The Performance of Dexcom G6 Glucose Monitoring in Critically Ill Patients

This is an investigator-initiated, prospective, observational accuracy-study. The aim of the study is to test the feasibility, safety, and accuracy of a subcutaneous continuous glucose monitoring system (the Dexcom G6 glucose monitoring system) in critically ill patients. A total of 40 adult patients admitted to the intensive care unit requiring intravenous insulin infusion to maintain blood glucose within target range will be enrolled. Subcutaneous glucose readings will be compared with routine arterial blood glucose measurements to determine accuracy.

The primary endpoint of the study is the mean absolute relative difference (MARD) in percent between CGM and arterial glucose values. Calculations show that 40 patients are required for a 95% confidence interval for the mean to have a range of ±1.7%, meaning that MARD can be estimated with high precision. Calculations of confidence intervals were done using normal approximation, assuming an SD of 5.5 percentage units. The following variables will be collected from the electronic medical record (Take Care), from the ICU-specific patient data management system (Clinisoft) and/or from the handheld Dexcom G6 monitor: Patient-related variables: Age, sex, height and weight Comorbidities (including diabetes status) Chronic medications Clinical variables: Date and time of ICU admission and discharge Reason for ICU admission Admission source (e.g. emergency department, other hospital, ward) Illness severity scores (e.g. SAPS [Simplified Acute Physiology Score], SOFA [Sequential Organ Failure Assessment]) Hemodynamic variables (e.g. heart rate, blood pressure) Blood gas results (including blood glucose concentration) Other routine laboratory results (e.g. serum creatinine, serum albumin, haematocrit) ICU mortality Treatment variables: Insulin doses (including continuous intravenous infusion rates, intravenous and subcutaneous insulin boluses) and times Doses and administration time for other medications given in the ICU (e.g. Vasopressors, paracetamol) Continuous Glucose Monitoring (CGM)-related variables: Date and time of sensor insertion Sensor insertion site Number and duration of disconnection episodes Reason for disconnection Date and time of sensor calibrations Sensor glucose values with date and time stamps Date and time of sensor removal Reason for sensor removal Complications at sensor insertion site (redness, swelling, infection, bruising) Exact date and time of sensor insertion will be manually recorded in a case report form at the bedside. Every time an arterial blood gas is obtained, the arterial blood glucose value and the corresponding CGM glucose value will be manually recorded along with date and time in the case report form. At the end of the study period, after removing the CGM sensor, CGM-data will be downloaded from the handheld Dexcom G6 monitor.

Patients admitted to the intensive care unit requiring insulin infusion to maintain blood glucose within target range

From date of sensor insertion until date of ICU discharge, date of removal of the arterial line, or date of death, whichever came first, assessed up to 10 days

From date of sensor insertion until date of ICU discharge, date of removal of the arterial line, or date of death, whichever came first, assessed up to 10 days

Numerical accuracy between subcutaneous sensor glucose values and arterial blood glucose values determined according to the International Organization for Standardization (ISO) criteria from 2013 (ISO 15197:2013)

From date of sensor insertion until date of ICU discharge, date of removal of the arterial line, or date of death, whichever came first, assessed up to 10 days

Numerical accuracy between subcutaneous sensor glucose values and arterial blood glucose values determined according to the Clinical and Laboratory Standards Institute (CLSI) Point of Care Testing 12-A3 (POCT12-A3) standard

From date of sensor insertion until date of ICU discharge, date of removal of the arterial line, or date of death, whichever came first, assessed up to 10 days

Clinical accuracy between subcutaneous sensor glucose values and arterial blood glucose values determined by Clarke error grid analysis

From date of sensor insertion until date of ICU discharge, date of removal of the arterial line, or date of death, whichever came first, assessed up to 10 days

Clinical accuracy between subcutaneous sensor glucose values and arterial blood glucose values determined by surveillance error grid analysis

From date of sensor insertion until date of ICU discharge, date of removal of the arterial line, or date of death, whichever came first, assessed up to 10 days

From date of sensor insertion until date of ICU discharge, date of removal of the arterial line, or date of death, whichever came first, assessed up to 10 days

From date of sensor insertion until date of ICU discharge, date of removal of the arterial line, or date of death, whichever came first, assessed up to 10 days

From date of sensor insertion until date of ICU discharge, date of removal of the arterial line, or date of death, whichever came first, assessed up to 10 days

Local reactions (e.g. allergic skin reactions, bruising) related to sensor insertion/sensor adhesives

From date of sensor insertion until date of ICU discharge, date of removal of the arterial line, or date of death, whichever came first, assessed up to 10 days

Inclusion Criteria: Age ≥18 years Admitted to ICU at the Karolinska University Hospital Solna Indwelling arterial catheter in situ or expected to be inserted in the next 2 hours Six hours or less since intravenous insulin infusion was commenced or intravenous insulin infusion expected to commence within the next 2 hours Vasopressor infusion ongoing or expected to commence within the next 2 hours Mechanical ventilation ongoing or expected to commence within the next 2 hours Patient expected to stay in the ICU until the day after tomorrow Exclusion Criteria: Pregnancy Unable to get consent from patient or next-of-kin Patients in whom death is considered imminent (within 24 hours)

Matuleviciene V, Joseph JI, Andelin M, Hirsch IB, Attvall S, Pivodic A, Dahlqvist S, Klonoff D, Haraldsson B, Lind M. A clinical trial of the accuracy and treatment experience of the Dexcom G4 sensor (Dexcom G4 system) and Enlite sensor (guardian REAL-time system) tested simultaneously in ambulatory patients with type 1 diabetes. Diabetes Technol Ther. 2014 Nov;16(11):759-67. doi: 10.1089/dia.2014.0238. Epub 2014 Sep 18.