Subjects of this study consisted of 333 women (aged 45-75 years) drawn from a large United Kingdom prospective study of diet and cancer, the European Prospective Investigation of Cancer and Nutrition-Norfolk study. Using newly developed gas chromatography/mass spectrometry and liquid chromatography/mass spectrometry methods incorporating triply (13)C-labeled standards, seven phytoestrogens (daidzein, genistein, glycitein, O-desmethylangolensin, equol, enterodiol, and enterolactone) were measured in 114 spot urines and 97 available serum samples from women who later developed breast cancer. Results were compared with those from 219 urines and 187 serum samples from healthy controls matched by age and date of recruitment. Dietary levels were low, but even so, mean serum levels of phytoestrogens were up to 600 times greater than postmenopausal estradiol levels. Phytoestrogen concentrations in spot urine (adjusted for urinary creatinine) correlated strongly with that in serum, with Pearson correlation coefficients > 0.8. There were significant relationships (P < 0.02) between both urinary and serum concentrations of isoflavones across increasing tertiles of dietary intakes. Urinary enterodiol and enterolactone and serum enterolactone were significantly correlated with dietary fiber intake (r = 0.13-0.29). Exposure to all isoflavones was associated with increased breast cancer risk, significantly so for equol and daidzein. For a doubling of levels, odds ratios increased by 20-45% [log(2) odds ratio = 1.34 (1.06-1.70; P = 0.013) for urine equol, 1.46 (1.05-2.02; P = 0.024) for serum equol, and 1.22 (1.01-1.48; P = 0.044) for serum daidzein]. These estimates of risk are similar to those established for estrogens and androgens in postmenopausal breast cancer but need confirmation in larger studies.