Deep Dive Quantifies Risk of COVID-19 Death in Both Diabetes Types

Miriam E. Tucker

May 20, 2020

The risk of death from COVID-19 is significantly increased in people with both main types of diabetes, although this is strongly influenced by age, as well as A1c and obesity, new pre-print data from the UK show.

The findings — which have not yet been peer-reviewed — appeared in two papers, both from the National Health Service (NHS) England, on May 20. They show an approximately two-to-threefold higher risk of mortality from the infection, with the highest risk in type 1 diabetes, even after adjustment for known confounders such as sex, ethnicity, deprivation, and pre-existing comorbidities.

"This is the most detailed diabetes data that's out there at the moment," Partha Kar, MD, National Specialty Advisor, Diabetes for NHS England, and a coauthor of both studies, told Medscape Medical News.

The new data follow an NHS announcement on May 14 that indicated diabetes is the most common underlying condition among those succumbing to COVID-19; 26% of people with the virus who had died in hospital in England also had diabetes.

This "sheds much-needed light on which groups of people with diabetes are more likely to experience poor outcomes if they catch coronavirus," said Bridget Turner, director of policy at Diabetes UK, in a statement.

It shows that the risk of death for people with diabetes "is higher than for people without the condition — with the risk for people with type 1 being higher than for those with type 2," and that a history of higher blood glucose levels, as well as obesity, seem to be contributing factors.

Turner stressed, however, it's important to note that "poorer outcomes are very strongly linked to older age."

The numbers of people with all types of diabetes dying in hospital from coronavirus under the age of 40 "were incredibly small, suggesting the risk for younger people is considerably lower," she added.

Kar, a consultant in diabetes & endocrinology at Portsmouth Hospitals NHS Trust, noted that the new data also "show what clinicians can do to improve the two modifiable factors, weight and blood glucose."

"There should be renewed focus on both. The pandemic isn't going anywhere in a rush...This shows the importance of what can be done if we can get on top of those things to reduce the risk."

And he stressed that there is a U-shaped association, too, between A1c and risk of COVID-19 death.

"You want to have an optimal [blood glucose] range — it's not about seeing how low [you can go], so that's an important message," he added.

Dying Risk Increased With Diabetes, But Age a Major Modifier

Previous studies from China and the United States have strongly linked in-hospital hyperglycemia with COVID-19 mortality but have not examined the potential effect of prior glycemic control on COVID-19 prognosis. 

The two new studies are the first to examine mortality in COVID-19 by diabetes type and the first to stratify the data by type, age, A1c, body mass index (BMI), and socioeconomic deprivation.

In the first study, Emma Barron, of Public Health England, York, and colleagues linked general practice data from March 1 to May 11, 2020 in England with in-hospital COVID-19 deaths from the COVID Patient Notification System, which was set up in March 2020.

Of 61,414,470 individuals in the GP registry alive as of February 19, 2020, 0.4% had a recorded diagnosis of type 1 diabetes and 4.7% of type 2 diabetes.

Of the total 23,804 COVID-19 deaths in England reported up to May 11, 2020, one third were in people with diabetes, including 31.4% with type 2 diabetes and 1.5% with type 1 diabetes. 

After multivariate adjustment, the odds of in-hospital COVID-19 death were 3.50 for those with type 1 diabetes and 2.03 for those with type 2 diabetes, compared to the population without known diabetes.

With further adjustment for cardiovascular comorbidities, the odds ratios were still significantly elevated in both type 1 (2.86) and type 2 (1.81) diabetes.

Age-wise, compared with people aged 60-69 years, the odds ratio for in-hospital death due to COVID-19 rose from 0.01 for those younger than 40 years to 9.14 for those aged 80 and above.

Kar stressed: "We have more in older age groups affected. Below the age of 40, the risk is really low. Below age 20, there were no deaths. And between 20 and 39, almost all [those who died] had comorbidities."

Odds were also higher for men versus women (1.94), those living in the most versus least-deprived quintiles of population (1.89), and for Asian and black ethnic groups (1.35 and 1.71, respectively) versus whites. 

U-Shaped Curves for A1c and BMI

The other paper analyzed national mortality data for people with diabetes from over 3 years ago, from January 1, 2017  to May 1, 2020, by Naomi Holman, PhD, of NHS England & NHS Improvement, London.

It showed a more than doubling in death rates for both diabetes types from the week ending April 3, 2020, compared to that which would have been expected based on previous years. 

Among 265,090 people with type 1 diabetes and 2,889,210 people with type 2 diabetes there were 418 and 9377 COVID-19–related deaths, respectively.

Older age, male sex, black ethnicity, and socioeconomic deprivation were again associated with increased COVID-19 mortality risk.

And U-shaped relationships were seen for both BMI and A1c in both diabetes types.

"The degree of hyperglycemia was strongly associated with risk of death related to COVID-19 after adjusting for other risk factors," say the researchers.

In type 2 diabetes, those with an A1c in the range of 59-74 mmol/mol (7.5%-8.9%) had a hazard ratio of 1.23 compared to those with an A1c of 48-53 mmol/mol (6.5%-7.0%).

The hazard ratio for those with an A1c ≥ 86 mmol/mol (10%) was 1.62.

A similar pattern was seen for type 1 diabetes, but the increased risk was only statistically significant for A1c ≥ 86 mmol/mol (10%) (hazard ratio, 2.19).

There was also a significantly increased risk of COVID-19 death among those with type 2 diabetes with A1c < 48 mmol/mol (6.5%) and a similar but nonsignificant risk at that A1c level in those with type 1 diabetes (hazard ratio, 1.22).

Kar emphasized: "Diabetes isn't a homogenous mass. It's very important to make the distinction between type 1 and type 2. They have different pathologies as far as the virus itself is concerned, whatever the mechanism. The treatment is very different. Clinicians need to understand that. This brings it more into sharp focus."

The U-shaped relationship was also seen with BMI: those with type 1 or type 2 diabetes who were underweight (BMI 20 kg/m2) had approximately double the risk of death from COVID-19 compared to those with a BMI of 25-29.9 kg/m2.

The researchers say, however, that confounding by factors associated with weight loss could play a role.  

For those with severe obesity (BMI ≥ 40 kg/m2) those odds ratios were 2.15 and 1.46 for type 1 and type 2 diabetes, respectively, compared with a BMI of 25-29.9 kg/m2.

"The elevated risk of COVID-19 in people with diabetes and severe obesity is marked and adds to evidence that obesity is an important risk factor for death from COVID-19 for which a number of possible mechanisms have been postulated," the researchers say.

Impaired renal function and history of previous hospital stays for stroke and heart failure also increased the COVID-19 mortality risk for both diabetes types, but neither history of myocardial infarction nor antihypertensive drug prescription did for either type of diabetes.

As has been reported previously, after adjustment for other risk factors, current smoking was associated with a lower risk of COVID-19 mortality in people with type 2 diabetes (0.63 vs nonsmokers).

Kar told Medscape Medical News that the NHS team next plans to look at ventilator use and at specific causes of death to see how much was due to thromboembolic or respiratory events, cytokine storm, and/or diabetic ketoacidosis.

Risk Stratification and Advice for Keeping People Safe at Work

As a result of these analyses, Diabetes UK is calling on the UK government to safeguard those with diabetes.

"Employers must put measures in place to keep people with diabetes safe, either by supporting people to work at home, or where this is not possible by putting people with diabetes on furlough, or by putting measures in place to allow stringent social distancing for those key workers who absolutely must be at work," said the organization.

Kar said, for now, "What this gives us is risk stratification in the population."

"If you have type 1 diabetes and you're a healthcare professional with an A1c that's not in the right zone, the question is should you be on the front lines? I think that's a question that has to now be raised since the risk is so high."

The corresponding author of both studies was Jonathan Valabhji, MBBS, MD, National Clinical Director for Diabetes and Obesity at NHS England & NHS Improvement, London, and both have been submitted for publication.    

NHS England & Improvement and Public Health England funded the first study, and NHS England & Improvement and NHS Digital funded the second.

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.