Phase-Related Resting Energy Expenditure in Critically Ill Adults: Metabolic Phenotypes and Determinants of Weight-Normalized Indices—A Retrospective Study

Precise measurement of resting energy expenditure (REE) is essential in the intensive care unit (ICU), where metabolic requirements evolve throughout the course of critical illness. Predictive equations frequently fail to capture this variability, and limited data describe phase-dependent changes in REE using indirect calorimetry (IC). This study aimed to evaluate phase-related variation in REE and metabolic phenotypes in mechanically ventilated adults and to identify clinical and physiological correlates of both absolute REE and REE normalized by ideal body weight (REE/IBW). Methods: We conducted an observational, retrospective cross-sectional study in two ICUs at different hospitals. A total of 149 mechanically ventilated adults with a valid IC measurement were included and classified by illness phase: acute (0–3 days), intermediate (4–14 days), or chronic (>14 days). Differences in metabolic and gas-exchange variables were assessed using ANOVA or Kruskal–Wallis tests. Two multivariable linear regression models were fitted, one using absolute REE and a second using REE/IBW, incorporating metabolic phenotype categories to account for body-size heterogeneity. Results: Metabolic profiles differed across illness phases. Median REE increased from the acute (1664 kcal/day) to the intermediate (1869 kcal/day) and chronic (2074 kcal/day; p = 0.024) phases. Hypometabolic profiles were more frequent in the acute phase (64%), whereas hypermetabolic profiles were more prevalent in later phases (48%). RQ values were higher in the chronic phase compared with the acute phase (median 0.99 vs. 0.80; p < 0.001). In multivariable analyses, illness severity scores showed weak or inconsistent associations with REE after adjustment for gas-exchange variables. VCO2 was independently associated with absolute REE (adjusted R2 = 0.83). In the REE/IBW model, VCO2, RQ, BMI, and metabolic phenotype were associated with normalized energy expenditure, with higher adjusted R2 (0.87) and lower prediction error metrics. Conclusions: Resting energy expenditure and metabolic phenotypes vary across phases of critical illness. Gas-exchange variables, particularly VCO2, were more closely associated with measured energy expenditure than severity scores. Normalization of REE by ideal body weight reduced variability and improved model performance, highlighting the analytical value of indirect calorimetry for characterizing phase-dependent metabolic patterns in critically ill adults.