The potential impact of admission insulin levels on patient outcome in the intensive care unit.

BACKGROUND: Blood levels of insulin in patients with critical illness at admission to the intensive care unit (ICU) and its association with in-hospital mortality are not fully defined. Our objective was to determine this association in a cohort of patients with critical illness who attended in a mixed ICU. METHODS: Prospective cohort was nested in a randomized clinical trial conducted in a 12-bed mixed ICU in a tertiary hospital in Medellin (Colombia). One hundred sixty consecutively admitted patients, 15 years or older, were analyzed. Blood insulin and blood glucose levels were measured at admission to the ICU, as well as Acute Physiology and Chronic Health Evaluation II and Sequential Organ Failure Assessment scores. A logistic regression model was created with in-hospital mortality as the outcome. RESULTS: In-hospital mortality was 57 (35.6%) of 160. Survivors had lower Acute Physiology and Chronic Health Evaluation II (median, 13 vs. 17) and lower insulin levels (median, 6.5 vs. 9 µU/mL) than did nonsurvivors. More women than men died (27 [48.2%] of 56 vs. 30 [28.8%] of 104), and 39% of the deaths (n = 22) occurred in patients with sepsis. Patients with insulin levels greater than 15 µU/mL had a higher mortality rate compared with patients with values of 5 µU/mL to 15 µU/mL (odds ratio, 3.57; 95% confidence interval, 1.18-10.8). CONCLUSION: At admission to the ICU, patients with critical illness showed hyperglycemia and relatively decreased insulin levels. High levels of insulin were independently associated with in-hospital mortality in this study population. LEVEL OF EVIDENCE: Prognostic study, level II.