The PlA1/A2 Polymorphism of Glycoprotein IIIa as a Risk Factor for Myocardial Infarction: A Meta-Analysis

Background The PlA2 polymorphism of glycoprotein IIIa (GPIIIa) has been previously identified as being associated with myocardial infarction (MI), but whether this represents a true association is entirely unclear due to differences in findings from different studies. We performed a meta-analysis to evaluate whether this polymorphism is a risk factor for MI. Methods Electronic databases (MEDLINE and EMBASE) were searched for all articles evaluating genetic polymorphisms of GPIIIa. For studies where acute coronary events were recorded in association with genetic analysis, pooled odds ratios (ORs) were calculated using fixed-effects and random-effects models. The primary outcome measure was MI, and a secondary analysis was also performed for acute coronary syndromes (ACS) more generally. Findings 57 studies were eligible for statistical analysis and included 17,911 cases and 24,584 controls. Carriage of the PlA2 allele was significantly associated with MI (n = 40,692; OR 1.077, 95% CI 1.024–1.132; p = 0.004) but with significant publication bias (p = 0.040). The degree of association with MI increased with decreasing age of subjects (≤45 years old: n = 9,547; OR 1.205, 95% CI 1.067–1.360; p = 0.003) and with adjustment of data for conventional cardiovascular risk factors (n = 12,001; OR 1.240, 95% CI 1.117–1.376; p<0.001). There was a low probability of publication bias for these subgroup analyses (all p<0.05). Conclusions The presence of significant publication bias makes it unclear whether the association between carriage of the PlA2 allele and MI is true for the total population studied. However for younger subjects, the relative absence of conventional cardiovascular risk factors results in a significant association between carriage of the PlA2 allele and MI.


Introduction
The fibrinogen receptor is the most abundant integrin on the platelet surface [1], and through binding a combination of fibrinogen, von Willebrand factor and fibronectin, its main function is to terminate haemorrhage following vascular injury. Despite this important physiological function, it also plays a pathological role when stimulated excessively or inappropriately, and is a principal mediator of acute coronary thrombosis [2].
The glycoprotein IIIa (GPIIIa) subunit of the fibrinogen receptor has a number of stable allelic variants resulting from single amino acid substitutions [3], of which the PlA1/A2 diallelic antigen system (involving a leucine-proline polymorphism at position 33, which alters the protein conformation and spatial orientation of the ligand-binding region) has been associated with an increased prevalence of cardiovascular disease. In 1996, Weiss et al reported an association between carriage of the PlA2 allele and cardiovascular disease, with the association most marked in subjects with unstable angina or myocardial infarction (MI) under 60 years of age (Odds ratio (OR) 6.2, 95% confidence interval (CI) 1. 8-22.4; p = 0.002) [4]. The subsequent literature has demonstrated marked inter-study variation in the level -and indeed presence -of such an association, which in part reflects generally underpowered investigations of an allele with a frequency in Caucasian populations of approximately 15 per 100 [5], falling to 1 per 100 in Oriental populations [6].
The identification of a contribution by a single gene polymorphism to a multifactorial, polygenic pathological process is challenging and requires a very large sample size [9]. Here we present the largest meta-analysis to date, involving 42,495 subjects, to address the question of whether carriage of the PlA2 polymorphism is an independent risk factor for acute coronary events.

Search strategy and selection criteria
Electronic databases (MEDLINE and EMBASE) were searched without language restriction up until 1 st April 2013 for all articles evaluating genetic polymorphisms in the platelet GPIIIa receptor. The Medical Subject Headings terms used for the primary search related to genetics (e.g. 'gene', 'polymorphism', 'mutation' and 'genotype') in combination with glycoprotein IIIa (e.g. 'glycoprotein IIIa', 'GP IIIa' and 'integrin beta 3'). Following removal of duplicates, a total of 2,288 articles were identified in the primary search. To further encompass all relevant literature, a secondary search of the references from reviews and included studies was performed.
All articles that investigated the association between carriage of the PlA2 polymorphism and acute coronary syndromes (ACS) were considered potentially eligible for inclusion and, based on analysis of title and abstract, 114 potentially suitable articles were identified. For inclusion into the meta-analysis, studies must have reported the distribution of the PlA1/A2 genotype in relation to the prevalence of a coronary event, either as raw data or calculated ORs. Both case-control and cohort studies were considered, with familial-based studies and studies without an English translation excluded.

Data extraction
Data were extracted from each study according to a predefined protocol: study design, number of cases/controls, geographic location and/or ethnicity, sex and clinical outcome. The primary outcome measure was MI with a secondary outcome being ACS more generally. For the secondary outcome measure, data on the incidence of ACS were included preferentially where MI incidence was also reported.
Ethnic group was recorded where explicitly stated within a study. Where genotype information was reported for .1 subpopulation as defined by geographic region or ethnic origin, each sub-population was considered separately in the analyses [9]. This was the case for studies by Herrmann et al and Kekomaki et al [72,73].
Where data adjusted for age, gender, ethnicity and cardiovascular risk factors were available, these were analysed in preference to raw data. Similarly, in studies which had more than one control group, subjects with coronary artery disease were used in preference to healthy subjects as controls.

Statistical analysis
Data were analysed using Comprehensive Meta-analysis software, version 2 (Biostat, USA). Pooled ORs were calculated using fixed-and random-effects models, along with the 95% CI to measure the strength of association. Fixed-effects summary ORs were calculated using the Mantel-Haenszel method [74], and the DerSimonian method was used to calculate random-effects summary Ors [75]. For data where more than one outcome was reported, combined effects were calculated as necessary [76]. Pooled ORs presented in the results were calculated using the fixed-effects model unless otherwise stated.
Tests for heterogeneity were performed for each meta-analysis, with significance set at p,0.05 [77]. I 2 was also calculated for each analysis, where $50% may represent substantial heterogeneity [78]. For assessment of publication bias, we utilised a funnel plot and Egger's regression asymmetry test [79]. In addition, the effect of individual studies on the summary OR was evaluated by reestimating and plotting the summary OR in the absence of each study.

Subgroup analyses based on subject demographics
Data were available to calculate pooled ORs based on subject ethnicity, sex and age ( Table 2). Analysis based on ethnicity was limited to Caucasians as the majority of studies did not explicitly state the ethnicity of participants. A non-significant pooled OR for the association of PlA2 carriage with MI was observed for the 11 available studies (n = 10,585; OR 1.050, 95% CI 0.962-1.146; p = 0.272) [71,81,88,[95][96][97]103,109,112,127,128].
Subgroup analyses based on subject age at first MI demonstrated an increased level of association with decreasing age ( Figure 4). For subjects #45 years old, carriage of the PlA2 allele produced a pooled OR for MI of 1.205 (n = 9,547; 95% CI 1.067-1.360; p = 0.003) with significant heterogeneity (I 2 = 70.3%, p, 0.001), and analysis using the random-effects model again increased the level of association (OR 1.356, 95% CI 1.044-1.762; p = 0.022) [72,82,84,86,91,97,107,116,[121][122][123]. Significant heterogeneity and an increased level of association as compared with the total population were also observed using the random-effects model for the #55 and #65 year-old subgroups.

Publication bias
Publication bias was assessed by plotting funnel plots and calculation of Egger's regression intercept. The funnel plot for the primary analysis was asymmetric with a significant Egger's regression intercept (p = 0.040), suggesting the likelihood of publication bias skewed towards studies that favour an association between carriage of the PlA2 allele and MI (Figure 7). This bias was independent of study size (,250 cases, p = 0.264; $250 cases, p = 0. 088), but dependent on study design (case-control studies, p = 0.021; cohort studies, p = 0.570). Further analysis of case-

Discussion
The data presented here appears to demonstrate an increased risk of MI in carriers of the PA2 allele, with an association that becomes stronger as age decreases. Significant publication bias identified in the primary analysis makes it unclear whether the association is true for the totality of the population studied. This bias is primarily driven by case-control studies and surprisingly independent of study size given that smaller studies are generally more prone to this effect. The presence of bias is consistent with the general findings of the two previous meta-analyses [7,8]. However despite these concerns, there remains a clear age effect with the association between carriage of the PlA2 allele and MI most evident for younger age cohorts.
The observed age-skewed risk profile may be explained by a relative (rather than absolute) decrease in the influence of genetic factors with age, given that the prevalence of conventional cardiovascular risk factors increases with age [131]. This hypothesis is supported by the increased association between carriage of the PlA2 allele and MI observed in the subgroup analysis using data adjusted for these risk factors. Similarly, subgroup analysis using unadjusted data and controls with coronary artery disease resulted in no significant association seen, and this is likely to be explained again by the dilution effect caused by the coexistence of conventional risk factors.

Aetiology of increased risk
It is unclear as to the mechanism by which carriage of PlA2 predisposes to increased cardiovascular morbidity and mortality. The aetiology of cardiovascular risk is not monogenic, but a complex polygenic interaction with environmental factors [131], with each additional factor contributing in an additive and sometimes in a synergistic manner [132]. This contribution can be clearly demonstrated in the increased risk that smoking adds to carriage of the PlA2 allele when compared to non-smoking PlA1 homozygotes [82,96,105,123]. It has been suggested that smoking and PlA2 increase cardiovascular risk via interacting mechanisms [133], but given the strong association of smoking with premature myocardial infarction [134], further data are required to test the strength of this association.
There remains the possibility of unidentified linkage disequilibrium with genes modulating other conventional cardiac risk factors, with elevated plasma lipids being previously linked to carriage of the PlA2 allele [133]. This hypothesis is however not supported by the data analysed within the present meta-analysis, with only one study finding higher triglycerides in carriers of the PlA2 allele [80] and conflicting reports on the levels of lipoprotein(a) [86,99]. Interestingly, Grove et al found that the association between carriage of the PlA2 allele and MI decreased as cholesterol levels increased, suggesting once again that the true effect of the PlA2 allele may be diluted and hence concealed by the concomitant presence of conventional risk factors [96].
The proximity of the PlA1/A2 epitope to the ligand binding site of GPIIIa has led investigators to consider how the single amino acid substitution of proline for leucine may affect the cycle of ligand association and dissociation with the fibrinogen receptor.  Studies have been inconclusive, with no difference observed in static systems but an enhancement of binding and outside-in signalling seen in cell culture under conditions of shear stress, thus potentially resulting in circulating platelets having a higher basal level of activation [2]. Similarly to plasma lipids, plasma fibrinogen concentration has been suggested as a potential modulator of the increased risk secondary to carriage of the PlA2 allele, but as with the lipid hypothesis, studies included in this meta-analysis do not support this association. Three studies reported a higher levels of fibrinogen in certain subgroups of individuals carrying the PlA2 allele [85,86,121] and two studies reported higher fibrinogen levels in PlA1 homozygous subjects [124,125].
Resistance to aspirin has been suggested as another potential mechanism by which carriage of the PlA2 allele may cause increased cardiovascular risk [135]. However, a recent large metaanalysis has suggested that this is not the case [136]. There is however significant inter-study heterogeneity, and the need for further studies in this regard remains.
A final avenue of investigation has been whether the PlA1/A2 antigens affect the degree or morphology of atherosclerosis. Carotid plaque morphology was examined by magnetic resonance imaging in 1,202 participants in the atherosclerotic risk in communities (ARIC) study [46]. Subjects who carried the PlA2 allele were found to have plaques with thinner fibrous caps, and these thinner caps represent the major precursor lesion for ACS [137]. However, this study was limited by a low frequency of the minor allele and technical constraints resulting in plaque morphology being assessed only in individuals with thick arterial walls.

Study limitations
A potential limitation of this meta-analysis is the presence of a mortality bias that may attenuate or entirely obscure any true association. Almost a third of individuals with a first major coronary event die out-of-hospital [138], and are not accounted for in the predominantly retrospective data presented in this metaanalysis. In fact, in most studies the subject must have survived a cardiac event for a number of months or even years to be available for inclusion. If carriage of the PlA2 allele is more likely to result in an immediately fatal cardiac event then the association would be attenuated, with the opposite effect observed if carriage of the PlA1 allele results in increased early mortality.
The potential presence of a mortality bias can be investigated by considering post-mortem data from out-of-hospital deaths. The Helsinki Sudden Death Study considered 700 white Finnish males (aged 33-70 years old) who had sudden, unexpected out-ofhospital deaths [139]. Carriage of the PlA2 allele was significant associated with acute coronary thrombosis in those diagnosed with sudden cardiac death (OR 3.4; 95% CI 1.5-6.3), with an increased association observed in those ,60 years old (OR 4.6; 95% CI 2.0-11.2). These results do suggest the presence of a mortality bias, but the data are not directly comparable to those presented within the present meta-analysis. Nevertheless, they do indicate that carriage of the PlA2 has an increased association with platelet-mediated thrombotic cardiac events, and are therefore consistent with the increased OR seen for the clinical outcome of MI or indeed ACS. Figure 5. Analysis of the association between carriage of the PlA2 allele and myocardial infarction based on the use of healthy controls or controls with known coronary artery disease. Analysis is of the PlA1/A1 versus PlA1/A2+PlA2/A2 genotype. doi:10.1371/journal.pone.0101518.g005 Table 3. Subgroup analyses of the association of carriage of the PlA2 allele and myocardial infarction by study characteristics.

Conclusions
In conclusion, carriage of the PlA2 allele is a significant risk factor for the development of acute coronary events in younger subjects, with data representing the total population subject to significant publication bias. This age-skewed risk profile appears to be the result of the relative impact of the polymorphism becoming attenuated as conventional cardiovascular risk factors develop with advancing age. The precise mechanism by which carriage of the PlA2 allele leads to increased cardiovascular risk remains unclear, however this should not impair its potential utility in risk stratification of younger subjects with modifiable risk factors.

Supporting Information
Checklist S1 PRISMA Checklist.