Related Experiment Videos
In vivo-in vitro correlation (IVIVC) modeling incorporating a convolution step
Journal of Pharmacokinetics and Pharmacodynamics
|July 27, 2001
Summary
In vivo-in vitro correlation (IVIVC) models can be improved by incorporating a convolution step. The odds model, including convolution, better describes plasma drug concentration data than the traditional identity model.
Area of Science:
- Pharmacokinetics and Pharmacodynamics
- Drug Development and Formulation
- Biostatistics and Modeling
Background:
- In vivo-in vitro correlation (IVIVC) models are crucial for predicting in vivo drug performance from in vitro data.
- Traditional IVIVC models often rely on deconvoluted data, which can be unstable and may not directly address primary interests in plasma drug concentration.
- The focus is typically on plasma drug concentration-time profiles (e.g., AUC, Cmax), not solely on the fraction dissolved.
Purpose of the Study:
- To describe and evaluate IVIVC models that incorporate a convolution step to directly model plasma drug concentration data.
- To compare the performance of the odds model (with convolution) against the identity model for IVIVC.
- To demonstrate an improved modeling approach for drug absorption and disposition.
Main Methods:
- Development and description of IVIVC models incorporating a convolution step, including odds, hazards, and reversed hazards models.
- Fitting the odds model and the identity model to plasma drug concentration/time data from two different products.
- Utilizing nonlinear mixed-effects modeling software for model fitting and comparison.
Main Results:
- The odds model, which includes a convolution step, demonstrated a reasonable fit to the plasma drug concentration data for both products.
- The odds model provided a significantly better fit compared to the identity model in both data sets.
- Incorporating a convolution step overcomes the limitations associated with analyzing deconvoluted data.
Conclusions:
- The odds model, incorporating a convolution step, offers a more robust and accurate approach for IVIVC compared to the identity model.
- This modeling strategy directly addresses the primary interest in plasma drug concentration and its functions.
- The findings support the use of convolution-incorporating models for improved prediction of in vivo drug behavior.