SAEVA Proceedings 2014 | Page 151

46TH  ANNUAL  CONGRESS  OF  THE  SAEVA        SKUKUZA      16-­‐20  FEBRUARY  2014   151     been explained by both a decrease in the quality of the oocytes produced by older mares and to a lesser extent by deterioration in the endometrial environment. Other risk factors, such as mare reproductive status, have been more difficult to identify and show variable impact on fertility between different studies. Aims A retrospective analysis of reproductive events affecting pregnant broodmares during unrelated abortion epizootics from confirmed EHV-1 infection was conducted on two Thoroughbred stud farms in geographically separated areas of South Africa. We then investigated how multiple risk factors interact to ultimately impact on the subsequent reproductive outcomes via a prospective cohort study in these same broodmare populations. Materials and methods The study population comprised a total of 416 Thoroughbred mares resident on either of two farms that had recently suffered epizootic equine herpesvirus type-1 (EHV-1) abortion that affected 9/30 (30.0%) and 18/316 (5.7 %) of their pregnant mares, respectively. We first characterised the reproductive efficiency in this population of mares. Thereafter, and of particular interest, was the use of the resultant breeding data to develop logistic regression models to explain the relative influence and interactions of the different reproductive variables in the different broodmare populations in the subsequent breeding seasons and to predict the probability of a mare: a) establishing a pregnancy, and b) Foaling successfully the next season. This multivariate model considered the following reproductive variables: (i) Month within the season, (ii) Mare age, (iii) Mare reproductive status, (iv) Incidence of EED, and (v) Number of oestrous cycles mated to the stallion in that season. The data from both farms were pooled for analysis because ‘farm’ was neither a confounder nor effect modifier when examining main and interaction effects. Statistical significance was defined as P<0.05 Results and discussion (i) The last month of breeding predicted efficiency, pregnancy and foaling. Interestingly, breeding in the first month of the season was associated with an improved probability of pregnancy among barren mares. (ii) As reported in previous studies, increasing mare age was associated both with an increased number of oestrous cycles required to establish a pregnancy and an increased rate of EED. (iii) Mare reproductive status predicted breeding efficiency and the last month of breeding, but not establishing pregnancy. (iv) EED predicted all the reproductive outcomes, and was strongly predictive of multiple breeding attempts and pregnancy, but it is both interesting and difficult to explain its association with reducing the probability of foaling by nearly 90%.   151