PSII-4 - The prediction of the onset of estrus using post-weaning changes in sow body temperature and posture

Abstract: There is interest in using changes in body temperature and/or behavior in sows as predictors of the onset of estrus. The objective of this study was to determine the accuracy of predicting the onset of estrus using post-weaning changes in either sow body temperature and/or posture. The study was carried out over a 10-d period following weaning involving 86 multiparous sows housed in individual stalls that were checked daily for standing reflex from day 3 to 7 post-weaning. Vulva surface temperature, measured with a thermal camera, was collected daily at 0615 (feeding time), 0700, 0730 (heat checking time), 0900, 1100, 1300, 1500, 1700, 1900, 2100, and 2300. Video recording of the sows was carried out continuously; the posture of each sow (lying, standing, or sitting) was recorded every 5 min. The percentage of observations that sows were active (i.e., sitting + standing) was calculated. Logistic regression analysis was used to develop models to predict the onset of estrus, using the Glmer function of R. Separate analyses were carried out for each of the 3 days that sows first exhibited standing estrus (day 4, 5, and 6 post-weaning). The data used was for the average percentage of observations of active postures and vulva surface temperature for each day starting from 2 days after weaning up to the first day of standing estrus for each sow. The dataset was randomly divided into 2 populations: training dataset (70% of sows), used to develop the prediction equations; testing dataset (30% of sows), used to validate the prediction equations. The dependent variable was if the sow was in estrus or not. The independent variables were changes in daily means for active postures (relative to day 2 post-weaning), changes in daily means for vulva temperature (between consecutive days), and the day post-weaning. There were differences in model accuracy statistics between the different models for each of the 3 days sows exhibited standing estrus (Table 1). Prediction equations based only on changes in posture generally had improved model accuracy statistics compared to those based only on changes in vulva temperature. Equations based only on day post-weaning generally had improved model accuracy statistics than those based on changes in either active postures or vulva temperature for sows first showing estrus on days 5 and 6 but not for those first showing estrus on day 4 post-weaning. Including changes in active postures and vulva temperature with day post-weaning resulted in little to no change in model accuracy statistics compared to the equations based only on day post-weaning (Table 1). These results suggest that changes in posture or body surface temperature, either individually or in combination, were not as accurate as using day post-weaning to predict the onset of estrus.