ESTIMATION OF FEMALE BODY FAT PERCENTAGE BASED ON BODY CIRCUMFERENCES

Introduction: The increase in the prevalence of obesity entails the need for accurate low-cost methods to estimate body fat percentage (BF%). Objective: The aim of this study was to develop mathematical models to estimate the BF% of young women from southern Brazil using inexpensive equipment, based on body circumferences (BC), considering the shortage of specific studies of this population. Methods: Subjects were women (n=130) aged 18 to 35 years (26.06 +/- 4.41 years), from the city of Curitiba. Body mass and stature were measured and used to estimate body mass index (BMI), while BC measurements of the arm, forearm, waist (narrowest point and umbilicus), abdomen, hip, thigh and leg were taken for use in the mathematical models. Dual-Energy X-Ray Absorptiometry (DXA) equipment recorded the BF%. Pearson's correlation was used for anthropometric variables and age in association with BF%. The indicators with best correlations were used to estimate linear regression mathematical models for prediction of BF%. The results of the two models and of nine anthropometric equations were compared to those obtained with the DXA using Pearson's correlation coefficient and paired t-test. Results: Age was not significantly correlated with BF% (r=0.113 and p=0.200). The BCs with the highest correlations with BF% were abdomen, hip and waist - umbilicus (r=0.697, 0.682, 0.660, respectively, and p<0.001). Eight equations showed positive correlation with BF%, but only the results obtained with the models estimated in this study did not differ from those of DXA, with paired t-test. Mathematical models were developed with three (r=0.744, r(2) =0.554; t=0.16 and p=0.869) and six (r=0.768, r(2) =0.591; t=-0.04 and p=0.967) anthropometric variables. Conclusion:The use of mathematical models developed for estimating BF% based on body circumference and body mass is considered feasible.