Intensive insulin therapy and brief hypoglycemia do not increase neurological injury markers in critically ill children

Introduction

Targeting glycemia to age-adjusted normal fasting blood glucose levels with intensive insulin therapy (IIT) improved short-term outcome of critically ill children as compared with conventional insulin therapy (CIT), but concomitantly increased the incidence of hypoglycemia [1]. Both hyperglycemia and hypoglycemia may adversely affect the developing brain of young children. We therefore studied the impact of targeting normoglycemia with IIT on circulating markers of brain injury in these patients.

Methods

This was a pre-planned analysis of all 700 pediatric critically ill patients included in a prospective, randomized, controlled study on IIT [1]. Patients were randomly assigned to the target of normal for age fasting blood glucose levels (2.8 to 4.4 mmol/l for age <1 year and 3.9 to 5.6 mmol/l for age ≥1 year) with insulin infusion throughout the ICU stay (IIT), or to insulin infusion only to prevent excessive hyperglycemia (CIT). Serum S100B and neuron-specific enolase (NSE), markers of astrocytic and neuronal damage, were measured by ELISA on fixed days (n = 700) and before and after hypoglycemia in a nested case-control design (n = 126).

Results

Patients admitted to the ICU after cardiac surgery had higher admission S100B and NSE levels than patients in the other diagnostic categories (P < 0.0001), with the highest levels observed for patients with the highest Risk Adjustment for Congenital Heart Surgery (RACHS) score (both P = 0.002). Admission levels of S100B and NSE were higher in patients needing intensive care for at least 3 days vs short-stay patients (both P < 0.0001) and in ICU nonsurvivors vs survivors (P = 0.002 and P = 0.0002). IIT did not affect the levels of S100B or NSE as compared with CIT. Patients experiencing hypoglycemia at any time during the ICU stay revealed higher S100B and NSE levels on admission (P < 0.0001 and P = 0.0007) than those without hypoglycemia. In the nested case-control study, S100B and NSE decreased after hypoglycemia (P = 0.001 and P = 0.009) in the cases, unlike in the controls on matched days.

Conclusions

IIT in the pediatric ICU did not evoke neurological damage, as evaluated by S100B and NSE, despite an increased incidence of brief hypoglycemia. Elevated markers in patients with hypoglycemia were not caused by hypoglycemia but explained by an increased risk of hypoglycemia in the most severely ill.