Poor sow longevity has both economic and welfare ramifications for the commercial swine industry. According to PigCHAMP™ reports, between years 2000 and 2010, the average culling frequency of breeding herd females in U.S. commercial swine herds has been 45.9% and sow mortality rate has risen to 8%. In some cases, individual herds have experienced culling and mortality rates above 50% and 15%, respectively. Since reproductive failure and leg problems (lameness and leg soundness) are the primary culling reasons for young sows, maintaining acceptable reproduction rates in younger females and selecting structurally sound replacement gilts are important factors in increasing sow productive lifetime. Lower replacement rates would not only improve the outlook of the swine industry, but also increase the profitability of pork producers in terms of reduced replacement costs. Furthermore, reduced number of gilt litters would improve herd productivity, as gilt litters tend to have lower number born and number born alive, and their offspring experience greater mortality and poorer average daily gain throughout the nursery and grow – finish phases of production.

 The objective of this study was to estimate the phenotypic and genetic associations of gilt compositional and structural soundness traits with reproductive and longevity traits (longevity defined as the ability to complete the fourth and the fifth parity), in order to determine factors measured or evaluated early in a sow’s life that are associated with superior sow productive lifetime.

 The study involved in total 1447 commercial females from two genetic lines. Gilts were on average 190 days of age and 124 kg body weight at the time of body composition and structural soundness evaluation. Evaluated compositional traits included body weight, loin muscle area, 10th rib backfat and last rib backfat. Structural soundness traits consisted of six body structure traits (body length, depth and width, rib shape, top line and hip structure), five leg structure traits per leg pair (front legs: legs turned, buck knees, pastern posture, foot size and uneven toes; rear legs: legs turned, weak/upright legs, pastern posture, foot size and uneven toes) and overall leg action. Two scorers independently evaluated all structural soundness traits on a nine-point scale. Studied lifetime reproduction traits included lifetime total number born (LNB), lifetime number born alive (LBA), number born alive per lifetime days (LBA/L) and percentage productive days from total herd days (PD%). Lifetime (L), herd days (HD) and removal parity (RP) were considered as longevity traits. The genetic relationships (genetic correlations) of herd days with other trait groups were not analyzed.

 The degree in which a trait is controlled by the genetics (heritability) and the genetic relationship between traits (genetic correlation) were estimated using two different genetic software programs. One of these software packages was capable of taking into account that not all animals had been removed by the end of data collection, thus their records were incomplete or what is commonly referred to as right censored. The model used to obtain the genetic parameter estimates for compositional traits included genetic line of the gilt (two genetic lines) and evaluation day (to account for differences in the 14 groups of gilts delivered to the farm) as fixed effects. The model used animal as a random effect, since each animal has a random sample of genes from each of its parents contributing to its genetic make up. Furthermore, standard formulas were applied to adjust 10th rib backfat, loin muscle area and the number of days to a constant body weight of 113.5 kg (NPPC, 2000). In the absence of an adjustment formula, last rib backfat measurements were adjusted during the genetic analyses to the constant average weight at evaluation (124 kg). Structural soundness traits were analyzed with an identical model to last rib backfat, except scorer was included as an additional fixed effect (to account for differences between two scorers). The model for longevity and lifetime reproduction traits included genetic line and herd entry group (contemporary group) as fixed and animal as a random effect.

 About 70% of the females were removed prior to the sixth parity. At the termination of data collection, 14% of the females were still alive and in production at the commercial sow herd. The parity distribution of the sows in the herd at the end of data collection ranged from the sixth to the ninth. Reproductive failure was the most frequent culling reason during the first three parities and it caused the loss of 16% of the research females before the fourth parity. Lameness or feet and leg problems were most distinctive prior to the third parity causing the removal of 7.5% of the young females. Litter performance became the most important culling reason by the fourth parity. The median survival times (time by which 50% of the females had been removed) were 546 herd days or 723 days of age, corresponding a mean removal parity of 3.7.
 By the fourth farrowing, the females averaged 818 days age and had been in the herd for 638 days. On average the fourth parity sows had 12.5 piglets born in total and 90% of the piglets were born alive. Number of piglets weaned was 9.4. The fifth farrowing occurred on average at the age of 967 days when the females had been for 787 days in the herd. Total number born was 11.1 piglets with 84% being born alive and number weaned was 8.4 piglets. There was no statistically significant difference between the lines regarding age of the females, cumulative non-productive days or litter size in the fourth or the fifth parity. A significant difference in body composition traits was observed between these maternal lines already upon the entry to the farm, and the same trend was apparent at every farrowing and weaning, parent line sows having thicker backfat and smaller loin muscle area (P ≤ 0.05).

 At the time of removal, females had farrowed on average 44.6 piglets of which 40.4 were born alive. On average, they had 0.039 live born piglets per a lifetime day and the percentage of productive days from total herd days was 61%. The lifetime reproductive performance of the parent line females, measured with the aforementioned traits, was significantly better in comparison to the performance of the grandparent line females (P ≤ 0.05).

 The heritability estimates obtained for longevity traits ranged from 0.12 to 0.16 and for lifetime reproduction traits from 0.13 to 0.17. Compositional traits had high heritabilities (0.50 – 0.70). For body structure traits (see the appendix for a visual description of evaluated traits), the heritability estimates were low to moderate (0.11 – 0.34). The majority of the heritability estimates for leg structure traits were relatively low (total range 0.07 – 0.29). The highest heritability estimates were obtained for weak front and rear pastern postures (0.28 and 0.29, respectively). The heritability of overall leg action was 0.12.
 There was a trend of greater backfat, larger loin muscle area and increased days to 113.5 kg body weight being genetically associated with improved longevity and lifetime. However, most of the genetic correlations were low and non-significant (P > 0.05). Larger loin muscle area was significantly associated with greater L, RP and LNB. On the other hand, increased days to 113.5 kg body weight was significantly associated with greater L, RP, LNB, LBA and PD%.

 As was the case with compositional traits, most of the genetic correlations of structural soundness traits with longevity measures and lifetime reproduction traits were low and non-significant (P > 0.05). In general, the associations of body structure traits with longevity measures and lifetime reproduction traits were favorable. Body length and rib shape had significant associations with all longevity measures and lifetime reproduction traits. Females with intermediate body length and more ideal rib shape remained in the herd longer and farrowed more piglets. Furthermore, greater body width was significantly correlated with greater L and RP. From the leg structure traits, slightly outwards turned front legs were significantly associated with greater LNB, LBA and LBA/L. Furthermore, less upright rear legs were associated with greater LBA/L and PD% and intermediate rear foot size with greater L and RP. Results implied that less upright pastern posture would be associated with improvements in longevity and lifetime reproduction traits. Overall leg action was unfavorably associated with longevity and lifetime reproduction, but these associations were non-significant.

 It is important to note, that the results obtained from this study need to be interpreted within the distributions of observations present in the dataset. The animals included into the study were preselected for their growth potential and structural soundness by the genetic supplier, and therefore the data set primarily consisted of well growing and relatively good structured animals.

 This study was conducted at a typical U.S. commercial farm offering the pork producers results that are obtained at a comparable environment to theirs. The removal reason frequencies across parities show that genetic improvements in both reproductive and structural soundness traits are needed to increase the profits of the producers. The genetic correlations obtained in this study indicate that in terms of improving sow lifetime reproductive performance and hence the profitability for pork producers, the most important gilt body composition, growth and structural soundness traits in commercial replacement gilt selection would be closer to intermediate growth rate and body length, more ideal rib shape (more of a barrel shape) and less upright rear legs.

Contact:  Dr. Kenneth J. Stalder, Professor, Iowa State University, Dept. of Animal Science, 109 Kildee Hall, Ames, IA 50011-3150, Phone: 515-294-4683, Email: stalder@iastate.edu