Factors associated with farm-level infection of porcine epidemic diarrhea during the early phase of the epidemic in Japan in 2013 and 2014.

The objective of this study was to investigate factors that caused rapid spread during the early phase of the porcine epidemic diarrhea (PED) epidemic in Japan in 2013 and 2014. Anonymized datasets from all pig farms were provided by Kagoshima (709 farms) and Miyazaki Prefectures (506 farms). Semi-parametric survival analysis was conducted using the first 180 days from the first case on December 3, 2013 in Kagoshima Prefecture. To compare the hazard between different farm management types, univariable survival analysis was conducted. As farm sizes varied among different farm types, bivariable survival analysis was conducted for farm size categories and farm density per km2 for each management type. A case-control study using a postal questionnaire survey was conducted in September 2014, and risk factor analysis was performed using generalized linear models with binomial errors. The hazard was significantly higher in farrow-to-finish farms than fattening farms [hazard ratio (HR) = 1.6, p < 0.01], but was not significantly different between reproduction and fattening farms (HR = 1.3, p = 0.16). In separate bivariable survival analyses for each farm type, large- and middle-scale farms had higher hazard than small-scale farms in fattening (HR = 5.8 and 2.6, respectively, both p < 0.01) and reproduction farms (HR = 4.0 and 3.6, respectively, both p < 0.01). In farrow-to-finish farms, large-scale farms had higher hazard than small-scale farms (HR = 2.8, p < 0.01), and higher farm density per km2 was also a risk factor (HR = 7.6, p < 0.01). In the case-control study, questionnaires were returned from 78 PED virus-infected and 91 non-infected farms. The overall response rate was 34%. Risk factors of the final model were occurrence of porcine reproductive and respiratory syndrome in the past 5 years [odds ratio (OR) = 1.97, 95% confidence interval (CI): 0.97-4.00, p = 0.054], use of a common compost station (OR = 2.51, 95%CI: 1.08-5.83, p = 0.03), and use of a pig excrement disposal service (OR = 2.64, 95%CI: 1.05-6.63, p = 0.04). High hazard in farrow-to-finish farms suggested transmission from slaughterhouses to susceptible suckling piglets. Hazard associated with large-scale farms and high density might be due to frequent vehicle entrance and transmission by roads. Improvement of farm hygiene management and avoidance of risky practices associated with contact with pig excrement were keys in preventing invasion of PED virus to a farm.