Resting energy expenditure in hemodialysis patients: which prediction equation estimates the best?

Abstract

In chronic kidney disease (CKD) there are many metabolic changes that can influence resting energy expenditure (REE). Protein energy wasting is highly prevalent in hemodialysis patients. Estimating REE is the first step in the process to advise an adequate nutritional therapy. There are many prediction equations for REE. The purpose of this study is to determine which equation predicts REE in a reliable way. In the VU University Medical Centre 56 hemodialysis patients (33 men), mean age 56±14.8 year, BMI (kg/m2) 27.0±6 indirect calorimetry was performed. Measured REE was compared with sixteen different prediction equations based on weight, height, sex, and/or age. The Root Mean Squared prediction Error (RMSE) are presented as well as the average difference between estimation and measurement (bias). 90–110% adequacy of measured REE was accepted as accurate prediction. The measured REE was 1526±299 kcal/d (mean±SD). The percentage accurate predictions, percentage bias and RMSE were: for original Harris Benedict equation (1919), Schofield equation based on weight and height and Cole equation respectively 48%, 50%, 57%; bias: +2.9%, +4.4%, -0.5%; RMSE: 288, 286,271 kcal/d. In this study Coles prediction equations provide an acceptable estimate of resting energy expenditure of HD patients. Further insight into explaining factors is warranted.