Validation of mechanical, electrical and thermal nociceptive stimulation methods in horses

To validate a model for investigating the effects of analgesic drugs on mechanical, thermal and electrical stimulation testing. To investigate repeatability, sensitivity and specificity of nociceptive tests. Randomised experiment with 2 observers in 2 phases. Mechanical (M), thermal (TL) and electrical (E) stimuli were applied to the dorsal metacarpus (M-left and TL-right) and coronary band of the left thoracic limb (E) and a thoracic thermal stimulus (TT) was applied caudal to the withers in 8 horses (405 ± 43 kg). Stimuli intensities were increased until a clear avoidance response was detected without exceeding 20 N (M), 60°C (TL and TT) and 15 V (E). For each set of tests, 3 real stimuli and one sham stimulus were applied (32 per animal) using a blinded, randomised, crossover design repeated after 6 months. A distribution frequency and, for each stimulus, Chi-square and McNemar tests compared both the proportion of positive responses detected by 2 observers and the 2 study phases. The κ coefficients estimated interobserver agreement in determining endpoints. Sensitivity (384 tests) and specificity (128 tests) were evaluated for each nociceptive stimulus to assess the evaluators' accuracy in detecting real and sham stimuli. Nociceptive thresholds were 3.1 ± 2 N (M), 8.1 ± 3.8 V (E), 51.4 ± 5.5°C (TL) and 55.2 ± 5.3°C (TT). The level of agreement after all tests, M, E, TL and TT, was 90, 100, 84, 98 and 75%, respectively. Sensitivity was 89, 100, 89, 98 and 70% and specificity 92, 97, 88, 91 and 94%, respectively. The high interobserver agreement, sensitivity and specificity suggest that M, E and TL tests are valid for pain studies in horses and are suitable tools for investigating antinociceptive effects of analgesics in horses.