Patterns of periparturient rumination and activity time in multiparous Holstein cows with and without dyscalcemia in early lactation.

Dyscalcemia, defined as reduced blood Ca at 4 DIM, is associated with reduced milk production and reproduction and an increased risk of negative health events. Cowside testing of blood Ca to diagnose dyscalcemia is difficult, and alternative methods to identify dyscalcemia are needed. Our objectives were to explore differences in periparturient rumination and activity time between cows with and without dyscalcemia and use activity and rumination variables to identify dyscalcemia. We performed a retrospective cohort analysis on data collected from multiparous Holstein cows (n = 182) from 2 herds in New York. Cows were affixed with ear or neck loggers to record daily activity (arbitrary units [AU]/d, defined by manufacturer) and rumination (min/d) times. Daily activity and rumination times were collected from 14 d before calving until 14 DIM. No cows received supplemental calcium or experienced clinical hypocalcemia during the study period. A blood sample was collected at 4 DIM and analyzed for total calcium concentration, and cows were subsequently classified as dyscalcemic if total calcium was ≤2.2 mmol/L (n = 57) or eucalcemic if total calcium was >2.2 mmol/L (n = 125). Linear mixed models were used to analyze differences in pre- and postpartum activity and rumination times between the calcemic groups. Logistic regression models were used to identify the probability of dyscalcemia from activity and rumination time variables from 0 to 4 DIM. Prepartum activity time was similar between eucalcemic and dyscalcemic cows (402.0 ± 10.4 AU/d and 395.1 ± 14.5 AU/d, respectively). Postpartum eucalcemic cows had greater activity time than dyscalcemic cows (436.1 ± 10.7 vs. 407.8 ± 14.7 AU/d, respectively). Prepartum rumination time was similar between eucalcemic and dyscalcemic cows (512.6 ± 9.8 min/d vs. 504.2 ± 14 min/d, respectively). Postpartum eucalcemic cows had greater rumination time than dyscalcemic cows (512.3 ± 10.5 min/d vs. 480.5 ± 15 min/d, respectively). Logistic regression models yielded AUC values ranging from 0.71 to 0.79, sensitivities of 17.5% to 40.3%, specificities of 91.2% to 94.4%, accuracy of 70.3% to 77.0%, positive predictive values of 59.0 to 76.0%, and negative predictive values of 72.0% to 78.0%. Our findings suggest that differences exist in postpartum activity and rumination times between cows that experience dyscalcemia and those that remain eucalcemic. Utilizing activity and rumination time data in the immediate postpartum period shows utility in identifying cows with dyscalcemia, which could aid in management decisions that ameliorate the associated negative outcomes; however, further work is needed to optimize their capabilities.