CAUSES AND PREDICTORS FOR 30-DAY RE-ADMISSIONS IN ADULT PATIENTS WITH DIABETIC KETOACIDOSIS IN THE UNITED STATES: A NATIONWIDE ANALYSIS, 2010-2014.

OBJECTIVE: We aimed to determine the causes and predictors for 30-day re-admission following a hospitalization for diabetic ketoacidosis (DKA) in the United States. METHODS: This retrospective cohort study analyzed data from the National Re-admission Database. We included adult patients with a primary discharge diagnosis of DKA, from 2010 to 2014. Our primary objective was to determine the frequency and causes for 30-day re-admission after an index hospitalization for DKA. We also performed multivariate regression analyses using covariates from the index admission to identify predictors for 30-day re-admissions. RESULTS: Among 479,590 admissions for DKA, 58,961 (12.3%) were re-admitted within 30 days. Recurrent DKA represented 40.8% of all-cause re-admissions. In multivariate analysis, end-stage renal disease (odds ratio [OR], 2.13; 95% confidence interval [CI], 2.00 to 2.27; P<.001), Charlson Comorbidity Index ≥3 (OR, 2.49; 95% CI, 2.42 to 2.58; P<.001), discharge against medical advice (OR, 1.97; 95% CI, 1.86 to 2.09; P<.001), and drug use (OR, 1.78; 95% CI, 1.71 to 1.86; P<.001) were the most significant predictors for 30-day re-admission. About 50% of patients were re-admitted within 2 weeks after discharge. CONCLUSION: In the U.S., about one in every eight patients with DKA is re-admitted within 30 days, with 40.8% representing recurrent DKA episodes. Patients with end-stage renal disease, high comorbidity burden, drug use, and/or leaving against medical advice represented the highest risk group for re-admissions. Future studies with interventions focusing on high-risk population are critically needed. ABBREVIATIONS: AKI = acute kidney injury; BMI = body mass index; CCI = Charlson Comorbidity Index; CI = confidence interval; DKA = diabetic ketoacidosis; DM1 = type 1 diabetes mellitus; DM2 = type 2 diabetes mellitus; ESRD = end-stage renal disease; ICD-9-CM = International Classification of Diseases, Ninth Edition, Clinical Modification; IQR = interquartile range; LOS = length of stay; NRD = National Re-admission Database; OR = odds ratio.

Duke Scholars

Author Joseph P. Mathew Anesthesiology, Cardiothoracic