Department of Animal Science
ORCID IDs
D. L. Morris https://orcid.org/0000-0001-6347-2804
J. L. Firkins https://orcid.org/0000-0003-3518-1940
C. Lee https://orcid.org/0000-0002-4522-1232
W. P. Weiss https://orcid.org/0000-0003-3506-4672
P. J. Kononoff https://orcid.org/0000-0001-6069-2174
Date of this Version
2021
Citation
J. Dairy Sci. 104:6727–6738
https://doi.org/10.3168/jds.2020-19684
Abstract
Measurement of urinary energy (UE) excretion is essential to determine metabolizable energy (ME) supply. Our objectives were to evaluate the accuracy of using urinary N (UN) or C (UC) to estimate UE and ultimately improve the accuracy of estimating ME. Individual animal data (n = 433) were used from 11 studies with Jersey cows at the University of Nebraska–Lincoln, where samples were analyzed after drying (n = 299) or on an as-is basis (n = 134). Dried samples resulted in greater estimated error variance compared with as-is samples, and thus only as-is samples were used for final models. The as-is data set included a range (min to max) in dry matter intake (11.6–24.6 kg/d), N intake (282–642 g/d), UE excretion (1,390–3,160 kcal/d), UN excretion (85–220 g/d or 20.6–59.5% of N intake), and UC excretion (130–273 g/d). As indicated by a bias in residuals between observed and predicted ME as dietary crude protein (CP; range of 14.9–19.1%) increased, the National Research Council dairy model did not accurately predict ME of diets, as dietary CP varied. The relationship between UE (kcal/d) and UN (g/d) excretion was linear and had an intercept of 880 ± 140 kcal. Because an intercept of 880 is biologically unlikely, the intercept was forced through 0, resulting in linear and quadratic relationships. The regressions of UE (kcal/d) on UN (g/d) excretion were UE = 14.6 ± 0.32 × UN, and UE = 20.9 ± 1.0 × UN − 0.0357 ± 0.0056 × UN2. In the quadratic regression, UE increased, but at a diminishing rate as UN excretion increased. As UC increased, UE linearly and quadratically increased. However, error variance was greater for regression with UC compared with UN as explanatory variables (8.42 vs. 7.42% of mean UE). The use of the quadratic regression between UN and UE excretion to predict ME resulted in a slope bias in ME predictions as dietary CP increased. The linear regression between UE and UN excretion removed slope bias between predicted ME and CP, and thus may be more appropriate for predicting UE across a wider range of dietary CP. Using equations to predict UE from UN should improve our ability to predict diet ME in Jersey cows compared with calculating ME directly from digestible energy.
Comments
© 2021 American Dairy Science Association®. Published by Elsevier Inc. and Fass Inc