AIMS: To assess validity and reliability of energy expenditure measurements with a short Douglas bag protocol compared to the standard metabolic monitor in a paediatric intensive care setting.
METHODS: 51 paired measurements were performed in 14 ventilated patients (age 0-18 years) with sepsis, trauma or following major surgery. Measured data were compared mutually and compared to Schofield equations using Bland-Altman analysis.
RESULTS: Comparing Douglas bag (3.21 +/- 1.43 MJ/day) and metabolic monitor (3.15 +/- 1.49 MJ/day) we found bias in energy expenditure of -0.06 (equal to -2%, NS) with limits of agreement of -0.5 to 0.4 MJ/day (equal to -16% to +13%). Intra-measurement variability (coefficient of variation) was within 10% for both methods. Both the metabolic monitor and Douglas bag showed significant bias compared to Schofield equations (3.39 +/-1.64 MJ/day) of -7% (P < 0.01) and -5% (P < 0.05), respectively, with wide limits of agreement: metabolic monitor vs. Schofield: -37% to +22%, Douglas bag vs. Schofield: -37% to +26%.
CONCLUSIONS: The Douglas bag method compared favourably to the metabolic monitor where Schofield equations failed to predict individual energy expenditure. Considering its low cost, this renders the short and simple Douglas bag method a robust measure and a routinely applicable instrument for tailored nutritional assessment in critically ill children.