Objective: To evaluate the accuracy and safety of continuous glucose monitoring system (CGMS) in subjects with normal glucose tolerance (NGT), impaired glucose regulation (IGR) and newly-diagnosed type 2 diabetes mellitus (T2DM).
Methods: A total of 162 subjects (53 NGT, 53 IGR and 56 newly diagnosed T2DM) at 22 - 72 years old were recruited. Each subject received a continuous glucose monitoring (CGM) of CGMS SYSTEM GOLD(TM) (Medtronic Minimed, Northridge, CA) for 3 consecutive days and was instructed to self-calibrate the interstitial glucose levels with finger-stick blood glucose measurements (self-monitoring of blood glucose, SMBG) 7 times daily. Means of median absolute difference (median AD) and median absolute relative difference (median RAD) were calculated to assess the difference between CGM and SMBG values. The correlation between CGM and SMBG values were analyzed with the liner regression analysis. The data were analyzed by the ISO criteria for home glucose meters and Clarke error grid analysis.
Results: All participants showed a good tolerance to the insertion of CGMS sensor and wearing the device. The mean duration of CGMS recording was (75.6 ± 8.3) h. A total of 4324 glucose meter values were paired with glucose measurements from CGMS. Overall, a good relationship (r = 0.943, P = 0.000) and no significant difference (P > 0.05) over a wide range (2.2 - 22.2 mmol/L) were found between CGM and SMBG values. The median AD was 0.5 (0.2 - 0.9) mmol/L and the median RAD 7.55% (3.33% - 14.09%) with 87.16% of sensor values meeting the ISO home glucose meter criteria. The Clarke error grid analysis showed that 99.58% of the readings from CGMS fell into the clinical acceptable zones including 87.5% values in zone A (clinical exact) and 12.1% values in zone B (benign error).
Conclusion: Both safe and well-tolerated, continuous glucose monitoring provides comparatively accurate blood glucose values to guide the diagnosis and treatment of diabetes.