The estrogenic and antiestrogenic effects of isoflavones, phytoestrogens contained in soy foods, have been proposed as mechanisms for the possible involvement of soy products in the development of breast cancer. We investigated the hypothesis that isoflavones reduce mammographic density, a predictor of breast cancer risk. We conducted a double-blind randomized trial in premenopausal women who received a daily 100 mg isoflavone supplement or a placebo over 12 months. Compliance with the study regimen was confirmed by urinary isoflavones and tablet counts. We used a computer-assisted method to measure mammographic density and paired t-tests to assess changes in mammographic characteristics from baseline to follow-up mammogram. Complete sets of mammograms were available for 30 women. The two groups differed by age and mammographic density at baseline, but were similar in body weight and nutritional intakes. We detected no significant changes either in the size of the dense areas or in the per cent densities. A non-significant decrease in breast area among intervention group subjects was probably the result of methodological issues in comparing mammograms taken under different conditions. In conclusion, our findings do not support the hypothesis that isoflavones decrease mammographic density during a one-year intervention. Although this exploratory study had limited power, it appears that isoflavones do not exert an estrogenic effect similar to hormone replacement therapy on mammographic density.