Czopowicz M, Kaba J, Szaluś-Jordanow O, Nowicki M, Witkowski L, Frymus T
Division of Infectious Diseases and Epidemiology, Department of Large Animal Diseases with the Clinic, Faculty of Veterinary Medicine, Warsaw University of Life Sciences - SGGW, Nowoursynowska 159c, 02-776 Warsaw, Poland.
Pol J Vet Sci. 2012;15(1):67-75. doi: 10.2478/v10181-011-0116-6.
The observational study was carried out in a population of Polish breeding goats in 2007 to determine the prevalence of fetal loss and identify risk factors contributing to its occurrence. The multivariate model allowing to predict the risk of the occurrence of fetal loss in a herd in a study population was developed. Data on the occurrence of fetal loss, as well as of 28 hypothesized risk factors were collected from goat owners using standardized questionnaire during face-to-face reviews on farms. Moreover, data on the herd-level seroprevalence of four abortifacient infections--Chlamydophila abortus, Leptospira spp., BVDV-1 and Neospora caninum--were included in the final analysis. Fetal loss was reported as occurring often in 12 of 49 goat herds (24.5%). The relationship between the hypothesized risk factors and the occurrence of fetal loss was verified in the multivariate logistic regression (alpha=0.05). Final analysis yielded four risk factors: regular veterinary supervision at least twice a year (OR 0.188; CI 95% 0.054 - 0.656), frequent occurrence of injuries and fractures (OR 3.172; CI 95% 1.081 - 9.310), frequent occurrence of respiratory signs in adult goats (OR 4.848; CI 95% 1.353 - 17.377) and presence of antibodies to C. abortus in a herd (OR 58.116; CI 95% 1.369 - 2466.438). The accuracy of the multivariate model was analyzed using receiver operating characteristic (ROC) curve technique. Area under the curve was 0.895 (CI 95% 0.801-0.981). For optimal cut-off value of 0.20-0.35 the multivariate model had sensitivity of 75.00% and specificity of 89.19% in predicting fetal loss in a herd.
