Dietary exposures and other factors of possible prognostic significance in relation to tumour size and nodal involvement in early-stage breast cancer.

In this case series study of 546 early-stage breast cancer patients we examined the relationship between tumour size and axillary nodal involvement separately in relation to other factors suggested to be of prognostic significance in breast cancer. Relationships were modelled using two-stage general linear models stratified by menopausal status. In stage I models, tumour size (diameter in cm) and nodal involvement, expressed as the loge (number of positive nodes total nodes examined), were regressed on an array of socio-demographic and reproduction-related variables as well as smoking and drinking. In stage II models we entered the same dependent variables, the linear predictor score from stage I modelling, estimated total calories consumed and calorie-adjusted fat, fibre and vitamin A intake, as well as serum cholesterol, body mass index, and oestrogen and progesterone receptor status. Using this very conservative analytical technique, we found that, among premenopausal women, for every gram increase in estimated calorie-adjusted daily fat intake there was a 0.09 mm increase in tumour diameter (T = 2.58, P = 0.01). Also among premenopausal women, we observed an inverse relationship between nodal involvement and fibre intake (g/d) and nodal involvement (B = -0.058, T = -2.11, P = 0.04). For postmenopausal women we observed an inverse relationship with serum cholesterol (mg/dl) (B = -0.005, T = -1.85, P = 0.07). Other findings of interest include the absence of background factor effects in premenopausal women, the lack of nutritional effects and the consistency of a protective effect of education among postmenopausal women, and a relationship between hormone receptor levels and tumour size among premenopausal women.(ABSTRACT TRUNCATED AT 250 WORDS)