X-ray repair cross-complementing protein 3 (XRCC3) belongs to DNA double-strand break repair pathway and XRCC3 rs861539 (C > T) polymorphism has been suspected with lung cancer risk. However, results from previous studies are inconclusive and affected by bias. Electronic databases of PubMed, EMBASE, China National Knowledge Infrastructure, and SinoMed were searched. References of relative reviews were also screened. Pooled odds ratios (ORs) and 95 % confidence intervals (CIs) were calculated to estimate the association strength. A number of 18 eligible studies with 6 studies of Asians, 11 of Caucasians, and 1 of African were extracted and analyzed, including 4,896 lung cancer cases and 6,360 controls. No significant correlation between XRCC3 polymorphism and lung cancer risk was observed in homozygote comparison (CC vs. TT; OR=0.877; 95 % CI, 0.659, 1.168), heterozygote comparison (CT vs. TT; OR=0.857; 95 % CI, 0.675, 1.089), dominant model (CC/CT vs. TT; OR=0.862; 95 % CI, 0.663, 1.123), or recessive model (CC vs. CT/TT; OR=1.047; 95 % CI, 0.956, 1.145). Subgroup analyses of ethnicity and controls did not reveal any significant association with lung cancer risk. No publication bias was detected. In this update meta-analysis of 18 studies and 11,256 participants, we find that XRCC3 rs861539 polymorphism does not contribute to lung cancer risk and there is no difference between Asians and Caucasians.