Molecular Biology Reports

, Volume 41, Issue 1, pp 317–324

Association between polymorphism in TRAF1/C5 gene and risk of rheumatoid arthritis: a meta-analysis

Authors

    • Department of RheumatologyShengjing Hospital of China Medical University
  • Wei Li
    • Editorial Department of Chinese Pediatric Emergency Medicine
  • Xinpeng Zhang
    • Department of NeurosurgeryThe Fourth People’s Hospital of Shenyang City
  • Xiaoli Zhang
    • Department of RheumatologyShengjing Hospital of China Medical University
  • Li Jiang
    • Department of RheumatologyShengjing Hospital of China Medical University
  • Yun Guo
    • Department of RheumatologyShengjing Hospital of China Medical University
  • Xiaofei Wang
    • Department of RheumatologyShengjing Hospital of China Medical University
Article

DOI: 10.1007/s11033-013-2864-0

Cite this article as:
Zhang, X., Li, W., Zhang, X. et al. Mol Biol Rep (2014) 41: 317. doi:10.1007/s11033-013-2864-0

Abstract

Rheumatoid arthritis (RA) is a common chronic inflammatory autoimmune disease. Single nucleotide polymorphisms of tumor necrosis factor-receptor associated factor 1/complement component 5 (TRAF1/C5) gene are suspected to be associated with the risk of RA. This meta-analysis was performed to study the relationship between the polymorphism rs10818488 in TRAF1/C5 gene with RA. We retrieved the relevant articles from PubMed, EMBASE and the China National Knowledge Infrastructure databases. Odd ratios were calculated to assess the association between TRAF1/C5 rs10818488 polymorphism and RA risk. Meta-analyses were performed on the total data set and separately for the major ethnic groups and RF and ACAP status. All analyses were performed using the Stata software. Eight articles were included in the present analysis. Meta-analysis showed a weak association between TRAF1/C5 rs10818488 polymorphism and RA in all subjects (OR = 1.13, 95 % CI = 1.01–1.27, Pheterogeneity < 0.001). Stratified analyses indicated that the TRAF1/C5 rs10818488 A allele was significantly associated with RA in Caucasians (OR = 1.29, 95 %CI = 1.14–1.47, Pheterogeneity = 0.026), Asians (OR = 0.92, 95 %CI = 0.86–0.99, Pheterogeneity = 0.378) and Africans (OR = 1.56, 95 %CI = 1.23–1.98, Pheterogeneity = 0.876), also significantly in positive ACPA and positive RF patients versus controls (ORs were 1.20 and 1.25, 95 %CIs were 1.08–1.33 and 1.14–1.37, P values for heterogeneity were 0.215 and 0.133, respectively). Genetic polymorphism rs10818488 in TRAF1/C5 gene might be associated with RA susceptibility.

Keywords

Rheumatiod arthritisTumor necrosis factor-receptor associated factor 1/complement component 5Single nucleotide polymorphismMeta-analysis

Introduction

Rheumatoid arthritis is a common chronic inflammatory autoimmune disease, affecting 1 % of the population worldwide [1]. Women are affected two to three times more often than men [2]. This disease might damage many tissues and organs, especially synovial joints involving autoimmune features, and eventually could cause severe disability and even early mortality [3]. Genetic factors have an important role in the development of RA, contributing 50–60 % of disease susceptibility [4]. In recent years, great progresses have been made in identifying susceptibility variants of RA using both genome-wide association study (GWAS) and candidate gene approaches [5].

Recently, tumor necrosis factor (TNF)-receptor associated factor 1/complement component 5 (TRAF1/C5) loci on 9q33-34 has been newly identified as a susceptibility gene for RA [6]. TNF is a critical cytokine in the pathogenesis of RA and using TNF antagonists is an effective treatment for RA [7, 8]. The TRAF1 gene encodes an intracellular protein that mediates TNFα signal transduction, which is considered to be associated with the T cell proliferation and activation [9]. TRAF1 could bind several intracellular proteins, including the nuclear factor-κB inhibitory protein TNFAIP3 [10]. C5 is a key member of the complement pathway, which has been suggested to play an important role in the development of RA for almost 40 years [11]. There is a study suggested that two different anti-C5 single-chain Fv (scFv) could represent potential therapeutic reagents to be used in RA patients [12]. Therefore, TRAF1/C5 could possibly contribute to the pathogenesis of autoimmune diseases RA.

In 2007, a multitiered case–control study and following stepwise replication of significant SNPs in three independent sample sets indicated that a polymorphism in the TRAF1/C5 region, that is rs10818488, could increase the susceptibility to and severity of RA. About the potential functional effect of rs10818488, this study suggested that this SNP maps to an intergenic region; 10 kb from both TRAF1 and C5 and is present in a transcription factor binding site, which may regulate the transcriptional activity of its neighboring genes, influencing the structure, function, and/or expression levels of TRAF1 and/or C5 [6]. The association of this polymorphism with RA has been vast replicated among European, Asian, North American, and African populations [6, 1319]. However, these reported results were contradictory and inconclusive. It is well known that many mutations in the same gene might contribute different role in people who have different ethnic background. According to Hapmap data, we could found that there are differences in the prevalence of the rs10818488 A allele in Caucasian population origin (52.5 %), Asian population (44.3 %) and Sub-Saharan African population (69.2 %). And significantly affected RA risks were found for this polymorphism among every population, but the effects are varied. It showed that there may be ethnicity difference for association between TRAF1/C5 rs10818488 with RA risks. Actually, several possible influential factors, such as small sample size, publication bias also could be explained for this discordance. Therefore, to comprehensively assess the relationships of TRAF1/C5 polymorphism (rs10818488) and the risk of RA, we perform an exhaustive meta-analysis. Additionally, we also explore the associations in different ethnic populations and different disease subgroups that were characterized by anticardiac phospholipids antibody (ACPA) and rheumatoid factor (RF).

Methods

Data sources

An exhaustive search was performed for articles on the association between TRAF1/C5 genotypes and RA published before August 2012. The PubMed, EMBASE and Chinese National Knowledge Infrastructure (CNKI) were used with the combination of terms “TRAF1/C5”, “polymorphism or SNP” and “RA or rheumatoid arthritis” without any restriction on language. The reference lists of the obtained articles were screened to identify other eligible references, which were then retrieved, and relevant original articles were also sought manually.

Study selection and data extraction

We evaluated potential relevant publications by examining their titles and abstracts and all studies matching the eligible criteria were retrieved. To select and retain a large group of homogeneous studies, we did not exclude any study on the basis of language.

Inclusion criteria

  1. (1)

    Evaluation of the rs10818488 and RA risk.

     
  2. (2)

    Using the methodology of a case–control study or cohort study.

     
  3. (3)

    There were sufficient published data for the computation of odds ratios (ORs) with 95 % confidence intervals (95 %CIs).

     

Exclusion criteria

  1. (1)

    The studies with cell line or animals would be excluded.

     
  2. (2)

    Duplicate studies were excluded to avoid giving double weight to a single result. In case of multiple reports of the same results, we considered only the estimates from the first publication.

     
  3. (3)

    Reviews, obviously unrelated articles and GWASs were excluded.

     

Validity assessment

According to inclusion criteria, a total of 11 articles were selected. Three studies were excluded because they were research on GWAS or because data overlap, data missing. Thus, eight studies were included in the final analysis [6, 1319]. Flow chart of the study selection process was shown in Fig. 1.
https://static-content.springer.com/image/art%3A10.1007%2Fs11033-013-2864-0/MediaObjects/11033_2013_2864_Fig1_HTML.gif
Fig. 1

Flow chart of the study selection process

For these eight studies, we performed an open assessment of the studies using a structured checklist based on Newcastle–Ottawa Scale (NOS) including selection of study subjects, comparability of cases and controls, and exposure of impact factors [20]. The maximum score of NOS is 9 and the studies with the score ≥7 were considered as high quality studies. All of eight studies included in the present meta-analysis were evaluated as high quality researches.

Data collection

The following information was sought from each publication included in the present meta-analysis: first author’s name, publication date, country origin, ethnicity, total number of cases and controls, as well as number of cases and controls with two alleles and three genotypes of TRAF1/C5 rs10818488 SNP. Finally we summarized the results of eight articles for our systematic review in Table 1.
Table 1

Characteristics of the studies included in the meta-analysis

Author (ref)

No of data sets

Country (ethnicity)

Number of cases

Number of control

Antibody status

Frequency of A allele

P (HWE)

% Weight

Statistical power

Number

%

Kurreeman et al. [6]

1

Netherlan (Caucasian)

290

254

89 %RF+

193

40

>0.05

6.59

0.82

2

Netherlan (Caucasian)

454

270

100 %RF + 81 %ACPA+

220

42

>0.05

7.10

0.52

3

Switzerland (Caucasian)

1500

1000

66 %RF + 62 %ACPA+

840

46

>0.05

8.72

0.75

4

American (Caucasian)

475

475

100 %RF+

356

38

>0.05

7.71

0.88

Morales et al. [13]

1

Columbia (South American)

274

421

NA

351

42

>0.05

7.11

0.42

Zervou [14]

1

Greece (Caucasian)

311

344

NA

190

28

>0.05

6.93

0.99

Stark et al. [15]

1

Slovakia(Caucasian)

520

303

53.8 %RF + 78.6 %ACPA+

234

39

>0.05

7.35

0.07

Nishimoto et al. [16]

1

Japan (Asian)

1504

752

NA

787

52

>0.05

8.61

0.05

2

Japan(Asian)

830

658

NA

724

55

>0.05

8.32

0.53

3

Japan(Asian)

1113

940

NA

1018

54

>0.05

8.63

0.17

 

4

Japan(Asian)

950

507

NA

571

56

>0.05

8.18

0.54

Zhu and Zhao[17]

1

China (Asian)

95

98

NA

60

31

>0.05

4.17

0.10

Mohamed et al. [19]

1

Egypt (North Africa)

172

160

86 %RF + 91.3ACPA+

87

27

>0.05

5.40

0.72

Karray et al. [18]

1

Tunisia (North Africa)

108

161

NA

133

41

>0.05

5.16

0.75

Statistical analyses

We first assessed Hardy–Weinberg equilibrium for each study using χ2 test in control groups. ORs corresponding to 95 %CIs were calculated to assess the association of TRAF1/C5 rs10818488 SNP and RA risk. Pooled ORs were obtained from combination of included studies by heterozygote comparison (AG vs. GG), homozygote comparison (AA vs. GG), dominant model (AA + AG vs. GG), recessive model (AA vs. AG + GG) and allelic model (A vs. G) respectively. For each genetic comparison model, subgroup analysis according to ethnicity was investigated to estimate ethnic-specific ORs for Asian, Caucasian and African. Meanwhile stratified analyses by ACPA and RF status were also applied for each genetic comparison model.

We investigated the between-study heterogeneity by the Cochran’s Q test and quantified by I2 (a significance level of P < 0.10 and/or I2 ≥ 50 %). To obtain summary statistics for ORs of TRAF1/C5 rs10818488 polymorphism and RA risk, we performed initial analyses with a fixed-effect model and confirmatory analyses with a random-effect model if there was significant heterogeneity. The random-effect model assumes that different studies show substantial diversity and assesses both within-study sampling errors and between-study variances. Fixed-effect model assumes that genetic factors have similar effects on RA susceptibility across all studies and that observed variations between studies are caused by chance alone. If study groups show no heterogeneity, the fixed and random effects models produce similar results, and, if not, the random-effect model usually produces wider CIs than the fixed-effect model. We assessed potential publication bias by examining inverted funnel plots and Egger’s test. This statistical test detects whether the intercept deviates significantly from zero in a regression of the standardized effect estimates against their precision. When the P value of Egger’s test was near 0.05, Fail-Safe number (m) was estimated to evaluate the influence of publication bias on the results. All of P values were two-sided and all analyses were performed using the Stata software version 11.0 (Stata Corp, College station, TX, USA).

Results

Detailed characteristics of each study are described in Table 1. Eight studies on TRAF1/C5 rs10818488 polymorphism and RA involved 8,488 cases and 6,182 controls. The types of control subjects used varied between studies and included randomly selected community controls and hospital patients.

The subjects in the study were population of Asian, Caucasian, South American and African. The frequencies of minor A allele of TRAF1/C5 rs10818488 SNP in controls were 40.2 % in population of Caucasian origin, and 49.6 % in Asian population, and 34.0 % in African population. Asian controls showed higher frequencies of minor A allele of TRAF1/C5 rs10818488 than other populations in these studies. The distribution of genotypes in the controls was all in Hardy–Weinberg equilibrium (HWE).

Association between A allele of rs10818488 in TRAF1/C5 gene with risk of RA

In total subjects

A summary of the meta-analysis findings of the relation between the TRAF1/C5 rs10818488 polymorphism and RA is provided in Table 2. Meta-analysis showed an association between the polymorphism and RA in all subjects (OR = 1.13, 95 %CI = 1.01–1.27, Pheterogeneity < 0.001) (Table 2).
Table 2

Association between A allele of TRAF1/C5 rs10818488 polymorphism with RA risk

Population

No. of study

Number of cases

Number of controls

A allele versus G allele (OR, 95 %CI, P value)

Test of heterogeneity

Fixed effect model

Random effect model

P value

Overall

8

8488

6182

1.07

1.02–1.12

0.009

1.13*

1.011.27

0.035

<0.001

Caucasian

4

3824

3067

1.26

1.17–1.36

<0.001

1.29*

1.141.47

<0.001

0.026

Asian

2

4492

2955

0.92*

0.860.99

0.018

0.92

0.86–0.99

0.022

0.378

African

2

280

321

1.56*

1.231.98

<0.001

1.56

1.23–1.98

<0.001

0.876

*The results were statistically significant (P < 0.05)

Stratified analyses by ethnicity

Meta-analysis was performed stratifying by ethnic group. Our analysis indicated that rs10818488 A allele was significantly associated with increased risk of RA in Caucasians and Africans (ORs were 1.29 and 1.56, 95 %CIs were 1.14–1.47 and 1.23–1.98, P values for heterogeneity were 0.026 and 0.876, respectively). In opposite, significantly decreased RA risk were found among Asians (OR = 0.92, 95 %CI = 0.86–0.99, Pheterogeneity = 0.378) (Table 2).

Stratified analyses by RF and ACPA status

Stratification of patients according to the presence of RF revealed a significant association between the A allele and RA in positive patients compared to negative individuals or controls (ORs were 1.17 and 1.25, 95 %CIs were 1.04–1.31 and 1.14–1.37, P values for heterogeneity were 0.521 and 0.133, respectively). The significant association of the A allele and RA also exists in the presence of ACPA positive patients versus controls (OR = 1.20, 95 %CI = 1.08–1.33, Pheterogeneity = 0.215). The stratification also showed significantly different RA risks for patients with RF-positive antibodies compared to ones with RF-negative antibodies (OR = 1.17, 95 %CI = 1.04–1.31, Pheterogeneity = 0.521) (Table 3).
Table 3

Stratified analysis of A allele of rs10818488 polymorphism and RA risk by antibody groups

Subgroup

No. of studies

OR

95 % CI

P for heterogeneity

ACPA+ versus ACPA−

2

1.08

0.95–1.22

0.450

ACPA+ versus control

2

1.20*

1.081.33

0.215

ACPA− versus control

2

1.12

0.99–1.26

0.868

RF+ versus RF−

3

1.17*

1.041.31

0.521

RF+ versus control

3

1.25*

1.141.37

0.133

RF− versus control

3

1.06

0.95–1.19

0.557

*The results were statistically significant (P < 0.05)

Association between the genotypes of TRAF1/C5 rs10818488 with RA risk

We also performed meta-analysis focus on the variances of genotypes which may have epidemic significance in TRAF1/C5 polymorphisms associates with RA. We divided four groups to evaluate the association between genotypes of rs10818488 polymorphisms with RA risk: AG versus GG, AA versus GG, AG + AA versus GG (dominant model), AA versus AG + GG (recessive model). A summary of the meta-analysis findings of the relation between the genotypes and RA is provided in Table 4. All the groups have no significant association between genotypes of rs10818488 polymorphism and RA risk in the total population.
Table 4

Association between the genotypes of rs10818488 polymorphism with RA risks

Genotypes

Population

No. of study

Association (OR, 95 % CI, P value)

Test of heterogeneity

Fixed effect model

Random effect model

P value

AG/GG

Overall

7

1.01 (0.92, 1.12) 0.784

1.08 (0.87, 1.34) 0.484

<0.001

 

Caucasian

2

1.52 (1.22, 191) < 0.001

1.54 (0.82, 2.91) 0.182

0.005

 

Asian

2

0.89 (0.79, 1.01) 0.063

0.89 (0.75, 1.06) 0.181

0.140

 

African

2

1.55 (1.08, 2.24) 0.019*

1.55 (1.08, 2.24) 0.019

0.997

AA/GG

Overall

6a

0.91 (0.81, 1.03) 0.120

0.99 (0.78, 1.27) 0.938

<0.001

 

Caucasian

2

1.45 (1.03, 2.05) 0.032

1.67 (0.65, 4.30) 0.284

0.010

 

Asian

1a

0.83 (0.72, 0.95) 0.009*

0.83 (0.71, 0.96) 0.015

0.312

 

African

2

2.31 (1.40, 3.81) 0.001*

2.31 (1.40, 3.82) 0.001

0.590

AA + AG/GG

Overall

7

1.01 (0.92, 1.10) 0.884

1.10 (0.87, 1.38) 0.430

<0.001

 

Caucasian

2

1.52 (1.23, 1.895) < 0.001

1.56 (0.79, 3.09) 0.205

0.002

 

Asian

2

0.87 (0.78, 0.98) 0.022*

0.87 (0.75, 1.02) 0.092

0.157

 

African

2

1.72 (1.22, 2.42) 0.002*

1.72 (1.22, 2.42) 0.002

0.808

AA/AG + GG

Overall

6a

0.95 (0.86, 1.04) 0.266

0.97 (0.84, 1.10) 0.603

0.115

 

Caucasian

2

1.21 (0.88, 1.66) 0.244

1.30 (0.72, 2.34) 0.383

0.084

 

Asian

1a

0.92 (0.82, 1.02) 0.101

0.92 (0.82, 1.02) 0.101

0.430

 

African

2

1.84 (1.18, 2.88) 0.007*

1.84 (1.18, 2.88) 0.007

0.701

*The results were statistically significant (P < 0.05)

aOne study in Chinese population was excluded because there was no information of AA genotype

Moreover, the stratified analyses by ethnicity indicated that in Asians AA and AA + AG genotypes were significantly associated with decreased risk of RA (ORs were 0.83 and 0.87, 95 %CIs were 0.72–0.95 and 0.78–0.98, P values for heterogeneity were 0.312 and 0.157, respectively). The overall ORs with its 95 %CIs showed statistically associations between rs10818488 polymorphism and RA risk in Africans for every genotypes (AA vs. GG: OR = 2.31, 95 %CI = 1.40–3.81, P = 0.590 for heterogeneity; AG versus GG: OR = 1.55, 95 %CI = 1.08–2.24, P = 0.997 for heterogeneity; AA + AG versus GG: OR = 1.72, 95 %CI = 1.22–2.42, P = 0.808 for heterogeneity; AA versus AG + GG: OR = 1.84, 95 %CI = 1.18–2.88, P = 0.701 for heterogeneity). There were no significant associations in other genotypes among other ethnicity groups. The details for each genotype stratified by ethnicity were described in Table 4.

Heterogeneity, sensitivity analysis and publication bias

Heterogeneity between studies was observed in overall comparisons and also subgroup analyses. Between-study heterogeneity was only found among the overall population and Caucasian populations in terms of the ORs of TRAF1/C5 rs10818488 A allele. Thus, meta-analyses were performed using random-effect models (Table 2). In the subgroup analyses stratified by ethnicity and autoantibody status respectively, it can be found that the heterogeneity of the subgroup reduced significantly.

The sensitivity analysis was conducted. Every single study involved in the meta-analysis was deleted each time to reflect the influence of the individual data set to the pooled ORs. This procedure did not change the pooled ORs supporting the robustness of our findings.

To explore potential sources of heterogeneity in this meta-analysis, a meta regression analysis for allelic model was implemented. The covariates included sample size, ethnicity, publication year and genotyping method (the assignments of variables were shown in Table 5). The results showed that all of the potential factors were probably not the major sources of heterogeneity (P value for regression all >0.05). The heterogeneity might attribute to other factors, the insufficient data limited to identify their sources only using meta regression.
Table 5

The covariates for the potential sources of heterogeneity and their assignments

Author (ref)

No of data sets

Year

Genotype method

Ethnicity

Case number

Control number

Kurreeman et al. [6]

1

2007 (x1 = 1)

SNPPER (X2 = 1)

Caucasian (X3 = 1)

290

254

2

2007 (x1 = 1)

SNPPER (X2 = 1)

Caucasian (X3 = 1)

454

270

3

2007 (x1 = 1)

SNPPER (X2 = 1)

Caucasian (X3 = 1)

1,500

1,000

4

2007 (x1 = 1)

SNPPER (X2 = 1)

Caucasian (X3 = 1)

475

475

Morales et al. [13]

1

2008 (x1 = 1)

Taqman (X2 = 2)

Caucasian (X3 = 1)

274

421

Zervou et al. [14]

1

2008 (x1 = 1)

Sdul restriction enzyme (X2 = 3)

Caucasian (X3 = 1)

311

344

Stark et al. [15]

1

2009 (x1 = 1)

Taqman (X2 = 2)

Caucasian (X3 = 1)

520

303

Nishimoto et al. [16]

1

2010(x1 = 2)

Taqman (X2 = 2)

Asian (X3 = 2)

1,504

752

2

2010 (x1 = 2)

Taqman (X2 = 2)

Asian (X3 = 2)

830

658

3

2010 (x1 = 2)

Taqman (X2 = 2)

Asian (X3 = 2)

1,113

940

4

2010 (x1 = 2)

Taqman (X2 = 2)

Asian (X3 = 2)

950

507

Zhu and Zhao [17]

1

2011 (x1 = 2)

PCR (X2 = 4)

Asian (X3 = 2)

95

98

Mohamed et al. [19]

1

2012 (x1 = 2)

PCR (X2 = 4)

African (X3 = 3)

172

160

Karray et al. [18]

1

2012 (x1 = 2)

PCR (X2 = 4)

African (X3 = 3)

108

161

No publication bias was detected by either the inverted funnel plot or Begg’s test. The shapes of the funnel plot for the comparison of the G allelic and the A allelic of TRAF1/C5 rs10818488 SNP seemed approximately symmetrical and P value of the Egger’ test was not statistical significance (data not shown).

Discussion

Recent large-scale genome wide association studies reveal a number of SNP markers that are associated with RA susceptibility, substantially improving our understanding of the genetic component of RA susceptibility. The association of rs10818488 A allele with increased RA risk was first described in 2007 [6], and has been validated among GWAS and familial study in Caucasian population [21, 22]. However, another GWAS conducted by Wellcome Trust Case Control Consortium showed no significant association between this polymorphism and RA risk [23]. Otherwise, there are studies in Asian populations suggesting the protective effect of rs10818488 A allele on RA risk [16, 17]. In addition, other reported results of associations between this polymorphism with RA susceptibility were contradictory and inconclusive in different racial populations [1315, 18, 19]. Recently there are published cohort studies suggesting that this polymorphism was associated with the death risk of RA [24, 25]. In order to system study whether TRAF1/C5 rs10818488 SNP is associated with risk for RA, we have performed this meta-analysis, which included 8,488 cases and 6,182 controls. Our results showed significant associations between TRAF1/C5 rs10818488 polymorphism and RA in all subjects. Stratified analyses suggested that the effects of TRAF1/C5 rs10818488 polymorphism on RA were significantly different among Caucasians, Asians and Africans. The present results also indicated the significant results in different status of ACPA and RF patients.

As we know, the incidence of most genetic polymorphisms could vary between different ethnic populations. In this meta-analysis, there are differences in the prevalence of the rs10818488 A allele among controls in Caucasian population origin (39.3 %), Asian population (49.6 %) and African population (34.0 %). In the stratified analyses by ethnicity, significantly affected RA risks were found for this polymorphism among every population, but the effects are varied heavily, that is A allele is risk factor in Caucasians, but protective factor in Asians. It showed that there may be ethnicity difference for association between TRAF1/C5 rs10818488 with RA risks. The reason may be that this polymorphism may have different effect in different populations, reflecting the diversities of the susceptible factors for different ethnicities. In addition, the observed different effects could be likely due to chance because studies with small sample size may have insufficient statistical power to detect a slight effect or may have generated a fluctuated risk estimate. So studies with larger sample size in different ethnicity populations are necessary to fully understand the relationship between the polymorphism and the risk of RA.

Rs10818488 maps to an intergenic region; 10 kb from both TRAF1 and C5 and is present in a transcription factor binding site, which may regulate the transcriptional activity of its neighboring genes. So formal testing of all known variation within this locus, both genetic and biological, will be necessary to pinpoint the precise biological process that is altered by the RA-associated variant(s) present in this region.

Since three studies have examined the relationship between TRAF1/C5 rs10818488 alleles and RA risk subgrouping by clinical characteristics, we were only able to incorporate three results for patients representing different RF status (RF-positive and RF-negative), two results for different ACPA status (ACPA-positive and ACPA-negative). Our study showed significant association between this SNP with RA risks in patients being RF-positive or ACPA-positive. This agree with the results of Kurreeman and Mohamed, who found rs10818488 A allele was significantly associated with risk of RA patients representing positive antibody [6, 19], and disagree with Stark who found no significant association [15]. The stratified analyses also showed increased RA risks for patients with RF-positive antibodies compared to those with RF-negative antibodies.

To our knowledge, this is the first meta-analysis to analyze the association between rs10818488 and RA with all genetic comparison models, which could provide most thorough information of rs10818488 SNP and RA risks. Pooled ORs were obtained by heterozygote comparison (AG vs. GG), homozygote comparison (AA vs. GG), dominant model (AA + AG vs. GG), recessive model (AA vs. AG + GG) and allelic model (A vs. G), respectively. There were no significant associations in the overall population for every type of genetic comparison. But the significant associations were found in Asian and African population respectively, which could have epidemical importance suggesting the genetic risks for RA in different ethnicities. Another important problem which should be paid attention is the observed different ethnic effects could be likely due to chance because this part of analyses included only two studies with small sample size and it may have insufficient statistical power to detect a slight effect or may have generated a fluctuated risk estimate. So larger sample size studies containing enough information including three genotypes are necessary to fully understand the relationship between rs10818488 polymorphism and risk of RA.

Despite our efforts in performing a comprehensive analysis, some limitations exist in our meta-analysis. First, our analysis used published international studies, which could arise publication bias, although the results for publication bias in our study were not statistically significant. Second, our ethnic-specific meta-analysis included data from Caucasian, Asian and African population, thus our results are applicable to these ethnic groups alone, and should be optimized by larger scale of populations. Third, lack of the original data of available studies limited our further evaluation of potential interactions, such as age, gender, family history, environmental factors and lifestyle.

In conclusion, our meta-analysis study showed that the TRAF1/C5 rs10818488 polymorphism is associated with RA susceptibility in major ethnic groups and different ethnic populations. As TRAF1/C5 is emerging as a novel common risk factor for diverse complex diseases and our study provided evidence showing the important role of TRAF1/C5 polymorphisms in the development of RA, genetic factors determining disease susceptibility may facilitate personalized medicine. The TRAF1/C5 locus warrants further investigation as a potential disease susceptibility locus in RA.

Acknowledgments

The authors declare that they have no competing interests.

Copyright information

© Springer Science+Business Media Dordrecht 2013