Disclosures: Dexcom and Medtronic provided sensor equipment and technical support, but the companies had no role in study design, data collection or analysis. Sadhu reports no relevant financial disclosures.
November 11, 2020
4 min read

CGM use ‘reliable, accurate’ for ICU in COVID-19

Disclosures: Dexcom and Medtronic provided sensor equipment and technical support, but the companies had no role in study design, data collection or analysis. Sadhu reports no relevant financial disclosures.
You've successfully added to your alerts. You will receive an email when new content is published.

Click Here to Manage Email Alerts

We were unable to process your request. Please try again later. If you continue to have this issue please contact customerservice@slackinc.com.

The use of continuous glucose monitoring technology in critically ill patients with COVID-19 and diabetes is reliable and accurate for real-time glucose trending, according to data from a pilot study.

Archana R. Sadhu

“As an endocrinologist, we can avoid going into the [hospital] room for the routine care of the patient; we can do a lot of things remotely, using video and electronic health records, but I felt there was more we could do to help the team, especially the nurses,” Archana R. Sadhu, MD, FACE, director of the system diabetes program and director of transplant endocrinology at Houston Methodist Hospital and assistant professor of medicine at Weill Cornell Medical College, New York, told Healio. “These patients were so insulin resistant, needing hundreds of units per day. Our staff used IV insulin to manage the hyperglycemia, but this requires hourly blood glucose monitoring. That is when the idea came — why can’t we use CGM to reduce blood glucose monitoring? At the same time, FDA allowed for some flexibility in using remote monitoring during the pandemic to protect health care workers. That was our ah-ha moment.”

Image of empty hospital room
Source: Adobe Stock

New insulin protocols

In a retrospective study, Sadhu and colleagues analyzed data from 11 adults positive for COVID-19 admitted to the ICU at Houston Methodist with a history of diabetes or hyperglycemia and requiring IV or subcutaneous insulin therapy. Participants received one of two CGM devices — the Dexcom G6 (n = 5) or the Medtronic Guardian Connect (n = 6). Researchers used Bluetooth transmission to connect sensors to an iPad or iPhone outside the patient’s room to display real-time sensor glucose readings. For the first 24 hours, sensor glucose values were documented but not used for clinical management, Sadhu said.

Researchers modified the institution’s usual insulin infusion protocol, from a target range of 140 mg/dL to 180 mg/dL to a broader target of 100 mg/dL to 200 mg/dL, and allowing blood glucose monitoring every 2 hours, Sadhu said.

“I modified the IV protocol to keep glucose in a slightly broader yet still safe range than we normally would, so nurses did not have to run in the room every time the glucose fluctuated beyond the narrow 140 mg/dL to 180 mg/dL range,” Sadhu said.

With CGM, blood glucose monitoring could be reduced to every 4 hours if the hourly sensor glucose values were in range and without system alarms or alerts.

To address concerns about accuracy and patient safety, changes in insulin dosing were made using only finger-stick blood glucose measurements. Nurses used log sheets to document hourly sensor glucose, finger-stick blood glucose and administration of known potential sensor interferents.

“Because of the lack of FDA approval for accuracy, I had the added practice where we were not adjusting insulin based on the sensor data,” Sadhu said. “We only used CGM to trend the glucose and then adjust insulin based on intermittent blood glucose readings. We were in a nice position to use the technology to its greatest advantages, and also be cautious about patient safety.”

Primary outcomes were feasibility and accuracy for trending blood glucose values. Secondary outcomes included reliability and nurse acceptance.

Researchers calculated mean absolute relative difference (MARD), Clarke error grid analysis (EGA) and Bland-Altman (B&A) plots for accuracy of paired sensor glucose and finger-stick blood glucose measurements. To determine accuracy, 437 paired sensor glucose and finger-stick blood glucose readings were analyzed.

Fewer finger sticks; staff acceptance

Researchers found that both CGM systems were feasible and reliable with good nurse acceptance. For Medtronic, the MARD was 13.1% with 100% of readings in zones A and B on Clarke EGA. For Dexcom, MARD was 11.1% with 98% of readings in zones A and B. B&A plots had a mean bias of –17.76 mg/dL (Medtronic) and –1.94 mg/dL (Dexcom).

“When we added the sensors, I was blown away to see that it was feasible, that we could adapt this to work, because all of this technology is built for outpatient use,” Sadhu said. “I’ll admit, in the beginning, there was a week where I lived in the ICU. There was a learning curve. We had issues like the Wi-Fi not working or needing compatible equipment. Those early obstacles were painstaking, but it was worth it thereafter.”

Researchers also demonstrated a 33.11% reduction in finger-stick blood glucose monitoring when CGM was used.

Sadhu said there are limitations to the technology, including a need for better integration into inpatient medicine. Staff will also need to adopt the technology, she said. A formal nursing survey was not completed; however, she said acceptance was high.

“The nurses, while they love this technology, they were a little skeptical,” Sadhu said. “Is it right? Can I rely on it? There will be a period of learning to trust and adopt it into their workflow. Most were thrilled and we had requests to use this with more patients, but we just didn’t have the equipment.”

The benefits, particularly during the pandemic, are “tremendous,” Sadhu said. As Houston Methodist prepares for a new surge of cases, she said additional research on CGM use in intensive care cannot come fast enough.

“Everyone is preparing for the worst and hoping for the best,” Sadhu said. “We hope this will go beyond the pandemic. We had MARDs of 11% and 13%, with inpatients who have numerous reasons to have erroneous or interfering variables for the technology. Compared with prior inpatient CGM trials with older models, it seems the technology had advanced considerably. As for how we use it, we still need to be cautious. First and foremost is patient safety. We have this blended strategy, using it for trends but adjusting insulin based on blood glucose readings, and it is ideal at this time. We can easily justify starting starting with this strategy now in the inpatient setting. These new versions [of CGM] can really pave the way for this to be a long-lasting practice.”

For more information:

Archana R. Sadhu, MD, FACE, can be reached at arsadhu@houstonmethodist.org.