Estradiol and breast cancer risk.

A144

Background Estradiol levels have been associated with breast cancer risk in large cohort studies. We are conducting a Phase II chemoprevention study in which Letrozole, a potent selective non-steroidal aromatase inhibitor, is compared to placebo in postmenopausal women not using hormonal agents whose eligibility is determined only by estradiol level. This analysis explores the correlates of estradiol and established breast cancer factors among women screened for trial eligibility. Methods Screened subjects (postmenopausal women not using hormonal agents or osteoporosis medications) completed a self-administered questionnaire regarding reproductive history, hormone or SERM use, and family cancer history. Estradiol levels were measured using the double extraction technique (Esoterix Inc, Calabasas, CA) and stratified as highest quartile ≥0.9 ng/dL or lower. Crosstabular analyses and Chi-Square tests (0.05 level of significance) were performed on the overall cohort and by cancer status to explore potential differences between the two estradiol groups with regard to the following variables: age, BMI, type of menopause, family breast cancer history and Gail score, and previous hormone use. Results Our sample consisted of 324 postmenopausal women (amenorrhea ≥ 12 mo, ≥ 55 yrs with hysterectomy ± oophorectomy and/or documented FSH) with a mean age of 57.7 years (range 40-76) and mean BMI 27.2. The majority were Caucasian (96.7%), mean estradiol was 0.98 ng/dL (range 0.5-17.0) and 25.5% had an estradiol ≥ 0.9 ng/dL. 72.4% experienced natural menopause (27.5% surgical) and 28.9% had an oophorectomy. 33.5% were breast cancer survivors. For the overall cohort and for women without a breast cancer history the high estradiol group was significantly associated with younger age <57 (p=0.0025), BMI ≥ 26 (p=0.0080), natural menopause (p=0.0379), and no oophorectomy (p=0.0017). These relationships did not hold for women with a breast cancer history. No significant relationship was found between estradiol and family history or Gail score. Conclusion In our cohort of postmenopausal women considering enrollment in a Phase II chemoprevention trial, estradiol levels were related to BMI, younger age and natural menopause, and not to Gail score, which incorporates several hormonal risk factors. Estradiol level was not explored in this analysis for its potential as an independent risk predictor. However, other data suggest that it could supplement existing risk models. Issues in measurement of estradiol and other estrogens at postmenopausal levels and other potential confounders must be explored further in larger datasets. Our intervention trial will evaluate the effect of an aromatase inhibitor in postmenopausal women with higher estradiol levels.