2006 - Brugge/Bruges - Belgium

PAGE 2006: Methodology- Other topics
Barry Weatherley

Population Analysis of Maraviroc Phase 1 Noncompartmental Pharmacokinetic Data

Weatherley B., L. McFadyen, P. Milligan

Clinical Pharmacology, Sandwich Laboratories, Pfizer Inc., United

PDF of poster

Introduction: Maraviroc, an antagonist of the human CCR5 receptor, is being developed for the treatment of HIV. Noncompartmental analysis (NCA) of Phase 1 studies showed the PK of maraviroc after single and multiple oral doses are not dose proportional. Food caused a reduction in the rate and extent of absorption with a greater effect at 100mg than at 600mg.

Objectives: To combine the NCA AUC and Cmax values from single and multiple dose Phase 1 studies to derive parametric models describing the effects of dose, food, formulation and steady state on the PK of maraviroc.

Methods: 330 AUC and 395 Cmax values from 134 healthy subjects in 5 Phase 1 studies were used. Single and multiple QD and BID solution and tablet doses ranged from 1 to 1200 mg. Sigmoid Emax models with intercept (E0) were applied to dose-normalised AUC (NAUC) or Cmax (NCMX). Additive and proportional residual error models with log-transformed data were used. Constraints were placed on changes in E0, Emax, Hill coefficient and ED50 arising from the dose, food, formulation and steady state covariates. Parameter estimation utilised mixed effect non-linear regression (NONMEM V).

Results: Population parameter estimates for the NAUC model (NAUC0-infinity, single dose solution, fasted) were E0 0.72 h.ng/mL/mg, Emax 9.42 h.ng/mL/mg, ED50 59.7 mg, Hill coefficient 0.96. Inter-subject variability in Emax was 25%. Proportional residual error was 33% and additive residual error was 0.14 h.ng/mL/mg. Food decreased E0 by 22% and increased ED50 by 511% vs fasted. Tablet decreased E0 by 22% and increased Emax by 37% vs solution. Tablet and food combined decreased E0 by 22% and increased ED50 by 511% vs fasted solution. Multiple dosing (BID, NAUC over dosing interval) increased E0 by 65% and increased Emax by 10% with ED50 unchanged vs single dose. Multiple dosing QD was identical to BID.

Parameter estimates for the NCMX model were E0 0.23 ng/mL/mg, Emax 2.06 ng/mL/mg, ED50 62.6 mg, Hill coefficient 2.23. Inter-subject variability in Emax was 24%. Proportional residual error was 15% and additive residual error was 0.89 ng/mL/mg. Food decreased E0 by 7% and increased ED50 by 732%. Tablet changes in E0 and Emax were small. Multiple dosing BID increased E0 by 42%, increased Emax by 12% and decreased ED50 by 7%. Multiple dosing QD parameters were between single and BID dosing.

Conclusion: The parametric models allow predictions of maraviroc AUC and Cmax under dosing, formulation and food conditions not studied.

Reference: PAGE 15 (2006) Abstr 1027 [www.page-meeting.org/?abstract=1027]
Poster: Methodology- Other topics