 |
|
||||
| Home | Login | |||||
| About | Editorial board | Articles
|
NSI Publications
|
Search | Instructions | Online Submission | Subscribe | Videos | Etcetera | Contact |
|
||||||||||||||||||||
|
|
|
A meta-analysis of interleukin-10-1082 promoter genetic polymorphism associated with atherosclerotic risk
Correspondence Address: Source of Support: None, Conflict of Interest: None DOI: 10.4103/0028-3886.132323
Objective: This meta-analysis was conducted to assess the relationship between interleukin-10-1082 G/A single nucleotide polymorphism with atherosclerosis (AS) risk. Materials and Methods: The databases of PubMed, EMBASE, Chinese National Knowledge Infrastructure and Wan-Fang were searched from January 2000 to January 2014. 16 studies (involving 7779 cases and 7271 controls) were finally included. Each eligible study was scored for quality assessment. We adopted the most probably appropriate genetic model (recessive model) after carefully calculation. Between study heterogeneity was explored by subgroup analysis and publication bias was estimated by Begg's funnel plot and Egger's regression test. Results: Statistically significant association was observed between AA genotype with overall AS risk, being mainly in coronary heart disease and stroke subgroups among Asian population, and peripheral artery disease (PAD) subgroup among Caucasians. Conclusions: Interleukin-10-1082 AA genotype is associated with increased overall AS risk. AA carriers of Asians seem to be more susceptible to coronary artery disease and stroke, and Caucasians are more susceptible to PAD. Keywords: Atherosclerosis, gene, interleukin-10, meta-analysis, single nucleotide polymorphism
Nowadays, coronary artery disease (CAD) and stroke remains the leading causes of death in developed countries, which are closely related to AS. Traditional risk factors such as smoking, obesity, dyslipidemia, hypertension, and diabetes mellitus are widely accepted. However, with the development of gene technology, more and more scientists have turned their focus on genetic risk factors and gene-environment interaction. As an important proinflammatory cytokine, interleukin-10 (IL-10) is widely known for its anti-inflammatory and B-cell- stimulating function. [1] It seems to be involved in both AS development and its acute complications, such as rupture and thrombosis. [2] Thus, the IL-10 gene appears to be a good candidate for AS studies. Currently, many epidemiological studies have focused on the association between IL-10-1082 G/A single nucleotide polymorphism (SNP) and AS risk, but the results are controversial. As far as we know, only one published meta-analysis suggested that IL-10-1082 G/A SNP was associated with CAD in Caucasians, [3] but not with the suitable genetic model. Accordingly, we designed this meta-analysis to assess the relationship between IL-10-1082 G/A SNP with AS risk.
Literature search strategy The electronic databases of PubMed, EMBASE, Chinese National Knowledge Infrastructure and Wan-Fang were searched from January 2000 to January 2014 with the following terms, "Interleukin-10" and "SNP" and "atherosclerosis." The retrieval process was conducted without limitation on language, and the relevant references were also searched manually. Study selection and data extraction We evaluated the titles and abstracts of all relevant publications, but excluded case reports, animal experiments, editorials, or review articles. To be included in this meta-analysis, studies must meet the following criteria: 1. Any case-control studies evaluate the association of IL-10-1082 G/A SNP and AS risk; 2. The study reported the characteristics of the study population; 3. Sufficient data could be derived from this study for our meta-analysis; 4. The method of data collection and analysis were statistically acceptable. Data were extracted by two reviewers (Huang and Jin) independently. Any discrepancy was discussed and resolved by consensus or a third reviewer (Li). The following information were extracted including the first author, publication year, ethnicity, number of cases and controls, type of disease, genotyping method, and genotype frequency. P values for Hardy-Weiberg equilibrium were recalculated in controls. Quality assessment The quality of included studies were assessed by two reviewers (Huang and Jin) independently following the criteria of Newcastle-Ottawa Scale (NOS). [4] The third reviewer (Li) checked the results, and a consensus was reached. The criteria of NOS for case-control study involved population selection, comparability, outcome, and ascertainment of exposure in cases and controls. Scores ranged 0-9. Studies with a score less than 6 were excluded. Statistical analysis This meta-analysis was performed using Stata 12.0 (StataCorp LP, College Station, TX). The strength of association between IL-10-1082 G/A SNP and AS was assessed by odds ratio (OR) and 95% confidence interval (CI). The Q statistic and I2 statistic [5],[6] were used to test between-study heterogeneity. The Mantel-Haenszel method for fixed effects was adopted if P > 0.1 and I2 <50%, [7] and the Der-Simonian-Laird method for random-effects was used if P < 0.1 and I2 >50%. [5] Subgroup analysis was used to explore potential heterogeneity. We also performed a sensitivity analysis to test the stability of the results. The dynamic changing of the correlation between IL-10-1082 SNP and AS risk was estimated by cumulative meta-analysis. Finally, publication bias was estimated by Begg's funnel plot and Egger's regression test. Genetic model determination To determine the overall gene effect, the model that includes gene was compared with the model without gene in allelic model (A vs. G). If the overall gene effect was statistically significant, further comparisons of OR1 (AA vs. GG), OR2 (GA vs. GG), and OR3 (AA vs. GA) were explored. [8] If OR1 and OR3 are significant and OR2 is nonsignificant, a recessive model (AA vs. GA + GG) is suggested. If OR1 and OR2 are significant and OR3 is nonsignificant, then a dominant model (AA + GA vs. GG) is implied. If OR2 and OR3 are significant and OR1 is nonsignificant, then a complete overdominant model (GA vs. AA + GG) is suggested. If OR, OR2 and OR3 are all significant, and the effect size of OR1 are twice than that of OR1 and OR2, then a codominant model (AA vs. GA vs. GG) is indicated. Multiple comparisons of genotype effects and possible inheritance model were listed in [Table 1].
Literature search and study selection The flow diagram of the study selection was shown in [Figure 1], and characteristics of included studies were listed in [Table 2]. After the initial search, a total of 251 articles in English and 5 in Chinese were identified. 235 articles were excluded after screening the titles and abstracts, 5 articles were excluded for low quality. Finally, 16 studies [9],[10],[11],[12],[13],[14],[15],[16],[17],[18],[19],[20],[21],[22],[23] catered to the inclusion criteria were included, involving a total of 7779 cases and 7271 controls. The involving AS diseases in this meta-analysis were CAD (including myocardial infarction, ischemic heart disease and acute coronary syndrome), stroke (including ischemic stroke and cerebral infarction) and peripheral artery disease (PAD) (including renal artery stenosis and abdominal aortic aneurysm). Among them nine studies were conducted in Caucasians, six studies were conducted in Asians, and one study was conducted in Latinos.
Determination of genetic model After preliminary calculation, 'A' allele was found to be the risk factor in allelic model (A vs. G), the pooled OR (95% CI, P value) was 1.22 (1.08, 1.38, P = 0.002) in the random-effect model. The pooled ORs of multiple comparisons OR1, OR2, OR3 were 1.39 (1.09, 1.78, P = 0.009); 1.12 (0.94, 1.34, P = 0.199), and 1.24 (1.05, 1.47, P = 0.012) respectively. The Wald test indicated that OR1 and OR3 were significant (P < 0.05), and OR2 was nonsignificant (P = 0.199), so the genetic model was most likely to be recessive. In order to find more evidence, OR1, OR2, OR3 were calculated in high quality studies (scores higher than 7). The pooler ORs were 1.39 (1.08, 1.79, P = 0.011); 1.12 (0.94, 1.34, P = 0.214); and 1.23 (1.02, 1.49, P = 0.030), respectively. Hence, the recessive model was determined, and genotype AA was compared with the combined genotype GA-plus-GG. Overall and subgroup analysis As shown in [Figure 2], in subgroup analysis based on ethnicity, statistically significant association was observed in Asians and the overall ethnicity, the pooled ORs (95% CI, P value) were 1.43 (1.19, 1.72, P = 0.000) and 1.28 (1.08, 1.52, P = 0.004), respectively.
Further subgroup analysis based on different AS diseases were performed in each ethnicity (CAD in Asians; stroke in Asians; CAD in Caucasians; stroke in Caucasians; PAD in Caucasians and CAD in Latinos). Statistically significant association was observed in stroke subgroup in Asians, and PAD subgroup in Caucasians [Figure 3]. The pooler ORs (95% CI, P value) were 1.55 (1.13, 2.12; P = 0.007) and 1.67 (1.04, 2.69; P = 0.034), respectively. However, we noticed the heterogeneity in CAD subgroup of Asians was acceptable (P = 0.276, I2 = 22.2%), so this subgroup was recalculated in fixed effect model, the pooled OR (95% CI, P value) was 1.37 (1.17, 1.70; P = 0.004).
When data sets were sorted according to genotyping method [Figure 4], heterogeneity in each subgroup was successfully removed (P > 0.1 and I2 < 50%), except those only contained one study. Statistically significant association could be observed in amplification refractory mutation system (ARMS) polymerase chain reaction (PCR) and PCR-sequence specific primer (SSP) subgroups. The pooled ORs (95% CI, P value) were 1.55 (1.28, 1.87; P = 0.000) and 2.49 (1.79, 3.46; P = 0.000), respectively.
Sensitivity analysis Sensitivity analysis was conducted by omitting each study in turn [Figure 5]. The results indicated that random-effects estimates before and after the removal of each study were similar at large. The most influencing study was the one conducted by Lio. The pooled OR (95% CI) was 1.28 (1.08, 1.52) before the removal of this study and 1.18 (1.04, 1.35) after the removal of this study, suggesting high stability of this meta-analysis.
Cumulative meta-analysis Cumulative meta-analysis was conducted by an assortment of the total number of sample size. As shown in [Figure 6], the inclination, though slightly undulated, toward significant association with overall AS risk.
Publication bias Potential publication bias was estimated by Begg's funnel plot and Egger's regression test. Dots seemed to be symmetrically distributed through visual assessment [Figure 7], and P values were 0.096 in Begg's test and 0.165 in Egger's test, suggesting no publication bias exist.
This meta-analysis was performed to make a better understanding of IL-10-1082 G/A SNP with AS risk. We found the risk of AA genotype carriers developing AS was 1.28-fold higher than 'G' allele (GG-plus-GA) carriers in overall ethnicity and 1.43-fold higher in Asian population. In order to make a more comprehensive understanding, further subgroup analysis was performed based on AS disease in each ethnicity. In Asians, the risk of AA carriers developing CAD and stroke were 1.37-fold higher and 1.55-fold higher than 'G' allele carriers (GA-plus-GG). While in Caucasians, the risk of developing PAD was 1.67-fold higher. This difference could be explained by gene-to-gene or/and gene-to-environmental interactions. Of course, potential publication bias or deficiencies of experiments could also lead to possible false positives. In this meta-analysis, heterogeneity was effectively removed in subgroup analysis based on genotyping method. Statistically significant association was observed in ARMS PCR and PCR-SSP groups, both indicated higher AS risk of AA genotype. Although heterogeneity shouldn't be the reason that certain PCR method can't be used, but the sensitivity and specificity of PCR technic should be further explored to reduce possible genotyping error. Human IL-10 gene has been mapped to chromosome 1q31-32 [24] with three important gene locus mutation in the upstream of the transcription start site (-1082G/A, -819C/T and -592C/A). [25] To our knowledge, IL-10 production is partly regulated by genes, but no matter in cytology experiments or epidemiological researches the results are controversial. In vitro experiment conducted by Turner, [25] the presence of GG genotype could down-regulate IL-10 expression. However, Eskdale's et al. study [26] suggested reduced IL-10 production of 'G' allele. As for epidemiology studies, Lio et al. [14] suggested that AA genotype was associated with lower IL-10 production and lower coronary AS risk, but Koch's et al. study [13] indicated that either 'G' or 'A' allele was not associated with coronary AS. So it's necessary to perform this meta-analysis and shed more light on IL-10-1082G/A SNP with AS. The strength of our meta-analysis could be summarized as follow. Firstly, we cautiously determined the probably most suitable genetic model through carefully calculation. Secondly, potential heterogeneity was successfully removed by subgroup analysis based on genotyping method. Thirdly, sufficient statistical methods were used to get credible conclusions. There were also some limitations need to be taken into consideration before accepting the results. First, significant heterogeneity was noticed in overall and subgroup analysis, so this part was pooled in the random-effect model, which may lead to inaccurate results. Second, because there is relatively lack of studies on -1082 G/A polymorphism and PAD in Asians, as well as in Latinos, so the results should be interpreted with caution. Third, the human IL-10 production is regulated by complicated gene-environment interaction, the effect of single genetic mutation is limited. So the evidence offered by this meta-analysis should be accepted with caution.
This meta-analysis provides evidence for IL-10-1082 G/A genetic polymorphism. AA genotype is associated with higher AS risk, but this genetic mutation may play a different role in different genetic and environmental background. We think the results are reliable. However, due to the limitations of this meta-analysis, more large sample and multicenter studies are required.
[Figure 1], [Figure 2], [Figure 3], [Figure 4], [Figure 5], [Figure 6], [Figure 7]
[Table 1], [Table 2]
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|||||
