Kinematic changes in dairy cows with induced, unilateral forelimb lameness during straight line walk

Early detection of lameness in dairy herds is essential to enable timely treatment of affected animals, thereby avoiding unnecessary costs and animal suffering. Since claw diseases most commonly affect the hind claws, specific kinematic changes in cows with forelimb lameness have not been investigated. However, in-depth knowledge on movement pattern alterations occurring during lameness of varying sources is essential to develop efficient lameness detection tools. In this study, 27 gait analysis trials consisting of > 2 000 strides were collected from 12 clinically sound dairy cows. The cows were equipped with nine body-mounted inertial measurement units (IMUs) and contributed with one baseline trial and one or two lameness trials each. A lameness induction method causing increased claw pressure was used to introduce mild, reversible, unilateral forelimb lameness. From the IMU data, 31 limb-and upper body movement parameters, mainly focusing on motion symmetry, were computed for each stride. Baseline and lameness data were compared in linear mixed models, where between-cow variability was accounted for. Twenty-two movement parameters differed between the two conditions (P ≤ 0.05). Forelimb lameness caused a more pace-like walk; the relative time between hoof-on of both hindlimbs and their respective ipsilateral forelimb decreased by 0.022 and 0.036 (ratio of stride duration), while the relative time between hoof-on of the hindlimb contralateral to the lame forelimb, and the lame forelimb increased by 0.050. The maximum protraction angle of the lame forelimb increased by 1.5°, while the protraction angle of the non-lame forelimb, and the retraction angle of the lame forelimb, decreased by 1.7° and 3.0°. All hindlimb protraction and retraction angles, except the protraction angle of the hindlimb contralateral to the lame forelimb, decreased by 1.2°-1.4°. Following signal decomposition of upper body vertical motion, the largest changes were detected for the head and neck; there were notable increases (0.13 and 0.11) of their first harmonic (asymmetric component) amplitudes, and decreases (0.10 and 0.050) of their second harmonic (symmetric component) amplitudes (relative to the range of motion). Changes in the within-stride differences in the withers' position during limb spread and midstance (respectively) were also detected. The vertical range of motion per stride increased for the head, neck, and back, but decreased for the pelvis. Although the investigated parameters hence show promise as lameness indicators, the movement changes occurring with the induced fore claw pain should be confirmed in clinical lameness cases, to ensure usability of the described pattern for early, automated forelimb lameness detection.