Nutrient-to-energy ratios have attracted wide interest as indices of dietary quality. They have been used in establishing standards of suitable quality (e.g., in the design of diets or in nutrition labelling). They have also been used as indices to be applied when considering whether the quality or the quantity of the diet is likely to be more limiting in particular situations (as in considering whether there may be greater benefit from specific nutrient intervention than from improvement in total food intake). In either mode of application, the underlying question is: “If an individual (or population of individuals) consumes this diet in amounts that will satisfy energy needs, will the concentration of nutrients also be high enough to meet his nutrient needs?”
Platt et al. (1) were largely responsible for the introduction of the ratio of protein energy to total energy (PE ratio) as a convenient and useful descriptor of one aspect of dietary quality in human nutrition. To take into account both the quality and concentration of the protein, they introduced the concept of net dietary protein calories as a percentage of total calories (NDPCals %).1 In their work it was applied to the assessment of the adequacy of intakes. For this purpose, the reference PE ratio was calculated as the simple ratio of protein requirements (expressed as equivalent energy) to energy requirements.
In the years that followed, the use of the PE ratio became the topic of much confusion and controversy. It now seems clear that it is a useful concept within limited ranges of application but that, since a reference PE ratio intended to be used as a standard of adequacy can be calculated in different ways, the proposed application of any derived reference ratio must be specified clearly and the limitations to its interpretation must be appreciated. This section attempts to link the principles and judgements that led to the estimation of protein and energy requirements in this report to the derivation and interpretation of reference PE ratios based on these requirement estimates. Emphasis continues to be placed upon the principles that should be applied.
1 Although the term “ratio” strictly implies a fraction, in existing usage numerical values are expressed as percentages.
The calculation of PE ratios which describe known diets is straightforward and noncontroversial. Confusion about the PE ratio relates to the reference criteria to be applied to these calculations and has arisen from three main sources:
This last point emphasizes the fact that reference PE ratios are situation-specific. Unless such ratios are empirically similar in all situations of interest (a hypothesis that can be examined through appropriate calculations), there can be no universal criterion of a suitable PE ratio (2–6).
10.2.1 Ratios applicable to a particular diet
In considering the nutritional suitability of the diet consumed by an individual, the natural reference standard is the ratio of that individual's protein and energy requirements. As is pointed out in this report (sections 2 and 11), the true requirement of a random individual is not known. It can be described only in probability terms as part of the distribution of requirements in a class of similar individuals. It is logical, therefore, that the appropriate PE ratio for an individual should be described as part of the distribution of possible satisfactory ratios of protein to energy requirements among similar individuals. This distribution would include all possible situations in which protein needs would be met when enough of the diet was ingested to meet energy needs. It would take into account the situation in which a random individual's protein needs were high and his energy needs were low as well as when another individual's protein needs were low and energy needs were high. It would consider the likelihood of these situations in any class of individuals who are similar in terms of the variables used to classify individuals for the description of requirements (section 2). From the distribution of possible PE ratios, probability statements may be derived concerning the probable adequacy or inadequacy of any particular PE ratio for an individual of the specified type (age, sex, activity, etc.).
Although the distribution of possible satisfactory PE ratios is not a standard statistical one, an approximate approach to its derivation can be developed. One such approach is described in Annex 9(A). Given reasonable estimates of both the average requirements for protein and energy and the correlation between requirements for protein and energy within groups of similar individuals, the equation presented in Annex 9(A) can be used. The equation defines a PE ratio such that, with any probability (or risk) one wishes to assign, there is confidence that a higher observed dietary ratio would meet or exceed the true requirement ratio of a random individual in the group—a “safe” PE ratio for that class of individual. This particular statistical and conceptual approach follows from earlier work on the PE ratio (6) and from application of Fieller's theorem (7).
The concept underlying this approach is shown in Fig. 4. The curve shown is the distribution for a specified class of individuals of all possible PE ratios that would meet the criterion of adequate protein intake when energy needs are met. This can be transposed into a probability curve describing the likelihood that a given PE ratio will or will not meet the protein requirements of a randomly selected individual when enough of the diet is ingested to meet his or her energy requirement. This concept is entirely analogous to that described for the distribution of protein requirements (section 2) and to the interpretation of requirement distributions in terms of statements of risk (sections 2 and 11). By applying the equation in Annex 9(A), the PE ratios associated with particular probabilities or risks can be estimated. By assigning a probability of 0.975 (a “risk” of 2.5 in 100) one can describe the “safe PE ratio” of the diet in a manner analogous to the “safe level of protein intake”. However, in practice, it must be recognized that the ratio is meaningless unless the energy needs are met by ingestion of enough of the diet; if the total intake of food is inadequate to meet energy needs, protein metabolism might be compromised and protein needs would not be fulfilled at the stipulated PE ratio. In addition, the derivation of the ratio requires that the situation for which energy needs are estimated should be defined (see later discussion). These are clear limitations to interpretation.
Fig. 4. Distribution of required PE ratios
10.2.2 PE ratios applicable to average diets of populations
Further adjustments have to be introduced when deriving reference PE ratios which can be used to judge the suitability of the average PE ratio of self-selected diets consumed by a group of individuals or a population.
Among individuals consuming self-selected diets, it must be expected that the PE ratio of ingested food varies from one person to the next. The average PE ratio for the group does not describe the diet consumed by all members of it. An example of this situation is shown in Fig. 5. Studies in North America and Guatemala suggest that the coefficient of variation of the PE ratio of ingested diets among comparable individuals may be some 10–15% (8,9). The approach to the development of a reference PE ratio described in the section above refers to the ratio that should characterize a particular diet, i.e., the one actually ingested by a particular individual. Clearly it is inappropriate to use this as a reference ratio in judging the mean of the distribution of ratios among self-selected diets. One approach would be to apply the concepts described above to the PE ratio of the diet for each individual and then to aggregate the probability (or risk) statements for the whole group (see discussion of this approach to the interpretation of protein requirements in section 11). Another approach would be to derive estimates of average requirements for energy and protein, the variabilities of each, and correlations between them, at the level of the group or population and then to apply the concept and approach developed above with the population as the unit of observation rather than the individual. For reasons described in section 11, such an aggregation of requirement estimates is quite complex. The derivation of population level PE ratios is well beyond the scope of this report.
Fig. 5. Comparison of required and ingested PE ratios in self-selected diets
The more overlap of these distributions there is, the greater will be the expected number of individuals with inadequate protein intakes, given the condition that all individuals satisfy their energy needs. A knowledge of both distributions would be required to predict the prevalence of inadequacy.
The important point that emerges from these considerations is that “per caput” or average dietary PE ratios must not be compared with reference ratios calculated by applying the equation given in Annex 9(A), which is based on information about the requirement distributions of individuals. A reference PE ratio derived in this way would be too low, since it does not take into account the variance of dietary PE ratios (or indeed the variance of reference PE ratios among classes of individuals). Earlier work, with a more restrictive model, suggested that the per caput reference PE ratio might be 15–20% higher than that provided by the equation given in Annex 9(A) (4). It now appears that even this may be an underestimate. Clearly any aggregate comparisons of actual and desirable PE ratios must be developed and interpreted with great care.
10.3.1 Impact of energy requirement on the reference PE ratio
If reference PE ratios are calculated as described above for individuals or classes of individuals, based on the requirement estimates presented in this report, it will be found that the requisite PE ratio rises with increasing age after infancy and early childhood. The reference ratio for the older adult (60+ years) is likely to be much higher than that for the preschool child. Although this may appear to contradict commonly held notions about which age groups are more vulnerable to “diet quality”, it is a perfectly logical conclusion from the requirement estimates presented in this report. The reference PE ratio does not rise because of a progressive increase in protein need per unit of body weight with increasing age; after early childhood there is relatively little change. Rather, the ratio rises because with increasing age there is a progressive fall in estimated energy needs in proportion to body size. The denominator of the ratio falls in relation to the numerator.
A similar phenomenon is observed in young adults as physical activity and hence energy requirements change. For very inactive individuals the reference PE ratio will be quite high; as activity increases, the ratio will fall. Accordingly, it should be recognized that the derived reference PE ratio will be very sensitive to measured or assumed levels of physical activity (5, 6) and hence to particular social situations.
In the present report it has been pointed out that energy requirements may be estimated in terms of the existing body size and composition and existing patterns of activity (the status quo approach) or they may be estimated in terms of “desirable” parameters of body size and activity (the normative approach). It is logical to assume that for any stable population, there will have been a variety of accommodations or adaptations that have established an equilibrium between energy intake and expenditure. Thus, it also seems logical to assume that existing energy intakes are adequate or nearly adequate, in most situations, to meet existing energy needs. It is when a normative approach to growth rate, body size in adults, and activity profiles for all ages is introduced that important differences between energy intake and energy requirement appear.
From the above it follows also that in calculating PE ratios there are two approaches that may be followed. The status quo approach would accept existing energy intake as an estimator of existing energy requirement and use this in the calculation of PE ratio. The normative approach would accept the energy requirement that would be associated with a desired state of growth rate, body size, and activity and use this in calculating the PE ratio. In so far as growth rate and body size descriptors differ in these two approaches, estimates of protein requirement as well as energy requirement would be affected. However, for differences in activity profiles, only the energy requirement estimates would be influenced.
These two approaches can be expected to yield quite different estimates of appropriate PE ratios. The status quo approach may yield ratios some 15–20% higher than the normative approach (4).
In the past decade or so, since the publication of the report of the Joint FAO/WHO Ad Hoc Expert Committee on Energy and Protein Requirements (10), it has been argued by many that both energy and protein intakes are low in comparison to requirement estimates for low income populations in developing countries. However, it also has been argued that if actions are taken to increase energy intakes to approximate requirement estimates, the accompanying increase in protein intake would meet or surpass requirement estimates (i.e., the PE ratio of existing diets is adequate but the total level of food intake is low). If one accepts the status quo approach to energy intake and requirement, the argument of relative adequacy may be fallacious. If existing energy intakes and the associated body sizes and activity profiles are deemed acceptable, then it is appropriate to assess protein intake without major regard to energy (or to calculate PE ratios using existing energy intakes as the estimate of energy requirement). This could lead to a different conclusion about the probable adequacy of existing diets, at least for some population groups.
Although recognizing the practical importance of these considerations, the Consultation did not attempt an analysis of dietary data to explore the implications. Rather, this discussion is presented to emphasize the need for caution in the calculation, application, and interpretation of PE ratios. The Consultation also emphasized the need to consider very carefully the goals of interventions (in terms of growth rates, body sizes, and activity profiles) in any consideration of the suitability of nutrient: energy ratios.
10.3.2 Catch-up growth as a special situation
During rapid growth there is a change in both energy and protein requirements per unit body size. Because these changes differ in relative magnitude, the appropriate reference PE ratio will change. The method used to calculate the reference PE ratio does not change, but the requirement estimates will differ. These changes were discussed in section 9. The important practical point is that while reference PE ratios may be expected to rise with increasing assumed growth rates, there may be practical external limitations on the rate of catch-up growth that can be achieved by children in the community; there would be little merit in establishing reference PE ratios for growth rates that cannot be attained (3, 11). Similarly, if the quantity of food offered to a child limits the rate of catch-up growth, a higher concentration of protein in the food may have no real value. A careful analysis of the situation is needed before reference PE ratios for catch-up growth can be calculated and interpreted meaningfully.
10.3.3 Adjustment for digestibility and amino acid composition of dietary protein
To be useful the PE ratio must take into account the digestibility and amino acid composition of dietary protein. How these affect protein needs or the efficiency of utilization of dietary protein has been discussed in section 7. Reference PE ratios could be adjusted upward to allow for the digestibility and amino acid composition of diets ingested by particular age groups. This was the method adopted in the past. However, in many applications a more useful approach would be to adjust the observed protein intake to a value that represents the intake of utilizable protein. The reference PE ratio would then continue to be expressed in terms of protein equivalent to egg or milk protein and the dietary protein would be adjusted to this base. This is analogous to the approach of Platt et al. (1) in their use of the term NDPCals %, although the methods of calculating the reference PE ratio and adjusting for protein quality are different in detail.
10.3.4 Other measures of dietary quality
This section has discussed only two parameters of the nutritional quality of diets, energy and protein, combined in the form of protein : energy ratios as an index of diet quality. It must be recognized that the concentration of all other necessary nutrients should be considered when assessing the adequacy of a particular diet or establishing standards of nutritional quality for foods and diets. It is potentially dangerous to consider energy and protein alone.
There is a special situation where the PE ratio (and other nutrient : energy ratios) of a diet may appear to be adequate, but where it can be predicted that insufficient amounts of the diet will be consumed to satisfy either energy or nutrient needs. This happens when young children are offered a too bulky diet with a low energy density (section 7.2). The volume of food to be consumed may exceed the capacity of the child and hence the PE ratio will be meaningless. The shortfall between the volume of food that a child needs to consume to satisfy requirements and the volume that it is capable of consuming depends on both the energy density of the food mixture and the pattern of the meals provided. The presumed limitation is the physical capacity of the gastrointestinal system to accept food offered at particular times. In this situation the constraints on food intake may be reduced either by changes in the feeding pattern or the energy density. It has been suggested that adding oil to increase the energy density would be a practical solution in many cases (12–14). If the PE ratio of the original diet were already low, adding oil would be expected to lower the ratio even further. Comparison of the resultant PE ratio with reference PE ratios might in some situations indicate that the addition was inadvisable. This would be an example of an appropriate use of PE ratios in assessing the quality of diets. This example also illustrates that much more than the PE ratio must be taken into account when assessing the quality of diets in the community.