Comparison of Equations to Predict Energy Requirements With Indirect Calorimetry in Hospitalized Patients.

Abstract

The accuracy of methods to determine resting energy expenditure (REE) contributes toward the adequate provision of nutrition support to hospitalized patients. Indirect calorimetry (IC) is considered the gold-standard method to determine REE. The aim of this study is to evaluate the degree of agreement between the REE measured by IC (REE-IC) and REE estimated by predictive equations in intensive care unit patients. The sample is made up of intensive care unit patients aged >18 years, both male and female, undergoing nutrition therapy. The predictive equations to estimate REE were the Harris Benedict (HB), Ireton Jones (IJ), and practical method (PM). Degree of agreement between REE-predictive equations and REE-IC was analyzed by the interclass correlation coefficient (ICC) and the Bland-Altman test. Average energy obtained by IC was significantly different from HB and IJ equations (P < .001). The HB equation significantly underestimated the REE-IC for body mass index (BMI) classification. Significant concordance was observed between the REE-IC and all estimate equations (P < .05). The IJ equation showed the greatest degree of concordance for BMI classification of underweight (ICC = 0.674; P = .011) and presented the least difference between the averages of the energy when compared with REE-IC (107.8 kcal/d; P < .05). The IJ equation showed better results with IC, with the greatest degree of concordance for BMI classification of underweight. Further research should develop others equations and validate tools to measure energy expenditure for accurate dietary recommendations for hospitalized patients undergoing nutrition therapy.