Doctors noticed something else as well: patients who had no history of diabetes sometimes developed severe diabetic symptoms while battling COVID, and some remained diabetic after COVID resolved. According to a meta-analysis of eight studies published in late November, as many as 14 percent of hospitalized COVID patients developed what appeared to be “new onset” diabetes.
Can the pandemic virus, SARS-CoV-2, directly trigger diabetes, or does something else explain these COVID-related cases? That question is the subject of a heated scientific debate and investigations that may ultimately lead to a better understanding of both ailments.
Muddying the waters are several confounding factors such as the fact that any acute illness can disturb glucose metabolism, that COVID treatment can also impact blood sugar, and conflicting data on whether or not SARS-CoV-2 can invade insulin-producing cells in the pancreas.
“The relationship between COVID-19 and diabetes is very complex,” says Francesco Rubino, chair of metabolic and bariatric surgery at King’s College London, “and it might involve more than one issue.”
To help solve the puzzle, Rubino—along with two King’s colleagues and Paul Zimmet, professor of diabetes at Australia’s Monash University and a co-author of the meta-analysis—have established an international registry called CoviDIAB to compile extremely detailed case histories of new-onset diabetes in COVID patients. Their announcement of the registry in an August issue of the New England Journal of Medicine drew responses from hundreds of clinicians around the world and, so far, about 165 completed case reports have been submitted. Preliminary analysis of the data will likely begin when at least 200 cases are registered, Rubino says.
In the meantime, he, Zimmet and other experts point to at least five explanations for the sudden appearance of diabetic symptoms in patients with COVID. All of them may be playing a role.
1. The virus may directly attack insulin-producing beta cells in the pancreas. Diabetes is fundamentally a disease of insufficient production of, or response to, insulin, the hormone that enables cells to use glucose as a fuel. In type 1 diabetes, which often strikes in childhood or adolescence, people lack the capacity to produce insulin because the beta cells in the pancreas have been destroyed by antibodies that target the body’s own proteins. In type 2 diabetes, the more common form, body cells have become less sensitive to insulin and beta cells are depleted or dysfunctional.
Perhaps the biggest point of contention about diabetes and COVID is whether or not SARS-CoV-2 directly attacks and destroys the specialized beta cells in the pancreas that produce insulin. There is evidence for and against this idea. For example, a study conducted last year at Cornell University showed that insulin-producing cells cultured in a lab express ACE2 receptors—the key doorway through which SARS-CoV-2 enters human cells—and that the virus can invade these cells. A 2010 study also found ACE2 on beta cells and suggested that the earlier SARS-CoV virus could use the receptors to enter and destroy those cells. Zimmet says that there is further evidence from postmortem studies of COVID patients showing the destruction of pancreatic beta cells. “I won’t say I am 100 percent convinced, but it’s a very, very plausible explanation,” he says.
Others are less persuaded. A study led by researchers at Vanderbilt University and published in Cell Metabolism in December searched for expression of the ACE2 protein in beta cells and found only negligible amounts. A second protein called TMPRSS2 that also plays a role in coronavirus entry to cells was largely absent as well. “We really thought this might explain how the virus got into beta cells, but we did not find the necessary proteins there” says Alvin C. Powers, director of the Vanderbilt Diabetes Center and a senior author of the study. “Negative findings are less exciting but very important. We are confident of our results.” He notes that a second study published in the same issue of Cell Metabolism came to the same conclusion.
2. The virus may indirectly attack insulin production. While scientists may disagree about whether SARS-CoV-2 can directly enter beta cells, there is evidence suggesting that it can attack other parts of the pancreas. Both of the Cell Metabolism studies found that viral entry proteins were expressed elsewhere in the pancreas and in the small blood vessels that nourish beta cells.
“One could envision a scenario in which the virus could affect these micro blood vessels and beta cells could die,” Powers suggests. Or it could infect other areas of the pancreas inducing inflammation that disrupts insulin production, he adds.
The virus might also bring on diabetes by attacking or inflaming other organs and tissues that are involved in glucose metabolism. ACE-2 receptors are plentiful in the intestines, blood vessels and liver, Rubino notes: “What happens if [viral infection] creates a dysfunctional intestine that doesn’t do its normal job in regulating blood sugar levels? What happens if the virus interferes with liver function, which is also so crucial?”
Even more concerning, Rubino says, is a scenario in which the virus enters several organs at once, creating multiple disruptions. “That could end up creating forms of diabetes that we haven’t seen before. Not type 1 or type 2 but something in between, something atypical.”
3. Acute illness and inflammation are causing symptoms of diabetes. Doctors have known for decades that any severe health event—pneumonia, heart attack, stroke, trauma—can cause blood glucose levels to spike, a condition called hyperglycemia that is a signature of diabetes. Stress-related hormones such as cortisol and adrenaline are believed to cause this elevation, which may subside when the patient recovers or may leave the patient permanently diabetic.
There is no doubt that severe COVID can impose the kind of stress that raises blood glucose in patients who have no history of diabetes, and sends it sky high in those who do.
Endocrinologist Alyson Myers sees this phenomenon daily in her role as medical director of inpatient diabetes at North Shore University Hospital in Manhasset, N.Y. Patients admitted there with COVID, she says, rarely have blood sugar levels in the normal range, which is below 140 milligrams of glucose per deciliter of blood. “They are usually coming in in the 200s,” whether they have a history of diabetes or not, and some arrive in an especially dangerous, hyperglycemic state called diabetic ketoacidosis, more typically seen in type 1 diabetes. “So, it’s not just new-onset, but new-onset of this severe form,” Myers says.
Hyperglycemia on admission is a predictor of mortality, Myers says, “so you want to get that sugar down as quickly as possible.” It’s not unusual for hospitalized COVID patients to be given very large doses of insulin, even if they never required it in the past.
4. Treating COVID with steroids raises blood sugar. A standard treatment for severe COVID-19 at Myers’ hospital and many others is a combination of the anti-viral drug remdesevir and high doses of a steroid drug such as dexamethasone, which tamps down inflammation. The latter drug, however, raises insulin resistance and may therefore make hyperglycemia even worse.
This treatment, too, is a reason that COVID patients may suddenly develop severe symptoms of diabetes. “Between the COVID and the steroids, their blood sugar is through the roof,” Myers says, “and we have to give them really high doses of insulin to combat that.”
5. New-onset diabetes might not actually be all that new. The fact that a patient has no recorded history of diabetes does not mean that they weren’t already diabetic or pre-diabetic or predisposed to the disease by virtue of genetics, obesity or some other factor.
All of these conditions are remarkably common. In the U.S., for example, the Centers for Disease Control and Prevention estimates that 10.5 percent of people have diabetes, one fifth of whom have not yet been diagnosed. Another 34 percent of the adult population has elevated blood sugar in the prediabetic range.
“Diabetes is typically a silent disease for a very long time,” Rubino says. “Estimates are that you may have it for five or more years without knowing it.”
One way to tell whether silent diabetes was already present in COVID patients, Rubino notes, is with a commonly used blood test called A1C that indicates average blood sugar levels for the previous three months: “A normal A1C allows you to be more confident that there wasn’t any diabetes two or three months ago.” Where available, A1C data will be an illuminating component of the CoviDIAB registry, as will follow-up data showing whether COVID-related diabetes vanishes as suddenly as it arose or if it persists.
Understanding precisely how the coronavirus disrupts glucose metabolism could help resolve longstanding questions about the role other infections play in diabetes. Viruses such as Coxsackie B and rubella are known to be associated with some cases of type 1 diabetes, but small data sets have made it difficult to pin down a mechanism. “With a pandemic we will probably see more cases than we’ve ever seen before,” Rubino says. “That’s why the story of COVID and diabetes is important for the understanding the role of viruses in causing diabetes.”