Exposure-response modelling approaches for determining optimal dosing rules in children

Area of Science:

• Pharmacometrics
• Pediatric Pharmacology
• Biostatistics

Background:

• Paediatric populations exhibit varying exposure-response relationships across different age groups.
• Current regulatory guidance suggests general age groupings, but their suitability for all drugs and diseases is uncertain.
• Accurate dosing in children requires understanding how drug effects change with age.

Purpose of the Study:

• To evaluate model-based approaches for quantifying age-related changes in exposure-response parameters.
• To develop an optimal dosing rule strategy based on age-varying parameters.
• To assess the performance of these methods in pediatric drug development.

Main Methods:

• Utilized Bayesian penalized B-splines and model-based recursive partitioning to model continuous age.
• Developed methods for deriving optimal dosing rules from age-dependent exposure-response models.
• Conducted simulation studies using linear and Emax exposure-response models, motivated by epilepsy drug development and in vitro cyclosporine data.

Main Results:

• Both Bayesian penalized B-splines and bootstrapped model-based recursive partitioning effectively estimated linear exposure-response parameters.
• These novel methods outperformed traditional linear models using categorical age covariates (ICH E11 groupings).
• Bayesian penalized B-splines demonstrated superior accuracy in estimating model parameters compared to recursive partitioning.

Conclusions:

• Model-based approaches provide a more precise way to characterize age-related exposure-response relationships in pediatric populations.
• Bayesian penalized B-splines offer a robust and accurate method for optimizing pediatric dosing strategies.
• These findings support the development of more tailored and effective dosing regimens for children.