In the context of nutrition, a working definition of adaptation might be: “a process by which a new or different steady state is reached in response to a change or difference in the intake of food and nutrients”. The word “new” includes the individual who is responding to a change, e.g., when a subject in a balance study moves from a high- to a low-protein intake. The word “different” is appropriate when comparisons are made between individuals or groups who are habitually exposed to different environmental or nutritional conditions. It might be useful to distinguish these as short-term and long-term adaptations.
Three general points are important in relation to both types of adaptation.
(a) The concept of a steady state is relative. No one is ever in an absolutely steady state, either in body weight or in nitrogen balance. During the day, when food is being eaten, nitrogen balance is positive; during the night, in the absence of food, it becomes negative (1). Over 24 hours these fluctuations even out. The situation is precisely analogous to the oxygen debt incurred during strenuous exercise, which is made up during rest. The time-scale over which a state may be considered steady will vary for different functions. It is also clear that in biological systems no so-called steady state is completely stable. This is illustrated by the slow changes in body composition and function that occur as adults age.
(b) Adaptations may be of fundamentally different kindsmetabolic, biological/genetic, and social/behavioural. The response to a change in protein intake is one of the best worked out examples of a metabolic adaptation, but even in this case we do not know the limits of human adaptability. Metabolic adaptations to different levels of energy and protein intake will be considered in the appropriate sections (4 and 5).
Reduction in physical activity as a consequence of a reduced energy intake could be regarded as a behavioural adaptation, with both good and bad effects. For example, the energy intakes of children may be adequate to support satisfactory growth rates, but only at the expense of a reduction in total energy expenditure, notably through less physical activity (2, 3). The result is to impair the child's capacity for exploration and play, and hence his mental, functional, and social development. This question is considered in more detail in section 4.
It has often been suggested that a decrease in body size might be an advantageous adaptation to shortage in food supply. The determinants of body size are complex and depend upon both environmental and genetic factors. In developing countries large numbers of children are small for their age compared with the NCHS standards (see section 3.2.1). This reduction in linear growth (“stunting”) is the result of environmental factors because it is reversible under favourable conditions. Whether it represents a handicap is hotly debated (see sections 3.2 and 3.4). It has also been suggested that selection might occur in favour of those who are genetically smaller and hence have lower needs. A full analysis of the significance of differences in body size would have to take account also of the secular changes that have occurred in a number of countries (4, 5) and of the established relationship between the heights of parents and their children (6).
(c) It follows from these considerations that adaptation implies a range of steady states and it is impossible to define a single point within the range that represents the “normal”. Different adapted states will carry different advantages and penalties. A decision about which is optimal or preferable can only be made in the light of a particular set of values. If the criteria are life expectancy and freedom from disease in the early years of life, then perhaps the nutritional state in industrialized societies might be preferred to that of developing countries, but there may be other criteria of optimal functional capacity.
The same point was made in the report of the 1971 Committee (7), in an apt quotation from Atwater & Benedict (8): “One essential question is, what level is most advantageous? The answer to this must be sought not simply in metabolism experiments and dietary studies, but also in broader observations regarding bodily and mental efficiency and general health, strength and welfare”.
The concept of a range of adapted states, each with advantages and disadvantages, produces a dilemma: it implies respect for different biological and cultural situations, but it may also encourage the acceptance of double standards and the endorsement of the status quo. To quote again from the report of the 1971 Committee: “when supplies are insufficient and purchasing power is low, consumption is likely to be less than requirements. In such circumstances ‘what is’ will not be ‘what should be’ ” (7, p. 15).
It follows that requirements cannot be specified on physiological grounds alone, such as the need to maintain balance. Consequently, in this report value judgements are made about the state that it is considered desirable to achieve. It is not expected that all those who use this report will make the same judgements. Our aim, therefore, has been to set out clearly the principles and the measurements on which the estimates are based, and to indicate as far as possible the areas of uncertainty, so that the estimates can be applied in a flexible way in different situations.
Body size is the major determinant of the absolute requirements for energy and protein. Variations in size are probably more significant quantitatively than the metabolic adaptations discussed in later sections. It is therefore necessary at the outset to define acceptable ranges of body size.
3.2.1 Children
Although the energy and protein requirements for the process of growth are relatively small compared with those for maintenance, except in the young infant, satisfactory growth is nevertheless a sensitive criterion of whether needs are being met. Therefore, the definition of satisfactory growth is the first step in estimating the requirements of infants and children. An example of the dilemma mentioned above is whether reference standards for the growth of children in industrialized countries should be accepted as universally relevant or whether local standards should be used (9). Children in many developing countries are smaller at birth than those in industrialized countries and grow at a slower rate during infancy and early childhood. The evidence suggests that in young children these differences are due primarily to environmental factors, including inadequate nutrition, and that genetic and ethnic factors are of lesser importance, so that young children of different ethnic groups should be considered as having the same or similar growth potential (10–12).
Even in a healthy privileged population there is a wide range of variation in the size of children. In such a population there is no indication that differences in size per se are related to health, wellbeing, or physiological function. However, in communities where children's growth is limited by environmental factors, there is evidence of an association between functional impairment and deficit in linear growth (13). In such situations, it is extremely difficult to separate the effects of undernutrition from those of other aspects of social deprivation. Therefore, it remains a matter for further research how far small size in children represents a handicap or an adaptation, whether minor limitations of genetic growth potential are harmful, and whether maximum growth is necessarily an indicator of optimal nutrition (14).
Nevertheless, the Consultation feels it desirable that the growth potential of children should be fully expressed, and estimates of energy and protein requirements should allow for this. Estimates of the requirements of children up to 10 years are therefore based on the reference growth standards published for international use by WHO (15), which are derived from the United States National Center for Health Statistics (NCHS) (16). The use of this particular reference population is recommended on the basis of a number of criteria (17). These standards are summarized in Annex 2(A).
Surprisingly, in contrast to adults (see below), there appear to be no recommendations for children concerning the ranges within which weight or height at any given age may be regarded as satisfactory. This is partly because information about the risk attached to given degrees of deficit is only just becoming available (18, 19), and partly because children start with different birth-weights, and therefore their attained weights will continue for some time to be above or below the median. In epidemiological studies of childhood undernutrition it is conventional to accept —2 SD from the median as the cut-off point between “normal” and “malnourished”, corresponding approximately to the 3rd centile or to 80% of the median for weight (15) and 90% for height. Similarly, +2 SD in weight for height may be taken as a cut-off point for obesity.
In relation to growth, two further points must be taken into account. Previous committees have based their estimates of the daily requirements for growth on increments in body weight of the reference population over intervals of 3 months for infants below 1 year and intervals of 1 year for older children. This procedure assumes that growth occurs at the same rate from day to day. It must be recognized that this is not so, and that even in normal children free from infection growth occurs in spurts. As a result, the variability in weight gain over short periods such as one month is extremely high. For example, in two longitudinal studies from birth to 3–6 months, the coefficient of variation of weight gain over 4-week periods was approximately 37% (20,21). The reasons for this variability in the rate of weight gain are not clear. The data of Fomon (personal communication,1980) show that it is greater than the variability in energy intake. One factor, therefore, may be day-to-day fluctuations in physical activity. From the point of view of protein requirements, the phenomenon could be regarded as analogous, on a longer time-scale, to the metabolic differences that have been observed between the day, when food is consumed, and the night, when it is not. Studies on adults have shown that during the night there is a negative nitrogen balance that is cancelled out by a positive balance during the day (1). In children it has been suggested that growth occurs in spurts in relation to food intake(22).
The variability in growth would not affect requirements averaged over a period of time if it were a consequence simply of fluctuations in food intake, so that intake and rate of growth each day were exactly matched. It seems unlikely, however, that this is the full explanation. Since this exact matching does not happen, the effect of the variability in growth may be to increase the growth component of the requirement, as a reduction in growth over one period will have to be compensated by an increased growth rate later on. As discussed in section 6.3.2, it is extremely difficult to estimate the quantitative effect of this variability in growth rate.
A second question that should be raised, although it cannot at present be answered, results from the fact that all allowances for growth are based on increments of body weight. It is conceivable that growth might be limited by special requirements of particular tissues; for example, the relative amounts of energy and protein needed to achieve a given gain in body weight might not be the same as the amounts needed to secure an appropriate increase in height. This is a subject on which further research is needed.
3.2.2 Adolescents
The desirable heights and weights of children over 10 years of age present special problems, because there is considerable variation between individuals and groups in the timing of the adolescent growth spurt, which starts at different calendar ages in boys and girls (6).
Furthermore, if children have been growing slowly from infancy, as happens in many developing countries, by 10 years of age the gap between their actual weight and their expected weight, based on that of adolescents in industrialized countries, will be very large. It is not known whether extra food at this stage can increase the extent and duration of the pubertal growth spurt.
For these reasons it is considered more realistic, after the age of 10 years, to relate requirements to the appropriate weight for height rather than weight for age. In order to maintain uniformity with the reference values for the earlier years of childhood, the standards published by WHO for height up to 18 years have again been chosen, but they do not include values of weight for height beyond 10 years. To provide these values the Consultation used data from a large sample of children measured in the United States of America earlier in this century (23). Annex 2(B) gives the median weight for height of boys and girls at each age, not only at the median height of the standards published by WHO, but also at different heights. By using actual height and median weight for height, it is possible to allow for the fact that puberty begins at different ages in different groups of adolescents. Again, as with younger children, there are no clear recommendations about the “acceptable” range of weight for height.
3.2.3 Adults
Adult groups in various parts of the world show substantial differences in height (24), and height variations are also common within countries and races.
In general, there is no reason to suppose that adults of either short or tall stature have a health risk attributable to their stature, except perhaps in relation to pregnancy and childbirth (25), and therefore no attempt is made in this report to define the height of a healthy reference population of adults. However, body weight, when expressed in relation to height, does influence health, and a range of desirable or acceptable weights for height has been proposed (26). These values, derived from actuarial analyses (27) and prospective epidemiological studies in Western communities, are set out in Annex 2(C). The upper limit of the acceptable range, at which there is an increased risk to health, has been reasonably well defined.
Unfortunately, the same cannot be said for the lower limit. It has been claimed (28) that the range has been set too low, and that it may, in fact, be beneficial to have a weight for height in excess of the desirable range shown in Annex 2(C). However, the evidence in support of this claim may be criticized: for example, the apparent increase in risk associated with being moderately underweight may, in the populations studied, be due to an association with smoking or chronic disease (29).
Long-term prospective studies on large numbers of adults in developing countries are not available. It must be recognized that in communities subject to infections, periodic food deprivation, and high energy demands for physical activity, a higher body weight than the average suggested in Annex 2(C) could be advantageous. At present the average weight of adults in many countries is below the mid-point of that range, and sometimes even below the lower limit (24). There is no direct evidence that this in itself is either beneficial or harmful. More information relating body weight and composition to health risk in these communities would be valuable. In the absence of such information, the range of weight for height shown in Annex 2(C) was accepted by the Consultation as appropriate for all populations.
A useful simplification is to express the weight for height as the body mass index (BMI) (Wt/Ht2, or Quetelet's index), since this function gives a measure of weight for height that is largely independent of actual height (30). The justification for using this index is twofold: the point made above, that stature is not considered to be related to health risks; and the point discussed in section 6 that, except in the very young and the elderly, height appears to have little effect on energy or protein requirements independently of its relationship to weight.
Estimates of requirements based on body weight are an approximation, since they do not take account of differences in body composition, which will determine true requirements. In recent decades the emergence of methods for estimating some body components in living subjects has resulted in observations on several thousand people ranging from newborn infants to the very old. At birth the neonate averages 14% body fat, which rises to about 23% at 12 months and declines to 18% at 6 years of age (31). During this period girls have slightly more body fat than boys (32), and this difference becomes more pronounced after 6 years (33). During adolescence the difference in the body fat content between the sexes becomes strikingly accentuated (34–36) and persists throughout adult life as shown by differences in the thickness of the subcutaneous fat layers (37). There is evidence that body composition is also influenced by genetic factors, since obesity tends to be familial (38) and monozygous twins are more concordant in fatness than dizygous twins (39).
Adolescence is also characterized by a major sex difference in the rate of acquisition of lean weight. Boys show a rapid and sustained spurt in lean weight, whereas there is a modest acquisition of body fat in the early phase of puberty, followed by a decline. In contrast, girls have a smaller spurt in lean weight, but they acquire more body fat. In adolescent boys the time of the spurt in lean weight has been found to coincide with the most rapid growth in height (36), and to continue until 20–25 years of age (34,35), whereas in girls the pubertal increase in lean weight ceases by about 18 years, in keeping with the marked decrease in the rate of gain in stature after menarche (6). During the second decade of life boys thus double their lean weight, while the increase in girls is only 1.5-fold. The end result at maturity is a fat-free weight of about 60 kg in males averaging 70 kg total body weight, and 42 kg in females averaging 63 kg. The variability in lean weight is less than that of total body weight (35,40).
The adult years are characterized by a decline in lean body mass in both sexes, which becomes obvious by the age of 40; by 85 years the lean body mass has reached a value about three-quarters of that characteristic of the young adult (35, 41-43). Simultaneous measurements of body nitrogen by neutron activation and body potassium as 40K show that, with advancing age, more potassium than nitrogen is lost (44), implying that the potassium-rich muscle mass is especially reduced. The relative loss of potassium is about 10% between the sixth and eighth decades (45). Autopsy data (46) confirm that subjects over 70 years of age have 40% less muscle than young adults, with a smaller reduction in the mass of visceral organs. Preferential loss of muscle with aging is also demonstrated by the decline in creatinine output (47) and the fall in 3-methyl histidine output (48). This age-related loss of lean body mass is commonly accompanied by an increase in body fat. Consequently, in relation to body weight, the lean tissue content of the body declines with age, and this accounts in part for a progressive fall in basal metabolic rate in relation to body size (49,50). In the elderly there is a decrease in the proportion of skeletal mass, as well as muscle (51).
Within the lean body mass there are also differences in the proportion of various tissues at different ages. From birth to maturity the brain increases its mass 5-fold, the liver, heart, and kidneys, which are even more metabolically active, increase 10- to 12-fold, while muscle multiplies its mass by about 40-fold.
Estimates of requirements for energy and protein are based primarily on metabolic and balance studies of limited duration. However, the estimated requirements should be enough to maintain health and sustain optimal bodily function, including physical and mental fitness.
During growth, in addition to the increases in weight and height, there are marked functional changes. In boys, aerobic capacity and heart volume in relation to lean body mass increase up to the age of 14–15 years (52,53). The natural peak in functional capacity coincides with high levels of spontaneous physical activity, and estimates of requirements must allow for this. Conversely, in the fourth to the fifth decade of life aerobic capacity starts to decline, with decreasing physical activity and energy requirements.
Intakes of energy or protein above as well as below those needed for optimal function may be detrimental if they exceed the adaptive capacity of the organism. Excessive energy intakes lead to obesity, with reductions in cardiorespiratory efficiency, physical performance, and endurance.
It has already been suggested (page 23) that stunting in linear growth may represent an adaptation that does not necessarily present any health hazard beyond early life. For example, cardiorespiratory function, physical performance, and muscular strength were found to be significantly better in stunted Tunisian children from a poor socioeconomic group than in children from affluent families, whose growth was closer to that of the standard in developed countries (54). Similarly, Italian children from poor families performed better in physical fitness tests than their counterparts from more prosperous families, in spite of their smaller size and lower habitual energy intakes (55). These findings suggest that habitual physical activity is a more important determinant of fitness than is body size per se. Nor do the effects of small stature necessarily carry penalties in adult life (56) except for tasks requiring a particular body build and strength. Thus high aerobic capacity related to body weight was found in Indian miners with very low weights and heights compared with their counterparts in other countries (57). On the other hand, in another study in India low weight and height were significant handicaps for obtaining employment in agriculture (58). Therefore, while stunting in height is a rather sensitive marker of socioeconomically disadvantaged populations, the consequences need to be carefully evaluated for evidence of any functional handicap.
In healthy people within the ranges of acceptable weight for height or weight for age proposed in section 3.2, the main determinants of requirements for both energy and protein are body weight and age, and in the case of energy, physical activity. Further considerations, discussed in section 9, apply for people outside these desirable ranges.
3.5.1 Relation to body weight
Protein. Within a given age range, the requirement for protein per kg of body weight is considered to be constant. Therefore, the primary expression of protein requirement is in grams of protein per kilogram. This principle applies to all ages, although absolute additions, in units of grams of protein per day, are made for pregnancy and lactation.
Energy. For energy the position is more complicated, because at a given age the main component of the energy requirement, the basal metabolic rate (BMR) varies not only with absolute body weight but also per kg. The total energy requirement per kg, therefore, cannot be taken as constant. Hence the primary expression of energy requirement is the total requirement per person, derived from that person's body weight.
With adults less than 60 years old the effect of age is relatively unimportant, since at a given weight the BMR decreases by only about 1% per decade. With children the change of BMR per kg with age is much greater—about 5% per year between 3 and 10 years. At present, we do not know to what extent this reflects an age-related change per se, or age-related changes in body weight. However, for this report the question is not of practical importance, since the BMR has not been used for estimating energy requirements in children below the age of 10 years.
The energy requirements of adolescents are treated in the same way as those of adults.
If at each age weight is the main determinant of requirements, the question then arises, what weight should be used? Within the acceptable range, many people will have weights that differ by 10% or more from the median. If the actual weight is used, it will tend to maintain the status quo. If the median of the reference range is used, the result will have a normative effect. The existing needs of those who are at the lower end of the weight range for age or height will be overestimated and the needs of those at the upper end will be underestimated. If the requirements so calculated are fulfilled, there will be a tendency for weight to move towards the median. This is what is meant by the term “normative”, as used in this report. It will be for the user to choose the most appropriate body weight for calculating requirements, depending on the circumstances and his aims.
3.5.2 Relation to age
For the tables in this report 6 main age ranges have been defined: 0–3 years; 3–10; 10–18; 18–30; 30–60; 60+; these are further discussed in section 3.6.1. The aim in selecting these ranges is to reflect the physiological characteristics of men and women, including the continual changes in rate of growth, body composition, physical activity, and patterns of food intake. These changes are particularly rapid at three periods—infancy, adolescence, and old age. Subdivisions are therefore necessary for infants and young children. In children up to 10 years of age no distinction is made between the sexes except for that which arises from differences in body weight. With adolescence important sex differences begin to appear in body composition and in the timing of the growth spurt (section 3.2.2). Since, however, the timing is very variable in relation to chronological age, subdivision into defined age groups would be impracticable. It would be desirable to divide further the age range from 60 years onwards, but unfortunately the information on which to base more accurate estimates related to age is too scanty.
In section 2 the term “class” was introduced to take account of the variability that exists between apparently similar individuals. In practice this concept of a homogeneous class is not entirely realistic. For the purposes of tabulation it is necessary to group together individuals, according to age, weight for height, etc. The wider the ranges chosen, the less homogeneous such groups will be. The problem then is to decide on operationally useful ranges. No hard-and-fast guideline can be laid down. A large number of narrow ranges is more precise than a small number of wide ones, but also more complex. The ranges in many of the tables (e.g., section 8) have therefore been chosen as a compromise between convenience and precision.
The variability of requirements within a group defined by any chosen range will clearly be greater than in a completely homogeneous class. This has implications for calculating the requirements of the group as a whole which are of great importance for the application of requirement estimates, as discussed in section 11.
The requirements specified for each range in the tables represent those of the individual or homogeneous class whose weight for age or weight for height is at the mid-point of the range. Such estimates do not necessarily correspond with the average requirement of the group as a whole, since this will depend on the distribution of individuals within the group.
A particular problem, analogous to that discussed for the individual in section 3.5, arises if a group of people has a mean weight which is markedly different from the mid-point of the reference range. For example, the data collected by Eveleth & Tanner (24) show that young Indonesian adults had on average a body mass index (BMI) of 19, which is close to the lower end of the desirable range defined in section 3.2. It would be possible to specify the requirements of this group either on the basis of their actual average weight, or on the basis of the average acceptable weight for height (BMI 22), which would be about 15% greater. As in the case of the individual, the choice will depend upon whether or not the user considers the actual situation to be satisfactory.
There is, moreover, an additional complication. In such a group with a mean BMI of 19, there will be some individuals whose weight for height is below the limits of the acceptable range. In section 9 we discuss adjustments to requirement estimates for people whose weight for age or for height is outside those limits. The concept of an acceptable range implies that some action is necessary to improve the situation of those who are outside it and are therefore in an unacceptable position. With a mixed group, some of whom are “inside” and some “outside”, there are again two choices of action: to concentrate on individual outsiders, or to take measures which will affect the whole group.
3.6.1 Expression of age ranges
Conflicting methods are often used for expressing the ages or age ranges to which values of body weight, BMR, etc., refer. In some tables, e.g., those of the NCHS, the value for body weight opposite the figure for 5 years, for example, means the weight at 5.0 years. Since few children will be measured exactly on their birthday, such values are usually obtained by interpolation. In tables in other reports, the entry for weight at 5 years may in some cases signify the average weight of children between 4.5 and 5.5 years (mean age 5.0 years), in others the average weight of children from 5 to 6 years (mean age 5.5 years).
For reasons of clarity and to comply with ordinary usage, age range (e.g., 5–6 years) are specified in this report, because this is how children are usually classified, for example in school.
When a value for body weight or height is given, it represents the value at the mid-point of the range, obtained by interpolation from the NCHS standards. The range 5–6 starts at 5 years, up to but not including 6 years. To write 5–5.99 implies an unrealistic degree of precision. Another useful solution is to designate the ranges as 5+, 6+, etc. (32).
When values (e.g., for BMR or protein requirement) are changing with age, it is clearly artificial to make dividing lines at particular ages at which abrupt changes are supposed to occur. In reality the values are continuous variables. If a more precise estimate is needed, it can be obtained by interpolation.
These are problems that relate more to applications than to the substance of this report, but it is necessary to outline them here in order to illustrate the different ways in which the figures and tables may be used.