1980 Volume 23 Issue 2 Pages 129-136
When the bed-side type of artificial pancreatic beta cell system developed originally by us was applied to diabetics whose insulin sensitivity was low or supernormal, it proved necessary to change parameters manually in order to decide the insulin infusion rates for adaptive control of blood glucose.
In the present study, a computer algorithm for self-adaptive control was establisehed. Firstly, the projected rate of change in blood glucose [ΔBGp (t)] calculated from the relationship between the rate of change in blood glucose concentration and blood glucose concentration during declines from the peak values after 50 or 100 g oral glucose loads in normal subjects, was put into the program. Under blood glucose regulation with the artificial pancreatic beta cell, the real rate of change in blood glucose concentration in each subject [ΔBG (t)] was calculated, and the index of insulin sensitivity was then estimated from the difference between [ΔBG (t)] and Δ(t). According to the calculated insulin sensitivity, the computer automatically changes the parameters which regulate the insulin infusion rate.
Validation of this algorithm was performed in depancreatized dogs which were made ketoacidotic by 3-hydroxybutyric acid infusion or made hyperepinephrinemic by epinephrine infusion. On application of this algorithm, adaptive control of blood glucose in the insulin resistant dogs was carried out by infusing insulin at 2.8 times or 2. 2 times the insulin doses required when 3-hydroxybutyric acid and epinephrine were not infused, respectively.
This artificial pancreatic beta cell system with a computer algorithm for adaptive control represents a suitable research tool for investigating the insulin sensitivity of diabetics or for analyzing the blood glucose regulatory mechanism.
