Estrogen has been proposed to trigger breast cancer development via an initiating mechanism involving its metabolite, catechol estrogen (CE). To examine this hypothesis, we conducted a multigenic case-control study to determine whether polymorphisms of the genes responsible for CE formation via estrogen biosynthesis (CYP17) and hydroxylation (CYP1A1) and CE inactivation (COMT) are associated with an elevated risk for breast cancer in Taiwanese women, and whether the association between genotype and risk may be modified by estrogen exposure. One hundred and fifty breast cancer patients and 150 healthy controls were recruited. PCR-based RFLP assays were used to determine the genotypes of estrogen-metabolizing genes. The breast cancer risk associated with individual susceptibility genotypes varied among the three genes and was highest for COMT, followed by CYP1A1 and CYP17. After simultaneous consideration of all three genes and other well-established risk factors of breast cancer, the COMT genotype remained the most significant determinant for breast cancer development and was associated with a 4-fold increase in risk (95% confidence interval, 1.12-19.08). Furthermore, a trend of increasing risk for developing breast cancer was found in women harboring higher numbers of high-risk genotypes (P = 0.006), including the high activity CYP17 (CYP17 A2/A2), high inducibility CYP1A1 (CYP1A1 MspI vt/vt), and low activity COMT (COMT L/L) genotypes. The association of risk with the number of susceptibility genotypes was stronger in women with prolonged estrogen exposure (indicated by a higher number of estrogen exposure years or a higher number of estrogen exposure years between menarche and first full-term pregnancy), women with higher estrogen levels (implied by early menarche), and women with a higher body mass index (> or = 22.5). On the basis of comprehensive profiles of estrogen metabolism, this study supports the possibility that breast cancer can be initiated by estrogen exposure.