Agreement Between Resting Energy Expenditure Predictive Formulas and Indirect Calorimetry in Non-Dialysis Dependent Chronic Kidney Disease

Abstract

Background and Aims: The gold standard method for measuring resting energy expenditure (REE) is indirect calorimetry (IC) using an expensive device that requires specialized training. To overcome the limitations of IC, REE prediction formulas are used in patients with chronic kidney disease (CKD). However, it is still controversial which of these formulas has greater accuracy compared to IC. We aimed to determine the accuracies of REE measured by IC and estimated by formulas in patients with CKD. Methods: Fifty-three patients with stage 4–5 CKD underwent IC and five current REE prediction formulas. Accuracy was measured by Lin’s correlation coefficient. Bland–Altman repeated measures analysis was used to assess the agreement of the formulas’ results with those of IC. Precision was measured by the predicted IC ± 10% and 20%. Systematic bias was assessed by the Student’s t-test, and linear regression was used to assess proportionality bias. Results: Patients had a mean estimated glomerular filtration rate (eGFR) of 12 ± 4 mL/min/1.73 m2, a mean age of 65 years, and 62% were male. The mean REE measured by IC was 1341 ± 37 Kcal/day, and the formula with the lowest mean bias (0.1509 [−653.5121; 398.9056]), best correlation (r = 0.789; p = 0.000), and best accuracy (85%) was the formula developed by Fernandes and Cols (REE (kcal/day) = 854 + (7.4 × body weight) + (179 × sex) − (3.3 × age) + (2.1 × eGFR) + 26 (if diabetes)). Conclusions: The Fernandes and Cols equation had good accuracy and was valuable for estimating energy requirements in the population studied.