Assessing the validity of a portable indirect calorimetry device in a pediatric sample

Background & aims Accurate and scalable tools to measure metabolic markers, such as resting energy expenditure (REE) and respiratory quotient (RQ) are essential for the success of personalized nutrition and weight management programs. These measures are particularly relevant for children, as their metabolism and energy needs fluctuate widely during development. Predictive equations are limited due to their poor precision at the individual level, particularly in children. While the Douglas Bag method remains the gold standard for measuring REE and gas exchange, it is impractical for routine pediatric use due to high monetary cost, long measurement time, and expertise requirements. The Breezing Med™ device could potentially serve as a portable, non-invasive alternative for accurately assessing metabolism, but it has not been validated in children. Methods The purpose of this study was to validate the Breezing Med™ device, a handheld, portable indirect calorimetry tool, against the gold-standard Douglas Bag method for measuring REE and gas exchange parameters in a pediatric sample. Nineteen children (mean age: 10.5 ± 3.7 years; range: 5–17 years) completed resting, fasted testing with the Breezing Med™ and Douglas Bag system. Agreement was assessed using concordance correlation coefficients (CCC), bias analyses, and Bland–Altman plots. Root mean square error (RMSE) and ±10% accuracy thresholds were also calculated to evaluate overall prediction error. Results were also compared to predictive energy expenditure equations that are commonly used for pediatric populations. Results Breezing Med™ demonstrated good agreement with Douglas Bag for REE (CCC = 0.88; bias correction factor = 0.93), although it systematically overestimated REE by 126.8 kcal/day (9.8%) and REE/kg by 4.5 kcal/kg/day (12.0%). Compared to the Douglas Bag, the Breezing Med™ overestimated VO₂ by 13.8%, while differences in VCO₂ were negligible (mean bias = −1.1 mL/min, −0.6%). In contrast, predictive equations showed smaller mean bias but greater variability, lower CCC values, and weaker overall agreement compared to Breezing Med™. Breezing Med™ demonstrated lower overall prediction error than predictive equations. RQ showed negative bias (mean difference = −0.10) with poor categorical agreement (κ_w= 0.12). Conclusions Breezing Med™ provides a scalable, non-invasive approach for measuring REE and gas exchange in children but appears to systematically overestimate VO₂ magnitudes. This led to overestimations of REE and poor RQ agreement in our sample. These findings suggest the device demonstrates good precision but reduced accuracy relative to the Douglas Bag method. Software calibration or correction factors would need to be developed to correct these overestimates before widespread use in clinical settings can be recommended. Testing on a larger and more diverse pediatric sample would be needed to develop and validate appropriate correction factors. With appropriate calibration, Breezing Med™ holds promise as a useful and accessible tool for individualized metabolic assessment, particularly in settings where traditional gold-standard methods are impractical.