GSTM1 Polymorphisms and Lung Cancer Risk in the Chinese Population: a Meta-Analysis Based on 47 Studies

Lung cancer is the most commonly diagnosed cancer as well as the leading cause of cancer death in males globally, with 1.6 million newly confirmed cases and 1.4 million deaths from lung cancer annually (Jemal et al., 2011). Human cancers can be initiated by DNA damage caused by environmental chemical agents, such as polycyclic aromatic hydrocarbons (PAHs), and some adverse habits including tobacco smoking and alcohol use (Neumann et al., 2005). Studies have shown that exposures to environmental and occupational PAHs are risk factors for lung cancer (Kriek et al., 1993; Li et al., 2004; Vineis & HusgafvelPursiainen, 2005). However, not all of those who have been exposed to the risk factors will develop lung cancer, suggesting that there is individual variation in cancer susceptibility in the general population (Neumann et al., 2005). To understand the contribution of genetic variations in lung cancer, genetic association approach has been widely used and has been fruitful. For example, studies have consistently associated the development of lung cancer with the genetic factors such as glutathione S-transferase M1 (GSTM1).


Introduction
Lung cancer is the most commonly diagnosed cancer as well as the leading cause of cancer death in males globally, with 1.6 million newly confirmed cases and 1.4 million deaths from lung cancer annually (Jemal et al., 2011). Human cancers can be initiated by DNA damage caused by environmental chemical agents, such as polycyclic aromatic hydrocarbons (PAHs), and some adverse habits including tobacco smoking and alcohol use (Neumann et al., 2005).
Studies have shown that exposures to environmental and occupational PAHs are risk factors for lung cancer (Kriek et al., 1993;Li et al., 2004;Vineis & Husgafvel-Pursiainen, 2005). However, not all of those who have been exposed to the risk factors will develop lung cancer, suggesting that there is individual variation in cancer susceptibility in the general population (Neumann et al., 2005). To understand the contribution of genetic variations in lung cancer, genetic association approach has been widely used and has been fruitful. For example, studies have consistently associated the development of lung cancer with the genetic factors such as glutathione S-transferase M1 (GSTM1).
Xin-Ping Chen 1,2& , Wei-Hua Xu 2& , Da-Feng Xu 3 , Xian-He Xie 4 *, Jia Yao 5 , Sheng-Miao Fu 2 * The association between GSTM1 gene and lung cancer has been investigated in numerous epidemiologic studies since glutathione S-transferase was first suggested as a potential marker for susceptibility to lung cancer in 1986 (Seidegard et al., 1986). Glutathione S-transferases consist five distinct families, namely alpha (GSTA), sigma (GSTS), mu (GSTM), pi (GSTP), and theta (GSTT) (Kiyohara et al., 2002). Located on the chromosome 1p13.3, the GSTM1 plays an important role in the xenobiotics' detoxification. The most common genotype of GSTM1 gene is homozygous deletion (null genotype), which has been suggested to be associated with the loss of enzyme activity, increased vulnerability to cytogenetic damage and resulted in the increased susceptibility to cancer (Hayes et al., 2005;McIlwain et al., 2006).
Recently, the role of GSTM1 polymorphism in the etiology of different types of cancer has drawn more and more attention, including lung cancer. A number of studies in China have been conducted to explore whether GSTM1 polymorphism is associated with lung cancer susceptibility, but provided controversial or inconclusive results. Therefore, we conducted a meta-analysis to more precisely define the effect of GSTM1 polymorphism on risk for lung cancer in Chinese populations.

Search strategy
We searched databases containing PubMed, Springer Link, Ovid, Chinese Wanfang Data Knowledge Service Platform, Chinese National Knowledge Infrastructure (CNKI), and Chinese Biology Medicine (CBM) up to 5th April 2014, using combination of the following terms: (1) GSTM1 or GST M1; (2) lung cancer or lung neoplasm or lung tumor; (3) polymorphism or variant or variation; and (4) Chinese or China. We limited the languages to English and Chinese. Besides, the references from retrieved articles were also searched.

Eligibility criteria
Studies were included in this meta-analysis if they met the following criteria: (1) case-control study or cohort study studying on associations between GSTM1 polymorphism and lung cancer susceptibility; (2) all patients with the diagnosis of lung cancer confirmed by pathological or histological examination; (3) sufficient published data about sample size, odds ratio (OR), and their 95% confidence interval (CI); (4) published in English or Chinese language; (5) all participants were Chinese. Studies were excluded when they were: (1) not case-control study or cohort study; (2) duplicate of previous publication; (3) based on incomplete data; (4) meta-analyses, letters, reviews, case reports, or editorial articles.

Data extraction
Data were independently extracted by two reviewers (Xin-ping Chen and Wei-hua Xu) using a standardized data extraction form. Discrepancies were resolved by discussion and if consensus was not achieved the decision was made by the all the reviewers. The title and abstract of all potentially relevant articles were screened to determine their relevance. Full articles were then scrutinized if the title and abstract were ambiguous. The following data were extracted from the identified studies: the first author, publication year, source of controls, geographic area, sample size, and the number of subjects with two GSTM1 genotypes. In this meta-analysis, the quality assessment of individual study was conducted according to the nine-star Newcastle-Ottawa Scale (Wells et al., 2009). For articles including different source of controls, data were extracted separately (Table 1).

Statistical analysis
Pooled odds ratios (ORs) and 95% confidence intervals (CIs) were used to assess the association of GSTM1 genetic polymorphism with lung cancer risk in Chinese population. Given that there was distribution of null/present heterozygote in only one study selected, the Hardy-Weinberg equilibrium (HWE) test could not be conducted. Cochrane's Q test was performed to test the between-study heterogeneity. If there was heterogeneity, then the random-effects model was chosen to pool the ORs with 95 % CIs, otherwise the fixed-effects model was used. Publication bias was investigated with the funnel plot, in which the Standard Error (SE) of log OR of each study was plotted against its OR. Funnelplot asymmetry was further assessed by the method of Egger's linear regression test (Egger et al., 1997). All the P values were two sided. P value less than 0.05 was considered statistically significant. All statistical analysis was conducted by using Stata version 10.0 (Stata Corp, College Station, Texas, USA).

Overall analysis
There was evidence of between-study heterogeneity in all included studies (χ 2 =88.54, p<0.001). Therefore, the random-effects model was used in overall analysis. The results showed that the pooled OR with 95% CI for lung cancer in Chinese with null GSTM1 was 1.45 (1.32-1.60) (Figure 1)

Subgroup analysis
In the subgroup analysis based on source of control, the results showed that the GSTM1 polymorphism was significantly related to lung cancer risk among populationbased population (OR = 1.55, 95%CI: 1.39-1.73), as well as among hospital-based studies (OR = 1.26, 95%CI: 1.08-1.46) ( Table 2). In addition, we also performed stratified analysis based on the quality score and geographic area, it revealed the similar results with all the studies (Table 2).

Sensitive analysis
To evaluate the stability of the results, we performed a sensitivity analysis by different model. All the results were not materially altered (Table 2). Hence, results of the sensitivity analysis suggest that the data in this metaanalysis are relatively stable and credible.

Bias diagnosis
The Begg's funnel plot and Egger's test were performed to access the publication bias of literatures. As showed in Figure 2, the shape of the funnel plots did

Discussion
Till date, a series of studies in China have focused on the relation between GSTM1 polymorphism and lung cancer risk. Nevertheless, the results were inconclusive and inconsistent. Some papers have reported that a statistically significant correlation was found between null GSTM1 and lung cancer risk. Conversely, the results from other studies suggested that the null GSTM1 was not associated with lung cancer risk. Therefore, we conducted this update meta-analysis by critically reviewing 47 individual studies on GSTM1 gene polymorphism with lung cancer risk in Chinese population. In the meta-analysis, we found that the GSTM1 null variant was significantly associated with lung cancer risk. However, our results showed a stronger association with lung cancer risk than those reported by the Carlste's study (Carlsten et al., 2008) on the GSTM1 polymorphism that included 19,638 cases and 25,266 controls of the world's overall populations (OR = 1.22, 95% CI = 1.14-1.30). Furthermore, our results are almost the same as those of Shi's results (Shi et al., 2008) (OR = 1.54, 95% CI = 1.31-1.80). His report only included 2235 cases and 2315 controls of Chinese population. To our knowledge, our study represented the first meta-analysis with a large sample size on the interaction of GSTM1 variant with lung cancer in Chinese population.
When we performed stratified analyses by quality score, geographic area and source of controls, significant association with susceptibility for the development of lung cancer was found in all the subgroups. With regard to heterogeneity, some of the factors extracted in this study were the main source of heterogeneity. But it might also make attributions for other unknown factors, such as dietary habits, dinking status, other environmental exposures (passive smoking and cooking oil fume), family history of cancer, other genetic-related respiratory diseases as well as other related genetic polymorphisms.
The pathways of carcinogen metabolism are complex, mediated by the activities of multiple genes. The effect of any single gene might have a limited impact on lung cancer risk than have so far been anticipated. Many controversial data are present in literature. Positive associations were found in certain populations and not confirmed in others. In addition to an expected interethnic variability in allele frequencies, variability has also been found within an ethnic group, resulting in heterogeneity in association studies. Gene-environment interactions could be a confounding factor in these studies, with controversial findings on cancer risk. Studies taking these factors into account may eventually lead to have a better, comprehensive understanding of the association between the GSTM1 polymorphism and lung cancer risk.
This study has some limitations. First, we didn't perform subgroup analysis on smoking status and other exposure history. Second, our results were based on unadjusted estimates. Therefore, the confounding factors might influence the estimates. Third, some publication bias was detected. Because the papers included in our meta-analysis were limited to those published in either English or Chinese only in the periods between 1989 and 2014, it is possible that some relevant published studies and unpublished studies that are likely to have null results were not included, which may have biased the results.
In summary, although studies investigating the association between GSTM1 polymorphism and the risk of lung cancer arrived at different conclusions (Piao et al., 2013;Shukla et al., 2013), this meta-analysis suggested that there was a significant association between null GSTM1 variant and lung cancer risk in the Chinese. Several recommendations on the future association studies of GSTM1-lung cancer can be made from this metaanalysis. First, a well thought-out study design is crucial for an association study. Second, it is important to make an effort to control risk factors, preferably in the design stage. Third, larger research articles in other populations with different environmental background are required. Lastly, care should be exercised in genotyping and in checking for abnormality, such as the deviation from HWP.