Abstract
Soy has a long history of consumption in Asia and was traditionally prepared by rinsing, cooking, and simmering, methods which remove estrogenic isoflavones (Isofls). Population studies have indicated that soy and/or Isofls may be associated with a decreased risk of breast cancer (BC), while in vitro and experimental data indicate dose-related proliferative effects of Isofls on breast cells. This review attempts to decipher the role of soy and Isofls in the risk of BC in women, since previous studies have suggested a lack of association with BC. Several dozen population studies conducted in Asian and Western countries were analyzed, as were data collected during in vitro animal and clinical trials of relevant doses of soy and Isofls. Although soy intake has been estimated well in Asian countries and could be related to preventive effects on BC risk, this has not been the case in the West, where the consumption of hidden soy is often omitted. However, in both cultures, the Isofl intake is misestimated, and the groups are misclassified. Indeed, in Asia, the origin of soy foods, i.e., homemade or industrial, has never been reported, and in the West, the amount of Isofls consumed in hidden soy has not been determined. Moreover, in most cohort studies, only a few subjects were exposed to active doses of Isofls on breast cells. Similarly, clinical interventions showed estrogenic effects of Isofls at relevant doses. Finally, population studies have not shown any convincing link between soy or Isofl intake and BC risk, likely because they have opposite effects on this pathology. Thus, based on in vitro, experimental, and clinical data, a deleterious effect of Isofls cannot be excluded when active doses are ingested, even if the soy food matrix can be protective.