Mean glycemic gap best predicts ICU mortality odds in adults with diabetes
Adults with diabetes admitted to two ICUs in China were less likely to survive if they had a larger gap between mean blood glucose 7 days after ICU admission and HbA1c-derived average glucose, according to study data.
“In this study, an elevated glycemic gap between the mean blood glucose level in the first 7 days after admission to ICU and HbA1c-derived average glucose was independently associated with 28-day mortality in critically ill patients with diabetes,” Li Jiang, MD, director of the intensive care unit at Xuanwu Hospital of Capital Medical University in China, and colleagues wrote in a study published in the Journal of Diabetes Investigation. “The predictive power was superior to the difference between the value of blood glucose at admission and HbA1c-derived average glucose.”
Researchers conducted a prospective observational cohort study of patients admitted to the ICU of two tertiary hospitals in Beijing between June 2017 and May 2020. Adults with type 2 diabetes estimated to stay more than 24 hours in the ICU without oral feeding were included in the study. Researchers obtained data on sociodemographics, insulin therapy before admission, comorbidities, blood glucose levels in the first 7 days of admission and Acute Physiology and Chronic Health Evaluation (APACHE) II score and HbA1c within 24 hours of admission. HbA1c was converted into HbA1c-derived average glucose to represent chronic average blood glucose within 3 months before ICU admission. Glycemic gap at admission was calculated as admission blood glucose minus HbA1c-derived average glucose. Mean glycemic gap was calculated as mean blood glucose level minus HbA1c-derived average glucose. The primary endpoint was mortality 28 days after admission.
There were 502 adults included in the analysis, of whom 192 died within 28 days of ICU admission and 310 survived. There were no significant differences in HbA1c and HbA1c-derived average glucose between the two groups. Those who died had higher levels of blood glucose at admission and mean glucose level than the survivor subgroup. Incidences of moderate hypoglycemia and severe hypoglycemia were more common among those who died and had a higher glycemic gap at admission and mean glycemic gap.
“Hypoglycemia may indicate the severity of acute illnesses and it seems prudent to prevent long-lasting hypoglycemia as much as possible by frequent and accurate blood glucose measurements and by use of a proper insulin protocol with safe and rational blood glucose range,” the researchers wrote.
Adults who died within 28 days of admission had lower daily intake of carbohydrates, higher daily doses of insulin and accumulated dosage of glucocorticoids during the first 7 days of admission compared with survivors. Ventilator-free hours and non-ICU hospitalization days were longer, and the duration of renal replacement therapy shorter for survivors compared with nonsurvivors.
Researchers plotted the area under the curve of APACHE II, glycemic gap at admission and mean glycemic gap within 7 days of admission. Mean glycemic gap had a greater predictive power for 28-day mortality (AUC = 0.82; 95% CI, 0.781-0.86) compared with APACHE II (AUC = 0.71; 95% CI, 0.663-0.757) and glycemic gap at admission (AUC = 0.601; 95% CI, 0.549-0.652). Adults with a mean glycemic gap higher than 3.6 mmol/L had significantly shorter survival odds compared with those with a mean glycemic gap lower than 3.6 mmol/L. HbA1c is not associated with survival likelihood 1 year after admission.
“More studies are required to evaluate various scoring tools for predicting mortality in ICU, such as the sequential organ failure assessment,” the researchers wrote.