(W-126) Population Pharmacokinetics of Atomoxetine at Steady-State Dosing for ADHD Treatment in Children and Adolescents

Addison Leabo, MS – Data Scientist 1, Department of Pediatrics, Children’s Mercy Research Institute, Kansas City, MO; J. Steven Leeder, PharmD, PhD – Executive Director (Interim), Department of Pediatrics, Children’s Mercy Research Institute, Kansas City, MO; Allison Dunn, PharmD, MS – Research Assistant Professor, Center for Translational Medicine, University of Maryland School of Pharmacy, Baltimore, MD

Objectives: Atomoxetine (ATX), a non-stimulant selective norepinephrine reuptake inhibitor, is indicated for treatment of attention-deficit/hyperactivity disorder (ADHD). Due to highly polymorphic cytochrome P450 2D6 (CYP2D6) and P450 2C19 (CYP2C19) activities, systemic exposure of ATX following oral administration is highly variable. ATX exposures are >10-fold higher in poor metabolizers compared to normal metabolizers when given the same weight-based dose¹. The purpose of this project is to evaluate steady-state ATX exposures in a large heterogenous population with different metabolic phenotypes to assess ATX dosing based on individual genetic makeup and stage of development.

Methods: Prospective pharmacokinetic (PK) and demographic data from 3 clinical studies (86 patients total) were combined in which ATX was prescribed to ADHD-confirmed subjects (6-17 years old) phenotyped for CYP2D6 and CYP2C19 activity. Oral doses for these studies were either weight-based or based on other PK models. Non-linear mixed effects modeling was used to characterize ATX plasma concentration. Multiple structural models were evaluated and selected based on decrease in the objective value function, reduction in between-subject variability, and goodness of fit. Patient-specific covariates were analyzed to explain variability in PK parameters including weight on disposition parameters and phenotypic based clearance effects. Precision and accuracy of the final covariate model was assessed using 200 bootstrap simulations and quantitative-predictive checks (QPC) on 1000 simulations of dose normalized Cmax. The final covariate model was used to evaluate the dose of ATX required to meet theoretical therapeutic goals over a wide range of weights and different metabolism phenotypes. All analyses were performed in Pumas v2.5.1 and R 4.2.3.

Results: A 2-compartment distribution model with 3 first-order transit absorption compartments best fit the data with estimated apparent central volume and clearance of 117L and 30.7L/hr, respectively. A combination additive and proportional error model best described residual unexplained variability, with estimates of 75.3nM and 35%, respectively. Including actual body weight and CYP phenotype in the model reduced between-subject variability by 47.0 and 55.5% for apparent central volume and clearance, respectively. Parameter estimates were physiologically plausible and estimated with high precision while QPC demonstrated adequate prediction of Cmax. Theoretically therapeutic doses of ATX in patients with a wide range of weights and different metabolism phenotypes were estimated.

Conclusions: A population PK model was developed to describe the steady state ATX exposure in children and adolescents across the spectrum of CYP2D6 and CYP2C19 activities. These results suggest that model-based individualized dosing of ATX can provide children and adolescents with ADHD with an adequate dose to achieve theoretical therapeutic concentrations.

Citations: < ![1] Brown, Jacob T., et al. "Single dose, CYP2D6 genotype‐stratified pharmacokinetic study of atomoxetine in children with ADHD." Clinical Pharmacology & Therapeutics 99.6 (2016): 642-650.