Battlefield Endocrinology

I have learned that we are all soldiers in this war against a viral enemy. As an endocrinologist, I did not consider myself “frontline,” but after witnessing the havoc of the coronavirus disease of 2019 (COVID-19) pandemic, I realized we all have a duty to fight.

The first consultation I received for an intensive care unit (ICU) COVID-19 patient required insulin infusion of 15 to 20 U/h. He had type 2 diabetes with many usual suspects for his severe insulin resistance: critical illness, acute respiratory distress syndrome (ARDS) with inflammatory storm, body mass index 43, hemoglobin A1c 9.3%, corticosteroids, vasopressors, and enteral nutrition. When I went to the ICU, I was quickly asked to see several others with extraordinarily similar circumstances. In that moment, I was awestruck. The ICU staff appeared as if they were on a battlefield caring for wounded combatants all injured simultaneously. Familiar physicians and nurses were now unrecognizable in their new armor of masks, gowns, and helmets; the personal protective equipment (PPE) of the health care worker (HCW). Open hallways where rounds were held were now crammed with tables of PPE, intravenous poles, medication pumps, and even ventilator control panels. Usually in the room, they were now strung outside with stretched tubing and electrical extensions in attempts to minimize the HCW in the line of fire. The noise of the alarms from the equipment completed the scene of battlefield medicine.

How can endocrinologists have an impact in this ICU setting? I realized that the orders we routinely place now increase the risk of virus transmission to the HCW. Evidence-based insulin infusion protocol aims for glucose targets of 140 to 180 mg/dL.^1,2 However, the protocol I have implemented for almost 15 years was now inadequate as it requires hourly blood glucose (BG) testing, and therefore, frequent exposure and consumption of PPEs. My first efforts were to revise this protocol to be more forgiving and protect the HCW, without harming the patient. While this was helpful, it still required at least 12 BGs per patient per day. I then began to consider how continuous glucose monitoring (CGM) systems could be added to our armamentarium. The reality though is that there is not a Food and Drug Administration (FDA) approved system for inpatient use. Despite advocacy to allow its use in the hospital on select patients, it is never recommended in critically ill patients.³ But as I watched each donning and doffing dwindling the PPE supplies, often just to obtain a BG, I could not convince myself that there was no role for CGM in this wartime.

Despite 15 years of inpatient clinical trials, there have not been sufficient data to show accuracy or improved patient outcomes with CGM.⁴ Some small studies have demonstrated benefits in hypoglycemia and septic shock,^5,6 but we are still left most often using a blood glucose meter, also not FDA approved in critically ill patients, with now the added liability for the HCW.⁷

I initiated discussions with our institution’s IRB and quality and safety committees regarding emergent approval for a feasibility and quality improvement study. In collaboration with two CGM manufactures, we obtained systems to begin our project. At this time, the FDA issued a statement that while still not “approved,” CGM is “allowed” in hospitals during the pandemic. Working closely with the ICU staff, IT, and the manufacturers, we were able to get the systems successfully adapted in the ICU environment to display CGM values and trends every five minutes outside the patient’s room. Our aim was to establish that the trend in CGM glucose was congruent with the trend in BG, thereby allowing us to extend the BG monitoring, when the trend remained in our target range. Due to accuracy concerns, we only used BG to dose insulin therapy. Patients were excluded if there was diabetic ketoacidosis, extracorporeal membrane oxygenation, pH < 7.0, or on more than one vasopressor.

We are gratified that the technology is not only emergently adaptable but also trends the BG reliably to increase the intervals between testing. In addition, it predicts negative trends, both hypoglycemia and hyperglycemia, earlier than the scheduled BG testing intervals would allow (Figure 1). Nurses performed sensors’ placements easily along with other bundled tasks. They were enthusiastic to incorporate CGM into their workflow to monitor for rapid unexpected changes and perform BG testing when most appropriate, instead of prespecified intervals. In many patients, if the BG was in our desired range and the CGM showed a stable trend, we were able to decrease BG to every four hours while on insulin infusion.

Figure 1.

Top: Remote monitoring with two different continuous glucose monitoring systems outside the patient room. Bottom: Trending of blood glucose values (red tear drops) and interstitial glucose (blue line).

In the future, I predict that CGM will be as essential and routine as cardiac monitoring in the hospital. During this pandemic, CGM is valuable and provides intelligence into glucose trends so that we may prevent unnecessary HCW risk and preserve PPEs. In battlefield endocrinology, this may be the most impactful wartime strategy we can contribute.