Amy Meng, Shringi Sharma, Brian Kirby, and Anita Mathias
Gilead Sciences, Inc., Foster City, CA, USA
Objectives: A joint semi-mechanistic population pharmacokinetic (PopPK) model of sofosbuvir (SOF) and its metabolites (GS-566500 and GS-331007) was developed, using pooled data across studies evaluating the safety and efficacy of sofosbuvir (SOF: Sovaldi®) and ledipasvir/sofosbuvir (LDV/SOF: Harvoni®) in HCV-infected adolescent subjects (12
Methods: Data from two Phase 2 studies in HCV-infected adolescent subjects were used for this analysis including 129 subjects with measurable plasma concentration data (40 subjects with 733 samples from the SOF study and 89 subjects with 1211 samples from the LDV/SOF study). A non-linear mixed-effects modelling approach with the first-order conditional estimation + interaction (FOCEI) method in NONMEM 7, version 7.4 (ICON, Maryland) was utilized. The impact of covariates, including age, body weight, sex, creatinine clearance (calculated CLCRSW), ribavirin (RBV) or ledipasvir (LDV) usage on the PK of SOF and GS-331007 was investigated. Based on the number of SOF samples below the limit of quantitation, the censored-data likelihood (M3) method was evaluated. Inter-individual variability was tested on central and peripheral clearance (CL/F and Q/F), central and peripheral volume (Vc/F and Vp/F), and absorption rate constant (Ka). Various models were tested to characterise the absorption profile.
Results: Plasma PK of SOF and its metabolites was described by a joint 1-compartment model with sequential first-zero order absorption for SOF, a 1-compartment model with first order absorption for GS-566500, and a 2-compartment model with first order absorption and lag time (Tlag) for GS-331007. The models were parameterized using CL/F, Q/F, Vc/F, Vp/F, Ka, Tlag, and relative absorbed fraction of each analyte (F). The orally administered SOF dose was divided into a fraction F1 leading to the parent and fractions F2 and F3 leading to GS-566500 and GS-331007, respectively, prior to reaching the circulation (accounting for the pre-systemic conversion of SOF to its metabolites in enterocytes via mucosal Cathepsin A). The sequential systemic conversion of SOF to its metabolites was also incorporated (CLSOF→CLGS-566500→CLGS-331007) to account for the Cathepsin A and Carboxylesterase 1 mediated SOF metabolism in hepatocytes. The typical estimated CL/F values were 122.7L/hr, 86.5L/hr, and 13.7L/hr for SOF, GS-566500, and GS-331007, respectively. The relative molar% of dose absorbed was 33.6% for SOF, 11.9% for GS-566500 and 54.5% for GS-331007 respectively in the SOF study and 57.3% for SOF, 7.6% for GS-566500 and 35.1% for GS-331007 in the LDV/SOF study. Statistically significant parameter-covariate relationships identified included LDV co-administration on SOF F1, and RBV co-administration, CLCRSW, and body weight on GS-331007 CL/F.
When administered with LDV (a P-gp inhibitor), relative bioavailability (F1) of SOF (a P-gp substrate) was increased by 70%. Of the covariates examined, co-administration of RBV increased GS-331007 CL/F by 62%. GS-331007 CL/F was increased by 33% and 9% between the 5th and 95th percentiles of CLCRSW (106-202mL/min) and body weight (45-91kg) values, respectively, suggesting minimal clinical impact of these covariates on GS-331007 PK. Finally, both the joint population PK model and the previously developed separate models for individual analytes performed similarly as evaluated based on model diagnostics, bootstrap output and predicted post-hoc exposures (AUCtau and Cmax). Similar analysis was performed for adult population, resulting in comparable exposures.
Conclusions: SOF, GS-566500, and GS-331007 PK in adolescent HCV-infected subjects can be comparably described by joint or individual population PK models, based on model performance and post hoc exposure estimates. This analysis illustrates that a joint parent-metabolite model can provide semi-mechanistic understanding, whereas if post hoc exposure estimates are the purpose of the modeling effort, individual models are appropriate for the analysis.
Reference: PAGE 27 (2018) Abstr 8770 [www.page-meeting.org/?abstract=8770]
Poster: Drug/Disease Modelling - Paediatrics