Santos Diana A, Silva Analiza M, Matias Catarina N, Magalhães João P, Minderico Cláudia S, Thomas Diana M, Sardinha Luís B
Exercise and Health Laboratory, CIPER, Faculdade de Motricidade Humana, Universidade de Lisboa, Lisbon, Portugal.
Exercise and Health Laboratory, CIPER, Faculdade de Motricidade Humana, Universidade de Lisboa, Lisbon, Portugal.
Nutrition. 2015 Jul-Aug;31(7-8):948-54. doi: 10.1016/j.nut.2015.02.003. Epub 2015 Feb 19.
Recently, new body indices, including body adiposity (BAI), a body shape (ABSI), and body roundness (BRI) indices have been developed to estimate adiposity. The aim of this study was to compare percent fat mass (%FM) with novel indices in an elite athlete population.
Using a cross-sectional design, %FM in 159 male and 50 female athletes using a four-component model was assessed. The %FM was compared with body mass index (BMI), BAI, ABSI, BRI, and with other %FM field methods (bioimpedance spectroscopy and skinfold prediction equation). These associations were determined using multilinear regression analysis, which resulted in predictive models of %FM in athletes. Cross-validation was performed using the prediction residual error sum of squares (PRESS) statistics method.
Although higher associations than other indices were observed, BRI still presented low coefficients of determination (men: R(2) = 0.36; women: R(2) = 0.25) when comparing with other field methods (R(2) range, 0.33-0.75). Using BAI as the independent variable, the R(2) was 0.07 for men and 0.14 for women. ABSI did not result in a significant association with %FM in women (R(2) = 0.05) while in men a significant association was found (R(2) = 0.22). The BMI model resulted in a R(2) = 0.20 for men and R(2) = 0.22 for women. Waist circumference and the sum of skinfolds were the anthropometric variables with the highest association with adiposity. New alternatives were presented with higher coefficients of determination (PRESS R(2) ranged from 0.47 to 0.71).
The newly developed body indices are limited in predicting %FM in elite athletes, particularly when compared with other commonly and readily available field methods like bioimpedance analysis or skinfold prediction models.
最近，已开发出包括身体肥胖指数（BAI）、身体形状指数（ABSI）和身体圆润度指数（BRI）在内的新身体指数来估计肥胖程度。本研究的目的是在精英运动员群体中比较体脂百分比（%FM）与这些新指数。
采用横断面设计，使用四成分模型评估了159名男性和50名女性运动员的%FM。将%FM与体重指数（BMI）、BAI、ABSI、BRI以及其他%FM现场测量方法（生物电阻抗光谱法和皮褶预测方程）进行比较。使用多元线性回归分析确定这些关联，从而得出运动员%FM的预测模型。使用预测残差平方和（PRESS）统计方法进行交叉验证。
尽管与其他指数相比观察到更高的关联性，但与其他现场测量方法（R²范围为0.33 - 0.75）相比，BRI的决定系数仍然较低（男性：R² = 0.36；女性：R² = 0.25）。以BAI作为自变量，男性的R²为0.07，女性为0.14。ABSI与女性的%FM没有显著关联（R² = 0.05），而在男性中发现有显著关联（R² = 0.22）。BMI模型男性的R² = 0.20，女性的R² = 0.22。腰围和皮褶总和是与肥胖程度关联最高的人体测量变量。提出了具有更高决定系数的新替代方法（PRESS R²范围为0.47至0.71）。
新开发的身体指数在预测精英运动员的%FM方面存在局限性，特别是与生物电阻抗分析或皮褶预测模型等其他常用且易于获得的现场测量方法相比。
