- Email: [email protected]
Dietary guidelines: some issues to consider before initiating revisions
Dietary guidelines: Some issues to consider before initiating revisions GEORGE H. BEATON, PhD
ABSTRACT This paper presents two main questions that need to be answered before deciding to revise dietary guidelines: Is there really a need to revise existing guidelines? and How is dietary risk management to be integrated into dietary guidelines? Any group assigned the task of evaluating and possibly revising existing guides must answer these questions. It is also critically important to specify the exact target group and purpose for any dietary guideline since the questions need to be answered in context. This paper addresses some of the implications of the recent report on Dietary Reference Intakes for macronutrients in its venture into the area of risk management. A major conclusion of the paper is that nutrition has entered a new era and that henceforth, as never before, risk management, the assessments of likelihood of benefit and risks of harm—as well as costs human and financial of our actions—and recommendations must take a very prominent position in our work. This is not a role for which most nutritionists are prepared and hence it must be seen as a major challenge in the preparation of food guides in the future. J Am Diet Assoc. 2003;103:S56-S59.
T
he symposium speakers offered a series of excellent papers discussing approaches and goals in developing food guides. Very appropriately, the emphasis was on foodbased guides (1) as compared with numeric targets (dietary goals and recommended intakes). Nevertheless, in developing food-based advice for the public, it remains necessary to take into account estimates of human nutrient needs and the expected relationship of food use and nutrient intake to health. In this article, I emphasize this aspect of the design and evaluation of food guides. In the symposium, two issues were not explicitly addressed. Both seem to have considerable relevance in Canada and the United States today: ■ How does one decide that an existing guideline needs revision? In parallel, one must ask about the criteria to be applied in deciding that a new or revised guideline is satisfactory. ■ How should “risk management” be integrated into the development of new guidelines or revision of old guides? Who is responsible for addressing content of a guideline in the risk management sense? IS THERE A NEED TO REVISE EXISTING GUIDELINES? Taking account of comments within the presented papers, as well as conversations during the breaks, six reasons for revision of existing dietary guidelines were identified. G. H. Beaton is professor emeritus in the Department of Nutritional Sciences, Faculty of Medicine, University of Toronto, Toronto, Ontario, Canada. Address correspondence to: George H. Beaton, PhD, 9 Silverview Dr., Willowdale, Ontario, Canada M2M 2B2. E-mail: [email protected]. Copyright © 2003 by the American Dietetic Association. 0002-8223/03/10312-2010$30.00/0 doi: 10.1016/j.jada.2003.09.038
S56 / December 2003 Suppl 2 Volume 103 Number 12
The Guidelines Need to Be Revised to Comply with the Emerging DRIs This reason assumes that existing guidelines are not in compliance with the Dietary Reference Intakes (DRIs). The new DRIs have a very different structure from the old Recommended Dietary Allowances and Reference Nutrient Intakes. There is no a priori reason to assume that existing food guides would be judged inappropriate if the new DRIs are correctly interpreted and applied (2). Obviously, this must be assessed by direct examination of the existing guideline. One possible indication for major change in existing guidelines will be discussed in the section of this article that focuses on risk management. The Guidelines Must Be Revised to Reflect the Entry of New/Fortified Foods in the Marketplace A specific example relates to the availability of calcium-fortified citrus drinks. It was suggested that the traditional milk product group should now be reconfigured to recognize, and include, other potentially important sources of calcium including the fortified citrus drinks. This proposal challenges the rationale that should support food guides. If the guides are seen as guides to sources of particular nutrients, as is often implied in promotional brochures, the proposal makes sense and traditional food groupings may be outlived. However, if food guides are intended to promote desirable patterns of food use, consistent with good health in the population, the suggestion to include a fruit drink in the milk group is illogical and even counterproductive. The resolution of this matter depends on the precise purpose of the food guide, a theme that was discussed in the symposium papers but sometimes is not explicitly declared. Guidelines in Canada and the United States Need to Be Harmonized to Facilitate Commerce and Health Promotion (Trans-Border Media Advertising) This may or may not be an important argument. It is invalid if it can be demonstrated that food selection practices (or food composition) are different between Canada and the United States in nutritionally important ways. Undoubtedly, senior government policy, influenced by trade considerations, will determine whether this argument is a major reason for revision of the food guides. Guidelines Need to Be Harmonized with Those Emerging in Europe Given recent developments in international trade regulations and emerging international nutrition standards, this suggestion may have as much— or more—validity than the previous proposal for harmonization between Canada and United States. Again, it is likely a policy issue to be settled at senior levels of government. The Guidelines Need to Be Revised Because Existing Guidelines Are Ineffective If there is evidence to this effect and also evidence that dietary guidelines are effective in any settings, this would be a very strong justification for attempted revision. There was such evidence (3) in relation to the old “Canada’s Food Rules” and radical revision was made. This author is not in a position to judge whether the particular guides that now exist are effective or not.
The Guidelines Must Be Revised Because the Public and Professionals Expect New Guidelines Approximately Every Five Years I suspect this is the most prevalent reason for revision and the weakest of all; in the United States, such review, but not necessarily revision, is mandated. There is an assumption that each change made is an improvement. The evidence to support such an assumption is dubious. There is a very old adage that seems to fit: “If it ain’t broke, don’t fix it.” From the list of six possible reasons for revision of the food guide, there seem to remain three possibly important reasons that would justify the undertaking of a revision of existing guides: ■ Evidence that existing guidelines, if followed, would promote dietary patterns that are nutritionally inadequate. ■ Evidence that changes are needed to address new positions on some aspects of the macronutrients or to ensure that, as much as possible, suggested dietary patterns are protective against nutrient inadequacy and also against excess intake. ■ Evidence that existing guides are ineffective and that more effective guides could be designed. Another reason of merit that was not voiced at the symposium is: ■ Evidence that we have reached a time when we can conclude that it would be more effective to abandon the notion of a single national food guide in favor of sets of guides developed for regions/groups with distinctly different food selection practices and consumption patterns. The first 2 of the original 6 reasons can be addressed in objective, controlled evaluations of existing guides. In turn, “passing” such assessments are appropriate criteria for any new proposal. An obvious example of the fourth of the original six reasons would be a food guide developed explicitly for vegetarians; guidelines in the United States for the substantial MexicanAmerican population in some regions; or, in Canada, guidelines for the growing Asian populations concentrated in some of our large cities. I am aware that group-specific interpretations of national guidelines are prepared and distributed, but I argue that this is not the same as developing de novo food guides for these substantial population groups. The third and fourth of the six reasons are obviously variants on a theme. It is beyond the competence of this author to specify how one could or should evaluate the effectiveness of an existing guide or of a proposed new guide(s). Ultimately, effectiveness implies demonstrable change toward the desired goal. Here this means adoption of patterns of food use that provide adequate nutrient intakes in the population (low prevalence of apparently inadequate intakes) and, more importantly, result in improvement in population health and reduction in chronic disease rates. Effectiveness then implies much more than knowledge transfer. It would seem that effectiveness assessment would have to address behavioral responses or something equivalent and, ultimately, impact on health. In turn, that implies the need for a realistic evaluation of outcomes of dietary change (in either direction) that can truly be expected. I phrase it that way intentionally, as we are all prone to exaggerate potential effects to promote whatever we think is good. There is now clear need to be more realistic in our assessments if we are to facilitate prudent judgments on some of the tradeoffs in public policy and allocation of assets that need to be considered. Preparation and dissemination of Food
Supplement to the Journal of THE AMERICAN DIETETIC ASSOCIATION / S57
Guides is not an academic exercise. Rather, it is a very important action with direct influence on resource allocation by government, by industry and by households. It behooves us all to ensure that the advice, if implemented, will provide benefits that justify the costs (monetary and otherwise). RISK MANAGEMENT AND DIETARY GUIDELINES Whether we realize it or not, nutrition has entered a new, long overdue phase. The nutrition community faces something that few in that community have ever considered very seriously. I refer to the need to approach our work—and any guidelines we promote—with a full awareness that today’s role is not just to try to avoid the potential harm of nutrient inadequacies but also to avoid potential harm of excess intakes. The DRI reports have brought this to the forefront by presenting estimates of Tolerable Upper Intake Levels (UL) (4). The UL was originally defined as “the highest level of daily nutrient intake that is likely to pose no risk of adverse health effects to almost all individuals in the general population” (4). That definition is repeated in all of the DRI reports and applied wherever there was sufficient evidence to justify setting a UL. In the most recent report addressing the macronutrients (5), a new reason for not setting a UL was invoked. Explicitly, evidence that would justify a UL for saturated and trans-fatty acids and for dietary cholesterol was presented and accepted by the scientific panel. However, no UL was presented “because trans-fatty acids are unavoidable in ordinary non-vegan diets, consuming 0 [the scientifically justified UL] would require significant changes in patterns of dietary intake. Such adjustments may introduce undesirable effects (eg, elimination of commercially prepared foods and dairy products and meats that contain trans-fatty acids may result in inadequate intakes of protein and certain micronutrients) and unknown and unquantified health risks” (5). Similar statements are presented for saturated fatty acids and cholesterol. This was an unfortunate decision on the part of the panel and the Institute of Medicine. A UL should have been presented based on scientific evidence. The risk-management considerations mentioned as justification for not setting the UL should have been weighed during the development of food-based dietary advice when the full range of alternative strategies would be considered. The approach to risk management involves modeling of expected intakes and looking at both potential inadequacies and excesses of nutrients with identified functional effects. In addition, one needs to think very carefully about what the actual risks (health consequences of inadequacy and excess for the nutrient in question) may be, and their importance to public health. The DRIs have mostly been designed to cover even minor health effects, creating a risk that such effects may or may not be deemed very important to the health of the group or population. Unless important risk is believed to exist, there may be little or no need for any attempt to raise or lower intake in comparison with intakes in the general diet or expected with compliance to existing guides. There is no question but that such modeling is made much more difficult with the introduction of voluntary fortification. The actions of the DRI panel that addressed macronutrients became even more questionable when the panel prepared the chapter on “Macronutrients and Healthful Diets,” the partial equivalent of the old US dietary goals. These were not based on the preceding scientific reviews. Rather, the goals seem to derive from a primary concern about diet and cardiovascular disS58 / December 2003 Suppl 2 Volume 103 Number 12
Table Comparison of macronutrient intakea goals for adults, United States and the World Health Organization (WHO) Nutrient
DRIb
WHO
Total fat Total carbohydrate Added sugar Total protein
20-35 45-65 ⬍25 10-35
15-30 55-75 ⬍10 10-15
a
Percent of energy intake. DRI⫽Dietary reference intakes; represents US recommendations.
b
ease and the assumption that the only effective way to minimize intakes of saturated and trans-fatty acids and dietary cholesterol would be to control total fat intake. It appears to have drawn heavily on clinical experience in developing dietary prescriptions for use in treating hyperlipidemias. Suggested carbohydrate and protein intakes more or less followed from that. This was a direct intrusion into risk management. If risk management was to be done in that setting, one might have expected greater discussion of other approaches to lowering of intakes of the two classes of fatty acids and dietary cholesterol. In the development of food guides, one must expect that there will be conflicts and overlaps between addressing needs for micronutrients and attempting to balance the macronutrients. The DRI report (5) has illustrated what scientific evidence does and does not establish and, by inference, where the opinions and judgments of informed people take over. What is not made clear is the point at which responsibility for judgments should pass from laboratory and clinical experts to those with expertise germane to foods and food consumption. In the coming months and years, it will be important to recognize these distinctions as a part of risk management in the design and use of food guides. It should be expected that epidemiology will play an ever-increasing role in the future development of food-based dietary guidelines. Earlier in this paper, I mentioned that two possible arguments for revising existing Canadian food guides were related to harmonization with either the US or European approaches. In this context, it is interesting to compare the numeric goals specified in the DRI report (5, Chapter 11) and in the very recent WHO report (6) (see Table), noting also that the WHO approach appears to have wide support in Europe. Superficially the two sets of goals seem similar and there should be no problem of compatibility. However, the DRI goals explicitly refer to limits for usual intake of the individual; the WHO goals explicitly refer to population or large group mean intakes. Therein lies the serious incompatibility. By the WHO criterion, Canada and the United States have achieved or nearly achieved goals for impact of fat intake on health. By comparison, applying the stated DRI goals, both countries face horrendous problems with 30% to 50% of the population consuming too much fat. Is it not reasonable to ask which of these judgments are more relevant to the planning of food-based dietary guidelines? In the absence of scientific evidence causally linking total fat to a detrimental health outcome (5), one must ask also whether we should focus on the identified risk factors (trans- and saturated fatty acids and dietary cholesterol)—rather than on total fat—in developing food-based guides? The importance of these two questions is why I termed the rejection of a UL to be an unfortunate decision by the DRI
panel and the Institute of Medicine; it seems to misdirect attention. Although not a customary inclusion in the definition of “foodbased” approaches to dietary guidelines, it is important to recognize that evidence relating to nutrient effects sometimes fails to explain well-documented and consistent evidence relating food use and chronic disease. The most obvious example is the relationship of consumption of fruits and vegetables to certain types of cancer. I think that in the future, we will see/must see a much greater attention directed to food-based evidence when developing food guides. For example, in addition to the fruit and vegetable evidence, one would have to weigh the limited epidemiologic evidence suggesting that high intakes of milk might have detrimental effects on osteoporosis at the same time it is argued that high intakes of calcium may be beneficial. We have much to learn about food intake patterns and chronic disease. Only a few centers seem to be addressing this; interest is higher in Europe than in North America, where reliance continues to be placed on laboratory studies and clinical trials rather than epidemiologic investigations and population health. Advocating direct pharmaceutical supplements might be a better choice than always demanding that nutrients come from foods although this is heresy in current thinking. These types of considerations go well beyond asking whether intakes are marginally inadequate in comparison to DRI estimates of need. We must now weigh predicted health outcomes and make our tradeoffs, not just retreat to old truisms. That is the difficult but extremely important era that nutrition has entered. Our paradigm has changed, and we must be prepared to change with it. That part of the challenge of developing food guides was not discussed in the symposium. CONCLUSION The original purpose of this paper was to stimulate discussion in the workshop following the symposium. The paper has been heavily revised for publication, particularly by adding specifics in the discussion of risk management, but hopefully continues to serve its intended purpose, now provoking a much broader audience into thinking and discussion. The paper is intentionally provocative but the message is clear: Nutrition is at a conceptual and philosophical turning point. The DRI reports are
not the end. They are but a piece of the beginning. We must now change how we think and teach about nutrition—and how we approach the design food guides. The future in our field will increasingly demand evidence that actions and expenditures yield substantial (in public health terms) benefits with minimal real risks and at a cost that can be justified by expected benefits. When we adopt “evidence-based’ approaches, we must all assess the continuity and consistency of the actual evidence. I fear that that has not been a universal practice in the past. I sincerely hope that the future offers better. Revision of food guides, if undertaken, may be a place to start. For planning, including the development of dietary guidelines, one needs to know not only existing patterns of food use and estimated requirements— one must also know the health conditions of the population and be able to make some inference about the likely improvement— or otherwise— of health that could be expected with a specific change in food use. All of the health outcomes in which we are interested are multifactorial. It was not the mandate of the DRI system to look at what is very loosely called attributable risk, but those planning for the improvement or preservation of population health must certainly address this. What real benefit is to be expected? Is it worth the cost? References 1. FAO/WHO Preparation and Use of Food-based Dietary Guidelines. WHO Technical Rpt Ser. No. 880, Geneva. 1998. 2. Institute of Medicine. Dietary Reference Intakes: Applications in Dietary Assessment. Washington, DC: National Academy Press; 2000. 3. McClinton PE, Milne H, Beaton GH. An evaluation of food habits and nutrient intakes in Canada: Design of effective food guides. Can J Public Health. 1971;62:139-146. 4. Institute of Medicine. Dietary Reference Intakes: A Risk Assessment Model for Establishing Upper Intake Levels for Nutrients. Washington, DC: National Academy Press; 1998. 5. Institute of Medicine. Dietary Reference Intakes: Energy, Carbohydrate, Fiber, Fat, Fatty Acids, Cholesterol, Protein, and Amino Acids [Prepublication copy, unedited proofs]. Vols 1 and 2. Washington, DC: National Academy Press; 2002. 6. WHO/FAO Diet, Nutrition and the Prevention of Chronic Diseases Report of a Joint WHO/FAO Expert Consultation. Geneva: The World Health Organization; 2003. WHO Tech Report Ser No. 919.
Supplement to the Journal of THE AMERICAN DIETETIC ASSOCIATION / S59
ABSTRACT This paper presents two main questions that need to be answered before deciding to revise dietary guidelines: Is there really a need to revise existing guidelines? and How is dietary risk management to be integrated into dietary guidelines? Any group assigned the task of evaluating and possibly revising existing guides must answer these questions. It is also critically important to specify the exact target group and purpose for any dietary guideline since the questions need to be answered in context. This paper addresses some of the implications of the recent report on Dietary Reference Intakes for macronutrients in its venture into the area of risk management. A major conclusion of the paper is that nutrition has entered a new era and that henceforth, as never before, risk management, the assessments of likelihood of benefit and risks of harm—as well as costs human and financial of our actions—and recommendations must take a very prominent position in our work. This is not a role for which most nutritionists are prepared and hence it must be seen as a major challenge in the preparation of food guides in the future. J Am Diet Assoc. 2003;103:S56-S59.
T
he symposium speakers offered a series of excellent papers discussing approaches and goals in developing food guides. Very appropriately, the emphasis was on foodbased guides (1) as compared with numeric targets (dietary goals and recommended intakes). Nevertheless, in developing food-based advice for the public, it remains necessary to take into account estimates of human nutrient needs and the expected relationship of food use and nutrient intake to health. In this article, I emphasize this aspect of the design and evaluation of food guides. In the symposium, two issues were not explicitly addressed. Both seem to have considerable relevance in Canada and the United States today: ■ How does one decide that an existing guideline needs revision? In parallel, one must ask about the criteria to be applied in deciding that a new or revised guideline is satisfactory. ■ How should “risk management” be integrated into the development of new guidelines or revision of old guides? Who is responsible for addressing content of a guideline in the risk management sense? IS THERE A NEED TO REVISE EXISTING GUIDELINES? Taking account of comments within the presented papers, as well as conversations during the breaks, six reasons for revision of existing dietary guidelines were identified. G. H. Beaton is professor emeritus in the Department of Nutritional Sciences, Faculty of Medicine, University of Toronto, Toronto, Ontario, Canada. Address correspondence to: George H. Beaton, PhD, 9 Silverview Dr., Willowdale, Ontario, Canada M2M 2B2. E-mail: [email protected]. Copyright © 2003 by the American Dietetic Association. 0002-8223/03/10312-2010$30.00/0 doi: 10.1016/j.jada.2003.09.038
S56 / December 2003 Suppl 2 Volume 103 Number 12
The Guidelines Need to Be Revised to Comply with the Emerging DRIs This reason assumes that existing guidelines are not in compliance with the Dietary Reference Intakes (DRIs). The new DRIs have a very different structure from the old Recommended Dietary Allowances and Reference Nutrient Intakes. There is no a priori reason to assume that existing food guides would be judged inappropriate if the new DRIs are correctly interpreted and applied (2). Obviously, this must be assessed by direct examination of the existing guideline. One possible indication for major change in existing guidelines will be discussed in the section of this article that focuses on risk management. The Guidelines Must Be Revised to Reflect the Entry of New/Fortified Foods in the Marketplace A specific example relates to the availability of calcium-fortified citrus drinks. It was suggested that the traditional milk product group should now be reconfigured to recognize, and include, other potentially important sources of calcium including the fortified citrus drinks. This proposal challenges the rationale that should support food guides. If the guides are seen as guides to sources of particular nutrients, as is often implied in promotional brochures, the proposal makes sense and traditional food groupings may be outlived. However, if food guides are intended to promote desirable patterns of food use, consistent with good health in the population, the suggestion to include a fruit drink in the milk group is illogical and even counterproductive. The resolution of this matter depends on the precise purpose of the food guide, a theme that was discussed in the symposium papers but sometimes is not explicitly declared. Guidelines in Canada and the United States Need to Be Harmonized to Facilitate Commerce and Health Promotion (Trans-Border Media Advertising) This may or may not be an important argument. It is invalid if it can be demonstrated that food selection practices (or food composition) are different between Canada and the United States in nutritionally important ways. Undoubtedly, senior government policy, influenced by trade considerations, will determine whether this argument is a major reason for revision of the food guides. Guidelines Need to Be Harmonized with Those Emerging in Europe Given recent developments in international trade regulations and emerging international nutrition standards, this suggestion may have as much— or more—validity than the previous proposal for harmonization between Canada and United States. Again, it is likely a policy issue to be settled at senior levels of government. The Guidelines Need to Be Revised Because Existing Guidelines Are Ineffective If there is evidence to this effect and also evidence that dietary guidelines are effective in any settings, this would be a very strong justification for attempted revision. There was such evidence (3) in relation to the old “Canada’s Food Rules” and radical revision was made. This author is not in a position to judge whether the particular guides that now exist are effective or not.
The Guidelines Must Be Revised Because the Public and Professionals Expect New Guidelines Approximately Every Five Years I suspect this is the most prevalent reason for revision and the weakest of all; in the United States, such review, but not necessarily revision, is mandated. There is an assumption that each change made is an improvement. The evidence to support such an assumption is dubious. There is a very old adage that seems to fit: “If it ain’t broke, don’t fix it.” From the list of six possible reasons for revision of the food guide, there seem to remain three possibly important reasons that would justify the undertaking of a revision of existing guides: ■ Evidence that existing guidelines, if followed, would promote dietary patterns that are nutritionally inadequate. ■ Evidence that changes are needed to address new positions on some aspects of the macronutrients or to ensure that, as much as possible, suggested dietary patterns are protective against nutrient inadequacy and also against excess intake. ■ Evidence that existing guides are ineffective and that more effective guides could be designed. Another reason of merit that was not voiced at the symposium is: ■ Evidence that we have reached a time when we can conclude that it would be more effective to abandon the notion of a single national food guide in favor of sets of guides developed for regions/groups with distinctly different food selection practices and consumption patterns. The first 2 of the original 6 reasons can be addressed in objective, controlled evaluations of existing guides. In turn, “passing” such assessments are appropriate criteria for any new proposal. An obvious example of the fourth of the original six reasons would be a food guide developed explicitly for vegetarians; guidelines in the United States for the substantial MexicanAmerican population in some regions; or, in Canada, guidelines for the growing Asian populations concentrated in some of our large cities. I am aware that group-specific interpretations of national guidelines are prepared and distributed, but I argue that this is not the same as developing de novo food guides for these substantial population groups. The third and fourth of the six reasons are obviously variants on a theme. It is beyond the competence of this author to specify how one could or should evaluate the effectiveness of an existing guide or of a proposed new guide(s). Ultimately, effectiveness implies demonstrable change toward the desired goal. Here this means adoption of patterns of food use that provide adequate nutrient intakes in the population (low prevalence of apparently inadequate intakes) and, more importantly, result in improvement in population health and reduction in chronic disease rates. Effectiveness then implies much more than knowledge transfer. It would seem that effectiveness assessment would have to address behavioral responses or something equivalent and, ultimately, impact on health. In turn, that implies the need for a realistic evaluation of outcomes of dietary change (in either direction) that can truly be expected. I phrase it that way intentionally, as we are all prone to exaggerate potential effects to promote whatever we think is good. There is now clear need to be more realistic in our assessments if we are to facilitate prudent judgments on some of the tradeoffs in public policy and allocation of assets that need to be considered. Preparation and dissemination of Food
Supplement to the Journal of THE AMERICAN DIETETIC ASSOCIATION / S57
Guides is not an academic exercise. Rather, it is a very important action with direct influence on resource allocation by government, by industry and by households. It behooves us all to ensure that the advice, if implemented, will provide benefits that justify the costs (monetary and otherwise). RISK MANAGEMENT AND DIETARY GUIDELINES Whether we realize it or not, nutrition has entered a new, long overdue phase. The nutrition community faces something that few in that community have ever considered very seriously. I refer to the need to approach our work—and any guidelines we promote—with a full awareness that today’s role is not just to try to avoid the potential harm of nutrient inadequacies but also to avoid potential harm of excess intakes. The DRI reports have brought this to the forefront by presenting estimates of Tolerable Upper Intake Levels (UL) (4). The UL was originally defined as “the highest level of daily nutrient intake that is likely to pose no risk of adverse health effects to almost all individuals in the general population” (4). That definition is repeated in all of the DRI reports and applied wherever there was sufficient evidence to justify setting a UL. In the most recent report addressing the macronutrients (5), a new reason for not setting a UL was invoked. Explicitly, evidence that would justify a UL for saturated and trans-fatty acids and for dietary cholesterol was presented and accepted by the scientific panel. However, no UL was presented “because trans-fatty acids are unavoidable in ordinary non-vegan diets, consuming 0 [the scientifically justified UL] would require significant changes in patterns of dietary intake. Such adjustments may introduce undesirable effects (eg, elimination of commercially prepared foods and dairy products and meats that contain trans-fatty acids may result in inadequate intakes of protein and certain micronutrients) and unknown and unquantified health risks” (5). Similar statements are presented for saturated fatty acids and cholesterol. This was an unfortunate decision on the part of the panel and the Institute of Medicine. A UL should have been presented based on scientific evidence. The risk-management considerations mentioned as justification for not setting the UL should have been weighed during the development of food-based dietary advice when the full range of alternative strategies would be considered. The approach to risk management involves modeling of expected intakes and looking at both potential inadequacies and excesses of nutrients with identified functional effects. In addition, one needs to think very carefully about what the actual risks (health consequences of inadequacy and excess for the nutrient in question) may be, and their importance to public health. The DRIs have mostly been designed to cover even minor health effects, creating a risk that such effects may or may not be deemed very important to the health of the group or population. Unless important risk is believed to exist, there may be little or no need for any attempt to raise or lower intake in comparison with intakes in the general diet or expected with compliance to existing guides. There is no question but that such modeling is made much more difficult with the introduction of voluntary fortification. The actions of the DRI panel that addressed macronutrients became even more questionable when the panel prepared the chapter on “Macronutrients and Healthful Diets,” the partial equivalent of the old US dietary goals. These were not based on the preceding scientific reviews. Rather, the goals seem to derive from a primary concern about diet and cardiovascular disS58 / December 2003 Suppl 2 Volume 103 Number 12
Table Comparison of macronutrient intakea goals for adults, United States and the World Health Organization (WHO) Nutrient
DRIb
WHO
Total fat Total carbohydrate Added sugar Total protein
20-35 45-65 ⬍25 10-35
15-30 55-75 ⬍10 10-15
a
Percent of energy intake. DRI⫽Dietary reference intakes; represents US recommendations.
b
ease and the assumption that the only effective way to minimize intakes of saturated and trans-fatty acids and dietary cholesterol would be to control total fat intake. It appears to have drawn heavily on clinical experience in developing dietary prescriptions for use in treating hyperlipidemias. Suggested carbohydrate and protein intakes more or less followed from that. This was a direct intrusion into risk management. If risk management was to be done in that setting, one might have expected greater discussion of other approaches to lowering of intakes of the two classes of fatty acids and dietary cholesterol. In the development of food guides, one must expect that there will be conflicts and overlaps between addressing needs for micronutrients and attempting to balance the macronutrients. The DRI report (5) has illustrated what scientific evidence does and does not establish and, by inference, where the opinions and judgments of informed people take over. What is not made clear is the point at which responsibility for judgments should pass from laboratory and clinical experts to those with expertise germane to foods and food consumption. In the coming months and years, it will be important to recognize these distinctions as a part of risk management in the design and use of food guides. It should be expected that epidemiology will play an ever-increasing role in the future development of food-based dietary guidelines. Earlier in this paper, I mentioned that two possible arguments for revising existing Canadian food guides were related to harmonization with either the US or European approaches. In this context, it is interesting to compare the numeric goals specified in the DRI report (5, Chapter 11) and in the very recent WHO report (6) (see Table), noting also that the WHO approach appears to have wide support in Europe. Superficially the two sets of goals seem similar and there should be no problem of compatibility. However, the DRI goals explicitly refer to limits for usual intake of the individual; the WHO goals explicitly refer to population or large group mean intakes. Therein lies the serious incompatibility. By the WHO criterion, Canada and the United States have achieved or nearly achieved goals for impact of fat intake on health. By comparison, applying the stated DRI goals, both countries face horrendous problems with 30% to 50% of the population consuming too much fat. Is it not reasonable to ask which of these judgments are more relevant to the planning of food-based dietary guidelines? In the absence of scientific evidence causally linking total fat to a detrimental health outcome (5), one must ask also whether we should focus on the identified risk factors (trans- and saturated fatty acids and dietary cholesterol)—rather than on total fat—in developing food-based guides? The importance of these two questions is why I termed the rejection of a UL to be an unfortunate decision by the DRI
panel and the Institute of Medicine; it seems to misdirect attention. Although not a customary inclusion in the definition of “foodbased” approaches to dietary guidelines, it is important to recognize that evidence relating to nutrient effects sometimes fails to explain well-documented and consistent evidence relating food use and chronic disease. The most obvious example is the relationship of consumption of fruits and vegetables to certain types of cancer. I think that in the future, we will see/must see a much greater attention directed to food-based evidence when developing food guides. For example, in addition to the fruit and vegetable evidence, one would have to weigh the limited epidemiologic evidence suggesting that high intakes of milk might have detrimental effects on osteoporosis at the same time it is argued that high intakes of calcium may be beneficial. We have much to learn about food intake patterns and chronic disease. Only a few centers seem to be addressing this; interest is higher in Europe than in North America, where reliance continues to be placed on laboratory studies and clinical trials rather than epidemiologic investigations and population health. Advocating direct pharmaceutical supplements might be a better choice than always demanding that nutrients come from foods although this is heresy in current thinking. These types of considerations go well beyond asking whether intakes are marginally inadequate in comparison to DRI estimates of need. We must now weigh predicted health outcomes and make our tradeoffs, not just retreat to old truisms. That is the difficult but extremely important era that nutrition has entered. Our paradigm has changed, and we must be prepared to change with it. That part of the challenge of developing food guides was not discussed in the symposium. CONCLUSION The original purpose of this paper was to stimulate discussion in the workshop following the symposium. The paper has been heavily revised for publication, particularly by adding specifics in the discussion of risk management, but hopefully continues to serve its intended purpose, now provoking a much broader audience into thinking and discussion. The paper is intentionally provocative but the message is clear: Nutrition is at a conceptual and philosophical turning point. The DRI reports are
not the end. They are but a piece of the beginning. We must now change how we think and teach about nutrition—and how we approach the design food guides. The future in our field will increasingly demand evidence that actions and expenditures yield substantial (in public health terms) benefits with minimal real risks and at a cost that can be justified by expected benefits. When we adopt “evidence-based’ approaches, we must all assess the continuity and consistency of the actual evidence. I fear that that has not been a universal practice in the past. I sincerely hope that the future offers better. Revision of food guides, if undertaken, may be a place to start. For planning, including the development of dietary guidelines, one needs to know not only existing patterns of food use and estimated requirements— one must also know the health conditions of the population and be able to make some inference about the likely improvement— or otherwise— of health that could be expected with a specific change in food use. All of the health outcomes in which we are interested are multifactorial. It was not the mandate of the DRI system to look at what is very loosely called attributable risk, but those planning for the improvement or preservation of population health must certainly address this. What real benefit is to be expected? Is it worth the cost? References 1. FAO/WHO Preparation and Use of Food-based Dietary Guidelines. WHO Technical Rpt Ser. No. 880, Geneva. 1998. 2. Institute of Medicine. Dietary Reference Intakes: Applications in Dietary Assessment. Washington, DC: National Academy Press; 2000. 3. McClinton PE, Milne H, Beaton GH. An evaluation of food habits and nutrient intakes in Canada: Design of effective food guides. Can J Public Health. 1971;62:139-146. 4. Institute of Medicine. Dietary Reference Intakes: A Risk Assessment Model for Establishing Upper Intake Levels for Nutrients. Washington, DC: National Academy Press; 1998. 5. Institute of Medicine. Dietary Reference Intakes: Energy, Carbohydrate, Fiber, Fat, Fatty Acids, Cholesterol, Protein, and Amino Acids [Prepublication copy, unedited proofs]. Vols 1 and 2. Washington, DC: National Academy Press; 2002. 6. WHO/FAO Diet, Nutrition and the Prevention of Chronic Diseases Report of a Joint WHO/FAO Expert Consultation. Geneva: The World Health Organization; 2003. WHO Tech Report Ser No. 919.
Supplement to the Journal of THE AMERICAN DIETETIC ASSOCIATION / S59