[Estimation of individual pharmacokinetic parameters using maximum a posteriori Bayesian method with D-optimal...

Area of Science:

• Pharmacometrics
• Bayesian Statistics
• Pharmacokinetic Modeling

Background:

• Accurate estimation of individual pharmacokinetic parameters is crucial for drug development.
• Sparse sampling strategies are desirable to reduce patient burden and costs.
• Existing methods like Multiple Linear Regression (MLR) have limitations in precision and flexibility.

Purpose of the Study:

• To develop and evaluate a Maximum a Posteriori Bayesian (MAPB) estimation method for individual pharmacokinetic parameters.
• To compare the performance of MAPB with MLR using a D-optimal sampling strategy.
• To assess the impact of sparse sampling on the accuracy and precision of MAPB.

Main Methods:

• Development of a MAPB estimation method.
• Utilized a D-optimal sampling strategy identified via WinPOPT software.
• Pharmacokinetic study of pioglitazone modeled using NONMEM.
• Performance evaluation using simulated data generated by Monte Carlo simulations.
• Comparison of MAPB with MLR in terms of accuracy and precision.

Main Results:

• MAPB accuracy and precision decreased with fewer samples per subject.
• MAPB estimation of clearance (CL) and volume of distribution (V) showed less bias with low inter-individual variability.
• MAPB demonstrated similar accuracy and precision to MLR for AUC estimation with a two-point design.
• MAPB outperformed MLR when sampling time was adjusted to one hour.

Conclusions:

• MAPB offers a viable alternative to MLR for pharmacokinetic parameter estimation with sparse sampling.
• MAPB provides enhanced sampling flexibility and richer pharmacokinetic information compared to MLR.
• The performance of MAPB is influenced by sampling design and data variability.