Laboratório de Avaliação Nutricional e Funcional, Departamento de Nutrição Social, Universidade Federal Fluminense, Rua Mário Santos Braga 30, Sala 415, Campus do Valonguinho, 24020-140, Niterói, Rio de Janeiro, Brazil.
Laboratório de Avaliação Nutricional e Funcional, Departamento de Nutrição Social, Universidade Federal Fluminense, Rua Mário Santos Braga 30, Sala 415, Campus do Valonguinho, 24020-140, Niterói, Rio de Janeiro, Brazil.
Clin Nutr. 2018 Feb;37(1):208-213. doi: 10.1016/j.clnu.2016.12.005. Epub 2016 Dec 18.
BACKGROUND & AIMS: Basal metabolic rate (BMR) is an important physiologic measure in nutrition research. In many instances it is not measured but estimated by predictive equations. The purpose of this study was to compare measured BMR (BMRm) with estimated BMR (BMRe) obtained by different equations.
A convenient sample of 148 (89 women) 20-60 year-old subjects from the metropolitan area of Rio de Janeiro, Brazil participated in the study. BMRm values were measured by an indirect calorimeter and predicted by different equations (Schofield, Henry and Rees, Mifflin-St. Jeor and Anjos. All subjects had their body composition and anthropometric variables also measured. Accuracy of the estimations was established by the percentage of BMRe falling within ±10% of BMRm and bias when the 95% CI of the difference of BMRe and BMRm means did not include zero.
Mean BMRm values were 4833.5 (SD 583.3) and 6278.8 (SD 724.0) kJday for women and men, respectively. BMRe values were both biased and inaccurate except for values predicted by the Anjos equation. BMR overestimation was approximately 20% for the Schofield equation which was higher comparatively to the Henry and Rees (14.5% and 9.6% for women and men, respectively) and the Mifflin-St. Jeor (approximately 14.0%) equations. BMR estimated by the Anjos equation was unbiased (95% CI = -78.1; 96.3 kJ day for women and -282.6; 30.7 kJday for men).
Population-specific BMR predictive equations yield unbiased and accurate BMR values in adults from an urban tropical setting.
基础代谢率（BMR）是营养研究中重要的生理指标。在许多情况下，BMR 并未直接测量，而是通过预测方程进行估算。本研究旨在比较通过不同方程获得的实测 BMR（BMRm）和估算 BMR（BMRe）。
本研究便利选取了巴西里约热内卢大都市区 148 名（89 名女性）20-60 岁的研究对象。通过间接测热法测量 BMRm 值，通过不同方程（Schofield、Henry 和 Rees、Mifflin-St. Jeor 和 Anjos）进行预测。所有研究对象均进行了身体成分和人体测量变量的测量。通过以下方法确定估算的准确性：BMRe 中有多少百分比落在 BMRm 的±10%范围内，以及当 BMRe 和 BMRm 平均值差值的 95%CI 不包括零时的偏差。
女性和男性的平均 BMRm 值分别为 4833.5（SD 583.3）和 6278.8（SD 724.0）kJday。除了 Anjos 方程预测的值外，所有其他方程预测的 BMRe 值都存在偏差和不准确。Schofield 方程的 BMR 高估约为 20%，而 Henry 和 Rees 方程（女性和男性分别为 14.5%和 9.6%）和 Mifflin-St. Jeor 方程（约为 14.0%）的 BMR 高估则相对较低。Anjos 方程预测的 BMR 无偏差（女性 95%CI = -78.1；96.3 kJ day，男性 95%CI = -282.6；30.7 kJday）。
特定于人群的 BMR 预测方程在城市热带环境中为成年人提供了无偏差和准确的 BMR 值。
