Population pharmacokinetics of artesunate and dihydroartesunate in adults and children following administration of a fixed dose combination formulation of chlorproguanil-dapsone-artesunate
Ann K. Miller (1), Mita M. Thapar (2), Siobhan C. Hayes (2), Colm Farrell (2), Bela R. Patel (1), Duane A. Boyle (1)
(1) GlaxoSmithKline, King of Prussia, PA, United States; (2) ICON Development Solutions, Marlow, Buckinghamshire, UK (mita.thapar@iconplc.com)
Objectives: To describe the population pharmacokinetics (PPK) for artesunate (ART) and its active metabolite dihydroartesunate (DHA) following the administration of a fixed‑dose combination of chlorproguanil-dapsone-artesunate (CDA) in healthy adults and adult/pediatric patients with P. falciparum malaria.
Methods: Plasma concentration-time data from four Phase I-III clinical studies including 61 healthy adults (HV),115 adult malaria patients (AP) and 1101 pediatric malaria patients (PP) with median age of 3 years (range; 1-14), were included in the PPK analyses. Separate PPK analyses were performed on log‑transformed data for ART and DHA using a non‑linear mixed-effect modeling (NONMEM) approach The exponents for allometric scaling for both oral clearance (CL/F) and apparent volume of distribution (V/F) were estimated. Separate random residual variability terms were applied to the adult and pediatric populations and a common inter-individual variability (IIV) term on the residual variability (eta on epsilon) was also included for both populations. Effects of covariates (e.g., age, gender, subject status) were evaluated on ART and DHA pharmacokinetics (PK).
Results: The final PPK model for ART was a one‑compartment model with two sequential first‑order absorption components and with first‑order elimination. The first and second absorption rate constants, Ka1 and Ka2 were estimated to be 0.58 hr-1 (IIV 51%) and 0.44 hr-1 (IIV 58%), respectively, with the change in absorption rate (MTIME) occurring at 2.4 hr post dose. Inter-occasion variability (IOV) in Ka was 41%. Allometric scaling exponents on CL/F and V/F were estimated to be 0.95 and 0.78, respectively. V/F was estimated to be 249 L (IIV 107%). Only subject status was found to significantly impact CL/F. The CL/F in HV was 1730 L/hr with the CL/F estimated to be 43% lower in AP and 2-fold higher in PP, with an associated IIV of 42%. All parameters for the structural and error models were estimated with a relative standard error (RSE) of less than 20% and 26%, respectively. Random residual variability estimates for pediatric and adult data were 111% and 59%, respectively, with an associated IIV of 11%.
The final PPK model for DHA was a one‑compartment model with an absorption (metabolic conversion) lag time, simultaneous zero‑order and first-order absorption with first‑order elimination. The fractions absorbed from the first‑order and zero‑order absorption processes were estimated to be 0.71 and 0.29, respectively. The lag-time for absorption was estimated to be 0.20 hr with a Ka of 0.50 hr-1 (IIV 24%) and a duration for zero-order input of 3.3 hr. CL/F was estimated to be 104 L/hr (IIV 43%) and V/F was 85.3 L (IIV 99%). All parameters for the structural and error models were estimated with RSE of less than 10% and 36%, respectively. Random residual variability estimates for pediatric and adult data were 150% and 59%, respectively, with an associated IIV of 25%. None of the covariate effects were found to have a clinically significant impact on DHA PK.
Conclusions: The final PPK models which incorporated the influence of weight on CL/F and V/F using allometry described ART and DHA PK data well following the administration of a fixed‑dose combination formulation of CDA. Additionally, subject status (HV, AP vs. PP) was found to be a predictor of ART oral clearance.