Advocating cardiovascular precision medicine with P2Y12 receptor inhibitors

Abstract

Antiplatelet therapy with P2Y₁₂-receptor inhibitors has become the cornerstone of medical treatment in patients with acute coronary syndrome treated with percutaneous coronary intervention (PCI). With over 100 million prescriptions filled since its approval, clopidogrel is the most widely used P2Y₁₂-receptor inhibitor. Dual antiplatelet therapy with clopidogrel plus aspirin has been associated with a lower rate of major cardiovascular events in patients after PCI than aspirin monotherapy. However, an alarmingly high number of clopidogrel-treated patients experience adverse thrombotic events. Insufficient P2Y₁₂-inhibition or high on-treatment platelet reactivity to adenosine diphosphate has stimulated the increased use of more potent P2Y₁₂ inhibitors: prasugrel (a third generation thienopyridine) and ticagrelor (a cyclopentyl-triazolo-pyrimidine). However, more potent platelet inhibition and low on-treatment platelet reactivity has resulted in increased major bleeding and higher costs. These limitations have suggested the need for an individualized antiplatelet approach in order to decrease thrombotic events and minimize bleeding. This model of personalized medicine integrates a patient’s demographic and biological data (pharmacodynamic, genomic, epigenomic, transcriptomic, and metabolic information) to target therapy in order to maximize efficacy while minimizing bleeding and costs. This review discusses the role of diagnostic tools such as platelet function and pharmacogenomic testing to personalize antiplatelet therapy.

Response variability to antiplatelet agents

Antiplatelet therapy with acetylsalicylic acid and a P2Y₁₂ receptor inhibitor [widely known as dual antiplatelet therapy (DAPT)] is the mainstay of secondary prevention in patients with acute myocardial infarction.^1–3 Although DAPT has been found to have greater efficacy, especially in patients after coronary stent implantation, a substantial proportion of patients continue to experience recurrent ischemic events despite such therapy. These clinical failures have been attributed to pharmacodynamic response variability primarily in clopidogrel-treated patients.^4–6

The observation that not all patients respond equally to antiplatelet treatment has prompted questioning and criticism of the non-selective ‘one-size-fits-all’ approach with clopidogrel therapy, as this treatment strategy completely ignores inter-individual pharmacodynamic variability.⁵ Pharmacodynamic response variability to P2Y₁₂ receptor inhibitors is a two-edged sword, placing hyperresponsive patients at potential risk of bleeding on the one hand and leaving hyporesponsive patients at risk for thrombotic events on the other.^7–9 The association between clopidogrel non- or poor responsiveness and high-on-treatment platelet reactivity (HPR) to adenosine diphosphate (ADP) with adverse ischemic events was first reported by Matetzky et al.⁴ in 2004 and Bliden et al.¹⁰ in 2007, respectively. The relation of HPR to thrombotic events in the percutaneous coronary intervention (PCI) population has been well established in multiple subsequent larger investigations. The strongest association has been observed for short-term thrombotic events, such as acute and sub-acute stent thrombosis, in patients after coronary stent implantation.^4,11–20 Several studies have demonstrated that testing of platelet function stimulated by ADP provides statistically significant prognostic information in clopidogrel-treated patients.^19,21–24 The earliest link between platelet function and ischemic risk was primarily observed in studies utilizing light transmission aggregometry (LTA) in patients undergoing PCI.^25–27 Subsequently, increased risk for cardiac ischemic events during short- and long-term follow-up in patients with HPR has been observed by other assays, including the VerifyNow^TM, the new generation impedance aggregometry test (Multiplate, MEA), the vasodilator stimulated phosphoprotein (VASP) phosphorylation assay,^{11,13,28–34} and thrombelastography.³⁵ On the other hand, low-on-treatment platelet reactivity (LPR) would indicate good or even hyper-responsiveness to P2Y₁₂ receptor inhibitor treatment, and might be related to bleeding events.^{29,30,36–39}

Factors associated with response variability to antiplatelet agents

Clopidogrel

Demographic, clinical, and genetic variables have been linked to HPR in patients treated with clopidogrel.⁴⁰ Obesity, renal dysfunction, diabetes, age, reduced left ventricular function, concomitant use of drugs, inflammation, and the presence of an ACS are clinical variables associated with HPR (Figure 1).^41–47 HPR is more prevalent in diabetic patients compared with patients without diabetes, especially in those requiring insulin therapy.^48,49 Mechanisms proposed for heightened platelet reactivity in patients with diabetes include changes in calcium metabolism, P2Y₁₂ receptor signalling up regulation, increased exposure to ADP, and increased platelet turnover.^50–52

Prasugrel and ticagrelor

The more potent P2Y₁₂-inhibitors, ticagrelor and prasugrel, were introduced to overcome the limitation of response variability to clopidogrel. Recent studies have however shown that the latter phenomenon may not be confined to clopidogrel treatment alone. In the acute phase of ST-elevation myocardial infarction (STEMI), 20–34.6% of the patients treated with prasugrel and 31.8–46.2% treated with ticagrelor exhibited HPR.^34,53–58 Recent studies report for prasugrel a HPR phenotype of at least 12% in a mixed cohort of ACS patients and patients that failed to respond to clopidogrel.⁵⁹ In contrast to these findings, other studies report only a marginal role of HPR in prasugrel- or ticagrelor-treated patients, but detected a remarkable occurrence of LPR in ACS patients. Notably, HPR in clopidogrel-pretreated patients predicted the pharmacodynamic response in prasugrel-treated patients. The TAILOR study and two other randomized studies have shown that up to 20% of patients continued to exhibit HPR despite switching from clopidogrel to prasugrel.^55,60,61 Furthermore, patients with HPR on clopidogrel showed a marked increased platelet inhibition in response to ticagrelor compared with prasugrel.⁶² Overall, the pharmacodynamic effects of prasugrel and ticagrelor seem to be similar, but time of pharmacodynamic assessment, the loading sequence and pretreatment with clopidogrel limit comprehensive conclusions on the comparative pharmacodynamic efficacy of the novel P2Y₁₂-inhibitors.^63,64

Interestingly, HPR in prasugrel and ticagrelor seems to differ from treatment failures observed in clopidogrel-treated patients. The approach of accusing inadequate biotransformation as major cause for inadequate response to antiplatelet treatment does not fit to the novel P2Y₁₂ inhibitors, as they are much less prone to alterations in activation as clopidogrel. To assess HPR patients, other techniques that investigate post-receptor signal transduction might be more suitable to elucidate this phenomenon. The idea that not only pro-aggregatory stimuli contribute to platelet hyper-reactivity but also the loss of function (LoF) of the inhibitory pathways regulated by nitric oxide (NO) and prostaglandin I₂ (PGI₂) might be attributable to this gives a more comprehensive insight into this phenomenon. It has been shown that especially the NO-soluble guanylyl cyclase (sGC) signalling pathway is vulnerable to impairment at various steeps and that attenuation of sGC is the main mechanism of platelet resistance to NO.⁶⁵ This diminished sGC activity in response to NO has been observed in various clinical states associated with increased cardiovascular risk, such as hypercholesterolaemia or diabetes and in cardiovascular disease states, including heart failure, ACS, aortic valve stenosis and newly diagnosed atrial fibrillation.⁶⁶ Variability in function of the platelet adenylyl cyclase (AC)/cyclic adenosine monophosphate system might partially explain the observations from the PLATO trial. The so-called ‘North American anomaly’ refers to the observation that the outcomes were superior with ticagrelor vs. clopidogrel across the entire international trial, but not in North America. The post hoc analysis suggested that the increased dosage of aspirin >300 mg in the USA in contrast to the other study population was the only factor to explain (out of 37 factors explored) this regional interaction; that ticagrelor was less effective and potentially more harmful vs. clopidogrel.^67,68 Due to its adenosine receptor agonistic properties, ticagrelor it is more vulnerable to attenuation of PGI₂ formation by aspirin, with consequent decreases in AC activation as clopidogrel. Nevertheless this issue still remains an area of debate and fruitful future research.⁶⁹

Drug–drug interactions

Omeprazole, calcium channel blockers, ketoconazole, and rifampicin have all been shown to influence the pharmacodynamic effect of clopidogrel.^43,70–80 Prasugrel and clopidogrel are thienopyridines, and prodrugs that require cytochrome P450 (CYP) enzymes for metabolic activation. It has been shown that drugs competing for metabolism by CYP enzymes alter the generation of prasugrel active metabolite, but only ritonavir has been shown to influence prasugrel antiplatelet effects significantly.⁸¹ In contrast, bioactivation is not necessary to produce an antiplatelet effect for the direct acting reversibly binding P2Y₁₂ receptor inhibitor, ticagrelor. Ticagrelor is metabolized primarily through CYP 3A4 and various drug–drug interactions have been reported for ticagrelor. Particular caution is recommended in concomitance with the use of simvastatin (increase in C_maxby 81%), digoxin (increase C_max of 75%), and CYP3A4 inhibitors (ketoconazole) or inducers (rifampicine).^78,79 Additionally it has to be considered that all medications influencing gut motility (especially morphine derivatives) might impair or delay absorption of P2Y₁₂ receptor inhibitors.^58,82–85

Genetic determinants of the response variability to clopidogrel

The fundamental principle of pharmacogenomics is to use genetic profiling to find the right drug and dose for the right patient.⁸⁶ As a prodrug, the pharmacokinetic and pharmacodynamic properties of clopidogrel are influenced by polymorphisms of genes encoding key CYP enzymes.^87,CYP2C19*2 is a loss-of-function allele associated with reduced generation of the clopidogrel active metabolite and pharmacodynamic effect (Figure 2). It is the most common loss-of-function allele, and is carried by about 30% of Caucasians and 50% of East Asians. CYP2C19*2 is associated with HPR and with increased risk for adverse cardiovascular events.^{43,72,88–95} In a collaborative meta-analysis of nine clinical studies involving 9685 patients (91% treated by PCI and 55% of whom had ACS), carriers of one or two CYP2C19*2 alleles exhibited significantly increased rates of cardiovascular events as compared with non-carriers [hazard ratio (HR): 1.55; 95% confidence interval (CI): 1.10–2.17; P = 0.01 and HR: 1.76; 95% CI: 1.24–2.5; P = 0.002, respectively]. There was an increased risk of stent thrombosis among carriers of one or two CYP2C19 LoF alleles (HR: 2.67; 95% CI: 1.69–4.22; P < 0.0001 and HR: 3.97; 95% CI: 1.75–9.02; P = 0.001, respectively).⁹⁸ On this basis, the U.S. Food and Drug Administration has approved a black box warning on the diminished effectiveness of clopidogrel in genetically predicted poor metabolizers (homozygous carriers of LoF alleles), and encouraging the identification of the genotype and consideration of alternative treatment in these patients.⁹⁷ Other identified CYP2C19 LoF variants (*3–*8) have a relatively low-allele frequency (<1%) in Caucasians.^98–100,CYP2C19*17 is a gain-of-function allele associated with greater activation of clopidogrel. There are conflicting data on the potential association of CYP2C19*17 with haemorrhagic events.^{36,93,94,101,102} Data on the clinical relevance of polymorphisms of other genes involved in the metabolism or action of clopidogrel (e.g. the intestinal efflux transport pump P-glycoprotein pump encoded by the ATP Binding Cassette Subfamily B Member 1 (ABCB1) gene,^{43,98,103–106} the Integrin Subunit Beta 3 (ITGB3) encoding the integrin Beta3 of the glycoprotein IIb/IIIa receptor (GPIIb/IIIa),^43,98 and the P2Y₁₂ receptor^43,98,107 are conflicting. Although the P-glycoprotein efflux and the CYP-P450 system are also involved in the transport and metabolism of prasugrel and ticagrelor, genetic polymorphisms of the responsive genes do not appear to significantly impact metabolism, pharmacodynamic effects or clinical outcomes of ticagrelor- or prasugrel-treated patients.^106,108,109

CYP2C19 genotyping

Until recently, CYP2C19 genotyping was available through a number of laboratories, but long waiting times made these tests unsuitable for use in the clinical practice of acute medicine. Now, tests are available whereby we can get results as fast as we get a troponin level. Currently, two assays fulfil the requirements for rapid diagnosis: the Spartan RX system (Spartan Bioscience Inc, Ottawa, Ontario) and the Verigene XP system (Nanosphere, Northbrook, IL, USA). Both systems provide information on *2, *3, and *17 carriage. The Spartan system provides data within 1 h, whereas the Verigene system requires 3 h.

The genotype is not however a perfect surrogate for the phenotype. Variable platelet reactivity to ADP was demonstrated in PCI patients heterozygous for LoF allele carriage.¹¹⁰ Overall, these patients had significantly higher on-treatment platelet reactivity than the wild type (*1/*1). Homozygous carriage of LoF has been associated with more uniformly higher on-treatment platelet reactivity. Carriers of *2/*17 and *3/*17 have largely unpredictable platelet reactivity (Figure 3).

It is ‘crucial’ to ascertain a gene-dose effect, because CYP2C19*2 heterozygotes represent 20–25% of the general population. Whereas the data argue for a complete resistance profile in CYP2C19*2 homozygotes, CYP2C19*2 heterozygotes display a more intermediate response to clopidogrel, although still significantly worse than that observed in wild-type patients. In the ELEVATE-TIMI 56 trial, tripling the maintenance dose of clopidogrel in patients carrying one LoF allele achieved levels of platelet reactivity similar to that seen with the standard 75-mg dose in non-carriers. In contrast, in patients with two loss-of-function alleles, doses as high as 300 mg daily did not result in comparable degrees of platelet inhibition (Figure 3).¹¹¹

It is thus becoming increasingly more difficult to ignore the potential of genetic testing. The successful validation and clinical application of such a point-of-care genetic test was presented in the RAPID GENE study, in which patients were randomized to undergo bedside testing and those who were identified as carriers of the allele were then treated with prasugrel instead of clopidogrel post-PCI. Seven days post procedure those identified by genetic testing were significantly less likely to have high on-treatment platelet reactivity than patients who did not undergo point-of-care testing.¹¹²

Is the genetically predicted metabolizer status a good marker of treatment response?

The response to this question was presented in the ARCTIC-Gene sub study. Whereas, we would expect that rapid metabolizers are super-responders to clopidogrel, and poor metabolizers should exhibit poor response to clopidogrel, the ARCTIC-Gene sub study showed that there is suboptimal concordance between the genotype and the phenotype, as 60% of poor metabolizers had a normal response to clopidogrel, whereas 32% of rapid metabolizers were poor responders to clopidogrel¹¹³ (Figure 4). In fact, it has been shown that genetic polymorphisms can explain only 12% of clopidogrel response variability,⁹³ perhaps explained by the interplay between many different environmental factors influencing platelet reactivity (comorbidities, cellular factors as well as gene-environment interactions and drug–drug interactions; Figure 1). As so many factors influence platelet reactivity, presumably controlling one or a few is unlikely to make a comprehensive difference as to the risk of thrombosis or bleeding.

Cost effectiveness of genotyping for the choice of the P2Y₁₂ receptor inhibitor

Some physicians argue for the use of prasugrel or ticagrelor in all ACS patients straight away. Nevertheless, cost-effectiveness analyses present some arguments against such an approach. Lala et al.¹¹⁴ developed a Markov model comparing the cost-effectiveness of those strategies (CYP2C19*2 genotype-guided strategy vs. the empiric use of clopidogrel or prasugrel) and captured adverse cardiovascular events as well as antiplatelet-related complications and quality-adjusted life-years (QALYs). The genetic testing-guided strategy yielded the most QALYs and was significantly less costly. Over 15 months, total costs were $18 lower with a gain of 0.004 QALY in the genotype-guided strategy vs. empiric clopidogrel, and $899 lower with a gain of 0.0005 QALY compared with empiric prasugrel prescription. The strongest predictor of the optimal strategy was found to be the relative risk of thrombotic events in carriers compared with wild-type individuals treated with clopidogrel. Even in the sensitivity analysis, testing the willingness to pay against demands of cost effectiveness, the genetic testing strategy had the highest probability of being cost-effective compared with both no-testing strategies at each threshold. A future task will be to team up with the large pharmaceutical groups; as to date, cost effectiveness might not be their primary objective. Precision medicine might not only help us to work cost effective but also help the industry to better define and position their platelet inhibitors.

Phenotyping

Although several test systems are available for monitoring the pharmacodynamic effect of antiplatelet therapy, they are all hampered by two major limitations.¹¹⁵ First, it is not possible to simulate the complexity of haemostasis in vitro, as those systems fail to include other clinical platelet activating factors that might influence outcome.¹¹⁶ This makes the interpretation of the testing result uncertain, because the in vitro observed antiplatelet effect only assures pharmacological efficacy but cannot prove the antiplatelet effect in vivo.¹¹⁷ Second, platelets are sensitive to manipulation and therefore in vitro activation.

LTA used to be the most widespread platelet function test and deemed to be the gold standard of platelet function testing.¹¹⁸ The first study to demonstrate the link between on-treatment platelet reactivity to ADP and thrombotic clinical event occurrence was the PREPARE POST-STENTING study.¹⁰¹ LTA was reported to predict ischemic events with a sensitivity between 60 and 79%, with a specificity of 59–82% and an area-under-the-curve (AUC) of the receiver operating characteristic (ROC) curve of 0.73–0.85%, OR of 3–35.^{1,0,17,21,25,119,120} LTA has also been shown to predict stent thrombosis and bleeding events.^12,14,16,17

Multiple electrode aggregometry (MEA) can effectively predict stent thrombosis [odds ratio (OR) = 9–37; area under the ROC curve: 0.78–0.92, sensitivity: 70–90% and specificity: 84–100%].^11,13,28 Additionally, MEA has been shown to predict major bleeding (ROC AUC: 0.61–0.74, sensitivity: 72–77%, and specificity: 62–66%.^{29,30,121,122}

The VerifyNow^TM P2Y₁₂ assay (Accumetrics, San Diego, CA, USA) has a higher reproducibility compared with other platelet function tests, and is capable of predicting MACE (OR = 1–6.5; ROC AUC: 0.56–0.87, sensitivity: 60–80%, and specificity: 63–92%) and major bleeding events (OR = 0.94; AUC = 0.84, sensitivity: 81%, and specificity: 80%).^36,123,124

The VASP phosphorylation levels (BioCytex, Marseille, France- VASP assay) is a specific assay for P2Y₁₂ signalling, because it solely evaluates the extent of P2Y₁₂-receptor activity. A positive VASP test is associated with an OR = 1.04–11.18^120,125 for stent thrombosis or MACE (area under the ROC curve: 0.55–0.79), with high sensitivity (70–100%),^13,126 but low specificity (25–37%).

For the PFA-100^TM (Dade Behring, Marburg, Germany), there have been less extensive studies. A usefulness of the device for prediction of MACE (OR = 3.2–22.9) in clopidogrel users^127–130 has been reported. A closure time ≤72s had a sensitivity of 86% and specificity of 76% to detect ischemic events.¹³⁰ There are conflicting data concerning the reproducibility of the test.^131–133

Other techniques with little or no evidence of predicting adverse cardiovascular events in antiplatelet therapy are: the Cone and Platelet Analyzer (DiaMed, Cressier, Switzerland),^134,135 Plateletworks (Helena Laboratories, Beaumont, Texas)²⁵ and the thromboelastography (TEG) haemostasis analyser (Haemoscope Corp., Niles, Illinois).^{10,119,136,137}

Studies using phenotyping for personalized antiplatelet treatment

Individualized antiplatelet therapy has been shown to improve clinical outcomes in some small studies. Guided therapy with up to four clopidogrel re-loadings in a group of patients initially not responding to therapy resulted in a reduction of major adverse cardiac events without an increase in major bleeding complications.^138–140 In line with this finding, intensified platelet inhibition with GPIIb/IIIa antagonists lowered the incidence of major adverse cardiac events without increasing bleeding rates.^141,142 In addition, some real-life registries such as the IDEAL-PCI registry¹⁴³ supports the idea that individualization of antiplatelet therapy minimizes early thrombotic events without increasing bleeding risk. In contrast, guided antiplatelet therapy did not improve patient outcomes in the Gauging Responsiveness With A VerifyNow^TMAssay-Impact On Thrombosis And Safety [GRAVITAS] trial (n = 2214).¹⁴⁴ Patients with HPR on clopidogrel treatment following stenting were randomized to 75 mg of clopidogrel or to receive a second clopidogrel loading dose of 600 mg and then a doubled maintenance dose of 150 mg of clopidogrel throughout the study. This strategy showed no difference in event rates during 6-month follow-up in a patient population at low-to-moderate thrombotic risk. The Testing Platelet Reactivity In Patients Undergoing Elective Stent Placement on Clopidogrel to Guide Alternative Therapy With Prasugrel [TRIGGER-PCI] trial, comparing prasugrel and clopidogrel in patients with HPR on clopidogrel after elective drug eluting stent (DES) implantation without procedural complications (low-thrombotic risk), was stopped prematurely for futility (n = 423), because an interim analysis indicated a low incidence of the primary efficacy endpoint. Nevertheless, platelet aggregation data from the TRIGGER-PCI trial demonstrated that HPR could be overcome by switching from clopidogrel to prasugrel.¹⁴⁵ The Double Randomization of a Monitoring Adjusted Antiplatelet Treatment vs. a Common Antiplatelet Treatment for DES Implantation, and Interruption vs. Continuation of Double Antiplatelet Therapy [ARCTIC] trial (n = 2440) randomized patients at low-to-moderate thrombotic risk to bedside platelet function monitoring vs. no monitoring. In the monitoring arm, antiplatelet therapy was intensified in patients with HPR on aspirin or clopidogrel (by increasing the dose of aspirin, or an additional loading dose of clopidogrel or an increased maintenance dose of clopidogrel, or switching to prasugrel, or by additional treatment with GP IIb/IIIa inhibitors). Although platelet reactivity was reduced in the monitoring arm, the therapy adjustment based on platelet function monitoring did not improve the composite endpoint of coronary adverse ischemic events.¹⁴⁶ These results were confirmed in a landmark analysis considering only patients after hospital discharge, excluding all peri-procedural myocardial infarctions.¹⁴⁷

Interestingly, the recently published ANTARCTIC trial (Assessment of a Normal vs. Tailored Dose of Prasugrel After Stenting in Patients Aged > 75 Years to Reduce the Composite of Bleeding, Stent Thrombosis and Ischemic Complications; with a relatively small sample size of 877 patients) suggested that platelet function monitoring with treatment adjustment 14 days after randomization did not improve the clinical outcome of elderly patients treated with coronary stenting for an acute coronary syndrome.¹⁴⁸ In patients with HPR, the 5mg prasugrel dose was increased to 10 mg. In patients with LPR, prasugrel 5 mg was replaced with clopidogrel 75 mg. Platelet function monitoring lead only in 4% of patients to treatment intensification (switch to prasugrel 10 mg) and in 39% of patients to treatment de-intensification, switch from prasugrel to clopidogrel. ANTARCTIC therefore, compared the effect of prasugrel 5 mg to prasugrel 5m/clopidogrel 75mg to show the superiority of the monitoring arm.

Based on these trials, some have concluded that the concept of individualized anti-platelet therapy has failed so far. These are factors which have been proposed to explain why these studies showed neutral results.^149,150

These include:

• the inclusion of patients at low-thrombotic risk (with the exclusion of STEMI patients)^{144,145,151–155}

• the limitations of platelet function testing (with different cut-offs for HPR; the low standardization of some methods, including the lack of standardized timing between pill ingestion and platelet function testing^156,157

• reasons related to the choice of a diagnostic strategy (phenotyping vs. genotyping vs. combined diagnostic strategy);^110,158

• reasons related to the protocol for personalized therapy (only a single switch to another drug or suboptimal dosing);^{140,144,145,153,159}

• reasons related to the definition of the primary endpoint (e.g. inclusion of cardiac biomarker rise in the definition of myocardial infarction in the ARCTIC study vs. the universal myocardial infarction definition in GRAVITAS and TRIGGER-PCI;^{144–146,150}

• reasons related to the time point of participants’ inclusion (randomization performed 12–24 h after PCI in the GRAVITAS trial vs. 2–7 h after the first clopidogrel maintenance dose intake, the day after successful PCI in the TRIGGER-PCI trial, excluding patients experiencing peri-procedural events, early after PCI or those with unsuccessful or complicated PCI procedures, therapy adjustment 14 days after randomization;¹⁴⁸

• reasons related to the statistical power of the studies (the incidence of the primary endpoint was lower than expected in TRIGGER-PCI: 0.4%,¹⁴⁵ and GRAVITAS: 2.3%.¹⁴⁴).

In this context, it must be emphasized that neither the phenotypic nor the genotypic approach alone produces solid evidence of efficacy. However, this lack of evidence is not evidence against the idea of individualized anti-platelet therapy. It is merely the assignment to us to further define clinical situations in which this approach might be beneficial. Especially, when concerning the antithrombotic therapy, the mechanistic point of view that the majority of study population (e.g. ACS patients) should benefit from the given intervention is delusive. For this purpose, we have to refine this approach and collect as much data on cofounders as possible to assemble this puzzle.

Precision medicine-implications for antiplatelet treatment

Antiplatelet treatment is the foundation of optimal medical treatment (OMT) in the prevention of secondary ischaemic events in patients with coronary artery disease (CAD). All recommendations for OMT with antiplatelet agents rely on guidelines that are based on the examination of current evidence within the concept of evidence-based medicine.¹⁶⁰ Like most medical treatments, antiplatelet therapy is associated with heterogeneity in benefit, but current treatment algorithms do not distinguish between patients with high and low potential to gain benefit from the treatment. Not only is this approach not patient-centred; it is also intrinsically inefficient in terms of cost, safety, and outcomes.¹⁶¹ Recent developments have led to a deeper understanding of risk based on genetics and objective assessment of platelet biology, opening a new field termed precision or personalized medicine. The idea behind this is the combination of clinical–pathological indices and patient’s ‘panomic’ data (including pharmacodynamic, genomic, epigenomic, transcriptomic, and metabolomics information) to create diagnostic, prognostic, and therapeutic strategies precisely tailored to individual patients.^162,163

Reactive vs. efficient medical treatments

According to a traditional treatment paradigm (‘reactive medicine’), we treat certain diseases according to our knowledge and experience, ideally adhering to guidelines. The antiplatelet agent so prescribed may not improve the patient’s condition or may be ineffective, causing an increase in disease severity. In that case, we react to this with a change of the antiplatelet agent, which might be ineffective again. This process of switching ineffective antiplatelet drugs is the centrepiece of reactive medicine. In fact this approach is intrinsically inefficient in terms of cost, safety, and outcomes. Nonetheless, we still prescribe antiplatelet agents to our patients in a reactive pattern, as shown in the example diagram describing the old paradigm of treatment with such drugs (Figure 5). In this paradigm, we first start from diagnosing patients’ CAD, and often treat it with PCI with stenting. Based on the diagnosis stable CAD, we prescribe clopidogrel. Since some patients receive a suboptimal treatment by being treated with an ineffective drug, the disease may progress and manifest itself, for example as a myocardial infarction. Patients are then switched to other drugs (prasugrel or ticagrelor), which might cause serious side effects in some patients (e.g. severe bleeding). Unfortunately, this whole process can be repeated several times, with a great deal of time loss and disease recurrences climbing to an unnecessary magnitude. Figure 6 shows a new paradigm, whereby a precision medicine-based methodology is introduced into the medical care process. The key point of avoiding the unnecessary process of first prescribing an ineffective antiplatelet drug is to deliver the right drug from the very beginning to individual patients based upon several characteristics (included in an algorithm or score), so that the repeated switching of antiplatelet agents is avoided and the efficiency and cost-effectiveness is greatly improved. One major goal would be to integrate pharmacogenomic information in a viable therapeutic algorithm to guide the dosing and type of antiplatelet therapy. Such an algorithm for personalized antiplatelet therapy in patients who are at high-thrombotic risk has been recently presented. This global risk algorithm is based on clinical (PREDICT score), biological (platelet function), and genetic (CYP2C19*2 carrier status) information.¹⁴⁹ It is important however to emphasize, that this algorithm is empirical, and has not yet been tested in clinical trials. Nevertheless, it provides an idea of how to integrate clinical, biological, and genetic information within the decision-making of antiplatelet therapy, and might be the closest we can get to precision medicine as of now in this field.

Unfortunately, the main obstacle in conducting studies in the field of precision medicine is lack of funding. Due to the objective of cost effectiveness, industry might not be the main point contact for such studies and only national or international not-for-profit agencies like the national institute of health or the European Research Council might be the only partners for these investigations.

Conclusion

Cardiologists have lagged behind our oncology colleagues in their acceptance of the concept of tailoring therapies to individual patients with the goal of providing the greatest benefit. We may have a greater willingness as in other disciplines to integrate information in patients who have specific molecular signatures into algorithms that afford therapies optimizing the benefit and the risk. Through this combined approach, a tailored antiplatelet treatment strategy based upon phenotypic and genotypic data has the potential to enhance outcomes. The randomized trials were underpowered or dampened by other limitations, and showed at least neutral results. Due to that, current guidelines do not recommend integrating genetic or phenotypic information into a clinical practice.^3,164 However, the results from these trials have been used to refute the utility of personalized antiplatelet therapy. Although the biologic underpinnings and observational data supporting personalization are robust, more evidence is needed to recommend personalization as standard of care.

Conflict of interest: E.G.: speaker honoraria from AstraZeneca, Baxter, Bayer, Boehringer Ingelheim, Bristol-Myers Squibb and Pfizer and has participated in advisory board meetings for AstraZeneca, Bayer, Boehringer Ingelheim and Bristol-Myers Squibb. R.De C.: grants and consultancy fees from Boehringer-Ingelheim, Bayer, BMS/Pfizer, Daiichi-Sankyo, Merck, Novartis, Lilly. E.P.N.: Speaker fees from Ely Lilli and fees for scientific activities with Astra Zeneca. J. M.S-M.: Grants and consultancy fees from Astra Zeneca, Eli Lilly, Daiichi Sankyo, Bayer, Roche.

References