Objective: To determine the association between hyperglycemia and outcome in children ventilated for meningococcal sepsis.
Design: Retrospective case notes review.
Setting: Eight bedded pediatric intensive care unit in London.
Patients: Consecutive children ventilated for meningococcal sepsis 2001-2004.
Interventions: None.
Measurements: Peak glucose for the entire admission was determined and mean glucose was calculated for the following three epochs: 1) first 24 hrs, 2) second 24 hrs, and 3) the entire pediatric intensive care unit admission. Patients were also grouped according to whether their blood glucose rose to >7 mmol/L (126 mg/dL), >10 mmol/L (180 mg/dL), or remained below these levels during the pediatric intensive care unit admission. Outcome measures were predicted mortality (based on pediatric risk of mortality score), ventilator free days at 30 days, nosocomial infection, use of renal replacement therapy, use of inotropes, and skin necrosis.
Main results: Ninety-seven patients were identified with a median age of 2.1 yrs and a median length of stay of 4 days. Four patients died. Peak glucose significantly correlated with predicted mortality and negatively correlated with ventilator free days at 30 days (p < 0.001 and p < 0.001, respectively). Patients who received renal replacement therapy or inotropic support, or developed a nosocomial infection or skin necrosis had significantly higher peak glucose than those who did not (p = 0.006, p < 0.0001, p = 0.022, and p < 0.0001, respectively). Patients who received renal replacement therapy or who developed skin necrosis had significantly higher mean blood glucose in the second 24 hrs of admission (p = 0.017 and p = 0.004, respectively). However, mean blood glucose in the first 24 hrs and over the entire admission did not correlate with outcome. Patients defined as hyperglycemic with blood glucose either >7 mmol/L or >10 mmol/L also had a significantly worse outcome than those who maintained blood glucose below these levels.
Conclusions: There was a significant association between hyperglycemia and outcome. Our results support a trial of tight glycemic control in this group of critically ill children.