Corresponding author: Hu Ding, PhD, MD, Department of Internal Medicine Tongji Hospital, Tongji Medical College, Huazhong University of Science & Technology, 1095# Jiefang Ave. 430030 Wuhan, People's Republic of China. Tel: 86-27-8366-3280; Fax: 86-27-8366-3280; E-mail: firstname.lastname@example.org
Our aim was to evaluate the effect of SNPs in the 9p21 locus on genetic susceptibility of coronary heart disease in the East Asian population. We searched PubMed, EMBASE and CNKI for publications relating to the association between SNPs within the 9p21 locus and coronary heart disease in the East Asian population. This meta-analysis was assessed by STATA 9.2. Twenty-one studies from 15 eligible papers composing 25,945 cases and 31,777 control subjects were included in the meta-analysis. The odds ratio (OR) and 95% confidence interval (CI) for the risk allele was 1.30 (1.25–1.35) with moderate heterogeneity. No publication bias was observed in this study. Sensitivity analysis further strengthened the validity of this association. In conclusion, SNPs within the 9p21 locus were strongly associated with the risk of coronary heart disease in the East Asian population with a similar risk OR to the Caucasian population.
Coronary artery disease (CAD) is a common complex disease, which is considered to be caused by both genetic and environmental factors (Wang, 2005). Genetic factors have been defined as an important risk contributor implicated in the pathogenesis of disease development of CAD (Topol et al., 2006). Recently, GWAS have identified several SNPs on chromosome 9p21 that were significantly associated with CAD (Helgadottir et al., 2007, Mcpherson et al., 2007). In subsequent replication studies, many other research groups had found a consistent association between 9p21 variants and CAD in a European-derived population (Schunkert et al., 2008). However, association signals within the 9p21 locus were weaker in South Asians (Pakistanis) than those of European origin [OR = 1.13 (95% CI, 1.05 to 1.22) vs. OR = 1.31, (95% CI, 1.26 to 1.37); Saleheen et al., 2010], and were nonreplicable in African Americans (Mcpherson et al., 2007), indicating genetic heterogeneity in different ethnic populations. Therefore, additional genetic studies would be necessary to further define the genetic effect of 9p21 variants on CAD in different descendants. For this purpose, we systematically reviewed the association of these SNPs on chromosome 9p21 with the risk of CAD in East Asian populations. A previous meta-analysis included 8 of the 21 studies included in the present meta-analysis (Palomaki et al., 2010). Thus, this study can be viewed as an extension and confirmation of the results of the previous meta-analysis.
Materials and Methods
We searched several databases, including PubMed, EMBASE and CNKI (National Knowledge Infrastructure), from June 2007 to July 2012 for all publications on the association between 9p21 and CAD/CHD. The key words were as follows: 9p21, CAD/CHD, MI, Chinese, Japanese, Korean and East Asian. Electronic searching was also supplemented by retrieving additional studies from the references, citations and from the PubMed option “Related Articles.” The language of the reviewed articles was limited to Chinese and English.
Studies included in the current meta-analysis had to meet all the following criteria: (1) evaluation of the SNPs within the 9p21 locus and coronary heart disease risk; (2) use of a case-control design, (3) sufficient information included to permit calculation of the odds ratio; (4) SNP genotyping was done at a high-quality level and (5) inclusion of the patients was carried out according to the standard diagnosis parameters. The exclusion criteria included: study not providing enough information (genotype frequency), or insufficient information for data extraction. For duplicate publications, the study with the smaller data set was excluded.
Two authors (L.P. and R.W) independently reviewed study eligibility and extracted the information and data from each study, including: author's last name, journal and year of publication, country of origin, selection and characteristics of CAD/CHD cases and controls, ethnicity of the study population, genotype method, genotypes and numbers of cases and controls. Disagreements and uncertainties were discussed and were resolved with consensus, involving another author (H.D.) if necessary.
The χ2 test was used to evaluate whether the observed frequencies of genotypes conformed to Hardy-Weinberg Equilibrium (HWE). Odds ratios with their confidence intervals were calculated for each study. Random effects methods were used to calculate pooled ORs and the associated 95% CIs. Heterogeneity between studies was evaluated using Cochran's Q statistic (Munafo & Flint, 2004) and the inconsistency index I2 (Higgins et al., 2003). Publication bias tests were performed by using the funnel plot, in which the standard error of log (OR) of each study was plotted against its log (OR). Funnel plot asymmetry was evaluated by Egger's linear regression test. To see whether the pooled effect estimate was influenced mainly by a certain individual study, sensitivity analysis was conducted by omitting one study at a time. All statistical analyses were performed with STATA 9.2 (StataCorp, College Station, TX, USA) using external packages (metan, metabias, funnel and metainf). A P-value <0.05 was considered statistically significant (two-tailed).
Extraction Process and Characteristics of the Studies
A total of 176 articles were obtained by literature searching of the PubMed, EMBASE and CNKI databases from June 2007 to July 2012, using different combinations of key terms. As shown in Figure 1, 24 potentially appropriate articles were selected for detailed review; none were meta-analyses. Nine articles were excluded: four duplicated publications (Zhang, Zhong, Chen et al., 2010, Zhang, Zhong, Wang et al., 2010, Cheng et al., 2011, Do et al., 2011); three were not association studies (Lin et al., 2010, Wang et al., 2010, Wang, Peng et al., 2011); one related to another disease (Wang, Meng et al., 2011); one was a review article (Roberts & Stewart, 2012). A total of 15 eligible articles including 21 data sets met the selection criteria (Fig. 1). Among these studies, ten were carried out in the mainland Chinese population, one among Chinese living in Taiwan, three in the Japanese population and two in the Korean population. The characteristics of selected studies are summarized in Table 1. The distribution of genotypes in the controls was consistent with HWE.
Table 1. Main characteristics of all studies included in the meta-analysis
Multiple SNPs (rs1333049, rs1333048, rs10757278, rs10757274, rs2383206, rs944797) within the 9p21 locus were involved in this analysis, and not all researchers used the same SNPs. These SNPs on chromosome 9p21 were highly correlated with each other. For simplicity, we present pairwise linkage disequilibrium (LD) coefficients r2 between rs1333049 and other SNPs. As previously reported, the SNP rs1333049 was in strong LD with rs1333048 (r2 = 0.98), rs10757278 (r2 = 0.95), rs10757274 (r2 = 0.96), rs2383206 (r2 = 0.94), rs944797 (r2 = 0.92) (Guo et al., 2011, Wang, Xu et al., 2011, Xie et al., 2011, Takeuchi et al., 2012), respectively. Given the high LD, we reported on a single SNP per data set in follow-up analysis.
To investigate the impact of the variants within the 9p21 locus on coronary heart disease risk, we performed a meta-analysis which included 57,722 individuals (25,945 cases and 31,777 control subjects). The risk allele frequency varied from 40% to 53% in East Asians. As shown in Figure 2, all the studies showed a trend of elevated OR for the risk allele. A random effect model was performed for the meta-analysis and generated a combined allelic OR of 1.30 for risk allele (95% CI, 1.25–1.35, P < 0.001) under an additive model with moderate heterogeneity (Q = 37.05, I2 = 0.46, P = 0.012). Next, we used the HETRED module of STATA 9.2 software to trace the origin of heterogeneity and found that the study of early-onset CAD (mean age = 48) conducted by Chen et al. (2009) was a key contributor. After omitting this study, heterogeneity of the present meta-analysis was significantly reduced (Q = 27.96, I2 = 0.32, P = 0.084). No differences in ORs were found among Chinese (19,067 cases and 18,527 controls), Japanese (5588 cases and 12,250 controls) and Koreans (1290 cases and 1000 controls) (ORs, 1.33, 1.26 and 1.22, respectively; P = 0.051); or among those with a phenotype of MI (3165 cases and 5947 controls) or CAD/CHD (22,270 cases and 25,276 controls) (ORs; 1.34, 1.29, respectively; P = 0.764).
Begg's funnel plot and Egger's test were performed to evaluate the publication bias of the studies. As shown in Figure 3, the shapes of the funnel plots did not reveal any evidence of obvious asymmetry in all comparison models. Then, Egger's test was used to provide statistical evidence of funnel plot symmetry. The results still showed no evidence of publication bias (t = 0.83, P = 0.419).
To further confirm the stability of our results, we conducted a sensitivity analysis by sequentially removing individual studies and we found that individual studies did not modify the estimates much, with pooled ORs ranging from 1.24 to 1.36 (Fig. 4).
The 9p21 variant has now been established as a novel independent risk factor for cardiovascular disease outcome, including CAD (Schunkert et al., 2008), stroke (Ding et al., 2009), intracranial aneurysms (Helgadottir et al., 2008), abdominal aortic aneurysm (Bown et al., 2008, Helgadottir et al., 2008), peripheral artery disease (Cluett et al., 2009) and heart failure (Yamagishi et al., 2009). In addition, recent studies demonstrated that it influences the progression of atherosclerosis development (Ye et al., 2008), significantly improves classification for recurrent MI or cardiac death combined with traditional risk factors (Buysschaert et al., 2010) and predicts the severity of coronary heart disease (Dandona et al., 2010). In this study, we performed a meta-analysis to examine the association between SNPs within the 9p21 locus and susceptibility to coronary heart disease in the East Asian population, pooling all available qualified data from 15 genetic studies comprising 25,945 cases and 31,777 control subjects. According to our findings, the presence of a risk allele at 9p21 locus increased the risk of coronary heart disease with risk OR 1.30 similar to that in the Caucasian population. However, there are some points, which should be addressed, concerning the ethnic diversity of the 9p21 locus. First, ethnic differences may contribute to these different results, since the distributions of the 9p21 locus were different between various ethnic populations. For instance, the frequencies of risk allele rs1333049*C within the 9p21 locus differs from 48% in the CHB population (Han Chinese in Beijing), 49% in the CEU population (US residents with northern and western European Ancestry), 24% in the LWK population (Luhya in Webuye, Kenya) to 18% in the YRI population (Yoruba people of Ibadan, Nigeria), based on the Hapmap database. Second, in Africans, who have had a much longer history with lower extent of LD (Reich et al., 2001), the 9p21 risk region may contain many more alleles and haplotype blocks (Ding et al., 2009) that produce minimal or no risk for CAD (Assimes et al., 2008).
The present meta-analysis should be interpreted within the context of its limitations. First, although the SNPs used in this study had high LD, ORs were estimated from the use of more than one SNP in the same meta-analysis, and therefore might provide us with less reliable evidence to make our conclusion. Second, moderate heterogeneity may have distorted the meta-analysis. However, most of the studies demonstrated the same directionality with respect to associations between the 9p21 variants and coronary heart disease. Last, detailed information on age, gender or other environmental factors was not available for analysis; the effects of such factors could therefore not be adequately addressed in the present meta-analysis.
In summary, we demonstrate that 9p21 is associated with coronary heart disease in a meta-analysis of 57,722 East Asians suggesting that SNPs in the 9p21 locus may be regarded as novel biomarkers for coronary heart disease. The clinical practical utility of our data requires further investigation in the future.
This study was funded by the research grants from National “973” project (No. 2012CB518004) and “863” project (No. 2012AA02A510), National Nature Science Foundation of China (81100066) and Key Project of Health Ministration.