Tumor Necrosis Factor and Interleukin-6 Gene Polymorphisms and Endometriosis Risk in Asians: A Systematic Review and Meta-Analysis

Authors

  • Jie Li,

    1. Guangxi Reproductive Medical Research Center, First Affiliated Hospital of Guangxi Medical University, Guangxi Zhuang Autonomous Region, China
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Yang Chen,

    1. Center for Genomic and Personalized Medicine, Guangxi Medical University, Nanning, Guangxi Zhuang Autonomous Region, China
    2. Institute of Urology and Nephrology, First Affiliated Hospital of Guangxi Medical University, Guangxi Zhuang Autonomous Region, China
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Shixiu Wei,

    1. Medical Research Center, Guangxi Medical University, Nanning, Guangxi, China
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Hongbo Wu,

    1. Guangxi Reproductive Medical Research Center, First Affiliated Hospital of Guangxi Medical University, Guangxi Zhuang Autonomous Region, China
    Search for more papers by this author
  • Chengjun Liu,

    1. Guangxi Jiangbin hospital, Nanning Guangxi Zhuang Autonomous Region, China
    Search for more papers by this author
  • Qiaoying Huang,

    1. Center for Genomic and Personalized Medicine, Guangxi Medical University, Nanning, Guangxi Zhuang Autonomous Region, China
    2. Medical Research Center, Guangxi Medical University, Nanning, Guangxi, China
    Search for more papers by this author
  • Liuming Li,

    Corresponding author
    1. Guangxi Reproductive Medical Research Center, First Affiliated Hospital of Guangxi Medical University, Guangxi Zhuang Autonomous Region, China
    • Corresponding author: Liuming Li, Guangxi Reproductive Medical Research Center, First Affiliated Hospital of Guangxi Medical University, Nanning, Guangxi Zhuang Autonomous Region 530021, China. Tel: 1 380 781 8680; E-mail: Liliuming1231@163.com; Yanling Hu, Medical Research Center, Guangxi Medical University, No. 22 Shuangyong Road, Nanning, Guangxi Zhuang Autonomous Region 530021, China. Tel: 1 321 134 8358; E-mail: ylhupost@163.com

    Search for more papers by this author
  • Yanling Hu

    Corresponding author
    1. Center for Genomic and Personalized Medicine, Guangxi Medical University, Nanning, Guangxi Zhuang Autonomous Region, China
    2. Medical Research Center, Guangxi Medical University, Nanning, Guangxi, China
    • Corresponding author: Liuming Li, Guangxi Reproductive Medical Research Center, First Affiliated Hospital of Guangxi Medical University, Nanning, Guangxi Zhuang Autonomous Region 530021, China. Tel: 1 380 781 8680; E-mail: Liliuming1231@163.com; Yanling Hu, Medical Research Center, Guangxi Medical University, No. 22 Shuangyong Road, Nanning, Guangxi Zhuang Autonomous Region 530021, China. Tel: 1 321 134 8358; E-mail: ylhupost@163.com

    Search for more papers by this author

Summary

A relationship between endometriosis and tumor necrosis factor (TNF-α) and interleukin-6 (IL-6) gene polymorphisms has been raised for Asians. However, this topic is controversial. This study was a meta-analysis to explore whether TNF-α/IL-6 gene polymorphisms were associated with a risk of endometriosis in Asians. By searching PubMed, HuGENet, and China National Knowledge Infrastructure (CNKI) databases, 17 studies were identified and included (3372 cases and 4008 controls). The odds ratio (OR) with 95% confidence interval (CI) was used to assess the association between TNF-α/IL-6 gene polymorphisms and endometriosis risk. An association of TNF-α gene -1031T/C polymorphism with endometriosis was found (TT + TC vs. CC: OR 0.50, 95% CI 0.30–0.82, I2 = 37.1%, P = 0.20; TT vs. CC: OR 0.50, 95% CI 0.30–0.82, I2 = 43.0%, P = 0.173; TC vs. CC: OR 0.49, 95% CI 0.29–0.83, I2 = 10.6%, P = 0.327). In addition, TNF-α-238A/G and IL-6 -174C/G gene polymorphisms were also likely to be associated with endometriosis in Asians. For the TNF-α-238A/G gene polymorphism, the OR was 1.577 (95% CI: 1.01–2.48). For the IL-6 -174C/G gene polymorphism, the OR was 1.554 (95% CI: 1.04–2.31). No associations were detected between the TNF-α-308A/G and IL-6 -634C/G polymorphisms and susceptibility to endometriosis. Our results indicate that the TNF-α gene -1031T/C polymorphism can reduce the risk of endometriosis, but for Asians, TNF-α-238A/G and IL-6 -174C/G gene polymorphisms may be a risk factor for endometriosis. No association was found for the TNF-α-308A/G and IL-6 -634C/G gene polymorphisms.

Introduction

Disease

In response to the influence of female hormones, the uterine cavity is lined with endometrial cells. Endometriosis is defined as the presence of endometrial cells at ectopic sites outside the uterine cavity, and it is a common gynecologic disorder in women (Govindan et al., 2007). Present in 15%–25% of women during their reproductive age, it can cause diverse clinical manifestations such as infertility, pelvic pain, and dysmenorrhea (Govindan et al., 2007; Kodati et al., 2008). However, the etiology and progression of endometriosis remain unclear, the diagnosis is difficult, and there is also no unified standard treatment for endometriosis. At present, some tentative treatment methods are used clinically, including hormonal treatments, pain medication, and surgery (Liu et al., 2013). To sum up, it is essential to explore the etiology and pathogenesis of endometriosis. Recently, a growing number of studies have been published (Mathur, 2000; Kennedy et al., 2001; Zondervan et al., 2001; Bischoff & Simpson, 2004) that suggest that mechanical, hormonal, immunologic, genetic, and environmental factors play a critical role in the etiology of endometriosis. In particular, many genetic association studies have concentrated on the associations between the development of endometriosis and some potential genetic polymorphisms (Chun et al., 2012; Gallegos-Arreola et al., 2012).

Gene

At present, studies on the relationship between gene polymorphism and the development of endometriosis are focused on the cytokines associated with immune regulation. Tumor necrosis factor (TNF-α) and interleukin-6 (IL-6) are mediators of the immune response; these proinflammatory cytokines have a broad range of biological activities (Chae et al., 2008; Mao et al., 2012). TNF-α and IL-6 gene polymorphisms have aroused widespread concern for endometriosis (Chae et al., 2008; Mao et al., 2012). Some previous studies (Eisermann et al., 1988; D'Hooghe et al., 2001; Bedaiwy et al., 2002; Koga et al., 2005; Cleophas et al., 2006) suggested that TNF-α and IL-6 were present in high concentrations in the peritoneal fluid of patients with endometriosis. As proinflammatory cytokines, TNF-α and IL-6 play a significant role in the regulation of cell proliferation, activation, motility, adhesion, and morphogenesis, which are associated with the inflammatory response, tissue repair, and neovascularization in endometriosis (Calhaz-Jorge et al., 2000).

The human TNF-α gene, which maps to Chromosome 6p21.3, is a significant source of genetic variability. A G to A substitution at position -238 (D'Alfonso & Richiardi 1994), a G to A substitution at position -308 (Wilson et al., 1992), and a C to T substitution at position -1031 (Higuchi et al., 1998) are in the proximal promoter of the TNF-α gene, and have already been examined for an association with endometriosis in previous studies. Also, some authors (Kitawaki et al., 2006; Wang et al., 2007; Chae et al., 2010) have suggested that the levels of IL-6 in the blood and peritoneal fluid of women with stages III–IV endometriosis are markedly increased. According to these controversial results, it is still unclear whether these polymorphisms in the TNF-α gene could affect the expression of the gene and could be related to the development of endometriosis.

The IL-6 gene is located on Chromosome 7p21. Within the IL-6 promoter region, some alleged single nucleotide polymorphisms (SNPs) are G-597A, G-572C, C-634G, and G-174C (Chae et al., 2008). Recently, several studies focused on the association between these polymorphisms in the IL-6 gene and endometriosis (Lee et al., 2002; Wieser et al., 2003; Bhanoori et al., 2005; Chae et al., 2010), especially the G-174C polymorphism and the C-634G polymorphism (rs1800796, previously denoted as -572G/C) (Wang et al., 2007). In different populations, an inconsistent correlation between these polymorphisms and endometriosis has been reported. At present, these studies have focused on Asians; the Caucasian population has not been investigated. Therefore, we have sufficient information and valid reason to explore in the Asian population the association between endometriosis risk and some of the polymorphisms. In addition, several genome-wide association studies (GWAS) have been performed for endometriosis in several loci (Adachi et al., 2010; Nyholt et al., 2012; Pagliardini et al., 2013), but TNF-α and IL-6 gene polymorphisms were not detected, which may be due to the strict significance threshold of 5 × 10−8 used to indicate genome-wide significance. Moreover, our large sample sizes were also worthy of research.

Therefore, to perform a comprehensive meta-analysis, it is essential to explore the association of the TNF-α gene and IL-6 gene polymorphisms and endometriosis in Asians. In this meta-analysis, we aimed to investigate whether TNF-α gene and IL-6 gene polymorphisms were associated with human endometriosis in Asians.

Materials and Methods

Search Strategy and Inclusion Criteria

The PubMed, HuGENet, and China National Knowledge Infrastructure (CNKI) databases were included in this search, with the combined keywords: “endometriosis,” “adenomyosis,” “EMT,” “EM,” “tumor necrosis factor,” “TNF,” “IL-6,” “interleukin-6,” and “polymorphisms.” The latest search was performed in May 2013, without any language restriction in the meta-analysis. In addition, the reference lists of relevant reviews and eligible articles were also checked. Articles published in the primary literature with human subjects were included. If the same case series were published in more than one article, the study with the most individuals investigated was selected and included only once in the analysis. The inclusion criteria were as follows: (a) case–control studies that quantitatively assessed the relationship of TNF-α/IL-6 polymorphism and risk of endometriosis in Asians; (b) studies that contained available data for examining the odds ratio (OR) with 95% confidence interval (CI); (c) studies including cases with endometriosis verified by laparotomy or laparoscopy and controls—patients without endometriosis verified by laparotomy or laparoscopy or healthy women with no genetic family history, no tumor, and without history of endometriosis; (d) the samples used in the measurements were blood. Major reasons for the exclusion of studies were as follows: (a) studies deficient in the raw data needed for statistical analysis and (b) duplicate studies.

All effective data from all eligible publications were independently extracted by two investigators and the two authors reached a consensus on all items. From each study, we collected the following characteristics: gene; the name of SNPs; authors; year of publication; country of the first or corresponding author; ethnicity (categorized as Caucasian or Asian); number of cases and controls; source of controls; genotyping methods; and evidence of Hardy–Weinberg equilibrium (HWE).

Statistical Analysis

In each case–control study, the OR and the corresponding 95% CI were used to assess the association between the TNF-α-1031T/C genotype and the risk of endometriosis in the dominant model (TT + TC vs. CC), recessive model (TT vs. TC + CC), homozygote comparison (TT vs. CC), and heterozygote comparison (TC vs. CC). TNF-α -238A/G, TNF-α -308A/G, IL-6 -634C/G, and IL-6 -174C/G genotypes were also evaluated in the same way. For each genetic analysis, the chi square (χ2)-based Q statistic was used to estimate the heterogeneity across all eligible comparisons and I2 values of 25%, 50%, and 75% were assigned as low, moderate, and high estimates. If I2 values were low or moderate (I2 < 50%), the fixed-effect model was considered; otherwise, the random-effect model was adopted among the statistics of meta-analysis.

In addition, we conducted a subgroup analysis to further explore the source of heterogeneity for IL-6 -634C/G genotypes. For the IL-6 -634C/G genotype, the subgroups of endometriosis stages III–IV cases were analyzed using the dominant model and recessive model and similar analyses were performed for group studies based on the case and control samples.

To assess publication bias, Begg's unweighted regression test and funnel plots were used (Egger et al., 1997). The evidence of population stratification was obtained by testing the distribution of genotypes in the control group using the χ2 method for HWE (P > 0.05) (Seagroatt & Stratton, 1998). Stata version 9.0 (Stata Corporation, College Station, TX, USA) was used in all statistical analysis. All the P values were two-sided.

Results

Characteristics of Eligible Studies

Using the keywords, 73 articles were retrieved in the initial search. In detail, 44 potentially relevant reports were for the TNF-α gene and 29 potentially relevant reports were for the IL-6 gene. After screening the titles and abstracts, 19 studies were excluded, which did not describe the association between TNF-α/IL-6 polymorphism and the risk of endometriosis. For further analysis, 54 full text articles were obtained. After thorough evaluation, 17 studies were found to be repeated in the database (Hsieh et al., 2002; Wieser et al., 2002, 2003; Asghar et al., 2004; Teramoto et al., 2004; Bhanoori et al., 2005; Kitawaki et al., 2006; Zhao et al., 2007; Chae et al., 2008; Lee et al., 2008; Lu et al., 2008; Hank Juo et al., 2009; Chae et al., 2010; Vijaya Lakshmi et al., 2010; Mao et al., 2011, 2012), 10 did not provide the raw data of TNF-α/IL-6 polymorphism that we needed for statistical analysis, and these studies reported specificity and sensitivity without the exact counts (Teramoto et al., 2004; Luo & He, 2006; Zhang et al., 2006; Chae et al., 2008; Lu et al., 2008; Zhang et al., 2008; Hank Juo et al., 2009; Zhou et al., 2010; Zhao et al., 2011; Kim et al., 2012), six studies were excluded because of duplication of data (Luo et al., 2006a, b; 2007a; Zhou et al., 2007a, b; Mao et al., 2012), one was not a study of the human (Zhen et al., 2002), and three studies were on Caucasians (Wieser et al., 2002, 2003; Zhao et al., 2007). Finally, 17 studies were selected (Hsieh et al., 2002; Lee et al., 2002; Asghar et al., 2004; Bhanoori et al., 2005; Kitawaki et al., 2006; Luo et al., 2007b; Wang et al., 2007; Lee et al., 2008; Chae et al., 2010; Che et al., 2010a, b; Vijaya Lakshmi et al., 2010; Zeng et al., 2010; Mao et al., 2011; Zeng & Huang, 2011; Mao et al., 2012) (Fig. 1). Also, the language was not restricted. In total, 7521 individuals were included in this case–control study, including 3372 surgically confirmed cases with histological evidence of advanced endometriosis, 3628 hospital patients without endometriosis (other gynecological diseases were not ruled out), and 380 healthy women with normal delivery and no history of endometriosis as clinical controls. The characteristics of the studies included in the meta-analysis are summarized in Table 1.

Table 1. The summary of studies of endometriosis risk with TNF-α/IL-6 gene promoter polymorphism
     Eligible subjectsGenotypes   
GeneSNP nameInvestigator, yearCountryPopulation of originCaseControlCaseControlSource of controlsMethodHWE
  1. PCR: single nucleotide polymorphism; RFLP: restriction fragment length polymorphism.

  2. SNP: single nucleotide polymorphisms, HWE: Hardy–Weinberg equilibrium.

TNF-α-1031T/C     TTTCCCTTTCCC   
  Mao et al. (2012)ChinaAsian4324992751342333215215Patients without endometriosis verified by laparotomy or laparoscopy.PCR0.6320
  Lee et al. (2008)KoreaAsian2462481408719149936Patients without endometriosis verified by laparotomy or laparoscopy.PCR-RFLP0.0517
  Asghar et al. (2004)JapanAsian8118063171118575Patients without endometriosis verified by laparotomy or laparoscopy.PCR0.5435
TNF-α-238A/G     AAAGGGAAAGGG   
  Asghar et al. (2004)JapanAsian1291751112707168Patients without endometriosis verified by laparotomy or laparoscopy.PCR0.7872
  Lee et al. (2002)KoreaAsian7020201456024178Patients without endometriosis verified by laparotomy or laparoscopy.PCR0.3694
  Zeng et al. (2010)ChinaAsian971410118606135Patients without endometriosis verified by laparotomy or laparoscopy.PCR-RFLP0.7963
  Che et al. (2010a)ChinaAsian200200014186010190The outpatient service charges no genetic family history, no tumor, and endometriosis.PCR0.7169
TNF-α-308A/G     AAAGGGAAAGGG   
  Asghar et al. (2004)JapanAsian1261790112505174Patients without endometriosis verified by laparotomy or laparoscopy.PCR0.8497
  Hsieh et al. (2002)ChinaAsian1201069258681880Patients without endometriosis verified by laparotomy or laparoscopy.PCR-RFLP<0.05
  Lee et al. (2002)KoreaAsian7020201060022180Patients without endometriosis verified by laparotomy or laparoscopy.PCR0.4131
  Luo et al. (2007b)ChinaAsian82801176402060Healthy women without gynecological disease.PCR-RFLP0.2013
  Zeng et al. (2010)ChinaAsian971411487538130Patients without a family history of genetic or endometriosis.PCR-RFLP<0.05
  Che et al. (2010b)ChinaAsian1001001128711089The outpatient service charges no genetic family history and no tumor and endometriosis.PCR0.2565
IL-6-634C/G     CCCGGGCCCGGG   
  Mao et al. (2011)ChinaAsian4324992741312727718834Patients without endometriosis verified by laparotomy or laparoscopy.PCR0.7843
  Chae et al. (2010)KoreaAsian3903512181482420412918Patients without endometriosis verified by laparotomy or laparoscopy.PCR-RFLP0.6796
  Kitawaki et al. (2006)JapanAsian20223613355141646210Patients without endometriosis verified by laparotomy or laparoscopy.PCR-RFLP0.1920
  Wang et al. (2007)ChinaAsian98995442270281Patients without endometriosis verified by laparotomy or laparoscopy.PCR-RFLP0.3197
IL-6-174C/G     CCCGGGCCCGGG   
  Bhanoori et al. (2005)South IndianAsian232210779146653151Patients without endometriosis verified by laparotomy or laparoscopy.PCR0.6088
  Lee et al. (2002)KoreaAsian70202016900202Patients without endometriosis verified by laparotomy or laparoscopy.PCRNA
  Wang et al. (2007)ChinaAsian989998009810Patients without endometriosis verified by laparotomy or laparoscopy.PCR-RFLP0.9597
Figure 1.

Process flow diagram of the selection of articles on genetic studies of the association of TNF-α/IL-6 gene polymorphisms and endometriosis.

For the TNF-α -1031T/C polymorphism, three studies were eligible (759 cases and 927 controls). For the TNF-α -238A/G polymorphism, four studies were eligible (496 cases and 718 controls). For the TNF-α -308A/G polymorphism, six studies were eligible (595 cases and 808 controls). For the IL-6 -634C/G polymorphism, four studies were eligible (1122 cases and 1185 controls). For the IL-6 -174C/G polymorphism, three studies were eligible (400 cases and 511 controls). In addition, two studies were excluded because the genotype distribution in controls disagreed with the HWE for the -308A/G polymorphism (P < 0.05) (Hsieh et al., 2002; Zeng et al., 2010). For the TNF-α -1031T/C, TNF-α -238A/G, IL-6 -634C/G, and IL-6 -174C/G polymorphisms, the genotype distribution in all control populations was in agreement with the HWE in eligible studies.

Meta-Analysis

TNF-α gene -1031T/C polymorphism

A significant association was found between the TNF-α gene -1031T/C polymorphism and endometriosis (fixed-effect model: TT + TC vs. CC: OR 0.50, 95% CI 0.30–0.82, I2 = 37.1%, P = 0.20; TT vs. CC: OR 0.50, 95% CI 0.30–0.82, I2 = 43.0%, P = 0.173; TC vs. CC: OR 0.49, 95% CI 0.29–0.83, I2 = 10.6%, P = 0.327). The OR for the risk of endometriosis associated with the T allele is shown in Table 2 and Figure 2.

Table 2. Meta-analysis of TNF-α/IL-6 gene polymorphisms and endometriosis
GeneSNP nameComparisonsOR95%CII2 (%)P for heterogeneity
  1. SNP: Single nucleotide polymorphisms; OR: odds ratios; CI: confidence intervals.

  2. NA: Not appropriate.

TNF-α-1031T/CTT+TC vs. CC0.500.30–0.8237.10.204
  TT vs. TC+CC0.960.79–1.1860.20.081
  TT vs. CC0.500.30–0.8343.00.173
  TC vs. CC0.490.29–0.8310.60.327
 -238A/GAA+AG vs. GG1.581.01–2.4835.60.198
  AA vs. AG+GG4.100.17–101.39NANA
  AA vs. GG3.970.16–98.13NANA
  AG vs. GG1.540.98–2.4245.50.138
 -308A/GAA+AG vs. GG1.000.66–1.560.00.509
  AA vs. AG+GG1.660.22–12.680.00.616
  AA vs. GG1.640.21–12.610.00.641
  AG vs. GG0.980.63–1.540.00.457
IL-6-634C/GCC+CG vs. GG0.940.64–1.340.00.499
  CC vs. CG+GG1.020.86–1.2175.10.007
  CC vs. GG0.910.64–1.312.30.381
  CG vs. GG0.820.56–1.190.00.939
 -174C/GCC+CG vs. GG1.551.04–2.3111.70.287
  CC vs. CG+GG1.200.43–3.380.00.547
  CC vs. GG1.210.40–3.68NANA
  CG vs. GG1.591.06–2.409.30.294
Figure 2.

Meta-analysis with the dominant model (TT + TC vs. CC) and fixed effects for the association between endometriosis risk and the TNF-α gene -1031T/C polymorphism. The name of the first author and year of publication of each study are shown. The results of individual and summary odds ratio (OR) estimates, 95% confidence interval (CI), and weight (W) of each study show the association between endometriosis risk and the TNF-α gene -1031T/C polymorphism. Horizontal lines represent 95% CI and dotted vertical lines represent the value of the summary OR.

-238A/G, -308A/G polymorphisms for the TNF-α gene and -634C/G, -174C/G polymorphisms for the IL-6 gene

Based neither on the fixed-effect model nor on the random-effect model, a significant association was detected in the analysis for the TNF-α -238A/G and IL-6 -174C/G gene polymorphisms in Asians. For the TNF-α -238A/G gene polymorphism, the summary OR was 1.58 (95% CI: 1.01–2.48) for Asians in the fixed-effect model (Fig. 3). For the IL-6 -174C/G gene polymorphism, the summary OR was 1.56 (95% CI: 1.044–2.313) for Asians in the fixed-effect model (Fig. 4). However, no association was observed between the TNF-α -308A/G and IL-6 -634C/G gene polymorphisms and endometriosis risk in the dominant model, recessive model, homozygote comparison, or heterozygote comparison. The meta-analysis of the association between TNF-α/IL-6 gene polymorphisms and endometriosis is summarized in Table 2. We did not detect any association in endometriosis stages III–IV-specific analysis for the IL-6 -634C/G gene polymorphism. The subgroup analysis for the association between the IL-6 -634C/G gene polymorphism and endometriosis is summarized in Table 3.

Table 3. Stratified analyses of the TNF-a/IL-6 gene polymorphisms on endometriosis risk
   Dominant modelRecessive model
GeneSNP nameEndometriosis stageOR(95% CI)I2 (%)P for heterogeneityOR(95% CI)I2 (%)P for heterogeneity
  1. Random-effects model was used when I2 > 50%; otherwise, fixed-effects model was used. NA: not appropriate.

IL-6-634C/GIII–IV0.70(0.42–1.15)0.00.6610.82 (0.65–1.02)38.10.199
Figure 3.

Meta-analysis with the dominant model (AA + AG vs. GG) for the association between endometriosis risk and the TNF-α -238A/G gene polymorphism in the Asian population.

Figure 4.

Meta-analysis with the dominant model (CC + CG vs. GG) for the association between endometriosis risk and the IL-6 -174C/G gene polymorphism in the Asian population.

To assess publication bias, the Egger's funnel plot was estimated and shown to be symmetrical. Using Begg's test, we did not detect a significant bias (P = 0.634).All the statistical results for the association between TNF-α/IL-6 gene polymorphisms and the risk of endometriosis are shown in Tables 2 and 3.

Discussion

As proinflammatory cytokines, TNF-α and IL-6 play a significant role in regulating cell proliferation, activation, motility, adhesion, and morphogenesis, which are associated with the inflammatory response, tissue repair, and neovascularization in endometriosis (Calhaz-Jorge et al., 2000). Some studies indicated that TNF-α and IL-6 were related to the pathogenesis of endometriosis (Calhaz-Jorge et al., 2000; Maas et al., 2001). However, in terms of nonsurgical diagnosis of endometriosis, whether these cytokines have any predictive value has not been determined.

The TNF-α -1031T/C polymorphism is relative to the transcription start site of the TNF gene, which can affect TNF regulation and transcriptional efficiency directly (Lee et al., 2008), thereby affecting the production of TNF-α and promoting the development of endometriosis. However, the association between the TNF-α gene -1031T/C polymorphism and endometriosis is controversial (Asghar et al., 2004; Lee et al., 2008; Mao et al., 2012). In our meta-analysis, strong evidence was found to support that the TNF-α gene -1031T/C polymorphism has a significant association with endometriosis, especially in the dominant model, in the homozygote comparison TT versus CC analysis, and in the heterozygote comparison TC versus CC analysis. These results suggest that the TNF-α gene -1031T/C polymorphism may reduce endometriosis risk. In addition, several researchers suggested that the TNF-α gene -1031T/C polymorphism could be associated with susceptibility to autoimmune diseases, such as Crohn's disease (CD) and ulcerative colitis (UC) (Negoro et al., 1999). As the value of the heterogeneity was low, a fixed-effect model analysis was conducted. The Q test of TNF-α gene -1031T/C polymorphism heterogeneity was not significant (TT + TC vs. CC: P heterogeneity = 0.204, I2 = 37.1%; TT vs. CC: P heterogeneity = 0.173, I2 = 43.0%; TC vs. CC: P heterogeneity = 0.327, I2 = 10.6%). In the fixed-effect model, the TNF-α gene -1031T/C polymorphism showed a protective association with endometriosis (TT + TC vs. CC: OR 0.50, 95% CI 0.30–0.82, P = 0.20; TT vs. CC: OR 0.50, 95% CI 0.30–0.82, P = 0.173; TC vs. CC: OR 0.49, 95% CI 0.29–0.83, P = 0.327). In conclusion, strong evidence was found to prove that the TNF-α gene -1031T/C polymorphism may reduce the risk of endometriosis.

As TNF-α and IL-6 are proinflammatory cytokines, TNF-α -238A/G and IL-6 -174C/G gene polymorphisms are associated with the inflammatory response, tissue repair, and neovascularization in endometriosis (Calhaz-Jorge et al., 2000), and also play a significant role in regulating cell proliferation, activation, and morphogenesis. In our meta-analysis, our results showed that the -238A/G and -174C/G polymorphisms had a significant association with endometriosis in Asians; the summary OR was 1.58 (95% CI: 1.01–2.48) for the TNF-α -238A/G gene polymorphism and the summary OR was 1.56 (95% CI: 1.04–2.31) the for IL-6 -174C/G gene polymorphism. These results suggest that the TNF-α -238A/G and IL-6 -174C/G gene polymorphisms are likely to be risk factors for the susceptibility to endometriosis for Asians. In this analysis, no association was detected for the TNF-α -308A/G gene polymorphism, excluding the studies that deviated from the HWE (Hsieh et al., 2002; Zeng et al., 2010). The results are supported by Lee et al. (2002), Asghar et al. (2004), and Che et al. (2010b). Thus, the TNF-α -308A/G polymorphism might not be a useful prognostic biomarker for endometriosis, especially in the Asian population. For the IL-6 -634C/G genotype, no association was found with endometriosis in Asians and the results are supported by Chae et al. (2010), who also found no association of the -634C/G gene polymorphism with endometriosis, or with any clinical subgroup of endometriosis. The subgroup of endometriosis stages III–IV cases was considered under the condition that one study (Chae et al., 2010) selected endometriosis stages III and IV as cases. This subgroup analysis could reduce the deviation due to inconsistent stages of endometriosis in different included studies, but we still did not detect any association between the subgroup and the IL-6 -634C/G genotype. Thus, in the Asian population, IL-6 -634C/G might not be a useful prognostic biomarker for endometriosis. For the TNF-α -308A/G and IL-6 -634C/G polymorphisms, no association was found with endometriosis in Asians. The Q test of heterogeneity might be one of the reasons for this; potential sources of heterogeneity were significant. Endometriosis susceptibility was affected by the prevalence of polymorphisms in the population and exposure to environmental factors. In addition, the number of studies on the relationship between TNF-α and IL-6 polymorphisms, and the risk of endometriosis, were too small, which has also affected the reliability of these results. In our meta-analysis, we only included publications on the Asian population and those on the Caucasian population were excluded, because we could not draw firm conclusions for the Caucasian population, since only three studies have been conducted (Wieser et al., 2002, 2003; Zhao et al., 2007). Thus, the different genetic and environmental backgrounds should be considered and ethnicity was also considered as a potential source of heterogeneity. In particular, further studies should investigate the Caucasian population and compare it with the Asian population.

In this analysis, there were 17 repeated studies in the database; most repeated studies had the same samples and sources, but the purposes of repeated studies were not the same except for the study population. For example, Mao et al. (2011) aimed to explore the association of the IL-6 gene -634C/G polymorphism with susceptibility to endometriosis and the case–control study was performed in 432 endometriosis patients and 499 control women to evaluate the SNP of IL-6 -634C/G. These samples were from October 2008 to April 2010 at Nanfang Hospital, Zhujiang Hospital, and the First People's Hospital of Foshan of China; Mao et al. (2012) aimed to explore the association of the TNF-α -1031T/C polymorphism and in combination with the IL-6 -634C/G gene polymorphism with susceptibility to endometriosis. In total, 432 endometriosis patients and 499 nonendometriosis women were included, but these samples were also from October 2008 to April 2010 at Nanfang Hospital, Zhujiang Hospital, and the First People's Hospital of Foshan of China. In this condition, we retained only one article and excluded the repeated studies, which could exclude data repeatability and lead to more accurate results.

As with all meta-analyses, possible limitations should be considered in this analysis. First, we cannot ignore the other potential sources of heterogeneity, and in subgroup analyses and meta-regression, we were unable to explore them further because some factors were not evaluated in most studies, such as the age of patients, the number of cases and controls, the number of cases of each endometriosis stage, and the lifestyle of the population. Second, the mechanism of endometriosis was not clear, and we should take SNP–SNP, gene–gene, and gene–environment interactions into account. Therefore, different genetic and environment backgrounds might have been a cause of discrepancy and the gene–environment interactions might have influenced the results of each study. In addition, the design of the included studies was different so that the collection of complete data was difficult.

In accordance with the revised American Society for Reproductive Medicine classification (ASRM), the classification of endometriosis patients was staged among the included 17 studies, but only stages III and IV were selected as cases in some studies, and several studies also included endometriosis with other gynecologic diseases as cases. This deviation might contribute to heterogeneity. However, considering the small sample size for the IL-6 -634C/G and IL-6 -174C/G polymorphisms, this might be the main potential source of heterogeneity, and there may not have been enough power to explore interactions. Lastly, because no prospective studies were found on this topic, in our meta-analysis, all of the included studies were retrospective case–control designs. Therefore, we cannot exclude the possibility of undetected bias.

Conclusion

The association of TNF-α/IL-6 gene polymorphisms with endometriosis risk was researched in our study. In this meta-analysis, we found strong evidence to support that the TNF-α gene -1031T/C polymorphism reduces the risk of endometriosis and may be a useful marker for the pathogenesis of endometriosis in Asians. The TNF-α -238A/G and IL-6 -174C/G gene polymorphisms are likely to be risk factors for susceptibility to endometriosis for Asians. On the other hand, no associations were detected between the TNF-α -308A/G and IL-6 -634C/G polymorphisms and endometriosis. Due to the small number of studies included, our results should be interpreted with caution and more larger-scale, well-designed, and population-based studies are necessary to explore the roles played by these polymorphisms of the TNF-α and IL-6 genes in the pathogenesis of endometriosis, especially for Caucasians.

Acknowledgements

This study was supported by grants from National Natural Science Foundation (81272853), Guangxi Health department (Gui Wei Zhong 2008035), National Natural Science Foundation (81060213), Guangxi Natural Science Foundation (2011GXNSFB018100), and Guangxi scientific research and technology development project (1298003-6-4).

Ancillary