A protocolized approach to identify and manage hyperglycemia in a pediatric critical care unit

Abstract

INTRODUCTION: Hyperglycemia is a risk factor for poor outcome in critically ill patients, and glycemic control may decrease morbidity and mortality in adults. There is limited information regarding hyperglycemia and its control in pediatric intensive care. OBJECTIVE: To determine prevalence and risk factors for hyperglycemia and evaluate our approach to glycemic control in critically ill children. DESIGN, SETTING, PATIENTS, AND MAIN OUTCOMES: A pediatric-specific protocol to identify and manage hyperglycemia was developed and instituted as standard practice in our pediatric intensive care unit, and was applicable to patients >6 months and >5 kg, without end-stage liver disease or type 1 diabetes mellitus. Triggers for routine blood glucose assessment were based on supportive measures including mechanical ventilation, vasopressor/inotrope infusions, and antihypertensive infusions. Hyperglycemic patients, defined by two consecutive blood glucose readings of >140 mg/dL (7.7 mmol/L), were treated with infused insulin to maintain blood glucose levels 80-140 mg/dL (4.4-7.7 mmol/L). We performed retrospective analysis 6 months after instituting this approach. Main outcomes were prevalence and risk factors for hyperglycemia, and effectiveness of our approach to achieve glycemic control. INTERVENTIONS: None. MEASUREMENTS/MAIN RESULTS: One hundred forty-five of 477 patients had blood glucose actively assessed, and 74 developed hyperglycemia and were managed with insulin. This approach to identify patients with hyperglycemia had a positive predictive value of 51% and negative predictive value of 94%. Hyperglycemia prevalence was 20%. Mechanical ventilation, vasopressor/inotropic infusion, continuous renal replacement therapy, high illness severity scores, and longer lengths of stay were associated with hyperglycemia. The average blood glucose of patients with hyperglycemia was 200 mg/dL (11 mmol/L), and on average, patients were treated with insulin for 6.3 days with 2.4 units/kg/day. Blood glucose levels were/dL (8.8 mmol/L) in 70% of insulin-treated days, 80-140 mg/dL (4.4-7.7 mmol/L) in 49% of insulin-treated days, and 4% of insulin-treated patients had any blood glucose measurements/dL (2.2 mmol/L). CONCLUSIONS: Hyperglycemia is prevalent in pediatric intensive care units and may be effectively identified and managed using a protocolized approach.