Indices of relative weight and obesity

Abstract

THE NEED for an index of relative body weight was recognized from the beginning of anthropometry, that is to say as soon as serious attention was given to the dimensions of the body and their biological and medical implications. Body weight in proportion to height or to some function of height is interesting because it should indicate something about ‘build’ or shape and about obesity or fatness. Various indices of relative weight have been espoused and applied for many years but as yet there is no agreement on any particular index. In part this reflects confusion-or at least lack of agreement-about what a relative weight index should represent and mean; in part the reason is a lack of “calibrating” data and of systematic examination of wide-ranging samples of data analyzed in parallel. The purpose of this paper is to provide a comparison of various indices of relative weight as applied to data on weight, height and body fatness of men in several countries in Europe, in Japan, men in South Africa, as well as of white men in the United States. In the present paper guidance in the analysis was provided by two assumptions. First, it is assumed that a major reason for the use of a relative weight index is to remove the dependency of weight on height. Second, it is assumed that in the selection of an index attention should be given to the degree to which the index may indicate relative obesity or body fatness. Superficially, it might seem simplest and most informative to express the weight of the individual as a percentage of the average weight of persons of the same height, age and sex in the population to which he belongs. That was the reasoning that led to publication of “standard height-weight” tables by the life insurance industry, beginning with the Medico-Actuarial Mortality Investigations of 1912.1 As originally published, the life insurance industry tables simply provided, for the two sexes, average weights, in pounds, at specified ages and heights, in inches. Those measures were recorded, ‘as customarily dressed in indoor clothing’, in connection with application for life insurance. Roughly, at least for men, it seems that the extra height added by the shoes may be compensated for by the extra weight added by the shoes and the rest of the ‘indoor clothing’, so that similar relationships should hold for barefoot height and nude weight or in light underclothing. We have published a metric system version of those 1912 tables based on smoothed plots of the discrete values in the original tables.2 In the present paper, ‘relative body weight’ means the body weight expressed as a percentage of those tabular ‘averages’. Later tabulations of similar ‘average’ weights and heights have been published by the life insurance industry, notably in 19592; the differences are not large and their meaning is obscured by such questions as whether ‘customarily dressed’ in the period 1935 through 1953, covered by the 1959 report, is comparable to the same term in the period 1898–1905 covered by the 1912 report. It should be noted that for some purposes it is immaterial whether a proposed table of ‘standard’ weights accurately portrays the average (or median) of the reference population. So long as the standard is accepted and all sets of data on individuals are referred to it, the validity of comparisons between individuals or groups will be independent of the ‘quality’ of the table. Populations differ from one another and populations change. Average values for weight and height for given age and sex for a given population do not necessarily apply to other populations or even to the same population at another time. Further, there is no present prospect of obtaining for any population true average values of weight for given height, age and sex. Certainly persons examined in connection with application for life insurance are far from being a random sample of the population. Apart from such considerations, scientifically minded investigators would much prefer an index of relative weight in purely physical units, unchanging from population to population or from time to time. Obviously we are concerned with weight per unit height but elementary consideration of the ratio, W/H, leads to the question as to why the weight, which is roughly equivalent to a volume, a three-dimensional or cubic unit should be standardized in terms of a single linear dimension. Obviously, if the body had the same form at different heights, weight would tend to be proportional to the third power of the height. That idea was incorporated by Livi in his indice ponderale, literally ‘ponderal index’ in English: the cube root of the body weight divided by the body length or height.3 In spite of the fact that it is easy to show that the body form does not remain constant with increasing length, the ponderal index, or the similar Rohrer index, W/H3, has been rather widely used. Florey4 suggests this reflects the use of the ponderal index in the popularization o