First Author: Augustin Casals, MD – Childrens Hospital Los Angeles Co-Author: Bashira A. Oloso, MS – Biostatistician, Children's Hospital Los Angeles Co-Author: Bradley De Souza, MB – Clinical Assistant Professor Pediatrics, Childrens Hospital Los Angeles Co-Author: Patrick A. Ross, MD – Professor of Clinical Anesthesiology and Pediatrics, Childrens Hospital Los Angeles Co-Author: Robinder Khemani, MD, MS – Childrens Hospital Los Angeles Co-Author: Anoopindar K. Bhalla, MD
Introduction: Accurate estimation of resting energy expenditure (REE) is challenging in critically ill children. REE is routinely estimated using Schofield’s equation (based on biometric data), though a volumetric capnography (VCO₂)-based method (Mehta’s equation) is reportedly more accurate. Underfeeding and overfeeding may influence recovery from critical illness, including weaning from mechanical ventilation (MV). We evaluated whether feeding classification - defined by the mean daily energy intake (EI) to REE ratio as underfed (UF, < 0.8), appropriately fed (AF, 0.8-1.2), or overfed (OF, >1.2) - was associated with spontaneous breathing trial (SBT) success in children with pediatric acute respiratory distress syndrome (PARDS). We hypothesized that AF patients, compared to UF and OF, would be more likely to pass their first SBT after passing a standardized oxygenation test, and that feeding classification would differ by REE estimation equation.
Methods: Post-hoc analysis of single site data collected from children 1 month to 18 years with PARDS enrolled in the Real-time Effort Driven ventilator management (REDvent) trial. Those on MV for < 3 days prior to first SBT and without VCO2 data were excluded. EI was derived from daily nutritional data. REE was estimated using Schofield and Mehta [MREEM = 5.5 * VCO2 (L/min) * 1440] equations. Feeding classification was determined from mean daily EI:REE ratio prior to first SBT.
Results: Of 247 REDvent patients, 66 met inclusion criteria. Using Schofield, 14 patients (21%) were AF, 7 patients (11%) OF, and 45 patients (68%) UF. By Mehta, 17 patients (26%) were AF, 12 patients (18%) OF, and 37 patients (56%) UF. Twenty-six patients (39%) passed their first SBT. SBT success did not differ significantly by feeding classification using either Schofield (43% AF, 43% OF, 38% UF, p = 0.93) or Mehta (47% AF, 25% OF, 41% UF, p = 0.51) equations. Bland-Altman analysis had a mean bias of 20.2% with wide 95% limits of agreement (-42.3, 82.7).
Conclusions: Feeding classification and REE, estimated by Schofield’s or VCO2-based equations, was not significantly associated with SBT success. Wide limits of agreement between equations used to estimate REE may influence clinical interpretation of appropriate nutrient delivery.
