Submitted to: International Congress on Obesity
Publication Type: Abstract Only
Publication Acceptance Date: 8/29/1998
Publication Date: N/A
Technical Abstract: This study reports the development of age, sex and ethnic-specific bio- electric impedance (BIA) prediction equations to estimate FFM, TBF, and %BF when applied to the BIA data in the National Health and Nutrition Ex- amination Survey III (NHANESIII). A means of addressing the international health concerns regarding obesity is by developing these BIA prediction equations to estimate body composition. Many of the published BIA predic- tion equations are from small samples and are affected by the problem of sample specifically. The validation sample (N=1842) consisted of data sets from New York City, North Dakota, and Ohio. The independent cross-valida- tion sample (N=548) consisted of data sets from Chicago, Maryland and Ohio. The subjects were separated into four age groups, i.e., 12-24.99, 25-44.99, 45-64.99 and 65+ years. The variable were age, stature, weight, BMI resistance, reactance, S2/R, and FFM, TBF, and %BF from DXA. A maximum R2 regression analysis was used to develop the prediction equations and the prediction equations were cross-validated. For the whites, the R2 values are between 0.84 to 0.98 for males and 0.76 to 0.96 for females. The RMSEs for white males and females range from about 1.9 kg to almost 3.0 kgs and 1.6 to 2.2 kgs respectively for FFM. For the black males and females for FFM, the R2 values ranged from 0.78 to 0.99 for males and from 0.54 to 0.99 for females. The RMSEs were from about 2.0 to 3.2 kgs for males and 1.4 to 3.5 kgs for females. TBF and %BF were calculated for each individual from the predicted FFM values, and the pure errors ranged from 2.0 to 3.0 kgs and 3.0 to 4.0% respectively. These BIA prediction equations are to be applied to the BIA data in the NHANES III survey to estimate reference values, FFM, TBF and %BF (Guo et al., under review).