Dietary fiber intake and risk of postmenopausal breast cancer defined by estrogen and progesterone receptor status—A prospective cohort study among Swedish women Reiko Suzuki 1 , Tove Rylander-Rudqvist 1 , Weimin Ye 2 , Shigehira Saji 3 , Herman Adlercreutz 4 and Alicja Wolk 1 * 1 The National Institute of Environmental Medicine, Division of Nutritional Epidemiology, Karolinska Institutet, Stockholm, Sweden 2 Department of Medical Epidemiology and Biostatistics, Karolinska Institutet, Stockholm, Sweden 3 Division of Clinical Trials and Research, Tokyo Metropolitan Cancer and Infectious Diseases Center, Komagome Hospital, Tokyo, Japan 4 Institute for Preventive Medicine, Nutrition, and Cancer, Folkh € alsan Research Center and Division of Clinical Chemistry, Biomedicum, University of Helsinki, Finland There is few data on the association between dietary fiber intake and estrogen receptor (ER)/progesterone receptor (PR)-defined breast cancer risk. We evaluated the association between dietary fiber and ER/PR-defined breast cancer risk stratified by post- menopausal hormone use, alcohol intake, and family history of breast cancer in the population-based Swedish Mammography Screening Cohort comprising 51,823 postmenopausal women. Fiber intake was measured by food-frequency questionnaire col- lected in 1987 and 1997. Relative risks (RRs) were estimated by hazard ratio derived from Cox proportional hazard regression models. During an average of 8.3-year follow-up, 1,188 breast can- cer cases with known ER/PR status were diagnosed. When com- paring the highest to the lowest quintile, we observed non-signifi- cant inverse associations between total fiber intake and the risk of all tumor subtypes; the multivariate-adjusted RRs were 0.85 (95% CI: 0.69–1.05) for overall, 0.85 (0.64–1.13) for ER1PR1, 0.83 (0.52–1.31) for ER1PR2 and 0.94 (0.49–1.80) for ER2PR2. For specific fiber, we observed statistically significant risk reduc- tions for overall (34%) and for ER1PR1 (38%) for the highest versus lowest quintile of fruit fiber, and non-significant inverse associations for other subtypes of cancer and types of fiber. Among ever-users of postmenopausal hormone (PMH), total fiber intake and especially cereal fiber were statistically significantly associated with 50% reduced risk for overall and ER1PR1 tumors when comparing the highest to the lowest quartile, but no association was observed among PMH never users. Our results suggest that dietary fiber intake from fruit and cereal may play a role in reducing breast cancer risk. ' 2007 Wiley-Liss, Inc. Key words: breast cancer; estrogen receptor; progesterone receptor; dietary fiber; risk Breast cancer is the most prevalent malignancy among women in Western countries. A number of studies indicate that a high level of endogenous/exogenous steroid hormones may play a criti- cal role in the development of postmenopausal breast cancer. 1–4 It has been postulated that dietary fiber intake may reduce estrogen levels through interference of fiber with their enterohepatic metab- olism. 5 Proposed mechanisms include increasing fecal excretion of estrogens, inhibition of intestinal reabsorption of estrogens by suppressing bacterial b-glucuronidase activity or binding unconju- gated estrogens to fiber in the colon. 6–10 It has been also reported an association of increased dietary fiber intake and fiber-related phytochemicals with a reduction of estrogen levels in blood 6,11–13 and with increased level of sex hormone-binding globulin (SHBG), which binds estradiol affecting its bioavailability. 14–16 Thus, it has been hypothesized that fiber intake may be associated with a decreased risk of breast cancer through the reduction of cir- culating estrogen levels. 17 Moreover, fiber might also indirectly reduce the risk of postmenopausal breast cancer by preventing obesity and type 2 diabetes—both being risk factors for estrogen receptor positive (ER1)/progesterone receptor positive (PR1) postmenopausal breast cancer 18,19 —through the control of insulin resistance. 20–22 Furthermore, a series of animal studies reported that the inhibitory effect of dietary fiber on the promotion of mam- mary carcinogenesis was not only through modulating the level of circulating 17b-estradiol, 8,23,24 but a potential effect of bioactive compounds with anticancer properties such as antioxidants, phy- toestrogens, phytic acid and protease inhibitors in dietary fiber was also discussed. 25–28 Epidemiological studies have evaluated intake of total fiber in relation to overall breast cancer risk in 10 cohorts, 29–38 in a meta- analysis of 12 older case–control studies, 39 and 8 more recent case–control studies, 40–47 but results are not consistent. To date, only 2 prospective studies with a limited number of cases have reported the association between intake of fiber and the risk of postmenopausal breast cancer defined by ER/PR status. 29,37 To evaluate the association between intake of dietary fiber (total, cereal, fruit and vegetable) and the risk of postmenopausal breast cancer and to assess whether the association differs across ER/PR status, we performed a large population-based prospective cohort study among Swedish postmenopausal women. Further- more, we also investigated whether the association was modified by use of postmenopausal hormones (PMH), levels of body mass index (BMI), alcohol intake, and family history of breast cancer. Methods The Swedish mammography cohort The Swedish Mammography Cohort (SMC) was previously explained elsewhere. 48 Briefly, the cohort was established in 1987–1990 in V€ astmanland County and Uppsala County in Swe- den. All women who were born between 1917 and 1948 in V€ astmanland County (n 5 41,786) and between 1914 and 1948 in Uppsala County (n 5 48,517) were invited to a mammography screening. A total 66,651 women completed a questionnaire (SMC-87) that elicited information on diet, weight, height, parity, age at first birth, family history of breast cancer and education level (response rate 74%). The information on age at menarche, age at menopause, history of use of oral contraceptives (OC) and PMH was obtained from supplemental questionnaires in Uppsala County in 1988–1990. In 1997, a second questionnaire (FFQ-97), which had addition- ally included questions about waist and hip ratio, details on repro- Grant sponsors: Swedish Cancer Foundation, Swedish Research Coun- cil/Longitudinal Studies. *Correspondence to: The National Institute of Environmental Medi- cine, Division of Nutritional Epidemiology, Karolinska Institutet, Nobelsv€ ag 13, P.O. Box 210, S-171 77 Stockholm, Sweden. Fax: +46-830-4571. E-mail: alicja.wolk@ki.se Received 30 March 2007; Accepted after revision 5 July 2007 DOI 10.1002/ijc.23060 Published online 31 August 2007 in Wiley InterScience (www.interscience. wiley.com). Abbreviations: BMI, body mass index; CIs, confidence intervals; ER, estrogen receptor; FFQ, food-frequency questionnaire; OC, oral contracep- tives; PMH, postmenopausal hormones; PR, progesterone receptor; SD, standard deviation. Int. J. Cancer: 122, 403–412 (2008) ' 2007 Wiley-Liss, Inc. Publication of the International Union Against Cancer