Coronary artery stent thrombosis: Incidence and risk factors

INTRODUCTION — Coronary artery stents are used in the majority of patients who undergo percutaneous coronary intervention (PCI), as they significantly reduce the need for repeat target vessel revascularization compared to balloon angioplasty. (See "Intracoronary stent restenosis", section on 'Incidence of restenosis'.)

Stent thrombosis is an uncommon but serious complication that almost always presents as death or a large non-fatal myocardial infarction (MI), usually with ST elevation. It is estimated that less than 10 percent of cardiac deaths after stent placement are attributable to stent thrombosis, with most of the remainder being due to disease progression [1].

Stent thrombosis can occur acutely (within 24 hours), subacutely (within 30 days), or as late as one year (late) or more (very late) after stent placement. Stent thrombosis within the first year appears to occur with similar frequency in patients with bare metal stents (BMS) or drug-eluting stents (DES), as long as patients are treated with dual antiplatelet therapy (aspirin plus a platelet P2Y₁₂ receptor blocker) for the recommended duration. The period of risk requiring dual antiplatelet therapy (DAPT) is longer with DES, due at least in part to delayed neointimal coverage, although this difference is decreasing with newer generation DES. (See 'Very late stent thrombosis' below.)

General issues related to stent thrombosis will be reviewed in this topic. The prevention and management of stent thrombosis are discussed separately. (See "Long-term antiplatelet therapy after coronary artery stenting in stable patients" and "Coronary artery stent thrombosis: Clinical presentation and management".)

DEFINITIONS — Although earlier clinical trials posed a problem for comparing stent thrombosis rates due to non-standardized definition, since 2007 most trials have adopted the criteria and classification proposed by the Academic Research Consortium (ARC) [2,3].

●Definite – Angiographic confirmation of a thrombus that originates in the stent or in the segment 5 mm proximal or distal to the stent, with or without vessel occlusion, which is associated with acute onset of ischemic symptoms at rest or ECG signs of acute ischemia or typical rise and fall of in cardiac biomarkers within 48 hours of angiography OR pathologic confirmation of stent thrombosis determined at autopsy or from tissue obtained following thrombectomy.

●Probable – Unexplained death occurring within 30 days after the index procedure, or an MI occurring at any time after the index procedure that was documented by ECG or imaging to occur in an area supplied by the stented vessel in the absence of angiographic confirmation of stent thrombosis or other culprit lesion.

In the discussion that follows, studies using definitions other than ARC will be noted.

INTRAPROCEDURAL STENT THROMBOSIS — Intraprocedural stent thrombosis (IPST) is the development of occlusive or non-occlusive new or increasing thrombus in or adjacent to a recently implanted stent before the percutaneous coronary intervention (PCI) procedure is completed [4,5]. IPST is not included in the academic research consortium definitions discussed above. In one rigorous study of IPST, it was further defined as any discrete, mobile, intraluminal filling defect, with defined borders, with or without associated contrast staining or a total occlusion with convex edges and staining [5].

Intraprocedural stent thrombosis, by definition, occurs sometime after an intracoronary stent is deployed and before the end of the PCI. The best available data on its incidence come from an analysis of events in the CHAMPION PHOENIX trial, which compared cangrelor to clopidogrel in patients undergoing PCI [5]. Among 10,939 patients included in the study, IPST developed in 0.8 percent. Similar incidences have been reported in other smaller series [4,6]. (See "Antithrombotic therapy for elective percutaneous coronary intervention: Clinical studies", section on 'Cangrelor'.)

The risk of IPST is significantly increased in patients with one or more of the following characteristics: ST- or non-ST elevation myocardial infarction (MI) (both compared with stable angina), thrombus at baseline, and bifurcation lesions [4,5].

The clinical presentation and management of IPST is discussed elsewhere. (See "Coronary artery stent thrombosis: Clinical presentation and management", section on 'Intraprocedural stent thrombosis'.)

MECHANISMS — Risk factors for the development of stent thrombosis fall into broad categories of procedural, patient, lesion, and stent characteristics as well as the cessation of antiplatelet therapy [7]. (See 'Risk factors' below.)

Each risk factor is likely to predispose the patient to stent thrombosis, which is characterized by platelet activation and aggregation, by one or more of the following mechanisms:

●Persistent slow coronary blood flow (local or arterial) as occurs with dissection or hypoperfusion.

●Prior to reendothelialization, exposure of the blood to prothrombotic subendothelial constituents, such as tissue factor, or to the stent itself.

●Failure to suppress platelet adhesion/aggregation at a time of prothrombotic risk, as occurs with premature cessation of antiplatelet therapy or drug resistance.

TIMING AND INCIDENCE — Most cases of stent thrombosis, occur within the first 30 days after placement, irrespective of stent type [8]. Reported rates tend to be lower in clinical trials (or meta-analyses of clinical trials) than in registries, as trials often enroll lower risk patients. In general, we tell our patients that the rate of stent thrombosis at one year is about one percent and that the yearly rate following one year is about 0.2 percent per year.

Most of the data for drug-eluting stent (DES) thrombosis across populations with a broad range of risk come from multiple registries in the era of first generation DES:

●In the Bern-Rotterdam registry, 8146 patients underwent PCI with a sirolimus- or a paclitaxel-eluting stent (SES or PES). At three years, the cumulative incidence of angiographic stent thrombosis was 2.9 percent, with early (acute and subacute) stent thrombosis occurring in 60 percent and late stent thrombosis (all events after 30 days) in 40 percent [9]. After one year, the rate of stent thrombosis was 0.6 percent per year.

●In the Swedish Coronary Angiography and Angioplasty Registry (SCAAR), 42,150 patients underwent PCI with either bare metal or drug-eluting stents [10]. During a mean follow-up of 661 days, the rate of definite stent thrombosis was 1.2 percent, of which 50 percent were acute or subacute. The rate after one year appeared constant out to three years at about 0.3 to 0.4 percent per year.

●In the Dutch stent thrombosis registry, 21,009 underwent PCI with either bare metal or drug-eluting stents [11]. During a median follow-up of 31 months, ARC definite stent thrombosis was present in 2.1 percent of which 32 percent was acute, 41 percent subacute, 13 percent late and 14 percent very late.

These data are limited by the restriction to first generation DES. (See 'Comparison of differing DES' below.)

Bare metal stents — Bare metal stent (BMS) thrombosis usually occurs within the first 24 to 48 hours (acute) or much less often within the first month (subacute) after stent placement [12-20]. In a pooled analysis of data from stent trials registries, stent thrombosis developed in 0.9 percent by 30 days and approximately 80 percent of these occurred within the first two days [14]. The incidence of stent thrombosis by 30 days has ranged between 0.5 to 2.5 percent with higher rates in patient populations at higher risk [12,13,21].

Thrombotic events with BMS are uncommon after 30 days of treatment with dual antiplatelet therapy [11,22-24]. This observation is consistent with angioscopic studies that showed complete re-endothelialization by three to six months [25,26]. In a report from the Swedish Coronary Angiography and Angioplasty Registry (SCAAR) of over 94,000 stent implantations, the cumulative incidence of stent thrombosis with BMS at one year was approximately 1.1 percent [27].

Very late stent thrombosis (after one year) is uncommonly seen with BMS and is often after a repeat procedure has been performed in the stented segment. In SCAAR, the rate of definite stent thrombosis at two years was 1.4 percent [27]. In our experience of adjudicating late and very late stent thrombosis after BMS, it is often difficult to determine the contributing role of restenosis.

Drug-eluting stents — The rates of stent thrombosis, particularly after one year, are lower with new-generation compared with first-generation DES [28].

Similar to BMS, most episodes of drug-eluting stent thrombosis occur in the first year and many of these within the first 30 days. In a meta-analysis of 10 randomized trials of 5030 patients comparing PES or SES (first generation DES) to BMS, the 30-day rate of stent thrombosis (ST) in patients receiving DES was 0.58 percent [21]. In a report of over 94,000 stent implantations in the Swedish Coronary Angiography and Angioplasty Registry (SCAAR), the cumulative incidence of stent thrombosis with DES at one year was approximately 0.4 to 0.8 percent [27]. In a 2006 meta-analysis of 19 randomized trials with 7060 patients, the risk of PES or SES thrombosis was 0.7 percent with one-year follow-up [29].

Events continue after one year (very late stent thrombosis) for at least five years. In clinical trial populations, the risk has been estimated at 0.2 to 0.3 percent per year after the first year [30], but a number of studies have demonstrated the risk beyond one year is higher (as much as 0.6 percent per year) in routine practice settings [10,31-34]. Ongoing inflammation and delayed healing with first generation DES were largely responsible for this higher risk of very late stent thrombosis. Much of the safety benefit for newer generation DES is due to the reduction in these very late events. (See 'Very late stent thrombosis' below and 'Comparison of differing DES' below.)

The following studies illustrate the range of risk beyond one year for patients who received differing DES:

●In a 2012 publication from SCAAR, the cumulative rates of definite stent thrombosis at two years for older DES (SES, PES, and the Endeavor zotarolimus [ZES]) and newer DES (everolimus and Endeavor Resolute ZES) were 1.3 and 0.6 percent, respectively [27].

●In the Bern-Rotterdam registry, the annual rate of stent thrombosis was 0.4 to 0.6 percent for up to four years in a population of 8146 individuals who underwent PCI with either SES or PES [32].

In a subsequent report from the Bern-Rotterdam registry (cohort III), the incidence rate of very late stent definite or probable stent thrombosis (ARC definition) per 100 person-years in a cohort of 12,000 patients with unrestricted use of DES (3819 SES, 4308 PES, 4212 EES), for these three DES were 1.6, 2.4, and 0.6 percent, respectively, during follow-up of up to four years.

In the TWENTE trial, which randomly assigned nearly 1400 patients to either ZES or EES, the rate of very late definite or probable stent thrombosis (between 12 and 24 months) was 0.3 percent in both groups [35]. (See "Intracoronary stents: Stent types", section on 'Everolimus- versus zotarolimus-eluting stents'.)

Impact of ACS — The intracoronary and systemic prothrombotic environment accompanying an acute coronary syndrome (ACS) has led to concerns regarding a possible increase in the risk of ST compared to patients with stable angina. This concern has been confirmed in analyses of the ACUITY AND TRITON-TIMI 38 trial databases [36,37]. The range of rates of early stent thrombosis (<30 days) in patients with stable angina or an ACS according to the Academic Research Consortium definition (definite or probable) has been summarized as follows [38]:

●Bare metal stents: 0 to 0.5, 1.4 to 1.6, and 0 to 2.9 percent (stable angina, NSTEMI and STEMI, respectively).

●Drug-eluting stents: 0.3 to 0.4, 1.2 to 1.9, and 0 to 3.1 percent (stable angina, NSTEMI and STEMI, respectively).

A 2012 meta-analysis of individual patient data from 11 trials found that the long-term (cumulative) rate of stent thrombosis in patients with ACS was not significantly different between first generation DES and BMS (5.8 versus 4.3 percent, respectively) [39]. We interpret the available evidence to suggest that patients with ACS are at greater risk of stent thrombosis than those with stable disease regardless of stent type.

In a network meta-analysis of randomized trials comparing first-generation DES, second-generation DES, or BMS in primary PCI for ST-segment elevation MI, second-generation DES was associated with significantly lower incidence of definite stent thrombosis at one year (odds ratio [OR] 0.3; 95% CI 0.11-0.83) and MI (OR 0.3; 95% CI 0.17-0.54) and target vessel revascularization at one year (OR 0.54; 95% CI 0.80-0.98) when compared with BMS [40].

RISK FACTORS — All risk factors for stent thrombosis likely work through one or more of the mechanisms discussed above. All of these promote platelet aggregation. (See 'Mechanisms' above.)

Early and late stent thrombosis — The single most important predictor of early and late stent thrombosis is absence of platelet P2Y₁₂ receptor blocker therapy at the time of the event [11]. In one report, the hazard ratio for premature discontinuation of clopidogrel was greater than 50 at nine months [41].

The relationship between cessation of dual antiplatelet therapy (DAPT) and cardiac events after percutaneous coronary intervention was studied using data from 5018 patients in the PARIS registry [42]. At two years, the overall incidence of cessation for any reason was 57.3 percent; the rates of discontinuation (clinician recommended), interruption (stopping for less than 14 days for surgery), and disruption (due to bleeding or noncompliance) were 40.8, 10.5, and 14.4 percent, respectively. Compared with those on DAPT, the hazard ratio for major adverse cardiovascular events (a composite of cardiac death, definite or probable stent thrombosis, spontaneous myocardial infarction, or clinically indicated target lesion revascularization) was significantly higher only for those with disruption (1.50); it was significantly lower for those with discontinuation (0.63). It was nonsignificantly higher for patients with interruption (1.41, 95% CI 0.94-2.12). (See "Long-term antiplatelet therapy after coronary artery stenting in stable patients", section on 'Patients needing temporary discontinuation'.) The impact of disruption on the risk of stent thrombosis was greater in the first 30 days compared with after 30 days.

However, similar to many other preventative interventions, the optimal use of DAPT does not ensure that a patient will not experience stent thrombosis. For example, in an observational study of over 10,000 patients undergoing drug-eluting stent (DES) placement, those with early stent thrombosis were on DAPT 86 percent of the time and those with late stent thrombosis were on DAPT 57 percent of the time [43]. Similar findings were noted in the PARIS registry, in which 80 percent of the cases of definite or probable stent thrombosis occurred in patients taking DAPT [42]. In these patients, it is likely that one or more predisposing risk factors leading to a prothrombotic environment make it difficult for DAPT to be fully effective. (See 'Risk factors' above.)

The timing of late (events from one month to one year) or very late stent thrombosis (events after one year) in relation to discontinuation of one or both antiplatelet agents was evaluated in a review of 161 published cases [44]. (See 'Timing and incidence' above.) At the time of stent thrombosis, 19 patients were on aspirin and thienopyridine, 94 were on aspirin only, 33 had aspirin and thienopyridine stopped simultaneously, and 15 had thienopyridine stopped first and then had aspirin discontinued. The following findings were noted in the 142 patients who had discontinued at least one drug:

●In 33 patients who stopped both antiplatelet agents simultaneously, the median time to stent thrombosis was seven days.

●In 15 patients who had stopped a thienopyridine with no ill effect and then subsequently stopped aspirin, the median time to stent thrombosis was seven days from the discontinuation of aspirin.

●In the 48 patients who stopped both agents, 75 percent of cases occurred within 10 days.

●In 94 patients with prior discontinuation of thienopyridine but on aspirin, the median time to event was 122 days. In this group, only 6 percent of cases occurred within 10 days.

The causes of and predictors of premature discontinuation were evaluated in a study of 1622 patients who received at least one DES. In 234 patients (14.4 percent), one or both antiplatelet drugs (aspirin or clopidogrel) were discontinued for at least five consecutive days during the first year after implantation [45]. Clopidogrel, as opposed to aspirin, discontinuation occurred in the majority (11.8 percent). Among the 218 patients who discontinued any agent and in whom the reason for discontinuation was known, the following observations were made:

●The major causes of discontinuation were bleeding events or invasive procedures, medical decisions, or patient decision (50 percent, 32 and 18 percent, respectively).

●Among those who discontinued due to bleeding events/invasive procedures, about half of these discontinuations may not have been necessary, as the bleeding was minor or the upcoming procedure was not likely to be associated with major bleeding.

●In patients who discontinued due to bleeding events/invasive procedures, predictors included renal impairment, prior major hemorrhage, or peripheral artery disease. Long-term anticoagulant therapy, undergoing procedures in a private hospital, and not receiving instructions were causes in those in whom a medical decision was the cause. In patients who stopped of their own accord, being an immigrant, or using psychotropic drugs were predictors.

The best available data on the risk factors for early (≤30 days) definite stent thrombosis in real world populations come from two case-control studies:

●In analysis of the 21,009 patients in the Dutch Stent Thrombosis Registry, of whom 437 (2.1 percent) presented with an ST, the most important risk factors were premature discontinuation of clopidogrel therapy (odds ratio [OR] 36.5), under-sizing (OR 13.46), dissection (OR 6.19), TIMI flow post-PCI (OR 5.24), CAD ≥50 percent proximal of the culprit lesion (OR 4.15), malignancy (OR 3.06), no aspirin at the time of the procedure (OR 2.82), and left ventricular ejection fraction <30 percent (OR 2.71) [11].

●In an analysis of clinical, angiographic, and genetic factors of ST in 123 patients and 246 controls on dual antiplatelet therapy in a French registry, nongenetic independent correlates of early stent thrombosis (not included in the genetic marker list below) were diabetes mellitus (OR 1.82, 95% CI 1.02-3.24), use of proton pump inhibitors (OR 2.19, 95% CI 1.29-3.75), and higher clopidogrel loading doses (OR 0.73, 95% CI 0.57-0.93) [46]. Significant independent genetic markers were CYP2C19 metabolic status (OR 1.99, 95% CI 1.47-2.69) and the presence of either the ABCB1 3435 TT (OR 2.16, 95% CI 1.21-3.88) or ITGB3 PLA2 carriage (OR 0.52, 95% CI 0.28-0.95) genotype.

Potential risk factors have been proposed from a number of studies

[11,14-17,19,20,23,29,32,37,41,47-54]. The cause-and-effect relationship of some factors, such as left ventricular dysfunction, is less certain:

●Acute coronary syndrome (ACS) and proximal left anterior descending coronary artery lesion (for early stent thrombosis) [34]. (See 'Impact of ACS' above.)

●Side-branch stenting, diabetes mellitus, and end-stage kidney disease (for late stent thrombosis) [34].

●Incomplete stent expansion. (See "Percutaneous coronary intervention with intracoronary stents: Overview".)

●Greater stent length.

●Residual plaque burden and small stent area on intracoronary ultrasound.

●Small vessel caliber.

●Residual thrombus or persistent dissection after stent placement.

●Inflow or outflow obstruction.

●Subtherapeutic periprocedural anticoagulation.

●Left ventricular dysfunction [11].

●Nonionic contrast media.

●Cocaine use.

●Emergent stent placement.

●Prior brachytherapy.

●Post-procedure TIMI flow grade <3.

●No aspirin at the time of the procedure.

●Malignancy.

●CAD ≥50 percent proximal of culprit lesion.

●Treatment of bifurcation lesions.

●Multivessel disease, as shown in the SYNTAX trial [54].

●High on treatment (oral antiplatelet therapy) platelet reactivity [55], including polymorphisms in the genes controlling hepatic enzymes involved in the metabolism of clopidogrel. (See "Clopidogrel resistance and clopidogrel treatment failure", section on 'Definitions' and "Clopidogrel resistance and clopidogrel treatment failure", section on 'Potential explanations'.)

●The use of clopidogrel rather than prasugrel or ticagrelor in patients with ACS [38,56].

●The presence of uncovered stent struts (as detected by optical coherence tomography) [57].

●Chronic kidney disease.

●Stent overlap.

One study found the following risk factors for definite stent thrombosis in second-generation DES based on the timing of stent thrombosis [40]:

●Early stent thrombosis: Current smoking, left ventricular ejection fraction <40 percent, prior PCI, stent overlap, severely calcified lesion, left main coronary artery lesion, proximal left anterior descending lesion, and post-diameter stenosis ≥20 percent.

●Late stent thrombosis: Age <70 years, STEMI at presentation, hemodialysis, left ventricular ejection fraction <40 percent, in-stent restenosis, and severely calcified lesion.

●Very late stent thrombosis: Proximal left anterior descending lesion and in-stent restenosis.

Although stent location probably does not increase the risk of stent thrombosis occurrence, stent location in the left main or proximal left anterior descending artery poses increased risk of an adverse outcome if stent thrombosis occurs and may be regarded as an indication for more aggressive prevention strategies.

Gastrointestinal bleeding is a risk of DAPT, and routine proton pump inhibitors are recommended in patients at higher risk for bleeding. Some studies have raised the possibility that some proton pump inhibitors interfere with clopidogrel's ability to inhibit platelet function due to interference with conversion to active metabolite. This issue is discussed in detail elsewhere. (See "Gastrointestinal bleeding in patients undergoing percutaneous coronary intervention", section on 'Prevention' and "Antithrombotic therapy for elective percutaneous coronary intervention: General use", section on 'Bleeding' and "Aspirin for the secondary prevention of atherosclerotic cardiovascular disease" and "NSAIDs (including aspirin): Primary prevention of gastroduodenal toxicity".)

Very late stent thrombosis — The specific risk factors for very late stent thrombosis are less well defined but relate to delayed neointimal coverage and ongoing vessel inflammation [49,58]. A probable explanation for an increased risk of thrombosis in first generation DES compared with bare metal stent (BMS) after cessation of antiplatelet therapy is delayed arterial healing, particularly incomplete neointimal coverage. However, rates of very late stent thrombosis with newer generation DES, such as everolimus-eluting stent or RESOLUTE-zotarolimus-eluting stent, are lower than earlier generation DES and possibly lower than BMS.

Documentation of very late stent thrombosis has come from real-time imaging studies using either angioscopy or optical coherence tomography (OCT), from postmortem examination of stented segments in patients with and without stent thrombosis, or from the histology of in-stent restenosis tissue obtained by atherectomy.

An angioscopic study of 37 consecutive stented lesions (15 sirolimus stents and 22 BMS) showed that neointimal coverage is not complete in most patients with SES, but complete with all BMS at three to six months [25].

Delayed neointimal coverage in patients with SES was confirmed in a longer-term study of 17 SES and 11 BMS in which angioscopy was performed at 4, 10, and 21 months [26]. There were two main findings:

●Neointimal coverage was complete in all but one of the patients with BMS at 3.6 months and in all at 10 months. In contrast, the majority of patients with SES had incomplete coverage at 21 months

●Mural red thrombi were seen only in patients with SES and incomplete neointimal coverage

Angioscopy is limited in the evaluation of the extent of neointimal coverage since it cannot quantify the proportion or thickness of coverage. These limitations are not present with OCT, an intravascular imaging technique that visualizes the surface of a stent with higher resolution than intravascular ultrasound (IVUS) [59,60]. (See "Intravascular ultrasound, optical coherence tomography, and angioscopy of coronary circulation", section on 'Optical coherence tomography'.)

The mechanisms underlying very late DES thrombosis were explored using OCT in a study of 64 registry patients. In this study, very late stent thrombosis occurred at a median of 4.7 years [60]. The leading associated findings in descending order were malapposition, neoatherosclerosis, uncovered struts, and stent underexpansion. These associations were similar in early- and new-generation DES.

These findings are consistent with those in two postmortem studies and analysis of atherectomy specimens [49,58,61]. A postmortem study compared 28 stented segments with thrombus to 34 without thrombus in 46 patients who died more than 30 days after DES implantation [49]. The most powerful histologic predictor of stent thrombosis was incomplete endothelial coverage (uncovered struts).

Similar findings were noted in the second postmortem study in which delayed arterial healing was more pronounced in 14 DES patients with late stent thrombosis than in nine patients without stent thrombosis [58]. In addition, patients with DES had a greater degree of delayed arterial healing and less complete re-endothelialization than patients with BMS.

Current smoking and total stent length >28 mm have been identified as risk factors for very late stent thrombosis [34]. Whether these play a role by influencing the rate of arterial healing is not known.

In-stent neoatherosclerosis has also been seen in many cases of very late stent thrombosis. Neoatherosclerosis in stents, similar to native coronary vessels, is characterized by macrophage-mediated degradation of collagen with necrotic core formation that predisposes to plaque rupture and thrombosis. In BMS, neoatherosclerosis rarely occurs before three years and is more common in those greater than six years old [62]. In DES, neoatherosclerosis is accelerated and can be a cause of thrombosis in stents less than two years after implantation [63]. The frequency of neoatherosclerosis development appears similar for first and second generation DES but may be associated with fewer unstable features with second generation devices [64]. The factors that predispose to development of neoatherosclerosis are unknown but may involve dysfunctional endothelium.

Other issues related to risk

On-label versus off-label use — The randomized trials demonstrating the lower rate of restenosis with DES compared to BMS were performed in patients with single, previously untreated coronary lesions (on-label use). In the SIRIUS trial, for example, the exclusion criteria included ostial, bifurcation, and unprotected left main lesions, and vessels with thrombus or severe calcification; in addition, multivessel stenting was not performed and patients with a recent MI or a left ventricular ejection fraction (LVEF) less than 25 percent were excluded [65].

There has been concern whether the risk and timing of stent thrombosis (ST) may be different for these excluded or so-called off-label indications (more complex disease). At least two studies have suggested that the risk of ST in patients who received BMS or DES for off-label indications is higher than for those with on-label indications at both early and later time points [66,67]. This is not surprising, as the baseline risk of patients with off-label indications is higher. For example, these individuals often have an acute coronary syndrome.

In previous, smaller studies, each of the more common off-label indications was found to be associated with an increased risk of stent thrombosis. These include multivessel stenting [68], bifurcation stenting [41,69-72], particularly with use of the crush technique [71,72], chronic total occlusions, and in-stent restenosis [69]. (See "Percutaneous coronary intervention of specific coronary lesions" and "Intracoronary stent restenosis".)

Nonanatomic factors — A number of non-coronary clinical factors are also associated with stent thrombosis. These include diabetes mellitus [9,41,47,69], being a Black person [73], chronic kidney disease [41,69], acute coronary syndrome at presentation [9], cocaine use [74,75], elevated C-reactive protein levels [76], and, in rare cases, a hypersensitivity reaction to the stent [77]. (See 'Impact of ACS' above and "Periprocedural complications of percutaneous coronary intervention", section on 'Hypersensitivity reactions'.)

Incomplete stent apposition — Incomplete stent apposition, also known as stent mal-apposition, is defined as the absence of stent strut contact with the underlying vessel wall (not overlying a side branch) [78]. This phenomenon, which is identified by intravascular ultrasound (IVUS) or optical coherence tomography, can occur at the time of stent implantation; it may resolve or worsen. It may also occur late (up to years) and in this setting has been attributed either to positive remodeling (an increase in vessel diameter) whereby the vessel pulls away from the stent, or to plaque or thrombus dissolution, such that a space forms between the stent and the vessel wall.

Two IVUS studies support an association between late incomplete stent apposition and very late stent thrombosis in patients with DES:

●In a study of 13 patients with very late stent thrombosis and 144 controls, incomplete stent apposition was significantly more frequent in patients (77 versus 12 percent) than controls [78].

●In a study of 23 patients with DES and seven with BMS very late stent thrombosis, stent malapposition was seen in 74 percent of DES but none of the BMS patients [79].

COMPARISON OF DES AND BMS — Both randomized trial and observational study data have demonstrated that the cumulative rate of stent thrombosis is similar for bare metal and the first generation drug-eluting stents (SES and PES) at up to five years [11,80].

Similarly, the rate of early stent thrombosis is comparable between BMS and SES or PES. There may be a slight predominance of BMS ST between 30 days and one year (late ST) and a slight preponderance of DES stent thrombosis beyond one year (very late stent thrombosis), particularly for early generation DES such as SES and PES [30]. There have been no randomized trials comparing second generation DES, such as everolimus-eluting stents (EES) or RESOLUTE-zotarolimus-eluting stents (R-ZES), that were large enough to be sufficiently powered to determine the relative rates of stent thrombosis between BMS and EES or R-ZES.

The relative rates of stent thrombosis between DES and BMS have been compared in meta-analyses. We advise caution in the interpretation of their conclusions, due to the relatively short duration of follow-up and the fact that the patients included in the trials are representative of only a portion of all patients treated with coronary artery stents.

Two large meta-analyses are referred to in the sections below:

●A 2012 comprehensive network meta-analysis that included 49 randomized trials (over 50,000 patients) and compared one DES to a BMS or to another DES (table 1) [81]. The primary endpoint in this meta-analysis was the one year rate of definite stent thrombosis according to Academic Research Consortium (ARC) criteria (see 'Definitions' above).

The benefit of this network meta-analysis is that indirect comparisons between DES that have not been directly compared in a randomized trial, but have been compared with a third stent, can be made [82]. However, our confidence is not high that comparisons made in this meta-analysis are representative due to the concerns about non-contemporaneous comparisons and other potential sources of bias.

●A 2012 analysis (Bangalore) of 77 randomized trials (of SES, PES, EES, ZES, and ZES-Resolute) with over 57,000 patients [83].

Late and one-year cumulative stent thrombosis — The overall risk of stent thrombosis at up to one year is low for both BMS and DES as long as patients are continued on dual antiplatelet therapy with both aspirin and platelet P2Y₁₂ receptor blocker for the recommended duration [84,85]. (See 'Timing and incidence' above.)

Studies comparing early generation DES (SES or PES) to BMS suggested that the risk of stent thrombosis at up to one year was comparable [24,29,86].

In the 2012 network meta-analysis, the rates of late definite stent thrombosis of SES, PES, and ZES were comparable to BMS; however, the rate was significantly lower with the CoCr-EES compared to BMS (odds ratio 0.27) [81]. Since only two of the studies in the meta-analysis directly compared CoCr-EES to BMS, this finding should not be considered definitive. Changes in the design of BMS over time and other unmeasured factors may have influenced the indirect comparison of CoCr-EES to BMS in this study more so than with earlier generation DES. In the larger 2012 meta-analysis (Bangalore), similar findings were noted, including a lower rate with EES (56 percent lower) compared to BMS [83].

Very late and cumulative long-term stent thrombosis — Very late stent thrombosis includes events after one year, while the cumulative rate of stent thrombosis includes all events. The rate of very late stent thrombosis with DES between one and four years is approximately 0.6 percent with the protocol definition [1] and approximately 0.9 percent with the ARC definition [30]. (See 'Timing and incidence' above.)

Meta-analyses using the ARC definitions of definite or probable stent thrombosis have concluded that there is no significant difference in the cumulative rate (at two or more years) of stent thrombosis between older DES (either SES or PES) and BMS [1,30,84,86-88]. However, some [1,9,87,89] but not all studies [30,58,81] have suggested that the rates of very late stent thrombosis are higher with these first generation DES.

The two large 2012 meta-analyses evaluated the rates of stent thrombosis at two years [81,83]. In the network meta-analysis, EES (CoCr-EES) had a significantly lower cumulative rate of stent thrombosis compared to BMS (odds ratio 0.35, 95% CI 0.17-0.69) [81]. The Bangalore meta-analysis came to a similar conclusion [83].

COMPARISON OF DIFFERING DES — Not all types of drug-eluting stents (DES) have been directly compared. The evidence suggests that use of newer generation DES leads to lower rates of stent thrombosis compared to older DES. Most of the data are for the everolimus-eluting stent (EES). There is less information available for R-ZES, but it is possible they also have lower rates than older DES. Potential factors responsible for lower rates of stent thrombosis with newer DES include more bio-compatible polymers, newer stent designs with thinner struts, and lower drug dose.

Some evidence suggests that the rate of stent thrombosis with bioresorbable polymer DES (see "Intracoronary stents: Stent types", section on 'Bioresorbable polymer drug-eluting stents') is similar to that with current-generation durable polymer DES [90].

Late and cumulative one-year stent thrombosis — With regard to late definite stent thrombosis, EES of both types outperformed older E-ZES and PES, but not SES. With regard to cumulative one-year definite stent thrombosis (the primary composite outcome), the following significant odds ratios (OR) were noted (other comparisons between SES, PES, the two types of EES, and the two types of ZES did not achieve statistical significance) [81]:

●CoCr-EES versus PES 0.28 (0.16-0.48)

●CoCr-EES versus SES 0.41 (0.24-0.70)

●CoCr-EES versus Re-ZES 0.14 (0.03-0.47)

●CoCr-EES versus PC-ZES 0.21 (0.10-0.44)

●PC-ZES versus SES 1.92 (1.07-3.90)

Similar findings of a lower rate of any stent thrombosis with EES, compared to SES, PES, and ZES, were found in the Bangalore meta-analysis [83]. No significant differences in the rates of stent thrombosis at one year were found in paired comparisons between other DES.

Very late stent and cumulative two-year thrombosis — Some [91-93], but not all [30,86,94], studies comparing rates of stent thrombosis between SES and PES after one year found lower rates with SES.

The current body of evidence from randomized trials suggests that long-term cumulative rates of stent thrombosis are lower with EES than SES or PES. There is some evidence to suggest this is also true for R-ZES:

●In the comprehensive network meta-analysis of randomized trials discussed above [81], no significant differences in the rates of definite stent thrombosis comparing one DES (two types of EES, two types of ZES, SES, and PES) to another at two years with one exception: the odds ratio comparing cobalt CoCr-EES to PES was significant (0.34, 95% CI 0.19-0.62). A significant risk reduction remained when the analysis evaluated two-year definite or probable stent thrombosis. Similar findings were noted in the Bangalore meta-analysis of randomized trials [83].

●In a 2012 meta-analysis of 11 trials (16,775 patients) that compared the risk of two-year definite stent thrombosis between EES and SES (five trials), PES (five trials) or R-ZES (one trial), the risk of definite stent thrombosis with EES compared with pooled DES occurred in 0.5 percent versus 1.3 percent of patients (relative risk 0.38, 95% CI 0.24-0.59) [8]. This translated into a 0.6 percent absolute risk reduction in favor of EES [95]. Similar results were obtained using the broader definition of definite or probable stent thrombosis. On further analysis, the lower risk with EES was found for early, late, cumulative one-year, and very late events.

The finding of a lower rate of long-term stent thrombosis with EES compared to first generation stents is supported by observational studies:

●In a study of 12,339 patients who received SES, PES, or EES, the cumulative incidence rate of definite stent thrombosis during four years of follow-up was lower with EES compared with PES or SES (adjusted hazard ratios 0.33, 95% CI 0.23-0.48 and 0.41, 95% CI 0.27-0.62, respectively) [96]. These findings remained significant for the analysis of definite or probable stent thrombosis. Differences in the rates of stent thrombosis were due mostly to a lower risk of stent thrombosis beyond one year (very late stent thrombosis).

●In a 2012 report from SCAAR in which newer DES (EES and Resolute ZES) were compared to older DES (SES, PES, and, according to their classification, Endeavor ZES), the cumulative risk of definite stent thrombosis at two years was lower with newer DES (adjusted hazard ratio 0.57, 95% CI 0.41-0.79) [10].

In a 2012 prospective observational study of 6166 unselected patients who received either EES or SES, the rates of stent thrombosis at two years were similar (hazard ratio 1.16, 95% CI 0.47-2.84, comparing EES to SES) [97].Rates of stent thrombosis between EES and R-ZES have been evaluated in the RESOLUTE All Comers and TWENTE randomized trials, neither of which was powered to detect differences in these rates [98,99]. (See "Intracoronary stents: Stent types", section on 'Everolimus- versus zotarolimus-eluting stents'.) In these two trials, the rates of definite or probable stent thrombosis were low and not significantly different in the two groups at one year or later. In a pooled analysis of studies with the R-ZES including 7618 patients, the five-year rate of stent thrombosis was 1.2 percent with an annual rate of 0.1 percent for very late stent thrombosis.

In a study of 8791 patients who were randomly assigned to either E-ZES (an older version of ZES) or a SES, the rates of definite or probable stent thrombosis were similar at three years (1.4 versus 1.8 percent) [100].

SUMMARY

●Timing and incidence – Stent thrombosis is an uncommon but potentially life-threatening complication of the placement of an intracoronary stent. While it can occur at any time after stent placement, the majority of events occur within the first month irrespective of stent type. Stent thrombosis with bare metal stents (BMS) or drug-eluting stents (DES) has been observed as long as five years after stent placement. The presentation is often with MI, usually with ST elevation, or death. (See 'Introduction' above and 'Timing and incidence' above.)

●Risk factors – The premature cessation of dual antiplatelet therapy is the most important risk factor for ST. (See 'Risk factors' above.)

●Timing

•The cumulative rates of ST at one year are comparable between BMS and most DES (with the possible exception of everolimus-eluting stents [EES], for which the rate is likely lower). (See 'Late and one-year cumulative stent thrombosis' above.)

•At two years, the cumulative rate of ST is approximately 1.5 to 2 percent, and similar for both BMS and first generation DES. It is likely that this risk is lower for EES. (See 'Very late and cumulative long-term stent thrombosis' above.)

•The cumulative long-term risk of ST appears comparable among all DES, except for possibly higher risk for PES and lower risk for EES. Given potential differences between DES for the period beyond two years, longer-term follow-up is needed. (See 'Very late stent and cumulative two-year thrombosis' above.)

●Comparison of differing DES – Data are limited to make inferences regarding the risk of stent thrombosis for RESOLUTE-zotarolimus-eluting stents (R-ZES), but available data suggest the long-term risk is similar for R-ZES and EES. (See 'Comparison of differing DES' above.)