Coronary angiography and subsequent percutaneous coronary intervention (PCI) is the appropriate treatment in many patients with moderate- and high-risk acute coronary syndromes (ACS). As ACS and complications after PCI are primarily platelet-driven, adequate antiplatelet therapy is of utmost importance in patients with ACS and after coronary stent implantation. The main antiplatelet agents currently used are salicylates (aspirin) and thienopyridines (clopidogrel, prasugrel). Both aspirin and clopidogrel inhibit platelet function irreversibly. Aspirin non-selectively acetylates the enzyme cyclooxygenase-1 (COX-1), thereby permanently inhibiting the ability of the platelet to synthesise the pro-thrombotic eicosanoid thromboxane A2. Clopidogrel and prasugrel inhibit the P2Y12 adenosine diphosphate (ADP) receptor, thereby preventing ADP-mediated platelet activation. After ACS and coronary stent deployment, the currently recommended treatment strategy is the use of dual antiplatelet therapy (DAPT) for at least one year, although the optimal duration of DAPT is subject to intensive investigation. Beyond this 12-month period, aspirin monotherapy is the currently recommended antiplatelet therapy.
Based on the superior efficacy associated with DAPT in various randomised clinical trials, an earlier unselective ‘one-size-fits-all’ strategy of DAPT was overwhelmingly adapted in clinical practice. However, continued occurrence of ischaemic events, particularly stent thrombosis (ST) and myocardial infarction (MI), despite DAPT at recommended doses, raised doubts about the appropriateness of a one-size-fits-all strategy. Given the incidence of ST events (2–4 %), approximately 100,000 patients worldwide develop an ST event annually, resulting in substantial mortality (in up to 40 % of cases).1,2 Numerous pharmacodynamic studies using multiple ex vivo methods indicating the ADP-induced P2Y12 receptor reactivity demonstrated the antiplatelet response variability to clopidogrel and the association of antiplatelet resistance/high on-treatment platelet reactivity to ADP to increased occurrence of ischaemic events, particularly in patients treated with PCI and stenting.3 Thus, persistent occurrence of ischaemic events and the irrefutable demonstration of clopidogrel response variability are two potent arguments against the widely practised non-selective or one-size-fits- all strategy of administering clopidogrel therapy and provided a strong rationale for monitoring platelet function during clopidogrel therapy.
Clopidogrel Response Variability
Järemo et al. reported inter-individual variability in antiplatelet response by measuring ADP-induced fibrinogen binding.4 In a subsequent important study, clopidogrel response variability was demonstrated by ADP-induced platelet aggregation and P-selectin and activated glycoprotein (GP) IIb/IIIa expression serially assessed at baseline and for 30 days following stenting (300 mg load and 75 mg/day maintenance dose); some patients had no demonstrable antiplatelet effect. The immediate concern at that time was that these latter patients were least protected from thrombotic event occurrence. In these potentially unprotected patients, the absolute difference between pre- and post-treatment platelet aggregation was ≤10 %. These patients were regarded as ‘resistant’.5 In this study, ~30 % of patients were resistant at days 1 and 5 post-stenting, and 15 % were resistant at day 30.
Since then numerous studies, using various laboratory methods to assess ADP-induced platelet function, such as turbidimetric aggregation, flow cytometry, to measure P-selectin and activated GP IIb/IIIa expression and vasodilator-stimulated phosphoprotein (VASP) phosphorylation levels, and point-of-care methods – VerifyNow P2Y12 assay, platelet mapping with thrombelastography and the Multiplate analyser – have been used to demonstrate clopidogrel response variability and resistance. Clopidogrel response variability and resistance are now accepted pharmacodynamic phenomena.3
Linking Clinical Outcomes to Clopidogrel Resistance/High On-treatment Platelet Reactivity
The pivotal properties of the P2Y12 receptor provide the rationale for ex vivo quantification of the intensity of the ADP-P2Y12 interaction as a means of identifying patients at increased thrombotic risk who require adjustment in antiplatelet therapy. To support the latter hypothesis, subsequent translational research studies aimed to link clopidogrel resistance or high on-treatment platelet reactivity (HPR) during clopidogrel treatment to ischaemic event occurrence in patients treated with PCI. Barragan et al. first demonstrated an association between a platelet reactivity index (PRI) >50 %, measured by VASP phosphorylation, and the occurrence of thrombotic events in a case-control study.6
At the same time, Matetzky et al., using aggregometry, observed that patients undergoing primary PCI for ST-segment elevation MI who were in the lowest quartile of clopidogrel responsiveness had the highest rates of ischaemic events during follow-up.7 Subsequently, it was suggested that the level of on-treatment platelet reactivity might be a superior risk predictor compared with clopidogrel responsiveness, because platelet reactivity to ADP was variable before clopidogrel treatment in patients on aspirin therapy.8 The important relationship between HPR, as measured by turbidimetric aggregometry, and the occurrence of ischaemic events in patients treated with stents was first prospectively demonstrated in the Platelet reactivity in patients and recurrent events post-stenting (PREPARE POST-STENTING) study (upper quartile, odds ratio [OR]: 2.6).9 In the Clopidogrel loading with eptifibatide to arrest the reactivity of platelets (CLEAR-PLATELETS) study, ADP-induced aggregation was measured serially over 18–24 hours in patients undergoing stenting.10 Patients who experienced peri-procedural MI had significantly greater mean aggregation compared with patients without MI; a threshold of 50 % mean platelet aggregation was associated with MI and was proposed as a therapeutic target. Multiple subsequent studies have consistently demonstrated that HPR is an important independent risk factor for the occurrence of thrombotic/ischaemic events after PCI (see Table 1).6,7,9–34 A recent consensus document has indicated that several platelet function tests are capable of identifying patients exhibiting a high on-treatment platelet reactivity status.3
Evidence to Support the Therapeutic Window Concept
In the PREPARE POST-STENTING study, a threshold of ~50 % maximal peri-procedural aggregation (20 μM ADP) was associated with six-month ischaemic event occurrence.9 Similarly, in the Clopidogrel effect on platelet reactivity in patients with stent thrombosis (CREST) study, ~40 % aggregation (20 μM ADP) was associated with stent thrombosis occurrence.22 In a third study, ~40 % pre-procedural platelet aggregation (5 μM ADP) among patients receiving long-term clopidogrel and aspirin therapy before stenting was associated with 12-month ischaemic event occurrence.12 These data suggest that adequate protection against ischaemic events with aspirin and clopidogrel therapy might be achieved by overall low to moderate levels of post-treatment platelet reactivity in the majority of patients. These findings have important implications for bleeding risks that may accompany markedly low levels of post-treatment platelet reactivity. There is some evidence from the recent studies to support the latter concept.
An increased responsiveness to clopidogrel measured by ADP-induced platelet aggregation using multiple electrode aggregometry (MEA) was associated with a 3.5-fold increased risk of procedure-related major bleeding in patients (n=2,533) undergoing PCI.35 In this study, more bleeding events were observed in patients with <188 aggregation unit (AU)-minutes, whereas the same investigators demonstrated a significant association of ischaemic events in patients with >468 AU-minutes. In a recent study by Campo et al., it was demonstrated that ≤86 P2Y12 reaction units (PRu) as measured by VerifyNow P2Y12 assay was significantly associated with one-month bleeding events, whereas ≥239 PRu was associated with one-month ischaemic event occurrence in 507 patients undergoing PCI.36 In another study by Gurbel et al., using the Thrombelastography Platelet Mapping assay in 225 patients undergoing stenting and treated with aspirin and clopidogrel, receiver-operating characteristic (ROC) curve analysis indicated that ADP-induced platelet-fibrin clot strength (MAADP) of >47 was best associated with three-year ischaemic event occurrence, whereas ≤31 MAADP was associated with bleeding events.37 Since new P2Y12 receptor blockers, such as prasugrel and ticagrelor, are associated with greater platelet inhibition in pharmacodynamic studies and more bleeding events in clinical trials, the concept of a therapeutic window based on platelet function measurement assumes significant importance in avoiding excessive bleeding risk.
High On-treatment Platelet Reactivity Defined by Receiver-operating Characteristic Curve Analysis
Importantly, studies have emerged that have used ROC curve analysis to define a threshold of on-treatment platelet reactivity associated with the optimal combination of sensitivity and specificity to identify thrombotic risk. It should be noted that such cut-off points may depend on the subset of patients studied. In fact, to date, cut-off values have been mainly investigated in patients undergoing PCI and there may be different targets, depending on the clinical setting and baseline risk profile.3 Most importantly, the observed cut-off values for platelet reactivity had a very high negative predictive value for thrombotic/ischaemic event occurrence, an observation of great potential clinical importance. However, the positive predictive value was fairly low for all assays. The latter observations are consistent with the fact that although a major determinant of thrombotic events, high on-treatment platelet reactivity is not the sole factor responsible for these events.
Several – some preventable – mechanisms may play a role in the occurrence of a heightened on-treatment platelet reactivity status, e.g. accelerated platelet turnover, genetic factors playing a role in the metabolisation of clopidogrel, a heightened baseline platelet reactivity, poor compliance, underdosing and drug–drug interactions.38
Platelet Function Testing in the Catheterisation Laboratory
Many studies have demonstrated that platelet function testing identifies patients with high on-treatment platelet reactivity and a higher risk of thrombotic events. It has also been shown that platelet reactivity decreases with higher doses of clopidogrel, switching to prasugrel or ticagrelor or with adding GP IIb/IIIa therapy. In addition, in small studies, tailoring antiplatelet therapy based on HPR was able to improve outcome after PCI.39–41 So it might be useful to test for HPR in these patients. However, very recently, the outcome of the Gauging responsiveness with a VerifyNow assay – impact on thrombosis and safety (GRAVITAS) trial was presented. In GRAVITAS, doubling the clopidogrel maintenance dose based on platelet function results did not reduce the incidence of thrombotic events after successful PCI.42 However, the GRAVITAS trial has some serious drawbacks: platelet function testing was done after successful PCI excluding the chance of preventing peri-procedural MI; the higher maintenance dose of clopidogrel was unable to prevent HPR in 40 % of the patients; the population included was mainly stable with a lower than anticipated event rate of “only 2.3 %” after six months of follow-up dose, making the trial severely underpowered (see Table 2). The proposed remedy of doubling the maintenance dosing of clopidogrel to 150 mg has been associated with only a moderate reduction in the prevalence of high on-treatment platelet reactivity and the use of more potent P2Y12 inhibitors, such as prasugrel or ticagrelor, might therefore be a better option.43–45
Thus the issue of tailoring antiplatelet therapy based on platelet function results prior to coronary stent implantation is certainly not a closed case and more work is urgently needed, preferably with potent platelet inhibitors. Probably the patients of interest are the higher-risk population, e.g. with multi-vessel stenting, left main stenting, ACS (including unstable angina, non-ST-elevation MI [NSTEMI] and ST-elevation MI [STEMI]), and patients with important risk-contributing clinical risk factors, such as diabetes mellitus, renal failure and reduced left ventricular ejection fraction, who have more to gain from stronger antiplatelet drugs.
On the other hand, it has been demonstrated that the majority of our patients undergoing PCI do not have HPR and that these patients have a low risk of thrombotic events. Applying stronger antiplatelet drugs to those patients will probably not reduce thrombotic risk but increase the risk of bleeding. Platelet function testing may therefore be used to reduce the thrombotic risk of our high-risk patients.
Genetics – The CYP2C19 Gene and Individualised Treatment with Clopidogrel
The conversion of clopidogrel to its active metabolite depends on the CYP2C19 gene. Carriers of at least one loss-of-function CYP2C19 allele (mostly the loss-of-function *2 variant) have reduced exposure to clopidogrel’s active metabolite, and consequently reduced platelet aggregation response compared with non-carriers.46
Recently, a meta-analysis in patients treated with clopidogrel (nine studies, 9,685 subjects) demonstrated that patients carrying one or two loss-of-function allele(s) had a higher risk of major adverse cardiovascular events (MACE) with hazard ratios (HR) of 1.55 and 1.76, respectively (see Figure 1).47 Furthermore, the HR for stent thrombosis for carriers of one loss-of-function allele was 2.67, and 3.97 for carriers of two loss-of-function alleles, compared with non-carriers (see Figure 2). The genetic post hoc analysis of the Platelet inhibition and patient outcomes (PLATO),48 Clopidogrel in unstable angina to prevent recurrent events (CURE)49 and Atrial fibrillation clopidogrel trial with irbesartan for prevention of vascular events (ACTIVE)-A49 trials and one case-control study50 were not included in this meta-analysis. Representing a population of which 64 % underwent PCI, PLATO showed a CYP2C19 effect only at 30 days (HR 1.37).48 A CYP2C19–clopidogrel interaction was not found and may have been missed in the ACTIVE-A and CURE trials because of the inclusion of a lower-risk population and a low proportion (<20 %) of patients that underwent PCI.49 In addition, a case-control study showed that stent thrombosis was significantly associated with the CYP2C19*2 allele (OR 1.70), corroborating the findings of the meta-analysis (see Table 3).50
In contrast to the *2 variant, the *17 variant is associated with increased CYP2C19 enzyme activity. Subsequently, the *17 variant has been shown to decrease both platelet reactivity51 and may reduce the risk of MACE52 in response to clopidogrel. However, patients with the CYP2C19*2/*17 genotype were found to exhibit increased platelet reactivity in response to clopidogrel, as compared with patients with the CYP2C19*1/*1 genotype.51 The heightening effect of CYP2C19*2 was only partly diminished by the concomitant presence of the CYP2C19*17 allele. This might be explained by the fact that carriage of the *2 allele leads to a complete loss of enzyme function, while CYP2C19*17 only enhances existing enzyme activity.
The ABCB1 gene is important because it encodes an intestinal efflux transporter also affecting clopidogrel. Four studies investigated the association between the ABCB1 C3435T polymorphism and adverse cardiovascular events in clopidogrel-treated patients.50,52–54 In the French registry of acute STEMI and NSTEMI (FAST-MI) cohort, TT genotype carriers showed a 72 % increased risk of death, non-fatal MI or stroke at one year compared with the CC genotype (see Figure 3).54 Also, the Trial to assess improvement in therapeutic outcomes by optimising platelet inhibition with prasugrel – Thrombolysis in myocardial infarction (TRITON–TIMI) 38 showed a 72 % increased risk of cardiovascular death, MI or stroke at 15 months compared with CC or CT patients (see Figure 4).53 No clopidogrel ABCB1 C3435T interaction was found in two other studies.50,52
Response to clopidogrel is most likely multigenic given its complex mode of action and hepatic activation. In fact, patients with two CYP2C19 loss-of-function alleles and at least one ABCB1 C3435T variant allele were shown to be at the highest risk of thrombotic events (HR 5.31).53,54
Recently, the common functional PON1 Q192R gene polymorphism was shown to determine pharmacokinetics, platelet response and clinical antithrombotic efficiency of clopidogrel among patients undergoing PCI. However, confirmation in other studies is awaited before PON1 polymorphism may be clinically relevant.
The question of whether CYP2C19 genotypes and/or platelet function testing should assist clinical decision making is heavily debated. The US Food and Drug Administration recently approved a new label for clopidogrel that includes a boxed warning stating that there is diminished effectiveness of clopidogrel in homozygous CYP2C19 loss-of-function allele carriers and alternative treatment should be considered in those patients. Following the new label, a high-quality report (American College of Cardiology Foundation [ACCF]/American Heart Association [AHA] clinical alert) was released providing nuanced guidance for clinicians on how to deal with the warning. It is recommended that genetic testing of CYP2C19*2 should be considered before initiating clopidogrel treatment in patients at moderate/high risk including patients undergoing elective high-risk PCI procedures, with other therapies such as prasugrel as a proposed alternative.55 Taking into account the knowledge about CYP2C19 loss-of-function variants and clopidogrel response, several strategies for antiplatelet treatment can be proposed. First, the decreased responsiveness to clopidogrel in loss-of-function allele carriers may be overcome by doubling the dose. There are indications that this may be a good strategy because a double dose for carriers of the *2 variant was shown to counterbalance the reduced platelet aggregation response to clopidogrel in a study of 60 patients undergoing elective PCI.56 Whether higher dosing for loss-of-function allele carriers translates into better outcomes remains to be determined.39 As previously stated in GRAVITAS, doubling the clopidogrel maintenance dose based on platelet function results did not reduce the incidence of thrombotic events after successful PCI.42 A second strategy could be treating all patients with prasugrel, of which the efficacy is independent of a patient’s CYP2C19 genotype. However, this would imply an enormous increase in drug cost, now generic clopidogrel is available. In addition, prasugrel is associated with an increased risk of bleeding compared with clopidogrel and is contraindicated for patients with a history of stroke or transient ischaemic attack (TIA). Finally, a genotype-guided antiplatelet strategy could be proposed in which patients are genotyped and those without a CYP2C19 loss-of-function allele receive clopidogrel, whereas prasugrel treatment would be initiated for patients (without history of stroke/TIA) carrying one or two loss-of-function alleles. Similarly, the allocation of prasugrel could be based on platelet function testing. By using modern techniques, genotype results can be obtained within a couple of hours.
It may be an option to use genotyping in clopidogrel-naive patients (STEMI, NSTEMI within four hours of a clopidogrel loading dose) and platelet function testing in those on adequate clopidogrel (75 mg for at least a week, 300 mg loading >24 hours or 600 mg loading >4 hours).
A cost-effectiveness study will soon be started at our department to evaluate this strategy. In addition, various trials are currently ongoing to investigate treatment strategies assisted by genetic testing (see www.clinicaltrials.gov NCT01134380, NCT00995514 and NCT01177592). It is highly questionable whether we should await the results from these studies to warrant genotyping-guided antiplatelet treatment. After all, there are data available showing there is a clopidogrel safety issue for patients carrying a CYP2C19 loss-of-function allele that can be solved using genotyping.
Obviously, the most important issue at the present time is the need to establish clinical utility (clinical implications with clear treatment decisions) to support the value of point-of-care platelet function testing in daily practice. The crucial question is whether dose or treatment adjustments on the basis of the results of platelet function testing improve clinical outcomes. Moreover, the magnitude of benefit and its clinical relevance depend highly on the absolute risk profile of the studied patient population. Crucial components in this clinical evaluation include an assessment of the absolute risk of the patient (multi-vessel coronary disease, left main intervention, STEMI, NSTEMI and important contributing clinical risk factors, such as diabetes mellitus, renal failure and reduced left ventricular ejection fraction).
No matter what level of evidence is required, it will be necessary to develop simple clinical algorithms to aid physicians in their interpretation and use of platelet function testing. The issue of personalised antiplatelet therapy on the basis of platelet function testing is important and worthy of effort and further study, as the cardiology community eagerly awaits the results of ongoing clinical trials.
DAPT is the cornerstone of treatment for patients with ACS and for those undergoing PCI. However, a wide inter-individual variability in the response to antiplatelet agents (and, in particular, clopidogrel) exists, resulting in a clinically relevant high on-treatment platelet reactivity state in a substantial subset of patients. At the present time, the routine measurement of platelet reactivity in the catheterisation laboratory has not been widely implemented because the management of patients who exhibit a high on-treatment platelet reactivity state is unknown. Moreover, the absolute risk profile of the studied patient population appears to be of crucial importance. Therefore, simple clinical algorithms to aid physicians in their interpretation and use of platelet function testing should be developed before the routine use of platelet function measurements in the routine care of patients with cardiovascular disease is recommended.