A comparison of the artificial pancreas (glucose controlled insulin infusion system) and a manual technique for assessing insulin sensitivity during euglycaemic clamping

Abstract

Two main methods are available for assessing insulin sensitivity with the hyperinsulinaemic euglycaemic clamp technique: one employs a glucose-controlled insulin infusion system (the Biostator) with automatic feedback control; the second depends on frequent glucose measurement and the use of an algorithm and a pocket calculator (‘manual’) to determine the glucose infusion rate. The amount of glucose infused is a measure of insulin sensitivity. The efficiency of the two methods was compared in nine normal subjects (seven lean, two obese). After an overnight fast subjects were infused with insulin at 50 mU · kg^-1 · h^-1 for 2 h; this rate was doubled during the first 10 min for the manual technique. Blood glucose averaged 4.7 ± 0.1 and 4.8 ± 0.1 mmol/l from 0 to 120 min for Biostator and manual techniques and did not deviate significantly from the desired level. Variability of the clamp was also similar over the same period (coefficient of variation 5.1 ±0.6% and 6.4 ±0.7%, Biostator and manual). Glucose infused to maintain steady state from 60 to 120 min was higher, however, with the manual than the Biostator method (5.7±0.6 versus 4.4 ± 0.6 mg·kg^-1·min^-1, p< 0.01) even when the loading dose was omitted, although the two methods correlated closely (p< 0.05). Glucose infusion rate varied more from minute to minute with the Biostator (coefficient of variation 28.8 ± 3% versus 12.2 ± 2.1%). Steady-state serum insulin levels (30–120 min) were the same during both methods. Thus both methods give effective clamping but the manual method is simpler and shows less variability in glu cose insulin infusion rate.