Population pharmacokinetic and exposure-response analysis for ambrisentan for the treatment of pulmonary arterial hypertension in paediatric subjects aged 8 to less than 18 years.
Malek Okour (1), Misba Beerahee (1), Colm Farrell (2), Mary Ann Lukas (1), Mita M Thapar (2)
(1) GlaxoSmithKline Research & Development Limited*, (2) ICON plc
Objectives:
The primary objectives of the analyses were:
- To evaluate the predictive performance of the previously developed adult ambrisentan population pharmacokinetic (PK) model using the data from 8 to <18-year-old subjects with PAH;
- To develop a population PK model for ambrisentan using data from 8 to <18-year-old subjects with PAH;
- To explore the exposure-response relationship between
- ambrisentan PK and change from baseline in 6 minute walk distance (6MWD) at 12 and 24 weeks;
- ambrisentan exposure and incidence of adverse events (AE) related to ambrisentan.
Methods:
NONMEM® program version VII level 4.1 was used for all analyses using PDx-Pop (Version 5.2) as an interface. The final Pop PK model was evaluated by performing a confidence interval prediction-corrected visual predictive check (CI pc-VPC)1.
Data:
The data used for development of population PK model was from a 24-week randomised, open label study aiming to evaluate safety, tolerability, and efficacy of high and low dose ambrisentan (adjusted for body weight) treatment groups in subjects aged 8 to <18 years with PAH. The population PK analysis dataset consisted of 211 observations (111 from low dose and 100 from high dose group) from 39 paediatric subjects with PAH of which 20 subjects were in the low dose group and 19 subjects were in the high dose group.
Results:
- A CI pc-VPC with the parameter estimates from the previously reported final population PK model in adult population demonstrated adequacy of the adult final population PK model in predicting the observed ambrisentan data in 8 to <18-year-old paediatric subjects with PAH.
- Ambrisentan PK in 8 to <18-year-old were adequately described by a two‑compartment model with first-order absorption and elimination and an absorption lag-time. Body weight was a significant covariate on the apparent clearance and volume parameters. No other covariates were identified as significant in the current model.
- For a subject of 70 kg, the estimated mean (95% CI) values were CL/F=1.17 (1.04, 1.33) L/hr, Vc/F=12.3 (8.94, 16.8) L, Vp/F=81.3 (50.2, 132) L, Q/F=0.457 (0.302, 0.691) L/hr, Ka=2.46 (1.49, 4.07) hr-1, ALAG=0.525 (0.393, 0.700) hr. The effect of weight on both CL/F and V/F were fixed to the theoretical values for allometric scaling. For the range of weights in the Study AMB112529 analysis dataset (20-77 kg), CL/F ranged from 0.457-1.26 L/hr. V/F scaled linearly with weight, ranging from 3.51-13.5 L across the weight range.
- Ambrisentan systemic exposure was neither strongly correlated to change from baseline in 6MWD at 12/24 weeks nor to the incidence of ambrisentan-related AEs.
Conclusions:
- Ambrisentan plasma concentrations in the 8 to < 18-year-old paediatric population were well interspersed with the historical adult data where ambrisentan PK were well described by a two‑compartment model with first-order absorption and elimination and an absorption lag-time.
- Ambrisentan systemic exposure (AUCss and Cmax,ss) observed in the 8 to <18-year-old paediatric population was comparable to the exposure observed in the adult population suggesting appropriateness of the weight-based dosing scheme in the 8 to <18-year-old population.
- Ambrisentan exposure-response analyses in adults and paediatric subjects were similar as neither analysis identified any strong relationships between exposure to ambrisentan and effects in 6MWD or incidence of ambrisentan-related AEs.
- Bergstrand M, Hooker AC, Wallin JE, et al (2011), Prediction-corrected visual predictive checks for diagnosing nonlinear mixed-effects models. The AAPS Journal; 13(2):143-51
*this study was sponsored by GlaxoSmithKline Pharmaceuticals.