Glucose control is an important part of supportive care for critically ill patients. Achieving optimal glucose control in such situations is challenging due to frequent fluctuations in blood glucose levels. These changes are often difficult to detect because the monitoring procedures are complex and require significant staff involvement, frequent blood draws, and consequent blood loss. Continuous glucose monitoring (CGM) is a simple and minimally invasive technique that has been approved and increasingly used by people with diabetes mellitus. However, its effectiveness in terms of glucose control management and accuracy in conditions with severe organ dysfunction has not been established.

The goal of this study is to assess the performance of CGM-guided glucose control in comparison to the standard glucose monitoring procedure. Additionally, the accuracy of CGM measurements under critical conditions will be evaluated against the standard of care.

This study will be an investigator-initiated, non-commercial, prospective, single-center, parallel-group, randomized controlled trial. Critically ill patients with hyperglycemia requiring intravenous insulin therapy will be randomly assigned to two groups: an intervention group that will receive intravenous insulin therapy aided by continuous glucose monitoring (CGM) measurements and a control group that will receive intravenous insulin therapy guided by arterial blood glucose measurements (RadiometerABL800). Patients will be enrolled within 48 hours after ICU admission. Intravenous insulin dosing will be adjusted according to the in-house glycaemic management protocol. After enrollment, patients will be monitored for maximal 10 days (duration of the sensor) or until stopping intravenous insulin therapy, ICU discharge or death, whichever occurs first, if these events happen before the sensor duration ends.

The primary outcome of the study will be the proportion of time spent within the target range of 7.8-10.0 mmol/L. Secondary outcomes will include mean glucose levels and other CGM metrics, daily vasoactive-inotropic score calculation, hospital length of stay, hospital-acquired infections, and acute renal failure.

Additionally, an accuracy analysis in extreme clinical conditions (pH < 7.20, post-resuscitation, ECMO support, severe haemodynamic instability, hypoxia) will be performed by comparing CGM measurements measured by DexcomG7 with arterial blood glucose measurements (RadiometerABL800) from the electronic health record. Satisfaction of health-care personnel will be evaluated. Sensor-related complications will be monitored.

Inclusion Criteria:

• age ≥ 18 years
• admission to the level 3 ICU
• two consecutive blood glucose measurements > 10.0 mmol/L
• intravenous insulin therapy

Exclusion Criteria:

• expected ICU stay < 48 hours
• pregnancy
• type 1 diabetes
• diabetic emergencies (DKA, DAHS)
• severe skin disease
• severe neutropenia (< 0.5 × 10^9/L)
• severe coagulopathy (thrombocytes < 20 × 10^9/L)
• manufacturer-defined conditions (hydroxyurea use, acetaminophen more than 4 g daily)

All IPD that underlie results in a publication will be shared.

Start date: 1 month after publication End date: no end date

IPD and supporting information will be stored and shared in the Repository of the University of Ljubljana, Slovenia. Access will be provided on request by the Repository.