Contents - Previous - Next

This is the old United Nations University website. Visit the new site at

Human nutrition

Current concepts concerning indispensable amino acid needs in adults and their implications for international nutrition planning

Vernon R. Young and Peter L. Pellett



Using a newly proposed amino acid scoring pattern, an assessment is made of the capacity of diets in developed and developing countries to satisfy human amino acid needs. The new adult pattern differs from that proposed by FAO/WHO/UNU (1985) for the adult but is similar to that for two- to five-year-old children. FAO food availability data were used to evaluate the dietary patterns in selected countries, including those of Africa, the Far East, and the Near East. Lysine appears to be the limiting amino acid in diets heavily dependent on cereals, characteristic of those in poor countries; the analysis reveals that for these countries the mean lysine supply is below the newly estimated requirements, implying that adults could be at risk of an inadequate intake, especially when total protein availability is low. Further, this analysis reveals that when animal foods contribute about one-third of the total protein supply or when animal foods together with pulses and soybeans account for about 40% of the total protein supply, the lysine requirement for the population will be met. For the US diet, the high proportion of the total protein contributed by animal foods, about 68%, provides a lysine intake that is well above the estimated requirement and that is almost constant across all age and sex groups. This reassessment leads to the recommendation that, especially in countries whose diet is mainly dependent on cereals, complementation to improve dietary protein quality should remain a consideration in the design and implementation of food, nutrition, and agricultural programmes and policies.


National and international estimates of adult human amino acid requirements have not changed greatly over the years [1-3]. Although the 1985 FAO/WHO/UNU report [3] drew its conclusions from the best published data available in 1981, the validity of the low current recommendations has been questioned [4]. Earlier international amino acid scoring patterns [5] derived essentially from consideration of infant and young-child needs, although adult needs were considered and tabulated. However, the FAO/WHO/UNU report proposed age-specific scoring patterns, and thus the extremely low estimates for adults have implications for nutrition policy that new data indicate to be quite inappropriate for the nutritional health of specific populations. For example, from the current international estimates alone, even poor cereal-based diets would provide indispensable (essential) amino acids well in excess (two- to threefold) of adult requirements. In consequence, such diets would be considered to be of high protein quality.

Table 1. Comparison of the FAO/WHO/UNU [3] adult requirement pattern of four critical amino acids with the contents in wheat, beef. and selected low-cost diets and those for developed and developing regions


FAO/WHO/UNU adult requirement pattern



Developed regions

Developing regions

Low-cost diets

Lysine 16 28 89 65 47 45 ▒ 9
SAA 17 40 40 36 37 31 ▒ 5
Threonine 9 29 46 40 40 34 ▒ 6
Tryptophan 5 11 12 12 12 10 ▒ 2

Data from Pellet and Young [6] except as otherwise indicated.

All values are expressed as milligrams amino acid per gram protein.

a. Mean ▒ SD for 19 low-cost diets tabulated by FAO/WHO/UNU [3]; adult requirement pattern from the same source.

Adult essential amino acid requirements, expressed per gram of protein, given in the FAO/WHO/UNU report [3] are shown in table 1 (see Table 1. Comparison of the FAO/WHO/UNU [31 adult requirement pattern of four critical amino acids with the contents in wheat, beef. and selected low-cost diets and those for developed and developing regions), and are compared with the amino acid composition of wheat, beef, and mixed diets reasonably typical of developed and developing regions. It can be seen that selected practical, low-cost diets appear to supply amino acids in considerable excess of adult needs. Moreover, a single cereal staple such as wheat would do so more than adequately when consumed at intake levels just sufficient to meet the needs for total nitrogen (protein). It is not surprising, therefore, that there is confusion in formulating nutrition policy on the basis of these data. For example, priorities given to cereal development in agricultural research could be significantly affected. If indispensable amino acid needs in the adult can be greatly exceeded where a diet is predominantly based on a single cereal, why should any priority be given to legume and animal protein development programme or even considerations of protein and amino acid complementation? Why not focus only on high-yield cereal crops and cheap, predominantly cereal-base' foods?

Amino acid scoring patterns

Amino acid data has been widely used to evaluate the protein values of foods and diets since the amino acid composition of egg was introduced as a standard [7 (see TABLE 2. Some national and international recommended amino acid scoring patterns). Because the relatively high levels of indispensable amino acids in egg proteins undervalued the. nutritional quality of many important food proteins values for human requirements were used as the basis for several subsequent international scoring system [3, 5, 8]. Although the 1974 National Academy of Sciences-National Research Council (NAS-NRC) [2, 9] scoring pattern was stated to be based on human needs, in practice the pattern was derived from data on egg and milk proteins. That is why the proposed scoring system [2] differed significantly for the sulfur amino acids from that of the 1973 FAO/WHO [5] pattern, although both apparently used the same human-requirement data in their derivations. The 1965 FAD/WHO [10] scoring system was based on the proportion of essential amino acids to total nitrogen in egg rather than on human needs. Because many objections were raised to this procedure, however, it was never widely used.

TABLE 2. Some national and international recommended amino acid scoring patterns


















(< 1 yr)

(2-5 yrs)

(6-12 yrs

(> 13 yrs)

Histidine - 14 17 26 14 19 -
Isoleucine 42 40 42 46 28 28 13
Leucine 48 70 70 93 66 44 19
Lysine 42 55 51 66 58 44 16
Total SAA 42 35 26 42 25 22 17
Total AAA 56 60 73 72 63 22 19
Threonine 28 40 35 43 34 28 9
Tryptophan 14 10 11 17 11 9 5
Valine41   50 48 55 35 25 13
Total (w/o histidine) 314 360 356   434 320 222 111

Values are milligrams of amino acid per gram of protein (N x 6.25).
SAA = methionine + cyste(i)ne; AAA = phenylalanine + tyrosine.

A common feature of all the scoring systems before the most recent 1985 one [3] was that a single reference pattern was used for all ages, despite data generally showing that infants needed some 35% or more of total amino acids in the form of indispensable amino acids while adults apparently needed only 15% or less [5]. Clearly, if the low percentage required by adults was adopted for scoring purposes, it would greatly overestimate the values of a protein for meeting the needs of children. In contrast, adoption of the children's pattern would appear to underestimate the value of a protein for adults.

In 1973, the available data for the indispensable amino acid requirements of infants, children, and adults (mg/kg/day) were individually tabulated for the first time [5]. Amino acid scoring patterns (mg/g protein or mg/g N) were then derived for each age group by dividing the amino acid requirements by the daily protein requirements. The pattern proposed for practical application, however, was developed by interpolating from only infant and schoolchild values. In practice, this gave an apparent extra margin of safety for the estimation of adult protein needs and in the assessment of nutritional quality of adult diets.

New data [11] for young children were available for the 1985 FAO/WHO/UNU report [3], and, as with the earlier procedure [5], amino acid requirements for various age groups were summarized from the published literature. They were then divided by the safe levels of intake for total protein for the same age categories to give the amino acid scoring pattern. Most of the requirement values used were essentially unchanged from values available to the earlier group. Final values [3] were generally lower than those tabulated in 1973 [5]. This was because the safe level of protein intakes recommended for adults and school-children had been increased, and therefore the indispensable amino acids required per unit weight gram of protein were correspondingly reduced.

More than 30 years have passed since the 1957 FAO committee on protein requirements [8] considered protein, or amino acid, scoring. Provided that an adjustment is made for the digestibility of ingested proteins [12, 13], the statement made by this earlier group concerning the role of scoring in relation to an amino acid requirement pattern remains valid.

The concept of a desirable pattern of essential amino acid has one great advantage. By comparison with such a pattern, data on the amino acid content of food combinations can be appraised in a wide range of situations in terms of possible defects of the diet and of methods of improving it. Comparisons, no doubt rough and approximate, can be made directly by the use of tables showing the contents of foods in essential amino acids, regardless of the proportions in which individual foods are included in the diet. On the other hand, even when the biological value of each dietary component is known, a deduction cannot be made about the biological value of the diet as a whole. [8]


Relationship to animal bioassay data

As the use of amino acid scoring systems has progressed over the years, the general expectation has developed that not only should the score be able to predict the potential nutritional value of a food or diet for humans, but it should (with or without digestibility considerations) also correlate directly with the results of animal assays, such as net protein utilization (NPU). Many such correlations were described [14-17] and, in effect, they became a test of validity of the various proposed scoring systems. The ability to predict NPU values and the correct limiting amino acids improved as the systems were developed further [15, 17]. It remains questionable, however. whether such correlations with animal assays are valid or even a desirable attribute of scoring systems when intended for application in human nutrition.

With hindsight, our view is that the systems developed to predict the nutritional value of protein sources for humans should not necessarily be expected to agree with values obtained in growing rats. The appropriate standard to assess dietary quality is the human amino acid requirement pattern, and. for animal bioassays to be useful, they should be designed to give predictions in line with those based on such requirements, rather than the reverse.

Current views on adult indispensable amino acid needs

Before considering further the relationship between human amino acid requirements and the nutritional value of specific food proteins and of complex diets, we present a synopsis of the arguments and concerns that we have with respect to estimates of such requirements in human adults [4], as proposed by various national [2] and international [3] groups. These various estimates were derived largely from results of nitrogen balance studies in healthy individuals (table 3). As we have discussed [4], however, significant problems are associated with deriving and interpreting data from such studies, particularly as applied previously to determine amino acid requirements in adults. Furthermore, from a series of experiments on the kinetics of amino acid metabolism in which rates of amino acid oxidation were estimated at various intake levels, it has been concluded that the current requirement estimations are far too low for leucine [18] valine [19], lysine [20], and threonine [21] if not for all of the indispensable amino acids except perhaps the sulfur amino acids (V. R. Young et al., unpublished data). We used a predictive approach to estimate the requirements for specific indispensable amino acids [4], based on considerations of obligatory nitrogen losses, and compared the results with those derived from (13)C-tracer studies (table 4). Reasonably good agreement exists between the estimated requirement levels as judged by the two approaches, further indicating that the requirement figures based on nitrogen balance accepted previously [2, [3] are too low.

TABLE 3. Indispensable amino acid requirements (mg/kg day) in adults as proposed by FAO/WHO/UNU [3] and NAS-NRC [2]

Indispensable amino acid

FAO/WHO/UNU (1985)

NAS-NRC (1974)















Total SAA




Total AAA















11 .2

SAA = methionine + cyste(i)ne; AAA = phenylalanine + tyrosine.

TABLE 4. Comparison of requirements for amino acids (mg/kg/day) in the adult, as estimated by venous methods


Method of estimation

N balance

Factorial (obligatory N)

13C-tracer studies








~ 30




~ 15




~ 20




~ 13

Source: Young et al [4]

A number of additional issues relating to interpretation and experimental design emerge from our tracer studies, and they have been raised usefully [22. 23]. These include concerns for (i) the effect of diurnal fluctuations in metabolism on the accuracy of determining daily amino acid oxidation, (ii) effects of a relatively high dietary nonessential nitrogen intake on amino acid metabolism, (iii) whether dietary balance among amino acids affects requirements, and (iv) the criteria used to assess nutritional adequacy All are important and are discussed in detail else where [4]. We feel that these issues are unlikely to affect significantly the magnitude of the difference between the estimates we have made, and hence the conclusions drawn from them, compared to the low values currently accepted by FAO/WHO/UNU [3].

We have therefore derived a revised adult amine acid requirement pattern (table 5). It is quite different from that proposed for adults but is similar to that for pre-school children. We further propose that this pattern can be applied to all age groups after infancy. A similar recommendation that the pattern for preschool children be used to score the protein value of foods and diets for consumption at all ages was made by a recent FAD/WHO expert consultation on protein quality evaluation [13].

TABLE 5. Recommended amino acid scoring pattern for individuals from preschool age through adulthood


Amount (mg/g protein)







Total SAA


Total aromatic








Derived from Young et al. [4] rounding to the nearest value of 5.

Our revised amino acid requirement pattern has not been universally accepted [24, 25]. However, our conclusions remain essentially the same, whether based on our pattern or that suggested by FAD/WHO [13] for application across all age groups above the age of one year.

International dietary planning implications

The new estimates of amino acid requirements for adults are, on average, higher by a factor of 2.5 than those proposed by FAO/WHO/UNU [3]. At first sight, these propositions appear to have dramatic implications for evaluation of the adequacy of all diets and possibly to re-create the global "protein problem." The practical implications, however, while significant for diets and food as well as agricultural policies in some developing countries, are less for the mixed diets of other developing countries and negligible for the developed regions. Problems may arise, however, whenever traditional diets cannot be consumed for economic, social, or environmental reasons.


Regional and socio-economic patterns of amino acid supply

When comparisons are made among diets in various countries, it has been demonstrated that as incomes increase, not only is more food available, but major changes occur in the pattern of foods selected, including significant increases in the availability of animal protein foods [26]. While data from food balance sheets are recognized as being limited, they can establish broad diet-health relationships, and the concentration of the protein and energy value of diets. Thus, table 6 (see TABLE 6. Per capita dietary composition in selected world regions) gives food composition data for selected diets in developed and developing regions of the world, based on the Fifth World Food Survey [27]. In developed regions the percentage of contribution of animal products relative to both protein and food-energy supplies is high. In contrast, the diets in developing regions depend much more on cereals for both protein and dietary energy.

TABLE 6. Per capita dietary composition in selected world regions


Developedmarket economies


Far East

Near East

Latin America

Least developed countries

Total food energy (kcal) 3,390 2,260 2,160 2,840 2,620 2,070
Total protein (g) 99 56 51 78 67 53
Animal protein (g) 56 11 8 19 28 10
Share of total DES (%)
Vegetable products 68.3 93.5 94.2 89.3 82.9 93.3
Animal products 31.7 6.5 5.8 10.7 17.1 6.7
Food groups            
Cereals 26.4 46.7 67.4 58.2 40.2 65.0
Pulse nuts seeds 2.4 7.0 5.9 3.8 4.9 6.5
Roots tubers 3.7 19.8 3.0 1.8 5.6 8.3
Meat 15.3 2.8 1.2 4.0 8.1 2.7
Milk 8.5 2.0 2.5 3.7 5.5 2.5
Eggs 1.6 0.2 0.3 0.4 0.8 0.2
Fish 1.6 0.8 0.7 0.3 0.7 0.5

Adapted from FAO [27].
a. DES = dietary energy supply. Values expressed as percentage of total energy supply. Only food groups providing significant levels of dietary protein are shown; thus, totals are somewhat less than 100%.

These differences in protein sources significantly affect the pattern of amino acids in the diet (see TABLE 7 Amino acid pattern (mg/g protein) for diets in selected regions and compared to human amino acid requirement patterns). The patterns presented in this table were derived by standard calculation procedures using published amino acid composition data [27, 28]. However, because such composition data (see TABLE 8. Amino acid composition (mg/g protein) of some major food protein sources) differ widely among food groups but are fairly constant within a given food group, it is possible to use a simplified approach to calculate dietary amino acid patterns [30]. For example, we found that the amino acid composition (mg/g protein) of the United States diet was almost identical, whether calculated from food balance sheet data using 144 food items or from a simplified version using only 20 representative food items [30].

TABLE 7 Amino acid pattern (mg/g protein) for diets in selected regions and compared to human amino acid requirement patterns



Requirement pattern

Developed countries


Far East

Near East

Latin America

A [3]

B [4]

C [13]

Isoleucine 48 42 47 44 47 13 35 28
Leucine 77 75 81 73 82 19 65 66
Lysine 65 47 46 41 56 16 50 58
Total SAA 36 34 40 36 34 17 25 25
Total AAA 81 72 86 79 79 19 65 63
Threonine 40 35 39 34 39 9 25 34
Tryptophan 12 12 12 12 11 5 10 11
Valine 53 48 57 49 5 13 35 35

Generated from FAO [27] data, as summarized in table 6 using amino acid composition data from FAO [28] and USDA [29]

TABLE 8. Amino acid composition (mg/g protein) of some major food protein sources











Isoleucine 40 42 46 41 42 52 60 47 63
Leucine 63 82 126 96 76 82 98 85 88
Lysine 26 36 38 34 72 87 79 100 70
Total SAA 35 36 32 48 19 37 34 45 56
Total AAA 81 80 58 81 77 75 96 90 98
Threonine 27 33 40 39 39 44 45 52 49
Tryptophan 11 13 6 19 10 12 14 11 16
Valine 43 58 51 55 46 56 67 52 72
Total indispensable amino acids 326 380 390 413 381 445 493 482 510

Source: FAO [28] and USDA [29]

The values for the United States were also much the same as those shown for the industrialized countries as a group. Furthermore, we observed that when USDA data [31] are used to calculate amino acid composition, not only are the values again similar, but they remain almost constant across some 22 different age and sex groupings despite wide variations in the proportions of protein foods consumed [30]. When the amino acid compositions of the major food groups (see TABLE 9. Indispensable amino acid composition mg/g protein) of selected food groups from Massachusetts nutrient data bank) are examined, this conclusion is not surprising. In the United States diet the major changes in consumption pattern, as affected by age and sex, involve differences in the proportions of animal protein. This is in marked contrast to the diets of the poor countries, where foods of animal origin may supply less than 10% of the protein, and cereals more than 60%. This simplified calculation was also suitable for evaluating diets of developing countries.

TABLE 9. Indispensable amino acid composition mg/g protein) of selected food groups from Massachusetts nutrient data bank


No. of items









Animal foods 1,233 85 ▒ 9 38 44 12 50 81 78 55
Legumes 85 65 ▒ 7 25 40 12 51 79 82 52
Cereals 260 29 ▒ 7 36 33 10 43 78 80 49
Others 887 48 ▒ 14 26 36 12 40 62 67 48

Amino acid composition data originate from FAO [28] and USDA [29].
a. Mean ▒ SD.

When the amino acid composition of regional diets is compared with the international requirement values for all regions, including Africa and the Near and Far East, the pattern exceeds the FAO/WHO/UNU 3] requirements for adults (column A, table 7) by a wide margin. When comparison is made with our newly proposed pattern (column B, table 7), or with the FAO/WHO/UNU child pattern (column C, table 7), the food availability amino acid pattern for developed countries and for Latin America is still above the requirement pattern for all amino acids. This is not the case for Africa, the Near East, or the Far East, where the lysine level is lower. Furthermore, from this analysis, the major difference between the patterns of diets in developing and developed countries is in their lysine content. This is primarily due to the much lower levels of lysine in cereal products than in animal foods.


Amino acid supply in selected countries

The data in table 10 (see TABLE 10. Food energy, protein, and indispensable amino acids in diets of 20 selected countries compared with GNP and the under-five mortality rate (MR)) allow comparison of the amino acid composition of diets with dietary patterns for 20 selected countries. These 20 countries were chosen from a stratified UNICEF [32] tabulation of 130 countries, with five countries selected at random (except for the United States) from each of four groups classified on the basis of their under-five-year mortality rates. This also, in effect, classified them on the basis of economic status, since mortality rates and gross national product (GNP) show a high correlation [33].

TABLE 10. Food energy, protein, and indispensable amino acids in diets of 20 selected countries compared with GNP and the under-five mortality rate (MR)


Under- five MR


Dietary energy (kcal day)

Protein (g/day)

AnimalProtein (g/day)

Percentage of total protein

Indispensable amino acid

Lys ma/day




Nut & oilseed





Sierra Leone 302 3 1 0 1.938 42 10 24 52 11 2 54 4 1 42 1 2 2.270
Rwanda 214 280 2,274 54 4 6 12 48 1 52 29 34 11 2.810
Yemen Arab Rep. 210 550 2,475 77 17 22 63 10 < 1 44 34 35 12 3.390
Bangladesh 1 30 1,837 40 5 13 79 2 _ 44 36 34 12 1,760  
Nigeria 182 730 2,178 55 11 19 50 10 5 42 36 37, 12 2,480
Pakistan 74 380 2, 1 80 59 1 3 24 66 5 I 40 34 33 11 2.240
Ghana 153 350 1,769 41 13 32 37 10 5 48 34 35 11 1 ,970
Indonesia 126 540 2,372 49 6 12 67 11 5 46 42 41 44 2.250
Tunisia 110 1 ,270 2,763 76 17 22 6 1 16 2 44 34 32 12 3,340
Guatemala 1 09 1, 1 60 2,138 52 1 3 24 58 12 < 1 46 33 38 10 2,440
Ecuador 1,150 2,114 50 26 51 31 6 1 62 37 40 12 3,100 3,100
Syna 71 1,620 3,005 85 22 26 51 7 4 46 34 33 12 3.910
Paraguay 64 1,240 2,839 50 33 4 i 27 18 1 61 34 37 11 4.800
Thailand 55 860 2,330 47 1 2 26 59 3 2 5 1 44 43 11 2,4(X)
Yugoslavia 31 2,110 3,550 103 39 37 51 4 < 1 52 35 35 12 5.360
Portuga l 22 1,970 3,204 85 35 41 43 3 < 1 59 36 38 11 5.020
Hungary 21 2,100 3,484 95 49 52 35 < 1 < 1 65 36 38 12 6.180
Greece 3,770 3,668 108 52 49 35 4 2 60 35 36 12 6.480  
Austria 13 9,140 3,575 91 57 63 24 1 1 68 36 39 12 6.190
United States 13 15.390 3.641 106 72 68 18 1 1 3 70 40 12 7 420

Mortality rates (MR) for children under five years of age and infants and GNP taken from UNICEF [32]. Energy and protein data from FAO [35].
Pulse protein percentages were calculated by including soybeans from the nut and oilseed category.

Between 35 and 70 food items were used in these calculations, depending on dietary complexity, as tabulated in the FAO food balance sheets [34, 35]. Also, the broad relationships, using simple linear regression analysis (see TABLE 11. Cross-country correlations among per capita GNP, contribution of dietary protein sources, and dietary lysine content for 20 countries), between wealth (GNP) and dietary protein values follow the earlier conclusions of PerissÚ et al. [26]. For example, as income increases, so does total protein consumption and the percentage of that protein derived from animal protein foods.

TABLE 11. Cross-country correlations among per capita GNP, contribution of dietary protein sources, and dietary lysine content for 20 countries




Protein (g/day)

Animal protein (%)

Cereal protein (%)

Pulse protein (%)

Lysine (mg/g) protein)

Kcalb 0.65**            
Protein (g/day)b 0.62** 0.93***          
Animal protein(%)b 0.75*** 0.71*** 0.701***        
Cereal protein (%)b -0.52* -0.44* -0.43 -0 60**      
Pulse protein (%)b -0.32 -0.34 -0.30 -0.51* -0.35    
Lys (mg/g protein)c 0.68*** 0.65** 0.62** 0.87*** -0.77*** -0.17  
Lys (mg/day)c 0.75*** 0.96*** 0.93*** 0.84*** -0.59** -0.31 0.80***

a. UNICEF [32]
b. Based on FAO food balance sheets [35]. Pulse protein includes soybeans from the nuts and oilseeds category.
c. Calculated from FAO food balance sheet data [35], using amino acid composition data from FAO [28] and USDA [29].

Additional correlation coefficients were also calculated using the other important indispensable amino acids. total SAA, threonine and tryptophan. When expressed on a mg/g protein basis, all cross-country correlations were non-significant (P > .05), except for pulse protein percent ages versus total sulfur amino acids, where R=-.47 (P<.05). For indispensable amino acids (mg/day), most of the relationships were significant (P < .01) but not to any greater degree than the relationship with total protein.

*P<.05. **P<.01. ***P<.001.

The percentages of protein supplied by the major food groups are also shown in table 10. In the published food balance sheet tabulations, pulses and nuts and oilseeds are presented as two separate groups. While pulse products are generally high in lysine, nuts and oilseeds are rather more varied and can include high-lysine foods, such as soybeans (about 65 mg/g protein), as well as low- to moderate-lysine products, such as peanuts and coconuts (about 35 mg/g protein). For this reason, for our calculation purposes we added soybeans to the pulse proteins and removed them from the nuts and oilseeds group. The combined percentages of the total protein from animal (AP%), cereal (CP%), and pulse (plus soybean, PP%) protein sources average almost 90% of the total dietary protein for the 146 countries tabulated by FAO in 1984 [35].

Diets vary considerably in the contributions to the total protein supply of the different food groups. In the United States, about 68% comes from animal protein and only 19% from cereals, while for Bangladesh the figures are 13% and 79%, respectively. The dietary pattern for Rwanda, with 48% of its protein from pulses, appears almost unique. The average percentage availability of pulse (including soybean) proteins in relation to total protein for the 146 countries tabulated by FAO [35] was much lower at 6%, with only six countries showing values in excess of 25%.

The conclusions to be drawn from table 10 confirm those based on the regional data presented in table 7: namely, lysine is the amino acid likely to be marginal in relation to our proposed amino acid requirement pattern of 50 mg/g protein, and also in reference to the value of 58 mg/g protein for pre-school children [3]. In addition, as total protein and animal protein availability decrease, so does the dietary lysine concentration. Cereal protein percentage is also a determinant of dietary lysine concentration but in an inverse manner (r = -0.77; P<0.001). In practice, it would appear that lysine needs may become difficult to meet when more than about half of the dietary protein is derived from cereals.

The important role of animal proteins as a source of lysine is further elaborated by the use of multiple regression analysis (see TABLE 12. Multiple regression analysis: Lysine value versus dietary protein sources in 20 countries). It can also be seen, however, that for the prediction of lysine concentration the best equation is:

Lysine (mg/g protein) = 0.57 AP% + 0.33 PP% + 31.3,

and this involves use of two variables, AP% and PP%. From consideration of r2, these two factors appear to explain 86% of the variation in the concentration of dietary lysine. Use of AP% alone is a less precise but still useful predictor, explaining some 76% of the variation in dietary lysine value. Addition of CP% does not improve the predictability (r2 = 0.86) but, in fact, reduces the overall variance ratio (F) value.

TABLE 12. Multiple regression analysis: Lysine value versus dietary protein sources in 20 countries

Dependent variable





Animal protein

Pulse proteina

Cereal protein


F value

Lysine 0.47***     37.5 .87 56.9***
(mg/g 0.57*** 0.33**   31.3 .93 52.5***
protein) 0.51* 0.25 -0.05 36.3 .93 33.1***
Lysine (mg/day) 87.0***     1,588 .97 268.3***
92.0*** 47.9**   1,214 .98 199.1***
93.1*** 64.8*** 22.0 443 .99 235.3***

a. Calculated as percentage of total protein for analysis against lysine measured in milligrams per gram of protein, and as gram per day for lysine measured in milligrams per day.

*P<.05. **P<.01. ***P<.001 (based on analysis of variance).

This analysis suggests that the lysine content of animal, pulse, and cereal proteins is sufficiently different that food group data can be used to predict dietary lysine content. From table 8, the lysine values of the cereals shown range from 26 to 38 mg/g protein, while for animal foods the range is 70 to 100 mg/g protein. In table 9, the mean indispensable amino acid values are tabulated for foods classified as animal foods, pulses, and cereals. The groupings follow standard practice in that animal foods, for example, would include meat, poultry, fish, eggs, and dairy foods. Pulses, however, are defined as various beans and peas together with soybeans from the nuts and oilseeds group. Mean values are also given from a category labelled "other", which includes nuts, oilseeds (except soybeans), fruits, and vegetables. These are less important in assessing protein values since they generally account for below 10% of total dietary protein [35].

Since food balance sheet data [35] allow easy calculation of dietary values for AP% and PP% in individual countries, we took the opportunity to use the multiple regression equation (table 12) to calculate lysine values (mg/g protein) for the dietary patterns in the 146 countries presented by FAO [35]. These estimated values show a wide range, as would be expected. Twenty-three countries (mostly in Europe or North America) exceeded 65 mg/g protein, and three (Burkino Faso, Nepal, and Pakistan) had values of 40 mg/g protein or less. It should be emphasized here that protein quality (score) is not the only determinant of overall protein nutritional value of a diet. The latter is also significantly affected by digestibility and the total quantity of protein. These are also likely to be low in the poorest countries.

These preliminary conclusions are not meant to imply a new manifestation of the protein problem but are a return to the status quo ante existing prior to an explicit statement of the low adult indispensable amino acid needs [3]. We do not believe that poor cereal-based diets can be made acceptable in their protein value by proposing lower values than are likely to be required to meet physiological needs. This conclusion is implicitly shared by the recent FAD/WHO working group on protein quality evaluation [13].


Daily lysine supply and requirement

The total daily lysine intake, or availability, is a composite of total protein (g/day) and its lysine content (mg/g protein). Since the amount of lysine is dependent on the quantity of animal protein present in the diet, significant correlations between total lysine and both total protein and animal protein would be expected. Lysine (in terms of mg/g protein and mg/day) is also highly correlated with wealth, but this is mainly dependent on the relationship with animal protein in the diet. By multiple regression analysis, total daily lysine availability can be predicted from daily availability of animal protein, pulse protein, and cereal protein (r2 = 0.98). For animal protein alone, predictability is still high (r2 = 0 94)

A safe level of lysine intake for a reference individual can be estimated from the desirable lysine concentration (50 mg/g protein) and recommendations for a safe level of protein intake [3]. For a 65-kg male age 18 to 60 years, the protein needs are 49 g daily; for a 55-kg female age 18 to 60 years they are 41 g [3]. An average safe level of lysine for both sexes would thus approximate 2,250 mg daily if the diet were completely digestible. To allow for a 90% digestibility, about 2,500 mg of lysine would be required daily. If the children's requirement pattern for all ages [13] were used, the figure would increase to 2,900 ma.

When food balance sheets for various countries are used to provide availability data, the lysine requirement values used for comparison should preferably be based on average requirement for protein (0.6 g/kg/ day) rather than on safe levels of intake (0.75 g/kg// day) [3]. Average needs for the reference adults of both sexes would thus be 36 g of protein and, hence, about 2,000 mg of lysine daily at 90% digestibility, or 2,100 mg if coarse grain cereals (85% digestibility) were specified.

Of the countries listed in table 10, only Bangladesh and Ghana would have their mean lysine supply below this average requirement for reference adults. Based on the use of the probability approach (see below), these two countries would show probabilities of inadequacy of 0.91 and 0.70 if a coefficient of variation for requirement of 12.5% (that used by FAO/WHO/UNU for protein) were specified. Average intakes of lysine in excess of 2,650 mg/day would be required before probabilities of inadequacy of below 0.025 were reached.

The countries identified as being at risk would differ in some cases from those with dietary lysine patterns below 50 mg/g protein. For example, in Sierra Leone the quality of protein in the average diet, as defined by lysine concentration, appeared to be adequate, but the quantity of total protein was low; in others (e.g.. Tunisia), a low lysine concentration combined with relatively generous intakes of total protein resulted in a higher lysine intake. For other countries (e.g.. Bangladesh and Ghana), the diet appears to be inadequate in terms of both quality and quantity.

For comparison, using the international recommendations [3], only 720 mg of lysine would be required at safe allowances of 45 g of protein daily with 580 mg at the average needs of 36 g of protein/day. All countries, even the poorest, would considerably exceed these values. Many, however, may be marginal when compared to the revised amino acid requirement values proposed above. Furthermore, a greater number would be estimated to be at risk if the children's value of 58 mg lysine/g protein [13] were used rather than our lower requirement estimate of 50 ma.


Utilizable protein

An alternative approach for assessing dietary protein value involves considering both quality and quantity, and this brings in the concept of utilizable protein. Quantity of protein is represented by daily protein availability (g/day) and quality by both digestibility and score. The digestibility was taken as 85% where the diets involved consumption of coarse grain cereals [3], while score was judged in relation to our new requirement pattern. In all cases, the limiting amino acid was lysine. The dietary supply of utilizable protein could then be compared with protein requirements. Again, average needs should be used rather than safe practical allowances. Thus, the average daily protein needs for our selected reference male and female adults would be 36 g of (highly digestible, high-quality) protein.

The probability of inadequacy of an average dietary intake can be estimated from knowledge of average requirement together with its coefficient of variation (CV) [36] For the protein requirement, FAD/WHO/ UNU used a CV of 12.5% [3] The algorithm specified by the NAS-NRC [36] committee was used to estimate the probability of inadequacy. When utilizable protein values are calculated using the data shown in table 10, the probabilities of inadequacy range from 0.023 for average intakes of 45 g/day (safe practical allowance, SPA) to 0.91 for 30 g/day (40 x 0.85 x 0.88) estimated for the average availability in Bangladesh.

Using these criteria, about half of the countries in table 10 might be at risk of inadequacy. For a somewhat larger CV (20%), as much as 50 g/day utilizable protein availability would be required for probabilities of inadequacy to be below the generally accepted 0.025. The value was also calculated for the 146 countries in the FAO [35] food balance sheets. The multiple regression equation was used to estimate lysine value and, hence, score. Although 44 countries had lysine values below 50 mg/g protein, few had utilizable protein supplies at or below the average adult requirements of 36 g/day. Probabilities of inadequacy were for India: 0.50; Liberia: 0.67; Sri Lanka: 0.81; and Zaire: 0.99. An additional 11 countries had average availabilities below 45 g of protein/day (safe level of protein intake) and could be considered marginal in that their probabilities of inadequacy were greater than 0.025. The diets of the remaining countries all provided mean utilizable protein in excess of even the safe intake by virtue of the higher levels of total protein.

Thus, the countries, and the individuals in those countries, most at risk are those with low intakes of diets based heavily on coarse grain cereals, where both the quantity and quality (including digestibility) of the protein available for consumption are likely to be low. The overriding importance of the total quantity of food, as represented by dietary energy supply, therefore remains the major cause for concern' but the lysine value of the diet may also be important and should not be dismissed from consideration.


The importance of animal and pulse proteins

The regression equations linking lysine with the dietary percentages of AP% and PP% can be used to indicate the minimum levels of these protein sources in the diet to meet the recommendations of 50 ma. It would appear that this value can be reached with about 33% animal protein, or 25% animal plus 15% pulse protein. To allow for 90% digestibility, the lysine requirement should be increased to 55 mg/g protein. The diet should therefore provide about 33% of the total protein from animal sources together with a further 10% from pulses.

If the young-child value of 58 mg/g protein is to be used for all ages as recommended [13], the proportions would increase to more than 45% animal protein, or 30% animal and 20% legume protein. These percentages would further increase slightly when digestibility is taken into consideration. Present values for the percentage of protein supplied from animal foods approximate 68% in North America, 42% in Latin America, 24% in the Near East, and 20% in Africa . Pulse protein levels approximate 3%, 12%, 8%, and 12%, respectively [35]. These data again indicate the probable marginal nature of lysine supply in both the Near East and African diets.

Provided total food energy needs are met, it is thus desirable for at least one-third of dietary protein to originate from animal sources when total protein is limited. This is important not only in the context of protein nutrition but also because animal foods supply vitamins and minerals in generous and readily available forms [6]. When animal protein percentages are less than this, the well-known concept of protein complementation should be applied by increasing the availability and consumption of pulse proteins. In general, at adequate levels of total protein, a combined percentage of about 45% of protein from the high-lysine animal and pulse sources will ensure adequate lysine intake.


Relationship to lysine fortification of cereals

The lysine concentration of diets in several regions of the developing world may limit the nutritional quality of protein. It might appear, therefore, that the largely negative outcomes of the major field trials of lysine-fortified cereals [37] sponsored by the US Agency for International Development in the early 1970s and conducted on children in Guatemala, Thailand, and Tunisia contradict our assessment of the international food, nutrition, and public health implications concerning lysine needs. A critical evaluation of these three major field trials was conducted by a task force of the US National Research Council [37], with the conclusion that the results of the trials cannot be used to resolve, unequivocally, the question of whether lysine fortification of cereals can have a beneficial effect in underprivileged populations. Indeed, the case was made by the task force that further evaluation of the available, but still incompletely analysed data should be undertaken and that additional field studies are needed.

We should also recognize variations within families in terms of food distribution. Women, in particular, may well be consuming diets with poorer lysine values than might be judged from the country-wide averages evaluated earlier. Of course, individuals consuming diets that we have characterized as being of poor quality have survived under these conditions. We do not know, however, whether this is because our predictions are invalid, whether a relatively benign physiological adaptation has occurred, or whether there has been a functionally costly accommodation to lower than apparently desirable amino acid (or lysine) intake levels [38-41]. Such responses or differences in various diets are difficult to assess with current diagnostic methods, and clearly this is an area that requires vigorous investigation.

Summary and conclusions

A partial motivation for our present analysis of the amino acid supply relative to human requirements arose from the personal contacts that we had with scientists working at the various international centres for agricultural research. To establish an effective linkage between nutrition and agricultural development [4244], the nutritional objectives must be based on a sound and reliable analysis of human needs and wants. In this context, we are aware that the recommendations for protein and amino acid requirements put forward by FAO/WHO/UNU [3] are now being used in significant policy-making decisions regarding the direction and nature of research at these centres, which may involve major changes in the thrust of the previous research programmes.

Appropriate priority may not be given to research on legumes or to improving the protein value of cereals because of the implications of the requirement data for adults [3] that such efforts are irrelevant to the future health and wellbeing of underprivileged adult populations. Because of the importance, therefore, of our new, but admittedly still tentative, estimates of amino acid requirements for adults [4] and of the policy implications of the recommendations [3], we undertook the present analysis. It is intended to provide a further basis to help resolve the current concerns and debate about the appropriate directions of major agricultural research programmes focused on improving human nutrition. The recommendation that the amino acid requirement pattern for the preschool child [13] be used to evaluate protein quality supports the argument that policy decisions based on the 1985 FAO/WHO/UNU recommendations are likely to be erroneous.

The present study revealed that the diets in a number of developing countries, Africa, and the Near and Far East may be marginal in both their lysine value and their utilizable protein. This is primarily due to the much lower levels of lysine in cereal products than in legumes and animal foods. We estimate that where diets supply about one-third of total protein from animal sources or about 45% from a combination of animal proteins and legumes, the needs for lysine by most individuals would be met adequately unless total protein supplies were very low. Thus, there appears to be a substantial rationale for continuing research directed towards increasing the production of legumes and of cereals with higher lysine content as well as for improving the output and efficiency of production of animal protein foods.

Decisions regarding the particular nature of the research programme should, however, be made within the context of the overall characteristics of the diet of the population of concern. Our analysis also suggests that with respect to legume proteins, emphasis should be given to their lysine and total protein contents rather than to the sulfur amino acids, which have traditionally been the focus of major concern.


  1. National Academy of Sciences-National Research Council, Committee on Amino Acids. Evaluation of protein nutrition. Washington, DC: National Academy of Sciences-National Research Council, 1959.
  2. National Academy of Sciences-National Research Council, Food and Nutrition Board. Improvement of protein nutriture. Washington, DC: National Academy of Sciences-National Research Council, 1974.
  3. FAO/WHO/UNU. Energy and protein requirements. Report of a joint FAO/WHO/UNU expert consultation. Geneva: World Health Organization, 19X5.
  4. Young VR, Bier DM, Pellett PL. A theoretical basis for increasing current estimates of the amino acid requirements in adult man, with experimental support. Am J Clin Nutr 1989;50:80-92.
  5. FAD/WHO. Energy and protein requirements. Report of a joint FAD/WHO ad hoc expert committee. Geneva: World Health Organization. 1973.
  6. Pellett PL, Young VR. The contribution of livestock products to human dietary needs with special reference to West Asia and North Africa. In: Thomson EF, Thomson FS, eds. Workshop on small ruminants in Mediterranean areas. Dortrecht: Kluwer Academic 27. Publishers, 1988:3-25.
  7. Block RJ, Mitchell HH. The correlation of the amino-acid composition of proteins with their nutritive value. Nutr Abstr Rev 1946;16:249-76.
  8. FAO. Protein requirements. Rome: Food and Agriculture Organization, 1957.
  9. Recommended dietary allowances. 9th rev. ed. Washington, DC: National Academy of Sciences, 1980.
  10. FAD/WHO. Protein requirements. Rome: Food and Agriculture Organization, 1965.
  11. Pineda D, Torun B, Viteri FE, Arroyave G. Protein quality in relation to estimates of essential amino acid requirements. In: Bodwell CE, Adkins JS, Hopkins DT, eds. Protein quality in humans: assessment and in vitro estimation. Westport, Conn, USA: AVI Publishing, 1981 :29-42.
  12. FAD/WHO. Energy and protein requirements: recommendations by a joint FAD/WHO informal gathering of experts. Food Nutr 1975; 1(2): 11-19.
  13. FAD/WHO. Report of a joint FAD/WHO expert consultation on protein quality evaluation. Rome: FAO, 1990 (in press).
  14. Hegsted DM. Assessment of protein quality. In: Improvement of protein nutriture. Washington, DC: Food and Nutrition Board, National Academy of Sciences, 1974:64-88.
  15. Kaba M, Pellett PL. Prediction of true limiting amino acids using available protein scoring systems. Ecol Food Nutr 1975; 4:100-16.
  16. Chavez JF, Pellett PL. Protein quality of some representative Latin American diets by rat bioassay. J Nutr 1976; 106:792-801.
  17. Pellett PL. Protein quality evaluation revisited. Food Technol 1978; 32:61-68.
  18. Meguid MM, Matthews DE, Bier DM, Meredith CN, Soeldner JS, Young VR. Leucine kinetics at graded 38. leucine intakes in young men. Am J Clin Nutr 1986;43:370-80.
  19. Meguid MM, Matthews DE, Bier DM, Meredith CN, Young VR. Valine kinetics at graded valine intakes in young men. Am J Clin Nutr 1986: 43:781-86.
  20. Meredith CN, Wen Z-W, Bier DM, Matthews DE, Young VR. Lysine kinetics at graded lysine intakes in young men. Am J Clin Nutr 1986; 43:787-94.
  21. Zhao X-L, Wen Z-W, Meredith CN, Matthews DE, Bier DM, Young VR. Threonine kinetics at graded threonine intakes in young men. Am J Clin Nutr 1986; 43:705-802.
  22. Millward DJ, Rivers JPW. Protein and amino acid requirements in the adult human. J Nutr 1986; 116:255961.
  23. Millward DJ, Rivers JPW. The nutritional role of indispensable amino acids and the metabolic basis for their requirements. Eur J Clin Nutr 1988; 42:367-93.
  24. Millward DJ. Letter. Am J Clin Nutr 1990;51:491-93.
  25. Young VR, Bier DM, Pellett PL. Letter. Am J Clin Nutr 1990;51:494-96.
  26. PerissÚ J, Sizaret F, Francois P. The effect of income on the structure of the diet. Nutr Newsletter 1969;7(3):110.
  27. FAO. The fifth world food survey. Rome: Food and Agriculture Organization, 1985.
  28. FAO. Amino acid content of foods and biological data on proteins. Rome: Food and Agriculture Organization, 1970
  29. US Department of Agriculture. Agricultural handbook no. 8. Washington, DC: Agriculture Research Service, United States Department of Agriculture, 1976-1986.
  30. Pellett PL, Young VR. Role of meat as a source of protein and essential amino acids in human protein nutrition. In: Pearson AM, Dutson TR, eds. Meat and health. London: Elsevier, 1990:329-70.
  31. US Department of Agriculture. Food intakes: individuals in 48 states, 1977-1978. Nationwide food consumption survey 1977-78. Report no. I-l. Hyattsville, Md, USA: Consumer Nutrition Division, Human Nutrition Information Service, US Department of Agriculture, 1983.
  32. UNICEF. The state of the world's children 1987. New York: UNICEF, 1987.
  33. Pellett PL. Determinants of nutritional status. Food Nutr 1987;13(1):2-14.
  34. FAO. Food balance sheets 1975-1977. Average and per caput food supplies 1961-1965 and 1967-1977. Rome: Food and Agriculture Organization, 1980.
  35. FAO. Food balance sheets 1979-1981. Rome: Food and Agriculture Organization, 1984.
  36. Nutrient adequacy: assessment using food composition surveys. Washington, DC: National Academy Press, 1986
  37. Committee on International Programs, Food and Nutrition Board, National Research Council. The results and interpretation of three field trials of lysine fortification of cereals. Report of a task force on amino acid fortification of cereals. Washington, DC: National Academy Press, 1984.
  38. Beaton GH. The significance of adaptation in the definition of nutrient requirements and for nutrition policy. In: Blaxter K, Waterlow JC, eds. Nutritional adaptation in man. London: John Libbey, 1985:215-32.
  39. Waterlow JC. What do we mean by adaptation? In: Blaxter K, Waterlow JC, eds. Nutritional adaptation in man. London: John Libbey, 1985:1-10.
  40. Waterlow JC. Metabolic adaptations to low intakes of energy and protein. Ann Rev Nutr 1986;6:495-526.
  41. Young VR, Gucalp C, Rand WM, Matthews DE, Bier DM. Leucine kinetics during three weeks at submaintenance-to-maintenance intakes of leucine in men: adaptation and accommodation. Humn Nutr 1987;41C: 1-18.
  42. Lunven P, Sabry ZI. Nutrition and rural development. Food Nutr 1981;7:13-21.
  43. Nygaard D, Pellett PL, eds. Dry area agriculture, food science and human nutrition. New York: Pergamon Press, 1986.
  44. Pinstrup-Andersen P. Food policy and human nutrition. In: Nygaard D, Pellett PL, eds. Dry area agriculture, food science and human nutrition. New York: Pergamon Press, 1986:67-82.

Contents - Previous - Next