Previous PageTable of ContentsNext Page

Observations on the Approach to the Problem of Protein Requirements

J.C. Waterlow
Department of Human Nutrition
London School of Hygiene and
Tropical Medicine

1. The question of protein quality, scoring, etc., which has received much attention from previous committees, is of less practical importance and should receive a lower priority than the question of the requirement in terms of reference protein. With most natural mixed diets, the range of variation in quality is not very large; even in extreme cases, e.g. cassava diets, I doubt whether there is much disagreement about the measures which should be taken to improve the quality. On the other hand, current ideas about protein needs range from 0.5 to more than 0.7 g reference protein per kg - a variation of over 40%. This range of uncertainty is obviously unacceptably large from a practical point of view. Therefore first priority should be given to reconsideration of requirements in terms of reference protein.

2. Recently several authorities have taken the line that the factorial method of estimating requirements should be abandoned. A recent UK committee (UK Department of Health and Social Security 1969) based its recommended allowances on what would be acceptable in the context of existing British dietary habits. Pirie (1969) in his book on Food Resources, says that in 'Britain, France and the USA... protein is supplying 10-12% of total calories. This may be taken as a reasonable value to aim at in all other countries'. Tremolières (1970) writes: 'Le pourcentage de calories protéiques (11 à 13%) apparaît comme l'index le plus fixe du comportement alimentaire des hommes; il est recommandé de prendre cet index en considération dans la détermination des taux calorico-azotés à recommander'. And again: 'Le besoin peut être considéré comme ce qui est jugé 'décent' d'être dans une certaine societé à un certain moment'.

Certainly one would not wish to deprive others of what we ourselves enjoy. However, it is not very satisfactory to base recommendations for the whole world on observations made in one type of society, as seems to follow from Pirie's proposal. In the case of infants this could introduce a serious error: in Western societies it is usual for the one year old child to receive about 3 g protein per kg daily, but there is abundant evidence that this is about twice the real requirement for health and growth at that age, unless, as does not occur in Western societies, there is a tremendously heavy infective load.

3. I suggest that the best solution is to return to the older concept of a physiological minimum, representing the level below which health certainly cannot be maintained. Between this and the levels quoted as satisfactory in the previous paragraph lies what was aptly called by an earlier UK Committee the 'area of ignorance' (UK Ministry of Health, 1964). Whereas the minimum is and must be based on physiological considerations, the point at which recommended allowances are set may reasonably be chosen on other, non-physiological, grounds, e.g. local food habits.

In the 1965 WHO/FAO Expert Group I took the line that only one level of requirements should be specified - either one gets enough or not enough. I now think that that was an oversimplification, which has not proved useful, for the reasons to be given below.

Agruments about different levels of requirements or recommended allowances tend to be at cross purposes because they confuse things which are different in kind. Terroine made the distinction quite clear with his terms 'besoin physiologique' and 'besoin hygiénique'. Since then we have tended to obscure the difference, first by omitting the idea of 'minimum' requirement, and then by taking the recommended allowance as the requirement plus various factors of safety.

4. I think that there are advantages in returning to the old distinction.
  1. It is of real practical importance to know the level of intake below which health cannot be maintained. This level presumably depends, as we have always supposed, on the obligatory loss of N, plus the amounts needed for growth, etc. To these should be added, where appropriate, losses imposed by infections.
  2. Once the minimum is established, it makes sense to ask the question: Does a level of intake above the minimum promote better health, efficiency, etc? Tremolières aptly quotes Atwater: 'A man may live and work and maintain bodily equilibrium on either a higher or lower nitrogen level or energy level. 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'.

On the approach adopted by the previous committee, this question could hardly even be posed, and it is significant that since 1965 there seems to have been less research done on this subject.

5. The first priority, therefore, is to establish more closely, if we can, the physiological minimum. I see no need to depart from the conventional approach of dividing this into three parts: the requirements to cover obligatory losses, growth and variability.

Since 1965 there has been a good deal of new research aimed at getting more accurate estimates of the various elements of the obligatory loss. Much of this was in fact inspired by the 1965 committee. Nevertheless, there are still some differences in the estimates accepted by various groups, as shown by the divergence between the recent US and UK reports (Table 1). In the attached paper, Payne reconsiders the magnitude of the elements of the obligatory loss, and arrives at an average figure for adult men of 0.36 g reference protein per kg - almost identical with the original estimate of the 1955 FAO Committee, described by Pirie (1969) as 'ungenerous'.

There is probably less divergence of view about the N requirements for growth.

The question of individual variation is also considered by Payne. He concludes that the range may well be smaller than has been thought but, like the 1965 Committee, recommends the addition of 20% as a safety factor to cover the needs of a group. This addition brings the group minimum for adult man to 0.43 g N per kg per day, and somewhat less for adult women. This is equivalent to 28 g reference protein, or 41.5 g mixed protein (NPU 67) for a 'reference' man weighing 65 kg.

6. It may be objected that since almost all diets in fact supply more protein than this, a figure of this kind serves no practical purpose, even if it does correctly reflect the minimum physiological requirements. Against this I would agrue:
  1. It is important that in times of famine or disaster we should know exactly where we stand with regard to the minimum needed to maintain life.
  2. Some dietary surveys which I believe to be accurate, e.g. in Jamaica, have shown that even in a country where overall supplies seem to be adequate, a proportion of individuals get less than this minimum figure. It may well be that the survey week happened to be a 'thin' week for these people; however, I believe that in such a country there are many people living dangerously close to the margin. Presumably they compensate for this by reduction in growth during childhood and perhaps by reduction in tissue mass in adult life.
7. Even if the concept of the minimum is accepted as useful, and the figure arrived at accepted as more or less correct, the problem remains of making the step from that to the recommended allowance (RA). The purpose of a recommended allowance is to provide a yardstick for diagnosis of the nutritional situation, and hence a basis for action. As Payne points out, the diagnosis will be insensitive if the recommendations are set either too high or too low. I accept that for practical reasons the RA or target must be substantially higher than the minimum, even at the level of the individual (the question of recommendations for countries, involving problems of distribution within the community and the family, has to be considered as a further step). What principles can we adopt for making the advance from minimum to target?
  1. Like previous committees, we could take the minimum and add some kind of safety factor. One rough and ready way of doing this might be as follows: in infants, other things being equal, growth provides a criterion of the adequacy of protein intake, and therefore for this age group we have better information than for any other. At one year, Payne's estimate of obligatory loss, excluding effects of infection, amounts to a mean of 0.9 g reference protein per kg per day. In fact, Holt's observations on breast-fed babies and ours on children who had recovered from malnutrition suggest a 'practical' requirement of 1.25 g milk protein per kg per day, below which no paediatrician would like to go.

If the NPU is taken as 90, both for breast and cows' milk, this reduces to 1.13 g reference protein per kg per day, i.e. 25% above the mean figure derived by the factorial method. This 25% represents the 'area of ignorance'. If the same correction were applied to Payne's figures for average obligatory losses in the adult man, it would come to 0.44 g reference protein, or 0.66 g mixed protein per kg per day. To use the same factor seems not unreasonable since even in the one year old child maintenance accounts for 6/7 of the requirement. This estimate is a little lower than that reached by the 1965 Committee. In communities where there is a heavy incidence of infection a further addition would have to be made, particularly for children.

  1. Secondly, one might decide that whatever may be the minimum, the RA must be based on different considerations - on what people actually eat or find acceptable, decent or 'convenable' (Tremolières). The further step might be taken, following Tremolières, of abandoning the idea of one reference man and one scale of recommended allowances. This would mean adjusting recommendations to particular circumstances. ('Le caractère régional parait indispensable' - Tremolières). This would make it difficult for FAO and WHO to provide practical guide-lines for general use. It also implies that in regions where the average body size is small, this should be accepted as the norm, and that the purpose of recommended allowances is merely to maintain the status quo. Such an approach would not be universally acceptable.
  2. The ideal method is to base recommendations on more positive information.
  1. On the range of individual variation
  2. On the extent of N losses which occur in ordinary life, over and above the conventional minimum
  3. On whether or not increasing the protein intake from say 50 to 60, 70 or 80 g per day produces any significant improvement in health and function.

This approach is not practicable at present because we do not have the information. These, however, are the points on which further research is needed, difficult though it may be.

  1. As I see it, the Expert Group has two functions:

To produce some guidance for world-wide use, the best it can in the light of existing knowledge;

To summarize that knowledge, and indicate the needs for future research.

The 1965 Report, whatever its faults, did succeed to some extent in this second objective, although it effectively blocked out one area of research - the effect, if any, of protein intakes above the so-called requirement.

My object in this paper is to emphasize that if, in revulsion against the factorial method, we adopt a purely pragmatic approach, the needs for research will be obscured, and we shall not only have made little progress but will hinder progress in the future.


Estimates of minimal nitrogen requirement for maintenance in adult males

Weight kg657065
BMR kcals/day150017501600
N loss in urine g/day3.00(1)3.50(1)3.20(1)
N loss in faeces g/day1.300.700.91
N loss by skin g/day1.301.400.13
Allowance for stress0.5600
Total g N/day6.165.604.24
Percentage addition for individual variation or margin of safety203020
Total g N/day7.407.305.05
As reference protein g/day46.545.531.5
As reference protein g/kg/day0.710.650.49

Note: (1) Calculated as 2 mg N per basal kcal

(2) Compiled by a subgroup of the Committee, but not adopted as a basis for recommended allowances


Food and Agriculture Organization. Protein Requirements, FAO Nutritional Studies (1955) No. 16, FAO, Rome

Pirie, N.W., Food Resources, Pelican Books, London (1969)

Tremolières, J. Papers submitted to WHO/FAO preliminary group (1970)

U.K. Recommended Intakes of Nutrients for the United Kingdom. Department of (1969) Health and Social Security, Reports on Public Health and Medical Subjects No. 120, H.M.S.O. London

U.K. Requirements of Man for Protein; Ministry of Health: Reports on Public Health (1964) and Medical Subjects, No. 111. H.M.S.O. London

U.S.A. Recommended Dietary Allowances. 7th revised edition 1968. A Report of the (1968) Food and Nutrition Board, National Research Council. Publication No. 1694. National Academy of Sciences Washington, D.C.

WHO/FAO Protein Requirements. Report of an Expert Group. World Health Organization (1965) Technical Reports Series No. 301. Geneva

Previous Page Top of Page Next Page