Correlations among abdominal anthropometry and % body fat Below are tables of Pearson product-moment correlations among anthropometric measurements made in the abdominal region and measures of % fat. Measures are 2 abdominal girths (natural waist and umbilicus), the abdominal skinfold thickness, and the biilicac diameter (width at the hips). The tables are derived from 2 studies conducted at NHRC: (1) the data set from which the original Navy equations were derived (Tables 1 & 2), and (2) the data set used to develop the 4-compartment equation briefly used by the Marine Corps. From these tables several facts should be evident. In the first data set the only measure of % fat is the 2-compartment model of Siri. In the second data set, there are 3 measures of body fat content: the Siri model, a DXA measurement, and the 4-compartment model of Friedl. It should be evident that: (1) abdominal girth is a slightly better indicator of fat content than waist girth, or abdominal skinfold thickness for men, (2) waist girth is a slightly better indicator of fat content (except that from DXA) than abdominal girth or abdominal skinfold thickness for women, and (3) bi-iliac diameter is not nearly as good an indicator of body fat content as either the girths or skinfold thickness.

Table 1. Results for men from NHRC equation development study

Table 2. Results for women from NHRC equation development study.

Table 3. Results for men from NHRC 4-compartment study

Table 4. Results for women from NHRC 4-compartment study

Figure 1. Relationship between Fat Content from DXA and from Siri

At right is a figure depicting the relationship between % fat from DXA and that from the Siri 2-compartment model. The solid line represents the regression of fat from Siri on fat from DXA. The dashed line is the line of identity. As can be seen the lines intersect at about 10% fat and the DXA fat estimate underestimates Siri fat by an increasing amount as the percent fat value increases. These values are from a Hologic 3500 DXA. Results from use of Lunar machines differ (they may actually over-predict as percent fat increases.) DXA is well suited to studies of change in body composition, but is not a replacement “gold standard”.