Automatic computation of the glycemic index: data driven analysis of the glucose standard

Abstract

The Glycemic Index (GI) is a tool for classifying carbohydrates based on their impact on postprandial glycemia, useful for diabetes prevention and management. This study applies a mathematical model for a data driven simulation of the glycemic response following glucose ingestion. The analysis is performed on a dataset of 35 healthy subjects undergone a standard 50 g oral glucose test. The results reveal a direct correlation between glucose response profiles and parameters describing glucose absorption, enabling the classification of subjects into three groups based on the timing of their glycemic peak: <30 min, 30-50 min, >50 min. These findings highlight the ability of a physiology-based mathematical model to capture inter-individual variability in postprandial glucose dynamics and represent a step toward simulation-based approaches for GI estimation.

Figures (13)

• Figure 1: Representation of the Dalla Man--Rizza--Cobelli model with six compartments (yellow boxes) and their relations.
• Figure 2: (a) Profile of gastric emptying rate $k_\mathrm{empt}$ in Eq. \ref{['eq:kempt']} for fixed values of $d$ and $b$ ($0.2$ and $0.9$, respectively). (b) Profile of $k_\mathrm{empt}$ in variation of $d$ with $b=0.8$. (c) Profile of $k_\mathrm{empt}$ in variation of $b$ with $d=0.1$. The three plots were generated with fictitious values for illustration purposes. In particular, $D=10^4$, $k_\mathrm{max}=0.05$, $k_\mathrm{min}=0.008$.
• Figure 3: Comparison between in vivo and estimated glucose concentration curves over the 35 considered subjects. (a) Mean and deviation of the in vivo glucose curves of our database. (b) Comparison between mean and deviation of in vivo curves (blue lines) and estimated curves (orange lines).
• Figure 4: Quantities of (a) glucose concentration, (b) plasma insulin concentration, (c) endogenous glucose production, (d) glucose rate of appearance, (e) insulin secretion, and (f) glucose utilization. The solid line represents the mean of 35 subjects, and the shaded area represents the SD.
• Figure 5: Quantities of (a) glucose concentration, (b) plasma insulin concentration, (c) endogenous glucose production, (d) glucose rate of appearance, (e) insulin secretion, and (f) glucose utilization of subjects 33 (blue line), 21 (red line), 6 (yellow line).