Data from the CORONADO (French Coronavirus SARS-CoV-2 and Diabetes Outcomes) study also revealed that body mass index (BMI) was independently associated with death or intubation at 7 days, while A1c and use of renin-angiotensin-aldosterone system (RAAS) blockers and dipeptidyl peptidase–4 inhibitors were not.
The presence of diabetes-related complications and older age also increased the risk of death.
The findings were published online Diabetologia by Bertrand Cariou, MD, PhD, of the department of endocrinology at the Hôpital Guillaume et René Laennec in Nantes, France, and colleagues.
First study to examine specific characteristics at time of admission
Previous studies have linked diabetes to worse outcomes in COVID-19, but this is the first to examine specific characteristics before and at the time of hospital admission that predict worse outcomes among people with diabetes, study coauthor Samy Hadjadj, MD, PhD, said in an interview.
“Before the CORONADO study it was ‘all diabetes [patients] are the same.’ Now we can surely consider more precisely the risk, taking age, sex, BMI, complications, and [obstructive sleep apnea] as clear ‘very high-risk situations,’” said Dr. Hadjadj, of the same institution as Dr. Cariou.
Another clinical message, Dr. Hadjadj said, is that, “even in diabetes, each increase in BMI is associated with an increase in the risk of intubation and/or death in the 7 days following admission for COVID-19. So let’s target this population as a really important population to keep social distancing and stay alert on avoiding the virus.”
But he urged caution regarding the A1c finding. “A1c might be associated with admission to hospital but other factors far beyond A1c drive the prognosis as soon as a patient is admitted. It’s surprising but reasonable speculation can explain this.”
And Dr. Hadjadj said that no obvious signals were identified with regard to medication use.
“Insulin is not suspected of having adverse effects closely related to COVID-19. RAAS blockers are not deleterious but indicative of hypertension, which is a comorbidity even in diabetes patients,” he said. (None of the patients studied were taking sodium-glucose cotransporter 2 inhibitors or glucagonlike peptide receptor agonists.)
Yet again, high BMI emerges as a major risk factor
The study included 1,317 patients with diabetes and confirmed COVID-19 admitted to 53 French hospitals during March 10-31, 2020. Participants included 88.5% with type 2 diabetes, 3% with type 1 diabetes, and 3.1% newly diagnosed on admission. Mean age was 69.8 years.
Diabetes-related disorders on admission were reported in 11.1% of participants overall. These included 132 episodes of severe hyperglycemia, including 40 of ketosis, of which 19 were ketoacidosis, and 14 hypoglycemic events. Severe anorexia was reported in 6.3%.
The composite primary endpoint, tracheal intubation for mechanical ventilation and/or death within 7 days of admission, occurred in 29% of patients (n = 382).
Of the secondary outcomes, 31.1% (n = 410) were admitted to ICUs within 7 days of hospital admission, including 20.3% (n = 267) who required tracheal intubation for mechanical ventilation.
On day 7, 10.6% (n = 140) had died and 18.0% (n = 237) were discharged.
In the univariate analysis, the primary outcome was more frequent in men (69.1% vs. 63.2%; P = .0420) and those taking RAAS blockers (61.5% vs. 55.3%; P = .0386). Median BMI was significantly higher in those in whom the primary outcome occurred (29.1 vs 28.1 kg/m2P = .0009),
Other characteristics prior to admission associated with risk of death on day 7 included age, hypertension, micro- and macrovascular diabetes-related complications, and comorbidities such as heart failure and treated obstructive sleep apnea.
Over 40% of those admitted had such complications. Of the patients analyzed, microvascular complications (eye, kidney, and neuropathy) were present in 47% and macrovascular complications (arteries of the heart, brain, and legs) were present in 41%.
Encouragingly, there were no deaths in patients aged under 65 years with type 1 diabetes, but only 39 participants had type 1 diabetes. Other work is ongoing to establish the effect of COVID-19 in this specific population, the researchers wrote.
Among prior medications, metformin use was lower in people who died, while insulin use, RAAS blockers, beta-blockers, loop diuretics, and mineralocorticoid-receptor antagonists were associated with death on day 7. The medication findings didn’t reach statistical significance, however.
When asked about the hint of a protective effect of metformin (odds ratio, 0.80; P = .4532), given that some experts have advised stopping it in the setting of COVID-19 because of the risk of lactic acidosis, Dr. Hadjadj said he wouldn’t necessarily stop it in all patients with COVID-19, but said, “let’s stop it in cases of severe condition.”