Gait screening using a mobile phone application : horse asymmetry levels and usability of a mobile phone application for gait analysis, during a long format FEI eventing competition

Abstract: Orthopaedic disorders are the most common reason for unplanned rest in the performing horse population. Locomotor disorders which slowly build up to clear lameness might be avoidable with easy access objective motion measurement devices. An early detection of a disorder affecting the locomotor apparatus will lead to an early diagnosis and directed treatment to avoid chronic, irreversible damages to the tissues, often caused by long drawn inflammation. During the last 30 years, biomechanical researchers have established and refined sensitive objective measures of lameness. Decreased loading of a painful front- or hind limb leads to a vertical motion asymmetry of the head or pelvis respectively at the trot. These asymmetries can be detected with high precision and accuracy, but the relevance of small amplitude asymmetries is not fully elucidated and clear thresholds for where an asymmetry is likely to reflect a painful condition are lacking. Several studies have investigated the prevalence of motion asymmetries in different horse populations, revealing that asymmetries above clinically used thresholds are very common, ranging from 73-90% in mature riding horses, youngsters, endurance horses, standardbred trotters and polo horses. These high prevalence numbers and the sensitive measurement devices available have raised the question whether these asymmetries are caused by pain or if it is a question of biological variation of the gait pattern. In the search for the relevance of mild and moderate motion asymmetries, the relation to a high performing locomotor system is of particular interest. Therefore, studying horses that are judges to be “fit to compete” for high level events could provide insight. The aim of this study was to describe vertical movement symmetry of head and pelvis in eventing horses as they trotted up during the horse inspection before and during an international FEI long format eventing competition. The asymmetry was described for horses that passed and did not pass the inspection. We investigated how the horses responded to the strenuous cross-country section of the competition by studying if the symmetry changed from the first to the second horse inspection (before and after cross country). In addition to the description of asymmetries on a group level, we also evaluated the usability of the mobile phone gait analysis tool, Sleip AI, which was used during the study. In total, 58 horses were recorded using a computer vision tool for gait analysis at the horse inspections at one FEI long format eventing competition. At the first horse inspection, 36% of the horses showed asymmetry of mild type or more, on one leg or more. In spite of the high exercise load during the cross country, 52% of the 44 horses presented at the second horse inspection (after the cross country) improved their total asymmetry, 36% had a worsening/increased total asymmetry and 11% remained unchanged in their total asymmetry level. No significant difference on a population level was found (p =0.77) using a paired non-parametric test (Wilcoxon signed rank test), when looking at all the groups, first compared to the second horse inspection. The level of measured movement asymmetry using the gait analysis app was higher in the horses which were given a nonpass (sent to holding) compared to the ones that were subjectively judged to be fit to compete and were given a “pass”, but this was not statistically tested due to the low number of horses sent to holding. Horses that got a straight away “pass” at the first horse inspection (n =53) had a median asymmetry score of 0.95 for total asymmetry. The runs that generated a “holding” decision (n =5), had the median score of 2.1 for total asymmetry at the first horse inspection. At the second horse inspection the “pass” (n =42) had a median score of 1.05 and the “holding” (n =2) a median score of 1.75. Some horses with high degree of asymmetry slipped through the “pass”. Objective gait analysis can be useful in order to pick these horses out for a second control. We observed that many of the outliers with a high degree of asymmetry were white/grey-coloured horses. When looking at the total asymmetry and the outliers in the “pass”-group at the first and second horse inspection 3/5 and 3/3 horses were white coloured. This might indicate white horses are visually more difficult to assess. A mean of 11.8 trot strides were recorded for the front limbs during the first horse inspection and 7.4 trot strides for the hind limbs. During the second horse inspection a mean of 12.2 trot strides were measured for the front limbs and 8.5 for the hind limbs. In order to render a more reliable interpretation of the gait analysis, more trot steps should be recorded. This could be made possible if the trot up was longer. To conclude, around one third of the eventers studied showed mild or more accentuated motion asymmetry at the day of the competition and surprisingly, the majority of this group of eventers improved their total motion asymmetry after the cross country. The mobile phone gait analysis application seemed to be a useful decision-tool for gait screening, but the horses should be trotted more strides in order to improve the reliability of the motion analysis.