Hyperglycemia at hospital admission – regardless of diabetes status – is a key predictor of COVID-19-related death and severity among noncritical patients, new research from Spain finds.
The observational study, the largest to date to investigate this association, wasNov. 23 in Annals of Medicine by Francisco Javier Carrasco-Sánchez, MD, PhD, and colleagues.
Among more than 11,000 patients with confirmed COVID-19 from March to May 2020 in a nationwide Spanish registry involving 109 hospitals, admission hyperglycemia independently predicted progression from noncritical to critical condition and death, regardless of prior diabetes history.
Those with abnormally high glucose levels were more than twice as likely to die from the virus than those with normal readings (41.4% vs 15.7%). They also had an increased need for a ventilator and intensive care unit (ICU) admission.
“These results provided a simple and practical way to stratify risk of death in hospitalized patients with COVID-19. Hence, admission hyperglycemia should not be overlooked, but rather detected and appropriately treated to improve the outcomes of COVID-19 patients with and without diabetes,” Dr. Carrasco-Sánchez and colleagues wrote.
The findings confirm those of previous retrospective observational, but the current study “has, by far, the biggest number of patients involved in this kind of study [to date]. All conclusions are consistent to other studies,” Dr. Carrasco-Sánchez, of University Hospital Juan Ramón Jiménez, Huelva, Spain, said in an interview.
However, a surprising finding, he said, “was how hyperglycemia works in the nondiabetic population and [that] glucose levels over 140 [mg/dL] … increase the risk of death.”
Pay attention to even mild hyperglycemia from admission
The study also differs from some of the priorin that it examines outcome by admission glycemia rather than during the hospital stay, therefore eliminating the effect of any inpatient treatment, such as , he noted.
Although blood glucose measurement at admission is routine for all patients in Spain, as it is in the United States and elsewhere, a mildly elevated level in a person without a diagnosis of diabetes may not be recognized as important.
“In patients with diabetes we start the protocol to control and treat hyperglycemia during hospitalization. However, in nondiabetic patients blood glucose levels under 180 [mg/dL], and even greater, are usually overlooked. This means there is not a correct follow-up of the patients during hospitalization.
“After this study we learned that we need to pay attention to this population … who develop hyperglycemia from the beginning,” he said.
The study was limited in that patients who had previously undiagnosed diabetes couldn’t always be distinguished from those with acute “stress hyperglycemia.”
However, both need to be managed during hospitalization, he said. “Unfortunately, there is high variability in inpatient glucose management. The working group of diabetes of the Spanish Society of Internal Medicine is working on specific protocols,” said Dr. Carrasco-Sánchez.
All-cause death, progress to critical care higher with hyperglycemia
The retrospective, multicenter study was based on data from 11,312 adult patients with confirmed COVID-19 in 109 hospitals participating in Spain’s SEMI-COVID-19 registry as of May 29, 2020. They had a mean age of 67 years, 57% were male, and 19% had a diagnosis of diabetes. A total of 20% (n = 2,289) died during hospitalization.
Overall all-cause mortality was 41.1% among those with admission blood glucose levels above 180 mg/dL, 33.0% for those with glucose levels 140-180 mg/dL, and 15.7% for levels below 140 mg/dL. All differences were significant (P < .0001), but there were no differences in mortality rates within each blood glucose category between patients with or without a previous diagnosis of diabetes.
After adjustment for confounding factors, elevated admission blood glucose level remained a significant predictor of death. Compared to < 140 mg/dL, the hazard ratios for 140-180 mg/dL and > 180 mg/dL were 1.48 and 1.50, respectively (both P < .001). (Adjustments included age, gender,, diabetes, , lymphopenia, (hemoglobin < 10 g/dL), serum , C-reactive protein > 60 mg/L, lactate dehydrogenase > 400 U/L and >1000 ng/mL.)
Length of stay was 12, 11.5, and 11.1 days for those with admission blood glucose levels > 180, 140-180, and < 140 mg/dL, respectively (P = .011).
Use ofand admission to intensive care also rose with higher admission blood glucose levels. For the composite of death, mechanical ventilation, and/or ICU admission, odds ratios for 140-180 mg/dL and > 180 mg/dL compared with < 140 mg/dL were 1.70 and 2.02, respectively (both P < .001).
The study was supported by the Spanish Federation of Internal Medicine. The authors have reported no relevant financial relationships.
A version of this article originally appeared on.