Evaluation of classification modes potentially suitable to identify metabolic stress in healthy dairy cows during the peripartal period S. Hachenberg, C. Weinkauf, S. Hiss and H. Sauerwein J Anim Sci published online Apr 27, 2007;
The online version of this article, along with updated information and services, is located on the World Wide Web at: http://jas.fass.org
www.asas.org
Downloaded from jas.fass.org by on May 14, 2011.
Page 1 of 31
Journal of Animal Science
Identification of metabolic stress in healthy cows
1 2 3
Evaluation of classification modes potentially suitable to identify metabolic stress in
4
healthy dairy cows during the peripartal period1
5 S. Hachenberg,* C. Weinkauf,* S. Hiss*, and H. Sauerwein2*
6 7 8
Institute of Animal Science, Physiology & Hygiene Unit, University of Bonn, 53115
9
Germany
10 11
1
This study was supported by the North Rhine-Westphalian Ministry of Environment,
12
Conservation, Agriculture and Consumers Protection (MUNLV) within the Inter-
13
Departmental Center of Sustainable Agriculture (USL). C. Weinkauf was supported through
14
the Deutsche Bundesstiftung Umwelt (DBU). The initiative for the experiment and the
15
subsequent administration and coordination of all works by Ute Müller is gratefully
16
acknowledged. We thank Birgit Mielenz, Karin Strack and Barbara Heitkönig for their
17
excellent technical assistance and Rupert Bruckmaier, at that time Institute of Physiology,
18
Technical University of Munich, Weihenstephan, Germany, for doing the IGF-I
19
measurements. Statistical advice from Dr. Rietz, Philosophical Faculty of Bonn University, is
20
gratefully acknowledged. Special thanks are given to the agricultural research center “Haus
21
Riswick” in Kleve, Germany (North Rhine-Westphalian Chamber for Agricultural Matters),
22
for providing the facilities for the animal experiment.
23 24
2
Corresponding author:
[email protected]
1 jas.fass.org by onas May 14, 2011. Published OnlineDownloaded First onfrom April 27, 2007 doi:10.2527/jas.2006-480
Journal of Animal Science
25
ABSTRACT: The transition of pregnancy to lactation with the concomitant negative energy
26
balance during early lactation requires substantial adaptive performance of the cow. Apart
27
from clinical disease problems, the identification of cows with suboptimal adaptation is
28
relevant in order to be able to adequately treat these animals or modify the ration. Effective
29
approaches are necessary to provide maximal information at the earliest time possible. We
30
therefore aimed to identify a measurement that, when applied at a defined point of time
31
relative to calving, comprised as much as possible other information on metabolic and health
32
status during early lactation. Blood samples were collected weekly from 4 wk ante partum to
33
12 wk post partum from 38 high yielding Holstein Frisian cows. Non-esterified fatty acids,
34
beta-hydroxybutyrate (BHB), IGF-I, and leptin were measured in serum and BCS was
35
recorded. Health status was characterized using the concentrations of haptoglobin (Hp) in
36
blood, the number of leukocytes and neutrophils (PMNL), as well as the activity of glutamate
37
dehydrogenase (GLDH) to evaluate liver impact. Using the factors related to fat mobilization,
38
the animals were classified according to their values recorded at one defined point of time or
39
time interval as being above or below certain thresholds. For each criterion, the groups
40
classified were compared with regard to the time course yielded from all recordings. From 7
41
criteria of classification, the most analogy with the variables of fat mobilization was obtained
42
when using NEFA and IGF-I (thresholds of 0.5 mM and 39 ng/mL in wk 1 post partum,
43
respectively). Both items were then combined to the criterion NEFA + IGF-I. Applying these
44
criteria, relations to indices of health and liver status were detectable on the basis of NEFA-
45
and NEFA + IGF-I-classes, which yielded differences in both GLDH and leukocyte numbers.
46
Animals with NEFA > 0.5 mM showed increased GLDH activity but decreased leukocyte
47
numbers. The time and effort required for measuring the IGF-I-concentration in addition to
48
NEFA is not justified for evaluating the metabolic status. Non-esterified fatty acid values 2 Downloaded from jas.fass.org by on May 14, 2011.
Page 2 of 31
Page 3 of 31
Journal of Animal Science
49
0.5 mM during the first wk of lactation were considered as the most suitable criterion for
50
identifying limited adaptive performance.
51 52
Key words: adaptation, cattle, fat mobilization, insulin-like growth factor-I, non-esterified
53
fatty acids, transition period
54 INTRODUCTION
55 56
Dairy cows need to mobilize adipose tissue to compensate for the relative deficit of
57
glucose occurring during the peripartal period due to the onset of lactation and the
58
concomitant negative energy balance. The rapid adaptation of key metabolic pathways is
59
central to successfully undergo the transition from pregnancy to lactation (Drackley et al.,
60
2001). Failure of adequate adaptation may result in imbalances leading to an increased risk of
61
digestive, metabolic, and/or infectious disorders (e.g., displaced abomasum, ketosis, or
62
mastitis; Ingvartsen, 2006).
63
Identification of animals having problems with adaptation is desirable and several
64
blood tests have been proposed for this (Jorritsma et al., 2003). However, most of the
65
investigations so far focused on finding “risk factors” related to various diseases. Differences
66
between well adapted animals and those with compromised adaptation might only be small,
67
and fixing a limit to identify the one or others is more difficult. Nevertheless, the detection of
68
cows with impaired capability of adaptation is important to treat individual cows adequately
69
during the peripartal period.
70
To evaluate the information yielded when classifying dairy cows according to
71
different criteria of fat mobilization in early lactation, blood samples of apparently healthy,
72
high yielding dairy cows were collected during the peripartal period. Threshold values from
73
defined times in relation to parturition were applied following mainly references from the
74
literature to retrospectively classify the cows showing values below or above these thresholds. 3 Downloaded from jas.fass.org by on May 14, 2011.
Journal of Animal Science
75
The 2 groups were then compared with regard to the entire time course of the other variables
76
to test which classification would yield the most analogy. Thereby an approach which
77
comprises maximal information with minimal effort at the earliest time possible for
78
identifying cows with suboptimal adaptive performance, albeit not clinically diseased, should
79
emerge.
80 MATERIALS AND METHODS
81 82
Animals and Diets
83
Holstein Frisian cows (n = 38), average age 3.9 ± 1.6 yr, were studied 4 wk ante
84
partum (a.p.) until 12 wk post partum (p.p.). They were housed in free stall barns in 2
85
different units (organic, n = 21; and conventional, n = 17) of the agricultural research center
86
from the North Rhine Westphalian chamber for agriculture matters (Riswick, Kleve,
87
Germany). We intended to integrate cows of comparable nutritional status and performance
88
regardless of how their feed was produced (i.e., grown according to conventional or organic
89
farming practice). Based on the metabolite and hormone profiles recorded in blood samples as
90
described below, no differences were detectable between the 2 groups. All cows were fed a
91
total mixed ration based on grass and corn silage for ad libitum intake; amounts offered and
92
refused were recorded daily to maintain approximately 10% (5 to 15%) orts. On average,
93
nutrient supply was similar in both groups and is shown in Table 1. During the dry period,
94
cows were fed a ration which satisfied the demand for maintenance and the additional
95
energetic needs of the conceptus and the mammary gland (i.e., 13 MJ NEL/day from wk 6 to
96
wk 4 a.p. and 18 MJ NEL/day from wk 3 a.p. until calving, respectively) according to the
97
recommendations of the German Society of Nutrition Physiology (GfE, 2001). Body
98
condition was scored every 2 wk during the entire experiment as described by Edmonson et
99
al. (1989) using 0.25 increments on the 1 to 5 scaling; milk yield and milk composition were
100
recorded every second wk by the regular milk recording organization (Landeskontrollverband 4 Downloaded from jas.fass.org by on May 14, 2011.
Page 4 of 31
Page 5 of 31
Journal of Animal Science
101
Nordrhein Westfalen e.V., Bischofstrasse 85, 47749 Krefeld, Germany). During the first 12
102
wk of lactation, daily milk yield was 39.2 ± 1.14 kg energy corrected milk.
103 104
Blood Collection and Laboratory Analyses
105
Blood samples were collected weekly via jugular venipuncture from 4 wk a.p. to 12
106
wk p.p., approximately 4 h after the morning feeding. For total blood leukocyte counts,
107
EDTA blood was used; for differential leukocyte counts, blood smears were prepared and
108
analyzed with the ADVIA 120-hematology-system (VLK; Laboratory for Veterinary
109
Medicine, Cologne, Germany). For all other measurements, serum was prepared by
110
centrifugation (20 min, 3000 x g, 4°C) of clotted blood and stored at -20°C until analysis. The
111
concentrations of haptoglobin (Hp) and leptin were measured by using double antibody
112
enzyme immunoassays as described by Hiss et al. (2004) and Sauerwein et al. (2004),
113
respectively. For progesterone (P4), the enzyme immunoassay protocol of Sauerwein et al.
114
(2006), which is based on the technique described by Prakash et al. (1987), was used. The
115
minimal detectable dose in the assay for Hp, leptin and P4 were 0.07 µg/mL, 0.3 ng/mL and
116
0.1 ng/ml, respectively. The intra- and inter-assay CV were 2.9 and 5.8% for Hp, 3.6 and
117
7.8% for leptin, and 9.3 and 17.2% for P4. The concentrations of IGF-I were measured by
118
RIA as described by Daxenberger et al. (1998). The sensitivity of the RIA was 1 ng/mL and
119
the intra- and inter-assay CV were 5.1 and 13.4 %, respectively.
120
Serum concentrations of NEFA were measured using an enzymatic test kit (Roche,
121
Mannheim, Germany; kit Nr. 1 383 175) in which the volumes were reduced for the use on
122
microtiter plates. Beta-hydroxybutyrate and GLDH were measured at VLK using kits from
123
Randox, Antrim, Ireland (kit-Nr. RB 1008) and from Roche (kit Nr. 197734), respectively.
124 125
Classification of the Animals According to Thresholds Set for BCS, for Blood Metabolites,
126
and for Hormones Related to the Mobilization of Body Fat 5 Downloaded from jas.fass.org by on May 14, 2011.
Journal of Animal Science
Page 6 of 31
127
The animals were retrospectively grouped according to the results from BCS and the
128
blood concentrations of metabolites and hormones related to fat mobilization to either being
129
below or above threshold values defined as explained below. Considering the metabolic and
130
physiological backgrounds, being below the threshold was classified as beneficial in case of
131
NEFA and BHB and the groups were termed using the suffix “-ok” (e.g., “NEFA-ok”).
132
Animals with values above the threshold were accordingly judged as being potentially at risk
133
for health impairments and the groups were termed using the suffix “-ltd” for limited adaptive
134
capability (e.g., “BHB-ltd”). For IGF-I, leptin, and BCS, the approach was vice versa (i.e.,
135
values below the threshold were judged as negative and the suffixes were accordingly used).
136
For the BHB threshold, the values recorded during wk 2 p.p. were used, because the peripartal
137
increase is well detectable at this time (Doepel et al., 2002). A concentration of 1200 mM
138
described previously as appropriate to identify fresh cows with subclinical ketosis (Enjalbert
139
et al., 2001) was used (i.e., cows with BHB concentrations
140
BHB-ltd and those with concentrations < 1200 mM as BHB-ok animals).
1200 mM were classified as
141
No defined threshold for plasma NEFA exists for postpartum cows (Nafikov et al.,
142
2006). For prepartum cows, a NEFA threshold of 0.5 mM was defined; cows exceeding this
143
threshold would have greater likelihood for developing postpartum disorders (Oetzel, 2004;
144
LeBlanc et al., 2005). We herein used the NEFA concentrations measured in wk 1 p.p.,
145
because the peripartal increase was then obvious occurring approximately 1 wk earlier than
146
for BHB (Doepel et al., 2002). To establish a threshold, we used a Receiver-Operating-
147
Characteristic curve [sensitivity vs. (1 – specificity); Perkins & Schistermann, 2006] in this
148
case in which the aforementioned BHB threshold provided the basis for the classification. The
149
cut-off value (0.5 mM NEFA) was defined at maximal values of both specificity (75%) and
150
sensitivity (79%) using maximum Youden-Index (sensitivity + specifity – 1). The area under
151
the curve was 0.786.
6 Downloaded from jas.fass.org by on May 14, 2011.
Page 7 of 31
Journal of Animal Science
152
For IGF-I, 2 different modes of setting a threshold were tested. First, the median of
153
IGF-I-concentrations (39.0 ng/mL) during the first wk p.p. was used to define the groups IGF-
154
I-ltd (< 39.0 ng/mL) versus IGF-I-ok ( 39.0 ng/mL). Second, the magnitude of the decrease
155
in the concentrations from wk 1 a.p. to wk 1 p.p. was considered; the ratio of the p.p./a.p.
156
values was calculated, expressed as percentage and the median of these percentages was used
157
to form the classes IGF-I-%-ldt (
158
IGF-I, 1 cow had to be excluded due to abnormally high values that could not be adequately
159
measured by diluting the sample, indicating matrix effects.
67.9) and IGF-I-%-ok (< 67.9). For all calculations using
160
To establish threshold values for leptin, an analogous approach as taken for IGF-I was
161
used. That is, we used the median recorded during wk 1 p.p. to classify the groups, leptin-ok
162
( 4.1 ng/ml) and leptin-ltd (< 4.1 ng/ml). In addition, we considered the percentage decrease
163
from wk 1 a.p. to wk 1 p.p. and used the median (71.4%) as threshold for the respective
164
Leptin-%-groups.
165
For BCS, again 2 different thresholds were tested: 1 used the categorization of the
166
animals according to their BCS during the last 2 wk a.p. using a threshold of 3.5 according to
167
Busato et al. (2002) and Lopez-Gatius et al. (2003), (i.e., classifying BCS-ltd:
168
ok: < 3.5). In addition, BCS losses ( -BCS) of more than 0.5 from wk -2 to -1 a.p. until wk 3
169
to 4 after calving were applied for a second threshold which yielded the groups -BCS-ltd:
170
0.5; -BCS-ok: < 0.5.
3.5 and BCS-
171 172
Statistical Analyses
173
All statistical analyses were performed with the SPSS program version 12.0 (SPSS
174
GmbH Software, Munich, Germany). For the groups classified at a defined point of time or
175
time interval according to the thresholds described above, the various blood concentrations as
176
dependent variable were compared between each of the “-ok” and the “-ltd” groups for the
177
entire time course recorded via the general linear model. Given that the classification was 7 Downloaded from jas.fass.org by on May 14, 2011.
Journal of Animal Science
178
done at a certain time and we considered the entire time course for the variables, we also
179
included analyses to compare the curves of variables which were used as criterion of
180
classification (e.g., the concentrations of NEFA during the study were compared also between
181
the NEFA-ok and the NEFA-ltd groups).
182
classification and also parturition, the latter being considered via nested periods (before and
183
after parturition) within sampling weeks. Sampling week was considered as repeated effect.
184
Three covariance structures (first order ante dependence, first order auto regressive, and
185
heterogeneous first order auto regressive) were initially tested and the 1 with the lowest
186
values of the Akaike’s information criterion was then applied.
Fixed effects tested were the respective
187
To evaluate the relation between NEFA and BHB as well as NEFA and IGF-I
188
concentrations, nonparametric coefficients of correlation (Spearmen) were calculated.
189
Variation in the data is expressed as SEM. Significance was declared at P < 0.05. In
190
consequence of the actual calving dates, the intervals of BCS-evaluation could not be met for
191
all cows. Eventually missing values were calculated by using the mean of the BCS score
192
before and after the interval without recordings.
193 RESULTS AND DISCUSSION
194 195
Fat Mobilization
196
The timely changes of the concentrations of NEFA, BHB, IGF-I, and leptin in serum
197
as well as BCS data recorded throughout the study, are shown in Figures 1 and 2,
198
respectively. For all variables recorded, effects of calving were observed (P < 0.001).
199
Mobilization of body fat associated with the negative energy balance p.p. could be recorded
200
using the variables described above. Elevated NEFA and BHB concentrations from fat
201
mobilization serving as energy source at a time in which intake of energy is limited (Pullen et
202
al., 1989; Bell, 1995) have been reported previously (Doepel et al., 2002; Pushpakumara et
203
al., 2003; Murondoti et al., 2004) and are related to a compromised metabolic status. The 8 Downloaded from jas.fass.org by on May 14, 2011.
Page 8 of 31
Page 9 of 31
Journal of Animal Science
204
maximal concentrations of NEFA (1.25 mM) during the first wk after parturition in our study
205
were less compared with the reports quoted above, but the temporal changes of the
206
concentrations of NEFA are in agreement with the works of Murondoti et al. (2004) and
207
Doepel et al. (2002), who reported maximal NEFA concentrations of 1.7 and 1.9 mM,
208
respectively, in wk 1 p.p.
209
The peripartal decrease of the concentrations of leptin we found is in line with the
210
observations of Block et al. (2001) and Chilliard et al. (2005). For IGF-I, the peripartal pattern
211
obtained in our study agrees with earlier works (Schams et al., 1991; Kim et al., 2004).
212
Cows were in proper body condition at the start of the study and the observed loss of
213
condition throughout the study can be judged as moderate (Lopez-Gatius et al., 2003). The
214
BCS at the time of calving observed herein corresponds well to the report by Kokkonen et al.
215
(2005) and showed no noticeable deflection from this during the first weeks of lactation. A
216
potential loss of information due to the every 2 wk scoring interval is considered as being of
217
minor importance in view of the relatively slow changes of body condition. In general, the
218
average mobilization of body fat, compared with other findings, was judged as moderate.
219 220
Health Status
221
Concentrations of Hp in serum, the number of leukocytes, the number of neutrophils,
222
as well as the activity of the liver specific enzyme GLDH during the experiment are presented
223
in Figure 3. For Hp, GLDH, and PMNL numbers, effects of calving (P < 0.001) were
224
observed. Leukocyte number was also affected (P < 0.01) by calving.
225
Immediately after calving, Hp reached maximal values corresponding to a 4.5-fold
226
increase of the prepartal concentrations. During the third wk of lactation, baseline values were
227
re-established. Haptoglobin is one of the major acute phase proteins in cattle, and increasing
228
concentrations are attributable to inflammatory reactions. Increasing concentrations of Hp at
9 Downloaded from jas.fass.org by on May 14, 2011.
Journal of Animal Science
Page 10 of 31
229
parturition agree with the observations from Uchida et al. (1993) and are probably related to
230
the concomitant hormonal changes and to tissue lesions occurring during birth.
231
The sequential changes of the pre- and post-partum activities of GLDH recorded in
232
our study were similar to those reported by Hoedemaker et al. (2004). Compared with healthy
233
non-lactating and non-pregnant cows (West, 1997), these values were increased and might be
234
attributed to liver cell damage caused by triacylglycerol accumulations as discussed elsewhere
235
(Staufenbiel et al., 1993). Nevertheless, the increase of GLDH activity observed herein is only
236
marginal compared with cows with clinical hepatic lipidosis or other liver disease (West,
237
1997).
238
Increasing numbers of PMNL towards the time of calving (Reisler et al., 2000;
239
Burvenich et al., 2003) were also found in the cows studied herein, which were without other
240
pathological findings. However, this increase was hardly reflected in leukocyte numbers. In
241
general, the criteria used to evaluate health were judged as being within the physiological
242
range of high yielding dairy cows.
243 244
Evaluation of the Classification According to Criteria Related to Fat Mobilization
245
The criteria related to fat mobilization were used to classify the cows according to
246
their values recorded at one defined point of time (NEFA, BHB, IGF-I, and BCS) or time
247
interval ( BCS, IGF-I-%, and Leptin-%) as being below or above a certain threshold (-ok or –
248
ltd groups). Between each of the 2 groups thereby defined, the temporal patterns of the
249
various variables tested were compared. Table 2 summarizes the results of these comparisons.
250
The presentation is limited to significant relationships. In addition, the effort required for
251
analysis as well as the earliest possible time of analysis is considered.
252
BCS a.p. and BCS loss. Differentiating the animals by their a.p. BCS values yielded
253
effects on leptin (P = 0.007); as expected, cows with BCS scores
254
relatively greater concentrations of leptin in serum than cows with lower scores. When 10 Downloaded from jas.fass.org by on May 14, 2011.
3.5 (BCS-ok) had
Page 11 of 31
Journal of Animal Science
255
animals were grouped according to the degree of BCS loss ( BCS), cows with a BCS decline
256
of more than 0.5 points also had lower concentrations of leptin (P = 0.017) and of IGF-I (P =
257
0.001) during the study. However, when considering that the time necessary to detect BCS
258
losses in early lactation is more than 2 wk p.p., BCS evaluation is of limited use when
259
requiring earlier information about adaptive status.
260
BHB and NEFA. Cows with BHB values greater than 1.2 M at the second wk p.p. had
261
elevated NEFA concentrations during the study (P = 0.01). Due to the high variability of the
262
concentrations of BHB throughout the study, the BHB threshold used did not allow for the
263
detection of differences in BHB serum concentrations between the BHB-ok and the BHB-ltd
264
group throughout the time interval recorded.
265
Increased NEFA as well as decreased IGF-I concentrations were observed in NEFA-
266
ltd cows, with major differences during the first 4 (NEFA: P < 0.04) and first 2 (IGF-I: P
