Proposed Use of Continuous Glucose Monitoring for Care of Critically Ill COVID-19 Patients

Abstract

Coronavirus disease 2019 (COVID-19) has disproportionately affected patients with diabetes. Mounting evidence has shown that adequate inpatient glycemic control may decrease the risk of mortality. In critically ill patients, insulin drips are the most effective means of controlling blood glucose. However, resource limitations such as the availability of protective equipment and nursing time have discouraged the use of insulin drips during COVID-19. In this commentary, we review existing evidence on the importance of glycemic control in COVID-19 patients with diabetes and propose a protocol for utilizing continuous glucose monitors (CGMs) to improve glycemic control by decreasing the need for bedside management in critically ill COVID-19 patients.

Since the early stages of the pandemic, observational studies of coronavirus disease 2019 (COVID-19) have repeatedly demonstrated worse outcomes in patients with diabetes. A study published in February 2020 from a series of patients in a Wuhan intensive care unit (ICU) found a higher prevalence of diabetes in nonsurvivors of COVID-19 (22%) than survivors (10%). ¹ Additionally, a retrospective study of English patients found a hazard ratio of death of 2.67 for patients with a hemoglobin A1c >7.5% compared with those without diabetes. ²

After diabetes was identified as a major risk factor for poor outcomes, initial hypotheses queried whether the physiologic dysfunction of chronic diabetes was the main driver of this increased risk. Several large studies of patients with diabetes admitted with COVID-19 failed to show an association between hemoglobin A1c and outcomes or between preadmission diabetes complications and COVID-19 mortality.^3,4 However, a recent study from Israel did find increased COVID-19 hospitalization risk in patients with diabetes who had an A1c ≥9%. ⁵ It therefore remains controversial whether preadmission glycemic control contributes to the excess risk of morbidity and mortality in COVID-19 patients with diabetes.

Before COVID-19, inpatient hyperglycemia had been associated with an increased risk of infections, hospital complications, increased length of stay, and overall mortality.^6,7 More recent studies suggest that poor glycemic control during hospitalization has an effect on mortality for patients with COVID-19 as well. In May, a study in Cell Metabolism by Zhu et al ⁸ retrospectively examined over 7000 COVID-19 cases at multiple sites, including 952 patients with pre-existing diabetes. After propensity matching, 250 patients with diabetes with an admission blood glucose of 70-180 mg/dL to 250 patients with diabetes with hyperglycemia (blood glucose [BG] >180 mg/dL), the authors found a hazard ratio of 0.14 for mortality at one month, offering slightly stronger evidence that control of hyperglycemia itself may contribute to an improved outcome.

Sardu et al ⁹ further explored the effects of BG control in patients with hyperglycemia and COVID-19. They offered insulin drips to 25 patients with COVID-19 pneumonia and hyperglycemia (BG >140 mg/dL) admitted to the floor, of which 15 chose a drip and 10 chose basal-bolus insulin. Although the average BG on admission was higher in patients who chose the insulin drip (221 vs 199 mg/dL), the average BG during the entire hospitalization was significantly lower in the insulin drip group (139 vs 192 mg/dL) despite a mean time on the drip of only 33 hours. Inflammatory markers in the insulin drip group were lower and more closely resembled those of patients without hyperglycemia at one and two weeks after admission. Although reasonably well-matched at the onset, half of the patients in the standard care group died compared with none in the insulin drip group (P = .005). The trial is nonrandomized and small but suggests that aggressive insulin control could potentially have an impact on mortality in patients with COVID-19.

Since Sardu et al’s study, a new treatment was shown to be effective for the treatment of COVID-19: dexamethasone. ¹⁰ As is well known, steroids exacerbate hyperglycemia. Although the initial trial did not delve into hyperglycemia with dexamethasone treatment, it is certain that glycemic control in critically ill patients will deteriorate without concerted effort.

The algorithmized intravenous (IV) insulin protocols, colloquially referred to as insulin drips, are the optimal management for diabetic crises as well as hyperglycemia in critical illness ¹¹ ; they are the most effective way to control blood sugars in the critical care setting. Target BG can be attained quickly and, with a half-life on the order of minutes, the treatment can be rapidly halted to avoid hypoglycemia.

Unfortunately, insulin drips are resource-intensive in the best of times. During COVID-19, they severely strain nursing time and personal protective equipment resources, not to mention the exposure risk to nurses. The algorithms (and patient safety) demand hourly bedside finger stick checks, meaning 24 room entries daily to manage the drip. The heavy resource burden of insulin drips has led to tension between health care worker time and safety versus optimal patient care.^12,13 However, in April 2020, the Food and Drug Administration (FDA) indicated no objection to the use of continuous glucose monitors (CGMs) for remote inpatient monitoring of BGs in COVID-19 patients. ¹²

CGMs allow patients to monitor their own BGs without confirmatory finger sticks. Although developed for the outpatient setting, with a little ingenuity (and help from manufacturers), these devices can work in the inpatient setting.^6,14 Combined with the positioning of the IV pole out of the room, CGMs could obviate the need for bedside glycemic care.

Reassuringly, prior studies of patients in the ICU have found no difference in glycemic control with CGM use or even some benefit in avoiding hypoglycemia. ¹⁵ A recent study in surgical patients demonstrated sufficient accuracy of the Dexcom G6, a CGM that is FDA-approved for BG monitoring without finger stick calibration, for IV insulin algorithms, with a mean absolute relative difference of <10% between the CGM readings and finger stick BGs. ¹⁶ Several health care centers in the United States (personal communication) and abroad ¹⁷ have already piloted CGM use for COVID-19 patients to good effect. The excellent review of inpatient CMG use during COVID-19 by Galindo et al proposes a practical guide for CGM initiation in noncritically ill patients. ¹²

Currently, there are no published guidelines for the use of CGMs in ICU settings, although there has long been interest in deploying them.^18,19 The rising number of COVID-19 patients requiring ICU admission, mounting evidence that improvement in glycemic control may improve clinical outcomes, and FDA’s nonobjection to utilize CGM technology urges clinicians and researchers to create such a protocol. Here, we propose a protocol for implementation in critically ill patients requiring insulin drips with COVID-19 (Table 1).

Table 1.

Protocol for ICU Insulin Drip Titration With Dexcom G6 CGM, Which Does Not Require Finger Stick Calibration.

Protocol for Dexcom G6 use for critically ill patients requiring IV insulin
Inclusion criteria	Patients with COVID-19 in the ICU with BG >180
Exclusion criteria	Shock requiring more than one pressor
	Patients on hydroxyurea
Protocol	1) Place Dexcom G6 CGM sensor and transmitter onto upper arm
	2) Set up CGM device
	3) Check FSBG at the time of sensor placement, one hour after sensor placement (while sensor is still in warm up there will be no readings), and two hours after sensor placement (when the sensor has warmed up)
	a) If CGM reading is off by >20%, follow FSBG every one to two hours until CGM readings within 20% of FSBG
	4) Check every six hours after two-hour warm-up placement for 22 more hours (24 hours total) for confirmation of CGM accuracy
	a) If CGM reading is off by >20%, follow FS BG every one to two hours until CGM readings within 20% of FSBG
	5) After 24 hours since CGM placement, check BG every six hours
	a) If CGM reading is off by >20%, follow FSBG every one to two hours until CGM readings within 20% of FSBG
	b) If double arrow down: do not use CMG reading and revert to every one hour finger sticks until resolved
	c) BG <200, ensure patient has glucose source (tube feeds or D10W)
	d) If alarming, verify with FSBG
	e) If hypoglycemia‚ follow hypoglycemia protocol
	f) If hyperglycemia‚ follow insulin GTT protocol
	g) If BG and arrow not available on CGM, check FSBG
	h) If clinical concern, check FSBG PRN
	6) Use CGM to monitor glucose data remotely every one hour
	a) Enter BG into insulin GTT calculator
	b) Adjust IV insulin per calculator if indicated
	7) If clinical status worsens but CGM already in place (eg, additional pressor), check FSBG at that time
	a) If CGM reading is off by >20%, follow FSBG every one to two hours until CGM readings within 20% of FSBG
	8) After 10 days, replace sensor with same transmitter
Imaging considerations	Ultrasound: no modifications
	X-ray, computed tomography, fluoroscopy: consider removing or place a lead apron over the device20,21
	Magnetic resonance imaging: remove device 21

BG, blood glucose; CGM, continuous glucose monitor; FSBG, finger stick blood glucose; GTT, glucose tolerance test; ICU, intensive care unit; IV, intravenous.

Similar to Galindo et al, our proposed protocol is based on our clinical experience as well as anecdotal correspondences with sites that have trialed CGMs for critically ill COVID-19 patients. There is a known lag time between serum and interstitial glucose measurement, ²² and it remains to be elucidated the effects of shock or pressors on this delay. It is therefore important for this protocol to address situations when CGM glucose measurements need to be confirmed by conventional glucometers and when it is safe to use CGM glucose value. Our protocol emphasizes checking fingerstick BGs not only when there is a change in the patient’s clinical status but also at times of rapid glucose shift (indicated with double arrows on the CGM receiver), which is particularly risky time for unacceptable CGM lag. ²² This protocol will need to be validated in prospective studies but can be used as a starting point in an unprecedented time of COVID pandemic. Expanding CGM technology to the inpatient setting can allow patients to maintain good glycemic control without sacrificing the safety of healthcare workers during the COVID-19 crisis.