Weight-height indices as estimators of fatness in men.

Weight-height (W-H) indices continue to be popular methods of assessing fatness in both clinical practice and population studies, even though skinfold thicknesses provide simple valid estimates. Different grades of obesity have been defined in terms of W/H2 values but these have not been related to fatness. So far, no allowance has been made for the changing body composition with age which alters the relationships between fatness and W/H2 or the other W-H indices. These relationships were calculated in an industrial sample of 138 adult men and 22-55 yr. W/H2 and relative weight (weight/desirable weight) were the best of the W-H indices of fatness, having the lowest correlations with height (r = 0.07), the highest correlation with percentage of body weight as fat (%F) (r = 0.75), and with fat mass (r = 0.88), and the lowest standard error of estimate (SEE) (4.1%F). However, W/H and W/H3 had very similar r and SEE with %F. Adding age to W/H2 improved the accuracy of the estimation of %F, (R2=0.66 SEE = 3.6%), and reduced the overestimation of %F in the young and underestimation in the middle-aged that occur with W/H2 alone. Validated equations for estimating %F and fat mass in men from W/H2 and age are presented. However, W-H indices should be used cautiously as they cannot distinguish between overweight due to muscle, bone, water or fat.