COVID-19 and Diabetes: Known Mechanisms and a 'New Beast'?

Miriam E. Tucker

April 14, 2020

Editor's note: Find the latest COVID-19 news and guidance in Medscape's  Coronavirus Resource Center.

Diabetes may be an independent risk factor for rapid progression and poor prognosis of COVID-19 through several known pathways and a possible new one: direct damage to pancreatic islets.

As with nearly every finding related to the novel 2019 coronavirus disease, the evidence thus far comes primarily from case series and anecdotal reports, and as such has obvious limitations.

In one of the latest case series, Chinese researchers compared 137 individuals without diabetes to 37 with diabetes, and then compared the subsets of those two groups without comorbidities (26 and 24 people, respectively).

In the latter comparison, those with diabetes still had a significantly greater risk of severe pneumonia, release of tissue injury-related enzymes, excessive uncontrolled inflammatory responses, and hypercoagulable state associated with dysregulated glucose metabolism, compared to those with no diabetes.

And those with diabetes also had higher levels of biomarkers suggesting an inflammatory "storm" preceding rapid deterioration of COVID-19, say Weina Guo, of Huazhong University of Science and Technology, Wuhan, China, and colleagues in their article, published online March 31 in Diabetes Metabolism Research and Reviews.

"While results of this study should be read in the light of some limitations — such as the low sample size and the large age difference between study groups when comorbidities were excluded — it still provides relevant insights that could inform about how COVID-19 interacts with pre-existing conditions," note Ernesto Maddaloni, MD, PhD, and Raffaella Buzzetti, MD, both of Sapienza University of Rome, Italy, in an accompanying editorial.

Possibly "Alarming" Interaction, Given Initial Symptoms May Be Silent

The editorialists note it has previously been observed that diabetes worsened outcomes of other similar viral infections, including the 2003 SARS coronavirus and 2009 H1N1 influenza.

Such an interaction in the current pandemic would be "alarming considering the high transmission rate of SARS-CoV-2 [the virus that causes COVID-19] and the global prevalence of diabetes," they emphasize

They also point to another concern from the Chinese article, namely that "the severity of COVID-19 in diabetes may be hidden by an initial milder presentation of SARS-CoV-2 infection."

Fewer patients with diabetes than those without experienced fever (59.5% vs 83.2%; P = .002), chills (56.8% vs 71.5%; P = .08), chest tightness (13.5% vs 29.2%; P = .05), and shortness of breath (13.5% vs 27%; P = .08).

"This phenomenon, which resembles the silent symptoms people with diabetes experience also in other conditions such as myocardial infarction, may cause a life-threatening delay in providing the needed care, finally resulting in poorer prognosis," they caution.

"Based on these results, it is crucial not to underestimate COVID-19 severity in patients with diabetes even in the absence of classical worrisome signs and symptoms and it may be useful to develop different clinical severity scores for patients with diabetes."

Is the Virus Attacking the Pancreas?

Guo and colleagues also report in their article that diabetes control worsened during hospitalization.

They write: "29.2% of the patients took insulin before and increased the dose of insulin after admission, and 37.5% of the patients took oral medicine before admission and started insulin therapy after admission, which meant that patients had poor glycemic control during hospitalization."

Although acute inflammatory states and acute stress responses can certainly raise glucose levels, the authors suggest another possibility: that the SARS-CoV-2 virus may actually damage the pancreatic islet cells.

They point to a study of the first SARS 2003 coronavirus outbreak in China in which diabetes developed within 2 weeks of hospitalization in 20 of 39 patients who did not previously have diabetes. For the most part, this was transient: only six patients still had diabetes at discharge and two still had it at 3 years. 

Immunostaining in a patient who had died showed strong staining of the angiotensin-converting enzyme 2 (ACE2) protein — the coronavirus binding site — in the islets but not the exocrine pancreatic tissue, "which means coronavirus might cause diabetes by damaging islets seriously," Guo and colleagues suggest.

Anecdotally, physicians report seeing both dramatic worsening of glycemic control in COVID-19 patients with pre-existing diabetes as well as new-onset cases.

On April 4, endocrinologist Joshua D. Miller, MD, medical director of diabetes care at Stony Brook Medicine, New York, tweeted the following: "Based on what we are seeing among acutely ill pts, I wonder if #COVID-19 causes insulin deficiency. So many pts requiring markedly elevated #insulin drip rates (some as high as 30-40 units/hr). I don't think just pressor/steroid related. Something else is going on here..."

And then in a subsequent tweet: "This is different than usual insulin resistance/pancreatic insufficiency in critical illness. It's a new beast."

And in response to an inquiry from Medscape Medical News, he wrote: "We've not seen this pattern of glycemia with associated insulin requirements before. The degree of glucose toxicity is profound and independent of preadmission diabetes control. Patients with seemingly well-controlled diabetes at home with A1cs in the single digits also experience similar degrees of dysglycemia. We are also seeing a subset of patients presenting in euglycemic DKA [with] normal lactate and other causes of ketoacidosis ruled out."

Miller said he's seeing the phenomenon in type 1 and type 2 diabetes, as well as new-onset diabetes in patients with COVID-19, and "lots of euglycemic [diabetic ketoacidosis], too."

And in an interview with Medscape Medical News, endocrinologist Jean-François Gautier, MD, PhD, of Lariboisière Hospital, Paris, said he's been observing the same thing.

At his center, which has been completely converted to provide care for COVID-19, "We have a lot of new-onset diabetes in infected patients...Our new-onset acute diabetes are all discovered in the context of COVID-19 positivity."

And, Gautier added, about 30% admitted to his hospital with COVID-19 already have diabetes.

Most have type 2 diabetes, but one patient admitted to the emergency department had severe respiratory distress and was found to have diabetic ketoacidosis. "So we started insulin right away. This was new-onset type 1 diabetes. The virus was the precipitating factor," he said.

Gautier is now part of a collaboration of about 10 to 15 hospitals throughout France collecting data on inpatients with COVID-19 and diabetes to assess determinants of severity, including outcomes and how many patients end up in intensive care. As of early April, they had included around 400 patients.

Probably Nothing New...

But Jorge Plutzky, MD, director of the Vascular Disease Prevention Program and director of preventive cardiology at Brigham and Women's Hospital, Boston, Massachusetts, isn't convinced that the phenomenon is novel.

"At the end of the day, it's not particularly surprising. We have seen that with other infections and other diseases like heart disease. People with diabetes tend to do worse and have worse outcomes. One of the drivers of that is thought to be a chronic inflammatory condition."

"The complications of COVID-19 are thought to be inflammatory, so potentially it would be expected that people with diabetes would do worse," he said.

And regarding the extreme hyperglycemia, Plutzky said: "Acute severe inflammatory states and acute stress responses will raise glucose. That often involves the liver, and we know that the infection is hypercoagulable as well."

"Whether it's an unusual mechanism that is targeting insulin release from the pancreas as opposed to just insulin resistance, which is a much more common state, it's very hard to know."

"I think it's an interesting observation but it's important to be cautious because the data are limited and changing very quickly, there are many different variables involved [and] we're looking at small datasets."

On the issue of the new-onset diabetes cases, Plutzky said: "You would need to look for that to be occurring in the absence of prediabetes. Those would be the people you would tip over in the setting of all else that's going on."

Could Targeting Inflammation Help Treat COVID-19 in Diabetes?

In Guo and colleagues' analysis excluding patients with comorbidities, the diabetes group was significantly older (61 vs 32 years) and had more nausea and vomiting (16.7% vs 0%) and higher mortality (16.7% vs. 0%). There were no differences by gender or other baseline symptoms. The diabetes group also had higher CT imaging scores.

Levels of lactic dehydrogenase, α-hydroxybutyrate dehydrogenase, alanine aminotransferase, and γ-glutamyltransferase were elevated in the patients with diabetes, indicating injury of the myocardium, kidney, and liver, the authors note. "This result is consistent with the extensive distribution of SARS-CoV-2 receptors ACE2, and can also partially explain why some patients died from multiple organ failure," they add.

In addition, levels of total protein, albumin, prealbumin, and hemoglobin were significantly lower in patients with diabetes versus without diabetes (all P < .01, P = 0.02 for prealbumin), which means patients with diabetes are more likely to be undernourished, they write.

Markers of inflammation, including C-reactive protein, erythrocyte sedimentation rate, and interleukin 6 (IL-6) were also higher in the diabetes group (all P < 0.01).

These findings point to potential treatments, say the editorialists.

The monoclonal antibody tocilizumab, approved for autoimmune disorders including severe rheumatoid arthritis and giant cell arteritis, targets overexpression of IL-6. It is being used off-label at some centers to treat patients with COVID-19 and is included in a number of ongoing trials in COVID-19.

If the observations in the current study are confirmed, they say, this drug could be particularly helpful in patients with diabetes, along with others targeting the same pathway, including siltuximab, or janus kinase inhibitors such as baricitinib, tofacitinib, and upadacitinib.

"Fast and efficient trials are urged to give evidence-based answers to all these unanswered questions," the editorialists write.  

Plutzky told Medscape Medical News his institution has protocols looking at these drugs in patients with COVID-19, including some with diabetes.

None of the sources in this story have reported any relevant financial relationships.

Diabetes Metab Res Rev. Published online March 31, 2020. Article, Editorial

For more diabetes and endocrinology news, follow us on Twitter and Facebook.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.