Meta-Analysis of the Association between Vitamin D Receptor Polymorphisms and the Risk of Autoimmune Thyroid Disease

The association between vitamin D receptor (VDR) polymorphisms (rs731236, rs1544410, rs2228570, and rs7975232) and the risk of autoimmune thyroid disease (AITD) had been investigated in previous studies. However, the results of these studies remained controversial. Thus, a meta-analysis was performed to derive a more precise conclusion. All related articles were systematically searched by PubMed, Embase, Google Scholar, and Chinese National Knowledge Infrastructure (CNKI). The pooled odds ratios (ORs) with 95% confidence intervals (CIs) were calculated to assess the strength of association. The overall results indicated that VDR rs731236 and rs2228570 polymorphisms were significantly associated with a reduced risk of AITD. However, a stratification analysis based on clinical types showed that VDR rs731236 and rs2228570 polymorphisms were associated only with a reduced risk of HT. A stratification analysis by ethnicity showed that VDR rs731236 polymorphism was significantly associated with a reduced risk of AITD in Asian and African populations. VDR rs2228570 polymorphism was associated with a reduced risk of AITD in Asian populations. VDR rs1544410 polymorphism was associated with a reduced risk of AITD in European and African populations, but with an increased risk of AITD in Asian populations. VDR rs7975232 polymorphism was significantly associated with an increased risk of AITD in African populations. In conclusion, the present study suggested that VDR rs731236, rs1544410, rs2228570, and rs7975232 polymorphisms were significantly associated with AITD risk. However, more well-designed studies should be performed to verify the current results.


Introduction
Autoimmune thyroid disease (AITD), mainly including Graves' disease (GD) and Hashimoto's thyroiditis (HT), is an organ-specific autoimmune disease and affects up to 5% of the general population [1,2]. Although the pathogenesis of AITD is still unknown, it is generally acknowledged that environmental factors and the intrinsic genetic predisposition of an individual play critical roles in the occurrence of the disease [3]. Environmental factors, such as viral infections, irradiation, drugs, and iodine intake, may involve interference with thyroid function, direct toxic effects on thyrocytes, immune stimulation, or other immunomodulatory effects [3,4]. However, it is difficult to directly link an environmental exposure with AITD due to the intervention of genetic factors. Increasing evidence suggests that single-nucleotide polymorphisms (SNPs) in AITD-related genes can influence individual predisposition to the disease [5][6][7].
Vitamin D is a fat-soluble vitamin and is activated in the liver and kidney [8]. Activated vitamin D [1,25(OH) 2 D] can promote the differentiation of monocytes and inhibit the maturation of dendritic cells [9]. Furthermore, it can also suppress the production of cytokines, such as interleukin-1, interleukin-2, interleukin-6, and tumor necrosis factor [10]. These cytokines play important roles in the development of lymphocytes, which are believed to be involved in the pathogenesis of autoimmune diseases. The immunomodulatory actions of 1,25(OH) 2 D are mediated by its binding to vitamin D receptor (VDR), which belongs to the family of trans-acting transcriptional regulatory factors and is widely expressed in various immune cell subsets, including lymphocytes, macrophages, and several endocrine cells [11]. The gene encoding VDR contains 14 exons and spans approximately 75 kilobases on chromosome 12q13.11. Many SNPs have been identified in the VDR gene [12]. Among them, the association of VDR rs731236, rs1544410, rs2228570, and rs7975232 polymorphisms with AITD risk has been widely reported . However, the results are inconsistent and ambiguous. Furthermore, considering that a single-center pilot study with small sample sizes may possess low statistical power, we performed a meta-analysis of all eligible studies to obtain a more precise conclusion.

Search Strategy.
All related articles were obtained by systematically searching PubMed, Embase, Google Scholar, and Chinese National Knowledge Infrastructure (CNKI). The search keywords were as follows: "vitamin D receptor OR VDR," "polymorphism OR genetic variation OR genetic variant," and "autoimmune thyroid disease OR AITD OR thyroid." There were no limitations on language and year of publication. The last search was updated on August 28, 2017. Furthermore, the references of all related articles were also retrieved to find other eligible studies.

Inclusion and Exclusion
Criteria. All eligible studies must meet the following inclusion criteria: (a) case-control studies; (b) evaluation of the association between VDR polymorphisms (rs731236, rs1544410, rs2228570, and rs7975232) and AITD risk; and (c) available genotype/allele frequencies.
In addition, the exclusion criteria were as follows: (a) letters, reviews, and case reports and (b) duplicate publication. If multiple studies had overlapping data, only those with complete data were included.

Data Extraction.
Two authors independently reviewed the related articles and extracted the following data: first author's name, year of publication, region, ethnicity, genotyping methods, the number of cases and controls, and genotype/allele frequency. Any disagreement was resolved by discussion with each other.

Statistical Analysis.
Hardy-Weinberg equilibrium (HWE) in the control group of each study was calculated by chi-square goodness-of-fit test, and P HWE < 0 05 was considered as a deviation from HWE. The strength of the association between VDR polymorphisms (rs731236, rs1544410, rs2228570, and rs7975232) and AITD risk was evaluated by the pooled odds ratios (ORs) with 95% confidence intervals (CIs). The significance of the pooled ORs was assessed by the Z test, and P Z < 0 05 was considered statistically significant. The chi-square based Q-test was used to investigate the between-study heterogeneity. If P H < 0 1 indicated the existence of between-study heterogeneity, the random-effect model was used to calculate the pooled ORs; otherwise, the fixed-effect model was applied for the analysis. A sensitivity analysis was conducted by omitting one study each time to estimate the stability of the result. Publication bias was determined by Begg's funnel plot and Egger's test. A symmetric funnel plot and P value of Egger's test more than 0.05 indicated the lack of publication bias. All statistical tests were performed using Review Manager 5.2 (The Nordic Cochrane Centre, The Cochrane Collaboration, Copenhagen) and

Study Selection and
Characteristics. The flowchart for identifying eligible studies is shown in Figure 1. A total of 139 articles were obtained through an initial search. After reviewing the titles and abstracts of these articles, we excluded 113 unrelated articles. The remaining articles were further checked by a full-text review. Finally, 22 articles met the inclusion criteria and were included in this metaanalysis. The main characteristics of all included articles are shown in Tables  3.2. Quantitative Synthesis. The association between VDR rs731236 polymorphism and AITD risk is shown in Table 2 and Figure 2. In the overall analysis, a significant association was found in homozygote comparison and recessive models (CC versus TT: OR = 0 67, 95% CI: 0.48-0.93, P Z = 0 02; CC versus CT + TT: OR = 0 80, 95%   The association of VDR rs1544410 polymorphism with AITD risk is presented in Table 3 and Figure 2. No significant association was observed in the overall analysis and stratification analysis by clinical types. However, a subgroup analysis by ethnicity showed that VDR rs1544410 polymorphism was associated with a reduced risk of AITD in European The association between VDR rs2228570 polymorphism and AITD risk is shown in Table 4  For VDR rs7975232 polymorphism, no significant association was observed in the overall analysis and stratification analysis by clinical types and P HWE (Table 5 and Figure 2). However, a stratification analysis by ethnicity showed that VDR rs7975232 polymorphism was associated with an

Sensitivity Analysis and Publication Bias.
A sensitivity analysis showed that the pooled OR values were not substantially changed after one study deletion each time, which suggests that results of this meta-analysis were stable. As shown in Figure 3, the shape of funnel plots was symmetric. In addition, all P values of Egger's test were more than 0.05, indicating the lack of publication bias ( Table 6).

Discussion
As an immune modulator, vitamin D is involved in the onset and development of AITD [35,36]. Low levels of vitamin D have been demonstrated in patients with AITD [36]. Furthermore, vitamin D deficiency was correlated with the duration of HT, which led to an increase in thyroid volume and in antithyroid antibodies levels [36]. Vitamin D exerts its biological effects by binding to VDR and activating VDRresponsive genes [37]. VDR is an intracellular receptor belonging to the steroid/thyroid nuclear receptor family and expressed in human immune cells including macrophages, dendritic cells, and T and B lymphocytes [35]. Therefore, the abnormal function of VDR, which is attributable to    [39]. In view of all this, these functional polymorphisms were speculated to be associated with AITD risk. Interestingly, some epidemiological studies confirmed the speculation and found significant association between these polymorphisms and AITD risk.  [19]. However, other studies including genome-wide association study showed that these polymorphisms did not influence individual susceptibility to AITD [13,18,20,23,24,40,41]. These inconsistent results may be due to the fact that a single-center pilot study with small sample sizes has low statistical power to detect a true association or that the genetic background of different populations changes the effect of low-penetration polymorphisms on AITD risk. In 2013, Feng et al. tried to clarify the association by metaanalysis and found that VDR rs1544410 and rs731236 polymorphisms were significantly associated with a reduced risk of AITD, while VDR rs7975232 and rs2228570 polymorphisms were not associated with AITD risk [42]. Due to the limited number of related studies, subgroup analyses by clinical types were not conducted. Furthermore, results of recent studies were still inconsistent with that of previous meta-analysis [13,15,16]. Therefore, an updated analysis was performed by combining the recent studies. Results indicated that VDR rs731236 and rs2228570 polymorphisms were significantly associated with reduced risk of AITD. A further stratification analysis based on clinical types showed that VDR rs731236 and rs2228570 polymorphisms were associated only with reduced risk of HT. In a stratification analysis based on ethnicity, VDR rs731236 polymorphism was associated with a reduced risk of AITD in Asian and African populations but not in European populations. VDR rs2228570 polymorphism was associated with a reduced risk of AITD only in Asian populations. It was worthy to note that VDR rs1544410 polymorphism was associated with a reduced risk of AITD in European and African populations but has an increased risk of AITD in Asian populations. A stratification analysis by P HWE values showed that the significant association of VDR rs2228570 polymorphism with a reduced risk of AITD was observed only in studies with P HWE ≤ 0 05, which indicated that the effect of VDR rs2228570 polymorphism on AITD risk needed to be interpreted cautiously. Heterogeneity was observed in the current meta-analysis. We tried to investigate the sources of heterogeneity by a stratification analysis based on clinical types, ethnicity, and HWE, but the investigation results were not satisfactory and could not provide a reasonable explanation for the sources of heterogeneity. In view of factors affecting vitamin D levels and methodological issues, heterogeneity may result from differences in economic and public health indexes among different countries, variations in environment and climate, age and gender mismatch in published studies, and variations in diagnostic criteria for GD/HT.
Although the current results showed the statistically significant associations of VDR polymorphisms with AITD risk, such associations had a small influence on the occurrence of AITD. In addition, several limitations impeding accurate assessment should be noted. Firstly, raw data such as gender, age, living style, and drug consumption could not be obtained from all included studies. Secondly, the number of studies was small in the subgroup analysis, especially in African populations. Last but not least, the present analysis did not consider the gene-gene and gene-environment interactions.
In conclusion, the present study suggested that VDR rs731236, rs1544410, rs2228570, and rs7975232 polymorphisms were significantly associated with AITD risk. However, more well-designed studies, especially studies on African populations, should be performed to verify the results.