Treatment of In-Stent Restenosis

Introduction

The development of percutaneous coronary intervention (PCI) with stent implantation has revolutionized the practice of cardiology over the course of the past decades. However, despite considerable technological advancements, in-stent restenosis (ISR) remains the most common cause of treatment failure after PCI. Moreover the high efficacy with contemporary devices—primarily drug-eluting stents (DES)—has facilitated the expansion of PCI to broader and increasingly more complex lesion and patient subsets. Accordingly, despite low rates of ISR in relative terms the absolute numbers of patients presenting with stent failure remain considerable. Importantly the management of this condition remains challenging, with high rates of subsequent events at medium- to long-term follow-up.

The term restenosis is used in a variety of settings across the field of interventional cardiology. Angiographic restenosis is commonly adjudicated as a binary event defined as a re-narrowing of more than 50% of the vessel diameter as determined by coronary angiography. As this definition is based on two-dimensional parameters accurate measurements are critically dependent on the acquisition of worst-view projections. Typically visual estimation of restenosis is employed in routine clinical practice in the catheterization laboratory. This requires the operator to develop a sense of what comprises a 50% diameter stenosis. In adjudication of ISR the basic frame of reference is the body of the stent—this is known as an in-stent analysis. However, restenosis also shows a predilection for occurrence at stent margins. Accordingly a frame of reference including both the body of the stent and 5-mm margin proximal and distal to the stent edges is also usually assessed—this is known as an in-segment analysis. It is important to recognize that the use of 50% diameter stenosis as a cut-off for determination of restenosis as a binary event is rather arbitrary. For this reason continuous parameters are also commonly employed as surrogate markers of restenosis. These parameters also offer the advantage of superior statistical power for comparison between treatments, which makes them particularly attractive for clinical trials as they reduce the sample size required. The most commonly used continuous parameters are minimal lumen diameter (MLD) or percentage diameter stenosis at follow-up angiography and late lumen loss (which is the difference between the MLD immediately postprocedure and that at follow-up angiography). Of these, percentage diameter stenosis and late loss are the most well-studied markers in clinical trials and mean values of these parameters correlate reliably with incidence of angiographic and clinical restenosis.

Intravascular imaging modalities such as intravascular ultrasound (IVUS) or optical coherence tomography (OCT) acquire data in three dimensions. Using these modalities restenosis is defined as a re-narrowing of more than 75% of the reference vessel area in cross-section. Visual estimation of stenosis is not usually employed and rapid online quantitative measurements are routinely available in the catheterization laboratory. Similarly, in autopsy studies restenosis is usually defined as a pathological vessel re-narrowing of more than 75% of the vessel area in cross-section. The term clinical restenosis is sometimes used to refer to restenosis of the treated lesion accompanied by requirement for re-treatment, for example, due to symptoms or signs of ischemia. Rates of clinical restenosis are usually considerably lower than rates of restenosis detected by imaging as not all restenotic lesions cause ischemia or elicit symptoms.

The principles underpinning the management of ISR are not dissimilar to those underlying the treatment of de novo coronary atherosclerotic lesions. The basic tenet of interventional treatment is that efficacy is optimized by maximizing acute gain and/or by minimizing late loss ( Figure 13-1 ). However, the major difference with restenotic in comparison with de novo lesions is the presence of an existing stent scaffold in the diseased coronary segment. This may offer certain mechanical advantages, if its structural integrity is intact, but also provides a challenge due to the potential disadvantages of implanting multiple stent layers.

Mechanisms of In-Stent Restenosis

Restenosis after PCI is well characterized as a distinct pathophysiological process rather than merely an accelerated form of postintervention atherosclerosis. Broadly speaking the contributing factors to restenosis after vascular intervention may be divided into five categories as follows:

• 1.

  acute or subacute prolapse of the disrupted plaque

• 2.

  elastic recoil of the vessel wall

• 3.

  constrictive vascular remodeling

• 4.

  neointimal hyperplasia (due to extracellular matrix deposition and smooth muscle cell hyperplasia)

• 5.

  de novo in-stent atherosclerosis (so-called neoatherosclerosis)

While implantation of a metal stent after balloon dilatation largely negates the impact of the first three processes on restenosis, the additional vessel injury imposed by stent implantation increases the extent of neointimal tissue formation during vessel healing. Moreover, although DES have reduced the incidence of restenosis dramatically, the delayed vessel healing seen with these stents may contribute to an increased incidence or accelerated course of de novo atherosclerotic disease within the stent in the months and years after intervention. Indeed, from a clinicopathological standpoint it appears that there are considerable differences between the restenosis that occurs after bare-metal versus after drug-eluting stenting ( Table 13-1 , Figure 13-2 ). In addition the availability of high-resolution intravascular imaging modalities such as optical coherence tomography (OCT) permits more detailed characterization of intravascular tissue and recognition of neoatherosclerotic changes in vivo ( Figure 13-3 ; Videos 13-1 to 13-4 ).

In terms of angiographic morphology of restenosis after stenting Mehran et al. developed the most widely accepted classification system for restenosis within bare-metal stents. This scheme is based on stenosis length (≤10 mm is classified as focal, >10 mm as diffuse), geographic localization of the neointima in relation to the stent and whether or not the restenosis is occlusive. Patterns are classified into 4 major groups: type I focal; type II diffuse within stent; type III diffuse within and beyond stent; and type IV occlusive (see case examples Videos 13-5 to 13-19 ). Importantly the pattern of restenosis at presentation is a predictor of subsequent outcome after re-intervention. In the original study target lesion revascularization rates were 19%, 35%, 50%, and 83% in groups I-IV, respectively (p < 0.001). While the majority of restenotic lesions within bare-metal stents are diffuse, in DES the majority are focal ( Table 13-1 , Figure 13-4 ). This may be because DES are generally very effective at suppressing neointimal overgrowth, which means that focal technical issues (e.g., stent fracture, local underexpansion) may play a relatively more important role in comparison with bare-metal stent restenosis.

The remainder of this chapter summarizes the spectrum of management options for patients presenting with restenosis following bare-metal or drug-eluting stent therapy. The principal randomized trials investigating outcomes of patients treated for ISR are summarized in E-Table 13-1 .

Balloon Angioplasty

The earliest approach to the treatment of ISR was with balloon angioplasty. Balloon dilatation directly targets the two principal causes of ISR—namely, neointimal hyperplasia and stent underexpansion. Mechanistically balloon angioplasty compresses and extrudes (axially and longitudinally) neointimal tissue and corrects underlying underexpansion of the stent backbone. One study with intravascular ultrasound (IVUS) suggested that the lumen enlargement with angioplasty was due to additional stent expansion and decrease in neointimal tissue volume in approximately equal measure.

The technical procedure is relatively straightforward and the rate of complications is low. In fact early reports suggested that balloon dilatation for ISR entailed less risk than standard angioplasty of de novo coronary lesions. In brief, an angioplasty balloon is advanced to the site of the ISR and expanded to nominal pressure. In general a balloon : vessel ratio of 1.1 : 1.0 is chosen. The length of balloon is usually targeted to treat only the restenotic segment of the existing stent rather than the entire stented segment. Failure to achieve full expansion of the angioplasty balloon with standard pressures of 8-14 atmospheres is recognized by a “dog-bone”–like appearance of the balloon. This should promote a switch to noncompliant balloons which facilitate high-pressure inflation up to approximately 25 atmospheres. In actual fact some operators recommend systematic use of noncompliant balloons for angioplasty of ISR though in this respect economic considerations may play a role in local practice. In occasional situations dilatation-resistant lesions may be successfully dilated with super-high-pressure balloons using inflation pressures of up to 40 atmospheres, though risk of perforation should be considered.

An important additional technical consideration in angioplasty for ISR is balloon slippage due to “watermelon-seeding.” This phenomenon is more often encountered in angioplasty for ISR as compared with de novo disease and occurs when, for example, upon balloon inflation the balloon slips proximally or distally to the intended site of dilatation. This is not just time-consuming but may lead to inadvertent trauma to nontarget regions of the treated vessel, so-called geographic miss, which may increase the risk of subsequent recurrent restenosis. Lesions with higher diameter stenosis and diffuse stenoses are more susceptible. Strategies to anchor the balloon and avoid “watermelon-seeding” include slow step-wise inflation of the balloon, sequential angioplasty starting with smaller balloon sizes and working up, the use of a “buddy wire” to stabilize the balloon, and employment of anti-slip cutting or scoring balloons as an alternative to standard balloons.

Though a limited role for balloon angioplasty alone remains in contemporary practice there are a number of drawbacks that mean that it is no longer the treatment of choice for ISR. First, residual stenosis after the procedure is relatively high (on the order of 20% and more). Second, some studies suggest that further early lumen loss occurs in the initial hours after angioplasty. Indeed, one study that systematically examined patients immediately after intervention and 30-60 minutes later showed a significant reduction in minimal lumen diameter during this time window. This is mainly due to re-intrusion of tissue back into the vessel lumen in the minutes and hours after angioplasty. Moreover higher early lumen loss was associated with subsequent clinical events. Third, randomized trial data have demonstrated superiority of alternative strategies.

Plaque Debulking With Atherectomy

Since the predominant cause of ISR is intimal hyperplasia, debulking techniques were frequently investigated for treatment of this condition. Atherectomy devices relieve coronary stenosis by removing rather than simply compressing coronary plaque. Devices may be broadly characterized as “remove and retrieve” type (e.g., directional atherectomy) or “disrupt and displace” type (e.g., rotational or laser atherectomy). Although initially targeted at primary treatment of de novo disease, their role evolved into that of an adjunctive therapy prior to stent implantation, as well as a useful tool for neointimal tissue removal in ISR. However the passage of time and the advent of newer and more effective devices have seen the use of these modalities decline significantly or in many instances fall completely out of use.

Plaque debulking with rotational atherectomy is done using a metal burr studded with diamonds, which is advanced to the site of the restenosis and rotated at high speed (150,000-200,000 rpm). By pulverizing stenotic plaque to microparticles of 20- to 50-µm diameter (small enough to pass through the coronary microcirculation) additional lumen enlargement is achieved during PCI. Although undoubtedly a useful tool in the armamentarium of the interventionalist performing PCI for ISR, studies investigating a strategy of systematic lesion preparation with atherectomy have yielded mixed results. The initial 200-patient single-center ROSTER randomized trial in patients with ISR was inconclusive. Although rotational atherectomy compared with angioplasty alone showed no evidence of increased acute gain, rates of repeat revascularization were improved at follow-up. On the other hand the 298-patient multicenter Angioplasty versus Rotational Atherectomy for Treatment of Diffuse In-Stent Restenosis Trial (ARTIST) showed higher rates of repeat revascularization with a strategy of rotational atherectomy and adjunctive low-pressure balloon angioplasty compared with standard balloon angioplasty alone. In light of the availability of other high-efficacy devices for the treatment of ISR rotational atherectomy has largely fallen out of use. However, selected cases remain where rotational atherectomy may be required for undilatable ISR due to heavily calcified in-stent plaques or stent underexpansion.

Laser atherectomy has also been investigated for the plaque debulking in ISR. Most commonly used catheters were based on XeCl excimer laser ablation using ultraviolet spectrum wavelengths. A multicenter registry study with excimer laser angioplasty and adjunctive balloon angioplasty showed this strategy to be safe and effective. A mechanistic IVUS-based registry examined the use of laser atherectomy followed by angioplasty for ISR. It showed that acute gain was achieved in three almost equal parts: tissue ablation, tissue extrusion by angioplasty, and additional expansion of the underlying stent. Moreover matched to lesions treated with angioplasty alone, laser angioplasty resulted in greater acute lumen gain, and a tendency for less frequent treatment failure. A registry report of laser versus rotational atherectomy with IVUS analysis showed that although significantly greater reduction in intimal hyperplasia volume was seen after rotational atherectomy, both strategies had similar long-term clinical outcomes. Unfortunately, randomized trial data comparing laser atherectomy with standard angioplasty were never published, and this technique is no longer performed.

Finally, directional atherectomy is of interest for being the most potent plaque-debulking technique in use. The basic principle of use is that plaque is removed from the vessel by a cutting device mounted on a positioning balloon catheter. Upon balloon inflation plaque is shaved into the windowed housing of the catheter and removed from the body. The principal scientific interest is the facilitation of histopathological analysis of excised plaque. Small-scale registries showed encouraging results and a comparison against rotational atherectomy suggested more potent and a lower incidence of subsequent target lesion revascularization with directional atherectomy. However, as with other debulking techniques, compelling randomized trial data against standard therapy were not realized and the device is no longer in widespread use, at least in the coronary arena.

Cutting and Scoring Balloon Angioplasty

Cutting balloons are comprised of standard balloon catheters mounted with lateral metallic blades known as athertomes ( Figure 13-5A ), which on inflation of the balloon incise into the treated stenotic plaque. There are two main advantages to their use:

• the incision of the blades into the stenotic plaque may favor subsequent extrusion

• the interaction of the blades with the plaque anchors the balloon in the plaque and prevents “watermelon-seeding”; this in turn might reduce problems related to geographic miss.

An initial large registry of patients treated with ISR from Lenox Hill, New York, compared outcomes of matched patients according to treatment with cutting balloon, rotational atherectomy, stenting, or plain angioplasty. Results suggested a clear edge for cutting balloon angioplasty in terms of angiographic and clinical outcomes at follow-up. However subsequent results from the restenosis cutting balloon evaluation trial (RESCUT) randomized trial were more disappointing. In 428 patients with restenosis randomized to treatment with cutting balloon versus plain balloon angioplasty, although cutting balloons showed less procedural balloon slippage, no advantage in terms of the primary endpoint of binary angiographic restenosis at 7-month angiographic follow-up was seen (cutting balloon angioplasty 29.8% vs. plain balloon angioplasty 31.4%; p = 0.82). Moreover case-control studies in which cutting balloon angioplasty was utilized prior to vascular brachytherapy did not suggest an advantage over standard balloon angioplasty lesion preparation.

Scoring balloons have a broadly similar mechanistic basis to cutting balloons. The main difference is that low-profile nitinol wires (on the order of 125 µm) in spiral formation are mounted on the surface of the balloon catheter instead of blades ( Figure 13-5B ). As a result the deliverability and flexibility of the catheters are increased, at the expense of a lesser degree of plaque incision. However, anchoring at the lesion and protection against “watermelon-seeding” are maintained. Although this approach is potentially attractive, published data in ISR are limited at present to case reports, although a randomized trial comparing scoring balloon angioplasty with plain balloon angioplasty is ongoing (ISAR-DESIRE 4; registered at clinicaltrials.gov, identifier: NCT01632371), and a drug-coated scoring bal­loon device is also being currently tested.

Vascular Brachytherapy

Intracoronary radiation therapy—commonly known as vascular brachytherapy—has also been successfully used to treat ISR. The therapy is delivered at the time of mechanical treatment of the stenosed stent and is termed brachytherapy due to the short distance from the radiation source to the target tissue. Radioactive material (typically in the form of seeds, less success with fluids) is delivered to the target lesion inside a specialized catheter, which is left to dwell in the coronary artery for a period of between 2-3 and 30-45 minutes. Both beta and gamma radiation sources have been successfully used. Gamma radiation sources have high energy (roughly 0.2-10 MeV), high tissue penetration, longer dwell times, and requirement for more stringent radiation protection protocols. Beta radiation has lower energy, lesser penetration, shorter dwell times, and a reduced requirement for radiation shielding. Preclinical investigation with both sources showed effective inhibition of neointimal hyperplasia in porcine models of coronary intervention.

Initial randomized trials with double-blind, most with sham-catheter, treatments showed improvement in both angiographic and clinical outcomes at 6 to 12 months' follow-up with both gamma and beta radiation sources in comparison with standard balloon angioplasty. There were two chief limitations of vascular brachytherapy, namely, requirement for specialized laboratory equipment with unwieldy treatment protocols and systematic impaired arterial healing following intervention. This latter issue foreshadowed many of the problems observed with early-generation DES technology and was associated with a spectrum of adverse events including geographic miss, edge restenosis, late stent thrombosis, and requirement for prolonged dual antiplatelet therapy and late “catch up” restenosis.

In many respects the advent of DES therapy heralded the end of the era of vascular brachytherapy. The combination of high acute gain with low late lumen loss seen with DES was particularly well suited to the challenging condition of ISR. Moreover, although DES therapy was not without its own safety concerns, the overall safety profile and the ease of therapy delivery made DES a superior treatment option in routine clinical practice.

Two important multicenter randomized trials directly compared outcomes of patients with ISR who were allocated to treatment with either vascular brachytherapy or repeat stenting with a DES. In the SISR trial patients treated with sirolimus-eluting Cypher stents had superior outcome to those treated with beta or gamma brachytherapy (see Figure 13-6A ). Similarly in the Paclitaxel-Eluting Stents versus Brachytherapy for In-stent Restenosis (TAXUS V ISR) trial patients randomized to paclitaxel-eluting Taxus stents had superior angiographic and clinical outcome to beta brachytherapy-treated patients (see Figure 13-6B ). Subsequent reports from both studies confirmed persistent advantage with DES out to 5 years. Although both studies enrolled only patients with bare-metal stent restenosis, and some encouraging observational data have been reported with brachytherapy for DES restenosis, there is no compelling reason to expect differential findings in patients with DES restenosis. Overall the lack of enthusiasm for brachytherapy coupled with concerns regarding delayed healing and a reduced commercial interest has led to extremely limited use of this treatment modality.