Modelling glucose dynamics during moderate exercise in individuals with type 1 diabetes

Area of Science:

• Biomedical Engineering
• Endocrinology
• Computational Biology

Background:

• Artificial pancreas systems require accurate glucose modeling for effective type 1 diabetes management.
• Simulating real-life scenarios, like exercise, is crucial for developing advanced artificial pancreas technology.
• Existing models may not fully capture glucose dynamics during moderate aerobic exercise.

Purpose of the Study:

• To propose and evaluate enhanced Bergman Minimal Models for simulating glucose dynamics during exercise in type 1 diabetes.
• To identify model parameters that physiologically represent the impact of exercise intensity and duration on glucose regulation.
• To provide a validated tool for in-silico testing of artificial pancreas systems during exercise.

Main Methods:

• Developed six variations of the Bergman Minimal Model to incorporate exercise effects.
• Estimated model parameters using a Bayesian approach with Markov chain Monte Carlo methods.
• Utilized clinical data from 17 adults with type 1 diabetes, including glucose, insulin, and oxygen consumption measurements during aerobic exercise.

Main Results:

• Identified a superior model incorporating glucose effectiveness and insulin action that increase proportionally with exercise intensity and duration.
• Model parameters were physiologically plausible and statistically validated.
• The selected model accurately reproduced results from two independent clinical studies.

Conclusions:

• The developed model accurately simulates the physiological effects of moderate exercise on glucose dynamics in individuals with type 1 diabetes.
• This enhanced modeling capability is vital for advancing artificial pancreas technology and exercise management strategies.
• The study provides a critical tool for in-silico testing and optimization of artificial pancreas systems during physical activity.