We assessed whether progesterone receptor (PgR) and Ki67 in primary tumors and/or matched metastases are predictors of clinical benefit from first-line endocrine therapy (ET) in advanced breast cancer. We evaluated patients treated at our institute with first-line ET (2002-2011), excluding those receiving concomitant chemotherapy or trastuzumab or pretreated with >2 lines of chemotherapy. A cut-off of 20 % immunostained cells was used for PgR and Ki67. The main endpoint was time-to-progression (TTP). Groups were compared by the log-rank test and Cox multivariate analysis. In the 135 assessable patients (93 % were receiving an aromatase inhibitor; biomarker assessment had been performed on primary tumors in 77 cases, on metastases in 23 and on both in 35), median TTP was 16 months (median follow-up 43 months). The overall discordance rate between primary tumors and metastases was 23 % for Ki67 and 31 % for PgR. A longer median TTP (24 vs. 12 months, P = 0.012) was seen for PgR >20 % in metastases. Ki67 showed a trend for TTP prediction in the entire case series (P = 0.062). Patients with high Ki67 and low PgR in metastases had a median TTP of only 5 months. High Ki67 in primary tumors (P = 0.026) or metastases (P = 0.01) predicted disease progression at the first evaluation. PgR in metastases remained a significant independent predictor of TTP at multivariate analysis (HR 2.45). In an ER-high population, PgR >20 % in metastases identified patients with a long TTP on endocrine treatment, while Ki67 >20 % was associated with an increased risk of non-response.