Use of Enterobacteriaceae analysis results for predicting absence of Salmonella serovars on beef carcasses

Previous work using a large data set (no. 1, n = 5355) of carcass sponge samples from three large-volume beef abattoirs highlighted the potential use of binary (present or absent) Enterobacteriaceae results for predicting the absence of Salmonella on carcasses. Specifically, the absence of Enterobacteriaceae was associated with the absence of Salmonella. We tested the accuracy of this predictive approach by using another large data set (no. 2, n = 2,163 carcasses sampled before or after interventions) from the same three data set no. 1 abattoirs over a later 7-month period. Similarly, the predictive approach was tested on smaller subsets from data set no. 2 (n = 1,087, and n = 405) and on a much smaller data set (no. 3, n = 100 postintervention carcasses) collected at a small-volume abattoir over 4 months. Of Enterobacteriaceae-negative data set no. 2 carcasses, > 98% were Salmonella negative. Similarly accurate predictions were obtained in the two data subsets obtained from data set no. 2 and in data set no. 3. Of final postintervention carcass samples in data set nos. 2 and 3, 9 and 70%, respectively, were Enterobacteriaceae positive; mean Enterobacteriaceae values for the two data sets were -0.375, and 0.169 log CFU/100 cm2 (detection limit = -0.204, and Enterobacteriaceae negative assigned a value of -0.505 log CFU/100 cm2). Salmonella contamination rates for final postintervention beef carcasses in data set nos. 2 and 3 were 1.1 and 7.0%, respectively. Binary Enterobacteriaceae results may be useful in evaluating beef abattoir hygiene and intervention treatment efficacy.