• Jotterand ChaparroCorinneCDepartment of Nutrition and Dietetics, School of Health Professions, University of Applied Sciences Western Switzerland, Geneva 1227, Switzerland; Pediatric Intensive Care Unit, Medico-Surgical Department of Pediatrics, Universit, Patrick Taffé, Clémence Moullet, Jocelyne Laure Depeyre, David Longchamp, Marie-Hélène Perez, and Jacques Cotting.
• Department of Nutrition and Dietetics, School of Health Professions, University of Applied Sciences Western Switzerland, Geneva 1227, Switzerland; Pediatric Intensive Care Unit, Medico-Surgical Department of Pediatrics, University Hospital of Lausanne, Lausanne 1011, Switzerland.
• J. Pediatr. 2017 May 1; 184: 220-226.e5.

ObjectiveTo determine, based on indirect calorimetry measurements, the biases of predictive equations specifically developed recently for estimating resting energy expenditure (REE) in ventilated critically ill children, or developed for healthy populations but used in critically ill children.Study DesignA secondary analysis study was performed using our data on REE measured in a previous prospective study on protein and energy needs in pediatric intensive care unit. We included 75 ventilated critically ill children (median age, 21 months) in whom 407 indirect calorimetry measurements were performed. Fifteen predictive equations were used to estimate REE: the equations of White, Meyer, Mehta, Schofield, Henry, the World Health Organization, Fleisch, and Harris-Benedict and the tables of Talbot. Their differential and proportional biases (with 95% CIs) were computed and the bias plotted in graphs. The Bland-Altman method was also used.ResultsMost equations underestimated and overestimated REE between 200 and 1000 kcal/day. The equations of Mehta, Schofield, and Henry and the tables of Talbot had a bias ≤10%, but the 95% CI was large and contained values by far beyond ±10% for low REE values. Other specific equations for critically ill children had even wider biases.ConclusionsIn ventilated critically ill children, none of the predictive equations tested met the performance criteria for the entire range of REE between 200 and 1000 kcal/day. Even the equations with the smallest bias may entail a risk of underfeeding or overfeeding, especially in the youngest children. Indirect calorimetry measurement must be preferred.Copyright © 2016 Elsevier Inc. All rights reserved.

18 keywords...

hide…