Asia Pacific J Clin Nutr (1995) 4: 137-139
Asia Pacific J Clin Nutr (1995)
Body mass index and nutritional
status: the effect of adjusting body mass index for the relative sitting
height on estimates of the prevalence of chronic energy deficiency,
overweight and obesity
Department of Human Sciences, Loughborough
University, Loughborough, Leics LE11 I 3TU, UK
Low body mass index (BMI) has been proposed as
a practical measure of adult chronic energy deficiency (CED),
although it was well-known limitations. One of these is that its
interpretation is complicated by the influence of body proportions,
in particular the relative leg length. This has been quantified
by examining data collected before 1970 of 349 adult Australian
Aborigines following a largely traditional way of life. These
Australian Aborigines exhibited low sitting height: stature ratios
(SH/S), 0.48± 0.02, (mean ± sd), range 0.41-0.54, ie they are relatively long legged, and
low BMI, 19.9± 3.2, range 12-30 kg/m2.
Thirty percent of individuals had BMI less than 18.5 kg/m2,
a suggested cut-off for CED. The regression of BMI on SH/S was
determined in men and women separately but found by covariance
analysis not to be different and a common equation for both sexes
was calculated. When BMIs were standardized to a SH/S of 0.52,
a value found in Europeans and other Indo-Mediterraneans, the
percentage classed as chronically energy-deficient fell to 7%.
In Asians and indigenous Americans with their high SH/S, the percentage
of the population with overweight and obesity could be overestimated
and the extent of CED underestimated. In populations with a mean
SH/S of 0.52, such as Europeans and Pacific peoples, standardizing
SH/S to 0.52 would not effect the prevalence of CED or overweight
and obesity but could move individuals across diagnostic boundaries
as there is variability in SH/S in all population groups. In conclusion,
when using BMI to assess energy nutritional status single cut-offs
are not applicable to all individuals and population groups without
allowance for the body form and type.
The body mass index (BMI, weight/height2,
kg/m2) is used widely as a measure of overweight and obesity.
On the basis of the mortality experience of insured North Americans,
the desirable range of BMI in men is regarded as 20-25 kg/m2
and in women 18.7-23.8 kg/m2 1. By common use,
BMI of 25-30 is described as overweight, and of over 30, as obesity.
More recently, it has been proposed that low BMI can be used as a
practical measure of energy undernutrition2,3. Three grades
of chronic energy defici 1000 ency (CED) have been defined, as BMI
less than 18.5, 17 and 16 kg/m2.
BMI has several well-known limitations such as being
a poor measure of energy stores in some age groups compared with other
anthropometric techniques, being a measure of size as much as composition4,5
and of being related differently to body composition in different
population groups5. Some of this variation is related to
differences in body proportions, particularly the ratio of sitting
height: stature (SH/S), also known as the relative sitting height.
In the Asia Pacific region, the widest range of SH/S in any geographical
area is found, being lowest in Australian Aborigines (0.475) and highest
in Japanese (0.55). This paper illustrates the effect of the SH/S
ratio causing an overestimation of the prevalence in CED in Australian
Aborigines. For Asians, the prevalence of CED will be underestimated
and those of overweight and obesity will be overestimated using the
Individual anthropometric data for 349 adult Australian
Aborigines were obtained from the Australian Institute for Aboriginal
and Torres Straits Islander Studies, Canberra, ACT, Australia. The
data was collected by the Australian physical anthropologist Andrew
A. Abbie, using Martin field anthropometry equipment, over a number
of years in groups living a largely traditional way of life or after
recent settlement. The data comes mainly from groups in central and
northern areas of Australia and were collected before 1970. The data
set used here has been described previously6. The original
data were cleaned and restructured by Freedman and Macho7
and further explored and restructed by the author.
Adults aged 20-65 years were selected for this analysis
although estimates of ages of all groups may have been only approximate.
Separate regression equations for the sexes were calculated for BMI
on SH/S. Testing by covariance analysis8 showed the slopes
and intercepts were not significantly different and the sexes were
combined. The combined equation was: BMI = 57.2* SH/S - 7.4, r2
= 015, standard error of the estimate = 2.9 kg/m2. This
equation was used to estimate the BMI at a SH/S ratio of 0.52, the
mean value in European populations including those used to determine
mortality rates of different weight-height and hence BMI groups. Individual
BMI were then standardized to SH/S of 0.52 by adding the residuals
to the estimated BMI at an SH/S of 0.52, using the following equation:
BMIstd = BMI0.52 + (BMI0-BMI1),
where BMIstd = standardized BMI
BMI0.52 = estimated BMI at SH/S of 0.52
BMI0 = actual BMI
BMI1 = estimated BMI at actual SH/S.
The proportions of men and women in the three groups
of CED or who were overweight or obese were calculated.
The ages and physical characteristics of the Australian
Aborigine men and women are shown in Table 1. The mean BMI and SH/S
are low compared with other population groups. The proportions in
different categories of BMI before and after standardization are shown
in Table 2. Standardization reduced the prevalence of CED (BMI<18.5
kg/m2) from 35 to 7%. The number of individuals classified
as Grade 3 (severe) CED, according to the classification of James,
Ferro-Luzzi and Waterlow2, fell from 30 (9%) to 1 (<1%).
Table 1. The age and physical characteristics
(mean ± SD) of the sample of Australian Aborigine men and women. 1000
|Sitting height, m
|Body mass index, kg/m2
|Sitting height: statue
Table 2. The effect of adjusting the body mass
index (BMI) to at relative sitting height of 0.52 in Australian Aborigines
on the distribution of BMI values.
CED = chronic energy deficiency.
The question of appropriate anthropometric reference
data and cut-off points for groups other than young adult Europeans
has always been controversial, particularly in the area of child growth
and nutritional status. Similar considerations apply in the case of
adults, where the use of BMI as a practical measure of CED and undernutrition
is being proposed. Here, the importance of standardization and cut-offs
arises from the narrow distribution of BMI around the cut-off points
which is illustrated by the steepness of the cumulative frequency
plot of BMI in most Third World groups. S 1000 mall changes in cut-offs
have marked effects on the numbers falling into different classes.
Conversely, small adjustments in BMI to standardize for variations
in, for example SH/S, also affect prevalences.
Australian Aborigines have a relative leg length greater
than that of most other groups and hence the effect demonstrated may
not be exceeded in other groups unless BMI are distributed more critically
around the cut-offs. In contrast, mean SH/S in Far East Asian groups
are higher than in other population groups, being 0.55 in Japanese,
0.54 in Chinese and Koreans and slightly less in Thai and Vietnamese9.
In these groups, standardization to the European mean of 0.52 would
reduce the BMI with the effects of reducing the estimated prevalence
of overweight and obesity and increasing that of CED. Of the other
groups, Africans have the lowest SH/S, the longest legs relatively
and, like the Australian Aborigines, the prevalence of CED may be
overestimated unless an allowance is made for SH/S. Even in groups
with SH/S of 0.52, where standardization will not affect the estimated
overall prevalence of overweight and obesity or CED, because of the
considerable intra-group variation in SH/S, individuals will move
across cut-offs and boundaries. As an example, in a sample of 420
Papua New Guinean adults of mean SH/S 0.52, standardization of individual
scores to 0.52 did not affect the estimated prevalence of CED but
more than 12% of individuals changed BMI by 1 unit or more (Norgan,
unpublished). In conclusion, low BMI has been a feature of many Australian
Aborigine groups due, in part, to a low SH/S, arising from their relatively
long legs. This is supported by paradoxically high adipose and muscle
mass status6. Therefore, single cut-offs of BMI are not
applicable to all population groups and allowance has to be made for
the body proportion in using BMI to assess energy nutritional status.
AcknowledgmentsThis work was carried out during a Sabbatical Leave at the Department
of Anatomy and Human Biology, University of Western Australia with
support from The Royal Society, The Wellcome Trust and the Clive and
Vera Ramaciotti Foundation. The individual data were made available
by the Australian Institute of Aboriginal and Torres Strait Islander
- FAO/WHO/UNU, Energy and Protein Requirements. Tech
Rep Ser No 724. Geneva: WHO,1985.
- James WPT, Ferro-Luzzi A, Waterlow JC. Definition
of chronic energy deficiency in adults. Report of a Working Party
of the International Dietary Energy Consultative Group. Eur J Clin
Nutr 1988; 42:969-981.
- Ferro-Luzzi A, Sette S, Franklin M, James WPT.
A simplified approach to assessing adult chronic energy deficiency.
Eur J Clin Nutr 1992; 46: 173-176.
- Norgan NG. Body mass index and body energy stores
in developing countries. Eur J Clin Nutr 1990; 44 (Suppl l ): 79-84.
- Norgan NG. Population differences in body composition
in relation to the body mass index. Eur J Clin Nutr 1994, 48, Suppl.
- Norgan NG. On the interpretation of low body mass
indices: Australian Aborigines. Am J Phys Anthropol 1994,94,229-237.
- Macho G, Freedman L. A Reanalysis of the Andrew
A Abbie Morphometric data on Australian Aborigines. Occasional Papers
Hum Biol No. 4. Canberra: Australian Institute of Aboriginal Studies,1987.
- Scedecor GW, Cochrane WG. Statistical Methods.
6t 2f6 h Edn. Iowa: Iowa State University Press,1967: 432-436.
- Pheasant S. Body space: anthropometry, ergonomics
and design. London: Taylor & Francis,1986.
Copyright © 1995 [Asia Pacific Journal of Clinical
Nutrition]. All rights reserved.
January 19, 1999
to the top