Genetic polymorphisms of cyclooxygenase-2 (Cox-2) gene have been implicated in the susceptibility to hepatocellular carcinoma (HCC), but the findings from published studies are conflicting and inconclusive. To obtain a more precise estimate of the association of Cox-2 polymorphisms with HCC risk, we performed a meta-analysis of eight eligible case-control studies identified through an extensive online database search of PubMed, Embase, Cochrane Library, Web of Science, China National Knowledge Infrastructure, Wanfang and Chinese Biomedicine Database; after exclusion, 2324 cases and 2604 controls were included. The pooled odds ratios with corresponding 95% confidence intervals were calculated to assess associations, using fixed- or random-effect models. In addition, subgroup analysis by ethnicity and sensitivity analysis were performed. Our results showed that the Cox-2 rs20417 (-765 G/C) polymorphism was not associated with HCC risk in the studied genetic contrast modes (C vs. G, GC vs. GG, and CC + GC vs. GG). No significant association was found with ethnic groups examined (P > 0.05). Similarly, no significant association of the Cox-2 rs5275 (+ 8473 T/C) polymorphism and HCC risk was found under any of the studied contrasts (C vs. T, TC vs. TT, CC vs. TT, CC + TC vs. TT, CC vs. TC + TT). The present meta-analysis, combining all currently available data, suggests no significant associations of either Cox-2 polymorphism with HCC risk. Further studies with a larger sample size are needed to determine the association in different ethnicities.
Keywords: Cox-2; hepatocellular carcinoma; meta-analysis; polymorphisms.