Previous PageTable Of ContentsNext Page

Chapter 1 Defining chronic energy deficiency

Chapter 1 Defining chronic energy deficiency

At the time of its convening, an initial task of the IDECG Working Party was to distinguish between acute energy deficiency (AED) and chronic energy deficiency (CED) in adults.

AED was regarded a priori as a state of negative energy balance, i.e. energy intake is less than the energy expenditure so that, despite changes in metabolic efficiency or physical activity patterns, there is a progressive loss of body weight and body energy stores. As this energy imbalance and weight loss continue, health and body function will be impaired over a period of time eventually leading to death. CED in the adult, on the other hand, is not a state of prolonged continuous loss of body weight and energy.

CED is defined as a "steady state" where an individual is in energy balance, i.e. the energy intake equals the energy expenditure, despite the low body weight and low body energy stores. Thus, by never growing to a normal size or having experienced one or more stages of energy deficiency, the individual has arrived at a reduced body weight with possibly limited physical activity, which have allowed the energy demands of a lower basal metabolic rate (BMR) and reduced amounts of activity to balance the lower intake.

AED may occur episodically among people who have a pre-existing condition of CED, for example, during a seasonal shortage of food in a community which is already deprived. The concept of a "steady state" is a theoretical one and the degree to which it is experienced by an individual depends on the time-scale considered. Normal fluctuations in energy balance occur over a 24-hour period, with eating by day and fasting by night.

Unsteady states also occur over a week with marked changes over the weekend in individuals in industrialized societies.

Festivals and religious occasions also involve phases of energy imbalance. Women in a community also show fluctuations in energy balance according to the phase of the menstrual cycle. In some developing countries, there are long term seasonal cycles, with weight being lost during the hungry season before the harvest and regained when the harvest has been completed. Since seasonal changes in energy balance could be functionally important, perhaps a steady state should be considered as involving a body weight and energy store which is maintained over a 2-3 month period.

There are various levels of severity for CED, but in all cases energy input equals energy output. For a person with an adequate amount of food, body weight and body energy stores are within the acceptable range of normality, health is not impaired and physiological function is not compromised. Energy is expended on bodily functions which include the maintenance of an optimum body temperature as well as the cost of storing the different forms of energy such as carbohydrate, fat, and protein after a meal. The energy used is sufficient for economically necessary and productive work and domestic chores as well as social and leisure activities. In such a situation, measuring either the average energy intake or expenditure over a suitable period of time should provide an estimate of the amount of energy required to maintain the stable body weight, level of physical activity and appropriate life-style.

The steady state will change when either energy supply is reduced or demands for energy output increase and this change is not counterbalanced on the other side of the equation. In response to this change, there is a fall in body weight but not essential energy output. Monitoring of individuals has shown that body weight does change as soon as energy imbalance occurs (Ferro-Luzzi, 1990). This means that during the acute energy deficiency phase body weight falls. When the adults do stabilize by matching their output to their energy intake, they will have a lower weight. In order to sustain productive and economically vital activities which may occur at the peak energy output periods such as planting or harvesting, body weight is sacrificed. During the initial period of body weight reduction much of the socially desirable and leisure activities may not be reduced or compromised and if so, only slightly.

Eventually, increased energy imbalance leads to marked changes in body weight as well as the quantity and quality of energy output. Therefore, long-standing energy deficiency is reflected by both changes in body weight and activity patterns. In normal circumstances, these sequences of events are most likely to occur over phases of a seasonal agricultural cycle. If the individual needs to continue working during a phase of seasonal food shortage, their weight will have dropped with little or no evidence of a decrease in physical activity. They may still, however, come into a steady state of chronic energy deficiency because their lower body weight will imply a lower BMR. The cost of moving also declines with the fall in body weight, so that the two savings in energy related to weight will allow a steady state of energy balance to be achieved once more. The costs, however, will be those associated with the lower body weight, i.e. lower energy stores and lower resistance to infection and environmental insults. If severe enough and over a long enough period of time, leisure and socially desirable activities will also begin to decrease.

Energy intake and CED

As nutritional science was developing, attempts were made to assess the adequacy of food supplies by measuring food intake and determining its adequacy in comparison to the assumed energy requirements. Nutritionists have usually sought to assess the adequacy of food supplies directly once they have calculated the food intake values and individual energy requirements. In practice, this involves measuring habitual food intakes, transforming the data into energy values, and relating them to needs estimated for the age, sex, optimum body weight and desirable activity levels of the adults. These issues have been addressed in the literature and various solutions have been proposed and utilized in several surveys (François, 1970).

As an example, the National Nutrition Monitoring Bureau (NNMB) in India1 has expressed a household's energy intake in terms of consumption units (CU) also referred to as adult equivalents which standardize the data according to the demographic characteristics of the family. This value is then compared with the recommended allowance of that "standardized" household member (Rao & Shastry, 1986). This recommended allowance (i.e. adequate intake) was set at 2 400 kcal/CU/day, and it was estimated that 30 percent of households had intakes below 2 000 kcal/CU/day while 10 percent had intakes of less than 1 600 kcal/CU/day. If the reference value of 2 400 kcal/CU/day is correct, this implies that there are very large numbers of households with inadequate intakes. The validity of this recommended allowance/CU/day depends on a value judgement and cannot, of course, take into account either the actual body size or the physical activity level of the individuals within a household, both of which influence the energy needs. Again, this method involves the collection of household food intake data, a time-consuming process which can be inaccurate. The use of household energy intakes and crude averages for the recommended allowances must, therefore, be considered only as a general approximation of the nutritional situation within a community. The costs involved in collecting such information on a large scale are also considerable.

[ 1. The National Institute of Nutrition (NIN), India analysed data from several series of surveys carried out by the National Nutrition Monitoring Bureau. These surveys are carried out each year on a representative sample of the rural areas from ten of the most populated states. A variety of information is collected including anthropometry of all household members, family food consumption and various socio-economic variables. For this BMI study, the NIN prepared two data sets, one the aggregated data from 1974-79 rounds and the other from 1988-90 which represented a total sample size of 48,252 and 21,361, respectively. ]

Further difficulties arise if one wants to assess individuals for CED by monitoring their individual intake rather than assessing CED at a household or national level. First, it is difficult to estimate energy intake with any degree of accuracy over a short period of time. Foremost is the difficulty in measuring and recording the food eaten and at the same time not influencing the intake by the mere measurement process. Other problems exist as well, for example, premenopausal women are likely to show a 5-10 percent fluctuation in their energy intake over the month's menstrual cycle. In addition, it must be recognized that in any group of individuals there will be a substantial range, perhaps 8-12 percent in BMR even when they are of the same age, sex, height and weight. Therefore, one adult with an intake of 2 400 kcal/day may have an intrinsic need for energy to meet a high BMR of perhaps 1 750 kcal/day, whereas another similar adult may need only 1 350 kcal/day. The 400 kcal difference would allow the adult with a low BMR to engage in up to 3 hours of extra physical activity per day. Thus, the difficulties in assessing an individual's energy intake and the concern that the measurements may provide a very inaccurate indication of whether the adult is receiving enough food led to the evaluation of alternative approaches.

Energy expenditure and CED

A Joint FAO/WHO/UNU Expert Consultation on Energy and Protein Requirements held in Rome, 5-17 October 1981, (WHO, 1985) suggested that energy requirements be based on estimates of energy expenditures rather than energy intakes and that the total energy expenditure of an individual be expressed as multiples of the measured or predicted basal metabolic rate (BMR). This ratio of the total energy expended to the BMR has been termed the physical activity level (PAL). Since BMR is based on body weight, the actual value obtained from this PAL ratio will reflect both the body weight and the level of physical activity of the individual. The minimum level of energy expenditure compatible with health, termed maintenance requirement, may be taken as corresponding to a PAL of 1.4 for the sake of simplicity. Thus, this value of 1.4 x BMR may be used as a cut-off point for assessing the prevalence of CED. Furthermore, studies have confirmed that both total energy expenditure and BMR are highly reproducible and stable (Shetty & Soares, 1988).

Given the usefulness of knowing the value of PALs, it would be desirable to have national surveys that indicate whether all individuals were able to sustain the levels of activity appropriate for their economic, social and educational welfare. Indeed, the monitoring of PAL seemed of such importance to the IDECG Working Party that they proposed monitoring both body weight and physical activity levels in defining CED. A classification system was proposed which depended upon demonstrating both a low body weight and a poor level of physical activity with a PAL below 1.4 (James et al., 1988). This would conform with current notions of how people adjust to inadequate energy intakes.

According to this method, BMIs were to be derived from the weights and heights of the individuals or group while the energy turnover or PAL was to be based on total energy expenditures assessed in relation to the BMR predicted from body weight. The IDECG Working Party outlined a sequential or stepwise assessment of CED for the epidemiological diagnosis of different grades of CED. An individual with a BMI > 18.5 was expected to have adequate energy reserves and was, therefore, considered normal. The next step involved an additional assessment of the PAL. Those individuals with a BMI of 17.0-18.5 and a PAL > 1.4 were classified as being normal while a presumptive diagnosis of CED grade I was made for those persons with a PAL < 1.4. Those with a BMI between 16.0-16.9 were categorized as having CED grade I if they had a PAL > 1.4 but grade II if the PAL was c 1.4. The different categories proposed by the IDECG Working Party are displayed in Table 1.1.

Later, detailed measurements of adults in Ethiopia and India suggested that the use of a cutoff point of 1.4 PAL can be very problematic. It is possible for women to sustain economically relevant and useful activities for up to 4.5 hours a day without exceeding a daily energy turnover value of 1.4 (Ferro-Luzzi et al., 1992). This is explained by the very low energy costs of activities performed sitting down which are undertaken during the rest of the day and amount to no more than 1.25 x BMR. This contrasts with a theoretical value of 1.4 for residual time decided upon by the Joint FAO/WHO/UNU Expert Consultation (ibid), suggesting perhaps, the need for a reexamination of the value recommended at that time. There was also a marked variation in individuals in terms of their activities from month to month and most of the population at one time or another had PAL values below 1.4. In women, the duration of work was only marginally reduced for those individuals with low BMI. Even women with a BMI c 16.0 had no apparent reduction in the time spent on these activities. Neither the intensity of work nor the time involved in social and other spontaneous activities was reduced. By contrast, the men did seem to display some evidence of behavioural adaptation and this may reflect their sensitivity to restricted intake of food which is greater than that of women who have a higher body fat content and energy stores for the same, albeit low, BMI. The combination of physical activity (as an indicator of activity turnover) with BMI for estimated degree of CED produced an unusual distribution that did not follow the classical Guassian model; a distribution unlike that of the BMI alone (See Figure 1.1).

TABLE 1.1 Epidemiological diagnosis of CED as proposed by IDECG Working Group





> 18.5

> 1.4 PAL




Grade III

Grade I








< 1.4 PAL

Grade III

Grade II

Grade I


Source: James et al., 1988.

Figure 1.1 - Grades of CED among Indian adults by BMI and BMI+PAL

Source: Ferro-Luzzi et al., 1992

Measures of food intakes of adults are difficult to match to estimates of food needs. It is also apparent that it is unrealistic to propose the global development of activity monitoring as an aid to the assessment of undernourished adults (Ferro-Luzzi et, al., 1992). The original concern that the developing countries may have large numbers of thin but very athletically fit individuals was also discounted. For the reasons noted above, it is proposed that the direct assessment of the nutritional status of a community in relation to the adequacy of food availability be made by concentrating on a simple anthropometric measure of body weight and height. Two issues are dealt with in the rest of this report: whether the body weight can be a useful index of a society's well-being with implications for food supplies and whether any cut-off points for nutritional adequacy are shown to be functionally meaningful.

CED as diagnosed by anthropometry

One of the simplest ways of diagnosing CED in adults would be to resort to the use of an uncomplicated, reliable and easily obtainable anthropometric measure such as height, body weight or some combined index of weight and height. Recent reports have advocated that height may be particularly useful as an index of socioeconomic conditions in developing societies because populations which are poorly fed and subject to repeated infections rarely grow well in either childhood or adolescence and fail to achieve an adult stature which is commensurate with their full genetic potential (Gopalan, 1987).

Measurements of height are relatively easy to make as are measurements of body weight. However, measuring the height alone of adults does not help with monitoring current nutritional conditions although it may reflect the unfavourable circumstances of their growth years. Long term energy deficiency in childhood does cause stunting and a reduction in the final adult stature attained, but the functional consequences of this short stature and the reduced work capacity and economic productivity seen in short adults cannot be alleviated. If it is undesirable for adults to be small and to not have fulfilled their genetic potential, then clearly attention needs to be directed towards alleviating those factors that lead to stunting in childhood. Thus, to concentrate on height of the adult as a primary measure of nutritional status of the community is inappropriate. The present concern is the need for an index for identifying chronically energy deficient adults in a community. On this basis a measure of body weight in relation to the height of adults was considered the more promising option.

If a simple, objective, and reliable definition of CED is based on anthropometry, it is important to evaluate the range of weight-height indices suitable for this purpose. While different types of indices may be used, the body mass index (BMI) may be ideal.

In broad terms, the body is composed of: active tissue mass, (lean body mass or fat free mass) and fat mass which is the body's principal energy store:

Body Weight = fat free mass + fat mass

As mentioned, an individual is said to be in a steady state situation, i.e. in "energy balance" when the input of food energy is equal to the output of energy expended by the individual. The latter is the sum of the energy expended at rest (BMR) plus the additional energy costs of digestion and absorption of nutrients and the costs of physical activity including work. Under these conditions, a change in body weight can be an indicator of a change in the steady state, i.e. energy balance of the individual. Although small, short-term fluctuations in body weight can also occur with changes in water balance, the body weight stability can be used in general as a measure of the steady state of energy balance of the individual.

The body weight of an individual is influenced by the person's height; body weight rises with increases in stature if normal body proportions are maintained. The issue of whether body weight differs between people with varying proportions of lean and fat tissues is dealt with in Chapter 2; for the present, weight is chosen as a simple index of both muscle and fat tissues. It is unreasonable to measure the body weight of an adult alone without taking into account their height. Therefore, a simple measure of weight in relation to height is needed. Once this is decided it is necessary to define some point below which the weight-for-height is not compatible with good health and effective ability to work. Measurements of the body weights and heights of individuals within a community could then be used to find the proportion of adults who are too thin or too obese. Before choosing acceptable or desirable weights, however, some simple indices of weight in relation to height must be established.

Requirements of an ideal nutritional anthropometric index

Two practical considerations determine the usefulness of an ideal nutritional anthropometric index for epidemiological purposes. They are as follows:

From a mathematical point of view, the index chosen should be highly correlated with body weight which is a proxy for body energy stores and relatively independent of height, i.e. the alterations in the index should not relate simply to differences in stature. This implies that the numerical value of the index of standard weight for height be the same for individuals at different heights.

Such an index will not only reflect the weights of the individuals or group vis-à-vis the range found in the population, but will also give some measure of the change in the body's composition, i.e. reflect changes in both the energy stores and in the active tissues of the body.

Types of weight-for-height indices

There are two basic types of weight-for-height indices in common use for community epidemiological measurements in adults.

1. Relative weight is the ratio of an individual's weight to a standard or expected weight for individuals of his or her sex and height, with age being used as an additional standardizing variable. Relative weight thus expresses the weight of a given individual as a percentage of the average weight of persons of the same height, sex and age. The standard or expected weights are frequently derived from a large population of people of the same height, sex and age. The use of relative weight provides a measure which is readily interpreted. For instance, to say that a group of people was 140-150 percent of the average weight for their age, sex and height conveys a meaningful image to a reader. However, the disadvantage of this approach is that the findings from various studies are difficult to compare when different standards have been employed. Differences among standards of various populations or ethnic groups can be large and hence relative weights are of limited use for international comparisons. In addition, acceptability of an international standard for relative weight is difficult to obtain.

2. A power-type index, which does not specifically employ a standard, is derived by the ratio

where p is a value, usually between 1.0 and 3.0.

Since these competing power type indices are purported to provide an index of "body bulk" or "energy reserves" and each has been used by different investigators, it is extremely important that their relative merits are examined critically. As noted above, the most desirable of these power indices should be maximally correlated with weight but minimally correlated with height, i.e. unbiased by stature, in all populations.

Weight/height ratio (W/H) is a ratio that shows consistently high correlations with body weight of the order of 0.95 to 0.99 in several studies. It is, however, correlated with height and in adult populations the correlations of W/H with height vary from 0.01 to 0.51 (Lee et al., 1981). This implies that W/H as an index is not independent of height.

Weight/height2 (W/H2), otherwise known as the body mass index (BMI) or Quételet's index also shows a consistently high correlation with body weight based on several different studies in different population groups. In a comparative series of analyses the correlation of W/H2 with height varied from 0.0 to 0.2. (Khosla & Lowe, 1967; Micozzi et al., 1986).2 Thus, this index is relatively independent of height and is less biased by height than W/H.

[ 2. Adolphe Quételet (1796-1876) wee a mathematician and statistician and Permanent Secretary of the Royal Observatory of Brussels. Ha proposed this indicator in the Physique sociale ou essay sur lo développement des facultés de l'homme. Bruxelles. C. Muquardt. 1869. ]

TABLE 1.2 Correlation between BMI and weight and height measurements from selected studies




Correlation with







0.83 to 0.86 (*)

-0.1 to 0.08




0.88 to 0.92

0.02 to 0.05




0.92 to 0.95 (*)

-0.01 to -0.12



















0.81 to 0.90 (*)

-0.01 to -0.12




0.85 to 0.92 (*)

-0.23 to -0.09






New Zealand7










(*) Note: Figures represent a range of correlations from several studies.


1 Khosla & Lowe, 1967

5 Lee et al., 1981.

2 Evans & Prior, 1969.

6 Goldbourt & Medalie, 1974.

3 Florey, 1970.

7 Watson a al., 1979.

4 Smalley et al., 1990.


Weight/height3 or an inversion of this index (H/W1/3) (also known as the Ponderal index) has been shown to have a substantially lower correlation with body weight than the other two indices (r = 0.6) and a negative correlation with height (r = -0.3) (Khosla & Lowe, 1967). A comparative analysis showed that in several populations W/H had the highest positive correlation with height, while there was a strong negative correlation for W/H3 (Lee et al., 1981). Yet, there were only modest correlations with body weight, and W/H3 was the least useful power type index for epidemiological uses.

It has been suggested, on the basis of a comparative analysis of the three power type indices, that W/H2 or BMI is both highly correlated with weight and consistently independent of height (Khosla & Lowe; 1967). It is, therefore, the index of choice for epidemiological purposes.

Several studies, summarized in Table 1.2, support the value of BMI as being highly correlated with body weight and at the same time independent of height. Zero order correlation analyses, as well as age-adjusted partial correlations of BMI with weight and height, show that BMI was highly correlated with weight in each ethnic and sex-specific population but was less influenced by height within each population group. The demonstration of such a relationship between BMI and body weight, unbiased by stature across a range of ethnic groups in populations worldwide, makes BMI a good choice for the anthropometric assessment of nutritional status of adults for epidemiological studies. BMI also has the advantage of being an index that has been reported in the literature over the past century, providing an excellent foundation for comparison purposes.

Previous PageTop Of PageTable Of ContentsNext Page