Estrogen receptor (ER) and progesterone receptor (PgR) levels have been reported to have prognostic significance with respect to disease-free survival in early-stage breast cancer patients. The current retrospective study was undertaken to determine whether ER and PgR levels, as well as other potential prognostic factors, might be related to a progression-free interval (PFI) during additive hormonal therapy in advanced-stage breast cancer patients. Eligibility requirements for this study included the following: histologically confirmed recurrent or metastatic breast cancer, known quantitative ER and PgR levels, postmenopausal status, treatment with either megestrol acetate or tamoxifen, and Eastern Cooperative Oncology Group (ECOG) performance status less than or equal to 2. The characteristics of the 105 patients included in these analyses were as follows: median age, 62 years; median disease-free interval (DFI), 523 days; median ER level, 44 fmol/mg; median PgR level, 52 fmol/mg; soft tissue-dominant disease, 37 patients (35%); bone-dominant disease, 36 patients (34%); visceral-dominant disease, 32 patients (31%); one site of disease, 60 patients (58%); two or more sites of disease, 45 patients (42%); treatment with megestrol acetate, 62 patients (59%); treatment with tamoxifen, 43 patients (41%). All of the independent variables listed immediately above were included in a multiple linear regression analysis in which PFI, expressed as log PFI, was the dependent variable. In this analysis, a positive linear relationship was observed between log PFI and the following independent variables: log ER, log PgR, and age (r2 = 0.329). An alternative model (r2 = 0.350) was derived, in which previous treatment with chemotherapy was negatively related to log PFI. However, it appears that previous treatment with chemotherapy could be a "proxy variable," because patients who had been treated with chemotherapy previously were significantly younger and had significantly lower ER (P = 0.0001) and PgR levels (P = 0.0004). None of the other independent variables were included in these models. If the assumption that PFI is a measure of the effectiveness of hormonal therapy is true, these results suggest that quantitative ER and PgR levels and age supersede other traditional predictor variables in predicting the hormonal responsiveness of individual breast carcinoma.(ABSTRACT TRUNCATED AT 400 WORDS)