992

Abstract

Introduction: Over the past decade, data from many sources have suggested benefits of controlling hyperglycemia in critical illness. It is unclear how these recommendations should be integrated into the care of critically ill children. We conducted a study of centers that incorporated a standard approach to glycemic control and report the ability of this protocol to identify and control hyperglycemia and induce hypoglycemia. Methods: Multicenter prospective observational trial. Six US pediatric ICUs (3 mixed cardiac/medical/surgical, 1 dedicated cardiac ICU, 2 medical/surgical ICU) adopted a standardized approach to identify hyperglycemia in children who were mechanically ventilated and/or on vasopressors and begin insulin therapy if blood glucose (BG) levels were >140 mg/dL, and use it to control BG 80-140mg/dL using a paper based algorithm. Data were collected on patients that did or did not develop hyperglycemia. Results: A total of 469 patients were evaluated, 156 that did not develop hyperglycemia (group 1), and 313 who developed hyperglycemia and were treated with insulin (group 2). Average age was 5 and 8 years old in groups 1 vs 2. Admission PELOD was 11.2 in group 1 and 17.0 in group 2 (p<0.01). Average ICU length of stay and ventilator days were 12.3 and 6.3 vs. 15.1 and 10.8 days(p=0.02 and <0.01), respectively. In those that developed hyperglycemia the average diagnostic BG levels was 227.9 mg/dL. With insulin treatment, average BG levels decreased to 180(SD4.1), 134(SD2.5), 131(SD2.7) and 131(SD2.8)mg/dL on days 1-4, respectively. Group 2 had higher mortality rates than group 2 (16% vs 6%).The rates of severe hypoglycemia did not significantly differ between groups with 0.52 events in 1000 ICU days in group 1 and 2.1 severe hypoglycemic episodes per 1000 ICU days in group 2 (p=0.26). Conclusions: We have assessed the safety and efficacy of a glycemic control protocol in a consortium of pediatric ICUs. This protocol, a simple, paper-based algorithm, effectively manage glucose levels with low rates of hypoglycemia. (Funding for this project was from NIDDK (R21DK081847; PI-Rigby) Clinical Trials: NCT01116674