Previous work (Marc et al., 2000) suggested that plasma cortisol responses to treadmill exercise or ACTH injection are a reliable marker for performance evaluation in warmblood horses. For practical purposes blood sample collections and treadmill exercise tests are somewhat troublesome and time consuming. The goal of this study was thus to evaluate the use of saliva for cortisol determination (by direct EIA) as a marker for performance and to investigate the reliability and repeatability of plasma cortisol responses to a single i. v. injection of ACTH (50 mug or 250 mug) Furthermore, the effect of training horses for 8 weeks 3 times per week covering the same distance (increasing from 3.5 km during the first week to 8 km during the last week) either by trotting (similar to 240 m/min) or by cantering (375 m/min) was investigated. For this purpose initially ten four-year-old Hannovarian geldings, all reared in the same State stud, were used. Mean overall correlation between salivary cortisol and plasma cortisol concentrations was 0.64 when samples of various points of time were used. However, in spite of attempts to standardize saliva sample collection, correlation between salivary cortisol levels and plasma cortisol levels at distinct points of time in different tests were low and significant (r = 0.85, p < 0.02) only in one test. Thus, salivary cortisol measurements for diagnostic purposes are not reliable or useful. The repeatability of plasma cortisol responses to ACTH for untrained and trained horses were r = 0.86 and r = 0.8 respectively (p <less than or equal to> 0.01 and p less than or equal to 0.05 respectively). Training horses either by trotting or cantering did not affect the cortisol response either to tread mill exercise or to stimulation by ACTH. It is concluded that the relationship between salivary cortisol levels and plasma cortisol levels is not close enough to allow the use of salivary cortisol determination as marker of the training status/fitness of horses. The repeatability of the cortisol response to ACTH is similar to the cortisol response to treadmill exercise. Based on plasma cortisol responses to ACTH or treadmill exercise training horses by cantering at low speed is not superior to training by trotting for the fitness of horses.
