This study aimed to investigate the associations among early postpartum estrous characteristics (EPEC) and reproductive outcomes in dairy cows fitted with automated monitoring devices (AMD). We hypothesized that EPEC within 41 DIM reflects cows' physiological resilience and serve as predictors of subsequent fertility. Furthermore, we trained and tested algorithms predicting the fertility potential of cows and compared the reproductive performance of cows classified as low, moderate, and high fertility. The study was conducted using data from 4,578 Holstein cows across 3 commercial dairy herds. Data regarding calving characteristics, postpartum health, milk yield, EPEC (estrus events, duration, rumination nadir, activity peak, and heat index), and the environment were collected. Reproductive outcomes of interest included pregnancy 77 ± 14 d after the first service, fertility class (high-fertility [HFERT] = pregnant to first service; low-fertility [LFERT] = nonpregnant after the first 3 services), and hazard of pregnancy up to 250 DIM. Using a training dataset, we developed an algorithm that was applied to the training and testing datasets to classify cows into bottom, moderate, and top fertility; the reproductive performance of these cows was then compared. Statistical analyses included logistic and Cox proportional hazard regressions. Overall, 38% of cows were pregnant 77 ± 14 d after first service, with a cumulative pregnancy of 67% by the third service. Occurrence of at least 1 estrus event within 41 DIM increased the odds of pregnancy to the first service (0 = referent, 1 = 1.20 [95% CI = 1.01-1.44], ≥2 = 1.16 [95% CI = 0.89-1.52]) and of a cow being HFERT (0 = referent, 1 = 1.26 [95% CI = 1.00-1.59], ≥2 = 1.09 [95% CI = 0.78-1.54]). However, more than 1 estrus event did not further improve fertility outcomes. The positive association between estrus within 41 DIM and hazard of pregnancy was reflected in shorter median days open (0 = 115.0 d, 1 = 94.0 d, ≥2 = 89.0 d). Lower rumination nadir was associated with increased odds of pregnancy to the first service and of a cow being HFERT. Similarly, cows in the lowest tertile of rumination nadir had the greatest hazard of pregnancy by 250 DIM. Cows in the testing dataset classified as bottom and top 25% percentiles of fertility had a 23-d difference in median days open (127 d vs. 104 d) and a 5.3-percentage point difference in censoring by 250 DIM (28.6% vs. 23.3%). This study demonstrates that EPEC are valuable indicators of reproductive potential in dairy cows. Integrating health, genetic, and environmental data with EPEC improves the prediction of fertility outcomes, providing opportunities to optimize reproductive management and efficiency in dairy herds. Finally, the results of our algorithms demonstrate the potential of EPEC to predict long-term reproductive performance of lactating Holstein cows.
Keywords: automation; estrus detection; fertility.
The Authors. Published by Elsevier Inc. on behalf of the American Dairy Science Association®. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).