Short communication: Assessing the accuracy of inline milk fat-to-protein ratio data as an indicator of hyperketonemia in dairy cows in herds with automated milking systems.

The objective of this study was to evaluate the accuracy of inline milk fat-to-protein (F:P) data to detect hyperketonemia (HYK) in herds with automated milking systems (AMS). The F:P ratio has been investigated as a tool for detecting HYK with moderate accuracy in past studies, but inline F:P data in AMS may also be useful for HYK screening. To assess the accuracy of these data in commercial settings, we monitored 484 cows from 9 AMS herds for their first 3 wk of lactation, taking blood samples once per week (n = 1,427). Positive cases of HYK were defined by whole-blood β-hydroxybutyrate (BHB) concentrations ≥1.2 or ≥1.4 mmol/L. Milk data were collected from the AMS software on each farm for each cow and converted into 4 different F:P values: (1) value from the same day as the BHB test; (2) 5-d centered-moving average (CMA); (3) 5-d backward-moving average (BMA); (4) 5-d forward-moving average (FMA). In linear regression models, all 4 values were associated with BHB, but slope estimates varied and R were low: same day (slope = 0.95, R = 0.07), CMA (slope = 1.05, R = 0.07), BMA (slope = 0.65, R = 0.04), and FMA (slope = 1.23, R = 0.09). In logistic regression models, the odds of having HYK (BHB ≥1.2 mmol/L) increased with every 0.1-unit increase from the mean F:P ratio (1.16) using same-day values (odds ratio = 1.35, 95% confidence interval = 1.25-1.47) and CMA (odds ratio = 1.39, 95% confidence interval = 1.27-1.51). The same increase in F:P from mean BMA (1.14) and FMA (1.17) was associated with 1.22 and 1.49 times the odds of HYK, respectively. For all 4 F:P variations, we evaluated the sensitivity, specificity, positive predictive value, and negative predictive value of different F:P thresholds with HYK status. As the F:P threshold increased from 1.17 to 1.50, sensitivity decreased (range: 77 to 9%) but specificity increased (range: 58 to 96%). Same-day and CMA F:P cutoffs at which a balance was reached between sensitivity and specificity ranged from 1.18 to 1.22; however, even at these values we found high rates of false positives and negatives (range: 31-39%). These results suggest that inline milk F:P data from inconsistently calibrated sensors should not be used alone to detect HYK in AMS herds.