Accuracy of fetal age estimates using transrectal ultrasonography for predicting calving dates in dairy cows in seasonally calving herds in New Zealand
AIM: To describe the accuracy of transrectal ultrasonography for predicting calving dates in dairy cows under typical New Zealand conditions and to assess potential risk factors for differences between predicted and actual calving dates.
METHODS: Data were collected from 116 seasonally calving herds over 2 years in a retrospective single cohort study. Transrectal ultrasonography was undertaken by experienced veterinarians (n=12) to determine if cows were pregnant, and if so to estimate fetal age. Predicted calving date was calculated by adding 282 days to the estimated conception date. Accuracy was assessed using differences between predicted and actual calving dates for each animal. Potential risk factors for animals calving >10 days before or after their predicted calving date were assessed using multinomial logistic regression models.
RESULTS: The study population comprised 83,104 cows over the 2 years of the study; 75,037 (90.3%) cows calved within 10 days of their predicted calving date, 3,683 (4.4%) calved >10 days earlier, and 4,384 (5.3%) >10 days later, than predicted. Risk factors for calving >10 days before or after the predicted calving date included having >1 artificial insemination (AI) before pregnancy diagnosis (p=0.03), where the cow’s most recent AI was <21 days before the end of the herd’s AI period (p<0.01), and where the diagnosis was made at the second or third herd-visit (p<0.01). The probability of calving being >10 days later than predicted also increased when the fetus was ≥13 weeks old at pregnancy diagnosis (p<0.01).
CONCLUSIONS AND CLINICAL RELEVANCE: In this study, >90% of cows diagnosed pregnant by veterinarians using transrectal ultrasonography calved within 10 days of the predicted calving date. In herds where herd reproductive performance is high, it would be expected that more cows would conceive to their first AI, and potentially fewer cows would have AI close to the end of the herd’s AI period, which would increase diagnostic accuracy. Where herd managers rely on accurate predicted calving dates they should be informed about realistic expected accuracy. For greatest accuracy, a complete AI history should be made available to the person performing the pregnancy diagnoses and cows at most risk of having inaccurate predicted calving dates should be identified.