Administration studies of levamisole in horses were carried out using two different levamisole preparations, namely, levamisole hydrochloride oral bolus and levamisole phosphate injectable solution. These preparations were analysed in detail for the presence of aminorex-like impurities. Both levamisole preparations were found to contain 1-(2-mercaptoethyl)-4-phenyl-2-imidazolidinone (I) and 4-phenyl-2-imidazolidinone (II) as degradation impurities, but neither aminorex nor rexamino was detected in these preparations. After the administration of these preparations to horses, aminorex, rexamino, in addition to levamisole and compound II, were detected in post-administration urine and plasma samples, among which compound II was found to have the longest detection time. Administration study of compound II was then performed on another horse to investigate whether it could be a metabolic precursor of aminorex and/or rexamino. However, no aminorex and rexamino was detected in the post-administration samples, suggesting that compound II was not a metabolic precursor of aminorex or rexamino. A metabolite (III) of compound II, tentatively identified to be a hydrolysis product of compound II, was observed instead. It has been established unequivocally that the normal use of levamisole products in horses can lead to the presence of aminorex, rexamino and 4-phenyl-2-imidazolidinone (II) in their urine and blood samples. As compound II has the longest detection time, the detection of aminorex (and in some cases rexamino) in some of the official samples from racehorses can be ascribed to the use of levamisole products as long as compound II is also present as a marker. These findings should be of direct relevance to the investigation of some of the cases of aminorex detection in official doping control samples from racehorses.