The COVID-19 pandemic has created a climate of clinical creativity. Given the urgent need to preserve personal protective equipment (PPE) and minimize risk of exposure to the novel coronavirus, some cutting-edge technology has jumped quickly from clinical trials into real-world hospital practice. One of those technologies is real-time continuous glucose monitoring (CGM).
Formerly confined to the outpatient setting except in the realm of research, CGM devices got the green light for in-hospital use in April 2020 when the FDA told manufacturers it would not object to such use during the pandemic. In response, the companies delivered two types of CGM devices to hospitals: the Abbott FreeStyle Libre 14-day Flash Glucose Monitoring System and the Dexcom G6.
While the factory-calibrated CGM devices are approved to inform outpatient decision making around insulin dosing without fingerstick testing, the devices are still not FDA approved for use in the hospital, noted Francisco J. Pasquel, MD, MPH, an assistant professor of medicine in the division of endocrinology at Emory University School of Medicine in Atlanta. “In the hospital, CGM devices are not approved yet because [of] the acuity of illness of patients and exposure to other medications,” he said. “We don't know if there are interfering medications or clinical scenarios where the devices may not be as reliable as in the outpatient setting.”
But with more hospitals adopting remote CGM, those answers may come sooner than was previously possible. “The devices are being used in more than 100 hospitals in the U.S.,” said Dr. Pasquel, who has collected national and international experiences including hospital CGM implementation through the Collaborative Open-Access Virtual Database for COVID-19 in Diabetes, online at covidindiabetes.org. “Before COVID, there were just a few groups testing the technology in the hospital. But now, multiple centers are using CGM—not necessarily for research. They're actually using it for caring for patients.”
He and other experts described current use of CGM in hospitals and the potential benefits and challenges of wider inpatient implementation.
While there isn't a large amount of data on using CGM in the hospital, especially on the general wards, there is reason to expect potential benefit, particularly during the pandemic.
In April 2020, the Diabetes Technology Society organized a consensus guideline panel of 24 international experts in the use of CGM. The guideline, published in September 2020 by the Journal of Diabetes Science and Technology, strongly recommended that clinicians consider initiating CGM in the hospital to reduce nurse contact for point-of-care glucose testing and use of PPE for patients on isolation with highly contagious infectious diseases, such as COVID-19.
However, the panel also said researchers must provide data to show that CGM improves patient-centered outcomes. Toward that goal, while bedside point-of-care glucose testing remains the standard of care in hospitalized patients, small trials in inpatient, non-ICU wards have found decreased hypoglycemia with CGM devices, according to an article published in the September 2020 Journal of Clinical Endocrinology & Metabolism.
Improved glycemic control in the non-ICU hospital setting was seen in preliminary results from a trial of the Dexcom G6 CGM, launched at Scripps Mercy Hospital in San Diego before the pandemic. It randomized 110 adults with type 2 diabetes admitted to a non-ICU floor to real-time CGM with the device (n=57) or usual care (n=53).
CGM data were wirelessly transmitted from the bedside and monitored by the hospital's telemetry service, which notified bedside nurses of glucose alerts and trends. A diabetes advanced practice nurse also reviewed daily glucose trends and provided recommendations to physicians for insulin adjustments as needed. Compared with the usual care group, the CGM group had significantly lower mean glucose levels and percentage of time in hyperglycemia and a greater time in the target range of 70 to 250 mg/dL (3.9 to 13.9 mmol/L), according to results published in the November 2020 Diabetes Care.
When COVID-19 hit, the Scripps Health research group put the trial on pause to roll out CGM as a standard of care for high-risk patients with diabetes on select hospital units, said lead author Addie L. Fortmann, PhD. An important advantage of this technology is that it provides a continuous picture of a patient's glucose trends every five minutes. It also reduces the number of times a bedside nurse needs to enter a room for point-of-care testing, which in turn can decrease nursing burden and exposure risk, while also preserving PPE, she said.
“When we have some of our highest-risk patients with comorbid diabetes and COVID in the hospital, more information is better. Being able to manage these patients' glucose more effectively with continuously streaming data is a huge benefit during this time,” said Dr. Fortmann, director of the diabetes service line at the Scripps Whittier Diabetes Institute in La Jolla, Calif.
The team decided to publish the results in case others wanted to do the same. “The study had not completed its entire enrollment, but . . . the thought was, since it's the G6 that's available now, if we pull out that subpopulation that had been started on the G6 and report on it, it might be valuable for other places across the country,” said senior author Athena Philis-Tsimikas, MD, an ACP Member and corporate vice president for the Scripps Whittier Diabetes Institute.
Currently, CGM is being used on three units at Scripps Mercy Hospital: a designated COVID-19 unit, a step-down ICU, and an ICU. “Patients with a diagnosis of diabetes who require insulin in the hospital are considered for CGM. Obviously, that gives us a very large pool to choose from, so we're risk stratifying to prioritize our highest-risk patients,” said Dr. Fortmann. “For example, patients with comorbid COVID and diabetes would be the top of our list to be placed on CGM.”
The protocol has three main components. The bedside nurse can access CGM values from a tablet, a 24/7 centralized monitoring center reaches out to nurses about patients trending toward hypoglycemia, and a centralized diabetes advance practice nurse reviews trends on a daily basis and makes recommendations to the hospitalist if insulin adjustments may be warranted, Dr. Fortmann noted. “Ideally, the CGM device remains on the patient for the duration of the hospitalization.”
Grady Memorial Hospital in Atlanta, where Dr. Pasquel and Emory investigators are conducting clinical studies using CGM in the non-ICU setting, has also rolled out the devices in the ICU. With regard to accuracy, CGM is “reliable in the hospital, but we're still not confident that it can completely replace point-of-care,” he noted.
To ensure CGM sensor values are within range of actual glucose values, hospitals are using a hybrid model with confirmatory point-of-care testing once or twice a day for non-ICU patients, according to a June 2020 paper in the Journal of Diabetes Science and Technology.
“The goal is to use sensor values but at the same time make sure intermittently that they are accurate,” said Dr. Pasquel, who was senior author of the article, which also provided recommendations for CGM use in the hospital during COVID-19. “Following the experience from other centers, before using sensor values, we confirm the initial results are within 20% of a matching point-of-care test, and then we document the device meets validation criteria.”
While CGM values have also partially replaced hourly point-of-care tests in the hospital's ICU, there's more frequent validation in critically ill patients. “We are validating every six hours because the patients are receiving insulin intravenously. So in the ICU is where you get the most benefit of this technology, because otherwise, you need to be checking point-of-care testing hourly,” said Dr. Pasquel.
CGM in the hospital is not without its drawbacks.
Cost is one, although whether the technology increases or reduces it remains unclear. At the start of the pandemic, manufacturers worked with the American Diabetes Association and other groups to donate thousands of CGM sensors, phones, and readers for hospitals to use during COVID-19, in addition to some sold at discounted prices. The prices of the devices have also decreased over the last few years, “And I expect that that cost will continue to decline,” Dr. Philis-Tsimikas said.
When assessing the costs of inpatient CGM implementation, it's important to account for savings as well, Dr. Pasquel noted. “If you are reducing the number of point-of-care tests and the amount of PPE, the use of this technology should be cost-effective,” he said, noting that the sensors in the G6 and the FreeStyle Libre last 10 and 14 days, respectively. “The cost of point-of-care testing is pretty high in the hospital, and instead of doing 24 point-of-care tests in the ICU, you would be doing only four to six with real-time CGM.”
Dr. Pasquel noted that while flash glucose monitoring with the FreeStyle Libre is less costly than real-time CGM, it is also less practical in the ICU because the sensor must be scanned to produce a value. “Real-time CGM [with the G6] is more convenient in the ICU because you can see real-time data outside of the patient room,” he said. (Dr. Pasquel has received research funds from Dexcom for investigator-initiated studies.)
But ACP Member Nita Thingalaya, MD, a practicing hospitalist who is board certified in obesity medicine, said she is not completely convinced that using CGM devices is cost-effective for hospitals.
“If the patient already has the device, that can be used in the hospital. It's a snap-on adhesive patch, which literally goes on the skin. However, the device is for personal patient use only,” she said. “The facility will have to provide a device for each patient requiring blood sugar checks. The feasibility of this approach depends on several factors, including if the patient meets the clinical criteria for its use.”
But patients who already use the devices at home should continue to use them during hospitalization if possible, and hospitals should create policies around how clinicians will access glycemic data (e.g., transferring information to/from the cloud), Dr. Thingalaya said. “How are we going to have access to these records, which are on the patient's smartphone? How are they going to be relayed to the hospitalists?”
The accuracy of CGM devices is another concern, and the most effective way to deal with discrepancies between CGM and point-of-care glucose values remains unclear. Clinicians know that the systems are based on different methodologies, but many are not sure how to address slight discrepancies, Dr. Philis-Tsimikas noted.
Other major challenges include getting hospital leadership buy-in and training staff before implementing the technology. “You have to have the staffing to educate a lot of people about this,” Dr. Philis-Tsimikas said. “You have to place the technology: to put it on someone, to connect it to the cloud, to connect it back to the nursing station. All of those things take time, money, trained staff, oversight.”
Since access to continuous 24-hour glucose trends can affect treatment decision making, she suggested having diabetes experts assist hospitalists and nurses. “It is slightly different than when you only had a point-of-care snapshot of only four to six numbers that you might have in a day. …You are being provided this enormous amount of data now, but in a regular hospital, there aren't people that really know how to use that.”
Fortunately, clinicians seem eager to learn, and one way for them to do that is to try out a CGM device on themselves, Dr. Philis-Tsimikas said. “We've had a number of hospitalist physicians and nurses that have asked if they can try it themselves, and the teams have happily placed those. You always learn something.”
Dr. Pasquel agreed. “That's the best way to learn the technology: to wear it,” he said. “It's extremely easy to place one and to use it with your phone. Once you see the trends, you may change your behavior after a burger with fries and a milkshake!”
Even without that hands-on experience, most physicians should not have issues understanding how the devices work. “I don't think physicians have difficulty with interpretation,” Dr. Thingalaya said. “I think the difficulty is in adoption and implementation.”
Ultimately, hospitals that have not implemented the technology through clinical trials may not yet be equipped to do so on the fly, according to Dr. Thingalaya. “The health system as a whole needs to have a policy in place. It's like having a telemetry monitor: If anything is being monitored, someone needs to be there to monitor it,” she said. “This is the way of the future, and I think it will come, but I think we're just not there yet.”
For hospitals who do want to take the plunge, resources on CGM implementation are available. Information on research publications and inpatient CGM protocols is online at covidindiabetes.org, and some researchers themselves are also happy to help.
“We've had 20 health systems reach out to us so far from across the United States asking for materials and protocols. I think they're all encountering many of the same challenges that we are, in terms of less familiarity with CGM amongst physicians and nursing staff, and questions about how to integrate this with routine care processes and the EMR,” said Dr. Fortmann. “I think we're a couple years ahead, so we can share those lessons learned, but we're all in this together, learning more each day.”