Epidemiological investigations in regard to porcine reproductive and respiratory syndrome (PRRS) in Quebec, Canada. Part 2: prevalence and risk factors in breeding sites.

Porcine reproductive and respiratory syndrome virus (PRRSV) is a major threat for swine industry and understanding factors involved in its epidemiology is undoubtedly essential for disease control. As a part of a larger project, a cross-sectional study was performed on breeding sites in a moderate density area of swine production in Quebec to estimate the prevalence of PRRSV infected sites and to evaluate if characteristics of sites and biosecurity practices, either as specific measures or as a global score, were associated with PRRSV status. A questionnaire and diagnostic procedures were performed on 54 breeding sites between September 2006 and August 2008. A biosecurity score that had been previously computed using two-step clustering procedure was used, classifying breeding sites into two biosecurity patterns (high vs. low) according to 21 specific biosecurity measures. The apparent prevalence of PRRSV infected sites was 74.0% (95% CI, 60.3-85.0). Univariable and multivariable logistic regression models with robust standard errors adjusting for potential clustering of sites due to same ownership were computed. In a first multivariable model evaluating characteristics of sites and specific biosecurity variables, four main effects were significantly associated (P<0.05) with PRRSV positive status: large pig inventory (OR: 10.7), proximity to closest pig site (OR: 7.3), absence of shower (OR: 8.7) and free access to the main entrance of the site by the rendering truck (OR: 7.0). In a second multivariable model including a global biosecurity score as a surrogate for a specific pattern of biosecurity measures, this score was not retained in the final model. The adjusted population attributable fractions were 16% for the proximity to closest pig site variable, 27% for the absence of shower variable, and 10% for the free access to main entrance of the site by the rendering truck. These two latter biosecurity measures, manageable directly on the site, should be prioritized and be part of any intervention strategy designed for PRRSV control.