Waist circumference as an indicator of adiposity and the relevance of body height.

Waist circumference (WC) is an obvious indicator of adiposity, but there is a confusing diversity of approaches to its quantitative use. Because taller people of any given relative fat content tend to have bigger WCs, indices of adiposity commonly take the form WC/height(q), where q is 1 or less. Sometimes the influence of height is regarded as insignificant, so that q is taken as zero. More attention has been given to such indices than to establishing how to use them for predicting adiposity. The usual approach has been empirical and statistical, but many published findings can be integrated using a more analytical approach. This leads to several hypotheses that are supported by published evidence, but which remain to be fully tested. Testing mainly requires data sets of the kind that have already been used for related purposes. The main hypotheses are as follow. For adults, the preferred index is WC/height(0.5), while for adolescents and children it is WC/height. However, to obtain equations for the prediction of percentage body fat, the latter should be regressed on the reciprocals of the squares of these indices, namely 1/(WC(2)/height) and 1/(WC/height)(2) respectively. These expressions decrease with increasing percentage body fat, but should do so in a linear manner. Two other hypotheses are that the non-fat content of the abdomen tends to increase with percentage body fat and that the ratio WC(2)/height is approximately proportional to the body mass index. The analysis is based mainly on the following ideas: firstly, that the area of fat or adipose tissue in a transverse computed tomogram of the abdomen equals the total area less the fat-free area, the former being the main determinant of WC and the latter being partly determined by fat-free body size as represented by height; secondly, that guidance can be usefully be sought in simple, dimensionally-correct models of body form, but that parameters in the resulting equations may become attenuated by the necessary use of regression analysis or the maximizing of correlations. There may be a better measure of fat-free body size with which to replace height in these indices, but the waist-to-hip ratio is unhelpful. It is hoped that this analysis can be usefully extended to the problem of estimating intra-abdominal (visceral) fat from waist circumference. Copyright 2010 Elsevier Ltd. All rights reserved.