Capecitabine/taxane combinations are highly active in metastatic breast cancer (MBC). We conducted a randomized, phase III, noninferiority trial comparing capecitabine plus paclitaxel (XP) with epirubicin plus paclitaxel (EP) as first-line therapy for MBC, regarding progression-free survival (PFS) as primary efficacy endpoint. Females who had received no prior chemotherapy for MBC were randomized to six 3-weekly cycles of XP (capecitabine 1000 mg/m(2) b.i.d., days 1-14; paclitaxel 175 mg/m(2) 3-h infusion, day 1) or EP (epirubicin 60 mg/m(2) 1-h infusion, day 1; paclitaxel as above). Secondary endpoints included response rate, overall survival, tolerability, and quality of life (QoL). Each arm included 170 patients, most of whom received all six cycles as planned. The difference in means of (logarithmic) PFS times (-0.205) did not meet the pre-defined level for noninferiority (-0.186). However, PFS was similar in the two arms [HR: XP vs. EP: 1.012 (95 % CI 0.785-1.304); median 10.4 months XP vs. 9.2 months EP]. Overall survival was also similar [HR 1.027 (95 % CI 0.740-1.424); median 22.0 vs. 26.1 months], and response rate was 47 % versus 42 %. Both regimens were tolerable: there were more grade 3/4 diarrhea and grade 3 hand-foot syndromes with XP and more grade 3/4 hematologic toxicities with EP. There were no major differences in QoL. Although, noninferiority of XP to EP was formally not proven, first-line XP was active and feasible. XP is a valid first-line alternative to anthracycline/taxane regimens, especially in patients previously treated with adjuvant anthracyclines.