The biological mechanisms underlying the relationship between mammographic density and breast cancer risk are unknown. Our objective was to examine the association between mammographic density and circulating factors that are putative breast cancer intermediate endpoints. Biologic data from a year-long aerobic exercise intervention trial conducted in 302 postmenopausal women aged 50-74 years were analyzed. Sex hormones, markers of glucose homeostasis, inflammatory markers, and adipokines were assayed in fasting blood drawn at baseline and after 1 year. Area and volumetric measurements of mammographic dense fibroglandular and nondense fatty tissue were made. Multiple linear regression was used to examine the association between the circulating factors and mammographic measures and partial correlations were estimated. Mammographic nondense volume was positively correlated with concentrations of estradiol (r = 0.28), estrone (r = 0.13), insulin (r = 0.41), glucose (r = 0.15), leptin (r = 0.49), and C-reactive protein (r = 0.22), and negatively correlated with sex hormone binding globulin (r = -0.30) and adiponectin (r = -0.12) but correlations became null after adjustment for overall body adiposity as represented by body mass index and waist circumference. With adjustment for overall adiposity, mammographic dense volume, a measure that represents fibroglandular tissue, was negatively correlated with leptin (r = -0.19) and C-reactive protein (r = -0.19). As expected, circulating factors originating from or correlated with adipose tissue were also correlated with mammographic measures of breast adipose tissue, but not after adjustment for overall body adiposity. Interpreting correlations between adiposity-derived factors and mammographic measures whose validity may be affected by adiposity is problematic. To rectify this problem, future studies with very good measures of the volume of fibroglandular tissue in the breast will be necessary.
Trial registration: ClinicalTrials.gov NCT00522262.