- Email: [email protected]
A) with recurrent spontaneous abortion: a meta-analysis
Human Immunology 73 (2012) 574–579
Contents lists available at SciVerse ScienceDirect
Association of tumor necrosis factor-␣ gene promoter polymorphisms (-308G/A, -238G/A) with recurrent spontaneous abortion: a meta-analysis Bingzhen Zhang, Tiecheng Liu, Zhiping Wang* Department of Epidemiology, Health Statistics, School of Public Health, Shandong University, Jinan, China
A R T I C L E
I N F O
Article history: Received 11 August 2011 Accepted 24 January 2012 Available online 14 February 2012
Keywords: Recurrent spontaneous abortion Tumor necrosis factor-␣ Gene polymorphism Meta-analysis
A B S T R A C T
Tumor necrosis factor-␣ (TNF-␣) gene promoter polymorphisms (-308G/A, -238G/A) have been associated with increased recurrent spontaneous abortion (RSA) risk, but the results of published articles are controversial. Hence, a meta-analysis was performed to assess the effect of TNF-␣ -308G/A, -238G/A polymorphisms on RSA risk. Heterogeneity testing and sensitivity analysis were performed using RevMan 5.0 software. Publication bias was assessed by the funnel plot method and modified Egger’s linear regression test. In 12 studies for the TNF-␣ -308G/A polymorphism, the summary odds ratio (OR) with the corresponding 95% confidence interval (95% CI) was 1.04 (95% CI: 0.86, 1.26) under a fixed-effect model in the overall population. In 5 studies for the TNF-␣ -238G/A polymorphism, the summary OR with the corresponding 95% CI was 1.11 (95% CI: 0.60, 2.03) under a random-effect model in the overall population. We could not identify the sources of heterogeneity for TNF-␣ -238G/A. In addition, no evidence of publication bias was detected. The results of this meta-analysis indicate that TNF-␣ -308G/A, -238G/A polymorphisms are not significantly associated with the risk of RSA in the overall population. However, more convincing evidence is required to draw a solid conclusion on the relation between the TNF-␣ -238G/A polymorphism and the risk of RSA. 䉷 2012 American Society for Histocompatibility and Immunogenetics. Published by Elsevier Inc. All rights reserved.
1. Introduction Recurrent spontaneous abortion (RSA) is defined as the occurrence of 3 or more clinically detectable pregnancy losses before the 20th week of gestation with the same partner [1] and is a major reproductive problem affecting 1–3% of otherwise healthy women [2]. The diagnosis of RSA involves multiple tests to detect parental chromosomal anomalies; maternal immunologic, hormonal, or thrombotic disorders; and anatomic abnormalities of the genital tract. Yet 40% of the affected women who do not exhibit any of the above-mentioned disorders are classified as unexplained RSA with unknown etiology [3]. Studies have indicated that the level of tumor necrosis factor-␣ (TNF-␣) apparently contributes to the pathogenesis of RSA [4,5] and higher serum levels of TNF-␣ were detected in RSA groups [6]. The TNF-␣ gene is located in 6p21.3 and has several functional polymorphism sites. Nucleotides -308 (rs1800629) and -238 (rs361525) in the promoter region of the TNF-␣ gene are 2 common functional polymorphisms that have been demonstrated to be associated with the production level of the cytokine [7–9]. The production of TNF-␣ is partly under genetic control [10], especially in the promoter region [11]. Although many studies have evaluated
the associations between TNF-␣ -308G/A, -238G/A polymorphisms and RSA, the results remain controversial. Hence, a meta-analysis was performed to assess the association of TNF-␣ -308G/A, -238G/A gene polymorphisms with the risk of RSA.
2. Subjects and methods 2.1. Search strategy Studies were identified by searches of PubMed, ISI Web of Science, CHINAINFO (Digital Journal Full-text Database), VIP (Database of Chinese Scientific and Technical Periodicals), the China Biology Medical Literature Database, and the China National Knowledge Infrastructure for relevant studies published in English or Chinese before April 2011 using the following search terms: cytokine gene polymorphisms and tumor necrosis factor gene polymorphisms, in combination with recurrent spontaneous abortion (RSA), recurrent pregnancy loss (RPL), recurrent miscarriage (RM), and recurrent fetal loss (RFL). In addition, the bibliographies of relevant studies, review articles, and meta-analysis articles were considered. The article search was performed independently by 2 investigators. 2.2. Inclusion criteria
* Corresponding author. E-mail address: [email protected] (Z. Wang).
The inclusion criteria were as follows: (1) RSA with clinical diagnosis criteria. RSA is defined as the occurrence of 3 or more clinically detectable pregnancy losses before the 20th week of
0198-8859/12/$36.00 - see front matter 䉷 2012 American Society for Histocompatibility and Immunogenetics. Published by Elsevier Inc. All rights reserved. doi:10.1016/j.humimm.2012.01.008
B. Zhang et al. / Human Immunology 73 (2012) 574–579
gestation. Furthermore, only unexplained or idiopathic abortions were included; (2) studies included data on the TNF-␣ -308G/A and/or -238G/A genotype(s); (3) case– control or cohort study published as an original study; (4) numbers in case and control groups for each genotype, or available data that can be used to calculate the numbers, should be reported; and (5) studies published in English or Chinese. The flow diagram for selection of studies is presented in Fig. 1.
2.3. Data extraction Two investigators independently extracted data and reached a consensus on all items. For each study, the following data were extracted: first author’s last name, country in which the study was performed, year of publication, mean ages in case and control groups if reported, and the frequency of each genotype in case and control groups
2.4. Statistical analysis Hardy–Weinberg equilibrium for TNF-␣ -308G/A, -238G/A genotype distributions in controls was tested using Pearson’s 2 test. The odds ratios (ORs) with the corresponding 95% confidence intervals (CIs) of RSA were considered under the dominant genetic models (AA plus GA vs GG). Heterogeneity among studies was assessed by the Q test and the I2 of Higgins and Thompson [12]. In the presence of heterogeneity (I ⬎ 50%), the random-effect model was adopted as the pooling method; otherwise, the fixed-effect model was used. The sensitivity analysis was performed by omitting each individual study, and an individual study was suspected of excessive sensitivity if the point estimate of its omitted analysis was outside the 95% CI of the combined analysis. We also conducted a stratified analysis for -308G/A based on the geographic region of the studies. In addition, a cumulative meta-analysis based on the size of the sample was conducted to evaluate the trends of summary OR with the corresponding 95% CI for the association, allowing us to see how the measured genetic effect changed when updated evidence accumulated. Publication bias was assessed by the funnel plot method and modified Egger’s linear regression test as proposed by Harbord et al. [13]. Data were analyzed by Review Manger (Cochrane Corp., Nordic Cochrane Center) 5.1.2 and Stata (Stata Corp., College Station, TX) version 11.1 software. All reported probabilities (p values) were 2-sided and a p value ⬍0.05 was considered significant.
575
3. Results 3.1. Characteristics of included studies Twelve studies [14 –25] fulfilled the inclusion criteria, including 12 studies [14 –25] on the -308G/A polymorphism (including 1,380 cases and 1,595 controls) and 5 studies [15,22–25] on the -238G/A polymorphism (including 884 cases and 912 controls) in the TNF-␣ gene. The TNF-␣ -308G/A, -238G/A genotype frequencies in controls were in Hardy–Weinberg equilibrium in all except 1 study [23]. Among these 12 studies, Daher et al. [17] and Prigoshin et al. [19] reported that only primary aborters were included in the cases; the other studies reported that both primary and secondary aborters were included in the cases [16,18,22,23] or did not clarify whether primary or secondary aborters were included [14,15, 20,21,24,25], so both primary and secondary aborters were included in the cases of this meta-analysis. General characteristics and the TNF-␣ -308G/A and -238G/A polymorphism genotype distributions in the published studies included in this meta-analysis are presented in Tables 1 and 2. 3.2. Quantitative synthesis 3.2.1. TNF-␣ -308G/A polymorphism and risk of RSA Fig. 2A illustrates the results of individual and summary ORs and the corresponding 95% CIs under the dominant genetic model in the overall population. Because no heterogeneity among studies was observed (I ⫽ 39%), the summary OR with 95% CI based on the fixed-effects model was 1.04 (95% CI: 0.86, 1.26). The cumulative meta-analysis indicated that the summary OR tended to 1 as the studies joined in one by one (Fig. 2B). In addition, we performed stratified analysis based on the geographic region of studies (because some studies did not report the ethnicity of the cases and controls, we could not perform stratified analysis based on ethnicity), but no significant association between the TNF-␣ 308 G/A polymorphism and RSA was observed in Europe, Asia, and other countries (Fig. 2C). 3.2.2. TNF-␣ -238G/A polymorphism and risk of RSA Fig. 3A illustrates the results of individual and summary ORs with the corresponding 95% CIs under the dominant genetic model in the overall population. Because heterogeneity among studies was observed (I ⫽ 74%), the summary OR with 95% CI based on the random-effects model was 1.11 (95% CI: 0.60, 2.03). However, after omitting the study of Liu et al. [24], the I2 values decreased (I ⫽ 45%) and the summary OR with 95% CI based on the fixed-effects model was 1.51 (95% CI: 1.16, 1.96; Fig. 3B). The results indicate that the TNF-␣ -238G/A polymorphism is associated with RSA when we decreased the heterogeneity of the studies. 3.3. Sensitivity analysis For each polymorphism, the results demonstrate that no evidence of any individual study having excessive influence on the summary effect was observed under the dominant genetic models (data not shown). 3.4. Publication bias evaluation Fig. 4A and 4B illustrate that no publication bias was detected for -308 G/A, -238G/A polymorphisms using the funnel plot method and the modified Egger linear regression test indicated that no significant publication bias was observed (p ⫽ 0.122 and 0.360, respectively). 4. Discussion
Fig. 1. The flow diagram for selection of studies.
The associations between the TNF-␣ gene promoter polymorphisms (-308G/A, -238G/A) and RSA have been widely studied. However, some studies have indicated that the polymorphisms were risk factors for RSA, whereas other studies drew converse
576
B. Zhang et al. / Human Immunology 73 (2012) 574–579
Table 1 Characteristics of studies on the association between the tumor necrosis factor-␣ -308G/A polymorphism and recurrent spontaneous abortion Author
Year
Babbage et al. [14] Baxter et al. [15] Reid et al. [16] Daher et al. [17] Pietrowski et al. [18] Prigoshin et al. [19] Kamali-Sarvestani et al. [20] Zhou et al. [21] Zammiti et al. [22] Finan et al. [23] Liu et al. [24] Palmirotta et al. [25]
2001 2001 2001 2003 2004 2004 2005 2006 2009 2010 2010 2010
Country
Age (mean)
UK UK UK Brazil Austria Argentina Iran China Tunisia Bahrain China Italy
Genotype (AA/GA/GG)
p for HWE
Case
Control
Case
Control
33 34 — 35 33 33 — 27.4 29.2 32.2 30.1 37
44 32 — — 58 — — — 29.0 31.8 28.8 38
1/12/30 3/22/51 2/6/7 1/11/36 2/33/133 0/6/35 0/14/117 0/8/42 14/39/319 8/32/164 0/22/110 0/13/87
1/16/56 5/40/93 1/13/29 1/18/89 4/41/167 0/5/49 0/21/122 0/5/45 5/47/222 4/32/212 1/13/138 3/21/76
0.906 0.787 0.744 0.932 0.432 0.721 0.343 0.71 0.187 0.041 0.276 0.313
HWE, Hardy–Weinberg equilibrium.
conclusions. Hence, we performed a meta-analysis to clarify the association between the 2 polymorphisms and the risk of RSA. In this meta-analysis, 12 studies (12 studies for -308G/A, 5 studies for -238G/A) on TNF-␣ gene promoter polymorphisms were performed to provide the most comprehensive assessment of the association between the 2 polymorphisms and the risk of RSA. The available data suggested that TNF-␣ -308G/A, -238G/A polymorphisms were not significantly associated with the risk of RSA when the effect of the polymorphisms was considered under the dominant genetic models in the overall population. The results of this meta-analysis were consistent only with the studies of Medica et al. [26] and Bombell and McGuire [27]. Medica et al. [26] did not observe any association between the TNF-␣ -308G/A polymorphism and RSA with 524 cases and 771 controls. Moreover, Bombell and McGuire [27] did not describe a significant effect of the -308G/A polymorphism on the risk of RSA with 630 cases and 875 controls. These results indicated that the TNF-␣ -308G/A polymorphism is not associated with the risk of RSA, consistent with the results of the cumulative meta-analysis. Moreover, the results were stable with the increase in sample size. In addition, we did not observe evidence of an association between the -308G/A polymorphism and RSA by stratified analysis based on geographic region. For the TNF-␣ -238G/A polymorphism, this was the first metaanalysis; no association was observed in the overall population. Because only 5 studies were included in this meta-analysis, we could not perform stratified analysis based on geographic region or ethnicity. Interestingly, among these 5 studies, Liu et al. [24] demonstrated that the G allele of the -238G/A polymorphism was a risk factor for RSA. When we excluded this study to decrease the heterogeneity of the studies, we determined that the AA plus GA versus GG genetic model of the -238G/A polymorphism was associated with a higher risk of RSA. We supposed that the -238G/A polymorphism might be associated with the risk of RSA; however, more convincing evidence, such as large sample size, number of studies, and ethnicities, is required to draw a solid conclusion. There were some limitations in this meta-analysis. First, the estimation on the -238G/A polymorphism in the overall population
was obtained by pooling the studies with a high degree of heterogeneity, but we could not perform meaningful stratified analysis because a small number of studies were included. Second, genotyping accuracy and quality control measures were not well described in some studies. Third, only studies published in English and Chinese were included in this meta-analysis, which might limit the results. A meta-analysis should be the appropriate method to obtain a more precise conclusion and allowed us to exclude modest associations for gene polymorphisms. We did not observe evidence that -308G/A, -238G/A polymorphisms in the TNF-␣ gene were associated with the risk of RSA in the overall population. The lack of association did not rule out the possibility that these polymorphisms were important in the pathogenesis of RSA, because genetic heterogeneous backgrounds might be to a certain extent responsible for the heterogeneity of the effect on RSA. However, differences in cytokine production induced by gene polymorphisms are not always indicative of immune/inflammatory events that occur at the placental interface during pregnancies/miscarriage, given that the cytokines (which generally act over a short interval) exerted their effects both in a paracrine and in an autocrine manner [25]. TNF-␣ gene -308G/A, -238G/A polymorphisms may be just regulators of the production of TNF-␣ at the genetic level, and the production of this cytokine may be regulated by other factors. Furthermore, Finan et al. [23] reported that significantly higher frequencies of TNF-␣ -1031C and -376A alleles were observed in RSA patients and significant differences were also noted in the distribution of -1031T/C and -376G/A genotypes between case and control subjects. Hence, the effect of TNF-␣ -1031T/C and -376G/A polymorphisms on the risk of RSA should be given more attention in the future. In conclusion, we did not observe any evidence that TNF-␣ gene -308G/A, -238G/A polymorphisms had a significant effect on the risk of RSA in the overall population. Because potential biases and confounders in this meta-analysis could not be ruled out completely, these results must be confirmed by further studies.
Table 2 Characteristics of studies on the association between the tumor necrosis factor-␣ -238G/A polymorphism and recurrent spontaneous abortion Author
Baxter et al. [15] Zammiti et al. [22] Finan et al. [23] Liu et al. [24] Palmirotta et al. [25]
Year
2001 2009 2010 2010 2010
HWE, Hardy–Weinberg equilibrium.
Country
UK Tunisia Bahrain China Italy
Age (mean)
Genotype (AA/GA/GG)
p for HWE
Case
Control
Case
Control
34 29.2 32.2 30.1 37
32 29.0 31.8 28.8 38
0/3/73 20/88/264 4/52/148 0/4/128 0/16/84
0/12/126 7/52/215 0/48/200 0/17/135 0/6/94
0.593 0.084 0.092 0.465 0.757
B. Zhang et al. / Human Immunology 73 (2012) 574–579
577
Fig. 2. Results of individual and summary odds ratios (ORs) with 95% confidence intervals (CIs) of recurrent spontaneous abortion (RSA) for the tumor necrosis factor-␣ (TNF-␣) -308 G/A polymorphism under the dominant genetic model. (A) The results of a meta-analysis for the TNF-␣ -308 G/A polymorphism in the overall population; (B) the results of a cumulative meta-analysis for the TNF-␣ -308 G/A polymorphism; and (C) the results of a stratified meta-analysis for the TNF-␣ -308 G/A polymorphism.
578
B. Zhang et al. / Human Immunology 73 (2012) 574–579
Fig. 3. Results of individual and summary odds ratios (ORs) with 95% confidence intervals (CIs) of recurrent spontaneous abortion (RSA) for the tumor necrosis factor-␣ (TNF-␣) -238 G/A polymorphism under the dominant genetic model. (A) The results of a meta-analysis for the TNF-␣ -238 G/A polymorphism in the overall population; and (B) the results of a meta-analysis for the TNF-␣ -238 G/A polymorphism after omitting the study by Liu et al. [24].
References
Fig. 4. Funnel plots of the odds ratios (ORs) of recurrent spontaneous abortion for tumor necrosis factor-␣ (TNF-␣) -308G/A, -238G/A polymorphisms for each study included in this meta-analysis. Funnel plot for (A) the TNF-␣ -308G/A polymorphism and (B) the TNF-␣ -238G/A polymorphism.
[1] Sierra S, Stephenson M. Genetics of recurrent pregnancy loss. Semin Reprod Med 2006;24:17–24. [2] Brown S. Miscarriage and its associations. Semin Reprod Med 2008;26:391– 400. [3] Stephenson MD. Frequency of factors associated with habitual abortion in 197 couples. Fertil Steril 1996;66:24 –9. [4] Arslan E, Colakog˘lu M, Celik C, GezginÈ K, Acar A, Capar M, et al. Serum TNF-alpha, IL-6, lupus anticoagulant and anticardiolipin antibody in women with and without a past history of recurrent miscarriage. Arch Gynecol Obstet 2004;270:227–9. [5] El-Far M, El-Sayed IH, El-Motwally Ael-G, Hashem IA, Bakry N. Tumor necrosis factor-alpha and oxidant status are essential participating factors in unexplained recurrent spontaneous abortions. Clin Chem Lab Med 2007;45: 879 – 83. [6] El-Far M, El-Sayed IH, El-Motwally AE, Hashem IA, Bakry N. Serum levels of TNF-alpha and antioxidant enzymes and placental TNF-alpha expression in unexplained recurrent spontaneous miscarriage. J Physiol Biochem 2009;65: 175– 81. [7] Wilson AG, Symons JA, McDowell TL, McDevitt HO, Duff GW. Effects of a polymorphism in the human tumor necrosis factor alpha promoter on transcriptional activation. Proc Natl Acad Sci U S A 1997;94:3195–9. [8] Knight JC, Udalova I, Hill AV, Greenwood BM, Peshu N, Marsh K, et al. A polymorphism that affects OCT-1 binding to the TNF promoter region is associated with severe malaria. Nat Genet 1999;22:145–50. [9] Koss K, Satsangi J, Fanning GC, Welsh KI, Jewell DP. Cytokine (TNF alpha, LT alpha and IL-10) polymorphisms in inflammatory bowel diseases and normal controls: differential effects on production and allele frequencies. Genes Immun 2000;1:185–90. [10] Skoog T, Eriksson P, Hoffstedt J, RydÊn M, Hamsten A, Armer P. Tumour necrosis factor-alpha (TNF-alpha) polymorphisms-857C/A and -863C/A are associated with TNF-alpha secretion from human adipose tissue. Diabetologia 2001;44:654 –5. [11] Wilson AG, Di Giovine FS, Blankemore M. Single base polymorphism in the tumor necrosis factor alpha gene detected by Nco´ restriction for the PCR product. Hum Mol Genet 1992;1:353. [12] Higgins JP, Thompson SG. Quantifying heterogeneity in a meta-analysis. Stat Med 2002;21:1539 –58. [13] Harbord RM, Egger M, Sterne JA. A modified test for small-study effects in meta-analyses of controlled trials with binary endpoints. Stat Med 2006;25: 3443–57. [14] Babbage SJ, Arkwright PD, Vince GS, Perrey C, Pravica V, Quenby S, et al. Cytokine promoter gene polymorphisms and idiopathic recurrent pregnancy loss. J Reprod Immunol 2001;51:21–7. [15] Baxter N, Sumiya M, Cheng S, Erlich H, Regan L, Simons A, et al. Recurrent miscarriage and variant alleles of mannose binding lectin, tumour necrosis factor and lymphotoxin alpha genes. Clin Exp Immunol 2001;126:529 –34.
B. Zhang et al. / Human Immunology 73 (2012) 574–579
[16] Reid JG, Simpson NA, Walker RG, Economidou O, Shillito J, Gooi HC, et al. The carriage of pro-inflammatory cytokine gene polymorphisms in recurrent pregnancy loss. Am J Reprod Immunol 2001;45:35– 40. [17] Daher S, Shulzhenko N, Morgun A, Mattar R, Rampim GF, Camano L, et al. Associations between cytokine gene polymorphisms and recurrent pregnancy loss. J Reprod Immunol 2003;58:69 –77. [18] Pietrowski D, Bettendorf H, Keck C, Bu¨rkle B, Unfried G, Riener EK, et al. Lack of association of TNFalpha gene polymorphisms and recurrent pregnancy loss in Caucasian women. J Reprod Immunol 2004;61:51– 8. [19] Prigoshin N, Tambutti M, Larriba J, Gogorza S, Testa R. Cytokine gene polymorphisms in recurrent pregnancy loss of unknown cause. Am J Reprod Immunol 2004;52:36 – 41. [20] Kamali-Sarvestani E, Zolghadri J, Gharesi-Fard B, Sarvari J. Cytokine gene polymorphisms and susceptibility to recurrent pregnancy loss in Iranian women. J Reprod Immunol 2005;65:171– 8. [21] Zhou Y, Li JY, Song LY, Zhou M, Chen SQ. Study on plasma TNF level and TNF alpha -308 polymorphisms in patients with unexplained repeated spontaneous abortion. J Mod Clin Med Bioeng 2006;12:32– 4.
579
[22] Zammiti W, Mtiraoui N, Finan RR, Almawi WY, Mahjoub T. Tumor necrosis factor alpha and lymphotoxin alpha haplotypes in idiopathic recurrent pregnancy loss. Fertil Steril 2009;91:1903– 8. [23] Finan RR, Al-Irhayim Z, Mustafa FE, Al-Zaman I, Mohammed FA, Al-Khateeb GM, et al. Tumor necrosis factor-alpha polymorphisms in women with idiopathic recurrent miscarriage. J Reprod Immunol 2010;84:186 –92. [24] Liu CM, Wang J, Zhou SR, Wang BB, Ma X. Association between -238 but not -308 polymorphism of tumor necrosis factor alpha (TNF-alpha)v and unexplained recurrent spontaneous abortion (URSA) in Chinese population. Reprod Biol Endocrinol 2010;8:114. [25] Palmirotta R, La Farina F, Ferroni P, Ludovici G, Nigro C, Savonarola A, et al. TNFA gene promoter polymorphisms and susceptibility to recurrent pregnancy loss in Italian women. Reprod Sci 2010;17:659 – 66. [26] Medica I, Ostojic S, Pereza N, Kastrin A, Peterlin B. Association between genetic polymorphisms in cytokine genes and recurrent miscarriage—a meta-analysis. Reprod Biomed Online 2009;19:406 –14. [27] Bombell S, McGuire W. Cytokine polymorphisms in women with recurrent pregnancy loss: meta-analysis. Aust N Z J Obstet Gynaecol 2008;48:147–54.
Contents lists available at SciVerse ScienceDirect
Association of tumor necrosis factor-␣ gene promoter polymorphisms (-308G/A, -238G/A) with recurrent spontaneous abortion: a meta-analysis Bingzhen Zhang, Tiecheng Liu, Zhiping Wang* Department of Epidemiology, Health Statistics, School of Public Health, Shandong University, Jinan, China
A R T I C L E
I N F O
Article history: Received 11 August 2011 Accepted 24 January 2012 Available online 14 February 2012
Keywords: Recurrent spontaneous abortion Tumor necrosis factor-␣ Gene polymorphism Meta-analysis
A B S T R A C T
Tumor necrosis factor-␣ (TNF-␣) gene promoter polymorphisms (-308G/A, -238G/A) have been associated with increased recurrent spontaneous abortion (RSA) risk, but the results of published articles are controversial. Hence, a meta-analysis was performed to assess the effect of TNF-␣ -308G/A, -238G/A polymorphisms on RSA risk. Heterogeneity testing and sensitivity analysis were performed using RevMan 5.0 software. Publication bias was assessed by the funnel plot method and modified Egger’s linear regression test. In 12 studies for the TNF-␣ -308G/A polymorphism, the summary odds ratio (OR) with the corresponding 95% confidence interval (95% CI) was 1.04 (95% CI: 0.86, 1.26) under a fixed-effect model in the overall population. In 5 studies for the TNF-␣ -238G/A polymorphism, the summary OR with the corresponding 95% CI was 1.11 (95% CI: 0.60, 2.03) under a random-effect model in the overall population. We could not identify the sources of heterogeneity for TNF-␣ -238G/A. In addition, no evidence of publication bias was detected. The results of this meta-analysis indicate that TNF-␣ -308G/A, -238G/A polymorphisms are not significantly associated with the risk of RSA in the overall population. However, more convincing evidence is required to draw a solid conclusion on the relation between the TNF-␣ -238G/A polymorphism and the risk of RSA. 䉷 2012 American Society for Histocompatibility and Immunogenetics. Published by Elsevier Inc. All rights reserved.
1. Introduction Recurrent spontaneous abortion (RSA) is defined as the occurrence of 3 or more clinically detectable pregnancy losses before the 20th week of gestation with the same partner [1] and is a major reproductive problem affecting 1–3% of otherwise healthy women [2]. The diagnosis of RSA involves multiple tests to detect parental chromosomal anomalies; maternal immunologic, hormonal, or thrombotic disorders; and anatomic abnormalities of the genital tract. Yet 40% of the affected women who do not exhibit any of the above-mentioned disorders are classified as unexplained RSA with unknown etiology [3]. Studies have indicated that the level of tumor necrosis factor-␣ (TNF-␣) apparently contributes to the pathogenesis of RSA [4,5] and higher serum levels of TNF-␣ were detected in RSA groups [6]. The TNF-␣ gene is located in 6p21.3 and has several functional polymorphism sites. Nucleotides -308 (rs1800629) and -238 (rs361525) in the promoter region of the TNF-␣ gene are 2 common functional polymorphisms that have been demonstrated to be associated with the production level of the cytokine [7–9]. The production of TNF-␣ is partly under genetic control [10], especially in the promoter region [11]. Although many studies have evaluated
the associations between TNF-␣ -308G/A, -238G/A polymorphisms and RSA, the results remain controversial. Hence, a meta-analysis was performed to assess the association of TNF-␣ -308G/A, -238G/A gene polymorphisms with the risk of RSA.
2. Subjects and methods 2.1. Search strategy Studies were identified by searches of PubMed, ISI Web of Science, CHINAINFO (Digital Journal Full-text Database), VIP (Database of Chinese Scientific and Technical Periodicals), the China Biology Medical Literature Database, and the China National Knowledge Infrastructure for relevant studies published in English or Chinese before April 2011 using the following search terms: cytokine gene polymorphisms and tumor necrosis factor gene polymorphisms, in combination with recurrent spontaneous abortion (RSA), recurrent pregnancy loss (RPL), recurrent miscarriage (RM), and recurrent fetal loss (RFL). In addition, the bibliographies of relevant studies, review articles, and meta-analysis articles were considered. The article search was performed independently by 2 investigators. 2.2. Inclusion criteria
* Corresponding author. E-mail address: [email protected] (Z. Wang).
The inclusion criteria were as follows: (1) RSA with clinical diagnosis criteria. RSA is defined as the occurrence of 3 or more clinically detectable pregnancy losses before the 20th week of
0198-8859/12/$36.00 - see front matter 䉷 2012 American Society for Histocompatibility and Immunogenetics. Published by Elsevier Inc. All rights reserved. doi:10.1016/j.humimm.2012.01.008
B. Zhang et al. / Human Immunology 73 (2012) 574–579
gestation. Furthermore, only unexplained or idiopathic abortions were included; (2) studies included data on the TNF-␣ -308G/A and/or -238G/A genotype(s); (3) case– control or cohort study published as an original study; (4) numbers in case and control groups for each genotype, or available data that can be used to calculate the numbers, should be reported; and (5) studies published in English or Chinese. The flow diagram for selection of studies is presented in Fig. 1.
2.3. Data extraction Two investigators independently extracted data and reached a consensus on all items. For each study, the following data were extracted: first author’s last name, country in which the study was performed, year of publication, mean ages in case and control groups if reported, and the frequency of each genotype in case and control groups
2.4. Statistical analysis Hardy–Weinberg equilibrium for TNF-␣ -308G/A, -238G/A genotype distributions in controls was tested using Pearson’s 2 test. The odds ratios (ORs) with the corresponding 95% confidence intervals (CIs) of RSA were considered under the dominant genetic models (AA plus GA vs GG). Heterogeneity among studies was assessed by the Q test and the I2 of Higgins and Thompson [12]. In the presence of heterogeneity (I ⬎ 50%), the random-effect model was adopted as the pooling method; otherwise, the fixed-effect model was used. The sensitivity analysis was performed by omitting each individual study, and an individual study was suspected of excessive sensitivity if the point estimate of its omitted analysis was outside the 95% CI of the combined analysis. We also conducted a stratified analysis for -308G/A based on the geographic region of the studies. In addition, a cumulative meta-analysis based on the size of the sample was conducted to evaluate the trends of summary OR with the corresponding 95% CI for the association, allowing us to see how the measured genetic effect changed when updated evidence accumulated. Publication bias was assessed by the funnel plot method and modified Egger’s linear regression test as proposed by Harbord et al. [13]. Data were analyzed by Review Manger (Cochrane Corp., Nordic Cochrane Center) 5.1.2 and Stata (Stata Corp., College Station, TX) version 11.1 software. All reported probabilities (p values) were 2-sided and a p value ⬍0.05 was considered significant.
575
3. Results 3.1. Characteristics of included studies Twelve studies [14 –25] fulfilled the inclusion criteria, including 12 studies [14 –25] on the -308G/A polymorphism (including 1,380 cases and 1,595 controls) and 5 studies [15,22–25] on the -238G/A polymorphism (including 884 cases and 912 controls) in the TNF-␣ gene. The TNF-␣ -308G/A, -238G/A genotype frequencies in controls were in Hardy–Weinberg equilibrium in all except 1 study [23]. Among these 12 studies, Daher et al. [17] and Prigoshin et al. [19] reported that only primary aborters were included in the cases; the other studies reported that both primary and secondary aborters were included in the cases [16,18,22,23] or did not clarify whether primary or secondary aborters were included [14,15, 20,21,24,25], so both primary and secondary aborters were included in the cases of this meta-analysis. General characteristics and the TNF-␣ -308G/A and -238G/A polymorphism genotype distributions in the published studies included in this meta-analysis are presented in Tables 1 and 2. 3.2. Quantitative synthesis 3.2.1. TNF-␣ -308G/A polymorphism and risk of RSA Fig. 2A illustrates the results of individual and summary ORs and the corresponding 95% CIs under the dominant genetic model in the overall population. Because no heterogeneity among studies was observed (I ⫽ 39%), the summary OR with 95% CI based on the fixed-effects model was 1.04 (95% CI: 0.86, 1.26). The cumulative meta-analysis indicated that the summary OR tended to 1 as the studies joined in one by one (Fig. 2B). In addition, we performed stratified analysis based on the geographic region of studies (because some studies did not report the ethnicity of the cases and controls, we could not perform stratified analysis based on ethnicity), but no significant association between the TNF-␣ 308 G/A polymorphism and RSA was observed in Europe, Asia, and other countries (Fig. 2C). 3.2.2. TNF-␣ -238G/A polymorphism and risk of RSA Fig. 3A illustrates the results of individual and summary ORs with the corresponding 95% CIs under the dominant genetic model in the overall population. Because heterogeneity among studies was observed (I ⫽ 74%), the summary OR with 95% CI based on the random-effects model was 1.11 (95% CI: 0.60, 2.03). However, after omitting the study of Liu et al. [24], the I2 values decreased (I ⫽ 45%) and the summary OR with 95% CI based on the fixed-effects model was 1.51 (95% CI: 1.16, 1.96; Fig. 3B). The results indicate that the TNF-␣ -238G/A polymorphism is associated with RSA when we decreased the heterogeneity of the studies. 3.3. Sensitivity analysis For each polymorphism, the results demonstrate that no evidence of any individual study having excessive influence on the summary effect was observed under the dominant genetic models (data not shown). 3.4. Publication bias evaluation Fig. 4A and 4B illustrate that no publication bias was detected for -308 G/A, -238G/A polymorphisms using the funnel plot method and the modified Egger linear regression test indicated that no significant publication bias was observed (p ⫽ 0.122 and 0.360, respectively). 4. Discussion
Fig. 1. The flow diagram for selection of studies.
The associations between the TNF-␣ gene promoter polymorphisms (-308G/A, -238G/A) and RSA have been widely studied. However, some studies have indicated that the polymorphisms were risk factors for RSA, whereas other studies drew converse
576
B. Zhang et al. / Human Immunology 73 (2012) 574–579
Table 1 Characteristics of studies on the association between the tumor necrosis factor-␣ -308G/A polymorphism and recurrent spontaneous abortion Author
Year
Babbage et al. [14] Baxter et al. [15] Reid et al. [16] Daher et al. [17] Pietrowski et al. [18] Prigoshin et al. [19] Kamali-Sarvestani et al. [20] Zhou et al. [21] Zammiti et al. [22] Finan et al. [23] Liu et al. [24] Palmirotta et al. [25]
2001 2001 2001 2003 2004 2004 2005 2006 2009 2010 2010 2010
Country
Age (mean)
UK UK UK Brazil Austria Argentina Iran China Tunisia Bahrain China Italy
Genotype (AA/GA/GG)
p for HWE
Case
Control
Case
Control
33 34 — 35 33 33 — 27.4 29.2 32.2 30.1 37
44 32 — — 58 — — — 29.0 31.8 28.8 38
1/12/30 3/22/51 2/6/7 1/11/36 2/33/133 0/6/35 0/14/117 0/8/42 14/39/319 8/32/164 0/22/110 0/13/87
1/16/56 5/40/93 1/13/29 1/18/89 4/41/167 0/5/49 0/21/122 0/5/45 5/47/222 4/32/212 1/13/138 3/21/76
0.906 0.787 0.744 0.932 0.432 0.721 0.343 0.71 0.187 0.041 0.276 0.313
HWE, Hardy–Weinberg equilibrium.
conclusions. Hence, we performed a meta-analysis to clarify the association between the 2 polymorphisms and the risk of RSA. In this meta-analysis, 12 studies (12 studies for -308G/A, 5 studies for -238G/A) on TNF-␣ gene promoter polymorphisms were performed to provide the most comprehensive assessment of the association between the 2 polymorphisms and the risk of RSA. The available data suggested that TNF-␣ -308G/A, -238G/A polymorphisms were not significantly associated with the risk of RSA when the effect of the polymorphisms was considered under the dominant genetic models in the overall population. The results of this meta-analysis were consistent only with the studies of Medica et al. [26] and Bombell and McGuire [27]. Medica et al. [26] did not observe any association between the TNF-␣ -308G/A polymorphism and RSA with 524 cases and 771 controls. Moreover, Bombell and McGuire [27] did not describe a significant effect of the -308G/A polymorphism on the risk of RSA with 630 cases and 875 controls. These results indicated that the TNF-␣ -308G/A polymorphism is not associated with the risk of RSA, consistent with the results of the cumulative meta-analysis. Moreover, the results were stable with the increase in sample size. In addition, we did not observe evidence of an association between the -308G/A polymorphism and RSA by stratified analysis based on geographic region. For the TNF-␣ -238G/A polymorphism, this was the first metaanalysis; no association was observed in the overall population. Because only 5 studies were included in this meta-analysis, we could not perform stratified analysis based on geographic region or ethnicity. Interestingly, among these 5 studies, Liu et al. [24] demonstrated that the G allele of the -238G/A polymorphism was a risk factor for RSA. When we excluded this study to decrease the heterogeneity of the studies, we determined that the AA plus GA versus GG genetic model of the -238G/A polymorphism was associated with a higher risk of RSA. We supposed that the -238G/A polymorphism might be associated with the risk of RSA; however, more convincing evidence, such as large sample size, number of studies, and ethnicities, is required to draw a solid conclusion. There were some limitations in this meta-analysis. First, the estimation on the -238G/A polymorphism in the overall population
was obtained by pooling the studies with a high degree of heterogeneity, but we could not perform meaningful stratified analysis because a small number of studies were included. Second, genotyping accuracy and quality control measures were not well described in some studies. Third, only studies published in English and Chinese were included in this meta-analysis, which might limit the results. A meta-analysis should be the appropriate method to obtain a more precise conclusion and allowed us to exclude modest associations for gene polymorphisms. We did not observe evidence that -308G/A, -238G/A polymorphisms in the TNF-␣ gene were associated with the risk of RSA in the overall population. The lack of association did not rule out the possibility that these polymorphisms were important in the pathogenesis of RSA, because genetic heterogeneous backgrounds might be to a certain extent responsible for the heterogeneity of the effect on RSA. However, differences in cytokine production induced by gene polymorphisms are not always indicative of immune/inflammatory events that occur at the placental interface during pregnancies/miscarriage, given that the cytokines (which generally act over a short interval) exerted their effects both in a paracrine and in an autocrine manner [25]. TNF-␣ gene -308G/A, -238G/A polymorphisms may be just regulators of the production of TNF-␣ at the genetic level, and the production of this cytokine may be regulated by other factors. Furthermore, Finan et al. [23] reported that significantly higher frequencies of TNF-␣ -1031C and -376A alleles were observed in RSA patients and significant differences were also noted in the distribution of -1031T/C and -376G/A genotypes between case and control subjects. Hence, the effect of TNF-␣ -1031T/C and -376G/A polymorphisms on the risk of RSA should be given more attention in the future. In conclusion, we did not observe any evidence that TNF-␣ gene -308G/A, -238G/A polymorphisms had a significant effect on the risk of RSA in the overall population. Because potential biases and confounders in this meta-analysis could not be ruled out completely, these results must be confirmed by further studies.
Table 2 Characteristics of studies on the association between the tumor necrosis factor-␣ -238G/A polymorphism and recurrent spontaneous abortion Author
Baxter et al. [15] Zammiti et al. [22] Finan et al. [23] Liu et al. [24] Palmirotta et al. [25]
Year
2001 2009 2010 2010 2010
HWE, Hardy–Weinberg equilibrium.
Country
UK Tunisia Bahrain China Italy
Age (mean)
Genotype (AA/GA/GG)
p for HWE
Case
Control
Case
Control
34 29.2 32.2 30.1 37
32 29.0 31.8 28.8 38
0/3/73 20/88/264 4/52/148 0/4/128 0/16/84
0/12/126 7/52/215 0/48/200 0/17/135 0/6/94
0.593 0.084 0.092 0.465 0.757
B. Zhang et al. / Human Immunology 73 (2012) 574–579
577
Fig. 2. Results of individual and summary odds ratios (ORs) with 95% confidence intervals (CIs) of recurrent spontaneous abortion (RSA) for the tumor necrosis factor-␣ (TNF-␣) -308 G/A polymorphism under the dominant genetic model. (A) The results of a meta-analysis for the TNF-␣ -308 G/A polymorphism in the overall population; (B) the results of a cumulative meta-analysis for the TNF-␣ -308 G/A polymorphism; and (C) the results of a stratified meta-analysis for the TNF-␣ -308 G/A polymorphism.
578
B. Zhang et al. / Human Immunology 73 (2012) 574–579
Fig. 3. Results of individual and summary odds ratios (ORs) with 95% confidence intervals (CIs) of recurrent spontaneous abortion (RSA) for the tumor necrosis factor-␣ (TNF-␣) -238 G/A polymorphism under the dominant genetic model. (A) The results of a meta-analysis for the TNF-␣ -238 G/A polymorphism in the overall population; and (B) the results of a meta-analysis for the TNF-␣ -238 G/A polymorphism after omitting the study by Liu et al. [24].
References
Fig. 4. Funnel plots of the odds ratios (ORs) of recurrent spontaneous abortion for tumor necrosis factor-␣ (TNF-␣) -308G/A, -238G/A polymorphisms for each study included in this meta-analysis. Funnel plot for (A) the TNF-␣ -308G/A polymorphism and (B) the TNF-␣ -238G/A polymorphism.
[1] Sierra S, Stephenson M. Genetics of recurrent pregnancy loss. Semin Reprod Med 2006;24:17–24. [2] Brown S. Miscarriage and its associations. Semin Reprod Med 2008;26:391– 400. [3] Stephenson MD. Frequency of factors associated with habitual abortion in 197 couples. Fertil Steril 1996;66:24 –9. [4] Arslan E, Colakog˘lu M, Celik C, GezginÈ K, Acar A, Capar M, et al. Serum TNF-alpha, IL-6, lupus anticoagulant and anticardiolipin antibody in women with and without a past history of recurrent miscarriage. Arch Gynecol Obstet 2004;270:227–9. [5] El-Far M, El-Sayed IH, El-Motwally Ael-G, Hashem IA, Bakry N. Tumor necrosis factor-alpha and oxidant status are essential participating factors in unexplained recurrent spontaneous abortions. Clin Chem Lab Med 2007;45: 879 – 83. [6] El-Far M, El-Sayed IH, El-Motwally AE, Hashem IA, Bakry N. Serum levels of TNF-alpha and antioxidant enzymes and placental TNF-alpha expression in unexplained recurrent spontaneous miscarriage. J Physiol Biochem 2009;65: 175– 81. [7] Wilson AG, Symons JA, McDowell TL, McDevitt HO, Duff GW. Effects of a polymorphism in the human tumor necrosis factor alpha promoter on transcriptional activation. Proc Natl Acad Sci U S A 1997;94:3195–9. [8] Knight JC, Udalova I, Hill AV, Greenwood BM, Peshu N, Marsh K, et al. A polymorphism that affects OCT-1 binding to the TNF promoter region is associated with severe malaria. Nat Genet 1999;22:145–50. [9] Koss K, Satsangi J, Fanning GC, Welsh KI, Jewell DP. Cytokine (TNF alpha, LT alpha and IL-10) polymorphisms in inflammatory bowel diseases and normal controls: differential effects on production and allele frequencies. Genes Immun 2000;1:185–90. [10] Skoog T, Eriksson P, Hoffstedt J, RydÊn M, Hamsten A, Armer P. Tumour necrosis factor-alpha (TNF-alpha) polymorphisms-857C/A and -863C/A are associated with TNF-alpha secretion from human adipose tissue. Diabetologia 2001;44:654 –5. [11] Wilson AG, Di Giovine FS, Blankemore M. Single base polymorphism in the tumor necrosis factor alpha gene detected by Nco´ restriction for the PCR product. Hum Mol Genet 1992;1:353. [12] Higgins JP, Thompson SG. Quantifying heterogeneity in a meta-analysis. Stat Med 2002;21:1539 –58. [13] Harbord RM, Egger M, Sterne JA. A modified test for small-study effects in meta-analyses of controlled trials with binary endpoints. Stat Med 2006;25: 3443–57. [14] Babbage SJ, Arkwright PD, Vince GS, Perrey C, Pravica V, Quenby S, et al. Cytokine promoter gene polymorphisms and idiopathic recurrent pregnancy loss. J Reprod Immunol 2001;51:21–7. [15] Baxter N, Sumiya M, Cheng S, Erlich H, Regan L, Simons A, et al. Recurrent miscarriage and variant alleles of mannose binding lectin, tumour necrosis factor and lymphotoxin alpha genes. Clin Exp Immunol 2001;126:529 –34.
B. Zhang et al. / Human Immunology 73 (2012) 574–579
[16] Reid JG, Simpson NA, Walker RG, Economidou O, Shillito J, Gooi HC, et al. The carriage of pro-inflammatory cytokine gene polymorphisms in recurrent pregnancy loss. Am J Reprod Immunol 2001;45:35– 40. [17] Daher S, Shulzhenko N, Morgun A, Mattar R, Rampim GF, Camano L, et al. Associations between cytokine gene polymorphisms and recurrent pregnancy loss. J Reprod Immunol 2003;58:69 –77. [18] Pietrowski D, Bettendorf H, Keck C, Bu¨rkle B, Unfried G, Riener EK, et al. Lack of association of TNFalpha gene polymorphisms and recurrent pregnancy loss in Caucasian women. J Reprod Immunol 2004;61:51– 8. [19] Prigoshin N, Tambutti M, Larriba J, Gogorza S, Testa R. Cytokine gene polymorphisms in recurrent pregnancy loss of unknown cause. Am J Reprod Immunol 2004;52:36 – 41. [20] Kamali-Sarvestani E, Zolghadri J, Gharesi-Fard B, Sarvari J. Cytokine gene polymorphisms and susceptibility to recurrent pregnancy loss in Iranian women. J Reprod Immunol 2005;65:171– 8. [21] Zhou Y, Li JY, Song LY, Zhou M, Chen SQ. Study on plasma TNF level and TNF alpha -308 polymorphisms in patients with unexplained repeated spontaneous abortion. J Mod Clin Med Bioeng 2006;12:32– 4.
579
[22] Zammiti W, Mtiraoui N, Finan RR, Almawi WY, Mahjoub T. Tumor necrosis factor alpha and lymphotoxin alpha haplotypes in idiopathic recurrent pregnancy loss. Fertil Steril 2009;91:1903– 8. [23] Finan RR, Al-Irhayim Z, Mustafa FE, Al-Zaman I, Mohammed FA, Al-Khateeb GM, et al. Tumor necrosis factor-alpha polymorphisms in women with idiopathic recurrent miscarriage. J Reprod Immunol 2010;84:186 –92. [24] Liu CM, Wang J, Zhou SR, Wang BB, Ma X. Association between -238 but not -308 polymorphism of tumor necrosis factor alpha (TNF-alpha)v and unexplained recurrent spontaneous abortion (URSA) in Chinese population. Reprod Biol Endocrinol 2010;8:114. [25] Palmirotta R, La Farina F, Ferroni P, Ludovici G, Nigro C, Savonarola A, et al. TNFA gene promoter polymorphisms and susceptibility to recurrent pregnancy loss in Italian women. Reprod Sci 2010;17:659 – 66. [26] Medica I, Ostojic S, Pereza N, Kastrin A, Peterlin B. Association between genetic polymorphisms in cytokine genes and recurrent miscarriage—a meta-analysis. Reprod Biomed Online 2009;19:406 –14. [27] Bombell S, McGuire W. Cytokine polymorphisms in women with recurrent pregnancy loss: meta-analysis. Aust N Z J Obstet Gynaecol 2008;48:147–54.