Pharmacokinetic-Pharmacodynamic modeling and prediction of antibiotic effects

Problems of emerging antibiotic resistance are becoming a serious threat worldwide, and at the same time, the interest to develop new antimicrobials has declined. There is consequently a need for efficient methods to develop new treatments that minimize the risk of resistance development and that are effective on infections caused by resistant strains. Based on in silico mathematical models, describing the time course of exposure (Pharmacokinetics, PK) and effect (Pharmacodynamics, PD) of a drug, information can be collected and the outcome of various exposures may be predicted. A general model structure, that characterizes the most important features of the system, has advantages as it can be used for different situations. The aim of this thesis was to develop Pharmacokinetic-Pharmacodynamic (PKPD) models describing the bacterial growth and killing after mono- and combination exposures to antibiotics and to explore the predictive ability of PKPD-models across preclinical experimental systems. Models were evaluated on data from other experimental settings, including prediction into animals. A PKPD model characterizing the growth and killing for a range of E. coli bacteria strains, with different MICs, as well as emergence of resistance, was developed. The PKPD model was able to predict results from different experimental conditions including high start inoculum experiments, a range of laboratory and clinical strains as well as experiments where wild-type and mutant bacteria are competing at different drug concentrations. A PKPD model, developed based on in vitro data, was also illustrated to have the capability to replicate the data from an in vivo study. This thesis illustrates the potential of PKPD models to characterize in vitro data and their usage for predictions of different types of experiments. The thesis supports the use of PKPD models to facilitate development of new drugs and to improve the use of existing antibiotics.