Use of microdosing to predict pharmacokinetics at the therapeutic dose: Experience with 5 drugs

Objectives: A volunteer trial was performed to compare the pharmacokinetics of 5 drugs-warfarin, ZK253 (Schering), diazepam, midazolam, and erythromycin-when administered at a microdose or pharmacologic dose. Each compound was chosen to represent a situation in which prediction of pharmacokinetics from either animal or in vitro studies (or both) was or is likely to be problematic. Methods: In a crossover design volunteers received (1) 1 of the 5 compounds as a microdose labeled with radioactive carbon (carbon 14) (100 I¼g), (2) the corresponding 14C-labeled therapeutic dose on a separate occasion, and (3) simultaneous administration of an intravenous 14C-labeled microdose and an oral therapeutic dose for ZK253, midazolam, and erythromycin. Analysis of 14C-labeled drugs in plasma was done by use of HPLC followed by accelerator mass spectrometry. Liquid chromatography-tandem mass spectrometry was used to measure plasma concentrations of ZK253, midazolam, and erythromycin at therapeutic concentrations, whereas HPLC-accelerator mass spectrometry was used to measure warfarin and diazepam concentrations. Results: Good concordance between microdose and therapeutic dose pharmacokinetics was observed for diazepam (half-life t1/2 of 45.1 hours, clearance CL of 1.38 L/h, and volume of distribution V of 90.1 L for 100 I¼g and t1/2 of 35.7 hours, CL of 1.3 L/h, and V of 123 L for 10 mg), midazolam (t1/2 of 4.87 hours, CL of 21.2 L/h, V of 145 L, and oral bioavailability F of 0.23 for 100 I¼g and t1/2 of 3.31 hours, CL of 20.4 L/h, V of 75 L, and F of 0.22 for 7.5 mg), and development compound ZK253 (F = <1% for both 100 I¼g and 50 mg). For warfarin, clearance was reasonably well predicted (0.17 L/h for 100 I¼g and 0.26 L/h for 5 mg), but the discrepancy observed in distribution (67 L for 100 I¼g and 17.9 L for 5 mg) was probably a result of high-affinity, low-capacity tissue binding. The oral microdose of erythromycin failed to provide detectable plasma levels as a result of possible acid lability in the stomach. Absolute bioavailability for the 3 compounds examined yielded excellent concordance with data from the literature or data generated in house. Conclusion: Overall, when used appropriately, microdosing offers the potential to aid in early drug candidate selection. © 2006 American Society for Clinical Pharmacology and Therapeutics.