Transmission models of ESBL-producing Escherichia coli in Dutch broiler production chain
MetadataShow full item record
Introduction: Extended-spectrum beta-lactamase (ESBL)-producing Escherichia coli in animals, especially in chickens, are considered a risk to human health, because this type of bacteria can be a continuous threat of the emergence of new bacteria that are resistant to antimicrobials. Antimicrobials are a kind of medicine that is widely used to treat human diseases such as penicillins, which are used to treat throat infections, meningitis, and syphilis. Disease caused by resistant bacteria cannot be cured with antimicrobials. Although the proportion of chickens with ESBL-producing E. coli drastically declined over the years in Dutch farms, chicken meat was still the most contaminated meat product among others. In addition, the bacteria can be found in all stages of the chicken production chain. The chain consists of mainly three different farm types: parent stock farms where chickens produce eggs, hatcheries where eggs are incubated, and broiler farms where chickens are fattened up for meat. Therefore, further control of ESBL-producing E. coli in the chicken production chain is important to reduce public health risks. Objective: The main objectives of this study were to evaluate the effectiveness of intervention scenarios to control the transmission of ESBL-producing E. coli in the chicken production chain and to estimate the risk to public health. Methods: In this study, we developed two different types of transmission models to describe the observed transmission dynamics within a chicken farm: one based on the immune development of chickens and one based on the infection characteristics of the bacteria. Both included the environmental contamination effect between production rounds and within flocks. The parameter values, which determine the spread of bacteria, were estimated by using the Approximate Bayesian computation (ABC) method. This method is a way to find the most appropriate parameter value by comparing the observed data with the simulated data. Then, we applied the models to the three stages in the chicken production chain and further added the effect of mixing eggs and chicks from different farms. The size of a flock and the number of farms were adjusted to the Dutch situation. Several intervention scenarios, including bird vaccination and farm disinfection, were applied to the models. Finally, using the simulated data, risk to human health was estimated. Results: Two models were developed based on two different assumptions, which were development of immunity of the chickens and difference between the infection characteristics of the bacteria. Both models were able to capture the observed transmission dynamics. The proportion of infected chickens on the day of slaughter was 10.59% and 13.56%, respectively. The proportion of the Dutch population becoming infected by consuming chicken meat was estimated at 0.14% and 0.18%, respectively. If the bacteria in the farm environment were eliminated by cleaning and disinfection, the infected proportion of chickens would be reduced to 0.61% and 0.52%, respectively, and that of humans to 0.01%. Conclusions: Both models were able to describe the observed transmission dynamics within and between the production stages equally well and estimate the outcome of the interventions quantitatively. Both indicated that improving farm management to eliminate the bacteria from the farm environment was the most effective intervention. According to our models, chicken meat was not a major source for transmitting the bacteria to humans.