An epidemiological investigation of the early phase of the porcine epidemic diarrhea (PED) outbreak in Canadian swine herds in 2014: A case-control study

A questionnaire was administered to participating producers from the 22 Canadian swine herds enrolled (n = 9 case and n = 13 control herds). Case herd producers were asked to provide information from the initial day of onset of clinical signs and 30 days prior to that day. Control herds were matched to a case herd on the basis of province, herd type and approximate size. The period of interest for a control herd was matched to the initial day of clinical signs of PED for the case herd, along with the 30 days prior to this day. The questionnaire questions focused on herd demographics, biosecurity protocols, live animal movements onto and off sites, deadstock movements, feed and people movements for both the case and control herds. The questionnaire for control herds were based on their matched case's period of interest, and together with case herds formed a matched stratum. Multivariable exact conditional logistic regression and mixed multivariable logistic regression models, with the matched stratum as a random effect, were used to assess the association between various risk factors and the odds of PED introduction into a herd. After adjusting for biosecurity practices, the odds of a PED occurrence was 38.1 (95% CI: 2.7-531.3) times greater for herds receiving feed from a single feed company that provided potentially contaminated feed (P = 0.007) than herds that did not. The number of live pigs delivered onto sites, semen deliveries and the frequency of deadstock pickups were not associated with PED status during the initial phase of the outbreak in univariable analyses. This study supports the role of potentially contaminated feed from a single feed company as a significant risk factor for PED viral introduction during the early phase of the Canadian outbreak.