A Comprehensive Model of Glucose-Insulin Regulation Including Acute and Prolonged Effects of Physical Activity in Type 1 Diabetes
Deichmann, J.; Bachmann, S.; Pfister, M.; Szinnai, G.; Kaltenbach, H.-M.
Show abstract
ObjectiveFor type 1 diabetic patients, accurate adjustment of insulin treatment to physical activity (PA) is a challenging open problem. Glucose uptake by the exercising muscles increases acutely, causing increased hepatic glucose production to maintain glucose homeostasis. Meanwhile, insulin sensitivity is elevated for a prolonged period to drive glycogen repletion during recovery. These processes strongly depend on PA duration and intensity, making their combined effects difficult to predict accurately. In this work, we develop a model of glucose-insulin regulation that captures PA from low to high intensity including acute and prolonged effects on glucose metabolism. MethodsWe extended an existing minimal model of glucose-insulin regulation to capture PA-driven changes in glucose metabolism. We incorporated the insulin-independent increase in glucose uptake and production, including the effects of glycogen depletion and of high-intensity PA on production. The model also captures the prolonged increase in insulin sensitivity. ResultsThe model accurately predicts glucose dynamics of published data during short and prolonged PA of moderate to high intensity and during subsequent recovery. In-silico full-day studies elucidate the effects of timing, duration and intensity of PA and of insulin bolus reduction on glucose levels during and after the activity. ConclusionThe proposed model captures the blood glucose dynamics during all main PA regimes. SignificanceMathematical models of glucose-insulin regulation are critical components of closed-loop insulin delivery and clinical decision support systems for achieving good glycemic control. The presented model shows potential for the development and assessment of algorithms targeting treatment adjustment to PA.
Matching journals
The top 2 journals account for 50% of the predicted probability mass.
Similar papers in this journal
- An approximate stochastic optimal control framework to simulate nonlinear neuromusculoskeletal models in the presence of noise 93%
- Kinetic and data-driven modeling of pancreatic β-cell central carbon metabolism and insulin secretion 93%
- ADP is the dominant controller of AMP-activated protein kinase activity dynamics in skeletal muscle during exercise 93%
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.