Assessing the predictive accuracy of oral glucose effectiveness index using a calibration model

Area of Science:

• Metabolic research
• Clinical diagnostics
• Epidemiological studies

Background:

• Reference methods for glucose effectiveness (GE) measurement, such as the glucose clamp and minimal model analysis (MMA) of frequently sampled intravenous glucose tolerance tests (FSIVGTT), are complex and not suitable for large-scale studies.
• Surrogate indices derived from oral glucose tolerance tests (OGTT) are widely used as a more feasible alternative to measure GE (oGE).
• The predictive accuracy of these OGTT-derived surrogate indices has not been rigorously validated against reference methods.

Purpose of the Study:

• To evaluate the accuracy of surrogate OGTT-derived indices (oGE) in predicting glucose effectiveness (GE) measured by the reference FSIVGTT (Sg_MM).
• To assess the predictive performance using calibration model analysis, including root mean squared error (RMSE) and cross-validation (CVPE).

Main Methods:

• 123 subjects with varying glucose tolerance (NGT, IFG/IGT, T2DM) underwent both FSIVGTT and OGTT.
• Calibration model analysis was employed to compare predicted Sg_MM from oGE with measured Sg_MM.
• Predictive accuracy was quantified using RMSE and CVPE.

Main Results:

• Insulin sensitivity and Sg_MM were reduced in subjects with T2DM and IFG/IGT compared to NGT.
• A modest but significant correlation was found between oGE and Sg_MM (r=0.25, p<0.001).
• Calibration modeling revealed a modest correlation (r=0.21, p<0.05) between measured and predicted Sg_MM, indicating poor predictive accuracy. CVPE and RMSE showed similar predictive abilities among surrogates.

Conclusions:

• While convenient and feasible for epidemiological research, OGTT-derived surrogate indices for GE demonstrate limited ability to accurately predict reference GE values.
• The predictive accuracy of these surrogate measures is insufficient for robust clinical or research applications requiring precise GE quantification.