Objective gait monitoring is increasingly accessible to trainers. A more comprehensive understanding of 'normal' gait adaptations is required. Forty two-year-old thoroughbred racehorses were recruited when entering training and followed for 22 months. Gait analysis was performed by equipping each horse with an inertial measurement unit with inbuilt GPS (GPS-IMU) mounted on the dorsum. Horses were exercised as per their regular training regimen. Data were analysed using a linear mixed model. For two-year-old horses, there was a non-linear pattern of stride duration (SD) over time (p < 0.001) with SD decreasing initially and then 'flattening off' over time (linear and quadratic coefficients -0.29 ms/week and 0.006 ms/week2). Horses showed an increase in SD of 2.21 ms (p < 0.001) per 100 m galloped, and over time, SD decreased by 0.04 ms (p < 0.001) with each 100 m galloped per week. Three-year-old horses overall showed no change in SD over time (p = 0.52), but those that had a period of time off showed a decrease in SD of -0.59 ms per week (p = 0.02). They showed an increase in SD of 1.99 ms (p < 0.001) per 100 m galloped, and horses that had a period of time off showed an increase in stride duration of 1.05 ms per 100 m galloped (p = 0.01) compared to horses which did not have time off. Horses demonstrate an adaptation to high-speed exercise over time. SD decreases with training when other factors are controlled for in naïve horses.
Keywords: adaptation; biomechanics; gait analysis; horse; training.