Reproductive performance of early-weaned female
swine according to their estrus pro®le and
frequency of arti®cial insemination
M.N. CorreÃa, T. Lucia Jr.
*, J.A.B. Afonso, J.C. Deschamps
PIGPEL, Centro de Biotecnologia, Faculdade de VeterinaÂria±Universidade Federal de PelotasCampus UniversitaÂrio, 96010-900 Pelotas-RS, Brazil Received 8 March 2001; accepted 20 December 2001
Abstract
We determined the estrus pro®le (weaning-to-estrus interval (WEI), estrus duration (ED), and frequency of estrus per detection period) in 184 female swine and estimated the effect of the WEI, ED and frequency of arti®cial insemination (AI) on farrowing rate (FR) and litter size. Estrus detection was done at 8:30 a.m. and 5:00 p.m. The WEI was categorized as short (<100 h), medium (100± 120 h) and long (>120 h). The ED was categorized as short (<60 h), medium (60±72 h) and long (>72 h). Mean lactation length was 14.6 days, mean WEI was 124.5 h and mean ED was 69 h. In each weaning group, females received either one or two AI, following a breeding schedule based on the estrus pro®le. In single-mated females, AI was performed 36 h after the beginning of estrus. In double-mated females, the ®rst AI was done 24 h after the beginning of estrus and the second AI occurred 12 h later. The period of estrus detection had no effect (P > 0:05) on WEI, ED, FR, total born (TB) and live born litter size (LB). Mean FR was 82.6%, mean TB was 10.0% and mean LB was 9.2%. Mean ED was shorter (P < 0:03) for females having medium and long WEI (67.0 and 65.4 h, respectively) than for those having short WEI (72.2 h). A linear regression analysis identi®ed a weak (R2 0:02) but signi®cant negative association between ED and WEI (P 0:05). The WEI did not
in¯uence FR (P > 0:05). Total litter size for females having short WEI (9.4) was lower (P < 0:03) than for those having long WEI (10.4). Also, LB for females having medium and long WEI (9.7±9.8) was higher (P < 0:05) than for those having short WEI (8.7). AI frequency had no effect on FR (P > 0:05). TB and LB litter size were lower (P < 0:05) for single-mated females (9.6 and 9.0, respectively) than for double-mated females (10.7 and 9.6, respectively). Double AI was associated with higher subsequent litter size. However, breeding schedules based only on estrus pro®le may not be precise in determining ideal breeding time, since females having short WEI had the longest ED and produced the lowest litter size. # 2002 Elsevier Science Inc. All rights reserved.
Keywords: Estrus pro®le; Arti®cial insemination frequency; Farrowing rate; Litter size
*Corresponding author.
E-mail address: pigpel@ufpel.tche.br (T. Lucia Jr.).
0093-691X/02/$ ± see front matter # 2002 Elsevier Science Inc. All rights reserved. PII: S 0 0 9 3 - 6 9 1 X ( 0 2 ) 0 0 8 6 8 - 3
1. Introduction
The factor having the highest in¯uence on the number of pigs weaned per female per year is the mean nonproductive days (NPD), which are those days when a breeding female is neither gestating nor lactating [1,2]. The weaning-to-estrus interval (WEI) is a
non-productive period with high impact on mean NPD[1,3]. The mean WEI in commercial
farms generally is between 5 and 7 days[4,5]. However, nearly 80% of all females are
detected in estrus no later than 6 days post-weaning [4,6,7], and usually have better
subsequent farrowing rates (FR) and litter sizes than those having longer WEI [4,7±9]. Several studies reported that maximum fertility would be achieved when females are
bred within a range of 28 h before ovulation, to 8 h after ovulation [10±13], which
generally occurs in the third part of the estrus. However, there is a negative association between WEI and estrus duration (ED) characterized by longer ED in females having
shorter WEI [6,10±12]. Thus, breeding females should be mated at different times
following the pattern of their WEI.
If the time of ovulation could be determined precisely, a single mating per estrus would probably be effective in reaching target FR and litter size[1], but as the time of ovulation is commonly unknown, most farms use multiple matings per estrus. Although used in many farms, triple mating per estrus does not improve subsequent FR and litter size when
compared to double mating [7,14±16]. Comparisons between single and double mating
identi®ed no differences in subsequent litter size for primiparous and multiparous females, although double mating improved litter size in parity-0 females[7,16]. Nevertheless, the effect of mating frequency in FR is controversial. Double mating can improve subsequent
FR regardless of parity [16], although there are reports that such an effect is not
characteristic [7]. Flowers and Esbenshade [15] reported improvement in both FR and
litter size after double mating in comparison with single mating.
In most of the studies conducted in this ®eld, mating frequencies have been determined by the breeding schedule practiced at farm level. That fact implies that such studies are prone to confounding because the females were not randomly assigned to receive a given mating frequency, and most single-mated females were mated only once because their estrus was either short or detected inef®ciently[16]. Even when mating frequencies were assigned to females[7], the breeding schedule did not consider the negative association between WEI and ED. Also, many studies compared distinct mating frequencies using
natural mating [7,9,16], without considering AI. Therefore, as the effect of mating
frequency on reproductive performance seems to depend upon ED [15], knowing the
estrus pro®le could help to optimize AI ef®ciency. This study described the estrus pro®le of early-weaned female swine in a commercial farm, and evaluated the effects of WEI, ED and of an assigned AI frequency (single or double) on FR and litter size of females inseminated following a breeding schedule based on the observed estrus pro®le. 2. Material and methods
This study was conducted during a 4-month period, on a commercial farm in the Rio Grande do Sul State (Brazil) having an average inventory of 1600 breeding females and a
target lactation length of 14 days. The females were all F1 (PIC Camborough 221). All
females were housed in individual cells, in a naturally ventilated barn throughout the entire breeding and gestation period. From the post-weaning period to the time of mating, the females were fed a lactation diet twice daily having 18% crude protein, with ad libitum access.
Estrus detection followed the management practice of the farm, based on positive response to back pressure by a technician in the presence of a boar, at two periods (8:30 a.m. and 5:00 p.m.). The start of the estrus was characterized as the mid-point between the ®rst positive response to back pressure and the previous period of estrus detection [6,17]. The end of the estrus was characterized similarly, using the interval between the last positive response to back pressure and the subsequent period of estrus detection.
The estrus pro®le was characterized according to the frequency distributions of both WEI and ED. Following its frequency distribution (Fig. 1), the WEI was grouped in three categories: short, when shorter than 100 h; medium, when between 100 and 120 h; and long, when longer than 120 h. The mean WEI for the short, medium and long groups were: 90:6 9:1; 116:5 5:7; and 198:0 62:0 h, respectively.
The ED was also categorized based upon its frequency distribution (Fig. 2) as: short, when shorter than 60 h; medium, when between 60 and 72 h; and long, when longer than 72 h. The mean ED was: 53:4 6:1 h for females having short estrus; 70:9 2:8 h for females having medium estrus; and 87:4 8:2 h for those having long estrus. The
distribution of both WEI and ED by category is shown inTable 1.
In each weaning group, one-half of the breeding females were selected at random to receive either one or two AI, with a dose of 5 109spermatozoa. For females receiving
only one AI, the single mating was conducted 36 h after the starting point of the estrus. For those receiving two matings, the ®rst mating was conducted 24 h after the starting point of the estrus and the second mating was conducted 12 h after the ®rst.
A total of 184 females were included in the study. While 96 females (52.2%) received one AI, 88 females (47.8%) received two AI. The differences in the number of females across groups are due to culling. Of all females, 10 (5.4%) were parity-1, 21 (11.4%) were
parity-2, 27 (14.7%) were parity-3 and 126 (68.5%) were parity-4. For further compar-isons, lactation length was categorized as: 13; 14±15; 16±17; and 18 days.
All statistical analyses were conducted using the SAS1software[18]. Chi-square tests
were used to compare frequencies of sows coming into estrus in each period of estrus detection and also to test differences in FR across the following independent variables: period of estrus detection, parity, lactation length, WEI, ED, and AI frequency. ANOVA was conducted to test the effects of period of estrus detection, parity and lactation length on WEI, as well as the effects of period of estrus detection, parity, lactation length and WEI on ED. A linear regression model estimated the strength of the effect of WEI on ED. ANOVA was also used to analyze how total born (TB) and live born (LB) litter size would be in¯uenced by the following independent variables: period of estrus detection, parity, lactation length, WEI, ED, and AI frequency. Interactions between the mentioned independent variables were also tested. All ANOVA models were generated through the GLM procedure, and the least square means were compared through the PDIFF procedure[18].
3. Results
The mean lactation length during the period studied was 14.6 days (Table 2). The mean WEI was 24.5 h (5.2 days) and the mean ED was 69.0 h (2.9 days).
Fig. 2. Frequency distribution of the ED (n 184). Table 1
Frequency of WEI and ED by category
Category WEI ED
Range (h) n % Range (h) n %
Short <100 85 46 <60 57 31
Medium 100±120 54 29 60±72 88 48
While 161 females (87.5%) were identi®ed in estrus at 8:30 a.m., 23 (12.5%) were detected in estrus at 5:00 p.m. (Table 3). The WEI, ED, FR and litter traits did not differ (P > 0:05) between time periods of ®rst estrus detection.
In this study, WEI and ED were not in¯uenced (P > 0:05) by either parity or lactation
length (Table 4). However, as shown inTable 5, mean ED for females having short WEI
was longer (P < 0:03) than for those having medium and long WEI (72.2 versus 67.0 and 65.4 h, respectively). According to a linear regression model, this association could be explained by the following equation: ED 73:4387 0:0356 (WEI). This model identi-®ed a weak (R2 0:02), although signi®cant linear association (P 0:05).
Out of all 184 females mated during the period of study, 152 (82.6%) eventually farrowed. The FR was not affected by WEI, ED (Table 6) or lactation length (Table 7).
Table 2
Descriptive statistics for parameters of reproductive performance
Parameter n Mean S.D. Minimum Maximum
Lactation length (day) 184 14.6 2.2 8.0 20.0
WEI (h) 184 124.5 53.4 47.5 359.5
ED (h) 184 69.0 13.4 24.0 104.5
TB litter size 152 10.0 2.8 3.0 16.0
LB litter size 152 9.2 2.8 1.0 15.0
Table 3
Influence of time of first estrus detection on parameters of reproductive performancea
Parameter Period of estrus detection
8:30 a.m. 5:00 p.m. Females in estrus (n) 161 23 Females in estrus (%) 87.5 12.5 WEI (h)b 131.3 6.4 131.8 12.2 ED (h)b 67.8 1.1 70.9 2.7 FR (%) 83.9 73.9 TB litter sizeb 10.1 0.3 10.3 0.6 LB litter sizeb 9.4 0.2 9.5 0.7
aNo statistically signi®cant differences observed for any variables (P > 0:05). bMean S:E:M.
Table 4
Mean S:E:M: for the WEI and ED by parity and lactation lengtha
Parity n WEI (h) ED (h) Lactation length (day) n WEI (h) ED (h) 1 10 135.4 17.3 74.8 4.3 13 58 136.3 8.0 68.3 2.0 2 21 143.0 11.8 65.0 2.9 14±15 63 127.2 8.3 67.7 2.0 3 27 125.1 10.4 68.5 2.5 16±17 44 123.4 9.3 71.1 2.3 4 126 121.9 5.3 68.6 1.4 18 19 138.6 13.1 69.7 3.2
Across parities (Table 7), the FR for primarous females (50.0%) was lower (P < 0:01) than for parity-3 (88.9%) and parity-4 females (84.9%). FR also did not differ across AI frequencies (P 0:20). For single-mated females the FR was 79.2%, whereas the FR for double-mated females was 86.4%.
Litter traits were not in¯uenced by ED (P > 0:05), although they differed as a function of WEI (Table 8). Total litter size for females having short WEI (9.4) was lower (P < 0:03)
Table 5
ED (mean S:E:M:) by WEI
WEI category n ED (h)
Short 85 72.2 1.4a
Medium 54 67.0 1.8b
Long 45 65.4 1.9b
Means having different superscripts differ across categories by at least P < 0:03. Table 6 FR by WEI and ED Category WEI ED n FR (%) n FR (%) Short 74 87.1 50 87.7 Medium 46 85.2 72 81.8 Long 32 71.1 30 76.9
Frequencies do not differ (P > 0:05). Table 7
FR by parity and lactation length
Parity n FR (%) Lactation length (day) n FR (%)
1 5 50.0a 13 46 79.3
2 16 76.2a,b 14±15 57 90.5
3 24 88.9b 16±17 34 77.3
4 107 84.9b 18 15 78.9
Frequencies having different superscripts in the same column differ by at least P < 0:01.
Table 8
TB and LB litter size by WEI and ED
Category WEI ED
n TB LB n TB LB
Short 74 9.4 0.3a 8.7 0.3c 50 10.3 0.4 9.8 0.5
Medium 46 10.6 0.4b 9.7 0.4d 72 10.1 0.3 10.0 0.5
Long 32 10.4 0.5a,b 9.8 0.5d 30 9.9 0.5 10.0 0.6
Means S:E:M: having different superscripts (a, b) across categories differ by P < 0:03. Means S:E:M: having different superscripts (c, d) across categories differ by P < 0:05.
than that observed for females having medium WEI (10.6). However, TB for females having long WEI (10.4) did not differ (P > 0:05) from that for females having either short or medium WEI. LB size was lower (P < 0:05) for females having short WEI (8.7) than for those having medium and long WEI (9.7 and 9.8, respectively). Litter traits estimated in this study were not in¯uenced (P > 0:05) by either parity or lactation length (Table 9).
Both TB and LB differed according to AI frequency (Table 10). Double-mated females
produced higher (P 0:02) TB (10.7) than those that were single-mated (9.6). The number of LB was also higher (P 0:04) for double-mated females (9.9) than for single-mated females (9.0).
4. Discussion
In the present study, mean litter size for double-mated females was higher than that observed for single-mated females by nearly one piglet. That ®nding indicates that litter size can bene®t from double AI instead of single AI, as reported in a study with natural
mating [14], but disagrees with studies that reported no improvement in litter size
associated with double natural mating for females having one or more parities [7,16].
In these studies, higher litter size following double mating was observed only for parity-0 females, which were not evaluated in the present study.
Estrus pro®le was used as a tool to help estimate the ideal breeding time. Ultrasound would provide more precise information about time of ovulation[6,10,11], but it is not used routinely in the ®eld due to the increased labor and management costs. Previous studies conducted in Europe reported that time of ovulation would be variable, occurring at either 35 8, 42 11 or 45 13 h after the beginning of estrus[6,11,12]. However, assuming that the observed estrus pro®le was precise in determining timing of ovulation, most ovulations would have occurred around 48 h after the starting point of estrus, since nearly
Table 9
TB and LB litter size by parity and lactation lengtha
Parity n TB LB Lactation length (day) n TB LB
1 5 10.2 1.2 9.8 1.3 13 46 10.1 0.5 9.4 0.5
2 16 11.3 0.7 10.9 0.7 14±15 57 10.4 0.5 9.8 0.5
3 24 10.2 0.6 9.7 0.6 16±17 34 10.3 0.6 9.8 0.6
4 107 10.1 0.3 9.3 0.3 18 15 11.1 0.8 10.6 0.8
aMeans do not differ (P > 0:05).
Table 10
TB and LB litter size by frequency of AI
AI frequency n TB LB
Single 76 9.6 0.3b 9.0 0.3d
Double 76 10.7 0.3a 9.9 0.3c
Means S:E:M: having different superscripts (a, b) across categories differ by P 0:02. Means S:E:M: having different superscripts (c, d) across categories differ by P 0:04.
70% of all females had ED of at least 60±72 h. As a consequence, single-mated females received their only AI at 36 h, nearly 12 h prior to the estimated ovulation time, whereas double-mated females were bred for the second time at the same period, after receiving the ®rst AI nearly 24 h prior to the estimated ovulation time. It is important to consider that the association between WEI and ED observed in this study was weaker than the ®gures reported in other studies conducted in Europe[10,11]. A negative association between WEI and ED was clearly characterized in Brazil only in primiparous early-weaned females treated with eCG post-weaning, while the association in untreated females was also weak
[17]. Although subsequent reproductive performance appears to be more in¯uenced by
mating frequency than by ED, the effect of mating frequency seems to depend on ED[15]. Thus, it is likely that the weak association between WEI and ED may explain why litter size was higher with double AI than for single AI, since the imprecise estimation of ovulation time would be more harmful for single-mated females than for those double-mated, which probably bene®ted from receiving the ®rst AI 24 h earlier. Also, it should be considered that all AI were conducted using semen extended for 24±48 h. Although fresh swine semen extended up to 48 h can be used without negative effects on subsequent fertility[19], its use after 48 h can be associated with reduced fertility rates, even when ovulation occurred 12± 24 h after AI[20]. As the prediction of ED based on the WEI was not precise, the effect of AI frequency on total litter size may have been in¯uenced by reduced semen viability since the single AI was performed 36 h after the estimated beginning of estrus. It is possible that semen aging could play a role in this process since aging of stored semen cannot be completely prevented, even with the use of long-term extenders[21]. The ®ndings of the present study indicate that the effect of AI frequency on reproductive performance depends on the precision of determining time of ovulation and that breeding schedules should not be based only on the association between WEI and ED, since this association may be in¯uenced by different periods of estrus detection, nutrition, genetics and management practices.
This study did not identify a signi®cant effect of AI frequency on FR, which agrees with
the ®ndings of Xue et al.[7]for natural mating. However, as improved FR were obtained
with double natural mating in other studies [15,16], the association between mating
frequency and FR still needs to be characterized clearly. Although the difference of nearly 7% points in FR in favor of double AI was not statistically signi®cant, such a difference may be important at ®eld level. Also, it is possible that signi®cant differences in FR according to AI frequency could not be detected with the sample size used in this study. The previous WEI had no signi®cant effect on FR, even though the FR of females having short and medium WEI was numerically higher than that for those having long WEI by more than 10% points. On the other hand, mean litter size was lower for females having
short previous WEI than for those having medium WEI. In previous studies [4,9,16],
females having short WEI presented improvement in both subsequent FR and litter size, but, in such studies, short WEI were at most 6 days long. In the present study, including females weaned on average after 14.6 days which can be considered as early weaning[22], short WEI lasted at most 4 days, medium WEI were between 4 and 5 days and the frequency of WEI longer than 6 days was almost 15%, which contrasts with reports that females weaned after 14 days would have higher probability of having a subsequent WEI longer than 6 days[9]. On the other hand, the distribution and the mean observed for the
WEI in the present study were similar to those observed in a previous study conducted in Brazil, including early-weaned females having a mean previous lactation length close to 18
days [17]. Although, some studies reported that early-weaned females would have
prolonged WEI, reduced FR and lower litter size in the subsequent parity in comparison
with those weaned after longer lactations [4,5,23,24], no differences in reproductive
parameters were observed across the range of lactation lengths analyzed in the present study. Thus, the association between previous WEI and subsequent performance observed in this study probably re¯ects management improvements achieved in many areas over the years that led to a substantial reduction in the mean WEI at ®eld level.
Acknowledgements
This research was funded by CNPQq (Conselho Nacional de Desenvolvimento Cien-tõ®co e TecnoloÂgico)ÐGrant #300702/97-3.
References
[1] Dial GD, Marsh WE, Polson DD, Vaillancourt J-P. Reproductive failure: differential diagnosis. In: Leman AD, Straw BE, Mengeling WL, D'Allaire S, Taylor DJ, editors. Diseases of swine. Ames, IA: Iowa State University Press, 1992. p. 88±137.
[2] Wilson MR, Friendship RM, MacMillan I, Hacker RR, Pieper R, Swaminathan S. A survey of productivity and its component interrelationship in Canadian swine herds. J Anim Sci 1986;62:576±82.
[3] Polson DD, Dial GD, Marsh WE. A biological and ®nancial characterization of nonproductive days. In: Proceedings of the Ninth International Congregation on Pig Vet Soc. Lausanne, Switzerland, 1990. p. 372. [4] Koketsu Y. Assessment of sows mating ef®ciency during the low productive period after early weaning: a
®eld study. Theriogenology 1999;51:1525±32.
[5] Koketsu Y, Dial GD. Factors in¯uencing the post-weaning reproductive performance of sows in commercial herds. Theriogenology 1997;47:1445±61.
[6] Weitze KF, Wagner-Rietschel H, Waberski D, Richter L, Krieter J. The onset of heat after weaning, heat duration and ovulation as major factors in AI timing in sows. Reprod Domest Anim 1994;29:433±43. [7] Xue JL, Dial GD, Trigg T, Davies PR, King VL. In¯uence of mating frequency on sow reproductive
performance. J Anim Sci 1998;76:2962±6.
[8] Vesseur PC, Kemp B, den Hartog LA. The effect of the weaning to oestrus interval on litter size, live born piglets and farrowing rate in sows. J Anim Physiol Anim Nutr 1994;71:30±8.
[9] Wilson MR, Dewey CE. The associations between weaning-to-estrus interval and sow ef®ciency. Swine Health Prod 1993;1:10±5.
[10] Kemp B, Soede NM. Relationship of weaning-to-estrus interval to timing of ovulation and fertilization in sow. J Anim Sci 1996;74:944±9.
[11] Nissen AK, Soede NM, Hyttel P, Schmidt M, D'Hoore L. The in¯uence of time of insemination relative to time of ovulation on farrowing frequency and litter size in sows, as investigated by ultrasonography. Theriogenology 1997;47:1571±82.
[12] Soede NM, Wetzels CCH, Zondag W, de Koning MAI, Kemp B. Effect of time of insemination relative to ovulation, as determined by ultrasonography, on fertilization rate and accessory sperm count in sows. J Reprod Fertil 1995;104:99±106.
[13] Waberski D, Weitze KF. Current timing of arti®cial insemination in pigs. Reprod Domest Anim 1996;31:527±30.
[14] Dewey CE, Martin SW, Friendship RM, Kennedy BW, Wilson MR. Associations between litter size and speci®c sow-level management factors in Ontario swine. Prev Vet Med 1995;23:101±10.
[15] Flowers WL, Esbenshade KL. Optimizing management of natural and arti®cial matings in swine. J Reprod Fertil (Suppl) 1993;48:217±28.
[16] Xue JL, Lucia T, Koketsu Y, Dial GD, Marsh WE. In¯uence of mating frequency and weaning-to-mating interval on sow reproductive performance. Swine Health Prod 1998;6:157±62.
[17] Lucia T, CorreÃa MN, Deschamps JC, Peruzzo IA, Matheus JEM, Aleixo JAG. In¯uence of equine chronic gonadotropin on weaning-to-estrus interval and estrus duration in early-weaned primiparous female swine. J Anim Sci 1999;77:3163±7.
[18] SAS1. SAS/STAT User's Guide (Release 6.03). SAS Inst. Inc., Cary, NC, 1991.
[19] Waberski D, Weitze KF, Lietmann C, LuÈbbert zur Lage W, Bortolozzo FP, Wilmen T, et al. The initial fertilizing capacity of long-term-stored liquid boar semen following pre- and post-ovulatory insemination. Theriogenology 1994;41:1367±77.
[20] Weitze KF, Habeck O, Willmen T, Rath D. Detection of ovulation in the sow using transcutaneous sonography. Zuchthygiene 1989;24:40±2.
[21] Johnson LA, Weitze KF, Fiser P, Maxwell WMC. Storage of boar semen. Anim Reprod Sci 2000;62: 143±72.
[22] Dial GD, Wiseman BS, Davies PR, Marsh WE, Molitor TW, Morrison RB, et al. Strategies employed in the United States for improving the health of swine. Pig News Inf 1992;13:111±23.
[23] Koketsu Y, Dial GD, King VL. In¯uence of various factors on farrowing rate on farms using early weaning. J Anim Sci 1997;75:2580±7.
[24] Xue JL, Dial GD, Marsh WE, Davies PR, Momont H. In¯uence of lactation length on sow productivity. Liv Prod Sci 1992;34:253±65.
