August 1981

Item 3.2.3 of the Provisional Agenda

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

Rome, 5 to 17 October 1981




S.G. Srikantia
University of Mysore


It has long been recognised that proteins differ in their ability to promote growth. The primary function of dietary protein is to supply nitrogen and aminoacids - both essential and non essential in amounts and proportions needed for the synthesis of tissue protein. The aminoacid content and profile is thus a critical determinant of protein quality and most methods which measure protein quality are directly or indirectly related to the efficacy with which they can satisfy aminoacid requirements.

All aminoacids have to be present simultaneously in adequate quantities and proper proportions at the site, for protein synthesis. A deficit in any one essential aminoacid (EAA) would limit protein synthesis proportionate to the extent of deficit (1). From this has developed the concept of the “most limiting aminoacid” and applied to judge dietary protein quality. If the composition of an “ideal” protein is known, which is completely utilised and has neither a deficit or excess of any EAA, the value of any given protein can be arrived at by calculating the extent of deficit of each EAA, in relation to that present in the ideal protein. The EAA in greatest deficit would become the limiting aminoacid and determine the protein value.

2. Use of standards

Egg protein has been used in the past as a standard, but this is not ideal. Although at levels of intake below requirement it is almost totally utilised, it is not so at higher levels. Egg protein has several EAA in excess, and at high intake levels, it can stand dilution with proteins of inferior quality without a compromise in its own quality. Recognising this, an expert committee of the FAO in 1957, proposed that an aminoacid scoring pattern in which the amounts of EAA in one gram of protein were about double that of the estimated adult human requirements should serve as an appropriate standard (2). It soon became apparent that the FAO pattern was not satisfacory. An FAO/WHO Committee which met in 1965 replaced the FAO pattern with that of egg/milk (3). A later FAO/WHO Committee in 1973 suggested the use of a new aminoacid pattern as the basic for computing aminoacid scores, instead of using a naturally occuring protein. The pattern was based upon current knowledge of human aminoacid requirements - largely that of adults.

The use of an aminoacid scoring system, though rational, has limitations. It does not take into account several factors which influence its practical application. Data on the aminoacid content after acid hydrolysis of a raw protein do not provide information on the availability of the aminoacid for absorption. Cooking and processing can adversely affect aminoacid availability, particularly of lysine and sulphur aminoacid, while severe heat treatment can lower availability of all aminoacids because of a decrease in digestibility of protein (5). Data on aminoacid composition do not provide information on the rate of release of aminoacids during digestion in the intestinal tract, the relative amounts of various EAA in the pool from which aminoacids are actively absorbed and of aminoacid imbalances which can influence utilisation of EAA (6).

Apart from these considerations, not all of the ingested protein is absorbed after degradation to aminoacids but some is absorbed in dipeptide form. The profile of dipeptides formed and the ratio of aminoacids to dipeptides in the intestinal lumen may have some significance with respect to protein quality.

The use of a reference protein or aminoacid scoring system for expressing the quality of other proteins has these limitations. In addition, the validity of the formulation and use of a reference protein or score depends upon the accuracy of knowledge of human aminoacid requirements. Requirements for adults and children are believed to be different and yet, so far a single standard has been used. This has been justified on the ground that a pattern which satisfies the needs of children will certainly satisfy those of an adult.

Recommended Dietary Allowances (RDA) based on requirement data should not be used to formulate practical diets, but used as guides to assess the probability of inadequate intake by population groups and to help countries plan their food supplies. The validity of the argument that an aminoacid score satisfactory for a child would meet the needs of an adult is not in question, but it may be debated whether there is not a need to formulate two separate patterns, since the use of the recommended pattern is most likely to overestimate RDA of approximately one half of a country's population with its implications for food supply.

3.Biological value

Among methods used to determine protein quality, N balance is one. From data on N balance, Net Protein Utilisation (NPU) which is the proportion of ingested protein retained can be calculated. By measuring True Digestibility (TD), the Biological Value (BV) can be arrived at - which is the proportion of the absorbed nitrogen retained. Biological Value and NPU may be considered as good measures of protein quality, since presumably, the retained N is used for protein synthesis. In experimental animals NPU can be directly estimated by carcass analysis and values are therefore likely to be more accurate than when BV and NPU are derived from N balance data, as it is done in human studies. The inaccuracies inherent in N balance studies are known, no matter how carefully conducted. NPU and BV thus measure the same parameter (N retained, except that BV is calculated from N absorbed and NPU from N ingested).

The concept of BV has the merit that it can be used to assess requirements of protein derived from foods with known quality differences, because BV is directly related to the efficiency of protein utilisation. It however has some serious limitations. It ignores the importance of factors which influence digestion of the protein and interaction of protein with other dietary factors before absorption. On theoretical grounds, the requirement of a protein which has a BV of 100, would be half that of another whose BV is only 50. The application of BV data for human protein requirements, is however, not this straightforward, because of methodological considerations. Conventionally, BV and NPU are determined using a single level of protein; more importantly, they are measured when the protein content of the diet is clearly below that of requirement, deliberately done to maximise existing differences in quality. Differences may however become considerably minimised, if not completely masked, when proteins are fed at levels above or close to requirement, since requirements of all EAA can be completely met even from a poor quality protein, when enough is consumed to satisfy the needs of its most limiting aminoacid. Thus what BV and NPU measure is the near maximal potential ability of the protein. That the utilisation of a protein - the % retained, falls with its increasing concentration in the diet was first shown over three decades ago (7) and subsequently repeatedly confirmed. BV can vary by a factor of two-form over 90% at low intakes (100 mg N/kg) to around 40% at high intakes (500 mg/kg) (8). In young men, BV of wheat gluten fell from 100 (intake 100 mg/kg) to 45 (intake 400 mg/kg) and 25 (intake 1.09/kg) when intakes progressively increased. (9). Similarly the BV of egg protein fell from a value of 100 at an intake of 200 mg/kg to around 60 and 70 when intakes increased to 400 and 500 mg/kg (10). As importantly, the dose-response relationship was found not to be linear through all ranges of intake, but curvilinear at both low and high intakes and at intakes approaching requirements. Differences in protein value as judged by BV were not evident when wheat gluten and egg were fed at levels below 200 mg/kg, but became clear at higher levels. Differences progressively increased as protein intake increased (11). The degree of carvature depends upon the most limiting aminoacid.

At intakes of protein approaching requirement, BV is considerably lower than maximal. To use values obtained under conditions designed to evaluate maximal potential, for purposes of calculating protein requirements would therefore have a high degree of inaccuracy built in. This would also be true when data on BV of two proteins obtained at levels meant to demonstrate maximal differences in quality are used to arrive at requirement of those two proteins, because proportionately between proteins seen at lower levels will not be maintained at higher levels. Requirements of any protein cannot be estimated from an extrapolation of the dose-response line obtained using low levels of the protein. If BV of a protein is to be used to compute human protein requirements, the value then, has to be measured at several intake levels close to requirement levels.

Another limitation of the use of BV as a measure of protein quality is that proteins which are completely devoid of one EAA can still have a BV of up to 40, because of the capacity of the organism to conserve and recycle EAA as an adaptation of inadequate intake of the aminoacid; also EAA needs for growth and maintenance are different (12).

Determination of BV of a single protein is of limited use for application to human protein requirements. No population derives all of its protein exclusively from a single food. Proteins come from a mixed bag of animal and vegetable foods or from a mixture of several vegetable foods. Mixtures of protein foods frequently promote better growth than anticipated from the performance of individual components of the mixture. This has been explained on the basis of a partial or complete correction of the constraint imposed by the limiting aminoacid present in individual proteins. That this may not be the sole explanation, is suggested by the observation that in some mixtures which promote better growth, levels of some EAA are lower than that seen in the better component of the mixture. A better aminoacid balance which improves utilisation of existing aminoacids has been suggested as a possible explanation.

When two or more protein sources are mixed, the outcome in terms of quality may be that the mixture has:

  1. a value which lies between those of the components predicted by changes in EAA composition,

  2. a value close to or identical with that of the better component or

  3. a value which is even higher than that of the better component.

There is no instance where the value falls below that of the poorest component. Various explanations, including changes in the ratio of total EAA to non EAA and digestibility of protein have been offered (13). The important point from the application angle, is that the determination of BV of a single protein has limited value, and that BV should be determined on combinations of proteins present in habitual diets. Such determination should be made at levels of protein intake needed for N balance in adults and for growth in children. It would be best to evaluate protein quality of the diet in the form in which it is actually eaten. This would be the procedure which would provide the most direct estimate of protein quality, but even such data suffer from some limitations. Such studies per force are done under controlled conditions, usually in a metabolic ward. There is evidence that some of the conditions under which such studies are done influence the results and that these conditions may not apply to real life situations.

4. Protein Calorie Interaction

One of them relates to an important facet of the protein-calorie interaction. Inadequate energy intake lowers the efficiency of protein utilisation and in most N balance studies, calorie adequacy is ensured. Although early data have suggested that excess calories promote better N retention, its practical implications had by and large been overlooked until recently. Results of some recent studies have reemphasised the importance of this interaction. The utilisation of both egg and polished rice protein in young men was about 30% higher and protein required to maintain N balance lower when fed at a calorie intake which was about 25% higher than that needed for maintenance (14). Similarly, when egg protein was fed at the “FAO/WHO 1973 safe level”, young men were in negative balance when their calorie intake was at maintenance level and they achieved positive balance only when calorie intakes were raised by 9 to 14% (15). Using two levels of protein at constant calorie intake and three levels of calorie at constant N intake, N balance was found to be influenced more by energy than by protein at marginal intakes (16). In all these studies, better N retention was due exclusively to decreased urinary N excretion. It needs to be ensured therefore that protein quality is always evaluated when calorie intake is at maintenance level and not in excess.

5. Frequency of food intake

The frequency of food intake appears to have an equally striking effect. In young women who consumed 62 g protein/day, 53 of which came from animal sources, the distribution of this protein into three meals instead of two, improved N utilisation through a reduction in urinary N (17). Similarly in 8 children who consumed 63 g of protein daily, 60% of which came from animal foods (2050 Kc) distribution of this into four meals instead of two, raised N retention almost three fold (18), again through a marked fall in urinary N. These findings are similar to that made in a single subject that at an intake of 1.14 g protein/kg, but not at a lower level, NPU was higher when the protein was consumed in four meals instead of two (19). The uniformity of ingestion of the day's total protein would appear to improve its utilisation. This factor is perhaps not taken care of sufficiently, in N balance studies, and emphasis the need to simulate as far as possible, the eating patterns of the populations for whose use N balance studies are carried out. There do not appear to be any data on what happens to N balance when exclusively vegetable proteins are fed in this fashion. The findings of such a study may have more than mere academic interest.

6. N utilisation from mixed proteins

A review of data on N balance studies in humans shows that at levels of intake of N between 70 and 100 mg/kg, through a wide variety of protein sources, such as egg, milk, meat and plant foods, either singly or in combination, efficiency of utilisation varies between a relatively narrow range of 60% and 72%. To maintain N balance, 77 mg N/kg/day was needed if the protein source was either of animal origin or a mixture of animal and vegetable protein, 93 mg/day of the source was a mixture of vegetable proteins and over 110 mg/day if a single vegetable protein was ingested (20). It is important to recognise that the differences between these figures is less than what would have been predicted on the basis of their known aminoacid content or BV obtained at low levels of intake, - a finding which emphasises the need for studies to be done directly on man, with diets which are habitually consumed.

It would be obviously be impossible to evaluate all combinations and permutations of even the major food protein sources and varying proportions of each of the components in mixtures which population groups consume the world over, let alone the minor sources which have some nutritional significance. It may however be possible to group the sources into major categories - those of animal origin such as milk, meat, fish and eggs - those of vegetable origin: - cereals - wheat, rice and barley, millets - maize, sorghum and finger millet, legumes, and oil seeds, look at the relative amounts of these which go into dietaries of large segments of the population in a broad way and determine the efficiency of utilisation of their protein as maesured by N balance at several levels close to requirement. This procedure will also take into account the valid criticism that biological value of a protein is not identical with its nutritive value since while the former is influenced solely by its aminoacid content and profile, the latter is influenced by additional factors outside those related to protein alone - a situation which from the practical view point is more realistic.

A recent committee of the National Academy of Sciences has in fact used this approach to formulate RDA for proteins (21). The committee has assumed that protein in diets of the type consumed in USA, are utilised to the extent of 75%. The committee states that “Further quality correlations are not required, except possible for young children who are subsisting almost exclusively on diets composed largely of cereal grains and root crops, an unlikely situation in the United States”. This approach has much to commend it and is perhaps at present the most realistic approach to the practical problem of recommending intakes which meet protein requirements. This would call for a new approach and each country would be required to carry out studies relevant to their food sources, the composition of habitual diets as well as cooking and eating practices. It would involve a considerable number of diet combinations to be evaluated.

N balance studies are time consuming, tedious and expensive. Development of methods which can rapidly evaluate N balances would therefore constitute an advance over existing methods. An adaptation of the conventional balance technique in which three or four levels of protein in the region of requirement are used, and small increments or decrements made either daily or once in two days is employed has been reported to yield data on BV for egg, spray dried whole milk and casein protein which are not statistically different from those obtained with observations made over longer duration (22). Values however tended to be slightly higher in the short term assay and is likely to be due to incomplete adaptation to rapidly changing protein levels. Wider application of this method may be expected to establish its usefulness and validity.

7. Summary

Several methods are currently in use for evaluating protein quality. Some are based upon chemical methods while others are biological assays. The biological value of a protein derived from nitrogen balance data, is at present considered as the best available measure of protein quality. The concept of a standard protein with an EAA profile that matches human needs, against which the quality of all proteins can be judged has the advantage of simplicity and ease of application, but its use to predict quality would be valid and appropriate only when protein digestion, and availability of aminoacid, both for absorption and utilisation are not critical.

Biological value measures the proportion of absorbed nitrogen which is retained and presumably utilised for protein synthesis and therefore reflects true protein quality. The practical application of data on BV which is determined by conventional methods, for purposes of correcting for differences in quality has limitations, since what is measured is maximal potential of quality and not a true estimate of quality at requirement level. It would be more appropriate to evaluate protein quality of diets as consumed by N balance technique in humans. Such a procedure calls for neither the definition and use of a standard, nor does it make use of the calculated biological value of the ingested protein based on an unrealistic intake level. It will also take care of the valid criticism that the biological value of a protein and its nutritive value are not identical.

To evaluate the large number of representative diets consumed by different population groups all over the world, never short term N balance techniques need to be developed and their validity established.

1. Block R.J. and Mitchell H.H., Nutr. Abs & Rev 16, 249, 1946

2. Report of the FAO Committee, FAO Nutritional Studies No. 16, 1957

3. Joint FAO/WHO Expert Group on Protein Requirements FAO Nut. meeting Rep. Series No. 37, W.H.O. TRS 301, 1965

4. Joint FAO/WHO Expert Committee, WHO TRS 522, 1973

5. Carpenter K.J., Nutr. Abst and Rev 43, 423, 1973

6. Harper, A.E. and Benevenga, N.J. in ‘Proteins as Human Food’ Ed: Lawrie, R.A. Av. Pub. 1970
Harper, A.E. in ‘Improvements of protein nutriture’, Committee on aminoacids, NAS, 1974
Scrimshaw, N.S., Brassani, R., Behar, M. and Viteri, F., J. Nutr., 66, 485 1958
Gopalan, C. Am.J.Clin.Nutr., 23, 35, 1970

7. Barnes, R.H., Bates, M.J. and Maack, J.E., J. Nutr., 32, 535, 1946

8. Bressani, R. in ‘Evaluation of Proteins for Humans’ loc.cit.

9. Inoue et al., Nutr.Rep.Inter., 10, 201, 1974

10. Young, V.R. et al., J. Nutr., 103, 1164, 1973

11. Inoue et al., 1973 loc cit

12. Said, A.K. and Hegsted, D.M., J. Nutr., 99, 474, 1969

13. Woodham, A.A. in ‘Nutritional Improvement of foods and feeds’, 1977, loc cit

14. Inoue et al., J. Nutr., 103, 1673, 1973

15. Garza, C.G., Scrimshaw, N.S. and Young, V.R., Am.J.Clin.Nutr., 29, 280, 1976

16. Calloway, D.H., J. Nutr., 105, 914, 1975

17. Leverton, R.M. and Gram, M.R., J. Nutr., 39, 57, 1949

18. I. Barja et al., Am.J.Clin.Nutr., 25, 506, 1972

19. Wu,H. and Wu D.Y., Proc.Soc.Exp.Biol.Med., 74, 78, 1950

20. Joint FAO/WHO Expert Committee W.H.O TRS 522, 1973

21. Recommended Dietary Allowances Nat.Res.Coun.Nat.Acad.Sci. Washington, 1980

22. Navarette, D.A., Nutr.Rep. Inter., 16, 695, 1977
R. Bessani et al., J.Fd.Sci., 44, 1136, 1979

Top of Page