A Prospective Comparative Study to Assess the Accuracy of Energy Expenditures Calculated by Bioimpedance Analysis and Indirect Calorimetry.

BACKGROUND: It is crucial to precisely measure resting energy expenditure (REE) to provide nutritional support, especially for obese individuals. Although indirect calorimetry (IC) is the gold standard, its limited accessibility necessitates the evaluation of alternative methods such as bioelectrical impedance analysis (BIA). OBJECTIVE: To compare the accuracy of REE measurements obtained using BIA and IC in obese adults. METHODS: This prospective cross-sectional observational study was conducted in the intensive care unit (ICU) of S. L. Raheja Hospital, Mumbai. Fifty adults aged 18-65 years with a body mass index (BMI) ≥25 kg/m² were evaluated. REE was measured using BIA (InBody, Seoul, Korea) and IC (COSMED Q-NRG, Rome, Italy) under standardized fasting and resting conditions. A paired t-test was used to compare mean REE values, and additional agreement and equivalence analyses were performed. RESULTS: The mean REE measured by IC was 1601.56 ± 376.72 kcal/day, while BIA yielded 1554.20 ± 355.70 kcal/day. The two approaches did not differ significantly (p = 0.05196). However, non-significance alone does not imply equivalence; therefore, further analyses were conducted. Bland-Altman analysis demonstrated a small mean bias between the two methods, with values within the 95% limits of agreement and no evidence of proportional bias. Lin's concordance correlation coefficient (CCC) indicated strong concordance, and the mean absolute percentage error remained within an acceptable clinical margin of ±10% relative to IC. A two one-sided test (TOST) procedure confirmed statistical equivalence between BIA and IC within the predefined ±10% equivalence margin. These findings collectively suggest that BIA provides a clinically comparable estimate of REE to IC, supporting its potential use when IC is unavailable. CONCLUSION: BIA demonstrated clinically acceptable agreement with IC for estimating REE in obese Indian adults. Using a predefined equivalence margin of ±10% (≈±160 kcal/day) of IC values, both the mean bias and Bland-Altman limits of agreement fell within this range, supporting statistical and clinical comparability between the two methods. Given its accessibility, portability, and cost-effectiveness, BIA may serve as a practical alternative to IC in clinical and research settings, although further validation across larger and more diverse populations is warranted.