Body composition and the breast cancer risk associated with dense breasts: a prospective cohort study of 106,813 predominantly premenopausal Korean women

Abstract Background Mammographic density is a strong breast cancer risk factor, yet higher body mass index is associated with lower density but higher postmenopausal risk. Because body mass index combines fat and lean mass, it is unclear which compartment underlies these opposing associations. It is also unknown whether a dense breast carries less risk when its density reflects little fat rather than abundant fibroglandular tissue, a question relevant to East Asian women, who often have dense breasts at low body mass index. Methods We followed 106,813 Korean women in a health-examination cohort by registry linkage (86% premenopausal; mean 6.5 years; 891 invasive breast cancers). Fat and skeletal muscle mass were measured by bioelectrical impedance and modelled jointly against an extremely dense breast (BI-RADS D versus A–C; recorded for 76,551 women, 713 cancers) and breast cancer. We tested whether the density–breast cancer association, adjusted for age, body mass index, waist circumference, and reproductive and lifestyle factors, varied across quartiles of fat mass or skeletal muscle percentage (muscle mass/body weight), or by menopausal status. Results Per SD, fat mass was strongly associated with lower odds of an extremely dense breast (odds ratio 0.41; 95% CI 0.40–0.42) and skeletal muscle mass weakly with higher odds (1.03; 1.01–1.05). Neither was significantly associated with breast cancer (hazard ratio [HR] 0.98; 0.90–1.05 and 1.07; 1.00–1.16). An extremely dense breast was associated with breast cancer (HR 1.46; 1.23–1.73). If density merely reflected little fat, the association should be weakest in women with the least fat and most muscle; instead, it was strongest in the lowest fat mass quartile (HR 1.74) and highest skeletal muscle percentage quartile (1.67), although not significantly (interaction P = 0.15 and 0.13). The association was stronger after than before menopause (HR 2.93 versus 1.38; interaction HR 2.66, 1.54–4.62; 69 postmenopausal cancers). Conclusions In this predominantly premenopausal cohort, body fat was inversely associated with density but not with breast cancer, and an extremely dense breast carried no less risk when reflecting low body fat. Although a modest modification cannot be excluded, density appears a valid risk marker across body habitus in East Asian women.