Epidemiology of Salmonella sp. in California cull dairy cattle: prevalence of fecal shedding and diagnostic accuracy of pooled enriched broth culture of fecal samples

BACKGROUND The primary objective of this cross-sectional study was to estimate the crude, seasonal and cull-reason stratified prevalence of Salmonella fecal shedding in cull dairy cattle on seven California dairies. A secondary objective was to estimate and compare the relative sensitivity (Se) and specificity (Sp) for pools of 5 and 10 enriched broth cultures of fecal samples for Salmonella sp. detection. METHODS Seven dairy farms located in the San Joaquin Valley of California were identified and enrolled in the study as a convenience sample. Cull cows were identified for fecal sampling once during each season between 2014 and 2015, specifically during spring, summer, fall, and winter, and 10 cows were randomly selected for fecal sampling at the day of their sale. In addition, study personnel completed a survey based on responses of the herd manager to questions related to the previous four month's herd management. Fecal samples were frozen until testing for Salmonella. After overnight enrichment in liquid broth, pools of enrichment broth (EBP) were created for 5 and 10 samples. All individual and pooled broths were cultured on selective media with putative Salmonella colonies confirmed by biochemical testing before being serogrouped and serotyped. RESULTS A total of 249 cull cows were enrolled into the study and their fecal samples tested for Salmonella. The survey-weighted period prevalence of fecal shedding of all Salmonella sp. in the cull cow samples across all study herds and the entire study period was 3.42% (N = 249; SE 1.07). The within herd prevalence of Salmonella shed in feces did not differ over the four study seasons (P = 0.074). The Se of culture of EBP of five samples was 62.5% (SE = 17.12), which was not statistically different from the Se of culture of EBP of 10 (37.5%, SE = 17.12, P = 0.48). The Sp of culture of EBP of five samples was 95.24% (SE = 3.29) and for pools of 10 samples was 100.00% (SE = 0). There was no statistical difference between the culture relative specificities of EBP of 5 and 10 (P > 0.99). DISCUSSION Our study showed a numerically higher prevalence of Salmonella shedding in the summer, although the results were not significant, most likely due to a lack of power from the small sample size. A higher prevalence in summer months may be related to heat stress. To detect Salmonella, investigators may expect a 62.5% sensitivity for culture of EBP of five, relative to individual fecal sample enrichment and culture. In contrast, culture of EBP of 10 samples resulted in a numerically lower Se. Culture of EBP of size 5 or 10 samples, given similar prevalence and limit of detection, can be expected to yield specificities of 95 and 100%, respectively.