Continuous interstitial glucose monitoring in diabetic and non-diabetic critically ill patients is simple and accurate: comparison with venous, arterial and capillary glucose measurements

Introduction: To reduce mortality, thigh glycemic control is recommended in critically ill patients due to their extreme glycemic variability. Continuous glucose monitoring (CGM) devices allows frequent determination of blood glucose levels; however, conflicting results have been reported from studies assessing their accuracy in critically ill patients. Aim of this study was to assess the repeatability and the analytical and clinical accuracy of FreeStyle Libre 2 (FSL-)CGM. Materials and methods: Prospective single-center observational study enrolling 40 critically ill patients. For four consecutive days, we measured three consecutive interstitial FSL-CGM-derived glucose levels, along with one arterial and venous blood gas analysis and a capillary-derived blood glucose level, obtaining a total of 480 FSL-CGM-derived glucose measurements and 160 measurements from arterial and venous blood gas analysis and from capillary glucose. Results: The mean blood glucose levels in the three daily timepoints from FSL-CGM were 130 ± 35, 131 ± 35 and 131 ± 35 mg/dL (p = 0.660). The Bland-Altman analysis comparing arterial BGA- and FSL-CGM-derived blood glucose levels had a bias of 10.3 mg/dL with limits of agreement from − 27.2 to 47.7. The mean absolute relative difference (MARD) between FSL-CGM and arterial blood gas analysis was 12 ± 10%. The Clarke, Parkes and Surveillance error grid analyses comparing arterial BGA- and FSL-CGM-derived blood glucose levels showed a good clinical accuracy. The presence of diabetes did not influence analytical accuracy, while the use of vasopressors was associated with a higher MARD. Conclusions: FSL-CGM demonstrated reproducibility and reliable analytical and clinical accuracy in critically ill patients, without difference between diabetic and non-diabetic patients, over a period of up to 96 h (4 days).