Pre-slaughter control of beef carcass contamination with Escherichia coli O157: A risk assessment approach

This thesis is an investigation of the extent to which factors in the pre-slaughter period of cattle production can be manipulated to reduce the number of cattle carcasses contaminated with Escherichia coli O157 and the quantity of E. coli O157 deposited on carcasses. The principal objective was to construct a simulation model representing the pre-slaughter phase of beef production and to use it to predict the relative advantages of various mitigation strategies against a pathogen excreted in bovine faeces. Because of uncertainty in key input variables, Monte Carlo simulation was used to generate model outcomes as probability distributions. Control measures based on testing of sale-lots or herds of origin for E. coli O157 and measures based on the amount of soiling of hides (tag) were simulated. In order to quantify the effect of fasting a field trial was performed where the rate of change in the concentration of E. coli biotype 1 in the faeces of fasted feedlot-cattle was found to be 0.024 log CFU/hour. A second simulation model was derived and used to provide estimates of herd-level sensitivity and specificity based on culture of faeces for E. coli O157. The herd-testing protocol was shown to have a poor ability to discriminate between infected and non-infected herds. Data were also obtained on the extent of tag accumulation on the hides of Ontario feedlot cattle presented for slaughter. The subjective rating system used to assess tag was found to have a statistical reliability of 0.84. Information from the literature plus that from the fasting, herd-testing, and tag score studies were combined with the simulation model to predict the performance of an abattoir with respect to E. coli O157 contamination of carcasses. The results highlight the ubiquitous nature of carcass contamination with E. coli O157, albeit at low levels. Results also suggest that a highly effective vaccine for use in cattle could greatly reduce the amount of contamination but the impact on human health outcomes would be dependent on extent of cross-contamination and opportunity for growth of the organism in beef product. Rapid tests applied to sale-lots of cattle departing the feedlot showed some promise while herd tests based on a limited sample size appear unattractive for reducing contamination.