Chapter 2 BMI as a reflection of the body energy stores
As noted in the previous chapter, the body is composed of active tissue mass (i.e. fat-free mass) and the mass of body fat which is the principal energy store.
There are several methods for estimating these two components; some methods are sophisticated and expensive. For example, the K40 natural radioactive potassium content of the body can be measured by carrying out a whole body count, from this the total potassium content can be estimated. From the amount of total potassium it is possible to estimate the mass of lean tissue since potassium is the chief ion within the cells of the active tissue mass. Methods that measure dilution of isotopic (tritiated or deuterated water) or chemical tracers (antipyrine) can provide estimates of total body water which in turn reflects largely the active tissue mass of the body since fat tissues contain little water.
Fat mass can be estimated by densitometric measurements since underwater weighing provides a measure of the density of the body which, in turn, reflects the contribution of fat mass to the body density. This is a relatively accurate method although it is a technique confined to laboratories with facilities for weighing humans underwater and, presents a certain amount of discomfort for the person being measured. There are simpler and less costly methods such as measuring skinfold thicknesses or body impedance which have been validated against a reliable estimate of fat mass. BMI has also been used as a simple anthropometric index that reflects the body's fat content and hence the body's energy stores. By comparing BMI with estimations of body fat stores, obtained through reliable methods such as densitometry, it has been shown that the BMI correlates well with body fat (Norgan & Ferro-Luzzi, 1982). (See Table 2.1).
An anthropometric study of 138 shipyard workers, aged 22-55 years in northern Italy also shows that the BMI is highly correlated with fat mass of the body (r = 0.87) and is, therefore, a reasonably good index of the body energy stores as fat (Norgan & Ferro-Luzzi, 1982). In international comparisons these relationships exist irrespective of the population group studied. It has also been shown that the BMI correlates well with body fat expressed as a percentage of body weight. The correlations in both males and females are, however, agedependent, with the highest correlations being found among individuals whose ages range between 26 and 55 years. The correlations are not as good among the young and the elderly. Nevertheless, once the age and gender-specific equations are used, the value for percent body fat predicted from BMI is close to that actually derived (Deurenberg et al., 1991). (See Table 2.2).
TABLE 2.1 Correlation between densitometric estimates of body fat composition and BMI from selected studies
Categories |
Number |
Correlations with BMI |
Males |
55 |
0.72 1 |
Females |
26 |
0.80 1 |
Students |
189 |
0.85 2 |
Executives |
249 |
0.67 2 |
Males |
245 |
0.71 3 |
Females |
324 |
0.82 3 |
Males |
141 |
0.77 4 |
Females |
135 |
0.76 4 |
Males |
474 |
0.71 5 |
Males |
150 |
0.90 6 |
Females |
213 |
0.95 6 |
Sources:
1 Allen et al., 1956. |
4 Roche et al., 1981. |
2 Keys et al., 1972. |
5 Revicki & Israel, 1986. |
3 Womersley & Durnin. 1977. |
6 Smalley et al 1990. |
Since measurements of subcutaneous skinfold thickness are used as a simple predictor of body fat stores, correlations have also been made between skinfolds and the BMI. There are good correlations between the BMI and the sum of several skinfold measurements (See Table 2.3).
The body fat stores in a woman are different from that of a man; thus, a BMI of 20 in a man indicates less body fat than it would in a woman. In either sex, BMI is an index of both the main energy stores, i.e. fat, and the active tissue mass of which muscle is the largest component. When the relationship between the body's fat content and BMI is calculated separately for both sexes, extrapolating the line to the point at which there would theoretically be no body fat reveals a BMI value in both sexes of about 15.5. This implies that women show a greater loss in body fat as a proportion of weight as they become more energy deficient although detailed studies on the body composition of very thin and undernourished men and women are not available.
In some age groups, the BMI is as much a measure of lean tissue mass as it is a measure of fatness. The BMI was found to be highly correlated with fat-free mass or active tissue mass of the body (r = 0.68) (Norgan & Ferro-Luzzi, 1982). BMI correlations with estimates of fat mass and lean tissue mass in individuals from developing countries, e.g. among 527 men and 546 women from Papua New Guinea and 159 men and 197 women from Ethiopia, show that BMI is reflective of both fat mass and fat-free masses (Norgan, 1990).
TABLE 2.2 Accuracy of BMI as a predictor of percent body fat
Percent body fat | |||
Number |
Observed mean BMI |
Difference with predicted BMI mean | |
Age groups |
|||
21-25 |
304 |
21.8 |
-0.1 |
26-35 |
49 |
24.7 |
-0.4 |
36-45 |
95 |
32.1 |
0.5 |
46-55 |
44 |
30.2 |
0.3 |
56-65 |
37 |
34.5 |
-0.3 |
> 66 |
50 |
34.1 |
-0.4 |
BMI groups |
|||
< 20 |
540 |
18.2 |
0.1 |
20-25 |
531 |
23.2 |
-0.1 |
26-30 |
109 |
32.4 |
0.3 |
_ 30 |
49 |
39.4 |
-0.5 |
Source: Deurenberg et al., 1991.
In some circumstances, the effectiveness of work performance depends upon a person's size or more specifically, on the amount of muscle mass. The power of the body to obtain leverage and the stabilizing effect of the body in a variety of tasks are dependent on weight as much as specific measures of lean tissue and body fat. Thus, there are advantages to having a combined index of lean and fat tissues.
Although a case can be made for the BMI as a useful general index of the mass of lean and fat tissue, it is also true that in individuals with the same BMI there will be different proportions of lean and fat tissues. Sometimes this variation is viewed as an indication that body weight or BMI is too crude an index of body energy to be used in classifying individuals as having chronic energy deficiency. It is true that in energy content terms, lean tissue, with its 70 percent water content and the rest principally protein or glycogen, has a much lower energy content on a weight basis than fat. So the purist is correct in claiming that ideally some good measure of the actual energy content of the body is needed. Nevertheless, it must be recognized that in functional terms which relate to energy use in physical activity it may be body bulk and muscle mass which is advantageous as well as having some energy reserves.
TABLE 2.3 Correlations BMI with sum of skinfold measurements in several populations
Categories |
Numbers |
Age |
Correlations with _ skinfolds (r) |
US students |
180 |
18-24 |
0.850 |
US executives |
249 |
49-59 |
0.777 |
Bantu |
116 |
31-60 |
0.732 |
Japanese farmers |
499 |
40-59 |
0.611 |
US railwaymen (sedentary) |
926 |
40-59 |
0.757 |
US railwaymen |
871 |
40-59 |
0.774 |
Finland (east) |
797 |
40-59 |
0.791 |
Finland (west) |
836 |
40-59 |
0 799 |
Crevalcore, Italy |
978 |
40-59 |
0.711 |
Montegiorgio, Italy |
636 |
40-59 |
0.797 |
Rome, railwaymen |
802 |
40-59 |
0.762 |
Source: Keys et al., 1972.
Body weight is easy to measure in comparison with the currently available methods for estimating body fat in difficult circumstances such as those found in developing countries. Thus, it has been difficult to train field workers in skinfold thickness measurements while impedance methods are only being validated now. Body weight has been widely used in North American and European countries and the body mass index is now internationally recognized as a valid measure of obesity in individuals despite the greater variation in body fat at the top end of the BMI scale. The application of the BMI index in developing countries is consistent with practices in developed countries. Incidentally, the BMI also allows for the simultaneous monitoring of the emerging problem of obesity in developing countries since it is a continuum index from grades of CED to grades of severe obesity.
Figure 2.1 - BMI and percentage body fat in adults from selected countries
Source: James et al., 1988
Data on BMI for men and women in Italy, Papua New Guinea, Ethiopia and Somalia are presented in Figure 2.1. Also included are data from 5,072 medically-screened and physically-fit British soldiers, most of whom were below 25 years of age with a presumed lean body mass at near maximum levels for their presumed genetic potential. Several important conclusions can be drawn from the data presented.
First, women in both developing and developed countries have a greater fat mass than men at each level of BMI. This is a clear biological difference between sexes found in other species as well. It is presumed to reflect the greater need for energy stores to maintain a buffer against a nutrition-induced failure in reproductive function and to allow the woman to cope with the substantial energy demands of lactation.
Second, the proportion of body fat in individuals from these developing countries (Papua New Guinea, Ethiopia and Somalia) was less than that of the Europeans (Italians and British) at equivalent BMIs. This is true for both men and women, but the effect is particularly marked in women. Perhaps the greater physical demand for activity of men and women in developing countries throughout their childhood, adolescence and adult life promote a greater muscle mass and a lower body energy content at identical body weights. Some genetic differences between races have been inferred from detailed studies of body form, however, there are increasing signs of the important long-term impact of environmental factors. Identifying these environmental factors may not be easy. There is mounting evidence that the nutritional level of an individual during foetal life can have a marked effect on the size and proportions of the body at birth. Breast-feeding rather than bottle-feeding also influences height growth and particularly the weight-for-height of children. These early events seem to programme the body's propensity to lay down lean and fat tissues later on in life, but other influences, such as mild protein and trace element deficiencies, e.g. zinc, may also be involved.
TABLE 2.4 Variation in body fat at a BMI of 20.0 among three rural populations
Papua New Guinea |
Ethiopia |
India | ||
Percent fat: |
1 |
7 |
12 | |
Fat mass (kg): |
||||
males |
1 |
4 |
6 | |
females |
6 |
8 |
8 | |
Source: Norgan, 1990.
An analysis by Norgan (1990) indicates that fatness or energy stores derived from BMI measures may vary in different population groups. Three rural communities in India, Ethiopia and Papua New Guinea were compared using a modelling approach and this indicated that the fat or energy stores at a given BMI differed markedly in the three population groups (See Table 2.4). Quite different levels of energy stores are thus encountered in different population groups (all from developing countries) for the same BMI. The differences between sexes in body energy stores represented by an equivalent BMI are also confirmed.
Naturally, these observations pose the question of whether different BMI values should be chosen as optimum for different populations and once more raise the issue of developing more specific measures of body fat. These questions were discussed many years ago when attempts were made to develop a rational approach to monitoring children's growth. Different countries developed their own national standards, however, after several decades it became apparent that well-fed children worldwide tended to grow at the same rate and that coherent objective analyses of comparative performances in different countries required only a single set of reference values. On the basis of this experience, it is probably wise to try to develop a single set of criteria using BMI as an index of chronic energy deficiency. This will maintain consistency with standard practice. Far more research will also be required before more sophisticated approaches to the BMI can be developed; meanwhile, its use as a simple index has barely begun.
Third, at normal ranges of BMI the relationship between BMI and percent body fat is approximately linear (though this tends to vary in different groups of individuals). However, at higher levels of BMI there tends to be a disproportionate increase in body fat.
These observations have considerable significance when BMI is used worldwide as an index of CED and may have to be considered when arriving at numbers of persons classified as CED in different populations in developing countries. Nevertheless, low BMI status in any community will reflect a low body energy store of individuals and a low fat free mass or lean body mass for a given stature. This is an equally important matter of concern and may be more typical of CED as a reflection of the state of adult nutrition in a community.
