OBJECTIVES: To determine whether surrogate end points [progression free survival (PFS), time to progression (TTP) and response rate (RR)] are correlated with overall survival (OS) in the first-line treatment of advanced or metastatic breast cancer (BC). METHODS: A systematic review of the literature was conducted to indentify randomized clinical trials (RCTs) that evaluate the efficacy of chemotherapy in first-line treatment of advanced or metastatic BC.  Searches were realized in MEDLINE and EMBASE databases from 1995 to April 2010. The nonparametric Spearman rank correlation coefficient (rs) was used as a measure of correlation between the difference (Δ) in surrogate outcomes (ΔPFS, ΔTTP and ΔRR) and the difference in OS (ΔOS). Correlation coefficients were compared using the normal approximation to the z-transformation of rs and its standard deviation. Linear regression analysis, through the origin of the plot, evaluating ΔOS as a function of differences in each surrogate outcomes was used to determine the proportion of variability explained (R2). Statistical analyses were performed using STATA software v.10. RESULTS: Thirty-four RCTs were included in the analysis, with a total of 11,398 patients evaluated. In the first-line therapy of advanced or metastatic BC, there was a weak significant association between ΔPFS and ΔOS [rs: 0.43 (Confidence Interval (CI) 95%: 0.04-0.71)]. When the analysis was performed including only RCTs in the metastatic stage, the rs data between ΔPFS and ΔOS increased statistically significant to 0.59 (CI95% 0.17 to 0.83). The surrogate outcomes that correlated better with the ΔOS, were ΔTTP [rs: 0,79 (CI95%: 0,43-0,94); R2:62%], and ΔRR [rs: 0,73 (CI95%: 0,55-0,85); R2: 53%]. CONCLUSIONS: In the first-line treatment of advanced or metastatic BC, TTP and RR may be appropriate surrogate end points for OS, although it is important to consider the magnitude of their variations.