A contemporary viewpoint on ‘aspirin resistance’

Abstract

Aspirin is an irreversible inhibitor of platelet prostaglandin synthase activity, and is the most widely prescribed drug for the secondary prevention of cardiovascular disease. In recent years, clinical and laboratory evidence has shown significant individual variability in the response to aspirin and its link to clinical outcome. The term ‘aspirin resistance’ has been introduced to describe situations when clinical or ex-vivo effects of aspirin are less than expected. The accumulating evidence of increased risk of major adverse clinical events (MACE) associated with ‘aspirin resistance’ in the settings of acute coronary syndrome (ACS), stroke, and peripheral arterial disease has stimulated the search for ways of overcoming aspirin resistance. Existence of the link between high on-treatment platelet reactivity and atherothrombotic events suggests the common mechanisms for atherosclerosis progression and thrombotic complications with the platelets, being a key cellular interface between coagulation and inflammation.

This review article provides a contemporary view on ‘aspirin resistance’ and discusses its definition, clinical importance, and possible mechanisms in light of recent data on the role of platelets in atherothrombosis.

Key words::

• Aspirin resistance
• atherothrombosis
• cardiovascular events
• inflammation
• platelet reactivity

Key messages

• Existence of the link between high on-treatment platelet reactivity and atherothrombotic events suggests the common mechanisms for atherosclerosis progression and thrombotic complications with the platelets, being a key cellular interface between coagulation and inflammation.

Introduction

Platelets have a vital homeostatic function in the prevention of excessive blood loss following injury by the facilitation of haemostasis. However, variation in platelet response to physiological or pathological stimuli may predispose to an increased risk of pathological bleeding or thrombosis.

Current therapeutic strategies for the prevention and treatment of arterial thrombosis are based on coagulation glycoproteins and on the known enzyme and receptor systems of platelets that are involved in the activation of these cells. Three broad groups of antiplatelet agents have been proven to be effective in a clinical setting, as follows: 1) cyclo-oxygenase-1 (COX-1) inhibitors (such as aspirin); 2) adenosine 5-diphosphate (ADP) receptor antagonists, irreversible (ticlopidine, clopidogrel, prasugrel) and reversible (ticagrelor); and 3) glycoprotein IIb/IIIa antagonists (abciximab, eptifibatide, tirofiban). Although aspirin, thienopyridines (ticlopidine, clopidogrel, prasugrel), and two-cyclopentyl-triazolo-pyrimidines (ticagrelor) inhibit a single pathway of platelet activation, their pronounced antithrombotic effect is explained by the fact that both thromboxane A₂ (TxA₂) pathway and ADP pathway are necessary for the amplification of platelet activation which is essential for the full platelet aggregation.

However, there is growing evidence that not all individuals respond equally to antiplatelet agents, a feature which has led to the introduction of the terms of aspirin and clopidogrel ‘resistance’ (1). Given that ‘resistance’ to antiplatelet drugs can be associated with increased risk of recurrent vascular events, the ‘problem’ of ‘variable platelet response to antiplatelet therapy’ attracts the attention of many investigators.

The aim of this review article is to provide a contemporary overview on concepts and controversies related to a ‘variable platelet response to antiplatelet therapy’.

Search strategy

We performed a comprehensive literature search by using electronic bibliographic databases (i.e. MEDLINE, EMBASE, DARE, Cochrane Database), scanning reference lists from included articles and hand-searching abstracts from national and international cardiovascular meetings. For the search, we had used the search terms ‘aspirin’, ‘acetylsalicylic acid’, ‘resistance’, ‘platelet’, ‘reactivity’, ‘atherothrombosis’, ‘atherosclerosis’, and ‘inflammation’.

Where necessary, study authors were contacted to obtain further data. We have essentially performed a semi-systematic review of the published literature, rather than conduct a formal Cochrane-style systematic review and critical appraisal.

Terminology

The term ‘aspirin resistance’ has been used to describe different manifestations of insufficient inhibitory effect of aspirin on platelets, as reflected by either the inability of aspirin to protect individuals from thrombotic complications (that is, clinically defined ‘aspirin resistance’) or persistently high platelet reactivity in laboratory tests, such as bleeding time, TxA₂ production, light aggregometry, etc. (that is, laboratory-defined ‘aspirin resistance’) (1). A pharmacological definition of ‘aspirin resistance’ focuses on the inability of aspirin to inhibit COX-1-dependent TxA₂ production and, consequently, the TxA₂-dependent pathway of platelet activation (2). Resistance to clopidogrel (or the other thienopyridines) is the failure to inhibit the platelet P2Y₁₂ receptor and, consequently, the P2Y₁₂-dependent pathway of platelet activation.

Some authors distinguish between pharmacodynamic ‘aspirin resistance’ (because of changes in the target enzyme for aspirin, COX-1, or because of transient inaccessibility of the enzyme due to blockade of the active site by non-steroidal anti-inflammatory agents) and pharmacokinetic ‘aspirin resistance’ (due to limited availability of the active drug at the level of its target). The in-vitro addition of aspirin to the blood sample would not significantly change the aggregation or thromboxane B₂ (TxB₂) levels in the first case and should significantly reduce the aggregation and TxB₂ concentration in the second scenario (3–5).

Laboratory-defined ‘aspirin resistance’ or aspirin non- responsiveness is present when in-vitro platelet reactivity is not properly blocked despite the use of aspirin (5–7). Despite attempts to dichotomize patients as either ‘responders’ or ‘non- responders’ using arbitrary cut-off points in platelet function tests, on-treatment platelet reactivity as a measure of effectiveness of antiplatelet therapy is more likely to have a normal, bell-shaped distribution, as many other biological processes (8–11).

‘High on-treatment residual platelet reactivity’ (2), found in vitro in patients on antiplatelet therapy, does not necessarily mean ‘resistance’ to the antiplatelet drug from a pharmacological point of view. The recommendation of the Working Group on antiplatelet drugs resistance (5) is to reserve the term ‘laboratory aspirin resistance’ to situations corresponding to the pharmacodynamic resistance, when an acetyl-salicylic acid (ASA)-specific laboratory test was used (platelet aggregation induced by arachidonic acid (AA) or the measurement of serum level of TxB₂) and in-vitro addition of ASA performed to exclude pharmacokinetic resistance. The results obtained with less specific methods used for monitoring the effects of aspirin reflect the reactivity of platelets as a whole and can only be partly attributed to the TxA₂ pathway of platelet activation. In this situation, the term ‘high on-treatment residual platelet reactivity’ would be more appropriate to describe insufficient inhibition of platelet function by aspirin.

Clinical ‘aspirin resistance’ or ‘treatment failure’ is the occurrence of thrombotic events in patients treated with antiplatelet agents (2,5,6). However, as thrombogenesis is a complex multifactorial process, even complete inhibition of a single pathway may not prevent recurrence of thrombosis in all cases. This limits the practical value of the clinical definition of ‘aspirin resistance’.

Assessment of platelet response to aspirin

The prevalence of ‘aspirin resistance’ reported in different studies varies significantly and grossly depends on the laboratory techniques used for its diagnosis. Unfortunately pharmacokinetic studies on aspirin are of limited use, as aspirin is rapidly absorbed from the stomach and acetylates platelet COX-1 in the pre-systemic circulation. Furthermore, low doses of aspirin, administered slowly, may reduce TxB₂ formation, even being hardly detected in the systemic circulation (12).

Aspirin efficacy is perhaps better determined by assays directly dependent upon COX-1 function, such as TxB₂ generation in serum and AA-induced platelet aggregation. Serum TxB₂ (a stable metabolite of TxA₂) reflects the total capacity of platelets to synthesize TxA₂ ex vivo in response to physical and chemical stimuli and can be estimated by measurement of TxB₂ in blood clotted at 37°C for 30–60 min.

The capacity of platelets to generate TxA₂ is approximately 1,000-fold greater than endogenous plasma level. Normal concentrations of circulating TxB₂ are very low, 1–2 pg/mL, and thus plasma samples must be purified and concentrated prior to analysis. Consequently, plasma assays are generally less sensitive and less specific for the estimation of the effect of aspirin. Also, stable levels of thromboxane production in vivo are maintained by different mechanisms, and reduction of thromboxane levels is usually only seen when > 95% inhibition of thromboxane generation is achieved in ex-vivo tests (13). However, even the adequate inhibition of aggregation may require 99% inhibition of TxA₂ production by platelets (14), and thus even a minimal residual capacity to generate TxA₂ may be sufficient to sustain TxA₂- dependent platelet activation (13). Low concentrations of TxA₂ (or other bioactive eicosanoids acting through TxA₂ receptors) from aspirin-treated stimulated platelets and from monocytes/macrophages may still activate the tyrosine-kinase-based signalling pathway. This pathway cannot trigger full platelet activation on its own but can synergize with the G_z-linked adrenaline receptor and thus activate platelets (15).

Indeed, in patients with stable coronary artery disease (CAD) prominent activity of platelet alpha-2 adrenoceptors was shown to be a considerable factor in high residual platelet reactivity (16). For example, Williams et al. (17) demonstrated that patients with acute coronary syndrome (ACS) have increased platelet sensitivity to adrenergic stimuli despite the effective suppression of platelet release response by aspirin. This may lead to false classification of some patients as ‘aspirin-resistant’ depending upon the particular assay used.

In healthy individuals and patients with CAD treated with aspirin, the rate of residual significant TxB₂ production is extremely low (18–20). Only about 2% or less of patients with CAD had evidence of incomplete inhibition of TxB₂ production, which was either because of under-dosing or non-compliance (18,19,21).

Levels of the urinary thromboxane metabolite—11-dehydro TxB₂—reflects time-integrated in-vivo TxA₂ biosynthesis (22,23). It is less specific for TxA₂, generated by platelet COX-1, as about 30% of the urinary metabolite is derived from extra-platelet sources or thromboxane generated by COX-2. The latter may also be present in platelets and megakaryocytes. Additionally, in conditions associated with inflammation or increased platelet turn-over, the contribution of extra-platelet sources of TxA₂ may even be higher (24–27).

The other most often used tests of platelet function, dependent on COX-1 activity, include agonist-induced platelet aggregation measured by light transmission (turbidimetric) aggregometry (LTA) in platelet-rich plasma (PRP), electrical impedance (whole blood) platelet aggregometry, semi-automated (platelet function analyser (PFA-100-Siemens Healthcare Diagnostics, Deerfield, IL, USA)) and automated (Ultegra rapid platelet function assay (RPFA) - Accumetrics, San Diego, CA, VerifyNow - Accumetrics, San Diego, CA) platelet aggregometry, and impedance aggregometry (Multiple platelet function analyzer (Multiplate® analyzer - Verum Diagnostica GmbH, Munich,Germany)) (Table I). Several agonists of varying concentration have been employed to stimulate platelet aggregation. However, the degree of involvement of COX-1- dependent pathway of platelet activation is not the same for different agonists. AA is the substrate for platelet COX-1, and aggregation response to AA-stimulation directly reflects platelet COX-1 activity. Activation-dependent changes in surface expression of P-selectin, CD63, activated GP IIb/IIIa receptor, and leucocyte-platelet aggregates in response to AA by flow cytometry may also be used to determine platelet inhibition by aspirin (28–32).

Table I. Platelet function tests commonly used for measurement of antiplatelet effect of aspirin.

Laboratory method	Principle of the method	Remarks
Aggregometry
Turbidimetric light transmittance aggregometry (LTA)	Analysis of light transmission in anti-coagulated platelet-rich plasma after stimulation of platelets with agonists (AA, ADP, epinephrine, collagen)	Lack of standardization. Low reproducibility. Depending on the type and concentration of agonist used, the aggregation response only partially and variably modulated by TxA2, synthesized from AA in the platelet membrane, or ADP, released from platelet granules
Impedance aggregometry	Measurement of electrical impedance between two platinum electrodes immersed in a sample of whole blood after platelets form aggregates on them following addition of an agonist	Drawbacks listed for LTA are also applicable to this technique. Few studies using this assay have been performed in patients on aspirin therapy
Point-of-care assays
Platelet function analyzer (PFA-100)	Whole blood assay. Assessment of closure time—in-vitro cessation of high shear blood flow through a 150-µm aperture in a membrane coated with collagen plus ADP or collagen plus epinephrine by platelet plug	Assay is sensitive to many variables including platelet count, red blood cells, platelet reactivity to collagen, plasma vWF, the effects of which on platelet aggregate formation can easily outweigh the effects of mild inhibition of platelet function
Rapid platelet function assay (Ultegra RPFA-ASA, VerifyNow Aspirin)	Whole blood assay. Measurement of the agglutination of fibrinogen-coated beads by platelets stimulated by agonist (AA, propyl gallate, ADP, TRAP cartridges)	Diagnostic criterion for aspirin resistance is based on a cut-off established by comparison with optical platelet aggregation in response to epinephrine after a single 325-mg dose of aspirin. Although sensitive to antiplatelet effect of aspirin, epinephrine-induced platelet aggregation is not highly specific for the pharmacological effect of aspirin
Multiple platelet function analyzer (Multiplate®analyzer)	Whole blood assay. Measurement of electrical impedance between two pairs of silver-coated sensor wires incorporated in a test cell, after platelets stick on them following addition of an agonist (ADP, ADP + prostaglandin E1, AA, TRAP-6, collagen, ristocetin)	Majority of studies using this assay have been performed in patients on both aspirin and clopidogrel therapy and showed that this assay can be used for predicting early stent thrombosis in patients after PCI
Other assays
Thromboelastography (TEG)	Whole blood assay. Measurement of viscoelastic changes that occur during clot formation in whole blood	It is influenced by platelet function, coagulation factors, natural inhibitors of coagulation and fibrinolysis. Largely platelet-independent unless platelet agonists used. Needs to be validated in clinical settings

Comparison of different laboratory methods in assessment of ‘aspirin resistance’ usually shows moderate reproducibility and very weak mutual correlations between them (33–36). This reflects their sensitivity to different pathways of platelet activation, both TxA₂-dependent and independent. Accordingly, the reported prevalence of ‘aspirin resistance’ varies widely (e.g. from < 1% to 61%) (37). A systematic review of 42 studies of aspirin use for secondary prevention showed a mean prevalence of ‘aspirin resistance’ of 24% (95% CI 20%–28%) (38). However, studies measuring platelet aggregation using AA-stimulated LTA reported a pooled unadjusted prevalence of ‘aspirin resistance’ of only 6% (95% CI 1%–12%). In studies using point-of-care platelet function-analysing devices, the unadjusted prevalence was significantly higher, being 26% (95% CI 21%–31%), which, to a great extent, can be explained by lower specificity of the point-of-care tests used for COX-1.

In a study that employed several testing approaches simultaneously, the prevalence of ‘aspirin resistance’ was relatively low with the use of more specific tests—for example, AA-induced LTA (4%) and VerifyNow Aspirin (6.7%)—and much higher with non-specific tests, ADP-induced platelet aggregation (51.7%) and PFA-100 (59.5%) (33).

Of note, interindividual variability of base-line platelet reactivity could affect the degree of platelet inhibition by aspirin (39). This is particularly important when tests that are less specific to COX-1 pathway are used to assess platelet response to aspirin (40). Factorial analysis has revealed the presence of a common factor (other than platelet COX-1) which explains almost half of the variation in platelet aggregation induced by collagen, ADP, and collagen-related peptide (21). Furthermore, poor compliance with the medication is a potential reason for ‘aspirin resistance’ and should be considered in patients with high on-treatment platelet responsiveness. The prevalence of ‘aspirin resistance’ in compliant subjects has been reported to be generally low (2% or less) (19).

Clinical implications of aspirin resistance

There is growing evidence that major adverse clinical events (MACE) in the settings of ACS, stroke/transient ischaemic attacks, and peripheral arterial disease can be predicted by the results of some tests evaluating on-treatment platelet reactivity.

Recently, several large meta-analyses have explored the relationship between laboratory data on platelet reactivity and risk of MACE (41). In a meta-analysis of 20 studies including 2,930 patients, 28% of participants were classified as ‘aspirin-resistant’ by laboratory methods. Overall, 39% of ‘aspirin-resistant’ patients compared with 16% of patients responsive to aspirin developed cardiovascular events (odds ratio 3.85; 95% CI 3.08–4.80). The odds ratio for increased mortality in aspirin-resistant patients was even higher, at 5.99 (95% CI 2.28–15.72).

The laboratory method used in various studies is also important—for example, 16% were classified as resistant to aspirin with LTA with relatively small heterogeneity of the results but with the highest odds ratio (3.85; 95% CI 2.5–5.88). In contrast, with the PFA-100, 33% of subjects were identified as aspirin-resistant, with a higher heterogeneity of the results and a lower predictive value for thrombosis (odds ratio for cardiovascular outcomes 2.94; 95% CI 1.88–4.55) (42). In another meta-analysis of 12 studies comprising 1,813 patients, the mean prevalence of aspirin resistance was 27%, although there was significant statistical heterogeneity amongst the studies (43). The pooled odds ratio of cardiovascular outcomes was 3.8 (95% CI 2.3–6.1) independently of the aspirin dose used (≤ 100 mg, 101–299 mg, and ≥ 300 mg).

Two other meta-analyses focused on studies where PFA-100 with the collagen/epinephrine cartridge was used for definition of non-responders. In the analysis by Crescente et al. (44) of 64 patient populations with 6,450 patients (2,283 subjects without history of vascular events and 4,167 patients affected by vascular events), the prevalence of aspirin non-responders was significantly higher in populations with vascular events (0.28 (95% CI 0.26–0.30) versus 0.23 (95% CI 0.21–0.25); P = 0.0002), and among them it was significantly higher in the acute phase of the disease. The factors associated with increased rate of poor responsiveness to aspirin were advanced age and type 2 diabetes. Pooled analysis found a relative risk of MACE amongst patients with ‘aspirin resistance’ of 1.63 (95% CI 1.16–2.28). Comparable results were obtained in another meta-analysis of studies based on the PFA-100 with an odds ratio of 2.1 (95% CI 1.4–3.4) for the increased recurrence of ischaemic events in aspirin non-responders (45). Taken together, these data indicate that the PFA-100 may be less useful for the prediction of MACE relevant to ‘aspirin resistance’.

In a large prospective study, Frelinger et al. (46) demonstrated that serum TxB₂ levels (a marker of COX-1-dependent platelet inhibition) and platelet aggregation with PFA-100 and collagen/ADP cartridge (i.e. aggregation non-specific to COX-1 pathway) were significantly correlated with subsequent MACE. These observations indicate that poor clinical outcomes in the aspirin-treated may be partly related to incomplete COX-1 inhibition, but also to COX-1-independent platelet hyper-reactivity.

In the Heart Outcomes Prevention Evaluation (HOPE) study (47), the 5-year follow-up of 976 patients with high risk of cardiovascular events revealed an increasing risk of MACE with each increasing quartile of 11-dehydro-TxB₂ levels, with patients in the upper quartile having a 1.8-fold higher risk of all MACE, as well as a 2-fold higher risk of myocardial infarction and a 3.5-fold higher risk of cardiovascular death compared to those in the lower quartile. This association was independent of conventional cardiovascular risk factors. Similarly, being in the upper quartile of LTA levels was an independent risk factor for developing cardiovascular events within 12 months (hazard ratio 7.76; 95% CI 1.72–35.3) (22).

In the pre-specified substudy of the Clopidogrel for High Atherothrombotic Risk and Ischemic Stabilization, Management and Avoidance (CHARISMA) (48), base-line urinary 11-dehydro-TxB₂ concentrations in the highest quartile were associated with an increased risk for MI, stroke, or cardiovascular death (adjusted hazard ratio 1.66; 95% CI 1.06–2.61). Increasing age, female gender, history of peripheral artery disease, current smoking, and oral hypoglycaemic or angiotensin-converting enzyme inhibitor therapy were also independently associated with higher urinary concentrations of 11-dehydro-TxB₂, whereas aspirin dose ≥ 150 mg, history of treatment with non-steroidal anti-inflammatory drugs, history of hypercholesterolaemia, and statin treatment were associated with lower concentrations. However, urinary 11-dehydro-TxB₂ is not a specific marker of platelet TxA₂ formation as it may also be produced by monocytes/macrophages in inflamed atherosclerotic plaques (49).

In a study of 1,069 patients taking clopidogrel and undergoing elective PCI, residual platelet reactivity has been detected using LTA, VerifyNow, and Plateletworks (Helena Laboratories, Beaumont, TX, USA) (50). Although the predictive accuracy of these tests was modest, high on-treatment platelet reactivity confirmed by any of these methods was associated with increased incidence of primary composite point of all-cause death, non-fatal acute MI, stent thrombosis, and ischaemic stroke. Many of the above studies had certain limitations, sometimes being underpowered and inadequately adjusted for potential confounders (age, gender, haemoglobin and platelet levels, hyperlipidaemia, etc.) or even compliance (41,51). Nevertheless the conclusions are generally uniform, that is, high platelet reactiveness on the antiplatelet therapy is linked to an increased risk of atherothrombosis (41).

However, there is still some debate on the relative importance of response to aspirin itself and high background platelet reactivity which cannot be sufficiently inhibited by aspirin. In fact, whereas the true ‘aspirin resistance’ appears to be a relatively rare phenomenon, residual on-treatment platelet reactivity can have greater importance. Indeed, in patients with ACS, high on-treatment platelet reactivity according to LTA with three agonists (AA, ADP, collagen) had the highest predictive value for cardiovascular death and non-fatal MI within 12 months (OR 4.7; 95% CI 2.9–7.7), whilst isolated platelet hyper-reactivity to only one agonist had no predictive value (52). In a study of 326 stable CAD patients, there was an association of ‘aspirin resistance’ with increased risk for all-cause death, MI, or stroke with both LTA with AA (i.e. COX-1-dependent pathway) and 10 µmol/L ADP (i.e. largely independent of TxA₂ production) (2,53). Also, Mueller et al. have demonstrated an 87% higher risk of arterial reocclusion after peripheral angioplasty in patients with ‘aspirin resistance’ as defined by response to ADP which parallels with appropriate aggregation in response to AA, an aspirin-specific stimulator (54). This observation indirectly confirms that background platelet reactivity may be more important rather than just the response to aspirin alone. Of note, in the cases of so-called ‘dual resistance’ to both aspirin and ADP receptor inhibitors, poor compliance with the medications should be considered (40,55–58).

‘Aspirin resistance’, residual platelet reactivity, or higher activity of inflammatory-coagulation response in severe vascular disease

Various genetic, biological, clinical factors affecting pharmacokinetics and pharmacodynamics of aspirin and modulating platelet sensitivity to the drug are likely to play a role in ‘aspirin resistance’ (Figure 1). Alongside well known factors contributing to poor response to aspirin treatment, such as factors reducing bioavailability of aspirin or alternative pathways of platelet activation, the variable effectiveness of aspirin in terms of clinical outcome and laboratory tests may be related to additional antithrombotic effects of aspirin such as inhibition of thrombin generation and attenuation of factor XIII activation (62–64). These effects of aspirin are linked with common genetic polymorphisms such as the Leu33Pro β3-integrin or Val34Leu factor XIII mutations (62–64). In addition, elevated levels of cholesterol (more than 240 mg/dL) appear to diminish the influence of aspirin on thrombin formation at the site of microvascular injury (62,64). Also, hypercholesterolaemia and smoking increase platelet response to other agonists, TxA₂ production from AA via lipid peroxidation, thus further contributing to ‘aspirin resistance’ (64).

Figure 1. Potential mechanisms of aspirin resistance.

There is also growing evidence that platelets play an intimate role in the inflammatory changes that lead to atherosclerosis, being a key cellular interface between coagulation and inflammation. Platelets are activated by inflammatory mediators and propagate inflammation by the further release of inflammatory substances, such as platelet factor 4 (PF4), IL-1β, RANTES (CCL5), and the surface expression of adhesion molecules such as GP Ib/IX/V, P-selectin, CD40L, and TNF superfamily 14 (TNFSF14) (66,67). Platelet receptors and released molecules also play an important role in acceleration of atherosclerosis by mediating leucocyte recruitment and adhesion to the vascular wall. Soluble CD40L (sCD40L) binds to platelet integrin glycoprotein IIb/IIIa, amplifying platelet activation in an autocrine fashion (66).

Platelet interactions with the monocytes—in the circulation or at the vessel wall itself—result in monocyte activation, which subsequently become more adhesive, more migratory, more procoagulant (increased tissue factor expression), more proinflammatory, and prone to macrophage differentiation.

Additionally, both monocytes and platelets contribute to an inflammatory phenotype of endothelium with enhanced expression of adhesion molecules and the release of IL-8, monocyte chemoattractant (MCP-1), matrix metalloproteases, and reactive oxygen species (66–69). This further promotes adhesion and activation of platelets and monocytes. At the same time, a dysfunctional endothelium, an early feature of atherosclerosis, produces less NO and prostacyclin, and releasing inflammatory molecules itself contributes to platelet activation and adhesion to the vascular wall.

Oxidized low-density lipoproteins (LDL) and high-density lipoproteins (HDL) have been shown to bind to platelet CD36, thus, in a dose-dependent way, provoking the release of significant amounts of CD40L as well as phosphatidylserine on the platelet surface (70–72). Stimulation both with thrombin and with aggregated IgG complexes induces similar release of sCD40L and RANTES from platelets, indicating common activation pathways of the coagulation system and the innate inflammatory response (73).

Whole blood assays of platelet function, examining platelet response in conditions of cell interactions, usually show a higher rate of increased on-treatment platelet reactivity than aggregation in PRP per se. For example, Chen et al. demonstrated that ‘aspirin resistance’ as measured by VerifyNow Aspirin was associated with impaired coronary flow reserve in patients who underwent elective PCI (74). In animal experiments, interactions between white blood cells (polymorphonuclear cells and monocytes) and platelets were suggested to play an important role in the pathogenesis of the ‘no-reflow’ phenomenon seen in interventional coronary procedures (75). In 1,600 healthy individuals with a family history of CAD, whole blood aggregation (but not aggregation in PRP depleted of leucocytes) increased with increasing quartile of leucocyte count both at base-line and after low-dose aspirin treatment, although the absolute magnitude of platelet reactivity was suppressed by aspirin therapy (76).

Tantry et al. (77) demonstrated incremental changes in platelet function, as measured by GP IIb/IIIa expression and increased release of RANTES, IL-8, and MCP-1, as CAD progresses from the asymptomatic state to stable angina and, finally, to unstable angina. The demonstration of significant correlations between inflammation, platelet function, and hypercoagulability markers has been interpreted by some as support for the hypothesis that reactive and activated platelets influence inflammation and other cells involved in low-grade inflammation of the vascular wall, thus creating a vicious cycle leading to a prothrombotic state, atherosclerosis progression, and/or plaque destabilization (78–80).

There is growing evidence that high platelet reactivity and atherosclerosis severity are closely related. Indeed, a high degree of platelet reactivity has been demonstrated in patients with more extensive coronary atherosclerosis and more severe peripheral arterial disease (81,82). Platelet activation reflects the stability of the CAD, with the greatest degree of platelet activation occurring in patients with onset of MI < 48 h (83). Frossard et al. (84) reported that the PFA-100 closure time (CT) at presentation may be an independent predictor of the severity of a STEMI (perhaps reflecting the ‘stability’ of a STEMI presentation), as measured by markers of myonecrosis. Also, Campo et al. (85) demonstrated that in patients undergoing PCI, base-line platelet reactivity (as measured by PFA-100 CT) influences the angiographic success of the procedure, the degree of ST-segment resolution, the extent of myonecrosis, and the short- and mid-term clinical outcomes.

Whilst ACS is independently associated with increased on-treatment platelet reactivity (86), this may be due to a higher pretreatment platelet activity in coronary ischaemic events (11). A significant association between interferon-inducible protein (IP-10), interferon-gamma (IFN-gamma), IL-4, and residual platelet reactivity can be demonstrated in patients with ACS (87). Elevated levels of CRP, leucocyte count, and fibrinogen are also significantly associated with high platelet reactivity in stable CAD patients on chronic clopidogrel treatment (88). The rate of platelet aggregation is also significantly related to monocyte count, high-sensitivity CRP (hs-CRP), mean platelet volume, vWF levels, and activity (89,90).

These findings indicate the strong impact of the systemic inflammatory properties on the pattern of platelet activity independently of the administration of aspirin or other antiplatelet agents. Indeed, there is a clear relationship between elevated inflammatory biomarkers and the risk of future atherothrombotic events (91–96), and dual antiplatelet therapy with aspirin and clopidogrel seems to bring more benefit to those with increased levels of inflammatory biomarkers (97,98).

Recent data show that obese individuals and patients with metabolic syndrome have greater native platelet reactivity and even retain greater reactivity after administration of aspirin (99). This parallels signs of chronic inflammation, such as increased leucocyte count and levels of CRP and sCD40L in these patients (100). Patients with diabetes have increased oxidative stress, inflammation, and endothelial dysfunction, and these may be related to higher pretreatment platelet activity, higher number of platelet-monocyte aggregates, enhanced biosynthesis of TxA₂ by monocytes, higher rate of non-responsiveness to antiplatelet therapy, and a higher rate of atherothrombotic complications in these patients (101–105).

It has been suggested that higher rates of non-responsiveness to antiplatelet therapy among diabetic patients may contribute to the lack of efficacy of aspirin for primary prevention of cardiovascular events in this population, and in contrast to the diabetes-free patients (106–108). Current prevention strategy in diabetic patients includes wide use of statins and inhibitors of the angiotensin system. The diverse effects of these agents include some anti-inflammatory properties and the ability to improve endothelial function. This may indirectly reduce platelet reactivity and blur the effect of aspirin on clinical outcomes in primary prevention amongst diabetics.

Shall we treat ‘aspirin resistance’?

The demonstration of the link between suboptimal platelet responsiveness to aspirin and increased risk of cardiovascular events prompted the concept of antiplatelet therapy monitoring. However, what should be the target of such monitoring? Shall we identify patients with true resistance to antiplatelet drugs, or shall we identify patients with high on-treatment platelet reactivity who may be at higher risk of ischaemic events?

We can only identify true non-responders using laboratory tests that are specific for the pharmacological target of an antiplatelet drug (30). The possibility of in-vitro testing of platelet resistance to aspirin has been demonstrated, but its clinical utility is still being established (4). For identifying patients with high residual platelet reactivity we need a well standardized test which is inexpensive, easy to perform, and reproducible, that may be used in different laboratories and results may be compared. This test must prove its utility in large prospective clinical studies.

To date, there are no robust clinical data obtained from large prospective trials with sufficient numbers of patients proving that the routine determination and/or monitoring of platelet function and consequent clinical decisions concerning antiplatelet therapy would lead to clinically relevant advantages and changes in patient management that are cost-effective. Consequently, no specific recommendations have been published for platelet function testing in routine clinical practice (5,109). However, class IIB recommendations from the American College of Cardiology/American Heart Association Task Force on Practice Guidelines have stated that platelet aggregation studies are warranted in patients undergoing PCI who are at risk of subacute stent thrombosis and in whom possible stent thrombosis might lead to devastating complications, with the option to increase the maintenance dose of clopidogrel from 75 to 150 mg/day in patients with < 50% blockade of platelet aggregation despite antiplatelet therapy (110).

The Task Force on Myocardial Revascularization of the European Society of Cardiology (ESC) and the European Association for Cardio-Thoracic Surgery (EACTS) states that monitoring of antiplatelet response by platelet function assays is currently used for clinical research, but not in daily clinical practice (111). At the same time, the use of a higher maintenance dose (150 mg) of clopidogrel is mentioned in patients with high thrombotic risk (e.g. in diabetics, patients after recurrent MI, after early and late stent thrombosis, for complex lesions, or in life-threatening situations should occlusion occur), undergoing elective PCI (111). Similarly, there are no recommendations for treatment of incomplete response to antiplatelet therapy ‘discovered’ using one of the platelet function tests (5).

On-going clinical trials aim to evaluate the effect of modifying therapy on clinical outcome for patients deemed non-responsive to antiplatelet therapy according to platelet function tests (Table II). In the ASCET study, 1,001 patients with documented coronary heart disease were randomized to either continued treatment with aspirin 160 mg/d or change to clopidogrel 75 mg/d. Clinical end-points (the composite of unstable angina, myocardial infarction, stroke, or death) were recorded for at least 2 years and related to the initial aspirin response, assessed by the PFA-100 method (113,118). In all, 26% of patients had high on-treatment platelet reactivity at randomization, but this did not significantly influence occurrence of the primary end-point (13.1% in non-responders versus 10.5% in responders; P = 0.41). Among aspirin non-responders who switched to clopidogrel, there was a 40% reduction in the combined end-point compared with those who stayed on aspirin, but this finding was not statistically significant, and there was no difference in the number of primary end-point events between the randomized groups (56/503 on aspirin versus 50/498 on clopidogrel; P = 0.57) (119). Recently published data of the BOCLA-Plan trial indicate that low responsiveness to aspirin or clopidogrel, defined with a platelet function test, can be overcome by escalating of the aspirin maintenance dose in case of low response to aspirin—or by an increase of the maintenance dose of clopidogrel, or switching to ticlopidine, or including prasugrel in the treatment plan, if low response to clopidogrel is detected (120). However, this study did not aim to investigate if the strategy to improve biochemical response provides improved MACE rates without causing an increased risk of bleeding. Future randomized clinical trials are needed to prove the clinical benefits of an individually tailored antiplatelet therapy.

Table II. Studies evaluating the effect of modifying therapy based on platelet responsiveness to aspirin (113–122).

Study	Design	Treatment arms	Outcomes measure
BOCLAplan	Design: Treatment, open label, single group assignment, safety/efficacy study Condition: CAD-PCI (either ACS or stable CAD) Enrolment: 500	Aspirin, clopidogrel: If the treatment with aspirin and/or clopidogrel was insufficient (platelet function testing) the dose was increased (aspirin 300 mg, aspirin 500 mg, clopidogrel 150 mg) or the drug was changed (clopidogrel to ticlopidine 2 × 250 mg, or prasugrel 10 mg, prasugrel 20 mg)	Whole blood LTA 48 hours after PCI to evaluate responders and low-responders to antiplatelet therapy. In the latter case alternative antiplatelet therapy according to the study plan was administered (time-frame: 2 years)
ASCET	Design: Treatment, randomized, open label, active control, parallel assignment, safety/efficacy study Condition: CAD Enrolment: 1,001	Aspirin (active comparator): aspirin 160 mg Clopidogrel (active comparator): clopidogrel 75 mg Patients were randomized to either continued treatment with aspirin 160 mg/d or change to clopidogrel 75 mg/d after initial determination of their platelet reactivity while on aspirin treatment (PFA-100^®)	Primary: composite end-point of all-cause death, non-fatal MI, ischaemic stroke, and unstable angina (time-frame: 2 years) Secondary: in-stent restenosis and/or thrombosis detected by coronary angiography (time-frame: 2 years)
RESIST	Design: Treatment, randomized, open label, parallel assignment, efficacy study Condition: stable angina–PCI Enrolment: 36 patients resistant to aspirin (VerifyNow Aspirin)	Conventional strategy: patient receives 325 mg aspirin orally and loading does of 600 mg clopidogrel at time of procedure Aggressive strategy (active comparator): patient receives 325 mg aspirin orally and loading dose of 600 mg clopidogrel at time of procedure with addition of IV GP IIb/IIIa inhibitor bolus intra-procedurally	Primary: elevation of cardiac enzyme (time-frame: 24 hours) Secondary: MACE (time-frame: 30 days): death, MI, stent thrombosis, urgent revascularization, bleeding
Dual antiplatelet therapy in patients with aspirin resistance following coronary artery bypass grafting	Design: Prevention, randomized, open label, parallel assignment, safety/efficacy study Condition: CAD Enrolment: 200 patients resistant to aspirin (multiple electrode aggregometry)	Patients with inappropriate response to aspirin 300 mg therapy after CABG randomized to receive aspirin monotherapy (no intervention) or dual antiplatelet therapy (active comparator): clopidogrel 75 mg in addition to aspirin	Primary: MACE events (death, non-fatal MI, stroke, cardiac rehospitalization) (time-frame: 6 months)
ARCTIC	Design: Treatment, randomized, open label, parallel assignment, efficacy study Condition: CAD, elective PCI-DES Enrolment: 2,500	First randomization: Monitoring arm (experimental): dose adjustment of both aspirin and clopidogrel in suboptimal responders identified based on a point of care assay (VerifyNow). Conventional arm (active comparator): fixed dose regiment of both aspirin and clopidogrel in all patients following DES implantation according to international guidelines Second randomization after 1 year of follow-up: Pursuit arm (experimental): pursuit of a dual oral antiplatelet therapy (aspirin and clopidogrel) beyond 1 year Interruption arm (active comparator): interruption of clopidogrel therapy.	Primary: composite end-point of death, MI, stroke, urgent coronary revascularization, stent thrombosis assessed at 1 year for the first hypothesis and between 6 up to 18 months of follow-up for the second hypothesis (time-frame: during the study—1 year in both ‘monitoring’ and ‘conventional’ arms and during the period from 6 up to 18 months in the ‘interruption’ and ‘pursuit’ arms)
TEG-CABG	Design: Treatment, randomized, open label, parallel assignment, safety/efficacy study Condition: elective/subacute multivessel CABG Enrolment: 375 TEG and multiplate aggregometry are performed to assess platelet reactivity and resistance to aspirin and clopidogrel	Clopidogrel+aspirin, hypercoagulable (experimental) Aspirin, hypercoagulable control (active comparator) Aspirin, normocoagulable control (active comparator)	Primary: graft patency at 3 months and 1 year (time-frame: 3 months, 1 year) Secondary: - rate of other thromboembolic events (e.g. MI, stroke, pulmonary embolus, etc.) (time-frame: 3 months, 1 year) - assessing coagulation profile pre- and postoperatively, including aspirin and clopidogrel resistance (time-frame: 1 year)
The ITALIC Study	Design: Treatment, randomized, open label, parallel assignment, safety/efficacy study Condition: Patients after DES implantation, excluding primary PCI for acute MI and left main PCI Enrolment: 3,200	Aspirin-clopidogrel regimen versus mono-aspirin regimen after 6 months of dual antiplatelet therapy in good aspirin responder patients	Primary: composite end-point of death, MI, repeat urgent revascularization, stroke requiring a new hospitalization, major bleedings at 12 months Secondary: Incidence the same composite end-point at 24 and 36 months and incidence of minor bleeding complications at 12, 24, and 36 months

Thus, the optimal treatment, if any, of aspirin resistance is unknown. Therefore, it is not currently recommended to test for aspirin resistance in patients or to change therapy based on these tests, in routine clinical practice (5). Patients that have experienced a recurrent vascular event require detailed evaluation of the causes of the initial and recurrent events. Whenever the suspicion of aspirin resistance is raised by laboratory tests, compliance and potentially modifiable causes of aspirin failure should be considered and optimized (e.g. concurrent intake of non-steroidal anti-inflammatory drugs, reduced absorption of aspirin, or inadequate dose of aspirin), and other risk factors potentially influencing platelet function (e.g. smoking, hypercholesterolaemia, inadequate glucose control, obesity) should be managed appropriately.

The experts of the Working Group on antiplatelet drug resistance admit the possibility of individual antiplatelet dosing in individual cases (e.g. patients with multiple cardiovascular risk factors and recurrent thrombotic events despite proven compliance to standard antiplatelet drugs doses). However, the Working Group states that those measures should be undertaken only as a research activity in academic centres with experience in platelet function testing, and currently there is not sufficient evidence of their efficacy (5).

New, more potent P2Y₁₂ receptor antagonists are currently available, and on-going clinical trials aim to justify an individualized approach to antiplatelet therapy, based on on-treatment platelet reactivity. Cholesterol-independent effects of statins, which include improvement of endothelial function, decrease of oxidative stress, and inflammation, are thought to inhibit indirectly platelet activity (48). Recent studies showed that statins significantly reduce platelet TxA₂ formation in patients taking low-dose aspirin (122), prompt overcoming of aspirin resistance in 65% of the patients after a therapy of 3 months (123), and may have a direct effect on platelet eicosanoid synthesis (124). Combined with aspirin, omega-3 polyunsaturated fatty acids facilitate aspirin-induced platelet inhibition (125), improving response to aspirin comparably with the aspirin dose increase (126). The addition of omega-3 polyunsaturated fatty acids to dual antiplatelet therapy significantly potentiates platelet response to clopidogrel after PCI (127).

Some existing evidence links inflammation and thrombosis, as well as anti-inflammatory properties of antiplatelet drugs (128), which supports the hypothesis that agents with both anti- inflammatory and antiplatelet effects may reduce vascular inflammation and platelet reactivity and limit acute and long-term thrombotic events.

Conclusion

Currently, extensive literature exists about aspirin resistance, its mechanisms, detection, and its treatment. Existence of a link between high on-treatment platelet reactivity and atherothrombotic events suggests common mechanisms for atherosclerosis progression and thrombotic complications, with the platelets being a key cellular interface between coagulation and inflammation. A change in the approach to prevention of atherothrombotic events may ultimately include all the components of atherothrombogenesis, including platelets, monocytes, and endothelial cells.

Nonetheless, tailoring antiplatelet therapy in accordance with the presence of aspirin resistance is one possible solution. However, the evidence in favour of this strategy is insufficient.

Acknowledgements

The work was supported by the European Society of Cardiology Atherothrombosis Fellowship.

Declaration of interest: The authors state no conflict of interest and have received no payment in preparation of this manuscript.