Back to IAHF Menu

Introduction 1

What are Dietary Reference Intakes? 1

Approach for Setting Dietary Reference Intakes, Including Tolerable

Upper Intake Levels 4

Model for the Derivation of Tolerable Upper Intake Levels 5

Risk Assessment and Food Safety 6

Application of the Risk Assessment Model to Nutrients 9

Steps in the Development of the Tolerable Upper Intake Level 13

Derivation of ULs: Summary of Progress to Date 18

References 21

Appendix I: Recommended Dietary Intakes for Individuals 23

Appendix II: Options for Dealing with Uncertainties 25

Appendix III: Reference Heights and Weights for Children and Adults 29

Appendix IV: Case Studies of Application of Model for Risk Assessment

of Upper Levels of Nutrients

A. Calcium 30

B. Folate 42

C. Riboflavin 54

The model for risk assessment of nutrients used to develop tolerable upper intake levels (ULs) is one of the key elements of the developing framework for Dietary Reference Intakes, which are dietary reference values for the intake of nutrients and food components by Americans and Canadians recently released by the U.S. National Academy of Sciences in a series of reports (IOM, 1997; 1998). The overall project is a comprehensive effort undertaken by the Standing Committee on the Scientific Evaluation of Dietary Reference Intakes (DRI Committee) of the Food and Nutrition Board (FNB), Institute of Medicine, National Academy of Sciences in the United States, with active involvement of Health Canada. The DRI project is the result of significant discussion from 1991 to l996 by the FNB regarding how to approach the growing concern that one set of quantitative estimates of recommended intakes, the Recommended Dietary Allowances (RDAs), was scientifically inappropriate to be used as the basis for many of the uses to which they had come to be applied. The lack of specific determinations of maximum or tolerable upper levels of intake was noted (IOM, 1994).

The two DRI reports issued to date provide recommended intakes (see Appendix I) and upper levels of intake for two groups of nutrients and food components: calcium and related nutrients, and folate, B vitamins, and choline. Currently, the DRI Committee and a panel of experts are reviewing dietary antioxidants and related compounds, with similar reviews planned for other micronutrients including trace elements and vitamins A and K, electrolytes and fluid, macronutrients, and other food components not traditionally classified as "nutrients", but purported to play a beneficial role in human diets. It is expected that when evaluation of all groups of nutrients and food components are completed as part of this ongoing process, the model will have been fully developed and validated.

Dietary Reference Intakes (DRIs) are reference values that are quantitative estimates of nutrient intakes to be used for planning and assessing diets for healthy people. They include both recommended intakes and ULs as reference values (see Figure 1). Although the reference values

are based on data, the data are often scanty or drawn from studies that had limitations in addressing the question. Thus, scientific judgment is required in setting the reference values:

• Recommended Dietary Allowance (RDA): the average daily dietary intake level that is sufficient to meet the nutrient requirement of nearly all (97 to 98 percent) healthy individuals in a group.

• Adequate Intake (AI): a value based on observed or experimentally determined approximations of nutrient intake by a group (or groups) of health people—used when an RDA cannot be determined.

• Tolerable Upper Intake Level (UL): the highest level of daily nutrient intake that is likely to pose no risks of adverse health effects to almost all individuals in the general population. As intake increases above the UL, the risk of adverse effects increases.

• Estimated Average Requirement (EAR): a nutrient intake value that is estimated to meet the requirement of half the healthy individuals in a group.

The development of DRIs expands on the periodic reports called Recommended Dietary Allowances, which have been published since 1941 by the National Academy of Sciences.

The Recommended Dietary Allowance (RDA) is the average daily dietary intake level that is sufficient to meet the nutrient requirement of nearly all (97 to 98 percent) healthy individuals in a particular life-stage (life-stage considers age and, when applicable, pregnancy or lactation) and gender group.

The process for setting the RDA depends on being able to set an Estimated Average Requirement (EAR). That is, the RDA is derived from the nutrient requirement; and if an EAR cannot be set, no RDA will be set. The EAR is the daily intake value of a nutrient that is estimated to meet the nutrient requirement of half the healthy individuals in a life-stage and gender group. Before setting the EAR, a specific criterion of adequacy is selected, based on a careful review of the literature. When selecting the criterion, reduction of disease risk is considered along with many other health parameters. The RDA is set at the EAR plus twice the standard deviation (SD) if known (RDA = EAR + 2 SD); if data about variability in requirements are insufficient to calculate an SD, a coefficient of variation for the EAR of 10 percent is ordinarily assumed (RDA = 1.2 × EAR).

The RDA for a nutrient is a value to be used as a goal for dietary intake by healthy individuals. The RDA is not intended to be used to assess the diets of either individuals or groups or to plan diets for groups.

The Adequate Intake (AI) is set instead of an RDA if sufficient scientific evidence is not available to calculate an EAR. The AI is based on observed or experimentally-determined estimates of nutrient intake by a group (or groups) of healthy people. For example, the AI for young infants, for whom human milk is the recommended sole source of food for the first 4 to 6 months, is based on the daily mean nutrient intake supplied by human milk for healthy, full-term infants who are exclusively breastfed. The main intended use of the AI is as a goal for the nutrient intake of individuals. Other uses of AIs will be considered by another expert group.

The Tolerable Upper Intake Level (UL) is the highest level of daily nutrient intake that is likely to pose no risks of adverse health effects to almost all individuals in the general population. As intake increases above the UL, the risk of adverse effects increases. The term tolerable intake was chosen to avoid implying a possible beneficial effect. Instead, the term is intended to connote a level of intake that can, with high probability, be tolerated biologically. The UL is not intended to be a recommended level of intake. There is no established benefit for healthy individuals if they consume nutrient intakes above the RDA or AI.

ULs are useful because of the increased interest in and availability of fortified foods and the increased use of dietary supplements. ULs are based on total intake of a nutrient from food, water, and supplements if adverse effects have been associated with total intake. However, if adverse effects have been associated with intake from supplements or food fortificants only, the UL is based on nutrient intake from those sources only, not on total intake. The UL applies to chronic daily use.

For many nutrients, there are insufficient data on which to develop a UL. This does not mean that there is no potential for adverse effects resulting from high intake. When data about adverse effects are extremely limited, extra caution may be warranted.

The scientific data used to develop recommended intakes and ULs have come from observational and experimental studies. Studies published in peer-reviewed journals were the principal source of data. Life stage and gender were considered to the extent possible; but for some nutrients, the data did not provide a basis for proposing different requirements or upper levels for men and women or for adults in different age groups.

Three of the categories of reference values (EAR, RDA, and AI) are defined by specific criteria of nutrient adequacy; the fourth (UL) is defined by a specific end point of adverse effect if one is available. In all cases, data were examined closely to determine whether reduction of risk of a chronic degenerative disease or developmental abnormality could be used as a criterion of adequacy. The quality of studies was examined, considering study design, methods used for measuring intake and indicators of adequacy, and biases, interactions, and confounding factors. After careful review and analysis of the evidence, including examination of the extent of congruence of findings, scientific judgment was used to determine the basis for establishing the values.

The term "tolerable" is chosen because it connotes a level of intake that can, with high probability, be tolerated biologically by individuals; it does not imply acceptability of that level in any other sense. The setting of a UL does not indicate that nutrient intakes greater than the RDA or AI are recommended as being beneficial to an individual. Many individuals are self-medicating with nutrients for curative or treatment purposes. It is beyond the scope of the model at this time to address the possible therapeutic benefits of higher nutrient intakes that may offset the risk of adverse effects. The UL is not meant to apply to individuals who are treated with the nutrient or food component under medical supervision.

The term adverse effect is defined as any significant alteration in the structure or function of the human organism (Klaassen et al., 1986), or any impairment of a physiologically important function, in accordance with the definition set by the joint World Health Organization, Food and Agriculture Organization of the United Nations, and International Atomic Energy Agency (WHO/FAO/IAEA) Expert Consultation in Trace Elements in Human Nutrition and Health (WHO, 1996). In the case of nutrients, it is exceedingly important to consider the possibility that the intake of one nutrient may alter in detrimental ways the health benefits conferred by another nutrient. Any such alteration (referred to as an adverse nutrient-nutrient interaction) is considered an adverse health effect. When evidence for such adverse interactions is available, it is considered in establishing a nutrient’s UL.

Like all chemical agents, nutrients can produce adverse health effects if intakes from any combination of food, water, nutrient supplements, and pharmacologic agents is excessive. Some lower level of nutrient intake will ordinarily pose no likelihood (or risk) of adverse health effects in normal individuals even if the level is above that associated with any benefit. It is not possible to identify a single "risk-free" intake level for a nutrient that can be applied with certainty to all members of a population. However, it is possible to develop intake levels that are unlikely to pose risks of adverse health effects to most members of the general population, including sensitive individuals. For some nutrients or food components these intake levels may, however, pose a risk to subpopulations with extreme or distinct vulnerabilities.

The possibility that the methodology used to derive ULs might be reduced to a mathematical model that could be generically applied to all nutrients was considered. Such a model might have several potential advantages, including ease of application and assurance of consistent treatment of all nutrients. It was concluded, however, that the current state of scientific understanding of toxic phenomena in general, and nutrient toxicity in particular, is insufficient to support the development of such a model. Scientific information regarding various adverse effects and their relationships to intake levels varies greatly among nutrients and depends on the nature, comprehensiveness, and quality of available data. The uncertainties associated with the unavoidable problem of extrapolating from the circumstances under which data are developed (for example, in the laboratory or clinic) to other circumstances (for example, to the healthy population) adds to the complexity.

Given the current state of knowledge, any attempt to capture in a mathematical model all the information and scientific judgments that must be made to reach conclusions regarding ULs would not be consistent with contemporary risk assessment practices. Instead, the model for the derivation of ULs consists of a set of scientific factors that always should be considered explicitly. The framework under which these factors are organized is called risk assessment. Risk assessment (NRC, 1983, 1994) is a systematic means of evaluating the probability of occurrence of adverse health effects in humans from excess exposure to an environmental agent (in this case, a nutrient or food component) (FAO/WHO, 1995; Health Canada, 1993). The hallmark of risk assessment is the requirement to be explicit in all the evaluations and judgments that must be made to document conclusions.

Top

Risk assessment is a scientific undertaking having as its objective a characterization of the nature and likelihood of harm resulting from human exposure to agents in the environment. The characterization of risk typically contains both qualitative and quantitative information and includes a discussion of the scientific uncertainties in that information. In the present context, the agents of interest are nutrients, and the environmental media are food, water, and nonfood sources such as nutrient supplements and pharmacologic preparations.

Performing a risk assessment results in a characterization of the relationships between exposure(s) to an agent and the likelihood that adverse health effects will occur in members of exposed populations. Scientific uncertainties are an inherent part of the risk assessment process and are discussed below. Deciding whether the magnitude of exposure is "acceptable" in specific circumstances is not a component of risk assessment; this activity falls within the domain of risk management. Risk management decisions depend on the results of risk assessments but may also involve the public health significance of the risk, the technical feasibility of achieving various degrees of risk control, and the economic and social costs of this control. Because there is no single, scientifically definable distinction between "safe" and "unsafe" exposures, risk management necessarily incorporates components of sound, practical decision making that are not addressed by the risk assessment process (NRC, 1983, 1994).

A risk assessment requires that information be organized in rather specific ways but does not require any specific scientific evaluation methods. Rather, risk assessors must evaluate scientific information using what they judge to be appropriate methods; and they must make explicit the basis for their judgments, the uncertainties in risk estimates, and when appropriate, alternative interpretations of the available data that may be scientifically plausible (NRC, 1994; OTA, 1993).

Risk assessment is subject to two types of scientific uncertainties: (1) those related to data and (2) those associated with inferences that are required when directly applicable data are not available (NRC, 1994). Data uncertainties arise when evaluating information obtained from the epidemiologic and toxicologic studies of nutrient intake levels that are the basis for risk assessments. Examples of inferences include the use of data from experimental animals to estimate responses in humans and the selection of uncertainty factors to estimate inter- and intraspecies variabilities in response to toxic substances. Uncertainties arise whenever estimates of adverse health effects in humans are based on extrapolations of data obtained under dissimilar conditions (for example, from experimental animal studies). Options for dealing with uncertainties are discussed below and in detail in Appendix II.

The organization of risk assessment is based on a model proposed by the NRC (1983, 1994); that model is widely used in public health and regulatory decision making. The steps of risk assessment as applied to nutrients are as follows (see also Figure 2):

•  Step 1. Hazard identification involves the collection, organization, and evaluation of all information pertaining to the adverse effects of a given nutrient. It concludes with a summary of the evidence concerning the capacity of the nutrient to cause one or more types of toxicity in humans.

•  Step 2. Dose-response assessment determines the relationship between nutrient intake (dose) and adverse effect (in terms of incidence and severity). This step concludes with an estimate of the UL—it identifies the highest level of daily nutrient intake that is likely to pose no risks of adverse health effects to almost all individuals in the general population. Different ULs may be developed for various life-stage groups.

•  Step 3. Intake assessment evaluates the distribution of usual total daily nutrient intakes among members of the general population.

•  Step 4. Risk characterization summarizes the conclusions from Steps 1 through 3 and evaluates the risk. Generally, the risk is expressed as the fraction of the exposed population, if any, having nutrient intakes (Step 3) in excess of the estimated UL (Steps 1 and 2). If possible,

characterization also covers the magnitude of any such excesses. Scientific uncertainties associated with both the UL and the intake estimates are described so that risk managers understand the degree of scientific confidence they can place in the risk assessment.

The risk assessment contains no discussion of recommendations for reducing risk; these are the focus of risk management.

A principal feature of the risk assessment process for noncarcinogens is the long-standing acceptance that no risk of adverse effects is expected unless a threshold dose (or intake) is exceeded. The adverse effects that may be caused by a nutrient or food component almost certainly occur only when the threshold dose is exceeded (NRC, 1994; WHO, 1996). The critical issues concern the methods used to identify the approximate threshold of toxicity for a large and diverse human population. Because most nutrients are not considered to be carcinogenic in humans, the approach to carcinogenic risk assessment (EPA, 1996) is not discussed here.

Thresholds vary among members of the general population (NRC, 1994). For any given adverse effect, if the distribution of thresholds in the population could be quantitatively identified, it would be possible to establish ULs by defining some point in the lower tail of the distribution of thresholds that would be protective for some specified fraction of the population. However, data are not sufficient to allow identification of the distribution of thresholds for all but a few, well-studied nutrients and compounds found in food (for example, acute toxic effects or for chemicals such as lead, where the human database is very large). The method for identifying thresholds for a general population described here is designed to ensure that almost all members of the population will be protected, but it is not based on an analysis of the theoretical (but practically unattainable) distribution of thresholds. By using the model to derive the threshold, however, there is considerable confidence that the threshold, which becomes the UL for nutrients or food components, lies very near the low end of the theoretical distribution, and is the end representing the most sensitive members of the population. For some nutrients, there may be subpopulations that are not included in the general distribution because of extreme or distinct vulnerabilities to toxicity. Such distinct groups, whose conditions warrant medical supervision, may not be protected by the UL.

The Joint FAO/WHO Expert Commission on Food Additives and various national regulatory bodies have identified factors (called uncertainty factors) that account for interspecies and intraspecies differences in response to the hazardous effects of substances and to account for other uncertainties (WHO, 1987). These factors are used to make inferences about the threshold dose of substances for members of a large and diverse human population from data on adverse effects obtained in epidemiological or experimental studies. These factors are applied consistently when data of specific types and quality are available. They are typically used to derive acceptable daily intakes for food additives and other substances for which data on adverse effects are considered sufficient to meet minimum standards of quality and completeness (FAO/WHO, 1982). These adopted or recognized uncertainty factors have sometimes been coupled with other factors to compensate for deficiencies in the available data and other uncertainties regarding data.

When possible, the UL is based on a no-observed-adverse-effect level (NOAEL), which is the highest intake (or experimental oral dose) of a nutrient at which no adverse effects have been observed in the individuals studied. This is identified for a specific circumstance in the hazard identification and dose-response assessment steps of the risk assessment. If there are no adequate data demonstrating a NOAEL, then a lowest-observed-adverse-effect level (LOAEL) may be used. A LOAEL is the lowest intake (or experimental oral dose) at which an adverse effect has been identified. The derivation of a UL from a NOAEL (or LOAEL) involves a series of choices about what factors should be used to deal with uncertainties. Uncertainty factors (UFs) are applied in an attempt both to deal with gaps in data and with incomplete knowledge regarding the inferences required (for example, the expected variability in response within the human population). The problems of both data and inference uncertainties arise in all steps of the risk assessment. A discussion of options available for dealing with these uncertainties is presented below and in greater detail in the Appendix II.

A UL is not, in itself, a description of human risk. It is derived by application of the hazard identification and dose-response evaluation steps (Steps 1 and 2) of the risk assessment model. To determine whether populations are at risk requires an intake or exposure assessment (Step 3, evaluation of their intakes of the nutrient) and a determination of the fractions of those populations, if any, whose intakes exceed the UL. In the intake assessment and risk characterization steps (Steps 3 and 4), the distribution of actual intakes for the population is used as a basis in determining whether and to what extent the population is at risk.

This section provides guidance for applying the risk assessment framework (the model) to the derivation of ULs for nutrients.

Although the risk assessment model outlined above can be applied to nutrients to derive ULs, it must be recognized that nutrients possess some properties that distinguish them from the types of agents for which the risk assessment model has originally developed (NRC, 1983). In the application of accepted standards for assessing risks of environmental chemicals to the risk assessment of nutrients and food components, a fundamental difference between the two categories must be recognized: within a certain range of intakes, many nutrients are essential for human well-being and usually for life itself. Nonetheless, they may share with other chemicals the production of adverse effects at excessive exposures. Because the consumption of balanced diets is consistent with the development and survival of humankind over many millennia, there is less need for the large uncertainty factors that have been used in the typical risk assessment of nonessential chemicals. In addition, if data on the adverse effects of nutrients are available primarily from studies in human populations, there will be less uncertainty than is associated with the types of data available on nonessential chemicals.

There is no evidence to suggest that nutrients consumed at the recommended intake (the RDA or AI) present a risk of adverse effects to the general population. It is clear, however, that the addition of nutrients to a diet, either through the ingestion of large amounts of highly fortified food or nonfood sources such as supplements, or both, may (at some level) pose a risk of adverse health effects. The UL is the highest level of daily nutrient intake that is likely to pose no risks of adverse health effects to almost all individuals in the general population. As intake increases above the UL, the risk of adverse effects increases.

If adverse effects have been associated with total intake, ULs are based on total intake of a nutrient from food, water, and supplements. For cases in which adverse effects have been associated with intake only from supplements and/or food fortificants, the UL is based on intake from those sources only, rather than on total intake. The effects of nutrients from fortified foods or supplements may differ from those of naturally occurring constituents of foods because of several factors: the chemical form of the nutrient, the timing of the intake and amount consumed in a single bolus dose, the matrix supplied by the food, and the relation of the nutrient to the other constituents of the diet. Nutrient requirements and food intake are related to the metabolizing body mass, which is also at least an indirect measure of the space in which the nutrients are distributed. This relation between food intake and space of distribution supports homeostasis, which maintains nutrient concentrations in that space within a range compatible with health. However, excessive intake of a single nutrient from supplements or fortificants may compromise this homeostatic mechanism. Such elevations alone may pose risks of adverse effects; imbalances among the concentrations of mineral elements (for example, calcium, iron, zinc, an copper) can result in additional risks (Mertz et al., 1994). These reasons and those discussed previously support the need to include the form and pattern of consumption in the assessment of risk from high nutrient intake.

This risk assessment model outlined in this paper is consistent with classical risk assessment approaches in that it must consider variability in the sensitivity of individuals to adverse effects of nutrients. A discussion of how variability is dealt with in the context of nutritional risk assessment follows.

Physiological changes and common conditions associated with growth and maturation that occur during an individual’s lifespan may influence sensitivity to nutrient toxicity. For example, (1) sensitivity increases with declines in lean body mass and with declines in renal and liver function that occur with aging; (2) sensitivity changes in direct relation to intestinal absorption or intestinal synthesis of nutrients (for example, Vitamin K, biotin); (3) in the newborn infant, sensitivity is also increased because of rapid brain growth and limited ability to secrete or biotransform toxicants; and (4) sensitivity increases with decreases in the rate of metabolism of nutrients. During pregnancy, the increase in total body water and glomerular filtration results in lower blood levels of water soluble vitamins dose-for-dose, and therefore, reduced susceptibility to potential adverse effects. However, in the unborn fetus this may be offset by active placental transfer, accumulation of certain nutrients in the amniotic fluid, and rapid development of the brain. Examples of life-stage groups that may differ in terms of nutritional needs and toxicological sensitivity include infants and children, the elderly population, and women during pregnancy and lactation.

Even within relatively homogeneous life-stage groups, there is a range of sensitivities to toxic effects. The model described below accounts for normally expected variability in sensitivity, but it excludes subpopulations with extreme and distinct vulnerabilities. Such subpopulations consist of individuals needing medical supervision; they are better served through the use of public health screening, product labeling, or other individualized health care strategies. (Such populations may not be at "negligible risk" when their intakes reach the UL developed for the healthy population.) The decision to treat identifiable vulnerable subgroups as distinct (not protected by the UL) is a matter of judgment and is made evident in the rationale provided for characterizing the UL.

In the context of toxicity, the bioavailability of an ingested nutrient can be defined as its accessibility to normal metabolic and physiological processes. Bioavailability influences a nutrient’s beneficial effects at physiological levels of intake and also may affect the nature and severity of toxicity due to excessive intakes. Factors that affect bioavailability include the concentration and chemical form of the nutrient, the nutrition and health of the individual, and excretory losses. Bioavailability data for specific nutrients must be considered and incorporated by the risk assessment process.

Some nutrients, such as folate, may be less readily absorbed when they are part of a meal than when taken separately. Supplemental forms of some nutrients, such as some of the B vitamins, phosphorus, or magnesium, may require special consideration if they have higher bioavailability and therefore may present a higher risk of producing adverse effects than equivalent amounts from the natural form found in food.

A diverse array of adverse health effects can occur as a result of the interaction of nutrients. The potential risks of adverse nutrient-nutrient interactions increase when there is an imbalance in the intake of two or more nutrients. Excessive intake of one nutrient may interfere with absorption, excretion, transport, storage, function, or metabolism of a second nutrient. For example, dietary interactions can affect the chemical forms of elements at the site of absorption through ligand binding or changes in the valence state of an element (Mertz et al., 1994). Phytates, phosphates, and tannins are among the most powerful depressants of bioavailability, and organic acids, such as citric and ascorbic acid, are strong enhancers for some minerals and trace elements. Thus dietary interactions strongly influence the bioavailability of elements by affecting their partition between the absorbed and the nonabsorbed portion of the diet. The large differences of bioavailability ensuing from these interactions support the need to specify the chemical form of the nutrient when setting ULs. Dietary interactions can also alter nutrient bioavailability through their effect on excretion. For example, dietary intake of protein, phosphorus, sodium, and chloride all affect urinary calcium excretion and hence calcium bioavailability. Interactions that significantly elevate or reduce bioavailability may represent adverse health effects.

Although it is critical to include knowledge of any such interactions in the risk assessment, it is difficult to evaluate the possibility of interactions without reference to a particular level of intake. This difficulty can be overcome if a UL for a nutrient or food component is first derived based on other measures of toxicity. Then an evaluation can be made of whether intake at the UL has the potential to affect the bioavailability of other nutrients.

Possible adverse nutrient-nutrient interactions, then, are considered as a part of setting a UL. Nutrient-nutrient interactions may be considered either as a critical endpoint on which to base a UL for that nutrient or as supportive evidence for a UL based on another endpoint.

In addition to nutrient interactions, other considerations have the potential to influence nutrient bioavailability, such as the nutritional status of an individual and the form of intake. These issues should be considered in the risk assessment. The absorption and utilization of most minerals, trace elements, and some vitamins are a function of the individual’s nutritional status, particularly regarding the intake of other specific nutrients such as iron (Barger-Lux et al., 1995; Mertz et al., 1994).

With regard to the form of intake, minerals and trace elements often are less readily absorbed when they are part of a meal than when taken separately or when present in drinking water (NRC, 1989). The opposite is true for fat-soluble vitamins whose absorption depends on fat in the diet. ULs must therefore be based on nutrients as part of the total diet, including the contribution from water. Nutrient supplements that are taken separately from food require special consideration, since they are likely to have different availabilities and therefore may represent a greater risk of producing toxic effects.

Top

Based on a thorough review of the scientific literature, the hazard identification step outlines the adverse health effects that have been demonstrated to be caused by the nutrient. The primary types of data used as background for identifying nutrient hazards in humans are as follows:

•  Human studies. Human data provide the most relevant kind of information for hazard identification, and, when they are of sufficient quality and extent, are given greatest weight. However, the number of controlled human toxicity studies conducted in a clinical setting is very limited for ethical reasons. Such studies are generally most useful for identifying very mild (and ordinarily reversible) adverse effects. Observational studies that focus on well-defined populations with clear exposures to a range of nutrient intake levels are useful for establishing a relationship between exposure and effect. Observational data in the form of case reports or anecdotal evidence are used for developing hypotheses that can lead to knowledge of causal associations. Sometimes a series of case reports, if it shows a clear and distinct pattern of effects, may be reasonably convincing on the question of causality.

•  Animal studies. The majority of the available data used in regulatory risk assessments comes from controlled laboratory experiments in animals, usually mammalian species other than humans (for example, rodents). Such data are used in part because human data on food derived substances, particularly nonessential chemicals, are generally very limited. Because well-conducted animal studies can be controlled, establishing causal relationships is generally not difficult. However, cross species differences make the usefulness of animal data for establishing ULs problematic (see below).

Six key issues that are addressed in the data evaluation of human and animal studies are the following:

1.  Evidence of adverse effects in humans. The hazard identification step involves the examination of human, animal, and in vitro published evidence addressing the likelihood of a nutrient or food component eliciting an adverse effect in humans. Decisions regarding which observed effects are adverse are based on scientific judgments. Although toxicologists generally regard any demonstrable structural or functional alteration to represent an adverse effect, some alterations may be considered of little or self-limiting biological importance. As noted earlier, adverse nutrient-nutrient interactions are considered in the definition of an adverse effect.

2.  Causality. Is a causal relationship established by the published human data? The criteria of Hill (1971) are considered in judging the causal significance of an exposure-effect association indicated by epidemiologic studies. These criteria include: demonstration of a temporal relationship, consistency, narrow confidence intervals for risk estimates, a biological gradient or dose response, specificity of effect, biological plausibility, and coherence.

3.  Relevance of experimental data. Consideration of the following issues can be useful in assessing the relevance of experimental data.

•  Animal data. Animal data may be of limited utility in judging the toxicity of nutrients because of highly variable interspecies differences in nutrient requirements. Nevertheless, relevant animal data are considered in the hazard identification and dose-response assessment steps where applicable.

•  Route of exposure. Data derived from studies involving oral exposure (rather than parenteral, inhalation, or dermal exposure) are most useful for the evaluation of nutrients and food components. Data derived from studies involving parenteral, inhalation, or dermal routes of exposure may be considered relevant if the adverse effects are systemic and data are available to permit interroute extrapolation.

•  Duration of exposure. Because the magnitude, duration, and frequency of exposure can vary considerably in different situations, consideration needs to be given to the relevance of the exposure scenario (for example, chronic daily dietary exposure versus short-term bolus doses) to dietary intakes by human populations.

4.  Mechanisms of toxic action. Knowledge of molecular and cellular events underlying the production of toxicity can assist in dealing with the problems of extrapolation between species and from high to low doses. It may also aid in understanding whether the mechanisms associated with toxicity are those associated with deficiency. In most cases, however, because knowledge of the biochemical sequence of events resulting from toxicity and deficiency is still incomplete, it is not yet possible to state with certainty whether or not these sequences share a common pathway. Iron, the most thoroughly studied trace element, may represent the only exception to this statement. Deficient to near-toxic exposures share the same pathway, which maintains controlled oxygen transport and catalysis. Toxicity sets in when the exposure exceeds the specific iron-complexing capacity of the organism, resulting in free iron species initiating peroxidation.

5.  Quality and completeness of the database. The scientific quality and quantity of the database are evaluated. Human or animal data are reviewed for suggestions that the substances have the potential to produce additional adverse health effects. If suggestions are found, additional studies may be recommended.

6.  Identification of distinct and highly sensitive subpopulations. The ULs are based on protecting the most sensitive members of the general population from adverse effects of high nutrient intake. Some highly sensitive subpopulations have responses (in terms of incidence, severity, or both) to the agent of interest that are clearly distinct from the responses expected for the healthy population. The risk assessment process recognizes that there may be individuals within any life-stage group that are more biologically sensitive than others, and thus their extreme sensitivities do not fall within the range of sensitivities expected for the general population. The UL for the general population may not be protective for these subgroups. As indicated earlier, the extent to which a distinct subpopulation will be included in the derivation of a UL for the general population is an area of judgment to be addressed on a case-by-case basis.

The process for deriving the UL is described in this section and outlined in Figure 3. It includes selection of the critical data set, identification of a critical endpoint with its NOAEL (or LOAEL), and assessment of uncertainty.

The data evaluation process results in the selection of the most appropriate or critical data set(s) for deriving the UL. Selecting the critical data set includes the following considerations:

•  Human data are preferable to animal data.

•  In the absence of appropriate human data, information from an animal species whose biological responses are most like those of humans is most valuable.

•  If it is not possible to identify such a species or to select such data, data from the most sensitive animal species, strain, or gender combination are given the greatest emphasis.

•  The route of exposure that most resembles the route of expected human intake is preferable. This includes considering the digestive state (for example, fed or fasted) of the subjects or experimental animals. Where this is not possible, the differences in route of exposure are noted as a source of uncertainty.

•  The critical data set defines a dose-response relationship between intake and the extent of the toxic response known to be most relevant to humans. Data on bioavailability are considered and adjustments in expressions of dose-response are made to determine whether any apparent differences in response can be explained. For example, it is known that different metal salts can display different degrees of bioavailability. If the database involves studies of several different salts (for example, iron or chromium valence states), and the effect of the nutrient is systemic, then apparent differences in the degree and/or form of the toxic response among different salts may simply reflect differences in bioavailability. Data on bioavailability are considered and adjustments in expressions of dose-response are made to determine whether any apparent differences in response can be explained.

•  The critical data set documents the route of exposure and the magnitude and duration of the intake. Furthermore, the critical data set documents the intake that does not produce adverse effects (the NOAEL), as well as the intake producing toxicity.

A nutrient can produce more than one toxic effect (or endpoint), even within the same species or in studies using the same or different exposure durations. The NOAELs and (LOAELs) for these effects will differ. The critical endpoint used to establish a UL is the adverse biological effect exhibiting the lowest NOAEL (for example, the most sensitive indicator of a nutrient or food component’s toxicity). The derivation of a UL based on the most sensitive endpoint will ensure protection against all other adverse effects.

For some nutrients, there may be inadequate data on which to develop a UL. The lack of reports of adverse effects following excess intake of a nutrient does not mean that adverse effects do not occur. As the intake of any nutrient increases, a point (A, see Figure 4) is reached at which intake begins to pose a risk. Above this point, increased intake increases the risk of adverse effects. For some nutrients, and for various reasons, there are inadequate data to identify point A, or even to make any estimate of its location.

Because adverse effects are almost certain to occur for any nutrient at some level of intake, it should be assumed that such effects may occur for nutrients for which a scientifically documentable UL cannot now be derived. Until a UL is set or an alternative approach to identifying protective limits is developed, intakes greater than the RDA or AI should be viewed with caution.

Uncertainty Assessment

Several judgments must be made regarding the uncertainties and thus the uncertainty factor (UF) associated with extrapolating from the observed data to the general population (see Appendix II). Applying a UF to a NOAEL (or LOAEL) results in a value for the derived UL that is less than the experimentally derived NOAEL, unless the UF is 1.0. The larger the uncertainty, the larger the UF and the smaller the UL. This is consistent with the ultimate goal of the risk assessment: to provide an estimate of a level of intake that will protect the health of the healthy population (Mertz et al., 1994).

Although several reports describe the underlying basis for UFs (Dourson and Stara, 1983; Zielhuis and van der Kreek, 1979), the strength of the evidence supporting the use of a specific UF will vary. Because the imprecision of these UFs is a major limitation of risk assessment approaches, considerable leeway must be allowed for the application of scientific judgment in making the final determination. Since data are generally available regarding intakes of nutrients and food components in human populations, the data on nutrient toxicity may not be subject to the same uncertainties as with nonessential chemical agents resulting in UFs for nutrients and food components typically less than 10. They are lower with higher quality data and when the adverse effects are extremely mild and reversible.

In general, when determining a UF, the following potential sources of uncertainty are considered and combined into the final UF:

•  Interindividual variation in sensitivity. Small UFs (close to 1) are used if it is judged that little population variability is expected for the adverse effect, and larger factors (close to 10) are used if variability is expected to be great (NRC, 1994).

•  Experimental animal to human. A UF is generally applied to the NOAEL to account for the uncertainty in extrapolating animal data to humans. Larger UFs (close to 10) may be used if it is believed that the animal responses will underpredict average human responses (NRC, 1994).

•  LOAEL to NOAEL. If a NOAEL is not available, a UF may be applied to account for the uncertainty in deriving a UL from the LOAEL. The size of the UF involves scientific judgment based on the severity and incidence of the observed effect at the LOAEL and the steepness (slope) of the dose response.

•  Subchronic NOAEL to predict chronic NOAEL. When data are lacking on chronic exposures, scientific judgment is necessary to determine whether chronic exposure is likely to lead to adverse effects at lower intakes than those producing effects after subchronic exposures (exposures of shorter duration).

ULs are derived for various life-stage groups using relevant databases, NOAELs and LOAELs, and UFs. In cases where no data exist with regard to NOAELs or LOAELs for the group under consideration, extrapolations from data in other age groups and/or animal data are made on the basis of known differences in body size, physiology, metabolism, absorption, and excretion of the nutrient.

If the data review reveals the existence of subpopulations having distinct and exceptional sensitivities to a nutrient’s toxicity, these subpopulations should be explicitly discussed and concerns related to adverse effects noted; however the use of the data is not included in the identification of the NOAEL or LOAEL upon which the UL for the general population is based.

The model described in this document has been applied to two groups of nutrients and food components as part of the continuing DRI process. The selection of a UF of approximately 1.0 for fluoride and magnesium is primarily based on the very mild (and in the case of magnesium, reversible) nature of the adverse effects observed. A slightly larger UF (1.2) was selected for vitamin D intake in adults and in other life-stage groups except infants as the short duration of the study used (Narang et al., 1984) and the small sample size supports the selection of a slightly larger UF. For vitamin D in infants, a larger UF (1.8) was selected due to the insensitivity of the critical endpoint, the small sample sizes of the studies, and the limited data about the sensitivity at the tails of the distribution. A UF of 2 was selected for calcium to account for the potential increased susceptibility to high calcium intake by individuals who form renal stones and the potential adverse effects in the range between normal phosphorus levels and levels associated with ectopic mineralization. The selection of a UF for phosphorus is due to the relative lack of human data describing adverse effects of excess phosphorus intakes.

With regard to the B vitamins and choline, because of lack of suitable data that met the requirements of the model, NOAELs (and LOAELs) could not be set for thiamin, riboflavin, vitamin B₁₂, pantothenic acid, or biotin. The UF for synthetic folic acid was 5, based primarily on the uncontrolled observation that millions of people to date had been exposed to self treatment with about 0.1 of the LOAEL as 400 m g in vitamin pills without reporting harm. For niacin as a supplement or food fortificant, the UF selected was 1.5, based on the transient nature of the adverse effect of flushing, and the consideration that it was applied to a LOAEL and not a NOAEL. The UF for both vitamin B₆ and choline was 2. In the case of vitamin B₆, there was less data available involving responses to pyridoxine doses under 500 mg/day, and thus more limited information upon which to base a UL. The UF of 2 for choline was selected because of the limited data regarding hypotension and the magnitude of the interindividual variation in response to cholinergic effects.

Derivation of a UL

UL values have been established for broad age groups for nutrients for which adequate data are available (see Table 1). Values are set at levels that are unlikely to pose risk to the most sensitive members of the general population and cannot be used to assess the prevalence of the population at risk for adverse effects as a result of excess intakes. The UL for magnesium is from supplement intake only, and for niacin and folate (in the form of folic acid) from fortified food and supplement intake only. In all three cases, the nutrient naturally found in foods is excluded from concern. The adverse effect or critical endpoint which is used for each nutrient is given in Table 2. Three case studies (calcium, folate, and biotin) are described in Appendix IV.

The UL is derived by dividing the NOAEL (or LOAEL) by a single UF that incorporates all relevant uncertainties for the lifestage category for which the data is available (see Table 1). The derivation of a UL involves the use of scientific judgment to select the appropriate NOAEL (or LOAEL) and UF. The risk assessment requires explicit consideration and discussion of all choices made, both regarding the data used and the uncertainties accounted for.

For infants, ULs were not determined for any of the B vitamins, choline, magnesium, phosphorus, or calcium because of the lack of data on adverse effects in this age group and concern regarding infants’ possible lack of ability to handle excess amounts. Thus, caution is warranted; food should be the source of intake of these nutrients by infants. For vitamin D and fluoride, due to the significant information on effects from various levels of intake by infants for these nutrients, ULs were developed.

Top

Table 1. Tolerable Upper Intake Levels (UL^a) for Certain Nutrients and Food Components