Expert Review of Cardiovascular Therapy
ISSN: 1477-9072 (Print) 1744-8344 (Online) Journal homepage: http://www.tandfonline.com/loi/ierk20
Long-term safety of drug-eluting stents Florian N Riede, Matthias Pfisterer & Raban Jeger To cite this article: Florian N Riede, Matthias Pfisterer & Raban Jeger (2013) Long-term safety of drug-eluting stents, Expert Review of Cardiovascular Therapy, 11:10, 1359-1378 To link to this article: http://dx.doi.org/10.1586/14779072.2013.837694
Published online: 10 Jan 2014.
Submit your article to this journal
Article views: 41
View related articles
Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=ierk20 Download by: [Central Michigan University]
Date: 06 November 2015, At: 00:20
THEMED ARTICLE y Stents + Devices
Review
Long-term safety of drug-eluting stents Expert Rev. Cardiovasc. Ther. 11(10), 1359–1378 (2013)
Downloaded by [Central Michigan University] at 00:20 06 November 2015
Florian N Riede, Matthias Pfisterer and Raban Jeger* Division of Cardiology, University Hospital, Basel, Switzerland *Author for correspondence: [email protected]
Stent implantation in coronary stenosis has revolutionized the treatment of coronary artery disease. The introduction of antirestenotic drug coatings further improved their efficacy in reducing target vessel revascularizations. With increasing use of drug-eluting stents (DES), stent thrombosis (ST) rose as potentially fatal major complication. Initially, the incidence of ST late after stent implantation seemed to be similar for DES and bare metal stents until several studies proved otherwise in first-generation DES. Since then, the design and components of DES have been changed and new polymers, drugs and different combinations of platelet inhibitors have been introduced to further improve the safety of DES. In this review, the authors focus on the relationship between DES, lesion anatomy, implantation technique and pharmacology to avoid the occurrence of ST. Furthermore, the relationship between dual antiplatelet therapy, bleeding rate and its significant impact on patient outcome is discussed. Finally, some promising future concepts are highlighted. KEYWORDS: drug-eluting stents • efficacy • future concepts • major bleeding • pathophysiology • risk factors • safety • stent thrombosis
The introduction of balloon coronary angioplasty, first performed by Andreas Gru¨ntzig on 16 September 1977 in Switzerland [1], was the birth of interventional cardiology. Percutaneous transluminal coronary angioplasty (PTCA) nowadays is one of the most frequently used major interventions in medicine. Soon after using balloon angioplasty on a big scale, it became evident that this novel technique harbors several risks and possible complications. Acute vessel closure was recognized as a pivotal complication in the early phase after intervention. Besides elastic recoil of the dilated vessel accounting for 40–50% of all cases, the occurrence of dissection, subintimal hemorrhage, platelet-mediated thrombus formation and vasoconstriction were reasons for acute vessel closure early after PTCA [2]. The process of slow narrowing of the vessel lumen was called restenosis manifesting later after PTCA. The pathophysiologic explanation for this phenomenon is an active response of the dilated vessel to the injury of the wall. In a self-perpetuating process, media muscle cells reallocate to the intima and start producing extracellular matrix, resulting in a narrowing of the vessel lumen [3]. Gru¨ntzig himself, one of the most experienced interventional cardiologists in the early 1980s of the last century, reported on a recurrence rate of 30% occurring in the first 6 months after the www.expert-reviews.com
10.1586/14779072.2013.837694
procedure [4]. The development of coronary stenting, which was introduced by Sigwart et al., was a treatment against acute dissections, elastic recoil of the vessel wall and occasionally arterial spasm after balloon angioplasty [5,6]. The stent acts as scaffold supporting the vessels’ integrity and by this counteracts acute vessel closure. However, it had also impact on restenosis rate. Two landmark studies comparing the effects of stent placement with balloon angioplasty, the BENESTENT trial [7] and the STRESS trial [8], both published in 1994, showed a lower rate of restenosis after 7, respectively, 6 months (22 vs 32% in the BENESTENT trial and 32 vs 42% in the STRESS trial). The acceptance of stenting as therapy for CAD was further increased, when evidence rose that stenting with combined platelet inhibition without anticoagulation was safe [9,10]. In an attempt to reduce in-stent neointimal hyperplasia, stents were coated with antiproliferative drugs, with the so-called drug-eluting stents (DESs) being the next breakthrough invention in interventional cardiology. Design of DES
Despite many inventions and modifications of the composition of stents, most modern DES consists of three parts: a stent body serving as
2013 Informa UK Ltd
ISSN 1477-9072
1359
Review
Riede, Pfisterer & Jeger
scaffold; a polymer covering the metal and serving as carrier for drugs; and a drug coating.
Downloaded by [Central Michigan University] at 00:20 06 November 2015
Stent body
Stent scaffolds differ in many engineering aspects and can be classified in a simplified way by materials and fabrication processes used; the mechanism of stent expansion (self-expansion, balloon expansion); and stent configuration and design [11]. The raw material mainly used determines the method of fabrication. Wires can be formed into stents by knitting, coiling and braiding. The self-expanding Wall Stent, for example, used a braided design and was made of multiple elgiloy wires (i.e., a cobalt– chromium alloy). On the contrary, the vast majority of modern balloon-expandable stents nowadays are produced by laser cutting from tubings. They are usually made of metal alloys that allow for thinner stent designs while preserving radial strength and radiopacity including materials such as stainless steel, cobalt–chromium and platinum–titan alloys. The raw material and its fabrication process determine the possible stent geometries: wires can be processed into coils, helical spiral and braided designs, tubes can be typically transformed into sequential ring geometries [11]. The latter accounts for most of the stents in use nowadays. It comprises a series of expandable seven-shaped structural elements, the so-called ‘struts’, joined by ‘bridges’ or ‘nodes’. In relation to the manner the elements are connected, two different cell configurations are possible: the closed cell and the open cell design. In a closed cell design, all internal infliction points of the struts are connected by bridges. The Palmaz-Schatz Stent was the first slotted-tube (a closed cell design) stent in clinical practice that confirmed the superiority of stenting versus balloon angioplasty in the treatment of CAD [12]. Like all stents with a closed cell design, the Palmaz-Schatz stent showed superior resistance to deformation and high radial strength but lacked flexibility [13]. The newer open cell or multicellular designs have less strut–strut intersections because not all of the internal inflection points of the struts are connected by bridges. They provoke less vascular injury and less intimal hyperplasia [14]. Strut thickness plays a pivotal role concerning the immediate and the longterm performance of stents. Thicker struts promise more radial support, better radiopacity and better wall support but show worse long-term efficacy concerning restenosis. This was shown in the ISAR-STEREO [15] and the ISAR-STEREO2 trials [16]. In the ISAR-STEREO trial, different stents from the same manufacturer with similar design, varying only in strut thickness (0.05 mm vs 0.14 mm), were randomly implanted in 651 patients. At 6 months, the late loss in large coronary arteries (>2.8 mm) was more pronounced in the thick-strut stent (Duet stent, Guidant) than in the thin-strut stent (Multilink stent, Guidant) (1.17 vs 0.94 mm, p = 0.001). Longitudinal stent deformation, defined as the shortening of a stent in the longitudinal axis after successful deployment, has recently been recognized as potential cause of stent malapposition and intraluminal protrusion of stent struts [17]. Overall, it occurs rarely, more frequently in ostial or bifurcation lesions and with multiple stents and it raises the risk of restenosis and ST [18]. 1360
Stent polymer coverage
To deliver a therapeutic drug at the site of the treated stenosis, it has to be brought to the stents’ surface. Drugs can be loaded either directly onto the struts (e.g., via impregnation onto microporous stent surfaces via solvent evaporation) or on a carrier vehicle matrix that is added to the device surface [19]. The carriers are usually polymeric molecules, which act like a sponge by increasing the surface area thereby enabling sufficient loading of the applied drug. The first polymers were durable. These polymers have been exclusively used in the firstgeneration DES (e.g., Taxus, Boston Scientific; Cypher, Cordis). The most commonly used permanent polymers are poly-n-butyl methacrylate, polyethylene-co-vinyl acetate and the tri-block copolymer poly(styrene-b-isobutylene-b-styren) (SIBS). By adding an antirestenotic drug together with one of these materials, a drug–polymer matrix can be formed and applied to the struts of the stent. Permanent polymers have been shown to trigger chronic hypersensitivity [20] and inflammation [21] reactions and thus induce restenosis and possibly ST. This bioresponsiveness of the persistent environment to the polymer ultimately determines the risk for adverse events. Biodegradable polymers have been developed in order to overcome hypersensitivity reactions leading to ST. Examples of biodegradable polymers are poly-L-lactic acid (PLLA), poly-L-lactide-cocaprolactone and poly DL-lactide-co-glycolide. They are degraded by hydrolysis and enzymatic activity and will leave a bare metal stent (BMS) after they have vanished through degradation. The noninferiority of a custom-made biodegradable polymer used in a rapamycin-eluting stent-platform has recently been shown in the ISAR-TEST 4 [22] and in the LEADER trials [23]. Polymer-free stent platforms have been introduced in recent years to overcome the undetermined effect of polymers on arterial wall healing, causing inflammation (even through degradation when using biodegradable polymers). In a randomized study, a custom-made polymer-free stent platform could recently prove its efficacy and safety over a period of 5 years in the ISAR-TEST three trial [24]. Drug coating
The first immunosuppressive agents used for a drug coating around the metal scaffold were sirolimus (previously called rapamycin) and paclitaxel. Sirolimus is a member of the limus family, that is, immunomodulators that inhibit the mammalian target of rapamycin. There are newer members of this family used in DES such as everolimus, zotarolimus and biolimus A9. By decreasing the positive and increasing the negative regulators of the cell cycle through interacting with the Cyclin Kinase pathway [25], the targeted cell will be arrested at the G0/ G1 phase in a cytostatic way. Cell migration and proliferation will both be inhibited. On the other hand, paclitaxel is a potent antimigratory and antiproliferative agent, acting through stabilization of cellular microtubules. Thus, it inhibits components of the cytoskeleton and mitotic spindle and impacts primarily the M-phase of the cell cycle. It works rather cytotoxic by leading to apoptosis and cell death. Both drugs mentioned Expert Rev. Cardiovasc. Ther. 11(10), (2013)
Long-term safety of DES
above, sirolimus and paclitaxel, work in DES as active coating. To deliver the drug at the site of the target lesion, it has to be bound to the stents surface, mostly by a polymer. Stent performance
Downloaded by [Central Michigan University] at 00:20 06 November 2015
The performance of a stent can be assessed by differentiating between its safety and its efficacy. Clinical trials designed to evaluate new percutaneous coronary intervention (PCI) devices; for example, new DES are designed to assess their efficacy and safety in order to allow a comparison with other devices. Thus, efficacy and safety can be regarded as the yin and yang of interventional cardiology, especially of DES. Efficacy
Review
to the reduced rates of restenosis in modern BMS. Stents with bioactive coatings, the so-called ‘bioactive stents,’ have been designed to enhance device integration with the surrounding biology. Titanium–nitride oxide covered bioactive stents demonstrated reduced rates of restenosis and platelet adhesion in a porcine model [40]. They showed a low incidence of major adverse events at 12 months in a real-world registry [41]. In comparison to modern zotarolimus-eluting stent (ZES), they remained inferior in a first randomized study [42] in terms of restenosis and late loss. Safety
Soon after introducing DES in clinical routine, there were first reports about abrupt vessel occlusion at the stent site potentially from a thrombotic genesis [43]. Late ST, a life-threatening complication of this technology, has emerged as a major concern in recent years. ST already known as complication after intracoronary brachytherapy [44,45], and its pathogenesis has been postulated to be related to delayed reendothelialization of the treated lesion (FIGURE 1) [46,47]. For a long time, it was thought that the benefit of DES would outweigh these risks; however, data from meta-analyses [48–50] and observational studies [51,52] questioned this paradigm. To prevent ST, prolonged dual antiplatelet therapy (DAPT) was recommended by the authorities and societies. Higher rates of major bleeding were encountered with DAPT,
Stents have been developed to overcome dissections following percutaneous transluminal coronary angioplasty (PTCA) [26]. Albeit BMSs have been demonstrated to reduce restenosis significantly more than balloon angioplasty, in-stent restenosis (ISR) occurred in about 10–60% of cases [7,8,27]. In the BENESTENT-1 trial, the initial 10% difference in target lesion revascularization (TLR) between the stent and PTCA groups remained unchanged at 5 years follow-up (17.2 vs 27.3%), emphasizing the prolonged benefit of BMS versus PTCA [28]. As a possible mean to prevent the development and to inhibit the recurrence of ISR, radiation therapy was used in coronary arteries (brachytherapy). In a metaanalysis by Oliver et al., assessing the performance of DES in comparison to intravascular brachytherapy, DES SES PES ZES EES BMS reduced the rate of revascularization (OR 0.51, 95% CI: 0.36–0.71) and binary restenosis (OR 0.57, 95% CI: 0.40–0.81) compared to brachytherapy, which itself demonstrated to be much more effective than BMS or PTCA alone [29]. Due to these results, the firstgeneration DES with polymer-based paclitaxel (Taxus, Boston Scientific) and sirolimus (Cypher, Cordis Corporation) coating quickly became the gold standard for interventional treatment of coronary artery disease [30–32]. DES showed a relative reduction in restenosis rate of 70–85% in comparison to BMS and an absolute rate of restenosis and target vessel revascularization (TVR) of under 10% [30–38]. In the more recent HORIZONS-AMI study, a newer PES was compared with an otherwise identical BMS from the same Figure 1. Scanning electron micrographs of 14-day endothelial coverage of manufacturer [39] in patients with Cypher (SES), Taxus Liberte´ (PES), Endeavor (ZES), Xience-V (EES) and Multi-Link STEMI. Compared with BMS, PES (BMS) in rabbit iliac arteries. reduced clinically driven target lesion BMS: Bare metal stent; EES: Everolimus-eluting stent; PES: Paclitaxel-eluting stent; revascularization at 12 months from SES: Sirolimus-eluting stent; ZES: Zotarolimus-eluting stent. 7.4 to 4.5% (p = 0.003), thus pointing Reproduced with permission from [47]. www.expert-reviews.com
1361
Review
Riede, Pfisterer & Jeger
thus bleeding arose as another safety aspect with the use of DAPT. Furthermore, it was widely debated, whether DES would lead to lower cardiac and all-cause mortality. Safety issues after stent implantation
Downloaded by [Central Michigan University] at 00:20 06 November 2015
All-cause & cardiac mortality of DES
Advances in interventional cardiology have led to a marked decline in post-procedural mortality, leading to a shift of the timing of mortality from the index PCI to later time points. The investigators of the EVENT registry, analyzing prospectively the contribution of cardiovascular death to 1-year mortality after PCI in 10,144 patients, found that overall 3% of patients died within the first year, 17% within the first month, 83% between 31 and 365 days after index PCI [53]. Cardiovascular death (CVD) was the predominant mode of death during the first 30 days by a factor of four. After the first month, the likelihood of non-CVD or death from unknown causes was very similar to the likelihood of CVD. Patients who survived the first year after index PCI received more often DES (88 vs 77%, p < 0.01) than deceased patients. Data from a singlecenter registry demonstrated significantly lower all-cause mortality with DES (8 vs 17%, p < 0.001; HR: 0.62, 95% CI: 0.53–0.73; p < 0.001) over a 4.5-year observation period, especially after adjustment for potential confounding factors [54]. The large Australian MIG registry reported an overall mortality rate at 30 days of 2.1% (79% cardiac), at 12 months of 3.9% (61%) and at long-term (3.2 ± 0.5 years) of 8.2% (50%) in a higher risk cohort [55]. DES deployment appeared protective against late mortality (HR: 0.85, 95% CI: 0.73–0.99; p = 0.04). After the first month, there was no difference in mortality rates between those who received a DES and those who received a BMS. The mortality benefit of DES over BMS was preserved in all age groups in a Dutch study, stratifying patients in quintiles based on age [56]. The recently published SORT OUT V trial showed similar rates of all-cause mortality (2.4 vs 2.2%, p = 0.67) and cardiac death (1.0 vs 1.1%, p = 0.71) at 12 months in biodegradable biolimus-eluting stents versus permanent polymer serolimus-eluting stent (SES). Thus, modern DES is safe in terms of mortality, although occurrence of very late ST remains a safety concern [57]. ST Definition & temporal patterns of ST
Clinically, ST are typically encountered as acute coronary syndromes consisting of acute chest pain in combination with ischemic ECG changes and raising heart enzymes (e.g., troponines) [58]. In order to categorize ST, it is important to differentiate several time segments relative to the index procedure. Widely used in clinical trials are the Academic Research Consortium (ARC) consensus criteria [59]. They have been issued to provide consistency in the reporting on the occurrence of safety endpoints of future stent trials and contain arbitrary chosen time intervals. Most ST events take place in the first 30 days after stent implantation and are defined as early ST. They are thought to 1362
happen mostly due to procedural or technical characteristics and compliance with dual antiplatelet therapy and are further stratified into acute ST (within the first 24 h) and subacute ST (24 h to 30 days). Late ST events are defined by occurring in the time interval between 30 days and 1 year after stent implantation and are less frequent. Very late STs are defined by the ARC criteria as occurring beyond 1 year. This interval has been questioned by various trials. In BASKET-PROVE, for example, late events were defined as those occurring 7–24 months after the intervention, on the basis of BASKETLATE data and on angiographic studies of the timing of restenosis [60]. From a pathophysiological point of view, late and very late ST are the most complex complications of stents. Their different risk factors and causes are discussed below. The ARC criteria define varying degrees of certainty in a tri-level classification. Definite ST requires angiographic or autoptical confirmation of ST. Probable ST is considered in cases with any unexplained death within a month after the stent implantation or with occurrence of acute ischemia in the territory of the implanted stent any time after the index procedure. Possible ST defines the weakest level of certainty and refers to any unexplained death 30 days or more after intracoronary stenting. Despite being neglected in many stent trials, there is evidence that possible ST represents true ST in over 50% of the cases [61]. Incidence of ST
Already during the BMS era, late ST had been observed. Two registries [62,63] reported rates of late ST of 0.76 and 0.3%, respectively, representing nearly one-third of overall ST events (average 1.6%). Early anecdotal reports of ST several months after DES implantation, describing late total occlusion soon after interruption of ticlopidine treatment [43], could be substantiated by subsequent trials. Iakovou et al. analyzed the incidence of ST in 2229 consecutive patients who underwent successful implantation of sirolimus- or paclitaxel-eluting at 9-month follow-up [64]. About 29 patients (1.3%) suffered from ST (9 [0.8%] with sirolimus and 20 [1.7%] with paclitaxel; p = 0.09). The post hoc analysis of the BASKET-LATETrial [51], one of the first trials comparing DES with BMS in a ‘real-world’ setting, showed an increased rate of thrombosisrelated events (2.6 vs 1.3%) in DES as compared with BMS in the 12 month following after the discontinuation of clopidogrel. The hypothesis of a higher risk of ST with DES was controversely debated and examined in subsequent trials and metaanalyses, which found a comparable risk of ST in DES and BMS [65–67]. The long-term risk of ST in DES was further analyzed in a 4-year cohort study assessing 8,146 patients treated with DES (PES and SES). The investigators reported an incidence of 1.0/100 patient-years and a cumulative incidence of 3.3% at 4 years follow-up [68]. The hazard of late ST (30 day to 1 year) amounted to 0.46% and the hazard of very late ST (1–4 years) to 0.57% per year. A large meta-analysis [69], comprising data from 30 studies (221,066 patients, 4276 ST) with DES used in 87% showed definite (probable) late ST in 0.5% Expert Rev. Cardiovasc. Ther. 11(10), (2013)
Downloaded by [Central Michigan University] at 00:20 06 November 2015
Long-term safety of DES
Clinical sequelae after ST
ST are serious complications after stent deployment manifesting as myocardial infarction and cardiac death. In 611 patients of the Japanese RESTART registry evaluating ST after SES implantation, the clinical sequelae differed among those with early ST, compared with those with late ST and very late ST [65]. Mortality rate at 1 year after ST was significantly higher in early and late ST (22.4 and 23.5%) compared with very late ST (10.5%). In clinical presentation, cardiac arrest took place two-times more often in early ST (11%) than in late ST (6%) and very late ST (5.8%). In a retrospective analysis of 431 cases with a first definite ST, the primary composite endpoint of cardiac death and recurrent definite ST occurred in 27.9% after a median follow-up of 27.1 months [72]. About 20.1% suffered a recurrent definite or probable ST, similarly 21.3% suffered a myocardial infarction at the longest available follow-up. Noteworthy, neither the category of ST (early versus late) nor type of the previously implanted stent (DES or BMS) did affect the clinical outcome. The timing-dependent differences in clinical presentation between early, late and very late ST suggest possible differences in the pathophysiological mechanisms responsible, which will be discussed further in this review. Pathophysiology: risk factors and causes for late and very late ST
Numerous risk factors and causes for ST have been identified factors (TABLE 1) [73]. Of note, data from large-scale registries demonstrate that the multivariate predictors of ST change during follow-up [68,70]. Procedural factors and lesion characteristics mostly account for ST in the first 30 days after the implantation of DES. Stent factors and patient factors mainly modulate the response of the vessel wall toward the artificial scaffold implanted. So these two risk factors are of importance in the genesis of late ST. www.expert-reviews.com
7 6 Cumulative incidence (%)
(0.4%) and very late ST in 0.6% (0.4%) after a median of 22 months. In this analysis, definite (probable) acute ST occurred in 0.4% (0.2%) and subacute ST in 1.1% (1.0%), suggesting that the majority of ST occurred within the first month. Similarly, in the Dutch ST registry of over 21,000 patients, most cases (>70%) of ST occurred within 30 days of the procedure [70]. Kimura et al. reported on the incidence of ST in the 5-year outcome data of the j-Cypher Registry [71]. Cumulative incidence of definite ST in 12,812 patients undergoing SES implantation was low (30 days: 0.3%; 1 year: 0.6%; and 5 years: 1.6%). Surprisingly, late and very late ST continued to occur without attenuation up to 5 years after sirolimus-eluting stent implantation (0.26% per year) representing continuous hazards for the patients. This was also demonstrated in the Bern-Rotterdam registry by showing a continuous hazard of ST until the end of the observation period (4 years) (FIGURE 2) [68]. Thus, the data of the registries suggested that the pathophysiology of very late ST may not be the same as early and late ST because the predictors for very late ST were quite different from those for early and late ST.
Review
5.7% 5
Definite and probable ST
4 3.3%
3
Definite ST
2 1 0 0
Months
1
2 3 Time since PCI (years) 1
Cumulative incidence 1.2 definite ST (%) Cumulative incidence 3.7 probable and definite ST (%) Patients at risk (n) 7537
4
12
24
36
48
1.6
2.1
2.7
3.3
4.1
4.6
5.2
5.7
7209
5157
2747
1051
Figure 2. Cumulative incidence of definite stent thrombosis in 8146 patients during a 4-year follow-up period. PCI: Percutaneous coronary intervention; ST: Stent thrombosis. Reproduced with permission from [1].
Stent factors: delayed & disturbed endothelial repair
Virmani et al. were among the first to report on incomplete stent strut coverage with focal platelet aggregates and persistent fibrin deposits within the necrotic core 16 months after SES implantation [74]. Although neointimal thickening was minimal at the sites of strut coverage, poor endothelial cell growth of the noncovered sites was documented in all such cases. In a number of autopsy cases, a lack of endothelial cell strut coverage was reported as the best correlate of late ST [75]. The measured parameter that best correlated with endothelialization was the ratio of uncovered to total stent struts per histological section. The odds ratio estimated for late ST in lesions with a rate of uncovered stent struts >30% was 9.0 (95% CI: 3.5–22.0). When neointimal thickness in thrombosed and patent DES lesions were compared, there was a marked shift toward less neointimal growth in DES with mural thrombus formation. In BMS, however, near-complete endothelialization has been postulated within 3–4 months. It was uncertain whether reendothelialization with DES is only delayed or persistently incomplete up to late time points. A recent autopsy study revealed that both SES and PES implantation sites heal over time nearly completely [76] in DES placed for >12 months with confirmed on-label use, whereas off-label indications for both stents resulted in incomplete healing, in DES even more than 1 year after implantation. Kubo et al. demonstrated significant differences in strut coverage between patients with unstable angina and stable angina in an optical coherence tomography (OCT) study. Markedly, less SES struts were fully covered 1363
Review
Riede, Pfisterer & Jeger
Downloaded by [Central Michigan University] at 00:20 06 November 2015
Table 1. Selected multifactorial causes of stent thrombosis. Factors
Causes
Stent factors
Incomplete endothelialization Hypersensitivity to drug coating or polymer
Patient factors
PCI for acute coronary syndrome/ST-segment elevation myocardial infarction Antiplatelet therapy and individual response Prior brachytherapy Comorbidities: diabetes mellitus, renal failure, malignancy Impaired left ventricular function
Lesion characteristics
Lesion/stent length Vessel/stent diameter Lesion complexity (bifurcation lesions, chronic total occlusions) Saphenous vein graft lesion
Procedural factors
Inadequate stent expansion/sizing Incomplete stent apposition Stent deployment in necrotic core Residual edge dissection
(28 vs 63%, p = 0.019) and more struts only partially covered (72 vs 37%, p = 0.019) by neointima in patients with unstable angina compared with those with stable angina. These findings were consistent with those of an autopsy study from Nakazawa et al. [77]. Stent factors: hypersensitivity reactions, chronic inflammation
Since October 2003, when the US FDA issued a warning on its website notifying physicians of possible generalized hypersensitivity reactions in 50 patients receiving SES, hypersensitivity reactions in association with DES came into focus of research. In 2004, the case of a 58-year-old man who died of late ST 18 months after having received two SES was reported. His coronary angiogram already showed vessel enlargement 8 months after implantation. An autopsy showed aneurysmal dilation of the stented arterial segment with a severe localized hypersensitivity reaction [78]. This case report enforced rumors about the allergic potential of DES. It was speculated that the polymers carrying the drugs would trigger an inflammatory response to DES implantation rather that the drug itself because an extensive eosinophilic infiltration specifically prominent around stent struts was found, whereas the proximal part of the stented artery also showed focal giant cell reaction surrounding a few polymer remnants that had become separated from the stent struts. Since autopsy studies are limited to postmortem specimens and therefore lack generalizability and coronary angiography and intravascular ultrasound (IVUS) do not delineate morphological changes of the vessel wall in response to DES implantation in sufficient detail, Cook et al. analyzed thrombus aspirates and then performed an IVUS examination in 28 patients presenting with very late ST and in 26 controls [79]. By combining intravascular imaging with histopathological analysis, they could demonstrate that incomplete stent apposition area was associated with an increase of total eosinophil count and fraction of white blood 1364
cells (WBCs). Furthermore, in specimens from very late DES ST, eosinophilic infiltrates amounted to 10% of WBCs. Conversely, in patients with spontaneous acute MI, early and late ST in BMS and early DES ST, only 1–3% of WBCs were eosinophils, resembling the proportions in the peripheral circulation. Thus, one can hypothesize that isolated components of DES give rise to a delayed-type hypersensitivity reaction (type IV) in certain patients that can finally lead to late, respectively, very late ST by fibrin and platelet deposition through inflammatory changes or by promotion of the coagulation pathway. Stent factors: stent thrombogenicity
Chronic inflammatory processes as described above can create a prothrombotic environment that can ultimately lead to late/very late ST. So far, stent thrombogenicity has been mainly attributed to the polymers used on DES. However, the antirestenotic drugs themselves may exert a prothrombotic effect in the targeted vessel as well. This has been demonstrated for the two first widely used drugs in DES, that is, sirolimus and paclitaxel [80]. They enhance endothelial tissue factor expression, which is a cell surface receptor for coagulation factor VII that serves the main activator of the coagulation cascade activating factors IX and X. Paclitaxel and sirolimus are both highly lipophilic drugs that easily penetrate into endothelial and underlying cells of the vessel wall. Through intracellular retention, they can act in a prothrombotic fashion by inducing tissue factor expression after deployment of SES or PES [81]. Patient factors: dual antiplatelet therapy
Premature discontinuation of antiplatelet therapy is one of the most important predictors of ST. After the ‘ESC DES fire storm’ in 2006, the general awareness of late ST being related to significant morbidity contributed to the shift from 6 to 12 months of DAPT after DES implantation as recommended by the FDA for safety reasons [82]. The current ACC/AHA guidelines from 2011 recommend 12 months of DAPT for DES [83], whereas the ESC guidelines from 2010 state that convincing data exist only for DAPT up to 6 months [84] in elective interventions. The Bern-Rotterdam registry, cited by the ESC committee, demonstrated in over 8000 patients that there is no difference in those treated for 3–6 months versus those treated for 1 year [85]. On the other hand, it suggested that prolongation of DAPT from 12 to 24 months would improve cumulative survival. The recently published PRODIGY trial randomized 2013 patients 30 days after stenting to 6 versus 24 months of clopidogrel-based DAPT [86]. The trial showed no benefit of prolonged DAPT in the incidence of ischemic events but noted an increased risk of major bleeding (HR 2.17; Expert Rev. Cardiovasc. Ther. 11(10), (2013)
Downloaded by [Central Michigan University] at 00:20 06 November 2015
Long-term safety of DES
95% CI: 1.44–3.22; p = 0.00018). In the EXCELLENT trial, 1443 patients receiving DES for stable CAD or acute coronary syndromes were randomized to 6 versus 12 months of DAPT with clopidogrel in combination with continued aspirin treatment [87]. Rates of target vessel failure at 12 months were 4.8% in the 6-month DAPT group and 4.3% in the 12-month DAPT group. The authors concluded that 6-month DAPT did not increase the risk of target vessel failure at 12 months after implantation of DES compared with 12-month DAPT. In an observational study examining 4666 consecutive patients receiving intracoronary stents (BMS and DES), the continued use of clopidogrel was a predictor of lower rates of death after 24 months in patients who were event free at 6 months (2.0% with vs 5.3% without, p = 0.03) and at 12 months (0% vs 3.5%, p = 0.004) [88]. These ‘real life’ data point to a possible benefit of extended duration of DAPT in DES in event-free (no death, myocardial infarction or revascularization) patients. Since the occurrence of major bleeding influences overall survival (see below), the analysis of event-free patients, alive at landmark analysis, could have an impact on survival by selection bias. Maybe by using a patient-tailored approach in choosing duration of DAPT, overall survival can be further improved [89]. Prasugrel as one of the new P2Y12 inhibitors proved its higher efficacy in comparison to clopidogrel in the TRITON-TIMI 38 trial, 13,608 patients with moderate-tohigh-risk acute coronary syndromes [90]. Prasugrel was associated with a highly significant 19% reduction in the primary composite endpoint of cardiovascular death, nonfatal myocardial infarction and nonfatal stroke from 12.1 to 9.9%, over a period of up to 15 months of follow-up. These benefits were observed at the cost of more serious bleeding. Prasugrel could even prove its superiority over clopidogrel in a cost-effectiveness study [91]. In the PLATO trial ticagrelor, another new P2Y12 inhibitor was tested with respect to the prevention of vascular events and death in patients presenting with an acute coronary syndrome and compared with clopidogrel [92]. At 12 months, the primary endpoint – a composite of death from vascular causes, myocardial infarction or stroke – had occurred in 9.8% of patients receiving ticagrelor as compared with 11.7% of those receiving clopidogrel (HR 0.84; 95% CI: 0.77–0.92; p < 0.001). Ticagrelor and prasugrel both seem to have the greatest benefit in the later phase by preventing recurrent ischemic events. Unfortunately, both trials were terminated after 12 and 15 months already. Thus, it remains elusive whether the two drugs prevent further events beyond the duration of the two trials. In the natural history study of coronary atherosclerosis by Stone et al., the 3-year cumulative rate of major adverse cardiovascular events was 20.4% [93]. The events were equally attributable to recurrence at the site of culprit lesions and at nonculprit lesions (12.9 vs 11.6% of patients). Similarly, in the 5-year follow-up of 428 patients of the BASKET-Trial, events remote from the stenting sites accounted for almost 40% of all cardiac events [94]. This documents the importance of progression of atherosclerosis in the whole coronary tree and the need to differentiate between new sites of www.expert-reviews.com
Review
atherosclerosis from target-vessel events. Thus, it can be questioned whether prevention of late ischemic events or late ST can be improved by prolongation of DAPT as it was shown for the initial 12 and 15 months for the two new P2Y12 inhibitors in the two trials (PLATO, TRITON-TIMI 38). Furthermore, the role of aspirin in the context of these drugs remains to be addressed in the future [89]. Patient factors: other patient factors
Several other patient risk factors are important in the genesis of ST. Prior brachytherapy itself bears the risk of ST through initiation of chronic inflammation by radiation injury [95]. This leads to a prothrombotic environment that bears the risk of late ST, even without the implantation of DES. Other comorbidities such as diabetes mellitus, renal failure, malignancy and an impaired left ventricular function are established as risk factors for ST [73]. Lesion characteristics: lesion/stent length
Despite the improved efficacy of newer DES devices, a long diseased coronary segment continues to be a major determinant of worse prognostic outcome. Many trials showed that greater lesion length and/or use of longer stents are associated with higher rates of restenosis with both BMS [96] and DES [97]. This was explained by the greater arterial injury and subsequently exaggerated intimal hyperplasia due to longer balloons and stents needed for treating longer lesions. In an observational study of 1911 consecutive patients with DES implantation (first generation SES and PES) followed for 19.4 months, Park et al. demonstrated a significant correlation between stent length and ST [98]. Total stent length was shown to be an independent predictor of total ST but not of late ST. In a prospective study by Iakovou et al. of 2229 patients receiving either PES or SES, stent length was recognized as a predictor of subacute thrombosis with the risk of thrombosis being 1.03-times greater for each 1-mm increase in length [51]. In a recent study by Claessen et al., third-generation PES was compared with second-generation everolimus-eluting stents (EES) [99]. The 2-year rates of major adverse cardiac event (MACE) with both EES and PES were highest in long lesions (>13.4 mm) and lowest in short lesions (3.0 mm), there was a small increase of risk in cardiac death/ within the target vessel was larger (8.3 ± 7.5 vs 4.0 ± 3.8 mm2; MI late after the intervention with DES but no difference in p = 0.03) in patients with very late ST than in those without. MACE rate and even in non-MI-related TVR. After 3 years of Late acquired ISA was reported in 2–5% of segments after follow-up, the clinical benefit of DES was maintained at no BMS [115] and in 7–21% after DES [116]. In a recent substudy overall increased risk of death or death/MI [102]. In those of the Harmonizing Outcomes with Revascularization and patients with small vessels, clinical TVR rates tended to be Stents in Acute Myocardial Infarction trial in patients with STsmaller after DES (9.9 vs 13.9% [BMS], p = 0.07), whereas segment elevation myocardial infarction, a modern PES reduced death/MI beyond 6 months was higher after DES (9.1 vs 3.8% restenosis by inhibiting neointimal hyperplasia, but had also a [BMS], p = 0.009), mainly due to increased late death/MI in higher rate of uncovered and malapposed stent struts as compatients with large stents (9.7 vs 3.1%, p = 0.006). In a multi- pared with the otherwise identical BMS (in 0.1 ± 0.2% of center registry by Yan et al, BMS implantation in large native BMS lesions vs 0.9 ± 2.1% of PES lesions; p = 0.0003) when coronary vessels ‡3.5 mm was associated with a low risk for ST examined 13 months after stent implantation [117]. The finding (1.0 vs 0.9%, p = 0.88), MACE (9.4 vs 9.4%, p = 0.90) and of ISA and its consequences are matter of an ongoing debate in repeat revascularization (4.8 vs 3.6%, p = 0.54) at 12 months asymptomatic patients. Cook et al. evaluated the impact of ISA that was comparable to DES [103]. In small arteries (
ISSN: 1477-9072 (Print) 1744-8344 (Online) Journal homepage: http://www.tandfonline.com/loi/ierk20
Long-term safety of drug-eluting stents Florian N Riede, Matthias Pfisterer & Raban Jeger To cite this article: Florian N Riede, Matthias Pfisterer & Raban Jeger (2013) Long-term safety of drug-eluting stents, Expert Review of Cardiovascular Therapy, 11:10, 1359-1378 To link to this article: http://dx.doi.org/10.1586/14779072.2013.837694
Published online: 10 Jan 2014.
Submit your article to this journal
Article views: 41
View related articles
Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=ierk20 Download by: [Central Michigan University]
Date: 06 November 2015, At: 00:20
THEMED ARTICLE y Stents + Devices
Review
Long-term safety of drug-eluting stents Expert Rev. Cardiovasc. Ther. 11(10), 1359–1378 (2013)
Downloaded by [Central Michigan University] at 00:20 06 November 2015
Florian N Riede, Matthias Pfisterer and Raban Jeger* Division of Cardiology, University Hospital, Basel, Switzerland *Author for correspondence: [email protected]
Stent implantation in coronary stenosis has revolutionized the treatment of coronary artery disease. The introduction of antirestenotic drug coatings further improved their efficacy in reducing target vessel revascularizations. With increasing use of drug-eluting stents (DES), stent thrombosis (ST) rose as potentially fatal major complication. Initially, the incidence of ST late after stent implantation seemed to be similar for DES and bare metal stents until several studies proved otherwise in first-generation DES. Since then, the design and components of DES have been changed and new polymers, drugs and different combinations of platelet inhibitors have been introduced to further improve the safety of DES. In this review, the authors focus on the relationship between DES, lesion anatomy, implantation technique and pharmacology to avoid the occurrence of ST. Furthermore, the relationship between dual antiplatelet therapy, bleeding rate and its significant impact on patient outcome is discussed. Finally, some promising future concepts are highlighted. KEYWORDS: drug-eluting stents • efficacy • future concepts • major bleeding • pathophysiology • risk factors • safety • stent thrombosis
The introduction of balloon coronary angioplasty, first performed by Andreas Gru¨ntzig on 16 September 1977 in Switzerland [1], was the birth of interventional cardiology. Percutaneous transluminal coronary angioplasty (PTCA) nowadays is one of the most frequently used major interventions in medicine. Soon after using balloon angioplasty on a big scale, it became evident that this novel technique harbors several risks and possible complications. Acute vessel closure was recognized as a pivotal complication in the early phase after intervention. Besides elastic recoil of the dilated vessel accounting for 40–50% of all cases, the occurrence of dissection, subintimal hemorrhage, platelet-mediated thrombus formation and vasoconstriction were reasons for acute vessel closure early after PTCA [2]. The process of slow narrowing of the vessel lumen was called restenosis manifesting later after PTCA. The pathophysiologic explanation for this phenomenon is an active response of the dilated vessel to the injury of the wall. In a self-perpetuating process, media muscle cells reallocate to the intima and start producing extracellular matrix, resulting in a narrowing of the vessel lumen [3]. Gru¨ntzig himself, one of the most experienced interventional cardiologists in the early 1980s of the last century, reported on a recurrence rate of 30% occurring in the first 6 months after the www.expert-reviews.com
10.1586/14779072.2013.837694
procedure [4]. The development of coronary stenting, which was introduced by Sigwart et al., was a treatment against acute dissections, elastic recoil of the vessel wall and occasionally arterial spasm after balloon angioplasty [5,6]. The stent acts as scaffold supporting the vessels’ integrity and by this counteracts acute vessel closure. However, it had also impact on restenosis rate. Two landmark studies comparing the effects of stent placement with balloon angioplasty, the BENESTENT trial [7] and the STRESS trial [8], both published in 1994, showed a lower rate of restenosis after 7, respectively, 6 months (22 vs 32% in the BENESTENT trial and 32 vs 42% in the STRESS trial). The acceptance of stenting as therapy for CAD was further increased, when evidence rose that stenting with combined platelet inhibition without anticoagulation was safe [9,10]. In an attempt to reduce in-stent neointimal hyperplasia, stents were coated with antiproliferative drugs, with the so-called drug-eluting stents (DESs) being the next breakthrough invention in interventional cardiology. Design of DES
Despite many inventions and modifications of the composition of stents, most modern DES consists of three parts: a stent body serving as
2013 Informa UK Ltd
ISSN 1477-9072
1359
Review
Riede, Pfisterer & Jeger
scaffold; a polymer covering the metal and serving as carrier for drugs; and a drug coating.
Downloaded by [Central Michigan University] at 00:20 06 November 2015
Stent body
Stent scaffolds differ in many engineering aspects and can be classified in a simplified way by materials and fabrication processes used; the mechanism of stent expansion (self-expansion, balloon expansion); and stent configuration and design [11]. The raw material mainly used determines the method of fabrication. Wires can be formed into stents by knitting, coiling and braiding. The self-expanding Wall Stent, for example, used a braided design and was made of multiple elgiloy wires (i.e., a cobalt– chromium alloy). On the contrary, the vast majority of modern balloon-expandable stents nowadays are produced by laser cutting from tubings. They are usually made of metal alloys that allow for thinner stent designs while preserving radial strength and radiopacity including materials such as stainless steel, cobalt–chromium and platinum–titan alloys. The raw material and its fabrication process determine the possible stent geometries: wires can be processed into coils, helical spiral and braided designs, tubes can be typically transformed into sequential ring geometries [11]. The latter accounts for most of the stents in use nowadays. It comprises a series of expandable seven-shaped structural elements, the so-called ‘struts’, joined by ‘bridges’ or ‘nodes’. In relation to the manner the elements are connected, two different cell configurations are possible: the closed cell and the open cell design. In a closed cell design, all internal infliction points of the struts are connected by bridges. The Palmaz-Schatz Stent was the first slotted-tube (a closed cell design) stent in clinical practice that confirmed the superiority of stenting versus balloon angioplasty in the treatment of CAD [12]. Like all stents with a closed cell design, the Palmaz-Schatz stent showed superior resistance to deformation and high radial strength but lacked flexibility [13]. The newer open cell or multicellular designs have less strut–strut intersections because not all of the internal inflection points of the struts are connected by bridges. They provoke less vascular injury and less intimal hyperplasia [14]. Strut thickness plays a pivotal role concerning the immediate and the longterm performance of stents. Thicker struts promise more radial support, better radiopacity and better wall support but show worse long-term efficacy concerning restenosis. This was shown in the ISAR-STEREO [15] and the ISAR-STEREO2 trials [16]. In the ISAR-STEREO trial, different stents from the same manufacturer with similar design, varying only in strut thickness (0.05 mm vs 0.14 mm), were randomly implanted in 651 patients. At 6 months, the late loss in large coronary arteries (>2.8 mm) was more pronounced in the thick-strut stent (Duet stent, Guidant) than in the thin-strut stent (Multilink stent, Guidant) (1.17 vs 0.94 mm, p = 0.001). Longitudinal stent deformation, defined as the shortening of a stent in the longitudinal axis after successful deployment, has recently been recognized as potential cause of stent malapposition and intraluminal protrusion of stent struts [17]. Overall, it occurs rarely, more frequently in ostial or bifurcation lesions and with multiple stents and it raises the risk of restenosis and ST [18]. 1360
Stent polymer coverage
To deliver a therapeutic drug at the site of the treated stenosis, it has to be brought to the stents’ surface. Drugs can be loaded either directly onto the struts (e.g., via impregnation onto microporous stent surfaces via solvent evaporation) or on a carrier vehicle matrix that is added to the device surface [19]. The carriers are usually polymeric molecules, which act like a sponge by increasing the surface area thereby enabling sufficient loading of the applied drug. The first polymers were durable. These polymers have been exclusively used in the firstgeneration DES (e.g., Taxus, Boston Scientific; Cypher, Cordis). The most commonly used permanent polymers are poly-n-butyl methacrylate, polyethylene-co-vinyl acetate and the tri-block copolymer poly(styrene-b-isobutylene-b-styren) (SIBS). By adding an antirestenotic drug together with one of these materials, a drug–polymer matrix can be formed and applied to the struts of the stent. Permanent polymers have been shown to trigger chronic hypersensitivity [20] and inflammation [21] reactions and thus induce restenosis and possibly ST. This bioresponsiveness of the persistent environment to the polymer ultimately determines the risk for adverse events. Biodegradable polymers have been developed in order to overcome hypersensitivity reactions leading to ST. Examples of biodegradable polymers are poly-L-lactic acid (PLLA), poly-L-lactide-cocaprolactone and poly DL-lactide-co-glycolide. They are degraded by hydrolysis and enzymatic activity and will leave a bare metal stent (BMS) after they have vanished through degradation. The noninferiority of a custom-made biodegradable polymer used in a rapamycin-eluting stent-platform has recently been shown in the ISAR-TEST 4 [22] and in the LEADER trials [23]. Polymer-free stent platforms have been introduced in recent years to overcome the undetermined effect of polymers on arterial wall healing, causing inflammation (even through degradation when using biodegradable polymers). In a randomized study, a custom-made polymer-free stent platform could recently prove its efficacy and safety over a period of 5 years in the ISAR-TEST three trial [24]. Drug coating
The first immunosuppressive agents used for a drug coating around the metal scaffold were sirolimus (previously called rapamycin) and paclitaxel. Sirolimus is a member of the limus family, that is, immunomodulators that inhibit the mammalian target of rapamycin. There are newer members of this family used in DES such as everolimus, zotarolimus and biolimus A9. By decreasing the positive and increasing the negative regulators of the cell cycle through interacting with the Cyclin Kinase pathway [25], the targeted cell will be arrested at the G0/ G1 phase in a cytostatic way. Cell migration and proliferation will both be inhibited. On the other hand, paclitaxel is a potent antimigratory and antiproliferative agent, acting through stabilization of cellular microtubules. Thus, it inhibits components of the cytoskeleton and mitotic spindle and impacts primarily the M-phase of the cell cycle. It works rather cytotoxic by leading to apoptosis and cell death. Both drugs mentioned Expert Rev. Cardiovasc. Ther. 11(10), (2013)
Long-term safety of DES
above, sirolimus and paclitaxel, work in DES as active coating. To deliver the drug at the site of the target lesion, it has to be bound to the stents surface, mostly by a polymer. Stent performance
Downloaded by [Central Michigan University] at 00:20 06 November 2015
The performance of a stent can be assessed by differentiating between its safety and its efficacy. Clinical trials designed to evaluate new percutaneous coronary intervention (PCI) devices; for example, new DES are designed to assess their efficacy and safety in order to allow a comparison with other devices. Thus, efficacy and safety can be regarded as the yin and yang of interventional cardiology, especially of DES. Efficacy
Review
to the reduced rates of restenosis in modern BMS. Stents with bioactive coatings, the so-called ‘bioactive stents,’ have been designed to enhance device integration with the surrounding biology. Titanium–nitride oxide covered bioactive stents demonstrated reduced rates of restenosis and platelet adhesion in a porcine model [40]. They showed a low incidence of major adverse events at 12 months in a real-world registry [41]. In comparison to modern zotarolimus-eluting stent (ZES), they remained inferior in a first randomized study [42] in terms of restenosis and late loss. Safety
Soon after introducing DES in clinical routine, there were first reports about abrupt vessel occlusion at the stent site potentially from a thrombotic genesis [43]. Late ST, a life-threatening complication of this technology, has emerged as a major concern in recent years. ST already known as complication after intracoronary brachytherapy [44,45], and its pathogenesis has been postulated to be related to delayed reendothelialization of the treated lesion (FIGURE 1) [46,47]. For a long time, it was thought that the benefit of DES would outweigh these risks; however, data from meta-analyses [48–50] and observational studies [51,52] questioned this paradigm. To prevent ST, prolonged dual antiplatelet therapy (DAPT) was recommended by the authorities and societies. Higher rates of major bleeding were encountered with DAPT,
Stents have been developed to overcome dissections following percutaneous transluminal coronary angioplasty (PTCA) [26]. Albeit BMSs have been demonstrated to reduce restenosis significantly more than balloon angioplasty, in-stent restenosis (ISR) occurred in about 10–60% of cases [7,8,27]. In the BENESTENT-1 trial, the initial 10% difference in target lesion revascularization (TLR) between the stent and PTCA groups remained unchanged at 5 years follow-up (17.2 vs 27.3%), emphasizing the prolonged benefit of BMS versus PTCA [28]. As a possible mean to prevent the development and to inhibit the recurrence of ISR, radiation therapy was used in coronary arteries (brachytherapy). In a metaanalysis by Oliver et al., assessing the performance of DES in comparison to intravascular brachytherapy, DES SES PES ZES EES BMS reduced the rate of revascularization (OR 0.51, 95% CI: 0.36–0.71) and binary restenosis (OR 0.57, 95% CI: 0.40–0.81) compared to brachytherapy, which itself demonstrated to be much more effective than BMS or PTCA alone [29]. Due to these results, the firstgeneration DES with polymer-based paclitaxel (Taxus, Boston Scientific) and sirolimus (Cypher, Cordis Corporation) coating quickly became the gold standard for interventional treatment of coronary artery disease [30–32]. DES showed a relative reduction in restenosis rate of 70–85% in comparison to BMS and an absolute rate of restenosis and target vessel revascularization (TVR) of under 10% [30–38]. In the more recent HORIZONS-AMI study, a newer PES was compared with an otherwise identical BMS from the same Figure 1. Scanning electron micrographs of 14-day endothelial coverage of manufacturer [39] in patients with Cypher (SES), Taxus Liberte´ (PES), Endeavor (ZES), Xience-V (EES) and Multi-Link STEMI. Compared with BMS, PES (BMS) in rabbit iliac arteries. reduced clinically driven target lesion BMS: Bare metal stent; EES: Everolimus-eluting stent; PES: Paclitaxel-eluting stent; revascularization at 12 months from SES: Sirolimus-eluting stent; ZES: Zotarolimus-eluting stent. 7.4 to 4.5% (p = 0.003), thus pointing Reproduced with permission from [47]. www.expert-reviews.com
1361
Review
Riede, Pfisterer & Jeger
thus bleeding arose as another safety aspect with the use of DAPT. Furthermore, it was widely debated, whether DES would lead to lower cardiac and all-cause mortality. Safety issues after stent implantation
Downloaded by [Central Michigan University] at 00:20 06 November 2015
All-cause & cardiac mortality of DES
Advances in interventional cardiology have led to a marked decline in post-procedural mortality, leading to a shift of the timing of mortality from the index PCI to later time points. The investigators of the EVENT registry, analyzing prospectively the contribution of cardiovascular death to 1-year mortality after PCI in 10,144 patients, found that overall 3% of patients died within the first year, 17% within the first month, 83% between 31 and 365 days after index PCI [53]. Cardiovascular death (CVD) was the predominant mode of death during the first 30 days by a factor of four. After the first month, the likelihood of non-CVD or death from unknown causes was very similar to the likelihood of CVD. Patients who survived the first year after index PCI received more often DES (88 vs 77%, p < 0.01) than deceased patients. Data from a singlecenter registry demonstrated significantly lower all-cause mortality with DES (8 vs 17%, p < 0.001; HR: 0.62, 95% CI: 0.53–0.73; p < 0.001) over a 4.5-year observation period, especially after adjustment for potential confounding factors [54]. The large Australian MIG registry reported an overall mortality rate at 30 days of 2.1% (79% cardiac), at 12 months of 3.9% (61%) and at long-term (3.2 ± 0.5 years) of 8.2% (50%) in a higher risk cohort [55]. DES deployment appeared protective against late mortality (HR: 0.85, 95% CI: 0.73–0.99; p = 0.04). After the first month, there was no difference in mortality rates between those who received a DES and those who received a BMS. The mortality benefit of DES over BMS was preserved in all age groups in a Dutch study, stratifying patients in quintiles based on age [56]. The recently published SORT OUT V trial showed similar rates of all-cause mortality (2.4 vs 2.2%, p = 0.67) and cardiac death (1.0 vs 1.1%, p = 0.71) at 12 months in biodegradable biolimus-eluting stents versus permanent polymer serolimus-eluting stent (SES). Thus, modern DES is safe in terms of mortality, although occurrence of very late ST remains a safety concern [57]. ST Definition & temporal patterns of ST
Clinically, ST are typically encountered as acute coronary syndromes consisting of acute chest pain in combination with ischemic ECG changes and raising heart enzymes (e.g., troponines) [58]. In order to categorize ST, it is important to differentiate several time segments relative to the index procedure. Widely used in clinical trials are the Academic Research Consortium (ARC) consensus criteria [59]. They have been issued to provide consistency in the reporting on the occurrence of safety endpoints of future stent trials and contain arbitrary chosen time intervals. Most ST events take place in the first 30 days after stent implantation and are defined as early ST. They are thought to 1362
happen mostly due to procedural or technical characteristics and compliance with dual antiplatelet therapy and are further stratified into acute ST (within the first 24 h) and subacute ST (24 h to 30 days). Late ST events are defined by occurring in the time interval between 30 days and 1 year after stent implantation and are less frequent. Very late STs are defined by the ARC criteria as occurring beyond 1 year. This interval has been questioned by various trials. In BASKET-PROVE, for example, late events were defined as those occurring 7–24 months after the intervention, on the basis of BASKETLATE data and on angiographic studies of the timing of restenosis [60]. From a pathophysiological point of view, late and very late ST are the most complex complications of stents. Their different risk factors and causes are discussed below. The ARC criteria define varying degrees of certainty in a tri-level classification. Definite ST requires angiographic or autoptical confirmation of ST. Probable ST is considered in cases with any unexplained death within a month after the stent implantation or with occurrence of acute ischemia in the territory of the implanted stent any time after the index procedure. Possible ST defines the weakest level of certainty and refers to any unexplained death 30 days or more after intracoronary stenting. Despite being neglected in many stent trials, there is evidence that possible ST represents true ST in over 50% of the cases [61]. Incidence of ST
Already during the BMS era, late ST had been observed. Two registries [62,63] reported rates of late ST of 0.76 and 0.3%, respectively, representing nearly one-third of overall ST events (average 1.6%). Early anecdotal reports of ST several months after DES implantation, describing late total occlusion soon after interruption of ticlopidine treatment [43], could be substantiated by subsequent trials. Iakovou et al. analyzed the incidence of ST in 2229 consecutive patients who underwent successful implantation of sirolimus- or paclitaxel-eluting at 9-month follow-up [64]. About 29 patients (1.3%) suffered from ST (9 [0.8%] with sirolimus and 20 [1.7%] with paclitaxel; p = 0.09). The post hoc analysis of the BASKET-LATETrial [51], one of the first trials comparing DES with BMS in a ‘real-world’ setting, showed an increased rate of thrombosisrelated events (2.6 vs 1.3%) in DES as compared with BMS in the 12 month following after the discontinuation of clopidogrel. The hypothesis of a higher risk of ST with DES was controversely debated and examined in subsequent trials and metaanalyses, which found a comparable risk of ST in DES and BMS [65–67]. The long-term risk of ST in DES was further analyzed in a 4-year cohort study assessing 8,146 patients treated with DES (PES and SES). The investigators reported an incidence of 1.0/100 patient-years and a cumulative incidence of 3.3% at 4 years follow-up [68]. The hazard of late ST (30 day to 1 year) amounted to 0.46% and the hazard of very late ST (1–4 years) to 0.57% per year. A large meta-analysis [69], comprising data from 30 studies (221,066 patients, 4276 ST) with DES used in 87% showed definite (probable) late ST in 0.5% Expert Rev. Cardiovasc. Ther. 11(10), (2013)
Downloaded by [Central Michigan University] at 00:20 06 November 2015
Long-term safety of DES
Clinical sequelae after ST
ST are serious complications after stent deployment manifesting as myocardial infarction and cardiac death. In 611 patients of the Japanese RESTART registry evaluating ST after SES implantation, the clinical sequelae differed among those with early ST, compared with those with late ST and very late ST [65]. Mortality rate at 1 year after ST was significantly higher in early and late ST (22.4 and 23.5%) compared with very late ST (10.5%). In clinical presentation, cardiac arrest took place two-times more often in early ST (11%) than in late ST (6%) and very late ST (5.8%). In a retrospective analysis of 431 cases with a first definite ST, the primary composite endpoint of cardiac death and recurrent definite ST occurred in 27.9% after a median follow-up of 27.1 months [72]. About 20.1% suffered a recurrent definite or probable ST, similarly 21.3% suffered a myocardial infarction at the longest available follow-up. Noteworthy, neither the category of ST (early versus late) nor type of the previously implanted stent (DES or BMS) did affect the clinical outcome. The timing-dependent differences in clinical presentation between early, late and very late ST suggest possible differences in the pathophysiological mechanisms responsible, which will be discussed further in this review. Pathophysiology: risk factors and causes for late and very late ST
Numerous risk factors and causes for ST have been identified factors (TABLE 1) [73]. Of note, data from large-scale registries demonstrate that the multivariate predictors of ST change during follow-up [68,70]. Procedural factors and lesion characteristics mostly account for ST in the first 30 days after the implantation of DES. Stent factors and patient factors mainly modulate the response of the vessel wall toward the artificial scaffold implanted. So these two risk factors are of importance in the genesis of late ST. www.expert-reviews.com
7 6 Cumulative incidence (%)
(0.4%) and very late ST in 0.6% (0.4%) after a median of 22 months. In this analysis, definite (probable) acute ST occurred in 0.4% (0.2%) and subacute ST in 1.1% (1.0%), suggesting that the majority of ST occurred within the first month. Similarly, in the Dutch ST registry of over 21,000 patients, most cases (>70%) of ST occurred within 30 days of the procedure [70]. Kimura et al. reported on the incidence of ST in the 5-year outcome data of the j-Cypher Registry [71]. Cumulative incidence of definite ST in 12,812 patients undergoing SES implantation was low (30 days: 0.3%; 1 year: 0.6%; and 5 years: 1.6%). Surprisingly, late and very late ST continued to occur without attenuation up to 5 years after sirolimus-eluting stent implantation (0.26% per year) representing continuous hazards for the patients. This was also demonstrated in the Bern-Rotterdam registry by showing a continuous hazard of ST until the end of the observation period (4 years) (FIGURE 2) [68]. Thus, the data of the registries suggested that the pathophysiology of very late ST may not be the same as early and late ST because the predictors for very late ST were quite different from those for early and late ST.
Review
5.7% 5
Definite and probable ST
4 3.3%
3
Definite ST
2 1 0 0
Months
1
2 3 Time since PCI (years) 1
Cumulative incidence 1.2 definite ST (%) Cumulative incidence 3.7 probable and definite ST (%) Patients at risk (n) 7537
4
12
24
36
48
1.6
2.1
2.7
3.3
4.1
4.6
5.2
5.7
7209
5157
2747
1051
Figure 2. Cumulative incidence of definite stent thrombosis in 8146 patients during a 4-year follow-up period. PCI: Percutaneous coronary intervention; ST: Stent thrombosis. Reproduced with permission from [1].
Stent factors: delayed & disturbed endothelial repair
Virmani et al. were among the first to report on incomplete stent strut coverage with focal platelet aggregates and persistent fibrin deposits within the necrotic core 16 months after SES implantation [74]. Although neointimal thickening was minimal at the sites of strut coverage, poor endothelial cell growth of the noncovered sites was documented in all such cases. In a number of autopsy cases, a lack of endothelial cell strut coverage was reported as the best correlate of late ST [75]. The measured parameter that best correlated with endothelialization was the ratio of uncovered to total stent struts per histological section. The odds ratio estimated for late ST in lesions with a rate of uncovered stent struts >30% was 9.0 (95% CI: 3.5–22.0). When neointimal thickness in thrombosed and patent DES lesions were compared, there was a marked shift toward less neointimal growth in DES with mural thrombus formation. In BMS, however, near-complete endothelialization has been postulated within 3–4 months. It was uncertain whether reendothelialization with DES is only delayed or persistently incomplete up to late time points. A recent autopsy study revealed that both SES and PES implantation sites heal over time nearly completely [76] in DES placed for >12 months with confirmed on-label use, whereas off-label indications for both stents resulted in incomplete healing, in DES even more than 1 year after implantation. Kubo et al. demonstrated significant differences in strut coverage between patients with unstable angina and stable angina in an optical coherence tomography (OCT) study. Markedly, less SES struts were fully covered 1363
Review
Riede, Pfisterer & Jeger
Downloaded by [Central Michigan University] at 00:20 06 November 2015
Table 1. Selected multifactorial causes of stent thrombosis. Factors
Causes
Stent factors
Incomplete endothelialization Hypersensitivity to drug coating or polymer
Patient factors
PCI for acute coronary syndrome/ST-segment elevation myocardial infarction Antiplatelet therapy and individual response Prior brachytherapy Comorbidities: diabetes mellitus, renal failure, malignancy Impaired left ventricular function
Lesion characteristics
Lesion/stent length Vessel/stent diameter Lesion complexity (bifurcation lesions, chronic total occlusions) Saphenous vein graft lesion
Procedural factors
Inadequate stent expansion/sizing Incomplete stent apposition Stent deployment in necrotic core Residual edge dissection
(28 vs 63%, p = 0.019) and more struts only partially covered (72 vs 37%, p = 0.019) by neointima in patients with unstable angina compared with those with stable angina. These findings were consistent with those of an autopsy study from Nakazawa et al. [77]. Stent factors: hypersensitivity reactions, chronic inflammation
Since October 2003, when the US FDA issued a warning on its website notifying physicians of possible generalized hypersensitivity reactions in 50 patients receiving SES, hypersensitivity reactions in association with DES came into focus of research. In 2004, the case of a 58-year-old man who died of late ST 18 months after having received two SES was reported. His coronary angiogram already showed vessel enlargement 8 months after implantation. An autopsy showed aneurysmal dilation of the stented arterial segment with a severe localized hypersensitivity reaction [78]. This case report enforced rumors about the allergic potential of DES. It was speculated that the polymers carrying the drugs would trigger an inflammatory response to DES implantation rather that the drug itself because an extensive eosinophilic infiltration specifically prominent around stent struts was found, whereas the proximal part of the stented artery also showed focal giant cell reaction surrounding a few polymer remnants that had become separated from the stent struts. Since autopsy studies are limited to postmortem specimens and therefore lack generalizability and coronary angiography and intravascular ultrasound (IVUS) do not delineate morphological changes of the vessel wall in response to DES implantation in sufficient detail, Cook et al. analyzed thrombus aspirates and then performed an IVUS examination in 28 patients presenting with very late ST and in 26 controls [79]. By combining intravascular imaging with histopathological analysis, they could demonstrate that incomplete stent apposition area was associated with an increase of total eosinophil count and fraction of white blood 1364
cells (WBCs). Furthermore, in specimens from very late DES ST, eosinophilic infiltrates amounted to 10% of WBCs. Conversely, in patients with spontaneous acute MI, early and late ST in BMS and early DES ST, only 1–3% of WBCs were eosinophils, resembling the proportions in the peripheral circulation. Thus, one can hypothesize that isolated components of DES give rise to a delayed-type hypersensitivity reaction (type IV) in certain patients that can finally lead to late, respectively, very late ST by fibrin and platelet deposition through inflammatory changes or by promotion of the coagulation pathway. Stent factors: stent thrombogenicity
Chronic inflammatory processes as described above can create a prothrombotic environment that can ultimately lead to late/very late ST. So far, stent thrombogenicity has been mainly attributed to the polymers used on DES. However, the antirestenotic drugs themselves may exert a prothrombotic effect in the targeted vessel as well. This has been demonstrated for the two first widely used drugs in DES, that is, sirolimus and paclitaxel [80]. They enhance endothelial tissue factor expression, which is a cell surface receptor for coagulation factor VII that serves the main activator of the coagulation cascade activating factors IX and X. Paclitaxel and sirolimus are both highly lipophilic drugs that easily penetrate into endothelial and underlying cells of the vessel wall. Through intracellular retention, they can act in a prothrombotic fashion by inducing tissue factor expression after deployment of SES or PES [81]. Patient factors: dual antiplatelet therapy
Premature discontinuation of antiplatelet therapy is one of the most important predictors of ST. After the ‘ESC DES fire storm’ in 2006, the general awareness of late ST being related to significant morbidity contributed to the shift from 6 to 12 months of DAPT after DES implantation as recommended by the FDA for safety reasons [82]. The current ACC/AHA guidelines from 2011 recommend 12 months of DAPT for DES [83], whereas the ESC guidelines from 2010 state that convincing data exist only for DAPT up to 6 months [84] in elective interventions. The Bern-Rotterdam registry, cited by the ESC committee, demonstrated in over 8000 patients that there is no difference in those treated for 3–6 months versus those treated for 1 year [85]. On the other hand, it suggested that prolongation of DAPT from 12 to 24 months would improve cumulative survival. The recently published PRODIGY trial randomized 2013 patients 30 days after stenting to 6 versus 24 months of clopidogrel-based DAPT [86]. The trial showed no benefit of prolonged DAPT in the incidence of ischemic events but noted an increased risk of major bleeding (HR 2.17; Expert Rev. Cardiovasc. Ther. 11(10), (2013)
Downloaded by [Central Michigan University] at 00:20 06 November 2015
Long-term safety of DES
95% CI: 1.44–3.22; p = 0.00018). In the EXCELLENT trial, 1443 patients receiving DES for stable CAD or acute coronary syndromes were randomized to 6 versus 12 months of DAPT with clopidogrel in combination with continued aspirin treatment [87]. Rates of target vessel failure at 12 months were 4.8% in the 6-month DAPT group and 4.3% in the 12-month DAPT group. The authors concluded that 6-month DAPT did not increase the risk of target vessel failure at 12 months after implantation of DES compared with 12-month DAPT. In an observational study examining 4666 consecutive patients receiving intracoronary stents (BMS and DES), the continued use of clopidogrel was a predictor of lower rates of death after 24 months in patients who were event free at 6 months (2.0% with vs 5.3% without, p = 0.03) and at 12 months (0% vs 3.5%, p = 0.004) [88]. These ‘real life’ data point to a possible benefit of extended duration of DAPT in DES in event-free (no death, myocardial infarction or revascularization) patients. Since the occurrence of major bleeding influences overall survival (see below), the analysis of event-free patients, alive at landmark analysis, could have an impact on survival by selection bias. Maybe by using a patient-tailored approach in choosing duration of DAPT, overall survival can be further improved [89]. Prasugrel as one of the new P2Y12 inhibitors proved its higher efficacy in comparison to clopidogrel in the TRITON-TIMI 38 trial, 13,608 patients with moderate-tohigh-risk acute coronary syndromes [90]. Prasugrel was associated with a highly significant 19% reduction in the primary composite endpoint of cardiovascular death, nonfatal myocardial infarction and nonfatal stroke from 12.1 to 9.9%, over a period of up to 15 months of follow-up. These benefits were observed at the cost of more serious bleeding. Prasugrel could even prove its superiority over clopidogrel in a cost-effectiveness study [91]. In the PLATO trial ticagrelor, another new P2Y12 inhibitor was tested with respect to the prevention of vascular events and death in patients presenting with an acute coronary syndrome and compared with clopidogrel [92]. At 12 months, the primary endpoint – a composite of death from vascular causes, myocardial infarction or stroke – had occurred in 9.8% of patients receiving ticagrelor as compared with 11.7% of those receiving clopidogrel (HR 0.84; 95% CI: 0.77–0.92; p < 0.001). Ticagrelor and prasugrel both seem to have the greatest benefit in the later phase by preventing recurrent ischemic events. Unfortunately, both trials were terminated after 12 and 15 months already. Thus, it remains elusive whether the two drugs prevent further events beyond the duration of the two trials. In the natural history study of coronary atherosclerosis by Stone et al., the 3-year cumulative rate of major adverse cardiovascular events was 20.4% [93]. The events were equally attributable to recurrence at the site of culprit lesions and at nonculprit lesions (12.9 vs 11.6% of patients). Similarly, in the 5-year follow-up of 428 patients of the BASKET-Trial, events remote from the stenting sites accounted for almost 40% of all cardiac events [94]. This documents the importance of progression of atherosclerosis in the whole coronary tree and the need to differentiate between new sites of www.expert-reviews.com
Review
atherosclerosis from target-vessel events. Thus, it can be questioned whether prevention of late ischemic events or late ST can be improved by prolongation of DAPT as it was shown for the initial 12 and 15 months for the two new P2Y12 inhibitors in the two trials (PLATO, TRITON-TIMI 38). Furthermore, the role of aspirin in the context of these drugs remains to be addressed in the future [89]. Patient factors: other patient factors
Several other patient risk factors are important in the genesis of ST. Prior brachytherapy itself bears the risk of ST through initiation of chronic inflammation by radiation injury [95]. This leads to a prothrombotic environment that bears the risk of late ST, even without the implantation of DES. Other comorbidities such as diabetes mellitus, renal failure, malignancy and an impaired left ventricular function are established as risk factors for ST [73]. Lesion characteristics: lesion/stent length
Despite the improved efficacy of newer DES devices, a long diseased coronary segment continues to be a major determinant of worse prognostic outcome. Many trials showed that greater lesion length and/or use of longer stents are associated with higher rates of restenosis with both BMS [96] and DES [97]. This was explained by the greater arterial injury and subsequently exaggerated intimal hyperplasia due to longer balloons and stents needed for treating longer lesions. In an observational study of 1911 consecutive patients with DES implantation (first generation SES and PES) followed for 19.4 months, Park et al. demonstrated a significant correlation between stent length and ST [98]. Total stent length was shown to be an independent predictor of total ST but not of late ST. In a prospective study by Iakovou et al. of 2229 patients receiving either PES or SES, stent length was recognized as a predictor of subacute thrombosis with the risk of thrombosis being 1.03-times greater for each 1-mm increase in length [51]. In a recent study by Claessen et al., third-generation PES was compared with second-generation everolimus-eluting stents (EES) [99]. The 2-year rates of major adverse cardiac event (MACE) with both EES and PES were highest in long lesions (>13.4 mm) and lowest in short lesions (3.0 mm), there was a small increase of risk in cardiac death/ within the target vessel was larger (8.3 ± 7.5 vs 4.0 ± 3.8 mm2; MI late after the intervention with DES but no difference in p = 0.03) in patients with very late ST than in those without. MACE rate and even in non-MI-related TVR. After 3 years of Late acquired ISA was reported in 2–5% of segments after follow-up, the clinical benefit of DES was maintained at no BMS [115] and in 7–21% after DES [116]. In a recent substudy overall increased risk of death or death/MI [102]. In those of the Harmonizing Outcomes with Revascularization and patients with small vessels, clinical TVR rates tended to be Stents in Acute Myocardial Infarction trial in patients with STsmaller after DES (9.9 vs 13.9% [BMS], p = 0.07), whereas segment elevation myocardial infarction, a modern PES reduced death/MI beyond 6 months was higher after DES (9.1 vs 3.8% restenosis by inhibiting neointimal hyperplasia, but had also a [BMS], p = 0.009), mainly due to increased late death/MI in higher rate of uncovered and malapposed stent struts as compatients with large stents (9.7 vs 3.1%, p = 0.006). In a multi- pared with the otherwise identical BMS (in 0.1 ± 0.2% of center registry by Yan et al, BMS implantation in large native BMS lesions vs 0.9 ± 2.1% of PES lesions; p = 0.0003) when coronary vessels ‡3.5 mm was associated with a low risk for ST examined 13 months after stent implantation [117]. The finding (1.0 vs 0.9%, p = 0.88), MACE (9.4 vs 9.4%, p = 0.90) and of ISA and its consequences are matter of an ongoing debate in repeat revascularization (4.8 vs 3.6%, p = 0.54) at 12 months asymptomatic patients. Cook et al. evaluated the impact of ISA that was comparable to DES [103]. In small arteries (
