An Updated Meta-analysis on the Association of X-Ray Repair Cross Complementing Group 1 Codon 399 Polymorphism with Hepatocellular Carcinoma Risk

Background: A number of studies have reported the association of X-ray repair cross-complementing group 1 (XRCC1) Arg399Gln polymorphism with susceptibility to hepatocellular carcinoma (HCC). However, the results were inconsistent and inconclusive. The aim of this study was to comprehensively explore the association of XRCC1 Arg399Gln variant with HCC risk. Materials and Methods: Systematic searches of PubMed, Elsevier, Science Direct, CNKI and Chinese Biomedical Literature Database were performed. Pooled odds ratio (OR) with 95% confidence intervals (CI) was calculated to estimate the strength of association. Results: Overall, we observed an increased HCC risk among subjects carrying XRCC1 codon 399 Gln/Gln, Arg/Gln and Gln/ Gln+Arg/Gln genotypes (OR=1.20, 95%CI: 1.05-1.38, OR=1.16, 95%CI: 1.05-1.28, and OR=1.14, 95%CI: 1.04-1.24, respectively) based on 20 studies including 3374 cases and 4633 controls. In subgroup analysis, we observed an increased risk of XRCC1 codon 399 Gln/Gln, Arg/Gln and Gln/Gln+Arg/Gln polymorphisms for HCC in hospital-based study (OR=1.25, 95%CI: 1.03-1.51, OR=1.21, 95%CI: 1.07-1.36 and OR=1.18, 95%CI: 1.06-1.31, respectively) and in Asian population (OR=1.19, 95%CI: 1.03-1.38, OR=1.17, 95%CI: 1.04-1.30 and OR=1.14, 95%CI: 1.04-1.25, respectively). Limiting the analysis to the studies with controls in agreement with Hardy-Weinberg equilibrium (HWE), we observed an increased HCC risk among Gln/Gln, Arg/Gln and Gln/ Gln+Arg/Gln genotype carriers (OR=1.17, 95%CI: 1.05-1.29, OR=1.12, 95%CI: 1.00-1.25 and OR=1.11, 95%CI: 1.02-1.21, respectively). Conclusions: This updated meta-analysis results suggest that XRCC1 Arg399Gln variants may contribute to HCC risk. Well-designed studies with larger sample size were required to further verify our findings.


Introduction
Liver cancer in men is the fifth most frequently diagnosed cancer worldwide and the second most frequent cause of cancer death.It is the seventh most commonly diagnosed cancer and the sixth leading cause of cancer death in women.An estimated 748,300 new liver cancer cases and 695,900 cancer deaths occurred worldwide in 2008.Half of these cases and deaths were estimated to occur in China.Among primary liver cancers, hepatocellular carcinoma (HCC) represents the major histological subtype, accounting for 70% to 85% of the total liver cancer burden worldwide (Jemal et al., 2011).
It has been known that chronic infection with hepatitis virus B (HBV) or hepatitis virus C (HCV) is the most common cause of HCC worldwide (Dominguez-Malagon et al., 2001).However, epidemiological data have exhibited that although a lot of people are exposed to these risk factors, only a minority of them develop HCC.

RESEARCH ARTICLE
An Updated Meta-analysis on the Association of X-Ray Repair Cross Complementing Group 1 Codon 399 Polymorphism with Hepatocellular Carcinoma Risk Ya-Dong Wang 1& *, Wen-Long Zhai 2& , Hai-Yu Wang 1 , Xiang-Qun Xia 1   This indicates that an individual susceptibility might play a certain role in HCC carcinogenesis (Pang et al., 2008;Clifford et al., 2010).Recently, Increasing evidence has been accumulated to support the hypothesis that common genetic polymorphisms in genes involved in DNA repair capacity may be of importance in determining individual susceptibility to develop HCC (Long et al., 2011;Huang et al., 2012;Wang et al., 2013).
XRCC1 is a major DNA repair gene involved in base excision repair (Vidal et al., 2001).A common single nucleotide polymorphism leads to amino acid substitutions in XRCC1 gene at codon 399 (exon 10, G-A, and Arg-Gln) (Shen et al., 1998).This mutation could alter XRCC1 function, diminish repair kinetics and influence susceptibility to adverse health effect, such as cancer.To date, numerous studies have investigated the association between XRCC1 Arg399Gln polymorphism and HCC risk (Yu et al., 2003;Long et al., 2004;Yang et al., 2004;Kirk et al., 2005;Li, 2005;Long et al., 2006;Borentain et al., 2007;Ren et al., 2008;Kiran et al., 2009;Wu, 2009;Jia et al., 2010;Zeng et al., 2010;Pan et al., 2011;Tang et al., 2011;Bo et al., 2012;Guo et al., 2012;He et al., 2012;Bose et al., 2013;Gulnaz et al., 2013;Mohana Devi et al., 2013).However, the results from epidemiological studies have been inconsistent and controversial.Although, in recent year, several meta-analyses on XRCC1 Arg399Gln polymorphism related to HCC risk have been reported, conclusions drawn from them were not consistent.Four published meta-analyses failed to observe any positive association between XRCC1 Arg399Gln polymorphism and HCC risk (Zhang et al., 2010;Liu et al., 2011;Xie et al., 2012;Zeng et al., 2013).Two articles (Chen et al., 2011;Pan et al., 2013) reported that XRCC1 Arg399Gln polymorphism was associated with the susceptibility to HCC in total population, but several papers published before 2011 were not included in Chen et al's study; In Pan et al's paper, one overlapping paper was not excluded and Long et al's paper (Long et al., 2006) with controls being not in accordance with Hardy-Weinberg equilibrium (HWE) was mistakenly included in Pan et al's paper.Li et al found that XRCC1 Arg399Gln polymorphism was associated with HCC risk in East Asian population (Li et al., 2013), however, two overlapping studies were not excluded in that paper.Two meta-analyses found that XRCC1 Arg399Gln polymorphism was associated with HCC risk in Chinese population (Duan et al., 2012;Liu et al., 2012), but several published papers were missing and overlapping articles were not excluded in their study.Ever since, new studies have provided additional data on the association between XRCC1 Arg399Gln polymorphism and HCC risk.Therefore, it is required to clarify the association between XRCC1 Arg399Gln variations and HCC risk objectively and comprehensively.In this study, we used the most updated data to address this issue by performing meta-analysis.

Literature and methods
We carried out a systematic search in Pubmed/Medline, Elsevier, Science direct, Chinese National Knowledge Infrastructure (CNKI) and Chinese Biomedical Literature Database, covering all papers published before December 1, 2013 with a combination of the following terms: "hepatocellular cancer" or "hepatocellular carcinoma" or "liver cancer" or "liver tumor", "polymorphism" and "XRCC1".Additional studies were identified by the references lists of the selected papers.
Data from studies were included in our meta-analysis only if the study met the following criteria: (1) The papers should include HCC risk and polymorphism of XRCC1 codon 399; (2) Only the case-control studies or cohort studies were considered; (3) The paper must offer the sample size, odds ratio (OR) with 95% confidence intervals (CI) or the information that can help infer the results; (4) When more than one article was identified for the same study population, we included the publication including more information.
Accordingly, articles that could not offer the numbers of case and control or other essential information were excluded; reviews and overlapping literatures were excluded, too.
In total, 42 published studies were identified with the association between XRCC1 codon 399 polymorphisms and HCC risk.We reviewed all papers in accordance with the criteria defined above, and nine meta-analysis papers, two reviews and eleven overlapping studies were excluded.Therefore, 20 original articles were determined to enter our study.

Data extraction
Data were carefully extracted and tabulated by two investigators first, and then inputted into an electric database.The following information was subtracted from each study: authors, publishing year, country, ethnicity of subjects, source of controls, the extract data of total and exposure number in case and control groups, and OR with 95%CI.If the study provided stratum information, the data coming from similar stratum were added up to make a full use of the data.Characteristics of individual studies were presented in Table 1.

Quantitative data synthesis
A meta-analysis of identified studies was conducted to estimate the strength of association between the polymorphisms of XRCC1 Arg399Gln and HCC risk.Data were combined using both a fixed-effects model and a random-effects model (DerSimonian et al., 1986).The Cochrane Q statistics test is used for the assessment of heterogeneity.The fixed-effects model is used when the effects are assumed to be homogenous, while the randomeffects model is used when they are heterogeneous.The funnel plot was drawn to assess publication bias, Egger's test and Begg's test were used to test for the funnel plot symmetry (Begg et al., 1994;Egger et al., 1997).We tested whether genotype frequencies of controls were in agreement with HWE using the χ 2 test.
All of the statistical analyses were performed with STATA10.0 software package (Stata Corporation, College Station, Texas) and Review Manager (Version 5.0.24.0, the Cochrane Collaboration).All the tests were two-sided, a P value of less than 0.05 for any test was considered to be statistically significant.

Meta-analysis databases
A database was established according to the extracted information from each article.Essential information was listed in Table 1, which showed first author, publishing year, source of control, country, ethnicity of subjects, the numbers of case and control.There were a total of 20 studies with 3374 cases and 4633 controls concerning the XRCC1 codon 399 polymorphisms related to HCC risk.

Test of heterogeneity
Table 2 showed the association of the XRCC1 codon 399 polymorphism with HCC susceptibility.The heterogeneity of XRCC1 codon 399 Gln/Gln versus Arg/ Arg, Arg/Gln versus Arg/Arg, and Gln/Gln+Arg/Gln versus Arg/Arg was analyzed for 20 case-control studies.
DOI:http://dx.doi.org/10.7314/APJCP.2014.15.11.4443 XRCC1 Arg399Gln Polymorphism and Hepatocellular Carcinoma Risk The results showed that XRCC1 codon 399 Gln/Gln versus Arg/Arg, Arg/Gln versus Arg/Arg and Gln/Gln+Arg/Gln versus Arg/Arg for total population, hospital-based study and Asian population, Gln/Gln+Arg/Gln versus Arg/ Arg for population-based study, and Arg/Gln verus Arg/ Arg and Gln/Gln+Arg/Gln versus Arg/Arg for control in agreement with HWE had heterogeneity with a P value less than 0.05, therefore, we analyzed the summary odds ratios for these with a random-effects model.Fixed-effects model was used to analyze the summary OR for the rest.

Bias diagnosis
Publication bias was examined by using funnel plot.The shape of the funnel plot seemed to be approximately symmetrical (Figure 2a, 2b, 2c).The results from Egger's test and Begg's test indicated that publication biases may not have a significant influence on our current metaanalysis results (Table 2).

Sensitivity analyses
Sensitivity analyses were conducted to identify the effects of the individual dataset on the summary odds ratio by sequential omission of each study.The overall effects were not modified when the studies were homogenous for Gln/Gln versus Arg/Arg, Arg/Gln versus Arg/Arg and Gln/ Gln+Arg/Gln versus Arg/Arg polymorphisms among total population by removing some eligible studies (Figure S1, S2, S3).

Total events
Heterogeneity: Chi² = 20.17,df = 14 (P = 0.13); I² = 31% Test for overall effect: Z = 3.17   .2014.15.11.4443 XRCC1 Arg399Gln Polymorphism and Hepatocellular Carcinoma Risk cases and 3265 controls.In addition, other three studies did not observe an effect modification with small sample sizes as well (Zhang et al., 2010;Xie et al., 2012;Zeng et al., 2013).Chen et al (Chen et al., 2011) found that XRCC1 Arg399Gln polymorphism was associated with the susceptibility to HCC on the base of 7 studies including 1342 cases and 2207 controls, but several papers published before 2011 (Wu, 2009;Jia et al., 2010;Zeng et al., 2010) were not included in their study.Pan et al also found a significant association between XRCC1 polymorphism and HCC risk containing 15 studies with 6556 individuals, but one overlapping paper was not excluded and one paper with controls being not in agreement with HWE was mistakenly included.Therefore, the conclusion drawn from Chen et al's paper and Pan et al's paper are not reliable.The present meta-analysis of 20 case-control studies including 3374 cases and 4633 controls might have sufficiently statistical power to detect the association of XRCC1 Arg399Gln polymorphism with HCC risk, owing to small random error in large sample sizes.
Similar to other systematic reviews and meta-analyses, there were some limitations in this present study.First, only published papers were included in this study.Therefore, publication bias is very likely to occur.To address this problem, Egger's test and Begg's test were conducted.Our results indicate that the likelihood of key publication bias might not be present in this meta-analysis.Secondly, each study had different eligibility criteria for subjects and different source of controls, which should be taken into account when explaining the combined estimates.When subgroup analysis was performed by source of control, we observed an association between XRCC1 codon 399 polymorphism and HCC risk in hospital-based study, but not in population-based study, maybe owing to too small sample size of population-base study (963 cases and 1019 controls in population-based study) to detect the modified effects.Thirdly, the summary ORs were based on individual unadjusted estimates, while a more precise analysis might be performed if detailed individual data were available, which could allow for an adjusted estimation by HBV status, age, alcohol consumption and environment factors.Therefore, it is required for authors of all of the published papers to share their data.
Considering that the frequency of the XRCC1 399 Gln allele variant is significantly different among different ethnic population.Allele frequency patterns of XRCC1 Arg399Gln polymorphism vary greatly between major ethnic groups.The frequency of Gln allele was more than 0.3 in Caucasians (Zhang et al., 2006;Zienolddiny et al., 2006), but less than 0.2 in Asian population (Long et al., 2004;Yin et al., 2007).When stratified by ethnicity, we observed an association of XRCC1 polymorphism with HCC risk among Asian population.Results from previous meta-analysis based on 13 studies including 3011 case and 3619 controls showed that XRCC1 codon 399 polymorphism was associated with HCC risk in East Asian population (Li et al., 2013), however, two overlapping studies (Chen et al., 2005;Han et al., 2012) were not excluded in that paper, therefore, the results from that paper are not accurate.Only two published studies included in this meta-analysis focused on the association of XRCC1 codon 399 polymorphism with HCC risk in Caucasian and African population, so, we did not perform meta-analysis in other ethnic population further.
It is widely acknowledged that, if the distribution frequency of genotypes in the control groups deviated from HWE, the results of genetic association studies might be spurious (Salanti et al., 2005).To address this problem, subgroup analysis was carried out in this study by HWE in controls.When the studies that were not in accordance with HWE were excluded from this present analysis, the results were persistent and robust, suggesting that this factor might not have significant effects on the overall estimates in the current meta-analysis.
In conclusion, this updated meta-analysis presented clear and comprehensive evidence that XRCC1 codon 399 variants were associated with the risk of HCC worldwide, especially in Asian population.Large scale studies with the pooling of individual study data should be taken into consideration in future association studies to confirm the results from this current meta-analysis.

Table 2 . Summery Odds Ratios on the Relation of Polymorphisms of the XRCC1 Arg399Gln to Hepatocellular Carcinoma risk
*HWE, Hardy-Weinberg Equilibrium