e at SciVerse ScienceDirect

Theriogenology 79 (2013) 882–889

Contents lists availabl

Theriogenology

journal homepage: www.theriojournal .com

Evaluation of three synchrony programs for pasture-based dairy heifers

S. McDougall a,*, F.M. Rhodes b, C.W.R. Compton a

aCognosco, Anexa Animal Health, Morrinsville, New ZealandbHamilton, New Zealand

a r t i c l e i n f o

Article history:Received 3 June 2012Received in revised form 20 December 2012Accepted 23 December 2012

Keywords:HeiferSynchronyOvsynchCosynchProgesteroneProstaglandin

* Corresponding author. Tel.: þ64 7 889 5159; faxE-mail address: smcdougall@anexa.co.nz (S. McD

0093-691X/$ – see front matter � 2013 Elsevier Inchttp://dx.doi.org/10.1016/j.theriogenology.2012.12.01

a b s t r a c t

The objective was to evaluate the efficacy and economic benefits of three synchronyprograms in 1137 heifers from 10 pasture-based dairy herds. Heifers were randomlyassigned to one of three treatments within each herd on Day �13 (Day 0 ¼ start of thebreeding program). They were treated with: (1) PGF2a on Days �13 and �2, with AI afterdetection of estrus between Days 0 and 3 (Double PG); (2) GnRH, PGF2a, and GnRH onDays �9, �2, and 0, respectively, with placement of an intravaginal progesterone(P4)-releasing device between Days �9 and �2, and set time AI on Day 1 (GPG þ P4); or(3) same as the GPG þ P4 group but with the set time AI on Day 0 (Cosynch þ P4). PlasmaP4 concentrations were determined on Days �20 and �13 to determine pubertal status.The Cosynch þ P4 treatment had a higher (P < 0.05) conception rate to AI (57% vs. 47%vs. 48% for Cosynch þ P4, GPG þ P4, and Double PG, respectively), 21-day in-calf rate (76%vs. 72% vs. 63% for Cosynch þ P4, GPG þ P4, and Double PG), and a shorter medianinterval from the start of the breeding program to conception (0, 14, and 19 days forCosynch þ P4, GPG þ P4, and Double PG). Heifers that had reached puberty beforebreeding, compared with those that had not, had higher (P < 0.05) in-calf rates to AI (53%vs. 47%) at 21 days (74% vs. 64%) and at 42 days (91% vs. 84%). Pubertal status wasassociated with herd, breed, age, and body condition score at the start of mating (P <

0.05). A partial budget model demonstrated that, compared with the Double PG program,there was an economic benefit from the Cosynch þ P4 (mean, NZ$25.73; 95% confidenceinterval, 2.99–50.69), but not the GPG þ P4 program (mean, NZ$�0.65; 95% confidenceinterval, �21.87 to 21.58). We concluded that the Cosynch þ P4 program resulted in thehighest fertility and economic benefit of the three programs evaluated, and that repro-ductive response was affected by pubertal status.

� 2013 Elsevier Inc. All rights reserved.

1. Introduction

Synchronization of estrus and ovulation of dairy heifers,in combination with AI, facilitates improving the rate ofgenetic gain by shortening the generation interval andobtaining replacements from dams with the highestgenetic merit in a herd [1]. However, the use of AI in dairyheifers has historically been limited in New Zealand. Forexample, in 2006 and 2007 only 118,226 yearlings weremated using AI compared with 2,904,667 cows, from a total

: þ64 7 889 3681.ougall).

. All rights reserved.3

population of 5.3 million dairy cattle [2]. The practicalitiesof inseminating heifers after either daily estrous detectionor some form of synchrony might be barriers to uptake ofthe technology.

Various synchrony programs for dairy heifers havepreviously been evaluated in pasture-based systems [3–6].However, these programs generally included the use ofestradiol, which can no longer be used in food-producinganimals because of the European ban, or involved detec-tion of estrus. Therefore evaluation of new programs wasrequired, including those that did not require estrousdetection and were practical under the extensive pasture-based management systems of New Zealand.

S. McDougall et al. / Theriogenology 79 (2013) 882–889 883

Use of fixed-time insemination (FTAI) removes therequirement for estrous detection, but conception rateshave been reported to be significantly lower using FTAIafter two injections of PGF2a 10 to 14 days apart (“DoublePG”) than for dairy heifers inseminated on detection ofestrus [7]. These low conception rates were attributed tovariations in time to estrus and ovulation depending on thestage of the estrus cycle at PGF2a administration, and thelack of efficacy of PGF2a in prepubertal animals [8,9].

Optimal synchrony of estrus and ovulation requirescontrol of follicle development and luteal function [10].This can be achieved in dairy cows using programsinvolving GnRH injections 9 days apart with PGF2a 2 daysbefore the final GnRH treatment, with FTAI 16 hours afterthe final GnRH treatment (GPG or Ovsynch), but suchprograms used in heifers do not achieve conception ratescomparable with AI after detection of estrus, because ofa lack of synchrony of estrus [11,12]. Insemination mightalso occur coincident with the final GnRH treatment(Cosynch). The Cosynch program reduces the number ofhandlings of cattle required, but has been reported to resultin lower conception rates compared with Ovsynch [13].However, in beef heifers, a Cosynch program resulted inpregnancy rates superior to a Double PG program [14].Addition of progesterone (P4) to GPG programs alsoresulted in significantly increased conception rates afterFTAI in beef heifers [15], and improved synchrony of estrusin dairy heifers [16].

Heifers that are prepubertal at the commencement ofsynchronyprogramsare less likely to exhibit estrus [17], andhave lower conception and pregnancy rates than postpub-ertal heifers [18]. In prepubertal beef heifers, inclusion ofP4 in the synchrony program significantly increased sub-mission and pregnancy rates [17], and conception rates [19].

The aim of the current study was to evaluate the efficacyand economic benefits of three synchrony programs inheifers from pasture-based dairy herds. The programs wereGPG or Cosynch, both with the addition (þ) of P4, andDouble PG. It was hypothesized that more heifers wouldconceive to AI and be pregnant by Day 21 of the breedingprogram after a GPG þ P4 program than after Double PG,and that conception rate to AI would be noninferior afterCosynch þ P4 compared with GPG þ P4.

2. Materials and methods

2.1. Heifers and treatments

The study was conducted after approval from the Ani-mal Ethics Committee of AgResearch Ruakura, Hamilton,New Zealand.

Dairy heifers (N ¼ 1137) from 10 herds (mean � SD,114 � 26 per herd) were enrolled on one calendar day foreach herd. All herds were spring-calving and locatedwithin the Waikato region of New Zealand. Blood samples(10 mL) were drawn from the tail vein into an evacuatedglass tube containing lithium heparin as an anticoagulant(Vacutainer; Becton Dickson, Franklin Lakes, NJ, USA) onDays �20 and �13 (where Day 0 ¼ start of the breedingprogram) for evaluation of P4 concentration in plasma byRIA (Coat-a-Count; DPC Corporation, Los Angeles, CA,

USA). Heifers were defined as prepubertal if the P4concentration in both samples was <1.0 ng/mL, andpostpubertal if P4 concentration in one or both sampleswas �1.0 ng/mL.

On Day �13, the body condition score (BCS) of eachheifer was assessed on a one to 10 scale [20], and tail paintapplied to aid estrus detection. Heifers were randomlyassigned, within sequentially presented blocks of threeheifers, to one of three treatment groups. Different colortail paints were used for identification of each treatmentgroup, but all heifers within a herd were managed asa single management group.

The three treatments were:

� Double PG: Heifers were treated with cloprostenol(Ovuprost 500 mg im; Bomac Laboratories Ltd., ManukauCity, New Zealand) on Days �13 and �2, with AI uponestrus detection between Days 0 and 3 (N ¼ 380; 33%);

� GPG þ P4: Heifers were treated with an intravaginalP4-releasing device (CueMate; Bomac Laboratories Ltd.)from Days �9 to �2, and gonadorelin (Ovurelin 100 mgim; Bomac Laboratories Ltd.) on Day �9, 500 mg clop-rostenol im onDay�2, and 100 mg gonadorelin on Day 0,with FTAI on Day 1 (N ¼ 383; 34%); and

� Cosynch þ P4: Heifers were treated in the same manneras for GPG þ P4 but with FTAI coincident with the finalgonadorelin treatment (N ¼ 374; 33%).

The timing of the treatments was such that the secondcloprostenol treatment for the Double PG group coincidedwith the cloprostenol treatment of the other two groups.Treatments were given between 9:00 AM and 1:00 PM.

All heifers were assessed once daily from Days 0 to 3 forloss of tail paint while yarded for drafting for AI. Herdowners decided which of the heifers in the Double PGgroup were in estrus and were to be inseminated each daybased on removal of tail paint or their own observations onpasture. When a heifer had been inseminated, no moreestrous observations were conducted.

Experienced AI technicians attended each herd atapproximately midday on Days 0 to 3. In seven herds, onetechnician undertook all inseminations, whereas in theremaining three herds, two technicians undertook insem-inations. Where two technicians were used, they under-took inseminations on each day of the program. Heifers inthe Double PG group detected in estrus between Days 0 and3 were inseminated on the day of detection. Heifers in theGPG þ P4 group in estrus at the time of the second GnRHinjection (Day 0) were recorded as in estrus, but notinseminated until Day 1.

Frozen semen from 40 sires was used for AI in the study(one to four sires were used in each treatment group). Afterthe last AI on Day 3, bulls were placed with the heifers fora mean (�SD) total duration of 80� 10 days. The number ofbulls required for each management group was calculatedassuming that 50% of the heifers would conceive to firstservice and that those not conceiving would return toestrus over a 5-day period on average 21 days later. Suffi-cient bulls were introduced such that no more than threeservices per bull per day would occur on average.

S. McDougall et al. / Theriogenology 79 (2013) 882–889884

Pregnancy diagnosis was conducted in all herds(mean � SD) 69� 2 days and 109� 5 days after the start ofthe breeding program. In seven herds, additional preg-nancy diagnosis took place 139 � 6 days after the start ofthe breeding program. The stage of gestation, in days, wasestimated at each examination if a heifer was diagnosedpregnant.

The date of birth and breed of individual heifers wasretrieved from an electronic database (LIC, Hamilton, NewZealand). These data were used to calculate the age at thestart of the breeding program (mean � SD ¼ 14.3 � 0.4months), and to classify heifers as Friesian (N ¼ 575; 51%),Jersey (N ¼ 159; 14%), or other (N ¼ 403; 35%).

2.2. Definitions and statistical methods

There were six primary outcome variables in this study.The in-calf to AI rate was defined as the proportion ofenrolled heifers that conceived to AI. The conception rate toAI was defined as the proportion of heifers conceiving to AIand was calculated only for heifers that were submitted forAI between Days 0 and 3. It was possible that some heifersin the Double PG group might have conceived to the firstnatural mating after the introduction of bulls, but naturalmatings were not recorded, and so these heifers wereexcluded from this analysis. The 21-day and 42-day in-calfrates were defined as the proportion of enrolled heiferspregnant at 21 or 42 days from the start of the breedingprogram, respectively. The final in-calf rate was defined asthe proportion of heifers present at the final pregnancy testthat were pregnant at the final pregnancy test. Finally, themedian interval in days from the start of the breedingprogram to conception was calculated by survival analysismethods. For heifers in the GPG þ P4 group, the repro-ductive performance measures were calculated using thefirst day of AI as the start point (Day 1 in the study), so asnot to bias the results for this group compared with thosein the other two groups that started 1 day earlier (Day 0 inthe study).

Potential predictor variables evaluated were treatmentgroup, herd of origin, age in months at the start of breeding(categorized into quartiles), pubertal status at commence-ment of the study, breed category, observation of estrus,and BCS (categorized into quartiles).

Balance of treatment groups for age, breed, pubertalstatus, and BCS of heifers was compared using chi-squaretests. Differences between treatment groups in the cate-gorical outcome variables were compared using two-sidedPearson chi-square test with Holm adjustment for multipletesting [21] to control the overall type I error rate. Multi-variable logistic regression using a forward stepwise addi-tion method was used to model the effect of treatmentgroup, and other variables associated by univariate analysis(P < 0.2), on the outcomes. Treatment group and herd oforigin were fitted as fixed effects and forced into eachmodel. Other variables were tested for significance and aspotential confounders and included if their Wald testprobability value was P< 0.05, or if their inclusion changedthe coefficients for the treatment effects by 15% or more.Interactions between treatment groups and other signifi-cant variables were also tested. Preliminary final models

were checked by plotting the deviance residuals against thefitted values and the Cook statistics against the standard-ized leverages, and if no values were of concern, weredeclared final. The mean effects of the other nontreatmentvariables in the final models were also estimated todescribe the change in outcome (on a percentage scale)associatedwith that variable. The noninferiority hypothesiswas tested using differences between the means and theconfidence intervals of the differences estimated from thefinal model, with the other included variables set to theirmedian value. The effect on the time to conception oftreatment group and potential confounding variables weretested by Kaplan–Meier survival analysis using the log-ranktest, and survival curves plotted.

A variable of secondary interest was pubertal status asdefined by plasma P4 concentrations. Associations wereevaluated between pubertal status and herd of origin,breed, and age and BCS at Day �13 using chi-square anal-ysis with adjustment for multiple testing, as previouslydescribed. The pubertal status (post vs. prepubertal) ofheifers was modeled by multiple logistic regressionmodelsand effects of significant variables estimated as previouslydescribed.

2.3. Power statistics

The first hypothesis was that more heifers wouldconceive to AI and be pregnant by Day 21 of the breedingprogram after a GPG þ P4 program than after Double PG. Itwas assumed that submission rates would be 70% and100%, and conception to AI rates would be 60% and 55%, forthe Double PG and GPG þ P4 groups, respectively. Thus itwas expected that approximately 42% and 55% of heiferswould be pregnant to AI in the Double PG and GPG þ P4groups, respectively. To detect this 13% difference,approximately 250 heifers per treatment group wererequired (a ¼ 0.05; a ¼ 0.2). To allow for loss to follow-up,approximately 300 heifers per treatment group wereenrolled.

The second hypothesis was that pregnancy rates by Day21 would be noninferior for the Cosynch þ P4 comparedwith the GPG þ P4 group. It was assumed that theconception rate of the two groups would be the same (55%)and an acceptable difference between the two groupswould be 10% (a ¼ 0.05; b ¼ 0.2). This required 420 heifersper treatment group (Pass 2008; www.ncss.com).

2.4. Partial budget

A partial budget model was prepared to estimate theeconomic effect of synchronization of estrus of heifers usingthe GPGþ P4 or Cosynchþ P4 programs comparedwith theDouble PG program. The model was built using input dataand distributions from the current study, other publisheddata, or the opinions of the authors (Table 1), and themethod of stochastic Monte Carlo simulation in the Micro-soft Excel add-in @Risk (Version 5; Palisade Corporation,Ithaca, NY, USA). This simulation modeling takes account ofthe variability of model inputs by assigning each a specificdistribution, and generates a distribution for the outcomesfromwhich mean economic effects are estimated.

Table 1Assumptions used to develop a stochastic partial budget model (in New Zealand $) in the Microsoft Excel add-in @Risk to compare the economic benefit ofestrus synchrony with artificial breeding of pasture-grazed dairy heifers using either GPG þ P4a or Cosynch þ P4b with a Double PGc treatment program.

Income variables Mean Distribution

Extra days in milk GPG þ P4d 3.9 RiskNormal(3.9, 1.5)Extra days in milk Cosynch þ P4d 6.4 RiskNormal(6.4, 1.5)Mean milk production per day (kg milk solidse) 1.2 RiskNormal(1.2, 0.12)Milk payment next season ($ per kg milk solidse) 5.50 RiskUniform(5.65, 5.75)Percentage in calf to AI, Double PG 31% RiskLogistic(0.31, 0.0136)Percentage in calf to AI, GPG þ P4 47% RiskLogistic(0.47, 0.0136)Percentage in calf to AI, Cosynch þ P4 57% RiskLogistic(0.57, 0.0135)Percentage calves female 50% RiskBinomial(100, 0.5)Extra AI heifer calves, GPG þ P4d 8%Extra AI heifer calves, Cosynch þ P4d 13%Male calf value $20.00 RiskUniform(15, 25)AI calf value (4 days of age) $200.00 RiskUniform(150, 250)

Cost variables Mean

Drug and treatment costs per head

Drug costs, Double PG $7.36Drug costs, GPG + P4 $40.00Drug costs, Cosynch + P4 $40.00Veterinary fees, Double PG $6.00Veterinary fees, GPG + P4 $11.00Veterinary fees, Cosynch + P4 $11.00

Overhead costs per head Mean Distribution

Average number of heifers per group 50Number of heifer yardings, Double PG 5Number of heifer yardings, GPG þ P4 4Number of heifer yardings, Cosynch þ P4 3Grazier yarding fees $50.00 RiskUniform(25, 75)Percentage of graziers charging yarding fees 25% RiskTriang(0.2, 0.25, 0.3)Visit and yarding fees plus travel, Double PG $222.50Visit and yarding fees plus travel, GPG þ P4 $290.00Visit and yarding fees plus travel, Cosynch þ P4 $277.50Difference in overhead costs per head, GPG þ P4d $1.35Difference in overhead costs per head Cosynch þ P4d $1.10

Distribution is the @Risk formula used to model variability; otherwise the mean value was used. Income variables are shown.a PGF2 on Days �13 and �2 with AI on detection of estrus from Day 0 to 3 of the breeding program.b GnRH, PGF2 , and GnRH on Days �9, �2, and 0 of the breeding program, respectively, with insertion of an intravaginal progesterone-releasing insert for

7 days commencing on Day �9, and fixed time AI on Day 1.c GnRH, PGF2 , and GnRH on Days �9, �2, and 0 of the breeding program, respectively, with insertion of an intravaginal progesterone-releasing insert

for 7 days commencing on Day �9, and fixed time AI on Day 0.d Compared with Double PG.e Milk solids ¼ kg of protein plus kg of fat.

S. McDougall et al. / Theriogenology 79 (2013) 882–889 885

In seasonal calving systems with a fixed start and end oflactation, the maximum number of days in milk occurswhen an animal calves at the start of the calving program(having conceived at the start of the breeding program) andeach day of delay to conception results in one day less inmilk. Hence, the additional number of days in milk in thelactation after synchrony for the GPG þ P4 or Cosynchþ P4heifers, relative to the Double PG heifers, was modeled bysubtracting the interval from the start of breeding to thedate of conception from the duration of the breedingprogram for each heifer, and including the mean and SEMof the differences between treatment groups in the model.

The economic benefits were assessed in terms of extramilk production, calculated from the estimated value ofmilk solids (kg of milk protein plus kg of milk fat; the basisof payment in the New Zealand dairy industry), mean dailyproduction of milk solids, and extra days in milk, plus thevalue of additional AI calves, calculated from the estimatedpercentage of extra female AI calves born and the valueof these calves at 4 days of age. Additional costs were

calculated from estimated treatment costs based on drugcosts and vet fees, plus management costs because ofyarding of heifers.

The means and 95% confidence intervals of the distri-butions of the economic benefit of each comparison werereported and the treatment comparison declared signifi-cantly different if the 95% confidence interval of thedifference did not include zero. Spearman rank correlationcoefficients were examined to determine the variables withgreatest influence on the net economic benefit.

Data were managed in a custom-built Microsoft Accessdatabase (Microsoft, Redmond, WA, USA), and analysisconducted using R (Version 2.8.1; R Development CoreTeam, 2008). Statistical significance was declared for testresults with P < 0.05.

3. Results

Nine heifers were removed from the analysis becausethey were infertile (freemartins, N ¼ 6), pregnant and

S. McDougall et al. / Theriogenology 79 (2013) 882–889886

aborted within 1 day after treatment with PGF2a (N¼ 1), orwere missing at the final treatment with PGF2a (N ¼ 2),leaving data for analysis from 1137 heifers. Six heifers werenot present for the final pregnancy diagnosis and were notincluded in results for the final in-calf rate.

Mean BCS of heifers at enrolment was 4.6, with 90% inthe range of 4.0 to 5.5 (inclusive). Paired plasma sampleswere available from 1123 heifers; 671 heifers (60%) hadplasma P4 concentrations �1.0 ng/mL in one or bothsamples.

There were no differences between treatment groups inage category (P ¼ 0.71), breed category (P ¼ 0.81), pubertalstatus (P ¼ 0.87), or BCS category (P ¼ 0.92).

3.1. Detection in estrus

A total of 248 of 380 heifers (65%) in the Double PGgroupwere detected in estrus by farmers and submitted forAI over Days 0 to 3. The percentage of heifers detected inestrus in this group was greater for those that were post-compared with prepubertal (87.3% vs. 34.2%; P < 0.01).

On Day 0, 107 of 380 (28%), 183 of 383 (48%), and 178 of374 (48%) of heifers were observed at the time of yarding tohave their tail paint disturbed for the Double PG, GPG þ P4and Cosynch þ P4 groups, respectively (P < 0.001).

3.2. Conception to AI

Results for the in-calf to AI rate and conception rate to AIfor the three treatment groups are summarized (Table 2).For both outcomes, heifers in the Cosynch þ P4 groupachieved higher values than those in the GPG þ P4 group(P < 0.05). In-calf to AI rate was also higher (P < 0.01) andconception to AI tended to be higher (P ¼ 0.069) for heifersin the Cosynchþ P4 group than for heifers in the Double PGgroup.

The Cosynchþ P4 group was declared noninferior to theGPG þ P4 group because the lower 95% confidence intervalof the difference in conception rate to AI was greater than�0.10, the a priori defined acceptable difference (difference(Coynsch þ P4) � (GPG þ P4) ¼ 0.11; 95% confidence interval,0.02–0.19). At the univariate level, the in-calf rate to AI

Table 2Number (%) of dairy heifers conceiving to AI or pregnant after three treatment p

Treatmentgroup

In-calf to AIa Conceptionrate to AIb

21-Dayin-calf ratec

Double PGg 119/380h (31.3) 119/248h,i (48.0) 238/380h (62GPG þ P4j 180/383i (47.0) 180/383h (47.0) 274/383i (71Cosynch þ P4k 213/374l (57.0) 213/374i (57.0) 284/374i (75

a Proportion of all heifers conceiving to AI.b Proportion of heifers artificially bred conceiving to AI.c Proportion of all heifers pregnant after 21 days of the breeding program.d Proportion of all heifers pregnant after 42 days of the breeding program.e Proportion of heifers examined that were pregnant at the end of the breedinf Median interval from start of breeding to conception.g PGF2 on Days �13 and �2 with AI on detection of estrus from Day 0 to 3 oh,i, and l Within a column, means without a common superscript differed (P <j GnRH, PGF2 , and GnRH on Days �9, �2, and 0 of the breeding program, res

for 7 days commencing on Day �9, and fixed time AI on Day 1.k GnRH, PGF2 , and GnRH on Days �9, �2, and 0 of the breeding program, resp

7 days commencing on Day �9, and fixed-time AI on Day 0.

varied among herds and was higher in pubertal comparedwith prepubertal heifers (53% vs. 47%; P ¼ 0.05), but onlytreatment group and herd were significant in the finalmodel. Conception rate to AI also varied among herds(range, 37%–63%) and was higher in heifers fromwhich tailpaint was removed 1 to 48 hours before AI compared with0 to 48 hours after AI (266/497 [53.5%] vs. 73/186 [39.2%];P < 0.001).

3.3. Pregnancy rates

Results for the 21-day, 42-day, and final in-calf rates forthe three treatment groups are summarized (Table 2).Heifers in the Cosynch þ P4 group had higher 21- and42-day in-calf rates than heifers in the Double PG group(P < 0.05), but did not differ from those in the GPG þ P4group. The 21-day in-calf rate varied among herds (range,53%–80%). In the final models, heifers that had reachedpuberty before breeding, compared with those that hadnot, had higher in-calf rates at 21 days (74% vs. 64%; P <

0.001) and 42 days (91% vs. 84%; P ¼ 0.03).The final in-calf rates were not different between

treatment groups, but varied among herds. For all finalmodels, there were no significant herd by treatmentinteractions or pubertal status by treatment interactions.

Survival curves for the interval from the start of thebreeding program to conception differed between thetreatment groups (P < 0.001; Fig. 1). The median intervalfrom the start of breeding to conceptionwas shorter for theheifers in the Cosynch þ P4 than the GPG þ P4 and DoublePG groups (P< 0.05), and tended to be shorter for heifers inthe Cosynchþ P4 group than the GPGþ P4 group (P¼ 0.06;Table 2).

3.4. Pubertal status

Overall, 671 of 1123 heifers (60%) were categorized aspostpubertal, but there was considerable variation amongherd of origin (range, 9%–93%). Factors associated withpubertal status were herd of origin (P < 0.001), breedcategory (P< 0.001), age (P¼ 0.002), and BCS at the start ofthe breeding program (P < 0.001; Table 3).

rotocols for synchronization of estrous and ovulation.

42-Dayin-calf rated

Finalin-calf ratee

Interval toconception (days)f

.6) 313/380h (82.4) 347/376 (92.3) 19h

.5) 335/383h,i (87.5) 352/381 (92.4) 14h

.9) 333/374i (89.0) 358/374 (95.7) 0i

g program.

f the breeding program.0.05).pectively, with insertion of an intravaginal progesterone-releasing insert

ectively, with insertion of an intravaginal progesterone-releasing insert for

Fig. 1. Kaplan–Meier survival curves of cumulative time to conception fromthe start of the breeding program (Day 0) for dairy heifers treated with threesynchronization protocols: (1) PGF2a on Days �13 and �2, with AI afterdetection of estrus between Days 0 and 3 (Double PG); (2) GnRH, PGF2a, andGnRH on Days �9, �2, and 0, respectively, with placement of an intravaginalprogesterone (P4)-releasing device between Days �9 and �2, and set timeAI on Day 1 (GPG þ P4); or (3) same as the GPG þ P4 group but with the settime AI on Day 0 (Cosynch þ P4). There were differences among treatmentgroups (P < 0.001).

S. McDougall et al. / Theriogenology 79 (2013) 882–889 887

3.5. Partial budget

The Cosynch þ P4 program had an economic benefitcompared with the Double PG program, but there was nobenefit from the GPG þ P4 program (Table 4). Additionalincome from greater milk yield because of earlier calvingand more AI calves in the Cosynch þ P4 compared with theDouble PG program accounted for the difference betweenthe two groups. Costs for the two P4-based treatments werevery similar. The variables most highly correlated with thenet benefit of the Cosynch þ P4 program were the addi-tional days in milk in the subsequent lactation (rs ¼ 0.83)through earlier conception pattern, daily milk production(rs ¼ 0.35), and value of an artificially-sired calf (rs ¼ 0.3).

Table 3Percentage of heifers (N ¼ 1123) that were postpubertal, based onprogesterone concentrations in two plasma samples, classified by breed,age, and body condition score at the start of the breeding program.

Variable Level Pubertal (%)

Breed Friesian 45.1a

Jersey 86.7b

Other 70.0c

Age (mo) 12.2-14.2 52.5a

14.3 59.1a

14.5 64.0b,c

14.8-15.9 65.9b,c

Body condition score 3.0-4.0 37.9a

4.5 66.8b

5.0 72.3b

5.5-7.5 74.3b

a–c Within an end point, means without a common superscript differed(P < 0.05).

4. Discussion

This study demonstrated that synchronization of heifersusing a combination of GPG and P4 with FTAI coincidentwith the final GnRH treatment (Cosynch þ P4) resulted insuperior reproductive and economic performance com-pared with a similar program but with FTAI delayed byapproximately 24 hours or to a Double PG program.

The conception rate to AI of 57% for heifers in theCosynch þ P4 group in the current study was comparablewith the 53% reported after a previous synchrony programwith FTAI in New Zealand [6], and was superior to resultsfor the GPG program in Holstein heifers in the United Stateswhen heifers detected in estrus before the final GnRH wereinseminated early and overall conception rates were 46%[11]. The 21-day in-calf rate of 76% for the Cosynch þ P4heifers was also better than the 32-day pregnancy rate of60% reported for dairy heifers in Florida after an 8-dayCosynch program using P4 inserts from Days 0 to 5 [22].The higher conception rates to AI in the Cosynch þ P4group compared with heifers in the GPG þ P4 group inthe current study were comparable with the finding ofLima et al. [23], in which a Cosynch þ P4 protocol resultedin improved reproductive outcomes compared with anOvsynch þ P4 protocol in dairy heifers in Florida.

In the current study, heifers observed to have lost tailpaint during the 48 hours before AI had higher conceptionrates to AI than those with loss of tail paint during the 48hours after AI. The interval from the end of a synchronyprogram to estrus was affected by the size of the dominantfollicle at the end of the program, with a larger dominantfollicle associated with a shorter interval to estrus [24]. It isnot clear if the shorter interval from the end of thesynchrony program to estrus was also associated witha shorter interval to ovulation. After a synchrony program,the interval between onset of estrus and ovulationappeared to be reasonably consistent at 1.3 days in a rangeof synchrony systems [25]. Thus it is likely that heifers inthe current study that were in estrus earlier relative to Day0, also ovulated sooner. This might partly explain the higherconception rates in the Cosynch þ P4 program comparedwith the GPG þ P4 program. In the latter program, nearlyhalf of the heifers had been detected in estrus by Day 0, yetby designwere not inseminated until 24 hours later. So, forthis group of heifers, inseminationmight have occurred toolate for optimum fertility. The interval between estrus andovulation has been reported as 38.5 � 3.0 hours (mean �SEM) in dairy heifers [26]. Because sperm capacitationtakes 4 to 12 hours, logically insemination should occur12 to 18 hours before ovulation to ensure functional spermare present at the time of ovulation [27]. In one study,heifers in estrus in the 24 hours preceding FTAI had higherconception rates than those in estrus 25 to 40 hours beforeFTAI (76% vs. 50%) [28]. In another study, conception ratesto FTAI were optimized when insemination occurred 6 to22 hours before ovulation [29].

Of the heifers in the Double PG group, 65% weredetected in estrus between 2 and 6 days after the finaltreatment with PGF2a. This appeared to be lower than insome previous reports, in which 86% were detected inestrus over 5 days after the final treatment [30]. This lower

Table 4Partial budget, in New Zealand dollars per heifer, on the use of two programs for synchronization of estrous compared with a double PGF2 program.

Treatment Group Additional income Additional costs Net benefit

Extra milk production Value of AI calves Total Treatment costs Management costs Total Mean 95% CI

GPG þ P4a 26.68 14.40 41.08 40.36 1.35 41.71 �0.65 �21.87 to 21.58Cosynch þ P4b 43.78 23.40 67.18 40.36 1.10 41.46 25.73 2.99–50.69

Abbreviation: CI, confidence interval.a GnRH, PGF2 , and GnRH on Days �9, �2, and 0 of the breeding program, respectively, with insertion of an intravaginal progesterone-releasing insert

for 7 days commencing on Day �9, and fixed time AI on Day 1.b GnRH, PGF2 , and GnRH on Days �9, �2, and 0 of the breeding program, respectively, with insertion of an intravaginal progesterone-releasing insert for

7 days commencing on Day �9, and fixed-time AI on Day 0.

S. McDougall et al. / Theriogenology 79 (2013) 882–889888

detection rate might have been related to lower sensitivityof detection of estrus by the observers, a 1-day shorterperiod of observation, or because only 60% of heifers in thecurrent study had reached puberty when the programstarted.

Pubertal status had a significant effect on reproductiveoutcomes in the current study; the percentage of heifersthat were postpubertal at the start of the study decreasedwith lower BCS and younger age at commencement of theprogram. In a previous study, the risk of conception to firstservice was not associated with withers height, age, BCS, orbody weight, but the proportion of heifers that wereprepubertal was not defined and it is likely that mostheifers were postpubertal, hence the biophysical measuresmight have been of lesser importance [31]. In the currentstudy, there were significant effects of herd on all outcomevariables, likely associated with the percentage of animalsthat were prepubertal in each herd. At a practical level, ourfindings demonstrated that optimal outcomes aftersynchrony programs will be achieved in heifers that arepostpubertal, and as such should be �14.5 months old andhave a BCS >4.0.

In addition to improved reproductive performance, theCosynch þ P4 program had a significant economic benefitcompared with the Double PG program. One potentialbenefit of use of AI in heifers is a shortening of the gener-ation interval with an associated faster rate of genetic gain.In a previous assessment of the economic benefits of arti-ficially inseminating heifers with or without synchrony, theoptimal economic decision was to select high genetic meritheifers and undertake AI on detection of estrus [1].However, in that analysis, the main benefit modeled wasthe value of the improved genetics of the calves, and itignored any potential benefits that synchrony might havein terms of earlier calving of the heifers with resultantincreased lactation length. The current study analyzed thereturns in terms of extra value associated with AI-siredcalves, and additional production in the first lactation.The economic benefit of the Cosynch þ P4 program wasbecause of the 100% submission rate and the higherconception rate that resulted in more AI-sired calves andearlier calving, and thus a longer first lactation lengthcompared with the other programs. Although the DoublePG programwas less expensive, the poorer pregnancy ratesresulted in poorer economic outcomes relative to theCosynch þ P4 treatment. In populations of heifers withfewer prepubertal animals, the Double PG systemmight bemore effective and less economically disadvantaged.

4.1. Conclusions

Synchrony of dairy heifers with GPG þ P4 and with FTAIat the time of the final GnRH injection (Cosynch þ P4)resulted in the highest biological and economic outcomesof the three programs tested. Pubertal status had a signifi-cant effect on reproductive performance, irrespective ofsynchrony program.

Acknowledgments

The contribution of the herd owners, staff, and thegraziers is gratefully acknowledged. The support of thetechnical team of Laura Clausen, Catherine Sheehan, AimeeWilson, and Cathy Yanez is also gratefully acknowledged.Progesterone RIA were undertaken by Angela Sheehan ofDairyNZ. Treatments and pregnancy testing were under-taken by the authors and by Dr Megan Clements andDr Geoff Plant of Anexa Animal Health.

The funding and product support provided by BayerAnimal Health New Zealand Ltd. and by DairyNZ via anOn-farm innovation fund grant (OF811) is acknowledged.Semen was provided by LIC, who were also helpful inorganizing inseminators for the study.

References

[1] Rendel JM, Xu ZZ, Burton LJ. Artificial insemination of yearlingheifers: an economic analysis. Proc Soc Dairy Cattle Vet NZVA 1998:255–62.

[2] Anonymous. New Zealand dairy statistics 2010/11. Hamilton, NZ:LIC and DairyNZ; 2011.

[3] Macmillan KL, Taufa VK, Day AM. Combination treatments forsynchronising oestrus in dairy heifers. Proc NZ Soc Anim Prod 1993;53:267–70.

[4] Hanlon DW, Williamson NB, Wichtel JJ, Steffert IJ, Craigie AL,Pfeiffer DU. The effect of estradiol benzoate administration onestrous response and synchronized pregnancy rate in dairy heifersafter treatment with exogenous progesterone. Theriogenology1996;45:775–85.

[5] Hanlon DW, Williamson NB, Wichtel JJ, Steffert IJ, Craigie AL,Pfeiffer DU. Ovulatory responses and plasma luteinizing hormoneconcentrations in dairy heifers after treatment with exogenousprogesterone and estradiol benzoate. Theriogenology 1997;47:963–75.

[6] Xu ZZ, Burton LJ. Reproductive performance of dairy heifers afterestrus synchronization and fixed-time artificial insemination. J DairySci 1999;82:910–7.

[7] Macmillan KL, Curnow RJ, Morris GR. Oestrus synchronisation witha prostaglandin analogue: 2. Systems in maiden heifers. NZ Vet J1978;26:93–103.

[8] Refsal KR, Seguin BE. Effect of stage of diestrus and number ofcloprostenol (ICI 80, 996) injections on intervals to estrus, LH peakand ovulation in heifers. Theriogenology 1980;14:37–48.

S. McDougall et al. / Theriogenology 79 (2013) 882–889 889

[9] Tanabe TY, Hann RC. Synchronized estrus and subsequent concep-tion in dairy heifers treated with prostaglandin F2a. I. Influence ofstage of cycle at treatment. J Anim Sci 1984;58:805–11.

[10] Patterson DJ, Kojima FN, Smith MF. A review of methods tosynchronize estrus in replacement beef heifers and postpartumcows. J Anim Sci 2003;81:E166–77.

[11] Schmitt EJ, Diaz PT, Drost M, Thatcher WW. Use of a gonadotropin-releasing hormone agonist or human chorionic gonadotropin fortimed insemination in cattle. J Anim Sci 1996;74:1084–91.

[12] Pursley JR, Wiltbank MC, Stevenson JS, Ottobre JS, Garverick HA,Anderson LL. Pregnancy rates per artificial insemination for cowsand heifers inseminated at a synchronized ovulation or synchro-nized estrus. J Dairy Sci 1997;80:295–300.

[13] Pursley JR, Mee MO, Wiltbank MC. Synchronisation of ovulation indairy cows using PGF2a and GnRH. Theriogenology 1995;44:915–23.

[14] Dahlen CR, Lamb GC, Zehnder CM, Miller LR, DiCostanzo A. Fixed-time insemination in peripuberal, lightweight replacement beefheifers after estrus synchronization with PGF2alpha and GnRH.Theriogenology 2003;59:1827–37.

[15] Martinez MF, Kastelic JP, Adams GP, Cook B, OlsonWO, Mapletoft RJ.The use of progestins in regimens for fixed-time artificial insemi-nation in beef cattle. Theriogenology 2002;57:1049–59.

[16] Rivera H, Lopez H, Fricke PM. Use of intravaginal progesterone-releasing inserts in a synchronization protocol before timed AIand for synchronizing return to estrus in Holstein heifers. J Dairy Sci2005;88:957–68.

[17] Lucy MC, Billings HJ, Butler WR, Ehnis LR, Fields MJ, Kesler DJ, et al.Efficacy of an intravaginal progesterone insert and an injection ofPGF2alpha for synchronizing estrus and shortening the interval topregnancy in postpartum beef cows, peripubertal beef heifers, anddairy heifers. J Anim Sci 2001;79:982–95.

[18] Richardson AM, Hensley BA, Marple TJ, Johnson SK, Stevenson JS.Characteristics of estrus before and after first insemination andfertility of heifers after synchronized estrus using GnRH, PGF2, andprogesterone. J Anim Sci 2002;80:2792–800.

[19] McGowan MR, Carroll CL, Davies FJ. Fixed-time insemination of Bosindicus heifers following the use of Syncro-Mate-B (Smb) tosynchronize estrus. Theriogenology 1992;37:1293–300.

[20] Roche JR, Dillon PG, Stockdale CR, Baumgard LH, VanBaale MJ.Relationships among international body condition scoring systems.J Dairy Sci 2004;87:3076–9.

[21] Holm S. A simple sequentially rejective multiple test procedure.Scand J Stat 1979;6:65–70.

[22] Rabaglino MB, Risco CA, Thatcher MJ, Kim IH, Santos JEP,Thatcher WW. Application of one injection of prostaglandin F2a inthe five-day Co-Synch þ CIDR protocol for estrous synchroniza-tion and resynchronization of dairy heifers. J Dairy Sci 2010;93:1050–8.

[23] Lima FS, Ayres H, Favoreto MG, Bisinotto RS, Greco LF,Ribeiro ES, et al. Effects of gonadotropin-releasing hormone atinitiation of the 5-d timed artificial insemination (AI) programand timing of induction of ovulation relative to AI on ovariandynamics and fertility of dairy heifers. J Dairy Sci 2011;94:4997–5004.

[24] Savio JD, Thatcher WW, Badinga L, de la Sota RL, Wolfenson D.Regulation of dominant follicle turnover during the oestrous cyclein cows. J Reprod Fertil 1993;97:197–203.

[25] Martinez MF, Adams GP, Kastelic JP, Bergfel DR, Mapletoft RJ.Induction of follicular wave emergence for estrus synchronizationand artificial insemination in heifers. Theriogenology 2000;54:757–69.

[26] Saumande J, Humblot P. The variability in the interval betweenestrus and ovulation in cattle and its determinants. Anim Reprod Sci2005;85:171–82.

[27] Vishwanath R, Nebel RL, McMillan WH, Pitt CJ, Macmillan KL.Selected times of insemination with microencapsulated bovinespermatozoa affect pregnancy rates of synchronized heifers. Ther-iogenology 1997;48:369–76.

[28] Smith RD, Pomerantz AJ, Beal WE, McCann JP, Pilbeam TE, Hansel W.Insemination of Holstein heifers at a preset time after estrous cyclesynchronization using progesterone and prostaglandin. J Anim Sci1984;58:792–800.

[29] Tenhagen BA, Kuchenbuch S, Heuwieser W. Timing of ovulationand fertility of heifers after synchronization of estrus withGnRH and prostaglandin F-2 alpha. Reprod Domest Anim 2005;40:62–7.

[30] Kaim M, Rosenberg M, Folman Y. Management of reproduction indairy heifers based on the synchronisation of estrous cycles. Ther-iogenology 1990;34:537–47.

[31] Donovan GA, Bennett FL, Springer FS. Factors associated with firstservice conception in artificially inseminated nulliparous Holsteinheifers. Theriogenology 2003;60:67–75.