Contents - Previous - Next

Low body mass index, physical work capacity and physical activity levels

Introduction
Work capacity or 'stress'
Low work capacity in a normal population
Relevance of form of activity
A theoretical analysis of effect; of BMI on activity
Proposed minimum level of BMI to effect work capacity and activity
BMI and the nature of work and activity
Conclusion
References
Discussion

J. V. G. A. Durnin

Institute of Physiology, University of Glasgow, Glasgow G12 8QQ, UK

In a normal population the distribution of body mass index (BMI) is such that a certain proportion of the population is likely to be at low values without necessarily being malnourished. However, although they may have low BMIs without being malnourished, they could certainly be physiologically and physically disadvantaged. An attempt is made to dissect out the probability of work capacity and physical activity being influenced by changes occurring in the human body with diminishing BMI.

The conclusion reached is therefore that before physical activity is affected, the BMI would probably have to be 17 or less, although it is possible that work capacity might be reduced before this level is reached. In any case, work requiring the use of the body mass - such as carrying loads, digging or shovelling earth or coal, pulling or cycling a rickshaw, stone splitting etc. - would impose a greater stress on people of low BMI.

Introduction

The body mass index (BMI) must have some sort of a distribution around the mean in a normal population, and although the distribution may be skewed towards the upper end of the range in most populations of reasonable nutritional status, there will be groups of individuals, albeit of decreasing proportions, towards the minimum levels of normal BMI.

A low BMI might therefore indicate simply the lower end of the distribution curve of a population of normal individuals and not necessarily reflect any abnormal nutritional state. However, if the BMI is low enough to really indicate under- or malnutrition, there is likely to be little argument that this will probably imply an effect on work capacity and, as a consequence, perhaps reduced levels of physical activity.

Work capacity or 'stress'

It is relatively easy to give a simplified connection between work or 'exercise' capacity, 'stress' and physical activity (Table 1). If an individual has an exercise capacity equivalent to 15 kcal/min, then a work situation which requires an energy expenditure of 7 kcal/min involves only modest stress to the individual since this level of activity needs <50% of his capacity. If, on the other hand, the exercise capacity is equivalent to 10 kcal/min then work at 7 kcal/min is at 70% of capacity and will be physically quite stressful. One could easily imagine that individuals with such restrictive capacities would tend to avoid any voluntary physical activity other than that of a light nature.

The questions confronting us are:

1. At what BMI levels are we likely to find low work capacities?

Table 1. Work capacity and 'stress' of work task

2. Would this situation necessarily imply malnutrition?

We could start our analyses by dealing with a situation where work capacity was equivalent to 10 kcal/min, or approximately 21/min of oxygen consumption or a VO₂ of about 35 ml/kg/min (assuming a body mass of between 55 and 60 kg).

Low work capacity in a normal population

This certainly represents a low work capacity but in the recent large-scale study in the UK the National Fitness Survey (Allied Dunbar, 1992) - about 10% of men between the ages of 35-55 years would come into the category of having this degree of low exercise capacity, although they might be assumed to be nutritionally reasonably fed, of average health and therefore probably physiologically normal. In the light of some surveys in the UK, we would have to assume that there existed substantial numbers of such individuals and that they were physiologically normal.

This is not, of course, to argue that all individuals within a distribution of 'normality' are equal. If we examine the distribution of genetically endowed cardiac output or haemoglobin levels, certain individuals are clearly disadvantaged compared with others. People are not equal. In relation to those at the lower end of the distribution curve in the present context, the crucial question is 'Is the situation due mainly to genetic factors with correspondingly considerable difficulty in effecting any improvement, or is this something which can be improved by nutritional or other intervention?'.

A third question we might ask is:

3. How often does work requiring >50% of the work capacity of an individual need to be done by people, either living and working in a rural environment in a developing country, or working in an impoverished urban setting?

The discussion of this general topic is being deliberately restricted to developing country situations, since, in a developed country, work which needs >50% of work: capacity, even of an individual with a very low work threshold, would be essential for only a very tiny proportion of the population.

Relevance of form of activity

I also intend to deal with two basically different types of activity, the first being exercise which does not involve working against a load - i.e. exercise such as in most forms of household tasks or most light agricultural work such as planting rice or groundnuts, weeding, walking etc., and the second type of work where a load factor would be present - i.e. carrying bags of grain, digging and shovelling etc.

In attempting to provide some answers to the first question - at what BMI levels are we likely to find low work capacities? - it would be helpful to have a reasonable quantity of experimental data which could be analysed. Unfortunately, very little exists in the published literature and much of the information has to be deduced indirectly. From some studies by Satyanaryana et al. (1977) in India, and Spurr (1987) in Columbia, and perhaps from the data of Desai et al. (1984) and Desai (1989) on adolescent boys in Brazil, it might appear that a BMI ~17 was a critical level for work capacity. The interesting studies in Guatemala described by Torun et al. (1989) where 'less lean body mass and muscle mass' had an influence on work output, reflect a different set of problems concerned with the nature of the work being done, and this will be discussed later.

A theoretical analysis of effect; of BMI on activity

1. Male 1.75m and 65 kg

Perhaps an attempt can be made to do a theoretical analysis of the likely effect of BMI on work or exercise capacity. Column I in Table 2 gives the approximate mass of the various organs and tissues in the body in a young adult man of 1.75 m in height and with a body mass of 65 kg, i.e. a physiologically normal and reasonably representative male. His BMI with these characteristics is ~21. In fact, these values for the mass of different organs and tissues depend on a very small number of dissections which have been carried out on fresh cadavers during the last 100 years or so. However, they are probably adequate for our present purposes. The metabolically most active tissues are present in relatively quite small masses, although this fact is not of major importance in the present context. Body fat will probably average 15% of body mass in a representative sample of relatively lean men. Muscle will comprise ~50% of the fat-free mass (FFM). With differing proportions of body mass, body composition and small differences in height, we can examine the possible effects on work capacity.

Table 2. Effect of variations of organ, tissue and body mass on body mass index (BMI) of an adult male

In relation to physical activity or work capacity, obviously the most important organ is skeletal muscle, although the heart and the oxygen-carrying capacity of the blood also play a significant part. The importance of skeletal muscle lies in more than just the mass of muscle - muscle fibre and the strength of muscular contraction have a role as well - but in the type of situation we are considering, which is mostly common forms of activity of only moderate intensity, the muscle mass is probably the most pertinent variable. Although it has little relevance to the subject matter of this paper, an age factor will influence the efficiency and the efficacy of muscular contraction; the proportion of the muscle fibre types alters with ageing, the muscle mass may diminish and the strength of contraction is reduced in the elderly. Although there is still uncertainty about the validity of the estimate of muscle mass, this is presumed to be reasonably correctly given by 50% of the FFM.

2. Male 1.7m and 60 kg

A reduction of body mass to 60 kg and in height to 1.7 m (shown in column 2, Table 2) leaves the BMI practically unchanged at 21. The change in body mass is unlikely to have a measurable influence on the mass of liver, gut, brain, nerve tissue, heart, kidneys, skin or bone. The mass of skeletal muscle may well have changed, as also will be the case with the fat depots in the body. We can surmise that in such an individual, this could mean that the amount of fat could become diminished to perhaps 10% of body mass, i.e. 6 kg. The FFM will then be 54 kg and muscle will account for about 27 kg - or almost exactly the same as in the original sample. Any likely error in these assumptions, as will also be the case with the next two examples, lies in the possibility that there may be a reduction in the mass of the skeleton and, to a small extent, the skin, and although these reductions would be small, the result would be a tendency to underestimate the muscle mass.

3. Male 1.7m and 52 kg

Column 3 in Table 2 shows the corresponding situation if we are dealing with an individual with a BMI of 18, and a body mass of 52 kg. (A very similar situation exists if we have a height of 1.65m and a weight of 49 kg.) Again, the mass of the metabolically active tissues will have altered, if at all, only fractionally. The fat mass, however, may well have decreased to almost minimal quantities and may be equivalent to only 6% of the total body mass, or about 3 kg. The FFM would then be 49 kg and muscle mass ~ 24.5 kg (50% of 52-49 kg). Again the reduction in muscle mass from the original example -from 27.5 to 24.5 kg - is small in absolute terms and unlikely to influence function. In fact, if we analyse these differences in absolute terms we are, of course, examining the wrong issue since proportionately there is a greater muscle/body mass ratio in the example in column 3 (and also 4) than is the case in column 1.

4. Male 1.7m and 49 kg

The final example (column 4, Table 2) deals with an even lower BMI of 17 in an individual of body mass 49 kg. The fat mass will probably be the same as before, so that the FFM will now become 46 kg, and muscle mass ~23 kg. This further reduction in muscle mass, but not in muscle/body mass ratio, is of such a degree that it seems improbable that muscular function will be jeopardized.

The examples I have chosen are selected and have some limitations. For instance, a height of 1.7 m may not be very appropriate for some of the body masses and tissue masses chosen, although these values would be quite commonly found in many population groups in India. But if we had chosen body heights of 1.65 m and 1.6 m, then comparable body masses and organ and tissue masses would have been correspondingly lower and the general argument would not have altered in any significant way.

Proposed minimum level of BMI to effect work capacity and activity

The foregoing has attempted some sort of analysis of the two questions posed at the beginning of this analysis. The two questions were:

1. How low does the BMI have to descend before it affects work capacity?
2. Does this situation necessarily imply malnutrition?

First, I have argued that the indirect evidence tends to suggest that BMI must be very low indeed (<17) before work capacity is affected. The probability is that physical activity of most forms will also not be diminished until BMI is again at very low levels. These conclusions are more or less supported by the data of Spurr (1987) and Satyanarayana et al. (1989). Secondly, malnutrition or undernutrition is not inevitably present with low BMIs. The proportion of men with low BMI in the sample of UK civilians and Armed Services (about 6000 men) (Durnin et al., 1984), those in the UK OPCS Survey (Knight, 1984), and those in the National Fitness Survey (Allied Dunbar, 1992) make the presence of undernutrition implausible.

BMI and the nature of work and activity

We might now examine the third aspect of this problem: what is the nature of work or physical activity that may be required of individuals with low BMIs and which might be stressful for them because their BMI was low?

If we examine the tables of the energy cost of the varied activities of occupational and non-working time, for industrialized but particularly for developing countries, which I initially assembled for the FAO/WHO/UNO Report on Energy and Protein Requirements (1985), and which were later expanded and included in Appendices 4.1 and 4.2 of Human Energy Requirements (James & Schofield, 1990), we can make some deductions concerning the influence of low BMIs on the physical stress related to these activities. The activities covered such things as household chores (various types of cooking), some occupational tasks such as carpentry, planting crops, and harvesting rice, walking; heavier activities such as pounding rice, clearing land, hoeing, cutting grass etc. and many leisure activities such as cycling, playing table tennis, football, dancing, swimming etc. The energy cost required by these types of activities would frequently be <5 times the basal metabolic rate (BMR) [i.e. the physical activity ratio (PAR) would be <51, and both this degree of strenuousness and the nature of the activity would make it unlikely that there would be a comparatively high physical stress to an individual with a low BMI (down to 17). A different set of criteria exists with some heavier activities such as cutting sugar cane, loading or carrying sacks of grain or cotton, digging earth or coal, using a sledge-hammer to hammer in fence posts, pulling or cycling a rickshaw, stone splitting, pushing a loaded wheel-barrow etc. Not only are these tasks which involve a high energy cost, they are also tasks which need the use of the body mass to exert part of the force required to accomplish the work. [A comprehensive but somewhat technical analysis of all the factors entailed in such work has been written by Grieve & Pheasant (1982).] A low body mass makes work of this nature much more physically demanding. It seems to me that some aspects of the Guatemalan findings can be better understood by being explained on this basis.

Therefore, the relevance of low BMI to physical activity, especially in the work situation, may be very much influenced by the type of activity; work which involves the use of the body mass as part of the force required by the work is more stressful to individuals with low BMI. Indeed, on occasions, the work could not be accomplished at all. An extreme but illustrative example from a normal population is that there are weights capable of being lifted by a weight-lifter of 100 kg body mass which are quite incapable of being moved by an equally fit weight lifter of 70 kg mass, so that even for healthy fit individuals of normal BMI, the nature of the work task is crucial.

Activity in leisure time is perhaps less likely to be affected by low BMI since heavy tasks would probably be infrequently chosen.

Conclusion

In conclusion, I should have to decide that low BMIs -but not very low BMIs - did not imply that work capacity were necessarily limited nor that physical activity were necessarily restricted. The possibility is that such negative effects would not appear until BMIs reached 17 or less. On the other hand, many types of heavy physical activity are such that individuals of low BMI are clearly at a disadvantage and these types of activity could well be important in some types of agricultural and constructional work.

References

Allied Dunbar (1992): UK national fitness survey. London: Health Education Authority.

Desai ID (1989): Nutritional status and physical work performance of agricultural migrants in southern Brazil. Proc. XIV Int. Cong. Nutr, Seoul, pp. 297-301.

Desai ID, Waddell C, Dutra S. De Oliveira SD, Duarte E, Robazzi ML, Romero LSC, Desai MI, Vichi FL, Bradfield RB & De Oliveira JED (1984): Marginal malnutrition and reduced physical work capacity of migrant adolescent boys in southern Brazil. Am. J. Clin. Nutr 40, 135-145.

Durnin JVGA, McKay FC & Webster CI (1984): A new method of assessing fatness and desirable weight for use in the Armed Services. Unrestricted report to Army Department, Ministry of Defence, London.

FAO/WHO/UNO (1985): Energy and protein requirements. Technical Report Series no. 724. Geneva: WHO.

Grieve D & Pheasant S (1982): Biomechanics. In The Body at Work, ed. WT Singleton, pp. 77-200. Cambridge: Cambridge University Press.

James WPT & Schofield EC (1990): Human energy
requirements. A manual for planners and nutritionists. Oxford: FAO and Oxford Medical Publications.

Knight I (1984): The heights and weights of adults in Great Britain. OPCS Social Survey Division. London: HMSO.

Satanarayana K, Naidu AN, Chatterjee B & Rao SM (1977): Body size and work output. Am. J. Clin. Nutr. 30, 322-325.

Satyanarayana K, Venkataramana Y & Rao SM (1989): Nutrition and work performance: studies carried out in India. Proc. XIV Int. Cong. Nutr., Seoul, pp. 302-305.

Spurr GB (1987): The effects of chronic energy deficiency on stature, work capacity and productivity. In Effects of chronic energy deficiency on stature, work capacity and productivity, eds B Schürch & NS Schrimshaw, pp. 95-134. Lausanne: IDECG.

Torun B. Flores R. Viteri F. Immink M & Diaz E (1989): Energy supplementation and work performance: Summary of Incap studies. Proc. XIV Int. Cong. Nutr., Seoul, pp. 306-309.

Discussion

Waterlow: Is it possible to analyse what component of an agricultural task is lifting or moving a load, e.g. 70% time engaged in brick carrying or 30% in ploughing?

Durnin: Yes. The method is explained in Grieve & Pheasant (1982).

Ferro-Luzzi: I accept your reasoning about the capacity for work relating to muscle mass, but I have difficulty understanding your static model showing how muscle mass is different in BMIs of 27 rasher then 21. It is not the same as someone metabolically losing fat. The lower the fat mass the more reliance on non-fat stores to mobilize energy, i.e. the Forbes point. Do you still believe that losing 3 kg between BMI 27 and 21 involves a 10% loss of lean body mass, and that this explains their loss of physical work capacity?

Durnin: There are two different situations; one is an individual losing mass from 65 to 55 kg where the criteria that Norgan mentioned need to be taken into account. The other is where you have two populations, one of 65 kg and the other of 55 kg. I think it is still possible to use this model in the second situation.

Norgan: Can you distinguish between aerobic work capacity and endurance work? For example, Spurr showed that aerobic capacity was maintained in mildly malnourished Columbians, but that endurance capacity was depleted. O'Dea's work on Aborigines of low BMI showed that they don't perform well on a bicycle ergometer, but have the endurance to walk for days.

Durnin: What I have been talking about relates mostly to aerobic work. The anaerobic problem is much more difficult to analyse. Spurr has very few subjects at a low BMI. The bicycle ergometer measurement of work capacity does involve the load you exert with the mass of the body. We found big differences in aerobic capacity measured by ergometer and on a treadmill.

Naidu: If you measured a series of different activities over 8 h, what does that show about endurance capacity?

Durnin: If you are only using 50% of your work capacity, then only very heavy work is compromised.

Shetty: In my analysis of individuals of similar BMI but different weights we showed differences in BMR and the reason for this is the difference in the composition of the FFM. In well-nourished individuals the ratio of non-muscle to muscle mass was 0.7 but in those with chronic energy deficiency (CED) the ratio of non-muscle to muscle was >1, i.e. muscle depleted. Spurr's study had very few individuals with a BMI <18.5 BMI. Only one small group of 17.5 showed any endurance effect. There is no evidence in the studies of Spurr or of Desai on adolescents, that endurance time is altered with low BMI. While you may think that perhaps we should go down to 17.5 for a cut-off point, Dias & Imminck (EJCN earlier 1992) showed that even 18.5 may be too low!

Ferro-Luzzi: Do we know the difference between someone of low BMI who is healthy and has both aerobic and endurance capacity such as a marathon runner, and someone of low BMI who has CED?

Durnin: There are only very scattered data on this issue in the literature.

Norgan: Imminck's data showing a possible cut-off point of BMI in Central America may have something to do with the difference in shape, i.e. the relative lengths of legs and trunk. Europeans and Indians have a sitting:height stature ratio of 0.52, while non-Europeans in Central and South America have a ratio of 0.53 or 0.54. Regression analysis shows a one unit change in BMI for every 0.01 change in sitting: stature ratio.

Scrimshaw: We generally associate short legs with malnutrition in early childhood. Secular trends support this. The Japanese were thought to have a genetically determined low limb to trunk ratio, but on migration to California the subsequent generations have the same ratios as Caucasians.

James: The fact that superbly trained athletes are ill so often raises doubts about their general 'fitness' and the desirability of a low BMI even in them. Do we know what percentage of the population we have who are fit and below 18.5 BMI?

Durnin: In our Army study it was 3-4%. The OPCS was much the same.

Naidu: The NCHS American survey data of adults between 20 and 60 had 8% of men and 16% of women with BMI below 18.5 across all racial and socio-economic groups, i.e. representative population sample.

James: If there is severe impairment in work capacity or endurance below a BMI of 17, perhaps with more subtle or refined analyses we might identify some impairment at a BMI of 18 or 19.

Sizaret: The frequency of obesity as measured by BMI >25 is over 40% in highland regions of Latin America such as Lima or in Chile. If these populations with short legs and longer trunks do tend to have higher BMIs then can we really compare the BMI populations in India, Africa and Latin America without taking into account these shape differences?

Naidu: I have analysed my own data with large numbers in four different regions of India and found that BMI is independent of leg length and sitting height in Indian populations.

James: In long-standing malnutrition, does this affect the leg length more than the trunk height?

Naidu: No, legs and trunk length are equally affected.

Norgan: During childhood different parts of the body develop at different rates with extremities developing earlier than the trunk and the final sizes depend on the timing of the malnutrition, so it is complicated, but generally you end up with a short adult of normal proportions.

Contents - Previous - Next