Measured and Predicted Resting Energy Expenditure in Malnourished Older Hospitalized Patients: A Cross-Sectional and Longitudinal Comparison

A number of equations have been proposed to predict resting energy expenditure (REE). The role of nutritional status in the accuracy and validity of the REE predicted in older patients has been paid less attention. We aimed to compare REE measured by indirect calorimetry (IC) and REE predicted by the Harris–Benedict formula in malnourished older hospitalized patients. Twenty-three malnourished older patients (age range 67–93 years, 65% women) participated in this prospective longitudinal observational study. Malnutrition was defined as Mini Nutritional Assessment Long Form (MNA-SF) score of less than 17. REE was measured (REE(measured)) and predicted (REE(predicted)) on admission and at discharge. REE(predicted) within ±10% of the REE(measured) was considered as accuracy. Nutritional support was provided to all malnourished patients during hospitalization. All patients were malnourished with a median MNA-LF score of 14. REE(measured) and REE(predicted) increased significantly during 2-week nutritional therapy (+212.6 kcal and +19.5 kcal, respectively). Mean REE(predicted) (1190.4 kcal) was significantly higher than REE(measured) (967.5 kcal) on admission (p < 0.001). This difference disappeared at discharge (p = 0.713). The average REE(predicted) exceeded the REE(measured) on admission and at discharge by 29% and 11%, respectively. The magnitude of difference between REE(measured) and REE(predicted) increased along with the degree of malnutrition (r = 0.42, p = 0.042) as deviations ranged from −582 to +310 kcal/day in severe to mildly malnourished patients, respectively. REE(predicted) by the Harris–Benedict formula is not accurate in malnourished older hospitalized patients. REE measured by IC is considered precise, but it may not represent the true energy requirements to recover from malnutrition. Therefore, the effect of malnutrition on measured REE must be taken into account when estimating energy needs in these patients.