EPR/81/4
August 1981
| FOOD AND AGRICULTURE ORGANIZATION OF THE UNITED NATIONS | ESN:FAO/WHO/UNU/ EPR/81/4 August 1981 |
| WORLD HEALTH ORGANIZATION | |
| THE UNITED NATIONS UNIVERSITY |
Item 2.1.3. of the Provisional Agenda
Joint FAO/WHO/UNU Expert Consultation on Energy and Protein Requirements
Rome, 5 to 17 October 1981
DETERMINATION AND EXPRESSION
OF
ENERGY REQUIREMENTS
by
D.W. Spady
University of Alberta, Edmonton Alberta, Canada
Introduction
This paper will discuss methods of determining and expressing energy requirements. The discussion is based upon a review of the relevant literature available to me and is somewhat theoretical in nature. There are no new data presented, nor are there presented any estimates of requirements.
ENERGY INTAKE
Energy requirements traditionally have been estimated by measuring the energy intake of individuals and extrapolating the results to populations. The time frame employed has varied from 24 hours to 7 days or more and the instruments used have included in part the following: dietary recall, weighed or nonweighed recorded diets, duplicate diets, prepackaged defined diets and liquid defined diets. Many of these have been designed primarily in a laboratory or clinical setting although they can be used in the field. These techniques have not been without their problems.
Disadvantages of Intake Data
(a) Accuracy and Precision
It is difficult to comment confidently about the accuracy of any of these techniques because there is no absolute standard against which to compare the results. For this reason, “accuracy” will refer to the degree of concordance that measures of energy intake have with measures of energy expenditure, when both are felt to employ satisfactory techniques of measurement.
Garrow (1978) in an extensive review of the methods of measuring energy intake concluded that even by using very motivated subjects and maintaining an accurate weighed measurement of food intake, the error in measurement of intake will commonly be 4% and may be as much as 10%; more commonly, the average error may be as much as 25%. Some factors accounting for these errors are: subject memory, motivation, and intelligence, skill of the interviewer, plate waste, methods of food preparation, homogeneity of nutrient composition of food and the use of food tables. Suffice to say that the measurements felt to be the most accurate are also those measured under conditions least representative of the routine environment of the individual. This criticism may be directed equally well to methods of measuring energy expenditure, thus, although the use of the more accurate methods of measurement may be acceptable for various metabolic investigations of energy balance, they cannot be considered representative of any population group.
Harries, Hobson and Hollingsworth (1962), in a review of the literature describing energy expenditure and 7 day energy intake survey records noted that there was an individual variation in recorded energy intake of at least 15%; variation in expenditure was only 10%. They concluded that “the use of intake data to estimate (energy) requirements is justifiable for mean values, but not for estimates of variation.”
(b) Relevance
Activity patterns, and hence energy needs, change with time. For this reason, intake data of thirty or more years ago may be invalid. This possibility has been suggested by Griffiths and Payne (1976) who noted that 4 to 5 year old children had an average energy intake of between 1100 and 1400 kcal, compared to a mean energy intake of 4 to 5 year old children in the 1930's of 1770 kcal. They suggested that the decrease in intake is associated with changes in patterns of behavior which may be reflected in altered patterns of activity and play of children. Spady (1976) found in 9 to 11 year old girls an average daily energy expenditure some 400 kcal less than the recommended energy intake, suggesting again that energy requirements are overestimates of energy expenditure. In older individuals living in developed countries we may see this discrepancy between actual energy intake and recommended energy requirements (as estimated from earlier studies on intake) being reflected in the increasing frequency of obesity and overweight. The obvious conclusion from this discussion is that if energy intake is to be used to estimate energy requirement, then up to date, representative surveys are essential.
ADVANTAGES OF THE USE OF INTAKE DATA
There are certain advantages to the use of dietary intake surveys for estimating energy requirements.
(a) Ease and Availability
Although accurate intake data are hard to obtain, when compared to the much greater difficulty of obtaining accurate data on energy expenditure, the choice is obvious. More individuals may be surveyed in the same time frame than would be possible if energy expenditure were measured. By employing good survey techniques and planning it may be possible to survey a wide variety of individuals, using sample sizes for each subgroup which would permit fairly easy statistical manipulation of the data to provide maximum amounts of statistically valid data.
(b) Growth
At certain stages of life when growth is rapid and stored energy may constitute a significant proportion of total energy intake, measurement of energy intake is more accurate in determining energy requirements than is the measurement of energy expenditure. In early infancy, 25% to 30% of total energy intake is diverted to new growth (Fomon, 1974; Brooke, Alvear & Arnold, 1979); in this circumstance, measurement of energy expenditure only as an estimate of energy requirements would lead to underestimates. A similar argument may be applied to children recovering from malnutrition (Spady, 1976) and theoretically also to pregnant and nursing mothers. At other ages, growth ceases to constitute a significant proportion of average daily energy requirement and then measurements of energy expenditure theoretically should be equally valid as indices of requirement.
ENERGY EXPENDITURE
Introduction
Energy intake is influenced to some degree by the availability of food, environment, and food habits. On the other hand, energy expenditure reflects the actual amounts of energy used by the individual and, with the exception of growing individuals mentioned previously, it should be a more desirable measure of the actual energy requirement of individuals.
Techniques of Measuring Energy Expenditure
Techniques available to measure energy expenditure include direct calorimetry using whole body calorimeters, and various methods of indirect calorimetry such as: ventilated hood systems, Douglas bags, Kofranyi-Michaelis respirometers and electrical devices, the most promising of which is the heart rate recorder. Direct calorimetry is the most accurate way of assessing energy expenditure and of much value in elucidating some of its determinants but of little value in determining the average energy used by individuals under the conditions of normal life, although newer, more portable direct calorimeters have been described (Webb, Annis & Trautman, 1980) and used in the field. For the most part, simultaneous measurements of expenditure by direct and indirect calorimetry agree to within the probable range of experimental error (Dauncey, 1980), although under certain circumstances, when subjects consume significantly less energy than they expend, the error can rise to between 8% and 25% (Webb, Annis, & Trautman, 1980). Notwithstanding this observation, indirect calorimetry methods, notably the K-M respirometers have been used to estimate energy expenditure for a great number of different activities and their use, together with accurately recorded activity diaries, when employed by experienced and motivated individuals, has been shown to reflect accurately the energy expenditure of individuals (Norgen, Ferro-Luzzi, & Durnin, 1974; Durnin & Brockway, 1959). In a well planned and implemented study, Durnin and Brockway concluded that the actual error in the determination of energy expenditure using respirometers and activity diaries was at least 5%; in other less experienced hands, the error could be much more.
Concordance Between Measured Energy Expenditure and Energy Intake
There are not many field studies which have measured simultaneously energy expenditure and energy intake. Among those carried out many have examined fairly regimented populations such as schoolboys (Banerjee & Saha, 1972), medical students (Banerjee & Mahindra, 1962), cadets and military students (Edholm, Fletcher et al, 1955). Harries, Hobson and Hollingsworth (1961) summarized data for many similar types of groups. The problem with these data lies in the fact that many people are not particularly regimented and therefore are more likely not to be measured; this tends to bias survey results. Few studies have examined the energy balance of more “free living” persons. A classic study of this type is that of Norgen, Ferro-Luzzi and Durnin (1974) among New Guinea natives. In this study, in one population group the degree of concordance of intake with expenditure was considerably less than desired.
Generally speaking, in these studies the concordance of intake and expenditure was very good, with errors of between 5 and 10% if measurements had been made for periods of greater than seven days; however, over shorter periods of time, correlations were poor and balances were characterized by wide variation in both intake and expenditure.
Most of the above studies used K-M respirometers and activity diaries. Other studies have used heart rate recorders. Acheson (1975) used both K-M respirometers and heart rate recorders to study the energy balance of men living in Antarctica and found that the heart rate method overestimated the standard (weighed food intakes over long periods of time) by 3%, the diary method, using measured values for activities, underestimated by 6% and by 11% if literature values were used. No one method was best for all subjects. Griffiths and Payne (1976) measured 5 – 7 day food intakes and estimated energy expenditure for 4 – 7 days with heart rate recorders in 4 and 5 year old children and found a very good concordance between intake and expenditure. Dauncey and James (1979) compared heart rate recorders with whole body calorimeters and found a marked variation in the degree of error obtained, the prime determinant of error being the manner in which the recorder was calibrated. The best predictions had an error of less than 5%. Heart rate recorders appear to be least accurate when measuring sedentary activities and most accurate when individuals were more active. It would appear that, in knowledgable hands and using properly calibrated regression equations, the use of heart rate recorders is as accurate as the use of K-M respirometers and activity diaries and has the tremendous advantage of great portability and lack of interference with normal daily activities. In addition, heart rate recorders may be used in children, something which other techniques cannot easily do. However, as with all the techniques of measuring energy expenditure it is time consuming and, if one accepts that measurement of either intake or expenditure is equally valid for determining energy requirements, as it appears to be for the most part, then the measurement of energy intake is more cost beneficial--since information on the consumption of other nutrients may be obtained simultaneously.
The Factorial Method
The energy requirements of an individual may be divided into maintenance energy, and energy for growth, activity, and on occasion, stress of illness. It has been postulated that the energy requirement of an individual could be determined by summing values for each of these parts. Such an approach has a certain appeal and in fact possibly may provide relatively accurate estimates for populations assumed to have certain sizes, weights and activity patterns; however, it is of little value in individuals, primarily because of the large interindividual variation in energy expenditure for basal metabolism and for activity. In addition it has the dangerous disadvantage of providing an illusion of precision to the estimate.
Maintenance energy refers to the energy necessary to maintain a sedentary and healthy individual at constant weight. It can be approximated to be 1.5BMR and estimated to be 105W7 5, where W is weight in kg (Payne & Water low, 1971) and where BMR is approximated by the equation 70W7 5 (Kleiber, 1975). For groups of children (Spady, 1980) and elderly men (Calloway & Zanni, 1980), measured estimates of maintenance requirements are very close to this theoretical value, although Spady noted that there was a definite sex difference; the average maintenance expenditure of boys was 109.5W7 5 and of girls was 95.6W7 5. These studies were in healthy individuals. Data of Schutz, Lechtig, and Bradfield (1980) describing the energy expenditure of poor women in Guatema la have been analyzed by myself and the resting energy expenditure is best approximated using a value of 60W7 5 instead of the corresponding value of 70W7 5. From these data, it is clear that in this population, using the formula of Kleiber would lead to overestimates of energy expenditure, consequently values of maintenance would also be excessive. Further data is required on larger population groups to assess the usefulness and validity of this particular value.
The energy cost of growth has been estimated by Payne and Water low (1971) to be approximately 5 kcal/gm weight gain. Measurements by Spady (1976) and Brooke (1979) have provided values close to this; however, it must be remembered that the value of 5 kcal/gm is a function of the nature of the tissue deposited and could vary by a factor of 2 or more. Again, for population estimates such a value may be useful in infancy but not on an individual basis.
Many of the comments made about maintenance and growth apply also to activity. Although the energy cost of specific activities has been measured many times in adolescent and adult populations, there are few data on the proportion of daily energy expenditure devoted to activity. In children, Spady (1980) estimated that between 25% and 31% of energy expenditure of 9 to 11 year old children was devoted to activity; these values are close to the FAO estimates of 26% in the last report. Griffiths and Payne (1976) found that “normal” 4 to 5 year old children expended about 25% of their energy expenditure on activity, a good agreement with the estimated 28% in the FAO report.
At the present time I am not satisfied with the factorial method of estimating energy expenditure. Much of my dissatisfaction is a function of the lack of data available for different population groups. The study mentioned above by Schutz et al illustrates my concern. The use of Payne and Water low's formula overestimates maintenance energy needs in this population, a kind of population very common in this world. This illustrates the need to obtain information about basal and maintenance metabolic rates for many different population subgroups and appropriate values used for factorial estimations in the future. Similar comments could be directed towards factorial estimations of energy devoted to activity.
The cost of growth may be higher in infants in poor countries, not because of any inherent metabolic differences, although genetic differences may exist, but because they get sick more often. This factor must be considered when estimating energy requirements in infancy.
METHODS OF EXPRESSING ENERGY REQUIREMENTS
GENERAL COMMENTS
It is my strong feeling, philosophically expressed throughout this review, that the energy requirements published by the FAO should not be used to determine the energy requirements of an individual, rather they should be used to make estimations of the energy requirements for population groups and subgroups. For individuals, FAO recommendations may be used as a starting point but it should be clear tha the estimated energy needs of an individual can be determined clinically only by extensive dietary investigation and consideration of the activity and present energy balance and health status of the person; even then the estimation probably will have an error of 5 to 10%.
The parameter used to express energy requirements should meet as closely as possible the following simple criteria:
At present, no parameter meets all of these requirements; there is no method known which can express energy requirements as a constant value at differing stages of the life cycle. In addition, no parameter considers the activity level of the individual, which can vary widely and randomly. For the above reasons therefore, all methods used require the primary subdivision of energy requirements into various age, sex and activity groups.
With these considerations in mind, what is available, what should be used, and what can be ignored. Available parameters include time, weight, height, weight/height (ideal weight), lean body mass, total body potassium, body surface area and metabolic body size.
Time
Time alone as a parameter is commonly used, however, implicit in its use but commonly not stated explicitly, is the fact that individuals are considered to have a standard weight and age and occasionally a standard height, to which this 24 hour energy requirement is assumed to apply. There are few “standard” individuals and the reverse of the above situation should be created, i.e. daily energy requirements should be expressed in terms of energy/parameter/24h (e.g. kcal/kg/24h). Although I doubt that this will change the accuracy of any of the predicted energy requirements, it will emphasize the basis upon which the energy requirements are expressed.
For the expression of energy costs of specific activities, the suggested manner is kcal/min/kg because few activities are carried on for an hour or more without interruption (Durnin, 1959).
Weight and Height
Although weight does not correlate well with energy intake or energy expenditure (Acheson, Campbell et al, 1980(A); Norgan et al, 1974; Edholm, Fletcher et al, 1955), traditionally, energy requirements have been expressed in terms of body weight and today, most authorities continue this practice. Few parameters offer any advantage over weight, it meets many of the criteria mentioned above and there are large amounts of data available which express energy intake and expenditure in terms of weight. One disadvantage of using weight alone to express energy requirements is that in chronically underweight or overweight populations, the use of energy requirements based on weight may tend to perpetuate a poor situation. A refinement would be to express energy requirements in terms of ideal weight/height, or in terms of height alone. There is some justification for this. In clinical medicine there is a tendency to determine the “ideal” weight of an individual based on his height and to use that value as the basis for suggested energy requirements. It is interesting to note that in the 1973 Protein and Energy Requirements report (pp 26) the expression of energy requirements in terms of expected weight/height of the population was advocated yet neither the reference man or reference woman has a height recorded and weight/height is not discussed elsewhere in that report.
The expression of energy requirements in terms of height alone has virtually all of the benefits of using ideal weight/height yet is mathematically and conceptually easier to grasp. Wait, Blair, and Roberts (1969) found that the total energy intake of well nourished children correlated well with body weight, height and body surface area. They felt that “height has an advantage as an index to need for it is influenced less by diet than is weight. It can, therefore, be used more effectively in estimating the needs of the heavy or slight child. Probably the most dependable index of need, particularly for girls, are calories per unit of height and age.” With this consideration in mind it is noteworthy that the latest Canadian Dietary Standards (DNHW, 1976) express energy requirements in terms of height as well as weight. Clearly, as in this entire discussion, consideration must be given for age, sex, and activity.
Other Modes of Expression
With the exception of growing individuals or those who are deliberately trying to lose weight, energy requirements should be equal to energy expenditure. Therefore, methods of estimating energy expenditure may also be used to express energy requirements. This is of value when estimating energy requirements from metabolic body size using the formula of Kleiber (1975) and hence energy requirements may be expressed in terms of kcal/kg7 5. Although there is a sound physiological argument to using the concept of metabolic body size to estimate energy expenditure, its use in expressing energy requirements offers little advantage over the conceptually much simpler weight or height parameters and I would be distressed to see energy requirements being expressed routinely in this manner.
A similar argument may be used to advocate the energy requirements in terms of body surface area; however, Kleiber, (1975, pp201) in an extensive review of the basis for using body surface area to estimate energy expenditure concluded “The surface area is unreliable mainly because the definition of an animal surface is vague. But even if the surface area could be defined and measured accurately, there is no theoretical basis for the hypothesis that the metabolic rate of homeotherms should be exactly proportional to their particular surface area rather than to a more general function of body size.” Durnin (1959) came to a similar conclusion.
Lean body mass, although being a good predictor of basal metabolic rate and hence maintenance energy requirement (Cunningham, 1980), again has little to offer as a manner of expressing energy requirements. Perhaps its most important detractor is that it is difficult to measure accurately without sophisticated laboratory facilities and hence would be of little use in surveys.
For the reasons outlined above, it is my feeling that energy requirements should be expressed in terms of weight and/or height. No other parameters offer significant advantages over these.
Donald W. Spady M.D., M.Sc., F.R.C.P. (C)
Department of Pediatrics
University of Alberta
Edmonton, Alberta, Canada
BIBLIOGRAPHY
Acheson, K.J., 1975: A Longitudinal Study to Assess the Current Field Techniques For Measuring Energy Balance in Man. Br. Antarct. Surv. Bull. Nos. 41 & 42, pp 87–97.
Acheson, K.J., I.T. Campbell, O.G. Edholm, D.S. Miller & M.J. Stock, 1980(A): The Measurement of Food and Energy Intake in Man--An Evaluation of Some Techniques. Amer. J. Clin. Nutr. 33:1147–1154.
Acheson, K.J., I.T. Campbell, O.G. Edholm, D.S. Miller & M.J. Stock, 1980(B): The Measurement of Daily Energy Expenditure--An Evaluation of Some Techniques. Amer. J. Clin. Nutr. 33:1155–1164.
Banerjee, B., S.K. Mahindra, 1962: Energy Intake and Expenditure of Medical College Women. J. App. Physiol. 17:971–973.
Banerjee, B., N. Saha, 1972: Energy Intake and Expenditure of Indian Schoolboys. Brit. J. Nutr. 27:483–490.
Brooke, O.G., J. Alvear, M. Arnold, 1979: Energy Retention, Energy Expenditure and Growth in Healthy Immature Infants. Pediat. Res. 13:215–220.
Calloway, D.H., E. Zanni, 1980: Energy Requirements and Energy Expenditure of Elderly Men. Amer. J. Clin. Nutr. 33:2088–2092.
Cunningham, J.J., 1980: A Reanalysis of the Factors Influencing Basal Metabolic Rate in Normal Adults. Amer. J. Clin. Nutr. 33:2372–2374.
Dauncey, M.J., W.P.T. James, 1979: Assessment of the Heart Rate Method For Determining Energy Expenditure in Man, Using a Whole Body Calorimeter. Brit. J. Nutr. 42(1) 1–13.
Dauncey, M.J., 1980: Metabolic Effects of Altering the 24h Energy Intake in Man, Using Direct and Indirect Calorimetry. Brit. J. Nutr. 43:257–269.
DNHW, 1975: Dietary Standard For Canada. Bureau of Nutritional Sciences, Department of National Health and Welfare, Ottawa.
Durnin, J.V.G.A., 1959: The Use of Surface Area and of Body Weight as Standards of Reference in Studies on Human Energy Expenditure. Brit. J. Nutr. 13:1:68–71.
Durnin, J.V.G.A., J.M. Brockway, 1959: Determination of the Total Daily Energy Expenditure in Man by Indirect Calorimetry: Assessment of the Accuracy of a Modern Technique. Brit. J. Nutr. 13:1:41–49.
Edholm, O.G., J.G. Fletcher, E.M. Widdowson, R.A. McCance, 1955: The Energy Expenditure and kFood Intake of Individual Men. Brit. J. Nutr. 9:286–300.
Fomon, S.J., 1974: Infant Nutrition. W.B. Saunders Co. Philadelphia, London, Toronto. 2nd ed.
Garrow, J.S., 1978: Energy Balance and Obesity in Man. 2nd ed. Elsevier/North-Holland Biomedical Press. Amsterdam, New York, Oxford.
Griffiths, M., P.R. Payne, 1976: Energy Expenditure in Small Children of Obese and Non-Obese Parents. Nature, 260:698–700.
Harries, J.M., E.B. Hobson, D.F. Hollingsworth, 1962: Individual Variation in Energy Expenditure and Intake. Proc. Nutr. Soc. 21:157–169.
Kleiber, M., 1975: The Fire of Life: an introduction to animal energetics. R.E. Krieger Pub. Co. Huntington, New York. 2nd ed.
Norgen, N.G., A. Ferro-Luzzi, J.V.G.A. Durnin, 1974: The Energy and Nutrient Intakes and the Energy Expenditure of 204 New Guinean Adults. Phil. Trans. R. Soc. Lond. B. 268:309–348.
Schutz, Y., A. Lechtig, R.B. Bradfield, 1980: Energy Expenditures and Food Intake of Lactating Women in Guatemala. Amer. J. Clin. Nutr. 33:892–902.
Spady, D.W., P.R. Payne, D. Picou, J.C. Waterlow, 1976: Energy Balance During Recovery From Malnutrition. Amer. J. Clin. Nutr. 29:1073–1088.
Spady, D.W., 1980: Total Daily Energy Expenditure of Healthy Free-Ranging School Children. Amer. J. Clin. Nutr. 33:766–775.
WHO, 1973: Energy and Protein Requirements. WHO Tech. Rep. Ser. No. 3,522, 1973.
Wait, B., R. Blair, L.J. Roberts, 1969: Energy Intake of Well Nourished Children and Adolescents. Amer. J. Clin. Nutr. 22:1383–1396.
Webb, P.J., J.F. Annis, S.J. Trautman, 1980: Energy Balance in Man as Measured by Direct and Indirect Calorimetry. Amer. J. Clin. Nutr. 33:1287–1298.
