Platelet activation and aggregation are central to the development of thrombotic complications during acute coronary syndromes (ACS) and following percutaneous coronary interventions (PCIs). Adenosine diphosphate (ADP) and thromboxane (TX) A2 are major secondary agonists released by platelets following activation.1 These secondary agonists play an important role in the amplification of platelet activation and aggregation and stable thrombus generation at the site of plaque rupture.
Simultaneous inhibition of the P2Y12 receptor by thienopyridines and TxA2 generation by aspirin is an effective antiplatelet treatment strategy to inhibit platelet function. Landmark clinical trials have demonstrated the central role of dual antiplatelet therapy in the treatment of ACS and in the prevention of complications during and after PCI.2–4 In recent years, despite the well-documented clinical efficacy of clopidogrel and aspirin treatment, the phenomenon of antiplatelet resistance or non-responsiveness has been repeatedly reported by various investigators and has been correlated in small studies with the occurrence of ischemic events.5–8
Definition of Antiplatelet Resistance
Platelet activation and aggregation are processes mediated by various receptor signaling pathways. A single treatment strategy directed against a specific receptor cannot overcome all thrombotic complications, and treatment failure following a single antiplatelet agent is not synonymous with drug resistance. It is our position that the optimal definition of resistance or non-responsiveness to an antiplatelet agent is the failure of the antiplatelet agent to inhibit the target of its action. The identification of resistance would therefore utilize a laboratory technique that detects residual activity of the target. In the case of clopidogrel resistance, there would be significant evidence of residual post-treatment P2Y12 activity and in the case of aspirin there would be residual post-treatment cyclo-oxygenase (COX)-1 activity.9
Clopidogrel—Mechanism of Action
Clopidogrel is rapidly absorbed from the intestine and converted by the hepatic cytochrome P450 isoenzymes to an active thiol metabolite.10,11 This short-lived active metabolite binds to the P2Y12 receptor specifically and irreversibly via a disulfide bridge between the reactive thiol group and two cysteine residues (cys17 and cys270) present in the extracellular domains of the P2Y12 receptor.12 Clopidogrel has also been reported to attenuate platelet–leukocyte aggregate formation, the level of C-reactive protein (CRP), P-selectin, and CD40L, and the rate of thrombin formation.9
Laboratory Evaluation of Clopidogrel Responsiveness
Since clopidogrel specifically inhibits the ADP receptor on platelets, ex vivo measurement of ADP-induced maximum platelet aggregation by light transmittance aggregometry (LTA) has been the most commonly used laboratory method to evaluate clopidogrel responsiveness and is considered the gold standard.13 Recently, it was suggested that the response to clopidogrel would be better demonstrated by measuring late platelet aggregation at six minutes after stimulation with ADP rather than maximum aggregation.14 However, a recent study from our laboratory based on the evaluation of both maximum and final aggregation from 100 consecutive patients undergoing stenting and treated with clopidogrel indicated that the two measurements were equivalent in determining the prevalence of clopidogrel non-responsiveness.15 Flow cytometric measurements of the expression of activated glycoprotein (GP) IIb/IIIa receptor and P-selectin expression after ADP stimulation can also identify clopidogrel non-responsiveness.13,16 In addition, measurements of ADP-induced platelet–fibrin clot strength by whole blood thrombelastography and the VerifyNow P2Y12 receptor assay using ADP as the agonist can also be used to measure clopidogrel responsiveness as point-of-care assays.17,18 The PFA-100 method using collagen–ADP-based cartridges and whole blood aggregometry are associated with inconsistent estimates of platelet reactivity to ADP. The phosphorylation state of vasodilator-stimulated phosphoprotein is a specific intracellular marker of residual P2Y12 receptor reactivity in patients treated with clopidogrel and can be measured by flow cytometry19 (see Figure 1).
Clopidogrel Responsiveness—Effect of Time of Treatment and Dose
Response variability and non-responsiveness to clopidogrel have been demonstrated mainly in patients undergoing coronary stenting.13,20 Clopidogrel responsiveness has been demonstrated to be dependent on the post-treatment assessment time and dose. In an early investigation, clopidogrel responsiveness was studied in patients undergoing stenting treated with a 300mg loading dose followed by 75mg per day maintenance dose. ADP-induced platelet aggregation and activation-dependent platelet surface marker expression (P-selectin and activated GPIIb/IIIa) were assessed at baseline and serially for 30 days following stenting. Response variability to clopidogrel was demonstrated, as measured by all markers, and a certain percentage of patients were found to have no demonstrable antiplatelet effect, where the difference between pre- and post-treatment ADP-induced platelet aggregation was ≤10%. We defined these patients as ‘clopidogrel-resistant’ or ‘non-responsive to clopidogrel therapy.’13
In the latter study, 63% of patients were resistant to clopidogrel treatment at two hours post-stenting, ~30% were resistant at day one and day five post-stenting, and 15% were resistant at day 30 post-stenting13 (see Figure 2). Therefore, clopidogrel ‘resistance’ in this study appeared to be time-dependent. Based on the results, we hypothesized that the occurrence of clopidogrel resistance might be related to the inadequacy of a 300mg loading dose to provide sufficient active metabolite generation to arrest platelet reactivity in selected patients, and that these resistant patients may be at particular risk for thrombotic complications.13
Since this initial description, the phenomenon of clopidogrel resistance has been confirmed by multiple investigators.20–28 The prevalence of clopidogrel non-responsiveness has been reported at 5–44%. As demonstrated in Table 1, the data are markedly concordant. This wide variation in prevalence may be primarily due to dosing; higher resistance estimates are present following the 300mg loading dose and lower estimates occur after the 600mg loading dose.20–28 In the pharmacodynamic study comparing 300 and 600mg clopidogrel loading doses, treatment with a 600mg loading dose during elective PCI reduced clopidogrel non-responsiveness to 8% compared with 28–32% after a 300mg loading dose (see Figure 2).
Moreover, the study demonstrated a narrower response profile following treatment with 600mg compared with 300mg clopidogrel.20 Similar increased responsiveness was also observed in the ISAR-CHOICE study, where a ceiling effect of platelet inhibition was observed with a 600mg clopidogrel loading dose whereas a non-significant increase in platelet inhibition was observed with a 900mg loading dose.29 In the ALBION study, there was a non-significant improvement in platelet inhibition following a 900mg loading dose compared with a 600mg dose.30 Two small studies have suggested that a 150mg maintenance dose is associated with lower post-treatment platelet function compared with a 75mg maintenance dose.31,32
The cumulative experience thus far demonstrates that clopidogrel treatment is associated with suboptimal inhibition as measured by ex vivo methods in a significant percentage of patients undergoing PCI. Post-treatment platelet aggregation is high in selected patients, there is wide response variability to therapy leading to an unpredictable antiplatelet effect, and a substantial number of patients exhibit either complete non-responsiveness or heightened platelet reactivity to ADP post-PCI. These non-responsive patients may be at particular risk for ischemic complications post-PCI.
Mechanism of Clopidogrel Resistance
Clopidogrel non-responsiveness is primarily a pharmacokinetic problem associated with insufficient active metabolite generation. Insufficient active metabolite generation may be secondary to limitations in intestinal absorption and functional and genetic variability in the hepatic cytochrome (CYP) P450 isoenzymes.33–35
First, the repeated demonstrations that a high loading dose of 600mg clopidogrel is associated with increased inhibition of ex vivo ADP-induced platelet aggregation in patients undergoing PCI and a decreased prevalence of non-responders supports insufficient active metabolite generation as a major factor in clopidogrel resistance.20,29,30 Second, recent studies involving the measurement of hepatic cytochrome (CYP) P450 activity suggest that individual variations in the activity of this enzyme play a major role.10,33–35 In a landmark investigation, Lau et al. demonstrated that pharmacological stimulation of CYP3A4 activity by rifampin or St Jonh’s Wort enhanced the inhibitory effect of clopidogrel, whereas agents that competed with clopidogrel for CYP 3A4 (e.g. erythromycin) attenuated the antiplatelet effect of clopidogrel.10,33,34 Lau et al. also demonstrated that lipophilic statins that compete with clopidogrel for CYP3A4 may also influence the antiplatelet effects of clopidogrel.33 Third, ketoconazole, a CYP3A4 inhibitor, did not have any effect on prasugrel active metabolite generation or prasugrel-induced platelet inhibition during co-administration. However, clopidogrel-induced platelet inhibition was reduced following a loading dose as well as after a maintenance dose in healthy volunteers during co-administration with ketoconazole.35 The latter effect on clopidogrel-induced platelet inhibition was accompanied by less active metabolite generation. This study further demonstrated the prominent role of CYP3A4 activity in clopidogrel response variability and non-responsiveness. Finally, in another study, prasugrel treatment was associated with superior active metabolite generation and platelet inhibition together with a lower incidence of non-responsiveness compared with clopidogrel treatment.36
Similar to statins, various proton pump inhibitors (PPIs) frequently used with clopidogrel and aspirin in the treatment of cardiovascular disease patients are also metabolized mainly by CYP3A4 and CYP2C19 isoenzymes. Therefore, recent studies focused on the relation of PPI treatment to clopidogrel metabolism and cardiovascular outcomes. In a randomized controlled study involving patients undergoing stenting, co-administration of omeprazole with aspirin and clopidogrel significantly reduced the antiplatelet effect of clopidogrel as measured by the VASP phosphorylation assay assay.37
Notably, in a retrospective analysis of 5,512 patients, one-year acute myocardial infarction rates were significantly higher in patients who received low-dose or high-dose PPIs in addition to clopidogrel compared with those receiving clopidogrel alone (3.08 or 5.03 versus 1.38%).38 Similarly, in another one-year retrospective cohort study investigating patients (n=16,690) who underwent stenting, there was a 50% increase in the incidence of a major cardiovascular event in patients receiving a PPI and clopidogrel compared with those receiving clopidogrel alone.39 These studies suggest that PPI exposure during clopidogrel treatment may reduce the antiplatelet effects of clopidogrel and impair its ability to prevent acute ischemic events.
It has been suggested that cigarette smoking induces CYP1A2 activity, thereby potentially affecting clopidogrel metabolism. A recent retrospective analysis of patients undergoing elective stenting treated with clopidogrel and aspirin reported that current smokers had greater platelet inhibition and a lower aggregation than non-smokers. This effect was more pronounced in patients who smoked more than a half-pack per day.40
Recent studies have focused on the influence of genetic polymorphisms on clopidogrel response variability, especially genes encoding CYP isoenzymes and the p-glycoprotein transporter. Among the genetic polymorphisms, mainly the common loss-of-function polymorphism of CYP2C19 was associated with decreased clopidogrel active metabolite exposure and less platelet inhibition. It was also demonstrated in a recent study that the 2C19*2 allele is associated with high on-treatment platelet reactivity and also long-term ischemic event occurrence.41–43
Finally, these data strongly indicate that clopidogrel non-responsiveness or resistance is an important phenomenon, and its relevance to clinical outcome is being explored in recent studies.
Aspirin is the most economical and effective antiplatelet drug prescribed for the treatment of cardiovascular and cerebrovascular diseases. The antiplatelet effect of aspirin is primarily due to irreversible acetylation of a serine residue (Ser530) in COX-1 in platelets that prevents the binding of arachidonic acid to the catalytic site.44 Subsequent generation of TxA2 and TxA2-induced platelet aggregation are inhibited for the life of the platelets.
Laboratory Evaluation of Aspirin Resistance (see Figure 3)
A reliable and specific laboratory method to identify aspirin resistance has not yet been uniformly accepted by investigators. Aspirin resistance has been regarded as the failure of aspirin to inhibit COX-1, the failure of aspirin to inhibit platelet function after stimulation by various agonists (‘biochemical aspirin resistance’), or the failure to protect against cardiovascular events (‘clinical aspirin resistance’ or, more specifically, ‘treatment failure’). Based on different ex vivo methods and criteria to define aspirin resistance, various studies have reported a wide variability in the occurrence of aspirin resistance (<1–55%)45–55 (see Table 2). Laboratory methods, including point-of-care methods using agonists such as ADP-, epinephrine-, or collagen-induced platelet aggregation, do not solely indicate COX-1 activity and thus are fundamentally flawed in specifically measuring the platelet response to aspirin (see Figure 2). More specific methods that indicate residual COX-1 activity include arachidonic acid-induced platelet aggregation and, more precisely, the measurement of the stable metabolite of TxA2 in serum during aspirin therapy.
Aspirin Resistance Is Assay-specific
Platelet aspirin resistance was found to be uncommon in compliant patients treated with high-dose aspirin (325mg per day) when measured by COX-1 specific assays such as using AA-induced light transmittance aggregometry (LTA) and thrombelastography (TEG).55 In another study of patients with a history of myocardial infarction, 9% were found to be non-compliant with aspirin therapy and only one patient was resistant to 325mg aspirin therapy as measured by AA-induced LTA.56 In a recent prospective, randomized, double-blind, double-cross-over investigation studying aspirin dosing using multiple COX-1-specific and -non-specific assays in patients with stable coronary artery disease (n=120), it was found that aspirin resistance was rare (1–6%) using COX-1-specific methods (AA-induced LTA, VerifyNow Aspirin assay with AA cartridges, and AA-induced aggregation measured by TEG) at all doses of aspirin compared with COX-1-non-specific methods (1–27%) (see Figure 4). Moreover, a dose-dependent platelet response to aspirin treatment was observed using methods where collagen was used as an agonist, such as collagen-induced LTA and PFA-100. The latter occurred in the presence of near complete inhibition of COX-1 enzyme activity as measured by AA-induced platelet aggregation, indicating that aspirin may have non-COX-1-mediated dose-dependent effects in platelets.57 More importantly, these dose-dependent effects were more pronounced in patients with diabetes, where treatment with higher doses was found to be effective in inhibiting platelet function.58
Similar results reporting a low prevalence of aspirin resistance using COX-1-specific assays have been reported.59 At the same time, further studies have found a correlation between adverse clinical events and aspirin resistance, as measured by PFA-100.60,61 Thus, taken together all of these studies indicate that non-compliance, various assessment methods, and underdosing may all be important factors responsible for the reported variability in aspirin resistance estimates in clinical studies.
It has also been demonstrated in recent studies that aspirin resistance may be associated with clopidogrel resistance. Moreover, aspirin-resistant patients have been shown to exhibit high platelet reactivity induced by various agonists such as collagen and ADP in addition to AA.28,62 Therefore, aspirin resistance as measured in these studies may mark a platelet reactivity phenotype indicative of high risk for ischemic events.
In addition to the mechanisms described above to explain the mechanism of antiplatelet resistance, the role of other cardiovascular risk factors that are known to augment platelet reactivity, such as diabetes, hyperlipidemia, and hyperglycemia, and non-adherence to the therapy—especially observed with aspirin—may affect antiplatelet responsiveness, the prevalence of antiplatelet resistance, and, ultimately, adverse clinical outcome.63,64
In conclusion, non-responsiveness to clopidogrel and aspirin is a reality. Very preliminary data from small studies suggest that high ex vivo platelet reactivity to ADP, incomplete P2Y12 receptor inhibition (clopidogrel non-responsiveness), and aspirin resistance are risk factors for post-stenting ischemic events including stent thrombosis. Aspirin resistance appears uncommon using assays that isolate the extent of COX-1 inhibition, whereas clopidogrel resistance is comparatively common early post-stenting. However, further work is required to solidify the link between high platelet reactivity/poor inhibition and the occurrence of adverse clinical events. Future studies should focus on the mechanism of antiplatelet resistance and its relation to clinical events. Standardized methods to define aspirin and clopidogrel resistance should be established. A focus on the development of simple, reproducible, and user-friendly point-of-care methods to determine aspirin and clopidogrel responsiveness may also assist in tailoring antiplatelet therapy to the individual patient. Treatment strategies to overcome non-responsiveness may include increased dosage of existing drugs or the use of newly developed more potent inhibitors. Finally, investigations that uniformly determine platelet reactivity following coronary stenting and employ treatment strategies to improve inhibition in patients with high post-treatment platelet reactivity are necessary to determine whether risk can be truly reduced by these measures.