| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
* Institute of Animal Science, Group of Animal Breeding, Swiss Federal Institute of Technology, CH-8092 Zurich, Switzerland
Division of Animal Nutrition and Physiology, Faculty of Veterinary Medicine, University of Berne, CH-3012 Berne, Switzerland
Institute of Animal Science, Research Station Chamau, Swiss Federal Institute of Technology, CH-6331 Huenenberg, Switzerland
Herbivore Research Unit, National Institute for Agricultural Research (INRA), F-63122 St-Genès-Champanelle, France
Corresponding author: J. W. Blum; e-mail:juerg.blum{at}itz.unibe.ch.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
100 cows if calving is highly seasonal and of 400 cows if calving is uniformly distributed. Using blood in addition to milk traits does not improve precision of estimation of herd-level EB, regardless of sample size.
Key Words: energy balance • dairy cow • estimation
Abbreviation key: AC = acetone, AP = absorbable protein, C30 = cows fed 30% of total DMI as concentrate, C50 = cows fed 50% of total DMI as concentrate, CF = crude fiber, CF-group = concentrate feeding group (C30 or C50), EB = energy balance, ECM = energy corrected milk, GH = growth hormone, GLDH = glutamate dehydrogenase, LDH = lactate dehydrogenase, p.p. = postpartum, T3 = 3,5,3'-triiodthyronine, T4 = thyroxine
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Various metabolic and endocrine blood and milk traits, such as NEFA, ketone bodies, insulin, IGF-1, thyroid hormones (Blum et al., 1985; Kunz et al., 1985; Ronge et al., 1988; Lucy et al., 1992; Gustafsson and Emanuelson, 1996), milk fat, protein, lactose, fat:protein ratio, and fat:lactose ratio (Macleod et al., 1984; Grieve et al., 1986; Steen et al., 1996; Heuer et al., 2000), have been shown to be related to EB. Automated blood and milk analysis and estimation of energy status with multivariate models might therefore have potential use in promoting health and productivity of dairy herds.
Based on such observations, the study aimed to estimate individual and herd-level EB using blood and milk traits in high-yielding dairy cows from wk 1 to 10 p.p. Heuer et al. (2000) reported that herd-mean EB can be adequately estimated with milk test-day information in large dairy herds, but that herd size limits the precision of prediction. In Switzerland, herds are rather small. For this reason it was tested, whether in small herds EB can be better estimated when—in addition to milk test-day information—various metabolic, enzymatic, and endocrine blood traits are incorporated into the models for EB estimation. Furthermore, the precision of prediction at herd-level EB with pooled herd samples was investigated.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Multiparous dairy cows (n = 90; 86 Holstein-Friesian and 4 Red Holstein) of parities 2 to 7 were studied from wk 2 antepartum to wk 20 p.p. at the research farm Chamau of the Swiss Federal Institute of Technology, Huenenberg, Switzerland. The herd-mean 305-d milk production was 9434 ± 1067 (mean ± SD) kg energy-corrected milk (ECM). The animals were housed in a free-stall barn and were held in free-stall groups of 12 cows. Each free-stall group was equipped with 12 mangers for forage and one automatic feeder for concentrate. Calving seasons were spring (April and May) and fall (November and December). Cows calved in four groups of equal size from fall 1998 to spring 2000.
The feeding plan is summarized in Table 1
. During the dry period, cows were fed a roughage mix containing 5 MJ NEL/kg DM. From wk 4 a.p. until calving, cows received the same roughage mix as p.p. containing 6 MJ NEL/kg DM and, additionally, concentrate in increasing quantities from 0 to 40% of the calculated amount p.p. After parturition, cows were fed a roughage mix ad libitum (6 MJ NEL/kg DM), which was freshly provided four times per day. Mangers for roughage were fixed on electronic balances. The concentrate was restricted provided by automatic feeders. Two groups (C30 and C50) were formed based on differences in the total amount of concentrate provided. From wk 1 to 10 p.p., C30 cows (n = 45) received concentrate in amounts corresponding to 30% of their individual DMI of the previous week and C50 cows (n = 45) in amounts corresponding to 50%, respectively. The two groups were balanced for parity, BCS a.p., BW a.p., calving season, and potential for milk production. The cows had access to roughage and concentrate during 24 h. Individual roughage and concentrate intake of each animal were continuously recorded over 24 h during the whole study period. Minerals and vitamins were fed according to calculated needs (Kessler, 1999), and sodium chloride was provided ad libitum.
|
. The roughage mix was composed of grass silage, whole plant corn silage, and grass hay and was kept at approximately 6 MJ NEL/kg DM. The concentrate contained 8 MJ NEL/kg DM, and its components were individually mixed according to lactation week and concentrate feeding group (C30 or C50) in order to attain the desired ratio between potentially absorbable protein (AP) and NEL in the total ration for each individual cow (Table 2). Feed was sampled on a weekly basis, and DM content was determined. Feed was analyzed for CP, crude fiber (CF), AP, and NEL at the Swiss Federal Research Station for Animal Production, Posieux, Switzerland, on a monthly basis (RAP, 1999).
|
Blood Traits
Blood samples were taken from the jugular vein between 0730 and 0900 using evacuated tubes containing dipotassium-EDTA (1.8 g/L) or no anticoagulant, from wk 1 to 10 p.p. on a weekly basis. Blood samples with the anticoagulant were immediately put on ice, whereas tubes without anticoagulant (for recovery of serum) were left at room temperature until clotting was finished (within 30 min). Tubes were then centrifuged for 20 min at 1500 x g. BHBA, sodium, potassium, chlorine, calcium, phosphorus, magnesium, lactate dehydrogenase (LDH), glutamate dehydrogenase (GLDH), and aspartate aminotransferase (AST) were determined in serum. Glucose, NEFA, cholesterol, creatinine, albumin, urea, insulin, IGF-1, growth hormone (GH), 3,5,3'-triiodothyronine (T3), and thyroxine (T4) were determined in plasma. The samples were stored at –20°C until analyzed at the Division of Animal Nutrition and Physiology, University of Berne, Switzerland, as described by Aeberhard et al. (2001b) and Bruckmaier et al. (1998). For the determination of plasma leptin concentrations in duplicates, the specific double-antibody radioimmunoassay described by Delavaud et al. (2000) was followed, with slight modifications. Instead of anti-ovine Ab 7137 leptin antiserum, Ab 8172 was used in a final dilution of 1:15'000. After addition of 125I-ovine leptin, the incubation continued for an additional 44 h instead of 20 h. As second antibody, goat anti-rabbit IgG was used instead of ram anti-rabbit IgG. It was diluted 1:50 in horse serum for standard curves, nonspecific binding tubes, and blanks and 1:50 in incubation buffer for the unknown plasma samples.
Milk Traits
Milk yield was measured twice daily. Time of milking was between 0430 and 0630 and between 1530 and 1730. Milk samples for determination of acetone (AC) were taken once per week (Tuesday morning) from wk 1 to 10 p.p., frozen immediately after milking and stored at –20°C until analyzed. Concentrations of AC were determined by flow injection analysis (Marstorp et al., 1983; Reist et al., 2000) at the Division of Animal Nutrition and Physiology, University of Berne, Switzerland. Milk samples for the determination of milk fat, protein, lactose, and urea were taken four times per week (Monday and Wednesday evening, Tuesday and Thursday morning) and analyzed at the laboratory of the Swiss Brown Cattle Breeders’ Federation, Zug, Switzerland. ECM was calculated as [(0.038 x g crude fat + 0.024 x g crude protein + 0.017 x g lactose) x kg milk] || 3.14. The 305-d standard lactation milk yields were calculated using the coefficients of the Holstein Association of Switzerland, Posieux, Switzerland.
Statistical Analysis and Model Development
Level of significance was set at P < 0.05 throughout the paper. S-PLUS 2000 Professional Release (MathSoft, 1999) statistical software was used. Mixed-effects models were fitted using the NLME 3.4 library (Pinheiro and Bates, 2000) for S-PLUS.
Incomplete test-day records and test-day records of cows under a medical or preventive treatment, which might have influenced metabolic or endocrine status, were omitted. Thus, 859 test-day records of 90 cows from wk 1 to 10 p.p. were available for calculations.
Descriptive statistical analysis revealed that a considerable number of traits was not normally distributed. Therefore, NEFA, creatinine, BHBA, leptin, insulin, IGF-1, GH, T4, LDH, GLDH, AST, milk AC, milk fat, milk protein, fat:protein ratio, and fat:lactose ratio were logarithmically transformed prior to analysis. Correlations between EB and the various blood and milk traits were calculated as Pearson’s correlations.
Linear mixed-effects models of the form yi = Xi ß + Zi bi + 
i, where yi is the response vector, ß is the vector of fixed effects, bi is the vector of random effects, Xi and Zi are the fixed effects and random effects regressor matrices, and i is the within-group error vector, were used to estimate EB considering blood and milk traits from wk 1 to 10 p.p. Models (n = 6) were fitted from varying basic models by backward elimination procedures, using a type I-error of P < 0.05. In each of these basic models, EB was the response variable and the animal was the repeatedly studied random effect, while the composition of the fixed effects, which are the physically measurable and recordable data to estimate EB, varied (Table 3). The models were fitted as follows:
|
3), and week of sampling at the categorical scale, and ECM and the different blood and milk traits at the continuous scale. 3) and week of sampling at the categorical scale, and ECM and milk traits at the continuous scale. Blood traits were not included in this model. To avoid colinearity, correlation structures of the fixed effects were examined prior to calculation. If the absolute value of the coefficient of correlation between two variables exceeded r = 0.5, the variable correlating stronger to EB was incorporated into the basic model and the other was omitted. The final models were crossvalidated and the residual plots of the final models were used as criterion for the model fit regarding homogenous variance assumption. To estimate precision of estimation of EB at the individual animal level, standard deviations (SD) of observed minus estimated EB (residuals) were calculated for the six final models.
Estimation of Herd-Level Energy Balance
Decreasing sample and herd size was simulated by random selection of cow test-days. In the first simulation, monthly test-days were randomly selected over the first 8 wk of lactation according to the current Swiss milk recording system in the field, resulting in data sets of 180 test-days. In the second simulation, one test-day per cow was randomly selected over the first 10 wk of lactation, resulting in data sets of 90 test-days. In the third simulation, one test-day of 40 randomly identified cows was randomly selected over the first 10 wk p.p., resulting in data sets of 40 test-days. Finally, in the fourth simulation, one test-day of 20 randomly identified cows was randomly selected over the first 10 wk p.p., resulting in data sets of 20 test-days. Each simulation was repeated 200 times, which was considered sufficient for stable SD, of the resulting herd means, according to Heuer et al. (2000). The precision of estimation of weekly EB with reduced data sets was expressed in 1 SD. These weekly SD and the average number of cows in the sample based on the 200 runs were computed for each of the four simulations with each of the six final models.
Finally, precision of estimation of mean overall EB from wk 1 to 10 p.p. was computed as described in the precedent section. Pooling of the samples was simulated in order to minimize laboratory expenditures with respect to a possible practical application. By pooling the individual samples of a herd, cow-specific information, e.g., lactation week and CF-group, got lost. Thus, only models 3, 4, and 6, which do not include the fixed-effects week of sampling, CF-group, and ECM, were considered for this part. Again, simulations were repeated 200 times. The precision of prediction of mean overall EB from wk 1 to 10 p.p. with reduced data sets was expressed in 1 SD of the difference between mean EB estimated with all data sets and mean EB predicted with the reduced data sets.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
. Apart from leptin, each of the determined metabolic, endocrine, and enzymatic blood traits correlated significantly with EB, as well as the determined milk components. Concentrations of NEFA correlated strongest (r = –0.685) with EB, followed by concentrations of glucose (r = 0.457), BHBA (r = –0.451), and T4 (r = 0.418). In milk, the order of correlation was fat:lactose ratio (r = –0.589), followed by fat (r = –0.565), fat:protein ratio (r = –0.496), and AC (r = –0.410).
|
(Models 1 to 6, respectively). The regression coefficients, 95% confidence intervals, P values, and R2 are shown. Residual plots of the six final models are presented in Figure 1, together with the appropriate SD.
|
|
|
|
|
|
|
) were highest in the most complete model (Model 1; Table 5), which included the fixed effects CF-group, sampling week, ECM, and all significant blood and milk traits, namely plasma NEFA, creatinine, T4, and the fat:lactose ratio in milk. Model 2 was derived from Model 1 by stepwise elimination of blood and milk traits, leaving only the three most informative ones: NEFA and T4 in plasma and the fat:lactose ratio in milk. In Model 2, R2 of the fixed effects accounted for 0.706 (Table 6) and 1 SD of observed minus estimated EB for 13.09 MJ NEL (Figure 1), respectively. Model 3 included as fixed effects blood and milk traits only, i.e., information about CF-group, sampling week, and ECM was omitted. EB was estimated best by plasma NEFA, cholesterol, albumin, and T4, the fat:lactose ratio in milk, and milk urea. The R2 of the fixed effects was 0.536 (Table 7), and precision of estimation was 16.52 MJ NEL (Figure 1). Model 4 was derived from Model 3 by stepwise elimination of blood and milk traits, leaving only the three most informative ones. These were the same as in Model 2, namely plasma NEFA and T4 and the fat:lactose ratio in milk. The R2 of the fixed effects accounted for 0.524 (Table 8) and 1 SD of the residuals for 16.70 MJ NEL (Figure 1), respectively. Model 5 included the fixed effects CF-group, sampling week, ECM, the fat:lactose ratio, and milk AC, but no blood traits. The R2 of the fixed effects was 0.695 (Table 9), and precision of estimation was 13.39 MJ NEL (Figure 1). Model 6 included milk protein, urea, and AC and the fat:lactose ratio as fixed effects. The R2 of the fixed effects accounted for 0.330 (Table 10) and 1 SD of the residuals for 19.76 MJ NEL (Figure 1).
Observed and estimated weekly EB with the six final Models and the complete test-days (n = 859) from wk 1 to 10 are presented in Table 11. For the reduced data sets simulating smaller herd sizes, average number of cows in the samples at the individual weeks are specified. Furthermore, precision of predicted herd-mean EB at decreasing sample (n) and herd size is shown.
|
. For each of the six final models, observed weekly EB from wk 1 to 10 p.p. is plotted against the observed EB minus 1 SD of the residual EB obtained with the reduced data sets. To correct for the differences between the six models with respect to the deviance between observed and estimated EB at particular weeks, the absolute value of this deviance (weekly residual mean) was further subtracted.
|
. Furthermore, SD of mean EB predicted with pooled samples of reduced data sets are depicted for each of these models. Again, a decrease in sample size or simulated herd size resulted in a loss in precision of prediction.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Correlation of Energy Balance with Blood and Milk Traits
Most coefficients of Pearson’s correlations between the various blood and milk traits and EB early p.p. were in the well-known directions and reflected the well-orchestrated endocrine changes and metabolic adaptations in early lactation, i.e., enhanced mobilization of depot fat, skeletal muscle breakdown and favored partitioning of absorbed nutrients to the mammary gland to provide sufficient substrates for milk synthesis (Krebs, 1966; Blum et al., 1983; Kunz et al., 1985; Bauman et al., 1988; Ronge et al., 1988; Chilliard et al., 1998; Chilliard, 1999). The negative Pearson’s correlation of EB to milk protein was probably due to an effect of stage of lactation, i.e., a concomitant increase in EB and decrease in milk protein concentration as the colostrum effect disappeared and, therefore, was not contradictory to the positive regression coefficient of milk protein concentration in the multivariate mixed-effects model estimating EB. The same effect explains the negative Pearson’s correlation of plasma albumin to EB, while the regression coefficient of albumin in the multivariate mixed-effects model was positive.
Model Development and Estimation of Individual and Herd-Level EB
Different models (n = 6) for estimation of EB using blood and milk traits were developed. In a first step, a best multivariate model with respect to precision of estimation of individual EB and R2 of the fixed effects was developed (Model 1). Further models with a reduced number of fixed effects and, partly, taking into account milk traits only, were developed with respect to practical applicability. Correlation of the fixed effects was examined prior to calculation in order to avoid colinearity. Whenever the absolute value of the coefficient of correlation between two variables exceeded r = 0.5, the variable correlating stronger with EB was incorporated into the basic model and the other was omitted. As a result, T4 displaced T3 and the fat:lactose ratio in milk displaced the fat:protein ratio and milk fat; although using the fat:lactose ratio instead of the fat:protein ratio for estimation of EB is unusual. Nevertheless, Steen et al. (1996) demonstrated in Norwegian dairy herds, that high fat:lactose ratios were closely associated with elevated ketone body status and, thus, provided evidence that they are connected with the energy status.
In Model 1, the regression coefficients of the fixed effects reflect metabolic adaptation after parturition. Typically, energy status improved considerably with lactation week (Heuer et al., 2000). The T4 was positively and NEFA, creatinine, ECM, and fat:lactose ratio were negatively associated with EB, which was more negative in C30 compared with C50 cows. The measure of determination of the complete model was high. The fixed effects explained most of the variability of EB, while the cow as random effect accounted only for a small R2. Nevertheless, precision of estimation of individual EB indicated in 1 SD of observed minus estimated EB was rather low. In Model 2, stepwise elimination of blood and milk traits (creatinine) leaving the three most informative ones (NEFA, T4, and fat:lactose ratio) resulted in the same high R2 of the complete model, and only in minor decreases in R2 of the fixed effects and in the precision of estimation of individual EB, respectively. For Model 1 as well as for Model 2, EB estimated by use of the complete data set (n = 859) was very close to the observed EB. However, decreased sample and herd sizes considerably decreased precision of prediction of herd-level EB, especially at the onset of metabolic adaptation after parturition in wk 1 p.p. A decrease in variance of EB from wk 1 to 10 p.p. and a more stable metabolic state toward wk 10 p.p. might have resulted in an increased precision of prediction.
Models 3 and 4 aimed to estimate EB by using blood and milk traits as the only fixed effects, i.e., information about CF-group, sampling week, and ECM was omitted. In Model 3, NEFA, albumin, fat:lactose ratio, and milk urea were negatively, and cholesterol and T4 were positively associated with EB, reflecting metabolic adaptation after parturition. Pearson’s correlation between milk urea and EB was positive, reflecting an increase of urea concomitant with DMI. However, a decrease in energy supply results in decreased ammonia fixation by microbes in the rumen as well as an increased AA catabolism in the liver and, consequently, leads to increased urea synthesis (Oltner and Wiktorsson, 1983), as evidenced by the negative association between EB and milk urea. The remaining three most informative traits in Model 4 were the same as in Model 2. In Models 3 and 4, the R2 of the complete models were again relatively high. However, the fixed effects, which are the physically measurable and recordable data to estimate EB, explained considerably less of the variability of EB, while the cow as random effect accounted for a greater part of variance of EB compared with Models 1 and 2. Mainly by elimination of the fixed-effect sampling week, EB could be less precisely estimated and the mixed model attributed some of the weekly variation of EB to the cow, explaining why the cow accounted for a greater part of variance of EB in Models 3 and 4. Hence, precision of estimation of individual EB was considerably lower in Models 3 and 4 than in Models 1 and 2. Mean estimated EB with Models 3 and 4 using the complete data set (n = 859) deviated considerably from mean-observed EB, especially in wk 2 and 4 p.p. and, again, toward wk 10 p.p. Again, these deviations from the observed means, which were stronger for Model 4 than for Model 3, were due to the elimination of the fixed-effect week in Models 3 and 4 compared with Models 1 and 2. In Models 1 and 2, the categorical variable sampling week corrected for the curvilinear trend of EB in the course of lactation, while Models 3 and 4 contained exclusively variables at the continuous scale and, therefore, failed to estimate mean EB at each individual week, even when considering the complete data set (n = 859). Once more, precision of prediction decreased considerably with a decrease in sample and herd size, and the increase in precision of EB from wk 1 to wk 10 p.p. can be explained with a decrease in variance of EB over this period due to a more stable metabolic state toward wk 10 p.p.
In Models 5 and 6, no blood traits were used to estimate EB. Model 5 comprised CF-group, sampling week, and ECM and was therefore similar to Models 1 and 2, while Model 6 comprised milk components and milk AC and, thus, was similar to Models 3 and 4. The regression coefficients of CF-group and ECM were in the same range as in Models 1 and 2. The fat:lactose ratio and milk AC were negatively associated with EB and reflected physiologic processes, e.g., mobilization of body fat, in early lactation (Steen et al., 1996; Aeberhard et al., 2001a). The fixed-effect milk corrected for the excursive progression of EB. Therefore, EB estimated using the complete data set (n = 859) with Model 5 was very close to the observed EB at each of the studied weeks, which is similar in Models 1 and 2. The R2 of the complete model and of the fixed effects was barely lower than those of the models containing blood traits in addition, and precision of estimation of EB at the individual animal level was in the same range as in Models 1 and 2. The EB at decreasing-herd and sample sizes was also very similar to Models 1 and 2. Therefore, use of blood traits in addition to milk traits does not result in higher precision of prediction of individual and herd-level EB, regardless of sample size. In Model 6, negative associations of milk protein, urea, AC, and fat:lactose ratio with EB reflected the physiological processes in early lactation, as discussed in the previous sections. The R2 of the complete Model and of the fixed effects were lower than in Models 1 to 5, but the cow accounted for considerably more of the variance than in Models 1 to 5. Precision of estimation of individual EB was lowest compared with the other five models. For the same reasons as in Models 3 and 4, estimated EB using the complete data set (n = 859) deviated considerably from mean observed EB at several weeks. Apart from the deviation of estimated from observed EB, especially in wk 2 p.p., precision of prediction of EB at herd level at reduced sample and herd size was as good for Model 6 as for any of the other models.
Prediction of mean overall herd-level EB over the first 10 wk of lactation by pooling samples was precise with each of the applicable models for this purpose (Models 3, 4, and 6), even with reduced data sets and smaller herd size. However, samples were selected randomly from wk 1 to 10 p.p. and were therefore uniformly distributed across this period. Accumulation of calvings either at the beginning or at the end of the sampling period would result in a marked loss in precision of prediction. Therefore, distribution of sampling has to be considered. Furthermore, no information about EB at a particular week can be obtained when pooling the samples. However, estimation of occurrence and degree of EB nadir might be of particular interest for taking management decisions.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
100 cows if calving is highly seasonal and of 400 cows if calving is uniformly distributed around the year. Estimation of EB at the individual animal level is not precise enough to allow practical application, neither by use of milk traits alone, nor by use of milk and blood traits. Estimation of overall mean herd-level EB over the first 10 wk of lactation using pooled samples is precise. However, information on herd-level EB in particular weeks is lost when samples are pooled. Remarkably, the use of blood traits in addition to milk traits does not result in higher precision of prediction of individual and herd-level EB, regardless of sample size.
|
ACKNOWLEDGEMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
|
FOOTNOTES |
|---|
2 Accepted as part of the Ph.D.-thesis of M. Reist by the Swiss Federal Institute of Technology, Zurich, Switzerland, September 2001.
Received for publication November 27, 2001. Accepted for publication March 11, 2002.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  M. van Straten, N. Y. Shpigel, and M. Friger Analysis of Daily Body Weight of High-Producing Dairy Cows in the First One Hundred Twenty Days of Lactation and Associations with Ovarian Inactivity J Dairy Sci, September 1, 2008; 91(9): 3353 - 3362. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A.-K. Nyman, U. Emanuelson, K. Holtenius, K. L. Ingvartsen, T. Larsen, and K. P. Waller Metabolites and Immune Variables Associated with Somatic Cell Counts of Primiparous Dairy Cows J Dairy Sci, August 1, 2008; 91(8): 2996 - 3009. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. Oikonomou, G. E. Valergakis, G. Arsenos, N. Roubies, and G. Banos Genetic Profile of Body Energy and Blood Metabolic Traits Across Lactation in Primiparous Holstein Cows J Dairy Sci, July 1, 2008; 91(7): 2814 - 2822. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. R. G. Wylie, S. Woods, A. F. Carson, and M. McCoy Periprandial Changes in Metabolite and Metabolic Hormone Concentrations in High-Genetic-Merit Dairy Heifers and Their Relationship to Energy Balance in Early Lactation J Dairy Sci, February 1, 2008; 91(2): 577 - 586. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Guinard-Flament, E. Delamaire, P. Lamberton, and J. L. Peyraud Adaptations of Mammary Uptake and Nutrient Use to Once-Daily Milking and Feed Restriction in Dairy Cows J Dairy Sci, November 1, 2007; 90(11): 5062 - 5072. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. Rezamand, T. A. Hoagland, K. M. Moyes, L. K. Silbart, and S. M. Andrew Energy Status, Lipid-Soluble Vitamins, and Acute Phase Proteins in Periparturient Holstein and Jersey Dairy Cows With or Without Subclinical Mastitis J Dairy Sci, November 1, 2007; 90(11): 5097 - 5107. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Hayhurst, M. K. Sorensen, M. D. Royal, and P. Lovendahl Metabolic Regulation in Danish Bull Calves and the Relationship to the Fertility of Their Female Offspring J Dairy Sci, August 1, 2007; 90(8): 3909 - 3916. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. P. W. de Roos, H. J. C. M. van den Bijgaart, J. Horlyk, and G. de Jong Screening for Subclinical Ketosis in Dairy Cattle by Fourier Transform Infrared Spectrometry J Dairy Sci, April 1, 2007; 90(4): 1761 - 1766. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Patton, D. A. Kenny, S. McNamara, J. F. Mee, F. P. O'Mara, M. G. Diskin, and J. J. Murphy Relationships Among Milk Production, Energy Balance, Plasma Analytes, and Reproduction in Holstein-Friesian Cows J Dairy Sci, February 1, 2007; 90(2): 649 - 658. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. Miglior, A. Sewalem, J. Jamrozik, D. M. Lefebvre, and R. K. Moore Analysis of Milk Urea Nitrogen and Lactose and Their Effect on Longevity in Canadian Dairy Cattle J Dairy Sci, December 1, 2006; 89(12): 4886 - 4894. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. R. Roche, J. M. Lee, and D. P. Berry Pre-conception energy balance and secondary sex ratio--partial support for the Trivers-Willard hypothesis in dairy cows. J Dairy Sci, June 1, 2006; 89(6): 2119 - 2125. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
U. J. Schroder and R. Staufenbiel Invited Review: Methods to Determine Body Fat Reserves in the Dairy Cow with Special Regard to Ultrasonographic Measurement of Backfat Thickness J Dairy Sci, January 1, 2006; 89(1): 1 - 14. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Roesch, M. G. Doherr, and J. W. Blum Performance of Dairy Cows on Swiss Farms with Organic and Integrated Production J Dairy Sci, July 1, 2005; 88(7): 2462 - 2475. [Abstract] [Full Text] [PDF]
|
|
–– = PM1 minus 1 SD (n = 180); 
–· = PM1 minus 1 SD (n = 40); –
– = PM1 minus 1 SD (n = 20).