Correction
11 Dec 2013: Wu Z, Lou Y, Jin W, Liu Y, Lu L, et al. (2013) Correction: Relationship of the p22phox (CYBA) Gene Polymorphism C242T with Risk of Coronary Artery Disease: A Meta-Analysis. PLOS ONE 8(12): 10.1371/annotation/7f489b96-e43f-4c6d-86d7-0689c8a55eba. https://doi.org/10.1371/annotation/7f489b96-e43f-4c6d-86d7-0689c8a55eba View correction
Figures
Abstract
Background
Observational and experimental studies have thus far been unable to resolve whether the CYBA C242T polymorphism is associated with coronary artery disease (CAD). Therefore, we undertook a comprehensive meta-analysis to more precisely evaluate the influence of this polymorphism on CAD and potential biases.
Methods
We screened MEDLINE, Embase, CNKI, Wanfang and CBM up to January 2013 and extracted data from 22 studies with 9,279 CAD patients and 9,349 controls. A random-effects model was exploited to synthesize the inconsistent outcomes of the individual studies, while addressing between-study heterogeneity and publication bias.
Results
The CYBA C242T polymorphism conformed to Hard-Weinberg Equilibrium for all studies (P>0.05). Overall comparison of the T allele with the C allele produced a non-significant risk estimate for CAD but with striking heterogeneity (T versus C: P = 0.87, OR = 0.99, 95%CI 0.89–1.11, Pheterogeneity<0.0001, I2 = 67.8%). However, subgroup analysis by ethnicity documented that the T allele carriers had a marginal risk increase (21%) of CAD among Caucasians (recessive genetic model: P = 0.05, 95%CI 1.00–1.46, Pheterogeneity = 0.15, I2 = 29.1%). Then data were divided into study design, the significance of CAD risk increase was substantially strengthened in matched case-control studies (allele comparison: P = 0.02, OR = 1.13, 95%CI 1.02–1.26, Pheterogeneity = 0.24, I2 = 21.6%).Further meta-regression analysis identified that a large proportion of heterogeneity was explained by body mass index (BMI) (P = 0.03, OR = 1.07, 95%CI 1.01–1.15) and study design (P = 0.03, OR = 1.30, 95%CI 1.02–1.64).There was no obvious publication bias as verified by funnel plot and Egger's linear regression test (t = −0.25, P = 0.81 for allele comparison).
Citation: Wu Z, Lou Y, Jin W, Liu Y, Lu L, Chen Q, et al. (2013) Relationship of the p22phox (CYBA) Gene Polymorphism C242T with Risk of Coronary Artery Disease: A Meta-Analysis. PLoS ONE 8(9): e70885. https://doi.org/10.1371/journal.pone.0070885
Editor: Andreas Zirlik, University Heart Center Freiburg, Germany
Received: February 6, 2013; Accepted: June 24, 2013; Published: September 5, 2013
Copyright: © 2013 Wu et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This work was supported by the science and technology fund of Shanghai Jiao Tong University School of Medicine (11XJ21001) and the National Natural Science Foundation of China (81201839). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Coronary artery disease (CAD) and myocardial infarction (MI) are principal public health contributors to worldwide death and disability, and as such are primary causes of economic burden [1]. About 1 of every 6 deaths was attributed to CAD in the United States in 2008 and as rough estimate 785,000 Americans will have a new coronary attack and 470,000 will have a recurrent attack annually [2]. The multiple potential pathologic pathways of CAD have been widely investigated, ranging from cholesterol accumulation to plaque rupture with thrombus formation [3]. Among these pathways, oxidative stress is thought to play a pivotal role in the pathophysiology of atherosclerosis, which is manifested as increased availability of reactive oxygen species (ROS) and reactive nitrogen species because of an imbalanced redox state. ROS has been implicated in a variety of biological functions, including induction of gene expression and promotion of endothelial cell and smooth muscle cell (SMC) proliferation. Imbalance of ROS production is involved in oxidation of low density lipoprotein, which consequently promotes the progression of atherosclerosis [4], [5].
NAD(P)H oxidase is the predominant cellular source of ROS in the context of atherosclerosis [6]. The phagocyte-specific cytochrome b, as an essential component of NAD(P)H oxidase, plays an important role in regulating NAD(P)H activity and ROS production. Phagocyte-specific cytochrome b is a membrane-associated heterodimer, consisting of 2 subunits (gp91phox and p22phox [CYBA]) [7]. P22phox is essential for vascular cell production of superoxide anion. Human p22phox is encoded by the CYBA gene which is located on 16q24. One of the most common polymorphisms of CYBA gene is C242T(rs4673), resulting in a non-conservative substitution of histidine for tyrosine at codon 72 and an alteration of NAD(P)H activity by disrupting the heme binding site [8].The C242T polymorphism has been demonstrated to be related to multiple inflammatory diseases [7] and metabolic disorders [9], [10]. Recent epidemiological studies suggested that this polymorphism could be associated with CAD [11], [12], but this could not be confirmed by other studies [13], [14]. The current discordance in reported studies may result from real racial diversity, restrictive sample size and unadjusted environmental confounders. Large meta-analysis makes it possible to achieve a more precise evaluation of this issue. Although a previous meta-analysis of the C242T polymorphism indicated a significant protective effect among Asians, they had a relative small sample size and merely concerned the ethnic difference [15]. In addition, some studies they included were departure from Hardy-Weinberg equilibrium (HWE) [8], [16]. None of these studies had comprehensively compared the effect of the C242T polymorphism on CAD risk in view of study differences. To maximize statistical power and robustly assess the relationship, we have incorporated the published studies to date and meta-analyzed the effect of the C242T polymorphism on CAD risk, while addressing heterogeneity and publication bias. We aimed to systematically estimate the CAD risk of the C242T polymorphism taking different study characteristics into account and go deeply into the origin of the heterogeneity.
Materials and Methods
Search strategy
We collected data by searching from PubMed/MEDLINE, Embase, CNKI (China Nation Knowledge Infrastructure Platform), Wanfang and CBM (China Biological Medicine Database) electronic databases up to January 2013. The search strategy followed the guidelines of the 2009 preferred reporting items for systematic reviews and meta-analyses (PRISMA) statement (Checklist S1) [17]. A combination of the following key words was used in the computer-based searches: ‘C242T’ or ‘rs4673’, ‘nicotinamide adenine dinucleotide phosphate oxidase’ or ‘NAD (P) H oxidas’, ‘cytochrome b’ or ‘CYBA’, ‘ph22phox’ and ‘coronary’ or ‘ischemic heart disease’ or ‘myocardial infarction’ or ‘atherosclerosis ’. We scrutinized the full texts of potentially appropriate articles and identified those containing useful data on the topic of interest for inclusion in the meta-analysis. Additional studies were identified by hand-searching and employing the Pubmed related search function. The bibliographies of relevant studies, reviews and previous meta-analyses were screened for additional useful studies which were on the same topic but not found by the initial search. If the allele or genotype frequency was not adequately reported, the original authors were contacted to request the raw data.
Selection criteria
All studies related to the association of the C242T polymorphism with CAD risk were potentially involved. The following criteria were adopted to the selection: (1) Articles published in English or Chinese journals or their supplements; (2) Studies providing adequate information on the C242T polymorphism by case-control status (studies without controls were excluded); (3) Published articles of human genetics (full texts or abstracts) without ethnicity restriction; (4) If articles involved more than one geographic or racial group, each group was considered as a separate study; (5) If multiple studies stemmed from the same population, only the largest scale study was included to avoid overlapping data; (6) The genotype frequency amongst control must conform to HWE.
Standard definitions were used to diagnosis disease outcomes [18].CAD was defined as previous MI, angina pectoris, or the presence of at least one or more major coronary epicardial coronary vessel with ≥50% luminal obstruction [19]. Acute coronary syndrome (ACS) was defined by unstable angina pectoris and fatal or non-fatal MI [20] and MI defined by a positive cardiac enzymes test in the setting of clinical symptoms and typical electrocardiographic changes [21].
Data Extraction
We undertook a standard data-collection protocol following the inclusion criteria described above. Two authors (Z.W. and Y. Lou.) independently extracted the relevant characteristics from each selected study and entered the data into separate databases in duplicate. Moreover, the results of data extraction were compared and any encountered divergences were resolved at a consensus meeting. The following information was gathered on the C242T polymorphism on the basis of different cohort: First author's name, publication year, ethnicity, country or area, study design, population source, disease outcomes, matching variables, clinical characteristics for case-patients and controls (such as age, body mass index [BMI], circulating lipid profiles and fibrinogen levels, the proportion of male, hypertension [HTN] , diabetes mellitus [DM] , smoking status and alcohol intake), the distribution of the C242T genotype both in patients and controls and conformity to genotype frequencies with HWE. Continuous variables manifested themselves as mean ± standard deviation (SD) or median (5th and 95th percentiles).
Quality Assessment
The Newcastle-Ottawa Scale (NOS) [22] was used to assess the methodological quality of all eligible studies. The NOS provides an easy and convenient tool for quality assessment of non-randomized studies based on three broad perspectives: Selection, Comparability, and Exposure for case-control studies. In our meta-analysis, age, gender, BMI and smoking were considered as important confounder factors. A study can be awarded a maximum of one star for each numbered item with the Selection and Exposure categories, and a maximum of two stars for Comparability. Studies can be awarded a maximum score of 9 stars with scores of 5 stars or more regarded to be of medium to high study quality. Quality assessment of each individual study was performed by three authors separately (Z.W., Y. Lou. and Q.C.). Authors W.J. and Y. Liu. were consulted for any encountered disagreement.
Statistical analysis
The synthetic studies' odd ratios (ORs) corresponding to 95% confidence intervals (CIs) for the relationship between the C242T polymorphism and CAD risk were calculated under different inheritance models, containing allele comparison (T versus C), dominant genetic model (CT+TT versus CC), recessive genetic model (TT versus CT+CC) and homozygote comparison (TT versus CC). We appraised the individual effect size together and modulated the study weights based on in-study variance by exploiting the DerSimonian & Laired method in a random-effects model. The possibility of heterogeneity was also assessed by using the Mantel-Haenszel model [23]. The between-study heterogeneity was estimated using Cochran's chi-square based Q statistic test [24] and P<0.1 was considered significantly heterogeneous among studies. The degree of consistency was estimated via index I2 statistic, which determines the percentage of the between-study variability caused by heterogeneity rather than chance. Large values of I2 strongly implied the presence of between-study heterogeneity [25], [26].The significance of the pooled OR was evaluated by the Z test, which was considered to be significant for P<0.05. Additionally, we carried out a meta-regression with restricted maximum likelihood estimation to investigate the extent to which covariates explained genetic heterogeneity among the individual ORs. To further assess the origins of inter-study heterogeneity, we examined pre-specified groupings of study characteristics with homogeneous effects, such as ethnicity (Caucasian, Asian and others), study design (matched and not mentioned), population source (hospital-based [H-B] and population-based [P-B]) and endpoints (CAD, ACS and MI).
We undertook a sensitivity analysis by successive one-study removal approach to identify the most influential studies, likely introducing bias to the overall estimation [27]. The visual funnel plot and Egger's test were used to appraise the probability of publication bias. The standard error of log (OR) of CAD risk under the allele comparison was plotted against its OR for each study. The publication bias could be visual by an asymmetric plot and be confirmed by Egger's linear regression test [28]. Egger's test was considered significant for P<0.05. A cumulative meta-analysis might indicate whether initial published research had an impact on subsequent publications and the evolution of the pooled estimates with ascending date of publication.
Consistency of the C242T polymorphism to HWE was examined via the chi-square test or Fisher's exact test based on a Web program (http://ihg2.helmholtz-muenchen.de/cgi-bin/hw/hwa1.pl). Data processing and statistical analyses were performed using Review Manager software release 5.0 (Oxford, England) and Stata 11.0 (Stata Corporation, College Station, Texas, USA). All P values were 2-sided.
Results
Flow of included studies
The flow diagram schematizing the selection process of included studies is shown in Figure 1. A total of 150 relevant publications were retrieved in accordance with our search strategy and inclusion criteria. After applying selection criteria, 124 studies were removed. The data from 26 studies depicting the association between the C242T polymorphism and CAD risk were pooled into the provisional data set. Among these studies, 2 data sets conducted in the Chinese [29] and Japanese population [10] were overlapped by other studies with larger cohort [11], [30]. Moreover, the genotyping frequency in the control population of Mata-Balaguer et al. [16] and Inoue et al. [8] did not conform to HWE (PHWE = 0.034 & 0.029); these two studies were precluded. Because the study by Saha et al. [31] contained data from two different racial descents (Chinese & Indian), we treated them as two independent studies when estimating the individual effect sizes. The Indian population was classified into the “others” group as its lineage was miscellaneous and cannot simply be grouped as Asian or Caucasian [32]–[34].Because the study population used in Li et al. [14] was of mixed background (82.9% of Caucasian), this study together with 2 Indian studies [31], [35], were collapsed into the “others” group.
In total, 21 publications involving 22 studies (20 articles written in English [11], [13], [14], [30], [31], [35]–[49] and 1 in Chinese) with sufficient information fulfilled our search strategy. Of the 22 qualifying studies, 13 used a Caucasian population [13], [36]–[39], [41]–[48], 6 used an Asian population [11], [30], [31], [40], [49] and 3 performed in others [14], [31], [35]. All studies were published between 1999 and 2012 and utilized a retrospective case-control design .P-B controls were adopted in 6 studies [31], [37], [41], [45], [49] and H-B in 16 studies [11], [13], [14], [30], [35], [36], [38]–[40], [42]–[44], [46]–[48]. CAD patients and controls were matched for age and gender in 11 studies [13], [31], [39], [41]–[43], [45]–[47], while the other 11 failed to mention this in their study design [11], [14], [30], [35]–[38], [40], [44], [48], [49]. CAD was employed as the primary outcome in 16 of the studies [13], [14], [30], [31], [35]–[38], [43]–[46], [48], [49], while MI and ACS were taken as endpoints in 4 [11], [39], [40], [47] and 2 studies [41], [42], respectively.
Quantitative data synthesis
Our meta-analysis encompassed 22 studies with 9,279 CAD patients and 9,349 controls. The selected characteristics of these eligible studies were shown in Table 1. The allele and genotype frequencies of the C242T polymorphism of each individual study which corresponded to HWE(PHWE>0.05) are also listed in Table 2. The combined overall frequency of the T allele was 23.8% in cases and 18.0% in controls. The frequency of the T allele among Caucasians (34.3% cases, 33.5% controls) was markedly higher than that among Asians (8.7% cases, 9.5% controls). The T allele proportion in the “other” group (35.9% cases, 64.1% controls) was similar to that in the Caucasian group. All the 22 studies were assessed according to the NOS scale and the results of the quality assessment are summarized in Table 3. Most studies (82%) scored 5 stars or more, suggesting a moderate to good quality.
Furthermore, we assessed the association of the C242T polymorphism with CAD risk under different genetic models. The summary results of the meta-analysis were shown in Table 4. There was no predominant association between the C242T polymorphism and CAD produced by comparing the T allele with C allele overall (allele comparison: P = 0.87, OR = 0.99, 95%CI 0.89–1.11; dominant model: P = 0.58, OR = 0.96, 95%CI 0.84–1.10; recessive model: P = 0.15, OR = 1.14, 95%CI 0.95–1.36; homozygote comparison: P = 0.20, OR = 1.14, 95%CI 0.93–1.40) (Figure 2). We identified a large between-study heterogeneity across all genotypic models (allele comparison: I2 = 67.8%, Pheterogeneity<0.0001; dominant model: I2 = 65.7%, Pheterogeneity<0.0001; homozygote comparison: I2 = 35.7%, Pheterogeneity = 0.05), with the exception of the recessive model (I2 = 27.9%, Pheterogeneity = 0.11) . Because every subject contributes two observations under the allele comparison and we cannot a priori hypothesis that these two observations are independent, we computed log ORs with standard errors in each study and then used them in our meta-analysis. As a result, the risk estimates did not change if the clustering of the alleles in each subject was taken into account.
‘Events’ indicates the total number of T allele. ‘Total’ indicates the total number of T allele plus C allele.
Sensitivity analysis
The striking heterogeneity identified across studies could influence the total outcome of the meta-analysis. Therefore, we performed a sensitivity analysis to look for the study or studies making the greatest contribution to heterogeneity. No single study or group of studies could be confirmed as disproportionately influencing the heterogeneity and ORs in total and subgroup analysis.
Cumulative analysis and publication bias
We did not find any clear evidence to suggest that the first published study affected and brought about the ensuing replication via cumulative meta-analysis (data not shown). In addition, funnel plot analysis suggested that the overall results were not asymmetric and Egger's regression asymmetry test further verified an absence of publication bias (t = −0.25, P = 0.81 for allele comparison) in the overall estimates (Figure 3).
Subgroup analysis
In view of a conspicuous heterogeneity present in total analysis, we undertook a panel of subgroup analyses on ethnicity, population source, study design and endpoint type. The initial aim was to evaluate the possible variability of overall estimates due to ethnicity and data was classified in accordance with the 3 racial descent groups: Caucasian (13 studies recruited 5,067 cases and 2,843 controls), Asian (6 studies recruited 3,777 cases and 6,128 controls), and others (3 study recruited 435 cases and 378 controls). The ORs of the C242T polymorphism appeared to decrease in the Asian group (allele comparison: P = 0.24, OR = 0.85, 95%CI 0.64–1.12, Pheterogeneity = 0.001, I2 = 75.6%; dominant model: P = 0.21, OR = 0.82, 95%CI 0.61–1.12, Pheterogeneity<0.0001, I2 = 77.4%) and in others group (allele comparison: P = 0.81, OR = 0.94, 95%CI 0.56–1.59, Pheterogeneity = 0.001, I2 = 84.8%; dominant model: P = 0.93, OR = 0.97, 95%CI 0.51–1.84, Pheterogeneity = 0.01, I2 = 80.1%) without significance. Interestingly, the risk estimates of the C242T polymorphism with CAD differed between Caucasians and Asians, with an increased OR of CAD among the former (allele comparison: P = 0.19, OR = 1.07, 95%CI 0.97–1.19, Pheterogeneity = 0.06, I2 = 41%; dominant model: P = 0.58, OR = 1.04, 95%CI 0.90–1.20, Pheterogeneity = 0.05, I2 = 43.5%). In particularly, we observed a marginal increased risk of CAD under the recessive genetic model (P = 0.05, OR = 1.21, 95%CI 1.00–1.46, Pheterogeneity = 0.15, I2 = 29.1%) and in homozygote comparison (P = 0.06, OR = 1.22, 95%CI 0.99–1.49, Pheterogeneity = 0.16, I2 = 28.3%) among Caucasians without distinct heterogeneity (Figure 4).
The TT homozygote shows a borderline increased risk of CAD under the recessive model. ‘Events’ indicates the total number of TT genotype. ‘Total’ indicates the total number of TT genotype plus TT+ CT genotype.
To address study design, stratification by the matched information for age and gender between cases and controls was performed, as a result, the T allele carriers showed a remarkable increased risk of CAD under most genetic models (allele comparison: P = 0.02, OR = 1.13, 95%CI 1.02–1.26, Pheterogeneity = 0.24, I2 = 21.6%; recessive model: P = 0.02, OR = 1.32, 95%CI 1.06–1.66, Pheterogeneity = 0.26, I2 = 19.1%, homozygote comparison: P = 0.03, OR = 1.31, 95%CI 1.03–1.66, Pheterogeneity = 0.29, I2 = 16%) but not under the dominant model (P = 0.22, OR = 1.09, 95%CI 0.95–1.26, Pheterogeneity = 0.23, I2 = 22.9%) in matched studies (Figure 5). Although the ORs of the C242T polymorphism seem to be reversed in studies which did not specify whether the matched case-control study design was used (allele comparison: P = 0.11, OR = 0.87, 95%CI 0.73–1.03, Pheterogeneity<0.0001, I2 = 75%; dominant model: P = 0.15, OR = 0.85, 95%CI 0.69–1.06, Pheterogeneity<0.0001, I2 = 75.5%; recessive model: P = 0.75, OR = 0.97, 95%CI 0.78–1.20, Pheterogeneity = 0.37, I2 = 8.1%, homozygote comparison: P = 0.89, OR = 0.98, 95%CI 0.71–1.34, Pheterogeneity = 0.07, I2 = 41.8%), this was not statistically significant. In further subgroup analysis, a lack of dramatic association of the C242T polymorphism with CAD risk was returned when using any genetic model in which data were categorized according to population sources and disease outcomes (Table 4).
The C242T polymorphism shows a significant increased risk of CAD under the allele comparison (T versus C). ‘Events’ indicates the total number of T allele. ‘Total’ indicates the total number of T allele plus C allele.
Meta-regression analysis
We performed a univariate meta-regression analysis for pre-defined underlying sources of heterogeneity, including ethnicity, population source, case definition, study-design, the baseline characteristics of total population (such as age, gender, BMI, the proportion of HTN, DM and smoking status). The adjusted R2 statistic indicates the proportion of between-study variance explained by the covariates. Accordingly, a large proportion of heterogeneity was significantly attributed to BMI (P = 0.03) and study design (P = 0.03) respectively, with CAD risk of the T allele being increasing in high BMI population (OR = 1.07, 95%CI 1.01–1.15) compared to low BMI population and in matched studies (OR = 1.30, 95%CI 1.02–1.64) compared to studies which did not mention whether the matched case-control study design was used (Table 5).
Furthermore, we compared the descriptive factors between Caucasian and non-Caucasian studies to investigate underlying confounders which may influence the results (Table 6).We found the distribution in study design and BMI was distinctly different between the two groups and could be a likely source of potential confounders. The proportion of Caucasian studies using matched study design (61.5%) was almost double that of non-Caucasian studies (33.3%). The out point of the BMI (25.8 kg/m2) was chosen as the median. Interestingly, most Caucasian studies (81.8%) had BMI≥25.8 kg/m2, while none of non-Caucasian studies had BMI≥25.8 kg/m2.
We took ethnicity and BMI into account by separately adding them into the multivariable regression. There was a decrease in the OR increase of population source, study design and BMI and a slight increase in OR increase of disease endpoint after adjustment for Caucasian/non-Caucasian, which was similar to that after adjustment for BMI. There were no substantial changes observed in the effect estimates of age, gender, HTN, DM and smoking status after adjusting for ethnicity and BMI. Nevertheless, the adjusted effects did not attain significance. The frequent reduction by adjusting may imply that confounding of the potential factors by ethnicity and BMI was likely to overestimate the influence of the study factors on the CAD risk estimates [50]. We also found that the adjusted R2 value significantly increased after adjustment for BMI.
Discussion
The correlation of the C242T polymorphism with CAD risk has yet to be clarified. Reckoning the CAD risk of the minor allele in independent case-control studies generated a conspicuous divergence, which was attributed to such causes as racial diversity, research strategy and limited sample size [51]. To address these issues, we commenced a meta-analysis and provided a comprehensive assessment of the publically available studies [52]. The current meta-analysis, including a total of 9,279 cases and 9,349 controls, is one of the largest systematic reviews addressing the association between the C242T polymorphism and CAD risk. Consequently, the overall comparison of T allele with C allele yielded a non-significant risk for CAD under different genetic models incorporating distinguishable heterogeneity.
Given that distinct heterogeneity was present and could not be completely eliminated [53], we conducted a subgroup analysis to investigate the potential effect of the C242T polymorphism on different groups with homogenous characteristics. As a result, we verified that ethnicity is an underlying source of between-study heterogeneity. In fact, the risk estimates of the C242T polymorphism on CAD among Caucasians were dramatically different to that among Asians. We observed a borderline increased risk of CAD under the recessive model and the homozygote comparison in the Caucasian group, while the opposite trend was observed among Asians. This discrepancy could be attributed to the high variation in C242T allele frequency across different ethnic populations. The T allele frequency among Caucasians was approximately fourfold higher than that among Asians. In addition, we speculate that the pleiotropic effect of the C242T polymorphism due to various genetic ancestral backgrounds could be another explanation. Our finding are inconsistent with a previous meta-analysis involving 6,273 cases and 5,045 controls [15]. The previous meta-analysis by Fang et al. [15] showed that the T allele carriers had a significant protective effect among Asians but not among Caucasians. This discrepancy is probably due to their relative small sample size and inadequate statistical power. Another explanation could be that the Fang et al. [15] meta-analysis included the studies by Mata-Balaguer et al. [16] and Inoue et al. [8], which significantly violated HWE. Deviations of HWE in control group are likely owing to genotyping errors, population classification and selection bias in the choice of control population and environmental exposures, resulting in reduced precision and raising doubt about the validity in estimation of effect sizes [54]. Hence, we excluded these two studies in order to avoid biasing the overall estimates in our meta-analysis. It is essential to set up an optimized databank of the C242T polymorphism concerning CAD in defined racial groupings to ascertain the reliability of our results.
As well as the obvious association to ethnicity, other environment exposures have confounding potential for the interpretation of between-study heterogeneity. When taking multiple study-level covariates from clinical characteristics into account, we identified study design and BMI as exhibiting significance and positivity in univariate meta-regression analysis, and as such should be cautiously deciphered. Moreover, we found that there was a larger proportion of the Caucasian studies using matched study design compared to non-Caucasian studies. The studies among Caucasians were also more often performed in high BMI populations than those among non-Caucasians. Then accounting for the effects of ethnicity and BMI in the multivariable meta-regression, study design was no longer detected as a significant source of heterogeneity. Therefore, study design and BMI are likely confounders. The estimated advantage of Caucasians over non-Caucasians may be partially accounted for by the effect of study design and BMI and therefore, the effect of ethnicity on the variation cannot be regarded as an independent factor. Although meta-regression suggests an ecological correlation rather than a causal inference, our result implies underlying confounding of the environmental factors with genotype. Large amounts of observational studies have verified that increased BMI contributes causally to CAD [55]–[57]. Considering a relative wide range of confidence intervals in our results, further carefully designed genetic association studies involving the interaction of BMI are needed to more accurately estimate the risk evaluation of CAD.
Although the relatively large sample size and conformation to HWE strengthen our meta-analysis [58], there are some methodology limitations [59]. First, large positive results are more likely to be accepted for publication than small negative results. Therefore, it was likely that the unpublished studies and “grey” literature (articles in languages other than English and Chinese) were not included. Although the Egger's test and funnel plots revealed that an absence of publication bias in our meta-analysis, this could not be ruled out completely. Second, some H-B studies recruited controls from an ill-defined background, which could not represent the true exposure experience of the source population and probably resulted in a biased estimate of the ORs. Third, our meta-analysis was specific to a single polymorphism on the CYBA gene and did not account for possible synergistic effects of other candidate genes or polymorphisms. Genetic susceptibility to CAD is most likely based on the interaction of polygenic and environmental factors [60], [61] and optimized models concerning these aspects are required.
Experimental studies have suggested the C242T polymorphism affects the heme-binding site and alters electron flow and superoxide production [62], which results in decreased basal and NAD(P)H-stimulated ROS production. Therefore, reduced ROS levels in the atherosclerotic vessels might be a sensitive signal of declined SMC survival. SMC-depleted plaques tend to be more vulnerable which is characterized by erosion or disruption of their caps [63], [64].Our results complement the previous evidences suggesting that a decrease in NAD(P)H activity, and thus decreased ROS production, could also accelerate the progression of atherosclerosis. Further confirmation of our results is needed before it is possibly to link p22phox to the development of atherosclerosis.
Taken together, our meta-analysis, comprising 18,628 participants, implied that the T allele might carry an increased risk of CAD. However, it is conceivable that the C242T polymorphism has a heterogeneous effect on CAD across different ethnicities, being moderate among Caucasians but lacking significance among Asians. Our meta-analysis also highlights that it is necessary to consider potential gene-gene and gene-environment interactions when trying to interpret and combine observed data. For now,it remains unclear whether the T allele carriers have positive or negative effects on CAD and further population surveys and functional researches will be required to elucidate the true relationship.
Supporting Information
Checklist S1.
The PRISMA checklist for this meta-analysis.
https://doi.org/10.1371/journal.pone.0070885.s001
(DOC)
Author Contributions
Conceived and designed the experiments: ZW Y. Lou YX GL . Performed the experiments: ZW Y. Lou WJ Y. Liu . Analyzed the data: ZW Y. Lou WJ Y. Liu LL QC YX GL. Contributed reagents/materials/analysis tools: ZW Y. Lou WJ Y. Liu LL QC YX GL. Wrote the paper: ZW Y. Lou.
References
- 1. Terkelsen CJ, Vach W (2010) Using biomarkers to prioritise patients with stable angina pectoris for bypass surgery. BMJ 340: b5414.
- 2. Roger VL, Go AS, Lloyd-Jones DM, Benjamin EJ, Berry JD, et al. (2012) Executive summary: heart disease and stroke statistics–2012 update: a report from the American Heart Association. Circulation 125: 188–197.
- 3. Topol EJ, Smith J, Plow EF, Wang QK (2006) Genetic susceptibility to myocardial infarction and coronary artery disease. Hum Mol Genet 15 Spec No 2: R117–123.
- 4. Griendling KK, FitzGerald GA (2003) Oxidative stress and cardiovascular injury: Part I: basic mechanisms and in vivo monitoring of ROS. Circulation 108: 1912–1916.
- 5. Griendling KK, FitzGerald GA (2003) Oxidative stress and cardiovascular injury: Part II: animal and human studies. Circulation 108: 2034–2040.
- 6. Griendling KK, Minieri CA, Ollerenshaw JD, Alexander RW (1994) Angiotensin II stimulates NADH and NADPH oxidase activity in cultured vascular smooth muscle cells. Circ Res 74: 1141–1148.
- 7. Dinauer MC, Pierce EA, Bruns GA, Curnutte JT, Orkin SH (1990) Human neutrophil cytochrome b light chain (p22-phox). Gene structure, chromosomal location, and mutations in cytochrome-negative autosomal recessive chronic granulomatous disease. J Clin Invest 86: 1729–1737.
- 8. Inoue N, Kawashima S, Kanazawa K, Yamada S, Akita H, et al. (1998) Polymorphism of the NADH/NADPH oxidase p22 phox gene in patients with coronary artery disease. Circulation 97: 135–137.
- 9. Matsunaga-Irie S, Maruyama T, Yamamoto Y, Motohashi Y, Hirose H, et al. (2004) Relation between development of nephropathy and the p22phox C242T and receptor for advanced glycation end product G1704T gene polymorphisms in type 2 diabetic patients. Diabetes Care 27: 303–307.
- 10. Shimokata K, Yamada Y, Kondo T, Ichihara S, Izawa H, et al. (2004) Association of gene polymorphisms with coronary artery disease in individuals with or without nonfamilial hypercholesterolemia. Atherosclerosis 172: 167–173.
- 11. Yamada Y, Izawa H, Ichihara S, Takatsu F, Ishihara H, et al. (2002) Prediction of the risk of myocardial infarction from polymorphisms in candidate genes. N Engl J Med 347: 1916–1923.
- 12. Cahilly C, Ballantyne CM, Lim DS, Gotto A, Marian AJ (2000) A variant of p22(phox), involved in generation of reactive oxygen species in the vessel wall, is associated with progression of coronary atherosclerosis. Circ Res 86: 391–395.
- 13. Nikitin AG, Chistiakov DA, Minushkina LO, Zateyshchikov DA, Nosikov VV (2010) Association of the CYBA, PPARGC1A, PPARG3, and PPARD gene variants with coronary artery disease and metabolic risk factors of coronary atherosclerosis in a Russian population. Heart Vessels 25: 229–236.
- 14. Li A, Prasad A, Mincemoyer R, Satorius C, Epstein N, et al. (1999) Relationship of the C242T p22phox gene polymorphism to angiographic coronary artery disease and endothelial function. Am J Med Genet 86: 57–61.
- 15. Fang S, Wang L, Jia C (2010) Association of p22phox gene C242T polymorphism with coronary artery disease: a meta-analysis. Thromb Res 125: e197–201.
- 16. Mata-Balaguer T, de la Herran R, Ruiz-Rejon C, Ruiz-Rejon M, Garrido-Ramos MA, et al. (2004) Angiotensin-converting enzyme and p22(phox) polymorphisms and the risk of coronary heart disease in a low-risk Spanish population. Int J Cardiol 95: 145–151.
- 17. Moher D, Liberati A, Tetzlaff J, Altman DG (2009) Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. PLoS Med 6: e1000097.
- 18. Vach W, Gerke O, Hoilund-Carlsen PF (2012) Three principles to define the success of a diagnostic study could be identified. J Clin Epidemiol 65: 293–300.
- 19. Wright RS, Anderson JL, Adams CD, Bridges CR, Casey DE Jr, et al. (2011) 2011 ACCF/AHA focused update incorporated into the ACC/AHA 2007 Guidelines for the Management of Patients with Unstable Angina/Non-ST-Elevation Myocardial Infarction: a report of the American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines developed in collaboration with the American Academy of Family Physicians, Society for Cardiovascular Angiography and Interventions, and the Society of Thoracic Surgeons. J Am Coll Cardiol 57: e215–367.
- 20. Joensen AM, Jensen MK, Overvad K, Dethlefsen C, Schmidt E, et al. (2009) Predictive values of acute coronary syndrome discharge diagnoses differed in the Danish National Patient Registry. J Clin Epidemiol 62: 188–194.
- 21. Anderson JL, Adams CD, Antman EM, Bridges CR, Califf RM, et al. (2011) 2011 ACCF/AHA Focused Update Incorporated Into the ACC/AHA 2007 Guidelines for the Management of Patients With Unstable Angina/Non-ST-Elevation Myocardial Infarction: a report of the American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines. Circulation 123: e426–579.
- 22.
Wells GA, Shea B, O'Connell D, Peterson J, Welch V, et al. The Newcastle-Ottawa Scale (NOS) for assessing the quality of nonrandomised studies in meta-analyses. http://wwwohrica/programs/clinical_epidemiology/oxfordasp Accessed 13 April 2013.
- 23. Cohn LD, Becker BJ (2003) How meta-analysis increases statistical power. Psychol Methods 8: 243–253.
- 24. Lau J, Ioannidis JP, Schmid CH (1997) Quantitative synthesis in systematic reviews. Ann Intern Med 127: 820–826.
- 25. Higgins JP, Thompson SG (2002) Quantifying heterogeneity in a meta-analysis. Stat Med 21: 1539–1558.
- 26. Higgins JP, Thompson SG, Deeks JJ, Altman DG (2003) Measuring inconsistency in meta-analyses. BMJ 327: 557–560.
- 27. Biros E, Norman PE, Jones GT, van Rij AM, Yu G, et al. (2011) Meta-analysis of the association between single nucleotide polymorphisms in TGF-beta receptor genes and abdominal aortic aneurysm. Atherosclerosis 219: 218–223.
- 28. Egger M, Davey Smith G, Schneider M, Minder C (1997) Bias in meta-analysis detected by a simple, graphical test. BMJ 315: 629–634.
- 29. He M, Jiang C, Zheng H, Wang F, Yang M, et al. (2004) [Study on the relationship between the polymorphism of P22phox C242T, vitamin E and coronary heart disease]. Wei Sheng Yan Jiu 33: 443–446.
- 30. He MA, Cheng LX, Jiang CZ, Zeng HS, Wang J, et al. (2007) Associations of polymorphism of P22(phox) C242T, plasma levels of vitamin E, and smoking with coronary heart disease in China. Am Heart J 153: 640 e641–646.
- 31. Saha N, Sanghera DK, Kamboh MI (1999) The p22 phox polymorphism C242T is not associated with CHD risk in Asian Indians and Chinese. Eur J Clin Invest 29: 999–1002.
- 32. Reich D, Thangaraj K, Patterson N, Price AL, Singh L (2009) Reconstructing Indian population history. Nature 461: 489–494.
- 33. Quintana-Murci L, Chaix R, Wells RS, Behar DM, Sayar H, et al. (2004) Where west meets east: the complex mtDNA landscape of the southwest and Central Asian corridor. Am J Hum Genet 74: 827–845.
- 34. Palanichamy MG, Sun C, Agrawal S, Bandelt HJ, Kong QP, et al. (2004) Phylogeny of mitochondrial DNA macrohaplogroup N in India, based on complete sequencing: implications for the peopling of South Asia. Am J Hum Genet 75: 966–978.
- 35. Narne P, Ponnaluri KC, Singh S, Siraj M, Ishaq M (2012) Relationship between NADPH oxidase p22phox C242T, PARP-1 Val762Ala polymorphisms, angiographically verified coronary artery disease and myocardial infarction in South Indian patients with type 2 diabetes mellitus. Thromb Res 130: e259–265.
- 36. Cai H, Duarte N, Wilcken DE, Wang XL (1999) NADH/NADPH oxidase p22 phox C242T polymorphism and coronary artery disease in the Australian population. Eur J Clin Invest 29: 744–748.
- 37. Fan M, Kahonen M, Rontu R, Lehtinen R, Viik J, et al. (2006) The p22phox C242T gene polymorphism is associated with a reduced risk of angiographically verified coronary artery disease in a high-risk Finnish Caucasian population. The Finnish Cardiovascular Study. Am Heart J 152: 538–542.
- 38. Gardemann A, Mages P, Katz N, Tillmanns H, Haberbosch W (1999) The p22 phox A640G gene polymorphism but not the C242T gene variation is associated with coronary heart disease in younger individuals. Atherosclerosis 145: 315–323.
- 39. Goliasch G, Wiesbauer F, Grafl A, Ponweiser E, Blessberger H, et al. (2011) The effect of p22-PHOX (CYBA) polymorphisms on premature coronary artery disease (</ = 40 years of age). Thromb Haemost 105: 529–534.
- 40. Katakami N, Kaneto H, Matsuoka TA, Takahara M, Imamura K, et al. (2010) Accumulation of gene polymorphisms related to oxidative stress is associated with myocardial infarction in Japanese type 2 diabetic patients. Atherosclerosis 212: 534–538.
- 41. Macias-Reyes A, Rodriguez-Esparragon F, Caballero-Hidalgo A, Hernandez-Trujillo Y, Medina A, et al. (2008) Insight into the role of CYBA A640G and C242T gene variants and coronary heart disease risk. A case-control study. Free Radic Res 42: 82–92.
- 42. Morgan TM, Krumholz HM, Lifton RP, Spertus JA (2007) Nonvalidation of reported genetic risk factors for acute coronary syndrome in a large-scale replication study. Jama 297: 1551–1561.
- 43. Najafi M, Alipoor B, Shabani M, Amirfarhangi A, Ghasemi H (2012) Association between rs4673 (C/T) and rs13306294 (A/G) haplotypes of NAD(P)H oxidase p22phox gene and severity of stenosis in coronary arteries. Gene 499: 213–217.
- 44. Nasti S, Spallarossa P, Altieri P, Garibaldi S, Fabbi P, et al. (2006) C242T polymorphism in CYBA gene (p22phox) and risk of coronary artery disease in a population of Caucasian Italians. Dis Markers 22: 167–173.
- 45. Niemiec P, Zak I, Wita K (2007) The 242T variant of the CYBA gene polymorphism increases the risk of coronary artery disease associated with cigarette smoking and hypercholesterolemia. Coron Artery Dis 18: 339–346.
- 46. Stanger O, Renner W, Khoschsorur G, Rigler B, Wascher TC (2001) NADH/NADPH oxidase p22 phox C242T polymorphism and lipid peroxidation in coronary artery disease. Clin Physiol 21: 718–722.
- 47. Vasiliadou C, Tousoulis D, Antoniades C, Marinou K, Tentolouris C, et al. (2007) Genetic Polymorphism C242T On The p22-phox Subunit Of NADPH Oxidase, As A Risk Factor For Myocardial Infarction: Effects On Platelets Activation And Oxidized LDL levels. J Am Coll Cardiol 49: 339A–340A.
- 48. Zafari AM, Davidoff MN, Austin H, Valppu L, Cotsonis G, et al. (2002) The A640G and C242T p22(phox) polymorphisms in patients with coronary artery disease. Antioxid Redox Signal 4: 675–680.
- 49. Lee WH, Hwang TH, Oh GT, Kwon SU, Choi YH, et al. (2001) Genetic factors associated with endothelial dysfunction affect the early onset of coronary artery disease in Korean males. Vasc Med 6: 103–108.
- 50. Minet L, Moller S, Vach W, Wagner L, Henriksen JE (2010) Mediating the effect of self-care management intervention in type 2 diabetes: a meta-analysis of 47 randomised controlled trials. Patient Educ Couns 80: 29–41.
- 51. Colhoun HM, McKeigue PM, Davey Smith G (2003) Problems of reporting genetic associations with complex outcomes. Lancet 361: 865–872.
- 52. Attia J, Thakkinstian A, D'Este C (2003) Meta-analyses of molecular association studies: methodologic lessons for genetic epidemiology. J Clin Epidemiol 56: 297–303.
- 53. Hemminki K, Lorenzo Bermejo J, Forsti A (2006) The balance between heritable and environmental aetiology of human disease. Nat Rev Genet 7: 958–965.
- 54. Schaid DJ, Jacobsen SJ (1999) Biased tests of association: comparisons of allele frequencies when departing from Hardy-Weinberg proportions. Am J Epidemiol 149: 706–711.
- 55. Manson JE, Colditz GA, Stampfer MJ, Willett WC, Rosner B, et al. (1990) A prospective study of obesity and risk of coronary heart disease in women. N Engl J Med 322: 882–889.
- 56. Whitlock G, Lewington S, Sherliker P, Clarke R, Emberson J, et al. (2009) Body-mass index and cause-specific mortality in 900 000 adults: collaborative analyses of 57 prospective studies. Lancet 373: 1083–1096.
- 57. Nordestgaard BG, Palmer TM, Benn M, Zacho J, Tybjaerg-Hansen A, et al. (2012) The effect of elevated body mass index on ischemic heart disease risk: causal estimates from a Mendelian randomisation approach. PLoS Med 9: e1001212.
- 58. Munafo MR, Flint J (2004) Meta-analysis of genetic association studies. Trends Genet 20: 439–444.
- 59. Stroup DF, Berlin JA, Morton SC, Olkin I, Williamson GD, et al. (2000) Meta-analysis of observational studies in epidemiology: a proposal for reporting. Meta-analysis Of Observational Studies in Epidemiology (MOOSE) group. Jama 283: 2008–2012.
- 60. Wu Z, Lou Y, Jin W, Liu Y, Lu L, et al. (2012) The Pro12Ala Polymorphism in the Peroxisome Proliferator-Activated Receptor Gamma-2 Gene (PPARgamma2) Is Associated with Increased Risk of Coronary Artery Disease: A Meta-Analysis. PLoS One 7: e53105.
- 61. Wu Z, Lou Y, Jin W, Liu Y, Lu L, et al. (2012) The C161T polymorphism in the peroxisome proliferator-activated receptor gamma gene (PPARgamma) is associated with risk of coronary artery disease: a meta-analysis. Mol Biol Rep
- 62. Rae J, Noack D, Heyworth PG, Ellis BA, Curnutte JT, et al. (2000) Molecular analysis of 9 new families with chronic granulomatous disease caused by mutations in CYBA, the gene encoding p22(phox). Blood 96: 1106–1112.
- 63. Tsai JC, Jain M, Hsieh CM, Lee WS, Yoshizumi M, et al. (1996) Induction of apoptosis by pyrrolidinedithiocarbamate and N-acetylcysteine in vascular smooth muscle cells. J Biol Chem 271: 3667–3670.
- 64. Ushio-Fukai M, Zafari AM, Fukui T, Ishizaka N, Griendling KK (1996) p22phox is a critical component of the superoxide-generating NADH/NADPH oxidase system and regulates angiotensin II-induced hypertrophy in vascular smooth muscle cells. J Biol Chem 271: 23317–23321.