Soluble and nuclear estrogen receptor (ER) content was measured by ligand binding assay, and estrogen and progesterone receptors by immunohistochemical assays (ER-ICA and PR-ICA) in 214 patients with breast cancer recruited at the "M. Ascoli" Cancer Hospital Centre in Palermo, Sicily, to assess the discriminant and predictive value of these parameters. On follow-up, data from both ER-ICA and PR-ICA showed a statistically significant difference, PR-positive patients having longer disease-free (DSF) and overall (OS) survival than PR-negative ones. Conversely, ER status did not correlate significantly with both DFS (P = 0.6) and OS (P = 0.2). In particular, PR-positive patients had 59 +/- 18 months DFS and 67 +/- 12 months OS, compared to 51 +/- 22 months DFS and 57 +/- 17 months OS of PR-negative cases. The present evidence implies that a PR-negative status identifies breast cancer patients with early relapse, as also suggested by previous studies. It also agrees with the results of ligand binding assay of ER, where ER status is a good discriminant and predictor of response to endocrine treatment, but is unable to anticipate early relapse in breast cancer patients. Evidence that PR status is a statistically significant prognostic indicator deserves further study to ascertain whether or not PR should be regarded as an ER-dependent parameter or be related to other biological variables such as growth factor (e.g., EGF), oncogene (e.g., Her2/Neu), or tumor suppressor gene (e.g., p53) products.