Accuracy of Predictive Formulas vs. Indirect Calorimetry in Estimating Energy Needs of Patients in Intensive Care...

Introduction: Accurately meeting the energy requirements of patients in intensive care units (ICUs) is crucial to prevent catabolism, muscle loss, and complications. We assessed their energy needs in this study using indirect calorimetry (IC) and predictive formulas, comparing the results with delivered energy intake and evaluating agreement. Materials and Methods: A total of 38 mechanically ventilated patients in seven ICUs at Kayseri City Hospital were included; eligible patients were ≥18 years old and mechanically ventilated for at least 24 h. Disease severity and nutritional risk were evaluated using validated indices (prognostic nutritional index (PNI) and Modified Nutrition Risk in the Critically Ill (mNUTRIC)), and basal energy expenditure (BEE) was measured by IC and calculated using the Harris-Benedict (HB) and ESPEN formulas. IC measurements lasted 15 min under resting conditions in conscious patients and, according to acute phase criteria, in unconscious patients in a quiet, temperature-controlled environment. Nutrition was provided enterally or parenterally based on patient condition and disease severity. Agreement between IC and predictive formulas was assessed using Bland-Altman analysis, a statistical method that evaluates agreement between two measurement techniques. Results: Estimated energy requirements differed significantly from delivered energy intake (p < 0.001). IC-derived values were significantly lower than those estimated by the HB equation and ESPEN recommendations (p < 0.001), suggesting that predictive equations may overestimate energy requirements in this population. By contrast, delivered energy intake was lower than IC-measured values, with a mean difference of approximately 503 kcal, indicating a potential risk of underfeeding in clinical practice. Weak correlations were observed between methods (IC vs. HB: r = 0.35, p = 0.003; IC vs. ESPEN: r = -0.21, p = 0.02), indicating limited agreement between predictive equations and IC measurements, and Passing-Bablok regression analysis further supported this lack of agreement between methods. Conclusions: The energy intake delivered to patients was lower than the calculated values. Indirect calorimetry is important for accurately monitoring and determining energy requirements based on delivered energy intake, and further research in this area is needed. These findings highlight the importance of individualized monitoring of energy expenditure in critically ill patients and suggest that reliance solely on predictive equations may lead to clinically relevant discrepancies in energy delivery.