Population pharmacokinetic modeling of risperidone and 9-hydroxyrisperidone to estimate CYP2D6 subpopulations in children and adolescents

Abstract

Aim: The study aims were to characterize risperidone and (±)-9-hydroxyrisperidone pharmacokinetic (PK) variability in children and adolescents and to evaluate covariate effects on PK parameters.

Methods: Steady-state samples were drawn at predose, 1, 2, 4, and 7 hours postdose; cytochrome P450 2D6 (CYP2D6) genotypes were available for 28 subjects. A nonlinear mixed-effects model (NONMEM) modeled the PKs of risperidone and (±)-9-hydroxyrisperidone; covariates included age, weight, sex, and CYP2D6 phenotype. The model included 497 observations [risperidone (n = 163), (+) and (-)-9-hydroxyrisperidone (n = 334)] from 45 subjects aged 3-18.3 (mean 9.6 ± 3.7) years, weighing 16.8-110 (43 ± 20.2) kg.

Results: A 1-compartment mixture model described risperidone and (±)-9-hydroxyrisperidone clearances for 3 CYP2D6 metabolizer subpopulations: extensive, intermediate, and poor. Weight significantly affected (±)-9-hydroxyrisperidone clearance. Clearance estimates in the mixture model were poor metabolizer 9.38 L/h, intermediate metabolizer 29.2 L/h, and extensive metabolizer 37.4 L/h.

Conclusion: Active moiety [risperidone plus (±)-9-hydroxyrisperidone] PK variability and the covariate effects were better explained with the addition of metabolite PK parameters. This model may aid the development of individualized risperidone dosing regimens in children and adolescents.

Figures

Fig 1

Fig. 1A Final model observed versus population predicted risperidone concentrations (Line of identity shown for clarity) Fig 1B Final model risperidone conditional weighted residuals (CWRES) versus population predicted risperidone concentrations (Regression line shown as dashed line) Fig 1C Final model risperidone conditional weighted residuals (CWRES) versus time after dose (h) (Regression line shown as dashed line)

Fig 1

Fig. 1A Final model observed versus population predicted risperidone concentrations (Line of identity shown for clarity) Fig 1B Final model risperidone conditional weighted residuals (CWRES) versus population predicted risperidone concentrations (Regression line shown as dashed line) Fig 1C Final model risperidone conditional weighted residuals (CWRES) versus time after dose (h) (Regression line shown as dashed line)

Fig 1

Fig. 1A Final model observed versus population predicted risperidone concentrations (Line of identity shown for clarity) Fig 1B Final model risperidone conditional weighted residuals (CWRES) versus population predicted risperidone concentrations (Regression line shown as dashed line) Fig 1C Final model risperidone conditional weighted residuals (CWRES) versus time after dose (h) (Regression line shown as dashed line)

Fig 2

Fig. 2A Final model observed versus population predicted 9-OH risperidone concentrations (Line of identity shown for clarity) Fig 2B Final model 9-OH risperidone conditional weighted residuals (CWRES) versus population predicted risperidone concentrations (Regression line shown as dashed line)

Fig 2

Fig. 2A Final model observed versus population predicted 9-OH risperidone concentrations (Line of identity shown for clarity) Fig 2B Final model 9-OH risperidone conditional weighted residuals (CWRES) versus population predicted risperidone concentrations (Regression line shown as dashed line)

Fig 3

Visual Predictive Check, risperidone observed data compared to the 95th, 50th, and 5th percentiles for 100 simulated data sets. No of simulated observations = 3633, % observations outside 90% CI = 0%. Observed data binned according to “ideal” sampling times after dose 1, 2, 4 and 7 h. Comparison of median (dashed line) and 5–95th percentile interval (solid black lines)

Fig 4

Visual Predictive Check, 9-OH risperidone observed data compared to the 95th, 50th, and 5th percentiles for 100 simulated data sets. No of simulated observations = 3794, % observations outside 90% CI = 3%. Observed data binned according to “ideal” sampling times after dose 1, 2, 4 and 7 h. Comparison of median (dashed line) and 5–95th percentile interval (solid black lines)