Click for next page ( 26


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 25
6 Macro cons tit u ents The literature concerning dietary macroconstituents (i.e., fats, protein, carbohydrates, dietary fiber, alcohol, and total caloric intake) and their associations with carcinogenesis was reviewed extensively in Chapters 4 through 11 of the committee's first report (National Research Council, 1982~. The paragraphs below describe the major gaps in knowledge, which are discussed in more detail in the first report, and provide recommendations for future research in those areas. Food is a complex mixture of chemicals, and the effects of individ- ual dietary components are dependent on many factors, including inter- actions among dietary constituents. However, for the sake of conve- nience, each macroconstituent is discussed separately. FATS After assessing the literature, the committee concluded that there is significant epidemiological evidence for an association between dietary fat and cancer at a number of sites, especially the breast and the large bowel. The evidence associating high fat intake with a high incidence of or mortality from these cancers was derived from both correlation and case-control studies in various populations. In most of the studies, it was not possible to identify clearly which compo- nents of fat were responsible for the observed effects. Where such a distinction was possible, however, total fat and saturated fat were implicated most frequently (National Research Council, 1982, Chapter 5~. The committee also noted that the epidemiological data are not entirely consistent, even though they point in the same direction. For example, the magnitude of the association between fat intake and breast cancer appears greater in the correlation data than in the case-control data, and some studies of large bowel cancer do not demonstrate an association with dietary fat. Possible explanations for these discrepancies were discussed in the first report. The committee was unable to reach a definitive conclusion about the relationship between serum cholesterol and cancer in humans, because the evidence is inconsistent and not sufficiently convincing to establish a causal relationship. Data on dietary cholesterol and cancer ri sk are also too limited to permit any inferences to be drawn (National Re search Council, 1982) . Numerous experiments on animals also indicate that dietary lipid influences tumorigenesis, especially in the breast and the colon. An 25

OCR for page 25
26 DIET, NUTRITION, AND CANCER: DIRECTIONS [OR RESEARCH increase in total dietary fat from 5% to 20% of the weight of the diet (i.e., from approximately 10% to 40% of total calories) appears to increase tumor incidence in each of these tissues. At low intakes of total fat, polyunsaturated fat appears to be more effective than satu- rated fat in enhancing tumorigenesis; however, the effect of polyun- saturated fat becomes less prominent as total dietary fat is increased to 20% of the diet, suggesting that total fat intake is the more significant factor. In general, the epidemiological data and the laboratory evidence are consistent (National Research Council, 1982~. The limited data on the possible mechanisms of action by which fat exerts its effects indicate that ingested fat affects tumor promotion rather than tumor initiation; however, an effect on initiation cannot yet be ruled out. The specific mechanism involved in tumor promotion is not known, although some evidence suggests that colon cancer i s associated with increased concentrations of bile acids in the feces (National Research Council, 1982~. The committee noted that of all the dietary factors that have been associated with cancers of various sites, fat has probably been studied most thoroughly and has produced the greatest frequency of direct associations (National Research Council, 1982~. Nevertheless, it is clear that there is a need for more accurate data on fat intake in specific populations and on its precise effect on tumorigenesis. For example, fat intake varies widely among individuals in the United States. We are not certain that there are differences in cancer incidence between individuals at the high and low ends of the intake spectrum. Unsaturated fat appears to exert a promoting effect on some experi- mentally induced tumors. Diets containing unsaturated fat may be hypo- cholesterolemic, and the results of some population studies suggest that individuals with low levels of serums cholesterol exhibit a greater incidence of tumors. These studies suggest that more data are needed to clarify the effects of unsaturated fat. For example, if unsatu- rated fat acts as a tumor promoter, what is its mechanism of action? Does it involve epoxidation, products of oxidation and degradation such as short-chain aldehydes, or effects on prostaglandin formation? There are many such unanswered questions. Therefore, the mechanists) by which different types of dietary fat exert their effects roust be clari- fied. For example, effects on the immune system require further study. lIeiniger (1981) demonstrated that some inhibitors of cholesterol biosyn- thesis suppress immune response. These finds ngs suggest that the effects of lipids and lipoproteins (dietary fat, lipoprotein, and apolipoprotei n) on the immune system require extensive investigation. Evidence implying that males with blood cholesterol levels less than 200 rng/100 ml are at increased risk of cancer, especially colorectal cancer, was derived largely from follow-up examinations of individual s in studies directed primarily toward cardiovascular dis- ease. Most of these were observational studies in which blood

OCR for page 25
Macroconstituents 27 cholesterol level s were measured and the subsequent mortality from cardiac and other diseases assessed. However, a few studies included deliberate attempts to lower blood cholesterol in subjects whose levels were high. The results from this latter type of study do not clearly indicate whether those with an increased risk of cancer were derived from the groups with initially low blood cholesterol or from groups in which blood cholesterol was lowered as a result of intervention to levels less than 200 mg/100 ml. Several investigators have suggested that the observed increase in mortality from cancer may be due to the presence of undetected cancer at the start of the study rather than to its development after blood cholesterol was measured. Thus, it is not clear whether low blood cholesterol is a consequence or possible cause of cancer (National Research Council, 1982~. If causal, the responsi- ble mechanism could be the excretion of high levels of cholesterol breakdown products in the intestine of persons with low blood choles- terol. This could occur despite the consumption of the standard North American high fat diet, which usually results in blood cholesterol levels that are higher than those in populations consuming diets with lower levels of fat. Metabolic studies are required to determine whether the excretion of cholesterol breakdown products is more active in individuals with low blood cholesterol levels. The effect of a high cholesterol diet on the risk of colon cancer is also unclear. ~ Canadian case-control study of diet and colorectal cancer showed that a much weaker effect resulted from high cholesterol intake than from high saturated fat intake (Jain et al., 1980~. It has been known for some time that a major dietary contributor to in- creased blood cholesterol is not cholesterol but, rather, high levels of fat (National Research Council, 1980b). Thus, further epidemiological and biochemical studies are required to explain the relationship of dietary fat and cholesterol intake to serum or plasma cholesterol levels and to the excretion of cholesterol and its metabolites (i.e., neutral and acidic steroids and their micro- bial by-products). We need to define the level to which the current high fat intake can be lowered to achieve a maximum reduction in the risk of cancer without concomitantly increasing the risk of other disease states. In this regard, the optimum proportions of saturated, monounsaturated, and polyunsaturated fats should be delineated. PROTEIN Epidemiological studies reviewed by the committee indicate possi- ble associations between high levels of dietary protein and increased risk of cancers at a number of different sites, including the breast, colon, pancreas, prostate, and endometrium. However, the litera- ture on protein is much more limited than that on fats. In addition, the high correlation between fat and protein intake in Western diets

OCR for page 25
28 DIET, NUTRITION, AND CANCER: DIRECTIONS FOR RESEARCH and the more consistent and often stronger association of these cancers with fat intake make it seem likely that dietary fat is the more active component. Nevertheless, the committee concluded that the evidence does not completely preclude an independent effect of protein (National Research Council, 1982, Chapter 6~. In laboratory experiments, the relationship between dietary protein and carcinogenesis appears to depend upon the level of protein intake. In most studies, carcinogenesis was found to be suppressed by diets containing levels of protein at or below the minimum required for opti- ~num growth. Chemically induced carcinogenesis is generally enhanced as protein intake is increased up to 2 or 3 times the normal requirement; however, higher levels of protein begin to inhibit carcinogenesis. A review of the preliminary data on possible mechanisms of action suggested that protein may affect both the initiation and the subse- quent growth and development of tumors (National Research Council, 1982). Thus, the association between dietary protein, especially different types of protein, and cancers of the breast, endometrium, prostate, colorectum, pancreas, and kidney needs further clarification. Studies should be specifically designed to determine whether the apparent effect of major dietary sources of protein, which contain a variety of other nutrients and nonnutritive components, is due to a direct asso- ciation of protein with cancer at these sites or reflects the action of another constituent of protein-rich foods. CARBOHYDRATES The committee found only extremely limited and inconclusive epidem- iological evidence concerning the role of carbohydrates (exclusive of dietary fiber) in the development of cancer in humans. The data from the few laboratory experiments on this subject could not be interpreted because of generally poor experimental designs and uncertainty about the actual carbohydrate content of the test diets (National Research Council, 1982, Chapter 7~. Thus, in contrast to lipids and protein, very little work has been d irected toward the study of carbohydrate intake, e specially the level s of dif ferent types of carbohydrate, and the occurrence of cancer. How- ever, one study has suggested that rats fed sucrose are more suscepti- ble to chemically induced tumors than are rats fed starch (Hoehn and Carroll, 197 9) . Thi s area requires further investigation. A complete understanding of the effects of carbohydrates on carcinogenesis may depend on a thorough knowledge of the effect s and interactions of each dietary component. DIETARY F IBER The association between dietary f iber and carcinogenesis, espe- cially in the colon, has been investigated at length in epidemiological

OCR for page 25
Macroconstituents 29 studies. However, both correlation and case-control studies have yielded inconsistent results. Many of these studies were based on total fiber consumption estimated by grouping foods (such as fruits, vegetable s, and cereals) according to their f iber content . However, in the only case-control study and the only correlation study in which the total fiber consumption was quantified rather than estimated from the fiber-rich foods in the diet, no association was found between total fiber intake and the risk of colon cancer. Thus, the committee concluded that the epidemiological evidence suggesting an inverse relationship between total fiber intake and the occurrence of colon - cancer is not compelling (National Research Council, 1982~. In the only study in which the effects of individual components of fiber were assessed, there was an inverse correlation between the in- cidence of colon cancer and the consumption of the pentose-containing fraction of fiber. Thus, it seems likely that further epidemiological study of fiber will be productive only if the relationship of cancer to specific components of fiber can be analyzed (National Resea rch Council, 1982, Chapter 8) . A few laboratory studies have also shown that some types of fiber (e.g., cellulose and bran) inhibit chemically induced tumorigenesis in the bowel. However, the data are somewhat inconsistent with respect to the type of fiber or specific chemical carcinogen. Moreover, the results of epidemiological and laboratory studies are difficult to equate, because most laboratory experiments have examined fiber-rich materials or their individual components, whereas most epide~iological studies have focused on fiber-containing foods whose exact composition has not been determined. Therefore, to obtain meaningful results, further information is needed on the basic chemistry and biological effects of fiber and its components (National Research Council, 1982~. Recently, more attention has been directed toward the physiological significance of dietary fiber, which generally includes indigestible carbohydrates and carbohydrate-like components of food such as cellu- lose, lignin, hemicelluloses, pentosans, gums, and pectins. Neverthe- less, because of the complex composition of dietary fiber, the physio- logical functions and metabolic activity of its individual components have not yet been studied sufficiently. Although epidemiological data concerning the role of total dietary fiber in the development of colon cancer are somewhat inconsistent, studies in animals have demonstrated that individual components of fiber, e.g., bran and cellulose, exert protective effects against the induction of cancer by chemicals, but that another type, agar, does not (Barbolt and Abraham, 1978; Freeman et al., 1978; Glauert et al., 1981~. The effects of fiber should be correlated with its structural properties. It would be useful to compare the effects of adding fiber- containing foods to the diet with the effects resulting from the addi- tion of specific dietary fibers, since current hypotheses concerning the

OCR for page 25
DIET, NUTRITION, AND CANCER: DIRECTIONS FOR RESEARCH effects of fiber are based on dietary patterns--not on the addition of specific fibers or fiber-containing foods. Studies are also required to separate the beneficial effects of high intakes of certain cruciferous vegetables on colon cancer from the effects, if any, of dietary fiber per se. There is little information concerning pathophysiological changes In the bowel that are associated with different types of high-fiber diets. The relationship of various types of fiber to the enterohepatic circulation of sterols, including sterol-derived hormone s, i s al so not understood . ALCOHOL There have been many studies concerning the effects of alcohol consumption on cancer incidence in human populations. In some countries, including the United States, excessive beer drinking has been associated with an increased risk of colorectal cancer, espe- c tally rectal cancer. Although it i s recognized that excessive alcohol consumption contributes to hepatic injury and cirrhosis, there is only limited evidence that this in turn leads to hepatocellular car- cinoma. Furthermore, excessive consumption of alcoholic beverages and cigarette smoking appear to act synergistically to increase the risk for cancer of the mouth, larynx, esophagus, and the respiratory tract (National Research Council, 1982, Chapter 11~. Although some reports have suggested that alcohol consumption per se is related to cancer, others have implicated specific alcoholic beverages as risk factors for cancers at certain sites, such as the esophagus and gastrointestinal tract. This suggests that nonalcoholic components of alcoholic beverages may be the responsible agents. Furthermore, it is not yet clear what role nutrient inadequacies, imposed by excessive alcohol consumption, play in the process of car- cinogenesis. TOTAL CALORIC INTAKE The committee found it especially difficult to separate the effect of caloric intake per se on carcinogenesis from the effects due to changes in the levels of the three macronutrients: fat, protein, and carbohydrates. It concluded that the epidemiological evidence support- ing total caloric intake as a risk factor for cancer is slight and largely indirect, because much of it is based on associations between body weight or obesity and cancer rather than on direct measurements of caloric intake. Studies that have evaluated both the caloric content of the diet and the intake of fat suggest that dietary fat is the more relevant variable (National Research Council, 1982, Chapter 4~. Similarly, studies in animals to examine the effect of caloric intake on carcinogenesis have been few and are difficult to interpret.

OCR for page 25
Macroconstituents 31 In these experiments, animals on calorie-restricted diets developed fewer tumors and their lifespan far exceeded that of animals fed ad libitum, thereby indicating a decrease in the age-specific incidence of tumors. For example, McCay et al. (1943) and Ross and Bras (1973) showed that underfed rats lived longer and developed fewer tumors than their littermates, which were fed ad libitum. In another experiment, Lavik and Baumann (1943) found that methylcholanthrene-treated rats on restricted caloric intake developed fewer tumors than did treated rats fed ad libitum. However, because the intake of all nutrients was simultaneously depressed in these studies, the observed reduction in tumor incidence or delayed onset of tumors might have been due to the reduction of other nutrients such as fat. It is also difficult to interpret experiments in which caloric intake has been modified by varying dietary fat or fiber, both of which may by themselves exert effects on tumorigenesis (National Research Council, 1982~. Neither the epidemiological nor the experimental studies permit a clear interpretation of the specific effect of caloric intake. How- ever, the effects of over- and underfeeding as such warrant further study, even though it may be difficult to separate caloric effects fray the effects of the specific nutrients that contribute to total caloric intake. RESEARCH RECOMMENDATIONS A better understanding of the relationship between certain macro- constituents (e.g., fat) and cancer has enabled the committee to make some more specific recommendations in this chapter than in other chap- ters. General Recommendations for Epidemiological Research ~ There is probably considerable interaction among the many compo- nents of the diet. Therefore, some potentially harmful substances may be "neutralized" by other dietary ingredients. For this reason, there is a need to evaluate the interrelationships among calories, protein, and fat (and its various components) and their effect on, for example, breast and colorectal cancer. 0 Simultaneously, the opportunity should be taken to evaluate the interrelationships between these nutrients and (1) the effects of hormonal status on breast cancer and (2) the effects of fiber, its components, and various micronutrients, especially vitamins and possible inhibitors in vegetables of the genus Brassica, on the eti- ology of colorectal cancer. o Research on interrelationships among macroconstituents should be designed so that it is possible to determine the overall effect of different groups of foods and not just individual foods or nutrients, as discussed in Chapter 4.

OCR for page 25
32 DIET, NUTRITION, AND CANCER: DIRECTIONS FOR RESEARCH The completed dietary studies on breast and colon cancer should be extended to examine other possibly diet-associated cancers that have been correlated with breast and colon cancer. In addition to the on- going investigations of prostate cancer, studies should be conducted on endometrial, ovarian, pancreatic, and renal cancers. More frequent monitoring of food intake, especially changes in intake of macronutrients in the average diet, is essential (see Chapter 4~. Reliable data bases for food composition should be developed for the analysis of macronutrients (see Chapter 4~. It is especially important that such data bases contain more information on the fiber content of each food and on the chemical composition of each type of fiber. It is now essential to give high priority to some long-term cohort studies that will test hypotheses about macronutrients and cancer (see Chapter 4~. Carefully planned intervention studies, involving changes in the macronutrient content of the diet, should be conducted in humans (see discussion of intervention studies in Chapter 4~. Such studies may be the only way to gain an understanding of the relative effects and the interrelationships among,Luacronutrients. In such studies, it may be conceptually easier to plan for the addition of constituents, e.g., specific types of fiber, to the diet. However, we should not overlook the need to evaluate the effect of reducing dietary fat by consumption of foods low in fat. The effects of dietary macroconstituents on the later stages of carcinogenesis need to be examined in laboratory studies. General Recommendations for Laboratory Research . o For animal experiments, the first priority is standardization of methodology. This is discussed in Chapter 5. The epidemiological data linking specific dietary components to cancers of the prostate, pancreas, and endometrium are limited. An expansion of the experimental data base, i.e., development of suitable animal models, is required in order to put these data into proper perspective. Specific Recommendations Fat: Epidemiological Studies . o More discriminating data are needed on the effect of the level and type of fat intake by humans. For example, we need to answer the

OCR for page 25
Macroconstituents 33 question, "What level of fat intake is associated with the maximum reduction in cancer incidence?" Is it 307 of calories, as recently recommended by this committee? Is it 25%, 20X, or lower levels? O More discriminating data are also needed on the effects of different types of fat. Studies should be conducted to answer the following questions: Is the finding that polyunsaturated fat increases tumor incidence in laboratory animals relevant to humans? Is this finding also relevant to the results emerging from intervention trials for cardiovascular disease? What should be the relative proportion of polyunsaturated, monounsaturated, and saturated fats in the optimal diet? Metabolic studies are required to evaluate the role of the break- down products of cholesterol in individuals with low blood cholesterol levels. o The mechanism underlying the reported association between low blood cholesterol and neoplasia should be determined. For example, it would be helpful to know whether hypocholesterolemic individuals are at high risk only if they consume a high fat diet and whether lowering fat intake in such individuals will reduce their risk. Further analysis of existing data on humans may help to answer these questions. Fat: Laboratory Studies ~ The relative roles of the level and type of fat (e.g., essential fatty acids ~ in al 1 phases of tumor formation should be studied. The stage at which dietary fat exerts its effects on the induction of tumorigenesis and its effects in the prepromotional, promotional, tumor development, and metastatic stages should be systematically investi- gated. Other aspects of lipid nutriture and carcinogenesis also need resolution. For example, further investigation is needed to determine the effects of bans unsaturated fats, lipid pyrolysis products, and lipid peroxides and to study the effects of lipids on membrane phe- nomena, on prostaglandin synthesis, and on immune phenomena. o The interplay among dietary lipids (including cholesterol), hypo- cholesterolemic agents, neutral and acidic fecal steroids, and gut microflora should be clarified with respect to their effects on tumori- genesis. o Studies should be conducted in animals to determine the threshold level (percent of calories) at which dietary fat begins to exert measurable effects on carcinogenesis. O The cholesterol vehicle (lipoprotein) should be examined for its ability to carry other substances that may affect tumor growth. Protein: Epidemiological Studies ~ The independent effects of the amount and type of protein and of their interaction with other macronutrients on the incidence of tumors in humans should be investigated.

OCR for page 25
34 DIET, NUTRITION, AND CANCER: DIRECTIONS FOR RESEARCH Protein: Laboratory Studies . The inf luence of the biological value of protein and the level and type of protein should be studied in different experimental systems. The effects of animal protein and vegetable protein should be compared. The effect of protein on different stages of carcinogenesis and the mechanism underlying this effect need further investigation. Carbohydrates: Laboratory Studies Simple sugars (mono- and disaccharides) should be compared with starches from various sources for their effects on tumor formation. Calorie s: Epiderniological S tudie s o The relationship between total caloric intake and cancer in humans should be studied further, together with the modifying inf lu- ence of energy expenditure through occupation or exerci se . Calo rie s: La boratory S tudie s The mechanism for the putative effect of caloric intake on car- cinogene si s need s to be determined . For example, to what extent are the calories provided by fat, protein, and carbohydrate responsible for the effect on carcinogenesi s? And i s the effect of the caloric contribution by each of the three macronutrients equivalent? Effects of age at which caloric restriction is instituted should be assessed. Carbohydrate i s usually used to replace fat or protein in the diet without considering how this may influence the outcome. However, the effect of replacing dietary fat with protein or carbohydrate in iso- caloric diet s has not been examined . Although thi s may be difficult to accomplish, it is important to attempt to study the individual effects of protein, carbohydrate, and fat in animals fed ad libitum. Fiber: Epidemiological Studies The content of f iber components in foods should be determined In order to assess dietary intake more accurately. O The inf luence of f iber on cancers other than colorectal cancer should be studied. Fi her: Laboratory S tudies The structure-function relationships of fiber (e.g., its pentose content and it s bile-acid binding capacity) in tumor formation should be studied.

OCR for page 25
Macroconstituents 35 The physiological properties of different components of fiber and their effects on the absorption and availability of nutrients should be systematically evaluated in metabolic studies. ~ The influence of fiber on tumors at sites other than the colon should be investigated. Alcohol: Epidemiological Studies Reliable methods are needed to quantitate alcohol intake. This should be followed by investigation of the effects of different intake levels on cancer risk. The influence of different alcoholic beverages (e.g., wine, beer, whisky, or liqueurs) on esophageal, gastric, and other cancers should be studied. Alcohol: Laboratory Studies Studies should be conducted to determine the influence of nonalcoholic components of alcoholic beverages on experimentally in- duced carcinogenesis. The association between carcinogenesis and nutrient deficiencies imposed by excessive alcohol intake should be evaluated.