Cluster analysis as a tool to assess litter size in conjunction with the amount of embryonic and fetal losses in pigs

High numbers of live born and vital piglets are required for an effective piglet production. It has been reported that embryonic losses in swine can be as high as 20 to 50%. Generally, up to 30% of embryonic losses are considered within the normal biological range. The aim of our study was first to determine embryonic and fetal losses in sows of the German Landrace breed, and second to identify sows, which are able to realize high numbers of intact embryos and fetuses in combination with low rates of embryonic and fetal losses. This study was conducted on a commercial farrow-finish operation and involved 64 gilts. Gilts were synchronized for ovulation and inseminated artificially (AI) twice at fixed times. Pregnant gilts were slaughtered on Day 30 (n=34) and Day 80 (n=30) after second AI. Corpora lutea (CL) and embryos/fetuses (E/F) were counted. The length of uterine horns was measured. Based on the difference between the number of CL and the number of E/F, embryonic and fetal losses were 36.9 and 37.9%, respectively. The results obtained regarding the number of CL, the number E/F (intact and total), the rate of embryonic or fetal losses and uterine space per E/F underwent a cluster analysis. The results show that several gilts are able to realize high numbers of intact E/F in combination with low rates of losses and limited uterine space per E/F. Numbers of 15 to 20 intact fetuses on day 80 of pregnancy are considered as a minimum for an adequate litter size at the end of pregnancy. In this study only 35% of gilts at Day 30 and 37% of gilts at Day 80 demonstrate this potential. Cluster analysis is a useful mathematical tool which helps to assess results of embryonic and fetal losses and to group gilts according to their performance. As a consequence cluster analysis provides clues which gilts or family structures should be analyzed more intensively.