How physiologically based models are setting doses for children

Before a drug is given to a child for the first time, someone has to decide the dose. Increasingly, that decision rests on a computer model of a child’s body.

What PBPK models do

Physiologically based pharmacokinetic (PBPK) models represent the body as a set of compartments matching real organs, connected by blood flow. Each organ has a size, a blood supply and, where relevant, a capacity to metabolise or excrete drugs. Combine that physiology with a drug’s chemical properties and the model predicts how the drug will be absorbed, distributed and cleared.

For children, the key ingredient is ontogeny: how organ size, body composition, blood flow and enzyme activity change from birth through adolescence. Population models built into commercial and academic platforms describe those changes for virtual children of different ages.

Where they are used

• First-in-paediatric dosing, predicting the dose that will match adult exposure before any child is treated.
• Drug interaction questions, where running clinical interaction studies in children would be impractical or unethical.
• Neonates and infants, where maturation effects are greatest and data are scarcest.

What regulators expect

Both the FDA and the EMA have issued guidance on how PBPK analyses should be reported and qualified. Regulators want to see that the model has been verified against observed data, ideally in adults and older children first, before it is trusted to predict for younger age groups.

The limits

A PBPK model is only as good as its inputs. For many enzymes and transporters, maturation in very young children is still poorly characterised, and predictions for preterm babies remain uncertain. That is why model predictions are usually confirmed with sparse sampling in a clinical study rather than relied on alone.