Comparison of predicted body fat from anthropometric methods: Body mass index and skinfold-thickness

http://dx.doi.org/10.5007/1980-0037.2008v10n1p1 The present study was designed to determine the level of agreement between four widely used skinfold-thickness equations and body mass index-based prediction equations (%BFBMI) for evaluating body fat percentage in university students. A total of 431 subjects (174 females and 257 males) aged 18-30 years old were included in the study. Skinfoldthickness equations proposed by Jackson and Pollock, Durnin and Womersley, Petroski and Guedes and Guedes were included in the study. The %BFBMI was predicted using Deurenberg’s age-and-sex specifi c equation. For the assessment of agreement, the paired t test, intraclass correlation coeffi cients and Bland-Altman plots were computed for the correlation between predicted percentage body fat according to each skinfold-thickness equation and %BFBMI. The results showed that the mean values of percent body fat predicted by each skinfold-thickness equation and from body mass index did not differ signifi cantly. Intraclass correlation coeffi cients ranged from 0.463 to 0.612. The Bland-Altman analysis indicated that the mean differences were close to zero, especially in males, but demonstrated a limited individual capacity of agreement, with upper and lower confi dence limits at -13.5% and 1.0%, respectively. In conclusion, these data provide evidence that, although the values of percentage body fat predicted by each skinfold-thickness equation and using body mass index were similar at a group level, individual biases were sometimes exceptionally high. It is recommended that comparisons between percentage body fat predicted by different anthropometric methods be interpreted with caution