PURPOSE: Hyperglycemia in critically ill, non-diabetic children is associated with increased mortality. The risk of hypoglycemia associated with insulin treatment, compared to the proposed benefit of controlling blood sugar levels in this population is unknown. Hypoglycemia occurs in critically ill pediatric patients independent of insulin therapy and may be associated with increased mortality. A standard computerized protocol with insulin to treat hyperglycemia may decrease the frequency of hypoglycemia as compared to physician titration of insulin.
METHODS: We performed a retrospective cohort study of all admissions to a 26-32 bed PICU during the years 2003-2005. Computerized hospital admission records and hospital laboratory databases were reviewed. Included were all patients admitted to the PICU for greater than 24 hours with at least a single blood glucose level. The associations between mortality, insulin use, or hypoglycemia (blood glucose < 60 mg/dL) were analyzed. Patients who received insulin under physician directed titration were compared to those who received insulin under the direction of a computerized insulin protocol.
RESULTS: In 2003, Hypoglycemia (<60mg/dL) occurred in 12.8 % of all patients, and in 9.1% of hyperglycemic patients (as defined by a single blood glucose > 150 mg/dl).Hypoglycemia (<60 mg/dL) was associated with increase in mortality (p<.001). Hypoglycemia (<40 mg/dL) occurred in 18% of patients treated with physician directed insulin therapy compared to 2% of patients treated with computerized protocol directed insulin therapy.
CONCLUSION: Hypoglycemia is associated with increased mortality. In our institution, a computerized insulin protocol may be associated with a decreased rate of hypoglycemia when compared to clinician directed titration of insulin.
CLINICAL IMPLICATIONS: The use of a computerized insulin protocol to standardize glucose control in critically ill children may minimize the incidence of hypoglycemia and allow for a reproducible method to study the risks and benefits of glucose control in critically ill children.
DISCLOSURE: Ellie Hirshberg, None.