Contents - Previous - Next

Energy-sparing mechanisms: reductions in body mass, BMR and activity: their relative importance and priority in undernourished infants and children

1. Introduction
2. Physical activity
3. Basal metabolic rate and body composition
4. Sequence of events during recovery
6. Future research
Discussion (summarized by A.M. Prentice)


* Division of Clinical Sciences. London School of Hygiene and Tropical Medicine, Keppel Street, London WC1E 7HT, U.K.


The few observational studies that have been published on the physical activity of young children suggest that a decrease in spontaneous activity is an early result of an inadequate energy intake. In the present state of knowledge, to get information about effects on metabolic rate and body composition we have to go back to the severely malnourished child and follow the changes that occur during recovery.

In severe malnutrition the normal pattern of organs and tissues is distorted; muscle is greatly depleted and brain well preserved; visceral organs occupy an intermediate position. Since the better-preserved organs have a high metabolic rate (MR), there will be a tendency in malnutrition for the MR per kg body weight to be greater than normal, as has indeed often been observed. If the MR per kg is low, this must mean that the specific metabolic rates of active tissues are depressed. It is suggested that the best way of expressing the MR in malnutrition is to relate it to total body potassium (TBK). Unless there is specific K depletion, there should be a close relation between TBK and the amount of intracellular protein in the body. Measurements made on this basis have confirmed that in malnutrition there is a reduction in MR per unit K, which is more severe in children with marasmus than in those with kwashiorkor.

Studies in Jamaica have shown that, as children recover from malnutrition, MR and whole-body protein turnover rise within 10-14 days to levels above normal. These changes are followed by restoration of muscle and fat and renewed skeletal growth.

We do not know whether this sequence of events occurs in the reverse order in the child who is becoming malnourished.

1. Introduction

To determine priorities in the responses of children to undernutrition would require serial observations over a period of time, without any intervention. To the best of my knowledge such information does not exist. To repeat the classical experiment of Keys on semi-starvation in adult volunteers, would certainly not be ethical on children. The effects of seasonal changes in food supply and in demand for physical work (e.g., FERRO-LUZZI, PASTORE and SETTE, 1988) might be regarded as Nature's experiment, but such studies seem to have been confined almost entirely to adults. As an exception, in a recent report on an Andean agricultural community (LEONARD, 1989), the seasonal decrease in food intake in young children was only 10-15%, compared with 30% in adults, and triceps skinfold thickness was much closer to the North American norm in the children than in the adults. These children, therefore, seem to have been protected against a reduction in food intake by the responses of their families. It would be interesting to know whether this protective behaviour towards children occurs in other communities subject to seasonal stress.

2. Physical activity

Three studies will be discussed only briefly, because they are considered in more detail by other contributors to this workshop. The first is the pioneer study by RUTISHAUSER and WHITEHEAD (1972) in Uganda, in which activity levels were measured by the diary method in 20 African and 5 European children. The African children spent less time than the Europeans on activities requiring high rates of energy expenditure. Total energy expenditure was estimated from the values obtained by PASSMORE and DURNIN (1955) for the cost of each activity in older children, and amounted on average to 78 kcal/kg/d in the Africans and 98 in the Europeans. Too much significance should not be attached to these figures, since the number of European children was small, and the energy costs of activities extrapolated from older children are probably too low (TORUN, CHEW and MENDOZA, 1983). The important point is that there was a different and lower pattern of physical activity in the African children. At the same time their growth was monitored; the average weight gain per year was 95% of the expected gain, and height gain was 106%. The conclusion seemed to be clear, that in these children growth was protected at the expense of a decrease in activity.

Another study, by TORUN (FAO, 1980), showed that when the energy intake of preschool children was reduced from 92 to 80 kcal/kg/d, there was no effect on weight gain but a fall of about 15% in energy expenditure. With a further reduction of energy intake to 71 kcal/kg/d, energy expenditure decreased even more, and at this point the children ceased to gain weight.

Thirdly, there is the famous group of studies by CHAVEZ and MARTINEZ (1984) comparing various aspects of behaviour in children with and without a food supplement. There were clear differences between the two groups of children in the level of spontaneous physical activity. Thus, all three studies agree in suggesting that a lower energy intake is accompanied by a lower level of physical activity, and there is a hint that the behavioural adjustment may precede and perhaps even prevent a falling off in growth. The evidence is discussed in more detail by Torun in this volume.

3. Basal metabolic rate and body composition

For a deeper analysis of changes in metabolic rate and in body composition, the only way forward seems to be to start at the other end, with the child who is already severely malnourished, and to examine the changes that occur during recovery. MY hypothesis is that the priorities for recovery will reflect in reverse order those that the child tries to maintain during the development of malnutrition.

Since this workshop is particularly concerned with energy expenditure, the first question is: what do we know about the energy expenditure of a severely malnourished child, for example a child with marasmus weighing 50% of expected weight-forage? When such a child first comes under our care it is virtually not eating and it is not growing, so that for the time being the energy cost of growth and the thermic effect of food can be ignored. From clinical observation such a child is almost totally inactive physically, so that its BMR will, without significant error, represent its total energy expenditure. TALBOT, in a classical paper as long ago as 1921, showed that per kg body weight the malnourished infant was likely to have a higher than normal metabolic rate, an observation that has been repeatedly confirmed, both in infants and in adults. The explanation lies, of course, in the abnormal pattern of the different tissues of the body, to which I shall return later.

To avoid the unsuitability of body weight as the basis of reference, some workers have used height, or the 3/4 power of body weight, the so-called metabolic weight. Table 1 shows that both these methods of expression indicate substantial reductions in BMR in the malnourished child. However, in my view neither solution is satisfactory. The reduction in relation to height or length simply reflects a low weight-for-height. As for the 3/4 power of the body weight, the physiological meaning of this long-established allometric relationship is still completely unclear (see BLAXTER, 1989). It is one thing to use it for comparisons among animal species, another to use it for comparisons between individuals.

Table 1. Basal metabolic rate in relation to height or weight0.75






MONCKEBERG et al. (1964)




BROOKE and COCKS (1974)





ABLETT and McCANCE (1971)




BROOKE and COCKS (1974)




The next step is to look more closely at the differences between a marasmic and a normal child in body composition. I use the term 'make-up' to refer to quantitative differences in the pattern of organs and tissues, as opposed to differences in the chemical composition of the body. Figure 1 shows that in malnutrition the brain forms a larger proportion of the body weight, in severe cases by a factor of almost 2. Muscle mass, as judged by creatinine excretion is, by contrast, greatly reduced (Table 2). These observations were made in Jamaica in the 1950s. They were later confirmed in a study in which muscle mass was measured directly, using 15N-creatine (REEDS et al., 1978). In the course of recovery the muscle mass doubled, while body weight increased only by 40%. All this confirms in more detail the earlier findings of KERPEL-FRONIUS and FRANK (1949) in children dying of malnutrition in Budapest after the last war. Brain and muscle are particularly important because of their size and their contrasting metabolic rates. Other smaller but functionally important and metabolically active organs, such as liver and kidney, appear to be relatively protected (KERPEL-FRONIUS and FRANK, 1949; GARROW, FLETCHER and HALLIDAY, 1965), though not to the same extent as the brain.

Table 2. Creatinine output of malnourished babies at intervals during recovery


Days after admission to hospital

Urinary creatinine, mg/24h

mg/cm length

mg/kg weight

Malnourished infants













Normal infants

1 year



2 years



Output per cm length is the better indicator of the degree of initial deficit, because during recovery length varies less than weight.

Modified from WATERLOW et al. (1961).

Figure 1. Relationship between weight deficit and brain weight as a percentage of body weight: r = -0.42, p < 0.005 From ALLEYNE et al. (1969), by permission.

It would be satisfactory if one could construct a balance sheet comparing, in a normal and a malnourished child, the contributions of all the different tissues to the total metabolic rate. HOLLIDAY (1978) has attempted this for normal children. His data indicate that in a child of 1.5 years weighing 11 kg, four organs - brain, liver, heart and kidney - would account for 14% of body weight and 86% of the metabolic rate. This must surely be an overestimate, since the four organs did not include the metabolically active gastrointestinal tract. Lymphoid tissue and bone marrow represent a mass of tissue that is difficult to quantify but may well make a significant contribution to total metabolism (WATERLOW and STEPHEN, 1968). The thymus, as one example of these tissues, is severely reduced in size in malnutrition (GOLDEN, JACKSON and GOLDEN, 1977).

Even though it is not possible to construct a complete balance sheet, some general conclusions can be drawn. If in the malnourished child the specific metabolic rates of brain and visceral tissues are normal or near-normal, then the whole-body metabolic rate per kg must be increased, as TALBOT (1921) observed. If, on the contrary, the whole-body metabolic rate per kg is normal or low, then it must follow that the specific metabolic rates of physiologically important tissues are depressed. A low BMR per kg was indeed found by MONTGOMERY (1962) in some of his patients, and has also been recorded by PARRA et al. (1973).

It is illuminating next to look at the problem from the point of view of the chemical rather than the anatomical composition of the body. Table 3 shows some elements of the body composition of two malnourished infants who died, compared with an estimate from the literature of the composition of a normal baby of the same weight, but obviously of younger age. The first point that I wish to emphasize is the reduction in the absolute amount of non-collagen protein in the wasted child. At this level of malnutrition, the cytoplasmic protein shrinks inside its scaffolding of collagen, which remains more or less intact. PICOU, HALLIDAY and GARROW (1966) observed, by analysis of cadavers, that the proportion of total N represented by collagen increased from about 25% in a normal infant to more than 40% in the malnourished. Therefore, even if the facilities were available, measurement of total body N by neutron absorption would not be particularly helpful.

Table 3. Comparison of the body composition of a normal, a stunted and a wasted child




Age (mo)




Weight (kg)




Weight/Height (%)




Height/Age (%)




% of body weight:









Total body content:

Non-collagen protein (kg)




Total K (mmol)




K/non-collagen protein (mmol/kg)




Data from PICOU et al. (1966).

The second point is that there is a reasonably close relationship between total body potassium and non-collagen protein. It seems to me, therefore, that total body K, as an indicator of cell mass should be the best reference base for expressing measurements of metabolic rate, since it is not affected by changes in extracellular fluid volume or by the relative increase in extracellular protein. The disadvantage is that it is affected by specific K depletion.

Table 4 shows the metabolic rate per unit total body K in malnourished children initially and after recovery.

Table 4. Metabolic rate (MR) in relation to total body potassium in malnourished and recovered children







kcal/mmol K.d.





Marasmic initial





p < 0.001




Oedematous initial





p < 0.05




* initial

To eliminate as far as possible the effects of specific K depletion, the initial measurements of total body K were made after one week on a maintenance diet (WATERLOW, GOLDEN and PATRICK, 1978). This allowed, to some degree, correction of water and electrolyte disturbances. In the marasmic group, the reduction in MR/mmol K was highly significant; in the oedematous children it was less so, perhaps because there was some residual K deficiency. In their original description of this study. BROOKE and COCKS (1974) observed that, in control and recovered children, the MR/mmol K was negatively correlated with body weight; this effect was eliminated when the values for TBK were raised to the 3/4 power. That transformation has not been made in Table 4.

Let us now return to the question of organ pattern. MONTGOMERY (1962) calculated that in malnourished children the MR/kg of lean tissue other than brain was reduced by 30%. BROOKE and COCKS (1974) made an estimate of MR that included brain but excluded muscle, and concluded that when the whole-body MR/mmol K was reduced by 13%, the non-muscle MR/mmol non-muscle K was reduced by 34%. Although these calculations involve a number of assumptions, their general tenor is clear. Such reductions in the metabolic rate of the vital organs must surely have very serious implications. The child is unable to maintain its body temperature (BROOKE, 1972) and it has a low rate of whole-body protein turnover (GOLDEN et al., 1977).

Two studies in vitro confirm the reduction in MR at the tissue level, at least in muscle. NICHOLS et al., (1968) showed a significant correlation between whole-body MR and oxygen uptake of muscle homogenates in the Cartesian diver. PARRA et al. (1973) showed reductions of energy sources in muscle at a time when metabolic rates were low (Table 5).

Table 5. Changes in basal metabolic rate and in energy substrates in muscle in malnourished and recovering children


10 days

20 days

BMR, kcal/kg/d




Muscle content mmol/kg fat-free massATP



1.92 *

Creatine phosphate



2.46 *




13.1 *

* Differences significant at p < 0.025.
Data from PARRA et al. (1973).

However, if a severely malnourished child does not die, the MR rises fairly quickly. Figure 2 illustrates some of the findings of MONTGOMERY (1962). In these children, the MR/kg body weight, initially extremely depressed, almost doubled in 3 weeks, to reach a level that was on average 50% greater than the normal BMR.

Figure 2. Changes in basal oxygen consumption during recovery in four typical malnourished children, showing the 'plateau' level. Changes in body weight are shown in the lower panel. From MONTGOMERY (1962), by permission.

4. Sequence of events during recovery

At the risk of some speculation, I will try to summarize what I believe to be the sequence of events during recovery from severe malnutrition. The first priority is to restore the rate of oxygen uptake in the physiologically important tissues, because without a power supply nothing can happen. This makes possible an increase in the rate of protein synthesis which, like the metabolic rate, may rise to 50% above the normal level (GOLDEN et al., 1977). Increased protein synthesis, in turn, allows a net gain of tissue protein. It seems likely that this occurs first in the tissues and organs with a high rate of protein turnover. It has been shown experimentally that in response to a meal there is an immediate increase in the rate of protein synthesis in liver, whereas the rate in muscle increases later (GARLICK et al., 1975). The data on creatinine output shown in Table 2 suggest that the restoration of muscle mass is indeed rather slow.

Serial measurements of skinfold thickness (Figure 3) indicate that fat deposition becomes important only at a relatively late stage of recovery. Skeletal growth also lags behind growth in body weight (WALKER and GOLDEN, 1988). It seems as if, through some regulatory mechanism whose nature is completely unknown, the child gives first priority to regaining normal weight-for-height. Only then is a rapid increase in body size permitted.

Figure 3. Increase in body weight (• - •), thigh skinfold thickness (•-•) and calculated thigh muscle thickness (+ - +) in a malnourished infant during recovery. From STANDARD et al. (1959), by permission.

I know of no definitive information about the restoration of physical activity, but I have the clinical impression that in the early stages of recovery the child is content to lie in its cot eating voraciously, sweating as it eats because of the thermic effect of the food, and between meals doing nothing but sleep. The urge to be physically active comes later.

In Table 6, this sequence of events is set out in the reverse order, reflecting what I believe to be the child's priorities for maintaining function and life in the face of an inadequate energy supply. This sequence, if it is correct, is different from that described by KEYS et al. (1950) in adults, who showed a drop in BMR of 16% in the first 2 weeks on a low energy intake.

Table 6. Possible sequence of responses to an inadequate energy intake in children


Reduction of physical activity


Cessation of growth in both weight and length/height


Loss of fat


Loss of muscle


Moderate decrease in mass of visceral organs


Decrease in the metabolic rate of visceral organs

Contents - Previous - Next