Altrenogest, a synthetic progestin widely used to suppress oestrus in mares, improves manageability and safety in training and competition. However, its administration can result in the detection of prohibited steroid impurities, including trendione, trenbolone, and epitrenbolone. Injectable formulations have been assumed to contain higher levels of these impurities, posing greater risk of anabolic effects compared to oral preparations. Differentiating the administration route is therefore critical for regulatory integrity, while ensuring access to this therapeutic agent for equine welfare and rider safety. This study investigated whether a metabolomic and machine learning approach could distinguish oral from intramuscular altrenogest administration in racehorses.
A controlled trial involving six horses (three oral, three intramuscular) was conducted by Charles Sturt University (A19050). Altrenogest was administered therapeutically for two weeks, with urine collected pre-administration, at multiple hourly intervals post-administration, and daily up to 21 days (32 samples total). Samples underwent solid phase extraction and were analysed by LC-QTOF-MS using both positive and negative electrospray ionisation with DDA acquisition. Data was processed with MS-DIAL, Python and R scripts, and MetaboAnalyst.
Untargeted metabolomics and statistical analysis identified five sulfated steroids—estrone sulfate, testosterone sulfate, 2-methoxyestradiol sulfate, pregnenolone sulfate, and cortisol sulfate—as potential biomarkers for administration route differentiation. A random forest classification model, trained on normalised peak area data, achieved an area under the curve (AUC) of 0.965 (95% CI: 0.931–0.995). Compound identities were confirmed using MS/MS spectra and retention time matching with purchased and synthesised reference standards.
This preliminary study demonstrates that metabolomic profiling combined with machine learning can distinguish oral from intramuscular altrenogest administration. While further validation with larger cohorts and race-day samples is required, the approach shows strong potential to resolve regulatory challenges associated with altrenogest use and its steroidal impurities, supporting both equine welfare and the integrity of the racing industry.