ABCB1 C3435T Polymorphism and Response to Clopidogrel Treatment in Coronary Artery Disease (CAD) Patients: A Meta-Analysis

Background A number of investigators have evaluated the association between the ABCB1 polymorphism and clopidogrel responding, but the results have been inconclusive. To examine the risk of high platelet activity and poor clinical outcomes associated with the ABCB1 C3435T polymorphism in CAD patients on clopidogrel, all available studies were included in the present meta-analysis. Methods We performed a systematic search of PubMed, Scopus and the Cochrane library database for eligible studies. Articles meeting the inclusion criteria were comprehensively reviewed, and the available data were accumulated by the meta-analysis. Results It was demonstrated that the ABCB1 C3435T variation was associated with the risk of early major adverse cardiovascular events (MACE) (T vs. C OR, 1.34; 95% CI, 1.10 to 1.62; P = 0.003; TT vs. CC: OR, 1.77; 95% CI, 1.19 to 2.63; P = 0.005; CT + TT vs.CC: OR, 1.48; 95% CI, 1.06 to 2.06; P = 0.02) and the polymorphism was also associated with the risk of the long-term MACE in patients on clopidogrel LD 300 mg (T vs. C: OR, 1.28; 95% CI, 1.10 to 1.48; P = 0.001; TT vs. CC: OR, 1.59; 95% CI, 1.19 to 2.13; P = 0.002; CT + TT vs.CC: OR, 1.39; 95% CI, 1.08 to 1.79; P = 0.01). The comparison of TT vs. CC was associated with a reduction in the outcome of bleeding (TT vs. CC: OR, 0.51; 95% CI, 0.40 to 0.66; P<0.00001). However, the association between ABCB1 C3435T polymorphism and platelet activity and other risk of poor clinical outcomes was not significant. Conclusions The evidence from our meta-analysis indicated that the ABCB1 C3435T polymorphism might be a risk factor for the MACE in patients on clopidogrel LD 300 mg, and that TT homozygotes decreased the outcome of bleeding compared with CC homozygotes.


Introduction
Clopidogrel inhibits the adenosine-diphosphate-induced platelet aggregation, reducing the cardiovascular complications in patients with coronary atherosclerotic heart disease (CAD), especially in those undergoing percutaneous coronary intervention (PCI) [1]. However, the pharmacodynamic response to clopidogrel varies greatly among patients [2], and patients with lesser degrees of platelet inhibition are more likely to experience recurrent ischemic events [3,4]. Although the mechanisms have not been fully clarified, genetic polymorphisms may play a vital role in individual susceptibility to drug response [5].
The ABCB1 (ATP-binding cassette, sub-family B, member 1, also called MDR1 or TAP1) gene encodes the intestinal efflux transporter P-glycoprotein, which modulates the absorption of clopidogrel [6]. The ABCB1 gene locates at 7, p21-21.1 [7], and more than 50 single-nucleotide polymorphisms (SNP) within this gene have been described in the literature. Among them, the ABCB1 C3435T (rs1045642) is extensively studied and some research has been shown that the ABCB1 C3435T genotype influences the impaired function of P-glycoprotein which can hinder the absorption of clopidogrel [8].
Antiplatelet response could be investigated through poor clinical outcomes and impaired response to antiplatelet therapy in the laboratory test. Simon et al. [9] first analyzed the effect of C3435T polymorphism on clinical outcomes in patients receiving clopidogrel and found that patients with TT genotype had a higher rate of subsequent cardiovascular events than those with CC genotype. One study [10] indicated that the ABCB1 C3435T polymorphism influenced ADP dependent platelet reactivity and showed that Tallele carriers were likely to have a poor response to antiplatelet therapy in the lab test. However, the results from different studies [9][10][11][12][13][14][15][16][17][18][19][20] were inconsistent. Thus, in the present study, a meta-analysis was performed to delineate the association between ABCB1 C3435T polymorphism and platelet activity as well as the risk of poor clinical outcomes in patients treated with clopidogrel.

Literature Search
Three electronic databases (PubMed, Scopus and the Cochrane library) were searched (the last search was updated in March 2012 with the following terms combined: antiplatelet, clopidogrel, aspirin, platelet activity, ABCB1, MDR1, multidrug resistance, polymorphism). All eligible studies were retrieved and their bibliographies as well as the previous meta-analysis were checked for other relevant studies.

Inclusion Criteria
The studies that met the following criteria were included: (1) published in English, (2) case-control studies on platelet activity and prospective cohort studies on clinical outcomes, (3) the evaluation of the ABCB1 C3435T polymorphism, platelet activity and the poor clinical outcomes in patients receiving clopidogrel, (4) availability of the genotype frequency on target population, and (5) the valid date, on publication or through corresponding by e-mail, to work out an odds ratio (OR) or P-value with 95% confidence interval (CI).

Data Extraction
Two reviewers independently extracted the data and reached a consensus on all items. The following information was achieved from each study: the first author's name, publication date,

Study Outcomes
The two parts of endpoints (high platelet activity and poor clinical outcomes) were studied. The poor clinical outcomes included major adverse cardiovascular events (MACE) which were composed of cardiovascular death, non-fatal myocardial infarction, non-fatal stroke, as well as all-cause mortality, MI, stroke, definite or probable stent thrombosis and major or minor bleeding.

Statistical Analysis
The observed genotype frequencies in controls or entire cohorts were tested to compare with the expected genotype frequencies by    Hardy-Weinberg equilibrium (HWE). Crude odds ratio (OR) with 95% confidence interval (CI) in each study was used to assess the strength of association between ABCB1 C3435T polymorphism and platelet activity as well as the poor clinical outcomes in patients who received clopidogrel. According to the method described by Woolf [21], the pooled ORs were assessed for allele comparison (T vs. C), dominant genetic model (CT + TT vs.CC), recessive genetic model (TT vs. CC + CT) and homozygote comparison (TT vs. CC), and its significance was evaluated by the Z-test. Heterogeneity between studies was diagnosed by the use of the x 2 -based Q statistic test, and regarded as significant if p value was less than 0.1 [22]. Meanwhile the statistic of I 2 was used to efficiently test for the heterogeneity, with I 2 less than 25%, 25-50%, and greater than 50% as low, moderate and high degree of inconsistency, respectively [23]. The fixed-effect method was adopted if the effects were appeared to be homogeneous, or the random-effect model was conducted. Subgroup analyses were applied to identify the heterogeneity. Sensitivity analyses were conducted by sequential omission of individual studies respectively to detect the potential influence of each study set to the pooled ORs. In addition, publication bias was carried out by the funnel plot, and the symmetry of the plot distribution indicated the absence of publication bias [24]. Funnelplot asymmetry was assessed with the Begg's [25] and Egger's [26] tests. All statistical tests were performed with the Stata (v.10.0, Stata Corporation) and Review Manager (v.5.1, The Cochrane Collaboration), and were considered significant if the 2-sided P value was less than 0.05.

Study Characteristics
A total of 113 studies on the ABCB1 C3435T polymorphism with respect to platelet activity and the poor clinical outcomes were found, of which 22 replicated studies were excluded. Additionally, 39 reviews, 1 meta-analysis and 3 studies not in English were excluded. Meanwhile, 29 irrelevant studies were excluded by reviewing the title and abstract and one trial [27] was identified by screening the bibliographies. And then, five studies [28][29][30][31][32] were excluded due to their insufficient genetic information, and three trials [27,33,34] on clinical outcomes were excluded because they were not cohort studies. The result on platelet activity test from Wang [11] was so suspicious that we excluded it from current meta-analysis on the polymorphism and the degree of platelet inhibition by test. Finally, twelve studies, of which four involved platelet activity and ten involved clinical outcomes, met the inclusion criteria (Fig. 1), and the main characteristics of them were summarized in Tables 1, 2 and 3. Various genotyping methods were applied including allele-specific polymerase chain reaction (PCR) [10,11], Taqman Assays [12][13][14][16][17][18][19][20], Affymetrix Assay and Illumina Infinium Beadchip Assay [15], as well as SNPlex [9]. Distribution of genotypes in the controls or the total of each cohort (Tables 4, 5, 6, 7, 8, 9 and 10) were all not deviated from HWE.

2.2.
Long-term major adverse cardiovascular events. The major adverse cardiovascular events (more than one year) had no significant association with ABCB1 C3435T polymorphism in all genotype genetic models (for CT + TT vs. CC: OR, 1.09; 95% CI, 0.77 to 1.54; P = 0.62; Fig. 3). The I 2 statistic indicated the between-study heterogeneity ( Table 11). The effect of the ABCB1 C3435T polymorphism was further evaluated in stratification analyses. According to the clopidogrel loading dose, we found that this polymorphism was associated with the risk of the MACE in patients treated with clopidogrel LD 300 mg in allele comparison (T vs. C: OR, 1.28; 95% CI, 1.10 to 1.48; P = 0.001), homozygote comparison (TT vs. CC: OR, 1.59; 95% CI, 1.19 to 2.13; P = 0.002), and dominant genetic model (CT + TT vs. CC: OR, 1.39; 95% CI, 1.08 to 1.79; P = 0.01; Fig. 3). However, no significant risk of the MACE in other subgroups with this polymorphism was observed in all comparisons. Meanwhile, the heterogeneity of each subgroup was decreased (Table 11).
2.3. Early major adverse cardiovascular events. Three studies provided data (MACE happened in about one month), and the heterogeneity was low for all comparisons except for that of TT vs. CC + CT (I 2 = 72%; P = 0.016) (  Fig. 4).
2.5. Ischemic stroke. The ischemic stroke rate in the five cohort studies was 0.69% (54 of the 7858). As described in Table 11, though no heterogeneity could be detected, the metaanalysis illustrated that ABCB1 C3435T polymorphism was unrelated to the rate of ischemic stroke in patients treated with clopidogrel (for CT + TT vs. CC: OR, 1.03; 95% CI, 0.54 to 1.96; P = 0.93; Fig. 6).
2.8. Bleeding. The bleeding rate in the five cohort studies was 7.82% (596 of the 7619). The comparison of TT vs. CC was associated with a significant reduction in the outcome of bleeding (TT vs. CC: OR, 0.51; 95% CI, 0.40 to 0.66; P,0.00001; Fig. 9).  No significance between ABCB1 C3435T polymorphism and other genetic models as well as heterogeneity were identified (Table 11).
In addition, the heterogeneity existed in allele comparison with regard to stent thrombosis (Table 11). When we stratified the trials by previous or current smoker percentage, the heterogeneity was not clear in the subgroup (percentage ,50%) (T vs. C: I 2 = 0%; P = 0.32) and the other (percentage .50%) (T vs. C: I 2 = 34%; P = 0.22).
Although we also found the heterogeneity in two genetic model contrasts of Myocadial infarction and one genetic model contrast of early MACE (Table 11), due to limited studies, we failed to explain the heterogeneity. Finally, the heterogeneity could not be detected significantly in other contrasts.

Sensitivity Analysis
Sensitivity analysis was performed in the ABCB1 C3435T dominant genetic model (CT + TT vs. CC). The significance of pooled ORs was not obviously affected by omission of individual studies except for MACE. One study [16] carried the greatest weight for long-term MACE. When it was excluded, the pooled pvalues were significant in all comparisons, whereas exclusion of any other did not influence the results. Similarly, exclusion of the study by Simon et al. [9], the pooled OR of the dominant genetic model in early MACE was not significant. Meanwhile, in this genetic model, the heterogeneity in our meta-analysis was not influenced excessively by exclusion of any single study.

Publication Bias
Funnel plot as well as Begg's and Egger's tests were carried out to access the publication bias of studies. Data showed that there was no evidence of publication bias in comparison of TT+TC vs. CC (Fig. 10).

Discussion
Clopidogrel, as a pro-drug, was known to require metabolic activation before inhibiting platelet aggregation. The ABCB1  C3435T had been revealed to be associated with loss of function of P-glycoprotein which decreased the active metabolite of clopidogrel. On the basis of antiplatelet responding in the laboratory test and poor clinical outcomes, several molecular cardiovascular studies were conducted to evaluate the association between the ABCB1 C3435T polymorphism and platelet response in CAD patients on clopidogrel, but the results were inconclusive. A former meta-analysis [35] showed that the association might exist between TT homozygotes of the ABCB1 C3435T polymorphism and risk of short-term recurrent ischemic events.
In the present meta-analysis, since we included newer studies [11,14,19,20] and conducted the research more meticulously with the subgroup study and more detailed trials which the former one had not included, new significance resulted. To begin with, the 3435T allele carrier was related with the risk of the early and longtern major adverse cardiovascular events in patients treated with clopidogrel LD 300 mg. However, we did not find the significant association in subgroup clopidogrel LD 600 mg and others. Simon et al. [9] first found that patients with TT genotype had a higher rate of subsequent cardiovascular events than those with CC genotype. The recent clinical trial showed that compared with a 300-mg loading dose, pre-treatment with a 600-mg clopidogrel loading dose before primary PCI was associated with improvement of angiographic results and 30-day major adverse cardiovascular events [36]. The platelet response in treatment might be influenced by both ABCB1 C3435T polymorphism and clopidogrel loading dose, which need further research to identify. Meanwhile another study [37] indicated that a 150 mg oral maintenance dose of clopidogrel resulted in more intense inhibition of platelet aggregation than a 75 mg maintenance dose, which suggested that the maintenance dose also interacted with the platelet activity. In addition, TT homozygotes decreased the outcome of bleeding compared with CC homozygotes in our meta-analysis. This was almost consistent with the result from the three respective trails [14,16,18]. However, one study [16] carried the greatest weight for this analyses and with limited studies included, the result should be interpreted with caution and further studies based on larger, stratified population should be examined.
Four studies [10,[12][13][14] on platelet activity tested by different methods were included in our research. This was the first metaanalysis that includes the studies on the polymorphism with the degree of platelet inhibition evaluated by empirical methods, though no significance was searched. Meanwhile, in the group of all-cause mortality, MI, Stroke, and stent thrombosis, the polymorphisms with them were also not significant. Various factors may inference these. Among them, we first pay attention to the evidence of heterogeneity, for which the reasons are unclear. It may be due to the following: the selection of methods; differences in age, gender, ethnicity, sample size; and the main clinical characteristics. For instance, the diabetes with insulin resistance lower the inhibition of platelet aggregation [38]. Various genotyping methods applied in different studies may also bring about the heterogeneity. Diverse definition of case in platelet activity tests were used, such as different time of evaluation, hence, the association mignt have been biased or simply lead to heterogeneity. If we carried out the subgroup by some of the above elements, the heterogeneity in some compares decreased.  Clopidogrel inhibits the platelet activity, and high platelet activity in patients treated with Clopidogrel indicates clopidogrel resistance or poor response to clopidogrel, which will lead to poor clinical outcome in the future. Though we have found some association between ABCB1 C3435T polymorphism and antiplatelet responding, we also should set our insights to the interaction between single-nucleotide polymorphisms (SNP) and other factors. As coexisting, rather than single, polymorphisms in different genes may be related to persistent platelet activation while on clopidogrel [39], so gene-gene interaction, such as P2Y12 or CYP2C19 with ABCB1 C3435T, should also be observed between ABCB1 C3435T polymorphism and antiplatelet responding. On the other hand, one recent research [40] showed that the combination of a calcium channel blocker and ABCB1 C3435T genotype influenced the change of 20 mmol ADP-induced maximal platelet aggregation (MPA) in smoking status receiving clopidogrel. In our meta-analysis, the small scale of population and inconsistent stratification standards in environmental exposures and genotypes lowered our statistical power to further explore the gene-environment interaction. As a result, we need to give careful consideration to more sophisticated gene-gene and gene-environment interactions in a future analysis, so as to obtain a more comprehensive understanding of the association between ABCB1 C3435T polymorphism and antiplatelet responding.
Besides, from the different aspirin doses or triple antiplatelet therapy our studies included, we believe that interaction between gene and drug combination may also exist. Prolonged use of aspirin may reduce the intestinal absorption of clopidogrel by inducing the expression of ABCB1 in human epithelial colorectal (Caco-2) cells [41]. Researchers recently started to focus on the interactions between genetic polymorphisms and clinical effect with triple antiplatelet therapy (cilostazol, clopidogrel and aspirin) [12]. All these have urged us to pay more attention to the interaction between gene and drug combination in using clopidogrel. However, since new antiplatelet medicine had gone to market, we discovered that ABCB1 genotypes were not significantly associated with clinical or pharmacological outcomes in patients treated with prasugrel [15] and the pharmacodynamic characteristics of ticagrelor were not influenced by CYP2C19 and ABCB1 genotypes [42]. These might overcome the difficulty in poor antiplatelet responding the ABCB1 genotypes associated and gave us fresh insight to the antiplatelet treatment in addition to aspirin and clopidogrel, however numerous trails should be arranged for further assessment.
Although considerable efforts have been put into the test, there are some limitations inherent in the study. First, the number of studies included are limited, especially for the information on the risk of MACE in patients treated with clopidogrel LD 300 mg and the outcome of bleeding. Thus the conclusion about these should be considered with caution. Second, detailed information such as the ethnicity and other main characteristic are not available in some studies, which further limit our evaluation. Third, gene-gene or gene-environment interaction, different loading or maintenance dose, influence from main clinical characteristics, standardized unbiased platelet activity evaluation and genotyping methods may affect the results. These variables can be planned more effectively by a separate analysis of these elements, to which we did not have access.
In summary, our meta-analysis indicated that the ABCB1 3435T allele carrier was related to the risk of major adverse cardiovascular events in patients on clopidogrel LD 300 mg, and TT homozygotes decreased the outcome of bleeding compared with CC homozygotes, whereas, the association between ABCB1 C3435T polymorphism and platelet activity as well as other risks of poor clinical outcomes were not significant. Thus, to validate our findings, additional larger studies need to focus on homogeneous cases along with standardized platelet activity evaluation and genotyping methods in further tasks.