Professional Documents
Culture Documents
Association of Production Diseases With Motor Activity-Sensing Devices and Milk Progesterone Concentrations in Dairy Cows
Association of Production Diseases With Motor Activity-Sensing Devices and Milk Progesterone Concentrations in Dairy Cows
net/publication/325470954
CITATIONS READS
0 32
9 authors, including:
Some of the authors of this publication are also working on these related projects:
All content following this page was uploaded by Jess A. Williams on 25 July 2019.
Theriogenology
journal homepage: www.theriojournal.com
a r t i c l e i n f o a b s t r a c t
Article history: We have previously established that the efficiency of identifying oestrus with activity-sensing devices
Received 13 January 2018 can be compromised by common production diseases; the present study was undertaken to determine
Received in revised form how these diseases may affect device readings. A total of 67 Holstein-Friesian cows, >20 days post-
24 May 2018
partum, were equipped with activity-sensing neck collars and pedometers, and simultaneous milk
Accepted 29 May 2018
Available online 31 May 2018
progesterone profiles were also monitored twice a week. The influences of common production stressors
on maximum activity and progesterone values were analysed. Approximately 30% potential oestrus
events (low progesterone value between two high values) remained unrecognised by both activity
Keywords:
Oestrus
methods, and progesterone values in these animals were higher on the potential day of oestrus when
Lameness both activity methods did not detect an event (0.043 ± 0.004 versus 0.029 ± 0.004 ng/mL; P ¼ 0.03). Data
Body condition score from a subset of 45 cows (two events each) were subjected to mixed models and multiple regression
modelling to investigate associations with production diseases. Cow motor activity was lower in lame
cows. Maximum progesterone concentrations prior to oestrus increased as time postpartum and body
condition score (BCS) increased. There were also fewer days of low progesterone prior to oestrus asso-
ciated with increases in BCS and maximum progesterone concentrations prior to oestrus. In conclusion,
lameness was associated with lower activity values, but this suppression was insufficient to account for
lowered oestrus detection efficiency of either device. However, associations were identified between
production diseases and progesterone profiles.
© 2018 Elsevier Inc. All rights reserved.
* Corresponding author.
E-mail address: hdobson@liv.ac.uk (H. Dobson).
https://doi.org/10.1016/j.theriogenology.2018.05.038
0093-691X/© 2018 Elsevier Inc. All rights reserved.
58 J. Williams et al. / Theriogenology 118 (2018) 57e62
2. Materials and methods The algorithm used to identify an oestrus was based on an increase
in steps per hour greater than 70% of the average for the previous 10
2.1. Animals days.
This study was performed under a UK Home Office license for 2.3. Hormone analyses
work on living animals and with the approval of the University of
Liverpool Ethical Review Process. HolsteineFriesian cows in their Milk samples for progesterone analysis were collected each
first to eleventh lactation entered the six-month study (March- Tuesday and Friday throughout the study. All samples were taken
eAugust) of a 200-cow herd that calved throughout the year on one immediately before milking and promptly stored at 20 C without
free-housed commercial UK dairy farm. Mean milk sales per cow preservative. Progesterone concentration was analysed as pregnane
were around 11,000 L annually, with average peak yields of 54 L per metabolites in 50 ml whole milk samples using an established EIA
day. At any one time, a group of approximately 80 cows in the assay [10]. For the present study, the minimum detectable amount
period four months after calving were housed in a cubicle shed was 0.02 ng/mL, and the intra- and inter-assay coefficients of
with concrete passage-ways. Milking of all cows took place three variation were 9.7 and 15.8%, respectively. When values were less
times a day starting at 05:30 h, 14.00 h and 20:30 h. All animals had than 0.2 ng/mL, luteal tissue was considered to be absent, whereas
continual access to a total mixed ration that was pushed up to a values greater than 0.3 ng/mL indicated the presence of luteal tis-
feed-fence five times a day. All cows were monitored after entering sue [10]. Progesterone profiles were produced for each study ani-
the study until confirmed pregnant after ultrasound examination mal and used to indicate potential oestrus periods to be identified
by the farm's veterinarian five weeks after artificial insemination or by neck-collars or pedometers. The term “potential oestrus event”
until the end of the study period. Therefore, many animals had used in data analysis was defined as each time milk progesterone
more than one potential oestrus event studied. concentration was below 0.2 ng/mL for 2e6 days, preceded and
Every two weeks, the same member of the research team followed by an increase above 0.3 ng/mL for a minimum of 7e10
assessed each cow's body condition score (BCS) using a 1e5 system days.
incorporating 0.5 scores [8], and lameness using a standardised 1e5
system [9]. For statistical analysis, individuals were categorised as
2.4. Data analyses
not lame (lameness score <1.5) or lame (lameness score >1.6;
n ¼ 24 cows). Once a month, individual cow 24 h milk yields
Sixty-seven cows were studied providing 164 potential oestrus
(subsequently converted to ECM: energy corrected milk (4% fat and
events for consideration. Data from four cows were excluded
3.3% protein calculated as 24 h milk production 0.383 fat
because the progesterone profiles were very irregular. Data were
% þ milk production 0.242 protein % þ 0.7832 milk pro-
also excluded from 34 potential events due to obvious false posi-
duction/3.1138), somatic cell counts (SCC), % fat and % protein were
tives within the neck collar or pedometer data (which mainly
recorded by Cattle Information Services (CIS; Rickmansworth, UK).
occurred during the luteal phase). The neck collar system mal-
The monthly value immediately prior to a potential oestrus event
functioned in 31 (18%) and the pedometers malfunctioned in 17
(NEAR_) was used for analysis. Treatments for lameness and
(10%). These malfunctions were due to mixtures of human error,
mastitis followed normal farm practice: regular or remedial foot-
e.g., power outage, malfunction of the equipment, failure to
trimming, and antibiotic treatment for udder infections.
correctly enter information on the computers. Statistical analysis
allowed for these missing data.
2.2. Activity-sensing devices Milk progesterone concentrations during all potential oestrus
events identified by either neck collars or pedometers were initially
On entry into the study, all cows were fitted with an activity- compared by Students t-test. Due to possible bias and problems
sensing neck-collar (Heatime™; SCR Engineers, Netanya, Israel; introduced either by over-representation or under-representation
distributed by FabDec, Ellesmere, UK) with data being downloaded of some cows while running Mixed Models, subsequent analyses
to a dedicated program (Dataflow, SCR Engineers) in 2-h blocks at were performed using the first two potential oestrus events in a
the milking parlour exit. As designated by the equipment provider, total of 45 cows which contributed to the final balanced data-set.
a value of greater than 4.7 standard deviations above the mean This comprised 10 primiparous animals, 14 cows that had had 2
activity of the previous eight two-hourly blocks was used to indi- calvings, 11 had 3 calvings, 4 had 4 calvings and six cows had had 5
cate oestrus - this algorithm precluded assessment of baseline or more calvings).
values. Also at recruitment, a leg pedometer (Afitag; SAE Afikim, All statistical analyses were performed with SAS software
Israel; distributed by Fullwood, Ellesmere, UK) was placed on all version 9.3, using the MIXED procedure for linear Mixed Models,
cows as an additional measure of cow activity. The pedometer data unless stated differently. A repeated effect of time (day postpartum)
were also collated remotely by a computer in 8-h blocks at milking within animals was tested. The correlations between test days were
using a dedicated program (Crystal; Fusion Electronics, Fullwood). accounted for by specifying a correlation structure (spatial power)
Table 1
Efficiencies of detection for neck-collars, pedometers or combined device data, with accompanying milk progesterone concentrations on day of potential oestrus
(mean ± SEM).
Device Total potential Events Efficiency Milk progesterone (ng/mL) for Milk progesterone (ng/mL) for P (between progesterone
events (n) recorded (n) (%) recorded event missed event concentrations)
among residuals to account for differing intervals between undetected event (overall mean ± SEM: 3.82 ± 0.43 days; P ¼ 0.23).
samplings. However, progesterone concentrations on the day of potential oestrus
Stepwise regression was used with a backward elimination events that were not identified by either neck collars or pedometers
approach to fit the final regression model. In each step a factor was tended to be greater than for events registered by each device
kept or excluded from the set of explanatory ones based on pre- (P ¼ 0.090 and 0.059, respectively; Table 1). When data were com-
specified exclusion criterion of non-significant (P > 0.20) effects. bined to examine only those events identified by the two devices
The residuals from the response variables generated from the simultaneously, progesterone concentrations were significantly
model were tested for normal distribution. Progesterone values at greater for events undetected by both devices (P ¼ 0.028; Table 1).
oestrus, and the number of prior days with low progesterone
values, deviated from a normal distribution and were log-
transformed. To facilitate interpretation of the data for these pa- 3.2. Associations between cow activity, progesterone measurements
rameters, non-transformed values are presented. and production diseases
Values for each milk variable were taken as those measured
nearest to each potential oestrus event; other dependent variables Possible associations between cow activity, progesterone
were: maximum neck-collar value, MAX_HT; maximum pedometer
value, MAX_PEDO; baseline pedometer value, BASE_PED; proges-
terone value at oestrus, PROG_OE; the number of days prior to 500
oestrus with low progesterone values, DAYS_PRI; maximum pro- A
(arbitray units)
The initial model used was: 300
400
score <1.5 or >1.6), Iq is the fixed effect of parity q (q ¼ 2 classes; 2
(arbitrary values)
measurements and production diseases (as defined previously) these animals tended to have more days with low progesterone
were then examined in a balanced data-set that included the two concentrations prior to oestrus (Fig. 2F).
first postpartum potential oestrus events from 45 cows (i.e., 90 Relationships between all variables were further considered by
events). examining partial correlations (MANOVA model), and the results
A total of 24 cows had a lameness score >1.6. From MIXED are summarized in Table 2. There were few relationships between
model analysis, lameness was associated with lower maximum and production disease variables; fat: protein ratio (NEAR_FP) was
baseline pedometer values (P < 0.05), and a tendency for lower positively associated with Energy Corrected Milk (ECM), and BCS
maximum neck-collar values; Fig. 1. Lameness was also associated was positively associated with oestrus number (PP_OE; P ¼ 0.01).
with lower progesterone concentrations at oestrus (P < 0.01; Maximum progesterone concentrations prior to oestrus were
Fig. 2A) and a similar tendency with maximum concentrations of positively associated with progesterone concentrations at oestrus,
progesterone prior to oestrus (Fig. 2B). Oestrus number (PP_OE) but negatively associated with number of days with low concen-
was positively associated with maximum progesterone concen- trations prior to oestrus (P < 0.03).
trations prior to oestrus but negatively associated with the number Maximum pedometer values were positively associated with
of days for which progesterone was low before oestrus (for both baseline pedometer values (P < 0.0001) and with maximum neck-
P ¼ 0.01; Fig. 2C and D). Finally, progesterone concentrations at collar values (P < 0.0001). Overall, only tendencies were found for
oestrus were higher in cows of parity >2 (P ¼ 0.0015; Fig. 2E), and relationships between activity variables and progesterone
0.07
A 1.6 B
0.05 1.2
p=0.008
1
0.04 p=0.13
0.8
0.03
0.6
0.02
0.4
0.01 0.2
0 0
lame, no lame, yes lame, no lame, yes
1.6 12
C D
Maximum prior progesterone ng/ml
1.4
Prior days with low progesterone
10
1.2 p=0.011
8
1
p<0.01
6
0.8
0.6 4
0.4
2
0.2
0
0 1 2 >=3
1 and 2 3
PP_OE PP_OE
0.07 12
E F
Prior days with low pogesterone
Progesterone at oestrus ng/ml
0.06 10
0.05 p=0.0015 p=0.13
8
0.04
6
0.03
4
0.02
2
0.01
0
0 parity > 2
parity > 2
Fig. 2. Effects of categorical variables on progesterone concentrations at potential oestrus (A and E), maximum concentrations of progesterone prior to potential oestrus (PP_OE; B
and C), and number of days with low progesterone concentrations prior to potential oestrus (D and F). Results are presented as least square means ± SEM.
J. Williams et al. / Theriogenology 118 (2018) 57e62 61
Table 2
Partial Correlation Coefficients from the Error SSCP Matrix (upper row) with corresponding P value presented below each variable. The Partial correlation coefficients were
adjusted for an overall parity effect (P < 0.00001).
r/p-value DAY_PP ECM NEAR_FP NEAR_SCC BCS lame PROG_OE DAYS_PRI MAX_PRIO MAX_HT MAX_PEDO BASE_PED
PP_OE 0.47 0.02 0.04 0.02 0.34 0.24 0.002 0.23 0.32 0.15 0.15 0.01
<0.001 ns ns ns 0.01 0.07 ns 0.09 0.02 ns ns ns
DAY_PP 1 0.02 0.10 0.02 0.13 0.19 0.10 0.14 0.01 0.27 0.05 0.05
ns ns ns ns ns ns ns ns 0.04 ns ns
ECM 1 0.36 0.09 0.06 0.17 0.02 0.13 0.03 0.09 0.05 0.20
0.01 ns ns ns ns ns ns ns ns 0.14
NEAR_FP 1 0.01 0.23 0.10 0.20 0.14 0.04 0.03 0.19 0.34
ns 0.08 ns 0.14 ns ns ns 0.15 0.01
NEAR_SCC 1 0.21 0.07 0.13 0.07 0.06 0.06 0.30 0.26
0.13 ns ns ns ns ns 0.03 0.05
BCS 1 0.18 0.07 0.28 0.26 0.01 0.005 0.10
ns ns 0.04 0.05 ns ns ns
Lame score 1 0.29 0.23 0.16 0.01 0.00 0.13
0.03 0.09 ns ns ns ns
PROG_OE 1 0.06 0.34 0.21 0.24 0.18
ns 0.01 0.13 0.08 ns
DAYS_PRI 1 0.28 0.16 0.05 0.05
0.03 ns ns ns
MAX_PRIO 1 0.07 0.06 0.06
ns ns ns
MAX_HT 1 0.57 0.22
<.0001 0.10
MAX_PEDO 1 0.58
<.0001
measurements. Few relationships were found between production detection devices coincides with visual observations that lame
diseases and activity measurements. cows walk less, lie down more, and have reduced oestrus intensity
Relationships between cow status, production diseases and once ovarian cyclicity has resumed after calving [10]. These re-
progesterone measurements were associated with BCS: higher BCS ductions may be due to the physical pain associated with lameness,
being associated with fewer days of low progesterone before oes- however, other (hormonal) influences are involved. Although
trus, and higher progesterone concentrations before oestrus (both ovarian follicular emergence, selection and physical growth are
P 0.05). Furthermore, increased lameness was associated with unaffected, lame cows have low LH pulse frequencies in the late
lower progesterone concentrations at the time of oestrus (P < 0.05), follicular phase, consequently low simultaneous plasma oestradiol
and there tended to be more days with low progesterone prior to concentrations, display less intense oestrus behaviour and have a
oestrus (P < 0.10). reduced incidence of LH surges and ovulations [11]. There were also
markedly lower progesterone concentrations during the luteal
4. Discussion phase when severely lame cows were compared to non-lame ani-
mals. In the present study, lameness was considered in two cate-
The present study confirms that a large proportion (~30%) of gories (lame/not lame) but there was still a tendency for lameness
periods with low milk progesterone values in postpartum cows are to be associated with lower maximum progesterone values prior to
not associated with increased motor activity, although progester- oestrus concurring with the hypothesis that lameness disturbs
one profiles indicated regular ovarian cyclicity [7]. That is, some normal ovarian function [10].
cows did not display signs of oestrus possibly as a behavioural We have already established that lame cows have a lower bite
trade-off while exposed to adverse conditions. Nevertheless, there rate at pasture and are more likely to have a low BCS [12]. The
was no difference in efficiency between neck-collars and pedom- negative correlation of BCS with the number of days with low
eters to detect periods of low progesterone. The maximum activity progesterone prior to oestrus (i.e., as BCS increases the follicular
values recorded in the present study were considerably greater phase becomes shorter) agrees with the hypothesis that negative
than the threshold values for both devices and so do not fully energy balance (low BCS) affects gonadotrophin pulsatility and
explain why many potential oestrus periods were not identified. ovarian steroid synthesis [13,14]. Similarly, a healthy follicle will
However, cows that were not detected did have marginally higher produce a good corpus luteum (as BCS increased, so did maximum
milk progesterone concentrations at a potential oestrus than those prior progesterone values). However, some cows cannot cope with
detected with increased motor activity. Moreover, following ad- negative energy balance after calving, and as suggested many years
justments for the effects of all concurrent variables, lameness was ago, any additional stressor (such as lameness) that disturbs this
associated with lower activity values. In addition, there were sig- knife-edge balance will ultimately lead to impaired reproductive
nificant correlations between three indices of progesterone con- function including poor oestrus expression [15,16].
centrations (lower values at oestrus, lower maximum values prior Milk fat and milk protein values were excluded from the final
to oestrus, and number of days with low values prior to oestrus) statistical model used in the present study as having non-
and the following: day postpartum (and postpartum oestrus significant influence. However, fat: protein ratios tended to be
number), parity, lameness and BCS. In this high-yielding herd, no positively correlated with BCS and progesterone values at oestrus.
associations with cow activity or progesterone variables were Previously, cows with a first test-day fat: protein ratio of >1.5 have
identified with ECM, although ECM was positively associated with been associated with higher risks for clinical disease, including
fat: protein ratios. ketosis which has a negative effect on oestrus detection [17e19].
The association of lameness with lower activities of both There is now the potential to harness the data from newly available
62 J. Williams et al. / Theriogenology 118 (2018) 57e62