Renal impairment or chronic kidney disease (CKD) affects the pharmacokinetics of renally eliminated drugs. There are hints that CKD has also an impact on hepatic metabolism. Using physiologically-based pharmacokinetic (PBPK) modeling approaches, concentration-time profiles of drugs can be simulated a priori in order to design clinical trials which are required in patients with CKD. In this thesis, physiological alterations during different CKD stages were quantified by an extensive literature search. Using PK-Sim® as PBPK modeling software, a parametrization of the physiological changes was performed, additionally distinguishing between age- and disease-related alterations by using an incorporated aging database. The uncertainty of each parameter was described by calculating a Taylor series expansion.
In order to qualify the parametrization, PBPK models of four probe compounds (gentamicin, gadodiamide, zanamivir, amikacin) which are eliminated solely by glomerular filtration were built. The mean prediction error (ME) and the root mean squared prediction error (RMSE) were calculated to assess the predictive performance of the diseased-informed fractional changes compared to uninformed simulations in which solely the glomerular filtration rate (GFR) was adjusted. Additionally, the activities of the cytochrome P450 (CYP) isoenzymes 3A4 and 1A2 were evaluated by including pharmacokinetic studies of probe substrates for these enzymes and estimating the differences in the respective clearances between diseased and healthy subjects. PBPK models for midazolam (substrate for CYP3A4) and theophylline (substrate for CYP1A2) were built and area under the curve (AUC) ratios of both observed and predicted AUCs were calculated. 
The pharmacokinetics of the probe drugs solely eliminated by glomerular filtration were well predicted for every CKD stage where pharmacokinetic studies have been conducted since ME calculations did not indicate over- or underprediction except for both informed and uninformed simulations of end-stage renal disease (ESRD) patients after administration of gentamicin. The RMSE did not show a clear picture when comparing disease-informed simulations to the uninformed ones. Whereas the informed simulations of ESRD patients were associated with a worse prediction in most of the cases, the informed simulations of the other CKD stages mostly depicted a clear improvement of predictive performance. In the case of CYP3A4/1A2-metabolized drugs, no significant changes during the different CKD stages were observed, which was supported by the comparison of predicted AUC ratios to the observed ones for midazolam and theophylline, respectively.
In summary, this thesis provides support for specific considerations regarding clinical trial design and pharmacotherapy for patients suffering from renal impairment.