Prospective Predictions of Human Pharmacokinetics for Eighteen Compounds

Abstract

Quantitative predictions of pharmacokinetics (PKs) and concentration-time profiles using in vitro and in vivo preclinical data are critical to estimate systemic exposures for first-in-human studies. Prospective prediction accuracies of human PKs for 18 compounds across all Biopharmaceutics Classification System/Biopharmaceutics Drug Disposition Classification System classes were evaluated. The a priori predicted profiles were then compared with clinical profiles. Predictions were conducted using advanced compartmental absorption and transit (ACAT) physiology based PK models. Human intravenous profiles were predicted with in vivo preclinical intravenous data using Wajima formulas. Human oral profiles were generated by combining intravenous PKs together with either physiologically based oral ACAT models utilizing solubility and permeability data or by using the average bioavailability (F) and absorption rate constant (ka) from preclinical species. Key PK parameters evaluated were the maximum plasma concentration (Cmax), the area under the plasma concentration-time curve (AUC), CL/F, and Vdss/F. A decision tree was provided to guide human PK and ACAT predictions. Our prospective human PK prediction methods yielded good prediction results. The predictions were within a twofold error for 80% (Cmax), 65% (AUC), 65% (CL/F), and 80% (Vz/F) of the compounds. The methods described can be readily implemented with available in vitro and in vivo data during early drug development.