Scale of the Problem
Treatment of patients with in-stent restenosis (ISR) remains a significant clinical problem.1 Bare-metal stents (BMS) continue to be used in a large number of patients undergoing percutaneous coronary interventions (PCIs).1–4 BMS are indicated in patients with high bleeding risks or those who are unable to maintain long-term dual antiplatelet therapy. Economic reasons and the fear of late stent thrombosis, particularly in complex clinical and anatomical scenarios, also help to explain incomplete drug-eluting stent (DES) penetration.1 In addition, although DES have drastically reduced the rate of ISR, these new devices are not immune to ISR, especially when used in off-label indications.1 As a matter of fact, the advent of DES generated a sense of confidence in the interventional cardiology community and interventions shifted to tackle increasingly complex lesions. This translated into higher rates of ISR compared with the single-digit figures obtained in early DES series. Overall, the annual number of patients presenting with DES ISR may approach 200,000 in the US alone.1 In Spain,4 ISR represented 6.3% of the total PCI activity in 2003 (20% DES penetration) and 5.5% of interventions in 2007 (58% DES penetration). In fact, an absolute increment in the number of patients requiring ISR treatment was detected.4
In-stent Restenosis Presentation and Anatomical Patterns
ISR has been classically considered a benign clinical entity.1–3 In fact, only patients with recurrent symptoms or inducible ischaemia require reinterventions. However, recent reports indicate that some patients with ISR present with unstable symptoms or even with a myocardial infarction.5 However, in our experience, ISR presenting as a relatively large myocardial infarction is rare2,3 and, in this context, the possibility of stent thrombosis should always be carefully excluded. Repeated PCI constitutes the mainstay of therapy in these patients. However, in cases with recalcitrant ISR or whenever multiple vessels may benefit from revascularisation, coronary surgery may be considered.1
ISR remains an angiographic definition: recurrent per cent diameter stenosis >50% at the stent site or at its edges (adjacent 5mm segments).1–3 Angiography is able not only to determine the severity of ISR but also to characterise different morphological patterns. Patients with diffuse ISR (>10mm) have a poor clinical and angiographic outcome after repeated interventions.1 The classification of Mehran et al.6 has been widely implemented (pattern I [focal], pattern II [diffuse], pattern III [proliferative] and pattern IV [total occlusion]), and these patterns are associated with progressively increasing requirements for target lesion revascularisation (19, 35, 50 and 83%, respectively). In our experience,7 the American College of Cardiology/American Heart Association (ACC/AHA) angiographic lesion classification is of value also in these patients. B2–C ISR lesions are more frequently associated not only with suboptimal immediate results but also with a higher restenosis rate and poorer long-term clinical outcome.
Angiography is unable to provide a precise visualisation of the stent struts; therefore, additional intracoronary diagnostic techniques that allow visualisation of the stent and to measure neointimal proliferation with accuracy are of particular interest in these patients. Intravascular ultrasound (IVUS) provides highly valuable anatomical information, providing unique insights on the exact mechanism of ISR. ISR is almost exclusively attributable to neointimal hyperplasia.3,8 However, many stents suffering from ISR are severely underexpanded. IVUS may be used to determine ISR severity and location and to guide and optimise the results of subsequent interventions.3,8 Compared with the anatomical information provided by IVUS, measuring the fractional flow reserve constitutes an elegant physiological assessment of ISR severity.9 Currently, IVUS remains the gold standard for mechanistic studies assessing minor differences in neointimal proliferation among different DES. Recently, however, optical coherence tomography has emerged as a valuable diagnostic tool in these patients.10 Due to its higher resolution, additional insights on neointimal material composition and the extent of stent coverage can be obtained. The trade-off of this technique is its poor tissue penetration, which may cause problems in large vessels or in assessing residual plaque behind the stent.
Recent studies suggest that heterogeneous material can be found (including not only neointimal hyperplasia but also fibroatheroma with necrotic core) obstructing the stent, particularly in patients with DES ISR. The unique ability of optical coherence tomography to accurately visualise stent coverage and heterogeneous tissue within the stent has generated major interest.10 However, its potential clinical use should be investigated in prospective studies.
Treatment of Bare-metal In-Stent Restenosis
Conventional Balloon Angioplasty
The use of balloon angioplasty (BA) in patients with ISR is straightforward and consistently associated with good initial results and a very low incidence of complications.11 Lumen enlargement results from the equal contribution of further stent expansion and tissue extrusion with some longitudinal redistribution.1 Despite favourable initial clinical results, the long-term outcome of BA in this setting is shadowed by a high restenosis rate, in particular in patients with diffuse ISR.1 However, the available evidence supports the use of BA as the primary treatment in patients with ‘focal’ ISR. Other clinical and angiographic determinants of an adverse prognosis are very similar to those found in patients with de novo lesions. In our experience,11 ISR patients with diabetes and those with a short (<4 months) time interval from stent implantation to repeat intervention had an increase of adverse cardiac events during follow-up.
It is important to optimise final results using adequate balloon/ artery ratios and high pressures. However, it is important to keep in mind that aggressive BA dilations can be associated with edge dissections. Particular attention should be paid to preventing balloon slippage, especially during aggressive dilations (‘watermelon-seeding phenomenon’). We found12 that this problem is more common in severe and diffuse lesions and when oversized balloons are used. Balloon slippage should not be considered simply a technical nuisance because it may be responsible for edge dissections, cumbersome procedures with suboptimal results and, more importantly, adverse clinical and angiographic outcomes.12 Side-branches emerging from ISR lesions rarely create significant problems.13 However, large side-branches should be adequately protected and treated if their ostium is compromised. Although small side-branches may experience transient flow deterioration, this is rarely associated with clinical consequences and systematically are patent at late follow-up.13
Non-stent-based strategies of local drug delivery at the treated coronary segment have been investigated for some years in order to reduce the exaggerated neointimal response to mechanical injury.14 Expectations were met, with promising results obtained with drug-eluting balloons at both bench and bedside.14 Balloons are more efficacious than other alternatives (nanoparticles, contrast media agents or injection catheters) to deliver drugs at the vessel wall. Many drugs were studied, but eventually lipophilic drugs – such as paclitaxel – were selected due to their capacity to be absorbed and retained by the arterial wall. Results of pre-clinical and clinical studies have been encouraging. Animal studies demonstrated that this therapy reduced neointimal proliferation. Subsequently, early clinical series also showed promising results.14
In a small randomised study, Sheller et al.15 enrolled 52 patients with BMS ISR who were allocated to BA or to paclitaxel-eluting balloon. At six months, late lumen loss was significantly reduced in the paclitaxel-eluting balloon group (0.74±0.86 versus 0.03±0.48; p=0.002). In addition, the retenosis rate was 45% after BA and only 5% after paclitaxel-eluting balloon. Finally, major clinical events at one year were significantly higher in the BA arm as the result of a higher need for target vessel revascularisation.15 Recently, the same German group reported another randomised trial16 where 131 patients with BMS ISR were allocated to either paclitaxel-eluting balloon or paclitaxel-eluting stents (PES). At six-month follow-up, in-segment late lumen loss was 0.38±0.61mm in the PES group versus 0.17±0.42mm (p=0.03) in the drug-coated balloon group, resulting in a binary restenosis rate of 20 versus 7% (p=0.06). At 12 months, the rates of major adverse cardiac events were 22 and 9%, respectively, primarily due to the higher need for target lesion revascularisation in the PES group. The authors concluded that treatment of coronary BMS ISR with the paclitaxel-coated balloon was at least as efficacious and as well tolerated as PES, and that in the treatment of ISR prevention of recurrences does not require a second stent implantation.
It is clear that the results of these studies are stimulating and indeed highly provocative.15,16 However, further studies including larger number of patients from multiple centres should be performed in order to establish the real clinical value of this promising new form of therapy in routine clinical practice.
Cutting Balloon and Debulking Techniques
The cutting balloon (CB) appears ideally suited to incise neointimal tissue favouring its subsequent extrusion. Furthermore, CB anchors to the lesions during inflation and reduces the risk of displacements and dissections at the stent margins. Observational studies17 suggested that CB was superior to conventional BA in the treatment of ISR. The Restenosis Cutting Balloon Evaluation Trial (RESCUT)18 was a multicentre, randomised European trial including 428 patients with all types of ISR. At late follow-up, the binary restenosis rate (29.8 versus 31.4%) and the rate of clinical events were equivalent in both groups. However, balloon slippage was less frequently seen in the CB group (6.5 versus 25%), which required less frequent stent implantation to obtain satisfactory angiographic results. Therefore, CB may be of particular benefit when it appears particularly important to avoid damaging the adjacent vessel wall.19 With this in mind, many investigators have systematically used this tool in patients with ISR before brachytherapy and before DES implantation.
After BA, the persistence of a significant amount of residual neointimal material within the stent is readily visualised with IVUS. This might explain the poor results of BA in patients with diffuse ISR. Therefore, ablative or debulking techniques have been advocated to remove intrastent tissue and to improve angiographic results.1 Initial studies demonstrating the ability of excimer laser and rotational atherectomy to ablate the tissue obstructing the stent generated great expectations. The use of laser plus BA resulted in greater lumen gain, more ablation/extrusion of neointima and larger final lumen areas.1 Nevertheless, despite these satisfactory mechanistic initial results, high recurrence rates were demonstrated and the technique was soon abandoned.1 Rotational atherectomy also showed initial promise in these patients. Compared with excimer laser, rotational atherectomy demonstrated superior ateroablative capabilities, especially when relatively large burrs were used.1 The Angioplasty versus Rotational atherectomy for the Treatment of diffuse In-Stent restenosis Trial (ARTIST) study20 was a large randomised trial (298 patients with ISR) comparing rotational atherectomy plus BA versus BA alone. A stepped-burr approach was selected followed by low-pressure BA.
However, at angiographic follow-up, the net angiographic gain was significantly larger and the binary restenosis rate significantly lower (51 versus 65%) in the BA arm. Moreover, both procedural complications and event-free survival at six months were also favourable to the BA arm.20 Conversely, in the smaller, single-centre Rotational Atherectomy versus Balloon Angioplasty for Diffuse In-Stent Restenosis (ROSTER) trial,21 where IVUS was systematically used to exclude patients with underexpanded stents, the results of rotational atherectomy were superior to those seen in the BA arm. Although some speculated that the strategy of low-pressure adjunctive BA in ARTIST was unable to properly expand the underlying stents, the reality is that the negative results of this large trial were considered as conclusive by the interventional cardiology community and the technique is no longer used in this setting. Finally, directional atherectomy, with even more efficient debulking proprieties, obtained favourable results in initial series.1 However, this device was too bulky and the limited length of the cutter prevented the adequate treatment of most patients with diffuse ISR.
Brachytherapy demonstrated its profound antiproliferative properties precisely in patients with ISR. In fact, this technique was never found to be useful in patients with de novo lesions. In patients with ISR, brachytherapy dramatically suppressed subsequent neointimal proliferation.1 Brachytherapy should be considered as the first major breakthrough against neointimal hyperplasia in the clinical arena. Multiple randomised trials (using beta and gamma sources) have clearly established its superiority over alternative strategies (mainly BA or debulking techniques) in patients with ISR. All of these clinical trials consistently demonstrated a lower rate of restenosis and a lower need for repeat revascularisation at follow-up.1 However, the logistics involved in these procedures were highly demanding. Major efforts were made to implement this complex and sophisticated technique in routine PCI centres. As previously stated, many investigators favoured the use of CB to safely pre-dilate these lesions. The ‘geographical misphenomenon’ described with this technique consisted of the appearance of edge restenosis at sites injured by the BA and not adequately protected by intracoronary radiation.22 To overcome this complication, the effectively irradiated area should be very large including the injured site and the adjacent normal coronary segments.
Another important problem was the appearance of cases of late thrombosis after these procedures.1 The profound ability to inhibit smooth-muscle cells appeared to be closely related to a delayed endothelisation process. Accordingly, additional stent implantation was strongly discouraged and prolonged dual antiplatelet therapy recommended. Last but not least, a late catch-up phenomenon was demonstrated after brachytherapy, accounting for a significant reduction in efficacy over time.23
Two large randomised studies compared brachytherapy with DES in patients with BMS ISR.24,25 In the Sirolimus-eluting stent with vascular brachtherapy for the treatment of In-Stent Restenosis from bare metal stents (SISR) trial,24 384 patients were assigned to brachytherapy or to SES implantation. At follow-up, rates of target vessel failure were doubled (21.6 versus 12.4%) in the brachytherapy group. The TAXUS-V trial25 randomised 396 patients with BMS ISR to brachytherapy or PES implantation. Again, target-vessel revascularisation rates were significantly higher in the brachytherapy group. Subsequent studies confirmed that the initial superiority of DES is maintained at very late follow-up. For all of these reasons, together with a reduced commercial interest in the technique, brachytherapy was virtually abandoned and is no longer widely available to assist the treatment of patients with ISR.26
BMS reduce the risk of restenosis compared with BA in de novo lesions. In addition, the larger acute angiographic gain obtained by repeat stenting may confer an advantage in patients with ISR. Initially, only cases experiencing dissections after BA for ISR were considered for BMS implantation.27 Subsequently, this strategy was selected for patients with ISR achieving suboptimal results after reinterventions.28 Finally, some early series suggested that ‘elective’ BMS implantation was of value in these patients.28 Results of repeat stenting in patients with ISR were superior in elective cases compared with those requiring re-stenting for suboptimal results or complications.28 Angiographic and IVUS studies demonstrated the superior initial results obtained with repeated stenting compared with other therapeutic alternatives in patients with ISR.1 In addition, mechanistic studies demonstrated that an ‘early lumen-loss phenomenon’, resulting from tissue reintrusion into the stent lumen, may further jeopardise the results of BA or ablation techniques. 29 In a dedicated, randomised, volumetric IVUS study,30 we demonstrated that this phenomenon was virtually abolished by repeated stenting.
The Restenosis Intrastent: Balloon Angioplasty Versus Elective Sirolimus-Eluting Stenting (RIBS-II) randomised trial2 (450 patients from 24 sites in Spain and Portugal) compared the use of BA and repeat BMS implantation in patients with ISR. Initial angiographic results were significantly better after BMS implantation. However, at six-month follow-up the minimal lumen diameter was similar in both groups as the result of a larger late loss in the BMS arm. Rates of recurrent restenosis and requirement for target vessel revascularisation were similar in both arms.2 However, subsequent subgroup analyses demonstrated the superiority of BMS over BA in two particular anatomical scenarios: the pre-specified subgroup patients with large vessels (>3mm in diameter) and patients with edge ISR.31
Accordingly, even in the DES era, BMS should continue to be considered in patients with large vessels and edge ISR and whenever suboptimal results are obtained.2 In fact, results of DES to tackle ISR-related edge dissections remain unknown. This therapy should also be considered in patients not suitable for long-term dual antiplatelet therapy.
The dramatic ability of DES to suppress neointimal proliferation in all type of lesion subset stimulated their early use (off-label) in patients with ISR. Early observational studies demonstrated the efficacy of DES in patients with BM ISR. In this setting, both sirolimus-eluting stents (SES) and PES were able to provide excellent clinical and angiographic results.1 Pre-dilation with short and undersized balloons appears desirable to prevent recurrences resulting from edge ISR. Underexpanded stents have frequently been found in these patients.32 Therefore, the use of IVUS appears indicated during these interventions to guarantee optimal expansion of DES.
The Intracoronary Stenting or Angioplasty for Restenosis Reduction – Drug-Eluting Stents for In-Stent Restenosis (ISAR-DESIRE) trial33 demonstrated that DES were superior to BA in reducing the need for new revascularisations at 12 months. In this study, Kastrati et al.33 found that the incidence of restenosis was significantly reduced after SES and PES compared with BA (14.3, 21.7 and 44.6%, respectively). The RIBS-II randomised study3 was a multicentre Spanish trial that demonstrated the superiority of SES over BA in patients with BMS ISR. Patients treated with SES had a lower restenosis rate (11%) and a better one-year event-free survival rate compared with those in the BA arm. The clinical benefit was the result of a significant reduction in the rate of target vessel revascularisation. The superiority of SES over BA was consistent across 10 pre-specified clinical and angiographic subgroups. A mechanistic IVUS sub-study of RIBS-II confirmed the striking antiproliferative efficacy of DES in these patients.3
Many studies have suggested that DES are associated with an increased risk of late thrombosis, mainly when used with off-label indications. However, at four-year clinical follow-up, patients treated with SES in the RIBS-II study had an excellent prognosis.34 After adjustment for potential confounders, DES implantation, time to ISR and the ACC/AHA and Mehran angiographic classifications were identified as independent predictors of adverse prognosis.34 In these patients, the potential value of an extended use of clopidogrel therapy beyond 12 months remains to be defined.
Direct comparisons of the efficacy of DES versus BMS in patients with BM ISR have not been performed. However, in a pooled analysis of RIBS-I and II studies, we found that after adjustment for potential confounders DES provided an improved clinical and angiographic outcome compared with BMS.35
Treatment of Drug-eluting In-stent Restenosis
ISR has not been eradicated by the use of DES.1 DES ISR has distinct angiographic patterns; namely, ISR is usually focal and frequently locates at the edges.1 As for patients with BMS ISR, outcome appears to be better for focal than for diffuse DES ISR. Edge ISR is frequently found when the adjacent coronary segment has been injured during pre-dilation and subsequently not covered by DES. Underexpansion appears to be a particularly important problem in patients suffering DES ISR. In rare cases, DES fractures appear to the responsible for ISR. Both problems can be suspected angiographically, but a precise diagnosis requires IVUS interrogation. IVUS appears well-suited to identifying areas of underexpansion and to optimising final results.32 More recently, optical coherence tomography has been used to study DES ISR. Heterogeneous backscatter is frequently detected in these patients.10
The best treatment for patients with DES ISR remains unknown. Results of repeated interventions appear to be poorer in patients with DES ISR than in patients with BMS ISR.36–42 Initial angiographic results are similar, but recurrence rates are much higher in patients with DES ISR. Early studies observed binary restenosis rates as high as 42%.38 However, recent reports suggest more favourable outcomes.39–41 Information is limited by small numbers and confounders, but BA appears to be associated with particularly poor results.38
However, this strategy is still selected by some investigators in patients with focal ISR due to the lack of scientific evidence in favour of more complex approaches. Results obtained by repeated DES implantation appear to be slightly better, but still are poorer than those obtained with DES in patients with BMS ISR. Steinberg et al.42 compared 119 patients treated with DES for DES ISR with a matched group of 119 patients receiving DES for BMS ISR. The rate of target-vessel revascularisation was two-fold higher (22.2 versus 10.3%) in patients with DES ISR. This emphasises that DES ISR represents a uniquely challenging anatomical scenario and effort should be made to optimise the results of the repeated interventions; in particular, underexpansion should be aggressively managed. Some investigators advocate a switch strategy (trying a new DES) during the repeated interventions.39–41 However, information is limited and consists of small series of patients. Currently, results are mixed and inconclusive.39–41 RIBS-III is a prospective, multicentre Spanish registry aimed to assess results of re-interventions in patients with DES ISR. The ISAR-DESIRE-II trial was presented at Transcatheter Cardiovascular Therapeutics (TCT) 2009 in San Francisco. In this study, the Munich group compared SES versus PES in patients with SES ISR. Eventually, results were similar with the two stents. Further studies are urgently needed to determine the best interventional strategy for patients suffering from DES ISR, a rare yet challenging condition.
Finally, in our experience repeat DES implantation for recurrences after stenting for ISR is feasible and safe.43 Many of these patients presented with unexpandable stents, resistant to very high dilation pressures, emphasising once more that prevention of this problem is of paramount importance. However, in selected cases this strategy, producing a triple metal layer in the coronary wall, is associated with good long-term clinical outcomes.43 Nevertheless, it is clear that other alternatives (in addition to recurrent repeat stenting) should be investigated for patients with recalcitrant ISR. As previously discussed, the use of paclitaxel-coated balloons appears attractive, but only anecdotal information is currently available in patients with DES-ISR.