Accuracy and Reliability of Intermittent Scanning and Real-Time Continuous Glucose Monitoring Systems in Diabetes Emergencies

Background:Diabetes care is a major challenge of patients treated in hospitals. A continuous glucose monitoring system (CGM) provides a more comprehensive assessment of glucose control than capillary blood glucose measurements. Especially in emergencies, data on CGM use in inpatients are limited. To evaluate real-world usability, accuracy of an intermittent scanning and a real-time CGM in patients admitted due to diabetes emergencies was assessed.Methods:In 151 patients admitted due to diabetes emergencies, this single-center prospective study investigated the mean absolute relative difference (MARD) in broad glycemic ranges. The CGM accuracy was evaluated by applying a modified version of the Food and Drug Administration (FDA) criteria for CGM use, Clark Error Grid (CEG), and Bland Altman analysis (BAA).Results:Analysis of 1,498 CGM-/POC-glucose (CGM-/POC-G) pairs revealed a MARD of 10.8% with stepwise improvement from the hypoglycemic to the hyperglycemic range. The CEG analysis showed that 99.1% of all glucose values fell within the optimal or acceptable zones. BAA indicated that 96.0% of CGM-G values fell within the limits of the POC-G values. Day-by-day analysis of overall MARD revealed the highest MARD on the first day of CGM use, followed by consistent and stable MARD levels from day 2 onward until the end of wear time. Applying a modified version of the %20/20 agreement rate of the FDA criteria, 90.7% of CGM-G laid within ±20 mg/dl/±20% agreement rule.Conclusion:This study indicates the usability of CGM for inpatient diabetes care by demonstrating a high accuracy and reliability of CGM measurement.