Correlations between periparturient serum concentrations of non-esterified fatty acids, beta-hydroxybutyric acid, bilirubin, and urea and the occurrence of clinical and subclinical postpartum bovine endometritis

BACKGROUND: Postpartum endometritis in cattle is a multifactorial disease with high economic impact. Both, clinical endometritis (CE) and subclinical endometritis (SCE) result in decreased reproductive performance. Results from in vitro studies led to the implication that non-esterified fatty acids (NEFA), beta-hydroxybutyric acid (BHBA), bilirubin, and urea could be used as predictors for endometritis in veterinary practice. In this field study, we set out to establish optimal predictor cut points of these metabolic parameters for the detection of CE and SCE. Serum samples were collected one week prior to parturition (wk -1), in the first week postpartum (wk +1) and between 28 and 35 days postpartum (wk +5) from 209 Holstein-Friesian cows. At wk +5, all cows were examined for signs of CE and SCE. RESULTS: Higher concentrations of urea at wk +1 were associated with increased odds of CE (OR = 1.7, P = 0.04) in primiparous (PP) cows. A predictor cut point of 3.9 mmol/L (sensitivity: 61%, specificity: 70%) was determined. In multiparous (MP) cows, the logistic regression model revealed that higher concentrations of NEFA at wk -1 were associated with increased odds of CE and SCE (healthy vs. CE: OR = 9.1, P = 0.05; healthy vs. SCE: OR = 12.1, P = 0.04). A predictor cut point of 0.3 mmol/L (sensitivity: 38%, specificity: 87% and sensitivity: 35%, specificity: 89%, respectively) was determined. Increasing concentrations of urea at wk +5 were associated with decreased odds of CE (healthy vs. CE: OR = 0.6, P = 0.01; SCE vs. CE: OR = 0.5, P = 0.03). A predictor cut point of 3.8 mmol/L (sensitivity: 52%, specificity: 81%) was determined. For BHBA and bilirubin relationships with CE or SCE were not detected. CONCLUSIONS: The corresponding combinations of sensitivity and specificity of the determined predictor cut points were not satisfactory for practical use. Thus, the analysed parameters, i.e. NEFA, BHBA, bilirubin, and urea, at the chosen time points, i.e. at wk -1, at wk +1, and at wk +5 relative to calving, are unsatisfactory for disease prediction. Further research is required to clarify the questions raised by the current study