Background and aims The last decade gave clear evidence that hyper/hypoglycemia and glucose variability are associated with increased mortality in critically ill patients. Continuous glucose monitor (CGM) is a new device in paediatric critical care units (PICU) with clear advantages in glucose monitoring. The aim of our study was to survey the incidence of glucose regulation disorders in our PICU and specify the association between the PRISM III score and the glycemic variability [mean amplitude of glycemic action (MAGE)].

Methods We evaluated 22 children: mean age: 1.3 years, mean length of PICU stay: 18 days; 20/22 patients were on invasive mechanical ventilation; 6/22 needed vasoactive agent therapy. CGM duration: 1–12 days. Interstitial glucose level was monitored by Guardian^® REAL Time CGM (Medtronic^®). Reference glucose values were obtained from blood gas analyzer or point-of-care glucose analyzer. We used Spearman correlation to evaluate the association between PRISM III and the MAGE.

Results Hypo- and hyperglycemia (CGM glucose < 55 mg/dl / CGM glucose > 180 mg/dl) were detected in 4.6% and 2.5% of measurements, respectively. The mean MAGE (meaningful excursion >45 mg/dl) and PRISM III were 78 mg/dl and 19. We found a significant correlation between PRISM III and MAGE (r = 0,55; p < 0.05). Pearson’s correlation coefficient (0.82) and Clarke Error Grid analysis (96% clinical accuracy) proved a good reliability of the CGM.

Conclusions Glucose homeostasis disorders are frequent in the PICU; hypoglycemia being more commonly detected. Increased PRISM III score contributes significantly to the elevation of glucose variability.