Heymsfield Steven B, Peterson Courtney M, Thomas Diana M, Heo Moonseong, Schuna John M, Hong Sangmo, Choi Woong
From the Pennington Biomedical Research Center, LSU System, Baton Rouge, LA (SBH, CMP, and JMS); Montclair State University, Montclair, NJ (DMT); Albert Einstein College of Medicine, Bronx, NY (MH); and Hanyang University, Seoul, Republic of Korea (SH and WC).
Am J Clin Nutr. 2014 Dec;100(6):1455-61. doi: 10.3945/ajcn.114.088831. Epub 2014 Oct 8.
Body mass index (BMI) is formulated on the assumption that body weight (BW) scales to height with a power of 2 (BW∝height(2)), independent of sex and race-ethnicity. Powers differing from 2 are observed in studies of selected samples, thus raising the question if BMI is a generalizable metric that makes BW independent of height across populations.
The objectives were to test the hypothesis that adult BW scales to height with a power of 2 independent of sex and race-ethnicity and to advance an understanding of BMI as a measure of shape by extending allometric analyses to waist circumference (WC).
We conducted cross-sectional subject evaluations, including body composition, from the NHANES and the Korean NHANES (KNHANES). Variations of the allometric model (Y = αX(β)) were used to establish height scaling powers (β ± SE) across non-Hispanic white and black, Mexican American, and Korean men and women.
Exploratory analyses in population samples established age and adiposity as important independent determinants of height scaling powers (i.e., β). After age and adiposity in the next series of analyses were controlled for, BW scaling powers were nonsignificantly different between race/ethnic groups within each sex group; WC findings were similar in women, whereas small but significant between-race differences were observed in the men. Sex differences in β values were nonsignificant except for BW in non-Hispanic blacks and WC in Koreans (P < 0.05). Nationally representative powers for BW were (NHANES/KNHANES) 2.12 ± 0.05/2.11 ± 0.06 for men and 2.02 ± 0.04/1.99 ± 0.06 for women and for WC were 0.66 ± 0.03/0.67 ± 0.05 for men and 0.61 ± 0.04/0.56 ± 0.05 for women.
Adult BW scales to height with a power of ∼2 across the 8 sex and race/ethnic groups, an observation that makes BMI a generalizable height-independent measure of shape across most populations. WC also follows generalizable scaling rules, a finding that has implications for defining body shape in populations who differ in stature.
体重指数（BMI）的制定基于这样一种假设，即体重（BW）与身高的平方成正比（BW∝身高²），与性别和种族无关。在对选定样本的研究中观察到幂次不同于2的情况，因此引发了一个问题，即BMI是否是一个可推广的指标，能使体重在不同人群中独立于身高。
目的是检验这一假设，即成年人体重与身高的平方成正比，与性别和种族无关，并通过将异速生长分析扩展到腰围（WC）来加深对BMI作为体型测量指标的理解。
我们对来自美国国家健康与营养检查调查（NHANES）和韩国国家健康与营养检查调查（KNHANES）的受试者进行了横断面评估，包括身体成分分析。使用异速生长模型（Y =αXβ）的变体来确定非西班牙裔白人和黑人、墨西哥裔美国人以及韩国男性和女性的身高缩放幂次（β±标准误）。
在人群样本中的探索性分析确定年龄和肥胖是身高缩放幂次（即β）的重要独立决定因素。在接下来的一系列分析中控制了年龄和肥胖因素后，每个性别组内不同种族/族裔群体之间的体重缩放幂次差异不显著；女性的腰围结果相似，而男性中观察到种族间存在虽小但显著的差异。除了非西班牙裔黑人的体重和韩国人的腰围外，β值的性别差异不显著（P < 0.05）。全国代表性的体重缩放幂次（NHANES/KNHANES），男性为2.12±0.05/2.11±0.06，女性为2.02±0.04/1.99±0.06；腰围的缩放幂次，男性为0.66±0.03/0.67±0.05，女性为0.61±0.04/0.56±0.05。
在8个性别和种族/族裔群体中，成年人体重与身高的缩放幂次约为2，这一观察结果使BMI成为大多数人群中可推广的、与身高无关的体型测量指标。腰围也遵循可推广的缩放规则，这一发现对于定义身高不同人群的体型具有重要意义。
