A number of case-control studies were conducted to investigate the association of IL6 gene polymorphisms with colorectal cancer (CRC). However, the results were not always consistent. We performed a systematic review and meta-analysis to examine the association between the IL6 gene polymorphisms and CRC. Data were collected from the following electronic databases: PubMed, EMBASE, Web of Science, BIOSIS Previews, HuGENet, and Chinese Biomedical Literature Database, with the last report up to July 2011. A total of 17 studies involving 4 SNPs were included (16 for rs1800795, 2 for rs1800796, 2 for rs1800797, and 1 for rs13306435). Overall, no significant association of these polymorphisms with CRC was found in heterozygote comparisons as well as homozygote comparison, dominant genetic model and recessive model. In subgroup analysis, among studies using population-based controls, fulfilling Hardy-Weinberg equilibrium, or using Taqman genotyping method, we did not find any significant association. However, the rs1800795 C allele was significantly associated with reduced risk for CRC among persons who regularly or currently took NSAIDs (four studies, OR = 0.750; 95 % CI, 0.64-0.88; P = 0.474 for heterogeneity test), and with increased risk for CRC among persons who drank (one study, OR = 1.97; 95 % CI, 1.32-2.94). Individuals with the rs1800795 C allele in the IL6 gene have a significantly lower risk of CRC, but in the setting of NSAIDs use. Further studies are merited to assess the association between the IL6 gene polymorphisms and CRC risk among persons who take NSAIDs, drink or smoke, etc.