Association of timing, duration, and intensity of hyperglycemia with intensive care unit mortality in critically ill children

Abstract

OBJECTIVE: To study the association of timing, duration, and intensity of hyperglycemia with pediatric intensive care unit (PICU) mortality in critically ill children. DESIGN: Retrospective cohort study. SETTING: PICU of a university-affiliated, tertiary care, children's hospital. PATIENTS: A total of 152 critically ill children receiving vasoactive infusions or mechanical ventilation. INTERVENTIONS: None. METHODS: With institutional review board approval, we reviewed a cohort of 179 consecutive children, 1 mo to 21 yrs of age, treated with mechanical ventilation or vasoactive infusions. We excluded 18 with <3 microg.kg(-1).min(-1) dopamine only, diabetes, or solid organ transplant and nine who died within 24 hrs of PICU admission. Peak blood glucose (BG) and time to peak BG during PICU admission, duration of hyperglycemia (percentage of PICU days with any BG of >126 mg/dL), and intensity of hyperglycemia (median BG during first 48 PICU hours) were analyzed for association with PICU mortality using chi-square, Student's t-test, and logistic regression. MEASUREMENTS AND MAIN RESULTS: Peak BG of >126 mg/dL occurred in 86% of patients. Compared with survivors, nonsurvivors had higher peak BG (311 +/- 115 vs. 205 +/- 80 mg/dL, p <.001). Median time to peak BG was similar in nonsurvivors (23.5 hrs; interquartile ratio, 5-236 hrs) and survivors (19 hrs; interquartile ratio, 6-113 hrs). Duration of hyperglycemia was longer in nonsurvivors (71% +/- 14% of PICU days) vs. survivors (37% +/- 5% of PICU days, p <.001). Nonsurvivors had more intense hyperglycemia during the first 48 hrs in the PICU (126 +/- 38 mg/dL) vs. survivors (116 +/- 34 mg/dL, p <.05). Univariate logistic regression analysis showed that peak BG and the duration and intensity of hyperglycemia were each associated with PICU mortality (p <.05). Multivariate modeling controlling for age and Pediatric Risk of Mortality scores showed independent association of peak BG and duration of hyperglycemia with PICU mortality (p <.05). CONCLUSIONS: Hyperglycemia is common in critically ill children. Peak BG and duration of hyperglycemia are independently associated with mortality in our PICU. A prospective, randomized trial of strict glycemic control in this subset of critically ill children who are at high risk of mortality is both warranted and feasible.