Lifetime changes in body fatness and breast density in postmenopausal women: the FEDRA study

Abstract

Background: High mammographic breast density (MBD) is an established risk factor for breast cancer (BC). Body fatness conveys an increased BC risk in postmenopause but is associated with less dense breasts. Here, we studied the relationship between body fatness and breast composition within the FEDRA (Florence-EPIC Digital mammographic density and breast cancer Risk Assessment) longitudinal study.

Methods: Repeated anthropometric data and MBD parameters (obtained through an automated software on BC screening digital mammograms) were available for all participants, as well as information on other BC risk factors. Multivariate linear regression and functional data analysis were used to longitudinally evaluate the association of body fatness, and changes thereof over time, with dense (DV) and non-dense (NDV) breast volumes and volumetric percent density (VPD).

Results: A total of 5,262 women were included, with anthropometric data available at 20 and 40 years of age, at EPIC baseline (mean 49.0 years), and an average of 9.4 years thereafter. The mean number of mammograms per woman was 3.3 (SD 1.6). Body fatness (and increases thereof) at any age was positively associated with DV and NDV (the association being consistently stronger for the latter), and inversely associated with VPD. For instance, an increase by 1 kg/year between the age of 40 years and EPIC baseline was significantly associated with 1.97% higher DV, 8.85% higher NDV, and 5.82% lower VPD.

Conclusion: Body fatness and its increase from young adulthood until midlife are inversely associated with volumetric percent density, but positively associated with dense and non-dense breast volumes in postmenopausal women.

Fig. 1

Function-on-Scalar models estimating, among 390 women from the FEDRA study in the age interval 63–68, the effect (β(t)) of the body mass index (BMI, kg/m²) over a time-varying function of: A the mammographic volumetric percent density; B the breast dense volume; C the breast non-dense volume. In Model 1 the BMI at age 20 was considered. In Model 2, the BMI at age 40 was considered. A positive value of β(t) indicates a positive association between BMI and mammographic density measures and vice versa. All the models were adjusted for menopausal status, age at menarche and birth index. The 95% confidence intervals were obtained through 500 bootstrap sampling