Outcomes and Healthcare Provider Perceptions of Real-Time Continuous Glucose Monitoring (rtCGM) in Patients With Diabetes and COVID-19 Admitted to the ICU (original) (raw)

Abstract

Objective:

We assessed the clinical utility and accuracy of real-time continuous glucose monitoring (rtCGM) (Dexcom G6) in managing diabetes patients with severe COVID-19 infection following admission to the intensive care unit (ICU).

Methods:

We present retrospective analysis of masked rtCGM in 30 patients with severe COVID-19. rtCGM was used during the first 24 hours for comparison with arterial-line point of care (POC) values, where clinicians utilized rtCGM data to adjust insulin therapy in patients if rtCGM values were within 20% of point-of-care (POC) values during the masked period. An investigator-developed survey was administered to assess nursing staff (n = 66) perceptions regarding the use of rtCGM in the ICU.

Results:

rtCGM data were used to adjust insulin therapy in 30 patients. Discordance between rtCGM and POC glucose values were observed in 11 patients but the differences were not considered clinically significant. Mean sensor glucose decreased from 235.7 ± 42.1 mg/dL (13.1 ± 2.1 mmol/L) to 202.7 ± 37.6 mg/dL (11.1 ± 2.1 mmol/L) with rtCGM management. Improvements in mean sensor glucose were observed in 77% of patients (n = 23) with concomitant reductions in daily POC measurements in 50% of patients (n = 15) with rtCGM management. The majority (63%) of nurses reported that rtCGM was helpful for improving care for patients with diabetes patients during the COVID-19 pandemic, and 49% indicated that rtCGM reduced their use of personal protective equipment (PPE).

Conclusions:

Our findings provide a strong rationale to increase clinician awareness for the adoption and implementation of rtCGM systems in the ICU. Additional studies are needed to further understand the utility of rtCGM in critically ill patients and other clinical care settings.

Keywords: COVID-19, rtCGM, type 2 diabetes, hyperglycemia, cardiac arrest

Introduction

As the COVID-19 pandemic continues to challenge US healthcare systems, ensuring the safety of healthcare providers is paramount as we strive to provide effective treatment to patients. Diabetes is one of the most common comorbidities associated with hospitalizations for SARS-CoV-2 infection1,2 and a significant risk factor for disease severity and mortality.3-5 Acute hyperglycemia has been proposed to be particularly dangerous during SARS-CoV-2 infection due to increased levels of inflammatory mediators6 and binding of SARS-CoV-2 to the angiotensin-converting enzyme 2 (ACE2).7

Furthermore, numerous studies have shown that achieving and maintaining adequate glycemic control in critically ill patients can profoundly impact clinical outcomes.8 This matter is especially important in critically ill diabetes patients and patients being treated with insulin infusions for hyperglycemic crises such as diabetic ketoacidosis (DKA) or hyperosmolar hyperglycemic state (HHS).9 Similarly, acute glycemic management may also play an important role in reducing viral load and infection duration in diabetes patients with COVID-19, even though the underlying mechanisms are still unknown.10

Insulin treatment with intravenous (IV) infusion is the preferred route of insulin delivery during critical care scenarios due to its rapid onset and short duration to match insulin requirements for rapidly changing glucose levels.11 However, utilizing arterial blood glucose testing often delays the detection of problematic glycemic patterns. This delay can impact the timeliness of achieving and sustaining optimal steady state insulin infusion.12,13 Although large randomized trials and real-world observational studies have demonstrated the glycemic benefits of real-time continuous glucose monitoring (rtCGM) in outpatient settings,14-18 the safety and efficacy of rtCGM in the intensive care unit (ICU) have not been well studied due to prior regulatory restrictions.

However, in response to the COVID-19 pandemic, the US Food and Drug Administration (FDA) recently issued a new policy to expand the availability and capability of noninvasive remote monitoring devices (eg, rtCGM) to facilitate inpatient monitoring while reducing patient and healthcare provider contact and exposure to SARS-CoV-2 infection for the duration of the public health emergency.19 rtCGM receivers are placed outside the doors of patients rooms so that nursing staff are able to closely monitor glycemic control to limit exposure with patients infected with SARS-CoV-2 and directly decrease the risk of infection for healthcare providers.20

Since rtCGM is an effective solution that can reduce exposure to SARS-CoV-2 infectious particles and minimize the utilization of personal protective equipment (PPE), we initiated a protocol for using rtCGM to manage glycemia in patients admitted to the ICU for severe SARS-CoV-2 infection at University Hospital (Stony Brook, NY) in April 2020. A recent consensus statement states the potential benefits of rtCGM in critically ill patients, but there are concerns about the accuracy of rtCGM due to potential effects from different states of shock, dehydration, hypoxemia, and various medications.21 Even though we shared similar concerns about the use of rtCGM in ICU settings, we hypothesized that rtCGM would be valuable for remotely managing glycemia in patients with diabetes with severe COVID-19 infections in the ICU.

We present how rtCGM was utilized and correlated with arterial blood glucose point-of-care (POC) measurements to manage COVID-19 patients with hyperglycemia in the ICU with the intention of testing whether we could minimize staff exposure to SARS-CoV-2 and conserve PPE.

Methods

Design and Participants

This was a retrospective analysis that evaluated the accuracy and clinical utility of rtCGM in managing 30 diabetes patients with severe SARS-CoV-2 infection following admission to the ICU at Stony Brook University Hospital. The protocol was approved by a local institutional review board (Office of Research Compliance at the Renaissance School of Medicine at Stony Brook University).

Inclusion criteria for the analysis were age ≥18 years, history of diabetes or presenting with new-onset DKA, admission to ICU, traditional blood glucose measurements ordered at least every two hours, lactate, and arterial blood gas (ABG)/venous blood gas (VBG) ordered as per routine standard of care in the ICU. Pregnant patients were excluded..

rtCGM Protocol

All patients were monitored using the Dexcom G6 system (Dexcom, Inc., San Diego, California). The system comprises three components: a disposable sensor inserted into subcutaneous tissue; a transmitter that is attached to the sensor; and a receiver (handheld or smartphone app) that displays interstitial glucose measurement data every five minutes. Data were uploaded to the Dexcom Clarity software for visualization and analysis. Arterial-line POC testing was performed using the Nova StatStrip glucose meter (Nova Biomedical, Waltham, MA).

After identifying eligible patients, members of the diabetes team placed the sensor and transmitter on the patient at bedside. The receiver was hung outside the patient’s room in view of the ICU nurses to minimize staff exposure to SARS-CoV-2. Nurses were trained to recognize and respond to alerts and problematic glycemic patterns. At the conclusion of rtCGM use due to patient discharge, patient expiration, or sensor expiration, nurses uploaded the rtCGM data to the device software (Dexcom Clarity) for analysis. During the first 24 hours of rtCGM, arterial-line POC testing was performed concurrently with rtCGM. rtCGM data were masked to treating clinicians in this phase, unless glucose levels fell below established thresholds. Alarms were set for hypoglycemia (<70 mg/dL [3.9 mmol/L]) in conjunction with trend arrows indicating a rapid rise or fall in glucose, which prompted physicians or nurses to reassess the patient to confirm the glycemic excursion/trend with POC and treat accordingly. After the first 24 hours of rtCGM, the data were unmasked and compared with arterial-line POC values. In patients where the rtCGM values were within 20% of POC values, rtCGM was used to adjust insulin (once every two hours for insulin infusion) in conjunction with confirmatory arterial-line POC testing.

Outcomes

Primary outcomes included (1) change in mean sensor glucose during rtCGM management compared with each patient’s pre-rtCGM use and (2) percentage differences between rtCGM and arterial-line POC glucose values during the first 24 hours of rtCGM management. A secondary outcome was the assessment of the perception of nursing staff regarding the utilization of rtCGM by using an investigator-developed questionnaire. The questionnaire was administered electronically using the Qualtrics platform (XM Qualtrics, Provo, UT) over a 30-day period during the Summer of 2020.

Statistical Analysis

We utilized exported data from the Dexcom Clarity web interface and aligned extracted arterial-line POC testing with respect to time. We plotted the data on the same (x, y) axes in order to compare the utility of the rtCGM to arterial-line POC in the ICU. Beyond interpreting the trends and general performance of the rtCGM instrumentation with respect to the arterial-line POC glucose values, we also calculated basic statistics to ascertain the percent absolute difference between POC glucose testing from the arterial line with the rtCGM glucose value in the closest corresponding time frame (since rtCGM is being measured every five minutes). We employed standard formulas to calculate the percent absolute difference between each arterial-line POC measurement and the closest rtCGM measurement. Percent absolute difference was defined by the formula:|V1−V2|12(V1+V2).

For each patient, we also calculated the overall mean and standard deviation of the percent absolute difference to characterize the general variability between the measurement modalities. Beyond evaluating interstital glucose values from rtCGM with arterial-line glucose values, we also calculated the mean sensor glucose values with respect to the number of daily POC glucose measurements before and during rtCGM management in an effort to gain insight into the clinical utility of rtCGM to help our staff minimize exposure to SARS-COV-2 by limiting unnecessary contact with patients with diabetes patients with severe COVID-19 infections who would normally require close glycemic management with arterial-line POC testing. We performed two-tailed paired t tests to study changes in mean sensor glucose (P = .0003) and the number of daily POC measurements (P = .08) as surrogate metrics to study the efficacy of glycemic management with rtCGM, limiting exposure, and potentially conserving PPE.

Results

Concordance Between rtCGM and Arterial-Line POC

rtCGM data showed high concordance with arterial-line POC in all but two patients (patients 16 and 17, as shown in Table 1). Since the percentage difference was greater than the established cutpoint of 20% between rtCGM and POC, rtCGM data were not used for insulin adjustments in patients 16 and 17 (Table 1). Discordance between rtCGM and POC values were observed in nine additional patients (patients 3, 5, 8, 9, 11, 15, 27, 28, and 29, as shown in Table 1). However, these differences between rtCGM and POC were not considered clinically significant because the POC values were obtained during periods of rapidly changing glucose, which is a known condition that causes a lag between blood glucose and interstitial glucose.

Table 1.

Percent Absolute Difference Between Each POC Arterial Glucose Value and the Closest rtCGM Glucose Value.

As shown in Figure 1, the rtCGM trend lines closely correlated with arterial-line POC values, even in patients with significant glycemic variability and extremely elevated glucose.

Figure 1.

Concordance between rtCGM and arterial-line POC values in selected patients. rtCGM: real-time continuous glucose monitoring.

Change in Mean Sensor Glucose

Within the full cohort, we observed a 14% reduction in mean sensor glucose during rtCGM management compared with pre-rtCGM management, P = .0003. (Figure 2). Reductions were observed in 23 of the 30 patients who were evaluated in this study.

Figure 2.

Overall change in mean sensor glucose level during rtCGM management. rtCGM: real-time continuous glucose monitoring.

Improvements in mean sensor glucose were observed in almost 77% of patients with diabetes and severe COVID-19 infections (n = 23 patients) with concurrent reductions in daily POC measurements in half of the patients in the study (n = 15 patients) due to rtCGM management as shown inTable 2.

Table 2.

Change in Mean Sensor Glucose Levels and Daily POC Glucose Measurements During rtCGM Management.

Two-tailed paired t tests were performed for changes in mean sensor glucose (P = .0003) and number of daily POC measurements (P = .08). During rtCGM management, all but three patients spent <1% of time above the low threshold target of <55 mg/dL (3.1 mmol/L). Among these patients, the percentage of time below the target ranged from 1.7% to 2.6%.

ICU Nurse Perceptions of rtCGM

Among the 66 ICU nurses surveyed, 35 (56%) reported interactions with rtCGM equipment and data as shown in Table 3. The majority of those nurses (63%) reported that the utilization of rtCGM helped improve clinical care for severe COVID-19 patients with diabetes, and 49% indicated that the use of rtCGM had reduced their use of PPE.

Table 3.

Responses to ICU Nurse Survey.

Discussion

Our findings demonstrate that use of rtCGM was useful in remotely observing and managing glycemic physiology by intensivists in critically ill patients with diabetes and severe COVID-19 infections. These COVID-19 patients with diabetes presented with marked increases in insulin requirements and significant fluctuations in blood glucose concentrations. The ability to continuously monitor and quickly respond to changing glucose in COVID-19 patients represents an important opportunity to potentially improve clinical outcomes and positively impact healthcare resource utilization. Our study emphasizes the use of rtCGM in these selected patients since COVID-19 patients with diabetes and/or uncontrolled hyperglycemia had a longer length of stay and significantly higher mortality than patients without diabetes or uncontrolled hyperglycemia as reported by Bode et al.4

The implementation of rtCGM also demonstrated the proof of concept of using remote monitoring in inpatients as a viable alternative to minimize unnecessary staff exposure to patients with active COVID-19 infections and reduce unnecessary use of PPE. Even though our focus is to help improve the management of diabetes in ICU patients with severe COVID-19 infections, the profound impact of protecting our staff, conserving PPE, and reducing the strain of allocating limited resources of as hospitals nationwide report PPE shortages during the peak of the pandemic cannot be understated. Although we observed increases in POC testing in 15 patients, we believe that this was due to our learning curve in understanding COVID-19 and a by-product of being in the patient room for non-diabetes care. Nevertheless, more than two-thirds of ICU nurses who responded to the survey indicated that rtCGM helped improve care for their COVID-19 patients.

In addition, rtCGM in the ICU greatly expanded our understanding of the impact of COVID-19 on glycemia. For example, one patient admitted to the ICU was acutely malnourished with markedly elevated insulin requirements and eventually required nutritional support with TPN. Since TPN therapy further increases the risk for hyperglycemia,12,13 we initiated rtCGM to guide insulin therapy adjustments that enabled us to reduce glucose levels from 400 mg/dL (22.2 mmol/L) to a range of 200-300 mg/dL (11.1-16.5 mmol/L) within the first 12 hours. Due to the ability to remotely monitor glucose in a near-continuous fashion in real-time with rtCGM, we were able to eventually lower the glucose range to 100-200 mg/dL (5.5-11.1 mmol/L) within seven days.

We also evaluated the accuracy of rtCGM in a 71-year-old male with an extensive past medical history of comorbidities and active problems, including coronary artery disease, atrial fibrillation, chronic kidney disease, hypertension, and hyperlipidemia during the course of two successive cardiac arrests and the time interval immediately following both, where the second arrest was fatal. In the hours prior to the first event, blood glucose levels remained relatively stable (220-240 mg/dL [12.2-13.2 mmol/L]). At the time of the event, glucose was at 222 mg/dL (12.3 mmol/L), but then decreased to 167 mg/dL and then slowly increased. At the time of the second event that was fatal, which occurred 50 minutes after the first cardiac arrest, glucose had risen to 198 mg/dL (9.3 mmol/L). After the patient expired, glucose levels gradually decreased over the first post-mortem hour to 141 mg/dL and then quickly declined to undetectable levels within the next 20 minutes. Although inconclusive, this glucose pattern suggests that factors other than hypoglycemia contributed to these events.

A key strength of our report is that we did not limit our investigation to the assessment of concordance between rtCGM and arterial-line POC values to compare instrument performances and clinical utility. Rather, our findings provide additional clinically actionable insights regarding the impact of COVID-19 on glycemia and the utility of rtCGM in identifying and addressing dangerous glucose excursions. Even though our questionnaire was not a validated instrument, the nurses’ responses to items related to PPE use also suggest that many nurses may have been unfamiliar with rtCGM and/or lacked trust in novel instrumentation to measure glucose. This may partially explain why POC testing frequency increased in a subset of patients despite readily available rtCGM glucose measurements.

We believe our findings support the usefulness of rtCGM in providing continuous and dynamic glucose measurements in individual patients during the devastating circumstances of the first wave of COVID-19 infections in our healthcare system. In terms of managing diabetes as a common comorbidity in severely-ill COVID-19 patients, our goal was to show how rtCGM can help identify highly complex patterns blood glucose metabolism and regulation in patients with diabetes and provide new insights about how to potentially manage diabetes complicated by various comorbidities and other conditions like COVID-19 infections that require hospitalization in the ICU setting. rtCGM may also be beneficial when critically high or low blood glucose levels are detected outside regularly scheduled point-of-care testing before follow-up appointments or preoperative evaluations, which may enable clinicians to identify potentially symptomatic physiologic abnormalities associated with rapidly changing glycemia. There are also many practical opportunities associated with remote monitoring with rtCGM that include the ability to remotely manage and proactively treat abnormal glycemic states due to worsening health, poor adherence, and other conditions that may require hospitalization.

Conclusion

Beyond the clinical implications of our findings in this small pilot study, our results highlight the need to increase awareness about the accuracy and value of rtCGM in the ICU and provide adequate training for all clinicians who will undoubtedly encounter rtCGM measurements that are used by patients to monitor their own health. Additional studies are needed to further understand the utility of rtCGM in critically ill patients and to determine what, if any, clinical conditions may limit their use. Nonetheless, we believe that the application of rtCGM in a limited study such as this is very useful to help guide further study if there are extenuating circumstances like COVID-19.

Acknowledgments

The authors thank the diabetes education team at Stony Brook Medicine for their dedication to patient care, especially amidst a global pandemic and Christopher G. Parkin, MS, CGParkin Communications, Henderson, NV, for editorial assistance in manuscript development.

Footnotes

Author Contributions: JM was responsible for the protocol design and study implementation. KC performed the statistical analysis. All authors reviewed the data. JM is guarantor of this work and, as such, had full access to all the data and takes responsibility for the integrity of the data and the accuracy of the data analysis.

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: JM reports receiving consulting fees from Medtronic Diabetes and Ascensia Diabetes Care. MC receives compensation for participation in the Dexcom speakers bureau.

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

References

• 1.Richardson S, Hirsch JS, Narasimhan M, et al. Presenting characteristics, comorbidities, and outcomes among 5700 patients hospitalized with COVID-19 in the New York city area. JAMA. 2020;323(20):2052-2059. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 2.Bhatraju PK, Ghassemieh BJ, Nichols M, et al. Covid-19 in critically ill patients in the seattle region — case series. N Engl J Med. 2020;382:2012-2022. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 3.Guo W, Li M, Dong Y, et al. Diabetes is a risk factor for the progression and prognosis of COVID-19. Diabetes Metab Res Rev. 2020;e3319. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 4.Bruce Bode B, Garrett V, Messler J, et al. Glycemic characteristics and clinical outcomes of COVID-19 patients hospitalized in the United States. J Diabetes Sci Technol. 2020;14(4):813-821. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 5.Fadini GP, Moieri ML, Longato E, Avogaro A. Prevalence and impact of diabetes among people infected with SARS-CoV-2. J Endocrinol Invest. 2020;43(6):867-869. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 6.Ceriello A, Zarich SW, Testa R. Lowering glucose to prevent adverse cardiovascular outcomes in a critical care setting. J Am Coll Cardiol. 2009;53(5 suppl):S9-S13. [DOI] [PubMed] [Google Scholar]
• 7.Brufsky A. Hyperglycemia, hydroxychloroquine, and the COVID-19 epidemic. J Med Virol. 2020;92(7):770-775. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 8.Yao RQ, Ren C, Wu GS, Zhu YB, Xia ZF, Yao YM. Is intensive glucose control bad for critically ill patients? A systematic review and meta-analysis. Int J Biol Sci. 2020;16(9):1658-1675. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 9.Kim N-Y, Ha E, Moon JS, Lee Y-H, Young E. Acute hyperglycemic crises with coronavirus disease-19: case reports. Diabetes Metab J. 2020; 44(2):349-353. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 10.Peric S, Stulnig TM. Diabetes and COVID-19. Wien Klin Wochenschr. 2020;1–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 11.Clement S, Braithwaite SS, Magee MF, et al. Management of diabetes and hyperglycemia in hospitals. Diabetes Care. 2004; 27(2):553-591. [DOI] [PubMed] [Google Scholar]
• 12.Pasquel FJ, Spiegelman R, Mccauley M, et al. Hyperglycemia during total parenteral nutrition: an important marker of poor outcome and mortality in hospitalized patients. Diabetes Care. 2010;33(4):739-741. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 13.Cheung NW, Napier B, Zaccaria C, Fletcher JP. Hyperglycemia is associated with adverse outcomes in patients receiving total parenteral nutrition. Diabetes Care. 2005;28(10):2367-2371. [DOI] [PubMed] [Google Scholar]
• 14.Beck RW, Riddlesworth T, Ruedy K, et al. Effect of continuous glucose monitoring on glycemic control in adults with type 1 diabetes using insulin injections: the DIAMOND randomized clinical trial. JAMA. 2017;317:371-378. [DOI] [PubMed] [Google Scholar]
• 15.Beck RW, Riddlesworth TD, Ruedy K, et al. Continuous glucose monitoring versus usual care in patients with type 2 diabetes receiving multiple daily insulin injections: a randomized trial. Ann Intern Med. 2017;167:365-374. [DOI] [PubMed] [Google Scholar]
• 16.Lind M, Polonsky W, Hirsch IB. Continuous glucose monitoring vs conventional therapy for glycemic control in adults with type 1 diabetes treated with multiple daily insulin injections: the GOLD randomized clinical trial. JAMA. 2017;317(4):379-387. [DOI] [PubMed] [Google Scholar]
• 17.Šoupal J, Petruželková L, Grunberger G, et al. Glycemic outcomes in adults with T1D are impacted more by continuous glucose monitoring than by insulin delivery method: 3 years of follow-up from the COMISAIR study. Diabetes Care. 2020;43(1):37-43. [DOI] [PubMed] [Google Scholar]
• 18.Charleer S, Mathieu C, Nobels F, et al. Effect of continuous glucose monitoring on glycemic control, acute admissions, and quality of life: a real-world study. Clin Endocrinol Metab. 2018;103(3):1224-1232. [DOI] [PubMed] [Google Scholar]
• 19.Dexcom. Press release. https://www.dexcom.com/hospitalfacts. Accessed April 12, 2020.
• 20.Shanghai International Forum for Infection Control and Prevention. Rational, scientific, and standardized protection: the core of infection prevention and control of COVID-19 in medical institutions [in Chinese]. 2020.https://mp.weixin.qq.com/s/G5Nwdd9kW9yVD-hTdwsKtg. Accessed August 31, 2020.
• 21.Wallia A, Umpierrez GE, Rushakoff RJ, et al. Consensus statement on inpatient use of continuous glucose monitoring. J Diabetes Sci Technol. 2017;11(5):1036-1044. [DOI] [PMC free article] [PubMed] [Google Scholar]