A threshold concentration for cortisol in equine urine was fixed at 1.0 microgram ml-1 by the racing authorities in 1994. In some circumstances, interlaboratory discrepancies were observed and structural cortisol modification was revealed. In order to elucidate the degradation process and to prevent it, an identification study of the produced compound was carried out. The modified substrate was characterised by the same molecular weight as cortisol and a shorter retention time under the conditions used for the cortisol quantification (M.A. Popot, PhD Thesis, King's College, London, 1996). To identify this isomer, HPLC-APCI-MS and MS-MS methods were applied to the cortisol post-administration extract diluted in the mobile phase which was either a mixture of methanol-water or labelled methanol-water (CH3OD-D2O). Stereochemical effects were studied under these conditions. Deuterium-hydrogen exchange was also monitored by HPLC-APCI-MS. Considering MS and MS-MS data, the hypothesis of isomerisation at C11 giving the 11 alpha-cortisol was rejected. Isotopic labelling has allowed determination of the number of labile hydrogen atoms of the modified cortisol. Cortisol and modified cortisol have the same number of mobile hydrogens. Therefore, the hypothesis of a reduction at C20 along with an oxidation at C11 has also been rejected. Deuterium-hydrogen exchanges could be a useful tool to elucidate the structure of compounds analysed by HPLC-APCI-MS in a complex matrix such as horse urine extract.