In a case-control study conducted among Asian-American women in Los Angeles County, we reported that the risk of breast cancer was significantly reduced in association with soy intake [Wu,A.H., Wan,P., Hankin,J. et al. (2002) Carcinogenesis, 23, 1491-1496]. In a subset of cases (n = 97) and controls (n = 97) we investigated the relationship between self-reported usual adult intake of soy isoflavones which was determined from a food frequency questionnaire and levels of plasma isoflavones (genistein and daidzein) and isoflavone metabolites (equol, dihydrogenistein and dihydrodaidzein) from a randomly timed blood specimen. In analyses conducted in cases and controls separately, levels of plasma genistein, daidzein and total isoflavones increased with increasing levels of self-reported intake of soy isoflavones. Breast cancer cases and control subjects did not differ in their respective associations between total plasma isoflavone levels and self-reported intake (P = 0.48). Among all subjects, there was a 3-fold difference in geometric mean plasma levels of total isoflavones [81.8 (95% CI = 53.4, 125.1) versus 26.4 nmol/l (95% CI = 16.6, 41.8)] between women in the highest quartile of soy isoflavone intake (>12.68 mg isoflavones/1000 kcal) compared with those in the lowest quartile of intake (<or=1.79 mg isoflavones/1000 kcal), a difference that was statistically significant (P = 0.002). The present study provides independent corroboration that breast cancer cases and control subjects can reliably recall their usual soy intake and that there is no evidence of selective recall biases between breast cancer cases and controls. These results further strengthen our previous observation of an inverse association between soy intake and breast cancer risk in the Los Angeles Asian Breast Cancer Study.