Recommended Dietary Allowances (RDAs), Recommended Dietary Intakes (RDIs), Recommended Nutrient Intakes (RNIs), and Population Reference Intakes (PRIs) are not “Recommended Intakes”

Recommended intakes of nutrients are hypothetical, perhaps even mythical, concepts used as yardsticks for the assessment of dietary surveys and food statistics; to provide guidance on appropriate dietary composition and meal provision; or for food labelling. The United States, Canada, and European countries, including The Netherlands, Germany, United Kingdom, France, and the Nordic countries, publish their own standards (1-6), and recently the European Union's Scientific Committee for Food has produced its own report, Nutrient and Energy Intakes for the European Community (7). Each committee has created its own terminology; and the many terms and their meanings have caused considerable confusion, which is exacerbated by the use of recommendations for purposes for which they were not intended. Many paediatricians and other health professionals think that the recommendations provide a reliable basis for the specific advice that they give families on the feeding of neonates, infants, children, and adolescents. Unfortunately, many users of these statistical terms do not appear to appreciate their real meanings and uses as expressed in the full reports: Instead they base their interpretation on the summary tables alone (8,9). They have therefore misunderstood the meaning of reference values, the underpinning assumptions, or the individual groups at which they are aimed. In addition, government agencies and industry are equally culpable in misusing reference values with their own objectives in mind, which are most often to increase the values rather than to reduce them (10). Finally, additional confusion arises from a failure to understand that, in contrast to labelling reference values, recommended intakes are not policy recommendations per se, but a set of reference standards developed by nutritional scientists that may be used to inform and to implement a sound public health policy (11). These are the points that we would like to clarify in this commentary. EVOLUTION OF DIETARY STANDARDS The first codified dietary recommendation was in the the British Merchant Seaman's Act in Britain in 1835, which mandated provision of lime or lemon juice to prevent scurvy in the Royal Navy (12). Probably the first true standard was published, in response to a request from the British Privy Council, to prevent starvation and associated diseases among the unemployed population during the economic depression of 1862. Several other recommendations were proposed during the following 50 years, notably by the food committees of the British Royal Society, of the League of Nations Health Organisation, and of the Canadian Council on Nutrition. From being recommendations for programmes to relieve starvation and illness resulting from economic and wartime crises, the dietary recommendations became, between 1920 and 1940, standards to maintain and improve the health of the population, with increased emphasis on meeting nutritional needs of infants, children, and pregnant women. Concurrently, from being observational standards, based on observed customary patterns of food consumption, they became based on scientific precepts and knowledge of human needs for essential nutrients and energy sources (11,13). A basic text that has influenced thinking is the Recommended Dietary Allowances (RDAs), prepared by the Food and Nutrition Board of the U.S. National Research Council and first published in 1941, “to serve as a guide for planning adequate nutrition for the civilian population” (14). According to the Food and Nutrition Board, RDAs are defined as the “levels of intake of essential nutrients that, on the basis of scientific knowledge, are judged... to be adequate to meet the known nutrient needs of practically all healthy persons” (2). Similarly, in 1983, the Canadian Ministry of Supply and Services defined their term Recommended Daily Nutrient Intakes as “the level of dietary intake thought to be sufficiently high to meet the requirements of almost all individuals in a group with specified characterisation” (15). Subsequently, a group of French experts applied a different concept, “Advisable Nutrient Intakes,” that “take into account not only scientific data on nutritional needs but also preferences and dietary habits, to the extent that these habits are not detrimental to health” (4). However, both concepts emphasise “appropriate” or “desirable” intakes, rather than specific needs; and as such they are akin to guidelines (11). The United Kingdom's Committee on Medical Aspects of Food Policy (COMA) published recommendations in 1979 on the amounts of energy and nutrients for groups of people in the United Kingdom. The definition of the RDA for a nutrient that was used in this report was “the average amount of the nutrient which should be provided per head in groups of people if the needs of practically all members of the group are to be met.” When these recommendations were reviewed in 1991, the British panel was aware of the continuing misuse and misinterpretation of the figures. To minimise this, it decided to set a range of intakes based as far as possible on the assessed range of requirements for each nutrient. These various levels were called Dietary Reference Values (DRVs). Three values were provided; the Estimated Average Requirement, the Reference Nutrient Intake, and the Lower Reference Nutrient Intake. According to the panel, intakes above the Reference Nutrient Intake would almost certainly be adequate, and the Lower Reference Nutrient Intake would represent the lowest intakes which would meet the needs of some individuals in the group; intake below this level would almost certainly be inadequate for most individuals (1). The Scientific Committee for Food (SCF) report followed the same thinking. But the committee preferred to call “the intake that is enough for virtually all healthy people in a group the population reference intake (PRI),” and used the term lowest threshold intake to characterise “the intake below which, on the basis of current knowledge, almost all individuals will be unlikely to maintain metabolic integrity according to the criteria chosen for each nutrient” (6). The Recommended Dietary Allowance from the United States, the Recommended Daily Nutrient Intakes from Canada, the Recommended Daily Amount and the Reference Nutrient Intakes from the United Kingdom, and the Population Reference Intakes from the European Union are similar concepts (Table 1). However, the rationale of the committees has evolved from the concept of “recommendation,” as being an implicit “amount that should be consumed,” to the concept of “reference values,” which reduces the chance of misunderstanding. Moreover, the use of the term “population” proposed by the European Union has the advantage of indicating that the proposed reference values are for groups of people and not for individuals. REFERENCES FOR FOOD LABELLING PURPOSES Consumers in the United States and Canada have long been accustomed to reading on the label of prepacked foodstuffs that one portion, or 100 g, of a food contains a particular percentage of the U.S. Recommended Daily Allowances. As a result of the Dietary Supplement Act of 1992, these values are no longer called RDAs but Reference Daily Intakes (RDIs), a change the U.S. Food and Drug Administration agrees is necessary to minimise confusion between RDAs and RDAs. Label reference values have been established for 19 vitamins and minerals. Their values have been set chiefly by selecting the highest RDA value from among the various sex and age groups in U.S. RDA tables published in 1968. The new U.S. regulation has also established label reference values for 8 other nutrients, including fat, cholesterol, and fibre. These values have been named Daily Reference Values (DRVs), but although U.S. regulatory requirements make it necessary to distinguish between the two sets of label reference values, in an attempt to avoid needless confusion, all reference values on food labels should be referred to as Daily Values, or DVs (16). Nevertheless, the use of many acronyms is confusing, and some familiarity with their derivation is necessary to understand that the DRVs (Dietary Reference Values) established by United Kingdom's Committee on Medical Aspects of Food Policy are quite different from the DRVs (Daily Reference Values) of the U.S. Food and Drug Administration regulations, or vice versa (Table 2). The European Union, like the Food and Drug Administration, ruled that food labelling should be standardised to help consumers; and hoped that such standardisation would also facilitate a single European market (16). However, contrary to other bodies, the European Union's Scientific Committee for Food has recommended that the Average Requirement for adult men (7) would be a more useful reference value than the PRI; because it would avoid the impression that an individual's intakes are systematically inferior to the needs of all members of the population. This recommendation presents a problem for products intended only for use by infants and young children. Therefore, the Scientific Committee for Food has proposed an additional set of reference values, based on the PRIs for children aged 6 months to 3 years (7,18). This approach resembles that advocated in the U.S. Dietary Supplement Act of 1992, which did not provide reference values for children younger than 4 years of age but included guidance on values that manufacturers may use on labels of foods for infants, young children, and pregnant and lactating women (16). ASSESSMENT OF VALUES-THE APPROACHES USED Because of the wide use of the RDIs, it is important to illustrate their limitations and applications by discussing the method that have been used to assess requirements. The physiological requirement for a nutrient should be the basis for calculating a reference intake. The ideal definition of a physiological requirement is the amount and chemical form of a nutrient that is needed systematically to maintain normal health and development without disturbance of the metabolism of any other nutrient. The corresponding dietary requirement would be the intake sufficient to meet the physiological requirement: Ideally this should be achieved without extreme homeostatic processes and excessive depletion or surplus in bodily depots (19). ANALOGY WITH INTAKES The earliest approach was empirically to base reference recommendations on the energy and nutrient intakes of groups of subjects apparently in good health. This method is weak, because it assumes that the subjects are in good health and are achieving their full genetic potential and that the diets are are quantitatively and qualitatively appropriate and free from adverse longterm effects. Reservations about this approach are strengthened by the current concern about the possible influence of early nutrition on metabolic “programming” and on the subsequent risk of hypertension, obesity, diabetes mellitus, and cardiovascular disease (20-23). It is commonly assumed, on the basis that breast-feeding is best for infants, that the natural intakes of breast-fed babies are an appropriate guide to optimal nutritional needs between birth and 6 months of age. However, there are uncertainties in this approach: We are not sure of the actual intakes of infants, because of the variability of milk supply (between 550 and 1100 ml/day) between women (24), because of the changing composition of the milk during the course of lactation during the day and even during a feeding, and because of difficulties in measuring these factors (25). Another major limitation of using maternal milk as a reference is its complex composition, which has evolved presumably to meet the specific requirements of the human infant and to ensure the optimum absorption and utilisation of its constituent nutrients. This complexity can not yet be emulated by synthetic feeds. Consequently, there are differences in the efficiency of substrate utilisation between breast- and formula-fed infants and, as a consequence, in the infants' dietary requirements. Examples of such differences are energy and protein requirements arising from variations in the energy cost of weight gain and body composition, and of iron requirements and bioavailability of iron from the two modes of feeding (26). Thus, the composition of human milk and the nutrient intake of breast-fed infants may not always be a useful guide for those babies who are not exclusively breast fed. FACTORIAL ANALYSIS The factorial approach calculates requirements by summarising the amounts needed for growth and maintenance and for physical activity in assessing energy requirements (27,28). The requirements for growth are derived from estimations of the composition of gain in body weight. However, very limited direct analytical information on body composition is available, although some data are available for newborns, adults, and a 4.5-year-old boy (29,30). One cannot reliably extrapolate from such information to the composition of the body and new tissue of children and adolescents in general. Such indirect methods as 40K and total body water measurements, bioelectrical impedance, and magnetic resonance imaging might fill some of these gaps in our knowledge. However, their limitations and those of the assumptions used to calculate body composition of infants, children, and adolescents should not be overlooked (30). Maintenance requirements are usually derived from the estimation of unavoidable losses in urine, faeces, sweat, shed skin, hair, and so forth. These losses depend not just on the dietary supply of the nutrient in question, but also on other metabolically interdependent nutrients. For example, nitrogen equilibrium in adults cannot be maintained simply by replacing observed obligatory losses (28); the supply of other nutrients is critical (e.g., energy, which affects adaptation to low protein supply) (31). Moreover, we do not know whether the metabolic adaptation needed to maintain equilibrium is functionally unimportant, or whether it might be detrimental to long-term health and survival (23,28). The situation is even more complex during childhood (32). Data on inevitable losses are scanty in newborns, toddlers, and infants. Furthermore, because such information is best gained under circumstances of negligible intakes of the nutrient under consideration, when homeostatic conservation is maximal, it is unlikely that such data could be acquired ethically. The interrelationship in infants and children between energy requirements and the efficiency of lean and new tissue synthesis, as well as the maturation of metabolism, are insufficiently understood to enable confident assessment of infants' and children's qualitative and quantitative protein requirements (33). BALANCE METHODOLOGY Balance studies at known intakes provide information about net whole retention and net intestinal absorption or secretion and whole-body retention of nutrients. The technique is demanding and, unless particular care is taken, technical difficulties in sampling can cause an overestimation of intakes and an underestimation of losses in faeces and urine. As a result, balance studies tend to overestimate net retention and thus to underestimate requirements (34). Subjects should be in equilibrium at the intake of the nutrient in question. This equilibrium depends on the previous supply of the nutrient relative to systemic requirements, which in turn influences intestinal uptake and transfer of nutrients and the rates at which they are subsequently utilised, catabolised, sequestered in body pools, or excreted. If systemic stores are large in relation to daily intake or requirement, their mobilisation will delay the onset of any adaptation that would be needed for balance-study equilibrium and for collection of reliable and meaningful data. This is the case with calcium, for which the daily intake is only a very minor fraction of the total calcium content of the body. A similar problem arises for phosphorus and magnesium. Consequently a very long time, possibly months, is taken to achieve equilibrium and to induce the adaptive responses detectable in conventional balance studies that are looking for increased intestinal and losses the and the systemic of is and adaptation to intakes can be a of these most reliable and available information depends on balance studies that can at be used to assess the of customary intakes, or at negligible intakes, to or obligatory losses under although the children in such the use of labels of or of nutrients can enable of the and the balance-study with these is to a method in requirements new methods OF of nutrient requirements should also into account and studies on specific The of scurvy in the British Navy is the and most of this approach (12). on human that scurvy could be or with as as per However, intakes are necessary to maintain a sufficient to prevent scurvy from for to months in an or to maintain the bodily by in a normal This that the requirement to the amount of nutrient needed to or just to avoid a will supply this nutrient at a adequate In years, has been on the possible of nutrients when in excessive all have a level of excessive intake at which (2). it is perhaps to set of intakes at levels well below those at which there is a risk of A major is the use of such as vitamins at intakes of to the reference intakes on the basis that such intakes could prevent the improve and prevent However, high of vitamins have and between at levels of intake used in of to assess requirements using balance the of customary intakes, and of or the metabolic adaptation of the individuals to different nutrient of this information our to assess both the of any particular intake and to the relative influences of intestinal and systemic factors and of dietary factors on the absorption and retention of the nutrient and to reliably such can be to other These factors the of a dietary the efficiency with which a nutrient is and used of this efficiency is needed to physiological requirements of a nutrient by factorial or with total feeding to an actual intake reference value or to facilitate the of data between influences on bioavailability are often and the term has been used to the of the or foodstuffs on nutrient the is true the more intestinal uptake and transfer of iron with that of The is that bioavailability can be only in and that such in one set of circumstances cannot be applied in other for foods are by other individuals under different about adaptation also when any increased intakes needed during and lactation are are no good data the that women with food intakes adequate for women should increase their intakes during for a increase in energy requirements. dietary surveys in intakes between pregnant and women A good is Because there are no specific to iron it has the the of the systemic iron is the efficiency of its uptake and transfer by the intestinal a of bioavailability to in for iron requirements far in of observed Such the increase of bioavailability of iron from standard during This of adaptation might be or even important than the of the dietary and the chemical form of iron of homeostatic adaptation has also to the of including in the of reference values a to for the maintenance of systemic important of such of low and for is that recommendations for some nutrients the amounts by the for the recommendations are being recently have some committees to to low levels of and to intakes during of bioavailability than those applied to intakes the United Kingdom's and European Union's reference values for iron and the of assessing it is not that values are and that, as a consequence, for even various on physiological requirements, these are expressed as quite different reference values (19). dietary reference requirements is a The values are and are often for rather than for They the often limited data available and the for an of data from groups of individuals to This is because there is no information on the population of requirements, which is to be but is in and influenced by homeostatic The influence of such factors as systemic and on the bioavailability and requirements for nutrients should be It is hoped that of the and of intestinal and systemic metabolic adaptation will help to clarify this Reference values are for infants and young children. They are often by between data from studies in infants and Furthermore, human milk is not as a standard for assessing the requirements of formula-fed infants, even the and longer term of breast-fed infants can be used as Reference intakes intended use than recommended intakes or and that these statistical terms in their They are for rather than for and, although they provide a of assessing the that members of a population are to nutrient reference values can not be used for the of nutrient in individuals. The of reference values can be established only by long-term of the or of nutrient and energy This will and of data and could be achieved by standardisation of both for reference values and their use for food labelling. to achieve such standardisation would be these current reference values are an of information for programmes and nutritional for Finally, if the of all these committees are appropriate sources of information for programmes in we with that they are for and be (11). has as are not for we that we are not even sure that they are for