Estrogen receptor (ER) and/or progesterone receptor (PR) expression has been associated with more favorable breast cancer prognosis. Results on the differential association of diet with ER and/or PR positive and negative tumors have been inconclusive. In a large case-control study conducted in Athens, Greece, we investigated whether diet is associated with the expression of ER-alpha or PR in mammary tumors of 421 women with histologically confirmed breast cancer. Diet was assessed through an extensive food frequency questionnaire and results were analyzed using multiple logistic regression. After controlling for non-nutritional variables and mutually adjusting for energy-generating nutrients and ethanol, carbohydrate intake was inversely associated with ER-alpha (P = 0.04) and PR (P = 0.10) expression. The odds ratios (OR) per one standard deviation increment were 0.69 with 95% confidence interval (95% CI) 0.48-0.98 for ER-alpha and 0.72 (95% CI 0.49-1.07) for PR expression. No consistent or statistically significant associations were noted for any of the other energy-generating nutrients or food groups examined. Although in these data no strong relations of qualitative aspects of diet with hormone receptor expression in breast cancer tumors were evident, the inverse association of carbohydrate intake with ER-alpha, and perhaps PR, expression merits further study in future investigations.