Vegetables, fruit, and cancer prevention: A reviewAmerican Dietetic Association. Journal of the American Dietetic Association; Chicago; Oct 1996; Steinmetz, Kristi A; Potter, John D.
Volume: 96 Issue: 10 Start Page: 1027 ISSN: 00028223
Historically, consumption of particular fruits and vegetables was thought to prevent or cure ailments ranging from headaches to heart disease. In fact, early medicine revolved largely around the prescription of specific food concoctions (plant foods in particular) for certain disorders 1,2. Until relatively recently, these attributes of vegetables and fruits were based more on beliefs than on scientific evidence, but during the past decades many studies have examined the relationship between vegetables, fruit, and health.
It has been estimated that up to 70% of all cancer is attributed to diet 3. Major dietary hypotheses include a relationship of high fat consumption to breast and colorectal cancers high alcohol intake to respiratory, gastrointestinal, breast, and liver cancers; and low fiber intake to colorectal cancer. Another dietary hypothesis that may be better supported by the scientific literature is that high consumption of vegetables and fruit is protective against cancer of a variety of anatomical sites. Whereas most of the earlier research on diet and cancer involved dietary components thought to increase risk, consumption of vegetables and fruit is thought to be protective.
This article summarizes the vast amount of research about vegetables, fruit, and cancer prevention. Refer to more comprehensive review articles 4,5 and to the original studies for more detailed information.
STUDIES SUGGESTING THAT VEGETABLES AND FRUIT ARE PROTECTIVE AGAINST CANCERCohort Studies
A prospective cohort study is in many ways the ideal method for studying dietary risk factors for diseases. The design is such that a large group of healthy persons (often 10,000 or more) provide information regarding some risk factor of interest (eg, they may fill out a mailed food frequency questionnaire) and are followed up for several years to see which persons contract specific diseases (often follow-up is implemented via cancer registries, death certificates, or follow-up questionnaires). The question of whether specific dietary patterns are related to the risk of getting a specific disease can then be addressed. Although relatively few cohort studies have been conducted because of the length of time and the very large number of participants required, several cohort studies initiated in the past decade are now being used to examine a number of disease end points.
Twenty cohort study analyses 6-28 have examined the relationship between vegetable and fruit consumption and different types of cancer. Of these, 19 studies 6-21,24-28 found an inverse association (protective association; ie, higher intake is associated with lower cancer risk) for at least one category of vegetable and/or fruit, and in 12 studies 6,7,13,15,17-20,24,25,27,28 statistically significant associations were shown. Of the 20 total studies, 3 reported on all cancer sites 7,10-12,20; 4 on lung cancer 6,8,18,24; 3 each on colorectal 9,19,25 and stomach cancers 14,15,21; 2 each on pancreatic 13,22,23, bladder 17,27, and breast cancers 26,28; and 1 on prostatic cancer 16. The cohort study evidence is most consistent for lung cancer; inverse associations for vegetable and/or fruit consumption have been shown in populations of Norwegian men 6, postmenopausal Iowa women 24, Seventh-Day Adventists 18, and members of a California retirement community 20.
At least 174 case-control studies worldwide have examined the relationship between vegetable and fruit consumption and cancer risk. Separate consideration of case-control studies by cancer site is useful when attempting to sort out potential causes and protective factors for cancer at different sites. Different cancers have very different nondietary risk factors and appear to have different dietary risk (and protective) factors as well. Organs and tissues in the body are exposed to harmful and beneficial substances to varying degrees and via various routes (eg, gastrointestinal tract, respiratory tract, blood, skin).
The studies that have investigated different types of cancer are presented in Table 1, along with an indication of the number that have shown a statistically significant inverse association for one or more vegetable and/or fruit categories. Stomach cancer has been studied the most, followed by cancers of the colon, esophagus, lung, oral cavity and pharynx, rectum, and breast; more than 12 studies have focused on each of these sites. Statistically significant inverse associations have been reported for one or more vegetable and/or fruit categories in 80% or more of the studies for cancers of the following sites: stomach, esophagus, lung, oral cavity and pharynx, rectum, bladder, cervix, endometrium, and larynx. Prostate cancer is the only cancer for which the majority of studies (100%) have not reported at least one statistically significant inverse association.
The number of studies reporting inverse, null, or positive associations for different types of vegetables and fruit are tallied in Tables 3 through 11 for cancers of the stomach, colon, esophagus, lung, oral cavity and pharynx, rectum, breast, pancreas, and bladder, respectively. Associations for 10 vegetable and fruit categories were tallied for each cancer site; then those food categories examined in three or more studies were included in the table for that cancer site. The 10 categories initially tallied included vegetables, fruit, raw vegetables, cruciferous vegetables, legumes, allium vegetables, green vegetables, carrots, tomatoes, and citrus fruit (specific food items included in each of these categories are listed in Table 2). The associations were tallied as inverse, null, or positive without regard to level of statistical significance. This type of tally is useful in summarizing the general direction of the results of a large number of studies. If there were no true relationship between vegetable and fruit intake and cancer risk, according to the laws of probability the number of studies finding inverse and positive associations would be evenly distributed regardless of level of statistical significance.
The entries in Tables 3 through 11 show a preponderance of inverse associations for the various vegetable and fruit categories. The evidence is particularly abundant for stomach cancer, the most common cancer worldwide, for which citrus fruit and vegetables, particularly raw and green vegetables, are consistently shown to be protective (Table 3). For colon cancer, the evidence for vegetables, again particularly the raw and green types, is similarly consistent (Table 4). Somewhat surprisingly, the case-control evidence does not support a protective effect of legumes against colon cancer, despite their high content of dietary fiber.
Lung cancer is the most common cause of cancer death in the United States in both men and women, and the case-control evidence for a protective effect of vegetables and fruit could hardly be more consistent (Table 6): 100% of studies show protective associations for vegetables, fruit, green vegetables, and tomatoes. Cigarette smoking is known to be the strongest and most common cause of lung cancer; in the majority of studies the vegetable and fruit associations were mathematically adjusted for the smoking habits of the subjects to reduce the possibility of observing a false protective effect of vegetable and fruit intake as a result of nonsmokers consuming more fruits and vegetables 29. Taken together, the case control and cohort study evidence regarding lung cancer convincingly indicates a protective effect of vegetables and fruit.
For cancers of the esophagus and oral cavity and pharynx, green vegetables and citrus fruit appear consistently to be protective; also, tomatoes appear to be protective against esophageal cancer and carrots against cancers of the oral cavity and pharynx, although conclusions are based on fewer studies (Tables 5 and 7). Most of the studies for these upper gastrointestinal cancers included adjustment for smoking habits and alcohol intake, which are the two major risk factors for these cancers in Western societies.
For rectal cancer, the evidence of a protective effect of cruciferous vegetables is most consistent (Table 8). For breast cancer, which is the most common cancer among US women and the second greatest cause of cancer death, the data generally show an inverse association for consumption of fruit, green vegetables, and carrots (Table 9). Studies of cancers of the pancreas (Table 10) and bladder (Table 11) similarly show predominantly inverse associations and no positive associations; for bladder cancer the evidence for protective effects of green vegetables and carrots is particularly consistent.
To summarize the 174 case-control studies, the evidence is convincing for a protective effect of vegetables and fruit against cancers of the lung, stomach, and esophagus. The evidence for other sites is either less abundant or less consistent, but suggests a probable protective effect against cancers of the oral cavity and pharynx, colon, breast, pancreas, and bladder. The exception is prostate cancer; several studies have been undertaken, yet none has shown a protective association.
Studies of Precursor Conditions
WHICH VEGETABLES AND FRUIT ARE MOST LIKELY TO BE BENEFICIAL?Because we are generally interested in preventing all types of cancer, not just cancer of a particular organ, it is important to evaluate which specific vegetables and fruits have most often been found to be inversely associated with the risk of any type of cancer. In developing dietary recommendations, this information is most useful. To this end, Table 12 displays a comprehensive tally of the 174 case-control and 20 cohort studies reviewed here. Vegetables in general, and raw vegetables in particular, are shown to be the outstanding categories: 85% of studies that looked at raw vegetables reported a protective association. Results for allium vegetables, carrots, green vegetables, cruciferous vegetables, and tomatoes are also fairly consistent; for each of these, 70% or more of studies showed a protective association.
Case-control studies of stomach cancera showing inverse, null, or positive associations for consumption of different types of vegetables and fruit.
Case-control studies of colon cancera showing inverse, null, or positive associations for consumption of different types of vegetables and fruit.
Case-control studies of esophagus cancera showing inverse, null, or positive associations for consumption of different types of vegetables and fruit.
HOW MUCH IS THE RISK OF CANCER REDUCED BY GREATER CONSUMPTION OF VEGETABLES AND FRUIT?Many of the cohort and case-control studies show an approximate halving of risk with high (vs low) intake of vegetables and fruit. A variety of factors inherent to the methodology of nutritional epidemiologic studies (eg, random subject error in reporting on diet, study populations with homogeneous diets) lead to conservative estimates of the risk or protection associated with dietary factors. Thus, it is possible that data consistently suggesting a halving of risk may in fact be signaling a stronger association.
MANY POTENTIALLY ANTICARCINOGENIC SUBSTANCES IN VEGETABLES AND FRUITA myriad of substances in vegetables and fruit have been shown or postulated to have anticarcinogenic properties 5, 64; many of these substances are listed in Figure 1. That there are so many of these substances lends biological plausibility to the findings of the aforementioned human observational studies. Some of the biological mechanisms by which these substances may help prevent cancer are given in Figure 2 and are discussed in relation to specific compounds in the following paragraphs. Figure 3 shows a simplified schematic of the cancer process and denotes points where this process may be interrupted by substances in vegetables and fruit.
Some of the substances listed in Figure I are widespread, whereas others are characteristic of particular classes of vegetables and fruit. For example, cruciferous vegetables, such as broccoli, cauliflower, brussels sprouts, and cabbage, are unique in their high content of dithiolthiones and isothiocyanates; these are organosulfur compounds that have been shown to increase the activity of enzymes involved in the detoxification of carcinogens and other foreign compounds 65-67. One particular isothiocyanate, sulphorophane, appears to be an exceptionally potent inducer of detoxification enzymes 68. Cruciferous vegetables also contain indole-3-carbinol, which has been shown to affect estrogen metabolism in human beings. Specifically, the estradiol hydroxylation pathway may be affected such that more of a less potent form of estradiol is formed, which may protect against estrogen-related cancers such as breast and endometrial cancers 69.
The allium vegetable family includes onions, garlic, scallions, leeks, and chives and is notable for its content of compounds such as diallyl sulfide and allyl methyl trisulfide. Allium compounds have been shown to induce enzymatic detoxification systems 70,71 . Furthermore, the antibacterial activity of these compounds may serve to inhibit the bacterial conversion of nitrate to nitrite in the stomach, thereby reducing the amount of nitrite available for reaction with secondary amines to form nitrosamines, which may be carcinogenic, particularly in the stomach 71.
Soybeans are relatively unique in their content of isoflavones, weak phytoestrogens that compete with more potent estrogens for binding to estrogen receptors 72-74. By this mechanism, isoflavones may inhibit estrogen-promoted breast cancer. Some phytoestrogens are structurally similar to tamoxifen, a drug that is being used to treat some types of breast cancer and is currently being tested for cancer prevention in high-risk women. One particular isoflavone, genistein, has been shown to inhibit the activity of tyrosine kinase, an enzyme involved in the transmission of signals from cellular growth factor receptors that is expressed at high levels in transformed cells 72. Genistein and another isoflavone, biochanin A, have also been shown to induce apoptosis of tumor cells 75. Soybeans further contain protease inhibitors, saponins, phytosterols, and inositol hexaphosphate, each of which has been shown to inhibit carcinogenesis in animals or in vitro (some of these compounds are also found in grains, vegetables, and/or other legumes) 72,73. The anticarcinogenic mechanism of protease inhibitors may be to increase the accuracy of DNA repair, to decrease the rate of cell division 64, or to inhibit proteases produced by tumor cells, thereby thwarting their destruction of the surrounding environment 76. Saponins may exert a preventive effect against colon cancer by binding bile acids and cholesterol in the colon and thereby reducing colonic cell proliferation 72. Additionally, soybeans and other legumes are especially rich in dietary fiber.
Case-control studies of cancers of the oral cavity or pharynxa showing inverse, null, or positive associations for consumption of different types of vegetables and fruit.
Case-control studies of rectal cancera showing inverse, null, or positive associations for consumption of different types of vegetables and fruit.
Citrus fruit is known for its high content of vitamin C, which, as an antioxidant, may protect cell membranes and DNA from oxidative damage. Vitamin C may further help prevent cancer via its ability to scavenge and reduce nitrite, thereby reducing substrate for the formation of nitrosamines 77. Vitamin C also plays a role in the synthesis of connective tissue protein, such as collagen; a deficiency of vitamin C may, therefore, affect the integrity of intercellular matrixes and have a permissive effect on tumor growth or hinder tumor encapsulation 78. Citrus fruit also contains coumarins (also found in some vegetables) and o-limonene (found specifically in citrus fruit oils), which have been shown to increase the activity of glutathione transferase, a detoxification enzyme 79,80.
Green leafy vegetables contain lutein, a carotenoid and xanthophyll pigment that has no vitamin A activity, but that, as an antioxidant, may protect against cancer via its ability to block damage by free radicals 81. Green leafy vegetables are also a rich source of folic acid, a deficiency of which may lead to chromosomal damage at sites thought to be relevant to specific cancers 82. Inadequate intake of folic acid may also lead to reduced methylation of DNA, which may permit a loss of the normal controls on the expression of genes related to cancer 83.
Orange vegetables, such as carrots, sweet potatoes, winter squash, and pumpkin, are relatively rich sources of beta carotene, as are some fruits, including papaya, mango, and cantaloupe. Beta carotene, like other carotenoids, is an antioxidant; through this function it may protect against free radical damage. Beta carotene can also be metabolized to vitamin A (retinol), which plays a role in differentiation of normal epithelial cells. Because lack of differentiation is a feature of cancer cells, adequate intake of vitamin A (from either carotenoids or retinol) may help avoid the development of cancer. Beta carotene may further inhibit cell proliferation 84 and may induce increased cell-to-cell communication, lack of which is a common feature of cancers85. Orange vegetables also contain alpha carotene, which is an antioxidant and a vitamin A precursor and which may inhibit cell proliferation 86. Tomatoes are particularly rich in a red pigment, lycopene, another antioxidant carotenoid 81.
Other potentially anticarcinogenic substances are not limited to one type of vegetable or fruit but are more widespread. For example, selenium is found in produce in amounts proportional to the selenium content of the soil in which it is grown. Selenium functions as a cofactor for glutathione peroxidase, an enzyme that protects against oxidative tissue damage. Selenium may further alter the metabolism of carcinogens via its role in the mixed function oxidase system in the liver 77. Vegetables also contribute vitamin E to the diet (although the major sources are vegetable oils and margarine); vitamin E is an antioxidant that protects polyunsaturated fatty acids in cell membranes from oxidation. Vitamin E further keeps selenium in the reduced state, thus facilitating the antioxidant capacity of selenium. Additionally, vitamin E has been shown to decrease the formation of nitrosamines in the stomach 87,88.
Flavonoids, such as quercetin and kaempferol, are polyphenolic antioxidants that occur in vegetables and fruit (tea and wine are also important sources) 89. In addition to being antioxidants, flavonoids may defend cells against carcinogens via their ability to increase the pump-mediated efflux of certain carcinogens from cells 90 or via induction of detoxification enzymes 64. Furthermore, quercetin may reduce cell proliferation 91.
Fruit, vegetables, and legumes are major sources of dietary fiber, which has been widely hypothesized to be protective against colon cancer 92-94. Dietary fiber may increase fecal bulk and decrease transit time; thereby, via dilution and a shorter period of contact, fiber may reduce the interaction between carcinogens and the intestinal epithelium. Fiber may also bind carcinogens and bile acids. Furthermore, certain types of dietary fiber are fermented by microflora in the colon, which leads to the production of short-chain fatty acids, one of which -butyrate- has been shown to be antineoplastic. The presence of short-chain fatty acids also lowers colonic pH and, subsequently, inhibits the conversion of primary to secondary bile acids, which stimulate colonic cell proliferation and are thought to promote carcinogenesis.
PUBLIC HEALTH IMPORTANCEThe evidence regarding vegetable and fruit consumption and cancer is generally consistent with a protective effect. The relative increase in cancer risk associated with low intake of vegetables and fruit is less than that associated with some cancer risk factors, for example, cigarette smoking or exposure to asbestos; yet the amount of cancer that could be prevented by increasing vegetable and fruit intake in the community could be quite appreciable because low intake is so widespread. Diet is something that can be changed, unlike certain other cancer risk factors, such as genetic factors. Also, on a practical note for nutritionists, it may be easier to convince people to eat more of something, rather than less of something.
Case-control studies of pancreas cancera showing inverse, null, or positive associations for consumption of different types of vegetables and fruit.
Case-control studies of bladder cancera showing inverse, null, or positive associations for consumption of different types of vegetables and fruit.
If increased consumption of vegetables and fruit is to be recommended, any adverse effects imposed by some substances in (or on) vegetables and fruit must be weighed against the probable benefits. Pesticides, aflatoxin, nitrate, and pickled vegetables are topics that have received attention. Some pesticides are known to be toxic. Although the research on the carcinogenicity of pesticides is too vast to be reviewed here, it can be noted that more than 99% of all pesticides in agricultural produce are produced by the plants themselves and that approximately two thirds of synthetic pesticides have been shown not to be carcinogenic in rodent tests 95,96. Pesticide residue on produce can be reduced by thorough washing or, alternatively, produce grown without the use of pesticides can be sought. Aflatoxin is a potent toxin produced by bacteria in crops that are not stored optimally; it may be involved in the etiology of primary liver cancer in developing countries 97. In the United States, the amount of aflatoxin in crops is federally regulated 95. Nitrate, which is on a metabolic pathway for the formation of nitrosamines, is present in large quantities in vegetables; yet the vitamin C, vitamin E, and phenolic compounds in the same vegetables, by reducing gastric pH and inhibiting the conversion to nitrite, probably counteract any risk posed by the presence of nitrate 88,98,99. Pickled vegetables have been shown to be positively associated with cancer risk in some studies in Asia, in contrast to the protective associations found for other fruits and vegetables 97.
Data from a 1991 survey of Americans' vegetable and fruit consumption 100 showed average intake to be 3.4 servings per day. In general, women reported consuming more than men, elderly persons more than young adults, and persons from higher socioeconomic groups more than those from lower groups. Only 23% of Americans reported eating five or more servings of vegetables and fruit per day. In the second National Health and Nutrition Examination Survey (1976-1980)101, the percentages of persons who consumed a cruciferous vegetable or a vegetable or fruit high in vitamin A or C were only 26%, 32%, and 42%, respectively, on a given day.
In 1991, the National Cancer Institute in the United States launched the 5 A Day for Better Health program to encourage Americans to eat five or more servings of vegetables and fruit every day. The program consists of a national media campaign, supermarket displays, promotional support, and community interventions through schools, worksites, and churches 100,102. This program faces the challenge of increasing vegetable and fruit consumption in a population in which, at baseline, only 8% thought they should eat five or more servings per day, and 66% thought two or fewer servings per day to be sufficient. Still, 41% believed that eating vegetables and fruit would help to prevent cancer 100.
In addition to playing a role in cancer prevention, more vegetables and fruit in the diet would likely provide benefits against other adverse health conditions as well. For example, the fiber in vegetables and fruit may help control diabetes and high serum cholesterol level and may prevent diverticulosis. Antioxidants may help prevent cataracts and the oxidation of cholesterol in the arteries. Potassium from vegetables and fruit may help prevent or control hypertension and reduce the subsequent risk of stroke and heart disease. The relatively low fat and energy contents of most vegetables and fruit may reduce the risk of obesity.
Many people may be motivated to eat more vegetables and fruit, but they may need practical advice on how to do so. Some suggestions that may be useful to dietitians in encouraging increased vegetable and fruit consumption are presented in Figure 4.
Table 12 Case-control and cohort studies of all types of cancera showing inverse, null, or positive associations for consumption of different types of vegetables and fruit. Vegetable or fruit category No. of studies % of Total studiesb Inverse Null Positive Inverse Null Positive Vegetables Fruit Raw vegetables Cruciferous vegetables Legumes Allium vegetables Green vegetables Carrots Tomatoes Citrus fruit 55 4 9 29 12 5 33 4 2 38 8 8 14 6 16 27 3 4 61 5 13 50 7 7 35 5 10 26 8 6 81 6 13 63 26 11 85 10 5 70 15 15 39 17 44 79 9 12 77 6 17 78 11 11 70 10 20 65 20 15 a Table summarizes results from 194 studies 6-28, 37, 103-262. Study results tallied include both statistically significant and nonsignificant associations. b Percentages may not add to 100% because of rounding.
CONCLUSIONThe scientific evidence regarding a role for vegetable and fruit consumption in cancer prevention is generally consistent and is supportive of current dietary recommendations. Continued attention to increasing vegetable and fruit intake is important.
Possible anticarcinogenic mechanisms of substances in vegetables and fruit.
K A. Steinmetz is a scientific officer for the World Cancer Research Fund, London, England, and a consulting nutrition epidemiologist in Forchheim, Germany. J D. Potter is a member and program head, Cancer Prevention Research, Fred Hutchinson Cancer Research Center, Seattle, Wash, and a professor of epidemiology, University of Washington, Seattle. Address correspondence to: Kristi A. Steinmetz, PhD, RD, Kersbacher Strasse 33A, 91301 Forchheim, Germany.
Reproduced with permission of the copyright owner. Further reproduction or distribution is prohibited without permission.