August 1981

Item 2.2.5 of the Provisional Agenda

Joint FAO/WHO/UNU Expert Consultation on Energy and Protein Requirements

Rome, 5 to 17 October 1981



I.T. Johnson, D.A.T. Southgate

A.R.C. Food Research Institute, Norwich, UK

1. Introduction

The energy value of the human diet is a fundamental nutritional concept of almost universal interest, and the accurate determination of such values is a prerequisite for effective nutritional programmes, and for the clinical management of such conditions as diabetes and obesity. The standard method of calculating energy values for diets of known composition is by the application of calorie conversion factors, the most important of which (4 Kcal per gm protein; 9 Kcal per gm fat; 4 Kcal per gm carbohydrate) are based on the work of Atwater carried out near the turn of the century (1). The use of such factors depends upon various assumptions, amongst which the most important is that the nutrients to which they are applied are in fact fully available for absorption and metabolism.

Dietary fibre, which comprises the plant cell wall materials and storage polysaccharides which are nutritionally unavailable to man, have been the subject of a surge of interest in recent years. Its relevance in the present context stems partly from its effect upon the accurate analysis of nutrient composition, and partly from its possible influence upon the behaviour of nutrients within the intestine, leading, perhaps to energy malabsorption. Both of these effects are potential sources of inaccuracy in the calculation of energy values, and recent work has led to the suggestion that the Atwater factors may need to be modified to take account of them (2).

The purpose of this paper is to examine and review the extent to which the presence of unavailable carbohydrates in the diet may render inaccurate the Atwater factors for fats and carbohydrate. The influence upon the absorption of protein is discussed in a separate paper.

2. Unavailable Carbohydrates: The calculation of potentially available energy

The total available energy of a food may be defined simply as its heat of combustion, minus the heat of combustion of the faecal and urinary residues to which it gives rise. Oxidisable urinary losses are derived from excreted protein. Faecal energy losses arise from several sources, including endogenous waste material such as mucus and exfoliated mucosal cells, bacterial cell mass from the intestinal flora, and finally undigested food residues themselves. All these may be contributed to or modified by the presence in the diet of unavailable carbohydrate, a term used here to include both dietary fibre, and some sugars, notably the oligosaccharides of the raffinose group, which are not hydrolysed or absorbed by the digestive/absorptive system of man (3).

Serious problem in the assessment of the energy value of diets are likely to arise where the carbohydrate content has been determined “by difference”, that is, by subtracting from 100 the sum of the water, ether extractives, protein and ash, and regarding the remainder as carbohydrate, without reference to the availability of its component compounds. Clearly, the most satisfactory approach to this problem would be the direct chemical analysis of all the carbohydrates present, so that an estimate of availability could be based on a knowledge of the fate of each component in the intestine. Such an analysis would take into account the possibility that some components of dietary fibre may be indirectly available to man via microbial digestion to fatty acids followed by absorption in the large intestine. Though this possibility cannot be entirely discounted, its importance is likely to vary considerably with the source of fibre (4), and with the intestinal flora of the individuals in question(5). Furthermore, even if all the products of microbial digestion of pentosans and cellulose in subjects consuming a high fibre western diet are available for absorption, the total contribution to the energy budget would be negligible for practical purposes(6). A good approximation to the total available carbohydrate value of a diet can therefore be achieved by the analysis of starch and free sugar only.

The practical application of calorie conversion factors to conventional western diets was investigated by Southgate and Durnin(7), who conducted a balance study for energy and the major nutrients using groups of young and old men and women consuming diets which contained different levels of unavailable carbohydrate. The results showed that the Atwater factors could be applied to a normal British diet, low in dietary fibre, with little error, even when no allowance was made for unavailable carbohydrate. As the level of fibre in the diet rose, so the apparent digestiblities of the nutrients fell, but even so, no real inaccuracy resulted so long as the factor for carbohydrate was reduced to 3.75 and used for available carbohydrate (starch and sugar) expressed as monosaccharides

The importance of making adequate correction for the unavailable carbohydrates in high fibre diets characteristic of the developing countries, is illustrated by the work of Watson et al (8, 9), who have compared the “available carbohydrate” composition of a range of Ghanaian foodstuffs with the carbohydrate content determined “by difference”, but excluding the “crude fibre” content. The work demonstrated that the application of the Atwater factors to carbohydrate content determined “by difference” resulted in an overestimation of available energy, ranging from 12 to 21 percent, depending on the composition of the diet. The discrepancy between the two values varied markedly between foodstuffs. For cereals, starchy roots and leafy vegetables, the differences were greatest in the case of foods which had been processed by means such as formentation or drying. Legumes and legume products also showed considerable differences, thought to be due to the high concentrations of gums and unavailable sugars which they contain.

3. Unavailable Carbohydrate: The Absorption of Potentially Available Energy

Apart from the overestimation of the available energy content of foodstuffs which can arise from a failure to allow for their unavailable carbohydrate content, inaccuracy would also result if the decline in the apparent digestibility of nutrients in high fibre diets were to reflect true energy malabsorption. It has long been claimed that if the levels of fat and nitrogen in the diet are kept constant, faecal losses of both nutrients rise with the addition of unavailable carbohydrate to the diet(10). Thus the apparent digestibility of these nutrients, defined as the difference between intake and faecal losses expressed as a percentage of intake, is seen to fall. True energy malabsorption would result if the addition of unavailable carbohydrates to the diet were to prevent the absorption of nutrients which would otherwise be available.

A number of possible mechanisms can be envisaged whereby true energy malabsorption might be brought about. It has been proposed, for example, that the physical properties of the intestinal contents may render readily digestible substrates inaccessible to digestive enzymes, or that the products of digestion may be unable to diffuse to the absorptive surface(6). This, coupled perhaps with a more rapid transit through the small intestine, might be expected to lead to a loss of nutrients into the large intestine where they would either be excreted intact, or used as substrates for bacterial growth.

Certainly there is ample evidence that such factors as the vegetable source of available carbohydrate (11), the physical form in which it is presented (12,13) and the presence of viscous fibre components (14) can all reduce the rate at which carbohydrate is digested and absorbed. However, human faeces never normally contain starch or sugars (7, 24). Moreover, it is likely that the capacity of the small intestine to digest and absorb carbohydrates (15) and fats (16) is several times higher than normal nutritional needs demand, and though the rate of absorption may be reduced, there is little evidence that this can lead to a reduction in total uptake. Another possibility is that a high intake of dietary fibre may lead to increased intestinal cell turnover or mucus secretion, and a consequent loss of endogenous fat and protein. This possibility awaits physiological confirmation however.

Recent studies tend to confirm that the consumption of unavailable carbohydrate leads to a decline in the apparent digestibility of other nutrients, but have not explained the mechanism underlying the effect. The work of Southgate and Durnin(7) showed that the faecal energy losses of subjects consuming a mixed British type diet were increased by the addition of higher levels of dietary fibre from a variety of fruit and vegetable sources. These losses were partly due to increased faecal excretion of fat and protein. Subsequent studies, using mixed diets of various kinds, confirm these observations (2,17), though a decrease in the apparent digestibility of fats is not always observed (18).

In other studies, the effect upon faecal composition of specific types of unavailable carbohydrate, given as supplements to subjects consuming controlled diets, have been observed. Prynne and Southgate (19) gave a 25 gm supplement of Isphagula husk to four subjects, and observed changes in faecal composition which varied between individuals, only one of which displayed any decrease in the apparent digestibility of energy and fat. Two other recent studies, both of which utilised supplements of refined cellulose (20, 21), have also failed to indicate a decrease in the apparent digestibility of fat.

There is a lack of physiological data as to the possible endogenous origin of energy rich faecal residues, but there is little evidence that potentially available nutrients do pass through the small intestine undigested in the presence of dietary fibre, and in fact, recent studies with ileostomy patients provide evidence to the contrary (22).

A more likely explanation perhaps is that the increased faecal excretion of nitrogen and fat is primarily of microbial origin, and represents the metabolic products of the intestinal microflora which is able to utilise unavailable carbohydrate (23).

This explanation is consistant with the observations of Walker, who showed that the faecal excretion of nitrogen and fat in children consuming a high-fibre diet, was stimulated not by protein or fat supplements, but by a daily supplement of fibre from oranges, which was essentially fat and protein free (24). Walker concluded from this and other observations that the extra energy rich material voided by consumers of high fibre diets did not represent a loss of potentially useful nutrients, and was therefore not of significance in estimating nutritional needs.

4. Conclusions

Two major sources of error may militate against the use of calorie conversion factors to achieve accurate estimations of nutritional energy values. These are, first, errors resulting from inaccurate analysis of the potentially available nutrients in the diet, and secondly, errors resulting from interactions between components of a mixed diet. Of these, the first seems likely to be by far the more important. The use of carbohydrate values determined “by difference” is a particularly serious source of inaccuracy, and would lead to considerable overestimations of the energy of diets rich in dietary fibre, as unmetabolisable sugars. The best solution would be a complete analysis of all the carbohydrate components of the diet, but a good approximation is achieved by the analysis of available carbohydrates, starch and free sugars.

Given sufficiently accurate analyses of the diet, there seems little reason to revise the conclusion of Southgate and Durnin (7) that for conventional British diets, the Atwater values, 4 and 9 for protein and fat, and 3.75 for monosaccharide sugars are adequate for practical purposes. The problem of diets high in unavailable carbohydrates remains however, and is not easily resolved. Diets which differ in their dietary fibre content tend to differ also in other aspects of their composition, and this may partly explain the fact that the reduced apparent digestibility of fats seen in studies using mixed diets is not always seen when controlled diets are augmented with specific types of unavailable carbohydrate. Much of the variability seen between different experiments and different individuals may reflect differences in the intestinal microflora of subjects, and their response to various types of unavailable carbohydrate. Clearly the situation is complex, and it is doubtful whether any simple modifications of the Atwater factors would introduce any greater accuracy into the determination of energy values. Considerably more information is required as to the details of the physiological mechanisms governing faecal composition.


1. Merrill, A.L., and Watt, B.K. (1955) Energy value of Foods, basis and derivation. U.S. Dept. of Agriculture. A.R.S. Handbook No. 74.

2. Calloway, D.H., and Kretsch, M.J. (1978) Protein and energy utilization in men given a rural guatemalan diet and egg formulas with and without added oat bran. Am. J. Clin. Nutr. 31, 1118–1126.

3. Cristofaro, E.,Mottu, F., and Wuhrmann, J.J. (1974) Involvement of the Raffinose family of oligosaccharides in flatulence; in Sugars in Nutrition. H.L. Sipple and K.W. McNutt Eds. Academic Press, London 1974.

4. Cummings, J.H., Southgate, D.A.T., Branch, W., Houston, H., Jenkins, D.J.A. and James, W.P.T. (1978) Colonic response to dietary fibre from carrot, cabbage, apple, bran, and guar gum. Lancet 1978, 1, 5–9.

5. Stephen, A.M., and Cummings, J.H. (1980) Mechanism of action of dietary fibre in the human colon. Nature 284, 283–284.

6. Southgate, D.A.T. (1973) Fibre and other unavailable carbohydrates and their effects on the energy value of the diet. Proc. Nutr. Soc. 32, 131–136.

7. Southgate, D.A.T. and Durnin, J.V.G.A. (1970) Calorie conversion factors. An experimental reassessment of the factors used in the calculation of the energy value of human diets. Br. J. Nutr. 24, 517–535.

8. Watson, J.D., Dako, D.Y. and Amoakwa-Adu, M. (1975) Available carbohydrates in Ghanaian Foodstuffs. Part II Sugars and Starch in Staples and other foodstuffs. Plant Foods For Man, 1, 169.

9. Watson, J.D. and Dako, D.Y. (1976) Available carbohydrates in Ghanaian foodstuffs. Part III: Computation of the energy content of diets using carbohydrate values obtained either “by difference” or determined “as available”. Plant Foods For Man, 2, 63.

10. Widdowson, E.M. (1955) Assessment of the energy value of human foods. Proc. Nutr. Soc. 14, 142–154.

11. Jenkins, D.J.A., Wolever, T.M.S., Taylor, R.H., Barker, H.M. and Fielden, H. (1980) Exceptionally low blood glucose response to dried beans: comparison with other carbohydrate foods. Brit. Med. J. ii, 578–580.

12. Haber, G.B., Heaton, K.W., Murphy, D. and Burroughs, L.F. (1977) Depletion and disruption of dietary fibre: effects on satiety, plasma-glucose, and serum-insulin. Lancet ii, 679–682.

13. O'dea, K., Nestel, P.J. and Antonoff, L. (1980) Physical Factors influencing postprandial glucose and insulin response to starch. Am. J. Clin. Nutr., 33, 760–765.

14. Jenkins, D.J.A., Wolever, T.M.S., Leeds, A.R., Gassull, M.A., Haisman, P., Dilawari, J., Goff, D.V., Metz, G.L. and Alberti, K.G.M.M. (1978) Dietary fibres, fibre analogues, and glucose tolerance: importance of viscosity. Brit. Med. J. i, 1392–1394.

15. Crane, R.K. (1977) Digestion and absorption: water-soluble organics. Inte. Rev. Physiol. Gastrointestinal Physiology II, 12, 325–365.

16. Booth, C.C., Alidis, D., and Read, A.E. (1961) Studies on the site of fat absorption. Gut. 2, 168.

17. Kelsay, J.L., Behall, K.M. and Prather, E.S. (1978) Effect of fibre from fruits and vegetables on metabolic responses of human subjects. 1) Bowel transit time, number of defecations, faecal weight, urinary excretions of energy and nitrogen and apparent digestibilities of energy nitrogen, and fat. Am. J. Clin. Nutr. 31, 1149–1153.

18. Farrell, D.J., Girle, L. and Arthur, J. (1978) Effects of dietary fibre on the apparent digestibility of major food components and on blood lipids in men. Australian Journal of Experimental Biology and Medical Science, 56, 469–479.

19. Prynne, C.J. and Southgate, D.A.T. (1979) The effects of a supplement of dietary fibre on faecal excretion by human subjects. Br. J. Nutr. 41, 495–503.

20. Mickelson, O., Makdani, D.D., Cotton, R.H. Titcomb, S.T., Colney, J.C. and Gatty, R. (1979) Effects of a high fibre bread diet on weight loss in college-age males. Am. J. Clin. Nutr. 32, 1703–1709.

21. Slavin, J.L. and Marlett, J.A. (1980) Effect of refined cellulose on apparent energy, fat and nitrogen digestibilities. J. Nutr. 110, 2020–2026.

22. Sandberg, A., Andersson, M., Hallgren, B., Hasselblad, K. Isaksson, B. and Hultén, L. (1981) Experimental model for in vivo determination of dietary fibre and its effect on the absorption of nutrients in the small intestine. Br. J. Nutr. 45, 283–294.

23. Cummings, J.H., Southgate, D.A.T., Branch, W.J., Wiggins, H.S., Houston, H., Jenkins, D.J.A., Jivraj, T. and Hill, M.J. (1979) The digestion of pectin in the human gut and its effect on calcium absorption and large bowel function Br. J. Nutr. 41, 477.

24. Walker, A.R.P. (1975) Effect of high crude fibre intake on transit time and the absorption of nutrients in South African Negro Schoolchildren. Am. J. Clin. Nutr. 28, 1161–1169.

Top of Page