|RICCI, ALESSANDRO - University Of Torino|
|PASCOTTINI, OSVALDO - University Of Guelph|
Submitted to: Journal of Dairy Science
Publication Type: Abstract Only
Publication Acceptance Date: 3/3/2018
Publication Date: N/A
Technical Abstract: A review of existing literature was conducted to determine the prevalence of purulent vaginal discharge (PVD) in dairy herds around the world and detection methodologies that influence prevalence estimates. Four databases (PubMed, Google Scholar, Web of Science, and Scopus) were queried with the search terms “endometritis”, “PVD”, or “Metricheck” and “dairy.” The abstracts of initial search results were scanned to ensure the study used dairy cows and one of three methods of PVD diagnosis: use of a gloved hand, vaginoscopy, or Metricheck. This resulted in collection of 42 manuscripts, 36 of which reported sufficient information about their methodology and PVD prevalence to be included in the subsequent analysis. Reports were from 5 geographic regions: North America, South America, Europe, Asia, and the South Pacific and included 85 observations on PVD prevalence. The metafor package in R statistical software was used to fit fixed, random, and mixed effects models to estimate average PVD prevalence and factors influencing this prevalence using a logistic transformation of the original proportion values weighted by the number of animals in the cohort. The fixed effects model estimated a prevalence of 0.235 (95% CI of 0.231-0.239) while the random effects model estimated a slightly higher effect size of 0.247 (95% CI 0.210 – 0.288). The full mixed effects model included fixed effects for parity, calving strategy (seasonal or continuous), region, method of diagnosis, and day of diagnosis. Only the day of diagnosis had a significant (P < 0.05) effect on the prevalence of PVD with a decrease in PVD prevalence of 0.4% per day in lactation. Average PVD prevalence estimated by the mixed-effects model was 0.249 (95% CI 0.200 – 0.307). All models included a large degree of heterogeneity indicating factors not included in this analysis account for most of the variation in PVD; however, our results show PVD prevalence does not vary with region of the world, parity, or method of diagnosis.