Complications and Outcomes in Patients With Diabetic Ketoacidosis and COVID19

Introduction: COVID-19 infection caused by the novel coronavirus SARS-CoV-2 has been associated with new-onset diabetes and diabetic ketoacidosis. There are currently limited reports observing the effect of COVID-19 infection on outcomes in patients hospitalized with diabetic ketoacidosis (DKA). Methods: This retrospective study used electronic medical records, between March 1, 2020 to September 14, 2020, and included patients aged ≥ 18 years with DKA. Clinical characteristics, hospital course, and complications (ICD-10 codes) were compared between patients admitted with DKA and confirmed COVID-19 infection (COVID+) to those with DKA without COVID-19 infection (COVID-). The American Diabetes Association criteria was used to define DKA. Results: Among 757 patients with DKA, 80 COVID+ patients with and 677 COVID- patients were included (mean age 49.1 years, female 52.7%, African-American 70.4%). Baseline characteristics were similar between groups. The COVID+ group had a lower admission glucose (mean glucose 413.1 ± 218.7 mg/dL vs 441.1 ± 222.6 mg/dL, p-value 0.05), and higher serum bicarbonate levels (18.2 mmol/L vs 16.6 mmol/L, p-value <0.05) compared to the COVID- group. Compared to the COVID- group, COVID+ required a longer time to achieve blood glucose < 250 mg/dl from admission (8.8 hours vs 7.1 hours, p-value <0.01), and had higher rates of acute respiratory failure (48.8% vs 9.5%, p-value <0.001), mechanical ventilation (32.5% vs 10.9%, p-value <0.001), sepsis (37.5% vs 15.7%, p-value <0.001), acute thrombotic events (11.3% vs 2.7%, p-value <0.05), and pneumonia (70% vs 10.3%, p-value <0.001). COVID+ had higher mortality (16.3% vs 3.1%, p-value < 0.001), ICU length of stay (median 5.5 days or 2–58 days vs median 2 days or 2–123 days, p-value <0.001), and total LOS (median 7.5 or 1–47 days vs median 5 days or 1–98 days, p-value <0.001). Both groups required similar total insulin doses throughout the entire hospitalization (mean 143 +/- (SD) vs 142 +/- SD units, p-value 0.088). Conclusion: Our data shows that COVID-19 infection is associated with increased mortality, worse clinical outcomes, and longer LOS among patients hospitalized with DKA, while presenting with less severe hyperglycemia and acidemia compared to those without COVID-19 infection.