In this study, with a larger number of proliferative BBD cases (n = 682) than the previous retrospective analysis in the same cohort (42
), we confirmed the originally observed significant inverse association between adolescent intake of fiber and nuts and risk of proliferative BBD. The unique feature of the current study is the addition of new prospective cases diagnosed after completion of the high school diet questionnaire. The results of the prospective only analysis were consistent with those of the larger combined analysis, suggesting that recall bias is not a plausible explanation of the inverse associations observed. Consistent with the original analysis (42
), with the addition of over 200 incident cases, no association was observed for fruit or vegetable consumption and proliferative BBD.
Our results suggest that fiber intake during adolescence is protective in the early breast carcinogenic process. Previous studies observed increased adolescent fiber intake to be associated with a reduced risk of breast cancer (37
), suggesting again the potential role of dietary intake in early life in breast cancer development. Inconsistent results, however, were reported in studies that have examined the relations of adult fiber intake with BBD (27
) or breast cancer (5
). No associations were observed between adult fiber intake and BBD risk overall or the proliferative subtypes in the only prospective analysis of adult diet and BBD in the same NHSII cohort (32
). In the current study, the results were unchanged after additional adjustment for adult fiber intake, suggesting that the observed association was independent of recent intake.
Dietary fiber intake has been hypothesized to be protective in mammary tumorigenesis through several biological mechanisms. Dietary fiber may increase excretion of estrogen by inhibiting deconjugation and reabsorption of estrogen from the gastrointestinal tract (4
). In addition, the protective effect may also be partly due to the anti-estrogenic effects of lignans and isoflavonoid compounds, which occur naturally in fiber-rich foods or arise as a result of bacterial action on such foods (54
). Dietary lignans were postulated to have an inhibitory effect on cell proliferation in breast tumors through decreased levels of circulating estrogens either by inhibiting the aromatase enzyme in biosynthesis of estrogen (58
) or by stimulating the synthesis of sex-hormone binding globulin (SHBG) (56
). Diet intervention studies have shown that the combination of low dietary fat and high dietary fiber was associated with reduced serum estrogen levels among premenopausal women (59
). Our results provide further support to the hypothesis that exposures during a susceptible period in women’s life play important roles in breast cancer development. It is possible that fiber intake during adolescence may set steroid hormone levels and endocrine profiles in adulthood and reduce risk of proliferative BBD and/or breast cancer. Future studies are needed to examine the associations between adolescent dietary fiber intake and adolescent and/or adult serum estrogen levels to confirm the hypothesized biological mechanisms.
Previous studies have shown that soluble and insoluble fiber may have different effects possibly through different biological mechanisms (22
). We do not have data on soluble or insoluble fiber, limiting our ability to assess their relations with proliferative BBD risk. We did, however, observe that, in general, higher intake levels of fiber from different sources were associated with lower risk of proliferative BBD. No statistically significant trends were observed for fiber from different sources possibly due to the relatively narrower range and lower variability of the intake levels for specific sources of fiber, compared to total fiber. The significant inverse association observed between total fiber and proliferative BBD could be due to the total amount of fiber with each source of fiber contributing a little, suggesting that dietary fiber itself, rather than a specific source of fiber, is important. Alternatively, results for fiber became somewhat attenuated when we additionally adjusted for nuts, suggesting that nut intake accounts for some of the fiber BBD association and a combination of nutrients including dietary fiber found in certain whole foods such as nuts act synergistically to produce the effect instead of just dietary fiber. Additional studies are warranted to further clarify the effects of different sources and types of fiber throughout the life course on breast proliferation and breast carcinogenesis.
High school intake of nuts was also related to a significantly reduced risk of proliferative BBD. Results were fairly consistent across the different types of nuts examined in this study. Results for nuts were essentially the same with additional adjustment for fiber, suggesting that in addition to fiber, the inverse associations between nut intake and proliferative BBD risk may also be attributable to nutrients other than fiber in nuts. A significant inverse association was observed between adolescent nut intake and proliferative BBD in the previous retrospective study by Baer et al. (42
), whereas no association was found between adult nut intake and breast cancer in the Malmo Diet and Cancer cohort (21
) Nuts are rich sources of unsaturated fat as well as a variety of other bioactive compounds (62
). Nut intake has consistently been reported to reduce the incidence of cardiovascular disease (63
). Given the scarce literature on nut intake and cancer risk, more experimental and epidemiological research is clearly needed to better understand whether the intake of this food group provides health benefits with respect to cancer and to elucidate the possible mechanisms of action.
We observed no association between adolescent intake of fruits and vegetables and proliferative BBD, although these food groups are also sources of fiber. Further examination of individual food items making up large percentages of fiber revealed an inverse association between apple intake and proliferative BBD, particularly in the prospective analysis. The associations between high school intake of fruits and vegetables and proliferative BBD have not been evaluated in previous studies except the original analysis in NHSII (42
). Some studies, however, have reported inverse associations between adult intake of fruits and vegetables and proliferative BBD risk (27
). Recently, a case-control study in China found inverse associations between adult intake of fruits and vegetables and risk of both proliferative fibrocystic breast conditions alone and with concurrent breast cancer (27
) and the risk of all three types of fibrocystic breast conditions (nonproliferative, proliferative, and atypia) (29
). Further, no significant association was observed for total crude fiber after adjustment for total fruit and vegetable intake (27
). One possible explanation of the inconsistent findings between the current study and the previous studies is that dietary intake during different periods in a woman’s lifetime may have different relations with BBD risk. Different study design may also partly account for the discrepant results, as our study is the only cohort study, while all the previous studies being case-control studies. Other factors may include different study populations and consequently different dietary patterns in these populations. For instance, an overall dietary pattern rich in fruits and vegetables in China could be quite different from those of western countries. Future large prospective studies are needed to clarify the role of fiber and fruit and vegetable intake during adolescence and adulthood in breast cancer development.
There are several major strengths in our study. The combined analysis included a large number of proliferative BBD cases, and this was the first prospective analysis to evaluate the relations between adolescent dietary intake and proliferative BBD risk. The centralized pathology review of BBD cases reduces the likelihood of misclassification, and we focused on a specific histological subtype of proliferative BBD, a marker of increased breast cancer risk. The similar results of the combined and prospective analyses suggest that any possible bias due to BBD diagnosis or changes in diet after BBD diagnosis on the recall of adolescent dietary intake should be minimal. We collected detailed information on potential confounding factors and adjusted for these factors in our analyses. Although the possibility of residual confounding cannot be completely ruled out, the almost identical results observed in the age-adjusted and multivariate-adjusted analyses suggest that it is unlikely that uncontrolled confounding could entirely account for the observed associations.
The study does have limitations. Although the HS-FFQ has been shown to be moderately reproducible and not strongly correlated with current diet as an adult (44
) and comparison with maternal report provided some form of validity, recall of adolescent diet in our study exceeded an average of 25 years (44
), and the validity of recall 15–35 years later has not been established. The estimates of the prospective analysis are less stable than those of the combined analysis due to the small number of prospective cases, and results need to be confirmed with longer follow-up and more cases. In addition, studies are needed to confirm our findings and to identify the biological mechanisms of action, particularly for nut intake. Furthermore, given the multiple comparisons of many foods tested in this study, the observed inverse association between adolescent apple intake and proliferative BBD in the prospective analysis could be a chance finding. Finally, the results of this study may not be readily generalizable to the general population of U.S. women, given that the reported intake levels of fiber in our study participants is higher than the intake levels for most Americans. In addition, the adolescent diets assessed in this study are from over 20 years ago. The diets of adolescent girls today may contain higher levels of fiber due to the addition of fiber to certain foods and increases in the consumption of whole grains in general. However, if adolescent fiber intake is associated with reduced risk of proliferative BBD through hormonal mechanisms as hypothesized, it is unlikely that the biological effects of fiber would have different effects in other populations of women.
In summary, our study observed significant inverse associations between adolescent dietary intake of fiber and nuts and risk of proliferative benign breast disease. Our results provide supportive evidence of the important role of dietary exposures during a unique period in a woman’s life in the earlier stage of breast carcinogenesis. These findings, if corroborated, may suggest a viable means for breast cancer prevention.