Background: Dietary vitamin D has been associated with lower mammographic breast density, a strong biomarker for breast cancer risk. Blood 25-hydroxyvitamin D [25(OH)D] is an integrated measure of vitamin D status (from food, supplements, and sun exposure) and varies with season. Our objective was to assess seasonal variations of breast density and compare such variations, if any, with that of 25(OH)D.
Methods: This cross-sectional study includes 741 premenopausal women recruited at screening mammography. Plasma 25(OH)D at recruitment was measured by RIA. Breast density was evaluated using a computer-assisted method. Seasonal variations were modeled using multivariate linear regression and semi-parametric cubic smoothing splines.
Results: Season was strongly associated with 25(OH)D (P < 0.0001). The highest smoothed mean 25(OH)D levels were seen at the end of July (81.5 nmol/L) and the lowest in mid-April (52.4 nmol/L). Breast density showed modest seasonal variations (P = 0.028). The lowest smoothed mean breast density was observed in early December (38.5%) and the highest at the beginning of April (44.3%). When a 4-month lag time was presumed, seasonal variations of breast density appeared to be a mirror image of those of 25(OH)D, and the correlation of daily smoothed estimates of mean breast density and 25(OH)D was negative and strong (r = -0.90).
Conclusion: In premenopausal women, changes in blood vitamin D seem to be inversely related to changes in breast density with a lag time of about 4 months. This finding encourages further investigation of the possibility that vitamin D could reduce breast density and breast cancer risk.