Identification of a NOAEL or LOAEL

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 ordinarily differ. The critical endpoint used to establish a UL is the adverse biological effect exhibiting the lowest NOAEL (e.g., the most sensitive indicator of a nutrient's toxicity). Because the selection of uncertainty factors (UFs) depends in part upon the seriousness of the adverse effect, it is possible that lower ULs may result from the use of the most serious (rather than most sensitive) endpoint. Thus, it is often necessary to evaluate several endpoints independently to determine which leads to the lowest UL.

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 (see Figure 4-2) 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 this point, or even to estimate 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 protec-

Observed Level of Intake

FIGURE 4-2 Theoretical description of health effects of a nutrient as a function of level of intake. The Tolerable Upper Intake Level (UL) is the highest level of daily nutrient intake that is likely to pose no risk of adverse health effects for almost all individuals in the general population. At intakes above the UL, the risk of adverse effects increases.

Observed Level of Intake

FIGURE 4-2 Theoretical description of health effects of a nutrient as a function of level of intake. The Tolerable Upper Intake Level (UL) is the highest level of daily nutrient intake that is likely to pose no risk of adverse health effects for almost all individuals in the general population. At intakes above the UL, the risk of adverse effects increases.


tive limits is developed, intakes greater than the Recommended Dietary Allowance (RDA) or Adequate Intake (AI) should be viewed with caution.

The absence of sufficient data to establish a UL points to the need for studies suitable for developing ULs.

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 L). 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. The greater the uncertainty, the larger the UF and the smaller the resulting 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 virtually all members 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. Because data are generally available regarding intakes of nutrients in human populations, the data on nutrient toxicity may not be subject to the same uncertainties as are data on non-essential chemical agents. The resulting UFs for nutrients and food components are typically less than the factors of 10 often applied to non-essential toxic substances. The UFs 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 in the final UF:

• Interindividual variation in sensitivity. Small UFs (close to 1) are used to represent this source of uncertainty 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).

• Extrapolation from experimental animals to humans. A UF to account for the uncertainty in extrapolating animal data to humans is generally applied to the NOAEL when animal data are the primary data available. While a default UF of 10 is often used to extrapolate animal data to humans for nonessential chemicals, a lower UF may be used because of data showing some similarities between the animal and human responses (NRC, 1994).

• LOAEL instead of 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 exposures are likely to lead to adverse effects at lower intakes than those producing effects after subchronic exposures (exposures of shorter duration).

Derivation of a UL

The UL is derived by dividing the NOAEL (or LOAEL) by a single UF that incorporates all relevant uncertainties. ULs, expressed as amount per day, are derived for various life stage groups using relevant databases, NOAELs, 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 or animal data are made on the basis of known differences in body size, physiology, metabolism, absorption, and excretion of the nutrient.

Generally, any age group adjustments are made based solely on differences in body weight, unless there are data demonstrating age-related differences in nutrient pharmacokinetics, metabolism, or mechanism of action.

The derivation of the UL involves the use of scientific judgment to select the appropriate NOAEL (or LOAEL) and UF. As shown in Figure 4-3, when using the same critical endpoint there is a greater level of uncertainty in setting the UL based on a LOAEL compared with a NOAEL. The risk assessment requires explicit consideration and discussion of all choices made regarding both the data used and the uncertainties accounted for. These considerations are discussed in the nutrient chapters.

Characterization of the Estimate and Special Considerations

If the data review reveals the existence of subpopulations having distinct and exceptional sensitivities to a nutrient's toxicity, these subpopulations are explicitly discussed and concerns related to adverse effects are 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.

Circumstances in Which No UL Is Established

There are two general conditions under which ULs are not established. In some cases, the availability of insufficient evidence regarding a



w CD




Increasing Intake -►

FIGURE 4-3 Effect of uncertainty assessment on the Tolerable Upper Intake Level (UL). Dashed line represents a hypothetical no-observed-adverse-effect level (NOAEL). Solid lines represent available data used to set the UL. Area containing diagonal lines represents theoretical range of uncertainty. LOAEL = lowest-observed-adverse-effect level; RDA = Recommended Dietary Allowance.

nutrient's capacity to cause adverse effects prohibits the application of the UL model. In other cases, the evidence is available, but meeting the UL derived from such evidence will necessarily result in the introduction of undesirable health effects because of the required adjustments in dietary patterns.

Insufficient Evidence of Adverse Effects

The scientific evidence relating to adverse effects of nutrient excess varies greatly among nutrients. The type of data and evidence of causation used to derive ULs have been described earlier in this chapter, but such data and evidence are simply unavailable for some nutrients. In some cases (e.g., the individual amino acids), some data relating to adverse effects may be available, but are of such uncertain relevance to human health that


their use in deriving ULs is scientifically insupportable. In every instance in which ULs are not derived because of lack of adequate evidence, the specific limitations in the database are described.

Offsetting Benefits Reduction

In the case of macronutrients, particularly, problems arise because of the adjustments in dietary patterns that would be required to meet a derived UL. For saturated and trans fatty acids and dietary cholesterol, for example, there is evidence that any intake greater than zero will increase serum levels of low density lipoprotein cholesterol, an established risk for cardiovascular disease. In such cases, the UL model calls for the establishment of a UL of 0. But it is clear that, because saturated fat and cholesterol are both unavoidable in ordinary diets, achieving such a UL will require extraordinary changes in patterns of dietary intake. Such extraordinary adjustments may introduce other undesirable health effects (e.g., elimination of foods containing saturated fats may result in a large excess intake of carbohydrate and insufficient intake of micronutrients). In addition, unknown and unquantifiable health risks may also be introduced. For these reasons, no UL will be proposed in circumstances in which implementation of measures to achieve the UL may lead to undesirable dietary adjustments. In all such cases, the basis for failing to propose a UL will be described.

Lack of ULs for Macronutrients and Implications

ULs were not set for macronutrients because (1) there was insufficient evidence for identifying an adverse effect, and therefore a LOAEL, upon which to determine a UL (e.g., protein), (2) data relating to adverse effects were available (e.g., amino acids), but were of uncertain relevance to human health because their use in deriving ULs was not scientifically supportable, (3) macronutrients are interrelated in providing energy and therefore it is not known whether the adverse effect is due to a high intake of one macronutrient (e.g., fat), due to a low intake of another macro-nutrient (e.g., carbohydrate, which is usually low in a high fat diet), or both (high fat, low carbohydrate diet), and (4) adjustments of dietary patterns to prevent exceeding a UL of near 0 g/d (e.g., trans and saturated fatty acids and cholesterol), resulting in inadequate intakes of certain micro-nutrients (e.g., iron and zinc). In addition, the UL method is not applicable to energy since any intake above the requirement would be expected to result in weight gain and an increased risk of premature mortality.

The failure to establish a UL for any nutrient should not be interpreted as a lack of concern for adverse health effects (i.e., it is not equiva-


lent to a recommendation that the nutrient can be consumed without limit). Lack of data regarding adverse effects is not evidence of safety. Indeed, in some cases (the previous example of saturated fat) there is clearly evidence of adverse health effects, but a UL is not established to avoid the need for drastic changes that may introduce undesirable health effects.

In every instance in which a UL is not established, it is necessary to offer specific advice regarding the need to avoid deficiency, or in some cases, to reduce intakes, consistent with the need to maintain healthy dietary patterns.


In order to assess the risk of adverse effects, information on the range of nutrient intakes in the general population is required. As noted earlier, in cases where the Tolerable Upper Intake Level (UL) pertains only to supplement use and not to usual food intakes of the nutrient, the assessment is directed at supplement intake only.


As described earlier, the question of whether nutrient intakes create a risk of toxicity requires a comparison of the range of nutrient intakes (from food, supplements, and other sources, or from supplements alone, depending upon the basis for the Tolerable Upper Intake Level [UL]) with the UL.

Figure 4-4 illustrates a distribution of chronic nutrient intakes in a population; the fraction of the population experiencing chronic intakes above the UL represents the potential at-risk group. A policy decision is needed to determine whether efforts should be made to reduce risk. No precedents are available for such policy decisions, although in the areas of food additives or pesticide regulations, federal regulatory agencies have generally sought to ensure that the 90th or 95th percentile of intake falls below the UL (or its approximate equivalent measure of risk). If this goal is achieved, the fraction of the population remaining above the UL is likely to experience intakes only slightly greater than the UL and is likely to be at little or no risk.

For risk management decisions, it is useful to evaluate the public health significance of the risk, and information contained in the risk characterization is critical for this purpose.

Thus, the significance of the risk to a population consuming a nutrient in excess of the UL is determined by the following:


Chronic Nutrient Intake Distribution

Population at Known Risk

Mean Intake UL NOAEL or


FIGURE 4-4 Illustration of the population at risk from excessive nutrient intakes. The fraction of the population consistently consuming a nutrient at intake levels in excess of the Tolerable Upper Intake Level (UL) is potentially at risk of adverse health effects. See text for a discussion of additional factors necessary to judge the significance of the risk. NOAEL = no-observed-adverse-effect level; LOAEL = lowest-observed-adverse-effect level.

1. the fraction of the population consistently consuming the nutrient at intake levels in excess of the UL,

2. the seriousness of the adverse effects associated with the nutrient,

3. the extent to which the effect is reversible when intakes are reduced to levels less than the UL, and

4. the fraction of the population with consistent intakes above the no-observed-adverse-effect level or even the lowest-observed-adverse-effect level.

Population at Known Risk

Mean Intake UL NOAEL or


FIGURE 4-4 Illustration of the population at risk from excessive nutrient intakes. The fraction of the population consistently consuming a nutrient at intake levels in excess of the Tolerable Upper Intake Level (UL) is potentially at risk of adverse health effects. See text for a discussion of additional factors necessary to judge the significance of the risk. NOAEL = no-observed-adverse-effect level; LOAEL = lowest-observed-adverse-effect level.

Thus, the significance of the risk of excessive nutrient intake cannot be judged only by reference to Figure 4-4, but requires careful consideration of all of the above factors. Information on these factors is contained in sections of the nutrient chapters that describe the bases for each of the ULs.


Dourson ML, Stara JF. 1983. Regulatory history and experimental support of uncertainty (safety) factors. Regul Toxicol Pharmacol 3:224-238.


FAO/WHO (Food and Agriculture Organization of the United Nations/World Health Organization). 1982. Evaluation of Certain Food Additives and Contaminants. Twenty-sixth report of the Joint FAO/WHO Expert Committee on Food Additives. WHO Technical Report Series No. 683. Geneva: WHO.

FAO/WHO. 1995. The Application of Risk Analysis to Food Standard Issues. Recommendations to the Codex Alimentarius Commission (ALINORM 95/9, Appendix 5). Geneva: WHO.

Health Canada. 1993. Health Risk Determination—The Challenge of Health Protection. Ottawa: Health Canada, Health Protection Branch.

Hill AB. 1971. Principles of Medical Statistics, 9th ed. New York: Oxford University Press.

Klaassen CD, Amdur MO, Doull J. 1986. Casarett and Doull's Toxicology: The Basic Science of Poisons, 3rd ed. New York: Macmillan.

Mertz W, Abernathy CO, Olin SS. 1994. Risk Assessment of Essential Elements. Washington, DC: ILSI Press.

NRC (National Research Council). 1983. Risk Assessment in the Federal Government: Managing the Process. Washington, DC: National Academy Press.

NRC. 1994. Science and Judgment in Risk Assessment. Washington, DC: National Academy Press.

OTA (Office of Technology Assessment). 1993. Researching Health Risks. Washington, DC: OTA.

WHO (World Health Organization). 1987. Principles for the Safety Assessment of Food Additives and Contaminants in Food. Environmental Health Criteria 70. Geneva: WHO.

WHO. 1996. Trace Elements in Human Nutrition and Health. Geneva: WHO.

Zielhuis RL, van der Kreek FW. 1979. The use of a safety factor in setting health-based permissible levels for occupational exposure. Int Arch Occup Environ Health 42:191-201.

Energy is required to sustain the body's various functions, including respiration, circulation, physical work, and maintenance of core body temperature. The energy in foods is released in the body by oxidation, yielding the chemical energy needed to sustain metabolism, nerve transmission, respiration, circulation, and physical work. The heat produced during these processes is used to maintain body temperature. Energy balance in an individual depends on his or her dietary energy intake and energy expenditure. Imbalances between intake and expenditure result in gains or losses of body components, mainly in the form of fat, and these determine changes in body weight.

The Estimated Energy Requirement (EER) is defined as the average dietary energy intake that is predicted to maintain energy balance in a healthy, adult of a defined age, gender, weight, height, and level of physical activity consistent with good health. To calculate the EER, prediction equations for normal weight individuals were developed from data on total daily energy expenditure measured by the doubly labeled water technique. In children and pregnant or lactating women, the EER includes the needs associated with the deposition of tissues or the secretion of milk at rates consistent with good health. While the expected between-individual variability is calculated for the EER, there is no Recommended Dietary Allowance (RDA) for energy because energy intakes above the EER would be expected to result in weight gain. Similarly, the Tolerable Upper Intake Level (UL) concept does not apply to



energy, because any intake above an individual's energy requirement would lead to undesirable (and potentially hazardous) weight gain.


Humans and other mammals constantly need to expend energy to perform physical work; to maintain body temperature and concentration gradients; and to transport, synthesize, degrade, and replace small and large molecules that make up body tissue. This energy is generated by the oxidation of various organic substances, primarily carbohydrates, fats, and amino acids. In 1780, Lavoisier and LaPlace measured the heat production of mammals by calorimetry (Kleiber, 1975). They demonstrated that it was equal to the heat released when organic substances were burned, and that the same quantities of oxygen were consumed by animal metabolism as were used during the combustion of the same organic substrates (Holmes, 1985). Indeed, it has been verified by numerous experiments on animals and humans since then that the energy produced by oxidation of carbohydrates and fats in the body is the same as the heat of combustion of these substances (Kleiber, 1975). The crucial difference is that in organisms oxidation proceeds through many steps, allowing capture of some of the energy in an intermediate chemical form—the high energy pyrophosphate bond of adenosine triphosphate (ATP). Hydrolysis of these high-energy bonds can then be coupled to various chemical reactions, thereby driving them to completion, even if by themselves they would not proceed (Lipmann, 1941). Typically, the rates of energy expenditure in adults at rest are slightly less than 1 kcal/min in women (i.e., 0.8 to 1.0 kcal/min or 1,150 to 1,440 kcal/d), and slightly more than 1 kcal/min in men (i.e., 1.1 to 1.3 kcal/min or 1,580 to 1,870 kcal/d) (Owen et al., 1986, 1987). One kcal/min corresponds approximately to the heat released by a burning candle or by a 75-watt light bulb (i.e., 1 kcal/min corresponds to 70 J/sec or 70 W).

Energy Yields from Substrates

Carbohydrate, fat, protein, and alcohol provide all of the energy supplied by foods and are generally referred to as macronutrients (in contrast to vitamins and elements, usually referred to as micronutrients). The amount of energy released by the oxidation of carbohydrate, fat, protein, and alcohol (also known as Heat of Combustion, or AH) is shown in Table 5-1.

When alcohol (ethanol or ethyl alcohol) is consumed, it promptly appears in the circulation and is oxidized at a rate determined largely by its concentration and by the activity of liver alcohol dehydrogenase. Oxi-



TABLE 5-1 Heat of Combustion of Various Macronutrients

Heat of


Atwater Factor^



kcalb/L O2

RQc (CO2/O2)






















Protein by



Protein through










a The energy derived by protein oxidation in living organisms is less than the heat of combustion of protein, because the nitrogen-containing end product of metabolism in mammals is urea (or uric acid in birds and reptiles), whereas nitrogen is converted into nitrous oxide when protein is combusted. The heat liberated by biological oxidation of proteins was long thought to be 4.3 kcal/g (Merrill and Watt, 1973), but a more recent demonstration showed that the actual value is 4.7 kcal/g (Livesey and Elia, 1988). b One calorie is the amount of energy needed to increase the temperature of 1 g of water from 14.5° to 15.5°C. In the context of foods and nutrition, "large calorie" (i.e., Calories, with a capital C), which is more properly referred to as "kilocalorie" (kcal), has been traditionally used. In the International System of Units, the basic energy unit is the Joule (J). One J = 0.239 calories, so that 1 kcal = to 4.186 kJ. A daily energy expenditure of 2,400 kcal corresponds to the expenditure of 10,000 kJ, or 10 MJ (Mega Joules)/d.

c RQ = respiratory quotient, which is defined as the ratio of CO2 produced divided by O2 consumed (in terms of mols, or in terms of volumes of CO2 and O2). d Atwater, a pioneer in the study and characterization of nutrients and metabolism, proposed to use the values of 4, 9, and 4 kcal/g of carbohydrate, fat, and protein, respectively (Merrill and Watt, 1973). This equivalent is now uniformly used in nutrient labeling and diet formulation. Nutrition Labeling of Food. 21 C.F.R. §101.9 (1991). e Alcohol (ethanol) content of beverages is usually described in terms of percent by volume. The heat of combustion of alcohol is 5.6 kcal/mL. (One mL of alcohol weighs 0.789 g.)

dation of alcohol elicits a prompt reduction in the oxidation of other substrates used for ATP regeneration, demonstrating that ethanol oxidation proceeds in large part via conversion to acetate and oxidative phos-phorylation. The phenomenon has been precisely measured by indirect calorimetry in human subjects, in whom ethanol consumption was found to primarily reduce fat oxidation (Suter et al., 1992). About 80 percent of the energy liberated by ethanol oxidation is used to drive ATP regeneration, so that the thermic effect of ethanol comes to about 20 percent (Siler et al., 1999). The thermic effect of food is the increase in energy expendi-


ture as measured by heat produced upon ingestion of that food. The thermic effect of alcohol is about twice the thermic effect of carbohydrate, but less than the thermic effect of protein (see later section, "Thermic Effect of Food").

Reported food intake in individuals consuming alcohol is often similar to that of individuals who do not consume alcohol (de Castro and Orozco, 1990). As a result, it has sometimes been questioned whether alcohol contributes substantially to energy production. However, the biochemical and physiological evidence about the contribution made by ethanol to oxidative phosphorylation is so unambiguous that the apparent discrepancies between energy intake data and body weights must be attributed to inaccuracies in reported food intakes. In fact, in individuals consuming a healthy diet, the additional energy provided by alcoholic beverages can be a risk factor for weight gain (Suter et al., 1997), as opposed to alcoholics in whom the pharmacological impact of excessive amounts of ethanol tends to inhibit normal eating and may cause emaciation.

Energy Requirements Versus Nutrient Requirements

Recommendations for nutrient intakes are generally set to provide an ample supply of the various nutrients needed (i.e., enough to meet or exceed the requirements of almost all healthy individuals in a given life stage and gender group). For most nutrients, recommended intakes are thus set to correspond to the median amounts sufficient to meet a specific criterion of adequacy plus two standard deviations to meet the needs of nearly all healthy individuals (see Chapter 1). However, this is not the case with energy because excess energy cannot be eliminated, and is eventually deposited in the form of body fat. This reserve provides a means to maintain metabolism during periods of limited food intake, but it can also result in obesity.

The first alternate criterion that may be considered as the basis for a recommendation for energy is that energy intake should be commensurate with energy expenditure, so as to achieve energy balance. Although frequently applied in the past, this is not appropriate as a sole criterion, as described by the FAO/WHO/UNU publication, Energy and Protein Requirements (1985):

The energy requirement of an individual is a level of energy intake from food that will balance energy expenditure when the individual has a body size and composition, and level of physical activity, consistent with long-term good health; and that would allow for the maintenance of economically necessary and socially desirable physical activity. In children and pregnant or lactating women the energy requirement includes the energy needs associated with


the deposition of tissues or the secretion of milk at rates consistent with good health (p. 12).

This definition indicates that desirable energy intakes for obese individuals are less than their current energy expenditure, as weight loss and establishment of a steady state at a lower body weight is desirable for them. In underweight individuals, on the other hand, desirable energy intakes are greater than their current energy expenditure to permit weight gain and maintenance of a higher body weight. Thus, it seems logical to base estimated values for energy intake on the amounts of energy that need to be consumed to maintain energy balance in adult men and women who are maintaining desirable body weights, taking into account the increments in energy expenditure elicited by their habitual level of activity.

There is another fundamental difference between the requirements for energy and those for other nutrients. Body weight provides each individual with a readily monitored indicator of the adequacy or inadequacy of habitual energy intake, whereas a comparably obvious and individualized indicator of inadequate or excessive intake of other nutrients is not usually evident.

Energy Balance

Because of the effectiveness in regulating the distribution and use of metabolic fuels, man and animals can survive on foods providing widely varying proportions of carbohydrates, fats, and proteins. The ability to shift from carbohydrate to fat as the main source of energy, coupled with the presence of substantial reserves of body fat, makes it possible to accommodate large variations in macronutrient intake, energy intake, and energy expenditure. The amount of fat stored in an adult of normal weight commonly ranges from 6 to 20 kg. Since one gram of fat provides 9.4 kcal, body fat energy reserves thus range typically from approximately 50,000 to 200,000 kcal, providing a large buffer capacity as well as the ability to provide energy to survive for extended periods (i.e., several months) of severe food deprivation. Large daily deviations from energy balance are thus readily tolerated, and accommodated primarily by gains or losses of body fat (Abbott et al., 1988; Stubbs et al., 1995). Coefficients of variation for intra-individual variability in daily energy intake average ± 23 percent (Bingham et al., 1994); variations in physical activity are not closely synchronized with adjustments in food intake (Edholm et al., 1970). Thus, substantial positive as well as negative energy balances of several hundred kcal/d occur as a matter of course under free-living conditions among normal and overweight subjects. Yet over the long term, energy balance is maintained with remarkable accuracy. Indeed, during long periods in the


life of most individuals, gains or losses of adipose tissue are less than 1 to 2 kg over a year (McCargar et al., 1993), implying that the cumulative error in adjusting energy intake to expenditure amounts to less than 2 percent of energy expenditure.

Components of Energy Expenditure Basal and Resting Metabolism

The basal metabolic rate (BMR) describes the rate of energy expenditure that occurs in the postabsorptive state, defined as the particular condition that prevails after an overnight fast, the subject having not consumed food for 12 to 14 hours and resting comfortably, supine, awake, and motionless in a thermoneutral environment. This standardized metabolic state corresponds to the situation in which food and physical activity have minimal influence on metabolism. The BMR thus reflects the energy needed to sustain the metabolic activities of cells and tissues, plus the energy needed to maintain blood circulation, respiration, and gastrointestinal and renal processing (i.e., the basal cost of living). BMR thus includes the energy expenditure associated with remaining awake (the cost of arousal), reflecting the fact that the sleeping metabolic rate (SMR) during the morning is some 5 to 10 percent lower than BMR during the morning hours (Garby et al., 1987).

BMR is commonly extrapolated to 24 hours to be more meaningful, and it is then referred to as basal energy expenditure (BEE), expressed as kcal/24 h. Resting metabolic rate (RMR), energy expenditure under resting conditions, tends to be somewhat higher (10 to 20 percent) than under basal conditions due to increases in energy expenditure caused by recent food intake (i.e., by the "thermic effect of food") or by the delayed effect of recently completed physical activity (see Chapter 12). Thus, it is important to distinguish between BMR and RMR and between BEE and resting energy expenditure (REE) (RMR extrapolated to 24 hours).

Basal, resting, and sleeping energy expenditures are related to body size, being most closely correlated with the size of the fat-free mass (FFM), which is the weight of the body less the weight of its fat mass. The size of the FFM generally explains about 70 to 80 percent of the variance in RMR (Nelson et al., 1992; Ravussin et al., 1986). However, RMR is also affected by age, gender, nutritional state, inherited variations, and by differences in the endocrine state, notably (but rarely) by hypo- or hyperthyroidism. The relationships among RMR, body weight, and FFM are illustrated in Figures 5-1 and 5-2 (Owen, 1988), which show that differences in RMR relative to body weight among diverse individuals such as men, women, and athletes mostly disappear when RMR is considered relative to FFM.


FIGURE 5-1 Resting metabolic rates (RMR) are contrasted against the weights of 44 lean (o) and obese (•) healthy women, 8 of whom were athletes (©), and 60 lean (A) and obese (A) healthy men. Reprinted, with permission, from Owen (1988). Copyright 1988 by W.B. Saunders.

Weight (kg)

FIGURE 5-1 Resting metabolic rates (RMR) are contrasted against the weights of 44 lean (o) and obese (•) healthy women, 8 of whom were athletes (©), and 60 lean (A) and obese (A) healthy men. Reprinted, with permission, from Owen (1988). Copyright 1988 by W.B. Saunders.

30 40 50 60 70 80 90 100

FIGURE 5-2 Resting metabolic rates (RMR) are contrasted against the fat-free masses (FFM) of 44 lean (o) and obese (•) healthy women, 8 of whom were athletes (©), and 60 lean (A) and obese (A) healthy men. Reprinted, with permission, from Owen (1988). Copyright 1988 by W.B. Saunders.

30 40 50 60 70 80 90 100

FIGURE 5-2 Resting metabolic rates (RMR) are contrasted against the fat-free masses (FFM) of 44 lean (o) and obese (•) healthy women, 8 of whom were athletes (©), and 60 lean (A) and obese (A) healthy men. Reprinted, with permission, from Owen (1988). Copyright 1988 by W.B. Saunders.


BEE has been predicted from age, gender, and body size. Prediction equations were developed for each gender (WN Schofield, 1985) by pooling and analyzing reported measurements made in 7,393 individuals. A recent re-evaluation of all available data performed by Henry (2000) has led to a new set of predicting equations.

Thermic Effect of Food

It has long been known that food consumption elicits an increase in energy expenditure (Kleiber, 1975). Originally referred to as the Specific Dynamic Action (SDA) of food, this phenomenon is now more commonly referred to as the thermic effect of food (TEF). The intensity and duration of meal-induced TEF is determined primarily by the amount and composition of the foods consumed, mainly due to the metabolic costs incurred in handling and storing ingested nutrients (Flatt, 1978). Activation of the sympathetic nervous system elicited by dietary carbohydrate and by sensory stimulation causes an additional, but modest, increase in energy expenditure (Acheson et al., 1983). The increments in energy expenditure during digestion above baseline rates, divided by the energy content of the food consumed, vary from 5 to 10 percent for carbohydrate, 0 to 5 percent for fat, and 20 to 30 percent for protein. The high TEF for protein reflects the relatively high metabolic cost involved in processing the amino acids yielded by absorption of dietary protein, for protein synthesis, or for the synthesis of urea and glucose (Flatt, 1978; Nair et al., 1983). Consumption of the usual mixture of nutrients is generally considered to elicit increases in energy expenditure equivalent to 10 percent of the food's energy content (Kleiber, 1975). Since TEF occurs during a limited part of the day only, it can result in noticeable increases in REE if energy expenditure is measured during the hours following meals.


Birds and mammals, including humans, regulate their body temperature within narrow limits. This process, termed thermoregulation, can elicit increases in energy expenditure that are greater when ambient temperatures are below the zone of thermoneutrality. The environmental temperature at which oxygen consumption and metabolic rate are lowest is described as the critical temperature or thermoneutral zone (Hill, 1964). Because most people adjust their clothing and environment to maintain comfort, and thus thermoneutrality, the additional energy cost of thermoregulation rarely affects total energy expenditure to an appreciable extent. However, there does appear to be a small influence of ambient temperature on energy expenditure as described in more detail below.


Physical Activity

The energy expended for physical activity varies greatly among individuals as well as from day to day. In sedentary individuals, about two-thirds of total energy expenditure goes to sustain basal metabolism over 24 hours (the BEE), while one-third is used for physical activity. In very active individuals, 24-hour total energy expenditure can rise to twice as much as basal energy expenditure (Grund et al., 2001), while even higher total expenditures occur among heavy laborers and some athletes.

The efficiency with which energy from food is converted into physical work is remarkably constant when measured under conditions where body weight and athletic skill are not a factor, such as on bicycle ergometers (Kleiber, 1975; Nickleberry and Brooks, 1996; Pahud et al., 1980). For weight-bearing physical activities, the cost is roughly proportional to body weight. In the life of most persons, walking represents the most significant form of physical activity, and many studies have been performed to determine the energy expenditures induced by walking or running at various speeds (Margaria et al., 1963; Pandolf et al., 1977; Passmore and Durnin, 1955). Walking at a speed of 2 mph is considered to correspond to a mild degree of exertion, walking speeds of 3 to 4 mph correspond to moderate degrees of exertion, and a walking speed of 5 mph to vigorous exertion (Table 12-1, Fletcher et al., 2001). Over this range of speeds, the increment in energy expenditure amounts to some 60 kcal/mi walked for a 70-kg individual, or 50 kcal/mi walked for a 57-kg individual (see Chapter 12, Figure 12-4). The exertion caused by walking/jogging increases progressively at speeds of 4.5 mph and beyond, reaching 130 kcal/mi at 5 mph for a 70-kg individual.

The increase in daily energy expenditure is somewhat greater, however, because exercise induces an additional small increase in expenditure for some time after the exertion itself has been completed. This excess post-exercise oxygen consumption (EPOC) depends on exercise intensity and duration and has been estimated at some 15 percent of the increment in expenditure that occurs during exertions of the type described above (Bahr et al., 1987). This raises the cost of walking at 3 mph to 69 kcal/mi (60 kcal/mi X 1.15) for a 70-kg individual and to 58 kcal/mi (50 kcal/mi X 1.15) for a 57-kg individual. Taking into account the dissipation of 10 percent of the energy consumed on account of the thermic effect of food to cover the expenditure associated with walking, then walking 1 mile raises daily energy expenditure to 76 kcal/mi (69 kcal/mi X 1.1) in individuals weighing 70 kg, or 64 kcal/mi (58 kcal/mi X 1.1) for individuals weighing 57 kg. Since the cost of walking is proportional to body weight, it is convenient to consider that the overall cost of walking at moderate speeds is approximately 1.1 kcal/mi/kg body weight (75 kcal/mi/70 kg or 64 kcal/mi/57 kg). The effects of varia-


tions in body weights and the impact of various physical activities on energy expenditure are considered in more detail in Chapter 12.

Physical Activity Level

The level of physical activity is commonly described as the ratio of total to basal daily energy expenditure (TEE/BEE). This ratio is known as the Physical Activity Level (PAL), or the Physical Activity Index. Describing physical activity habits in terms of PAL is not entirely satisfactory because the increments above basal needs in energy expenditure, brought about by most physical activities where body weight is supported against gravity (e.g., walking, but not cycling on a stationary cycle ergometer), are directly proportional to body weight, whereas BEE is more nearly proportional to body weight075. However, PAL is a convenient comparison and is used in this report to describe and account for physical activity habits. The effect of variations in activities on PAL is described in Chapter 12.

Total Energy Expenditure

Total Energy Expenditure (TEE) is the sum of BEE (which includes a small component associated with arousal, as compared to sleeping), TEF, physical activity, thermoregulation, and the energy expended in depositing new tissues and in producing milk. With the emergence of information on TEE by the doubly labeled water (DLW) method (Schoeller, 1995), it has become possible to determine energy expenditure of infants, children, and adults under free-living conditions. TEE from doubly labeled water does not include the energy content of the tissue constituents laid down during normal growth and pregnancy or the milk produced during lactation, as it refers to energy expended during oxidation of energy-yielding nutrients to water and carbon dioxide.

It should be noted that direct measurements of TEE represent a distinct advantage over previous TEE evaluations, which had to rely on the factorial approach and on food intake data, which have limited accuracy due to the inability to reliably determine average physical activity cost and nutrient intakes.

Estimated Energy Requirement

Information on energy expenditure obtained by DLW studies conducted by a number of research units (see Appendix I) are used in this report to estimate energy requirements, taking into account estimates of the energy content of new body constituents during growth and preg-


nancy and of the milk produced during lactation. Energy expenditure depends on age and varies primarily as a function of body size and physical activity, both of which vary greatly among individuals. Recommendations about energy intake vary accordingly, and are also subject to the criterion that an individual adult's body weight should remain stable and within the healthy range.


Reported Energy Intake

The reported energy intakes of weight-stable subjects (i.e., those in energy balance) could, in principle, be used to predict energy requirements for weight maintenance. However, it is now widely recognized that reported energy intakes in dietary surveys underestimate usual energy intake (Black et al., 1993).

The most compelling evidence about underreporting has come from measurements of total energy expenditure (TEE) by the doubly labeled water (DLW) method (Schoeller, 1995). The use of a measure or estimate of TEE to validate instruments that measure food intake is dependent on the principle of energy balance. That is, in weight-stable adults, energy intake must equal TEE. By comparing reported energy intake to TEE, the accuracy of food intake reporting can be assessed (Goldberg et al., 1991a).

A large body of literature documents the underreporting of food intake, which can range from 10 to 45 percent depending on the age, gender, and body composition of individuals in the sample population (Johnson, 2000). Underreporting tends to increase as children grow older (Livingstone et al., 1992b), is worse among women than in men (Johnson et al., 1994), and is more pronounced among overweight and obese than among lean individuals (Bandini et al., 1990a; Lichtman et al., 1992; Prentice et al., 1986). Low socioeconomic status, characterized by low income, low educational attainment, and low literacy levels increase the tendency to underreport energy intakes (Briefel et al., 1997; Johnson et al., 1998; Price et al., 1997; Pryer et al., 1997). Ethnic differences affecting sensitivities and psychological perceptions relating to eating and body weight can also affect the accuracy of reported food intakes (Tomoyasu et al., 2000). Finally, individuals with infrequent symptoms of hunger under-report to a greater degree than those who experience frequent hunger (Bathalon et al., 2000).

There is some evidence suggesting that underreporters often fail to report foods perceived to be bad or sinful, such as cakes/pies, savory


snacks, cheese, fried potatoes, meat mixtures, soft drinks, spreads, condiments, and generally foods known to be high in fat (Bingham and Day, 1997; Krebs-Smith et al., 2000). Reported intakes of added sugars are also significantly lower than that consumed, due in part to the frequent omission of snack foods from 24-hour food recording (Poppitt et al., 1998).

Finally, there is no objective evidence for the existence of "small eaters," individuals who can survive long term on the low energy intakes that they report in dietary surveys (Black, 1999; Lichtman et al., 1992; Prentice et al., 1986). Clearly, it is no longer tenable to base energy requirements on self-reported food consumption data.

Factorial Approach

Previous Recommended Dietary Allowances for energy (NRC, 1989) used the factorial method to estimate TEE. This method calculates TEE using information on the amount of time devoted to different activities and the energy costs of each activity throughout a theoretical 24-hour period. The factorial method allowed theoretical estimation of TEE for a defined activity pattern (using measured average costs of standard activities and theoretical activity duration). Thus, mean expected energy requirements for different levels of physical activity were defined.

However, there are recognized problems with the factorial method and doubts about the validity of energy requirement predictions based on it (Roberts et al., 1991). The first problem is that there are a wide range of activities and physical efforts performed during normal life, and it is not feasible to measure the energy cost of each. Another concern with the factorial method is that the measurement of the energy costs of specific activities imposes constraints (due to mechanical impediments associated with performing an activity while wearing unfamiliar equipment) that may alter the measured energy costs of different activities. Although generalizations are essential in trying to account for the energy costs of daily activities, substantial errors may be introduced. In addition, energy expenditure during sleep, once considered to be equivalent to basal metabolic rate (BMR), is generally somewhat lower (-5 to -10 percent) than BMR (Garby et al., 1987).

Also, and perhaps most importantly, the factorial method only takes into account activities that can be specifically accounted for (e.g., sleeping, walking, household work, occupational activity, and so on). However, 24-hour room calorimeter studies have shown that a significant amount of energy is expended in spontaneous physical activities, some of which are part of a sedentary lifestyle (Ravussin et al., 1986; Zurlo et al., 1992). In addition, some individuals manifest a substantial amount of fidgeting. Together these were reported to average about 350 kcal/d, ranging from


140 to 700 kcal/d (Ravussin et al., 1986). Thus, the factorial method is bound to underestimate usual energy needs (Durnin, 1990; Roberts et al., 1991).

Most comparisons of the factorial approach with DLW determinations of TEE have shown significantly higher measured values for TEE than predicted by the factorial method (Haggarty et al., 1994; Jones et al., 1997; Roberts et al., 1991; Sawaya et al., 1995). In two direct comparisons of factorial energy requirement estimates with DLW, one confirmed that the factorial method underestimated energy needs (Leonard et al., 1997), while the other found no difference between the methods in an elderly population with a mean age of 70 years (Morio et al., 1997).

Measurement of Energy Expenditure by Doubly Labeled Water

The DLW method is a relatively new technique that measures TEE in free-living individuals. It was originally proposed and developed by Lifson for use in small animals (Lifson and McClintock, 1966; Lifson et al., 1955), but has been adapted for human studies and extensively used (Schoeller et al., 1986). Two stable isotopic forms of water (H218O and 2h 2O) are administered, and their disappearance rates from a body fluid (i.e., urine or blood) are monitored for a period of time, optimally equivalent to 1 to 3 half lives for these isotopes (7 to 21 days in most human subjects). The disappearance rate of 2^O relates to water flux, while that of H218O reflects water flux plus carbon dioxide (CO2) production rate, because of the rapid equilibration of the body water and bicarbonate pools by carbonic anhydrase (Lifson et al., 1949). The difference between the two disappearance rates can therefore be used to calculate the CO2 production rate, and with knowledge of the composition of the diet, TEE can be calculated.

To predict TEE from a measurement of CO2 production, it is necessary to have an estimate of the average respiratory quotient (RQ = ratio of CO2 produced to the O2 consumed) of the subject during the period of measurement. This is because the energy released per liter of CO2 varies with the RQ and hence with the substrate mix oxidized by the body (Elia, 1991). The ratio of the CO2 produced to the O2 consumed by the biological oxidation of a representative sample of the diet is commonly referred to as the food quotient, or FQ (Flatt, 1978).

Short-term measurements of RQ by indirect calorimetry are not useful for the DLW technique because RQ varies markedly during the day, particularly after meals. It is therefore more accurate to estimate the average RQ from information on the subjects' dietary intake. When energy balance prevails, the average RQ is equal to the FQ. If substantial gains or losses of body constituents are known to occur during the period of measurement,


appropriate adjustments can be made in estimating the average RQ. Although food reports are inaccurate for measuring total energy intake, FQ calculations from food records can be used because FQ has a relatively small effect on DLW measurements of TEE.

Several validations of the DLW study have been conducted in which DLW-derived estimates of TEE were compared with measurements of TEE in whole-body calorimeters (Table 5-2). Although studies in whole-body calorimeters do not mimic normal life conditions, they do allow for an exact comparison of the DLW method with classic calorimetry, which is considered the most reliable measurement of energy expenditure. As shown in Table 5-2, there is a close agreement between means for the CO2 production rate determined by the two methods in all the validation studies. The precision of DLW measurements, as assessed by the variability of individual DLW measurements from the calorimetry assessments, ranged from -2.5 to 5.9 percent in the different studies. These validation studies show that the DLW method can provide an accurate assessment of the CO2 production rate and hence TEE in a wide range of human subjects.

One particular advantage of the DLW method is that it provides an index of TEE over a period of several days. Because 1 to 3 half-lives of isotope disappearance are needed for changes in isotopic abundance to be measured accurately by mass spectrometry, optimal time periods for DLW measurements of TEE range from 1 to 3 weeks in most groups of individuals (Schoeller, 1983). Thus, in contrast to other techniques, DLW can provide TEE estimates over biologically meaningful periods of time that can reduce the impact of spontaneous daily variations in physical activity. Moreover, because DLW is noninvasive (requiring only that the subject drink the stable isotopes and provide at least three urine samples over the study period), measurements can be made in subjects leading their normal daily lives. A critical mass of DLW data has now accumulated on a wide range of age groups and body sizes, so that the estimated energy requirements provided in this report could be based on DLW measurements of TEE.

The available DLW data (Appendix I) are not from randomly selected individuals, except in the recent study of Bratteby and coworkers (1997), and they do not constitute a sample representative of the population of the United States and Canada. However, the measurements were obtained in men, women, and children whose ages, body weights, heights, and physical activities varied over wide ranges. At the present time, a few age groups are underrepresented and interpolations had to be performed in these cases. Thus, while the available DLW data do not yet provide an entirely satisfactory set of data, they nevertheless offer the best currently available information.


A second potential criticism of using DLW-derived estimates of TEE as a basis for estimating energy requirements is that the approach assumes that TEE is relatively unaffected by fluctuations in energy balance. Although there is some capacity for TEE to increase or decrease spontaneously when energy intakes increase or decrease, these changes are small and attenuate the effect of energy imbalances only modestly (Levine et al., 1999; Roberts et al., 1990). Indeed, overfeeding studies show that overeating is inevitably accompanied by substantial weight gain, and that reduced energy intake induces weight loss (Saltzman and Roberts, 1995). Thus, although there may be some adaptive capacity to alter TEE in response to changes in dietary energy intake, the DLW-based evaluation of TEE at approximate weight maintenance provides an appropriate estimate of energy expenditure from which energy requirements for maintaining energy balance can be derived.

Body Mass Index


A growing literature supports the use of the body mass index (BMI, defined as weight in kilograms divided by the square of height in meters) as a predictor of the impact of body weight on morbidity and mortality risks (Seidell et al., 1996; Troiano et al., 1996). As an index of healthy weight and as a predictor of morbidity and mortality risk, it has supplanted weight-for-height tables, which were derived primarily from white populations and relied on questionable estimates of frame size (NHLBI/NIDDK, 1998). BMI, although only an indirect indicator of body composition, is now used to classify underweight and overweight individuals.

While sophisticated techniques are available to precisely measure fat-free mass (FFM) and fat mass (FM) of individuals, these techniques are used mainly in research protocols. For most clinical and epidemiological applications, body size is judged on the basis of the BMI, which is easy to determine, accurate, and reproducible. The main disadvantages of relying on BMI are that (1) it does not reliably reflect body fat content, which is an independent predictor of health risk, and (2) very muscular individuals may be misclassified as overweight (Willett et al., 1999).

The National Institutes of Health (NIH) clinical guidelines on the identification, evaluation, and treatment of normal, overweight and obese adults and the World Health Organization have defined BMI cutoffs for adults over 19 years of age, regardless of age or gender (NHLBI/NIDDK, 1998; WHO, 1998). Underweight is defined as a BMI of less than 18.5 kg/m2, overweight as a BMI from 25 up to 30 kg/m2, and obese as a BMI of 30 kg/m2 or higher. A healthy or desirable BMI is considered to be from 18.5 up to


TABLE 5-2 Comparison of Carbon Dioxide Production Rates Measured by the Doubly Labeled Water Method and Indirect Calorimetry in Humans




Time (d)

Coward et al., 1984

Adults, in energy balance^



Klein et al., 1984

Adults, in energy balance



Schoeller and Webb, 1984

Adults, in energy balance



Roberts et al., 1986

Preterm infants, growing



Schoeller et al., 1986

Adults, in energy balance "Low" dose "High" dose


4 4

Jones et al., 1987

Infants, after surgery



Westerterp et al., 1988

Adults, in energy balance Sedentary Active

5 4


Riumallo et al., 1989




Seale et al., 1990

Adults, in energy balance



Ravussin et al., 1991

Obese adults, in energy balance



Schulz et al., 1992

Adults, in energy balance



Seale and Rumpler, 1997

Adults, in energy balance



a Calculations for pool: I = 2-pool model using measured pool sizes as proposed by Coward et al. (1984) and detailed by Roberts et al. (1986), S = single-pool model as described by Lifson et al. (1955) and Lifson and McClintock (1966), F = 2-pool model with fixed ratio of 1.03 between pool sizes as described by Schoeller et al. (1986). b Calculations for fractionated water loss: 50 = assumed to be 50 percent of total water output, 25 = assumed to be 25 percent of total water output, M = measured or calcu-



ti/2 (d)

















5.9 ± 7.6





-1.4 ± 4.8





5.0 ± 9.5





1.7 ± 4.5





-0.9 ± 6.2





1.4 ± 3.9





-1.0 ± 7.0







-1.04 ± 0.63




-2.5 ± 5.8







lated from data on water balance, P = assumed to be proportional to carbon dioxide output (Jones et al., 1987; Schoeller et al., 1986).

c Growth correction: L = no change or linear change in pool sizes, E = exponential change in pool sizes.

d Energy balance indicates that induction of positive or negative energy balance was not part of study protocol.


25 kg/m2, a view adopted in this report. Although the healthy BMI range is the result of a consensus, there are reasons to suggest that slightly different mortality-based BMI ranges may be appropriate for different populations (NHLBI/NIDDK, 1998).

In establishing the 2000 Dietary Guidelines, the U.S. Departments of Agriculture and of Health and Human Services set the "healthy weight" upper limit at a BMI of 24.99 kg/m2 for adult men and women because mortality increases significantly beyond this point (USDA/HHS, 2000). Although the incidence of diabetes, hypertension, and coronary heart disease begins to increase even below this cutoff, a BMI of 24.99 kg/m2 is considered a reasonable upper limit of healthy weight. The lower BMI limit of 18.5 kg/m2 is not as well substantiated. The point at which low BMI poses a health risk is poorly defined. The abil

Quick Weight Loss Action

Quick Weight Loss Action

Why Indulge In Self-Pity When You Can Do Something About Your Weight Now. Say Goodbye to Your Weight Problems That Have Only Make Your Life Nothing But Miserable. Have you often felt short-changed because of your weight or physical appearance?

Get My Free Ebook

Post a comment