Postprint version. Published in Cancer Causes Control, Volume 19, January 1, 2008, pages 481-489.
NOTE: At the time of publication, the author Marilyn Tseng was not yet affiliated with Cal Poly.
The definitive version is available at https://doi.org/10.1007/s10552-008-9109-x.
Objective Whether dietary patterns, rather than single foods or nutrients, are associated with breast density is not known. We investigated this in the Minnesota Breast Cancer Family Study. Methods Participants completed a 153-item food frequency questionnaire and provided screening mammograms for breast density assessment using a computer-assisted method. We used multivariate linear regression to quantify dietary pattern–breast density associations. Results Among 3,147 women with dietary information, three dietary patterns emerged from principal components analysis: a fruit–vegetable–cereal pattern, a salad–sauce–pasta/ grain pattern, and a meat–starch pattern. Among 1,286 women with breast density estimates, the fruit–vegetable–cereal and salad–sauce–pasta/grain patterns were inversely associated with percent breast density only in stratified analyses. The fruit–vegetable–cereal pattern was inversely associated with breast density among premenopausal women (b = -0.13, p = 0.09; interaction p = 0.009) and current smokers, (b = -0.30, p = 0.02; interaction p = 0.05), while the salad–sauce–pasta/grain was inversely associated with breast density among current smokers (b = -0.27, p = 0.06; interaction p = 0.006). Conclusion Overall our results do not provide strong evidence for associations of dietary patterns with breast density. Suggestive inverse associations for the fruit–vegetable– cereal and salad–sauce–pasta/grain dietary patterns among smokers are consistent with previous reports and leave open the possibility that some dietary patterns influence breast density in population subsets. Nevertheless, these findings require confirmation, and their underlying reasons have yet to be clarified.