Diets high in fiber or certain fiber fractions have been hypothesized to protect against breast cancer through mechanisms that include inhibition of intestinal reabsorption of estrogen excreted via the biliary system (1, 2), reduction in estrogen synthesis by inhibition of human estrogen synthetase (particularly by lignin; Ref. 3), and a decrease in the glycemic index (particularly by cereal fiber; Ref. 4). However, few cohort studies have examined dietary fiber in relation to breast cancer risk, and none has examined fiber fractions (1). Therefore, we examined the relationships between total dietary fiber and dietary fiber fractions, and breast cancer risk in the largest prospective cohort study of this issue to date.

The design of our study has been described in detail elsewhere (5). In brief, 89,835 women ages 40–59 years were recruited into the Canadian National Breast Screening Study between 1980 and 1985 from the general Canadian population. Information was obtained from participants on demographic characteristics, lifestyle factors, menstrual and reproductive history, and use of oral contraceptives and replacement estrogens. Starting in 1982, a questionnaire regarding diet and physical activity was distributed to all of the new attendees at all of the screening centers, and to women returning to the screening centers for rescreening. By that time, some women had already been enrolled in the study and were not seen again at the screening centers. The intake of total dietary fiber and other nutrients was computed by multiplying the frequency of consumption of each unit of food by the nutrient content of the indicated portion size and summing overall foods (6). Dietary fiber fractions were estimated from food intake using values from the revised versions of McCance and Widdowson’s food composition tables (7). For total dietary fiber, the Pearson correlation coefficient between values obtained from food frequency questionnaires and diet records was 0.70 (6).

Outcome (incident breast cancer or death) was ascertained by computerized record linkage to the National Mortality Database, the Canadian Cancer Database, and the Ontario Cancer Registry. Cases were women who were diagnosed during follow-up with incident in situ or invasive carcinoma of the breast. The linkages yielded data on mortality and cancer incidence to December 31, 2000 for women in Ontario, December 31, 1998 for women in Quebec, and December 31, 1999 for women in other regions in Canada. Analyses were based on the 49,536 women with complete information on all of the covariates. Cox proportional hazards models were used to estimate hazard rate ratios and 95% confidence intervals. For tests of trend in risk across successive levels of categorical variables, median values of each category were fitted in the risk models as successive integers.

During an average of 16.2 years of follow-up (801,079 person-years), 2,536 incident breast cancer cases were diagnosed. Because the results were very similar in age-adjusted and multivariate-adjusted models, only the multivariate-adjusted rate ratios are shown in Table 1. We observed no association between breast cancer risk and total dietary fiber intake, or intakes of any of the specific fiber fractions, including soluble and insoluble fiber, fiber from cereals, fruit, and vegetables, lignin, and cellulose. We also found no association with any of the fiber fractions when we examined them as continuous variables or after categorizing them by deciles. For lignin, which showed a test for trend that approached statistical significance in the analysis of quintile levels, the relative risk for the highest versus lowest decile level was 0.90 (0.73–1.11; P for trend = 0.09). For a two-sided 5% significance level, the statistical power to detect a hazard ratio of 0.8 between the highest and lowest quintile levels of exposure was 92%. The results were not altered appreciably by exclusion of the 602 cases that occurred during the first 3 years of follow-up, exclusion of the 348 cases of in situ breast cancer, or by mutual adjustment for the various dietary fiber fractions.

This is the first prospective cohort study of fiber fractions and breast cancer risk. The prospective design of our study eliminates the potential for recall bias. However, the possibility of error in the measurement of fiber intake cannot be discounted. Nondifferential misclassification would have tended to attenuate true associations and could have resulted from inaccurate recall of past diet or from changes in diet over time. Although we adjusted our estimates for many potentially confounding variables, uncontrolled confounding from dietary (or other) factors cannot be excluded. Our null findings are consistent with the results of the few previous cohort studies of total fiber intake, which showed null or weak inverse associations (1). Although there were suggestions that certain fiber fractions, particularly lignin, might be associated with a reduction in risk, true associations, if they exist, appear likely to be weak.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1

This study was supported in part by the National Cancer Institute of Canada.

Table 1

Multivariate-adjusteda rate ratiosb for the highest versus the lowest quintile levelsc of total dietary fiber and fiber fractions in relation to risk of incident breast cancer

Total fiberSoluble fiberInsoluble fiberCereal fiberFruit fiberVegetable fiberLigninCellulose
Quintile 1 1.00 (referent) 1.00 (referent) 1.00 (referent) 1.00 (referent) 1.00 (referent) 1.00 (referent) 1.00 (referent) 1.00 (referent) 
Quintile 2 1.01 (0.88–1.15) 0.96 (0.84–1.10) 0.96 (0.84–1.09) 0.95 (0.83–1.08) 1.03 (0.90–1.18) 1.01 (0.88–1.15) 1.04 (0.91–1.18) 0.93 (0.81–1.06) 
Quintile 3 1.00 (0.86–1.15) 0.91 (0.79–1.06) 0.88 (0.76–1.01) 1.05 (0.92–1.19) 1.10 (0.96–1.26) 0.95 (0.82–1.09) 0.93 (0.81–1.07) 1.05 (0.91–1.20) 
Quintile 4 0.88 (0.75–1.02) 0.89 (0.76–1.04) 0.90 (0.78–1.04) 0.94 (0.82–1.08) 1.03 (0.89–1.19) 0.94 (0.80–1.10) 0.94 (0.82–1.09) 0.89 (0.76–1.03) 
Quintile 5 0.92 (0.78–1.09) 0.90 (0.75–1.08) 0.89 (0.76–1.03) 0.90 (0.78–1.04) 1.07 (0.92–1.25) 0.90 (0.75–1.08) 0.89 (0.76–1.03) 0.97 (0.83–1.14) 
P for trend 0.16 0.23 0.13 0.13 0.51 0.21 0.06 0.64 
Total fiberSoluble fiberInsoluble fiberCereal fiberFruit fiberVegetable fiberLigninCellulose
Quintile 1 1.00 (referent) 1.00 (referent) 1.00 (referent) 1.00 (referent) 1.00 (referent) 1.00 (referent) 1.00 (referent) 1.00 (referent) 
Quintile 2 1.01 (0.88–1.15) 0.96 (0.84–1.10) 0.96 (0.84–1.09) 0.95 (0.83–1.08) 1.03 (0.90–1.18) 1.01 (0.88–1.15) 1.04 (0.91–1.18) 0.93 (0.81–1.06) 
Quintile 3 1.00 (0.86–1.15) 0.91 (0.79–1.06) 0.88 (0.76–1.01) 1.05 (0.92–1.19) 1.10 (0.96–1.26) 0.95 (0.82–1.09) 0.93 (0.81–1.07) 1.05 (0.91–1.20) 
Quintile 4 0.88 (0.75–1.02) 0.89 (0.76–1.04) 0.90 (0.78–1.04) 0.94 (0.82–1.08) 1.03 (0.89–1.19) 0.94 (0.80–1.10) 0.94 (0.82–1.09) 0.89 (0.76–1.03) 
Quintile 5 0.92 (0.78–1.09) 0.90 (0.75–1.08) 0.89 (0.76–1.03) 0.90 (0.78–1.04) 1.07 (0.92–1.25) 0.90 (0.75–1.08) 0.89 (0.76–1.03) 0.97 (0.83–1.14) 
P for trend 0.16 0.23 0.13 0.13 0.51 0.21 0.06 0.64 
a

Multivariate models included study center, treatment allocations, age in 5-year age groups, body mass index, cigarette smoking (duration in years), education level (less than high school, high school, and university), vigorous physical activity (h per day in tertiles), oral contraceptive use (never + 4 levels of duration), hormone replacement therapy (never + 4 levels of duration), parity (quintiles), history of benign breast disease, history of breast self-exam, family history of breast cancer, menopausal status (pre, peri, post), intakes of energy (continuous), alcohol, calcium, vitamin C, vitamin E, folic acid, and saturated fat (quintiles).

b

95% confidence intervals are shown. All Ps are from two-sided tests.

c

Cut-point values (g/day) for the quintiles are as follows: total fiber (15.2, 18.6, 21.7, 25.8), soluble fiber (4.6, 5.6, 6.5, 7.8), insoluble fiber (2.8, 3.6, 4.5, 5.5), cereal fiber (2.6, 3.6, 4.4, 5.6), fruit fiber (2.0, 3.3, 4.7, 6.6), vegetable fiber (5.4, 7.0, 8.6, 11.0), lignin (1.0, 1.3, 1.6, 2.1), and cellulose (3.3, 4.4, 5.4, 6.9).

We thank Statistics Canada, the provincial and territorial Registrars of Vital Statistics, and the Cancer Registry directors for their assistance in making the cancer incidence and mortality data available. We also thank Dr. Charles B. Hall for his valuable input regarding power calculations.

1
Willett W. C. Diet and breast cancer.
J. Intern. Med.
,
249
:
395
-411,  
2001
.
2
Gerber M. Fiber and breast cancer: another piece of the puzzle–but still an incomplete picture.
J. Natl. Cancer Inst.
,
88
:
857
-858,  
1996
.
3
Adlercreutz H., Bannwart C., Wahala K., Makela T., Brunow G., Hase T., Arosemena P. J., Kellis J. T., Jr., Vickery L. E. Inhibition of human aromatase by mammalian lignans and isoflavonoid phytoestrogens.
J. Steroid Biochem. Mol. Biol.
,
44
:
147
-153,  
1993
.
4
Slavin J. L. Mechanisms for the impact of whole grain foods on cancer risk.
J. Am. Coll. Nutr.
,
19
:
300S
-307S,  
2000
.
5
Terry P. D., Jain M., Miller A. B., Howe G. R., Rohan T. E. Dietary carotenoids and risk of breast cancer.
Am. J. Clin. Nutr.
,
76
:
883
-888,  
2002
.
6
Jain M. G., Harrison L., Howe G. R., Miller A. B. Evaluation of a self-administered dietary questionnaire for use in a cohort study.
Am. J. Clin. Nutr.
,
36
:
931
-935,  
1982
.
7
Paul A. A., Southgate D. A. T. .
McCance and Widdowson’s The Composition of Foods
, Her Majesty’s Stationary Office London  
1978
.