FACTORS CONTRIBUTING TO
VARIATION OF DURATION OF ESTRUS AND
TIME OF OVULATION IN A COMMERCIAL SOW HERD

B.A. Belstra, W.L. Flowers, K.J. Rozeboom, and M.T. See

Summary

Despite considerable variation in duration of estrus (DE) and the onset of estrus-to-ovulation interval (OEOI), sows generally ovulate at about 70 to 80% of their individual DE. Unfortunately, DE and time of ovulation can only be established retrospectively and multiple inseminations are necessary to ensure that at least one insemination occurs near ovulation because the viability of gametes is brief. A sow’s weaning-to-estrus interval (WEI) has been found to be a prospective indicator of time of ovulation since it tends to be inversely related to DE. The influence of other factors on DE and OEOI, such as the short lactation lengths (< 21 d) and different sow genotypes currently in use in the swine industry, has gone largely uninvestigated. The DE and time of ovulation of 86 weaned sows (parity 1 to 10; 13 to 19 d lactation) in a large commercial herd was monitored via the back pressure test and transabdominal real-time ultrasonography, respectively, at 0200, 0800, 1400 and 2000 h, beginning 2 d postweaning. Seventy-five of the 86 sows (87%) returned to estrus normally, within 2.5 to 7.0 d postweaning. Neither lactation length nor parity affected their WEI (range, 2.5 to 7.0 d; mean ± SEM, 4.4 ± 0.1 d), DE (12 to 90 h; 59.5 ± 1.6 h), OEOI (18 to 72 h; 45.0 ± 1.6 h), or the percentage of DE at which ovulation occurred (DE%, 27 to 160%; 76.0 ± 2.8%). There was only a weak negative correlation between WEI and DE (r = - 0.25, P < 0.03). One of the three sow genotypes studied tended to ovulate earlier than each of the other two genotypes (67.6 ± 5.4 vs. 78.0 ± 3.7 and 81.5 ± 6.3%, P < 0.10). These data suggest that the short lactation lengths recently adopted by the US swine industry have not altered the temporal relationships between estrus and ovulation suggested in the literature and that sow genotype may be an important source of variation in DE%.

Introduction

The strategy of administering 2 to 3 inseminations, one every 12 to 24 hours after onset of estrus, is standard practice in the swine industry. Such multiple insemination regimens are necessary because the life span of ova and spermatozoa in the sow’s reproductive tract is relatively short (approximately 8 and 24 h, respectively) and ovulation occurs at some unknown time during estrus, which is a relatively long but highly variable event (24 to 96 h). Multiple insemination regimens increase the odds that at least one insemination will occur near enough to ovulation to fertilize a large percentage of the ova (³ 90%) and result in satisfactory sow reproductive performance.

Now that real-time ultrasonography has been applied to the study of ovulation in swine two useful temporal relationships between estrus and ovulation have been revealed (Figure 1., Kemp and Soede, 1996). First, regardless of differences in duration of estrus (DE) between sows, ovulation generally occurs at about 70 to 80% of a sow’s DE (DE%). Second, a negative relationship between a sow’s weaning-to-estrus interval (WEI) and DE has been suggested. A short WEI (3 to 4 d) tends to result in long DE and a long WEI (³ 6 d) tends to result in a short DE.

Figure 1. Effect of weaning-to-estrus interval (WEI) duration of estrus (DE), onset of estrus-to-ovulation interval (OEOI) and percent of estrus at which ovulation occurred (%). Based on data from sows weaned after » 24 d lactations, Kemp and Soede (1996).

Collectively, these findings have been used to suggest that the timing of insemination regimens should be adjusted based on the WEI of the sows being mated in an effort to concentrate inseminations around the anticipated time of ovulation. Flowers (1998b) tested such a strategy under controlled conditions in two different commercial sow herds and found that it produced different results in each WEI category in each herd, apparently because of differences in DE in each WEI category between herds. Since variation in DE between sows and between sow herds seems to be a major factor affecting variation in the time of ovulation and the success of a given insemination regimen, we wanted to identify factors that contributed to variation of DE. Average lactation length has decreased from approximately 25 d in 1990 to 18 d in 1999 and 72% of herds now have an average lactation length £ 22 d (Figure 2., PigCHAMPÒ, 2000). This gradual transition has been largely driven by the potential to produce more litters and pigs/sow/year but has also been associated with increased sow reproductive problems (Dial, 1995). We hypothesized that the general temporal relationships between WEI, DE and time of ovulation illustrated above might not apply to sows weaned after short lactation lengths and that this might explain some of the reproductive failure exhibited by early-weaned sows. Thus, the objective of this study was to examine the effect of lactation length, as well as other potential factors such as parity, genotype and WEI on DE and time of ovulation in a commercial herd.

Figure 2. Frequency distribution of average lactation length in US herds based on data from PigCHAMPÒ (2000).

Materials and Methods

Eighty-six weaned sows from a 2,500 sow commercial herd representing lactation lengths of 13 to 19 d, parities of 1 to 10 and 3 genotypes were used. From the day of weaning (0 d) to 2 d postweaning sows were checked for estrus by the back pressure test during nose-to-nose contact with a mature boar at 0800 and 1400 h. From 2 d to 10 d postweaning detection of estrus was performed by the same means, but at 0800, 1400, 2000 and 0200 h and as sows exhibited estrus, the follicular status of their ovaries was examined by transabdominal real-time ultrasonography scans with an Aloka 500V equipped with a 3.5 MHz convex linear transducer at these same times until ovulation was confirmed. The mid-point between estrus checks and ultrasonography scans was considered the time of ovulation and onset/end of estrus, respectively. All sows that had not expressed estrus by 6 d were scanned to determine their ovarian status. Sows that failed to develop follicles and return to estrus by 10 d postweaning were considered anestrous. Seventy-five of the 86 weaned sows (87%) returned to estrus normally, within 2.5 to 7.0 d postweaning. The remaining 11 sows were excluded from the analysis because, 5 returned to estrus £ 2 d postweaning (3 had multiple follicular cysts) and 6 remained anestrus for ³ 10 d postweaning. The GLM procedure of SAS was used to examine the effects of lactation length (13, 14, 15, 16, 17, 18, 19 d), parity (1, 2, ³ 3), genotype (A, B, C) and WEI (3, 4, 5, 6 d) on DE, OEOI and DE% (SAS, 1990). No significant interactions were found between any of the independent variables and no interactions were included in the final model used for analysis. Data are expressed as LS means ± SEM.

Results and Discussion

There was considerable variation in both DE (range, 12 to 90 h; mean ± SEM, 59.5 ± 1.6 h) and OEOI (18 to 72 h; 45.0 ± 1.6 h) among the 75 sows analyzed (Figure 3.). However, this variation is comparable to several previous reports in the literature (see reviews, Flowers, 1998a; Soede and Kemp, 1997). It is this wide range in DE and OEOI that makes multiple insemination regimens necessary to achieve satisfactory reproductive performance.

Figure 3. Frequency distribution of duration of estrus (DE) and onset of estrus-to-ovulation interval (OEOI).

Even though sows did ovulate near 70% of their DE on average (76.0 ± 2.8%), as suggested in the literature, it is important to note the range in DE% (27 to 160%, Figure 4.). Again, this amount of variation has also been described in the literature, but it is hard to comprehend the usefulness of the 70 to 80% assumption when there is this amount of variation in DE% among 75 sows. Even though 38% of the sows did ovulate at 61 to 80% of their DE, a similar proportion (32%) ovulated at 81 to 100% of their DE. In addition, the percentage of DE at which ovulation occurred exceed 100% in 4 of the 75 sows, meaning that ovulation occurred after the end of estrus. Since detection of estrus is subjective, the DE of these sows may have been longer and they may not have ovulated after the end of estrus.

Figure 4. Frequency distribution of the percentage of estrus at which ovulation occurred (DE%).

It is possible that some of this variation in DE, OEOI and DE% could be due to mis-interpretation of estrus signs and(or) ultrasonography scans by the technician. However, it seems unlikely that this large amount of variation could be explained primarily by human error. Neither lactation length (13 to 19 d) nor parity (1, 2, ³ 3) had an effect on DE, OEOI or DE% (Figures 5. and 6.). Lactation length was analyzed as a continuous variable but is graphically displayed as three classes in figure 5.

Figure 5. Effect of lactation length on duration of estrus (DE), onset of estrus-to-ovulation interval (OEOI) and percent of estrus at which ovulation occurred (DE%).

Figure 6. Effect of parity on duration of estrus (DE), onset of estrus-to-ovulation interval (OEOI) and percent of estrus at which ovulation occurred (DE%).

Differences between three sow genotypes (A, B, C) with 0%, 100% and 75% heterosis, respectively, were examined. Even though there was no difference (P > 0.10) in DE or OEOI between sow genotypes, since genotype C sows had the longest DE and shortest OEOI compared to the other genotypes they tended to (P < 0.10) ovulate at a lower DE% than the other two genotypes (Figure 7.). This trend is interesting but does not suggest that genotype specific insemination regimens are necessary in this herd because the OEOI of genotype C sows was only 3 to 4 h shorter than genotype A and B sows. However, this finding may indicate that genotype can be an important source of the variation in estrus and time of ovulation characteristics between sows.

Figure 7. Effect of genotype on duration of estrus (DE), onset of estrus-to-ovulation interval (OEOI) and percent of estrus at which ovulation occurred (DE%).

Finally, when sows were compared by WEI, as Kemp and Soede (1996) did (Figure 1.), the relationships between WEI, DE and time of ovulation in the present study were similar to their findings for some but not all WEI categories (Figure 8.). Sows that had a WEI of 4 d did have a shorter DE than sows with a WEI of 3 d (57.5 ± 2.1 vs. 66.0 ± 3.2 h, P < 0.04) and even though their OEOI was not significantly shorter (P > 0.10) they did not have different DE% (P > 0.10). Sows that had a 5 and 6 d WEI did not exhibit an additional reduction of DE or OEOI compared to sows that had a 4 d WEI. Furthermore, one might argue that sows that had a WEI of 6 d had a longer OEOI and ovulated later in estrus (DE%) compared to the other WEI categories. However, these differences were not significant (P > 0.10) and may be attributed to the smaller number of observations (n = 6) in the 6 d WEI category. Overall, the relationships between WEI, DE and time of ovulation in the present study were similar to those reported by Kemp and Soede (1996) but also underscore the variation in these characteristics between herds. Furthermore, such differences may explain why very different reproductive results can occur for sows in each WEI category in different herds when the same insemination regimen is applied (Flowers, 1998b).

Figure 8. Effect of weaning-to-estrus interval (WEI) on duration of estrus (DE), onset of estrus-to-ovulation interval (OEOI) and percent of estrus at which ovulation occurred (DE%). ^{a, b} Duration of estrus columns lacking a similar superscript letter are different (P < 0.05).

This study will be repeated in the same commercial herd and twice in a second herd to examine seasonal and herd differences in estrus and ovulation characteristics.

Implications

These data suggest that the short lactation lengths recently adopted by the US swine industry do not alter the general temporal relationships between estrus and ovulation reported in the literature. While there does not appear to be any difference between the estrus and ovulation characteristics of parity 1 to 10 sows in the same herd, gilts were not examined. Furthermore, the repeatability of estrus and ovulation characteristics over a female’s lifetime from gilt to mature sow has, to our knowledge, not been reported. Sow genotype may be an important source of variation in estrus and ovulation characteristics and merits further investigation. Relationships between weaning-to-estrus interval, duration of estrus and time of ovulation are likely somewhat herd specific.

Literature Cited

Dial, G.D. 1995. Optimizing breeding performance in early weaning systems: factors

affecting the reproductive response to lactation length. Recent Advances

in Swine Production Health. pp 101-112.

Flowers, W.L. 1998a. Insemination programs for swine to increase fertility. J. Anim. Sci.

76(Suppl. 3):39-46.

Flowers, W.L. 1998b. Management of reproduction. In: J. Wiseman, M.A. Varley and

J.P. Chadwick (Eds.) Progress in Pig Science. pp 383-405. Nottingham University

Press, Thrumpton, Nottingham, UK.

Kemp, B. and N.M. Soede. 1996. Relationship of weaning-to-estrus interval to timing of

ovulation and fertilization in sows. J. Anim. Sci. 74:944-949.

PigCHAMP. 2000. Global Benchmarking in Swine Herds. PigCHAMPÒ, Inc. p. 7.

SAS. 1990. SAS/STATÒ User’s Guide, (Version 6, 4^th ed.). SAS Institute Inc., Cary, NC.

Soede, N.M. and B. Kemp. 1997. Expression of oestrus and timing of ovulation in pigs.

J. Reprod. Fertil., Suppl. 52:91-103.