| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
DIETARY POLYPHENOLS AND HEALTH: PROCEEDINGS OF THE 1ST INTERNATIONAL CONFERENCE ON POLYPHENOLS AND HEALTH |
1 From U557 INSERM (UMR INSERM/INRA/CNAM) and Unité de Surveillance et d'Epidémiologie Nutritionnelle, InVS/CNAM, Institut Scientifique et Technique de la Nutrition et de l'Alimentation/CNAM, Paris
2 Presented at the 1st International Conference on Polyphenols and Health, held in Vichy, France, November 18–21, 2004. 3 Address reprint requests to S Hercberg, U557 INSERM, CNAM, 5 rue Vertbois, F-75003 Paris. E-mail: hercberg{at}cnam.f.
ABSTRACT
An important question being raised by nutritionists today is whether available scientific data support an important role for polyphenols in the prevention of pathologic conditions that represent an important public health burden, such as cardiovascular diseases, cancers, and osteoporosis. More broadly, when can we consider scientific knowledge sufficient to allow specific public health implications and recommendations? The history of the relationship between ß-carotene and cancer illustrates the complexity of the research process leading to the demonstration of a causal relationship between nutritional factors and the prevention of disease. The ß-carotene story, which has developed in the past 30 y, is particularly significant and illustrative because of apparent controversies that are far from resolved. This is an extremely interesting example from which many lessons can be learned. For ß-carotene, we need to collect sufficient information from experimental, clinical, and epidemiologic research before we support any specific public health recommendations. The same principles must be applied to recommendations regarding polyphenols (in particular, which polyphenols, at which doses, to achieve which benefits for which populations). If these questions are not answered, then we run the risk of needing to renounce recommendations regarding polyphenols in the future, damaging the credibility of nutritional recommendations for public health.
Key Words: ß-Carotene • cancers • recommendations • intervention trial • polyphenols
INTRODUCTION
Although their existence has been known for many years, the interest in polyphenols has increased sharply in the past 2 decades. This "boom" is reflected by the major increase in the number of scientific articles published in the past 20 y; when performing a MEDLINE search with the keyword polyphenols, we found that 7 articles had been published in 1981, 15 in 1990, 53 in 1995, 125 in 1998, and 437 in 2002. In particular, the number of articles on the potential benefits of polyphenols in preventing chronic diseases has increased. Industry has not been oblivious to this accumulation of scientific data, and the result is that today we can buy "over-the-counter" supplements containing polyphenols and products fortified with polyphenols; indeed, the food industry would like to make claims regarding the supposed health benefits of polyphenols in their products.
Currently, however, the main question is whether available scientific data justify an important role for polyphenols in disease prevention through nutrition, particularly with respect to pathologic conditions with an important public health burden, such as cardiovascular diseases, cancers, and osteoporosis. More generally, when can we consider scientific knowledge sufficient to allow specific public health applications and recommendations?
In this respect, the history of the relationship between ß-carotene and cancer illustrates the complexity of the research process leading to the demonstration of causal relationships between nutritional factors and disease prevention. This is an extremely interesting example from which many lessons can be learned, as suggested by several authors (1–4).
The ß-carotene story that developed in the past 30 y is particularly illustrative of the intricacies of nutritional research because of apparent controversies that are far from resolved. The story has its ups and downs and could be titled "The rise and fall of ß-carotene," "ß-Carotene and health: the search for the holy grail," "Song of the sirens," "Dr. Jekyll and Mr. Hyde," or "The never-ending story."
1970–1980: THE RISE OF ß-CAROTENE
Background
The story began 
30 y ago, between 1970 and 1980, with the conjunction of mechanistic theories regarding the role of free radicals considered capable of inducing numerous pathologic processes (especially carcinogenesis and atherosclerosis) and the interest in nutrients such as ß-carotene that, because of their antioxidant capacity, could prevent such harmful effects. Evidence from experiments with animals indicating that ß-carotene could have anticancer activity began to accumulate, and data from observational epidemiologic studies of cancer and vegetable, fruit, and ß-carotene intake became available.
A catalyst was necessary to make sense of all of these elements and to provoke a major interest in ß-carotene. This came in the form of an article by Richard Peto and coworkers (5) in the March 1981 issue of Nature, "Can dietary ß-carotene materially reduce human cancer rates?"
The role of free radicals in carcinogenesis
In the 1970s, the indispensable but potentially dangerous role of oxygen had been known for more than a century (6). Jules Verne, the French writer and visionary, wrote in 1865, in his novel "From the earth to the moon," that "oxygen ... this gas without a smell, eminently vital, can cause the most serious disorders in the organism."
Oxygen toxicity in laboratory animals was first described in 1878 (7) and was established in 1899 (8). The first experiments investigating a free radical reaction were reported in 1894 by Fenton (9). However, it was not until the late 1940s to early 1950s that retrolental fibroplasia among premature newborns was recognized as being attributable to oxygen toxicity (10), and it was not until the late 1960s to early 1970s that newborn bronchopulmonary dysplasia and adult respiratory distress syndrome were appreciated by the medical community (11). Moreover, the presence of free radicals in biologic systems was not generally considered likely until the discovery of superoxide dismutase in 1969 (12), although the basis of oxygen toxicity and X-irradiation was proposed to be a common free radical mechanism and the radical theory of aging was hypothesized in the 1950s (13).
It was at a meeting of the World Health Organization in the mid-1960s that a Russian toxicologist, Professor Sanojki, referred to various degenerative diseases as being "rusting diseases" and linked their cause to free radicals (4). Several studies performed between 1972 and 1980 suggested that free radical production directly and indirectly played a major role in carcinogenesis.
Animal studies supporting the anticancer activity of ß-carotene
After the first reports of antitumor effects of vitamin A in animal models, which appeared in the 1960s, investigators were curious to know whether ß-carotene, the precursor of vitamin A, had antitumor effects. In 1973, Dorogokupla (14) showed that subcutaneous tumors induced with injection and topical application of dimethylbenzanthracene developed at a slower rate among rodents fed a diet supplemented with "unlimited amounts of red carrots," compared with animals fed the unsupplemented diet. In 1977, Epstein (15) reported that tumors induced with ultraviolet radiation appeared at a slower rate among hairless mice given injections of a ß-carotene solution, compared with mice given injections of a placebo solution. More and more experiments suggested that carotenes, irrespective of their vitamin A activity, were able to prevent or slow the growth of tumors induced with ultraviolet light or dimethylbenzanthracene. From a mechanistic viewpoint, in vitro and in vivo studies showed that carotenoid pigments were able to scavenge highly reactive species such as oxygen and other free radicals and thereby prevent harmful effects of these species (16).
Epidemiologic data on ß-carotene intake and the risk of cancer
In the early 1980s, there was already substantial epidemiologic evidence indicating associations between high fruit and vegetable consumption, high estimated ß-carotene intake in the diet and/or high blood concentrations of ß-carotene, and lower incidences of cancers, particularly lung cancer (16). The first epidemiologic studies investigated the relationship between ß-carotene and cancer almost by accident, because dietary questionnaires designed for more general purposes happened to ask about the main local ß-carotene source and it was realized later that consumers of those particular foods had lower cancer risks.
Five prospective and 15 retrospective case-control, questionnaire-based studies of populations in 8 different countries were published at the end of the 1970s (5). A relative risk of cancer (lung, stomach, colorectal, or other) of 1.5–2.0 was typically observed in comparisons of groups with low and high quartiles or quintiles of ß-carotene intake, as estimated with dietary questionnaires.
Using data from experimental studies, animal studies, and observational epidemiologic studies, Peto et al (5) highlighted in 1981 the potential public health significance of ß-carotene and the need for future controlled trials. In their article, Peto et al (5) concluded, "it is most unlikely that this inverse cancer risk association will disappear entirely with future observational studies, but the inverse association may be an artifact, due merely to association of ß-carotene ingestion with some truly protective dietary habit(s) or component(s) or avoidance of some truly harmful habits or components." Randomized, population-based trials are direct tests of hypothesis-generating consistent associations noted in observational studies.
1980–1990: THE SEARCH FOR THE HOLY GRAIL CONTINUES
During the 1980s, much experimental evidence accumulated in favor of cancer-preventing effects of dietary ß-carotene and of ß-carotene administered through injection or topical application. In National Cancer Institute preclinical studies, ß-carotene and other carotenoids inhibited tumorigenesis in head/neck, mammary, colon, bladder, and skin cancer models in vivo and in mammary organ cultures in vitro.
During this period, overwhelming observational evidence provided support for the association between greater consumption of carotenoid-rich foods (specifically, higher ß-carotene intake) and lower cancer risk. Finally, > 125 case-control or cohort studies relevant to the association between ß-carotene (assessed with dietary intake or biochemical measurements) and cancer were conducted between 1980 and 1990, with various measures in diverse populations. Most of the studies consistently showed protective effects (1). Relative risk increases of 50-150% (ie, relative risks of 1.5–2.5) were typically reported for the lowest vegetable/fruit or ß-carotene intake categories, compared with the highest. These observed associations were relatively strong and occurred in studies of men and women (17), different racial groups (18), and current smokers, former smokers, and nonsmokers (19) and therefore appeared to be quite generalized (1). These results could have substantial potential public health implications. Taken together, at the end of the 1980s, the investigation of carotene-rich vegetables, ß-carotene intake, and serum or plasma ß-carotene concentrations in relation to cancer (especially lung cancer) provided the most persuasive evidence available in the diet-cancer epidemiologic literature for a protective association, in terms of both magnitude and consistency (1, 3).
1990–2000: THE SONG OF THE SIRENS AND THE BEGINNING OF THE FALL
At the beginning of the 1990s, only data from observational research and experimental work were available for formulation of nutritional recommendations regarding ß-carotene intake. However, on the basis of the criteria of causality, such as 1) consistency, 2) strength of the association, 3) dose-response gradient, and 4) biologic plausibility, the evidence could have been considered sufficient to support additional recommendations and specific claims regarding either carotenoid-rich foods or ß-carotene intake in particular. It is certain that the then-available data contributed to the increasing popularity of vitamin supplements during the 1980s.
In the 1990s, many impatiently but confidently awaited the results of intervention trials that had begun in the 1980s. Well-designed, randomized intervention trials are considered to provide highly relevant, specific, convincing evidence of causality between nutrient intake and pathologic conditions (for instance, cancer risk) and thus play an important role in the development of nutrition-related recommendations. Randomized trials avoid most of the biases inherent in observational studies. Randomized intervention trials addressing the role of supplementation with antioxidant micronutrients in cancer prevention were started in the 1980s, and some were completed by the mid-1990s. The results of 5 large intervention trials were published between 1993 and 2000.
The Nutritional Intervention Trials in Linxian (20) were developed in a general Chinese population. In that series of studies involving a population of 29 584 subjects (mostly nonsmokers), the use of daily doses of 15 mg ß-carotene, 30 mg vitamin E, and 50 µg selenium during a > 5-y period significantly reduced the overall mortality rate by 9%, the rate of death resulting from all cancers by 13%, and the rate of death resulting from gastric cancer by 21%. Because of the nature of the combination micronutrient supplement, intervention effects from this trial could not be attributed with certainty to any one of the 3 nutrients. These results were consistent with those of observational studies of the protective effects of antioxidant nutrients, especially ß-carotene.
A few months after the publication of that trial, other trials did not confirm those encouraging results. However, it is important to bear in mind that the Chinese population of the Linxian area is known to have high incidences of gastric and esophageal cancers and high frequencies of nutritional deficiencies.
The Alpha-Tocopherol and Beta-Carotene (ATBC) Cancer Prevention Study began in Finland in 1985 (21). The subjects were 29 133 male heavy smokers, 50–59 y of age, enrolled in a randomized, double-blind trial. Supplementation consisted of 20 mg ß-carotene/d or 50 mg 
-tocopherol/d, in a 2 x 2 factorial design. The results for ß-carotene were surprising. No benefits with respect to the prevention of lung cancer were seen. Instead, a significant 16% increase in the incidence of lung cancer was noted among those receiving the ß-carotene supplements, suggesting an adverse effect. This 16% increase (95% CI: 2–33%) in lung cancer incidence was clearly inconsistent with the 2-fold risk reduction attributed to high ß-carotene intake in numerous observational studies, and it essentially ruled out a primary preventive effect on lung cancer among smokers subjected to a regimen of daily intake of a 20-mg ß-carotene supplement for 5–8 y. The significance of this unexpected finding was increased by the fact that the ß-carotene group also experienced an 8% increased overall mortality rate, including an apparent increase in the rate of death attributable to ischemic heart disease.
The interpretation of 2 decades of research on the relationship between ß-carotene from vegetables and fruit and lung cancer was suddenly called into question by these findings. Thus began the "fall of ß-carotene."
A few months later, similar results were reported from the Carotene and Retinol Efficacy Trial (CARET). This multicenter trial, which included 18 314 heavy smokers and professionally exposed subjects (22), was terminated 21 mo earlier than planned. After 4 y, increases in the lung cancer incidence and total mortality rate were observed in the group receiving supplementation with 30 mg ß-carotene/d and 25 000 IU (13 664 retinol equivalents) retinyl palmitate/d. A total of 388 subjects developed lung cancer, ie, a 28% increase in lung cancer incidence among participants who received the ß-carotene/retinyl palmitate combination daily for an average of 4 y, compared with participants who received a placebo.
Until these CARET results were announced and published in early 1996, the ATBC study findings were viewed cautiously and, by some, with skepticism. Thereafter, the concordant data from the ATBC study and the CARET created a striking apparent contradiction with previous results from observational epidemiologic studies.
In contrast to findings from the ATBC study and the CARET, the Physicians' Health Study (23), which began in 1982 and involved 22 071 US male subjects (primarily nonsmoking physicians), 40–84 y of age, showed no effective difference in lung cancer incidences between the ß-carotene group (50 mg on alternate days) and the placebo group after 12 y of supplementation. This was based on only 66 and 71 lung cancer cases in the ß-carotene and placebo groups, respectively, and represented a nonsignificant 7% reduction. No adverse or beneficial effects were observed in the 11% of smokers in the ß-carotene group. In the Women's Health Study (24) as well, no effect of vitamin E (600 IU) or ß-carotene (50 mg on alternate days) supplementation was observed among 39 876 apparently healthy, female, health care professionals, 
45 y of age, after a treatment duration of 2.1 y and a median total follow-up period of 4.1 y.
The results of trials published in the 1990s did not support an association or role for supplemental ß-carotene, at the doses and durations of supplementation tested, in cancer prevention in populations. These results were seemingly at odds with those from observational epidemiologic studies. These trial results make it highly unlikely that pharmacologic doses of supplemental ß-carotene are beneficial for prevention of most lung cancers, and they provide strong evidence for adverse effects (eg, increased tumor promotion or progression) among smokers.
These studies raised the issue of interpretation of cohort and case-control studies (eg, is it the ß-carotene in the diet?) and reopened the question of the safety of ß-carotene supplements, which had long been considered resolved. The results require that caution be exercised in recommending supplemental ß-carotene and argue against changing dietary recommendations from foods to nutrient specificity at this time.
After publication of the negative effects of 2 of these trials, there was a total change in the view of the role of ß-carotene. Some researchers in the scientific community called for an end to additional chemoprevention studies involving ß-carotene and expressed concern about the dangers and liabilities of its unregulated use (3, 25). Results from the ATBC study and the CARET led the Food and Drug Administration to disallow health claims for ß-carotene related to cancer prevention, under the Dietary Supplement Health and Education Act of 1994 (26), in part because no recognized scientific body (such as the National Institutes of Health or the National Academy of Sciences) had provided "authoritative statements" supporting such claims.
Therefore, the hopes raised in the 1980s concerning the protective effect of ß-carotene totally disappeared at the end of the 1990s; it was as if Dr Jekyll became Mr Hyde. ß-Carotene not only was considered not to be protective but also was potentially deleterious, with an increased risk of lung cancer among smokers. Nevertheless, the ß-carotene experience seems to have had minimal impact on the dietary supplement industry, which has seen its sales double in the past 5 y to > $13 billion from > 20 000 products (27).
2004: THE NEVER-ENDING STORY
In this context, a new revival was recently provoked by the presentation of preliminary results of the Supplémentation en Vitamines et Minéraux Antioxydants (SU.VI.MAX) study (28), a randomized, double-blind, placebo-controlled, primary prevention trial among 13 017 middle-aged subjects who received a combination of 120 mg vitamin C, 30 mg vitamin E, 6 mg ß-carotene, 100 µg selenium, and 20 mg zinc, compared with placebo. After a mean follow-up period of 7.5 y, 88 men in the intervention group developed cancers, compared with 124 men in the placebo group, which yields a relative risk of 0.69. This difference was reflected in higher incidences of digestive, respiratory, and skin cancers in the placebo group. Among women, no statistically significant difference in the occurrence of cancer was observed between the groups. A total of 174 deaths occurred during the trial, 76 in the intervention group and 98 in the placebo group. No overall significant difference in rates for any cause of death was observed between the groups. Again, a sex-group interaction was observed, showing a significant protective effect (relative risk: 0.63) of antioxidants among men and no effect (relative risk: 1.03) among women.
Therefore, a combination of antioxidants including ß-carotene, at doses achievable through the diet, may have protective effects on mortality rates and on the total number of cancers among apparently healthy men, with no evident increase in cancer risk. These doses correspond to those in a healthy diet with a high consumption of fruits and vegetables, confirming the results of prospective observational studies.
Baseline ß-carotene and vitamin C status among men in the SU.VI.MAX study was lower than that among women. The ineffectiveness of supplementation among women might have been attributable to their better baseline antioxidant status. The improved antioxidant nutritional status achieved with supplementation among men, which proved protective against cancer, supports the results of other prospective observational studies of the benefits of consuming large amounts of fruits and vegetables. This reinforces the general recommendations of a lifelong diversified diet rich in foods that are good sources of antioxidant nutrients.
The SU.VI.MAX. study has not led to a total "rehabilitation" of ß-carotene as a nutrient capable of reducing cancer incidence. The design of the study does not enable us to confirm a specific protective role of ß-carotene, but it does raise some new issues. The apparent discrepancies between observational and randomized studies and among the randomized trials could be explained by the choice of study population (general or high-risk subjects, well-nourished or deficient, with adequate or insufficient antioxidant status), the different doses of supplementation (nutritional or higher), the number of antioxidants tested (1, 2, or more), and the type of administration (alone or in balanced association).
The different studies available, and especially the intervention trials, suggest that the effects of ß-carotene differ according to the dose and the type of population studied. On the basis of available evidence, it currently seems wise to advocate a diet rich in fruits and vegetables, rather than consumption of specific ß-carotene supplements, to decrease cancer risk. Taking into consideration the possible prooxidant role of high doses of ß-carotene (29), it is recommended that long-term use of such supplements at high doses be limited, especially among subjects with high risks of cancer, such as smokers.
WHAT LESSONS FOR POLYPHENOLS?
Returning to research on polyphenols, we should bear in mind that, compared with ß-carotene, research on the health effects of polyphenols started more recently and fewer data are available. This relatively late interest is explained largely by the complexity of their chemical structures, the numerous different phenolic compounds and potentially active metabolites that exist, and the scarcity of intake data.
As was the case for ß-carotene, experimental studies with animals and cultured human cell lines tend to support a role for polyphenols in the prevention of cardiovascular diseases, cancers, neurodegenerative diseases, diabetes mellitus, and osteoporosis. We must be cautious, however, because most of those studies were conducted at doses or concentrations that probably do not occur in the human body under normal conditions. Clinical studies have led to contradictory results. Convergent data from epidemiologic studies showed an inverse relationship between the consumption of foods rich in polyphenols and the risk of myocardial infarction or cancer; however, those were observational studies. More human studies are needed to provide clear evidence of protective effects on health and to allow better evaluation of the potential risks of too-high polyphenol intake.
First, reliable food composition data are needed for estimation of the consumption of polyphenols, for study of their association with disease risk in an epidemiologic setting. This is a difficult task, because polyphenols in foods are not easy to measure, measurements are expensive, and the dataset will be large if all different compounds are included. Second, valid biomarkers should be identified, if possible, to provide a more independent estimate of the intake and metabolism of polyphenols. Some work on food composition tables and biomarkers is currently underway, as noted during this conference, but not enough well-designed, epidemiologic, observational studies have yet been performed. Much work on the bioavailability of polyphenols is being performed, which is an important step in the possible "rise of polyphenols." The results of these studies should lead to additional investigations, because epidemiologic studies should be focused on polyphenols that are effectively bioavailable. If observational studies indicate possible beneficial effects of all or specific polyphenols, then small intervention trials should be started with either polyphenol supplements or specific foods especially rich in the compound of interest, with the use of surrogate endpoints such as established cardiovascular and cancer risk factors, intima/media thickness, and flow-mediated dilation. It will be important to take genetic factors into account in such studies. It is possible that some people are more or less sensitive to potential beneficial or deleterious effects of polyphenols. If all of these types of studies support a protective effect, then large-scale, population-based, intervention studies should be launched to determine whether there is a true causal beneficial effect. However, these will occur in the distant future, because such studies represent the ultimate verification of effects.
In conclusion, we should take into consideration the lessons of the story of ß-carotene. We need to collect enough information from experimental, clinical, and epidemiologic research before we can support any specific recommendations regarding polyphenols (in particular, which polyphenols, at which doses, to achieve which benefits for which populations). To do this, we need more research, more funds, more patience, and more exchanges within the scientific community and among all research disciplines (basic and applied research). If we are not rigorous in finding clear answers for all of the aforementioned points, then we run the risk of needing to renounce recommendations in the future, damaging the credibility of nutritional recommendations from a public health viewpoint.
ACKNOWLEDGMENTS
I thank Fernando Viteri, University of California, Berkeley, and Children's Hospital Oakland Research Institute, for his many helpful comments during the preparation of this article.
REFERENCES
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
