Exposure of food-producing animals to antimicrobials: antimicrobial use versus internal concentrations: a modelling approach

Abstract

The aim of this thesis was to assess methods to investigate and describe exposure of food-producing animals to antimicrobials. High antimicrobial use for a period of many years (also due to high prophylactic use and use as growth promotors) has resulted in development of resistant and multi-resistant microorganisms in animal husbandry. Significant positive associations have been established between antimicrobial consumption and the probability for existence of antimicrobial resistant microorganisms in food-producing animals. Therefore, it is clear there is a hazard in the form of a reservoir of resistant microorganisms in food-producing animals. However, the exposure (and thus risk) that humans get from this reservoir is still unclear. First the methods applied by the Netherlands Veterinary Medicines Institute (SDa) for the monitoring of antimicrobial consumption in food-producing animals in the Netherlands using Defined Animal Daily Doses were investigated and assessed. Then, these methods were compared for pig farms with the monitoring method for antimicrobial consumption used in Denmark. After understanding the methods for monitoring of antimicrobial use, the internal exposure was assessed through modelling approaches. Both too high and too low (sub-therapeutic) dosing of antimicrobials may needlessly accelerate the development of resistance. Therefore, although little explored, a possible method to reduce antimicrobial resistance in livestock may be to optimize treatment regimens as is also done in human medicine. In order to optimize the treatment regimen, predictions must be made about the efficacy of the antimicrobial at the site of infection and about the antimicrobial concentrations achieved at the site of resistance development, which typically is the gut. Pharmacokinetic modelling approaches, such as physiologically based pharmacokinetic (PBPK) models, that simulate internal concentration-time profiles of antimicrobials in food-producing animals may be the tool needed for the assessment of antimicrobial concentrations in the gut, without the need to perform large-scale animal experiments. Two PBPK models were developed, one for cattle and one for piglets, using the antimicrobial oxytetracycline as a pilot compound. For calves and adult cattle, the models proved to be robust in modelling slow- and immediate-release oxytetracycline concentrations after intramuscular administration in edible tissues and the blood compartment. Therefore, we can conclude the model parametrization was performed accurately and that the equations used are correct. The model for sheep was successfully re-build based on the model script provided in literature and adapted to a PBPK model for piglets. Kinetics of oxytetracycline in piglets, however, proved different from kinetics in sheep and cattle. Therefore, the piglets PBPK model still requires additional data for parametrization and validation. In conclusion, this research showed the potential of modelling approaches to determine the exposure of food-producing animals to antimicrobial substances. The SDa system provides relatively detailed information about the use of antimicrobials in farm animals in the Netherlands. Combining these data with internal exposure simulations generated by PBPK modelling may lead to a method of accurately describing exposure of animals to antimicrobials and make accurate predictions of the hazard and risk for antimicrobial resistance development and spread in animals and between animals and humans.