Sub-cutaneous fat thickness measured by magnetic resonance imaging, ultrasound, and calipers.

Sub-cutaneous fat thickness was measured at 12 sites on the body surface of 24 males and 26 females using magnetic resonance imaging (MRI), skinfold calipers, and A-mode ultrasound. The mean of the 12 fat thickness measurements and individual site thicknesses were compared between individuals using analysis of variance. In males, the mean thicknesses for ultrasound and calipers were similar (P greater than 0.05) but both were less than the MRI (P less than 0.001). MRI and ultrasound were similar in females but less than calipers (P less than 0.001). A good between-subject correlation was found between all three methods in the males but only the calipers and MRI were well correlated in females. Within-subject correlations are poor for all measures and in both sexes. Factor loadings for a varimax rotation of two principal components indicate that the fat is distributed in 1 of 2 patterns: either principally on the trunk or on the limbs. The principal component analysis and the result of canonical correlations obtained from the factor loadings confirm the findings of the analysis of variance, in that a general level of fatness is measurable by all three methods over a range of subjects. However, the pattern of fat thicknesses measured over a number of specific sites by one method of measurement is unlikely to be duplicated by either of the other two methods on the same individual.