The authors have declared that no competing interests exist.
Conceived and designed the experiments: YJW XY LX FLT. Performed the experiments: YJW XY XXW YZY ZXZ SMZ LX FLT. Analyzed the data: YJW XY XXW. Contributed reagents/materials/analysis tools: YJW XY XXW MTQ. Wrote the paper: YJW XY XXW. Access to full-text article: YJW XY.
Genetic variations in vitamin D receptor (VDR) may contribute to tuberculosis (TB) risk. Many studies have investigated the association between VDR BsmI gene polymorphism and TB risk, but yielded inconclusive results.
We performed a comprehensive meta-analysis of 15 publications with a total of 2309 cases and 3568 controls. We assessed the strength of the association between VDR BsmI gene polymorphism and TB risk and performed sub-group analyses by ethnicity, sample size and Hardy–Weinberg equilibrium (HWE). We found a statistically significant correlation between VDR BsmI gene polymorphism and decreased TB risk in four comparison models: allele model (b vs. B: OR = 0.78, 95% CI = 0.67, 0.89; Pheterogeneity = 0.004), homozygote model (bb vs. BB: OR = 0.61, 95% CI = 0.43, 0.87; Pheterogeneity = 0.001), recessive model (bb vs. Bb+BB: OR = 0.70, 95% CI = 0.56, 0.88; Pheterogeneity = 0.005) and dominant model (bb+Bb vs. BB: OR = 0.77, 95% CI = 0.61, 0.97; Pheterogeneity = 0.010), especially in studies based on Asian population. Sub-group analyses also revealed that there was a statistically decreased TB risk in “small” studies (<500 participants) and studies with PHWE>0.5. Meta-regression and stratification analysis both showed that the ethnicity and sample size contributed to heterogeneity.
This meta-analysis suggests that VDR BsmI gene polymorphism is associated with a significant decreased TB risk, especially in Asian population.
According to the latest information of the tuberculosis (TB) epidemic provided by the World Health Organization (WHO) Global Tuberculosis Report 2012, the TB mortality and incidence rates have been decreasing for several years. However, the global burden of TB remains enormous, especially in South-East Asia and South Africa, the number of new cases in 2011 is 8,700,000 including 1,131,000 HIV-infected patients; and 1,400,000 people died from TB, including 1,000,000 HIV-negative patients and 430,000 HIV-positive patients
The interaction of 1,25-dihydroxyvitamin D3 with the vitamin D receptor (VDR) is able to activate monocytes, stimulate cell-mediated immunity and suppress lymphocyte proliferation
To confirm the association between vitamin D receptor BsmI gene polymorphism and TB risk, we performed this meta-analysis by calculating the estimate of overall TB risk and evaluated influence of ethnicity, sample size and Hardy–Weinberg equilibrium (HWE).
According to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA), a systematic literature searching was performed on PubMed, EMBASE and Chinese National Knowledge Infrastructure (CNKI) up to March 19, 2013. In order to identify as many relative articles as possible, the search strategy was based on combinations of “Vitamin D receptor” or “VDR”; “BsmI”, “rs1544410”; “polymorphism”, “variant”, “locus” or “SNP”; “tuberculosis”. Furthermore, we performed manual search of references of relative articles and reviews. To minimize potential publication bias, no restrictions were placed on language, sample size and time period.
Studies were selected according to the following inclusion criteria: (a) investigation on the association of VDR BsmI gene polymorphism with TB risk; (b) case-control studies; (c) genotype distribution information in cases and controls; (d) genotype frequencies found in the control group deviate from Hardy-Weinberg equilibrium (HWE). The major reasons for exclusion of studies were (a) overlapping data; (b) case-only studies, reviews, letters and editorials; (c) studies without detailed BsmI genotype frequencies.
Data of each eligible study was extracted by two reviewers (Wu and Yang) independently with a standard data-collection method. The following data was extracted: first author, year, country, ethnicity, source of controls, TB diagnosis standards, number of cases and controls, genotype frequency in cases and controls. Different ethnicities were categorized as Asian and African. According to the source of controls, all eligible studies were defined as hospital-based (HB) and population-based (PB). The Hardy–Weinberg equilibrium (HWE) was examined by Chi-square test (p<0.05 was considered as significant disequilibrium) based on BsmI genotyping distribution in controls
The strength of the association between VDR BsmI gene polymorphism and TB risk was evaluated by calculating pooled odds ratio (OR) with 95% confidence intervals (95% CI). The pooled ORs were calculated for five comparison models: allele model (b vs. B), homozygote model (bb vs. BB), heterozygote model (Bb vs. BB), dominant model (bb+Bb vs. BB) and recessive model (bb vs. Bb+BB). The statistical heterogeneity between studies was checked using Chi-square based Q test and considered significant at P<0.1
Study selection process was shown in
Firstauthor | Year | Country | Ethnicity | Control | Diagnosis standards | HIVInfection | HWE | Cases | Controls | ||||
BB | Bb | bb | BB | Bb | bb | ||||||||
2004 | Malawi | African | PB | Bacteriology | Positive | Yesa | 212 | 123 | 10 | 314 | 192 | 39 | |
2004 | India | Asian | PB | Clinical symptoms, bacteriology,X-ray | Negative | Yes | 16 | 24 | 6 | 18 | 27 | 19 | |
2004 | West Africa | African | PB | Clinical symptoms, bacteriology,X-ray | Mixed | Yes | 20 | 108 | 215 | 39 | 208 | 387 | |
2006 | South Africa | African | HB | Clinical symptoms, bacteriology | Negative | Nob | 6 | 38 | 60 | 9 | 32 | 76 | |
2007 | West Africa | African | PB | Clinical symptoms, bacteriology | Mixed | Yes | 146 | 141 | 33 | 152 | 152 | 38 | |
2008 | India | Asian | PB | Clinical symptoms, bacteriology,X-ray | Negative | No | 23 | 16 | 12 | 16 | 17 | 27 | |
2009 | India | Asian | PB | Clinical symptoms, bacteriology,X-ray | Positive | Yes | 40 | 47 | 20 | 39 | 62 | 45 | |
2009 | India | Asian | PB | Clinical symptoms, bacteriology,X-ray | Negative | Yes | 27 | 22 | 16 | 16 | 23 | 21 | |
2009 | India | Asian | PB | Clinical symptoms, bacteriology,X-ray | NA | No | 16 | 14 | 10 | 15 | 13 | 21 | |
2009 | Iran | Asian | PB | Clinical symptoms, bacteriology,X-ray | NA | No | 7 | 67 | 43 | 13 | 21 | 26 | |
2010 | Iran | Asian | PB | Bacteriology,X-ray | NA | No | 23 | 86 | 55 | 0 | 29 | 21 | |
2011 | India | Asian | PB | Clinical symptoms, bacteriology,X-ray | NA | No | 73 | 102 | 53 | 274 | 568 | 211 | |
2011 | India | Asian | PB | Clinical symptoms, bacteriology,X-ray, PPD | Negative | No | 32 | 52 | 17 | 57 | 134 | 34 | |
2011 | Istanbul | Asian | PB | Clinical symptoms, bacteriology,X-ray | NA | Yes | 28 | 68 | 32 | 5 | 38 | 37 | |
2011 | Korean | Asian | PB | Bacteriology | NA | Yes | 2 | 13 | 135 | 0 | 8 | 75 |
PB: population-based; HB: hospital-based; a: studies with PHWE>0.05; b: studies with PHWE<0.05.
We observed a significant decreased risk of TB susceptibility in allele model (b vs. B: OR = 0.78, 95% CI = 0.67, 0.89; Pheterogeneity = 0.004), homozygote model (bb vs. BB: OR = 0.61, 95% CI = 0.43, 0.87; Pheterogeneity = 0.001,
b vs. B | bb vs. BB | Bb vs. BB | bb+Bb vs. BB | bb vs. Bb+BB | |||||||
N | OR | Ph | OR | Ph | OR | Ph | OR | Ph | OR | Ph | |
15 | 0.78(0.67,0.89) |
0.004 | 0.61(0.43,0.87) |
0.001 | 0.86(0.67,1.09) | 0.016 | 0.77(0.61,0.97) |
0.010 | 0.70(0.56,0.88) |
0.005 | |
4 | 0.92(0.81,1.06) | 0.351 | 0.81(0.50,1.30) | 0.113 | 0.98(0.80–1.20) | 0.769 | 0.93(0.77,1.12) | 0.777 | 0.80(0.54,1.16) | 0.055 | |
11 | 0.70(0.57,0.85) |
0.01 | 0.52(0.32,0.85) |
0.001 | 0.77(0.53,1.12) | 0.012 | 0.66(0.46,0.94) |
0.009 | 0.65(0.49,0.88) |
0.015 | |
4 | 0.94(0.84,1.05) | 0.336 | 0.82(0.56,1.19) | 0.116 | 0.88(0.73,1.05) | 0.353 | 0.86(0.73,1.02) | 0.622 | 0.92(0.65,1.30) | 0.038 | |
11 | 0.68(0.56,0.82) |
0.068 | 0.51(0.31,0.86) |
0.005 | 0.84(0.53,1.33) | 0.007 | 0.68(0.44,1.03) | 0.006 | 0.61(0.49,0.77) |
0.346 | |
8 | 0.75(0.61,0.92) |
0.009 | 0.50(0.32,0.79) |
0.020 | 0.89(0.74,1.06) | 0.497 | 0.76(0.60,0.97) |
0.150 | 0.63(0.44,0.90) |
0.007 | |
7 | 0.80(0.64,1.00) | 0.050 | 0.78(0.45,1.37) | 0.013 | 0.99(0.55,1.78) | 0.002 | 0.85(0.51,1.40) | 0.005 | 0.78(0.57,1.07) | 0.080 |
N: number of studies included; OR: odds ratio; Ph: p value for heterogeneity;
OR with statistical significance;
:studies with more than 1000 participants;
:studies with less than 1000 participants;
:studies with PHWE>0.05;
:studies with PHWE<0.05.
Furthermore, we performed sub-group analyses to explore the effect of ethnicity, sample size and HWE. As for ethnicities, a decreased TB risk was found in Asian population in four comparison models: allele model (b vs. B: OR = 0.70, 95% CI = 0.57, 0.85; Pheterogeneity = 0.010), homozygote model (bb vs. BB: OR = 0.52, 95% CI = 0.32, 0.85; Pheterogeneity = 0.001), recessive model (bb vs. Bb+BB: OR = 0.65, 95% CI = 0.49, 0.88; Pheterogeneity = 0.015) and dominant model (bb+Bb vs. BB: OR = 0.66, 95% CI = 0.46, 0.94; Pheterogeneity = 0.009). In Africans, however, no significant association except only a reduced trend was found in each model.
When stratified by sample size, we found a decreased risk of TB in “small” studies for three comparison models: allele model (b vs. B: OR = 0.68, 95% CI = 0.56, 0.82; Pheterogeneity = 0.068), homozygote model (bb vs. BB: OR = 0.51, 95% CI = 0.31, 0.86; Pheterogeneity = 0.005), and recessive model (bb vs. Bb+BB: OR = 0.61, 95% CI = 0.49, 0.77; Pheterogeneity = 0.346). In a coincidence, three studies
Further, in the stratified analyses by HWE, we observed the BsmI polymorphism was correlated with a decreased risk of TB in studies with PHWE>0.05 for four comparison models: allele model (b vs. B: OR = 0.75, 95% CI = 0.61, 0.92; Pheterogeneity = 0.009), homozygote model (bb vs. BB: OR = 0.50, 95% CI = 0.32, 0.79; Pheterogeneity = 0.020), recessive model (bb vs. Bb+BB: OR = 0.63, 95% CI = 0.44, 0.90; Pheterogeneity = 0.007) and dominant model (bb+Bb vs. BB: OR = 0.76, 95% CI = 0.60, 0.97; Pheterogeneity = 0.150).
Heterogeneity between studies in each model is shown in
Sensitivity analysis was performed to assess the influence of each individual study on the pooled OR by deleting one single study each time. The results showed that no individual study affected the pooled OR significantly, suggesting stability of this meta-analysis (
A: Funnel plot analysis of homozygote model (bb vs. BB). Egger’s test p = 0.617, Begg’s test p = 0.921; B: Funnel plot analysis of dominant model (bb+Bb vs. BB). Egger’s test p = 0.685, Begg’s test p = 1.000; the circles represent the weight of individual study.
Tuberculosis is second only to HIV as the greatest killer due to a single infectious agent worldwide. Gene mutation could affect the function of VDR gene and might be associated with TB risk. In recent years, given the potential roles of VDR playing in the etiology of TB, more studies have been conducted to identify whether the VDR BsmI gene polymorphism was the genetic determiner of TB. However, these studies yielded different or even controversial results. In this meta-analysis, we found that the b allele was associated with a significant decreased risk of TB in four main comparison models. Therefore, it was first reported that VDR BsmI gene polymorphism could predict the susceptibility of TB.
Linkage disequilibrium measure is used to describe the association of alleles of adjacent polymorphisms with each other. BsmI (B>b), ApaI(A>a) and TaqI(T>t) are located near the 3′UTR of VDR gene through the strong linkage disequilibrium, and the extended genotype baT is associated with increased level of VDR
In this meta-analysis, we demonstrated that VDR BsmI gene polymorphism (B>b) contributed to TB risk in allele model (b vs. B), homozygote model (bb vs. BB), recessive model (bb vs. Bb+BB) and dominant model (bb+Bb vs. BB), which was in inconsistent with several previous independent studies
As we know, gene polymorphisms are complicated and fluctuating, which mainly attributed to different ethnicities. Moreover, the burden of TB is highest in Asia and Africa geographically. Therefore, we performed a sub-group analysis on ethnicities in this meta-analysis. It showed that the VDR BsmI gene polymorphism had a decreased risk of TB in Asians rather than Africans, especially in the same four comparison models above mentioned (
During other sub-group analyses, we also found that sample size greatly affected the association between VDR BsmI gene polymorphism and TB risk. As shown in
HWE is essential for a sound case–control study. It is probable that studies without HWE in controls have selection bias or genotyping error, which may cause misleading results. In this meta-analysis, we found the BsmI polymorphism was significantly associated with a decreased risk of TB in studies with PHWE>0.05. And five studies with PHWE>0.05 were based on Asians
For heterogeneity, we found ethnicity and sample size were the source of heterogeneity, which was in consistent with sub-group analyses. When studies were stratified by ethnicity and sample size, we conducted heterogeneity test to explore the source of heterogeneity. The obvious heterogeneity existed in Asians and “small” studies. The racial differences might be responsible for the foremost source of heterogeneity. In addition, the “small” studies may be affected by the small-study effect, in which effects reported are larger, and lead to between studies variance. Thirdly, we observed that eligible studies were conducted in seven countries in Asia and Africa respectively. The environmental factors which may influence VDR gene expression are quite different in these countries. For example, the intensity and hours of sunlight in South Africa are stronger and longer than Korean by geography. Fourthly, the different experimental designs, diagnosis standards, HIV status and even socialized medical care also may contribute to the heterogeneity. Thus this kind of heterogeneity is hard to exclude, because recruitment of enough TB cases came from specific regions is difficult.
As for the aforementioned publication bias detected by Begg’s funnel plot and Egger’ test, the funnel plot was roughly symmetrical and no publication bias had been detected. Although Merza’s study was the only study reported increased TB risk in homozygote, heterozygote and dominant model, the effect for the publication bias was small. So we had not ruled out the Merza’s study. The results of this meta-analysis are relatively stable and reliable.
Although a previous meta-analysis conducted by Gao et al suggested the BsmI polymorphism was related to host susceptibility to TB in Asians
In conclusion, results from this meta-analysis demonstrate that VDR BsmI gene polymorphism is associated with decreased TB risk, especially in Asian population. To confirm this association, future large scale case-control studies are required to validate these findings.
(TIF)
(TIF)
(DOC)