New prediction equations for resting energy expenditure in older hospitalized patients: Development and validation

• Published in: Nutrition (Burbank, Los Angeles County, Calif.) Los Angeles County, Calif.), 2023-11, Vol.115, p.112188-112188, Article 112188
• Main Authors: Kawase, Fumiya, Masaki, Yoshiyuki, Ozawa, Hiroko, Imanaka, Manami, Sugiyama, Aoi, Wada, Hironari, Kobayashi, Shinya, Tsukahara, Takayoshi
• Format: Article
• Language: English
• Subjects: Accuracy; Age; Aging; Agricultural cooperatives; Bias; Body composition; Body measurements; Body weight; Calorimetry; Circumferences; Clinical outcomes; Clinical trials; Datasets; Energy; Energy expenditure; Females; Generalized linear models; Hospitalization; Hospitalized patients; Hospitals; Indirect calorimetry; Internal medicine; Males; Malnutrition; Mortality; Musculoskeletal system; Nutrition; Older patients; Prediction accuracy; Predictions; Regression analysis; Regression models; Resting energy expenditure
• ISSN: 0899-9007; 1873-1244
• DOI: 10.1016/j.nut.2023.112188

•New prediction equations were developed for older hospitalized patients, including those ages 70 to 102 y.•Height, body weight, age, sex, and calf circumference were used in the equations.•Considering height, body weight, age, sex, and calf circumference improved accuracy. Accurate resting energy expenditure (REE) prediction is needed to prevent over- or underfeeding in older hospitalized patients. However, few validated REE prediction Equations are known for such patients. Therefore, this study aimed to develop new REE prediction Equations and evaluate their validity. This single-center, cross-sectional study enrolled 134 patients ages ≥70 y. For holdout validation, patients were randomized in a 3:1 ratio; for the development data set, a new Equation was developed according to the measured REE using indirect calorimetry. The new and existing Equations were compared using the validation data set. Mean patient age was 87.4 ± 6.9 y, and 34.3% were male. Two Equations were developed in multivariable regression models: Equation 1: REE (kcal/day) = 313.582 + Height (cm) × 3.973 + Body weight (kg) × 5.332 - Age (y) × 5.474 - (0 if male; 1 if female) × 20.012 + Calf circumference (cm) × 12.174; and Equation 2: REE (kcal/day) = 594.819 + Height (cm) × 3.760 + Body weight (kg) × 8.888 − Age (y) × 6.298 − (0 if male; 1 if female) × 16.396. The mean relative bias (95% CI) with measured REE as a reference had a small bias for Equations 1 and 2 (−0.1 [−4.1 to 3.9]% and −0.2 [−4.4 to 4.1]%, respectively); however, the Harris-Benedict, Food and Agriculture Organization of the United Nations/World Health Organization/United Nations University, Ganpule, and body weight × 20 Equations had larger biases (−6.2 [−10.3 to −2.0]%; 5.3 [1.3 to 9.3]%; −13.9 [−18.6 to −9.3]%; and −11.6 [−16.1 to −7.1]%, respectively). New prediction Equations using height, body weight, age, sex, and calf circumference improve REE prediction accuracy in older hospitalized patients.