Authors: Suzanne Neubauer, PhD, RD, CNSD and Patricia Bebo, MS, RD
In January of 1999, the New England Journal of Medicine published a study on women which concluded that dietary fiber, from fruits, vegetables and cereals, provided no protective effect against the development of colorectal adenomatous polyps or cancer (1). The popular press seized upon the conclusions and brought into question years of dietary advice given by the medical community and several nationally recognized organizations, including the American Dietetic Association and the American Cancer Society. The public began to question the real necessity of striving to eat the recommended 25-35 grams of fiber per day, even though the authors maintained that there were still very substantial and valid reasons to increase fiber intake.
Colorectal cancer is the fourth most common cancer in the United States. It is the only cancer that strikes men and women almost equally. In 2000, colorectal cancer will account for 10% of all new cancers and 10% of all cancer deaths in the United States (2). Worldwide there is a 20-fold variance in incidence rates of colorectal cancer. In general, the variance has been attributed to dietary and environmental differences. In a 1971 landmark study, Burkitt (3) concluded that the low incidence of colorectal cancer in Africa was due to high dietary fiber intake. Thus began an odyssey of epidemiological studies around the world to verify Burkitt’s findings. Several studies published in the last decade are reviewed here in an effort to determine whether the generally accepted theory that high fiber diets exert a protective effect against the development of adenomatous polyps and colorectal cancer is still valid.
Causes of Colon Cancer
Colorectal cancer is a slow developing cancer that may be asymptomatic and therefore remain undiagnosed for several years after initial development. The cancer can develop in any of the four sections of the colon or the rectum. Adenomatous polyps are thought to be precursors of cancer development and they may provide information on the initiation phase of malignant cell proliferation. Ninety-five percent of colorectal cancers are adenocarcinoma (2).
Dietary Fiber: Intake vs. Outcome
Several studies have investigated total dietary fiber intake, fiber intake from specific food sources, and the components of dietary fiber to uncover clearly defined relationships to colorectal adenomatous polyps and/or cancer (1, 4-14).
Total Dietary Fiber
Platz et al. (4) investigated dietary fiber and its relationship to colorectal adenoma in men. This prospective cohort study was based on data collected from the Health Professionals Follow-up Study. The cohort was followed for 8 years and diet was assessed at baseline with a 131-item semi-quantitative food frequency questionnaire (SFFQ). When comparing the highest quintile of intake (32.3 g/day) to the lowest (11.6 g/day), total dietary fiber had a significant inverse relationship to the risk of colorectal adenoma of the distal colon. Increasing intakes of total dietary fiber also had a significant inverse association with the development of incident polyps of the distal colon. Similarly, a case-control study conducted by Neugut et al. (5) found that intake of total dietary fiber, highest quartile (> 18.8 g/day) compared with the lowest quartile (<11.2 g/day), had a significant inverse relation to the risk of recurrent colorectal adenomatous polyps in women but not in men.
In addition, Giovanucci et al. (6), reporting results of an earlier evaluation of data from the Health Professionals Follow-up Study, and Negri et al. (7), whose casecontrol study investigated cancer of the colon and rectum, both concluded that total dietary fiber exerted a protective effect against the development of adenomatous polyps and colorectal cancer. In contrast, Peters et al. (8) published results from a case-control study of men and women from Los Angeles County, California. The study subjects were age, sex, race and neighborhood matched. The diet of the subjects was assessed over the previous 15 years. They found no relationship between total dietary fiber intake and the risk of colon cancer.
Generally, most case-control studies assess diet from two to three years previous to diagnosis in order to get the most accurate recall possible. This is the case in the Neugut et al. (5) and Negri et al. (7) studies. The probability of recall bias in the Peters et al. (8) study, assessed over 15 years, seems quite likely. Overall, these studies suggest that total dietary fiber is protective against the development of colorectal adenomatous polyps and cancer.
In an effort to further delineate what specific food groups provide a protective effect, several studies (4, 6-12) investigated the food sources of dietary fiber. The main dietary contributors of fiber are fruits, vegetables, legumes and cereals/grains.
Negri et al. (7) conducted a large retrospective case-control study in which cases and controls were compared from several teaching and general hospitals in Italy. The groups were unmatched but multivariate analysis of the data controlled for several potentially confounding variables. The study had 96% participation and diet was assessed during 2 years pre-diagnosis, using a 78-item interview administered food frequency questionnaire (FFQ). High fruit fiber intake (13.38 g/d) vs. low intake had a significant inverse relationship to cancer of the colon, rectum, and colon/rectum.
Sandler et al. (9) reported similar results in women. Researchers recruited cases and controls from patients who underwent a colonoscopy. Diet was assessed, during 1 year pre-diagnosis, via a 117-item quantitative FFQ. Fruit intake (high > 22.3 servings /week vs. low < 8.4 servings/week) had a significant inverse relationship to the risk of colorectal adenoma. In addition, the authors also reported a significant inverse trend with intake (high > 8.81 g/d vs. low < 3.99 g/d) of fiber from fruits and vegetables combined. In agreement with these findings, Platz et al. (4) reported significant findings in men with fruit fiber intake (8.4 g/d vs. 1.3 g/d) and adenomatous polyps of the distal colon.
The findings of Peters et al. (8) and Le Marchand et al. (10) do not support the findings of Negri et al. (7) and Sandler et al. (9). Both studies were case-control designs. Peters et al. (8) reported no protective effect of fruit against the development of colon cancer and Le Marchand et al. (10) had similar findings when looking at the relationship of non-starch polysaccharide (NSP) from fruit to colorectal cancer in men and in women.
It may be noted that there is a mixture of results in the literature, and that these studies show associations – not cause and effect. As discussed previously, the results from Peters et al. (8) may suffer from dietary recall bias. The other studies discussed are comparable. There does appear to be an association between increasing fruit intake and fruit fruit fiber in particular, and decreased incidence of adenomatous polyps and/or colorectal cancer. More research is necessary in this area to clarify if any particular constituent of fruit is more protective than another.
Le Marchand et al. (10) conducted a large case-control study using subjects from Hawaii. The researchers matched cases and controls for age, sex and ethnicity. The subjects’ diets were assessed using a 282-item FFQ and the study had a 66% participation rate. The reference period for the FFQ was 3 years prior to onset of symptoms for the cases and before the interview for controls. Dietary fiber from the various food sources was measured using enzymatic chemical methods and reported as non-starch polysaccharide (NSP) using these methods.
NSP from vegetables had a significant inverse relationship with the risk of developing colorectal cancer. Comparison between high (5.3 g/d) vs. low (2.7 g/d) intake for men and women showed a significant inverse trend. Le Marchand et al. (10) then analyzed NSP intake by subsite of cancer occurrence. In men, NSP from vegetables had a significant inverse trend in relation to cancer of the rectum and, in women, NSP from vegetables showed a significant inverse trend in relation to cancer of the right colon and left colon. In agreement with Le Marchand et al. (10), Caygill et al. (11) assessed data available through the Food and Agricultural Organization and reported an inverse trend in colorectal cancer with respect to intake in the current diet as compared to intake 20 years previous (slope -2.23). Thun et al. (12) reported on results from the Cancer Prevention II study. In this report the large cohort was used for a case-control analysis. Vegetable, citrus and high-fiber grains together had an inverse association to the risk for colon cancer.
The findings of Fuchs et al. (1) do not agree with the previously discussed findings on vegetable fiber. The authors analyzed data from the Nurses’ Health Study, a prospective cohort study, which compiled data from women over 16 years. Diet was analyzed three times, in 1980, 1984 and 1986. The SFFQs had 61, 121 and 136 items, respectively. Comparisons of vegetable fiber were made between high (10.0 g/d) and low (2.7 g/d) intake. Vegetable fiber was associated with a significant increase in the risk of colorectal cancer. Their results were based on analysis of the food intake information from the 1980 SFFQ that had only 61 items. The cohort studies and most of the case-control studies reviewed used questionnaires with over 100 food items. This may have resulted in an underrepresentation from a variety of vegetables sources, which in turn may have underestimated intake and increased the relationship to risk.
With the exception of the Nurses’ Health Study (1), the studies reviewed reported strong significant inverse trends, which strongly suggests that vegetable fiber is protective against colorectal adenomas and cancer.
As mentioned, Thun et al. (12) reported a decreased risk of colon cancer with a combination intake of vegetable, citrus and high fiber grains. Several studies (1, 4, 6- 8, 10,11) looked at cereal and grain independent of other food sources of fiber.
Giovanucci et al. (6) analyzed data compiled in the Health Professionals Followup Study. The study investigated dietary fiber and its relationship to colorectal adenoma in men. Diet was assessed at baseline with a 131 item SFFQ. Fiber from grains had a significant inverse relationship with the development of colorectal adenomas in men. The results were based on a comparison of high intake (> 10.6 g/ d) vs. low intake ( < 2.8 g/d). Platz et al. (4), reporting on the same cohort after data was collected for an additional 6 years, found no association between grain fiber intake and colorectal adenomas. There are several possible explanations for the different results. The earlier study assessed the data using a limited number of confounders. Also several additional risk factors have been identified in the years since Giovanucci et al. (6) reported their findings. Platz et al. (4) adjusted for these confounders in their assessment of the cohort’s diet recalls. The smaller number of cases reported at the 2 years follow-up vs. 8 years also may have contributed to inexact conclusions.
Caygill et al. (11) reported a significant inverse association between cereal intake and the risk of colorectal cancer in women. In addition to the null findings of Platz et al. (4), other studies (3, 8,10) also reported no association between intake of fiber from cereals and grains and the risk of colorectal adenomas and cancer.
In contrast, Negri et al. (7), analyzing data from a case-control study previously outlined, reported a significant positive relationship between grain fiber, high intake (6.54 g/d) compared to low intake (3.3 g/d), and cancer of the colon and colon/ rectum combined.
Given the substantial studies with null results and the remainder with varying positive and negative results there is little evidence that cereal fiber exerts a protective effect with respect to the development of colorectal adenomas and cancer.
Fiber from food sources has been researched extensively. However the measurable components, i.e. soluble, insoluble, pectin, cellulose, hemicellulose and lignin, have not been as extensively covered in the literature. The components may hold a clue to further explain the relationship between plant foods, mainly fruits, vegetables and whole grains, and colorectal cancer. Four recently published studies (4,7,10,13) reported findings on the effect of fiber components and the risk of colorectal adenoma or cancer in men and women.
Slattery et al. (13) published results of a population based case-control study conducted in the United States. Diet was assessed for two years prior to diagnosis for cases or date of selection for controls. Soluble fiber had a significant inverse relationship in men over 67 years old when comparing the highest quintile of intake (> 11.7g/d) with the lowest quintile of intake ( < 5.6 g/d). In addition, soluble fiber was inversely associated with cancer of the proximal colon in women.
In agreement with Slattery et al. (13), Platz et al. (4) found soluble fiber, high intake (9.4 g/d) versus low intake (3.4 g/ d), had a significant inverse relationship with adenomatous polyps of the distal colon. Two other case-control studies (7,10) also reported inverse relationships. Pectin. Only Slattery et al. (13) measured pectin intake and found pectin had a significant inverse trend in relation to the risk of colon cancer. Specifically, it was found protective in all male subjects, in males over 67 years old, and in men with cancer of the proximal colon.
Slattery et al. (13) reported that insoluble fiber exerted a 30% decreased risk of development of colon cancer for all female subjects and a 50% decreased risk for cancer of the proximal colon in woman. In addition, Negri et al. (7) also showed an inverse relationship in men and women for colorectal cancer. In contrast, Platz, et al. (4) found no inverse relationship. Cellulose. Platz et al. (4) found increased consumption of cellulose (8.6 g/d versus 3.0 g/d) was inversely related to the risk of development of adenomatous polyps in the distal colon. In agreement with Platz et al. (4), LeMarchand et al. (10) reported a significant inverse relationship of cellulose to colorectal cancer for men and women. Negri et al. (7) also found an inverse relationship with the development of colorectal cancer.
The remainder of the components, hemicellulose and lignin either were not measured or showed no significant trend in relation to colorectal adenoma or cancer. Given the small sampling of studies available for the components of fiber and their relationship to colorectal adenomas and cancer, it may be cautiously concluded that souble fiber, including pectin, as well as insoluble fiber, including cellulose, exerts a protective effect. However, further research must be conducted to clarify the relationships.
The key to developing an overall conclusion about the relationship between dietary fiber and colorectal adenomas and cancer is to look at the studies in aggregate. No single study can account for all the variation in human diets, accurately measure all variables and account for all confounders (14). Therefore, one must be cautious when interpreting the results of a highly publicized study such as Fuchs et al. (1).
Table 1 summarizes the findings of all the studies reviewed in this paper. The scope of this paper did not permit a comprehensive review, however, the studies selected were those felt to be best designed and those whose primary objective was to identify the relationship between dietary fiber and colon cancer. The table includes both significant and non-significant results. The inclusion of all studies is important because epidemiologic studies report results as trends. An inverse or positive non-significant result often lends support to significant findings or leads researchers to expand investigations based on trends. Therefore, they should be considered when assessing results from several studies.
This review of the current literature presented both case-control and cohort study designs in approximately equal proportions. Eight of the 15 studies presented were case-control and 6 of the 15 were cohort designs. The results may be viewed in aggregate because the information is reported across an equal variety of epidemiologic methodologies.
Total dietary fiber was inversely associated with colorectal adenomas and/or cancer in 73% of the studies, fruit fiber in 77% of the studies, and vegetable fiber in 66% of the studies (Table 1). Cereal/grain fiber was found protective in 36% of the studies but the null hypothesis or no effect was found for cereal fiber in 45% of the studies. Therefore, based on the aggregate review of these studies, intakes of fruit fiber, vegetable fiber and total dietary fiber are inversely related to the development of colorectal adenomas and cancer. The results for cereal fiber are not consistent, therefore no conclusion may be made.
Although the studies are few, it may also be cautiously concluded that soluble and insoluble components of fiber exert a protective effect against the development of colorectal adenomas and cancer (Table 2). An important point to make is that fruits, vegetables and cereal/grains are the main sources of these fiber components in the human diet.
In an early finding from the Nurses’ Health Study (17), intake of red meat was significantly associated with the risk of colon cancer. Although Fuchs et al. (1) did not find a relationship between fiber and colon cancer, women who reported higher fiber intake had fewer servings of beef, pork or lamb/week. Increasing fruit, vegetable and grain intake and decreasing red meat intake is still an appropriate recommendation.
As our knowledge of food composition continues to expand, we may find other components, e.g., antioxidants, isoflavones, carotenoids, in these foods to be anticarcinogenic (19). Although additional research is warranted to further delineate the specific components of each food that may impart a protective effect upon the endothelial lining of the colon and rectum, food pattern analyses are still useful in that they allow for interaction effects of multiple qualities in whole foods. Steinmetz and Potter (19) list the many possible anticarcinogenic mechanisms of substances in fruits and vegetables, including the alteration of bacterial flora, bile acid composition, pH, and fecal bulk. Fruit and vegetable fiber and cereal fiber likely complement each other in the prevention of cancer. Cereal fiber also binds bile acids reducing transit time and increasing stool bulk. Bile acids are then fermented to short-chain fatty acids which lowers colonic pH and inhibits the conversion of primary to secondary bile acids. The secondary bile acids are thought to promote carcinogenesis.
Implications for Practice
Based on the results reported in this paper, dietetics professionals should remain committed to the 5 a Day for Better Health program. Both vegetarians and non-vegetarians with adequate intake will benefit from fruit and vegetable fiber. Continued advocacy of a diet high in fiber, with a dietary goal of 20-35 grams per day, is recommended.
Used with permission from the Vegetarian Nutrition DPG.