High mammographic breast density (H-MBD) has been associated with increased breast cancer (BC) risk, even after adjustment for established BC risk factors. Only a few studies have examined the influence of diet on MBD. In a longitudinal study in Florence, Italy, we identified about 2,000 women with a mammogram taken 5 years after enrollment, when detailed information on dietary and lifestyle habits and anthropometric measurements had been collected. Original mammograms have been identified and retrieved (1,668; 83%), and MBD was assessed by 2 experienced readers, according to Wolfe's classification and a semiquantitative scale. By logistic analysis, we compared women with H-MBD (P2 + DY according to Wolfe's classification) with those with low-MBD (N1 + P1). H-MBD was confirmed to be inversely associated with BMI, number of children and breast feeding, while it was directly associated with higher educational level, premenopausal status and a previous breast biopsy. In multivariate analyses adjusted for non-dietary variables, H-MBD was inversely associated with increasing consumption of vegetables (p for trend = 0.005) and olive oil (p for trend = 0.04). An inverse association was also evident between H-MBD and frequent consumption of cheese and high intakes of beta-carotene, vitamin C, calcium and potassium (p for trend < or = 0.05). On the other hand, we found a positive association with increasing consumption of wine (p for trend = 0.01). This large longitudinal study, the first carried out in Mediterranean women, suggests that specific dietary components may play a key role in determining MBD in this population, thus possibly modulating BC risk.