OBJECTIVE: To examine the clinical effectiveness and cost-effectiveness of oral capecitabine for locally advanced and metastatic breast cancer in relation to its licensed indications. DATA SOURCES: Twenty-three electronic databases and other databases of ongoing research and Internet resources, bibliographies of retrieved articles and industry submissions. REVIEW METHODS: Two reviewers independently screened and assessed all titles and/or abstracts including economic evaluations. Randomised controlled trials (RCTs) and observational studies that investigated capecitabine monotherapy, in patients pretreated with an anthracycline-containing regimen or a taxane, or capecitabine in combination with docetaxel, in patients pretreated with an anthracycline-containing regimen, were included. The economic evaluation was based on data reported in the manufacturer's submission. RESULTS: For capecitabine monotherapy, 12 uncontrolled observational studies were identified. The methodological quality of the studies was low. Capecitabine demonstrated antitumour activity, but was associated with a particular risk of hand-foot syndrome and diarrhoea. Economic evaluation was hampered by the poor quality of the published studies, but compared indirectly with vinorelbine, capecitabine was associated with lower costs and improved patient outcomes. For capecitabine in combination with docetaxel, one RCT was identified. Combination therapy was superior to single-agent docetaxel in terms of survival, time to disease progression and overall response. Adverse events occurred more frequently with combination therapy. The economic evaluation demonstrated an overall improved QALY score for combination therapy with a slight reduction in costs. CONCLUSIONS: No conclusions could be drawn regarding the therapeutic benefit of capecitabine monotherapy; RCTs are required. Capecitabine appeared cost-effective compared with vinorelbine, but serious doubts remain; the poor quality of the trials may invalidate this conclusion. Based on limited evidence, combination therapy was more effective than single-agent docetaxel and likely to be cost-effective, but was associated with higher incidences of hand-foot syndrome, nausea, diarrhoea and stomatitis.