Soy-based dietary supplements have been proposed as natural alternatives to menopausal hormone therapy, but their effect towards breast cancer risk is controversial and has little been evaluated. We investigated the association between the use of those supplements and breast cancer risk in a large prospective cohort of women aged over 50 years.

A total of 76,442 women from the E3N cohort, born 1925–1950, were followed between 2000 and 2011 (11.2 years on average; 3608 incident breast cancers). Use of soy-based dietary supplements was assessed every 2–3 years. Hazard ratios (HRs) of breast cancer and their 95% confidence intervals (CI) associated with current, past, and never use of soy supplements were estimated using multivariable Cox models.

HRs associated with current use of soy supplements were 0.92 (95% CI=0.76 to 1.11) for breast cancer overall, 0.78 (95% CI=0.60 to 0.99) for ER+ breast cancer and 2.01 (95% CI=1.41 to 2.86) for ER- breast cancer, compared to never use. No association was observed for past use. We found effect modification by family history of breast cancer (P=0.03) and menopausal status (P=0.04): HRs for current use were 1.36 (95% CI=0.95 to 1.93) among women with a family history of breast cancer in first-degree relatives, and 0.82 (95% CI=0.65 to 1.02) among women with no such history; they were 1.06 (95% CI=0.87 to 1.30) more than 5 years after menopause onset and 0.50 (95% CI=0.31 to 0.81) in perimenopause or within 5 years after menopause.

We report opposite associations of soy-based dietary supplements with ER+ and ER− breast cancer risks. Our results suggest caution in using them in women with a family history of breast cancer. Whether their risk profile may be more favorable among premenopausal or recently postmenopausal women deserves further investigation.