Accuracy of Prediction Formulae for the Assessment of Resting Energy Expenditure in Hospitalized Children.

Abstract

The resting energy expenditure (REE) of ill children is commonly estimated from prediction formulae developed in healthy children. The aim of the present study was to evaluate the accuracy of commonly employed REE prediction formulae versus indirect calorimetry in hospitalized children. We performed a cross-sectional study of 236 infants, children, and adolescents consecutively admitted to the Intermediate Care, Nephrology, Intensive Care, Emergency, and Cystic Fibrosis Units of the De Marchi Pediatric Hospital (Milan, Italy) between September 2013 and March 2015. REE was measured by indirect calorimetry and estimated using the World Health Organization (WHO), Harris-Benedict, Schofield, and Oxford formulae. The mean (standard deviation) difference between the estimated and measured REE was: -1 (234) kcal/day for the WHO formula; 82 (286) kcal/day for the Harris-Benedict formula; 2 (215) kcal/day for the Schofield-weight formula; -2 (214) kcal/day for the Schofield-weight and height formula; and -5 (221) kcal/day for the Oxford formula. Even though the WHO, Schofield, and Oxford formulae gave accurate estimates of REE at the population level (small mean bias), all the formulae were not accurate enough to be employed at the individual level (large SD of bias). The WHO, Harris-Benedict, Schofield, and Oxford formulae should not be used to estimate REE in hospitalized children.