This paper systematically reviews the results from epidemiologic studies investigating the hypothesis that breast cancer risk in postmenopausal women increases with increasing concentrations of estradiol in blood and with increasing urinary estrogen excretion rates. Data from 29 epidemiologic studies of endogenous hormones and postmenopausal breast cancer were used. The ratio of the average estrogen concentration in the women with breast cancer to that in the women without breast cancer (and its 95 percent confidence interval [CI]) was calculated for each study, and the results were summarized by calculating weighted averages of the log ratios. In six prospective studies of serum estradiol concentration, 329 women who subsequently developed breast cancer had, overall, a 15 percent (CI = 6-24 percent, P = 0.0003) higher mean concentration of estradiol in their blood than the 1,105 women who remained free of cancer. The results of these prospective studies did not differ significantly from each other (chi2 for heterogeneity = 8.7; degrees of freedom = 5; P > 0.1). Similar differences in mean estrogen levels were seen in the case-control studies which reported either estradiol concentrations in the blood or urinary estrogen excretion. However, the case-control studies showed significant heterogeneity among their results. The data from the prospective studies strongly suggest that breast cancer risk in postmenopausal women is associated with relatively high concentrations of endogenous estradiol.