Percutaneous coronary intervention (PCI) is one of the most common procedures performed in US and European hospitals. In the United States, 954,000 PCI procedures were performed in 2010, and in over 90% of these procedures, at least 1 coronary stent was implanted.¹ Stent thrombosis (ST) is the most feared complication of coronary stenting and is characterized by rapid accumulation of thrombus within or adjacent to the stent. ST is typically associated with the abrupt onset of an acute coronary syndrome (ACS), manifesting as severe unstable angina, acute myocardial infarction, or sudden cardiac death. ST should be distinguished from progressive in-stent restenosis, resulting in eventual vessel occlusion, typically presenting as progressive exertional angina, but occasionally as ACS.²

In this chapter, we will review the classification of ST, the pathophysiology and predictors of ST, its consequences and treatment, and ongoing efforts to prevent this rare but major complication of percutaneous coronary revascularization. Data will be presented separately for bare metal stents (BMS; in aggregate for all types) and for first-generation (G1) and second-generation (G2) drug-eluting stents (DES). The G1-DES data refer specifically to paclitaxel-eluting stents (PES; Boston Scientific, Natick, MA) and sirolimus-eluting stents (SES; Cordis, Miami, FL), whereas G2-DES data reflect the experience accumulated predominantly with everolimus-eluting stents (EES; Abbott Vascular, Santa Clara, CA; or Boston Scientific), zotarolimus-eluting stents (Endeavor [E-ZES] and Resolute [R-ZES]; Medtronic, Santa Rosa, CA). All these DESs have durable polymers (Table 60-1). A special section is devoted to DESs with bioabsorbable polymers or polymer-free DESs and to bioresorbable vascular scaffolds toward the end of this review.

ST has been defined by the Academic Research Consortium (ARC) with respect to both its timing and certainty of its occurrence, in order to standardize interpretation of clinical trials and registries (Table 60-2).³ The ARC classification only considers events that occur after PCI is completed (after the patient has left the catheterization suite). More recently, an additional category of ST has been proposed—that which occurs after the stent has been implanted but before the PCI is completed, known as intraprocedural stent thrombosis (IPST). IPST is defined as the new appearance or worsening of preexisting thrombus within or adjacent to the recently implanted stent, with or without associated clinical symptoms. Diagnosis of IPST requires frame-by-frame angiographic analysis, usually in an angiographic core laboratory. The frequency of IPST is increased in patients undergoing stent implantation for ST-segment elevation myocardial infarction (STEMI), those with high white blood cell count, and patients receiving treatment for thrombotic or bifurcation lesions and with bivalirudin monotherapy (vs heparin plus glycoprotein IIb/IIIa inhibitors).⁴ IPST is an independent predictor of subsequent mortality, reinfarction, and out-of-lab ARC ST. IPST heralds a much worse outcome at both 30 days and 1 year, even when normal epicardial blood flow is achieved at the end of the procedure.^4,5

Before the introduction of DES in 2003, acute (<24 hours) and subacute (1-30 days) ST was noted in 0.9% of more than 6000 patients treated with BMS, with approximately 50% of the events occurring within the first day of stent implantation.⁶ The most important parameters associated with early ST were residual dissection, longer stent(s), and smaller minimal in-stent luminal diameter. Thus, most early ST was caused by inadequate deployment, which could be improved by more liberal use of intravascular ultrasound to avoid stent underexpansion and uncovered edge dissections or residual disease.⁷ The introduction of G1-DES did not reduce the incidence of early ST, especially if antiplatelet therapy is interrupted within the first 30 days.^8,9 Similarly, the risk of early ST with G2-DES was reported to be 0.5%, compared with 0.7% for both BMS and G1-DES (P = .074), in a large German registry of PCI.¹⁰

In contrast, incidence of late (1-12 months) and very late (>1 year) ST differs significantly among the different stent types. G1-DES are associated with a significantly higher risk of very late definite ST than either of the other types of stents (BMS or G2-DES), and it appears that this excess risk persists over many years after stent implantation. Among more than 18,000 patients treated in the German Registry, placement of G1-DES increased the risk of ST between years 1 and 3 nearly 5-fold compared with BMS or G2-DES.¹⁰ In fact, the risk of very late ST with G2-DES may be similar to BMS, although more studies are required in this regard to be certain of these observations.

Acute and subacute ST differ significantly from late and very late ST with respect to pathophysiologic mechanisms. For the former, a number of factors contribute, including technical issues such as stent underexpansion, untreated disease at the edges, and residual uncovered dissections; patient- and lesion-related characteristics, such as presentation with ACS, small vessel size, long lesion length, and plaque characteristics such as presence of thrombus; and the inherent thrombogenicity of the implanted stent, determined by its architecture and presence or absence of polymer and antiproliferative drugs.¹¹ All these variables further interact with the potency and continuity of dual antiplatelet therapy (DAPT), whose premature discontinuation appears to be the most powerful predictor of early ST.¹²

Late and very late ST is characterized by a number of pathologic processes and conditions: delayed arterial healing and endothelial coverage after DES; hypersensitivity reactions to the polymer (particularly relevant to SES); vessel toxicity with positive remodeling and fibrin deposition behind the struts (particularly relevant to PES); and bifurcation stenting, neoatherosclerosis,¹³ and strut fracture (which affects both BMS and DES). Neoatherosclerosis is a recently described phenomenon that underlies both restenosis and ST (Fig. 60-1).¹⁴ Neoatherosclerosis may be incited by the persistent inflammatory stimulus provided by the permanent presence of the drug-eluting polymer. In vivo studies of patients with ST using optical coherence tomography (OCT) highlighted the role of lipid-laden intima with (explaining clinical presentation as ACS) or without rupture, suggestive of neoatherosclerosis, as the underlying cause.¹⁵ In a series of 15 patients with ST, OCT identified an obstructive thrombus in all subjects, marked stent underexpansion (60% of predicted area), and neoatherosclerosis in one-fourth of cases.¹⁶ In contrast to early ST, interruption of DAPT plays a less important role in late ST (although it may be important if there is an additional trigger to increase platelet reactivity, such as trauma or the need for unplanned surgery). At least for G1-DES, the risk of late and very late ST appears to remain constant over 4 to 5 years from stent implantation, at a rate of 0.3% to 0.6% per year.^17,18

G2-DESs, particularly EES with durable fluorinated polymer, appear to be less susceptible to late and very late ST. In a pathologic analysis of patients (after excluding those with stent duration of implantation >3 years), morphometric assessment was available in 73 SES, 85 PES, and 46 EES stents, half of which were implanted for off-label indications.^19,20 Remarkably, the incidence of uncovered struts in each of these groups was 18.0%, 18.7%, and 2.6%, respectively (P <.0005), despite a similar average implant duration of approximately 9 months in the 3 groups. The 3 groups also had similar rates of restenosis underlying ST (12%-17%). Neoatherosclerosis was observed in 35%, 19%, and 29% of patients in the SES, PES, and EES groups, respectively (P = .91), while hypersensitivity reactions were seen exclusively with SES. The overall inflammation score was significantly lower for EES (0.26 vs 1.00 for each of G1-DES; P <.0005). It is important to emphasize that these data were obtained from autopsies, but only one-third of the deaths were stent related, whereas one-third were a result of noncardiac causes and the remainder were cardiac deaths unrelated to ST. Thus, only one-third of these patients fulfilled the criteria for definite ST according to the ARC definition (see Table 60-2).

Multiple observational series and randomized clinical trials have analyzed the independent predictors of ARC ST for BMS and DES, with models including demographic and baseline characteristics, procedural variables, and medication regimen and adherence.^8,12,21-25 Almost invariably, early discontinuation of DAPT—defined as <4 weeks for BMS and <3 to 6 months for G1-DES—was associated with a 5- to 90-fold increase in hazard of ST (Table 60-3). Other important predictors emerging in some of these analyses were chronic kidney disease, diabetes mellitus, bifurcation stenting, stent length, and intracoronary irradiation provided for BMS in-stent restenosis.

A more recent analysis of nearly 19,000 patients stented between 1998 and 2011 demonstrated that, using BMS as reference, G1-DES, diabetes mellitus, current smoking, prior myocardial infarction (MI), presentation with STEMI, complex lesion morphology, and increasing residual stenosis were independently associate with higher rates of definite ST up to 3 years from procedure. More than 50% of all events occurred within the first month.¹⁰

Following the first report of 4 cases of late and very late ST occurring after G1-DES,²⁶ Stone et al²⁷ performed a meta-analysis of 9 randomized clinical trials comparing SES with BMS (n = 1748 patients, 4 trials) or PES with BMS (n = 3513, 5 trials). The rates of protocol-defined ST up to 4 years were 1.2% versus 0.6% for SES and BMS, respectively (P = .20), and 1.3% versus 0.9% for PES and BMS, respectively (P = .30). The study highlighted the excess of ST events occurring beyond 1 year from PCI in the DES-treated patients, corresponding to approximately 1 additional event for approximately 500 patient-years. The majority of patients were not on DAPT at the time of the event. Notably, the incidence of death or MI was similar in DES and BMS patients. In this regard, the beneficial effects of DES in terms of preventing restenosis (which, when severe, may lead to MI) may offset the rare but serious consequences of ST.²⁸

In a consecutive series of nearly 6000 patients treated between 2007 and 2010 and followed for 2 years, definite or probable ST occurred in 1.9%. Three-quarters of the events occurred within 30 days, and only 16% occurred beyond 1 year. Cardiogenic shock, STEMI presentation, lack of DAPT, diabetes mellitus, and stent length and diameter were independent predictors of ST.²⁹

Recent data suggest that G2-DES, particularly EES, have lower rates of ST than G1-DES. In an analysis of 4 randomized trials of EES and PES (n = 6789), ST occurred by 2 years in 0.7% of EES patients and 2.3% of PES patients (P = .0001). The reduction in ST events with EES was apparent in all intervals—early, late, and very late. DAPT discontinuation beyond 6 months was associated with increased risk of ST in PES patients only.³⁰ Supporting these data are the results from the XIENCE V USA postapproval registry of 8061 patients. The incidence of definite or probable ARC ST was 0.8% at 1 year, and the independent predictors for its occurrence were discontinuation of DAPT before 30 days (hazard ratio [HR], 8.63; 95% confidence interval [CI], 2.69-27.73; P = .0003), chronic kidney disease, and total stent length.³¹ Black race may also be related to higher risk of ST, independent of medication compliance.³²

The most comprehensive analysis to date of the relationship between high on-treatment platelet reactivity (HPR) and ST is derived from the Assessment of Dual Antiplatelet Therapy with Drug-Eluting Stents (ADAPT-DES) study. Patients (n = 8583) with uncomplicated DES implantation had platelet function testing at an average of 20 hours after PCI using the VerifyNow device (Accriva, San Diego, CA). Half of the patients had ACS, and one-third had 3-vessel coronary artery disease (CAD) or left main trunk involvement. An average of 1.8 lesions per patient were treated using 1.7 stents. Nearly two-thirds of the stents used were EES, whereas 30% were G1-DES. At 1 year, the compliance was 84% with DAPT (aspirin and clopidogrel) and 95% with aspirin. Forty-two percent of patients had platelet reactivity units (PRU) >208, consistent with HPR on clopidogrel. Definite or probable ST occurred in 0.84% of patients by 1 year and in 1.07% of patients by 2 years. There was a significant excess of definite or probable ST in those with PRU >208, (HR, 1.98; 95% CI, 1.30-3.01; P <.0001; Fig. 60-2).³³ Eighty percent of the events were definite ST; very late ST accounted for only a fourth of all events. HPR was predictive for 2-year definite or probable ST (HR, 2.08; 95% CI, 1.22-3.17; P = .008) and even more so for definite ST (HR, 2.84; 95% CI, 1.48-5.48; P = .002), but was not an independent predictor of mortality at 1 or 2 years, reflecting possibly the interplay between higher bleeding risk and lower ischemic risk associated with more effective platelet inhibition.

At the clinical level, the PARIS Registry evaluated in detail the compliance of patients to DAPT as well the reasons and timing of its discontinuation. Among 5033 patients enrolled in 5 countries after successful DES or BMS implantation for any clinical indication, the rate of discontinuation of DAPT was 2.6% at 30 days, 11.8% at 6 months, and 19.9% at 1 year. As expected, the rate of physician-recommended discontinuation was different at each time point: 0.3%, 7.6%, and 12.3%, respectively. At 1 year, the rate of death was 2.2%, spontaneous MI 1.9%, ST 1.9%, and revascularization 4.5%. There was no significant difference in protocol-defined ST between patients on and off DAPT (1.9% vs 2.7%, P = .14), but the incidence of MI was higher in those off DAPT (4.8% vs 1.4%, P <.001). After adjusting for baseline characteristics, patients off DAPT were more likely to experience death (P = .002), MI (P <.001), and ST (P = .009), particularly in the first 7 days after PCI, and only when discontinuation was the result of disruption (interruption due to noncompliance or bleeding), rather than physician-recommended interruption.³⁴ A recent evaluation of DAPT disruption in the R-ZES program found that discontinuation beyond 1 month after PCI is associated with a very low risk of ST.³⁵

The duration of DAPT after stenting is dictated both by clinical presentation and by stent type. European and American professional guidelines concur that ACS patients should receive DAPT for up to 1 year regardless of whether PCI was performed or DES was implanted.^36,37 In contrast, for patients undergoing stent implantation for non-ACS indications, the duration of DAPT prescription varies with stent type. Consensus has evolved that BMS does not require more than 1 month of DAPT.³⁸ For DES, considerable uncertainty remains. Concerns of increased ST rates with G1-DES led to the recommendation (not founded on randomized trials) that at least 1 year of DAPT should be used. Furthermore, the ongoing hazard of very late ST with G1-DES has prompted the evaluation of longer DAPT therapy, even up to 36 months from PCI.

Several randomized trials have been performed to determine whether patients might benefit from DAPT usage after DES for longer than 1 year. The Optimal Duration of Clopidogrel Therapy With DES to Reduce Late Coronary Arterial Thrombotic Events (DES LATE) trial enrolled 5045 patients free of any complications 12 months after DES implantation and randomized them to aspirin alone or to DAPT for at least an additional year. Two-thirds of the patients received G1-DES, and the same proportion had ACS. The primary end point was a composite of cardiac death, myocardial infarction, or stroke 24 months after randomization. There was no difference in the incidence of the primary end point between the 2 groups (2.4% vs 2.6%, respectively; P = .75). Definite ST occurred in 0.5% and 0.3%, respectively (P = .34), whereas Thrombolysis In Myocardial Infarction (TIMI) major bleeding was noted in 1.1% and 1.4%, respectively (P = .20; Fig. 60-3).³⁹

These results confirmed a previous report from the same group.⁴⁰ The Prolonging Dual Antiplatelet Treatment After Grading Stent-Induced Intimal Hyperplasia Study (PRODIGY) randomized 2013 patients to 1 of 4 stents (EES, PES, ZES, or BMS) and to short (6 months) or long (24 months) DAPT. Aspirin was continued indefinitely in all patients. After excluding failed PCI and early complications, 1970 patients were randomized to the 2 DAPT regimens. Three-fourths of patients had PCI for ACS. The primary end point (all cause death, MI, or stroke) occurred in 10.0% and 10.1% of patients in the short and long DAPT groups, respectively (P = .98), and there was no interaction between the rate of events and clinical presentation or stent used (BMS vs DES). ST was not different between the groups (P = .80; Fig. 60-4). Type 2 or greater Bleeding Academic Research Consortium classification (BARC) bleeding was less common in the short DAPT group (3.5% vs 7.4%; P = .0002), with all the difference occurring, as expected, beyond 6 months from randomization.⁴¹ In a more detailed analysis of the event rates according to stent type, the investigators found that E-ZES patients had significantly better outcomes with shorter DAPT (P_INT = .004).⁴²