Modelos preditivos da composiÃÃo corporal utilizando a antropometria e a absortometria de raios-x de dupla energia




Several specific and generalized equations to predict body mass and/or body composition components, fat percentage (%F), fat mass (FM) and lean mass (LM) are found in the literature. Simple methods, e.g. anthropometry, as well as more sophisticated techniques, e.g. the dual-energy X-ray absorptiometry (DXA), which has been recognized as the gold standard method, have been used to assess body composition. The differences among populations, the increase in the prevalence of overweight, obesity and chronic-degenerative diseases in young adult men lead to a necessity for research in different regions in several phases of a manâs life. Thus, the objective of this study is to describe the aspects involved in the developmental and validation process as well as the possibilities of applying predictive equations for body fat, establishing a comparison of %F, FM and LM values, estimated by predictive models and measured by DXA. In this cross-sectional study. Forty-five men ranging from 20-30 years were assessed at the Human Performance Laboratory of Pernambuco University- Recife/PE and at the Albuquerque do à Laboratory, Recife-PE, during the december/2004 of august/2005. The following anthropometric measurements were used: total body mass, height, waist circumference, hip circumference, central (SS: subscapular, SI: supra-iliac, mid-axillary and ABD: abdominal) and peripheral skinfold thickness (TR: triceps, BI: biceps, TH: thigh and LE: leg), anthropometric index: body mass index (BMI), waist-hip ratio (WHR), waist-height ratio (WHtR) and %F, FM and LM of upper and lower-body, trunk and total body measured by DXA. The analyses were carried out in the statistical package SPSS, version 11.0. Statistical significance was set at 5%. Lohman (1992) cross validation criteria, besides the analysis of residual scores proposed by Bland e Altman (1986), the simple linear regression, and the Receiver Operating Characteristic (ROC) were used. Out of seven predictive models included in the cross validation process, only the Durnin and Womersleyâs was validated. There was a strong association between WHtR and %F of upper body, lower body, trunk and total body; the greatest area under the ROC curve was WHtR, 0.87 (0.72-0.99). Therefore, Durnin and Womersleyâs; equation was recommended to be used in the studied group. WHtR was the most indicated anthropometric measure to explain %F variation by compartment, presenting the greatest explanatory power of cardiovascular risk. LM demonstrated lower variability in relation to %F and FM, being suggested as the explanatory variable for predictive models and the component of body composition indicated as a form to reduce body proportion.


anthropometry predictive models validaÃÃo validation dual-energy x-ray absorptiometry absortometria de raios-x de dupla energia nutricao composiÃÃo corporal â modelos preditivos antropometria

Documentos Relacionados