Incidence and long-term outcomes of critically ill adult patients with moderate-to-severe diabetic ketoacidosis: Retrospective matched cohort study

The objective of this study was to describe the clinical outcomes and treatment intensity of adult intensive care unit (ICU) patients with moderate-to-severe diabetic ketoacidosis (DKA). We aimed also to compare their clinical course with matched non-DKA ICU controls and to identify prognostic factors for mortality and hospital readmission within 1 year. This is a retrospective matched cohort study. The settings are 2 tertiary teaching hospitals in Edmonton, Canada. Patients were adults with moderate-to-severe DKA admitted from January 2002 to December 2009. Control patients were defined as randomly selected age, sex, and Acute Physiology and Chronic Health Evaluation II score–matched nondiabetic ICU patients (1:4.5 matching ratio). Diabetic patients were stratified according to severity of exacerbation. None. From 2002 to 2009, the incidence of DKA per 1000 admissions was 4.59 (95% confidence interval [CI], 3.64-5.71). Severe DKA was associated with higher Acute Physiology and Chronic Health Evaluation II and Sequential Organ Failure Assessment scores in the first 3 days of ICU stay as compared with moderate DKA. Mechanical ventilation was received in 39%, vasopressors in 17%, and renal replacement therapy in 12% of DKA patients, respectively. One-year mortality and readmission rates were 9% and 36%. By logistic regression, death and/or readmission occurring in 1 year was independently associated with insulin use (odds ratio, 4.79; 95% CI, 1.14-20.05) and treatment noncompliance (odds ratio, 3.33; 95% CI, 1.04-10.64). Compared with matched non-DKA patients, those with DKA had lower mortality and were more likely to be discharged home. Diabetic ketoacidosis necessitating ICU admission is associated with considerable resource utilization and long-term risk for death. Interventions aimed to improve compliance with therapy may prevent readmissions and improve the long-term outcome.