The fatness-specific bioelectrical impedance analysis equations of Segal et al: are they generalizable and practical?

The fatness-specific bioelectrical impedance analysis (BIA) equations of Segal et al (Am J Clin Ntr 1988;47: 7-14; Segal equations) have been shown to be generalizable across sex, ethnicity, age, and degrees of adiposity. However, these fatness-specific equations require an a priori determination of percentage body fat (%BF) by using a skinfold equation or densitometry to categorize subjects into obese or nonobese groups. These procedures negate the use of BIA as a fast and simple method. It was hypothesized that the average of the Segal nonobese and obese fatness-specific equations (BIA average method) could be used in lieu of the skinfold method for categorizing subjects who are not obviously lean or obese. In phase 1 these three methods were compared for a subsample of 59 women who were not obviously lean or obese. The %BF of 75% of these subjects was accurately estimated within 3.5%BF by using the BIA average method whereas only 71% and 46% were accurately estimated by fusing the densitometric and skinfold methods, respectively. In phase 2, the predictive accuracy of the Segal fatness-specific equations, used in combination with the BIA average method, was compared with other BIA equations published previously for 602 American Indian, Hispanic, and white women and men. The Segal fatness-specific equations yielded the smallest prediction error (SEE = 2.22 kg for women and 3.59 kg for men) and the %BF of 70% of the subjects was accurately estimated within 3.5%BF compared with 24-59% for other BIA equations. Therefore, we recommend using the Segal fatness-specific and average equations to assess body composition in heterogeneous populations.