Overall Assessment and Major Conclusions
This chapter describes the committee's process for integrating the evidence relating dietary components to chronic diseases and presents its major conclusions concerning the role of diet in health. It is prefaced by a brief description of the special features of this study.
Special Characteristics of the Study
Over the past half century, extensive epidemiologic, clinical, and experimental research has shown that diet is one of many factors that play an important role in the etiology and pathogenesis of major chronic diseases (AHA, 1988; Ahrens et al., 1979; DHEW, 1979; NRC, 1982; Page et al., 1961; U.S. Senate, 1977). In recent decades, scientists have identified many dietary factors that influence the incidence and course of specific chronic diseases and have attempted to define the pathophysiological mechanisms (AHA, 1982; Ahrens et al., 1979; Goldstein and Brown, 1984; Levy et al., 1979; NRC, 1980, 1982). Simultaneously, scientists, public health policymakers, the food industry, consumer groups, and others have been engaged in a debate about how much and what kind of evidence justifies giving dietary advice to the public and how best to control risk factors on which there is general agreement among scientists (Ahrens, 1985; Council on Scientific Affairs, 1979; Blackburn, 1979; CAST, 1977; Connor, 1979; Habicht et al., 1979; Grobstein, 1983; Gussow and Thomas, 1986; Harper, 1978; Hegsted, 1978; NRC, 1980, 1982; O'Connor and Campbell, 1986; Olson, 1979; Palmer, 1983; U.S. Senate, 1977).
The present study was launched in an attempt to address the critical scientific issues that have been under debate, many of which are fundamental to nutrition policy on reducing the risk of chronic diseases. The committee recognized at the outset that the absence of consensus on certain diet-disease interrelationships derives partly from a lack of knowledge and partly from the absence of generally accepted criteria for interpretation and acceptability of the abundant though incomplete evidence on diet and chronic diseases. Several reports have addressed the importance of dietary factors in public health. With the exception of the recent Surgeon General's Report on Nutrition and Health (DHHS, 1988), however, they have focused primarily on identifying dietary risk factors for single diseases (e.g., American Diabetes Association, 1987; AHA, 1986, 1988; NIH, 1984a,b, 1985; NRC, 1982). With the exception of Ahrens et al. (1979), very few reports dealing with general health maintenance have documented in detail the criteria for acceptability of the evidence or provided a detailed basis for their conclusions. This report attempts to cross the boundary be-
tween identifying dietary risk factors for single diseases and determining how these risk factors influence the spectrum of chronic diseases and conditions, including atherosclerotic cardiovascular diseases, hypertension, obesity, cancer, osteoporosis, diabetes mellitus, hepatobiliary disease, and dental caries. The report complements and extends past efforts of government agencies and voluntary health and other scientific organizations by presenting an in-depth analysis of the overall relationship between diet and the major chronic diseases (e.g., AHA, 1988; DHHS, 1988; NRC, 1982; USDA, 1985; USDA/DHHS, 1980).
In the foregoing chapters, the committee reviews the evidence on all major public health conditions in which diet is believed to play an important role. In this chapter, it presents its overall conclusions about the effect of nutrients, foods and food groups, and dietary patterns on chronic diseases.
In Chapter 2, the committee presents criteria for assessing the data from single studies and explains the process for evaluating the cumulative evidence. The committee first considers the special strengths and weaknesses of each kind of epidemiologic and laboratory study on diet and chronic diseases and then evaluates the total evidence against the commonly used criteria for assessing causality, i.e., strength of association, dose-response relationship, temporally correct association, consistency of association, specificity of association, and biologic plausibility. It emphasizes, however, that these criteria alone do not define acceptability of the evidence.
Special attention is given to dietary interactions and competing risks, which are important considerations both for arriving at conclusions and for formulating dietary recommendations. For example, although diets containing high levels of plant foods have been associated with a lower risk of certain cancers, such diets, because of their high fiber content, could in principle initially inhibit the absorption of essential minerals such as calcium, thereby possibly enhancing other risks. Such potential competing risks and dietary interactions were considered in drawing the conclusions presented in this chapter.
The committee recognizes that genetically dependent variability among individuals and variability due to age, sex, and physiological status may affect physiological requirements for nutrients as well as responses to dietary exposures and, thus, the risk of chronic diseases. Therefore, to the extent possible, it addresses not only the risk to the general population, but also the feasibility of defining risks to subpopulations and individuals that may differ in susceptibility. Recognizing the limitations of the data on diet-disease relationships, the committee wishes to emphasize the necessarily interim nature of its conclusions.
Criteria and Process for Interpreting and Integrating Evidence
Chapter 2 explains the many limitations to drawing conclusions about the association between dietary factors and chronic diseases. The term insufficient data could perhaps be applied to most issues concerning nutrition and health. In particular, it characterizes many of the relationships between diet and certain chronic diseases. The lack of certainty about causal associations and mechanisms of action is common and stems in part from attempts to relate a complex mixture such as diet to complex, multifactorial chronic diseases for which the pathophysiological, environmental, and genetic predisposing factors are imprecisely understood. Although this is a cause for concern, and therefore warrants further research (see Directions for Research in Chapter 28), it is not unusual in questions pertaining to human health.
In some cases, there is conclusive evidence that a particular dietary factor plays a role in the etiology of a particular chronic disease, but that is the exception rather than the rule. Despite such limitations, a large body of evidence has emerged in the past four decades concerning chronic diseases and their relationship to general dietary patterns or specific dietary components.
Studies in Humans
The strengths and limitations of different types of studies in humans and the methods of assessing dietary intake in such studies are described in Chapter 2. In general, the accuracy of assessing dietary intake is limited by the need to rely on the subjects' memories, the potential for misclassification, the bias of the subject or the investigator, the difficulty of precisely quantifying dietary exposure in years past (which would reflect the long latency period of most chronic diseases), the difficulty of standardizing the methodology of data collection, variation in accuracy of recall between subjects and controls, the likelihood of dietary modification by subjects over time, and the limitations of food composition data.
The committee recognizes that ecological correlations of dietary factors and chronic diseases among populations cannot be used alone to estimate the strength of the association between diet and disease. In general, due to the limitations summarized above, correlations among individuals in a population, including analyses of case-control and cohort studies, are likely to underestimate the strength of the association. Many prospective studies, such as the Framingham (Dawber et al., 1982) and Tecumseh (Nichols et al., 1976) studies, in which dietary practices of individuals were related to disease precursors or outcome (e.g., serum lipid levels or heart attacks), have failed to demonstrate the hypothesized relationship between the diet of an individual and the risk of disease. The absence of established relationships in such studies is probably due to a limited capacity to characterize the diet of an individual, the difficulty of taking into account the large day-to-day variability in dietary intake, and the variability in response (e.g., in serum total cholesterol levels) among people with similar dietary intakes. In contrast, the effect of diet has been more consistently demonstrated in comparisons of population groups with substantially different dietary practices (e.g., vegetarians and nonvegetarians, or Mediterraneans and northern Europeans). In the committee's judgment, when findings from studies within populations differ from those between populations, the latter assume greater importance because of the odds against identifying correlations between dietary factors and chronic disease within a population whose diet is fairly homogeneous.
Results of intervention studies that randomly allocate people to different diets are often considered ideal for assessing causal associations. These received special attention. The committee recognizes that rigid criteria for selecting participants in such studies lead to greater homogeneity in the study samples. In general, long exposures are required for the effects of dietary factors on disease risk to become manifest, and it is difficult to control the diets of noninstitutionalized populations for extended periods. The results of small-scale, short-term clinical investigations conducted under controlled conditions may have limited applicability to the general population or to chronic diseases with long latency periods.
Despite these limitations, in the committee's judgment, repeated and consistent findings of associations between certain dietary factors and certain diseases evaluated against the criteria described in Chapter 2 indicate that such associations are likely to be real and indicative of cause-and-effect relationships.
Studies in Animals
Animal experiments are an important counterpart to epidemiologic and clinical research on nutrition and chronic diseases. As described in Chapter 2, such studies can control genetic variability as well as dietary exposure and permit more intensive observation. However, in assessing the results of animal experiments, one must consider variability among species in diet and nutrient requirements, in absorption and metabolic phenomena, and thus in the comparability of their exposure and disease outcomes to humans.
The committee placed more confidence in data derived from studies on more than one animal species or test system, on results that have been reproduced in different laboratories, and on data that indicate a dose-response relationship.
Integrating the Overall Evidence
The committee recognizes that an a priori weighting scheme could be helpful in evaluating the many types of health-related data but concluded that its application to studies on diet and chronic diseases is not feasible in view of the limitations mentioned above. Thus, in addition to the criteria summarized above and discussed in more detail in Chapter 2, the committee based its conclusions on the totality of the evidence. It took the general view that the strength of the evidence should be evaluated on a continuum from highly likely to very inconclusive. The strength, consistency, and preponderance of the data and the degree of concordance in epidemiologic, clinical, and laboratory experiments determined the strength of the conclusions in the report.
In Section II of this report (Evidence on Dietary Components and Chronic Diseases), the criteria described in Chapter 2 provide the basis of a review of the evidence by nutrients. The 13 chapters in that section (6 through 18) summarize the relevant epidemiologic, clinical, and laboratory data pertaining to each nutrient or dietary factor and specific chronic diseases, including cardiovascular diseases, specific cancers, diabetes, hypertension, obesity, osteoporosis, hepatobiliary disease, and dental caries. Nutrient interactions and mechanisms of action are discussed where applicable.
Section III (Impact of Dietary Patterns on Chronic Diseases) briefly reassembles the evidence
relating nutrients to specific chronic diseases or conditions and presents conclusions about the role of dietary components and patterns in the etiology of those diseases. Where data permit, the potential for reducing the risk of each disease by changes in dietary patterns is also discussed. This section goes beyond examining data on individual nutrients to consider the evidence on foods, food groups, and dietary patterns. Some of the evidence was obtained directly and some was obtained by extrapolation from data on nutrients or other dietary components. For example, much of the epidemiologic evidence on diet and colon cancer pertains to consumption of vegetables and diets with a high plant-fiber content or low levels of fat. Thus, it is possible to draw direct conclusions about dietary patterns rather than just about individual nutrients or nonnutritive components in these foods. In contrast, many metabolic studies on diet and osteoporosis have involved the measurement of calcium intake rather than the consumption of dairy products or other calcium-containing foods. Thus, extrapolation of the data from such studies is necessary to arrive at conclusions about dairy foods.
Major Conclusions and their Bases
Following are the general conclusions drawn from the committee's in-depth review followed by specific conclusions pertaining to the major dietary components and specific chronic diseases. Each section begins with a brief discussion of the findings that served as a basis for the conclusions.
In the United States during this century, there have been noticeable changes in per capita availability of foods, in eating patterns, and in chronic disease trends (see Chapters 3 and 5). Although average per capita availability of calories does not appear to have varied substantially since 1909, the percentage of total calories from fat in this period increased by 11%, while calories from carbohydrates decreased. Since the 1960s, the per capita supply of fat has steadily increased, but remarkable changes have occurred in the sources and therefore the types of fat available. Use of whole milk has declined, whereas consumption of low-fat milk and whole-milk cheeses has increased. Beginning in the 1940s, the use of butter and lard greatly decreased, while the use of margarine and salad and cooking oils dramatically increased. Most of these changes have resulted in increased per capita availability of polyunsaturated and monounsaturated fatty acids and a decreased supply of saturated fatty acids. Currently, the percentages of calories from fat by specific groups of fatty acids in the U.S. food supply are: approximately 17% monounsaturated (mainly oleic acid), 15% saturated, and 7% polyunsaturated (mainly linoleic acid) (see Chapter 3).
The use of eggs has also declined since 1947, resulting in lower average supplies of cholesterol. The use of poultry has increased steadily and markedly, while that of beef has declined slightly since the mid-1970s.
Complex carbohydrates in the food supply declined from 1909 to 1967, because of the decreased use of grain products and potatoes, but have increased by 9% since 1967. The proportion of carbohydrates from sugars in the food supply increased from approximately 33% in 1909 to slightly more than 50% in 1980. Use of sucrose declined remarkably since its peak use in 1971. Since then it has been partially replaced with corn syrups, including high-fructose corn syrup.
The average availability of protein in the U.S. population has remained at 11% of calories since 1909, but the proportion of protein from animal sources increased 16% by 1982 as a result of decreased use of flour, cereal products, and potatoes and increased use of meat, poultry, fish, and dairy products.
Per capita consumption of vegetables increased by 5% between 1970 and 1983. This includes a notable increase in the per capita intake of broccoli and cauliflower. An increased per capita availability of vitamin A since 1967 resulted from higher amounts of carotenoids in newer varieties of carrots and sweet potatoes. Growing supplies of citrus fruits since 1967 and increasing fortification of foods with vitamin C led to an increased availability of vitamin C in the food supply, and the large increase in the use of salad and cooking oils greatly increased the availability of vitamin E.
Estimates of nutrient availability in the food supply are higher than amounts actually consumed by the population, since food supply data are measured at the wholesale/retail level and do not take into account food wastage or nutrient losses that occur during food processing, marketing, and food preparation. Also, changes in the food supply over time are difficult to compare with data from national surveys that measure actual consumption of foods because of differences in survey methods.
Nevertheless, as described in Chapter 3, both food supply data and the national food consumption surveys have provided useful insights into food and nutrient availability and intake, as well as eating patterns.
The cross-sectional data on food consumption patterns in the national surveys do not permit simple correlations to be made between trends in eating patterns and trends in chronic disease incidence and mortality. The latter are reviewed in Chapter 5. Age- and sex-adjusted mortality from coronary heart disease in the United States has declined more than one-third in the last two decades. Even larger declines have been observed in mortality from stroke and other hypertension-related causes of death. On the other hand, recent surveys in the United States report increases in the prevalence of overweight both in men and women, especially in the younger age groups.
Cancers are responsible for approximately 22% of all deaths in the United States, and total cancer mortality has remained essentially unchanged in recent years in this country. However, the incidence and the mortality rates for specific cancers have shown noticeable changes with time. For example, the rates for lung cancer, one of the major causes of cancer mortality in the United States, have begun to decline for males following decades of increase; however, the incidence and mortality rates for lung cancer among females are increasing. The incidence of breast cancer in women has increased in the last 20 to 30 years, whereas mortality appears to have slightly decreased in premenopausal women and slightly increased in the postmenopausal group. In contrast, stomach cancer incidence and mortality have been declining sharply in both sexes for nearly half a century. Similarly, the incidence of (but not mortality from) endometrial cancer has risen in the past decade in the western part of the United States.
Approximately 15 to 20 million Americans are afflicted with osteoporosis in the United States, which has one of the highest rates of hip and other fractures in the world. These numbers are expected to increase steadily. The incidence of noninsulin-dependent diabetes mellitus (NIDDM), which is the seventh leading cause of mortality in the United States, is estimated to have increased sixfold in the past 50 years. By contrast, in the last 10 to 15 years, there has been a large and unprecedented decline in the prevalence of dental caries among U.S. children.
The committee analyzed these trends in the major chronic diseases as well as trends in eating patterns (Chapters 3 and 5). It reviewed the epidemiologic, clinical, and laboratory evidence pertaining to dietary factors and chronic diseases (Chapters 6 through 26) and attempted to put into perspective the role of diet versus the role of other environmental and genetic factors in the etiology of these diseases (Chapters 4 and 5).
Following are the general conclusions drawn from the committee's in-depth review as well as the specific conclusions pertaining to the major dietary components and specific chronic diseases.
· A comprehensive review of the epidemiologic, clinical, and laboratory evidence indicates that diet influences the risk of several major chronic diseases. The evidence is very strong for atherosclerotic cardiovascular diseases and hypertension and is highly suggestive for certain forms of cancer (especially cancers of the esophagus, stomach, large bowel, breast, lung, and prostate). Furthermore, certain dietary patterns predispose to dental caries and chronic liver disease, and a positive energy balance produces obesity and increases the risk of NIDDM. However, the evidence is not sufficient for drawing conclusions about the influence of dietary patterns on osteoporosis and chronic renal disease.
· Most chronic diseases in which nutritional factors play a role also have genetic and other environmental determinants, but not all the environmental risk factors have been clearly characterized and susceptible genotypes usually have not been identified. Furthermore, the mechanisms of genetic and environmental interactions involved in disease are not fully understood. It is evident that dietary patterns are important factors in the etiology of several major chronic diseases and that dietary modifications can reduce such risks. Nevertheless, for most diseases, it is not yet possible to provide quantitative estimates of the overall risks and benefits.
Fats, Other Lipids, and High-Fat Diets
There is a substantial body of evidence pertaining to fats and other lipids and their impact on health. Most of this evidence concerns atherosclerotic cardiovascular diseases, several forms of cancer, and to a lesser extent, obesity. For atherosclerosis, the evidence is derived from decades of
study, including extensive epidemiologic investigations in many parts of the world, laboratory experiments in different animal species, clinical studies, and intervention trials. In these studies, investigators have examined the effect of the type and amount of various dietary fatty acids, cholesterol, and other dietary components on blood lipid and lipoprotein profiles, on the development of atherosclerotic lesions, and on the occurrence of coronary events as well as the effects of dietary intervention alone or in combination with cholesterol-lowering drugs. Although most of these investigations have drawbacks, as discussed in Chapters 2, 7, and 19, overall they provide an extensive and reliable body of evidence from which to draw conclusions.
The information pertaining to fats and cancer risk (Chapters 7 and 22) pertains primarily to cancers of the colon, prostate, and breast is derived from ecological, case-control, and prospective studies in humans in many parts of the world and from extensive laboratory experiments on spontaneous and chemically induced cancers in rodents. The data base on fat and cancer is limited compared to that on coronary heart disease, especially with respect to the relative effects of different types and amounts of fats and because of the absence of intervention trials. Nonetheless, it is sufficient to serve as the basis for certain conclusions.
There have been relatively few studies in humans on the effect of fat intake per se on obesity (Chapters 6, 7, and 21), and these are compromised by the difficulty of separating the effect of fat from the effects of total calories and other macronutrients. In contrast, there is a substantial body of laboratory evidence on this subject.
The following conclusions derive from the committee's extensive review of the data described in Chapters 6 (Calories), 7 (Fats and Other Lipids), 19 (Atherosclerotic Cardiovascular Diseases), 21 (Obesity and Eating Disorders), 22 (Cancer), and 25 (Hepatobiliary Disease).
· There is clear evidence that the total amounts and types of fats and other lipids in the diet influence the risk of atherosclerotic cardiovascular diseases and to a less well-established extent, certain forms of cancer, and possibly obesity. The evidence that the intake of saturated fatty acids and cholesterol are causally related to atherosclerotic cardiovascular diseases is especially strong and convincing.
· In several types of epidemiologic studies, a high-fat intake is associated with increased risk of certain cancers, especially cancers of the colon, prostate, and breast. The epidemiologic evidence is not totally consistent, but it is supported by experiments in animals. The combined epidemiologic and laboratory evidence suggests that a reduction of total fat intake is likely to decrease the risk of these cancers.
· High-fat intake is associated with the development of obesity in animals and possibly in humans. In short-term clinical studies, a marked reduction in the percentage of calories derived from dietary fat has been associated with weight loss.
· Although gallbladder disease is associated with obesity, there is no conclusive evidence that it is associated with fat intake.
· Intake of total fat per se, independent of the relative content of the different types of fatty acids, is not associated with higher blood cholesterol levels and coronary heart disease (CHD). A reduction in total fat consumption, however, facilitates reduction of saturated fatty acid intake; hence in addition to reducing the risk of certain cancers, and possibly obesity, it is a rational part of a program aimed at reducing the risk of CHD.
· Clinical, animal, and epidemiologic studies demonstrate that increased intakes of saturated fatty acids (12 to 16 atoms in length) increase the levels of serum total and low-density-lipoprotein (LDL) cholesterol and that these higher levels in turn lead to atherosclerosis and increase the risk of CHD. Saturated fatty acid intake is the major dietary determinant of the serum total cholesterol and LDL cholesterol levels in populations and thereby of CHD risk in populations. Lowering saturated fatty acid intake is likely to reduce serum total and LDL cholesterol levels and, consequently, CHD risk.
· The few epidemiologic studies on dietary fat and cancer that have distinguished between the effects of specific types of fat indicate that higher intakes of saturated fat as well as total fats are associated with a higher incidence of and mortality from cancers of the colon, prostate, and breast. In general, these findings are supported by data from animal experiments.
· Clinical and animal studies provide firm evidence that omega-6 PUFAs when substituted for SFAs result in a lowering of serum total cholesterol and LDL cholesterol and usually also some lowering of high-density-lipoprotein (HDL) cholesterol levels.
· Laboratory studies in rodents suggest that diets with high levels of vegetable oils containing omega-6 PUFAs promote certain cancers more effectively than diets with high levels of saturated fats, whereas there is some evidence that diets with a high content of omega-3 PUFAs may inhibit these same cancers. However, these findings are not supported by the limited number of epidemiologic studies that have distinguished between the effects of different types of fat. There are no human diets that naturally have very high levels of total PUFAs, and there is no information about the long-term consequences of high PUFA intakes.
· Fish oils containing large amounts of omega-3 PUFAs reduce plasma triglyceride levels and increase blood clotting time. Their effects on LDL cholesterol vary, and data on the long-term health effects of large doses of omega-3 PUFAs are limited. Limited epidemiologic data suggest that consumption of one or two servings of fish per week is associated with a lower CHD risk, but the evidence is not sufficient to ascertain whether the association is causal or related to the omega-3 PUFA content of fish.
· Clinical studies indicate that substitution of MUFAs for SFAs results in a reduction of serum total cholesterol and LDL cholesterol without a reduction in HDL cholesterol.
· Clinical, animal, and epidemiologic studies indicate that dietary cholesterol raises serum total cholesterol and LDL cholesterol and increases the risk of atherosclerosis and CHD. There is substantial inter- and intraindividual variability in this response. High dietary cholesterol clearly seems to contribute to the development of atherosclerosis and increased CHD risk in the population.
· Clinical studies indicate that TFAs and their cis isomers have similar effects on plasma lipids. Animal studies do not indicate that TFAs have a greater tumor-promoting effect than their cis isomers.
Carbohydrates, Vegetables, Fruits, Grains, Legumes, and Cereals
and Their Constituents
The committee's conclusions on carbohydrates and foods containing complex carbohydratesi.e., vegetables, fruits, grains, legumes, and cereal productsderive from a review of direct and indirect evidence. Many epidemiologic studies in different parts of the world have focused on diets high in plant foods in general or in green, yellow, and cabbage-family vegetables and citrus fruits in particular. They have concentrated on the incidence of or mortality from different forms of cancer (especially cancers of the lung, large bowel, stomach, and esophagus) and on coronary heart disease. In some of these studies, the diets of vegetarians and nonvegetarians have been compared. Many clinical metabolic studies have focused on the effects of refined sugars and specific starches on blood glucose and insulin sensitivity. In a series of experiments in animals, investigators have examined the effects of certain components of these foods (e.g., different fibers, vitamins, minerals, and nonnutritive components) on plasma cholesterol levels or on different types of cancer. The indirect evidence on plant foods and chronic disease risk is derived from epidemiologic studies that have focused on diets that are high in fat and animal proteins and that usually tend to be low in carbohydrates and plant foods. Both the direct and indirect evidence has been useful in drawing conclusions, although the committee notes the difficulty in comparing the results of epidemiologic studies, which generally pertain to foods, to metabolic studies or animal experiments, which often deal with single nutrients.
The following conclusions are based on a review of the evidence throughout the report, especially Chapters 7 (Fats and Other Lipids), 9 (Carbohydrates), 10 (Dietary Fiber), 11 (Fat-Soluble Vitamins), 12 (Water-Soluble Vitamins), and 22 (Cancer).
· Diets high in plant foodsi.e., fruits, vegetables, legumes, and whole-grain cerealsare associated with a lower occurrence of coronary heart disease and cancers of the lung, colon, esophagus, and stomach. Although the mechanisms underlying these effects are not fully understood, the inverse association with CHD may be largely explained by the usually low SFA and cholesterol content of such diets. Such diets are also low in
total fat, which is directly associated with the risk of certain cancers, but rich in complex carbohydrates (starches and fiber) and certain vitamins, minerals, trace elements, and nonnutritive constituents, and these factors probably also confer protection against certain cancers and CHD.
· Compared to nonvegetarians, complete vegetarians and lacto-ovovegetarians have lower serum levels of total and LDL cholesterol and triglycerides. These lower levels may be the combined result of lower intakes of saturated fatty acids and total fat and higher intakes of water-soluble fiber (e.g., pectin and oat bran). In clinical and animal studies, such fiber has been found to produce small reductions in serum total cholesterol independently of the effect due to fat reduction.
· Populations consuming high-carbohydrate diets, which are high in plant foods, have a comparatively lower prevalence of NIDDM, possibly because of the higher proportion of complex carbohydrate intake and lower prevalence of obesitya risk factor for NIDDM. In clinical studies, such diets have been shown to improve glucose tolerance and insulin sensitivity.
· Epidemiologic studies indicate that consumption of carotenoid-rich foods, and possibly serum carotene concentration, are inversely associated with the risk of lung cancer.
· Laboratory studies in animals strongly and consistently indicate that certain retinoids prevent, suppress, or retard the growth of chemically induced cancers at a number of sites, including the esophagus, pancreas, and colon, but especially the skin, breast, and bladder. However, most epidemiologic studies do not show an association between preformed vitamin A and cancer risk or a relationship between plasma retinol level and cancer risk.
· Epidemiologic studies suggest that vitamin C-containing foods such as citrus fruits and vegetables may offer protection against stomach cancer, and animal experiments indicate that vitamin C itself can protect against nitrosamine-induced stomach cancer. The evidence linking vitamin C or foods containing that vitamin to other cancer sites is more limited and less consistent.
· Some investigators have postulated that several other vitamins (notably vitamin E, folic acid, riboflavin, and vitamin B12) may block the initiation or promotion of cancer, but the committee judged the evidence too limited to draw any conclusions.
· Epidemiologic and clinical studies indicate that a diet characterized by high-fiber foods may be associated with a lower risk of CHD, colon cancer, diabetes mellitus, diverticulosis, hypertension, or gallstone formation, but there is no conclusive evidence that it is dietary fiber, rather than the other components of vegetables, fruits, and cereal products, that reduces the risk of those diseases. Although soluble fibers can decrease serum cholesterol and glucose levels, and certain insoluble fibers inhibit chemically induced tumorigenesis, it is difficult to compare the effects of specific dietary fibers tested in the laboratory with the effects of fiber-containing foods or of other potentially protective substances present in these foods.
· Although human and animal studies indicate that all fermentable carbohydrates can cause dental caries, sucrose appears to be the most cariogenic. The cariogenicity of foods containing fermentable carbohydrates is influenced by the consistency and texture (e.g., stickiness) of the food as well as by the frequency and sequence of consumption. Sugar consumption (by those with an adequate diet) has not been established as a risk factor for any chronic disease other than dental caries in humans.
Protein and High-Protein Diets
Compared to the association of fats and complex carbohydrates with chronic disease risk, the association between high-protein diets or protein per se with such risk has received little attention. Much of the evidence on protein and chronic diseases derives indirectly from epidemiologic studies that examined the effects of high fat diets on the risk of atherosclerotic cardiovascular diseases or cancer. In contrast, many animal experiments have measured the effect of high animal-protein intake on serum total cholesterol or on tumor yield. These studies are reviewed in Chapters 8 (Protein), 19 (Atherosclerotic Cardiovascular Diseases), and 22 (Cancer). The committee also examined the limited data on the effect of protein intake on urinary calcium excretion and its possible consequences for osteoporosis, as well as the basis for the more recent interest in the potential adverse effects of a high-protein intake on chronic renal disease. These topics are examined in Chapters 8 (Protein), 13 (Minerals), and 23 (Osteoporosis). Following are the committee's major conclusions.
· In intercountry correlation studies, diets high in meata major source of animal proteinhave a strong positive association with increased athero-
sclerotic coronary artery disease and certain cancers, notably breast and colon cancer. Such diets are often characterized by a high content of SFAs and cholesterol, which probably accounts for a large part of the association with CHD, and by a high content of total fat, which is directly associated with the risk of these cancers. However, these diets also tend to have low levels of plant foods, the consumption of which is inversely associated in epidemiologic and animal studies with the risk of heart disease and certain cancers. Total serum cholesterol can be reduced in people with high blood cholesterol by replacing animal foods in their diet with plant foods.
· High protein intake can lead to increased urinary calcium excretion. The impact of this finding on the development of osteoporosis in the general population is unclear.
· The data linking elevated intakes of animal protein to increased risk of hypertension and stroke are weak, and no plausible mechanisms have been posited for either effect.
The associations among intake of energy-yielding foods, energy expenditure, and obesity and their relation to chronic diseases have been studied for decades, as evidenced by a voluminous literature on these topics. More recently, investigators have been examining the relationship between genetic factors and energy balance. These topics are reviewed in Chapters 6 (Calories) and 21 (Obesity). There are extensive data from epidemiologic, clinical, and animal studies concerning the effect of modifying caloric and nutrient intake on body weight and adiposity. The clinical data pertain to over- and underfeeding of nonobese and obese individuals and the effect of starvation on body weight and body composition. Numerous animal experiments have been conducted in nonobese and spontaneously obese rodents. Studies in both humans and animals demonstrate the difficulty of accurately measuring total food and caloric intake, assessing energy balance, and separating the effects of caloric intake per se from the effect of specific macronutrients on body weight. The committee has analyzed the association between energy balance (energy intake and energy expenditure) and body weight and obesity and has examined obesity as an independent risk factor for atherosclerotic cardiovascular diseases, hypertension, NIDDM, and certain cancers.
The following conclusions are based on this assessment.
· Positive energy balance can result from increased energy intake, reduced energy expenditure, or both, and over the long term can lead to obesity and its associated complications.
· Although data from clinical and animal studies demonstrate that overfeeding leads to obesity, increased body weight in cross-sectional and longitudinal population surveys of adults cannot be accounted for by increased energy intake. Thus, it is likely that obesity develops in adult life either because of reduced physical activity, or overfeeding, or both. Obesity is enhanced not only by this energy imbalance but also by a genetic predisposition to obesity and altered metabolic efficiency.
· Epidemiologic studies indicate that increased energy expenditure is inversely associated with the risk of CHD.
· Epidemiologic and clinical studies and some experiments in animals demonstrate that obesity is associated with an increased risk of NIDDM, hypertension, gallbladder disease, endometrial cancer, and osteoarthritis. It may also be associated with a higher risk of CHD and postmenopausal breast cancer.
· Studies in humans suggest that fat deposits in the abdominal region pose a higher risk of NIDDM, CHD, stroke, hypertension, and increased mortality than do fat deposits in the gluteal or femoral regions.
· Experience in long-term management of obesity indicates that neither frequent fluctuations in body weight nor extreme restrictions of food intake are desirable.
· Long-term follow-up studies indicate that extreme leanness is associated with increased mortality and that the causes of that mortality are different from those associated with excess weight.
· The specific causes of obesity are not well known, although some obese people clearly consume more energy compared to people of normal weight, whereas others are very sedentary or may have increased metabolic efficiency. Compared to maintenance of stable weight, weight gain in adult life is associated with a greater risk of cardiovascular disease, NIDDM, hypertension, gallbladder disease, and endometrial cancer. Certain risk factorse.g., high serum cholesterol, elevated serum glucose, and high blood pressurecan be curtailed by weight reduction in overweight adults.
The extensive data on the health effects of alcohol consumption are examined in Chapters 16 (Alcohol), 19 (Atherosclerotic Cardiovascular Diseases), 20 (Hypertension), 22 (Cancer), and 25 (Hepatobiliary Disease). In these chapters, special note is made of the high incidence of alcoholism in the United States and the difficulty of obtaining accurate measures of alcohol intake. Attention is given to separating the effects of mild and moderate levels of alcohol consumption from those of excessive intake or alcohol abuse. The effects of alcohol consumption have been studied in relation to malnutrition, obesity, carcinogenesis, hypertension, cardiovascular diseases, total mortality, cirrhosis of the liver, several diseases of the nervous system, and adverse pregnancy outcome.
Following are the committee's major conclusions related to alcohol.
· When consumed in excess amounts, alcohol replaces essential nutrients including protein and micronutrients and can lead to multiple nutrient deficiencies.
· Sustained heavy intake of alcoholic beverages leads to fatty liver, alcoholic hepatitis, and cirrhosis. It also increases the risk of cancers of the oral cavity, pharynx, esophagus, and larynx, especially in combination with cigarette smoking, whereupon the effects on cancer risk become synergistic. There is some epidemiologic evidence that alcohol consumption is also associated with primary liver cancer and that moderate beer drinking is associated with rectal cancer. The association of alcohol consumption with increased risk of pancreatic or breast cancer is less clear.
· Excessive alcohol consumption is associated with an increased incidence of CHD, hypertension, stroke, and osteoporosis.
· Alcohol consumption during pregnancy can damage the fetus, cause low infant birth weight, and lead to fetal alcohol syndrome. No safe level of alcohol intake during pregnancy has been determined.
Salt and Related Compounds
Most studies pertaining to salt and human health have focused on hypertension or gastric cancer. The epidemiologic evidence pertaining to hypertension is extensive and comes from longitudinal and cross-sectional studies in large and varied populations throughout the world and from clinical metabolic studies among smaller groups. These data are complemented by a considerable body of evidence derived from studies of animals. In contrast, the data on salt and gastric cancer derive primarily from correlation studies in different parts of the world and a limited number of case-control studies of cancer risk in migrants. These have generally been assessments of the role of salt-pickled and salt-cured foods rather than of sodium chloride intake per se on cancer risk. The effects of several other dietary elements (e.g., potassium, calcium, magnesium, and PUFA) on hypertension have also been examined, but these data are limited compared to those on sodium chloride.
The following conclusions are based on the evidence concerning salt and related compounds and their relation to chronic diseases. This evidence is reviewed in Chapters 15 (Electrolytes), 20 (Hypertension), and 22 (Cancer).
· Blood pressure levels are strongly and positively correlated with the habitual intake of salt. In populations with a sustained salt intake of 6 g or more per day, blood pressure rises with age and hypertension is frequent, whereas in populations consuming less than 4.5 g of salt per day, the age-related rise in blood pressure is slight or absent and the frequency of hypertension is uniformly low. Clinical studies demonstrate that once hypertension is established, it cannot always be fully corrected by resumption of a moderately low (<4.5 g/day) salt intake.
· Although clinical and epidemiologic studies indicate that some people are more susceptible to salt-induced hypertension than others, there are no reliable markers to predict individual responses. Epidemiologic evidence suggests that blacks, people with a family history of hypertension, and all those over age 55 are at a higher risk of hypertension.
· Epidemiologic and animal studies indicate that the risk of stroke-related deaths is inversely related to potassium intake over the entire range of blood pressures, and the relationship appears to be dose dependent. The combination of a low-sodium, high-potassium intake is associated with the lowest blood pressure levels and the lowest frequency of stroke in individuals and populations. Although the effects of reducing sodium intake and increasing potassium intake would vary and may be small in some individuals, the estimated reduction
in stroke-related mortality for the population is large.
·A high salt intake is associated with atrophic gastritis in epidemiologic and animal studies, and there is also epidemiologic evidence that a high salt intake and frequent consumption of salt-cured and salt-pickled foods are associated with an elevated incidence of gastric cancer. The specific causative agents in these foods have not been fully identified.
Minerals and Trace Elements
The association of calcium, and to a lesser extent phosphorus and magnesium, with human health has received considerable attention in recent years. Most of the attention has been directed to the role of calcium and magnesium in regulation of blood pressure, the association of calcium and phosphorus intake with peak bone mass and development of osteoporosis, and the relationship of calcium to colon cancer. This evidence derives from epidemiologic studies, cross-sectional studies, many clinical metabolic studies, a few recent intervention trials with calcium supplements, a large number of animal experiments pertaining to calcium intake and bone diseases, and a few experiments on calcium intake and carcinogenesis. These studies are reviewed in Chapters 13 (Minerals), 20 (Hypertension), 22 (Cancer), and 23 (Osteoporosis).
Trace element nutrition has also received much attention in the past decade, especially the role of iron, zinc, selenium, and fluoride intake in human health (see Chapter 14, Trace Elements). With the exception of epidemiologic and laboratory studies on selenium and cancer and on fluoride and dental caries and a few studies on zinc and tumorigenesis, the data pertaining to trace element ingestion and chronic diseases are extremely limited and thus limit the committee's ability to arrive at conclusions. Of particular importance in the committee's assessment were potential antagonistic and synergistic interactions, which occur often among trace elements but about which knowledge is incomplete.
The following conclusions are based on this review.
· Epidemiologic, clinical, and animal studies suggest that sustained low calcium intake is associated with a high frequency of fractures in adults, but the role of dietary calcium in the development of osteoporosis and the potential benefits of calcium supplementsin amounts that exceed the Recommended Dietary Allowances (RDAs)in decreasing the risk of osteoporosis are unclear.
· Some epidemiologic studies have shown an association between calcium intake and blood pressure, but a causal association between low calcium intake and high blood pressure has not been established.
· A few data from epidemiologic and animal studies suggest that a high calcium intake may protect against colon cancer, but the evidence is preliminary and inconclusive.
· Unequivocal evidence from epidemiologic and clinical studies indicates that fluoridation of drinking water supplies at a level of 1 ppm protects against dental caries. Such concentrations are not associated with any known adverse health effects, including cancer.
· Low selenium intake in epidemiologic and animal studies and low selenium levels in human sera have been associated with an increased risk of several cancers. Moreover, some studies in animals suggest that diets supplemented with large doses of selenium offer protection against certain cancers. These data should be extrapolated to humans with caution, however, because high doses of selenium can be toxic.
· The data on most trace elements examined in this report (e.g., copper and cadmium) are too limited or weak to permit any conclusions about their effects on chronic disease risk.
Claims for the health benefits of dietary supplements have drawn substantial attention in recent decades. Patterns of supplement use have been surveyed in small-scale studies as well as in recent nationwide surveys conducted by the Food and Drug Administration, the U.S. Department of Agriculture, and the National Center for Health Statistics (see Chapter 18). These surveys and other investigations have focused on the use of multiple vitamin-mineral supplements and individual nutrients in the general population and among population subgroups such as the elderly, children, medical students, and health professionals. Attempts have also been made to ascertain the motivation for supplement use and the contribution of supplements to total dietary intake of vitamins and minerals. In comparison to fairly extensive data on the patterns of supplement use, information on their health effects is meager,
especially with regard to the long-term effects of multiple nutrient supplements. Several professional societies, notably the American Medical Association, the American Dietetic Association, the American Institute of Nutrition, and the American Society for Clinical Nutrition, have taken positions on the use of dietary supplements. The committee has based the following conclusion on the evidence reviewed in Chapter 18 (Dietary Supplements).
· A large percentage of people in the United States take dietary supplements, but not necessarily because of nutrient needs. The adverse effects of large doses of certain nutrients (e.g., vitamin A) are well documented. There are no documented reports that daily multiple vitamin-mineral supplements, equaling no more than the RDA for a particular nutrient, are either beneficial or harmful for the general population. The potential risks or benefits of the long-term use of small doses of supplements have not been systematically examined.
Coffee, Tea, and Other Nonnutritive Dietary Components
In the United States, nearly 3,000 substances are intentionally added to foods during processing. Another estimated 12,000 chemicals, such as vinyl chloride and acrylonitrile, which are used in food packaging and which inadvertently enter the food supply, are classified as indirect additives. Furthermore, minute quantities of many thousands of naturally occurring toxicants and environmental contaminants are found in foods. The health effects of these nonnutritive substances are reviewed in Chapter 17 (Coffee, Tea, and Other Nonnutritive Dietary Components) along with coffee and tea. The committee emphasizes the difficulty of assessing the long-term effects of this large group of substances. Many of them have not been tested or tested only in short-term experiments, and there are very few epidemiologic data. The committee also stresses the importance of considering dietary interactions (both synergistic and antagonistic), since many of these compounds occur simultaneously in foods. Chapter 17 includes a discussion of the potential health effects of mutagens in foods. Most of the evidence on nonnutritive components pertains to cancer risk.
The major conclusions pertaining to nonnutritive components are presented below.
· Coffee consumption has been associated with slight elevations in serum cholesterol in some epidemiologic studies. Epidemiologic evidence linking coffee consumption to the risk of CHD and cancer in humans is weak and inconsistent.
· Tea drinking has not been associated with an increased risk of any chronic disease in humans.
· The use of such food additives as saccharin, butylated hydroxyanisole, and butylated hydroxytoluene does not appear to have contributed to the overall risk of cancer in humans. However, this lack of evidence may be due to the relatively recent use of many of these substances or to the inability of epidemiologic techniques to detect the effects of additives against the background of common cancers from other causes. The association between food additives and cancer is also complicated by the long latency period between initial exposure to a carcinogen and the subsequent development of cancer.
· A number of environmental contaminants (e.g., some organochlorine pesticides, polychlorinated biphenyls, and polycyclic aromatic hydrocarbons) cause cancer in laboratory animals. The committee found no evidence to suggest that any of these compounds individually makes a major contribution to the risk of cancer in humans; however, the risks from simultaneous exposure to several compounds and the potential for adverse effects in occupationally exposed people have not been adequately investigated.
· Certain naturally occurring contaminants in food (e.g., aflatoxins and N-nitroso compounds) and nonnutritive constituents (e.g., hydrazines in mushrooms) are carcinogenic in animals and thus pose a potential risk of cancer in humans. Naturally occurring compounds shown to be carcinogenic in animals have been found in small amounts in the average U.S. diet. There is no evidence thus far that any of these substances individually makes a major contribution to cancer risk in the United States.
· Most mutagens detected in foods have not been adequately tested for carcinogenic activity. Although mutagenic substances are generally suspected of having carcinogenic potential, it is not yet possible to assess their contribution to the incidence of cancer in the United States.
· Overall, there is a shortage of data on the complete range of nonnutritive substances in the diet. Thus, no reliable estimates can be made of
the most significant exposures. Exposure to nonnutritive chemicals individually, in the minute quantities normally present in the average diet, is unlikely to make a major contribution to the overall cancer risk to humans in the United States. The risk from simultaneous exposure to many such compounds cannot be quantified on the basis of current evidence.
AHA (American Heart Association). 1982. Rationale of the diet-heart statement of the American Heart Association. Report of Nutrition Committee. Circulation 65:839A-854A.
AHA (American Heart Association). 1986. Dietary guidelines for healthy American adults: a statement for physicians and health professionals by the nutrition committee, American Heart Association. Circulation 74:1465A-1468A.
AHA (American Heart Association). 1988. Dietary guidelines for healthy American adults: a statement for physicians and health professionals by the nutrition committee, American Heart Association. Circulation 77:721A-724A.
Ahrens, E.H., Jr. 1985. The diet-heart question in 1985: has it really been settled? Lancet 1:1085-1087.
Ahrens, E.H., Jr., W.E. Connor, E.L. Bierman, C.J. Glueck, J. Hirsch, H.C. McGill, Jr., N. Spritz, L. Tobian, Jr., and T.B. Van Itallie. 1979. Symposium report of the task force on the evidence relating six dietary factors to the nation's health. Am. J. Clin. Nutr. 32:2621-2748.
American Diabetes Association. 1987. Nutritional recommendations and principles for individuals with diabetes mellitus: 1986. Nutr. Today 22:29-35.
Blackburn, H. 1979. Diet and mass hyperlipidemia: a public health view. Pp. 309-347 in R.I. Levy, B.M. Rifkin, B. Dennis, and N. Ernst, eds. Nutrition, Lipids, and Coronary Heart Disease: A Global View. Raven Press, New York.
CAST (Council for Agricultural Science and Technology). 1977. Dietary Goals for the United States: A Commentary. Report. No. 71. CAST, Ames, Iowa. 18 pp.
Connor, W.E. 1979. Too little or too much: the case for preventive nutrition. Am. J. Clin. Nutr. 32:1975-1978.
Council on Scientific Affairs. 1979. American Medical Association concepts of nutrition and health. J. Am. Med. Assoc. 242:2335-2338.
Dawber, T.R., R.J. Nickerson, F.N. Brand, and J. Pool. 1982. Eggs, serum cholesterol, and coronary heart disease. Am. J. Clin. Nutr. 36:617-625.
DHEW (U.S. Department of Health, Education, and Welfare). 1979. Healthy People: the Surgeon General's Report on Health Promotion and Disease Prevention. DHEW (PHS) Publ. No. 79-55071. Office of the Assistant Secretary for Health and Surgeon General, Public Health Service, U.S. Department of Health, Education, and Welfare. U.S. Government Printing Office, Washington, D.C. 177 pp.
DHHS (U.S. Department of Health and Human Services). 1988. The Surgeon General's Report on Nutrition and Health. DHHS (PHS) Publ. No. 88-50210. Public Health Service, U.S. Department of Health and Human Services. U.S. Government Printing Office, Washington, D.C. 712 pp.
Goldstein, J.L., and M.S. Brown. 1984. Progress in understanding the LDL receptor and HMG-CoA reductase, two membrane proteins that regulate the plasma cholesterol. J. Lipid Res. 25:1450-1461.
Grobstein, C. 1983. Should imperfect data be used to guide public policy? Science 83 4:18.
Gussow, J.D., and P.R. Thomas. 1986. Dietary goals and guidelines: a progressive jump or jumping the gun? Pp. 110-155 in The Nutrition Debate: Sorting Out Some Answers. Bull Publishing, Palo Alto, Calif.
Habicht, J.P., H. Orlans, W.H. Stewart, D.M. Hegsted, R.N. Brandon, and M.R. McHugh. 1979. Symposium: translation of scientific and nutrition findings to social policy. Fed. Proc. 38:2551-2569.
Harper, A.E. 1978. Dietary goalsa skeptical view. Am. J. Clin. Nutr. 31:310-321.
Hegsted, D.M. 1978. Dietary goalsa progressive view. Am. J. Clin. Nutr. 31:1504-1509.
Levy, R.I., B.M. Rifkin, B.H. Dennis, and N. Ernst, eds. 1979. Nutrition, Lipids and Coronary Heart Disease: A Global View. Raven Press, New York. 566 pp.
Nichols, A.B., C. Ravenscroft, D.E. Lamphiear, and L.D. Ostrander. 1976. Daily nutritional intake and serum lipid levels: the Tecumseh Study. Am. J. Clin. Nutr. 29:1384-1392.
NIH (National Institutes of Health). 1984a. Health Implications of Obesity: National Institutes of Health Consensus Development Conference Statement. Ann. Intern. Med. 103:1073-1077.
NIH (National Institutes of Health). 1984b. Osteoporosis: Consensus Development Conference Statement, Vol. 5. Publ. No. 421-132:4652. National Institute of Arthritis, Diabetes, and Digestive and Kidney Diseases and the Office of Medical Applications of Research. U.S. Government Printing Office, Washington, D.C. 6 pp.
NIH (National Institutes of Health). 1985. Lowering blood cholesterol to prevent heart disease: consensus conference. J. Am. Med. Assoc. 253:2080-2090.
NRC (National Research Council). 1980. Toward Healthful Diets. Report of the Food and Nutrition Board, Division of Biological Sciences, Assembly of Life Sciences. National Academy of Sciences, Washington, D.C. 24 pp.
NRC (National Research Council). 1982. Diet, Nutrition, and Cancer. Report of the Committee on Diet, Nutrition, and Cancer, Assembly of Life Sciences. National Academy Press, Washington, D.C. 478 pp.
O'Connor, T.P., and T.C. Campbell. 1986. Dietary guidelines. Pp. 731-771 in C. Ip, D.F. Birt, A.E. Rogers and C. Mettlin, eds. Dietary Fat and Cancer. Alan R. Liss, New York.
Olson, R.E. 1979. The U.S. quandary: can we formulate a rational nutrition policy? Pp. 119-133 in M. Chou and D.P. Harmon, Jr., eds. Critical Food Issues of the Eighties. Pergamon Press, New York.
Page, I.H., E.V. Allen, F.L. Chamberlain, A. Keys, J. Stamler, and F.J. Stare. 1961. Dietary fat and its relation to heart attacks and strokes. Circulation 23:133-136.
Palmer, S. 1983. Diet, nutrition, and cancer: the future of dietary policy. Cancer Res. 43:2509s-2514s.
USDA (U.S. Department of Agriculture). 1985. Report of the Dietary Guidelines Advisory Committee on the Dietary Guidelines for Americans. Human Nutrition Information
Service, U.S. Department of Agriculture, Hyattsville, Md. 19 pp.
USDA/DHHS (U.S. Department of Agriculture/Department of Health and Human Services). 1980. Nutrition and Your Health: Dietary Guidelines for Americans. Home and Garden Bulletin No. 232. U.S. Department of Agriculture and Department of Health and Human Services, Washington, D.C. 20 pp.
U.S. Senate. 1977. Dietary Goals for the United States, 2nd ed. Report of the Select Committee on Nutrition and Human Needs. Stock No. 052-070-04376-8. U.S. Government Printing Office, Washington, D.C. 83 pp.