ARTICLE ABSTRACT
Background: The TP53 rs1042522 polymorphism (c.215C>G, Arg72Pro) has been extensively investigated as a potential risk factor for colorectal cancer, but the results have thus far been inconclusive.Methods: We searched multiple electronic databases to identify studies investigating the association between the Arg72Pro polymorphism and colorectal cancer. Individual study odds ratios (OR) and their confidence intervals were estimated using allele-frequency, recessive, and dominant genetic models. Summary ORs where estimated using random effects models.Results: We identified 23 eligible case-control studies, investigating 6,514 cases and 9,334 controls. There was significant between-study heterogeneity for all genetic models. The control group in one of the studies was not in Hardy-Weinberg equilibrium; only three studies reported that genotyping was blinded to case/control status and five studies used tumor tissue for case genotyping. Overall, we did not identify any association between rs1042522 and colorectal cancer risk under an allele-frequency comparison (OR, 0.99; 95% confidence interval, 0.89–1.09). Likewise, no association was evident under dominant or recessive models. Studies using tumor tissue for case genotyping found a protective effect for the Pro allele, compared with studies using somatic DNA (Pinteraction = 0.03). Results were also inconsistent between different genotyping methods (Pinteraction = 0.03).Conclusion: We did not identify an association between TP53 rs1042522 and colorectal cancer. Published results seem to be driven by technical artifacts rather than true biological effects.Impact: Future genetic association studies should use more rigorous genotyping methods and avoid the use of tumor tissue as a source of DNA to prevent genotype misclassification due to loss of heterozygosity. Cancer Epidemiol Biomarkers Prev; 19(7); 1840–7. ©2010 AACR.
