Lung Cancer (2008) 59, 155163

available at www.sciencedirect.com

journal homepage: www.elsevier.com/locate/lungcan

CYP1A1 and GSTM1 polymorphisms and lung cancer risk in Chinese populations: A meta-analysis
Xiuquan Shi a, Suhua Zhou a, Zhongxu Wang b, Zongcan Zhou c, Zengzhen Wang a,
a

Department of Epidemiology and Health Statistics, School of Public Health, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, PR China b National Institute for Occupational Health and Poison Control, Chinese Center for Disease Control and Prevention, Beijing, PR China c Department of Toxicology, School of Public Health, Health Science Center, Peking University, Beijing, PR China Received 12 June 2007; received in revised form 3 August 2007; accepted 6 August 2007

KEYWORDS
CYP1A1; GSTM1; Polymorphism; Lung cancer; Susceptibility; Meta-analysis

Summary Genetic polymorphisms of cytochrome p450 (CYP1A1) and glutathione S-transferase M1 (GSTM1) genes are thought to have signicant effects on the metabolism of environmental carcinogens and thus on cancer risk, but the reported results are not always consistent. In this meta-analysis, we assessed reported studies of associations between polymorphisms of these two genes and risk of lung cancer in Chinese populations. Through a systematic literature search for publications between 1989 and 2006, we summarized the data from 46 studies on polymorphisms of MspI and exon7-Val of CYP1A1 and GSTM1 and lung cancer risk in Chinese populations, and found that compared with the wild-type homozygous genotype (type A), lung cancer risk for the combined variant genotypes (types B and C) was 1.34-fold (95% condence interval [CI] = 1.081.67) (Z = 2.64, P = 0.008); the risk for the combined variant genotypes (Ile/Val and Val/Val) of CYP1A1 exon7 was 1.61-fold (95% CI = 1.242.08) (Z = 3.62, P < 0.001), compared with the Ile/Ile genotype; and that the risk for the GSTM1 null genotype was 1.54-fold (95% CI = 1.311.80) (Z = 5.32, P < 0.001), compared with the GSTM1 present genotype. Therefore, in 46 published studies in Chinese populations, we found evidence of an association between the CYP1A1 variant and GSTM1 null genotypes and increased risk of lung cancer. 2007 Elsevier Ireland Ltd. All rights reserved.

1. Introduction
Environmental chemical pollutants (ECPs) are increasingly present in our living environment as a result of the development of the modern industry and urbanization. Many ECPs are widely spread and difcult to be degradated in the environment. Therefore, ECPs could have a long-term effect on

Corresponding author. Tel.: +86 27 83692725; fax: +86 27 83692725. E-mail address: zzhwang@263.net (Z. Wang).

0169-5002/$ see front matter 2007 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.lungcan.2007.08.004

156 human health. Among the ECPs are the most-studied pollutants, polycyclic aromatic hydrocarbons (PAHs), which have been found in cigarette smoke and in polluted indoor and outdoor air and shown to be associated with risk of many diseases including cancer. Lung cancer is a serious threat to public health, ranking number 1 in cancer-related deaths. Human cancers can be initiated by DNA damage caused by environmental chemical agents, such as PAHs, and some adverse habits including tobacco smoking and alcohol use [1]. Studies have shown that exposures to environmental and occupational PAHs are risk factors for lung cancer [24]. 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 [1]. For example, phase I and phase II enzymes, such as CYP1A1 and GSTM1, respectively, that can metabolize xenobiotics in humans are polymorphic [58]. Both biological and biochemical evidence indicates that genetic polymorphisms of these genes can inuence the balance between metabolic activation and detoxication of some toxicants, such as benzo[a] pyrene, and thus they are relevant to individual susceptibility to lung cancer. CYP1A1 and GSTM1 are good candidate genes, because they are modiers of risk of lung cancer due to their allelic variants that alter the inducibility of the enzyme by the inducers. The effect of metabolic polymorphisms on lung cancer risk has been shown to depend on the level of exposure to xenobiotics in some subgroups of individuals. For example, it is suggested that the effect of the CYP1A1 polymorphism is greater in non-smokers than in smokers and in women than in men [9,10]. It is conceivable that individuals who have inherited specic variants in these genes, such as CYP1A1 and/or GSTM1, may become susceptible to chemical carcinogens and thus at a high risk of developing lung cancer. In this report, we tested the hypothesis by performing a meta-analysis that the inter-individual susceptibility to lung cancer is associated with genetic variation in metabolic enzymes. We summarized reported casecontrol studies on three most-studied polymorphisms (i.e., MspI-restriction fragment length polymorphism and exon7-Val polymorphism of CYP1A1 and GSTM1) in Chinese populations. Because a single study may have been underpowered in detecting the effect of low penetrance genes, particularly in assessing doseresponse relationships, a quantitative synthesis of accumulated data from published studies may enhance statistical power to detect the association between genetic polymorphisms and lung cancer risk. Meta-analyses of studies on these two genes in other ethnic groups have been reported elsewhere (Raimondi et al. [6]; Houlston [11]; Ye et al. [12]; Vineis et al. [13]).

X. Shi et al. China/Chinese; lung cancer and CYP1A1 and China/Chinese; lung cancer and P4501A1 and China/Chinese. We evaluated potentially relevant publications by examining their titles and abstracts and then procured the most relevant publications for a closer examination. Besides the database search, the references lists of the selected papers were also screened for other potential articles that may have been missed in the initial search. The search and evaluation were conducted from January to March 2007. The following criteria were used for the literature selection for the meta-analysis: (1) the articles should be published in either English or Chinese between January 1989 and December 2006; (2) the articles should describe studies only in Chinese populations; (3) only the casecontrol studies and cohort studies were considered; (4) the paper should clearly describe lung cancer diagnoses and the sources of cases and controls; (5) the authors must offer the size of the sample, odds ratios (ORs) and their 95% condence intervals (CIs) or the information that can help infer the results in the papers; (6) the denition of the exposure/risk genotypes was similar in all papers; (7) the methods of data collection and analysis should be statistically acceptable; (8) those publications that presented data allowing such outcomes to be derived were also included. Accordingly, the following exclusion criteria were also used: (1) the articles had studied ethnic populations other than Chinese; (2) the design and the denition of the exposure were obviously different from those of the selected papers; (3) not offering the source of cases and controls and other essential information; (4) reviews and repeated literatures were also excluded. After the searching, we selected 49 published papers dealing with casecontrol and cohort studies of the polymorphisms. We reviewed all papers in accordance with the criteria dened above, and we then excluded 14 papers because their study designs were different from others or they did not list data clearly enough for further analysis or repeated literatures. For the included studies, we used Jadad scale to appraise their qualities of study designs (hospital versus population-based study, response rates, lack of bias, control of confounding, etc.) and gave a matched score to each paper, in the nal the weighted score was 3.65 for Jadad score, and thus, the qualities of the included studies were acceptable. Among the 35 qualied literatures, 15 studies focused on the CYP1A1 MspI polymorphism, 11 studies focused on the CYP1A1 exon7 polymorphism, and 20 papers reported on the polymorphism of GSTM1. Apparently, there were 11 articles that had reported two kinds of polymorphisms.

2. Materials and methods

2.1. Literature search strategy for identication of the studies
We carried out a search in the Medline and Chinese National Knowledge Infrastructure (CNKI), covering all papers published between 1989 and 2006, with a combination of the following keywords: lung cancer and GSTM1 and

CYP1A1 and GSTM1 polymorphisms and lung cancer risk in Chinese

157

2.2. Data extraction

To minimize the bias and to improve the reliability, two reviewers checked all potentially relevant studies independently. Data on the following characteristics were also extracted: rst author, dates on which the study was published, journal, the exact data of total and exposed number in case and control groups, and odds ratios and their condence intervals. A standardized procedure was used to estimate odds ratios from the exact data of subject numbers of the cases and controls displayed in several of the papers. We tabulated the data and then computed the odds ratios. If the study provided stratum information, the data from similar stratum were added up to make a full use of the data. From the selected reports, data were abstracted in duplicate, using a standardized form. At rst, we intended to analyze the genegene interactions (i.e., between CYP1A1 MspI and GSTM1 polymorphism and between CYP1A1 exon7 and GSTM1 polymorphism), but we only selected 4 and 5 papers of CYP1A1 and GSTM1 in Chinese populations, respectively, on the basis of the inclusion and exclusion criteria. Because the number of studies was limited for the genegene interaction analysis, we did not present the data in this report.

models incorporate an estimate of the between-study variance and tend to provide wider condence intervals when the results of the constituent studies differ among themselves [15], in the absence of between-study heterogeneity, the two methods provide identical results. We also did cumulative meta-analysis to evaluate whether the summary OR for the allele contrasts changed over time as more data were accumulated. Inverted funnel plots and the Eggers test were used to provide diagnosis of publication biases. All of the statistical analyses were performed with Statistical Analysis System software (version 8.1; SAS Institute, Cary, NC) and Review Manager (version 4.2.8, The Cochrane Collaboration). All the tests were two-sided.

3. Results
3.1. Literature search and meta-analysis databases
According to the extracted information from each article we established a database. All essential information is listed in Tables 13. For example, Table 1 shows the cancer type of the studies, the publication year, rst author, and the numbers of cases and controls for each CYP1A1 MspI genotype. There were a total of 15 casecontrol studies concerning the CYP1A1 MspI polymorphism [3,1629] (Table 1), 11 casecontrol studies [17,20,25,3037] concerning the CYP1A1 exon7 polymorphism (Table 2), and 20 casecontrol studies [3,16,21,23,24,29,31,3749] concerning the GSTM1 polymorphism (Table 3). Other necessary information is also listed in the forest plots in our meta-analysis. No qualied researches (only three reviews) were acquired before 1998, which suggested the researches between polymorphisms and lung cancer were started very late in China.

2.3. Methods for quantitative synthesis

ORs and their 95% CIs in each study were calculated directly from the data given in the articles. In the analysis of pooled data, two different approaches were used: a xed effect model and a random-effect model. A xed effect model was rst performed, assuming the same homogeneity of the effect size across all of the studies, followed by a random-effect model under the assumption of heterogeneity between studies. Both within-study variability and between-study variability were considered in the randomeffect model, and the effect size was assumed to be a draw from a distribution with specic ORs. Finally, risk coefcients (combined ORs) and their 95% CIs were presented. Tests for heterogeneity between studies were performed with the Chi-square test. The funnel plot was drawn to assess publication biases, and the test suggested by Egger et al. [14] was used to test for the funnel-plot symmetry, in which a regression model was built, using the standardized estimate of the size effect as a dependent variable and the inverse of the standard error as an independent variable. If the intercept was signicantly different from zero, the estimate of the effect was considered biased. We evaluated the risk of lung cancer associated with combined variant genotypes (i.e., type B and type C for MspI and Ile/Val and Val/Val for exon7 in CYP1A1) versus their wild-type homozygote and the null genotype of GSTM1 versus its present homozygote. For each polymorphism, we estimated the between-study heterogeneity across the eligible comparisons using the Chi-square test, and the heterogeneity was considered signicant if P < 0.05. Values from a single study were combined using models of both random (DerSimonian and Laird) effects and xed (MantelHaenszel) effects. Although random-effect

3.2. Test of heterogeneity

Table 1 shows the association between the CYP1A1 MspI polymorphism and lung cancer risk. We analyzed the heterogeneity for all 15 casecontrol studies and the test value of Chi-square was 37.39 with 14 degree of freedom (d.f.) and P < 0.001 in a random-effect model. For the association between the CYP1A1 exon7 polymorphism and lung cancer risk, the Chi-square value for the heterogeneity of all 11 casecontrol studies was 22.73 with 10 d.f. and P = 0.010 in a random-effect model (Table 2). Similarly, for the association between genetic polymorphism of GSTM1 and lung cancer risk (Table 3), the Chi-square value for the heterogeneity of the 20 casecontrol studies was 31.75 with 19 d.f. and P = 0.030 in a random-effect model. Another index for the heterogeneity test is the I-square value, and if the I-square value is lower than 70%, the heterogeneity of the studies is considered acceptable. In this meta-analysis, I-square values were 62.6%, 56.0%, and 40.1% for the CYP1A1 MspI, CYP1A1 exon7 and GSTM1 polymorphisms, respectively. So we can use random-effect models to combine all studies although they had some between-study heterogeneity, because the random-effect model assumes that the studies included in the meta-analysis are a random sample of a hypothetical population and thus have withinand between-study variability.

158

X. Shi et al.

Table 1 Meta-analysis with a random-effect model for the association between lung cancer risk and the CYP1A1 MspI polymorphism (the combined types B and C versus type A)

3.3. Quantitative data synthesis

The risks (or ORs) of cancer associated with genetic polymorphisms were estimated for each study. Overall, the eligible studies included 5252 cases (lung cancer patients) and 6626 controls, and the group of exposure (i.e., variant genotypes) included 3143 cases and 3272 controls. There were significant differences in the frequencies of variant genotypes between cases and controls.

For the CYP1A1 MspI polymorphism, the data available for our meta-analysis were obtained from 15 casecontrol studies of 1841 cases and 2413 controls, of which 1212 cases and 1414 controls had the combined variant genotypes (type B and type C, the exposure group) and 629 cases and 999 controls had wild-type homozygote (type A) of the CYP1A1 gene. The overall OR for the combined types B and C versus type A genotypes was 1.34 (95% CI = 1.081.67) as estimated in a random-effect model (Z = 2.64, P = 0.008).

Table 2 Meta-analysis with a random-effect model for the association between lung cancer risk and the CYP1A1 exon7 polymorphism (the combined Ile/Val and Val/Val genotypes versus the Ile/Ile genotype)

CYP1A1 and GSTM1 polymorphisms and lung cancer risk in Chinese

159

Table 3 Meta-analysis with a random-effect model for the association between lung cancer risk and the GSTM1 polymorphism (null genotype versus present genotype)

For the CYP1A1 exon7 polymorphism, the data available for this analysis were obtained from 11 casecontrol studies of 1176 cases and 1898 controls, of which 616 cases and 727 controls had the combined variant genotypes (Ile/Val and Val/Val) and 560 cases and 1171 controls were wildtype homozygote (Ile/Ile) for exon7 of the CYP1A1 gene. The overall OR for the combined Ile/Val and Val/Val versus Ile/Ile genotypes was 1.61 (95% CI = 1.242.08) in a random-effect model (Z = 3.62, P < 0.001). For the GSTM1 polymorphism, the data available for this analysis were obtained from 20 casecontrol studies of 2235 cases and 2315 controls, of which 1315 cases and 1131 controls had the GSTM1 null genotype (i.e., the exposure group) and 920 cases and 1184 controls had the GSTM1 present genotype. The overall OR for the null versus present genotype was 1.54 (95% CI = 1.311.80) in a random-effect mode (Z = 5.32, P < 0.001). In order to compare the difference and explore the sensitivity of the analysis, we also reported the results of the xed effect models as the following: the combined ORs and 95% CIs were 1.40 (1.231.59), 1.58 (1.351.86), and 1.49 (1.321.68) for the CYP1A1 MspI, CYP1A1 exon7 and GSTM1 polymorphisms, respectively, similar to the results obtained from the random-effect models.

Fig. 1 Funnel plot of the meta-analysis of lung cancer risk and the CYP1A1 MspI polymorphism.

3.4. Bias diagnostics

In the funnel plot analysis of publication biases (the contrast of homozygous genotype plotted against the precision), the shape of the funnel plot appeared to be approximately symmetrical, and the magnitude of the main ORs was in dispersion on the right side of 1, but there was some uncertainty for the symmetrical degrees were not content (Figs. 13).

Fig. 2 Funnel plot of the meta-analysis of lung cancer risk and the CYP1A1 exon7 polymorphism.

160

X. Shi et al. polymorphism. For the exon7 polymorphism, the combined variant genotypes (Ile/Val and Val/Val) versus wild-type homozygote (Ile/Ile), OR = 1.61 (95% CI = 1.242.08) which was obtained from 11 studies of 1176 cases and 1898 controls. Both of the results implied that CYP1A1 variant genotypes were the hazard factors for lung cancer. Our results are consistent with that reported study of 1950 cases and 2617 controls in United States by Marchand et al. [10]. Whereas Houlston [11] suggested that there was no evidence of lung cancer risk associated with the combined variant (MspI and exon7) and CYP1A1 homozygote, but his report included 2058 cases and 2765 controls of different ethnic populations. Vineis et al. [13] also did not nd evidence of lung cancer risk associated with the MspI combined variant and CYP1A1 homozygote; because his report only included 478 cases and 638 controls in Japanese (others were Caucasians). It was inferred that the major reason of different results was due to low penetrance of the CYP1A1 gene in Caucasians and other non-Chinese ethnicity [18]. Inter-individual variability in the GST enzyme (a phase II enzyme) activity is believed to confer differential susceptibility to environmental cancers, such as lung cancer [50,51]. Some genetic variants in the GST genes, such as the GSTM1 null polymorphism, are known to abolish enzyme activities [7]. Because individuals with the GSTM1 null genotype are reportedly to have higher levels of PAH-dGMP adducts in their lung tissue than those with the GSTM1 present genotype, such genetic variants have been extensively studied as candidates for lung cancer susceptibility. Gao et al. [38] rst reported a possible relationship between the GSTM1 deciency and lung cancer risk (OR = 2.56 and 95% CI = 1.195.49) in a Chinese study of 46 cases and 70 controls. This study stimulated a serious of reports that further tested this hypothesis. But some earlier studies had very limited power to demonstrate such a moderate effect because of their small sample sizes that yielded apparently conicting results [3,16,21,23,24,29,31,3749]. The present meta-analysis of 20 casecontrol studies of 2235 cases and 2315 controls had a much improved statistical power. Our results from the meta-analysis in Chinese populations of the GSTM1 polymorphism showed a strong association with lung cancer risk than those reported by Houlston [52] (OR = 1.13, 95% CI = 1.031.25). However, his report included 3301 cases and 5744 controls of different ethnic populations, mainly Caucasians. Furthermore, our results are almost the same as those of the Ye Zs study [12] on the GSTM1 polymorphism that included 7333 cases and 8518 controls of east Asians, mainly included Japanese populations. The discrepancies in the results of meta-analyses often reect the differences in genetic backgrounds and exposure history in the study populations included in the analyses. Given that exposure to tobacco carcinogens is a major risk factor for lung cancer, the hypothesis that the modulation of carcinogen metabolism is under genetic control is a plausible mechanism for explaining inter-individual susceptibility. Our meta-analysis only focused on Chinese populations, whereas Houlstons analysis [11] included different ethnic populations, including Europeans (especially Caucasians), Africans, Americans and Asians (Japan). It is

Fig. 3 Funnel plot of the meta-analysis of lung cancer risk and the GSTM1 polymorphism.

The Eggers test is based on a linear regression of the standard normal deviate against its precision. In our analysis we used the inverse of the standard error as the independent variable and the standardized estimate of the size effect (log OR upon its standard error) as the dependent variable. The estimate of the effect is considered biased, if the intercept is signicantly different from zero. The test results were as the following: t = 0.26 with 14 d.f., P = 0.799 for the CYP1A1 MspI polymorphism; t = 0.23 with 10 d.f., P = 0.823 for the CYP1A1 exon7 polymorphism, and t = 2.03 with 19 d.f., P = 0.058 for the GSTM1 polymorphism. Furthermore, the fail-safe number was 91, 79, and 224 for these polymorphisms, respectively, at a 0.01 level. These results are consistent with the standard of greater than ve times of included-article number. Therefore, both the Eggers tests and the fail-safe numbers suggested that publication biases may not have a signicant inuence on the results of the CYP1A1 gene, but there was some uncertainty for the results of the GSTM1 gene because the P value was much approached to 0.05. It was obvious that smaller studies tended to report a higher risk than larger studies, and the results of the large studies were more stable.

4. Discussion
It is well known that CYP and GST genes are large families of endoplasmic and cytosolic enzymes that catalyze the activation and detoxication, respectively, of reactive electrophilic compounds, including many environmental carcinogens (e.g., benzo[a] pyrene and other PAHs). Our results of this meta-analysis suggest that genetic variations of CYP1A1 and GSTM1 may contribute to susceptibility to lung cancer in Chinese populations. CYP1A1 is a phase I enzyme, it plays an essential role in the metabolic activation of major classes of tobacco procarcinogen such as aromatic amines and PAHs. So it may affect the metabolism of the environmental carcinogens and alter susceptibility to lung cancer. In our meta-analysis of CYP1A1 gene in Chinese populations, the combined variant genotypes (type B and type C) versus wild-type homozygote (type A), OR = 1.34 (95% CI = 1.081.67) which was obtained from 15 studies of 1841 cases and 2413 controls for the MspI

CYP1A1 and GSTM1 polymorphisms and lung cancer risk in Chinese obvious that different ethnicities may have different susceptibility to lung cancer. For example, it was reported that the 1.9 allele of human CYP1A1 MspI polymorphisms was associated with an increased incidence of Kreyberg Type I bronchogenic carcinoma in Japanese populations and this allele was about 3 times more frequent in Japanese than in Caucasians of Norway and the USA [53]. In Houlstons analysis [52] for the GSTM1 gene, the ethnicities of cases and controls were probably mixed in a number of the studies included in his analysis. The frequency of the GSTM1 polymorphism varies considerably between ethnic groups; therefore, a failure to match cases and controls by ethnicity represents a source of bias in meta-analyses. It is possible that an association mediated by linkage disequilibrium may be conned to certain ethnic populations; hence, population stratication would mask such a genetic effect. 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. Lung cancer has some known major environmental determinants other than tobacco smoke, and large studies with detailed exposure information are needed to evaluate reliably any moderate genetic effects. Although different results in published meta-analyses were partly explained by the different ethnic populations included in the analyses, there must be other potential confounders that were not fully presented in the published papers. For instance, although subjects in these studies are approximately similar in the age, gender, and smoking, which will remove most confounders by using this matched method in case and control groups, other environmental exposure, such as life styles, and other related diseases, should also be considered. In our analysis, based on the results of the Eggers test and the funnel plot, it is likely that the genetic susceptibility to lung cancer associated with GSTM1 may have been overestimated because the publication bias cannot thoroughly be excluded. It is also unclear to what extent it might have been exaggerated in the published cases-control studies on the GSTM1 polymorphism and lung cancer risk. In addition to logistic regression analysis that was used to control the confounders, other statistical methods, such as Bayesian hierarchical models and the multilevel models, should also be considered in future analyses. 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 2006, 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. Therefore, although the test for publication bias was not statistically signicant, possible bias, especially the outcome-reporting bias, still could not be ruled out.

161

type of GSTM1 could alter susceptibility to lung cancer in some extent in Chinese populations.

Conict of interest
The authors promised there were not any possible conicts of interest in this research.

Acknowledgements
The authors wish to thank Dr. Qingyi Wei (Department of Epidemiology, The University of Texas, M.D. Anderson Cancer Center, Houston, TX) for his critical review and scientic editing of this manuscript, and this study was supported by a grant of Chinese National Key Basic Research and Development Program (2002CB512910).

References
[1] Neumann AS, Sturgis EM, Wei QY. Nucleotide excision repair as a marker for susceptibility to tobacco-related cancers: a review of molecular epidemiological studies. Mol Carcinog 2005;42:6592. [2] Kriek E, Van Schooten FJ, Hillebrand MJ, Van Leeuwen FE, Den Engelse L, De Looff AJ, et al. DNA adducts as a measure of lung cancer risk in humans exposed to polycyclic aromatic hydrocarbons. Environ Health Perspect 1993;99:715. [3] Li WY, Lai BT, Zhan XP. The relationship between genetic polymorphism of metabolizing enzymes and the susceptibility to lung cancer [in Chinese]. Chin J Epidemiol 2004;25:10425. [4] Vineis P, Husgafvel-Pursiainen K. Air pollution and cancer: biomarker studies in human populations. Carcinogenesis 2005;26:184655. [5] Mohr LC, Rodgers JK, Silvestri GA. Glutathione S-transferase M1 polymorphism and the risk of lung cancer. Anticancer Res 2003;23:211124. [6] Raimondi S, Paracchini V, Autrup H, Barros-Dios JM, Benhamou S, Boffetta P, et al. Meta- and pooled analysis of GSTT1 and lung cancer: a HuGE-GSEC review. Am J Epidemiol 2006;164:102742. [7] Reszka E, Wasowicz W. Signicance of genetic polymorphisms in glutathione S-transferase multigene family and lung cancer risk. Int J Occup Med Environ Health 2001;14:99113. [8] Mathijssen RH, Van Schaik RH. Genotyping and phenotyping cytochrome P450: perspectives for cancer treatment. Eur J Cancer 2006;42:1418. [9] Seow AL, Ng DP, Tan KW, Zhao B. CYP1A1 polymorphism, environmental exposures and lung cancer risk among non-smoking Chinese women in Singapore. In: AACR Meeting Abstracts. 2004. p. 1221. [10] Marchand LL, Guo C, Benhamou S, Bouchardy C, Cascorbi I, Clapper ML, et al. Pooled analysis of the CYP1A1 exon 7 polymorphism and lung cancer (United States). Cancer Causes Control 2003;14:33946. [11] Houlston RS. CYP1A1 polymorphisms and lung cancer risk: a meta-analysis. Pharmacogenetics 2000;10:10514. [12] Ye Z, Song HL, Higgins JP, Pharoah P, Danesh J. Five glutathione S-transferase gene variants in 23,452 cases of lung cancer and 30,397 controls: meta-analysis of 130 studies. PLoS Med 2006;3:e91. [13] Vineis P, Veglia F, Benhamou S, Butkiewicz D, Cascorbi I, Clapper ML, et al. CYP1A1 T3801C polymorphism and lung cancer: a pooled analysis of 2451 cases and 3358 controls. Int J Cancer 2003;104:6507.

5. Conclusion
In summary, our meta-analysis evaluated the relationship between genetic polymorphisms and lung cancer risk and revealed that variant genotypes of CYP1A1 and null geno-

162
[14] Egger M, Davey SG, Schneider M, Minder C. Bias in meta-analysis detected by a simple, graphical test. BMJ 1997;315:62934. [15] Hu ZB, Ma HX, Chen F, Wei QY, Shen HB. XRCC1 polymorphisms and cancer risk: a meta-analysis of 38 case-control studies. Cancer Epidemiol Biomarkers Prev 2005;14:18108. [16] Hu YL, Gao Y, Zhang Q. Genetic polymorphisms of CYP1A1, 2D6 and GSTM1 related with susceptibility to lung cancer [in Chinese]. Tumor 1998;18:26971. [17] Song N, Tan W, Tang HJ, Lin DG. Impact of cytochrome P450 gene mutations on the risk of development of lung cancer in a Chinese population [in Chinese]. Chin J Cancer 1999;18: 4958. [18] Song N, Tan W, Xing DY, Lin DG. CYP1A1 polymorphism and risk of lung cancer in relation to tobacco smoking: a case-control study in China. Carcinogenesis 2001;22: 116. [19] Yin L, Pu Y, Liu TY, Tung YH, Chen KW, Lin P. Genetic polymorphisms of NAD(P)H quinine oxidoreductase, CYP1A1 and microsomal epoxide hydrolase and lung cancer risk in Nanjing, China. Lung Cancer 2001;33:13341. [20] Zhou XW, Shi Y, Zhou YK. The relationship between CYP1A1 genetic polymorphism and susceptibility to lung cancer [in Chinese]. J Environ Occup Med 2002;19(355-7): 87. [21] Wang JW, Deng YF, Li L, Kuriki K, Ding J, Pan X, et al. Association of GSTM1, CYP1A1 and CYP2E1 genetic polymorphisms with susceptibility to lung adenocarcinoma: a case-control study in Chinese population. Cancer Sci 2003;94:44852. [22] Cai XL, Chen SD, Wang BG, Chen MX, Xian XZ. Genetic polymorphisms of CYP1A1 MspI susceptibility of lung cancer of Guangdong population [in Chinese]. Jiangsu J Prev Med 2003;14:12. [23] GuYF, Zhang SC, Lai BT, Wang H, Zhan XP. Relationship between genetic polymorphism of metabolizing enzymes and lung cancer susceptibility [in Chinese]. Chin J Lung Cancer 2004;7:1127. [24] Liang GY, Pu YP, Yin LH. Studies of the genes related to lung cancer susceptibility in Nanjing Han population, China [in Chinese]. Hereditas (Beijing) 2004;26:5848. [25] Dong CT, Yang Q, Jiang B, Dong QN. Study on the relationship between polymorphisms of CYP1A1 gene and susceptibility of lung cancer in Sichuan population [in Chinese]. Chin J Lung Cancer 2004;7:3842. [26] Chen SD, Ye WY, Chen Q. Relationship between CYP1A1 polymorphism, serum selenium and lung cancer [in Chinese]. Chin J Public Health 2004;20:7967. [27] Li Y, Tang XY, Ma XT. Cytochrome P24501A1 polymorphisms and susceptibility to lung cancer: a case-control study on Chinese population in Henan [in Chinese]. Henan Med Res 2005;14:2279. [28] Wang DQ, Chen SD, Wang BG, Cai XL, Xian XZ. A case control study on relationship between lung cancer and genetic polymorphism of CYP1A1, CYP2E1 and GSTM1 in Han nationality, in Guangzhou area [in Chinese]. Bulletin Chin Cancer 2006;15:57982. [29] Qian BY, Han HW, Gu F, He M, Li HX, Song FJ, et al. Case-control study genetic polymorphism in CYP1A1 and GSTM1 and smoking and susceptibility to lung cancer [in Chinese]. Chin J Clin Oncol 2006;33:5002. [30] Hu YL, Zhang Q. Genetic polymorphisms of CYP1A1 and susceptibility of lung cancer [in Chinese]. Chin J Med Gene 1999;16:268. [31] Chen SQ, Xu L, Ma GJ, Wu JZ, Xue KX. Identication of genetic polymorphism of CYP1A1 and GSTM1 and lung cancer patients by using allele-specic PCR and multiplex differential PCR [in Chinese]. Carcinog Teratog Mutag 1999;11:119 21.

X. Shi et al.
[32] Xue KX, Xu L, Chen SQ, Ma GJ, Chen JZ. The relationship between polymorphisms of the CYP1A1 and GSTM1 genes and the susceptibility to lung carcinoma in Chinese population [in Chinese]. Carcinog Teratog Mutag 1999;11: 326. [33] London SJ, Yuan JM, Coetzee GA, Gao YT, Ross RK, Yu MC. CYP1A1 I462V genetic polymorphism and lung cancer risk in a cohort of men in Shanghai, China. Cancer Epidemiol Biomarkers Prev 2000;9:98791. [34] Zhang LZ, Wang X, Shi YX. Susceptibility to lung cancer is associated with genetic polymorphism of cytochrome P450 (CYP1A1) [in Chinese]. Acta Acad Med Xuzhou 2001;21: 810. [35] Xue KX, Xu L, Chen SQ, Ma GJ, Wu JZ. Polymorphisms of the CYP1A1 and GSTM1 genes and their combined effects on individual susceptibility to lung cancer in a Chinese population [in Chinese]. Chin J Med Gene 2001;18:1257. [36] Zhang LZ, Wang X, Hao XZ, Shi YX, Liu ZH. Relationship between susceptibility to lung cancer and genetic polymorphism in CYP1A1, GSTM1 [in Chinese]. Chin J Clin Oncol 2002;29:53640. [37] Yang XR, Wacholder S, Xu Z, Dean M, Clark V, Gold B, et al. CYP1A1 and GSTM1 polymorphisms in relation to lung cancer risk in Chinese women. Cancer Lett 2004;214:197204. [38] Gao JR, Ren CL, Zhang Q. CYP2D6 and GSTM1 genetic polymorphism and lung cancer susceptibility [in Chinese]. Chin J Oncol 1998;20:1856. [39] Lan Q, He XZ, Costa DJ, Tian L, Rothman N, Hu G, et al. Indoor coal combustion emissions, GSTM1 and GSTT1 genotypes, and lung cancer risk: a case-control study in Xuan Wei, China. Cancer Epidemiol Biomarkers Prev 2000;9:6058. [40] Chen S, Xue K, Xu L, Ma G, Wu J. Polymorphisms of the CYP1A1 and GSTM1 genes in relation to individual susceptibility to lung carcinoma in Chinese population. Mutat Res 2001;458:417. [41] Chan Y, Wang X, Wang XY, Liang ZQ. A study of genetic polymorphism of GSTM1 gene in normal population and lung cancer population in Yunnan [in Chinese]. Acta Yunnan Normal Univ 2002;22:524. [42] Chan-Yeung M, Tan-Un KC, Ip MS, Tsang KW, Ho SP, Ho JC, et al. Lung cancer susceptibility and polymorphisms of glutathione-Stransferase genes in Hong Kong. Lung Cancer 2004;45:15560. [43] Luo CL, Chen Q, Cao WF, Chen SD. Combined analysis of polymorphism of GSTM1 and mutagens and p53 gene in the patients with lung cancer [in Chinese]. Chin J Clin Oncol 2004;31:121820. [44] Wang N, Zhuang DG, Wu YM. Research on relationship between GSTM1 gene deletion and lung cancer genetic susceptibility [in Chinese]. Acta Henan Univ Sci Technol (Med Vol) 2005;23:78. [45] Zeng M, Chen SD, Xie CM, Wang BG, Feng YX, Chen MX, et al. Case control study on relationship between lung cancer and its susceptibility marker [in Chinese]. Chin J Public Health 2005;21:7714. [46] Ye WY, Chen SD, Chen Q. Interaction between serum selenium level and polymorphism of GSTM1 in lung cancer [in Chinese]. Acta Nutrimenta Sinica 2005;27:1720. [47] Li DR, Zhou QH, Yuan TZ, Guo ZL, Zhu W, Wang YP, et al. Study on the association between genetic polymorphism of CYP2E1, GSTM1 and susceptibility of lung cancer [in Chinese]. Chin J Lung Cancer 2005;8:149. [48] Li Y, Tang XY, Ma XT, Chen J. Glutathione S-transferase M1 polymorphisms and susceptibility to lung cancer [in Chinese]. J Med Drug Forum 2005;26:102. [49] Wang QM, Lu QF, Zen HN, Bao M, Zhang HJ. The relationship between CYP2C9 and GSTM1 genetic polymorphism and lung cancer susceptibility [in Chinese]. Cancer Res Prev Treat 2006;33:810. [50] Bartsch H, Nair U, Risch A, Rojas M, Wikman H, Alexandrov K. Genetic polymorphism of CYP genes, alone or in combination,

CYP1A1 and GSTM1 polymorphisms and lung cancer risk in Chinese

as a risk modier of tobacco-related cancers. Cancer Epidemiol Biomarkers Prev 2000;9:328. [51] Perera FP, Mooney LA, Stampfer M, Phillips DH, Bell DA, Rundle A, et al. Associations between carcinogenDNA damage, glutathione S-transferase genotypes, and risk of lung cancer in the prospective physicians health cohort study. Carcinogenesis 2002;23:16416.

163

[52] Houlston RS. Glutathione S-transferase M1 status and lung cancer risk: a meta-analysis. Cancer Epidemiol Biomarkers Prev 1999;8:67582. [53] Nebert DW, Petersen DD, Puga A. Human AH locus polymorphism and cancer: inducibility of CYP1A1 and other genes by combustion products and dioxin. Pharmacogenetics 1991;1:6878.