Prepartum Blood Biochemistry and Periparturient Disease Risk in Dairy Cows: A Retrospective Two-Farm Study

Periparturient diseases are a major cause of economic loss and impaired welfare in dairy cattle. We screened 29 prepartum biochemistry and hematology parameters for association with retained placenta (RP), dystocia, abortion/stillbirth, ketosis, and milk fever in 419 calving events from two Korean dairy farms (DF1, n = 248; DF2, n = 171). Each calving was sampled twice prepartum on alternating biweekly schedules, resolved into an early (week -4/-3) and late (week -2/-1) window. Parameters were compared between affected and unaffected calvings using linear models adjusted for parity, body condition score (BCS), and farm, with Benjamini-Hochberg false discovery rate (FDR) correction, and were then entered into multivariable logistic models. After adjustment and FDR correction, two markers remain associated with disease: lower prepartum calcium for milk fever and elevated direct bilirubin with reduced cholesterol and triglycerides for ketosis, both detectable in both windows three to four weeks before calving. Multivariable models discriminated milk fever (AUC = 0.81) and ketosis (AUC = 0.75), with comparable discrimination when models were applied to the other farm. A prepartum bilirubin signal for RP, consistent with an emerging bilirubin-RP association reported via plasma metabolomics, was detectable but did not survive confounder adjustment and is reported as hypothesis-generating. Using only routine clinical biochemistry and a practical biweekly schedule, these parameters were associated with subsequent milk fever and ketosis weeks before calving, identifying them as candidate markers that require prospective validation.