase for several well-known metabolites such as etiocholanolone, 5 alpha-androstane-3 beta, 17 alpha-diol, 5 beta-androstane-3 alpha, 17 alpha-diol and 5-androstene-3 beta, 17 alpha-diol Finally, these results allowed to confirm the suitability of the developed strategy and give to this new MEPS application a promising interest in the field of GC-MS based steroid profiling and metabolomic (C) 2010 Elsevier B V All rights reserved