Nutritional factors related to breast cancer were investigated by means of a hospital-based case-control study in Milan (Italy) and Montpellier (France). Liposoluble vitamins, cholesterol and triglycerides were measured in blood samples taken from interviewed subjects (319 cases and 344 controls). In addition serum zinc and copper was assessed in the Italian samples and serum fatty acids and malonyl-di-aldehyde in the French samples. A significant difference was found between cases and controls in total fat and cholesterol intake in both populations, and in saturated fatty acid and mono-unsaturated fatty acid consumption in the French samples. No difference emerged in liposoluble vitamin consumption in both populations nor in zinc and copper consumption in the Italian samples. A statistically significant higher serum level of cholesterol and plasma level of vitamin E was observed in cases compared to controls in both populations. The difference in plasma vitamin E was confirmed after adjustment on total cholesterol and triglycerides. Similarly, zinc serum level was higher in Italian cases than in Italian controls, while malonyl-di-aldehyde was lower in French cases than French controls. A multivariate analysis was performed after classification of cases and controls according to quantile distribution of controls. Nutrient consumption and relevant blood markers were directly or partially correlated in both populations. All known risk factors plus age, serum total cholesterol and triglycerides were used as covariates. The odds ratio values for the highest quantiles are: Dietary cholesterol, OR = 1.9 (1.1-3.4); total dietary lipids, OR = 1.9 (1.0-3.4); plasma vitamin E, OR = 4.2 (1.9-9.0); serum zinc, OR = 12.2 (5.4-27.7); serum malonyl-di-aldehyde, OR = 0.56 (0.33-0.97).