Trevor N. Johnson(1), Thomas Kerbusch(3), Peter A. Milligan(3), Barry Jones(4), Geoffrey T. Tucker(1,2) and Amin Rostami-Hodjegan(1,2)
Simcyp Ltd(1) John Street Sheffield S2 4SU, UK Academic Unit of Clinical Pharmacology(2), University of Sheffield, , Royal Hallamshire Hospital, Sheffield S10 2JF, UK. (3)Clinical Pharmacology, Pfizer Ltd, Sandwich, Kent, UK.(4)Pharmacokinetics Dynamics and Metabolism, Pfizer Ltd, Sandwich, Kent, UK.
Purpose. To assess, by simulation, factors that influence the detection of mDDIs in phase 2/3 clinical trials using population pharmacokinetics (POPPK).
Method. Steady state plasma concentrations of a hypothetical drug in the presence and absence of enzyme inhibitors, were generated from in vitro data using the Simcyp program. Population (Caucasian, 50% male, 20-50 y) size was varied from 80 – 2000. The compound was metabolized mainly by CYP3A4 with a contribution from CYP2D6. Concomitant medications (COMEDs) with different Iu/Ki ratios (Iu = population average unbound plasma concentration, Ki = inhibition constant)(0.006, 0.026, 0.38, 3.3, 11, 22 for CYP3A4; 0.06, 0.9, 6.75, 13.5 for CYP2D6) were evaluated. The frequency of COMED was varied from 1.25% – 10%. The extent of interaction was determined using NONMEM (p > 0.001 backward; p > 0.01 forward selection of covariate).
Results. No false negative (Iu/Ki ≥ 0.38) or false positive (Iu/Ki < 0.38) interactions were detected using a population size of 2000. However, at Iu/Ki = 0.38 (e,g. fluconazole 50 mg/day) and COMED level of 2.5%, a statistically significant interaction could only be detected using > 480 subjects.
Conclusion. Simulations are recommended to define the size of study populations necessary to detect mDDIs with confidence using the POPPK approach.
Reference: PAGE 14 (2005) Abstr 831 [www.page-meeting.org/?abstract=831]
Poster: poster