External Evaluation of Risperidone Population Pharmacokinetic Models Using Opportunistic Pediatric Data

Area of Science:

• Pharmacokinetics and Pharmacodynamics
• Pediatric Pharmacology
• Drug Metabolism and Disposition

Background:

• Risperidone is an antipsychotic medication approved for schizophrenia in adolescents and for autistic disorder and bipolar mania in children and adolescents.
• Off-label use of risperidone extends to younger children for various psychiatric conditions, necessitating reliable dosing strategies.
• Several population pharmacokinetic (PopPK) models for risperidone and its metabolite 9-OH-risperidone exist, but their external validation in pediatric populations is crucial.

Purpose of the Study:

• To externally evaluate the predictive performance of published risperidone PopPK models using opportunistically collected pediatric data.
• To identify a robust PopPK model suitable for precision dosing of risperidone in pediatric patients.
• To assess the utility of opportunistic data for evaluating external PopPK models.

Main Methods:

• Utilized 103 risperidone and 112 9-OH-risperidone concentrations from 62 pediatric patients (aged 0.16–16.8 years).
• Assessed five published PopPK models (four parent-metabolite, one parent-only) using statistical criteria, goodness-of-fit plots, prediction-corrected visual predictive checks (pcVPCs), and normalized prediction distribution errors (NPDEs).
• Evaluated model accuracy, precision, bias, and generalizability to the external pediatric dataset.

Main Results:

• All tested models demonstrated similarly precise predictions for risperidone and 9-OH-risperidone, with Root Mean Square Errors (RMSE) within specific ranges.
• One model, a one-compartment mixture model, showed significantly lower bias for risperidone (Mean Percent Error [MPE] of 1.0%) compared to others.
• A different model, developed with fewer data and a similar population, exhibited lower bias for 9-OH-risperidone (MPE: 17%). However, none of the models were fully generalizable, and all showed modest predictive performance.

Conclusions:

• Opportunistically collected pediatric data may not be ideal for external evaluation of risperidone PopPK models due to heterogeneity.
• Existing risperidone PopPK models exhibit modest predictive performance in this pediatric cohort, suggesting uncaptured inter-individual variability.
• Further development and validation of PopPK models are needed for reliable precision dosing of risperidone in diverse pediatric populations.