The prognostic value of p53 protein accumulation in breast cancer, especially as detected by methods other than immunohistochemistry, has not been established unequivocally. A sensitive immunofluorometric assay of p53 protein employing DO-1 and CM-1 antibodies was used in this study to assay extracts of 171 breast carcinomas from northern Italy. p53 over-expression, demonstrated in 36 (21%) tumours, was associated with lack of oestrogen receptor (ER) expression but was not related to patient age, stage, lymph node status, tumour size, histologic type, grade or progesterone receptor (PR) expression status in contingency tables. An increased risk for cancer relapse of p53-positive patients compared to p53-negative patients was determined using multivariate Cox regression analysis, which also showed that p53 protein over-expression was an independent predictor of reduced disease-free survival in node-positive and ER+ patients but not in node-negative or ER- individuals. The equivalent analysis for assessing the impact of p53 status on overall survival was not statistically significant, possibly reflecting the short patient follow-up. Our results suggest that an immunoassay of p53 protein, applicable to cytosolic extracts prepared for steroid hormone receptor analyses, may provide information for breast cancer prognosis.