Central obesity and breast density in premenopausal US Chinese women Free

B96

Previous studies suggest that central adiposity increases breast cancer risk in premenopausal women. But overall adiposity, strongly correlated with central adiposity, is inversely associated with breast cancer risk in premenopausal women, and inversely associated with breast density (BD), a strong risk factor for breast cancer. How central adiposity is associated with BD independently of overall adiposity is unclear. Objective. We examined the association of waist circumference (WC), a measure of central adiposity, with BD independent of body mass index (BMI) as a measure of overall adiposity in a sample of premenopausal US Chinese women. Methods. From 11/05-10/07, we recruited 386 premenopausal Chinese women in the Philadelphia region into a longitudinal breast density study. Baseline data collection included questionnaires, anthropometry, fasting serum samples, and mammograms. In preliminary analyses, BD was assessed in 300 women using a computer-assisted method. Anthropometry was available for 290 women. We used multivariate linear regression to quantify associations between anthropometric measures and BD. Because WC and BMI were highly correlated, we created a variable representing WC adjusted for BMI using the residual method. Results. Mean (±SD) WC in our sample was 79.6±7.7 cm, mean BMI was 23.6±2.9 kg/m2, and mean %BD was 47.1±16.3%. With adjustment for age, menopausal stage, live births, and BMI, WC was significantly associated with non-dense breast tissue size (p=0.004) and significantly inversely associated with %BD (p=0.01), but it was not associated with dense breast tissue size (p>0.9). Central adiposity has been hypothesized to increase risk by increasing insulin resistance or by reducing adiponectin levels. However, in regression models including 194 women with available insulin, glucose, and adiponectin results, insulin resistance estimated as a homeostasis model assessment (HOMA) score was inversely rather than positively associated with %BD (p=0.15) - an effect driven by its weak correlations with BMI (r=0.20, p=0.01) and BMI-adjusted WC (r=0.13, p=0.07). Adiponectin level was positively rather than inversely associated with %BD (p=0.02) - an effect driven by its strong inverse correlation with BMI (r=-0.30, p<0.0001). Conclusion. Despite a hypothesized association between central adiposity and breast cancer risk, WC was not associated with dense area and was inversely associated with %BD in our sample, even after adjustment for BMI. Our findings suggest that central adiposity has effects on BD that are similar to but independent of overall adiposity; that insulin resistance and adiponectin level are not important determinants of BD; and that if central adiposity increases breast cancer risk, it is through a mechanism without direct effects on breast cell proliferation as indicated by mammographic density.