The selection of chemotherapeutic regimens is challenging for metastatic breast cancer (MBC) patients whose diseases have failed to respond to anthracyline and taxane. Capecitabine has advantages of oral administration and favorable toxicity profiles. This study was conducted to evaluate the efficacy of capecitabine and to identify the subgroup of patients who would potentially have benefit from capecitabine monotherapy in patients with anthracycline- and taxane-pretreated MBC. Female patients with MBC who had been previously treated with anthracycline and taxane received oral capecitabine 2500 mg/m(2) divided in two doses daily for 2 wk with 1-wk rest period. Between September, 1999, and December, 2002, a total of 38 patients were enrolled. Among the 36 evaluable patients, one patient achieved a complete response (CR), 9 patients had partial responses (PRs), and 13 patients had stable diseases (SDs). Response rate was 26% [95% confidence interval (CI), 12-40%] and the tumor control rate (TCR, CR+PR+SD) was 61% (95% CI, 45-77%). The median follow-up duration was 27.8 mo. The median duration of response was 8.9 mo, the median time to progression was 4.6 mo, and the median overall survival was 18.1 mo. The major toxicities were hand-foot syndrome, diarrhea, and emesis. There was no treatment-related death. The predictors of better overall survival were positivity of hormone receptor, disease-free survival longer than 1 yr, non-refractoriness to anthracycline, and fewer number (<or= 3) of involved organs. Capecitabine monotherapy is effective and well tolerated for MBC patients who had previously been treated with anthracycline and taxane. The TCR could predict overall survival as well as the objective response in this study, suggesting a possible role of TCR as a surrogate marker for survival in MBC patients on salvage chemotherapy. The patients who have relatively slow growing tumor and less tumor burden could have benefit from capecitabine monotherapy following anthracycline- and taxane-based chemotherapy.