Population Pharmacokinetics of Eltrombopag in Healthy Subjects and Patients with Chronic Idiopathic Thrombocytopenic Purpura
Ekaterina Gibiansky (1), Jianping Zhang (2), Daphne Williams (3), Zhao Wang (1), Daniele Ouellet (2)
(1) ICON Development Solutions, Ellicott City, MD, USA; (2) GlaxoSmithKline, Research Triangle Park, NC, USA; (3) Bristol-Myers Squibb, Princeton, NJ, USA
Objectives: Eltrombopag is an orally bioavailable small molecule agonist of thrombopoietin receptor (TPO-R) that has been recently approved for treatment of chronic idiopathic (immune) thrombocytopenic purpura (ITP). The aim of this analysis was to develop a population pharmacokinetic (PK) model of eltrombopag, predict steady-state exposure at therapeutic doses, and identify and quantify main demographic/covariate factors influencing eltrombopag exposure.
Methods: 111 subjects from 3 Phase 1 studies (dense data) and 88 patients from a Phase 2/3 study (sparse data) contributed 4093 plasma eltrombopag concentrations. Dosing in the studies ranged from 5 to 200 mg QD as a single dose, or as multiple doses administered for 5 days to 6 weeks duration. The analysis was performed using a mixed-effects modeling approach with the first-order conditional method (FOCEI) of NONMEM. The full model approach was implemented for covariate modeling, followed by elimination of insignificant or poorly estimated covariates. Visual predictive check and non-parametric bootstrap stratified by major covariates were implemented for model evaluation. Estimates of individual eltrombopag steady-state exposure were obtained by simulations and summarized for subpopulations identified by the model.
Results: PK of eltrombopag was described by a 2‑compartment linear model with dual sequential first-order absorption, absorption lag-time, and inter-occasion variability in absorption. Mean (95% CI) parameters of a typical 70 kg Caucasian male ITP patient not taking corticosteroids were estimated as CL/F=0.668 (0.561, 0.775) L/hr, Vc/F=8.76 (8.14, 9.38) L, Vp/F=11.3 (10.1, 12.5) L, and Q/F=0.399 (0.361, 0.437) L/hr. Inter-individual variability was 40.6%, and 37.4% in CL/F and V/F, respectively, with correlation of R=0.743. Inter-occasion variability in Ka (CV=127%) was much higher than inter-individual variability, which was therefore dropped from the model. In the ITP patients that received 50 mg QD dosing, mean (95% CI for the mean) steady-state exposure was estimated as AUCτ=108 µg*hr/mL (88, 134) and Cmax=8.01 (6.73-9.53) µg/mL. Weight increased CL/F, Vc/F, Q/F, and Vp/F equally with a power coefficient of 0.62 ((0.45-0.78) for CL/F and Q/F, and (0.25-0.98) for Vc/F, and Vp/F). For the range of weights in the analysis (43 to 122 kg), CL/F, Vc/F, Q/F, and Vp/F increased with body weight from 26% lower to 41% higher values than for 70-kg individual. The mean (95%CI) CL/F was 33% (26%, 41%) lower in Asians compared to other races, 26% (7%, 45%) lower in patients taking corticosteroids concomitantly, 19% (7%, 31%) lower in females compared to males; and 17% (0, 34%) higher in healthy subjects compared to ITP patients. Age and mild renal impairment did not influence the eltrombopag PK.
Conclusions: The developed population pharmacokinetic model identified and quantified patient characteristics predictive of eltrombopag exposure, and enabled further analysis to characterize pharmacokinetic-pharmacodynamic relationships.
