Association between Osteoprotegerin gene polymorphisms and risk of coronary artery disease: A systematic review and meta-analysis

A systematic search assembling the following terms: genetic polymorphism, single nucleotide polymorphism, gene mutation, genetic variants, coronary atherosclerosis, myocardial ischemia, acute coronary syndrome, coronary artery disease, myocardial infarction, ischemic heart disease, TNFRSF11B, osteoprotegerin, was conducted in PubMed, Web of Science, Cochrane Library, Chinese National Knowledge Infrastructure (CNKI) and Chinese Wan Fang databases up to May 1 2017 to identify all potentially relevant studies. Hand-searching was carried out to determine other potential eligible studies through scanning the references cited in the retrieved articles. The full-text articles were further reviewed to determine whether they were included in the final analysis strictly based on eligibility criteria. If two reviewers disagreed, all the authors critically evaluated the studies to judge of the inclusion or exclusion of a certain study.

All the eligible articles were supposed to meet the following major inclusion criteria: i) assessment of the association between OPG gene polymorphisms and CAD; ii) case-control or cohort studies; iii) data provided by articles about allele frequency should be sufficient for calculating genotypic odds ratio (OR) with corresponding to 95% confidence interval (95% CI) in cases and controls. Moreover, only when a single nucleotide polymorphism (SNP) in the OPG gene was reported by at least two articles would it be analyzed by meta-analysis. Studies were excluded when they were i) duplicated data; ii) case report, review articles and editorial comment. The diagnosis of the CAD case was based on the WHO criteria for CAD as previously described (stenosis ≥50.0% of the diameter in at least one major coronary artery based on computer-assisted assessments) [17]. All healthy control subjects were identified according to patient history, serum biochemistry examination and electrocardiogram (ECG) test.

Data extraction was performed independently by two authors using a standardized data extraction form including following elements: 1) author’s name, year of publication; 2) patient characteristics of each group; 3) number of participants in case and control groups; 4) study type; 5) genotyping method; 6)p value of Hardy-Weinberg equilibrium (HWE) test in the control; 7) OR and 95% CI for association with CAD. Assessment of the quality of studies was performed using the Newcastle-Ottawa Scale (NOS) as previously described [18]. Briefly, two authors of this article separately evaluated the qualities based on eight items and scored the studies from 0 to 9 points. Studies with a score not less than seven points were considered to be of high quality. Discrepancy was resolved as described above.

First, the genotype frequencies of the OPG gene polymorphism among the controls of all included studies were assessed under HWE using a χ² goodness-of-fit test. Odds ratios with corresponding 95% CIs were used to estimate the strength of association between OPG gene polymorphisms and CAD. The between-study heterogeneity across all eligible comparisons was tested by the χ²-based Q statistic. Heterogeneity was considered significant when the p value was less than 0.10. When heterogeneity existed, the random effects model was performed to calculate the pooled OR of each eligible study, otherwise, the fixed effect model was used. Generally, we assessed the association between the OPG gene polymorphism and CAD under five genetic models: allele, homozygote, heterozygote, dominant and recessive models. Sensitivity analysis was conducted through omitting one study at a time to examine its influence on the overall estimate to evaluate the stability of the meta-analysis. Publication bias was also analyzed using the Egger’s linear regression test and funnel plots and publication bias was considered present when the p value was less than 0.05. All statistical analyses were done using STATA version 11.0 (STATA Corporation, College Station, TX, USA). All p values were two-tailed.

As shown in Figure.1, 36 potentially eligible records were initially identified through literature search. Thirty articles were excluded, including seven articles that were duplicated, five articles that were reviews, 12 articles that did not involve CAD, four articles that did not conform to the diagnostic criteria of CAD, one article that lacked normal controls, and one article that did not provide sufficient data for the distribution of the genotype. Finally, six articles in accordance with the inclusion criteria were included in this meta-analysis [12, 13, 14, 15, 16, 19]. To be specific, two studies involved the G209A polymorphism, three studies with T245G polymorphism, three studies with T950C polymorphism, and five studies with G1181C polymorphism.

Figure 1

The characteristics of the included studies are summarized in Table 1. Overall, four studies were conducted in Asians, and the other two studies were carried out in Caucasians. The polymerase chain reaction-restriction fragment length polymorphism (PCR-RFLP) method was used to detect the gene polymorphisms in five out of six studies. The average score of NOS was at 9. The genotype distribution of the controls in all studies was consistent with HWE.

Table 1

Main characteristics of studies included in the meta-analyses.

Meta-analysis of the G209A polymorphism was involved with two studies consisting of 215 CAD cases and 98 controls. There was no association between the G209A polymorphism and the risk of CAD when pooling all the data in the meta-analysis (AA/AG vs. GG: OR = 1.005, 95% CI = 0.564-1.792, p = 0.986) (Figure 2A) (Table 2). For the T245G polymorphism, three studies with 393 CAD cases and 410 controls were included for final meta-analysis, but the results showed no relationship between the T245G polymorphism and the risk of CAD either (TT/TG vs. GG: OR = 0.664, 95% CI = 0.247-1.785, p = 0.417) (Figure 2B) (Table 2). For the T950C polymorphism, 1105 CAD cases and 906 controls were included in the meta-analysis. A significant association was found between the T950C polymorphism and risk of CAD under the dominant model (CC/CT vs. TT: OR = 1.327, 95% CI = 1.090-1.617, p = 0.005) (Figure 2C), allele model (C vs. T: OR = 1.264, 95% CI = 1.108-1.442, p <0.001), homozygote model (CC vs. TT: OR = 1.615, 95% CI = 1.242-2.100,p <0.001), recessive model (CC vs. CT/TT: OR = 1.376, 95% CI = 1.097-1.726, p = 0.006) (Table 2). Meta-analysis of the G1181C polymorphism was involved with five studies consisting of 1137 CAD cases and 1024 controls and the results indicated that the G1181C polymorphism was significantly associated with the risk of CAD under the dominant model (CC/CG vs. GG: OR = 1.268, 95% CI = 1.064-1.511, p = 0.008) (Figure 4D). In addition, a statistically significant association also existed between the G1181C polymorphism and risk of CAD under the heterozygote model (CG vs. GG: OR = 1.243, 95% CI = 1.027-1.504, p = 0.026) (Table 2).

Figure 2

Table 2

Main results of the meta-analyses of the pooled odds ratios.

The result of the sensitivity analysis showed that the pooled ORs of the G209A, T245G, T950C and G1181C polymorphisms were not considerably affected by eliminating any individual study (Figure 3). The funnel plots were symmetrical by visual inspection (Figure 4) and Egger’s test also suggested no publication bias (p >0.05). These results confirmed that this meta-analysis was robust.

Figure 3

Figure 4