The fertility of cows plays an important role in the productivity of dairy herds. Pregnancy loss (PL) is one of the major sources of decreased fertility, and this produces adverse economic effects for dairy farms [35,36]. The reported incidence of PL ranges from 0.4 to 10.6% [8] and both infections and non-infectious factors are known to play a role for this. While the infectious causes of PL have been a primary focus of attention, non-infectious PL is actually more common in endemic situations. Various causative factors, including external, maternal and genetic factors, have been reported for PL in dairy cattle. These include heat stress [11], the season [24], milk production [12], cow parity [21], the serum progesterone level after conception [26], the inseminating bull [23], twin pregnancy [23] and the herd [25]. However, other investigations have reported that milk production [31], cow parity [26] and clinical disease [23] were not associated with PL. Further elucidation of the factors that cause PL may enable the development of more effective management regimens for efficient herd reproduction.

PL may result in the retention of fetal membranes and the development of endometritis, which subsequently reduces reproductive performance in dairy cows. Furthermore, a prolonged postpartum interval in cows with PL can lead to increased culling [1]. However, the effects of PL on reproductive performance, culling and economic loss have not been fully determined and there have been no detailed studies of PL in Korean dairy herds. Therefore, the objectives of this study were to determine the effects of herd, cow parity, the insemination protocol and season on the incidence of PL, and to determine the downstream effects of PL on the rate of endometritis and culling, the reproductive performance and the economic viability of dairy herds.

The overall incidence rate of PL was 6.9% in a population of 1,001 pregnant dairy cows. The incidence was greater (p < 0.01) during the second trimester (3.4%) than during the first trimester (1.3%), although the incidence during the second trimester was not significantly different from the rate during the third trimester of gestation (2.3%, Table 2). The average time at which PL occurred was 154.5 ± 63.2 days of gestation. The GLIMMIX procedure demonstrated that the estimated standard error was 0.0398 ± 0.1085, indicating no influence of the herds on the incidence of PL. The procedure identified cow parity as the risk factor for PL (Table 3). Based on the odds ratio, the likelihood of PL in cows with parities of 1 or 2 was decreased by 0.6 or 0.5 fold of that for the cows with a parity of 3 or higher, respectively (p < 0.05). However, the herd, the insemination protocol and the season were not found to be risk factors for PL.

Following PL, the mean incidence rate of endometritis was 23.2% and this rate was higher (p < 0.05) for the cows with PL during the third trimester (45.5%) than for the cows with PL during the first (7.7%) and second trimesters (14.7%, Table 4). The mean culling rate following PL was 46.4% and the time of PL did not affect the culling rate (Table 4). The overall mean intervals from PL to the first service and conception were 63.4 ± 5.2 and 101.8 ± 10.8 days, respectively. The interval from PL to first service was longer (p < 0.01) for the cows with PL during the third trimester than for the cows with PL during the first and second trimesters, while the interval from PL to conception did not differ according to time of PL (Table 4). Thus, the overall mean extended calving interval in cows with PL was 256 days.

The economic loss resulting from each case of PL in these herds was estimated at approximately $2,333 due to the effects on the cost of nutrition, the average growth of calves, the labor and medical costs, and culling (Table 5).

In this study, we determined the risk factors for PL and the downstream effects of PL on reproductive performance, as well as assessing the economic impact of PL in seven dairy herds. Cow parity was positively correlated with the frequency of PL. Furthermore, the risks of endometritis and of a longer interval from PL to first service were greater when PL occurred during the third trimester of gestation. The overall mean economic loss resulting from PL was estimated at approximately $2,333.

The overall incidence of PL (6.9%) in this study is in agreement with several previous papers [11,21,25,33,34], which reported PL rates of 6.4~8.6%. It is notable, however, that the reported rate and incidence of PL does vary widely between studies, and 6.9% is higher than some previously reported rates (0.4 and 3.6%; [13,28]) and it is lower than the rates (10.2 and 10.6%) reported in other studies [23,24]. These discrepancies may reflect differences in the geography, the study population, the case definitions and the procedures among the studies. The incidence of PL peaked during the second trimester in this study, and this is consistent with the results of Thurmond et al. [35]. Yet Forar et al. [9] found that the risk of PL was highest in the first trimester of gestation and then it progressively decreased as gestation progressed. On the other hand, Markusfeld-Nir [24] demonstrated that the greatest risk of PL was observed in the first trimester, whereas the second trimester had the lowest incidence of PL. The reasons for these differences are unclear, but they may reflect variations in the detection time between studies and/or a discrepancy in the detection time versus the actual time of PL.

Cow parity was positively linked to PL in this study. The incidence of PL was greatest in the cows with a parity of 3 or higher, which is similar to the results of Thurmond et al. [35], who reported that PL increased after 4 calvings. Similarly, Humblot [16] showed that frequency of embryonic mortality increased with parity (1st to 3rd parity), although others have reported no effect of parity on PL [21,26]. The influence of parity on PL is unclear. One possibility is that high milk production at the third parity compared to previous parities, which mobilizes more body fat and results in severe loss of body condition (BCS) [22], might be linked to the increased PL. Silke et al. [33] showed that BCS loss between days 28 and 56 of gestation was a risk factor for PL. Likewise, a drop in BCS from the previous parturition to 30 days postpartum has been shown to increase PL, with a 1 unit decrease in score increasing the pregnancy loss by a factor 2.4 [23]. In agreement with previous reports [9,11], there were no significant differences in the incidence of PL among the seven dairy herds in this study. However, some previous investigations have found a herd effect [16,25]. Moreover, as indicated by Labernia et al. [21], the management characteristics and man-cow interactions that tend to vary from farm to farm may influence the rate of PL. Various protocols for the synchronization of estrus and/or ovulation have been used to optimize the service rate in dairy herds. We found no significant effect of the previous synchronization of estrus and ovulation on the incidence of PL, and this was consistent with the results of López-Gatius et al. [23], who demonstrated no significant effect of previous estrus synchronization on PL. Several other studies found that the PL rates were not different between the cows inseminated following timed AI and the cows inseminated upon detecting estrous; these results also concur with our present findings [4,6,14,30]. However, in a report by Cartmill et al. [5], a tendency for an increased PL was observed with timed AI. PL was not associated with the season in our investigation, and this is inconsistent with previous studies. For example, it has been reported that the risk for PL was increased during the warm period (May to September) versus the cool period (October to April) due to heat stress [11]. Similarly, in another report, PL was increased during June to October, and it peaked in July and September [3]. On the other hand, Markusfeld-Nir [24] demonstrated that the incidence of PL was greater during autumn and early winter than during the summer months, and this is in agreement with the report of Grimard et al. [12]. These differences in the effect of season on PL might be due to different climates (temperature-humidity), geography or other environmental factors among studies.

There are very few published reports about the effects of PL on the subsequent incidence of reproductive disturbances and/or reproductive performance. However, PL was associated with an extension of 80 days from first service to conception and an extension of 77 days in the calving interval [10]. The mean incidence of endometritis following PL was 23.2% and this was more common in those cows that suffered PL during the third trimester than in the cows with PL during the first and second trimesters. This may reflect the more delayed uterine involution and increased incidence of retained placenta associated with PL in the third trimester, as compared to the earlier trimesters of gestation [15,29]. Endometritis following PL may lead to increased culling and it may extend the period of open days. Interestingly, the culling rates were not different among cows that experienced PL during the first, second and third trimesters of gestation. However, the overall mean culling rate of 46.4% for the cows with PL was higher than the general rate of 27% for all the herds. The mean interval from PL to the first service was 63.4 days and this interval was longer for the cows with PL during the third trimester than for the cows with PL during the first and second trimesters. Again, this might be associated with delayed uterine involution and/or the increased incidence of endometritis, as indicated by Kim and Kang [18], and Borsberry and Dobson [2]. In fact, earlier re-pregnancy may be important for PL cows if further economic loss is to be minimized. Our results showed that the overall mean interval from PL to conception was 102 days, thereby culminating in a calving interval that was 256 days longer than that of the cows without PL.

A delayed calving interval and increased culling are the main sources of economic loss in dairy herds. Here, the economic loss resulting from each PL was estimated at approximately $2,333 due to the effects on the cost of nutrition, the average growth of calves, the labor and medical costs, and culling. In view of this economic impact, it would be desirable to reduce the interval from PL to conception. On the other hand, earlier culling might be recommended to prevent a further economic loss when PL occurs in the later period of gestation.

In conclusion, we have demonstrated here that cow parity is positively correlated with PL, and that the risks of endometritis and a longer interval from PL to first service are greater when PL occurred during the third trimester of gestation. Yet irrespective of when PL occurs, it produces a serious economic impact on dairy herds. Therefore, reducing nutritional stress by improving the management of cow parity and minimizing the interval from PL to conception will be necessary strategies for reducing the adverse economic effects of PL in dairy herds.