Association between SMAD3 gene polymorphisms and osteoarthritis risk: a systematic review and meta-analysis

Objective Several studies have been performed to investigate the association between SMAD3 gene polymorphism and osteoarthritis (OA), but the results were inconclusive. This study aims to determine whether SMAD3 polymorphism is associated with risk of OA. Method A comprehensive literature search in PubMed, Embase, and ISI Web of Science for relevant studies was performed. After extracting data from eligible studies, we chose the fixed or random effect model according to the heterogeneity test. Estimation of publication bias and sensitivity analysis were conducted to confirm the stability of this meta-analysis. Results In total, 10 studies from 6 articles with 5093 OA patients and 5699 controls were enrolled in this meta-analysis. The combined results revealed significant association between SMAD3 rs12901499 polymorphism and the risk of OA (allele model: OR 1.21, 95% CI 1.07–1.38). Subgroup analysis revealed that G allele increased the risk of OA in Caucasians, but not in Asians (allele model: Caucasians: OR 1.31, 95% CI 1.18–1.44; Asians: OR 1.24, 95% CI 0.95–1.61). And the pooled results revealed significant association between SMAD3 rs12901499 polymorphism and both knee and hip OA (knee OA: OR 1.18, 95% CI 1.04–1.34; hip OA: OR 1.31, 95% CI 1.18–1.44). Conclusion The current meta-analysis revealed that the G variant of SMAD3 rs12901499 polymorphism increased the risk of OA in Caucasians. Further well-designed studies with larger sample size in different ethnic populations are required to confirm these results. Electronic supplementary material The online version of this article (10.1186/s13018-018-0939-2) contains supplementary material, which is available to authorized users.


Background
Osteoarthritis (OA), a late-onset musculoskeletal disease in the elder, is featured by the gradual degradation of articular cartilage with further lesion to the synovium, subchondral bone, or the other joint tissues. The osteoarthritis could cause chronic joint pain with swelling and restricted range of motion through the pathologic process including narrowing of joint space and osteophyte formation [1][2][3]. It was estimated that over 15% of the population suffered from OA, and the number tends to be doubled by 2020 due to the increasing elder population [4,5]. Though the mechanism of osteoarthritis still has not been fully clarified, a large number of risk factors have been reported, including age, sex, obesity, trauma in the joint, environmental factors, and genetic factors [3,6,7].
The SMAD3 (SMAD family member 3) gene, located on chromosome 15q22.33, acts a critical role in the joint homeostasis [8,9]. It is known as a downstream mediator in the TGF-B (transforming growth factor-b) signaling pathway which plays a key role in anabolism of chondrocytes [9]. The genetic variations of TGF-B signaling have been reported significant relationship to the OA [10]. In TGF-β signaling pathway, SMAD3 translocates into the nucleus and regulates target gene transcription and produces the phenotype in cartilage by interaction with DNA and transcription factors [11]. Several studies have been performed to investigate the association between SMAD3 polymorphisms and OA susceptibility, but the results remained unclear. Valdes [12] revealed that the SMAD3 gene rs12901499 polymorphism was associated with hip and knee OA in European populations. However, the results reported by other subsequent studies remain inconsistent and inconclusive [13][14][15][16][17]. In this study, we therefore performed a meta-analysis to evaluate whether the SMAD3 gene polymorphisms are associated with the risk of OA.

Literature search strategy
We conducted a comprehensive literature search using the electronic databases PubMed, Embase, and ISI Web of Science for relevant studies published in English (last search was updated on April 10, 2018). The search strategy was based on the following keywords: ("SMAD3" OR "SMAD family member 3") AND ("polymorphism" OR "variant" OR "SNP") AND ("osteoarthritis" OR "OA"). References of clinical trials and review articles were also searched manually for additional articles. All the literature search was performed according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (Additional file 1: Table S1) [18].

Inclusion criteria
Two researchers screened the relevant investigations and further determined the eligible studies which met the following inclusion criteria: (1) case-control or cohort design; (2) evaluating the association between SMAD3 polymorphisms and knee/hip OA susceptibility; (3) patients with OA were diagnosed based on clinical manifestation and radiographic findings, or received total joint arthroplasty because of primary OA; and (4) enough data on genotype or allele frequency for calculation of odds ratio (OR) and corresponding 95% confidence interval (CI). The animal model research, review, case reports, or the studies without sufficient data were excluded. If several articles reported findings for repeated study populations, we only selected the most recent study or the one with the largest sample size.

Data extraction
For each eligible study, two independent investigators extracted the following data: first author's name, publication year, country and ethnicity of study population, study design, OA sites, sample size, demographics of enrolled subjects, genotyping method, studied polymorphisms, and genotype distributions.

Quality assessment
According to the Newcastle-Ottawa Quality Assessment Scale (NOS) [19], the quality score of each study was based on three categories: selection (4 items, 1 point each), comparability (1 item, up to 2 points), and exposure/outcome (3 items, 1 point each). Each Fig. 1 Flow chart of the study selection process study scored from 0 point (worst) to 9 points (best), and scored 6 or less were classified as low quality, whereas studies scoring 7 or higher were defined as high quality.

Statistical analysis
All statistical analyses were conducted with STATA version 12.0 (STATA Corporation, College Station, TX, USA), and p value < 0.05 was considered significant except for the I 2 statistic. To assess the correlation between SMAD3 polymorphisms and OA susceptibility, we calculated pooled ORs with 95% CI and analyzed five genetic models: allele model, dominant model, recessive model, homozygous model, and heterozygous model.
Heterogeneity between studies was measured using Q and I 2 statistics [20]. If I 2 > 50% and p value of Q statistic < 0.10, the DerSimonian-Laird random effect model was applied to calculate pooled ORs and 95% CIs [21].
Otherwise, a fixed effect model was used as the pooling method [22].
Subgroup analysis was conducted by ethnicity, OA site. Sensitivity analysis was also performed by removing individual study sequentially in order to evaluate the stability of pooled results. We evaluated the publication bias by funnel plot and Egger's regression test. Figure 1 presented the selection process and reasons for exclusion. One hundred fifty-five articles were retrieved totally from a systematic literature search, 35 articles were removed because of duplications, and 101 articles were excluded after review of title and abstract; only 19 full-text articles remained for further evaluation. Subsequently, 4 articles were excluded because of inadequate data, 1 article was excluded due to overlapped   [12][13][14][15][16][17]. One article provided by Valdes included 5 different study populations; these 5 studies were analyzed independently. Therefore, 10 independent studies from 6 articles were included in our meta-analysis. Table 1 shows the main characteristics of included studies, and Table 2 presents the genotype and allele distributions of the SMAD3 rs12901499 polymorphism. A total of 5093 OA patients and 5699 controls were included in this study, which involved 5 Caucasian and 5

Characteristics of included studies
Asian populations. Patients diagnosed with OA were recruited according to clinical and radiographic results or ascertained by total joint arthroplasty. The genotype distribution of the control group showed conformation to Hardy-Weinberg equilibrium in all the included studies. As for the sites of OA, 9 studies examined knee OA, and 4 studies examined hip OA. Regarding the NOS scale, the quality of all the included studies was fairly high (Additional file 2: Table S2). Eventually, all the 10 studies from the articles as stated above were included in the meta-analysis for further research.
Association between SMAD3 rs12901499 polymorphism and OA susceptibility Table 3 and Fig. 2 summarize the meta-analysis results on the association between SMAD3 rs12901499 polymorphism and risk of OA. Because genotype distribution data  (Table 4).

Sensitivity and publication bias analysis
With the aid of funnel plots and Egger's test (Table 3; p egger = 0.692), we find no significant publication bias (Fig. 3). Furthermore, by using sensitivity analysis (Fig. 4), the combined results remained stable after removing individual studies. The robustness of summarized estimate was shown by the data above.

Discussion
OA, known as a degenerative disease in the aging population, is the most universal cause of joint disease which could finally result in physical disability [23]. Although OA is considered as a multifactorial disease, genetic factors are reported as vital determinants in the pathogenesis of this disorder [6,7]. Enormous attention has been paid to the association between gene SNPs and risk of osteoarthritis, and SMAD3 SNP rs12901499 was studied by several researchers. In different studies, the results ranged from no association to the strong linkage between the SNPs and the disorder [12,14,16,17]. The inconsistent findings on the associations between OA and SMAD3 rs12901499 polymorphism in Asians from relatively underpowered studies above may be attributed to factors like small sample size and different population. So we conducted this systematic review and metaanalysis to draw a more definitive conclusion.
To obtain compelling evidence of the linkage between SMAD3 rs12901499 polymorphism and risk of OA, we enrolled a total of 5093 OA patients and 5699 controls from 10 studies in this meta-analysis. The pooled results showed there is an association between OA and rs12901499 polymorphism in the overall population. And subgroup analysis stratified by ethnicity demonstrated that G allele increased the risk of OA in Caucasian. While in Asians, no association was found between SMAD3 rs12901499 polymorphism and OA risk. The results seem intriguing that there is association in the overall population but without positive result in Asian. It could be probably statistically insufficient when the majority of cases of the study were originated from Caucasian (6100/10792). And one other possibility may relate to the different expression patterns of Smad3 in non-homogeneous ethnic populations.
Sensitivity analysis and bias estimation warrant the stability of our meta-analysis.
Based on the present analysis, we propose that patients harboring the G allele of SMAD rs12901499 polymorphism experience an increased susceptibility to OA in Caucasian, though the mechanism underlying the association between  [17]. Through an in vivo study, Wu et al. found that loss of Smad3 could enhance the BMP signaling to induce articular chondrocytes hypertrophy and lose its normal phenotype [24]. Further functional studies are required to identify the role of SNP rs12901499 in OA susceptibility. The meta-analysis presented does have some limitations. Firstly, the genotype distribution data was not available in Valdes' study and only allele model among overall population was analyzed to assess the association. Secondly, OA was considered as a multifactorial disease; however, the interactions between the gene and environment were not fully addressed in this meta-analysis, which may magnify the role of Smad3 polymorphism in osteoarthritis. Thirdly, heterogeneity, which comes from different designation about interventions, participants, or outcomes in amount of studies, was inevitably existed though partially addressed by subgroup analysis stratified by ethnic groups and OA sites. Fourthly, since the data is not completely available, the subgroup analysis stratified by every potential confounders in including BMI, age, and gender could not be performed, which may lead to relatively inaccurate pooled results. Last, though there was no obvious publication bias revealed by funnel plot and Egger's test, the selection bias could not be completely removed because only studies published in English were included.

Conclusion
In conclusion, the present meta-analysis demonstrated that the G variant of SMAD3 rs12901499 polymorphism increased the risk of OA in Caucasians. By contrast, SMAD3 rs12901499 polymorphism was not associated with OA risk in Asians. Due to the limitations of our study, further well-designed studies with larger sample size in different ethnic populations should be performed to confirm these results.