Return
to Menu Page Return
to Menu Pageamong the people who ate the highest amounts of these foods, the risk of developing colon cancer was one-third that of people who ate the lowest amounts (Morgan et al., 1988). Furthermore, consumption of beans and lentils at least three times per week was associated with approximately 50 percent lower risk of developing prostate cancer than consumption of beans and lentils less than once a month (Mills et al., 1989b).
The association of vegetarianism with decreased risk of certain cancers appears to have correlates with biological parameters in Seventh-Day Adventists. Vegetarian Seventh-Day Adventists appear to have less colonic mucosal cell turnover than nonvegetarians and those at increased risk of colon cancer (Lipkin et al., 1985). This point is significant, because it is believed that decreased cell proliferation of the colonic mucosa may be a hallmark of decreased risk of colon cancer (Lipkin, 1974).
Macrobiotic diet. No studies to date have examined directly the role of a macrobiotic diet in chronic disease prevention. However, a number of studies have examined associations between macrobiotic diets and biological risk parameters such as blood pressure, ~cholesterol levels, and estrogen metabolism. The earliest of these studies were surveys of blood pressure and blood lipid levels, which were conducted in the Boston macrobiotic community. One of these studies showed that young adults eating a macrobiotic diet had blood pressure levels of 106 mm Hg systolic and 60 mm Hg diastolic, which was significantly lower than would be expected in the general population (Sacks et al., 1974). In fact, in comparison with people of similar age and sex in Framingham, MA, the subjects' systolic blood pressure was an average of 11 mm Hg lower and diastolic blood pressure 14 mm Hg lower (Sacks et al., 1975). In addition, those in the macrobiotic community who ate some animal food tended to have higher blood pressures than others in the macrobiotic community who did not (Sacks et al., 1974).
Blood lipid levels, a general indicator of coronary heart disease risk, also were substantially lower among the people eating macrobiotically than in the Framingham comparison group. In fact, average plasma total cholesterol levels among the macrobiotic vegetarians was 126 mg/dL versus a total average plasma cholesterol level of 184 mg/dL in the age-and sex-matched controls in the Framingham population (Sacks et al., 1975). Levels of ~low-density lipoprotein (LDL) cholesterol--the type of cholesterol that promotes heart disease--also were substantially lower in those eating a macrobiotic diet, averaging 73 mg/dL in the macrobiotic vegetarians and 118 mg/dL in the controls (Sacks et al., 1975). Although levels of high-density lipoprotein (HDL) cholesterol--the type of cholesterol that protects against heart disease--were also lower in the macrobiotic vegetarians (43 mg/dL vs. 49 mg/dL), the ratio of total to HDL cholesterol, a measure of the relative atherogenicity (i.e., ability to form plaques) of the blood lipid profile, was substantially lower among the macrobiotic vegetarians than in the comparison group (2.9 vs. 3.8). These differences persisted even when adjusted for weight differences. The relatively favorable blood cholesterol profile of macrobiotic vegetarians has been confirmed in several other surveys (Bergan and Brown, 1980; Knuiman and West, 1982; L. Kushi et al., 1988; Sacks et al., 1985).
In recent years, there has been increasing interest in the role of fat-soluble antioxidants in atherogenesis. It has been hypothesized that oxidation of LDL particles may be a critical step in the uptake of LDL by macrophages, as well as for some other mechanisms that ~increase the atherogenicity of LDL cholesterol (Steinberg and Witztum, 1990). The relative proportion of antioxidants to circulating LDL has been suggested as an additional measure of the atherogenicity of blood lipid levels (Berry, 1992). A study of macrobiotic vegetarians demonstrated that they had not only lower LDL levels but also higher plasma levels of antioxidants relative to cholesterol compared to nonvegetarians (Pronczuk et al., 1992).
The favorable cardiovascular disease profile of macrobiotic vegetarians is likely to be largely due to the relative avoidance of meat and dairy products. Indeed, when the diet of macrobiotic vegetarians was supplemented with 250 grams of beef per day, plasma total cholesterol increased by about 19 percent after 4 weeks, from an average of 140 mg/dL to 166 mg/dL (Sacks et al., 1981). Comparisons of macrobiotic vegetarians with lactovegetarians and nonvegetarians also indicate a direct relationship between average blood total cholesterol levels and dairy product intake (Sacks et al., 1985).
In the context of cancer risk, studies comparing women eating a macrobiotic diet with women eating a typical American diet demonstrate substantial differences in estrogen ~metabolism (Goldin et al., 1981, 1982). In fact, women eating a macrobiotic diet had substantially higher fecal excretion and lower urinary excretion of estrogens, with somewhat lower serum levels of estradiol. This point is significant because many cancers, especially breast cancer, are growth dependent on hormones such as estrogen. The altered estrogen metabolism profile of women eating macrobiotically may reflect a lower risk of breast cancer (Goldin et al., 1981, 1982).
Furthermore, in subsequent studies it was demonstrated that women eating macrobiotically had dramatically higher urinary excretion of lignans, such as enterolactone and enterodiol, and of isoflavonoids, such as daidzein and equol, than women consuming a lacto-ovo-vegetarian diet or an omnivorous diet. Women with breast cancer had the lowest levels of these phytoestrogens (Adlercreutz et al., 1986, 1987). These differences appeared to be related to greater intake of whole grains, legumes, and vegetables by the macrobiotic women.
It has been hypothesized that such a fiber-rich diet may, by the presence of these lignans and ~other weak estrogens (i.e., phytoestrogens) in the intestinal tract, stimulate the synthesis of sex hormone binding globulin in the liver and may thus decrease levels of free estradiol in the plasma (Adlercreutz et al., 1987). This may, in turn, reduce the risk of breast as well as other hormone-dependent cancers. It has also been suggested that these phytoestrogens (see the "Herbal Medicine" chapter) may actually compete with endogenous estradiol on the cellular level, further reducing the cellular proliferation and, hence, the potentially carcinogenic effects of estradiol (Price and Fenwick, 1985; Tang and Adams, 1980).
In addition to the general macronutrient differences (lower fat, higher complex carbohydrate intake) between macrobiotic diets and the standard American diet, certain foods in the standard macrobiotic diet may have specific anticancer effects. Examples of such foods, which are absent in the typical American diet, include various soyfoods, such as miso and tofu, and sea vegetables (see the "Diets of Other Cultures" section for a discussion of the health benefits of these foods).
Health risks associated with strict vegetarian diets. Except for vitamin B12 deficiency, diets ~that exclude meat or animal products do not produce deficiencies in adults if they are correctly followed. Nevertheless, there are reports in the literature that have associated some forms of vegetarianism with high risks of deficiencies in children and pregnant women (Debry, 1991). The nutrition of children on a vegetarian diet is considered to be adequate and well balanced when the diet contains dairy products and eggs. It has been suggested that a severe or strict vegetarian diet is not suitable for infants or toddlers. For example, serious deficiency states (e.g., rickets, osteoporosis, anemia, growth retardation) have been described in children subsisting on such regimens (Lentze, 1992). However, it is interesting to note that in a study that examined the maternity care records of 775 vegan mothers living on a commune in southern Tennessee, there was only one case of preeclampsia (Carter et al., 1987). The authors concluded that it is possible to sustain a normal pregnancy on a vegan diet, and the source of protein (i.e., animal or vegetable) does not seem to affect birth weight, as long as vegan mothers receive continuous prenatal care, supplement their diets with prenatal vitamins, calcium, and iron, and apply "protein-complementing" nutritional principles.
~Diets of other cultures. A cultural diet is defined as the diet of any group of people who share beliefs and customs. By this definition, everyone in the United States is a member of some cultural group. For many cultural groups, food plays an important role in maintenance of both spiritual and physical health. The following is a brief overview of several cultural diets--Asian, Mediterranean, and traditional Native American Indian--that are thought to provide some protection against many of the nutritionally related chronic illnesses prevalent among users of the mainstream diet in the United States today. Although there are many more cultural diets than are covered here, solid scientific research has not yet been collected to establish whether they provide any particular health benefits.
Asian diet. This diet is consumed predominantly by people living in China, Southeast Asia, Korea, and Japan. Rice is a staple and the center of the meal, and there is little or no use of dairy products. Soybean products are important sources of protein and calcium. Dishes incorporate many different ingredients and may be stir-fried or steamed. This diet, in its traditional form, is low in fat and high in carbohydrates and sodium (Kittler and Sucher, 1989).~Mediterranean diet. The Mediterranean Basin is geographically defined as an inland sea that touches three continents--Europe, Asia, and Africa--and is surrounded by 15 almost contiguous countries: Spain, France, Italy, the former Yugoslavia, Albania, Greece, Turkey, Syria, Lebanon, Israel, Egypt, Libya, Tunisia, Algeria, and Morocco. Divided by language and, historically, by political and religious conflict, the Mediterranean countries have for centuries been joined by a similar diet of daily staples.
The Mediterranean diet consists of a daily intake of grains, potatoes, pasta, greens and other vegetables, fruit, beans and other legumes (e.g., lentils, split peas), nuts, cheese, and yogurt. Fish, poultry, eggs, sweets, and red meat are eaten less frequently. However, olive oil and garlic are almost always consumed in abundance (Spiller, 1991). In the case of Spain, France, and Italy, it is their southernmost parts that are considered Mediterranean, defined by their use of olive oil.
Another important aspect of the Mediterranean diet is its emphasis on less refined complex carbohydrates (e.g., pasta) in place of sugar and the highly refined starches generally ~consumed in the United States, even though direct evidence for benefit in reducing disease risk is limited. Anticipated reductions in colon cancer by diets high in grain fiber diets have been difficult to document epidemiologically, although inverse associations with vegetables have been seen repeatedly. However, reduced constipation and reduced risk of colonic diverticular disease are clear benefits (Willett et al., 1990).
It is interesting to note that in the northern areas of many of the European Mediterranean countries, where there is more use of butter, other animal fats, and meat, there is also a higher incidence of cancer (La Vecchia, 1993).
Traditional Native American Indian diet. Many foods used throughout the world today were probably first used by Indians of North, Central, and South America--for example, beans, corn, cranberries, peanuts, peppers, potatoes, pumpkin, squash, and tomatoes. Today, Native American Indians live in areas that are vastly different from one another, so there is no single typical diet. In fact, traditional diets are prepared infrequently except for ceremonial occasions. This is true even for the Arizona Hopi, who still live in old villages that their ~ancient ancestors inhabited (Kuhnlein and Calloway, 1977). Nevertheless, in many American Indian diets, corn is the staple food. It is eaten fresh roasted or boiled, as hominy, or as cornmeal in a variety of dishes. Meat is eaten when it can be obtained by hunting or fishing, but because it is so expensive to buy, it is used sparingly. Milk and dairy products are not used often because of a high incidence of lactose intolerance (lactose is the primary sugar in milk). Berries, wild plants, and roots are used when available (Robinson and Lawler, 1982).
Research base. The following provides an overview of research on the effectiveness of some components of the Asian, Mediterranean, and traditional Native American Indian diets in lowering some risk factors for disease.
Asian diet. In a cohort study of 265,000 people in Japan, consumption of miso soup (a food made from soybeans) appeared to reduce the risk of breast cancer (Hirayama, 1986) and stomach cancer (Hirayama, 1981). A similar inverse association was seen between stomach cancer and tofu intake (Hirayama, 1971). Furthermore, miso has been observed to inhibit ~formation of mammary tumors in rodents (Baggott et al., 1990) and may have antioxidant properties as well (Santiago et al., 1992).
Antioxidant properties have been proposed as a principal mechanism by which dietary compounds such as beta-carotene, vitamin E, indoles, and others exert cancer-preventive effects (Steinmetz and Potter, 1991a, 1991b). It has also been suggested that sea vegetables, perhaps through their high concentration of alginic acid, a type of dietary fiber, may decrease the risk of breast cancer (Teas et al., 1984; Yamamoto et al., 1987). Beans and bean products, especially those derived from soybeans (e.g., miso, tofu, tempeh), also contain protease inhibitors (Messina and Barnes, 1991), isoflavonoids (Adlercreutz et al., 1987), and other compounds that may play roles in cancer prevention (Axelson et al., 1984).
Mediterranean diet. The high consumption of olive oil is considered a major contributor to the disease-preventive aspects of this diet. Olive oil is a monounsaturated fat, meaning that somewhere along the fat, or fatty acid, molecule there is a single site not completely ~"saturated" with hydrogen atoms. Substituting monounsaturated fats, such as olive oil, for saturated fat in the diet has been shown to reduce LDL cholesterol without affecting HDL cholesterol, thus providing an improved ratio (Mensink and Katan, 1992). In addition, monounsaturated fats in the diet have been found to reduce blood sugar and triglycerides in adult-onset diabetics (Garg et al., 1992).
In one of the first studies of its kind, researchers in France placed approximately 300 patients who had recently had a heart attack (myocardial infarction) on a Mediterranean type of diet and compared their incidence of having a second myocardial infarction with that of a control group of patients who were placed on the standard therapeutic diet. The experimental group consumed significantly more bread and fruit, a margarine with a fatty acid composition comparable to that of olive oil, and significantly less butter, cream, and meat than the control group. After a followup of about 27 months, there were only 3 cardiac deaths and 5 nonfatal myocardial infarctions in the experimental group versus 16 cardiac deaths and 17 nonfatal myocardial infarctions in the control group (de Lorgeril et al., 1994). It is interesting to note that the patients on the Mediterranean type of diet had increases in ~blood levels of vitamin E and C while controls did not.
Garlic, a staple of the Mediterranean diet, also has been implicated as a major disease-preventive food. A growing number of reports in the medical literature suggest that garlic supplementation may be effective in decreasing serum cholesterol levels by as much as 15 to 20 percent and thus may have a protective effect against cardiovascular disease (Kleijnen et al., 1989; Turner, 1990). Many of these studies have been faulted for having methodological problems, although a recent meta-analysis of the various studies reporting a cholesterol-lowering effect found that garlic did appear to significantly reduce total serum cholesterol (Silagy and Neil, 1994).
There are also reports suggesting that garlic may prevent the development of cancer in humans (Dorant et al., 1993). Lin and colleagues (1994) reported that processed garlic effectively reduced the amount of DNA damage caused by N-nitroso compounds, which are found in many foods such as cooked meat and have been implicated as carcinogens (cancer-causing compounds).~Traditional Native American Indian diet. In the case of the Hopi and Papago tribes, studies have shown that traditional foods have mineral content superior to federally provided commodity foods (Calloway et al., 1974). Followers of traditional Native American diets have found ways of maximizing available nutrients; an example is in the techniques of processing the corn used in tortillas, a staple in diets derived from the Mexican and Central American tradition. The corn is soaked in lime, which softens the skin of the corn kernels as well as increasing the calcium content of the resulting tortillas (Katz, 1987). Traditional lime soaking also liberates bound niacin in the corn. Because milled corn has been substituted for lime-soaked corn in Native American Indian diets, niacin deficiency has become a problem, and incidences of niacin-deficiency-induced pellagra have increased. Although few studies have been done on the possible disease-preventive aspects of the traditional Native American Indian diet, health surveys have found that heart disease and cancer, two diet-related diseases, are virtually nonexistent in some Indian populations, such as the Navajo (Reese, 1972).
Barriers and Key Issues Related to Alternative Diet and Nutrition Research~This chapter has so far dealt primarily with basic and clinical research relating to diet and nutrition interventions either for preventing or treating illness. However, to discuss such research without mentioning outside factors that will affect how research results are evaluated or disseminated to the public provides only a small part of the overall picture. Because nutritional and dietotherapy interventions affect an array of biochemical and physiological processes in the body, evaluating the interventions' effectiveness requires equally complex methodologies. Also of concern to those who work in this field is to research and develop "alternatives" to institutionalized nutrition and feeding programs that directly contribute to diet-and nutrition-related chronic disease. Finally, there is the issue of dissemination; no matter how good the research, it is valuable only if it reaches those who will benefit from it. Research data should be disseminated not only to the doctors or other health personnel who may prescribe such therapies, but also to the eventual target audience (i.e., the patients). Some of these issues are discussed below.
Study Design
~In most instances, it is virtually impossible to conduct a double-blind study of a dietary regimen. Patients obviously will know whether they are being fed a normal diet or a modified diet. Therefore, a single-blind design (in which the evaluator of the data is "blind" to who receives which treatment) is more appropriate for most dietary studies. However, in some instances a double-blind study is appropriate, such as when all subjects are given a tasteless, colorless pill. Also, investigators must consider the possibility of a negative placebo ("nocebo") effect in the control group as well as a placebo effect in the treatment group. In other words, patients who think they are not getting the therapy may not get better because of this knowledge--just as patients who think they are receiving an effective treatment may get better spontaneously, independent of the therapy. (See the "Mind-Body Interventions" chapter for a discussion of placebo and nocebo effects.)
The recommendation of the recent OTA report Unconventional Cancer Treatments offers an example of one possible methodological approach to evaluating many nutritional interventions. As a practical approach to evaluating the treatments they had examined, the OTA report's authors proposed a "best case review" conceptually similar to approaches ~used by NCI for evaluating biological response modifiers (BRMs) (Office of Technology Assessment, 1990). Best case reviews are discussed in detail in appendix F.
BRMs are agents that exert anticancer effects in a novel manner. Unlike conventional cytoxic therapy, which kills tumor cells with slightly less chemical toxicity than chemotherapy, BRMs stimulate, or potentiate, the body's immune system to overcome or at least restrain the invading tumor. Most BRMs currently being investigated are genetically engineered copies of peptides found naturally in the human body; among them are tumor necrosis factor, interleukins, and interferons. At an early stage in this clinical research, NCI researchers recognized that because BRMs act biologically and not chemically, they had to be evaluated using a procedure that is significantly different from the one used for standard cytotoxic agents (Oldham, 1982). With this in mind, in 1978 NCI created a special Biological Response Modifiers Program to coordinate research to identify, study, and clinically evaluate BRMs.
Clinical evaluation of BRMs is difficult because determining the optimal dose or dosing ~schedule for the agent is critically important. Also important is the identification of responsive tumor types and even of the stage of disease and metabolic condition of patients who are treated (Creekmore et al., 1991; Hawkins et al., 1986; Oldham, 1985). Consequently, experts in BRM research advise against testing BRMs in large, controlled clinical trials (Phase III trials) until after these parameters have been optimized in careful individual tests of the proposed therapy in patients believed to be the most likely to show a favorable response (Creekmore et al., 1991; Hawkins et al., 1986; Oldham, 1985). This type of evaluation is inherently exploratory and observational; therefore, it cannot be conducted in the same rigidly stipulated way as a Phase III trial.
Because such studies demand the continual application of good clinical and scientific judgment, they are demanding of the time and energy of investigators, who typically are experts keenly interested in the therapy under evaluation. This time-consuming process is necessary because the consequence of proceeding prematurely to a Phase III trial is an inconclusive or falsely negative result. Since dietary and nutritional interventions often affect biology in ways at least as complicated as the anticancer BRMs currently being evaluated by ~NCI, recommendations concerning the evaluation of BRMs should apply to the evaluation of many diet and nutrition interventions.
Furthermore, careful scientific judgment is needed as to when sufficient evaluation has gone on and confirmation is in order. For example, there is general agreement in the scientific community that the time is now at hand to "confirm" the efficacy of megadose antioxidants in the prevention of coronary artery disease progression in a large Phase III trial (Steinberg, 1993). This assessment is correct even though such trials are costly, and there is always a risk of a false negative result if all the relevant parameters have not yet been fully optimized. The interaction of a variety of factors in producing a therapeutic outcome would be expected to be of particular importance in biological or nutritional therapies (Christensen, 1993; Weglicki et al., 1993).
An excellent example of the superiority of good scientific judgment over the premature use of the controlled clinical trial is the discovery by George C. Cotzias that L-dopa is an effective treatment for Parkinson's disease (Cotzias et al., 1967). When Cotzias began his ~studies, L-dopa was regarded as an interesting therapeutic idea that had been determined to be without utility in a series of careful clinical trials that included controlled double-blind, Phase III methodology (Fehling, 1966; Lasagna, 1972). Had Cotzias not used good scientific judgment and persistent curiosity by testing L-dopa in careful, uncontrolled protocols that included larger doses than in the earlier, methodologically flawed trials, it is entirely possible that recognition of L-dopa's enormous value would have been delayed for decades or never recognized at all.
Phase III trials of orthomolecular therapy, for example, may also require a similarly innovative approach. One possibility, after preliminary, dose-optimization studies have been completed, is to randomly allocate suitable patients to conventional or orthomolecular programs. Much of the evaluation would have to be "open"; however, symptom-rating scales could be scored by observers unaware, or "blinded," to which patients received which treatment.
First and foremost, however, is the requirement for an openminded approach by intelligent ~and skeptical clinical investigators whose effort is respected by their peers. Such studies also will require good-faith cooperation from academic medical units and the support provided by adequate research funding. Evaluation of the research will have to include a recognition that initially negative results do not prove a therapy valueless: the therapy may merely have been incorrectly tested. Investigators should strive to develop hypotheses that can be objectively tested. Even when testing mechanisms are novel, they must be rigorous and reproducible by independent investigators, and their results must be convincing to openminded but skeptical reviewers. These objectives can and must be achieved. The cost of not achieving them may be the unnecessary delay or, worse, the complete dismissal of an effective treatment for a previously untreatable debilitating illness.
Alternatives to Federal and Other "Institutionalized" Programs That Influence Diet and Contribute to Chronic Disease
Not even increased openmindedness in mainstream research or increased Federal funds for research may be enough to get effective alternative diet and nutrition prophylactic and ~therapeutic treatments into more general use. Indeed, without some significant changes in Americans' beliefs and expectations about food and nutrition, promising research results--whether alternative or conventional--will have little impact. For example, studies indicate that Americans are quite aware of the relationship between nutrition and health (Cotugna et al., 1992). However, during the past two decades they have made little apparent progress toward meeting the RDAs (U.S. Department of Agriculture and U.S. Department of Health and Human Services, 1990), which recent research indicates may already be too low for many vitamins and nutrients. Although there has been an increase in consumption of low-fat milk and a decrease in the consumption of meat and eggs during the past decade, USDA's 1987-88 national food consumption survey (U.S. Department of Agriculture, 1988) indicated that Americans, on average, eat only one serving of fruit or fruit juice and two servings of vegetables per day. This amounts to roughly half the recommended Federal Government minimum (Patterson et al., 1990) and much less than the minimum advocated by many others. Furthermore, the consumption of saturated fat by women has consistently remained around 13 percent of total calories (Welsh, 1991).
~Numerous Government programs and information dissemination channels exist that potentially could have a major positive influence on American dietary habits. Unfortunately, many, if not most, are having a negative rather than a positive impact on Americans' dietary knowledge, beliefs, and practices. These programs and channels include
• Government feeding and food support programs,
• Public education and the mass media,
• School-based and worksite programs, and
• Health care provider settings.
Government feeding and food support programs. The Food and Nutrition Service of USDA administers 14 food assistance programs that aim to "provide needy people with access to a more nutritious diet, to improve the eating habits of the Nation's children, and to stabilize
~farm prices through the distribution of surplus foods" (U.S. Department of Agriculture, 1993a). More than 25.4 million people participated in the Food Stamp Program in 1992, and more than 5.6 million participated in the Special Supplemental Food Program for Women, Infants, and Children (WIC). In addition, approximately 25 million children participate in the National School Lunch Program (NSLP) each day, and an average of 900,000 people participate daily in the Nutrition Program for the Elderly.
Studies have reported that participation in the Food Stamp Program or the size of the food stamp benefits, or both, have had a positive impact on the availability of nutrients. However, the effect of these programs on nutrient intake is negligible. Only WIC was found to increase intake of numerous nutrients, including iron, calcium, and vitamin C, among pregnant women (Rush et al., 1988) and preschool children (Rush et al., 1988). Moreover, when participants in NSLP were compared to nonparticipants, increased nutrient intakes of vitamin A, vitamin B6, calcium, and magnesium (Hanes et al., 1984)--vitamins typically deficient in the school-aged population (Nelson et al., 1981)--were observed in the NSLP participant population.~Whether these various feeding programs provide adequate nutrition for Americans deserves critical analysis. For example, NSLP has been criticized for maintaining its outdated purpose of preventing nutritional deficiencies without including food patterns that would prevent such prevailing chronic diseases as heart disease, hypertension, cancer, and atherosclerosis (American School Food Service Association, 1991; Citizens' Commission, 1990). In light of the findings on the positive effects of fruits and vegetables (Steinmetz and Potter, 1991a, 1991b) and the negative effects of saturated fats (Willett, 1990) on health and chronic disease, it may be necessary to modify NSLP to at least meet the Dietary Guidelines for Americans (U.S. Department of Agriculture and U.S. Department of Health and Human Services, 1990).
In September 1993, assistant secretary of agriculture Ellen Haas announced plans to improve NSLP by doubling the amount of fresh fruits and vegetables supplied to schools and reducing the amount of fat in commodity foods (U.S. Department of Agriculture, 1993b). This modification is urgently needed to update NSLP so that it will be in line with the current scientific findings on diet and disease. Similar modifications should be made ~across all USDA feeding and food assistance programs to help Americans consume a better diet.
Furthermore, for the agricultural system to meet even the current USDA dietary guidelines, adjustments are required in the mix and output of farm products. Appropriate new food policies need to be in place to support such changes. For example, in the commodity area, the price and income support programs put a premium on milk fat, and surplus disposal operations are designed to increase the supply of high-fat butter and cheese on the market at artificially low prices. This system runs counter to encouraging better diet and nutrition in the population.
Public education and role of the mass media. The primary goal of public nutrition education is to bring about behavioral changes in individuals', groups', and populations' dietary patterns presumed to be detrimental to health. The educational program is designed to provide enough knowledge so that healthy choices in nutrition can be made. The mass media, including magazines, newspapers, and television, are a major source of nutrition information ~for the public (American Dietetic Association and International Food Information Council, 1990). An analysis of mass initiatives in the area of promoting healthy diet and nutrition choices indicates that such campaigns can be useful in setting the stage for behavior change (DeJong and Winsten, 1990). One example is a joint NCI-Kellogg (the cereal manufacturer) initiative that promoted consumption of a high-fiber diet through advertising and food labeling on cereal boxes. On the basis of purchase data from supermarkets in the Baltimore and Washington, DC, metropolitan area, the purchase of high-fiber cereals increased 37 percent in the 48 weeks of the initiative (Levy and Stokes, 1987).
Mass media-based education programs not only affect the intermediary steps to behavior change but also have proved to have a more direct influence on health, such as affecting changes in eating patterns and disease risk factors. For instance, the Finnish North Karelia Project showed that education programs using mass media strategies can markedly reduce certain coronary heart disease risk factors (Vartiainnen et al., 1991).
In recent years, the social marketing approach, which draws marketing techniques from the ~private sector and focuses on the thorough understanding of consumer needs and opinions, has been used increasingly in health promotion campaigns (J. Ling, 1992). This approach can yield promising new insights into consumer behavior and into product and strategy design (Walsh et al., 1993), thereby enhancing the efficacy of a health promotion initiative.
The National Heart, Lung, and Blood Institute's (NHLBI's) National High Blood Pressure Education Program (NHBPEP) and National Cholesterol Education Program are examples of Federal public education programs that extensively use mass media and social marketing strategies to convey health messages to the public. Both programs employ the strategy of focusing on raising knowledge and awareness on two tiers: among the public and among health care professionals. From NHBPEP's inception in 1972, awareness, treatment, and control rates for high blood pressure have increased dramatically (Rocella and Lenfant, 1992), and age-adjusted stroke mortality has fallen nearly 57 percent (Rocella and Horan, 1988).
Unfortunately, few nutrition messages in the mass media promote a healthy diet. Indeed, ~advertisements for foods that are high in sodium, fat, or sugar often compete directly against nutrition messages designed to help people make better food choices. For example, breakfast cereals, snacks, and fast foods are among the most heavily advertised products on television programs aimed at children (Cotugna, 1988), and the television "diet" consists of foods primarily of low nutritional value (Story and Faulkner, 1990). Television viewing also appears to affect food consumption. Studies have reported, for instance, that the amount of time spent watching television directly correlates with the request, purchase, and consumption of foods advertised on television (Clancy-Hepburn et al., 1974; Gorn and Goldberg, 1982; Taras et al., 1989). Consequently, the mass media, the food industry, the Government, and health professionals should collaborate to broadcast health promotion messages more extensively. One such example is the airing of public service announcements on Saturday morning children's television by a major fruit-processing company, which was prompted by the "Five-a-Day for Better Health" campaign initiated by NCI and the Produce for Better Health Foundation.
School-based and worksite programs. Schools are an ideal setting in which to model and ~encourage healthy lifestyle behaviors. More than 95 percent of American youth aged 5 to 17 are enrolled in schools (U.S. Department of Education, 1990). School-based nutrition education and physical activity programs appear to be ideal venues for effecting change in lifestyle-related risk factors for heart disease, cancer, and obesity. Children eat one to two meals per day in school, and the cafeteria can be a learning laboratory where students can practice and experience positive nutrition habits they learn from the school curriculum. Previous studies in school-based cardiovascular research, including the "Know Your Body" program (Walter et al., 1988), have shown that health promotion in schools can have a favorable impact on nutrition knowledge and diet-related skills (Contento et al., 1992) as well as on specific outcomes, such as blood cholesterol level, carbohydrate intake, fitness, blood pressure, and smoking status (Stone et al., 1989). In addition, school-based health promotion programs have had positive impacts on obesity (Resnicow, 1993).
Additional research of longer duration that includes multiple components (school food service, curriculum, family outreach) is needed to determine the degree to which schools can affect the exercise and diet habits of children. These studies could be similar to the ongoing ~NHLBI-funded "Child and Adolescent Trial for Cardiovascular Health" and the "Eat Well and Keep Moving" project at the Tesseract schools in Baltimore, MD.
Worksites are another important channel for promoting nutrition. Nearly 70 percent of adults between the ages of 18 and 65 are employed (U.S. Bureau of the Census, 1986). Thus worksites provide access to large numbers of people and offer the opportunity to make environmental and social norm changes that support healthy eating (Sorensen et al., 1986). Indeed, worksite educational programs have been shown effective in weight control (Sherman et al., 1989), cardiovascular risk reduction (O'Brien and Dedmon, 1990), smoking cessation (Windsor and Lowe, 1989), and cancer screening (Heimendinger et al., 1990). By modifying cafeteria menus and policy for meals served at corporate functions to support healthier choices and allowing time at work for nutrition education activity, worksites can be promising vehicles in modifying employees' eating habits. However, efforts should be made to overcome low participation rates and high dropout rates in worksite programs (American Dietetic Association, 1986).
~Education of health care providers and patient counseling. Patients place a great deal of credibility in the nutrition advice given to them by their physicians (American Dietetic Association, 1990). Many physicians, however, fail to provide such advice to their patients. A study by the University of Minnesota found that only 10 percent of surveyed physicians gave nutrition advice to more than 80 percent of their patients (Kottke et al., 1988). Although many physicians view nutrition as an effective tool that should be used in medical practice, there are significant barriers that keep physicians from adequately counseling their patients on issues of diet and nutrition. These barriers include lack of time, adequate staffs, and insufficient insurance coverage (Glanz and Gilboy, 1992). In addition, physicians' perceived inability to effectively alter their patients' lifestyle practices contributes to this problem (Wechsler et al., 1983). Another major contributor to physicians failing to give their patients nutrition counseling may be their own lack of nutrition knowledge. More than one study of practicing physicians has found that only about half of those surveyed felt prepared to provide dietary counseling to their patients (Kimm et al., 1990).
This lack of adequate nutrition knowledge among physicians may be partly due to a deficient ~nutrition curriculum in U.S. medical schools. In fact, a survey of 45 U.S. medical schools by the National Academy of Sciences found the state of nutrition education in medical schools to be largely inadequate to meet the needs of patients and the medical profession (National Research Council, 1985). Improved standards for nutrition education in U.S. medical schools appear to be necessary if the vast preventive and therapeutic role of nutrition in health care is to be exploited to the fullest.
In the context of alternative diets or therapies, an additional barrier to effective counseling by conventional health care providers is the sometimes outright hostility toward alternative therapies held by these providers. For example, surveys have indicated that many cancer patients do not tell conventional providers that they are pursuing alternative therapies, in part, because the conventional health care provider is often unsupportive or skeptical of such therapies (Eisenberg et al., 1993). In extreme cases, some conventional providers will refuse to treat a patient whom the provider knows is seeing an alternative practitioner for the same condition. Such attitudes are substantial hurdles to overcome for both the adequate and objective evaluation and dissemination of effective alternative therapies.~Another problem physicians may encounter when trying to give nutrition counseling to patients is that American society is becoming increasingly multicultural. Health professionals often have difficulty communicating with clients whose cultural heritage is different from their own. This problem is particularly acute when a physician is dealing with someone who comes from a cultural group or society where health and religion are intertwined (Kittler and Sucher, 1989). For example, "looking good" is a common goal in many technologically developed societies, but various cultural groups have a different view. Mexican Americans, African Americans, and other ethnic groups do not share the typical American concepts of appropriate body size, particularly for adult women (Massara, 1980; Schreiber and Homiak, 1981; Stern et al., 1982). It has been suggested that the "mainstream" standards for weight in adult women are, in fact, based more on the value of thinness--which is related to youth and higher socioeconomic status (Cassidy, 1991; Sobal, 1991)--than on science or epidemiology (Ritenbaugh, 1982).
Therefore, successful nutritional counseling depends on culturally sensitive communication strategies; health care practitioners must be both knowledgeable about general ethnic, ~regional, and religious food habits and aware of individual practices and preferences. Health care professionals can improve cross-cultural counseling through a four-step process of self-evaluation, preinterview research, indepth interviewing, and unbiased data analysis. A detailed description of the rationale for these steps can be found in Kittler and Sucher (1990).
However, the success of cross-cultural counseling cannot always be measured by a patient's adherence to a diet. Differences in worldview, traditional food habits, and factors that influence dietary adaptation may be of greater consequence to a client than the health implications of the diet. The best chances of compliance occur when the health care practitioner is aware of personal cultural assumptions and is knowledgeable about the cultural heritage of a patient and its specific influences on the patient's food habits, and when diet modifications are made with consideration for individual cultural and personal preferences.
There are some examples of successful intervention programs that have been based on ~indepth studies of the total context of ethnic food consumption. Hall (1987) noted that materials designed for Mexican-American diabetics not only had to be translated into Spanish but also had to be redesigned, incorporating culturally relevant concepts, methods, meal plans, and activities, to be effective. The same study also found it advisable to incorporate recommendations for traditional home remedies that have been shown scientifically to be of value in the treatment of diabetes. For example, a diabetes intervention directed toward Mexican Americans may include the use of cooked prickly pear cactus (nopales), which has long been used in Mexican folk medicine to control diabetes (Frati-Munari et al., 1983). Traditional remedies, however, are encouraged only as complements to biomedical treatments; the Hall program suggests that prescribed diabetes medications be taken with traditional herbal teas (Hall, 1987).
Another excellent intervention program is a physician-based system of dietary risk assessment and intervention, designed for use with low-literacy, low-income southern populations (Ammerman et al., 1991, 1992). This program focuses on the top 20 contributors of saturated fat and cholesterol that, based on the National Health and Nutrition ~Examination Survey II (NHANES II) data, are commonly found in the diets of African-American populations in the South. Attention is also given to traditional southern food preparation practices, such as baking with lard, frying with vegetable shortening, and seasoning vegetables with meat fat (Ammerman et al., 1991). All assessment and intervention materials are based on food rather than on nutrients. Diet change recommendations are linked with recipes in a southern-style cookbook.
For those who do little cooking at home (a growing population), information is provided on how to eat sensibly at fast-food restaurants. Low-cost dietary alternatives and southern food preferences are emphasized throughout the materials, which are written at the fifth-to sixth-grade reading level. The goal of the program is to reduce saturated fat and cholesterol intake while preserving ethnic eating patterns--that is, to adapt the traditional diet rather than introduce a radical transformation of eating patterns (Ammerman et al., 1992). Evaluation of the program shows promising results, both in the physicians' administration of the program and in changes in patients' attitudes. Currently underway is a 5-year, randomized clinical trial of the effectiveness of the program in lowering cholesterol among patients in rural Virginia ~and North Carolina (Ammerman et al., 1992). In addition, a "northern" version of this approach and these materials is being tested.
Programs of the type described by Hall and Ammerman cannot be developed in the absence of the necessary data specific to the ethnic group being targeted. Multidisciplinary research with a nutritional-anthropological focus is necessary to explore these issues further.
Research Needs and Opportunities in Diet and Nutrition
It is virtually impossible to list all of the research opportunities in diet and nutrition that should be pursued more extensively and vigorously. Rather, this section presents broad areas where the data indicate that more intense efforts might yield significant results.
Optimal Levels of Vitamins, Minerals, and Other Nutritional Supplements
Although there have been many studies to determine the effect of a single vitamin deficiency, ~few studies have attempted to determine the optimal dietary requirement for most vitamins and minerals. There is increasing evidence that the consumption of nutrients at RDA levels is not adequate for promoting optimal health. Thus, nutritional supplementation above the RDAs for many vitamins and minerals may be indicated. The following areas of research in vitamin and nutritional supplementation are likely to yield significant results:
Research is needed to determine how and at what levels such antioxidants as vitamins E and C and beta-carotene provide optimal immune enhancement.
• There is growing evidence that some carotenoids can directly affect cancer cells. Mechanisms may include free radical and other charged particle quenching, which would result in less damage to DNA; decreased adenylate cyclase activity, which would decrease proliferation of the cancer; generation of regulatory proteins that could alter cell cycles and metabolism; and other mechanisms as yet unknown (Bendich, 1991). Carotenoids have been shown to protect cells from mutagens. Further, research into the direct effects of antioxidants on cancer cells and on DNA repair mechanism is needed.~Research is needed into the role that many vitamins play aside from being enzyme cofactors. In view of the interest in free radicals or reactive molecular intermediates in the pathogenesis of a variety of medical and neurological diseases, there is a wealth of opportunities for research on the ability of vitamins and other nutritional supplements to prevent or reverse the effects of these types of molecules.
Research is needed in the United States to provide data from controlled studies to verify mostly European work on the clinical efficacy of minerals such as magnesium and selenium in the treatment of disease. More extensive intervention studies are needed to determine whether adding mineral supplements will improve the preventive effects of other dietary interventions (e.g., salt and fat restriction) against cardiovascular disease, and to verify the promising effects from Europe in prevention of abnormalities in pregnancy. In particular, the extended study of magnesium treatment of bronchial asthma is also indicated. There are already many clues on magnesium's mechanisms of action, and there are many data on clinical efficacy in a number of clinical diseases or complaints. It would be more feasible--not to mention less expensive--to set up double-blind intervention studies supplementing ~subjects' existing diets with magnesium or selenium, rather than try to get them to completely change their diet to include more foods containing these minerals.
A recent National Institutes of Health (NIH) report (1994) on calcium recommended that optimal levels of intake by women to prevent osteoporosis should be 1,500 mg per day. Because the usual American intake of magnesium is no more than 300 mg per day, such a level of calcium intake would constitute a calcium-magnesium ratio of 5:1. It is noteworthy that in Finland, where the prevalence of osteoporosis is high (Simonen, 1991), the average calcium-magnesium intake ratio is 4:1, a ratio that has been associated with the highest death rate in young to middle-aged men from ischemic heart disease in the world (Karppanen et al., 1978). An NIH consensus development conference on magnesium, similar to the one held recently on calcium, would provide a much-needed forum for elucidating other avenues of research that may be warranted on this important mineral.
Alternative Dietary Lifestyles and Cultural Diets
~Studies of Seventh-Day Adventists, macrobiotic vegetarians, and populations eating Asian and Mediterranean types of diets indicate that these groups are at lower risk of heart disease and some cancers. Likewise, studies usually demonstrate that blood pressure and blood lipid levels fall when participants follow a vegetarian diet (Cooper et al., 1982; Kestin et al., 1989; Margetts et al., 1986; Rouse et al., 1986). Prospective studies of vegetarian groups other than California Seventh-Day Adventists have found a decreased risk of heart disease and cancer as well as a decreased risk of death from all causes (Burr and Butland, 1988; Burr and Sweetnam, 1982; Chang-Claude et al., 1992; Frentzel-Beyme et al., 1988). Vegetarian populations, including Seventh-Day Adventists, also appear to be at decreased risk of other diseases such as gallbladder disease (Pixley et al., 1985) and diabetes mellitus (Snowdon and Phillips, 1985). Furthermore, such vegetarian diets as advocated under the Pritikin program appear to improve control of diabetes (Barnard et al., 1983).
Accordingly, studies of populations eating vegetarian and some cultural diets provide evidence that alternative dietary patterns may have a major impact on disease risk. Specifically, such studies show evidence of potentially profound implications for the risk of ~developing heart disease, certain cancers (such as colon or prostate cancer), and other chronic illnesses, such as diabetes. For example, in addition to containing large amounts of antioxidants, beans, leafy green vegetables, whole grains, many fruits, and fish are very rich in magnesium (Seelig, 1980). The Pritikin, Mediterranean, Seventh-Day Adventist, and macrobiotic diets are, thus, rich in magnesium. Studies have shown that magnesium supplementation of animals on atherosclerosis-inducing diets protects against arterial damage. Animal studies also have shown that magnesium deficiency increases hypercholesterolemia (especially the LDLs) and increases vulnerability to oxidative damage (Rayssiguier et al., 1989, 1993). It has recently been shown that magnesium repletion and vitamin E are mutually enhancing in protecting against magnesium deficiency-and stress hormone-induced cardiac necrosis (Freedman et al., 1990, 1991; Guenther et al., 1992, 1994a; Weglicki et al., 1992), and that vitamin E and magnesium deficiency shortens the time needed to induce atherosclerosis (Guenther et al., 1994b).
There is experimental evidence that magnesium deficiency (at least in very young rodents) can cause leukemias and lymphomas (Averdunk and Guenther, 1985; Battifora et al., 1968, ~1969; Bois, 1968; Bois and Beaulnes, 1966; Bois et al., 1969; Hass et al., 1981a, 1981b; Jasmin, 1963; McCreary et al., 1967), especially when there is also deficiency in such antioxidants as vitamins C and E, and in selenium (Aleksandrowicz, 1975). The protective role of magnesium against certain diseases may be supported by epidemiological findings about geographic areas low in magnesium where there is a high prevalence of human and cattle lymphoid neoplasms, leukemias, and gastric cancers (Aleksandrowicz, 1973; Aleksandrowicz and Skotnicki, 1982; Seelig, 1979, 1993).
The studies of vegetarian groups and Asian and Mediterranean populations are congruent with the growing body of studies in other populations that indicate a potentially profound role for dietary factors in the etiology of various chronic illnesses. These include growing evidence of the undesirable health effects of meat and high-fat dairy intake and the health promotion effects of abundant consumption of vegetables, fruits, monounsaturated fats, garlic, and whole cereal grains.
The translation of the findings of these animal and human studies into therapeutic approaches ~may alleviate the burden of some of these diseases. Equally important are measures that are being taken by industry and the Federal Government to support healthful dietary habits. Studies need to be undertaken on a wide variety of alternative diets that have been found to be beneficial, including
• vegetarian diets,
• ultra-low-fat diets,
• high-polyunsaturated-fat diets,
• Mediterranean-type diets, and
• diets rich in soy foods, such as East Asian diets.
Initiatives also are needed along the following lines:~ • There is a need for a more critical examination of dairy products on health, especially regarding fractures.
• There is a need for a more critical look at the effects of meat consumption on health--for example, on coronary heart disease, colon cancer, and fractures.
• More detailed data are needed on the effects of fruits, vegetables, monounsaturated fats, and garlic on cancer, coronary heart disease, cataracts, stroke, and so forth.
• There is a need for more information on the long-term effects of overrefined carbohydrates (e.g., sugar) on the human metabolism and immune functioning.
Qualitative research is needed on various aspects of cultural diets and the effects of cultural beliefs on health and illnesses. Such studies might include the following:
• Qualitative research on ethnic concepts of appropriate body shape and size by gender, ~age, and socioeconomic status.
• Qualitative research on ethnic definitions of health and approaches to health.
• Qualitative research on ethnic attitudes and approaches to dieting.
• Research on how the types of programs described by Hall and Ammerman can be adapted for use with other ethnic populations, and evaluation of such programs in meeting biomedical and nutritional goals (e.g., reduction of cholesterol levels).
The goal in all this research should be to elucidate categories and concepts of importance to members of the public and to determine which of their traditions should be encouraged. These data are critical in achieving the larger biomedical nutritional goal of a well-nourished and healthy population by using terms that can be understood by the lay public, especially members of minorities.
~Studies on the Relationship Between Energy Consumption and Disease
Data that have been accumulating since the early part of this century indicate that overconsumption of energy may contribute to chronic illness, while restriction of energy may promote health and prolong life. Moreschi (1909) demonstrated that underfeeding could impede the growth of tumors. Rous (1914) confirmed and expanded those findings, but no further progress occurred until 1935 when McKay demonstrated a broad disease preventive effect, as well as extension of lifespan, as a result of caloric restriction (McKay et al., 1935).
In 1940, Albert Tannenbaum demonstrated that energy restriction per se in rodents can inhibit tumor initiation and growth, that increased caloric use stimulated by exogenous thyroid may inhibit certain cancers and metastases, and that fats can promote the growth, in many circumstances, of already initiated tumors (Tannenbaum, 1940, 1942a, 1942b, 1945a, 1945b). Later, Jose reported that Australian Aborigines who became malnourished upon weaning, and who regularly developed a decreased ability to produce antibodies, unexpectedly showed increased proliferative responses of T lymphocytes upon stimulationwith certain phytomitogens (Jose et al., 1969).
As research progressed, it was demonstrated that protein-energy restriction could cause in animals the same enhanced response to phytomitogens seen in Australian Aborigines. In addition, experiments revealed that even at presumably dangerously low protein levels (3 to 5 percent), cell-mediated immunity remained intact and in some cases appeared to be greatly increased, as in the development of cell-mediated responsiveness to stimulation with minute doses of antigen. Additional effects were augmentation of delayed allergic reactions, increased capacity for lymphoid cells to initiate graft-versus-host reactions, up-regulated cellular immune responses against syngeneic and allogeneic tumor cells, and increased capacity to resist certain types of viral infections (Good et al., 1977, 1980).
A direct attempt to follow a high-quality energy-restricted diet as a health measure has been advocated for a number of years by R.L. Walford, M.D. at UCLA (Walford and Crew, 1989). Recently Walford participated in a 2-year experiment in which he monitored the health of eight humans growing and recycling all food in the 3.15-acre hermetically sealed ~experimental ecological enclosure called Biosphere 2 in Oracle, AZ. The eight Biosphere 2 subjects underwent 24 months of moderate caloric restriction (1,700 to 2,400 kcal per day, despite a heavy work load) with a very high quality semi-vegetarian diet in rigidly controlled circumstances. Physiologic changes in the eight volunteers over the 2-year stay were dramatic, with blood glucose dropping 15 percent, cholesterol dropping to an average of 125 mg/dL, and blood pressure dropping to low normal (Walford et al., 1992). Blood white cell counts also decreased, which, along with the decrease in glucose, mirrored changes seen in restricted monkeys and restricted rodents. Further tests are ongoing, but preliminary results thus suggest that humans respond, at least initially, much like all other mammals tested on dietary energy restriction. A great deal more study is needed to see if the immune-enhancing, life-prolonging effect of an energy-restricted diet in lower animals also is manifest in humans.