Continuous Glucose Monitoring in Hospitalized Patients: A Systematic Meta-analysis and Review

Background: Hyperglycemia and hypoglycemia is associated with negative outcomes in hospitalized patients with diabetes mellitus (DM). Novel technologies that continuously measure glucose have emerged, but are only scarcely implemented. Continuous glucose monitoring (CGM) is still not deemed safe and reliable by clinicians for in-hospital use. Aim: The aim of this systematic meta-analysis and review was to assess safety and accuracy of the use of CGM in hospital settings. Methodology: A systematic search was conducted in PubMed, Embase, and the Cochrane Library for full-text studies evaluating safety and accuracy of CGM devices in hospitalized adult patients. Furthermore, a statistical analysis and meta-regression was conducted on continuous data (mean glucose levels) and dichotomous data (hypoglycemic events) extracted from the included studies. Results: 49 studies were included evaluating subcutaneous CGM (SC-CGM), microdialysis CGM (MD-CGM), intravascular CGM (IV-CGM), transdermal CGM (TD-CGM), and closed-loop systems in hospitalized patients. 18 studies presented adequate data for a statistical meta-analysis. SC-CGM (Chi² = 4.04, (P = 0.77); I² = 0%) and closed-loop CGM (Chi² = 1.64, (P = 0.20); I² = 39%) showed no significant heterogeneity and significant effect (P<0.00001) and (P=0.02), respectively, compared to conventional glucose measurements. When comparing dichotomous data, the overall risk of detecting hypoglycemic events was increased in the intervention group RR: 1.86 [0.95, 3.65] P=0.07. Conclusions: SC-CGM and MD-CGM appear less safe and accurate compared to conventional glucose measurement, while the opposite is true for TD-CGM, IV-CGM and closed-loop CGM systems. CGM has the potential to decrease glycemic fluctuations and prevent episodes of hypoglycemia, but more studies on safety and accuracy are needed. Disclosure N. Hashemi: None. K. Houlind: None. N. Ejskjaer: None.