Relative Hyperglycemia Is a Marker of Disease Severity in COVID-19

Patient in hospital bed with COVID, two nurses in PPE stand beside bed writing on clipboard

As in other illnesses, hyperglycemia on admission has been shown in COVID-19 to increase the risk of adverse outcomes during hospitalization. An additional measure has previously been validated to predict severe acute illness: the “glycemic gap,” the difference between current plasma glucose and estimated average glucose based on hemoglobin A1c (HbA1c).

Now, Brigham and Women’s Hospital researchers have extended the glycemic index’s utility to COVID-19. In a large multicenter cohort of adults with diabetes or hyperglycemia who were hospitalized for COVID-19, a higher glycemic gap was a stronger predictor of in-hospital mortality, mechanical ventilation, or ICU admission than either admission plasma glucose or HbA1c alone. Marie E. McDonnell, MD, diabetes section chief in the Division of Endocrinology, Diabetes and Hypertension at the Brigham, Donald C. Simonson, MD, MPH, ScD, a senior researcher in the Division, Nadine E. Palermo, DO, associate director of acute diabetes care in the Diabetes Management Program, and colleagues published the findings in The Journal of Clinical Endocrinology & Metabolism.


At the Brigham and four other academic hospitals, researchers retrospectively obtained data on 1,786 patients admitted between March 1, 2020, and February 28, 2021. They had PCR-confirmed COVID-19, a related diagnosis code, and either established diabetes or hyperglycemia measured at least twice during hospitalization.

Patient Characteristics

The cohort was 51% male, and the mean age was 66 years. The mean body mass index was 31.5 kg/m2, and diabetic ketoacidosis (DKA) was present on admission in 2.4% of patients.

Patients had evidence of both acute and chronic hyperglycemia, as the mean admission glucose was 12.0 mmol/L and the mean HbA1c was 8.07%. The mean glycemic gap was 1.7 mmol/L, indicating the cohort was more acutely hyperglycemic at admission than the mean HbA1c indicated.

Primary Outcome

The primary outcome of the analysis was death during hospitalization, which occurred in 10.6% of cases. In an adjusted multivariable model, the statistically significant independent predictors of in-hospital mortality were:

  • Age—adjusted odds ratio (aOR), 1.03 per year
  • Glycemic gap above 0—aOR, 1.06 per mmol/L
  • BMI—aOR, 1.04 per kg/m2
  • DKA on admission—aOR, 3.56

Protective factors were:

  • Higher estimated glomerular filtration rate—aOR, 0.97 per mL/min/1.73 m2
  • Outpatient use of any diabetes medication—aOR, 0.41

Secondary Outcomes

Secondary outcomes were the need for mechanical ventilation (MV), required for 23% of the cohort, and ICU admission, required for 39%.

Significant independent predictors were:

  • Glycemic gap above 0—aOR, 1.06 per mmol/L for MV; 1.06 per mmol/L for ICU admission
  • BMI—aOR, 1.03 per kg/m2 for MV
  • DKA on admission—aOR, 13.6 for MV; all patients with DKA were admitted to the ICU

Protective factors were:

  • Younger age—aOR, 0.99 per year for MV
  • Any diabetes medication—0.69 for ICU admission
  • Male sex—aOR, 0.63 for MV; 0.72 for ICU admission
  • Use of angiotensin-converting enzyme inhibitor or angiotensin receptor blocker—aOR, 0.54 for MV

Guidance for Clinicians

The glycemic gap can be calculated easily at admission for people with diabetes or stress hyperglycemia. These findings suggest it could be applied to triage patients among different levels of care (e.g., floor or ICU). Whether the glycemic gap actually contributes to poor clinical outcomes remains to be determined.

Leave a Reply