NONMEM improves group parameter estimation for the minimal model of glucose kinetics

Area of Science:

• Endocrinology
• Pharmacokinetics
• Mathematical Modeling

Background:

• The minimal model of glucose kinetics is crucial for interpreting blood glucose and insulin levels after intravenous glucose tolerance tests (IVGTT).
• Standard nonlinear least squares fitting can lead to high variability and unidentifiable parameters in group studies.
• Modifications like tolbutamide administration improve parameter stability but hinder pancreatic secretion modeling.

Purpose of the Study:

• To evaluate the effectiveness of the nonlinear mixed effects model (NONMEM) for analyzing glucose kinetics data in a group setting.
• To compare the parameter estimation variability between the standard method and NONMEM.

Main Methods:

• Utilized an intravenous glucose tolerance test (IVGTT) on 20 healthy volunteers.
• Applied both standard nonlinear least squares fitting and the NONMEM approach for parameter estimation.
• Employed the NONMEM loss function for simultaneous estimation of population parameters.

Main Results:

• NONMEM significantly reduced the variability of group parameter estimates.
• Asymptotic parameter coefficients of variation with NONMEM were at least twenty times smaller than those from the classical method.
• NONMEM allows for simultaneous estimation of population means, variances, and covariances.

Conclusions:

• NONMEM offers a substantial improvement in the precision of group parameter estimates for glucose kinetics.
• This approach enhances the identifiability and reliability of the minimal model in group analyses.
• NONMEM is a valuable tool for analyzing IVGTT data in research settings involving treatment effects.