Measure and characterization of lameness in gestating sows using force plate, kinematic, and accelerometer methods

The objective was to assess sows' lameness by measuring weight distribution on limbs using a force plate made up of 4 individual platforms each resting on 4 single-ended beam load cells. The weight was recorded at an average rate of 14 readings per s over a 15 min period. Ten sows (5 lame sows and 5 sound sows) were weighed twice on 2 different days to assess the repeatability of the measure. Sixty-one sows were then selected in 2 different sites and visually scored for lameness, using a 3-point scoring system (0=normal gait; 1=abnormal gait, and/or stiffness; and 2=shortened stride, and/or the sow puts less weight or avoids putting weight on 1 leg). Various measures were recorded from each sow using the force plate (percentage of weight, the ratio between the weights applied by contralateral legs, weight shifting, and amplitude of weight bearing and weight removing), kinematics (speed, stride length, swing time, stance time, foot height, and carpal and tarsal joints angle average and amplitude), and accelerometers (time spent standing during 24 h, frequency of stepping behavior during feeding, and latency to lie down after feed delivery). The within-sow CV for each measure taken from the force plate were lower than 15%, which reflects a good repeatability. Among force plate measures, only the weight shifting frequency and the ratio between the weights applied by contralateral legs differed between lameness scores. Sows that scored 2 had a higher frequency of weight shifting for fore legs (P=0.0003) and hind legs (P=0.0007) than sows scored 0 and 1. The ratio between the weights applied by contralateral legs decreased with the increase of lameness score for the hind limbs (P=0.014). However, these measures also differed between sites (P<0.01). These differences may be due to various reasons, including but not limited to genetics and housing systems. Nevertheless, the results suggest that force plate measures such as the asymmetry in the weight applied between a pair of legs and weight shifting are good indicators of lameness. Multivariate analysis on fore and hind legs showed independency between variables related to animals in movement (measures from kinematics) and variables related to static animals (measures from the force plate and accelerometers). Therefore, both static and dynamic methods need to be used to detect various lame sows.