Coronary Stent Thrombosis in the Current Era: Challenges and Opportunities for Treatment
- First Online:
- Cite this article as:
- Zwart, B., van Werkum, J.W., Heestermans, A.A.C.M. et al. Curr Treat Options Cardio Med (2010) 12: 46. doi:10.1007/s11936-009-0055-z
- 43 Views
The introduction of the drug-eluting stent has raised concerns regarding the occurrence of stent thrombosis (ST), particularly late (and very late) thrombosis. This renewed attention shows that ST remains a major concern after implantation of both bare metal and drug-eluting stents. Cardiologists should be aware of this dreadful complication, because it is associated with substantial morbidity and mortality. Numerous clinical, procedural, and angiographic risk factors have been identified. Moreover, the influence of novel determinants, such as high on-treatment reactivity, genetic predisposition, and the stent’s direct effects on the (healing of the) vessel wall, now are recognized. Consequently, the pathophysiology of ST has evolved into a complex multifactorial model. This broader understanding of the pathophysiology of ST enables cardiologists to perform extensive risk stratification to identify patients at higher risk and provides clues to important treatment options. The core of primary prevention after stent implantation, as well as secondary prevention after ST, should consist of a) the prevention of modifiable risk factors and b) optimal individualized treatment for each patient. Future developments, such as genetic bedside testing, point-of-care platelet testing, and sophisticated imaging modalities, might aid in this approach.
The treatment of patients with coronary artery disease has changed dramatically since Andreas Grüntzig  introduced percutaneous coronary intervention (PCI) in 1977. By means of a simple expanding balloon at the site of coronary narrowing, it was possible to reduce the narrowing and relieve angina pectoris. However, initial treatment with plain-old balloon angioplasty was limited by early elastic recoil, negative remodeling, and intimal hyperplasia, leading to restenosis (35%–45%). The high percentage of acute closures (5%–8%) was another major limitation, although it was drastically reduced by the introduction of coronary stents.
Coronary stenting initially was a bail-out indication for treating acute complications of balloon angioplasty. The most important bail-out indications for coronary stent implantation were to provide a mechanical scaffold for the vessel wall, to seal dissections, and to prevent elastic recoil. Studies in the mid-1990s demonstrated that coronary stenting also reduced angiographic restenosis, from the 30% to 40% rate seen with balloon angioplasty to 10% to 20%. However, the broad use of coronary stents introduced a new complication of coronary stent implantation: coronary stent thrombosis (ST).
Coronary stent thrombosis
PCI with stent implantation induces mechanical laceration and fissuring of the atherosclerotic plaque as well as denudation of the arterial endothelium. Hence, it is not surprising that the studies done in the mid-1990s reported ST rates as high as 20% to 25%, because little concomitant anticoagulation was administered . Acute thrombotic closure of the coronary stent results in a life-endangering condition with catastrophic consequences. It usually presents as an acute myocardial infarction or sudden death. Initially, this problem was tackled with complex anticoagulation regimes such as aspirin, heparin, and warfarin therapy; however, these treatments led to unacceptably high rates of major bleeding, vascular complications, and prolonged hospital stays. The development and introduction of new antiplatelet agents, such as the thienopyridines (eg, ticlopidine and clopidogrel) and the glycoprotein IIb/IIIa inhibitors (abciximab, tirofiban, eptifibatide), as well as advances in techniques (eg, application of high-pressure stent dilatation) led to a breakthrough in the general use of coronary stents.
Improvements in strut configuration and thickness, as well as new materials, enhanced deliverability and reduced vessel damage. These advancements led to the introduction in 2003 of drug-eluting stents (DES), which release drugs that reduce neointimal formation through the arrest of cell proliferation. DES reduced rates of in-stent restenosis significantly. By that time, it was generally believed that ST occurring after the first month was rare.
Nevertheless, soon after the introduction of DES, the interventional community was alarmed by a suspected high incidence rate of late ST (beyond 1 month). Although late ST was not investigated extensively in bare metal stents (BMS), the introduction of DES brought this subject to the attention of cardiologists [3•, 4]. Consequently, in recent years, renewed attention has been paid to late ST, with experts concluding that ST continues to occur at a stable rate after the first month [5, 6•].
Given the fact that ST often has devastating clinical consequences, there has been considerable interest in identifying patients at high risk for this catastrophic event. Multiple studies and registries have investigated the relationships between patient, lesion, and procedural factors and ST, but all have been hampered by methodologic challenges, primarily the low incidence of ST in contemporary patient series [7–10•]. Despite these important limitations, several predictors have been found, providing much insight into this complex pathophysiology.
The primary aim of this review is to provide a better understanding of the underlying mechanisms responsible for ST and, ultimately, to establish an optimal strategy to reduce this dreadful complication.
Definition of stent thrombosis
Despite being a quantitatively minor problem, ST has a major clinical impact because of the high risk of myocardial infarction and death. However, the true impact and incidence of ST have been neglected for more than 10 years, partly because its definition has varied greatly among randomized clinical trials and observational registries. To allow a fair comparison (ie, provide consistency across studies) between the true rates of ST across different trials and registries, a new uniform definition of ST was proposed recently by the Academic Research Consortium (ARC), a collaboration of research organizations from Europe and the United States .
- ST is classified according to 1) its timing and 2) the level of evidence for its presence. Timing is defined in four categories, implying different pathophysiologic mechanisms (Table 1). The level of evidence is stratified into three categories, indicating varying degrees of certainty: possible, probable, and definite ST (Table 2).Table 1
Temporal categories of stent thrombosis
Time after stent implantation
< 24 h
>24 h but < 30 days
>30 days but < 1 year
>1 yearTable 2
Degrees of certainty for stent thrombosis
Angiographic confirmation of stent thrombosis: Intracoronary thrombus in or < 5 mm proximal/distal to the stent and Clinical or biochemical changes compatible with cardiac ischemia, or angiographic thrombus
Pathologic confirmation of thrombus (at autopsy or after thrombectomy)
Any unexplained death <30 days after coronary stenting
Any MI in the territory of the implanted stent
Any unexplained death >30 days after coronary stenting
BMS versus DES
The pivotal trials investigating the safety of DES initially reported on a follow-up of 1 to 2 years. After their use was embraced by interventional cardiologists because of reductions of in-stent restenosis rates, DES came under fire from the US Food and Drug Administration in 2003 because of a supposed surplus of late, and particularly very late, ST accompanying the use of DES [3•,4].
Recently, several reviews and meta-analyses thoroughly addressed this issue [12,13]. Although several confounding factors (eg, off-label use of DES, indications for index PCI, subtypes of stents, and the complexity of lesions treated) varied broadly across the studies, complicating a head-to-head comparison, the following general conclusions can be drawn: 1) mortality rates with BMS and DES are similar, 2) the incidence of early ST is the same for BMS and DES, and 3) the use of BMS might impose a slight excess risk of ST within the first 6 months, whereas DES are associated with a moderate increase in ST after 1 year. However, the latter does not translate into higher mortality rates, which can be explained by the reduced need for revascularization associated with the use of DES. Some authors report even lower mortality rates with DES compared with BMS .
In an important meta-analysis published in 2009, Brar et al.  compared DES with BMS in 33,873 patients with myocardial infarction treated with primary PCI. The authors concluded that the use of DES in myocardial infarction appears safe and efficacious and is not associated with an increase in ST. These data were confirmed by a large study from Shishehbor et al. .
Incidence of coronary stent thrombosis
In the BMS era, the incidence of ST declined quickly, from initial rates of approximately 24% to only 1% to 5% in the first month.
The incidence of ST beyond 1 month after implantation was not recognized until the publication of case reports of late ST associated with brachytherapy [17,18]. Since then, late BMS thrombosis received researchers’ attention, although relatively few studies on the subject were published. These studies reported a yearly incidence of less than 1% for late ST in BMS [5,19].
The aforementioned meta-analyses estimated the incidence of late ST in both BMS and DES, despite different definitions of ST among the studies. The overall incidence of late and very late ST from registries and trials was estimated at 0.5% to 1.5% during the first year and 0.5% per year thereafter.
Outcome after stent thrombosis
Although several studies addressed the important question regarding clinical outcome after ST, they provided little consistency regarding mortality rates [20–22]. However, recently published studies assessed the outcome in larger cohorts of patients with ST.
de la Torre-Hernandez et al.  analyzed 23,500 patients after stent implantation, 301 of whom developed definite ST. Mortality at 1 year follow-up was 16%, and recurrent ST occurred in 4.6% of patients. In addition, several risk factors for mortality were identified, including older age, left ventricular ejection fraction less than 45%, nonrestoration of Thrombolysis in Myocardial Infarction (TIMI) flow grade 3, and additional stenting.
van Werkum et al. [24••] performed a long-term follow-up in a consecutive cohort of 431 patients with definite ST. After a median follow-up of 27.1 months, the primary end point (a composite of cardiac death and definitive ST) occurred in 111 patients (25.8%), with a cardiac mortality rate of 12.3%.The cumulative incidence rates of definite recurrent ST, definite or probable recurrent ST, any myocardial infarction, and any target vessel revascularization were 18.8%, 20.1%, 21.3%, and 32.0%, respectively. This unexpectedly high recurrence rate was confirmed by another study by Lemesle et al. , who reported a recurrence rate of 36% in patients successfully treated for a first ST during a median follow-up of 40 months.
Pathophysiology and predictors of stent thrombosis
- The pathophysiology of ST has evolved from the identification of single causative factors to a complex multifactorial model (Fig. 1). Numerous risk factors have been identified during the past years Historically, these predictors can be classified as clinical, procedural, or lesion related. More recently, researchers recognized the involvement of novel determinants, including a heightened platelet reactivity status despite antiplatelet therapy, impaired responsiveness to antiplatelet therapy, genetic predisposition, and direct effects of the stent on the vessel wall. The many factors contributing to ST are depicted in Fig. 1.
Although early and late ST share several common risk factors, the impact of these factors varies. Early and late ST represent unique pathophysiologic characteristics, which are discussed separately.
Early stent thrombosis
Numerous studies have reported the predominance of mechanical and anatomic etiologies underlying early ST, including bifurcation and restenotic lesions, number of stents implanted, small vessels, lesion diameter and complexity (types B and C), undersizing and/or underexpansion of the coronary stent, (residual) dissections, postprocedural TIMI flow grade less than 3, absence of glycoprotein IIb/IIIa inhibitor treatment, and lack of intravascular ultrasound (IVUS) guidance [7–10•]. Disturbances in coronary flow (increased shear stress) at bifurcations and restenotic lesions may activate platelets and contribute to delays in arterial healing. Moreover, a large number of stents and long total stent length delay the process of endothelialization, which in turn increases the risk of ST.
Another very important nonmechanical cause of early ST is premature cessation of clopidogrel therapy, which is associated with hazard ratios up to 90 .
Late and very late stent thrombosis
Late ST seems less strongly linked to mechanical factors related to the index PCI procedure. In addition, the influence of clopidogrel cessation on late and very late ST is less well established, and when an association has been found, the risk has been consistently lower than that for early ST [10•,21,26–30].
Nevertheless, several clinical predictors are strongly related to late and very ST, including acute coronary syndrome as the indication for the index PCI, active malignancy, diabetes, low ejection fraction, and renal failure [10•,20]. In addition, the stent itself is involved in the development of ST. Several autopsy studies described the histopathologic changes of the vessel wall following stent implantation, presumably as a result of the direct effects of the coronary artery . Chronic inflammation surrounding the stent is found in a substantial number of patients. This phenomenon has been exclusively related to the use of polymers and other components in DES. Within months, this inflammatory process can induce late malapposition and vascular remodeling. Besides the negative influence of this process, DES induce delayed healing and incomplete endothelialization, resulting in impaired coverage of stent struts, even beyond 1 year after stent implantation . These bare stent struts are thought to induce thrombus formation to the vessel wall.
Although (the extent of) these findings might be biased by the fact that they are from autopsy studies representing a small and selected patient population, they were reproduced recently in studies using intravascular imaging technologies (eg, IVUS and optical coherence tomography) and histopathologic findings from thrombi obtained with thrombectomy catheters [31–33••]. Cook et al. [33••] found both histopathologic (thrombus) signs of inflammation and IVUS evidence of vessel modeling. Importantly, histopathologic analysis of harvested thrombus showed infiltration of eosinophils and macrophages following implantation of sirolimus-eluting and paclitaxel-eluting stents, respectively, suggesting a hypersensitivity reaction.
The importance of concomitant antithrombotic therapy
Various antithrombotic regimens have been studied for their ability to minimize the incidence of acute and subacute ST while reducing the risks of hemorrhagic complications. Current guidelines recommend dual-antiplatelet therapy (aspirin and clopidogrel) for all patients undergoing coronary stent implantation. Dual-antiplatelet therapy is absolutely necessary to prevent ST, although the optimal duration of this therapy after coronary stent implantation remains unclear. It is hoped that ongoing trials such as Double Randomization of a Monitoring Adjusted Antiplatelet Treatment Versus a Common Antiplatelet Treatment for DES Implantation, and Interruption Versus Continuation of Double Antiplatelet Therapy (ARCTIC, NCT00827411) and Safety and Efficacy of Six Months Dual Antiplatelet Therapy After Drug-Eluting Stenting (ISAR-SAFE, NCT00661206) will reveal the optimal duration of dual-antiplatelet therapy after DES implantation.
High on-treatment platelet reactivity
Despite the standard “one-size-fits-all” dual-antiplatelet therapy regimen with aspirin and clopidogrel, it has become clear that many patients still suffer from recurrent atherothrombotic coronary events. As a result, the number of studies and papers focusing on this so-called treatment failure has grown exponentially since 2003. Throughout the past few years, multiple studies have demonstrated that the response to a fixed dose of antiplatelet therapy (clopidogrel and aspirin) is highly variable and the responsiveness to clopidogrel, as measured with adenosine diphosphate (ADP)-induced aggregation, follows a bell-shaped gaussian distribution. Consequently, many patients receiving combination therapy with aspirin and clopidogrel fail to obtain the optimal benefit from it. This phenomenon has been termed resistance to antiplatelet therapy in the medical literature. However, the somewhat confusing term resistance implies that these drugs do not reach their pharmacologic target at all, which is not the case in most instances. Hence, the alternative term high on-treatment platelet reactivity was introduced recently. To date, six studies have demonstrated a clear association between the magnitude of on-treatment platelet reactivity and the occurrence of ST. Thus, high on-treatment platelet reactivity has emerged as another important risk factor for ST. Moreover, consistent findings across multiple studies indicate that high on-treatment platelet reactivity is associated particularly with early ST [34,35]. Whether high on-treatment platelet reactivity also is associated with late ST needs to be explored in sufficiently powered studies.
The pathophysiology of arterial thrombosis is very complex, and multiple factors have been identified as being involved. Nonetheless, these known factors alone do not fully explain an individual’s risk profile, and it is likely that multiple genetic polymorphisms are involved. Likewise, there is much interindividual heterogeneity in the response to antithrombotic therapy [36•, 37, 38], and it has been demonstrated clearly that interindividual variations in metabolism, transporters, and drug targets are important determinants of drug efficacy. Because this straightforward principle of pharmacogenetics can affect any step in modulating the pharmacokinetics and pharmacodynamics, this concept might be more relevant than complex genetics leading to the development of cardiovascular diseases .
During the past four years, the impact of a genetic mutation related to clopidogrel metabolism has been elucidated. In 2006, Hulot et al.  identified the CYP2C19*2 mutation responsible for the biotransformation of the prodrug into its active metabolite. This cytochrome P-450 enzyme is involved in the biotransformation of clopidogrel into its active metabolite. Carriers of the *2 alleles exhibit higher platelet reactivity [40, 41]. Of equal importance, the CYP2C19*2 mutation is not uncommon, with carrier frequencies varying between 20% and 50%, depending on the population and ethnicity. In addition, a dose response is evident: homozygotes for CYP2C19*2/*2 respond even less well to clopidogrel therapy than heterozygotes for CYP2C19*2, who in turn respond less well than wild-type homozygotes. Importantly, several large trials published in 2009 demonstrated that genetic variation has an effect on the pharmacologic and clinical response to clopidogrel. Several studies correlated these genetic variations to clinical outcomes, including cardiovascular death, myocardial infarction, and ST. In their subanalysis of the Trial to Assess Improvement in Therapeutic Outcomes by Optimizing Platelet Inhibition With Prasugrel (TRITON)-TIMI 38, Mega et al.  found that CYP2C19*2 carriers had a threefold risk of ST. Sibbing et al.  found CYP2C19*2 carriers to have a fourfold risk of ST.
Other CYP mutations have been associated with high on-treatment platelet reactivity. Harmsze et al.  investigated a large cohort of clopidogrel-treated patients undergoing elective PCI and discovered that CYP2C9*3 carriers exhibited higher platelet reactivity and more often were poor responders.
These findings represent a major step forward in this field of research and might have important treatment implications. Future trials will have to prove the clinical relevance of genetic testing.
Triggering mechanisms of stent thrombosis
Several superimposing factors have been investigated in myocardial infarction, including time of day, physical exercise, infection, and emotional stress. However, only case reports have been published so far on a possible relationship between triggering mechanisms and ST [44, 45]. In our own institutional experience, however, triggering mechanisms seem to play an important role (unpublished data). All consecutive patients with definite ST in our large cohort were interviewed about the possible performance of vigorous physical exercise, the presence of an infection, or the presence of acute emotional stress preceding the onset of symptoms accompanying the ST. In 23% of these patients, a trigger was identified. Importantly, a clear circadian variation with a steep morning peak also was identified .
It should be kept in mind that curative management strategies for patients presenting with ST are mainly empirically based, as only a few studies have evaluated (observationally) the available treatment modalities . Nonetheless, data from numerous registries and case-control studies do offer certain guidance for managing patients with ST. For instance, evidence from histopathologic studies indicates that patients in whom ST is highly suspected should be treated with emergent PCI, not thrombolytic therapy [47–49].
Because in most cases ST presents as myocardial infarction, or even cardiogenic shock, the credo “time is muscle” is of utmost importance, and patients should be transported immediately to an interventional center upon diagnosis. The diagnosis of ST is confirmed by coronary angiography, although it must be noted that, in general, the identification of a thrombus on a conventional angiogram may be a difficult challenge.
Given the relatively large thrombus load associated with ST, thrombus aspiration might be beneficial in obtaining effective reperfusion , although more clinical data are urgently needed.
Subsequent balloon dilatation should be performed; if stent malapposition (eg, due to undersizing or late malapposition) is identified, additional balloon dilatation with well-sized balloons and higher balloon pressure is advisable. However, if the stent appears well expanded and no residual dissection is present, another stent should not be implanted, as several studies have demonstrated that implanting an additional coronary stent at the time of the first ST is associated with an increased risk of recurrent ST and even death [23,24••].
Dr. ten Berg has received speaker’s fees from Eli Lilly, Sanofi-Aventis, Merck Sharp & Dohme, and The Medicines Company. Dr. van Werkum has received speaker’s fees from Accumetrics and Siemens and has served on an advisory board for The Medicines Company. No other potential conflicts of interest relevant to this article were reported.