Nutritional factors have been suggested to play an important role in the prognosis of breast cancer through their effect on tumor characteristics. This study evaluated four tumor characteristics and prognosis in relation to premorbid diet and body size. From a cohort of 89,835 women in the National Breast Screening Study (NBSS) in Canada, data on 676 incident cases of invasive carcinoma of breast, on whom we had dietary information, were used. A high energy intake lowered the likelihood of being ER positive and PR positive but after adjusting for ER status, was still associated with a higher risk of dying of breast cancer. Total fat and various types of fats were associated with a greater likelihood that a woman would be ER and PR positive, however the likelihood of dying from breast cancer was higher with higher fat consumption. There was no significant effect of higher intakes of beta carotene or vitamin C on ER status, nodal status or tumor size, but a significantly lower risk of dying from breast cancer was observed. Higher intake of carbohydrates and calcium was associated with a lowered frequency of ER and PR positive status but also with a lower risk of dying. Of the five indicators of body size studied, higher triceps skinfold thickness was associated with a slightly lower chance of being ER positive, PR positive, and node negative, and a significantly higher likelihood of dying. It appears that while there are significant associations between some of the diet and body size variables and tumor characteristics, the effect of most nutritional factors on prognosis in breast cancer may not be mediated via their effect on tumor characteristics.