Sage Journals: Discover world-class research

Abstract

Traditionally, the care of critically ill patients with diabetes or stress hyperglycemia in the intensive care unit (ICU) demands the use of continuous intravenous insulin (CII) therapy to achieve narrow glycemic targets. To reduce the risk of iatrogenic hypoglycemia and to achieve glycemic targets during CII, healthcare providers (HCP) rely on hourly point-of-care (POC) arterial or capillary glucose tests obtained with glucose monitors. The burden of this approach, however, was evident during the beginning of the pandemic when the immediate reduction in close contact interactions between HCP and patients with COVID-19 was necessary to avoid potentially life-threatening exposures. Taking advantage of the advancements in current diabetes technologies, including continuous glucose monitoring (CGM) devices integrated with digital health tools for remote monitoring, HCP implemented novel protocols in the ICU to care for patients with COVID-19 and hyperglycemia. We provide an overview of research conducted in the ICU setting with the use of initial CGM technology to current devices and summarize our recent experience in the ICU.

Keywords

CGM intensive care unit critically ill continuous glucose monitoring inpatient hospital

Introduction

Improving glycemic control is paramount to prevent complications in the intensive care unit (ICU).¹ To achieve glycemic targets, the use of continuous intravenous insulin (CII) therapy is generally indicated in the ICU setting, as it allows for rapid adjustments of insulin doses to keep glucose levels within a narrow target. This approach is efficacious but demands a high burden for staff as it usually requires hourly testing with point-of-care (POC) glucose values to make therapeutic adjustments.² Despite close monitoring, the risk of iatrogenic hypoglycemia is still high and similar to the increased risk for poor outcomes observed with hyperglycemia, the incidence of hypoglycemic events is also associated with high morbidity and mortality.^3,4 This association has been observed in ICU and non-ICU settings.⁵ Hence, different strategies have been investigated to monitor glucose levels closely to guide insulin therapy decisions and maintain acceptable glucose ranges.

The first commercially available continuous glucose monitoring (CGM) devices were introduced in the 1990’s (Medtronic MiniMed, Northridge, CA),⁶ and soon after their commercialization the initial experience with CGM in the ICU was published.⁷ Efforts investigating CGM in the ICU are still ongoing, but the acceleration in the use of the technology in clinical practice emerged as a response to COVID-19 when the Food and Drug Administration (FDA) indicated no objection to the use of these devices in the hospital.

We present here a review of the literature related to the use of CGM to care for patients admitted to critical care units and provide an overview of the accuracy data, implementation approaches including remote monitoring, as well as a discussion of the potential barriers for implementation.

Glucose Monitoring and Glycemic Targets in the ICU

Glycemic targets for ICU patients have been debated for the last 2 decades. The initial findings by Dr. Van den Berghe and her colleagues from the University Hospital of Leuven in Belgium that showed a reduction in mortality in surgical ICU patients with intensive insulin treatment revolutionized critical care practice. The authors reported an impressive survival advantage (32% reduction, after adjustment for repeated interim analyses) among patients treated to a euglycemia target (BG 80-110 mg/dL (4.4-6.1 mmol/L), compared to those treated to a conventional target (CII started for BG >215 mg/dL (>12 mmol/L) with a CII target of 180-200 mg/dL [10.0-11.1 mmol/L]). The incidence of severe hypoglycemia (<40 mg/dL [<2.2 mmol/L]) was 5% (39/765) with intensive insulin therapy vs <1% (6/783) with conventional control.¹

The findings from the single-center Leuven trial motivated the design of additional studies aiming at reproducing such findings, with some studies stopped due to unacceptable risk of hypoglycemia. The NICE SUGAR trial, the largest multinational trial (enrolling mostly medical ICU patients) comparing intensive insulin therapy (target: 81 to 108 mg/dL [4.5 to 6.0 mmol/L]) vs a conventional target (<180 mg/dL [10 mmol/L]) showed, however, an unexpected increase in mortality with intensive insulin therapy. The treatment effect difference in mean glucose was narrower in this study (115 ± 18 vs. 144 ± 23 mg/dL [6.4 ± 1 vs. 8 ± 1.2 mmol/L]) and the incidence of severe hypoglycemia (<40 mg/dL (2.2 mmol/L), was slightly higher (6.8% vs 0.5%) than the Leuven trial. Differences in targets, types of infusion-pumps, sampling sites (capillary vs intravenous), accuracies of glucometers, and nutritional strategies were suggested as potential explanations for the discrepancy between NICE SUGAR and the Leuven trial results.³ Since the NICE SUGAR trial, there has been a general agreement in targeting glucose levels between 140 and 180 mg/dL (7.8-10 mmol/L) in the ICU with consideration of lower targets (110-140 mg/dL [6.1-7.8 mmol/L]) for surgical patients. There is, however, lack of support for “euglycemia targets” (BG <110 mg/dL [6.1 mmol/L]).⁸ The Critical Care Society recommends [quality of evidence: very low] that a BG ≥ 150 mg/dL (8.3 mmol/L) should trigger initiation of insulin therapy, titrated to keep BG < 150 mg/dL (8.3 mmol/L) for most adults ICU patients and to maintain BG values absolutely <180 mg/dL (10 mmol/L) using a protocol that achieves a low rate of hypoglycemia (BG ≤ 70 mg/dL [3.9 mmol/L]).⁹

The elimination of iatrogenic hypoglycemia as a confounding factor of the potential treatment effect of intensive glucose control remains an important gap in knowledge where the use of technology may have a role. The availability of glucose levels every 1 to 5 minutes, data trends, algorithms to assess the rate of change in glucose values, alarms for prevention of hypoglycemia, and remote monitoring provide an unprecedented opportunity to design and examine protocols that could finally determine the best treatment targets in the ICU.

In 2010, Holzinger et al reported a significant reduction of severe hypoglycemia (1.6 vs. 11.5%) in critically ill patients (N = 124) using the Real-Time CGM device (discontinued) vs standard of care. Continuous glucose monitoring data were used at least every 2 hours to guide insulin infusion adjustments with glucose target goal of 80 to 110 mg/dL (4.4-6.1 mmol/L).¹⁰ In addition, circulatory shock requiring norepinephrine therapy had no influence on the accuracy and reliability of the device.¹¹ In 2014, Boom et al tested the safety, efficacy, workload, and costs of using the FreeStyle Navigator device (discontinued) compared to standard of care. The investigators used a computerized algorithm for insulin dosing based on CGM data in the intervention group, targeting glucose levels between 90 and 162 mg/dL (5-10 mmol/L). Alarms were set at 90 mg/dL (5 mmol/L) and hypoglycemia defined as <40 mg/dL (2.2 mmol/L), and values were entered in the algorithm for insulin adjustment. The system required multiple calibrations during the day. The authors reported non-inferiority in glycemic control using CGM vs POC, with a potential reduction in nursing workload and costs related to glycemic control. No episodes of severe hypoglycemia were observed in either group, likely due to the higher target range and the use of a computerized algorithm.¹²

Agus et al, randomized 980 pediatric patients to intensive (target 80-110 mg/dL [4.4-6.1 mmol/L]) versus standard care to assess the impact of glycemic control on clinical outcomes. The authors used CGM (Guardian REAL-TIME, discontinued) to prevent and detect severe hypoglycemic episodes early, but no decisions regarding insulin dosing or glucose rescue were made based on CGM.¹³ The incidence of severe hypoglycemia (<40 mg/dL [2.2 mmol/L]) was 3% in the intensive group (16 of 490 patients) compared with 1% in the standard-care (5 of 490 patients) (P = .03). The overall rate of hypoglycemia (<60 mg/dL [3.3 mmol/L]) was 19% versus 9% (P < .001).¹³ No differences in health-care-related infections (primary endpoint), mortality, length of stay, or organ failure, were observed. Although the relevance of achieved glucose targets (and difference between groups in a narrow range in patients without diabetes) is uncertain, the findings of a sub-analysis of this trial suggest severe hypoglycemia may be prevented with protocols using alarms based on the rate of glucose change, or can be treated early to reduce their duration, particularly when using higher alarm thresholds.¹⁴ The multiple calibrations required during the protocol make the specific approach less attractive for inpatient use. An additional randomized control trial (RCT), by Agus et al, using CGM (G4 Platinum, Guardian REALTime, and Enlite) to signal impending hypoglycemia, was stopped due to futility and potential harm with a lower glycemic target, with an unexpected increase in infectious outcomes and higher risk of severe hypoglycemia (5.2% vs 2.%, P = .03).¹⁵

More recently, Beardsall et al, compared the use of real-time CGM to optimize nutritional delivery (dextrose infusion) and guidance on using insulin delivery in preterm infants to maintain glucose levels between 2.6 and 10 mmol/L (46.8-180 mg/dL) vs standard of care. Three endpoints of equal importance were examined (time in the range between 2.6 and 10 mmol/L (46.8-180 mg/dL) and between 4 and 8 mmol/L (72-144 mg/dL), and mean sensor glucose concentration). Overall, CGM use reduced exposure to prolonged or severe hyperglycemia and hypoglycemia in this pediatric population.¹⁶

In summary, studies using CGM appear to show the use of this technology can reduce the incidence of severe hypoglycemia when targeting lower glucose levels and may be associated with reduced costs. While we wait for additional clinical trial data (with a clinically relevant separation in glycemic targets), it appears reasonable to maintain a glucose target of 140 to 180 mg/dL (7.8-10 mmol/L) for most patients (for insulin adjustment) and to use time-in range as a metric to evaluate overall glycemic control. Considering higher thresholds for hypoglycemia prevention (eg, 85 mg/dL (4.7 mmol/L) in the non-ICU¹⁷ and perhaps at <100 mg/dL [5.5 mmol/L] in the ICU)^18
-20 may be more appropriate as there is not clear evidence yet that targeting “euglycemia” improves clinical outcomes in children or adults.

Previous and Current CGM Technologies

In 1999, the FDA approved the first CGM system for commercialization by Medtronic.²¹ Today, there is a wide variety of CGM devices available in the market. There are intravascular sensors (IV-CGM), subcutaneous (SC-CGM), and transdermal CGM sensors.

In the United States, the use of CGM systems in the ICU is infrequent. The GlucoScout and OptiScanner 5000 are IV-CGM approved by FDA but are not commonly used. None of the SC-CGM have yet been approved for inpatient monitoring in the United States. Currently, the only CGM device clinically approved in Europe, for use in the ICU is the IVBG System. OptiScanner 5000 and Sentrino are also approved but for use in non-critically ill patients in Europe.²²

Newer devices have innovative features that make them more reliable, accurate, and easier to use compared with older devices. A non-exhaustive list of CGM devices examined in the last 2 decades is included in Table 1. Older CGM devices needed frequent calibration with fingerstick glucose measurements to have accurate readings. In recent years, newer CGM systems have been paired with smartphones via mobile apps that facilitate glucose monitoring, visualization of current trends, and arrows that depict the rate of glucose change, access summarized information, and data sharing.^23,24 Table 2 lists characteristics of currently available subcutaneous CGM devices and their limitations.

Table 1.

CGM Devices, Manufacturers, Regulatory Status, and Availability.

Type of sensor	Device	Manufacturer	Year	Availability	FDA approval
Intravascular sensor	Eirus System	Maquet Getinge Group—Rastatt, Germany	2013	Discontinued	No
	GluCath	Medtronic Minimed—Northridge, CA, USA	2004	Discontinued	No
	GlucoClear	Edwards Lifescience—Irvine, CA, USA	2009	Discontinued	No
	IVBG System	Edwards Lifescience—Irvine, CA, USA	2009	Discontinued
	GlySure	GlySure—Abingdon, Oxfordshire, UK	2012	Discontinued	No
	GlucoScout	International Biomedical—Autin, TX, USA	2014	Available	Yes
	Optiscanner 5000	OptiScan Biomedical Corporation—Hayward, CA, USA	2017	Available	Yes
	CMI system	Cascade Metrix Inc—Fishers, IN, USA	2015	Discontinued	No
	GlucoSet	GlucoSet—Trondhjem, Sor-Trondelag, Norway	2011	Discontinued	No
	Diramo System	Flowsion—Nordborg, Syddanmark, Denmark.	2009	Discontinued	No
	MicroEye	Probe Scientific—Bedford, England	2012	Discontinued	No
Subcutaneous sensor	FreeStyle Navigator	Abbott Diabetes Care—Chicago, IL, USA	2008	Discontinued	Yes
	FreeStyle Libre 14 days		2014	Discontinued	Yes
	FreeStyle Libre 2		2020	Available	Yes
	FreeStyle Libre 3		2022	Available	Yes
	CGMS System Gold	Medtronic Minimed—Northbridge, CA, USA	1999	Discontinued	Yes
	Guardian REAL—TIME		2006	Discontinued	Yes
	Sentrino		2012	Discontinued	Yes
	Enlite		2013	Available	Yes
	Guardian Sensor 3		2016	Available	Yes
	Guardian Sensor 4		2020	Available in Europe	No
	DGMS	San MediTech—Hozhou, Zhejiang, China	2012	Discontinued	No
	GlucoDay	A. Menarini Diagnostics—Florence, Italy	2010	Available in Europe	No
	GlucoMen Day	A. Menarini Diagnostics—Florence, Italy	2020	Available in Europe	No
	G4	Dexcom—San Diego, CA, USA	2012	Discontinued	Yes
	G5		2015	Discontinued	Yes
	G6		2018	Available	Yes
	G7		2023	Available	Yes
	Eversense Implantable CGM	Senseonic, Inc.—Germantown, MD, USA	2022	Available	Yes
Transdermal sensors	Symphony	Echo Therapeutics—Philadelphia, PA, USA	2012	Discontinued	No

Abbreviations: CGM, continuous glucose monitor; FDA, Food and Drug Administration, IVBG, Intravenous blood glucose; CMI, Cascade Metrix, Inc; CGMS, Continuous Glucose Monitoring System; DGMS, Dynamic Glucose Monitoring System.

Table 2.

Characteristics of Currently Available Subcutaneous CGM Devices and Their Limitations.

Hallmarks	Freestyle Libre 2	Freestyle Libre 3	G6	G7	MiniMed Guardian Sensor 3	MiniMed Guardian Sensor 4	Eversense E3
Sensor wear	14 days	14 days	10 days	10 days	7 days	7 days	180 days
Glucose range	40-400 mg/dL (2.2-22.2 mmol/L)	40-400 mg/dL (2.2-22.2 mmol/L)	40-400 mg/dL (2.2-22.2 mmol/L)	40-400 mg/dL (2.2-22.2 mmol/L)	40-400 mg/dL (2.2-22.2 mmol/L)	40-400 mg/dL (2.2-22.2 mmol/L)	40-400 mg/dL (2.2-22.2 mmol/L)
Warm-up	1 h	1 h	2 h	30 min	2 h	2 h	24 h
Sensor placement	Back of the upper arm	Back of the upper arm	Abdomen	Abdomen Back of the upper arm (>6 yo)Upper buttock (2-6 yo)	Upper arm	Upper buttock and back of upper arm (7-17 yo). Upper arm and abdomen (>18 yo)	Implantable Upper arm
Apps/software	FreeStyle Libre 2 appLibreView	FreeStyle Libre 3 appLibreView	Dexcom G6 AppClarityFollow App	Dexcom G7 AppClarityFollow App	Guardian Connect AppCareLink	Guardian Connect App CareLink	Eversense App
Calibration	Factory calibrated	Factory calibrated	Factory calibrated	Factory calibrated	Yes, 3/day	Factory calibrated	Yes, 2/day for the first 21 days, after 1/day
Alarms	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Monitoring	Flash glucose monitoring (on-demand)	Continuously (every min)	Continuously (every 5 min)	Continuously (every 5 min)	Continuously(every 5 min)	Continuously (every 5 min)	Continuously (every 5 min)
Substance inferences	>500 mg Ascorbic acid (falsely raise sensor readings)	>500 mg Ascorbic acid (falsely raise sensor readings)	Hydroxyurea,Acetaminophen (>4 g/day)	HydroxyureaAcetaminophen (>4 g/day)	Acetaminophen	Acetaminophen	MannitolTetracycline
Contraindications	CT scan, X-ray, or diathermy treatment.	CT scan, X-ray, or diathermy treatment.	CT scan, X-ray, or diathermy treatment, MRI^a	CT scan, X-ray, or diathermy treatment, MRI	CT scan, X-ray, or diathermy treatment, MRI.	CT scan, X-ray or diathermy treatment, MRI	MRI Compatibility—remove transmitter before procedure.
Tested in the ICU	No	No	Yes^18,19,25	No	Yes^25,26	No	No

Abbreviations: CGM, continuous glucose monitor; CT, computed tomography; ICU, intensive care unit; MRI, magnetic resonance imaging.

No interference observed after X-ray or CT scans in 1 report.²⁷

Accuracy

The accuracy and reliability of CGM in the critical care setting has been a focus of many CGM studies. Understanding the functionality and limitations of CGM during the care of critically ill patients is a priority, as misleading glucose values could lead to inadequate or inappropriate treatment and potentially devastating consequences. Previous studies evaluating CGM use in ICU populations have provided information on accuracy metrics, most commonly reporting the mean or median absolute relative difference (MARD or MedARD) between sensor and reference glucose values, as well as error grid data (ie, Clarke or surveillance) for assessment of clinical reliability. Although there have been a small number of studies evaluating the accuracy and feasibility of IV-CGM,^28
-31 most studies have evaluated SC-CGM devices given higher rates of complications (eg, thrombus formation, infection) associated with IV-CGM use.³²

Previous studies using subcutaneous CGM devices have shown differences in accuracy, with reported MARD or MedARD values ranging from 6.6% to 15.6%.^12,33,34 These differences may be explained in part by variations in clinical factors, frequency of CGM calibration and type of reference glucose measurement used for comparison. Despite the variability in accuracy metrics, error grid analyses in many of these studies indicated good correlation between sensor and reference glucose pairs with ≥98% of values falling within acceptable treatment zones for clinical management.^11,33 Currently, definitive accuracy studies comparing a gold standard method with available SC-CGM devices have not yet been completed. However, utilization of newer CGM devices for clinical care during the COVID-19 pandemic has provided additional accuracy data and further insight into the feasibility of CGM use in critical illness. Accuracy analyses using the Dexcom G6 in critically ill patients with COVID-19 reported MARD values of 11.1%¹⁹ and 12.6%,¹⁹ again with ≥98% of paired glucose values falling in Clarke Error Grid zones A and B.^19,25 Although recent data continue to support the potential for CGM implementation in ICU practice, it highlights the need for ongoing assessment of sensor function in a setting where acute clinical changes may impact device reliability.

Randomized Clinical Trials Using CGM in the ICU

We reviewed systematically PubMed to identify randomized clinical trials testing the impact of CGM in the ICU on glycemic metrics. Using the following search strategy: ((((continuous glucose monitoring) OR (CGM)) AND (((critically ill) OR (intensive care unit)) OR (ICU))) AND ((((hyperglycemia) OR (glycemic control)) OR (hypoglycemia)) OR (diabetes))) AND (randomized clinical trial) we identified 54 articles (last search conducted on September 25, 2022). Out of these 54 articles, 16 were confirmed to be randomized clinical trials using GCM in the ICU. In general, studies mostly showed non-inferiority compared to standard monitoring, with several studies showing a reduction in the incidence of hypoglycemia. A summary of these studies is listed in Table 3. Additional descriptions of key studies are discussed above in the Glucose Monitoring and Glycemic Targets in the ICU section.

Table 3.

Clinical Trials Testing the Impact of CGM on Glycemic or Clinical Endpoints in the ICU.

Author	ICU population	Intervention (device)	Primary endpoint	Results	Comments/limitations
Perri et al³⁵	47 InfantsWith weight <1500 g fed on parenteral nutrition.	Strict glycemic control vs standard glycemic values (Enlite sensor)	Number of severe dysglycemia episodes defined as <2.61 mmol/L (47 mg/dL) >10 mmol/L (180 mg/dL)	Reduced dysglycemia episodes 1 (IQR: 0-2) vs 3 (IQR: 1-7); P = .005 with a NNT to avoid dysglycemia of 3.2 (95% CI: 1.8-16.6) and increased time in range in intervention group.	This study demonstrated that insulin infusion guided by CGM improves glycemic control in preterm infants by reducing the number of dysglycemic events.Small population. Single centered study
Beardsall et al¹⁶	180 InfantsWithin 24 h. of birth, birth weight < 1200 g, (gestational age of 33± 7 days)	RT-CGM vs POC (Guardian 2)	Proportion of time that sensor glucose was in target range (2.6-10 mmol/L [46.8-180 mg/dL]) for the first 7 days of life.	Increased time in range in CGM groupAdjusted mean difference of 8.9% (95% CI: 3.4-14.4), P = .005.	CGM led to prevention and early detection of hypoglycemia / Potential recruitment bias from short recruitment window. Data loss due to technical difficulties with sensor insertion.
Beardsall et al³⁶	21 InfantsWith <1200 g at birth and <48 h of age	CGM + closed loop vs CGM alone + paper algorithm (Enlite)	Percentage of time in range (72-144 mg/dL [4.0-8.0 mmol/L])	Increased time in range in patients with closed loop from 26% (6-64) in control group vs 91% (78-99) in close loop group; P < .001	There were no serious adverse events and no difference in total insulin infused.Fist study RCT to evaluate closed loop control of CGM on preterm infantsLimitations: Small sample size. Short duration (48-72 h). (Pilot).
Lu et al³⁷	144 Adults>18 and <65 years oldICU expected stay >72 hAdmission glucose or 2 random BG > 180 mg/dL (10 mmol/L)	CGM vs POC (DGMS)	Percentage of time in target range (8.0-10.0 mmol/L)	Improve time in range with CGMTIR 29.0% (23.3%-38.0%) in control group vs 51.5% (30.2%-67%) in CGMS group; P < .001.	Severe medical conditions and high proportion of patients with severe acute pancreatitis made glucose control more difficult.
Preiser et al³⁸	77 AdultsExpected ICU stay >3 days, hyperglycemia (BG >150 mg/dL [8.3 mmol/dL]) up to 6 h after admission and/or on CII	CGM vs Intermittent glucose monitoring IGM (GlucoClear)	Proportion of time in range (90-150 mg/dL [5-5.3 mmol/L]) and severe hypoglycemia events	Decreased incidence of severe hypoglycemia in CGM group (39.5% in the control group vs. 20.5% in the CGM group)	Decreased incidence and severity of hypoglycemic events and therefore safer glycemic control was achieved with CGM.No data reported on insulin treatment.
van Steen et al³⁹	107 AdultsAge ≥ 18 yearsExpected ICU stay >48 hIn need of enteral feeding	Low-Carbohydrate Formula (LCF) vs Standard Formula (FreeStyle Navigator II)	Glucose variability defined as mean / median absolute glucose change in mmol/L/hGlucose time in range (6-9 mmol/L [108-162 mg/dL)	Glucose variability (mmol/L/h) in LCF 0.8 (0.7–1.2) vs standard care 0.9 (0.6-1.2). P = .501	All patients received the CGM system. Blinded subcutaneous CGM compared outcomes with glucose measurements of blood gas analysis and POC.Limitations: Single centered study. Enteral formulas were different in nutritional contents.
Agus et al¹⁵	713 Children2 weeks to 17 years of ageReceiving vasoactive support or mechanical ventilation	Tight vs higher glucose target (Dexcom G4, Guardian REAL-Time and Enlite)	Number of ICU free days to day 28 (inverse equivalent of 28-day hospital mortality-adjusted ICU length of stay)	19.4 days (IQR: 0-24.2) in the lower-target group vs 19.4 days (IQR 6.7 - 23.9) in the higher target group. P = .58	CGM and computer-guided insulin adjustment were used to minimize hypoglycemia.CGM was effective in detecting impending hypoglycemia—study stopped early due to low likelihood of benefit and possible harm (risk of severe hypoglycemia)
De Block et al⁴⁰	35 Adults>18 years and < 75 years (66 ± 10)APACHE score ≥ 20 (28 ± 6)	RT CGM vs blinded CGM (GlucoDayS)	Percentage of time in the target range (80-110 mg/dL[4.4-6.1 mmol/L]) of glycemia	% of time in target range 79 ± 7 in RT-CGM vs 91 ± 8 in blinded CGM; not statistically significant	RT CGM did not improved glucose control or variabilityLimitations: Small sample size.
Boom et al¹²	178 AdultsExpected stay in the ICU >24 h and in need of insulin	CGM vs POC to guide insulin treatment (FreeStyle Navigator)	Incidence of severe hypoglycemia (2.2 mmol/L [40 mg/dL])	There was no incidence of severe hypoglycemia events.	In the intervention group the CGM data was used to feed a computerized glucose regulation algorithm.No incidence of hypoglycemia (unable to test differences between groups)
Leelarathna et al⁴¹	24 AdultsWith hyperglycemia (≥10 mmol/L [180 mg/dL]) or on insulin therapy	Closed loop therapy vs local protocol (FreeStyle Navigator)	Percentage of time in range (6.0-8.0 mmol/L [108-144 mg/dL])	Increased time in range in closed-loop therapy group.Time in range in the closed loop therapy was 54.3% (44.1 to 72.8) vs 18.5% (0.1 to 39.9) in the control group, P = .001	The closed loop system consisted of a CGM (FreeStyle Navigators) a computer running a model predictive algorithm and 2 syringe pumps.Limitations: Small sample group. Single centered group. Short duration. Subspecialized patients.
Agus et al¹³	980 Children0 to 36 months undergoing surgery with cardiopulmonary bypass	Tight vs standard glycemic control (Guardian REAL time)	Rate of heath care-associated infection in the cardiac ICU	The rate of health care associated infection was not significantly decreased with thigh glycemic control (8.6 vs 9.9 per 100 patients-day, P = .67)	CGM used as a tool to detect hypoglycemia, (higher in the intervention arm)
Sampaio et al⁴²	20 AdultsWithin 24 h of acute myocardial infarction	Insulin glargine + regular insulin vs standard care -NPH insulin + regular insulin (CGMS Gold, Medtronic)	Glycemic controlPercentage of time in range (80-180 mg/dL [4.4-10 mmol/L])	Percentage of time in range (80-180 mg/dL [4.4-10 mmol/L]) by CGMS was 73%-18% for insulin glargine and 77%-11% for standard care	Therapy was guided by POC, and Blinded CGM was used to analyze glycemic parametersCGM monitoring demonstrated that a significant number of hyper and hypoglycemia events went unnoticed even with frequent POC measurements.Limitations: Small sample size (pilot)
Steil et al¹⁴	311 Children≤ 3 years old undergoing cardiac surgery	Tight glycemic control vs standard care (Guardian REAL time)	Severe hypoglycemia incidence	% of severe hypoglycemia in TGC was 3.2% (95% [CI] 1.2 to 7.5%) with 1 instance of BG < 40 mg/dL [2.2 mmol/L] observed in the STD arm.	CGM RT failed to detect hypoglycemia in a substantial number of patients.
Holzinger et al¹⁰	124 Adults≥18 yearsMechanically ventilated	RT-CGM vs POC (Guardian, Medtronic)	Percentage of time at a glucose level <110 mg/dL (6.1 mmol/L)	The percentage of time <110 mg/dL was 55% ± 18% in the control group vs 59% ± 20% in the intervention group, P = .076	Blinded CGM in the control group.CGM reduced the incidence of hypoglycemic events. With no improvement on glycemic control compared to POC.Limitations: Sample size. Single centered. Short duration of intervention (72 h). The same titration algorithm was used for both groups
Beardsall et al⁴³	389 Infants<24 h of ageVery low birth weight < 1500 g	Early insulin (fixed dose of CII [0.05 U kg/h] + IV 20% dextrose) vs standard care (CGMS Gold)	Mortality at the expended date of delivery	Mortality on the expected day was 9% in the control group vs 14% in the intervention group. OR 0.61 (95% CI 0.33 to 1.15); P = .2Absolute difference of 5.1% (95% CI: –1.4 to 11.5%) P = .16.	Limitations:Study stopped early for concerns about futility regarding the primary outcome and potential harm

Abbreviations: RCT, randomized controlled trial; BG, blood glucose ; CGM, continuous glucose monitor; CI, confidence interval; ICU, intensive care unit; IQR, interquartile range; IV, intravascular; OR, odds ratio; POC, point-of-care; RT, real time; TIR, time in range; APACHE, Acute Physiological and Chronic Health Evaluation; NPH, neutral protamine Hagedorn; CGMS, Continuous Glucose Monitoring System; DGMS, Dynamic Glucose Monitoring System; TGC, tight glycemic control; STD, standard care.

Real World Experience During Covid-19

Early during the COVID-19 pandemic, it was clear that (1) patients with hyperglycemia/diabetes had worse outcomes and (2) there was an urgent need to find alternative strategies to monitor and treat patients with diabetes to avoid the risk of infection of healthcare workers and improve glycemic control. In April 2020, the FDA announced that they would not object to the use of non-invasive remote CGM to aid glucose monitoring during the pandemic.² To reduce HCP exposure to the virus, personal protective equipment (PPE) waste, and reduced staff workload multiple centers implemented the use of remote CGM in the ICU. Several studies emerged to describe the experience gained with this technology while caring for critically ill patients with COVID-19.^18
-20,25,44 Studies consistently reported the feasibility of CGM use in this setting aimed at reducing the burden of critical care nurses caring for patients with COVID-19. Sadhu et al,²⁵ Agarwal et al,¹⁹ Davis et al,¹⁸ and Faulds et al²⁰ (except for Sadhu et al that used G6 and Guardian Connect, all other groups used a G6 CGM device only) reported an estimated reduction of POC testing of 33.1%, 60%, 68%, and 71%, respectively.

The use of remote CGM has emerged as a promising strategy to monitor patients remotely and to guide insulin infusion, particularly in extreme situations as experienced early during the COVID-19 pandemic. Additional research is needed to confirm the safety of such approaches and the impact on additional metrics, including cost-effectiveness, staff burden, and workload.^19,20,25

Implementation of CGM in the ICU, Role of Hybrid Protocols

The COVID-19 pandemic created a need for rapid deployment of traditionally outpatient CGM systems in a novel inpatient environment. While the FDA provisional allowance opened the possibility of using CGM in the inpatient setting, there still existed concerns surrounding safety and implementation. Health systems worked to craft protocols that would help mitigate safety concerns and facilitate CGM use among nursing staff. Hybrid protocols combining intermittent POC glucose testing with CGM were widely used during the COVID-19 pandemic. Most protocols required more stringent initial sensor-POC pair validation before allowing intermittent non-adjunctive CGM (ie, use of CGM to make therapeutic decisions without the need of confirmatory POC testing) with less frequent ongoing POC testing and validation. Such protocols helped ensure a degree of accuracy while still facilitating reduction in required frequency of POC testing. Table 4 describes key elements of COVID-19 CGM protocols with the majority using a threshold of 20% (for glucoses >100 mg/dL [5.5 mmol/L]) or 20 mg/dL (1.1 mmol/L) (for glucose <100 mg/dL [5.5 mmol/L]) for validation.^18
-20,45 Five of the 6 studies described more frequent POC testing after initial placement^{18,20,25,44,45} and in 3 protocols the CGM was in fact not used until 24 hours after placement.^25,44,45 Figure 1 shows a summary of combined experience gained at The Ohio State University and Grady/Emory University with currently available technologies. Using a real-time CGM device, a smartphone with mobile apps can be located outside of the patient room. Data can also be shared at the nursing station and with ICU or diabetes team staff members (via an app for data sharing). Healthcare provider can customize real-time alarms (for low or high glucose or signal loss). Additional steps for implementation have been previously described in more detail.^2,20,46

Table 4.

Hybrid CGM + POC Protocols Used During the COVID-19 Pandemic

Study	Device	Initial validation	Criteria non-adjunctive use	Ongoing validation/POC testing	Alarms
Chow et al⁴⁵	Dexcom G6	CGM not used first 24 h	20%	Every 2 h	<70 mg/dL (3.9 mmol/L) & urgent low soon alert
Davis et al¹⁸	Dexcom G6	2 POC-CGM pairs	20%	Every 6 h	<100 mg/dL (5.5 mmol/L), >300 mg/dL (16.7 mmol/L) & urgent low soon alert
Agarwal et al¹⁹	Dexcom G6	Not described	20%	Minimum daily	<100 mg/dL (5.5 mmol/L), >250 mg/dL (13.9 mmol/L) & urgent low soon alert
Sadhu et al²⁵	Dexcom G6 & Medtronic	CGM not used first 24 h	glucose 100-200 mg/dL (5.5-11.1 mmol/L)	Every 4 h glucose 100-200 mg/dL (5.5-11.1 mmol/L); every 2 h glucose <100 mg/dL (5.5 mmol/L) or >200 mg/dL (11.1 mmol/L)	Not discussed
Longo et al⁴⁴	Dexcom G6	CGM not used first 24 h	<35 mg/dL (1.9 mmol/L)	Minimum daily	Not discussed
Faulds et al²⁰	Dexcom G6	2 POC-CGM pairs	20%	Every 4 h	<100 mg/dL, (5.5 mmol/L) >300 mg/dL (16.7 mmol/L) and urgent low soon alert

Abbreviations: CGM, continuous glucose monitor; COVID-19, coronavirus disease; POC, Point-of-care.

Figure 1.

Implementation of CGM in the ICU Setting. Based on our recent experience, 4 main components are useful for implementation in the ICU including: (a) Devices for real-time CGM used in the ICU. (b) Protocol for hybrid monitoring: before sensors are validated POC test are conducted hourly to guide CII. Once initial sensor readings are available, they can be compared with a simultaneous POC test to determine accuracy and decision to initiate CGM guided insulin titration. A confirmatory validation can be performed 1 hour after the first POC-CGM comparison. Every 4 to 6 hours confirmation of sensor accuracy is recommended. After discontinuation of CII nonadjunctive CGM can be used for insulin management with multiple daily injections (with periodic validation, eg, QACHS or Q6h). (c) Data sharing for real-time remote monitoring (eg, nursing station, primary team) and download of summary data for interpretation (eg, population management platform). (d) EHR documentation and periodic validation. Abbreviations: CGM: Continuous glucose monitoring; CII: continuous intravenous insulin; EHR: Electronic Health Record; ICU, intensive care unit; MDI: multiple daily injections; POC: point of care; Q1hr: every hour; Q6hr: every 6 hours; QACHS: before meals and at bedtime.

Beyond Glycemic Control

The use of CGM certainly holds the potential to improve outcomes beyond glycemic control including reducing nursing burden, increasing nursing satisfaction, improving health system workflows, decreasing patient discomfort, and subsequently improving patient satisfaction. The current standard of care POC glucose testing can be a burdensome task for nurses and nursing support personnel. Boom conducted a time-in-motion analysis and found that each POC test took approximately 3 minutes to perform and the use of CGM significantly reduced nursing workload by a total of 19 minutes in 24 hours. This reduction in nursing personnel cost contributed to a significant reduction in hospital costs associated with CGM compared to traditional POC testing.¹² While there are few studies examining nursing perspectives on the implementation of CGM, Faulds et al found a high degree of acceptance and a high degree of perceived accuracy among nurses²⁰ and Chow et al reported that 63% of nurses believed CGM systems improved the care they provided their patients.⁴⁵ Even though frequency of POC was not significantly reduced overall, due to the nature of the protocol.

While improvement in glycemia associated with CGM could have huge impact on patient clinical outcomes, there are additional potential patient benefits for consideration. Point of care glucose monitoring is a painful procedure for patients and can disrupt a patient’s sleep and recovery especially when performed hourly in association with CII. The longer wear time of a CGM systems means patients may only need to have a sensor insertion performed once over the course of a hospitalization. Moreover, CGM results in minimal blood loss, particularly in comparison to use of arterial or venous samples. These and other outcomes beyond glycemia should be studied and incorporated into clinical trials whose primary emphasis is glycemic control.

Limitations and Barriers to Implementation

The use of CGM in the ICU may assist glycemic control by providing more information compared to intermittent POC. However, there are limitations to this technology that are important to consider, especially in the ICU. Critically ill patients frequently require vasopressors, present with hypoxia or hypothermia, are exposed to multiple medications and intravenous solutions, have variable parental/enteral nutrition support, or require unexpected procedures. Factors that cause a rapid rate of change in glucose may magnify any potential discrepancy between interstitial and blood glucose measurements. Limitations include (1) circumstances that affect accuracy, and (2) implementation barriers, resources, documentation, and costs.

Hypoperfusion may produce falsely low readings and the mechanical compression of the sensor can result in signal loss.¹⁸ However, the accuracy of devices appears to be preserved among patients receiving vasopressors.^11,47 Medications with substances interfering with sensor readings may produce falsely elevated or decreased glucose values (summary in Table 2). Patients admitted to the ICU require multiple radiology procedures, but information about their potential effects on sensor accuracy is limited. In a small observational study, the accuracy of the device after CT and X-rays was preserved.²⁷ According to device manuals specifications, devices need to be removed before magnetic resonance imaging (MRI).⁴⁸ However, Freestyle 2 and 3 maintain functionality and data integrity after MRI, as well as the G6.⁴⁹ During the perioperative period, a few studies have provided good accuracy and no interference with electrocautery/diathermia.⁵⁰ However, for patients undergoing Coronary Artery Bypass Graft (CABG), the CGM devices were associated with intermittent interference and potentially decreased accuracy after the procedure.⁴⁷ Patients’ safety and well-being are priorities. Therefore, further study of this potential limitation is needed.

There are also several barriers to implementation, including the need for approval and collaboration with multiple stakeholders, which may differ by institution. Buy-in from the end-users (nursing staff) is essential to ensure adoption of CGM into the clinical workflow in order to effectively reduce the burden associated with hourly glucose monitoring.⁵¹ In addition, due to high nursing staff turnovers, identifying “nurse champions” that can help new staff with training/troubleshooting will promote continued use of the technology. Developing a training curriculum and additional tools (summary sheets, presentations, videos, remote support) may also facilitate the training process and sustained effective CGM use.

Due to potential known and unknown CGM interferences, documentation of sensor accuracy provides another layer of safety when using this technology in the hospital. It is critical to have procedures in place to track inaccuracy events, decisions for sensor replacement, or discontinuation of CGM therapy. Involvement of information technology services early during the implementation process may help facilitate the implementation of a system for proper tracking of CGM use in the hospital.¹⁸ Recent efforts are focusing on the creation of data standards and workflow guidance for integration of CGM data in the electronic health record (EHR).⁵²

Future Directions

The use of CGM is changing the life of patients living with diabetes. In the outpatient setting, this tool helps to monitor glucose levels without the need for fingersticks, provides trends to predict changes in glucose, alarms for prevention, and detection of severe hypo or hyperglycemia, can promote behavior change, and is widely accepted by users. In the ICU, this technology thus far has been useful to reduce the burden of nurses taking care of ICU patients while allowing the continued use of the standard of care therapy during COVID-19. Studies conducted in ICU populations show this technology may significantly reduce the time spent in hypoglycemia. The use of alarms and visualization of trends can both help prevent the incidence or reduce the time spent in hypoglycemia that already occurred.⁵³ In addition, the information can help monitor the recovery after a hypoglycemic episode and prevent recurrence in high-risk patients.

With the expanding knowledge about the safe and helpful use of technology, we are hopeful the path for regulatory approval will become more efficient. These technologies continue to advance rapidly for outpatient use and translation of each new device is not a viable path for inpatient use of CGM.

The fact that the technology provides continuous data with visualization of trends as well as alarms for hypoglycemia prevention and early detection, suggests the criteria for accuracy should not be as stringent as it currently is for POC. This is also relevant when POC is used as reference, as single POC values may significantly differ from interstitial glucose levels during a rapid rate of glucose change.

Current approaches with hourly fingersticks are not just a significant burden to ICU staff but are suboptimal compared to the wealth of data and mitigation elements (ie, alarms, trends, digital sharing) provided by CGM devices. Once the accuracy of devices designed for the hospital is established for the ICU, electronic health records (EHR) integration and protocols for CGM use (ie, hybrid protocols) are proved to be effective, this technology has the potential to transform ICU glycemic care. The available and emerging data and experience gained with CGM in the ICU suggest this technology should at least be approved as adjunctive to POC in validated hybrid protocols until more definitive technology is available and ready to replace inconvenient hourly fingesticks.

Footnotes

Authors’ Contributions

L.G.A., E.F., and M.C.P.G. prepared the tables and wrote the first draft. G.M.D. and K.D. critically reviewed and edited the manuscript. E.F. and F.J.P. prepared the Figure. F.J.P. conceptualized the approach and wrote and edited different sections of the manuscript. All authors reviewed and approved the manuscript.

Abbreviations

ICU, intensive care unit; HCP, healthcare providers; POC, point-of-care; CGM, continuous glucose monitoring; CII, continuous intravenous insulin; FDA, Food and Drug Administration; MARD or MedARD, mean or median absolute relative difference; IV-CGM, intravascular CGM; SC-CGM, subcutaneous CGM; CABG, coronary artery bypass graft; EHR, electronic health record.

Declaration of Conflicting Interests

The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: EF discloses research support with Dexcom Inc. KD discloses research support from Novo Nordisk, Sanofi, Viacyte, Abbott; consulting with Eli Lilly, Novo Nordisk, Boehringer-Ingelheim, and Tolerion; and honorarium from UptoDate, Medscape, and Elsevier. GD discloses research support from Insulet and honorarium from Medscape. FJP discloses unrestricted research support from Insulet, Dexcom and Ideal Medical Technologies and personal fees from Boehringer Ingelheim, Dexcom, and Medscape for consulting activities.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iDs

Eileen Faulds

Kathleen Dungan

Francisco J. Pasquel

References

den Berghe

Wouters

Weekers

, et al. Intensive insulin therapy in critically ill patients. N Engl J Med. 2001;345(19):1359-1367. doi:10.1056/NEJMoa011300.

Pasquel

Lansang

Dhatariya

Umpierrez

GE.

Management of diabetes and hyperglycaemia in the hospital. Lancet Diabetes Endocrinol. 2021;9(3):174-188. doi:10.1016/s2213-8587(20)30381-8.

NICE-SUGAR Study Investigators; Finfer

Chittock

, et al. Intensive versus conventional glucose control in critically ill patients. N Engl J Med. 2009;360(13):1283-1297. doi:10.1056/NEJMoa0810625.

Egi

Bellomo

Stachowski

, et al. Hypoglycemia and outcome in critically ill patients. Mayo Clin Proc. 2010;85(3):217-224. doi:10.4065/mcp.2009.0394.

Turchin

Matheny

Shubina

Scanlon

Greenwood

Pendergrass

ML.

Hypoglycemia and clinical outcomes in patients with diabetes hospitalized in the general ward. Diabetes Care. 2009;32(7):1153-1157. doi:10.2337/dc08-2127.

Mastrototaro

The MiniMed Continuous Glucose Monitoring System (CGMS). J Pediatr Endocrinol Metab. 1999;12(suppl 3):751-758.

Chee

Fernando

van Heerden

PV.

Closed-loop control of blood glucose levels in critically ill patients. Anaesth Intensive Care. 2002;30(3):295-307. doi:10.1177/0310057X0203000306.

American Diabetes Association. The American Diabetes Association (ADA) has been actively involved in the development and dissemination of diabetes care standards, guidelines, and related documents for many years. Introduction. Diabetes Care. 2009;32(suppl 1):S1-S2. doi:10.2337/dc09-S001.

Jacobi

Bircher

Krinsley

, et al. Guidelines for the use of an insulin infusion for the management of hyperglycemia in critically ill patients. Crit Care Med. 2012;40(12):3251-3276. doi:10.1097/CCM.0b013e3182653269.

10.

Holzinger

Warszawska

Kitzberger

, et al. Real-time continuous glucose monitoring in critically ill patients: a prospective randomized trial. Diabetes Care. 2010;33(3):467-472. doi:10.2337/dc09-1352.

11.

Holzinger

Warszawska

Kitzberger

Herkner

Metnitz

Madl

Impact of shock requiring norepinephrine on the accuracy and reliability of subcutaneous continuous glucose monitoring. Intensive Care Med. 2009;35(8):1383-1389. doi:10.1007/s00134-009-1471-y.

12.

Boom

Sechterberger

Rijkenberg

, et al. Insulin treatment guided by subcutaneous continuous glucose monitoring compared to frequent point-of-care measurement in critically ill patients: a randomized controlled trial. Crit Care. 2014;18(4):453. doi:10.1186/s13054-014-0453-9.

13.

Agus

Steil

Wypij

, et al. Tight glycemic control versus standard care after pediatric cardiac surgery. N Engl J Med. 2012;367(13):1208-1219. doi:10.1056/NEJMoa1206044.

14.

Steil

Langer

Jaeger

Alexander

Gaies

Agus

MS.

Value of continuous glucose monitoring for minimizing severe hypoglycemia during tight glycemic control. Pediatr Crit Care Med. 2011;12(6):643-648. doi:10.1097/PCC.0b013e31821926a5.

15.

Agus

Wypij

Hirshberg

, et al. Tight glycemic control in critically ill children. N Engl J Med. 2017;376(8):729-741. doi:10.1056/NEJMoa1612348.

16.

Beardsall

Thomson

Guy

, et al. Real-time continuous glucose monitoring in preterm infants (REACT): an international, open-label, randomised controlled trial. Lancet Child Adolesc Health. 2021;5(4):265-273. doi:10.1016/S2352-4642(20)30367-9.

17.

Singh

Satyarengga

Marcano

, et al. Reducing inpatient hypoglycemia in the general wards using real-time continuous glucose monitoring: the glucose telemetry system, a randomized clinical trial. Diabetes Care. 2020;43(11):2736-2743. doi:10.2337/dc20-0840.

18.

Davis

Faulds

Walker

, et al. Remote continuous glucose monitoring with a computerized insulin infusion protocol for critically ill patients in a COVID-19 medical ICU: proof of concept. Diabetes Care. 2021;44(4):1055-1058. doi:10.2337/dc20-2085.

19.

Agarwal

Mathew

Davis

, et al. Continuous glucose monitoring in the intensive care unit during the COVID-19 pandemic. Diabetes Care. 2021;44(3):847-849. doi:10.2337/dc20-2219.

20.

Faulds

Boutsicaris

Sumner

, et al. Use of continuous glucose monitor in critically ill COVID-19 patients requiring insulin infusion: an observational study. J Clin Endocrinol Metab. 2021;106(10):e4007-e4016. doi:10.1210/clinem/dgab409.

21.

Clarke

Foster

JR.

A history of blood glucose meters and their role in self-monitoring of diabetes mellitus. Br J Biomed Sci. 2012;69(2):83-93.

22.

Galindo

Umpierrez

Rushakoff

, et al. Continuous glucose monitors and automated insulin dosing systems in the hospital consensus guideline. J Diabetes Sci Technol. 2020;14(6):1035-1064. doi:10.1177/1932296820954163.

23.

Cappon

Vettoretti

Sparacino

Facchinetti

Continuous glucose monitoring sensors for diabetes management: a review of technologies and applications. Diabetes Metab J. 2019;43(4):383-397. doi:10.4093/dmj.2019.0121.

24.

Hester

Zabala

Winskel

Pasquel

. Chap 9: digital health apps for people with diabetes. In: Klonoff

Kerr

Weitzman

, eds. Diabetes Digital Health and Telehealth. New York: Academic Press; 2022:101-108.

25.

Sadhu

Serrano

, et al. Continuous glucose monitoring in critically ill patients with COVID-19: results of an emergent pilot study. J Diabetes Sci Technol. 2020;14(6):1065-1073. doi:10.1177/1932296820964264.

26.

Bridges

Preissig

Maher

Rigby

MR.

Continuous glucose monitors prove highly accurate in critically ill children. Crit Care. 2010;14(5):R176. doi:10.1186/cc9280.

27.

Migdal

Spanakis

Galindo

, et al. Accuracy and precision of continuous glucose monitoring in hospitalized patients undergoing radiology procedures. J Diabetes Sci Technol. 2020;14(6):1135-1136. doi:10.1177/1932296820930038.

28.

Sechterberger

van der Voort

Strasma

DeVries

JH.

Accuracy of intra-arterial and subcutaneous continuous glucose monitoring in postoperative cardiac surgery patients in the ICU. J Diabetes Sci Technol. 2015;9(3):663-667. doi:10.1177/1932296814564993.

29.

Munekage

Yatabe

Sakaguchi

, et al. Comparison of subcutaneous and intravenous continuous glucose monitoring accuracy in an operating room and an intensive care unit. J Artif Organs. 2016;19(2):159-166. doi:10.1007/s10047-015-0877-2.

30.

Schierenbeck

Franco-Cereceda

Liska

Evaluation of a continuous blood glucose monitoring system using central venous microdialysis. J Diabetes Sci Technol. 2012;6(6):1365-1371. doi:10.1177/193229681200600615.

31.

Steen

Rijkenberg

Limpens

van der Voort

Hermanides

DeVries

JH.

The clinical benefits and accuracy of continuous glucose monitoring systems in critically ill patients—a systematic scoping review. Sensors. 2017;17(1):146. doi:10.3390/s17010146.

32.

Strasma

Finfer

Flower

, et al. Of an intravascular fluorescent continuous glucose sensor in ICU patients. J Diabetes Sci Technol. 2015;9(4):762-770. doi:10.1177/1932296815585872.

33.

Tingsarat

Buranasupkajorn

Khovidhunkit

Boonchaya-Anant

Laichuthai

Accuracy of continuous glucose monitoring in the medical intensive care unit. J Diabetes Sci Technol. 2022;16(6):1550-1554. doi:10.1177/19322968211027590.

34.

Leelarathna

English

Thabit

, et al. Accuracy of subcutaneous continuous glucose monitoring in critically ill adults: improved sensor performance with enhanced calibrations. Diabetes Technol Ther. 2014;16(2):97-101. doi:10.1089/dia.2013.0221.

35.

Perri

Tiberi

Giordano

, et al. Strict glycaemic control in very low birthweight infants using a continuous glucose monitoring system: a randomised controlled trial. Arch Dis Child Fetal Neonatal Ed. 2022;107(1):26-31. doi:10.1136/archdischild-2020-320540.

36.

Beardsall

Thomson

Elleri

Dunger

Hovorka

Feasibility of automated insulin delivery guided by continuous glucose monitoring in preterm infants. Arch Dis Child Fetal Neonatal Ed. 2020;105(3):279-284. doi:10.1136/archdischild-2019-316871.

37.

Zuo

Guo

Wen

Kang

Continuous glucose monitoring system can improve the quality of glucose control and glucose variability compared with point-of-care measurement in critically ill patients: a randomized controlled trial. Medicine. 2018;97(36):e12138. doi:10.1097/MD.0000000000012138.

38.

Preiser

Lheureux

Thooft

Brimioulle

Goldstein

Vincent

JL.

Near-continuous glucose monitoring makes glycemic control safer in ICU patients. Crit Care Med. 2018;46(8):1224-1229. doi:10.1097/CCM.0000000000003157.

39.

van Steen

Rijkenberg

Sechterberger

DeVries

van der Voort

PHJ.

Glycemic effects of a low-carbohydrate enteral formula compared with an enteral formula of standard composition in critically ill patients: an open-label randomized controlled clinical trial. JPEN J Parenter Enteral Nutr. 2018;42(6):1035-1045. doi:10.1002/jpen.1045.

40.

De Block

Gios

Verheyen

, et al. Randomized evaluation of glycemic control in the medical intensive care unit using real-time continuous glucose monitoring (REGIMEN trial). Diabetes Technol Ther. 2015;17(12):889-898. doi:10.1089/dia.2015.0151.

41.

Leelarathna

English

Thabit

, et al. Feasibility of fully automated closed-loop glucose control using continuous subcutaneous glucose measurements in critical illness: a randomized controlled trial. Crit Care. 2013;17(4):R159. doi:10.1186/cc12838.

42.

Sampaio

Franco

Goldberg

Baptista

Eliaschewitz

FG.

Glucose control in acute myocardial infarction: a pilot randomized study controlled by continuous glucose monitoring system comparing the use of insulin glargine with standard of care. Diabetes Technol Ther. 2012;14(2):117-124. doi:10.1089/dia.2011.0157.

43.

Beardsall

Ogilvy-Stuart

Frystyk

, et al. Early elective insulin therapy can reduce hyperglycemia and increase insulin-like growth factor-I levels in very low birth weight infants. J Pediatr. 2007;151(6):611-617, 617.e1. doi:10.1016/j.jpeds.2007.04.068.

44.

Longo

Elias

Khan

Seley

JJ.

Use and accuracy of inpatient CGM during the COVID-19 pandemic: an observational study of general medicine and ICU patients. J Diabetes Sci Technol. 2022;16(5):1136-1143. doi:10.1177/19322968211008446.

45.

Chow

Kelly

Rieff

, et al. Outcomes and healthcare provider perceptions of real-time continuous glucose monitoring (rtCGM) in patients with diabetes and COVID-19 admitted to the ICU. J Diabetes Sci Technol. 2021;15(3):607-614. doi:10.1177/1932296820985263.

46.

Galindo

Aleppo

Klonoff

, et al. Implementation of continuous glucose monitoring in the hospital: emergent considerations for remote glucose monitoring during the COVID-19 pandemic. J Diabetes Sci Technol. 2020;14(4):822-832. doi:10.1177/1932296820932903.

47.

Perez-Guzman

Duggan

Gibanica

, et al. Continuous glucose monitoring in the operating room and cardiac intensive care unit. Diabetes Care. 2021;44(3):e50-e52. doi:10.2337/dc20-2386.

48.

Matievich

Kiaie

Dunn

. Safety and functional integrity of CGM sensors when used during radiologic procedures under high exposure conditions [published online ahead of print June 30, 2022]. J Diabetes Sci Technol. doi:10.1177/19322968221106206.

49.

Thomas

Welsh

Gray

JM.

Safety and functional integrity of continuous glucose monitoring components after simulated radiologic procedures. J Diabetes Sci Technol. 2021;15(4):781-785. doi:10.1177/1932296820920948.

50.

Tripyla

Herzig

Joachim

, et al. Performance of a factory-calibrated, real-time continuous glucose monitoring system during elective abdominal surgery. Diabetes Obes Metab. 2020;22(9):1678-1682. doi:10.1111/dom.14073.

51.

Faulds

Jones

McNett

, et al. Facilitators and barriers to nursing implementation of continuous glucose monitoring (CGM) in critically ill patients with COVID-19. Endocr Pract. 2021;27(4):354-361. doi:10.1016/j.eprac.2021.01.011.

52.

Nguyen

DuBord

, et al. The launch of the iCoDE Standard Project. J Diabetes Sci Technol. 2022;16(4):887-895. doi:10.1177/19322968221093662.

53.

Spanakis

Urrutia

Galindo

, et al. Continuous glucose monitoring-guided insulin administration in hospitalized patients with diabetes: a randomized clinical trial. Diabetes Care. 2022;45(10):2369-2375. doi:10.2337/dc22-0716.