Behavioural indicators of pain in horses undergoing surgical castration

Finding reliable pain indicators is paramount for equine welfare and little is available on post castration behaviour modifications in the horse. We aimed to investigate reliability and changes over time of behavioural pain indicators shown by horses undergoing routine castration. Eight stallions of different breeds, 2 to 4 years old, were castrated with closed technique in general anaesthesia. The subjects were placed in an observation box for 2 days before and 3 days after intervention. Their behaviour was video-recorded from a distance for 15min at each of the following intervals: before surgery, 4, 8, 16, 24 and 40 hours after. Two blind observers, using Solomon software, analysed duration and frequency of pain related behaviours. Observers significantly agreed (Kendall’s coefficient of concordance for k related samples from a continuous field, W<0.05) assessing some behaviours (agitation, reluctance to move, kicking the abdomen, lethargy, lowered head carriage, flank watching, rolling, attention and curiosity) but not others (apathetic glance, abnormal walking). Compared to basal, 8h after castration, horses showed significantly more (Wilcoxon test) agitation, reluctance to move, lowered head carriage and flank watching (P<0.05). A discriminant analysis was performed on concordant behaviours and 100% of the observations were correctly classified by the model as basal or 8h. 16 hours after intervention the behaviour of horses was comparable to basal condition. We conclude that inter-observer reliability of assessors evaluating pain behaviour cannot be assumed and should always be evaluated. Altered behaviours occurred predominantly 8h post castration, suggesting that this time is critical for pain evaluation and treatment.