Mammographic densities and circulating hormones: a cross-sectional study in premenopausal women.

Progestogens appear to influence breast density more than estrogens in postmenopausal women taking hormone replacement therapy (HRT), but little is known about the effect of circulating hormones on mammographic density among premenopausal women. This cross-sectional study explores the relationship of body weight and sex steroids with breast density. Luteal serum samples were analyzed for progesterone, estrone, estradiol, and sex hormone-binding globulin (SHBG). Mammograms were assessed for density using a computer-assisted method. We performed mediation tests using multiple linear regression models. Significant associations of SHBG and estradiol with percentage density disappeared after adjustment for body weight and other covariates, whereas the relationship between progesterone and breast density remained borderline significant. The mediation tests indicated that progesterone has a direct and an indirect effect on mammographic density. Our finding that progesterone shows a stronger association with percentage of mammographic density than estrogen agrees with clinical reports describing denser mammographic patterns among women taking HRT, although these women differ in menopausal status.

The effect of migration on cancer incidence among Japanese in Hawaii.

OBJECTIVES: This analysis compared cancer incidence trends among Japanese in Japan, and Japanese and Caucasians in Hawaii, between 1960 and 1997, and estimated the impact of migration on the incidence of different cancers. METHODS: Incidence information was obtained from 8 volumes of Cancer Incidence in Five Continents. The migration effect was estimated from the areas under the incidence curves as the ratio of the geographic and the ethnic difference in cumulative cancer incidence. RESULTS: Among the 5 more common cancers, the migrant effect was strongest for colon and stomach cancers, prostate and breast cancers were affected to a lesser degree, and lung cancer risk differed little between Japanese in Japan and Hawaii. Migration led to lower risk of stomach, esophageal, pancreatic, liver, and cervical cancers, but to higher rates for all other cancers. The large variation in time for migrants to adopt the host population's cancer risk suggests that risk factors have organ-specific effects, or operate at different times in life. Although the available incidence rates are limited by under-reporting and early detection efforts, mortality rates confirm the significant differences in cancer risk. CONCLUSIONS: The persistent difference in cancer incidence several generations after migration supports the idea that living in the host country is not, alone, sufficient to modify cancer risk for all cancer sites to the level of the host population. Although the migration effect can be partially explained by known etiologic factors, a large proportion of the changing risk remains unexplained.