Endocrine Abstracts

The mean absolute relative difference (MARD) on CGM was reported with the various values approximately from 5% to 10%. The MARD in patients with type 1 diabetes was reported to be higher than that in patients with type 2 diabetes. Therefore, as one of the reasons why the reported MARD values varied, it cannot be denied that glycemic variability affects calibration accuracy. We examined relationships between calibration accuracy and glycemic variability on CGM. This is a prospective study where personal-CGM (GUARDIAN CONNECT) and professional-CGM (iPro2) were worn in parallel for 6 days (CGM attachment: day 1) and “blood glucose levels used to calibrate” (BGc) were obtained using “glucometers that were compliant with ISO15197:2013” (ISO) or “those not compliant with ISO15197:2013” (not-ISO) using a randomized crossover design. 25 type 2 diabetic inpatients were almost equally allocated to 2 groups (n = 13 and n = 12). ISO provided 11 BGc during days 2–3 and non-ISO provided those during days 4–5 in group 1 and vice versa in group 2. Every pre- and post-prandial and bedtime (7 times) BGc were obtained on days 2 and 5, and every pre-prandial and bedtime (4 times) BGc were obtained on days 3 and 4. Over days 2–3 and days 4–5, correlations between mean of 11 MARDs (Mean-MARD) and mean of two 24-h coefficient of variation [CV] (Mean-CV) were analyzed, on personal-CGM and professional-CGM. For ISO measurement, Mean-MARD correlated to Mean-CV on professional-CGM (r = 0.47 P = 0.02); however, Mean-MARD did not correlate to Mean-CV on personal-CGM. For not-ISO measurement, Mean-MARD did not correlate to Mean-CV on personal-CGM and professional-CGM. The results for ISO measurement show that inter-individual variability of accuracy for real-time calibrations may weaken the relationship between calibration accuracy and glycemic variability. The results for not-ISO measurement show that low calibration accuracy may lead to larger inter-individual variability of accuracy for calibrations, weakening the relationship between calibration accuracy and glycemic variability. In this study, professional-CGM calibrated using ISO may have revealed the relationship between calibration accuracy and glycemic variability. Patients using personal-CGM should consider possibility of having low calibration accuracy caused by high glycemic variability, that may be masked by real-time calibration.