Cocaine is one of the most widespread illegal stimulants utilized by the human population throughout the world. The aim of this study was to establish the highest no-effect dose (HNED) of cocaine on the spontaneous locomotor activity (SLA) of horses in a behavior chamber, and thereby to determine the maximal acceptable threshold of the urinary drug concentration in horses. Twelve English thoroughbred mares received 0.02, 0.03, 0.04, 0.08 or 0.12 mg kg(-1) cocaine i.v. or saline solution (control). It was noted that doses above 0.04 mg kg(-1) induced a significant increase in SLA (P < 0.05, Tukey's test). No significant increase in SLA was seen in the mares that received 0.03 mg kg(-1), but the animals showed important behavioral changes that did not occur after the 0.02 mg kg(-1) dose. It was concluded that the HNED of cocaine for horses in a behavior chamber is 0.02 mg kg(-1). After injection of this dose in five horses, urine samples were collected at predetermined intervals through vesical catheterization. The concentrations of cocaine, norcocaine, benzoylecgonine and ecgonine methyl ester were quantified by liquid chromatography/electrospray ionization tandem mass spectrometry. Cocaine and norcocaine concentrations remained consistently below the level of detection. Benzoylecgonine reached a mean (+/- SEM) maximum concentration of 531.9 +/- 168.7 ng ml(-1) after 4 h, whereas ecgonine methyl ester peaked 2 h after injection at a concentration of 97.2 +/- 26.5 ng ml(-1). The maximum admissible concentration for cocaine and/or metabolites in the urine of horses is difficult to establish unequivocally because of the substantial individual variation in the drug elimination pattern observed in horses, which can be inferred by the large standard error of the means obtained.