In a previous study of a series of 105 patients with primary breast cancer we found that the progesterone receptor (PgR) status was an important prognostic factor for early recurrences. 95 patients from the same series were followed-up for a median of 9.5 years and reassessed for the prognostic value of PgR status by univariate and multivariate statistical methods. In univariate analysis, the disease-free interval was only related to the lymph-node status. For overall survival, PgR and combined PgR-ER (oestradiol receptor) status had a prognostic value (P = 0.035 and 0.05, respectively). Moreover, PgR status was found to be discriminant for the survival of the node-negative patients (P = 0.017). In multivariate analysis, ER and PgR status were not significant, indicating that receptor status is not a powerful predictor of the course of breast cancer.