1274-PUB: Clinical Comparison of Hospitalized Patients with Diabetic Ketoacidosis Admitted to Intensive Care Unit (ICU) vs. Non-ICU Areas

Abstract

Many patients with DKA are admitted to Intensive Care Units (ICUs) due to restriction of insulin infusions outside of ICU areas. Our previous retrospective study showed that 37.0% of patients did not require admission to ICUs. Notably, reduced ICU admissions decrease costs and preserve bed availability for critical patients. Aim: To compare the clinical characteristics and hospital care among patients with DKA admitted to the ICU to those admitted to the non-ICU areas. Methods: A secondary analysis of collected data from the previous DKA study included 170 patients admitted from April-December 2017. Results: Of 170 emergency department (ED) patients, 96 (56.5%) were admitted to an ICU and 74 (43.5%) to non-ICUs; 72 (42.4%) had T1D, 82 (48.2%) had T2D, and 16 (9.4%) had other types of DM. Mean age was 46.9±16.0 years old. The length of ED stay was significantly shorter (4.68±2.4 vs. 9.5±4.4 hours, p<0.001) and the length of hospital stay significantly longer (6.2±11.2 vs. 3.1±3.9 days, p=0.03) in the ICU group than in the non-ICU group. Acute kidney injury (p=0.01), stroke and deep vein thrombosis (p=0.05), and substance abuse (p=0.03) were more common in the ICU group. Mean blood glucose, pH, anion gap, ketone, potassium, and lactate levels were significantly higher, while bicarbonate levels were significantly lower in the ICU group. Insulin (96.9%) and fluid (97.9%) infusions were more commonly used in the ICU group. Patients in the ICU group received more endocrine (p<0.001) and dietitian (p<0.006) consultations than those in the non-ICU group. Conclusions: Patients in the ICU group had more severe metabolic acidosis and other critical illnesses. Patients with uncomplicated DKA can be safely treated in the ED. Insulin drips should not be the criteria for ICU admission. Further study along with these data can help with the development of a severity scoring tool to identify the appropriate patients for ICU admissions. Disclosure A. Liu: None. K. Carmichael: None. M. Schallom: None. C. Arroyo: None.