Capecitabine 1000 mg/m(2) bid × 14 days every 21 days (14/21) has been reported to have similar efficacy but more favorable toxicity profile than the approved dosage of 1250 mg/m(2). However, a dose-toxicity relationship of capecitabine in breast cancer patients has not been fully elucidated. We performed a systematic review and meta-analysis to compare a safety profile between capecitabine starting dose of 1000 and 1250 mg/m(2) bid. Studies were identified using PubMed, ASCO, and San Antonio Breast Cancer Symposium abstract databases through December 2015. Eligible trials included phase II/III trials of capecitabine monotherapy at 1000 or 1250 mg/m(2) bid (14/21) for breast cancer patients that reported adequate safety data for all (grade 1-4) or high (grade 3-4) grade hand foot syndrome (HFS), diarrhea, fatigue, nausea, vomiting, stomatitis, neutropenia, thrombocytopenia, or anemia, as well as dose reductions, treatment discontinuation or treatment-related deaths. The summary incidence was calculated using random-effects models. A total of 4833 patients from 34 trials were included. 1218 and 3615 patients were treated with capecitabine 1000 and 1250 mg/m(2) bid, respectively. A significantly lower incidence of dose reduction (15.9 vs. 39.0 %; P = 0.007), high-grade HFS (12.0 vs. 19.0 %; P = 0.01), diarrhea (5.3 vs. 9.1 %; P = 0.01), and neutropenia (1.8 vs. 7.3 %; P < 0.01), and all-grade neutropenia (5.8 vs. 25.4 %; P = 0.01) was seen in capecitabine 1000 mg/m(2) compared to 1250 mg/m(2). Capecitabine monotherapy at 1000 mg/m(2) bid (14/21) has a clinically meaningful and significantly better toxicity profile compared to 1250 mg/m(2) bid (14/21).