The PlA1/A2 Polymorphism of Glycoprotein IIIa as a Risk Factor for Stroke: A Systematic Review and Meta-Analysis

Background The PlA1/A2 polymorphism of glycoprotein IIIa (GPIIIa) has been reported to be associated with risk of stroke in some studies, although other studies suggest no such association. This meta-analysis and systematic review was conducted to investigate the hypothesis that carriage of the PlA2 allele is a risk factor for stroke. Methods Electronic databases (MEDLINE and EMBASE) were searched for all articles evaluating carriage of the PlA2 allele and the incidence of stroke. Pooled odds ratios (ORs) were calculated using fixed-effect and random-effect models. Findings A total of 35 articles were eligible for inclusion, of which 25 studies were suitable for statistical analysis. For carriage of the PlA2 allele, OR 1.12 (n = 11,873; 95% CI = 1.03–1.22; p = 0.011) was observed for the incidence of stroke in adults, with subgroup analyses identifying the association driven by stroke of an ischaemic (n = 10,494; OR = 1.15, 95% CI = 1.05–1.27; p = 0.003) but not haemorrhagic aetiology (n = 2,470; OR = 0.90, 95% CI = 0.71–1.14; p = 0.398). This association with ischaemic stroke was strongest in individuals homozygous for the PlA2 allele compared to those homozygous for wild-type PlA1 (n = 5,906; OR = 1.74, 95% CI = 1.34–2.26; p<0.001). Subgroup analysis of ischaemic stroke subtypes revealed an increased association with stroke of cardioembolic (n = 1,271; OR 1.56, 95% CI 1.14–2.12; p = 0.005) and large vessel (n = 1,394; OR = 1.76, 95% CI 1.34–2.31; p<0.001) aetiology, but not those of small vessel origin (n = 1,356; OR = 0.99, 95% CI 0.74–1.33; p = 0.950). Egger's regression test suggested a low probability of publication bias for all analyses (p>0.05). Conclusions The totality of published data supports the hypothesis that carriage of the PlA2 polymorphism of GPIIIa is a risk factor for ischaemic strokes, and specifically those of cardioembolic and large vessel origin.


Introduction
The platelet fibrinogen receptor is integral to primary haemostasis, since it regulates platelet aggregation and the formation of stable thrombus following vascular injury [1,2]. The mature fibrinogen receptor is formed from the dimerisation of two glycoprotein (GP) subunits, GPIIb and GPIIIa [3], and has a number of stable allelic variants based on single amino acid substitutions [4]. The PlA2 polymorphism of GPIIIa is formed from the substitution of leucine by proline at position 33, and is one of the more studied variants due to reports that it may be associated with cardiovascular disease.
In 1996, carriage of the PlA2 allele of GPIIIa was reported to be associated with myocardial infarction (MI) when compared to subjects homozygous for the PlA1 wild-type [5]. Subsequently, there has been much inter-study disagreement as to the presence of any association, although we have recently performed a metaanalysis which clearly demonstrates a significant association between carriage of the PlA2 allele and MI in younger subjects (#45 years old: n = 9,547; odds ratio (OR) 1.205, 95% confidence interval (CI) 1.067-1.360; p = 0.003) [6].
MI and ischaemic stroke are modulated by the same cardiovascular risk factors, although differing arterial specificity for these risk factors underlies subtle differences in their respective pathophysiology [7]. Divergence in aetiology of these pathologies is further illustrated by the observation that conventional cardiovascular risk factors only account for a small proportion of all stroke risk, and familial studies suggest that genetic factors may be more important [8]. Indeed, the major common loci associated with MI have been found to not give a risk profile of similar magnitude for ischaemic stroke, although it has been suggested that they may play a role restricted to specific stroke subtypes [9].
The PlA2 polymorphisms of GPIIIa may transcend these pathophysiological discrepancies as, lying adjacent to the ligand binding site of the fibrinogen receptor, it is positioned to modulate the final common pathway of platelet-mediated thrombosis. Indeed, a meta-analysis published in 2010 considering 187 candidate genes for ischaemic stroke identified 14 studies examining the PlA2 polymorphism, and suggested the presence of a possible association [10], but with no subgroup analyses performed.
Here we present the most comprehensive meta-analysis to date of the association between carriage of the PlA2 allele and ischaemic stroke, with inclusion of 25 studies and 11,873 subjects. Subgroup analyses based on subject demographics and ischaemic stroke subtypes are presented. Additionally, we report the first meta-analysis examining the possible association of the PlA1/A2 polymorphism with haemorrhagic stroke in 1,621 subjects, to further explore the relationship between this polymorphism and cerebrovascular events.

Search strategy
Electronic databases (MEDLINE and EMBASE) were searched up until 30 th July 2013 for all articles evaluating carriage of the PlA2 polymorphism of GPIIIa and the incidence of stroke. The medical subject headings (MeSH) terms used in the primary literature search were 'stroke', 'cerebrovascular', 'subarachnoid', 'subdural', and 'extradural' in combination with 'glycoprotein IIIa', 'integrin beta3' and their associated synonyms. No language restrictions were in place during the primary search. A secondary search of the literature was performed by hand, based on all potentially relevant articles identified from the primary search.

Inclusion criteria and data extraction
Articles were considered suitable for inclusion in the systematic review if they studied association between carriage of the PlA2 polymorphism and risk of stroke. Studies performed wholly in children (,18 years) were excluded, although studies that considered 'young adults' (often defined as 15-44 years) were included. Non-English language articles were included if sufficient information could be extracted from the abstract.
In the case of studies suitable for statistical analysis, the following data were extracted where available: the number of cases/controls for each genotype (or calculated OR where raw data were not available), stroke subtype and population demographics (age, sex, ethnicity and the presence of traditional cardiovascular risk factors). For duplicated studies, the most complete dataset was utilised. Where more than one control group was available, those with cardiovascular risk factors were selected in preference to healthy subjects. Where genotype information was reported for .1 subpopulation as defined by geographic region or ethnic origin, each subpopulation was considered separately in the analyses [11].

Statistical analysis
Data were analysed using Comprehensive Meta-analysis version 2 software (Biostat Inc., NJ, USA). Pooled ORs were calculated using fixed-and random-effects models, along with the 95% CIs to measure the strength of association. Fixed-effects summary ORs were calculated using the Mantel-Haenszel method [12], and the DerSimonian method was used to calculate random-effects summary ORs [13]. For data where more than one outcome was reported, combined effects were calculated as necessary [14]. Pooled ORs presented in the results were calculated using the fixed-effects model unless otherwise stated.
Tests for heterogeneity were performed for each analysis, with significance set at p,0.05 [15]. I 2 was also calculated for each analysis, where $50% may represent substantial heterogeneity [16]. For assessment of publication bias, we utilised a funnel plot and Egger's regression asymmetry test [17]. 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.

Ischaemic stroke subgroup analyses based on subject demographics
Data were available to calculate pooled ORs based on subject age, ethnicity and sex (Table 3). No subgroup analyses based on the presence or absence of traditional cardiovascular risk factors were possible. Of these subgroup analyses, only Caucasian ethnicity was found to result in a significant association between

Publication bias
Publication bias was assessed by construction of funnel plots and calculation of Egger's regression intercept. All analyses demonstrated a low probability of publication bias (p,0.05). Furthermore, subgroup analyses based on study size did not suggest increased publication of positive results by smaller studies. For studies with ,250 cases, OR was 1.15 (n = 4,496; 95% CI 0.99-1.33; p = 0.078), and for studies with $250 cases, OR was 1.16 (n = 5,797; 95% CI 1.03-1.31; p = 0.017) ( Table 3).

Discussion
The results of this study demonstrate that carriage of the PlA2 allele of GPIIIa is a strong risk factor for ischaemic stroke. The validity of this association is strengthened by the finding that subjects homozygous for the mutant allele have almost five times the risk of stroke compared to those subjects who are heterozygous.
An age-skewed risk profile caused by a relative decrease in the influence of genetic factors with increasing age might have been expected, given the results of our recent meta-analysis of the association between PlA2 carriage and risk of MI, which demonstrated a stronger association in younger patients [6,65]. However the association in young adults was found to be of similar magnitude to that in the primary analysis of ischaemic stroke, notwithstanding the non-significant CI. The apparent absence of an age-dependent risk profile may be partially attributed to an underpowered subgroup (441 young adult cases) but also to the fact that unlike in MI, conventional risk factors only account for a small proportion of all stroke risk and that genetic factors are thought to be more important [8]. The association between stroke and carriage of the mutant allele is therefore not obscured in older individuals in the same way as in the MI population.
Subgroup analyses were performed for ethnicity, sex and study size, the results for sex and study size being consistent with the primary analysis suggesting that these factors are not major modulators of the increased risk associated with the polymorphism. Caucasian ethnicity produced a significantly higher association, suggesting that there may be ethnic differences in the impact of carriage of the PlA2 allele. Analysis based on ethnicity was clearly limited by the underlying low prevalence of the PlA2 allele in non-Caucasian populations. The prevalence of the allele in Asians is low, and this was emphasised in three studies of Chinese populations that identified no carriers of the mutant allele across their 554 cases [59][60][61]. Data were not available to perform either quantitative or qualitative analyses based on the presence or absence of conventional cardiovascular risk factors.

Ischaemic stroke subtypes
Subgroup analyses based on ischaemic stroke subtype identified an increased association between carriage of the PlA2 allele and stroke of CE or LVD aetiology. These findings are based on three studies per subtype and contain low numbers of cases (520 and 452 The PlA2 Allele and Risk of Stroke PLOS ONE | www.plosone.org respectively). However, much published data suggest that platelet activation modulates CE and LVD stroke rather than SVD. In a prospective study of 54 cases of non-CE stroke, those with LVD were found to have significantly higher platelet expression of CD63 during the acute phase, and significantly higher CD62P (Pselectin) expression up to 30 days post event, thus indicating higher levels of platelet activation [66]. The capacity of antiplatelet agents taken within 7 days of an ischaemic stroke to modulate stroke severity has been found to be much greater in LVD than in other stroke subtypes, in a study of 1,622 patients [67]. Moreover, despite the general assumption that CE thrombi are generally platelet-poor, autopsy data from 17 patients who died within 30 days of an ischaemic stroke found that, in CE thrombi, there was a greater proportion of red blood cells and a reduced proportion of fibrin, but no significant difference in the proportion of platelets, when compared to LV thrombi [68].
A further divergence in the aetiology of ischaemic stroke between LVD and SVD may arise because of differences in wall shear stress in the respective vessels. The mechanism by which platelets aggregate is dependent on shear stress, with transient aggregation occurring in the absence of activation at high shear rates [1]. Interestingly, in vitro investigations into the effect of the proline substitution on ligand binding have observed increased adhesion to fibrinogen and von Willebrand factor (vWF) in cell culture under conditions of shear stress [69], compared to static systems where no significant differences in maximal binding (B max ) or dissociation constant (K d ) were observed [70,71]. It was previously thought that shear stress was relatively constant throughout the vascular system, but more recent publications suggest that there is significant variability and that the common carotid artery is the site with highest shear [72]. Such differences in shear stress between large and small vessels, with resultant divergent effects on platelet activation, may contribute to the increased association of stroke observed, in carriers of the PlA2 allele, in CE and LVD when compared to SVD.

Haemorrhagic stroke subgroup
There is a suggestion, from the results of this study, of an apparent decrease in the risk of haemorrhagic stroke in subjects carrying the PlA2 allele. This however was non-significant, but it should be noted that this analysis was based on a relatively small pool of 2,470 subjects, and merits further investigation in future studies. If confirmed, this would give further insight into the mechanism by which the polymorphism affects platelet function. As discussed above, the amino acid substitution lies adjacent to the ligand binding site of the fibrinogen receptor and so may modulate normal platelet function; however, whether this is truly so remains contentious. Ex vivo studies have provided conflicting data on the effect that carriage of the PlA2 allele has on platelet fibrinogen binding and on platelet aggregation, in both healthy and ischaemic stroke cohorts [70,73,74] [75]. The PlA1/A2 polymorphism has also been hypothesised to modulate efficacy of anti-platelet drug therapy, and hence to influence cardiovascular risk whilst on such therapy; however, a recent meta-analysis suggests that this is unlikely to be the case [76].
Observations from in vivo studies appear to be more consistent, with carriage of the PlA2 allele resulting in a reduction in bleeding time both pre-and post-aspirin therapy in healthy subjects [77,78]. Additionally, carriage of the mutant allele was found to reduce peri-operative bleeding following coronary artery bypass grafting in aspirin-naïve subjects [79].
The totality of these experimental findings, when viewed in the context of the clinical data reported in the present study, suggests that carriage of the PlA2 allele does indeed result in modulation of platelet function leading to increased activation and aggregation.
The clinical manifestation of this modulation is an increased propensity for thrombus formation. The discrepancy with ex vivo findings may possibly be due to the impact of the polymorphism being dependent on interaction of platelets with the vasculature, and this hypothesis warrants further investigation in future in vitro studies.

Study limitations
There are a number of limitations to the study presented here that are inherent to the data available. Due to the increase in the prevalence of traditional cardiovascular risk factors with age, it is preferable to include data that are adjusted for these factors as the contribution of a single gene polymorphism to a multifactorial, polygenic process is likely to be subtle [11] [65]. Data were not available to perform such analysis nor for subgroup analyses based on non-Caucasian ethnicity or female sex, and the data for young adults were limited.
The majority of strokes occur out-of-hospital with many patients not surviving the index event [80]. Inclusion into the studies reported here necessitated survival for a period of time post-event, and so the presence of a mortality bias may attenuate the true association. There is no post-mortem study equivalent for the Helsinki Sudden Death Study for stroke [81] , but a polymorphism in GPIIb has been shown to increase the risk of fatal stroke and to decrease post-stroke survival suggesting the possibility of similar findings with carriage of the PlA2 allele [82].

Conclusions
We have demonstrated that carriage of the PlA2 polymorphism of GPIIIa is associated with an increased risk of ischaemic stroke in adults, and this risk is higher in homozygous subjects and those with strokes of CE and LVD aetiology. This risk is likely to be modulated by increased platelet activity that, in conjunction with interaction with the vascular wall, results in increased thrombus formation. The magnitude of the increased risk in PlA2 carriers is similar to that seen in MI, and suggests that PlA1/A2 genotyping may add usefully to risk stratification for patients at risk of stroke.

Author Contributions
Conceived and designed the experiments: CNF AF. Performed the experiments: CNF BHE. Analyzed the data: CNF BHE. Wrote the paper: CNF AF.