- Email: [email protected]
Interactions between the associations of parity, lactation length, and weaning-to-conception interval with subsequent litter size in swine herds using early weaning
Preventive Veterinary Medicine 37 (1998) 113±120
Interactions between the associations of parity, lactation length, and weaning-to-conception interval with subsequent litter size in swine herds using early weaning Yuzo Koketsu*, Gary D. Dial Department of Clinical and Population Sciences, Commonwealth Avenue, room 257, College of Veterinary Medicine, University of Minnesota, St. Paul, MN 55108, USA Accepted 28 July 1998
Abstract Sixteen farms having an average lactation length between 14.9 and 18.9 days were selected based upon producers' willingness to keep records of production and lactation feed intake. Data contained 9162 subsequent litter-size records abstracted from computerized production information-system files and lactation feed-intake records. Two-way interactions between the associations of parity and lactation length, and parity and weaning-to-conception interval with subsequent litter size were found in the same statistical model. Subsequent litter size did not increase in parities 1 and 2 as lactation length increased ± but subsequent litter size in sows of parities 3±6 increased. In only parity 1, sows with weaning-to-conception interval 6±12 days produced 0.5 fewer pigs at subsequent farrowing than those with weaning-to-conception interval 1±5 days. However, in multiparous sows, no difference in subsequent litter size was found between weaning-to-conception intervals 1±5 and 6±12 days. # 1998 Elsevier Science B.V. All rights reserved. Keywords: Sows; Litter size; Lactation duration; Parity
1. Introduction Progressive producers in North America are increasingly using segregated early weaning (SEW) to increase overall productivity in breeding and fattening herds (Dial et al., 1992). In the SEW practice, piglets are weaned early (at age 10±20 days) from their * Corresponding author. Tel.: +1 612 625 4767; fax: +1 612 625 8750; e-mail: [email protected] 0167-5877/98/$ ± see front matter # 1998 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 7 - 5 8 7 7 ( 9 8 ) 0 0 1 0 3 - 2
114
Y. Koketsu, G.D. Dial / Preventive Veterinary Medicine 37 (1998) 113±120
sow and moved to an off-site nursery while passive immunity still exists (Dial et al., 1992). This SEW practice improves growth performance (due, at least in part, to reducing pathogen transfer from sows to piglets). Furthermore, shortened lactation length increases the number of litters per female per year (Xue et al., 1993) and pigs weaned per farrowing crate per year (King et al., 1998) in breeding herds. However, uterine involution is not complete until 21±28 days after farrowing (Palmer et al., 1965), while short lactation length is suggested to decrease embryonic survival rates (Varley and Cole, 1976). Thus, subsequent litter size fell from 11.5 to 11.2 pigs when lactation length was reduced from 28 to 21 days (Xue et al., 1993). Although early weaning might not improve litter size at each farrowing at the sow level, farms using early weaning appropriately can increase number of litters and improve farrowing-facility efficiency at the herd level over those of farms using conventional weaning. Lactation length (Varley and Cole, 1976), parity (Rasbech, 1969), weaning-toconception interval (Clark and Leman, 1987; Wilson and Dewey, 1993), lactation energy intake (Kirkwood et al., 1988), and conception month (Xue et al., 1994) are associated with subsequent litter size. Although separate reports on each factor for subsequent litter size are available, these factors and interactions have not been well investigated in the same statistical model. Furthermore, previous reports have used relatively longer average lactation lengths and different ranges of lactation length. The early-weaned sow's response to such factors might be different from that of sows on farms using conventional weaning (3±5 weeks). Our objectives in this study were to observe the `effects' (associations) of possible factors on subsequent litter size, and the interaction between these factors on subsequent litter size of sows located on commercial farms using early weaning. 2. Materials and methods 2.1. Definitions of production parameters Litter size was defined in this study as the total pigs born/litter at the subsequent farrowing. Weaning-to-conception interval was defined as the interval from weaning to the conception mating, which was back-calculated from the subsequent farrowing date in the PigCHAMP (a computerized record-keeping system). Farrowing-to-conception interval was defined as the sum of the weaning-to-conception interval and lactation length. 2.2. Data collection and database setup Sixteen commercial farms in the Midwest USA were selected for this study based upon their willingness to keep records of production events, feed intakes during lactation and their herds having an average lactation length lower than 19 days. Day 19 was arbitrarily chosen as the selection criterion of early weaning because the minimum average lactation length was 18.9 days in a summary of swine breeding herds in North America (Stein et al., 1990). On all 16 farms, breeding females were retained in buildings, and they were
Y. Koketsu, G.D. Dial / Preventive Veterinary Medicine 37 (1998) 113±120
115
fed corn±soybean-meal diets. Group weaning systems were being practiced. At weaning, lactating sows in the same group were weaned and moved to a breeding section on the same day. Their replacement gilts were purchased as crossbred gilts from breeding companies (10 farms; 6376 sows) or were raised by their rotational breeding programmes (six farms; 2786 sows). Data collected from the commercial farms included: (a) reproductive performance recorded in the PigCHAMP information system; and (b) daily feed intake of sows during lactation. The amount of feed provided twice daily to each lactating sow was recorded during a one-year period (1991). The farms were visited at quarterly intervals to determine the accuracy of the amounts being fed to sows as recorded by producers, and to determine whether management practices, facilities, health status or diet had changed. During the initial visit before the project started, farm workers were trained to weigh 0.9, 1.8 and 2.7 kg of feed using a scoop and then reexamined at subsequent visits in the following procedure. Farm workers were requested to scoop amounts of 0.9, 1.8, and 2.7 kg of feed into pails. Pails were weighed using a standard spring scale. Mean measurement errors (% of intended amount) for scoop-amount weighing 0.9, 1.8 and 2.7 kg were 3.9, ÿ1.3 and ÿ4.1%, respectively. Sows with lactation length 1±7 and >28 days were excluded from the data set based on the logic that sows on commercial farms having poor performance are more likely to be weaned within seven days after farrowing and subsequently culled. Furthermore, sows with lactation longer than 28 days would be expected to have been used as nurse sows for litters other than their own. Thus, sows with lactation length between 8 and 28 days from farms having an average lactation length <19 days were used in this study. Twelve records of sows having weaning-to-conception intervals >180 days were also excluded. 2.3. Statistical analysis Descriptive statistics were obtained using the Univariate procedure in SAS (1988). The general linear model procedure in SAS was used for statistical models. The farm effect in the statistical model was treated as a block to adjust for farm-to-farm variation. The farm effect was also confirmed to be a significant random effect (p<0.001) in order to extend inferences to a wider set of farms than these 16 farms. The number of pigs born alive at the current farrowing was used as a covariate to adjust for each individual sow's current litter size. Each factor was grouped and used as a categorical variable to obtain least-squares means. We categorized the continuous data because we designed this project like a cohort study in epidemiology. We also intended to obtain a cross-classified table when an interaction was found. Thus, three lactation feed intake (ADFI) groups were constructed, namely <4.2 (low), 4.2±5.6 (medium), and >5.6 kg/day (high), which were divided using the 25th and 75th percentiles of ADFI. Four parity groups were constructed, namely 1, 2, 3±6 and 7. Six lactation-length groups were constructed, namely 8±10, 11±13, 14±16, 17±19, 20±22 and 23±28 days. Conception months were arbitrarily categorized into four periods (seasons). December to March, April to May, June to August and September to November were defined as winter, spring, summer and autumn, respectively. Three weaning-to-conception intervals were formed, namely 1±5, 6±12, and 13 days. Farrowing-to-conception interval was divided into eight groups so that each group had
116
Y. Koketsu, G.D. Dial / Preventive Veterinary Medicine 37 (1998) 113±120
>200 sows/group and groups were formed at intervals of three days (except for the first and last two groups), namely 12±17, 18±20, 21±23, 24±26, 27±29, 30±32, 33±47, and 48 days. The first statistical model included live pigs born at current farrowing, groups of parity, lactation length, lactation feed intake, farrowed season and weaning-to-conception interval, and the farm effect. All possible two- and three-way interaction expressions were also included in the model. Backward elimination (p>0.100) was used to make a final model. Using the same method, a separate model that included farrowing-toconception interval replacing weaning-to-conception interval and lactation length was constructed. Fisher's protected least-significant difference tests (Steel and Torrie, 1980) were performed to compare least-squares means of total pigs born at subsequent farrowing between the various factor groups. 3. Results Average lactation length (SE) of the 16 farms ranged from 14.90.10 to 18.90.20 days. Our final data set contained 9162 subsequent-litter-size records and 8206 lactation feed-intake records. All 16 farms had sows weaned in each of the six lactation-length groups. Summary statistics for subsequent total pigs born and various possible affecting factors for subsequent litter size are presented in Table 1. The low-feed-intake group (<4.2 kg) had a 0.25 or 0.28 smaller subsequent litter size than the other groups (p0.02) (Table 2). Sows conceived during the spring produced smaller (p0.04) subsequent litter sizes than those during the winter and autumn months (Table 2). Litter sizes produced by sows conceived during the summer were intermediate among the four periods (Table 2). As farrowing-to-conception interval increased from 12±17 to 48 day groups, subsequent litter size increased by 1.5 pigs (Table 2). No interaction between the association of lactation length and feed intake during lactation with subsequent litter size was found (p0.31). An interaction between the associations of parity and lactation length with subsequent litter size was found (p0.004). In parities 1 and 2, no differences in subsequent litter size were found between any lactation-length groups (p>0.30) (Table 3). In contrast to low-parity sows, sows of Table 1 Descriptive statistics for litter size and its possible risk factors Item Subsequent total pigs born Lactation length (days) Parity Lactation ADFI a (kg/day) Weaning-to-conception interval (days) Farrowing-to-conception interval (days) Pigs born alive at current farrowing a
ADFI is average daily feed intake.
n 9162 9162 9162 8206 9162 9162 9162
Mean 11.15 17.39 2.94 5.02 11.57 29.0 10.22
SE 0.031 0.036 0.020 0.012 0.193 0.194 0.028
Percentile 10th
90th
7 13 1 3.7 4 19 7
15 22 6 6.4 28 45 13
Y. Koketsu, G.D. Dial / Preventive Veterinary Medicine 37 (1998) 113±120
117
Table 2 Least-squares means (LSM) of total pigs born at subsequent farrowing by the factors Risk factor
n
LSM
SE
Lactation ADFI (kg/day) <4.2 4.2±5.6 >5.6
1893 4021 2292
11.10 a 11.35 b 11.38 b
0.11 0.09 0.11
Conception season December to March April to May June to August September to November
3338 1516 2003 2305
11.44 a 10.95 b 11.28 a,b 11.45 a
0.11 0.14 0.13 0.15
Farrowing-to-conception interval (days) 12±17 18±20 21±23 24±26 27±29 30±32 33±47 48
338 1613 2530 2058 935 271 652 765
10.41 a 10.96 b 10.90 b 10.98 b 11.04 b 11.39 b,c 11.98 c 11.86 c
0.23 0.11 0.09 0.09 0.14 0.31 0.22 0.20
d
a,b,c d
Means within the risk-factor group lacking a common superscript letter differ (p<0.05). ADFI is average daily feed intake.
parity 3 had increased subsequent litter size as lactation length increased beyond 11±13 days (Table 3). Furthermore, in lactation-length groups 14±28 days, parity-1 sows had lower subsequent litter sizes than parity 2 and 3±6 sows (p<0.022). In addition, no association between parity and subsequent litter size was found in the lactation length 13 days groups (p>0.38). An interaction between the associations of parity and weaning-to-conception interval with subsequent litter size was also found (p0.001). In only parity-1 sows, the weaningTable 3 Least-square means (LSM) of total pigs born at subsequent farrowing by parity and lactation length (LL) LL (days)
Parity 1 n
8±10 11±13 14±16 17±19 20±22 23±28 a,b,c d,e,f
54 310 960 731 500 206
2 LSM 10.74 10.71 10.80 d 10.77 d 10.56 d 10.56 d
SE 0.50 0.19 0.11 0.12 0.14 0.24
n 22 194 605 495 397 144
3±6 LSM 11.61 10.95 11.26 e 11.37 e 11.31 e 11.30 e
SE 0.71 0.23 0.14 0.15 0.17 0.28
n 50 342 1,407 1,175 871 239
7 LSM
SE b,c
11.83 10.72 a 11.39 b,e 11.40 b,e 11.60 b,e 12.15 c,f
n
0.45 9 0.18 52 0.10 122 0.11 135 0.12 102 0.22 40
LSM
SE a,b
11.89 10.64 a 11.52 a,b,e 11.41 a,b,e 11.93 b,e 12.43 b,f
Lactation-length means within a column lacking a common superscript letter differ (p<0.05). Parity-specific means within a row lacking a common superscript letter differ (p<0.05).
1.05 0.44 0.31 0.31 0.36 0.65
118
Y. Koketsu, G.D. Dial / Preventive Veterinary Medicine 37 (1998) 113±120
Table 4 Least-square means (LSM) of total pigs born at subsequent farrowing by parity and weaning-to-conception interval (WCI) LL (days)
Parity 1 n
1±5 6±12 13 a,b,c d,e,f
2 LSM
SE
1543 10.49 b,d 0.13 498 9.98 a,d 0.13 720 11.56 c 0.15
n
3±6 LSM
1304 11.13 a,e 297 10.84 a,e 256 11.93 b
7
SE
n
LSM
0.16 0.17 0.24
3,146 11.39 a,e 502 11.24 a,f 436 11.92 b
SE
n
0.10 369 0.12 42 0.17 49
LSM
SE
11.23 a,e 0.26 11.34 a,b,f 0.37 12.35 b 0.51
Weaning-to-conception-interval means within a column lacking a common superscript letter differ (p<0.05). Parity-specific means within a row lacking a common superscript letter differ (p<0.05).
to-conception interval 6±12 days group had smaller litter size than the 1±5 days group (p<0.001) (Table 4). As parity increased from 1 to 7, subsequent litter size in weaning-toconception interval 6±12 days increased by 1.4 piglets (Table 4). Parity was not associated with subsequent litter size in the weaning-to-conception interval 13 days group (p>0.11) (Table 4). 4. Discussion Our findings show that the negative associations between low parity, short lactation length and subsequent litter size are consistent with previous studies (Rasbech, 1969; Svajgr et al., 1974). Furthermore, our finding indicates an interaction between the associations of parity and lactation length with subsequent litter size ± consistent with previous observations suggesting that young animals are not affected by lactation-length changes (Walker et al., 1979; Clark and Leman, 1987). The mechanism for this interaction is not clear in pigs. It is possible that the duration of uterine involution affects the relationship between parity, lactation length and subsequent litter size. The endometrium in the uterus is regenerated between 14 and 21 days after farrowing in pigs (Jainudeen and Hafez, 1993). This repairing process of the endometrium in young sows may be faster than those of old parity sows ± as observed in cows (Zain et al., 1995). However, it is not recommended to use early weaning at 8±16 days for parity-1 sows, because short lactation length is also associated with low farrowing rate in young sows (Koketsu et al., 1997). It is speculated that young females are still developing their immature endocrine systems even though they can repair their uteri faster than can older parity sows. Subsequent litter size for lactation length 8±10 days was larger than that in lactation length 11±13 days in Table 3. Some statistical comparisons between means in lactationlength groups with few observations might lack statistical power. In our study, subsequent litter size did not increase between 4.2±5.6 and >5.6 kg/day groups, but did increase between <4.2 and 4.2±5.6 kg/day. Greater energy and feed intake during lactation is associated with higher embryo-survival proportions during early gestation (Baidoo et al., 1992) and greater subsequent litter size (Kirkwood et al., 1988).
Y. Koketsu, G.D. Dial / Preventive Veterinary Medicine 37 (1998) 113±120
119
Thus, our finding on lactation feed intake is consistent with their reports. Furthermore, no interaction (p0.31) between lactation feed intake and parity indicates that sows have a similar capability to increase litter size responding to lactation feed intake increase. Our finding indicates that increasing feed intake during lactation can reduce the negative association between short lactation length and subsequent litter size. The small litter size from sows conceived during the spring was not reported previously. The exposure of females to elevated ambient temperature within three weeks after mating decreases embryo survival percentage (Warnick et al., 1965). It is possible that high room temperatures during the spring influenced embryo survival and litter size in our study. Alternatively, our finding on litter size during the spring may be associated with other management factors which we did not measure. By contrast to weaning-to-conception interval groups, no substantial fluctuations or dips of subsequent litter size in farrowing-to-conception interval groups were found in our results. Subsequent litter size continuously increased as farrowing-to-conception interval increased. Positive effects of prolonged weaning-to-conception interval (13 days) on subsequent litter size might alleviate negative effects of short lactation length on litter size. Prolonged weaning-to-conception interval is related to low feed intake, short lactation length and low parity (Koketsu et al., 1996). Furthermore, Kemp and Soede (1996) suggested that small litter size occurring in a certain weaning-to-conception period such as 6±12 days are caused by a short duration of oestrus followed by suboptimal breeding timing. Presumably, sows which conceived between 6 and 12 days after weaning may not have inherent reproductive problems, but are affected by short oestrus and suboptimal breeding management following short lactation length and low feed intake, especially for parity-1 sows. This study is not a controlled experiment in the laboratory, but an observational study using field data. Thus, our findings should not be interpreted as causality but associations. Furthermore, our results could be biased by management factors or culling policy which we did not measure. However, even with these limitations, this research provides veterinarians with information regarding the effects of parity, lactation feed intake, lactation length and weaning-to-conception interval and those interactions on subsequent litter size. Acknowledgements The authors thank Dr. Vickie L. King for her statistical consulting. This research was supported in part by the Minnesota Agricultural Experiment Station. References Baidoo, S.K., Kirkwood, R.N., Aherne, F.X., Foxcroft, G.R., 1992. Effects of feed intake during lactation and after weaning on sow reproductive performance. Can. J. Anim. Sci. 72, 911±917. Clark, L.K., Leman, A.D., 1987. Factors that influence litter size in swine: Parity 3 through 7 females. J. Am. Vet. Med. Asso. 191, 49±58.
120
Y. Koketsu, G.D. Dial / Preventive Veterinary Medicine 37 (1998) 113±120
Dial, G.D., Wiseman, B.S., Davies, P.R., Marsh, W.E., Molitor, T.W., Morrison, R.B., Thawley, D.G., 1992. Strategies employed in the United States for improving the health of swine. Pig News Info. 13, 111±123. Jainudeen, M.R., Hafez, E.S.E., 1993. Gestation, Prenatal Physiology, and Parturition. In. (Ed.). Hafez. E.S.E. Reproduction in Farm Animals. Lea and Febiger. Philadelphia, 6th, 213±236. Kemp, B., Soede, N.M., 1996. Relationship of weaning-to-estrus interval to timing of ovulation and fertilization in sows. J. Anim. Sci. 74, 944±949. Kirkwood, R.N., Mitaru, B.N., Gooderatine, A.D., Blair, R., Thacker, P.A., 1988. The influence of dietary energy intake during successive lactations on sow prolificacy. Can. J. Anim. Sci. 68, 283±290. King, V.L., Koketsu, Y., Reeves, D., Xue, J.L., Dial, G.D., 1998. Management factors associated with swine breeding herd productivity in the United States. Prev. Vet. Med. 35, 255±264. Koketsu, Y., Dial, G.D., Pettigrew, J.E., King, V.L., 1996. Feed intake pattern during lactation and subsequent reproductive performance of sows. J. Anim. Sci. 74, 2875±2884. Koketsu, Y., Dial, G.D., King, V.L., 1997. Influence of various factors on farrowing rate on farms using early weaning. J. Anim. Sci. 75, 2580±2587. Palmer, W.M., Teague, H.S., Venzke, W.G., 1965. Macroscopic observations on the reproductive tract of the sow during lactation and early postweaning. J. Anim. Sci. 24, 541±545. Rasbech, N.O., 1969. A review of the causes of reproductive failure in swine. Bri. Vet. J. 125, 599±614. SAS, 1988. SAS/Stat User's Guide (Release 6.03). SAS Inst. Inc., Cary, NC. Steel, R.G.D., Torrie J.H., 1980. Principles and Procedures of Statistics: A Biometrical Approach (2nd Ed.) McGraw-Hill Publishing Co., New York, 181±183. Stein, T.E., Duffy, S.J., Wickstrom, S., 1990. Differences in production values between high- and lowproductivity swine breeding herds. J. Anim. Sci. 68, 3972±3979. Svajgr, A.L., Hays, V.W., Cromwell, G.L., 1974. Effect of lactation duration on reproductive performance of sows. J. Anim. Sci. 38, 100±105. Varley, M.A., Cole, D.J.A., 1976. Studies in sow reproduction 6. The effect of lactation length on preimplantation losses. Anim. Prod. 22, 209±214. Walker, N., Watt, D., MacLeod, A.S., Johnson, C.L., Boaz, T.G., Calder, A.F.C., 1979. The effect of weaning at 10, 25 or 40 days on the reproductive performance of sows from the first to the fifth parity. J. Agri. Sci. 92, 449±456. Warnick, A.C., Wallace, H.D., Palmer, A.F., Sosa, E., Duerre, D.J., Caldwell, V.E., 1965. Effect of temperature on early embryo survival in gilts. J. Anim. Sci. 24, 89±92. Wilson, M.R., Dewey, C.E., 1993. The association between weaning-to-mating interval and sow productivity. Swine Health Prod. 1, 10±15. Xue, J.L., Dial, G.D., Marsh, W.E., Davies, P.R., Momont, H.W., 1993. Influence of lactation length on sow productivity. Livest. Prod. Sci. 34, 253±265. Xue, J.L., Dial, G.D., Marsh, W.E., Davies, P.R., 1994. Multiple manifestations of season on reproductive performance of commercial swine. J. Am. Vet. Med. Asso. 204, 1486±1489. Zain, A.E.D., Nakao, T., Abdel Raouf, M., Moriyoshi, M., Kawata, K., Moritsu, Y., 1995. Factors in the resumption of ovarian activity and uterine involution in postpartum dairy cows. Anim. Reprod. Sci. 38, 203±214.
Interactions between the associations of parity, lactation length, and weaning-to-conception interval with subsequent litter size in swine herds using early weaning Yuzo Koketsu*, Gary D. Dial Department of Clinical and Population Sciences, Commonwealth Avenue, room 257, College of Veterinary Medicine, University of Minnesota, St. Paul, MN 55108, USA Accepted 28 July 1998
Abstract Sixteen farms having an average lactation length between 14.9 and 18.9 days were selected based upon producers' willingness to keep records of production and lactation feed intake. Data contained 9162 subsequent litter-size records abstracted from computerized production information-system files and lactation feed-intake records. Two-way interactions between the associations of parity and lactation length, and parity and weaning-to-conception interval with subsequent litter size were found in the same statistical model. Subsequent litter size did not increase in parities 1 and 2 as lactation length increased ± but subsequent litter size in sows of parities 3±6 increased. In only parity 1, sows with weaning-to-conception interval 6±12 days produced 0.5 fewer pigs at subsequent farrowing than those with weaning-to-conception interval 1±5 days. However, in multiparous sows, no difference in subsequent litter size was found between weaning-to-conception intervals 1±5 and 6±12 days. # 1998 Elsevier Science B.V. All rights reserved. Keywords: Sows; Litter size; Lactation duration; Parity
1. Introduction Progressive producers in North America are increasingly using segregated early weaning (SEW) to increase overall productivity in breeding and fattening herds (Dial et al., 1992). In the SEW practice, piglets are weaned early (at age 10±20 days) from their * Corresponding author. Tel.: +1 612 625 4767; fax: +1 612 625 8750; e-mail: [email protected] 0167-5877/98/$ ± see front matter # 1998 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 7 - 5 8 7 7 ( 9 8 ) 0 0 1 0 3 - 2
114
Y. Koketsu, G.D. Dial / Preventive Veterinary Medicine 37 (1998) 113±120
sow and moved to an off-site nursery while passive immunity still exists (Dial et al., 1992). This SEW practice improves growth performance (due, at least in part, to reducing pathogen transfer from sows to piglets). Furthermore, shortened lactation length increases the number of litters per female per year (Xue et al., 1993) and pigs weaned per farrowing crate per year (King et al., 1998) in breeding herds. However, uterine involution is not complete until 21±28 days after farrowing (Palmer et al., 1965), while short lactation length is suggested to decrease embryonic survival rates (Varley and Cole, 1976). Thus, subsequent litter size fell from 11.5 to 11.2 pigs when lactation length was reduced from 28 to 21 days (Xue et al., 1993). Although early weaning might not improve litter size at each farrowing at the sow level, farms using early weaning appropriately can increase number of litters and improve farrowing-facility efficiency at the herd level over those of farms using conventional weaning. Lactation length (Varley and Cole, 1976), parity (Rasbech, 1969), weaning-toconception interval (Clark and Leman, 1987; Wilson and Dewey, 1993), lactation energy intake (Kirkwood et al., 1988), and conception month (Xue et al., 1994) are associated with subsequent litter size. Although separate reports on each factor for subsequent litter size are available, these factors and interactions have not been well investigated in the same statistical model. Furthermore, previous reports have used relatively longer average lactation lengths and different ranges of lactation length. The early-weaned sow's response to such factors might be different from that of sows on farms using conventional weaning (3±5 weeks). Our objectives in this study were to observe the `effects' (associations) of possible factors on subsequent litter size, and the interaction between these factors on subsequent litter size of sows located on commercial farms using early weaning. 2. Materials and methods 2.1. Definitions of production parameters Litter size was defined in this study as the total pigs born/litter at the subsequent farrowing. Weaning-to-conception interval was defined as the interval from weaning to the conception mating, which was back-calculated from the subsequent farrowing date in the PigCHAMP (a computerized record-keeping system). Farrowing-to-conception interval was defined as the sum of the weaning-to-conception interval and lactation length. 2.2. Data collection and database setup Sixteen commercial farms in the Midwest USA were selected for this study based upon their willingness to keep records of production events, feed intakes during lactation and their herds having an average lactation length lower than 19 days. Day 19 was arbitrarily chosen as the selection criterion of early weaning because the minimum average lactation length was 18.9 days in a summary of swine breeding herds in North America (Stein et al., 1990). On all 16 farms, breeding females were retained in buildings, and they were
Y. Koketsu, G.D. Dial / Preventive Veterinary Medicine 37 (1998) 113±120
115
fed corn±soybean-meal diets. Group weaning systems were being practiced. At weaning, lactating sows in the same group were weaned and moved to a breeding section on the same day. Their replacement gilts were purchased as crossbred gilts from breeding companies (10 farms; 6376 sows) or were raised by their rotational breeding programmes (six farms; 2786 sows). Data collected from the commercial farms included: (a) reproductive performance recorded in the PigCHAMP information system; and (b) daily feed intake of sows during lactation. The amount of feed provided twice daily to each lactating sow was recorded during a one-year period (1991). The farms were visited at quarterly intervals to determine the accuracy of the amounts being fed to sows as recorded by producers, and to determine whether management practices, facilities, health status or diet had changed. During the initial visit before the project started, farm workers were trained to weigh 0.9, 1.8 and 2.7 kg of feed using a scoop and then reexamined at subsequent visits in the following procedure. Farm workers were requested to scoop amounts of 0.9, 1.8, and 2.7 kg of feed into pails. Pails were weighed using a standard spring scale. Mean measurement errors (% of intended amount) for scoop-amount weighing 0.9, 1.8 and 2.7 kg were 3.9, ÿ1.3 and ÿ4.1%, respectively. Sows with lactation length 1±7 and >28 days were excluded from the data set based on the logic that sows on commercial farms having poor performance are more likely to be weaned within seven days after farrowing and subsequently culled. Furthermore, sows with lactation longer than 28 days would be expected to have been used as nurse sows for litters other than their own. Thus, sows with lactation length between 8 and 28 days from farms having an average lactation length <19 days were used in this study. Twelve records of sows having weaning-to-conception intervals >180 days were also excluded. 2.3. Statistical analysis Descriptive statistics were obtained using the Univariate procedure in SAS (1988). The general linear model procedure in SAS was used for statistical models. The farm effect in the statistical model was treated as a block to adjust for farm-to-farm variation. The farm effect was also confirmed to be a significant random effect (p<0.001) in order to extend inferences to a wider set of farms than these 16 farms. The number of pigs born alive at the current farrowing was used as a covariate to adjust for each individual sow's current litter size. Each factor was grouped and used as a categorical variable to obtain least-squares means. We categorized the continuous data because we designed this project like a cohort study in epidemiology. We also intended to obtain a cross-classified table when an interaction was found. Thus, three lactation feed intake (ADFI) groups were constructed, namely <4.2 (low), 4.2±5.6 (medium), and >5.6 kg/day (high), which were divided using the 25th and 75th percentiles of ADFI. Four parity groups were constructed, namely 1, 2, 3±6 and 7. Six lactation-length groups were constructed, namely 8±10, 11±13, 14±16, 17±19, 20±22 and 23±28 days. Conception months were arbitrarily categorized into four periods (seasons). December to March, April to May, June to August and September to November were defined as winter, spring, summer and autumn, respectively. Three weaning-to-conception intervals were formed, namely 1±5, 6±12, and 13 days. Farrowing-to-conception interval was divided into eight groups so that each group had
116
Y. Koketsu, G.D. Dial / Preventive Veterinary Medicine 37 (1998) 113±120
>200 sows/group and groups were formed at intervals of three days (except for the first and last two groups), namely 12±17, 18±20, 21±23, 24±26, 27±29, 30±32, 33±47, and 48 days. The first statistical model included live pigs born at current farrowing, groups of parity, lactation length, lactation feed intake, farrowed season and weaning-to-conception interval, and the farm effect. All possible two- and three-way interaction expressions were also included in the model. Backward elimination (p>0.100) was used to make a final model. Using the same method, a separate model that included farrowing-toconception interval replacing weaning-to-conception interval and lactation length was constructed. Fisher's protected least-significant difference tests (Steel and Torrie, 1980) were performed to compare least-squares means of total pigs born at subsequent farrowing between the various factor groups. 3. Results Average lactation length (SE) of the 16 farms ranged from 14.90.10 to 18.90.20 days. Our final data set contained 9162 subsequent-litter-size records and 8206 lactation feed-intake records. All 16 farms had sows weaned in each of the six lactation-length groups. Summary statistics for subsequent total pigs born and various possible affecting factors for subsequent litter size are presented in Table 1. The low-feed-intake group (<4.2 kg) had a 0.25 or 0.28 smaller subsequent litter size than the other groups (p0.02) (Table 2). Sows conceived during the spring produced smaller (p0.04) subsequent litter sizes than those during the winter and autumn months (Table 2). Litter sizes produced by sows conceived during the summer were intermediate among the four periods (Table 2). As farrowing-to-conception interval increased from 12±17 to 48 day groups, subsequent litter size increased by 1.5 pigs (Table 2). No interaction between the association of lactation length and feed intake during lactation with subsequent litter size was found (p0.31). An interaction between the associations of parity and lactation length with subsequent litter size was found (p0.004). In parities 1 and 2, no differences in subsequent litter size were found between any lactation-length groups (p>0.30) (Table 3). In contrast to low-parity sows, sows of Table 1 Descriptive statistics for litter size and its possible risk factors Item Subsequent total pigs born Lactation length (days) Parity Lactation ADFI a (kg/day) Weaning-to-conception interval (days) Farrowing-to-conception interval (days) Pigs born alive at current farrowing a
ADFI is average daily feed intake.
n 9162 9162 9162 8206 9162 9162 9162
Mean 11.15 17.39 2.94 5.02 11.57 29.0 10.22
SE 0.031 0.036 0.020 0.012 0.193 0.194 0.028
Percentile 10th
90th
7 13 1 3.7 4 19 7
15 22 6 6.4 28 45 13
Y. Koketsu, G.D. Dial / Preventive Veterinary Medicine 37 (1998) 113±120
117
Table 2 Least-squares means (LSM) of total pigs born at subsequent farrowing by the factors Risk factor
n
LSM
SE
Lactation ADFI (kg/day) <4.2 4.2±5.6 >5.6
1893 4021 2292
11.10 a 11.35 b 11.38 b
0.11 0.09 0.11
Conception season December to March April to May June to August September to November
3338 1516 2003 2305
11.44 a 10.95 b 11.28 a,b 11.45 a
0.11 0.14 0.13 0.15
Farrowing-to-conception interval (days) 12±17 18±20 21±23 24±26 27±29 30±32 33±47 48
338 1613 2530 2058 935 271 652 765
10.41 a 10.96 b 10.90 b 10.98 b 11.04 b 11.39 b,c 11.98 c 11.86 c
0.23 0.11 0.09 0.09 0.14 0.31 0.22 0.20
d
a,b,c d
Means within the risk-factor group lacking a common superscript letter differ (p<0.05). ADFI is average daily feed intake.
parity 3 had increased subsequent litter size as lactation length increased beyond 11±13 days (Table 3). Furthermore, in lactation-length groups 14±28 days, parity-1 sows had lower subsequent litter sizes than parity 2 and 3±6 sows (p<0.022). In addition, no association between parity and subsequent litter size was found in the lactation length 13 days groups (p>0.38). An interaction between the associations of parity and weaning-to-conception interval with subsequent litter size was also found (p0.001). In only parity-1 sows, the weaningTable 3 Least-square means (LSM) of total pigs born at subsequent farrowing by parity and lactation length (LL) LL (days)
Parity 1 n
8±10 11±13 14±16 17±19 20±22 23±28 a,b,c d,e,f
54 310 960 731 500 206
2 LSM 10.74 10.71 10.80 d 10.77 d 10.56 d 10.56 d
SE 0.50 0.19 0.11 0.12 0.14 0.24
n 22 194 605 495 397 144
3±6 LSM 11.61 10.95 11.26 e 11.37 e 11.31 e 11.30 e
SE 0.71 0.23 0.14 0.15 0.17 0.28
n 50 342 1,407 1,175 871 239
7 LSM
SE b,c
11.83 10.72 a 11.39 b,e 11.40 b,e 11.60 b,e 12.15 c,f
n
0.45 9 0.18 52 0.10 122 0.11 135 0.12 102 0.22 40
LSM
SE a,b
11.89 10.64 a 11.52 a,b,e 11.41 a,b,e 11.93 b,e 12.43 b,f
Lactation-length means within a column lacking a common superscript letter differ (p<0.05). Parity-specific means within a row lacking a common superscript letter differ (p<0.05).
1.05 0.44 0.31 0.31 0.36 0.65
118
Y. Koketsu, G.D. Dial / Preventive Veterinary Medicine 37 (1998) 113±120
Table 4 Least-square means (LSM) of total pigs born at subsequent farrowing by parity and weaning-to-conception interval (WCI) LL (days)
Parity 1 n
1±5 6±12 13 a,b,c d,e,f
2 LSM
SE
1543 10.49 b,d 0.13 498 9.98 a,d 0.13 720 11.56 c 0.15
n
3±6 LSM
1304 11.13 a,e 297 10.84 a,e 256 11.93 b
7
SE
n
LSM
0.16 0.17 0.24
3,146 11.39 a,e 502 11.24 a,f 436 11.92 b
SE
n
0.10 369 0.12 42 0.17 49
LSM
SE
11.23 a,e 0.26 11.34 a,b,f 0.37 12.35 b 0.51
Weaning-to-conception-interval means within a column lacking a common superscript letter differ (p<0.05). Parity-specific means within a row lacking a common superscript letter differ (p<0.05).
to-conception interval 6±12 days group had smaller litter size than the 1±5 days group (p<0.001) (Table 4). As parity increased from 1 to 7, subsequent litter size in weaning-toconception interval 6±12 days increased by 1.4 piglets (Table 4). Parity was not associated with subsequent litter size in the weaning-to-conception interval 13 days group (p>0.11) (Table 4). 4. Discussion Our findings show that the negative associations between low parity, short lactation length and subsequent litter size are consistent with previous studies (Rasbech, 1969; Svajgr et al., 1974). Furthermore, our finding indicates an interaction between the associations of parity and lactation length with subsequent litter size ± consistent with previous observations suggesting that young animals are not affected by lactation-length changes (Walker et al., 1979; Clark and Leman, 1987). The mechanism for this interaction is not clear in pigs. It is possible that the duration of uterine involution affects the relationship between parity, lactation length and subsequent litter size. The endometrium in the uterus is regenerated between 14 and 21 days after farrowing in pigs (Jainudeen and Hafez, 1993). This repairing process of the endometrium in young sows may be faster than those of old parity sows ± as observed in cows (Zain et al., 1995). However, it is not recommended to use early weaning at 8±16 days for parity-1 sows, because short lactation length is also associated with low farrowing rate in young sows (Koketsu et al., 1997). It is speculated that young females are still developing their immature endocrine systems even though they can repair their uteri faster than can older parity sows. Subsequent litter size for lactation length 8±10 days was larger than that in lactation length 11±13 days in Table 3. Some statistical comparisons between means in lactationlength groups with few observations might lack statistical power. In our study, subsequent litter size did not increase between 4.2±5.6 and >5.6 kg/day groups, but did increase between <4.2 and 4.2±5.6 kg/day. Greater energy and feed intake during lactation is associated with higher embryo-survival proportions during early gestation (Baidoo et al., 1992) and greater subsequent litter size (Kirkwood et al., 1988).
Y. Koketsu, G.D. Dial / Preventive Veterinary Medicine 37 (1998) 113±120
119
Thus, our finding on lactation feed intake is consistent with their reports. Furthermore, no interaction (p0.31) between lactation feed intake and parity indicates that sows have a similar capability to increase litter size responding to lactation feed intake increase. Our finding indicates that increasing feed intake during lactation can reduce the negative association between short lactation length and subsequent litter size. The small litter size from sows conceived during the spring was not reported previously. The exposure of females to elevated ambient temperature within three weeks after mating decreases embryo survival percentage (Warnick et al., 1965). It is possible that high room temperatures during the spring influenced embryo survival and litter size in our study. Alternatively, our finding on litter size during the spring may be associated with other management factors which we did not measure. By contrast to weaning-to-conception interval groups, no substantial fluctuations or dips of subsequent litter size in farrowing-to-conception interval groups were found in our results. Subsequent litter size continuously increased as farrowing-to-conception interval increased. Positive effects of prolonged weaning-to-conception interval (13 days) on subsequent litter size might alleviate negative effects of short lactation length on litter size. Prolonged weaning-to-conception interval is related to low feed intake, short lactation length and low parity (Koketsu et al., 1996). Furthermore, Kemp and Soede (1996) suggested that small litter size occurring in a certain weaning-to-conception period such as 6±12 days are caused by a short duration of oestrus followed by suboptimal breeding timing. Presumably, sows which conceived between 6 and 12 days after weaning may not have inherent reproductive problems, but are affected by short oestrus and suboptimal breeding management following short lactation length and low feed intake, especially for parity-1 sows. This study is not a controlled experiment in the laboratory, but an observational study using field data. Thus, our findings should not be interpreted as causality but associations. Furthermore, our results could be biased by management factors or culling policy which we did not measure. However, even with these limitations, this research provides veterinarians with information regarding the effects of parity, lactation feed intake, lactation length and weaning-to-conception interval and those interactions on subsequent litter size. Acknowledgements The authors thank Dr. Vickie L. King for her statistical consulting. This research was supported in part by the Minnesota Agricultural Experiment Station. References Baidoo, S.K., Kirkwood, R.N., Aherne, F.X., Foxcroft, G.R., 1992. Effects of feed intake during lactation and after weaning on sow reproductive performance. Can. J. Anim. Sci. 72, 911±917. Clark, L.K., Leman, A.D., 1987. Factors that influence litter size in swine: Parity 3 through 7 females. J. Am. Vet. Med. Asso. 191, 49±58.
120
Y. Koketsu, G.D. Dial / Preventive Veterinary Medicine 37 (1998) 113±120
Dial, G.D., Wiseman, B.S., Davies, P.R., Marsh, W.E., Molitor, T.W., Morrison, R.B., Thawley, D.G., 1992. Strategies employed in the United States for improving the health of swine. Pig News Info. 13, 111±123. Jainudeen, M.R., Hafez, E.S.E., 1993. Gestation, Prenatal Physiology, and Parturition. In. (Ed.). Hafez. E.S.E. Reproduction in Farm Animals. Lea and Febiger. Philadelphia, 6th, 213±236. Kemp, B., Soede, N.M., 1996. Relationship of weaning-to-estrus interval to timing of ovulation and fertilization in sows. J. Anim. Sci. 74, 944±949. Kirkwood, R.N., Mitaru, B.N., Gooderatine, A.D., Blair, R., Thacker, P.A., 1988. The influence of dietary energy intake during successive lactations on sow prolificacy. Can. J. Anim. Sci. 68, 283±290. King, V.L., Koketsu, Y., Reeves, D., Xue, J.L., Dial, G.D., 1998. Management factors associated with swine breeding herd productivity in the United States. Prev. Vet. Med. 35, 255±264. Koketsu, Y., Dial, G.D., Pettigrew, J.E., King, V.L., 1996. Feed intake pattern during lactation and subsequent reproductive performance of sows. J. Anim. Sci. 74, 2875±2884. Koketsu, Y., Dial, G.D., King, V.L., 1997. Influence of various factors on farrowing rate on farms using early weaning. J. Anim. Sci. 75, 2580±2587. Palmer, W.M., Teague, H.S., Venzke, W.G., 1965. Macroscopic observations on the reproductive tract of the sow during lactation and early postweaning. J. Anim. Sci. 24, 541±545. Rasbech, N.O., 1969. A review of the causes of reproductive failure in swine. Bri. Vet. J. 125, 599±614. SAS, 1988. SAS/Stat User's Guide (Release 6.03). SAS Inst. Inc., Cary, NC. Steel, R.G.D., Torrie J.H., 1980. Principles and Procedures of Statistics: A Biometrical Approach (2nd Ed.) McGraw-Hill Publishing Co., New York, 181±183. Stein, T.E., Duffy, S.J., Wickstrom, S., 1990. Differences in production values between high- and lowproductivity swine breeding herds. J. Anim. Sci. 68, 3972±3979. Svajgr, A.L., Hays, V.W., Cromwell, G.L., 1974. Effect of lactation duration on reproductive performance of sows. J. Anim. Sci. 38, 100±105. Varley, M.A., Cole, D.J.A., 1976. Studies in sow reproduction 6. The effect of lactation length on preimplantation losses. Anim. Prod. 22, 209±214. Walker, N., Watt, D., MacLeod, A.S., Johnson, C.L., Boaz, T.G., Calder, A.F.C., 1979. The effect of weaning at 10, 25 or 40 days on the reproductive performance of sows from the first to the fifth parity. J. Agri. Sci. 92, 449±456. Warnick, A.C., Wallace, H.D., Palmer, A.F., Sosa, E., Duerre, D.J., Caldwell, V.E., 1965. Effect of temperature on early embryo survival in gilts. J. Anim. Sci. 24, 89±92. Wilson, M.R., Dewey, C.E., 1993. The association between weaning-to-mating interval and sow productivity. Swine Health Prod. 1, 10±15. Xue, J.L., Dial, G.D., Marsh, W.E., Davies, P.R., Momont, H.W., 1993. Influence of lactation length on sow productivity. Livest. Prod. Sci. 34, 253±265. Xue, J.L., Dial, G.D., Marsh, W.E., Davies, P.R., 1994. Multiple manifestations of season on reproductive performance of commercial swine. J. Am. Vet. Med. Asso. 204, 1486±1489. Zain, A.E.D., Nakao, T., Abdel Raouf, M., Moriyoshi, M., Kawata, K., Moritsu, Y., 1995. Factors in the resumption of ovarian activity and uterine involution in postpartum dairy cows. Anim. Reprod. Sci. 38, 203±214.