Abstract #M50
Section: Animal Health (posters)
Session: Animal Health Posters 1
Format: Poster
Day/Time: Monday 7:30 AM–9:30 AM
Location: Exhibit Hall A
Session: Animal Health Posters 1
Format: Poster
Day/Time: Monday 7:30 AM–9:30 AM
Location: Exhibit Hall A
# M50
Haptoglobin critical thresholds for predicting health disorders during the transition period.
A. L. Kerwin*1, D. V. Nydam2, W. S. Burhans3, S. K. Wall4, K. M. Schoenberg4, K. L. Perfield4, T. R. Overton1, 1Department of Animal Science, Cornell University, Ithaca, NY, 2Department of Population Medicine and Diagnostic Sciences, College of Veterinary Medicine, Cornell University, Ithaca, NY, 3Dairy-Tech Group, South Albany, VT, 4Elanco US Inc, Greenfield, IN.
Key Words: haptoglobin, metabolic health, transition cow
Haptoglobin critical thresholds for predicting health disorders during the transition period.
A. L. Kerwin*1, D. V. Nydam2, W. S. Burhans3, S. K. Wall4, K. M. Schoenberg4, K. L. Perfield4, T. R. Overton1, 1Department of Animal Science, Cornell University, Ithaca, NY, 2Department of Population Medicine and Diagnostic Sciences, College of Veterinary Medicine, Cornell University, Ithaca, NY, 3Dairy-Tech Group, South Albany, VT, 4Elanco US Inc, Greenfield, IN.
The objective was to 1) establish cow-level thresholds for haptoglobin (Hp) concentrations to predict health disorders and 2) to investigate the magnitude of association between elevated Hp and health disorders within 30 DIM. Data were collected from 72 commercial dairy farms and Hp and β-hydroxybutyrate (BHB) concentrations were measured in 988 Holstein cows, 0 to 12 DIM. Associations between the health disorder of interest and Hp were assessed in SAS v. 9.4, PROC GENMOD, using a Poisson distribution, log link function, p-scale option, exchangeable correlation matrix, herd as a random effect and included biologically plausible covariates and 2-way interactions. Manual backward stepwise elimination removed covariates and interactions if P > 0.10. If Hp was significant, an ROC analysis with PROC LOGISTIC was used to determine the critical threshold for predicting the health disorder by identifying the point on the curve with the maximum Youden’s index. Based on the threshold, Hp was categorized and the multivariable model was reassessed to determine risk ratios. Haptoglobin was not effective for predicting herd reported mastitis, DA, or ketosis (P > 0.10). The Hp threshold for predicting herd reported metritis (MET) was 1.52 g/L (AUC = 0.8; Se: 67%, Sp: 84%), for predicting herd removal (CULL) was 0.68 g/L (AUC = 0.7; Se: 72%, Sp: 68%), and for association with hyperketonemia (HYK; BHB ≥1.0 mmol/L) was 0.74 g/L (AUC = 0.6; Se: 45%, Sp: 72%). The Hp threshold for predicting any significant health disorder (DIS; MET, CULL, HYK) was 0.74 g/L (AUC = 0.6; Se: 45%, Sp: 74%). Risk ratios for variables remaining in the health disorder models are reported in the table. Elevated Hp was found to be associated with HYK, MET, and herd removal within 30 DIM.
Table 1 (Abstr. M50). Risk ratios (RR) for variables associated with health disorders
1Multiparous vs. primiparous.
2Lame vs. not lame.
Disorder | Variable | Threshold | P-value | RR | 95% CI |
MET | |||||
Hp (g/L) | 1.52 | <0.01 | 8.6 | 3.0 to 24.5 | |
HYK | |||||
Hp (g/L) | 0.74 | <0.01 | 1.8 | 1.4 to 2.3 | |
Parity1 | <0.01 | 2.1 | 1.5 to 2.8 | ||
CULL | |||||
Hp (g/L) | 0.68 | <0.01 | 4.9 | 2.4 to 9.9 | |
Lameness2 | <0.01 | 3.5 | 1.6 to 7.6 | ||
DIS | |||||
Hp (g/L) | 0.74 | <0.01 | 1.8 | 1.5 to 2.3 | |
Parity1 | <0.01 | 1.6 | 1.2 to 2.3 | ||
Lameness2 | 0.04 | 1.3 | 1.0 to 1.6 |
Key Words: haptoglobin, metabolic health, transition cow