| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Bovine Research Australasia, PO Box 660, Camden 2570, NSW, Australia
2 EpiCentre, Institute of Veterinary, Animal, and Biomedical Sciences, Massey University, Palmerston North, New Zealand
Corresponding author: A. R. Rabiee; e-mail: ahmadr{at}dairydocs.com.au.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Pregnancy rates for Ovsynch programs did not differ significantly from those with natural breeding programs [predicted Bayesian relative risk (RR) = 1.04, 95% Bayesian credible interval = 0.36 to 3.23]. Results of Ovsynch vs. PGF2
programs showed that the risk of conception (predicted Bayesian RR = 0.89, 95% Bayesian credible interval = 0.31 to 2.64), and pregnancy rates predicted Bayesian RR = 1.11, 95% Bayesian credible interval = 0.61 to 2.13) did not differ significantly between the Ovsynch group and cows in the PGF2 group. Comparisons between Ovsynch and Select Synch demonstrated that the risk of conception (predicted Bayesian RR = 0.94, 95% Bayesian credible interval = 0.52 to 1.59), and pregnancy rates (predicted Bayesian RR = 1.08, 95% Bayesian credible interval = 0.38 to 3.09) did not differ significantly between the 2 groups. Examination of Ovsynch vs. modified Ovsynch programs showed that the risk of pregnancy in cows synchronized with modified Ovsynch was similar to those treated with Ovsynch (predicted Bayesian RR = 0.89, 95% Bayesian credible interval = 0.71 to 1.12).
Meta-analyses identified that the conception and pregnancy rates obtained with the prostaglandin, Select Synch, and modified Ovsynch (including presynch and CoSynch) programs were comparable with the Ovsynch program. Modifications to the Ovsynch program such as presynchronization and timed artificial insemination at the time of second GnRH injection (CoSynch) may be an alternative for reproductive management of dairy herds where detection of estrus is less than optimal. The findings of this study demonstrate that the Ovsynch program could benefit dairy operations because it allows for timed artificial insemination of lactating cows without detection of estrus. There was, however, little or no significant improvement in pregnancy rates using Ovsynch over other programs and the costs of labor and hormone administration should be considered when selecting this form of reproductive technology for routine use.
Key Words: Ovsynch • dairy cattle • meta-analysis • Bayesian statistics
Abbreviation key: CI = confidence or credible interval, ECP = estradiol cypionate, MCMC = Markov chain Monte Carlo, RR = relative risk, TAI = timed artificial insemination.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
-induced luteolysis for estrus synchronization, low rates of estrus detection persist. Breeding on observed estrus is more effective in achieving conception than is timed AI (TAI) after PGF2 administration (Fetrow and Blanchard, 1987; Stevenson et al., 1987).
The variability in onset of ovulation after the end of a synchrony will determine whether an animal can be inseminated at prearranged times. To achieve maximum precision in ovulation synchrony, it is necessary to have a recently selected dominant follicle present by the completion of the treatment program. New wave emergence needs to be synchronized during treatment, because both the stage of follicular wave and the duration of dominance cause variation in the duration of the follicular phase. Gonadotropin-releasing hormone has been used to cause predictable new wave emergence in some cattle synchronization protocols, that is, either in combination with PGF2 (Pursley et al., 1995, 1997a,Pursley et al., b; Twagiramungu et al., 1995; Schmitt et al., 1996) or as part of a progesterone regimen (Ryan et al., 1995).
Systematic breeding programs aim to optimize reproductive management on the dairy farm by eliminating the need to detect estrus before breeding (DeJarnette et al., 2001). Incorporation of ovulation synchronization in dairy herd reproductive management programs allows producers to minimize labor requirements for estrus detection while improving overall reproductive performance (Pankowski et al., 1995). New technology has advanced the development of systems to improve pregnancy rates by synchronizing estrus and follicle development with regression of the corpus luteum, precisely controlling the time of ovulation to provide a fixed timed insemination and improve embryo survival (Thatcher et al., 2000, 2001a,b, 2002).
Several synchrony programs have been developed in recent years based on timed insemination. One of these programs, Ovsynch (Pursley et al., 1997a,b), is claimed to be more successful for the insemination of cows at fixed time for first service. The Ovsynch program consists of 2 injections of GnRH, 7 d before, and 48 h after an injection of PGF2. Cows are inseminated 16 to 25 h after the second injection of GnRH. This system synchronizes follicle maturation with regression of the corpus luteum before the GnRH-induced ovulation and timed insemination. Several studies during the past few years have compared the Ovsynch program with other synchrony programs such as double PGF2 (Pursley et al., 1997a; de la Sota et al., 1998; Keister et al., 1999; Stevenson et al., 1996, 1999, 2000; Cartmill et al., 2001a), progesterone-based programs (Geary et al., 1998; Williams et al., 2002), Select Synch (Burke et al., 1996; Stevenson et al., 1999, 2000; Cartmill et al., 2001b), Heat Synch (Bartolome et al., 2002; Pancarci et al., 2002), and natural breeding (Pursley et al., 1997b; Aréchiga et al., 1998; Keister et al., 1999; Cordoba and Fricke, 2001).
Meta-analysis is a technique for combining data from several trials, and is particularly useful when there is conflict in the literature and when sample sizes of individual studies are insufficient to detect a statistically significant result (L’Abbe et al., 1987; Peto, 1987). The objective of this study was to use meta-analytical methods to evaluate the efficacy of the Ovsynch program in improving conception and pregnancy rates compared with other synchrony programs in lactating dairy cows. This is an epidemiological approach to evaluate the efficacy of synchrony programs currently used in commercial dairy farms around the world. It is not the intention of this paper to provide physiologic or pharmacologic explanations for the success or failure of a particular program, as there are several comprehensive reviews on Ovsynch and other synchrony programs that provide these details (Thatcher et al., 2000; 2001a,b; 2002).
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Synchrony Programs
The synchrony programs investigated and compared in this study are described below.
Ovsynch (GnRH-PGF2-GnRH).
Gonadotropin-releasing hormone (GnRH) injection followed by an injection of PGF2 7 d later, and a second GnRH injection given 48 h after PGF2.
Natural breeding.
Animals visually monitored for estrus and insemination on detection of estrus; or standard reproductive management for each trial, which consisted of mating using natural service bulls, or AI after detection of estrus aided by the use of Kamar heatmount detectors (Immucell, Portland, OR) or tail paint. Some programs included the occasional use of PGF2 and GnRH when determined to be appropriate by the herd veterinarian or manager.
PGF2.
Single, double, and triple PGF2 injection 11 to 14 d apart, followed by insemination at the subsequently detected estrus that may have followed each injection.
Select Synch (GnRH-PGF2).
Gonadotropin-releasing hormone injection followed in 7 d with an injection of PGF2.
Heat Synch (GnRH-PGF2-ECP).
Gonadotropin-releasing hormone injection followed in 7 d by an injection of PGF2 and estradiol cypionate (ECP) injection given 24 h after PGF2.
Modified Ovsynch.
Includes presynch Ovsynch (1 or 2 injections of PGF2 given 14 d apart, with the first injection of the Ovsynch given 12 to 14 d after the second PGF2 or injections of PGF2 and GnRH 7 d apart), or presynchrony using controlled internal drug release insertion 7 d before the first injection of the Ovsynch program, and other minor modifications such as insemination at the time of second GnRH injection (CoSynch) and varying doses of GnRH and PGF2.
We defined pregnancy rate as the proportion of treated or control cows that became pregnant by the date after intervention defined for each trial by those researchers. We defined conception rate as the proportion of inseminated cows only that became pregnant in either group. Overall pregnancy rate was the proportion of cows that became pregnant for cows that entered the trial by the end of breeding season or program as defined for each trial.
Trials Included/Excluded in the Study
Trials were included or excluded in the study based on criteria developed by the first 2 authors. Trials included in the analysis were those in which: (1) an adequate description of the randomization process was provided, (2) animals were lactating at the time of study enrollment, and (3) sufficient data were included to determine rates of conception and pregnancy. Of all the papers describing the results of an Ovsynch program, only those trials that met the defined inclusion criteria were included in the meta-analyses described. Published data in journal papers were crosschecked with conference papers to avoid repetition in presented data. All trials that compared Ovsynch with other synchrony programs, including those not considered for the meta-analysis, are listed in Table 1
. Trials were ineligible for inclusion in the analysis if they were not randomized clinical trials (n = 6), if the trial was compromised with supplementary treatment (n = 8), if the subjects were beef (n = 5) or dairy heifers (n = 7) rather lactating dairy cows, or if the paper provided insufficient data on the variables being measured (n = 12).
|
administration, breeding type (at detected estrus or TAI). Where the papers provided results from different farms, animals, or series of studies, these were considered separate trials. In some cases, information was only extracted from first or second round of synchrony programs. Results reported for each outcome were examined and, where appropriate, pooled to give a summary estimate of relative risk and a 95% credible interval (CI).
Statistical Analyses
Conception rates, individual service pregnancy rates and overall pregnancy rates from all trials selected for inclusion were used to calculate the relative risk (RR) of conception and pregnancy in Ovsynch-treated cows, compared with other synchrony programs. Variations among trial-level relative risks were assessed using a 
2 test of heterogeneity. In this case, the null hypothesis was that the effect of treatment was the same across k trials and the null hypothesis was rejected if the heterogeneity test statistic was greater than a critical value that separated the upper 10% of a 2 distribution with (k – 1) degrees of freedom. We used an level of 0.10 because of the relatively poor power of the 2 test to detect heterogeneity among small numbers of trials (Egger and Smith, 2001).
Heterogeneity of results among trials was quantified using the I2 statistic, (Higgins et al., 2003). The I2 statistic describes the percentage of total variation across studies that is due to heterogeneity rather than chance. Where Q is the 2 heterogeneity statistic and k is the number of trials, I2 was calculated as:
 | ([1]) |
Uncertainty intervals for I2 (dependent on Q and k) were calculated using the method described by Higgins and Thompson (2002). Negative values of I2 were put equal to zero, consequently I2 lay between 0 and 100%. A value greater than 50% may be considered substantial heterogeneity.
To estimate the effect of an Ovsynch synchrony program compared with other programs on reproductive outcomes, we used Bayesian meta-analysis models (Domenici et al., 1999). Here, the observed number of events in the control group of each trial, OCi, was assumed to be from a binomial distribution defined by parameters 
Ci (representing the underlying probability of a positive reproductive outcome in the control group) and nCi (representing the number of subjects in the control group). The observed number of events in the treatment group, OTi, was similarly defined by parameters,pTi and nTi:
 | ([2]) |
 |
Using a fixed-effects Bayesian approach, we assumed that the log of the relative risk of a positive reproductive outcome following treatment, log (Ti /Ci), followed a normal distribution with mean 
and variance proportional to the number of subjects enrolled in the trial. Within the Bayesian paradigm, an uninformed normal prior centered at zero was assigned to .
To account for heterogeneity among trial results, the fixed-effects model was extended to a random-effects model where the log of Ci was assumed a function of individual trial-level effects 
i and the log of Ti assumed to be a function of individual trial-level and treatment effects (i + i).
 | ([3]) |
In this case, the effect of treatment, i, was parameterized as having a normal distribution with mean µ and variance 
2. This allowed the effect of treatment observed across i trials (i) to vary around a "true" mean (µ) with variance 2. An uninformed normal prior centered at zero was used for µ and an uninformed inverse gamma distribution used for 2. An uninformed normal prior centered at zero was assigned to the individual trial-level effect parameter, i.
To estimate values of the various parameters, a joint probability distribution was formed by combining the prior distributions with the likelihood, given the observed data. Markov chain Monte Carlo (MCMC) methods implemented within the WinBUGS package (Spiegelhalter et al., 2002) were used to obtain samples from the joint posterior distribution to estimate (in the case of the fixed-effects models) and µ and 2 (in the case of the random effects models). Our aim was to provide summary measures of treatment effect relevant to clinical decision making; thus (in addition to reporting the credible intervals of the relative risk estimates for the fixed-and random effects models), summary measures of effect have been reported as the 95% CI of the relative risk estimates derived from µ and 2. In practical terms, this predicted distribution should be thought of as the relative risk estimate of a successful reproductive outcome when an Ovsynch or other synchrony programs are used in a future (or unobserved) trial (Parmigiani, 2002).
For the Bayesian analyses, the MCMC sampler was run for 40,000 iterations and the first 1000 burn-in samples were discarded. Convergence of the posterior sampling distribution was visually assessed using cumulative path plots and quantified using the Raftery and Lewis convergence diagnostic (Raftery and Lewis, 1992a, b). Posterior sample sizes were determined by running sufficient iterations to ensure that the MCMC standard error of the posterior means were at least one order of magnitude smaller than their posterior standard deviation. Parallel chains of the MCMC sampler were run using diverse initial values to ensure that convergence was achieved to the same distribution (Gelman, 1996).
Sensitivity to the distributional assumptions was assessed by assuming different distributions for the trial effects and comparing subsequent inferences. For example, a second series of analyses was conducted where the variance of treatment effect (2) was assumed to arise from a uniform distribution. Within a model, we determined how sensitive the combined estimate was to any one trial. This was achieved by leaving one trial out, calculating the combined effect of the remaining trials, and comparing the results with the combined effect based on all trials.
Results of each model were summarized as forest plots. Here, each study was represented by a black square and a horizontal line, corresponding to the point estimate and the 95% CI of the individual trial-level relative risk. Pooled relative risk estimates (and their 95% CI) for the fixed-effects, random effects, and prediction model were denoted by diamonds on the same plot.
Publication Bias
Funnel plots (Light and Pillemer, 1984) were constructed to identify the presence of publication bias for each outcome. For each study included, estimates of the precision were plotted on the horizontal axis and study standard error on the vertical axis. The nature of sample distribution means that the plot will be naturally funnel-shaped with an apex pointing upwards. If publication bias is present, the plot will have a deficiency, usually in the area occupied by small negative studies. It is based on the premise that the precision in estimating the treatment difference will increase as the sample size of the study increases (Whitehead, 2002).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
, Select Synch, Heat Synch, and modified Ovsynch.
Details of data extracted from each paper are presented in Table 1
. The treatment protocol in synchrony programs included the dose and days of GnRH and PGF2 administration, duration of treatment, days of intravaginal controlled internal drug release insertion; and day, time, and type of insemination (at observed estrus or TAI) for each synchrony program.
Nine trials comparing Ovsynch with natural breeding extracted from 7 papers were used in the meta-analysis of conception rates in lactating dairy cows. A total of 2345 cows were used in these studies. Figure 1 gives the crude RR of conception and pregnancy rates for individual trials. Bayesian (pooled and predicted) RR are provided to summarize treatment effects across all included trials. The Ovsynch synchrony program did not produce a significant increase in pregnancy rate for all trials combined (predicted Bayesian RR = 1.04; 95% Bayesian CI = 0.36 to 3.23; Figure 1) and the results of this comparison were heterogeneous (I2 =77%, uncertainty intervals = 56 to 88%, Table 2). Sensitivity analyses conducted following the exclusion of studies of Cordoba and Fricke (2002) and Jobst et al. (2000) showed that the results of remaining studies were homogeneous (I2 = 34%; uncertainty intervals = 0 to 72%).
|
|
programs were used in the meta-analysis of conception rate, pregnancy rate, and overall pregnancy rates in lactating dairy cows (Figures 2 to 4), representing 4970 cows. The risk of conception and pregnancy rates in the Ovsynch-treated group did not differ from those in the PGF2 program (predicted Bayesian RR = 0.89; 95% Bayesian CI = 0.31 to 2.64, Figure 2, and predicted Bayesian RR = 1.11; 95% Bayesian CI = 0.61 to 2.13, Figure 3, respectively). Both fixed- and random-effects models showed that the results of conception and pregnancy rates were heterogeneous (Table 2, Figures 2 and 3). The results of overall pregnancy rate were homogeneous (I2 = 45%, uncertainty intervals = 0 to 77%, Table 2).
|
|
|
and predicted Bayesian RR = 1.08; 95% Bayesian CI = 0.38 to 3.09, Figure 6, respectively). Both fixed- and random-effects models showed that the results of pregnancy trials on Ovsynch vs. Select Synch were heterogeneous (I2 = 82%, uncertainty intervals = 60 to 92%, Table 2). Sensitivity analyses conducted following the exclusion of pregnancy rate data in the studies of Jobst et al. (2000) showed that the results of remaining studies were homogeneous (Table 2).
|
|
). Fixed- and random-effects models showed that the results of studies on Ovsynch vs. modified Ovsynch were homogeneous (I2 = 0%, uncertainty intervals = 0 to 25%, Table 2).
|
. This includes the pooled RR, predicted RR (Bayesian models), 95% CI, variance of trial-specific treatment effect and its 95% CI, heterogeneity 2, degrees of freedom (df), significance of 2, and I2 and its uncertainty intervals. Results of sensitivity analysis on selected synchrony program comparisons are presented in Table 2.
Funnel plots (Figures 8 and 9) of 2 comparisons were constructed to demonstrate use of funnel plots to assess publication bias. Figure 8 was constructed for Ovsynch vs. natural breeding; the plot was not symmetrical and the small number of studies made it difficult to conclude whether publication bias was present. Figure 9 was constructed for Ovsynch vs. modified Ovsynch, and it displays some evidence of a symmetrical funnel plot, suggestive of a lack of publication bias in this group of studies.
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Only papers that investigated and reported fertility data, such as conception and pregnancy rates, were included to evaluate the efficacy of programs to improve the reproductive management of the herd. Although other papers that reported the effect of Ovsynch program on follicular development and corpus luteum function are important, these studies did not provide fertility data and were not used. We intended to evaluate the time-associated fertility events (e.g., 21-d conception rates) for the trials included in this study. Those fertility measures would be ideal for use in meta-analysis. However, very few studies reported these time-associated incidences for pregnancy and conception rates; of those that reported these time-associated incidences, the time measured from the insemination to pregnancy differed. Most published papers referred to conception/pregnancy to the first and second insemination or mating. Therefore, the conception and pregnancy rates as defined earlier in this paper were evaluated for each study and included in the analysis.
Determinations of the degree of similarity or difference between the results of apparently similar studies and examination of possible sources of variability are primary issues in the conduct of a meta-analysis (Greenland, 1994; Olkin, 1994). Consequently, we examined sources of heterogeneity in the study results. The relatively small number of trials from which data suitable for analysis were available in some cases resulted in relatively large predicted Bayesian CI of RR, and reduced our ability to draw more definite conclusions. Funnel plot analysis of results in this study showed that the results of a number of smaller trials, and trials with negative or positive outcomes of programs may not have been published. The plots showed little evidence of a funnel shape, were difficult to interpret, and reflected a need for more studies.
Ovsynch vs. Natural Breeding
Results of pregnancy rates for the Ovsynch and natural breeding programs did not differ significantly (predicted Bayesian = 1.04, Bayesian CI = 0.36 to 3.23, Figure 1, Table 2). Heterogeneity was influenced largely by the results of Cordoba and Fricke (2002) and Jobst et al. (2000), who reported lower pregnancy rates than other studies. There was no notable difference in study design between these and other trials included in the meta-analysis, except that in these studies, insemination was predominantly based on tail-paint removal (Cordoba and Fricke, 2002) and Heatwatch (Jobst et al., 2000) rather than visual observation. Sensitivity analyses conducted following the exclusion of these 2 studies (Jobst et al., 2000; Cordoba and Fricke, 2002) showed that the remaining studies were homogeneous (I2 = 34%, uncertainty intervals = 0 to 72%). In the natural breeding group, when inseminated at observed estrus, pregnancy rates were lower, and in some trials were comparable with the Ovsynch program. Insemination of cows based on the tail-paint removal method resulted in a greater pregnancy rate than those based on observed estrus. Therefore, the advantage of Ovsynch in increasing the number of cows submitted is diminished by the lack of a sensitivity of some heat-detection methods.
The homogeneity of the available trial data, following the removal of studies of Cordoba and Fricke, (2002) and Jobst et al. (2000), indicated that pregnancy rate was significantly improved in the Ovsynch group compared with those in the naturally bred group (Pursley et al., 1997b; Aréchiga et al., 1998; Keister et al., 1999). These data also suggest that reproduction of lactating dairy cows can be managed effectively without estrus detection by using the Ovsynch protocol.
Ovsynch vs. PGF2
The Ovsynch and PGF2 synchrony programs produced similar outcomes for conception and pregnancy data. The PGF2 programs included single, double, or triple prostaglandin injections with some modifications in the breeding protocol. These studies were, therefore, not identical, and not conducted under similar conditions. For example, the studies of de la Sota et al. (1998) and Aréchiga et al. (1998) were conducted under summer heat conditions. However, we consider that the studies were sufficiently similar to be compared with a view to estimating measures of interest in the meta-analysis. The risk of conception and pregnancy did not differ significantly with cows synchronized with PGF2 programs compared with those treated with Ovsynch protocol (Figures 2, 3, and 4).
In most of the PGF2 program studies, the breeding programs were based on insemination at or after detected estrus, whereas in the Ovsynch program, all inseminations were based on TAI breeding protocol, regardless of estrus detection. Consequently, it could be expected that the Ovsynch program would produce greater pregnancy rates. However, there was not a significant improvement in the risk of pregnancy with the Ovsynch breeding program. In this study, all conception rate data extracted from PGF2 trials were based on insemination at or after detected estrus, except one study (Stevenson et al., 1999, trial 2; Table 1). Stevenson et al. (1999) reported that conception rates were lower after TAI, but conception rates were greater when AI occurred after observed estrus. The pooled estimates in this study showed that conception and pregnancy rates were slightly and nonsignificantly better in cows treated with the Ovsynch protocol compared with those in the PGF2 programs.
The results of studies reporting conception and pregnancy rates were heterogeneous (Table 2). Heterogeneity was influenced largely by the results of studies of Alnimer et al. (2002), Jemmeson et al. (2000), and Britt and Gaska (1998) for conception rate data, and de la Sota et al. (1998), Tenhagen et al. (2001), Alnimer et al. (2002), Britt and Gaska (1998) for pregnancy rate data. Sensitivity analyses conducted following the exclusion of these trials from the study showed that heterogeneity persisted in the remaining studies. This heterogeneity could not be readily attributed to consistent differences among trial protocols.
The risk of pregnancy by the end of breeding program (overall pregnancy) with the Ovsynch protocol was comparable with PGF2 programs (Figure 4). The results of studies reporting the overall pregnancy were homogeneous (I2 = 45%, uncertainty intervals = 0 to 77%, Table 2). These results show that the Ovsynch program may be as effective as PGF2 programs in improving pregnancy rates, although the Ovsynch protocol would be able to provide a very predictable time of AI after the start of treatment (Stevenson et al., 1996; Alnimer et al., 2002). These results show that the Ovsynch program could benefit dairy operations by allowing TAI of lactating cows without detection of estrus, producing similar pregnancy rates to those achieved by PGF2 programs.
The PGF synchrony programs varied greatly; for example, some were single injections of PGF, and others were double or triple PGF programs. Injection intervals varied, as did breeding programs, for example, single AI vs. TAI vs. multiple AI. The approach used by the original authors of these papers in comparing some of these multiple AI PGF programs with the Ovsynch program, especially with a single treatment and only one AI, is questionable. We note the limitations of these approaches, and suggest that further meta-analyses with greater numbers of studies available may exclude some studies based on more equitable comparison.
Ovsynch vs. Select Synch
The risk of conception with Ovsynch treated cows did not differ from those treated with Select Synch protocols (predicted Bayesian RR = 0.94, Bayesian CI RR = 0.52 to 1.59, Figure 5), and the results were homogeneous (I2 = 34%, uncertainty intervals = 0 to 75%, Table 2). These data suggested that Select Synch programs could produce a similar conception rate to the Ovsynch program. However, the breeding program with Select Synch depends on estrus detection, which could be a practical disadvantage compared with the Ovsynch program.
The risk of pregnancy did not differ between Ovsynch cows and those treated with Select Synch protocols (predicted Bayesian = 1.08, Bayesian CI = 0.38 to 3.09). Heterogeneity in the results of studies reporting pregnancy rates (I2 = 82%, uncertainty intervals = 60 to 92%, Table 2) was influenced largely by the results of Jobst et al. (2000) who reported lower pregnancy rate with Ovsynch program than other trials. Differences in the study design between these studies and other trials were the dose and type of GnRH preparation used in the studies of Jobst et al. (2000). Sensitivity analyses showed that results of the remaining trials were homogeneous (I2 = 52%, uncertainty intervals = 0 to 84%, Table 2). An improvement in the risk of pregnancy can be achieved with Ovsynch program if cows are inseminated on detected estrus (Cartmill et al., 2001a), a program similar to the Select Synch program breeding protocol. These data demonstrate that the reproductive performance of cows treated with Ovsynch and Select Synch protocols are comparable.
Ovsynch vs. Modified Ovsynch
The risk of conception and pregnancy in cows synchronized with modified Ovsynch did not differ from those in the Ovsynch program (predicted Bayesian RR = 0.89, Bayesian CI = 0.71 to 1.12, Figure 7), and the results of these studies were homogeneous (I2 = 0%, uncertainty intervals = 0 to 25%). A conception or pregnancy rate of individual studies using modified Ovsynch were comparable, and in some studies greater than an Ovsynch program. Most of studies we defined as modified Ovsynch were based on presynchronization of the estrus cycle, with single or double PGF2 or an insertion of a controlled internal drug release device before the implementation of the Ovsynch protocol (Keister et al., 1999; Cartmill et al., 2001a; Fricke, 2001; Moreira et al., 2001). Small numbers of studies on each modification were a limitation and prevented analysis of these as separate groups. If the studies kept the core hormonal therapies of the Ovsynch program (GnRH-PGF2 -GnRH) and similar timing of breeding with modification at either the beginning or end of the program, these were pooled to provide a single point estimate. Increased pregnancy rates were detected when cows that were presynchronized before the implementation of Ovsynch were compared with cows that had no pre-synchronization treatments (Figure 7). The likely reason for increased pregnancy rates in presynchronized cows was the increased frequency of cows initiating the synchronization protocol at favorable stages of the estrous cycle (Thatcher et al., 2001a). Effective resynchronization programs place cows between d 5 to 10 of the cycle at the time of the GnRH injection depending upon which day the cows expressed estrus after the injection of PGF2. Days 5 to 10 of the cycle are considered the optimal time to begin the Ovsynch program (Thatcher et al., 2001a,b). These modifications to the Ovsynch program produced a comparable, and in some cases slightly higher, pregnancy rate compared with those treated with the Ovsynch program. The CoSynch program protocol (modified Ovsynch) requires cows to be handled only 3 times and, thus, would be easier to incorporate into an AI program. Therefore, modifications such as presynchronization of the Ovsynch protocol and altering the time of AI (CoSynch) could be suitable TAI programs for commercial dairy herds.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Progress toward improving reproductive efficiency in lactating dairy cows may be achieved by combining TAI with a protocol for synchronization of ovulation that can be initiated at any stage of the estrous cycle. Overall, conception/pregnancy rates obtained with PGF, Select Synch, and modified Ovsynch programs were similar to the Ovsynch program. However, synchronization of ovulation using Ovsynch protocols can provide an effective way to manage reproduction in lactating dairy cows by eliminating the need for estrus detection. The modifications in Ovsynch program such as presynchronization and TAI at the time of second GnRH injection (CoSynch) can be a useful alternative for reproductive management of dairy herds with less than optimal estrus detection rates. Advantages of modified Ovsynch included data that are more homogeneous, greater submission rates than natural breeding, and avoidance of estrus detection. However, data on conception or pregnancy density (time-associated indices) are also required to allow the comparison of the efficacy of different synchrony programs.
|
ACKNOWLEDGEMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Received for publication December 19, 2004. Accepted for publication March 28, 2005.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
and (or) GnRH on pregnancy rates of Ovsynch- and Cosynch-treated lactating Holstein cows. Anim. Reprod. Sci. 77:51–60.[Medline]
. Anim. Reprod. Sci. 67:27–35.[Medline]
, gonadotropin-releasing hormone, and timed artificial insemination. J. Dairy Sci. 83:2366–2372.[Abstract]
. Vet. Rec. 145:551–554. or GnRH + prostaglandin F2 + GnRH. Theriogenology 52:901–911.[Medline]
as a postpartum reproductive management tool for lactating dairy cows. J. Dairy Sci. 78:1477–1488.[Abstract]
and GnRH. J. Dairy Sci. 85:2403–2406.. Theriogenology 44:915–923.
. J. Dairy Sci. 79:402–410.[Abstract]
. J. Dairy Sci. 82:506–515.[Abstract]
. Theriogenology 28:937–946.
, norgestomet and gonadotropin-releasing hormone. J. Dairy Sci. 83:2008–2015.[Abstract]
and human chorionic gonadotropin. Theriogenology 56:91–104.[Medline]
This article has been cited by other articles:
|
|
 |
|
  R. H. Miller, H. D. Norman, M. T. Kuhn, J. S. Clay, and J. L. Hutchison Voluntary Waiting Period and Adoption of Synchronized Breeding in Dairy Herd Improvement Herds J Dairy Sci, March 1, 2007; 90(3): 1594 - 1606. [Abstract] [Full Text] [PDF]
|
|
