Over the course of a few short hours, as a dairy cow transitions from pregnancy to lactation, her demand for calcium (Ca) more than doubles to support colostrum and subsequent milk production. This leads to a drop in blood Ca concentration, often leaving the cow in a hypocalcemic state. Through the coordinated response of several hormones and tissues, normocalcemia can be resumed in several hours. Unfortunately, this is not the case for all cows, and some experience prolonged or delayed reductions in blood Ca beyond the first 24 h of lactation. While not always accompanied by clinical signs of disease, these delayed or prolonged reductions in blood Ca are associated with decreased production, increased risk of additional disease events, and early herd removal. Presently, the only method of identifying cows with delayed or prolonged hypocalcemia is by collecting a blood sample at 4 days in milk (DIM) and quantifying Ca in a laboratory setting. This method is laborious, cost prohibitive, and time consuming, often taking 24 h to analyze the collected sample, closing the appropriate time window for treatment. To identify cows suffering from prolonged and delayed reductions in blood Ca, alternative methods to blood collection must be identified. We hypothesize that patterns of rumination time before and after calving will provide a non-invasive method for detection of cows that develop delayed or persistent hypocalcemia. Our objective is to determine if patterns of rumination times in the periparturient period differ between cows that experience delayed or persistent hypocalcemia compared to cows that achieve normocalcemia by 4 DIM.

Within the 1-year funding period, we specifically aim to:
1) Determine if reductions in prepartum rumination time are indicative of Ca status at 4 DIM.
2) Describe the association between the rate of change in rumination time across the first 4 days after calving and the resumption of normocalcemia by 4 days in milk.

All aims will be completed via retrospective analysis of data previously collected from multiparous Holstein cows (n = 453) from 2 herds in Cayuga County, New York. Daily rumination times (min/d) were collected from 14 days prior to calving to 21 days after parturition, and blood samples were collected and analyzed for Ca at 4 DIM. Our first aim will be achieved by analyzing prepartum rumination time and blood Ca status at 4 DIM. Cows will be classified as having high or low blood Ca at 4 DIM using a diagnostic cut point of 2.2 mmol/L. We will create linear regression models to analyze the association between prepartum rumination time and blood Ca status at 4 DIM. Our second aim will be completed using rumination data collected during the first 4 DIM and blood Ca at 4 DIM. The rate and magnitude of rumination time change in the days following calving will be calculated for all cows used in Aim 1. We will determine the association between the rate and magnitude of rumination time change and blood Ca status at 4 DIM using support vector machine analysis, a machine learning tool.