Accuracy and Time Delay of Glucose Measurements of Continuous Glucose Monitoring and Bedside Artificial Pancreas During Hyperglycemic and Euglycemic Hyperinsulinemic Glucose Clamp Study.

Abstract

Glucose values of continuous glucose monitoring (CGM) have time delays compared with plasma glucose (PG) values. The artificial pancreas (STG-55, Nikkiso, Japan) (AP), which measures venous blood glucose directly, also has a time delay because of the long tubing lines from sampling vessel to the glucose sensor. We investigate accuracy and time delay of CGM and AP in comparison with PG values during 2-step glucose clamp study. Seven patients with type 2 diabetes and 2 healthy volunteers were included in this study. CGM (Enlite sensor, Medtronic, Northridge, CA, USA) was attached on the day before the experiment. Hyperglycemic (200 mg/dL) clamp was performed for 90 minutes, followed by euglycemic (100 mg/dL) hyperinsulinemic (100 μU/mL) clamp for 90-120 minutes using AP. CGM sensor glucose was calibrated just before and after the clamp study. AP and CGM values were compared with PG values. AP values were significantly lower than PG values at 5, 30 minute during hyperglycemic clamp. In comparison, CGM value at 0 minute was significantly higher, and its following values were almost significantly lower than PG values. The time delay of AP and CGM values to reach maximum glucose levels were 5.0 ± 22.3 (NS) and 28.6 ± 32.5 ( P < .05) min, respectively. Mean absolute rate difference of CGM was significantly higher than AP (24.0 ± 7.6 vs 15.3 ± 4.6, P < .05) during glucose rising period (0-45 min); however, there were no significant differences during other periods. Both CGM and AP failed to follow plasma glucose values during nonphysiologically rapid glucose rising, but indicated accurate values during physiological glucose change.