IDF21-0095 Continuous glucose monitoring calibrated with point of care devices or glucometers for self monitoring of blood glucose

Abstract

Background: Iatrogenic hypoglycemia is an important complication for patients with diabetes. To detect a risk of hypoglycemia accurately during hospitalization, increasing accuracy of continuous glucose monitor (CGM) is necessary. Aim: We examined differences in calibration accuracy between point of care devices and “glucometers that were compliant with ISO15197:2013” (ISO compliant glucometers) when using personal CGM during hospitalization. Method: This was a prospective study where personal-CGM (GUARDIAN CONNECT) was worn for 6 days (CGM attachment: day 1) and blood glucose levels used to calibrate were obtained using point of care devices (StatStrip Xpress Glucose/Ketone) or ISO compliant glucometers (ACCU-CHEK Guide) using a randomized crossover design. Sixteen inpatients with type 2 diabetes were equally allocated to 2 groups. Point of care devices provided 11 blood glucose levels used to calibrate during days 2-3 and ISO compliant glucometers provided 11 blood glucose levels used to calibrate during days 4-5 in group 1. ISO compliant glucometers provided 11 blood glucose levels used to calibrate during days 2-3 and point of care devices provided 11 blood glucose levels used to calibrate during days 4-5 in group 2. In each group 1 and group 2, point of care devices and ISO compliant glucometers were allocated from “fixed 2 point of care devices” and “fixed 2 ISO compliant glucometers” using a randomized quadruple (2 × 2) crossover design. Mean absolute relative difference (MARD) was calculated for every calibration. Results: Blood glucose levels used to calibrate obtained using point of care devices was significantly higher than blood glucose levels used to calibrate obtained using ISO compliant glucometers (173.6 ± 69.9 mg/dL vs. 158.2 ± 47.9 mg/dL, p = 0.02: unpaired t-test). MARD was significantly higher on calibration using point of care devices than on calibration using ISO compliant glucometers (7.6 ± 7.5% vs. 5.8 ± 5.8%, p = 0.01: unpaired t-test). The study patients correlated to distributions of MARD on calibration using point of care devices (η2 = 0.19, p = 0.003: correlation ratio). The study patients also correlated to distributions of MARD on calibration using ISO compliant glucometers (η2 = 0.23, p < 0.001: correlation ratio). Discussion: StatStrip Xpress Glucose/Ketone is cheaper than other point of care devices in Japan, therefore, StatStrip Xpress Glucose/Ketone should be thought to be superior in terms of being able to measure the wide range of glucose levels (10–900 mg/dL) cheaply. ISO compliant glucometers may calibrate personal CGM more accurate than StatStrip Xpress Glucose/Ketone during hospitalization. The correlation between study patients and distributions of MARD may suggest the individual differences in CGM sensor accuracy between patients.