J. Anim Sci. 2007. 85:3110-3115. doi:10.2527/jas.2007-0104
© 2007 American Society of Animal Science
Technical note: High-performing swine herds improved their reproductive performance differently from ordinary herds for five years1
Y. Koketsu2
School of Agriculture, Meiji University, Kawasaki 214-8571, Japan
 |
Abstract
|
|---|
The objectives of this study were to determine changes in herd productivity and the performance of female pigs over time in commercial swine herds. Annual measurement data from 1999 to 2003 were obtained from the record files of 113 herds in Japan. Two groups were formed according to the 25th percentile of pigs weaned/mated females per year (PWMFY) in 2003; the 2 groups were high-performing herds (those constituting the top 25%) and the remaining ordinary herds. The effects of group based on PWMFY in 2003, year, and the group x year interaction on repeated measures between 1999 and 2003 were analyzed by using mixed-effects models. A regression analysis was also used to compare key measurements in productivity between the 2 groups, with years as a continuous variable. Variance components were obtained to determine herd repeatability of PWMFY for the 2 herd groups. The average female inventory increased from 290 ± 31 to 355 ± 42 females for these 5 yr. The PWMFY also changed from 20.9 ± 0.21 to 21.2 ± 0.30 pigs. An interaction between year and group was detected (P < 0.05) for PWMFY. In the regression comparison, high-performing herds increased their PWMFY by 0.31 ± 0.09 pigs each year, whereas the ordinary herds did not increase. The number of pigs weaned per sow increased by 0.07 ± 0.02 pigs each year in high-performing herds and increased by 0.03 ± 0.01 pigs each year in ordinary herds. In high-performing herds, for each year, the percentage of sows mated by 7 d after weaning increased by 0.92 ± 0.25%, the percentage of reserviced females decreased by 0.63 ± 0.26%, and culling rate increased by 1.53 ± 0.50%. Repeatability of PWMFY for high-performing herds and ordinary herds was 28.8 and 54.0%, respectively. This study shows that productivity in high-performing herds was improved compared with that of ordinary herds.
Key Words: commercial farm • herd repeatability • management • pig • repeated measure
|
INTRODUCTION
|
|---|
Annual measurements of herd productivity and the performance of female pigs in data collected from producers in North America and other countries have been studied to provide benchmarks of productivity and performance and to enhance our knowledge of production systems for the swine industry (Wilson et al., 1986
; Stein et al., 1990; Koketsu, 2000). However, comparisons among herds for a single year do not necessarily describe the changes in productivity and performance in herds (McGilliard et al., 1990). Previous studies using dairy data have suggested that production measurements vary considerably over years within a herd, whereas herd productivity is not random, but is repeatable to a certain degree (McGilliard et al., 1990; Rougoor et al., 1999). An analysis of serial data for a relatively long period can provide a view of changes in the reproductive performance of females and in production management and production systems, but these have not been reported in commercial swine herds.
For best-practices benchmarking, the characteristics of highly efficient herds have been described by using annual productivity (Stein et al., 1990; Koketsu, 2000). However, little research has been conducted on how the productivity of high-performing or ordinary herds changes and how performance measurements vary in swine over time.
A high repeatability in financial and production measurements was reported in dairy herds (McGilliard et al., 1990; Rougoor et al., 1999). No research has reported on how repeatable herd measurements are in swine, although repeatability values for individual sows have been reported for reproductive performance, such as the number of pigs born alive and the weaning-to-service intervals (Le Cozier et al., 1997).
The objectives of this study were to observe changes in herd productivity and performance measurements, to quantify how high-performing herds improved, and to determine the repeatability of those measurements in commercial swine herds by using 5-yr repeated measures data.
|
MATERIALS AND METHODS
|
|---|
Data
Animal Care and Use Committee approval was not obtained for this study because the data were obtained from the existing PigCHAMP database (Meiji University, Kawasaki, Japan). The database was created in the following manner. Each year, all producers using PigCHAMP were requested to mail their data files to the School of Agriculture, Meiji University, every time they renewed the yearly maintenance contract. In 2004 the database included approximately 1.7% of all Japanese herds, with approximately 4% of female inventories in Japan; the country had 7,770 breeding herds and 917,000 females in February of 2004 (Ministry of Agriculture, Forestry and Fisheries of Japan, 2005).
Data Selection
By August 31, 2004, 134 record files had been mailed to the university. Of the 134 herds, 5 herds were grow-finish operations and were removed. Herds that were studied were required to have no missing data for 5 yr to allow a balanced data set for year-to-year comparisons. A period of 5 yr was arbitrarily chosen to keep at least 110 herds (out of the 129) with 5 consecutive years of data. Of the 129 breeding herds, 16 had no complete reproductive data for the 5 yr from 1999 to 2003 and were removed; 113 (87.6%) were selected for further analysis.
Definitions and Calculations
Females included maiden gilts, mated gilts, and sows, whereas mated females included mated gilts and sows (PigCHAMP, 1996). The average mated female inventory as a measurement of annual pig year was calculated as the total days that mated females were fed in a herd (pig days) during a 1-yr period divided by 365.
Culled females included females shipped to a slaughterhouse or euthanized in a barn. Culling rate was defined as the number of culled females multiplied by 100 and then divided by the average mated female inventory. Farrowing percentage was calculated as the percentage of the number of farrowed sows divided by the number of mated gilts and sows. Preweaning mortality risk was calculated as the percentage of the number of pigs that died before weaning divided by the number of pigs born alive in farrowed and weaned litters.
Herd Productivity and Performance Measurements
The best indicator of herd productivity is the number of pigs weaned/mated female per year (PWMFY), which is the product of the number of litters/mated female per year and the number of pigs weaned per sow (Dial et al., 1992). The number of litters/mated female per year is improved by decreasing the nonproductive female days (NPD), whereas the number of pigs weaned can be improved by decreasing preweaning mortality and increasing the numbers of pigs born alive per sow (Dial et al., 1992). The NPD was defined as the average number of days when mated females were neither gestating nor lactating (Koketsu, 2005). The body of measurements was chosen from a previous report (King et al., 1998), including lactation length, parity measurements, and culling rate. The percentage of sows mated by 7 d after weaning, weaning-to-first-service interval, and percentage of reserviced females were also added because these were related to breeding management, such as estrous detection, mating, and pregnancy check in breeding herds (Koketsu, 2000).
Herd Category
Herds were ranked and categorized into 2 groups according to PWMFY in 2003. For comparison, herds in the upper 25th percentile of this ranking and the remaining herds were then designated as high-performing herds and reference (ordinary) herds, respectively.
Statistical Analysis
The observational unit was the herd for 1 yr. All analyses were performed with SAS (SAS Inst. Inc., Cary, NC). Repeated measures data between 1999 and 2003 were analyzed by using mixed-effects models with REPEATED and RANDOM statements and a first-order autoregressive option (Littell et al., 2006). First, the effects of the group based on PWMFY in 2003, the year, and the group x year interaction for each measurement were analyzed. Estimate statements were used to compare the means of the herd groups each year and to compare the means of the year subgroups. A method of comparison using regression, with years as a continuous variable, was used in the repeated measures to quantify a year trend (Littell et al., 2002) whenever a year effect or a group x year interaction was found (P < 0.05). A power calculation for the linear regression was also used (PS software, Vanderbilt University, 2004).
Repeatability for Herd Measurements
Herd repeatability was defined as the correlation between repeated measures in the same herd. Herd repeatability (rep) of variables within a herd over time was estimated from herd (Vh) and residual (Vr) components of the variance by using the following formula: rep = Vh/(Vh + Vr). The values of the variance components were obtained by using the NESTED procedure, with herd as a class (Le Cozier et al., 1997). Residual was the residual variance, expressed as a percentage of the total variance.
|
RESULTS
|
|---|
The comparison between the high-performing and ordinary herds for herd productivity and the performance of females from 1999 to 2003 is shown in Tables 1 and 2. The average female inventory increased from 290 ± 31 to 355 ± 42 females for these 5 yr. The number of pigs weaned/mated female per year changed from 20.9 ± 0.21 to 21.2 ± 0.30 pigs. Year effects were found on the number of pigs weaned per sow, preweaning mortality risk, percentage of sows mated by 7 d, and culling rate (P < 0.05). Group effects were found on PWMFY, number of pigs weaned per sow, preweaning mortality risk, NPD, farrowing percentage, percentage of reserviced females, and weaning-to-first mating interval (P < 0.05). An interaction between the year effect and the group effect was found with PWMFY (P = 0.04).
Regression comparisons in measurements using years as a continuous variable are presented in Table 3. No quadratic expression of the year was found in any measurement. The high-performing herds increased PWMFY by 0.31 ± 0.09 pigs each year, whereas ordinary herds did not increase. The number of pigs weaned per sow increased by 0.07 ± 0.02 pigs each year in high-performing herds, and increased by 0.03 ± 0.01 pigs each year in ordinary herds for the 5 yr. Preweaning mortality risk decreased by 0.26 ± 0.16% in high-performing herds, and decreased 0.22 ± 0.10% each year in ordinary herds. The slope of preweaning mortality risk was not different from 0 in the 31 high-performing herds (P = 0.11, power = 0.26).
In the high-performing herds for each year, the percentage of sows mated by 7 d after weaning increased by 0.92 ± 0.25%, the percentage of reserviced females decreased by 0.63 ± 0.16%, and the culling rate increased by 1.53 ± 0.50%. In the ordinary herds, the culling rate decreased by 0.99 ± 0.31% each year, whereas the percentage of reserviced females did not improve.
The herd repeatability values for PWMFY of high-performing herds and ordinary herds were 28.8 and 54.0, respectively (Table 4). Additionally, 66.6% of 31 herds were high-performing herds (those in the top 25%) in the first year and the fifth year.
View this table:
[in this window]
[in a new window]
|
Table 4. Comparisons in the repeatability (rep) of herd performances for 5 consecutive years between 2 herd groups
|
|
|
DISCUSSION
|
|---|
This study showed that high-performing herds continuously improved their herd productivity, measured as PWMFY, but that ordinary herds did not. High-performing herds in this study improved PWMFY by 0.31 ± 0.09 pigs each year. This improvement may be explained by the increased numbers of pigs weaned per sow, increasing percentages of sows mated by 7 d, and decreasing percentages of reserviced females each year in the high-performing herds. Both herd groups steadily improved the number of pigs weaned for these 5 yr by decreasing preweaning mortality risk, although the number of pigs born alive did not increase. As the number of pigs weaned per sow increased by 1 pig, PWMFY increased by 2.17 pigs (Koketsu, 2005). However, ordinary herds did not increase the number of pigs weaned per sow or PWMFY as the high-performing herds did.
Decreasing percentages of reserviced females and increasing culling rates implied imposing a strict culling policy in the high-performing herds. The increasing percentages of sows mated by 7 d postweaning may be also attributed to an accurate heat check. Both the percentages of sows mated by 7 d and the percentage of reserviced females were related to herd fertility and PWMFY (Dial et al., 1992; Koketsu, 2007). However, ordinary herds could not improve either of these percentages or PWMFY as much as high-performing herds. To improve herd fertility and PWMFY, it has been recommended that an accurate heat check using a boar be performed, that early pregnancy be detected by real-time ultrasonography, and that an appropriate culling policy be implemented (Dial et al., 1992; Miller et al., 2003).
The repeatability for herd fertility, a measurement of within-herd variability across time (e.g., average days open, services per conception), was reported to be between 0.59 and 0.72 in dairy herds (McGilliard et al., 1990). The repeatability for fertility in the ordinary swine herds was similar to that reported previously in dairy herds. The repeatability (28.8) in the high-performing herds was relatively lower for PWMFY than in the ordinary herds. The high-performing herds changed or improved every year, whereas the ordinary herds did not improve greatly. It may be difficult for a herd to be high performing over a long period.
In conclusion, productivity in high-performing herds was improved, at least in part, through breeding and lactation management, compared with ordinary herds. Annual benchmarking, which objectively measures the herd productivity and management system, is also recommended to improve herd productivity. Finally, it is noteworthy that this study is an observational study using records from commercial herds. Results could be biased by health status, genetics, and management changes, which were not measured. However, even with such limitations, this study provides valuable information for swine producers, extension specialists, and veterinarians on herd measurements over a number of years.
|
Footnotes
|
|---|
1 Appreciation is expressed to the staff of Global Pig Farms Inc. (Setagun, Gunma, Japan) for their technical assistance and cooperative producers to the commercial swine producers for providing their records. This research was supported by a Research Project Grant from Ministry of Education, Culture, Sports, Science and Technology, High-Tech Research Center Project (2006 to 2008). 
2 Corresponding author: koket001{at}isc.meiji.ac.jp
Received for publication February 16, 2007.
Accepted for publication June 13, 2007.
|
LITERATURE CITED
|
|---|
Dial, G. D., W. E. Marsh, D. D. Polson, and J. P. Vaillancourt. 1992. Reproductive failure: Differential diagnosis. Pages 88–137 in Diseases of Swine. 7th ed. A. L. Leman, B. E. Straw, W. L. Mengeling, S. D’Allaire, and D. J. Taylor, ed. Iowa State Univ. Press, Ames.
King, V. L., Y. Koketsu, D. Reeves, J. L. Xue, and G. D. Dial. 1998. Management factors associated with swine breeding-herd productivity in the United States. Prev. Vet. Med. 35:255–264.[CrossRef][Medline]
Koketsu, Y. 2000. Productivity characteristics of high-performing commercial swine breeding farms. J. Am. Vet. Med. Assoc. 216:376–379.[CrossRef][Medline]
Koketsu, Y. 2005. Six component intervals of nonproductive days in breeding-female pigs on commercial farms in Japan. J. Anim. Sci. 83:1406–1412.[Abstract/Free Full Text]
Koketsu, Y. 2007. Longevity and efficiency associated with age structures of female pigs and herd management in commercial breeding herds. J. Anim. Sci. 85:1086–1091.[Abstract/Free Full Text]
Le Cozier, Y., J. Dagorn, J. Y. Dourmad, S. Johansen, and A. Aumaître. 1997. Effect of weaning-to-conception interval and lactation length on subsequent litter size in sows. Livest. Prod. Sci. 51:1–11.[CrossRef]
Littell, R. C., G. A. Milliken, W. W. Stroup, R. D. Wolfinger, and O. Schabenberger. 2006. SAS System for Mixed Models. SAS Inst. Inc., Cary, NC.
Littell, R. C., W. W. Stroup, and R. J. Freud. 2002. SAS System for Linear Models. SAS Inst. Inc., Cary, NC.
McGilliard, M. L., V. J. Conklin, R. E. James, D. M. Kohl, and G. A. Benson. 1990. Variation in herd financial and production variables over time. J. Dairy Sci. 73:1525–1532.[Abstract]
Miller, G. M., S. M. Breen, S. L. Roth, K. L. Willenburg, S. Rodriguez-Zas, and R. V. Knox. 2003. Characterization of image and labor requirements for positive pregnancy diagnosis in swine using two methods of real-time ultrasound. Swine Health Prod. 11:233–239.
Ministry of Agriculture, Forestry and Fisheries of Japan. 2005. Available: http://www.maff.go.jp/toukei Accessed Dec. 24, 2006.
PigCHAMP. 1996. PigCHAMP Reports Manual. University of Minnesota, St. Paul.
Rougoor, C. W., R. B. M. Huirne, A. A. Dijkhuizen, and F. Mandersloot. 1999. Repeatability in ranking of dairy farms on technical and economic performance over years. Prev. Vet. Med. 38:25–34.[CrossRef][Medline]
Stein, T. E., S. J. Duffy, and S. Wickstrom. 1990. Differences in production values between high- and low-productivity swine breeding herds. J. Anim. Sci. 68:3972–3979.[Abstract]
Vanderbilt University. 2004. PS: Power and Sample Size Calculation. Vanderbilt University, North Nashville, TN. Available: http://biostat.mc.vanderbilt.edu/twiki/bin/view/Main/PowerSampleSize Accessed Jan. 24, 2007.
Wilson, M. R., R. M. Friendship, I. McMillan, R. R. Hacker, R. Piper, and S. Swaminathan. 1986. A survey of productivity and its component interrelationships in Canadian swine herds. J. Anim. Sci. 62:576–582.[Abstract/Free Full Text]
This article has been cited by other articles:
|
|
|
|
 
Y. Sasaki and Y. Koketsu
Mortality, death interval, survivals, and herd factors for death in gilts and sows in commercial breeding herds
J Anim Sci,
November 1, 2008;
86(11):
3159 - 3165.
[Abstract]
[Full Text]
[PDF]
|
|
|
Top
Abstract
INTRODUCTION
