[PDF] Predicting nutrient digestibility in high-producing dairy cows

View - Download Predicting nutrient digestibility in high-producing dairy cows

Previous PDF

Next PDF

1123J. Dairy Sci. 101:1123-1135

https://doi.org/10.3168/jds.2017-13344

© 2018, THE AUTHORS. Published by FASS Inc. and Elsevier Inc. on behalf of the American Dairy Science Association

This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/3.0/).

ABSTRACT

Our objective was to determine the effects of dry

matter intake (DMI), body weight (BW), and diet characteristics on total tract digestibilities of dry mat ter, neutral detergent fiber, and starch (DMD, NDFD, and StarchD, respectively) in high-producing dairy cows. Our database was composed of 1,942 observa tions from 662 cows in 54 studies from Michigan, Ohio, and Georgia. On average, cows ate 23 ± 4.5 kg of dry matter/d, weighed 669 ± 79 kg, and produced 38 ±

10 kg of milk/d. Diets were 31 ± 5% neutral deter

gent fiber, 27 ± 6% starch, 2.6 ± 1.2% fatty acids, and 17 ± 1.4% crude protein. Digestibility means were

66 ± 6, 42 ± 11, and 93 ± 5% for DMD, NDFD, and

StarchD, respectively. Forage sources included corn si lage, alfalfa, and grasses. Corn source was classified by its ruminal fermentability. Data were analyzed using a mixed effects model, including diet chemical composi tion, forage source, and corn source, all expressed as percentage of dry matter, except for DMI, which was expressed as percentage of BW (DMI%BW); location and 2-way interactions were fixed effects. Cow, block, period, treatment, and study were included as random effects. Best fitting candidate models were generated using backward and stepwise regression methods. Ad ditionally, the simplest model was generated using only DMI and location as fixed effects and all random effects. Candidate models were cross-validated across studies, and the resulting predictive correlation coef ficients across studies (PC) and root mean square error of prediction (RMSEP) were compared by t -test. For each nutrient, the digestibility model that resulted in the highest PC and lowest RMSEP was determined to be the best fitting model. We observed heterogeneous coefficients among the different locations, suggesting that specific location factors influenced digestibilities. The overall location-averaged best fitting prediction

equations were: DMD = 69 - 0.83 × DMI%BW (PC = 0.22, RMSEP = 5.39); NDFD = 53 + 0.26 × grass %DM - 0.59 × starch %DM + 3.06 × DMI%BW - 0.46 × DMI%BW2

(PC = 0.53, RMSEP = 9.70); and StarchD = 96 + 0.19 × HFERM%DM - 0.12 × starch %DM - 1.13 × DMI%BW (PC = 0.34, RMSEP =

4.77); where HFERM%DM is highly-fermentable corn

source as percentage of DM. Our results confirm that digestibility is reduced as DMI increases, albeit at a lower rate than that reported in National Research Council. Furthermore, dietary starch depresses NDFD.

Whereas DMD can be predicted based on DMI only,

the best predictions for NDFD and StarchD require diet characteristics in addition to DMI.

Key words:

model, intake, NDF, starchINTRODUCTION The primary goal of any nutrition program is to for mulate diets that meet the nutrient requirements of the animal. To design an effective program, data on both DMI and nutrient digestibility are needed, as a major factor influencing nutrient utilization in dairy cattle is the relationship between intake and digest ibility (Colucci et al., 1982). In lactating dairy cows, DMI is driven by milk production but can be limited by physical fill and metabolic effects (Allen, 2000). Increased DMI increases passage rate of digesta and, thus, depresses digestibility of nutrients (Tyrrell and Moe, 1975). However, a clear understanding of animal and dietary factors that influence intake, digestibility, and their interactions is still lacking.

The equation used by the NRC (2001) to estimate

digestibility is based on TDN content of the diet at maintenance and intake expressed as a multiple of main tenance. This equation may not adequately describe digestibility of high-producing dairy cows because the equation used by NRC (2001) was based primarily on data derived at lower levels of intake. Moreover, starch content is likely more important than TDN content in determining the digestibility discount (Ferraretto et al.,

2013).

Previous meta-analyses have focused on both dietary and animal characteristics to model digestibility in lac -Predicting nutrient digestibility in high-producing dairy cows

R. A. de Souza,* R. J. Tempelman,* M. S. Allen,* W. P. Weiss,† J. K. Bernard,‡ and M. J. VandeHaar*1

*Department of Animal Science, Michigan State University, East Lansing 48824 †Department of Animal Science, The Ohio State University, Wooster 44691 ‡Department of Animal & Dairy Science, University of Georgia, Tifton 31793

Received June 16, 2017.

Accepted September 26, 2017.

1

Corresponding author: mikevh@msu.edu

1124DE SOUZA ET AL.

Journal of Dairy Science Vol. 101 No. 2, 2018

tating dairy cows. Huhtanen et al. (2009) summarized the effects of feeding level and diet composition on di gestibility. They reported that TDN could be calculated for lactating cows using OM digestibility at a mainte nance level determined either in vivo in sheep or by us ing in vitro methods, DMI, CP content, and proportion of forage in the diet. Nousiainen et al. (2009) analyzed data from lactating cows and determined effects of for age quality, the proportion of concentrate, CP content, and fibrous by-products on OM and NDF digestibility. These findings suggest that inclusion of both dietary and animal factors were critical for accurate modeling of digestibility. However, these meta-analyses included only treatment means of the contributing factors avail able in the studies summarized and not data from indi vidual cows. In human medical research, meta-analyses are commonly performed using individual observations (Riley et al., 2010), but this is not common in animal science. Riley et al. (2010) summarized the benefits of using individual observation data over treatment means data; contextualizing the advantages cited by Riley et al. (2010) to the current study, when treat ment means are used the ability to quantify variability among animals within the same diet is reduced such that important information about the effect of DMI on digestibility is lost. The objective of our study was to develop equations to estimate total tract digestibility of DM ( DMD NDF ( NDFD ), and starch (

StarchD

) using recent data derived from individual observations of high- producing lactating cows. We hypothesized that diet, especially dietary starch, and animal characteristics, specifically DM as a percentage of BW, together would provide accurate predictions of nutrient digestibility.

MATERIALS AND METHODS

Our analysis was based on individual observations

of lactating Holstein dairy cows from 54 studies, 85% of which were conducted in the last 15 yr; a few of which have not yet been published. A complete list of the sources of these studies is provided in the Supple mentary Material ( https:// doi .org/ 10 .3168/ jds .2017 -13344 ). These st udies were conducted at Michigan

State University (

MSU ), The Ohio State University, and the University of Georgia. The final database con tained 1,942 observations from 662 cows on 195 differ ent treatments.

The variables included in the database were DMD,

NDFD, StarchD, BW, metabolic BW (

MBW , kg 0.75 milk yield ( MY ), DMI, dietary ingredients, and dietary chemical composition (Table 1). Additional information on studies included experimental design and treatment arrangement.Among the 54 studies, the methods for total-tract digestibilities determination included total fecal collec tion (n = 16 studies at Ohio State University; Weiss and Shockey, 1991), the use of chromic oxide as an external marker (n = 12 studies at MSU; Voelker and Allen, 2003), and the use of indigestible NDF (n = 19 studies at MSU; Cochran et al., 1986) and indigestible ADF (n = 7 studies at University of Georgia; Cochran et al., 1986) as internal markers. Feed fractions used in the analysis included NDF, nonforage NDF ( nfNDF forage NDF ( fNDF ), starch, total fatty acids ( FA), and CP, all expressed as a percent of DM. Values for these fractions were used as reported in the original paper, using standard methods. When neither nfNDF nor fNDF was reported in a publication, the nfNDF was estimated based on the concentrate composition using values from NRC (2001) and fNDF was deter mined by subtracting nfNDF from NDF. When the FA concentration was not reported, it was estimated based on ether extract [using FA = ether extract - 1, NRC (2001)] or on diet composition. The descriptions of ingredients and chemical composition of the diets are presented in Table 1. The forage sources in these studies were mainly corn silage, alfalfa hay and silage, wheat straw, orchardgrass silage, and ryegrass silage. The diets contained different amounts of each forage source in various combinations.

Forages were categorized as corn silage (

CS%DM ; in clusion of corn silage in the diet, expressed as a percent of DM), grass (

Grass%DM

; inclusion of wheat straw, orchardgrass silage, and ryegrass silage in the diet, ex pressed as a percent of DM), and alfalfa (

Alfalfa%DM

inclusion of alfalfa hay and silage in the diet, expressed as a percent of DM). Grass and alfalfa were classified by maturity as immature (grass: <55% NDF; alfalfa: <40% NDF), mid-mature (grass: 55-60% NDF; alfalfa:

40-46% NDF), and mature (grass: >60% NDF; alfalfa:

>46% NDF). The only exception for this classification was that wheat straw, in the Grass%DM category, was classified as mature independent of its NDF content. Whereas NDF content is not a perfect predictor of ma turity, it was the best predictor of maturity reported in all the studies used in our analysis. Additionally, the forages were grouped according to their type: silage or hay. Nonforage fiber sources ( NFFS ) included alfalfa meal, beet pulp, soy hulls, wheat middlings, whole cot tonseed, brewers wet grain, and distillers grain; they were all grouped as NFFS (inclusion of NFFS in the diet, expressed as s percent of DM).

Regarding the concentrate ingredients, corn was

categorized according to the ruminal fermentability of the source, such that high-moisture corn and steam- flaked corn were classified as highly fermentable

HFERM%DM

; inclusion of highly fermentable corn Journal of Dairy Science Vol. 101 No. 2, 2018PREDICTING NUTRIENT DIGESTIBILITY 1125

Table 1

. Mean, SD, and 95% CI of animal characteristics, total-tract digestibility, and diet ingredients and chemical composition based on 52 studiesVariablen Mean SD95% CI

Michigan

Ohio

Georgia

Upper

Lowern Mean SD n Mean SD n Mean SD

Animal description

BW, kg1,942 669 79 830 513 1,362 676 79 416 624 65 164 636 75 Milk yield, kg/d1,882 38 10 58 19 1,320 40 10 399 36 8.1 163 34 8.0 DMI, kg/d1,942 23 4.5 31 12 1,362 24 4.7 416 22 4 164 23 3.6 DMI, %BW1,942 3.5 6.5 4.6 1.7 1,362 3.5 0.65 416 3.5 0.65 164 3.6 0.56

Total-tract digestibilit

y, %

DM1,942

66 6.4 77 55 1,362 66 5.8 416 66 3.2 164 62 6.5

NDF1,932

42 11 63 19 1,352 40 11 416 49 9.0 164 44 10

Starc h1,708 93 5.4 98 79 1,258 94 4.3 394 92 4.8 56 90 8.5

Ingredient composition, %DM

F orage Corn silage1,608 31 9.8 36 24 1,073 29 7.1 392 35 14 143 35 8.1 Grass 1

427 21 21 43 4.8 285 17 2147 33 23 95 24 17

Alfalfa

2

1,701 22 13 25 14 1,296 24 13 323 22 13 100 6.6 0.92

T otal forage1,942 50 7.6 56 43 1,362 48 7.6 416 54 7.3 164 49 3.7 NFFS 3

1,336 12 7.9 16 6.9 873 11 7.8 319 13 7.9 144 17 6.1

Grain 4

1,942 24 9.5 31 17 1,362 25 10 416 22 8.5 164 23 5.5

Protein supplement

5

1,934 13 4.8 17 9.1 1,362 14 3.9 408 11 5.5 164 7.3 3.2

Chemical composition, %DM

NDF1,942

31 5.2 33 27 1,362 30 4.5 416 34 5.6 164 35 3.6

Forage NDF1,942

22 3.6 24 19 1,362 22 3.3 416 23 4.4 164 23 3.2

Nonforage NDF1,942 9.0 4.6 11 5.5 1,362 8.0 4.4 416 11 4.9 164 11 2.9 Starc h1,942 27 5.9 30 24 1,362 27 6.2 416 26 5.0 164 26 3.6 Fatt y acid1,942 2.7 1.2 3.5 11 1,362 2.6 1.2 416 2.9 1.3 164 3.1 0.69

CP1,942

17 1.4 18 16 1,362 17 1.3 416 16 1.7 164 17 1.6

1 Sum of wheat straw, orchardgrass silage, and ryegrass silage. 2

Sum of alfalfa hay and silage.

3

Nonforage fiber source: sum of alfalfa meal, beet pulp, soy hulls, wheat middlings, whole cottonseed, brewers wet grain, and distillers grain.

4 Sum of dry corn, high-moisture corn, steam-flaked corn, and oats. 5

Sum of soy meal, soy plus, blood meal, and urea.

1126DE SOUZA ET AL.

Journal of Dairy Science Vol. 101 No. 2, 2018

in the diet, expressed as a percent of DM) and ground dry corn was classified as moderately fermentable

MFERM%DM

; inclusion of moderately fermentable corn in the diet, expressed as a percent of DM).

All statistical analyses were performed using SAS

v. 9.4 (SAS Institute Inc., Cary, NC). For numerical stability, all covariates (X) were rescaled according to Kutner et al. (2005) using the formula: X* = [X - mid point (X)]/[0.5 × range (X)]. All covariates were jointly checked for multicollinear ity and Pearson correlations between the covariates were computed. When covariates were highly corre lated or had variance inflation factors ( VIF ) greater than 10 (multicollinearity analysis), the covariate with lesser interest was removed from the analysis. For ex ample, if the MY and DMI had VIF greater than 10 the MY was removed because our focus was on DMI. This process was performed until all covariates in the model have VIF lower than 10. Subsequently, the full models containing the linear, quadratic, and cubic ef fects of all dietary variables, DMI, location, all possible

2-way interactions between dietary variables, DMI, and

location, cow, block, period, treatment, and study were created. The generic full model was 0 1 8 2 1 8 3 1 8 1 22
33
18 18 where Y lcbpts is the nutrient digestibility (DMD, NDFD, or StarchD); 0 is the intercept; Diet d are the diet covariates [ d = Grass%DM, Alfalfa%DM, CS%DM,

HFERM%DM, MFERM%DM, NDF (%DM), starch

(%DM), FA (%DM), and CP (%DM)], all expressed as percent of DM; d are the linear coefficients for the Diet d d are the quadratic coefficients for the Diet d d are the cubic coefficients for the Diet d 1 2 , and 3 are the linear, quadratic, and cubic coefficients for the intake; ? d is the coefficient for the interaction between Diet d and intake;

Location

l is the effect of location l l = Michigan, Ohio, and Georgia); dl is the linear coefficient on Diet d for

Location

l . All other terms were random effects, including Cow c for cow c (c = 1 to 662), Block b for block b b = 1 to 94, referring to the original design structure of the study),

Period

p for period p p

1 to 135), and

Treatment

t for treatment t t = 1 to 195) all nested within studies, Study s for study s (s = 1 to

52), with

quotesdbs_dbs46.pdfusesText_46

[PDF] Le roman de renard

[PDF] LE ROMAN DE RENART

[PDF] le roman de renart fiche de lecture 5ème

[PDF] le roman de renart question reponse

[PDF] le roman de renart questionnaire corrigé

[PDF] le roman de renart résumé 5eme

[PDF] le roman de renart résumé de chaque chapitre

[PDF] le roman de renart résumé par aventure

[PDF] le roman de renart résumé par chapitre wikipedia

[PDF] le roman de renart(la peche d ISENGRIN)

[PDF] le roman definition

[PDF] le roman est il le reflet de la société

[PDF] le roman est un miroir que l'on promène le long d'un chemin explication

[PDF] le roman est-il le reflet de la réalité dissertation

[PDF] le roman est-il le reflet de la société ?