The effect of subclinical ketosis in early lactation on reproductive performance of postpartum dairy cowsJ. Dairy Sci. 90:2788–2796
American Dairy Science Association, 2007.
The Effect of Subclinical Ketosis in Early Lactation on Reproductive
Performance of Postpartum Dairy Cows
R. B. Walsh,*1 J. S. Walton,† D. F. Kelton,* S. J. LeBlanc,* K. E. Leslie,* and T. F. Dufﬁeld*
*Department of Population Medicine, and
†Department of Animal and Poultry Science, University of Guelph, Ontario, Canada, N1G 2W1
d). To further investigate this, the effect of elevatedcirculating BHBA was permitted to vary with time. The Data generated from 796 Holstein cows enrolled in effect decreased with time, such that the daily probabil- a clinical trial to investigate the health effect of a mo- ity of pregnancy increased similar to nonsubclinically nensin controlled release capsule were analyzed to in- ketotic cows by approximately 160 d in milk. From this vestigate the association between circulating serum analysis, both the relative circulating concentration of β-hydroxybutyrate (BHBA) concentration in the peri- BHBA and the duration of elevated circulating BHBA parturient period and subsequent reproductive perfor- were negatively associated with the probability of preg- mance. Overall, accounting for both repeated measures nancy at ﬁrst service.
within cow and clustering at the herd level, non-preg- Key words: subclinical ketosis, β-hydroxybutyrate,
nant cows after ﬁrst insemination tended to have in- time to pregnancy creased circulating BHBA concentrations from 3 wkbefore calving to 9 wk after calving relative to pregnantcows. Including the interaction between the week of sample collection and pregnancy outcome, non-preg- The concept of homeorhesis or prioritization of energy nant cows had higher circulating BHBA concentrations distribution to meet speciﬁc physiological demands was in the second week after calving than cows diagnosed reﬁned in 1980 (Bauman and Currie, 1980). Parti- pregnant after ﬁrst artiﬁcial insemination. Within indi- tioning of available energy to milk production early in vidual weeks, cows with circulating BHBA concentra- lactation, at the expense of reproduction, has formed a tions ≥1,000 mol/L in the ﬁrst week postpartum were conceptual framework to address the effect of negative less likely to be diagnosed pregnant after ﬁrst insemina- energy balance on future reproductive performance, as tion. In the second week postpartum, the cows with measured by the time to establishment of luteal activ- circulating BHBA concentrations ≥1,400 mol/L were ity, probability of pregnancy after ﬁrst AI [pregnancy signiﬁcantly less likely to be pregnant after ﬁrst artiﬁ- risk (PR)/AI], and time to conception.
cial insemination. A dose response relationship was Cows experiencing delayed onset of luteal activity found when a comparison of the probability of preg- postpartum typically have lower DMI, produce less nancy after ﬁrst insemination and duration of elevated milk, and subsequently lose more body condition early circulating ketone bodies was investigated. The proba- in lactation (Staples et al., 1990). At the whole-animal bility of pregnancy was reduced by 20% in cows diag- level, the magnitude and duration of negative energy nosed subclinically ketotic in either the ﬁrst or second balance are inﬂuenced by feed intake and milk produc- week postpartum. Nevertheless, cows above the sub- tion. At the organ and molecular level, variation in clinical ketosis threshold in both the ﬁrst and second energy utilization efﬁciency, ranging from rumen func- week postpartum were 50% less likely to be pregnant tion to the ability of the liver to repartition mobilized after ﬁrst insemination. Similarly, the median time to fat, affect energy balance status.
pregnancy increased in cows experiencing elevated Previous research has reported an association be- BHBA concentrations in either (124 d) or both (130 d) tween calculated negative energy balance and repro- the ﬁrst and second week postpartum relative to cows ductive traits, including days to ﬁrst luteal activity, never experiencing elevated BHBA concentrations (108 ﬁrst service conception risk, days from calving to AI,and days from calving to conception (Villa-Godoy etal., 1988; Staples et al., 1990; Reist et al., 2003b). Thephysiological state of energy deﬁciency impairs hypo- Received August 28, 2006.
Accepted January 31, 2007.
thalamic responsiveness to circulating estradiol-17β, 1Corresponding author: firstname.lastname@example.org resulting in reduced GnRH pulse frequency and con- EFFECT OF SUBCLINICAL KETOSIS ON REPRODUCTION comitant reduction in LH required for follicular devia- collected from a coccygeal blood vessel into a 10-mL tion and eventual ovulation (Jolly et al., 1995; Dawuda vacuum tube (Monoject red stopper blood collection et al., 2002; Butler, 2003).
tubes; Sherwood Medical, St. Louis, MO). Blood was Energy balance is dependent on many factors. Vari- collected at the time of monensin CRC or placebo oral ous metabolic and endocrine blood and milk measures administration 3 wk before calving and at wk 1, 2, 3, such as NEFA, ketone bodies, glucose, insulin, milk 6, and 9 postpartum. Blood samples were stored, on fat, or fat:protein ratio are indirect measures of energy ice, in an insulated cooler from time of collection until balance. Circulating ketones increase when the maxi- sample processing occurred. Within 5 h postcollection, mal ability of the liver to oxidize NEFA and store triac- blood samples were centrifuged at 733 × g for 10 min.
ylglycerol is exceeded. Associations between elevated The samples of serum were submitted to the Clinical circulating ketone concentration and periparturient Pathology Laboratory (Department of Pathobiology, uterine disease were described by Reist et al. (2003a); Ontario Veterinary College) for the measurement of Ca, however, investigation of thresholds was based on the P, total protein, urea, glucose, and BHBA concentra- distribution of circulating ketone concentrations or an- tions and determination of aspartate aminotransferase ecdotal experience rather than the associated change activity using an automated analyzer (Dacos 2 ana- in the risk of disease (Dufﬁeld et al., 1998). Similarly, lyzer; Coulter Electronics, Hialeah, FL).
the effect of postpartum circulating ketone concentra- Peripartum disease information including dystocia tion on PR/AI, time to commencement of luteal activity, (veterinary-assisted parturition), retained placenta and time to pregnancy has been described later in lacta- (failure to pass the fetal membranes within 24 h), milk tion with variable results (Andersson and Emanuelson, fever (veterinary diagnosed), metritis (inﬂammation of 1984; Cook et al., 2001). A circulating BHBA concentra- the uterus <15 DIM), clinical ketosis, displaced aboma- tion threshold immediately postpartum, deﬁned by de- sum (left or right displacement diagnosed by a veteri- monstrable reduction in the PR/AI, has not been val- narian), respiratory illness, and lameness was captured in on-farm data sheets, veterinary records, and on-farm The objectives of this retrospective analysis were to computer record systems (Dufﬁeld et al., 1999). Repro- investigate relationships between the magnitude and ductive data, including DIM at ﬁrst insemination, con- duration of increased serum BHBA measured at −3 wk ception date, number of inseminations, herd removal and in each of the ﬁrst, second, third, sixth, and ninth date, and pregnancy status at removal, were recorded weeks postpartum and the PR/AI, the time from calving using the same methods. There was minimal use of to ﬁrst AI, and time from calving to pregnancy.
ovulation synchronization protocols. Pregnancy diagno-sis was performed at least 40 d after insemination by MATERIALS AND METHODS
rectal palpation. Five herds had ﬁrst lactation animalsexcluded, because they had been exposed to a monensin Data were collected as part of a clinical trial to deter- premix within 4 wk of parturition.
mine the effect of prepartum administration of monen-
sin as a controlled release capsule (CRC) on postpartum
Data Management and Statistical Analysis
health, energy balance, and reproductive performance(Dufﬁeld et al., 1998, 1999). Brieﬂy, 25 Holstein herds Statistical analysis was performed using Intercooled around Guelph, Ontario, Canada, that were enrolled in Stata 9.1 (StataCorp LP, College Station, TX) and SAS milk recording through Ontario DHI were studied from (SAS Inst. Inc., Cary, NC). Descriptive statistics were March 1995 to December 1995. Herd size ranged from used to characterize animals diagnosed pregnant after 25 to 160 lactating animals, with 305-d rolling herd ﬁrst insemination, relative to animals diagnosed not average for milk production from 7,000 to 10,000 kg.
pregnant. Data were collected as part of a clinical trial; Five herds were fed a TMR, and the remaining herds therefore, treatment remained in all analyses.
had a component feeding system. A technician visited The shape of the BHBA curve between weeks was each farm weekly at a consistent time of day. At each assessed using PROC MIXED in SAS (SAS Institute, weekly visit, the technician administered the randomly 2004), accounting for the random effect of herd and assigned treatment (monensin CRC or placebo); re- repeated measures of cows within herd. After trans- corded animal data including lactation, season of calv- forming the BHBA data into the natural logarithm ing (winter: December through February; spring: scale, variables offered to the model included treat- March through May; summer: June through August; ment, parity, periparturient disease, the pregnancy sta- fall: September through November), and BCS (Fergu- tus diagnosed after ﬁrst AI, and all 2-way interactions.
son et al., 1994; ≤2.5 thin; ≥2.75 to ≤3.5 fair; ≥3.75 fat); Residual analysis was performed at the cow level. The and collected blood samples. Each blood sample was distribution of residuals was symmetrical with long Journal of Dairy Science Vol. 90 No. 6, 2007 WALSH ET AL.
tails, thus violating the normality of residual assump- Table 1. Characterization of 796 Holstein dairy cattle from 25 On-
tario herds diagnosed pregnant or non-pregnant after ﬁrst insemina-
tion. To investigate the effect of inﬂuential observations (extreme values in the independent or dependent vari-ables), they were removed from the analysis. Removal of inﬂuential observations improved the normality of residual assumption without affecting the direction of Days to ﬁrst insemination1 80.4 (77.3 to 83.5) 79.4 (77.1 to 81.8) the reported coefﬁcients indicating that the observed 3.44 (3.35 to 3.43) 3.39 (3.43 to 3.5) effect was not solely due to the effect of extreme observa- tions. Final models included all observations. Least Periparturient disease3 Retained placenta squares means of weekly BHBA concentration were estimated for pregnant and non-pregnant cows at each weekly measurement. To appropriately account for the Displaced abomasum repeated measures within cow, a postestimation adjust- ment was performed on the reported probability value using the Bonferroni-Sidak adjustment to account for Cystic ovarian disease correlation between sampling intervals.
The effect of BHBA concentration at each week on Mean (95% conﬁdence interval).
probability of pregnancy at ﬁrst insemination was as- Proportion of randomly assigned treatments with the monensin controlled release capsule (CRC).
sessed using multivariable logistic regression account- 3Proportion (n).
ing for correlation of cows within herd using a general-ized estimating equation with an exchangeable correla-tion structure. Cow level variables offered to the model to modiﬁed Cox-Snell residuals assessed goodness of included parity, season of calving, initial BCS, BCS at ﬁt. Animals with survival times less than 30 DIM were 60 DIM, change in BCS, season of calving, and peripar- excluded from this analysis, because they were not at tum disease information. Herd level variables included risk for pregnancy, and they did not complete the sam- farm type (free stall vs. tie stall) and herd size. Treat- pling protocol. The proportionality assumption was as- ment with monensin was retained in all models, be- sessed graphically by plotting the logarithm of the haz- cause the data were generated as part of a monensin ard function against the logarithm of time. The Akaike clinical trial. Initial screening of variables was per- information criterion was used to assess the overall ﬁt formed using a manual backward stepwise procedure of the ﬁnal model. Delta-beta-like residuals were used with signiﬁcant variables (P < 0.1) remaining in the to identify points of leverage, whereas deviance residu- model. This procedure was repeated for each week post- als were calculated and sorted to identify possible partum in which serum BHBA measurements were available. Initially, serum BHBA was offered as a con-tinuous variable. Subsequently, based on the appear- ance of a smoothed scatterplot of serum BHBA andprobability of pregnancy at ﬁrst AI, realistic thresholds The prevalence of dystocia, twins, peripartum dis- were created at 200 mol/L increments from 1,000 to ease, reproductive abnormalities, and other health 1,600 mol/L. Ultimately, sequential multivariable lo- events in the present data were reported previously gistic regression, accounting for the effect of clustering (Dufﬁeld et al., 1999). Brieﬂy, treatment with a monen- at the herd level, was applied based on the predicted sin CRC reduced the risk of displacement of the aboma- thresholds to evaluate their potential signiﬁcance.
sum by 40% [odds ratio (OR) = 0.59; P = 0.004] and the
The effect of prolonged exposure to elevated ketone risk of multiple illnesses by 40% (OR = 0.6; P = 0.01).
bodies on PR at ﬁrst insemination was investigated The data and outcome of ﬁrst insemination were re- by generating a composite variable representing the corded for 796 cows. The PR/AI at ﬁrst insemination additive effect of the number of weeks with serum was 34.8%. Further characterizations of animals based BHBA ≥1,000 mol/L in the ﬁrst week and ≥1,400 mol/ on pregnancy result to ﬁrst AI are presented in Table L in the second week.
1. Median time from calving to ﬁrst insemination was Finally, time from calving to ﬁrst insemination and 74 d, ranging from 31 to 243 d. Days in milk at ﬁrst time to pregnancy were modeled using semiparametric AI, parity, season of calving, and periparturient disease survival analysis. This method accounted for animals were signiﬁcant in a model to determine risk factors that did not become pregnant. Robust standard errors for PR/AI. A BCS of 3.25 to 3.75 was assigned to 86% were generated to account for clustering of animals of animals independent of treatment. Treatment did within herds. Plotting the cumulative hazard relative not affect the probability of pregnancy at ﬁrst AI (mo- Journal of Dairy Science Vol. 90 No. 6, 2007
EFFECT OF SUBCLINICAL KETOSIS ON REPRODUCTION Figure 1. Least squares mean serum BHBA concentrations and
standard error accounting for the random effect of herd and repeated Figure 2. Predicted probability of pregnancy generated through
measures with unequal spacing and adjusting for treatment and multivariable logisitic regression accounting for the correlation of parity in Holstein cows that were subsequently diagnosed pregnant cows within herds relative to serum BHBA concentration determined or non-pregnant after ﬁrst insemination. Each cow was sampled once in the second week postpartum in 796 Holstein dairy cows.
in the third week before calving, in each of the ﬁrst 3 wk, and againin the sixth and ninth weeks after calving. *P < 0.05.
1,064.0 ± 30.1, 66 to 8,976, respectively). The predicted nensin-treated group, 35.2%; placebo group 34.5%; probability of pregnancy, relative to circulating BHBA OR = 1.03; P = 0.84).
concentration for the second week postpartum is pre- Cows diagnosed non-pregnant after ﬁrst AI tended to sented in Figure 2. Serum BHBA in wk 3, 6, and 9 were have increased circulating BHBA concentrations from 3 not signiﬁcantly associated with PR.
wk before calving to 9 wk after calving relative to cows The proportion of animals classiﬁed as subclinically diagnosed pregnant after ﬁrst AI (P = 0.09) accounting ketotic (SCK) at the predetermined thresholds by ﬁrst
for monensin treatment and parity. A tendency for an service conception is presented in Table 2. From these interaction between the week of sample and pregnancy models, cows with serum BHBA ≥1,000 mol/L in wk status after ﬁrst AI was identiﬁed (P = 0.1). This inter- 1 or ≥1,400 mol/L in wk 2 were signiﬁcantly less likely action term was signiﬁcant at wk 2. Least squares to be diagnosed pregnant after ﬁrst insemination. There means were calculated at each sampling period for preg- were a total of 288 animals with serum BHBA ≥1,000 nant and non-pregnant cows and compared adjusting mol/L in wk 1, and 150 cows were ≥1,400 mol/L in for repeated measures within cow. Nonpregnant cows wk 2. A total of 104 cows were classiﬁed as SCK in both had increased circulating BHBA concentrations rela- tive to pregnant cows in the second week after calving Given the available data, evaluation of the effect of (P = 0.03; Figure 1). Treatment with the monensin CRC prolonged exposure to elevated ketones was limited to reduced the BHBA estimate at each sampling point (P analysis of the number of weeks above a series of thresh- < 0.001). Periparturient disease did not signiﬁcantly olds. A dose response curve was identiﬁed when model-alter the shape of the BHBA curve. Neither initial BCS, ing the number of weeks positive without specifying ﬁnal BCS (thin, fair, fat), nor BCS change (≥1 unit) was the order of weeks positive and using a 1,000 mol/L signiﬁcant in a model of factors affecting the probability threshold in wk 1 and the 1,400 mol/L threshold in of pregnancy at ﬁrst AI.
wk 2 (Table 3).
Within individual weeks after calving, an association Survival analysis was used to measure the time from between serum BHBA concentration and the PR/AI was calving to ﬁrst insemination and from calving to preg- identiﬁed, accounting for the effect of correlation of cows nancy. The daily probability of AI was reduced in cows within a herd. The PR/AI was reduced by 2% (OR = classiﬁed as SCK in either wk 1 or 2 relative to cows 0.98, P = 0.06) in wk 1 and 3% (OR = 0.97, P = 0.01) never classiﬁed as SCK (P = 0.04; Table 4). Further, in wk 2 for each 100 mol of BHBA/L increase. Although median time to ﬁrst AI was delayed (P < 0.01) for cows the reported OR appears exceedingly small, the range calving in the spring relative to cows calving in the fall of serum BHBA reported is wide in both wk 1 and wk [median, 95% conﬁdence interval (CI) spring = 81 d
2 (mean ± SE; range: 1,032.4 ± 27.6, 101 to 9,382 vs.
(95% CI = 78 to 85 d; fall 70 d, 68 to 72 d)].
Journal of Dairy Science Vol. 90 No. 6, 2007 WALSH ET AL.
Table 2. Investigation of the effect of serum BHBA threshold on the probability of pregnancy after ﬁrst
insemination in 796 lactating Holstein dairy cows from 25 Ontario herds
1Odds ratio (OR) determined by logistic regression controlling for the effect of treatment with monensin controlled release capsule and clustering at the herd level. Each week modeled individually.
Median time to pregnancy for animals never classi- probability of pregnancy, measured using the adjusted ﬁed as SCK was 108 d (95% CI = 102 to 112) relative hazard ratio (HR), between SCK and non-SCK cows
to 124 d (95% CI = 116 to 132 d) for animals above the was equivalent (Table 6; Figure 3).
threshold in either wk 1 or 2 and 130 d (95% CI = 115to 136 d) for animals greater than the BHBA threshold in both wk 1 and 2. The proportional hazard assumptionwas violated in this data set. Therefore, the effect of Postpartum hyperketonemia has repeatedly been as- exceeding the BHBA threshold was permitted to vary sociated with reduced reproductive performance in lac- over time. This correction signiﬁcantly improved the tating cows. The current analysis offers several unique ﬁt to the model. Controlling for season of calving and opportunities. First, the large number of individual monensin treatment, the daily probability of pregnancy cows that were followed continuously from before calv- was reduced in cows classiﬁed as SCK in either or both ing, weekly for the ﬁrst 3 wk after calving, and again wk 1 or 2 relative to cows never classiﬁed as SCK (Table in the sixth and ninth weeks allows investigation of 5). The effect of elevated BHBA concentration remained the differences in energy status between cows over the until approximately 140 DIM, after which the daily transition period and within individual weeks. A com-prehensive investigation of the effect of magnitude andduration of hyperketonemia on reproductive perfor- Table 3. Effect of prolonged elevation of serum BHBA concentration
mance would best be calculated using the cumulative on the probability of pregnancy after ﬁrst insemination in 796 lactat- exposure to elevated circulating BHBA concentrations.
ing Holstein dairy cattle from 25 Ontario herds Unfortunately, the sampling frequency combined with a lack of information on the typical duration of ketosis prevented this calculation in this data. Second, this > Threshold either week2 data set predates ovulation synchronization protocols > Threshold both weeks3 and as such offers a unique opportunity to investigate Never subclinically ketotic the association between SCK and time to insemination 1Odds ratio (OR) adjusting for the effect of clustering at the herd and other reproductive performance traits. Third, the large number of herds enrolled permits description of 2≥1,000 mol of BHBA/L in wk 1 or ≥1,400 mol of BHBA/L in wk the average effect of SCK in lactating dairy cows on 3≥1,000 mol of BHBA/L in wk 1 and ≥1,400 mol of BHBA/L in reproductive performance controlling for differences in Journal of Dairy Science Vol. 90 No. 6, 2007 EFFECT OF SUBCLINICAL KETOSIS ON REPRODUCTION Table 4. Final Cox proportional hazard model of the effect of the number of weeks cows were classiﬁed as
subclinically ketotic (SCK) based on circulating serum BHBA concentration on days from calving to ﬁrst
insemination in 796 Holstein dairy cows from 25 herds in Ontario, Canada
Numbers of weeks above SCK threshold Season of calving 1HR = hazard ratio.
2≥1,000 mol of BHBA/L in wk 1 or ≥1,400 mol of BHBA/L in wk 2.
3≥1,000 mol of BHBA/L in wk 1 and ≥1,400 mol of BHBA/L in wk 2.
4Relative to cows randomly assigned to receive the placebo device.
There was a tendency for increased circulating BHBA physiologic lesion associated with negative energy bal- concentrations in non-pregnant cows after ﬁrst AI from ance and leads to delayed resumption of luteal activity 3 wk before calving to 9 wk after calving relative to (Reist et al., 2000; Butler, 2003).
cows diagnosed pregnant. The delayed effect of negative Within individual weeks, the probability of preg- energy balance or hyperketonemia on reproductive per- nancy after ﬁrst AI decreased linearly with increasing formance is well documented. Both the duration and circulating BHBA concentration in both wk 1 and 2. A magnitude of negative energy balance are associated 3% reduction in probability of pregnancy for each 100 with reduced hypothalamic GnRH pulse frequency and mol of BHBA/L increase was signiﬁcant in wk 2 (Fig- perpetuation of an inappropriate negative feedback ure 2). This difference is relatively small; however, the loop between estradiol concentration and the required range of values observed in the second week makes LH release necessary to support follicular deviation and this ﬁnding very large at BHBA concentrations >1,400 ovulation. This failure of communication along the hy- mol/L. Pregnancy risk at ﬁrst insemination was not pothalamic-pituitary-ovarian axis is the most common signiﬁcantly affected by BHBA concentrations in wk Table 5. Final Cox proportional hazard model of the effect of number of weeks cows were classiﬁed as
subclinically ketotic (SCK) based on circulating serum BHBA concentrations on days from calving to concep-
tion in 796 Holstein dairy cows from 25 herds in Ontario, Canada
Number of weeks above SCK threshold Season of calving 1HR = hazard ratio.
2TVC = time varying covariate estimate varying with the natural logarithm of time.
3≥1,000 mol of BHBA/L in wk 1 or ≥1,400 mol of BHBA/L in wk 2.
4≥1,000 mol of BHBA/L in wk 1 and ≥1,400 mol of BHBA/L in wk 2.
5Relative to cows randomly assigned to receive the placebo device.
Journal of Dairy Science Vol. 90 No. 6, 2007
WALSH ET AL.
Table 6. Adjusted hazard ratio (AHR) representing the relative effect
predicted independent of threshold used. This associa- of classiﬁcation as subclinically ketotic (SCK; serum BHBA concen-tration ≥1,000 mol/L in wk 1 or ≥1,400 mol/L in wk 2) in either tion was signiﬁcant at ≥1,000 mol/L of BHBA. The or both of the ﬁrst 2 wk postpartum on the daily probability of preg- low number of cows above the higher threshold in- nancy at selected time intervals after calving creases the likelihood of a type II error in this analysis.
In the second week after calving, the probability of preg-nancy decreased with increasing BHBA concentration > SCK threshold > SCK threshold thresholds, with a 40 to 50% reduction at thresholds≥1,400 mol/L.
The duration of exposure to elevated circulating BHBA affects the PR/AI. Cows identiﬁed as SCK in either of the ﬁrst 2 wk were 17% less likely to be preg- nant after ﬁrst AI, whereas cows that remained above 1Hazard ratio × time varying covariateln(time).
the SCK thresholds in both weeks were 53% less likelyto be pregnant. The probability of pregnancy after ﬁrstinsemination in cows classiﬁed as SCK in either week −3, 3, 6, or 9. A strong association between elevated was not different (P = 0.28) from cows never classiﬁedcirculating ketone concentration in the ﬁrst week post- as SCK but was greater than cows classiﬁed as SCK in partum and risk of uterine disease was reported (Reist both weeks (P = 0.02; Table 3). There are few other et al., 2003a). Signiﬁcant associations between propor- studies with which to compare these ﬁndings. The mag- tion of lactating cows cycling by 30 DIM, PR/AI, and nitude of and the rate of recovery from negative energy proportion pregnant by 80 DIM have been reported with balance were associated with time to pregnancy (Reist hyperketonemia from the second to ﬁfth week after et al., 2003a). Similarly, most of the variation in onset calving (Plym Forshell et al., 1991; Reist et al., 2000; of ovarian activity and time to pregnancy was attrib- Koller et al., 2003).
uted to the maximum ketone concentration recorded, Within individual weeks and using thresholds, cows independent of the week postpartum when it occurred with circulating BHBA concentrations ≥1,000 mol/L (Reist et al., 2000; Koller et al., 2003).
(40% of cows) in the ﬁrst week and ≥1,400 mol/L (19% Time from calving to ﬁrst insemination is a reﬂection of cows) in the second week were at signiﬁcantly in- of management practice regarding the voluntary wait- creased risk of non-pregnancy after ﬁrst insemination ing period and estrus detection efﬁciency and accuracy.
(Table 2). In the ﬁrst week after calving, a 25% reduc- Subclinically ketotic cows in either of the ﬁrst 2 wk tion in the probability of pregnancy was consistently were inseminated 8 d later than animals that werenever SCK. Delayed insemination between cows classi-ﬁed as SCK in either week relative to those classiﬁedas SCK in both weeks was not anticipated. This outcomemay reﬂect a differential effect on reproductive efﬁ-ciency for cows that experience a ﬁrst elevation in circu-lating ketone concentrations in the ﬁrst week (in-creased risk of uterine disease; Reist et al., 2003a) andsecond week (increased risk of delayed ovulation; Reistet al., 2000) relative to cows that recover from hyperke-tonemia before the second week of lactation (Table 3).
There was a signiﬁcant effect of SCK on time from calving to pregnancy. Concentrations of BHBA thatwere ≥1,000 mol/L in wk 1 and ≥1,400 mol/L in wk2 were associated with increased time to pregnancy.
This effect was not consistent over time. Animals thatwere SCK in either week had an adjusted HR of 0.55at 60 DIM and 0.75 at 120 DIM. Animals SCK in bothweeks had an associated HR of 0.52 at 60 DIM and 0.84 Figure 3. Survival curves of time to pregnancy for 796 lactating
at 120 DIM. In other words, the negative effect of SCK Holstein cows classiﬁed as subclinically ketotic (SCK) based on serum early in lactation appears to diminish as cows progress BHBA concentration determined in either the ﬁrst week (>1,000 through lactation. Moreover, it is important to note that mol/L) or the second week postpartum (≥1,400 mol/L) or both ofthe ﬁrst 2 wk postpartum relative to cows not SCK in either of the there was no signiﬁcant difference in time to pregnancy ﬁrst 2 wk.
between cows classiﬁed as SCK in either week relative Journal of Dairy Science Vol. 90 No. 6, 2007 EFFECT OF SUBCLINICAL KETOSIS ON REPRODUCTION to those classiﬁed as SCK in both weeks (P = 0.28; effect is unknown. Monensin (22 mg/kg of DM) de- creases days to ﬁrst ovulation and decreases the num- Despite reports on the association between energy ber of large follicles before 30 DIM with no change in balance and commencement of luteal activity, PR/AI, ﬁrst service conception risk (Tallam et al., 2003). Both and time to ﬁrst insemination and to pregnancy, the clinical and subclinical endometritis increases time to underlying mechanisms are still not fully clear. The pregnancy (LeBlanc et al., 2002). This association may results of the current analysis provide further evidence represent the link between monensin treatment and a of residual effects of health problems in the transition reduced daily probability of pregnancy (HR).
period on subsequent reproductive performance. Spe-ciﬁcally, SCK in early lactation is associated with im- paired reproductive performance 50 to 100 d later. Thisdelayed effect on reproductive traits has been attrib- Cows diagnosed as SCK in either of the ﬁrst 2 wk uted to increased risk of delayed commencement of ovu- postpartum, using a threshold of serum BHBA concen- lation (Butler, 2003). More recently, Leroy et al. (2006) tration ≥1,000 mol/L in the ﬁrst week postpartum or found that in vitro maturation of follicles in media con- ≥1,400 mol/L in the second week postpartum, were ditions analogous to those follicular ﬂuid concentra- 20% less likely to be pregnant at ﬁrst insemination.
tions of glucose and BHBA experienced during SCK The probability of pregnancy was reduced by 50% in signiﬁcantly reduced the ability of fertilized ova to be- cows experiencing SCK in both of the ﬁrst 2 wk. Using come morula and hatched blastocysts, indicating a di- time to pregnancy as the outcome, SCK in either the rect toxic effect of BHBA and NEFA on maturing ova.
ﬁrst 2 wk postpartum decreased the daily probability of The peak herd level prevalence of SCK has been re- pregnancy (HR) until approximately 165 d postpartum.
ported from the second to ﬁfth week postpartum (An-dersson and Emanuelson, 1984). Attempts have been made to understand factors that inﬂuence when the Financial support was provided by Elanco Animal peak level of SCK occurs within individual herds. One Health, the University of Guelph Doctor of Veterinary such study by Heuer et al. (2001) reported that season Science program, and Ve´toquinol. The technical assis- of calving and pen number (speciﬁcally management tance provided by Eleanor Robinson is greatly ap- practices such as time away from feed) explained 17% of the variation in milk acetone concentrations. Thus,application of this data would best identify herds at risk for reproductive failure rather than cows.
Similar to the present study, others reported no asso- Abe, N., I. J. Lean, A. Rabiee, J. Porter, and C. Graham. 1994. Effects of sodium monensin on reproductive performance of dairy cattle.
ciation of BCS at calving, or change in BCS early in II. Effects on metabolites in plasma, resumption of ovarian cy- lactation, with any reproductive parameters (Ruegg clicity and oestrus in lactating cows. Aust. Vet. J. 71:277–282.
and Milton, 1995). The distribution of BCS and BCS Andersson, L., and U. Emanuelson. 1984. An epidemiological study of hyperketonemia in Swedish dairy cows: Determinants and the loss in this data set was similar between animals that relation to fertility. Prev. Vet. Med. 3:449–462.
were pregnant to ﬁrst insemination and those that were Bauman, D. E., and W. B. Currie. 1980. Partitioning of nutrients not pregnant. Insufﬁcient variation in body condition during pregnancy and lactation: A review of mechanisms involv-ing homeostasis and homeorhesis. J. Dairy Sci. 63:1514–1529.
data has been previously cited as a partial explanation Butler, W. R. 2003. Energy balance relationship with follicular devel- for the inconsistent association between BCS and en- opment, ovulation and fertility in postpartum dairy cows. Livest.
ergy balance (Heuer et al., 2000).
Prod. Sci. 83:211–218.
Cook, N. B., W. R. Ward, and H. Dobson. 2001. Concentrations of A treatment effect of Na monensin on time to preg- ketones in milk in early lactation, and reproductive performance nancy has not been reported. Further, the CRC formu- of dairy cows. Vet. Rec. 148:769–772.
lation was not associated with alterations in commence- Dawuda, P. M., J. R. Scaife, J. S. Hutchinson, and K. D. Sinclair.
2002. Mechanisms linking under-nutrition and ovarian function ment of luteal activity, days to ﬁrst insemination, or in beef heifers. Anim. Reprod. Sci. 74:11–26.
days to conception, despite consistent reductions in se- Dufﬁeld, T. F., K. E. Leslie, D. Sandals, K. Lissemore, B. W. McBride, rum BHBA (Abe et al., 1994; Lean et al., 1994; Dufﬁeld J. H. Lumsden, P. Dick, and R. Bagg. 1999. Effect of a monensin-controlled release capsule on cow health and reproductive perfor- et al., 1999). No previous study, investigating the effect mance. J. Dairy Sci. 82:2377–2384.
of monensin, has included ketone information in the Dufﬁeld, T. F., D. Sandals, K. E. Leslie, K. Lissemore, B. W. McBride, J. H. Lumsden, P. Dick, and R. Bagg. 1998. Effect of prepartum same model. Inclusion of BHBA in the same model re- administration of monensin in a controlled-release capsule on moves the effect on PR/AI achieved through the im- postpartum energy indicators in lactating dairy cows. J. Dairy provement of energy status that monensin treatment Ferguson, J. D., D. T. Galligan, and N. Thomsen. 1994. Principal permits and assigns it to BHBA concentration. Thus, descriptors of body condition score in Holstein cows. J. Dairy Sci.
the underlying mechanism to explain this treatment Journal of Dairy Science Vol. 90 No. 6, 2007 WALSH ET AL.
Heuer, C., W. M. van Straalen, Y. H. Schukken, A. Dirkzwager, and Reist, M., D. K. Erdin, D. von Euw, K. M. Tschu¨mperlin, H. Leuen- J. P. T. M. Noordhuizen. 2000. Prediction of energy balance in a berger, H. M. Hammon, C. Morel, N. Ku¨nzi, and J. W. Blum.
high yielding dairy herd in early lactation: Model description and 2003a. Use of threshold serum and milk ketone concentrations precision. Livest. Prod Sci. 65:91–105.
to identify risk for ketosis and endometritis in high-yeilding dairy Heuer, C., A. Wangler, Y. H. Schukken, and J. P. Noordhuizen. 2001.
cows. Am. J. Vet. Res. 64:188–194.
Variability of acetone in milk in a large low-production dairy Reist, M., D. K. Erdin, D. von Euw, K. M. Tschu¨mperlin, H. Leuen- herd: A longitudinal case study. Vet. J. 161:314–321.
berger, H. M. Hammon, C. Morel, C. Philipona, Y. Zbinden, N.
Jolly, P. D., S. McDougall, L. A. Fitzpatrick, K. L. MacMillan, and Kunzi, and J. W. Blum. 2003b. Postpartum reproductive function: K. W. Entwistle. 1995. Physiological effects of undernutrition on Association with energy, metabolic and endocrine status in high postpartum anoestrus in cows. J. Reprod. Fertil. Suppl.
yielding dairy cows. Theriogenology 59:1707–1723.
Reist, M., A. Koller, A. Busato, U. Kupfer, and J. W. Blum. 2000.
Koller, A., M. Reist, J. W. Blum, and U. Ku ¨ pfer. 2003. Time empty First ovulation and ketone body status in the early postpartum and ketone body status in the early postpartum period of dairy period of dairy cows. Theriogenology 54:685–701.
cows. Reprod. Domest. Anim. 38:41–49.
Ruegg, P. L., and R. L. Milton. 1995. Body condition scores of Holstein Lean, I. J., M. Curtis, R. Dyson, and B. Lowe. 1994. Effects of sodium cows on Prince Edward Island, Canada: Relationships with yield, monensin on reproductive performance of dairy cattle. I. Effects reproductive performance, and disease. J. Dairy Sci. 78:552–564.
on conception rates, calving-to-conception intervals, calving-to- SAS Institute. 2004. SAS User's Guide: Statistics. 9th ed. SAS Inst.
heat and milk production in dairy cows. Aust. Vet. J. 71:273–277.
Inc., Cary, NC.
LeBlanc, S. J., T. F. Dufﬁeld, K. E. Leslie, K. G. Bateman, G. P. Keefe, Staples, C. R., W. W. Thatcher, and J. H. Clark. 1990. Relationship J. S. Walton, and W. H. Johnson. 2002. Deﬁning and diagnosing between ovarian activity and energy status during the early post- postpartum clinical endometritis and its impact on reproductive partum period of high producing dairy cows. J. Dairy Sci.
performance in dairy cows. J. Dairy Sci. 85:2223–2236.
Leroy, J. L. M. R., T. Vanholder, G. Opsomer, A. Van Soom, and A.
Stata Corporation. 2005. Stata Statistical Software: Release 9.1.
de Kruif. 2006. The in vitro development of bovine oocytes after Stata Corporation, College Station, TX.
maturation in glucose and β-hydroxybutyrate concentrations as- Tallam, S. K., T. F. Dufﬁeld, K. E. Leslie, R. Bagg, P. Dick, G. Vessie, sociated with negative energy balance in dairy cows. Reprod.
and J. S. Walton. 2003. Ovarian follicular activity in lactating Domest. Anim. 41:119–123.
Holstein cows supplemented with monensin. J. Dairy Sci.
Plym Forshell, K., L. Andersson, and B. Pehrson. 1991. The relation- ships between the fertility of dairy cows and clinical and biochemi- Villa-Godoy, A., T. L. Hughes, R. S. Emery, L. T. Chapin, and R. L.
cal measurements, with special reference to plasma glucose and Fogwell. 1988. Association between energy balance and luteal milk acetone. Zentralbl. Veterinarmed. A 38:608–616.
function in lactating dairy cows. J. Dairy Sci. 71:1063–1072.
Journal of Dairy Science Vol. 90 No. 6, 2007
THE EFFECTS OF ADVERTISING USING GLOBAL APPEAL VS CHINA- SPECIFIC APPEALS: A STUDY OF CHINESE CONSUMERS IN HONG KONG AND CHINA Fung Sze Man A Project Submitted in Partial Fulfillment of the Requirements for the Degree of Master of Arts in Communication Supervisor: Dr. Vivian Sheer
Curriculum PERSONAL INFORMATION: Name: Lynn B. Gerald, PhD, MSPH RANK/TITLE: Canyon Ranch Endowed Chair/Professor Health Promotion Sciences Business Address: University of Arizona Mel and Enid Zuckerman College of Public Health 1295 N. Martin Ave. Tucson, Arizona 85724 HOSPITAL AND OTHER (NON ACADEMIC) APPOINTMENTS: