Key points are not available for this paper at this time.
Background: Computational pharmacokinetics increasingly relies on real-world clinical data rather than theoretical simulations. This study integrated authentic patient measurements from PK-DB clinical trials with FDA regulatory databases to characterize drug disposition patterns. Methods: Concentration-time profiles from 75 patients receiving caffeine (100-400 mg, n = 37) or paracetamol (500-1000 mg, n = 38) were analyzed using noncompartmental and one-compartment modeling approaches. FDA regulatory API queries provided contextual regulatory information. Parameters were estimated by a nonlinear least squares optimization of actual clinical measurements. Results: Caffeine exhibited dose-linear kinetics with apparent oral clearance of 2.96 L/h and a 4.0-hour elimination half-life. Paracetamol's elimination was faster (t½ 1.8 h) with an apparent oral clearance of 14.29 L/h. One-compartment models fit patient data excellently (R² > 0.98; AAFE 1.08-1.50 for all doses). The correlation between dose and AUC reached an impressive r = 0.9997 for caffeine, which was a confirmation of linear pharmacokinetics. The parameters derived from the models were in agreement with published population studies; thus, the approach was validated. Conclusions: The combination of curated clinical databases along with empirical compartmental modeling leads to the derivation of pharmacokinetically interpretable parameters that are based on actual measurements. This framework allows for evidence-based dose optimization and is applicable to more than 150 compounds in PK-DB, thus providing a significant advantage over simulation-dependent methodologies in translational pharmacometrics.
Agboola et al. (Tue,) studied this question.