Steven Sun, PharmD: No relevant disclosure to display
Objectives: Pexmetinib is a novel dual inhibitor of p38 mitogen-activated protein kinase and endothelial tyrosine kinase receptor, Tie2, under prior investigation in low or intermediate-1 risk myelodysplastic syndrome (MDS) (NCT00916227 and NCT01496495). The objectives of this analysis were to describe the population pharmacokinetics (popPK) of pexmetinib and to evaluate potential covariates that may be important predictors of pexmetinib pharmacokinetics (PK).
Methods: Data from two Phase 1 healthy volunteer studies and two Phase 1 studies in low or intermediate-1 risk MDS patients were included in this analysis. Two different formulations of pexmetinib have been used in patients in clinical development, a neat powder-in-capsule (PIC) formulation and an enabled liquid-filled capsule (LFC) formulation. Upon selection of a base model accounting for formulation effect via a relative oral bioavailability term (F), a forward selection approach was utilized for the univariate and multivariate screen (p < 0.05) followed by a backward deletion approach (p < 0.01). First order conditional estimation with interaction was used in NONMEM v7.5.0 for pexmetinib popPK model development.
Results: 113 participants received the PIC formulation and 65 participants received the LFC formulation. A 2-compartment model with a sequential zero- and first-order absorption (2.13 hr; 0.44 hr-1, respectively) and a relative F term to model formulation effect was used to describe the pooled clinical PK data (3737 concentrations from 178 subjects) from both formulations simultaneously. The relative F of pexmetinib was decreased by approximately 92% when taking the PIC formulation compared to taking the LFC formulation. Subject’s health status (healthy or diagnosed with MDS) was an independent predictor of both apparent clearance (CL/F) and central volume of distribution (Vc/F), whereas weight was an independent predictor only for Vc/F. CL/F decreased by 32% and Vc/F increased by 170% in healthy subjects compared to those diagnosed with MDS. The effect of weight on Vc/F was 1.73, suggesting that the heaviest patients in study had a 2.4 higher fold Vc/F compared to the typical patients in study. All model parameters were estimated with good precision (relative standard error < 45%). Our final popPK model estimates for CL/F, Q/F, Vc/F, and apparent peripheral volume of distribution were 16.4 L/hr, 9 L/hr, 13 L, and 95.6 L, respectively.
Conclusion: A population PK model has been developed to describe two different formulations of pexmetinib tested in clinical studies. Although weight was retained as a significant covariate on volume of distribution, the clinical implications of this are thought to be minimal, and no dosing adjustment is needed on the basis of patient weight. In general, the model fit the pooled clinical data well; however, other absorption models may be explored to further characterize the formulation effect.
