Institute of Physical Activity and Nutrition, School of Exercise and Nutrition Science, Deakin University, Burwood, Melbourne, Victoria, Australia.
Chemistry and Molecular Biosciences, The University of Queensland, Brisbane, Queensland, Australia.
Am J Clin Nutr. 2023 Jun;117(6):1164-1173. doi: 10.1016/j.ajcnut.2023.04.010. Epub 2023 Apr 11.
The aging process alters the resting metabolic rate (RMR), but it still accounts for 50%-70% of the total energy needs. The rising proportion of older adults, especially those over 80 y of age, underpins the need for a simple, rapid method to estimate the energy needs of older adults.
This research aimed to generate and validate new RMR equations specifically for older adults and to report their performance and accuracy.
Data were sourced to form an international dataset of adults aged ≥65 y (n = 1686, 38.5% male) where RMR was measured using the reference method of indirect calorimetry. Multiple regression was used to predict RMR from age (y), sex, weight (kg), and height (cm). Double cross-validation in a randomized, sex-stratified, age-matched 50:50 split and leave one out cross-validation were performed. The newly generated prediction equations were compared with the existing commonly used equations.
The new prediction equation for males and females aged ≥65 y had an overall improved performance, albeit marginally, when compared with the existing equations. It is described as follows: RMR (kJ/d) = 31.524 × W (kg) + 25.851 × H (cm) - 24.432 × Age (y) + 486.268 × Sex (M = 1, F = 0) + 530.557. Equations stratified by age (65-79.9 y and >80 y) and sex are also provided. The newly created equation estimates RMR within a population mean prediction bias of ∼50 kJ/d (∼1%) for those aged ≥65 y. Accuracy was reduced in adults aged ≥80 y (∼100 kJ/d, ∼2%) but was still within the clinically acceptable range for both males and females. Limits of agreement indicated a poorer performance at an individual level with 1.96-SD limits of approximately ±25%.
The new equations, using simple measures of weight, height, and age, improved the accuracy in the prediction of RMR in populations in clinical practice. However, no equation performs optimally at the individual level.
衰老过程会改变静息代谢率（RMR），但它仍然占总能量需求的 50%-70%。老年人的比例不断上升，尤其是 80 岁以上的老年人，这就需要一种简单、快速的方法来估计老年人的能量需求。
本研究旨在生成和验证专门用于老年人的新 RMR 方程，并报告其性能和准确性。
本研究的数据来源于一个国际数据集，其中包括 1686 名年龄≥65 岁的成年人（38.5%为男性），使用间接测热法测量 RMR。使用多元回归法从年龄（岁）、性别、体重（kg）和身高（cm）预测 RMR。在随机、性别分层、年龄匹配的 50:50 分割和留一法交叉验证中进行双重交叉验证。将新生成的预测方程与现有的常用方程进行比较。
新的男性和女性≥65 岁的预测方程在与现有方程的比较中，尽管略有改善，但总体性能有所提高。其描述如下：RMR（kJ/d）=31.524×W（kg）+25.851×H（cm）-24.432×年龄（y）+486.268×性别（M=1，F=0）+530.557。还提供了按年龄（65-79.9 岁和>80 岁）和性别分层的方程。对于≥65 岁的人群，新创建的方程估计 RMR 的人群平均预测偏差约为 50 kJ/d（约 1%）。在≥80 岁的成年人中，准确性降低（约 100 kJ/d，约 2%），但仍在男女两性的临床可接受范围内。一致性界限表明，个体水平的性能较差，1.96-SD 界限约为±25%。
新的方程使用体重、身高和年龄的简单测量值，提高了临床实践中人群 RMR 预测的准确性。然而，没有一个方程在个体水平上表现最佳。
