Anatomical distribution of subcutaneous fat and its description by multivariate methods: how valid are principal components?

Principal components analyses of subcutaneous fat measurements covering arm, trunk, and leg areas of the body were performed on 44 samples varying in sex, age, and ethnicity of subjects; measurement technique; and number of sites studied. Three major components emerged, similar in all samples: A first component of obesity, a second component of anatomical fat distribution contrasting trunk and extremity fat, and a third distributional component contrasting leg and arm fat. These components did not appear to vary with age (school children to older adults), sex, ethnicity (Black-White), technique (caliper, ultrasound, or X-rays), or number of sites (five to 13). The three major components are similar to vectors or obesity and relative fat patterning reported earlier by anthroposcopic (Skerlj et al., 1953) and graphic (Garn, 1955) methods. One of the pattern components (trunk-extremity) appears to correspond to a body build continuum associated with adult chronic diseases (Vague, 1956). The stability of these components among samples that differ in several characteristics suggests the validity of these components in describing aspects of body build of anthropological and epidemiological importance.