If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
A meta-analysis was conducted on 19,723 pregnancy diagnosis records in dairy cows.
•
Four periods were defined due to the embryonic development up to 90 days of gestation.
•
Transrectal ultrasound is the dominant method of pregnancy diagnosis.
•
Pregnancy losses ranged from 27% in early embryonic to 2% in later fetal stages.
Abstract
Pregnancy losses are a biological challenge and economically important in dairy herds. A meta-analysis was conducted to quantify losses in 4 periods from 19 to 90 d of pregnancy corresponding to the physiological development of gestation in dairy cows. A total of 19,723 diagnostic records from 46 studies were included. Pregnancy losses averaged 27%, 13%, 7%, and 2% in the early embryonic (19 to 32 d), late embryonic (30 to 45 d), early fetal (45 to 60 d), and later fetal (60 to 90 d) stages. These results provide a formal synthesis of the incidence of pregnancy losses in dairy cows.
Graphical Abstract
Graphical AbstractSummary: A meta-analysis was performed on a total of 19,723 diagnostic records of pregnancy loss from 46 studies on dairy cows. Pregnancy losses averaged 27%, 13%, 7%, and 2% in the early embryonic (19 to 32 days), late embryonic (30 to 45 days), early fetal (45 to 60 days), and later fetal (60 to 90 days) stages. These results provide a formal synthesis of the incidence of pregnancy losses in dairy cows.
Reproductive efficiency is an important determinant of dairy herds' profitability because of its direct link to an optimal milk production cycle and culling. A combination of management practices that achieve a high insemination rate and probability of pregnancy per insemination is critical for reproductive performance (
). This includes prompt and accurate identification of open cows following insemination. However, as fertility and reproductive performance improve through management and genetics, pregnancy and calving rates may be limited by substantial pregnancy losses (PL) (
). Numerous studies have reported PL at different stages of gestation, documenting greater apparent losses in the first month after practical diagnosis of pregnancy (i.e., from approximately 28 d), with reduced rates of loss later (
). The periods investigated vary among studies, leading to inconsistency in the apparent PL. Pregnancy loss before d 24 is classified as early embryonic loss and those between d 24 and 42 to 50 as late embryonic losses, whereas those detected after d 50 are considered fetal losses (
). Several techniques exist to detect PL including direct (transrectal ultrasonography or palpation) and indirect (measures of pregnancy-specific glycoproteins or progesterone) methods (
). Because PL is common, cows correctly diagnosed pregnant between approximately 25 and 60 d of gestation must undergo one or more confirmation tests to confirm the pregnancy status and to identify cows experiencing PL for timely rebreeding. Accordingly, it is important to provide valid benchmark data for farmers, veterinarians, and advisors to identify and investigate elevated rates of PL. Two narrative review articles (
) summarized a variety of studies that addressed PL in dairy cows. The objective of the present meta-analysis is to further characterize PL in dairy cows by providing an updated and newly quantitative summary estimate of PL during specific periods of gestation in dairy cattle.
We conducted a meta-analysis. We began with data from the review articles (
), including only studies on dairy cows. All published trials were re-examined to confirm the reported PL values and risk periods used. We searched for subsequent publication of the cited unpublished data in the reviews. If not published, estimated PL values were kept as reported in the review paper. We used the PubMed (https://pubmed.ncbi.nlm.nih.gov/) and CAB (https://www.cabdirect.org/cabdirect) search engines in July 2022 using “pregnancy OR fetal OR embryonic AND loss OR death” and “dairy AND cattle OR cow OR bos taurus OR bovine” in the title and abstract for PubMed and in the title for CAB to identify additional data sources as indicated in Table 1. We set 4 risk periods for PL: early (19 to 32 d) and late (30 to 45 d) embryonic periods and early (45 to 60 d) and late (60 to 90 d) fetal periods. These periods were based on the reported data for first and second pregnancy diagnoses, which explains the small overlap between the first 2 risk periods. After eliminating duplicate articles and studies included in the 2 above-mentioned review articles, the literature search returned a total of 142 original research and review articles. Articles with no experimental data on PL or pregnancy diagnoses (n = 58) or in which pregnancy diagnoses spanned 2 or more periods of interest without subdivided data (n = 56) were excluded. Data from trials reporting induced or specifically disease-associated PL were also excluded (n = 14) except when data for control groups were available, in which case only the control group data were used. The complete list of the search results with the reason of exclusion or inclusion in the meta-analysis is available at https://doi.org/10.5683/SP3/4NAUFI. When one of the pregnancy diagnoses occurred within ± 3 d of a defined period the data were attributed to the most relevant subgroup. A list of the studies included in the meta-analysis is in Table 1.
Table 1Sources of data included in a meta-analysis on the risk of pregnancy loss in lactating dairy cows
Use of pregnancy specific proteins and progesterone assays to monitor pregnancy and determine the timing, frequencies and sources of embryonic mortality in ruminants.
Effect of progesterone (P(4)) intravaginal device (CIDR) to reduce embryonic loss and to synchronize return to oestrus of previously timed inseminated lactating dairy cows.
Resynchronization of ovulation and timed insemination in lactating dairy cows, II: Assigning protocols according to stages of the estrous cycle, or presence of ovarian cysts or anestrus.
Effect of a rapid resynchronization of nonpregnant cows with estradiol cypionate (ECP) and PGF2a on pregnancy rates (PR) and pregnancy loss (PL) in lactating dairy cows.
Effect of addition of a progesterone intravaginal insert to a timed insemination protocol using estradiol cypionate on ovulation rate, pregnancy rate, and late embryonic loss in lactating dairy cows.
Short communication: Blood samples before and after embryonic attachment accurately determine non-pregnant lactating dairy cows at 24 d post-artificial insemination using a commercially available assay for pregnancy-specific protein B.
Effect of a deslorelin implant in a timed artificial insemination protocol on follicle development, luteal function and reproductive performance of lactating dairy cows.
Is twin pregnancy, calving and pregnancy loss predictable by serum pregnancy-specific protein B (PSPB) concentration 28–35 days after AI in dairy cows?.
Factors associated with pregnancy-associated glycoprotein (PAG) levels in plasma and milk of Holstein cows during early pregnancy and their effect on the accuracy of pregnancy diagnosis.
) of R (version 4.2.1; R Foundation for Statistical Computing). A fixed and random effects model for meta-analysis of single proportions was built to calculate the pooled effect size using the number of lost pregnancies (events) and the number of pregnant cows at first diagnosis (total), using the inverse variance method. The inconsistency of results among trials was quantified using both Cochran's Q test and the I2 statistic. Greater I2 values denote variation in true effect sizes and indicate heterogeneity in the data (
). A forest plot was used to display the pooled risks of PL and their 95% CI within each period of interest and for the whole period (19 to 90 d). Funnel and influential case diagnostics were conducted at the subgroup level to analyze publication bias to assess balance between results of large and small studies (results not shown). These analyses were conducted on previously collected data and do not contain any experiments with animal subjects, so did not require IACUC approval.
The data included 46 data sets derived from 2 review articles supplemented with 14 additional publications. Transrectal ultrasound was the dominant method of pregnancy diagnosis, whereas pregnancy-specific glycoprotein measurements were used in some studies. The meta-analysis (Figure 1) showed that during the early embryonic stage, PL averaged 27% (95% CI = 21 to 33%) and was highly variable as indicated by a prediction interval of 13 to 53%. One study (
Use of pregnancy specific proteins and progesterone assays to monitor pregnancy and determine the timing, frequencies and sources of embryonic mortality in ruminants.
) measured progesterone at 23 d to classify pregnancy, which is of dubious accuracy; this was also one of the larger studies in the data set with 1,870 cows, increasing its weight in the meta-analysis. Nevertheless, its estimated PL aligned with the others in the 19 to 32 d risk period (Table 1). Other studies in the same risk period used quantification of interferon-stimulated gene (ISG) 15 for the first classification of pregnancy, which may produce approximately 15% false-negative results (
). The incidence of PL decreased in the subsequent periods and was estimated at 13% (95% CI 10 to 16%), 7% (95% CI 4 to 11%), and 2% (95% CI 1 to 3%) for the late embryonic, early fetal, and later fetal periods, respectively. The I2 statistic was greater than 50% for all periods, indicating substantial heterogeneity among studies within a period. Combined, using 19,723 dairy cows diagnosed pregnant between 19 and 90 d, we obtained a summary estimate of PL of 10% over the whole period. There was substantial heterogeneity of the data [I2 (95% CI) = 96.4% (95.8% to 97.0%); Q statistics: χ2 = 1,265.64, df = 32, P < 0.0001]. However, the overall PL might underestimate the real value because none of the studies included in this meta-analysis cover the entire period to 90 d. For instance, PL between d 28 and 110 of gestation was 12.3% (
). The funnel plots showed no evidence of publication bias. Three data points appeared to be possible outliers, but removing them from the data set did not lead to significant changes in the results of the meta-analysis.
Figure 1Forest plot of the incidence of pregnancy loss in dairy cows in different stages of gestation. Events: number of lost pregnancies. Total: number of pregnant cows at the first pregnancy diagnosis. Weight (common): weights to all studies based entirely on the amount of information captured by each study. Weight (random): weights assigned under random effects. IV, Fixed + Random, 95% CI: inverse variance of the summary estimate of pregnancy losses of fixed effect and random effects model (represented by squares) and their 95% CI (represented by the upper and lower limits of the line connected to the square). The size of the square reflects the relative weighting of the study to the overall effect size estimate. The vertical lines represent the overall random (dotted) and fixed (dashed) effect size estimate. The diamonds at the bottom represent the 95% CI for the subgroup and overall estimates. Tau2: the variance of the distribution of true effect sizes. Chi2: Cochran's Q test for heterogeneity. I2: percentage of variation among studies that is due to heterogeneity rather than chance.
The objective of this meta-analysis was to quantify PL in dairy cows based on a synthesis of available data. Despite the slight overlaps, the risk periods for PL in this meta-analysis generally align with the physiological periods of pregnancy development and loss described by
. Our random effects model provides the average risk of PL across all studies within a risk period and underlines the high between-study heterogeneity (i.e., the value of I2 indicates variability in risk estimates as a consequence of true differences between trials rather than due to chance). Ultrasonography and rectal palpation represent the most common and practical methods of pregnancy diagnosis in the field, and ultrasonography from 28 d after insemination is widely used as a reference test. However, as quantified in this analysis, there is meaningful PL during the first month after early pregnancy diagnosis (i.e., from 28 to 60 d), which supports the need for confirmatory pregnancy tests when early pregnancy diagnosis methods are used. Because pregnancy diagnosis from 28 d is widely practiced and because PL diminishes after 60 d, numerous studies have pregnancy diagnosis and confirmation performed at ~30 and ~60 d, but these were excluded using our criteria based on reproductive biology. From a physiological point of view, we believe that the distinction at 45 d should be made. This point represents the end of the differentiation stage and the start of the fetal period from d 45 of pregnancy to parturition (
The benefits of a meta-analysis include providing benchmark values of the effect size or risk, increasing statistical power by combining larger quantities of data than available in any one study, and quantifying heterogeneity among studies. If sufficient comparable covariate data are available, meta-regression can identify sources of variation in outcomes. For PL, however, there are numerous causes of these PL that were previously summarized and discussed (
), but many of which are difficult to study, which probably interact, and which were not reported consistently in the studies included here. Therefore, additional research is needed on the mechanisms of PL and methods to attribute cause(s) of PL in specific cases. The current meta-analysis summarizes PL in the first trimester gestation and highlights the high variability of PL in dairy cows during the early embryonic period. The benchmark risks quantified here can inform monitoring of PL in dairy herds to identify problems and successes. Study of mechanisms of PL in specific periods will support development of methods to reduce pregnancy loss.
Notes
This study received no external funding.
These analyses were conducted on previously collected data and do not contain any experiments with animal subjects, so did not require IACUC approval.
The authors have not stated any conflicts of interest.
References
Alnimer M.A.
Lubbadeh W.F.
Effect of progesterone (P(4)) intravaginal device (CIDR) to reduce embryonic loss and to synchronize return to oestrus of previously timed inseminated lactating dairy cows.
Resynchronization of ovulation and timed insemination in lactating dairy cows, II: Assigning protocols according to stages of the estrous cycle, or presence of ovarian cysts or anestrus.
Effect of a rapid resynchronization of nonpregnant cows with estradiol cypionate (ECP) and PGF2a on pregnancy rates (PR) and pregnancy loss (PL) in lactating dairy cows.
Effect of addition of a progesterone intravaginal insert to a timed insemination protocol using estradiol cypionate on ovulation rate, pregnancy rate, and late embryonic loss in lactating dairy cows.
Use of pregnancy specific proteins and progesterone assays to monitor pregnancy and determine the timing, frequencies and sources of embryonic mortality in ruminants.
Short communication: Blood samples before and after embryonic attachment accurately determine non-pregnant lactating dairy cows at 24 d post-artificial insemination using a commercially available assay for pregnancy-specific protein B.
Factors associated with pregnancy-associated glycoprotein (PAG) levels in plasma and milk of Holstein cows during early pregnancy and their effect on the accuracy of pregnancy diagnosis.
Effect of a deslorelin implant in a timed artificial insemination protocol on follicle development, luteal function and reproductive performance of lactating dairy cows.
Is twin pregnancy, calving and pregnancy loss predictable by serum pregnancy-specific protein B (PSPB) concentration 28–35 days after AI in dairy cows?.
00137-5&id=fx1.jpg){kind=link}
00137-5&id=gr1.jpg){kind=link}