PopPK Modeling Strategies

Under construction

General PopPK models

Absorption models

Bioavailability

Each component in the bioavailability equation can vary due to drug-specific physicochemical properties or physiological factors. In turn, these variations can contribute to inter-individual variability (IIV) in F.

When intravenous (IV) data is unavailable for modeling oral pharmacokinetic (PK) data, a common practice is to attribute the IIV associated with F to the apparent clearance (CL/F) and apparent volume of distribution (V/F) parameters. By doing so, any correlation between these two parameters can also be estimated.

Alternatively, one can directly introduce an IIV on F by fixing its value to 1 (renaming it Frel, “rel” for “relative”) and estimating the inter-individual variability in Frel itself. While this approach may not always improve predictive performance, it can be valuable if:

  • F-specific covariates are suspected, or
  • Residual unexplained variability (RUV) must be accounted for in future simulations.

Distribution models

The distribution of drug is often modeled using compartments. The most common distribution models are one and two-compartment models.

Elimination models

Ususally, elimination is modeled using a first-order elimination rate, meaning that the rate of elimination depends on the current drug concentration.

Specific PopPK models

  • Monoclonal antibodies: TMDD models and their approximations

Using PopPK Models for a New Indication

  1. Start with an Existing Model: You have a population pharmacokinetic (PopPK) model for Drug X used in Indication A.
  2. Gather New Data: New data is collected for Drug X in a different setting, Indication B.
  3. Evaluate the Existing Model: Test the current PopPK model with the new data from Indication B without making any changes.
  4. Assess Model Fit
  • If the model fits well: Use the existing model to estimate individual drug exposure.
  • If the model does not fit well: Update the PopPK model by combining the old data with the new data.
    • If the old data is unavailable, perhaps a prior can be used.
  1. Re-evaluate the Updated Model
  • If the updated model fits the new data well: Use it to estimate individual drug exposure.
  • If it still doesn’t fit: Add more covariates, e.g. indication.
  1. Refine the Model if Needed
  • If the updated model still doesn’t fit well: Consider improving the basic structure of the model to better capture the drug’s behavior.

By following these steps, you can effectively use an existing PopPK model for a new Indication setting, ensuring accurate estimates of drug exposure.

References