Saphenous vein grafts (SVG) have been extensively used in coronary artery bypass surgery (CABG) as additional conduits to arterial grafts. However, it is well known that vein grafts may develop degenerative processes that may result in stenosis or occlusion1 in 12–20% of patients at the end of the first year and approximately 50% by 10 years.2 The main causes of graft disease are intimal hyperplasia, mostly within the first 12 months, and atherosclerotic plaque in the following years.3 Surgical reoperation of these patients is feasible but associated with a higher mortality rate and a poorer clinical outcome compared with the first operation.4 Consequently, percutaneous coronary intervention (PCI) has become an attractive therapeutic alternative and has been widely performed during the last few decades. Currently, it covers almost 10% of the total PCI procedures performed in the US.5
In the first performed balloon angioplasty, Andreas Gruentzig rapidly noted the poor outcome of percutaneous intervention in SVGs. In the first 50 cases reported, he described a very high restenosis rate (60%) in the subset of SVG-treated lesions, which led him to note: “the different kind of disease may explain the high incidence of recurrence in graft stenoses.”6 These disappointing results led Gruentzig to wonder about the role of balloon angioplasty in SVGs. Today, PCI of SVG stenoses is accompanied by less favourable short- and long-term outcomes compared with PCI of native vessels. PCI of SVG carries a significant risk of major adverse clinical events (MACE), mainly myocardial infarction (MI) or reduced antegrade flow (no-reflow phenomenon) basically caused by distal embolisation.7 Furthermore, a high restenosis rate and atherosclerotic disease progression in the graft are responsible for a poor outcome in mid- and long-term follow-up.
Several technical improvements, such as the use of stents,8,9 distal protection devices and the direct stenting technique, have contributed to ameliorate acute angiographic and clinical outcomes; however, long-term results are still not similar to those of native vessels. Controversial data are emerging on the potential role of drug-eluting stents (DES) in this setting: they seem to reduce short-term target vessel revascularisation (TVR), but their impact on long-term clinical outcome is still debated.10,11 The aim of this article is to examine the best available therapeutic options for patients undergoing PCI of SVG lesions.
Prevention of No-Reflow and Periprocedural Myocardial Infarction
Role of Distal Protection
Graft thrombosis and intimal hyperplasia are the main factors responsible for the few cases of SVG closure within 30 days and within one year after operation, respectively. The development of atherosclerotic plaques is instead the main mechanism of late (over one year) vein graft failure, the most frequent cause of recurrence of symptoms in patients treated with CABG. Plaques are usually concentric, diffuse, lipid-rich and with a high content in thrombotic material and inflammatory cells. All these features make SVG plaque friable and prone to thrombotic occlusion and PCI-induced distal embolisation. This embolisation of atherothrombotic debris is the pathophysiological factor that is responsible for most of the acute angiographic and clinical complications of the procedure.7,12–17 Embolic protection devices (EPD) have been developed to reduce this risk. These have been tested in a randomised clinical trial and are currently widely used in clinical practice.
There are two types of EPDs: filters and occlusion-aspiration devices. Each of them has some peculiarities that may represent an advantage over the others under specific angiographic conditions. There is not a single EPD device suitable for every treatable lesion; in selecting the type of device the operators should carefully take into consideration the anatomy of the graft. Distal filter devices assure distal perfusion and allow injection of contrast media during PCI, while trapping most particulate debris (see Figure 1). Distal filters include the FilterWire-EZ (Boston Scientific®), the Interceptor Plus Coronary Filter System (Medtronic®) and the Spider (eV3®). Limitations to their use include the need for a suitable landing zone, the frequent need to pre-dilate the lesion to allow the passage of the filter device, the risk of embolisation while crossing the lesion and the possible failure to capture particles <100μm.
Proximal and distal occlusion-aspiration devices halt antegrade flow during intervention. The stagnant column of blood containing particulate debris and humoral mediators is aspirated before relieving distal occlusion, thus allowing the retrieval of particles of any size. Distal occlusion devices include the PercuSurge GuardWire (Medtronic®) and the TriActive system (Kensey Nash®). Disadvantages account for inadequate vessel imaging during the procedure, possible embolisation while crossing the lesion with EPD, distal ischaemia causing intolerance in some patients and possible balloon-induced injury of the SVG wall.
The Proxis embolic protection system (St Jude Medical®) is a proximal occlusion device. The advantages of this device include the prevention of embolisation also while crossing the wire, the protection of all distal side branches (Y conduits) and the possibility of aspiration of large thrombus by using larger lumen catheters. It can also be used to increase guide support by anchoring the guiding catheter during inflation of device balloon. Limitations of this device are the requirement of a landing zone precluding its use in proximal lesions and a potential low tolerability in terms of the prolonged balloon inflation and the consequent induced myocardial ischaemia.
The benefit of EPDs in SVG PCI procedures has been well established, and there are no trials demonstrating the superiority of a specific device among the others. The Saphenous vein graft Angioplasty Free of Emboli Randomized (SAFER) trial18 compared the distal occlusion device (PercuSurge GuardWire, Medtronic®) with conventional guidewire in SVG PCI procedure with stent implantation in 801 patients. This trial was halted early because of a significant benefit associated with the use of GuardWire in term of MACE (death, MI, emergency by-pass and TVR) at 30 days. The use of GuardWire led to a 6.9% absolute (42% relative) reduction in the 30-day primary end- point (9.6 versus 16.5%; p=0.004), with a significant reduction in the incidence of MI (8.6 versus 14.7%; p=0.008) and no-reflow phenomenon (3 versus 9%; p=0.001), mirrored by an increase in post-procedural Thrombolysis In Myocardial Infarction (TIMI) III flow (98 versus 95%; p=0.04).
After SAFER, many trials sought to compare different EPDs in order to identify a potential superiority of a device over the others. Five trials (FilterWire EX Randomised Evaluation [FIRE],19 Protection During Saphenous Vein Graft Intervention to prevent Distal Embolization [PRIDE],20 Saphenous Vein Graft in a Distal Embolic Randomized [SPIDER],21 Proximal Protection During Saphenous Vein Graft Intervention [PROXIMAL]22 and Assessment of the Medtronic Ave Interceptor Saphenous Vein Graft Filter System [AMEthyst])23 compared the different types of EPD. These non-inferiority trials demonstrated the substantial equivalence of all the studied devices in terms of MACE at 30 days (see Figure 2).
According to the available data, EPDs play a pivotal role in reducing the incidence of MACE in the short term and, therefore, received a class Ib recommendation in the European Society of Cardiology (ESC)24 and American College of Cardiology (ACC)/American Heart Association (AHA)25 guidelines in PCI procedures of degenerated SVG.
Despite these recommendations, the utilisation of these devices is far from being common. In an analysis of more than 19,000 SVG PCI procedures in the ACC/National Cardiovascular Data Registry from January 2004 to March 2006, EPD usage was limited to 22% of the whole group.26
Direct stenting, the technique of stent implantation without previous balloon pre-dilatation, is advisable in cases of SVG PCI. The rationale of this technique is that pre-dilatation seems to be an important cause of distal embolisation. A study performed at Washington Heart Center27 in 2003 underlined that direct stenting during SVG PCI could reduce the incidence of peri-procedural MI (10.7 versus 18.4%; p<0.024) and lower the amount of maximum creatine kinase-MB isoenzyme (CK-MB) release (9.5 versus 19.6ng/ml; p<0.001). In 2008, Okabe et al. compared direct stenting with any technique associated with the use of EPDs. The authors reported no significant difference in terms of CK-MB rise, thus indicating direct stenting as a possible alternative to EPDs.28 However, a post hoc analysis focusing on patients treated with direct stenting and included in the SAFER trial showed that EPDs use was associated with a clear clinical benefit (MACE at 30 days with EPD 6% versus without EPD 14%; p<0.001),29 suggesting that EPDs are advisable anyway.
Adjunctive Medical Therapy
Salloum et al.30 investigated the amount of vasoconstrictor molecules release during SVG PCI. Laboratory measurements in blood selectively aspirated through the graft before and after PCI showed a significant 10-fold increase of serotonin and a strong release of endothelin subsequently to the PCI. These data support the concept of a potential role of vasodilator agents in the treatment and prevention of no-reflow phenomenon.
Many studies performed in patients undergoing PCI of native coronaries showed that the intra-coronary administration of vasodilators such as calcium-channel blockers (diltiazem and verapamil) or adenosine and nitroprusside resulted in a reverse from no-reflow to normal flow in around 90% of cases.31,32 Nicardipine, a dyhydropiridine calcium-channel blocker, exerted promising effects. In an observational study conducted in 83 consecutive SVG PCI procedures, the use of nicardipine immediately before direct stenting demonstrated an acceptable no-reflow rate of 2.4% and a global in-hospital MACE rate of 3.8%.33
IIb/IIIa glycoprotein inhibitors are widely used in managing no-reflow phenomenon during PCI in native vessels. However, these drugs seem to offer no benefit in the prevention and treatment of no-reflow phenomenon in the setting of SVG intervention. In a post hoc analysis of the SAFER trial, patients receiving IIb/IIIa inhibitors showed a two-fold rate of MACE at 30 days compared with those not receiving the drug and independently from the use of Percusurge.34 A meta-analysis performed by Roffi et al. in 2002,35 pooled the data of patients treated with PCI in SVG and enrolled in IIb/IIIa inhibitors trials. This meta-analysis showed a trend towards a worse outcome associated with the use of IIb/IIIa inhibitors. According to these results it seems that the use of IIb/IIIa inhibitors does not offer any significant benefit in the setting of SVG PCI.
Drug-eluting Stents in Saphenous Vein Graft Intervention
SVG PCI shows a higher but also delayed (>12 months) rate of restenosis compared with PCI in native coronary arteries. DES showed a superior efficacy for reducing restenosis rate and TVR compared with bare-metal stents (BMS) in native coronary arteries. However, in most of the DES pivotal trials, patients with SVG disease were excluded from randomisation and, to date, the regulatory authorities consider SVG PCI to be an off-label indication to the use of DES. From a theoretical ground, some points seem not to recommend the use of DES in this setting: the larger size of the vessel and, therefore, the potential lower benefit of the device and the need to avoid high pressure of balloon inflation during stent deployment in order to avoid the risk of distal microembolisation (but with a consequent risk of stent incomplete apposition). On the other hand, the high restenosis rate after BMS implantation in SVG could make the use of DES desirable.
Currently, there are a few small and specific studies addressing the matter of the potential benefit of DES implantation in SVG intervention, but the available data are mostly conflicting and inconclusive.
A recent registry by Latib et al. compared 127 patients treated with DES with 131 patients who received BMS for SVG stenoses.36 At two years there were no statistical differences in death (8.7 versus 7.8%), MI (6.3 versus 9.4%) or TVR (19.7 versus 24.2%) between DES and BMS, respectively. In the cohort of patients enrolled in the Strategic Transcatheter Evaluation of New Therapies (STENT) registry,37 1,380 patients received a SVG intervention: 820 were treated with DES and 348 with BMS.
At nine-month follow-up the use of DES was associated with fewer MACE (14 versus 21%; p=0.001), a lower composite of death or MI (8.7 versus 14%; p=0.006) and TVR (hazard ratio [HR] 0.36; p<0.001). At two years, DES use was associated with a lower incidence of death, but the composite of death and MI and the benefit in reducing TVR and stent thrombosis found at nine months were no longer present.
However, these two registries shared a selection bias: the patients treated with DES versus BMS were significantly different in terms of several clinical features at entry. In the Latib series in particular, patients treated with DES had a significantly higher risk feature, in that they more frequently had diabetes, (33.1 versus 15.3%), had older grafts (11.6±5.3 versus 9.6±5.2 years), had more total occlusions (7.7 versus 1.2%), smaller graft size (3.16±0.66 versus 3.44±0.76mm) and longer lesions (34.1±25.1 versus 22.7±11.6mm stent length). After adjusting for baseline clinical differences, DES showed a significant reduction in two-year TVR (HR 0.31, 95% confidence interval [CI] 0.14–0.66; p=0.002). Similarly, the STENT registry BMS group had more emergent procedures – approximately twice as many ST-segment elevation MIs – larger vein graft diameter (3.7 versus 3.3mm; p=0.001) and more no reflow (6.9 versus 3.3%; p=0.003) compared with the DES group.
The STENT registry showed that DES implantation was associated with a reduction of TVR at nine months that disappeared at two years. This late ‘catch-up’ phenomenon has been described in other trials. The Reduction of Restenosis In Saphenous vein grafts with Cypher sirolimus-eluting stent (RRISC) trial,38 a small randomised trial comparing sirolimus-eluting stents (SES) and BMS in SVG PCI showed at six-month angiographic follow-up less TVR and restenosis among DES-treated patients (14 versus 33%; p=0.03 and 5 versus 27%; p=0.01, respectively). However, at 32 months, late catch-up was observed and TVR rates were similar between the two arms (34 versus 38%; p=NS).
Short-term intravascular ultrasound (IVUS) studies showed that DES reduce intimal hyperplasia in SVG early after stent implantation, but their impact on the development of ‘new’, in-segment, atherosclerotic plaques and disease progression in non-treated segments have not yet been well characterised. This matter is relevant, as the non-target site could be a potential cause for this catch-up phenomenon. The Stent or Surgery (SoS) trial,39 a prospective, randomised trial of 80 patients with SVG lesions (39 treated with BMS and 41 with paclitaxel-eluting stents [PES]) reported a highly significant reduction in target lesion revascularisation (TLR) (5 versus 28%; p=0.003) but only a modest and non-significant trend in reduction in TVR (15 versus 31%; p=0.08) after 1.5 years in patients treated with DES, suggesting that many patients (about 10%) may require a subsequent non-target graft intervention after the procedure.
This late catch-up phenomenon is debated as it is not evident in each study. In particular, a matched case-control study from Thoraxcenter (Rotterdam, Holland) analysed 250 patients divided into two groups treated with DES versus BMS at four-year follow-up.40 In this real-world population there was no sign of catch-up phenomenon in TVR following DES implantation in SVG, and both total and cardiac death were identical in the two groups.
A recent meta-analysis presented by Navarese et al. at the European Association of Percutaneous Cardiovascular Interventions (EAPCI) Congress (EuroPCR 2010), pooled randomised and non-randomised studies and analysed multiple end-points. No differences were found in terms of MI (odds ratio [OR] 0.92, 95% CI 0.64–1.33; p=0.67) between DES and BMS, while DES showed a clear benefit in TVR (OR 0.53, 95% CI 0.39–0.72; p=0.0001), which was confirmed by a separate sub-analysis of randomised trial and observational studies.41 Finally, mortality showed opposite findings when pooling separate randomised and non-randomised studies.
In summary, the data suggest that there is no evidence of superiority of DES over BMS in SVG intervention. To exhaustively address these controversial findings, randomised trials powered for mortality and MI with adequate long-term follow-up are highly warranted.
Emerging Techniques and Devices
Undersized Drug-eluting Stent Implantation
A recent Korean-American study retrospectively analysed the intra-hospital outcome of 209 patients treated with IVUS-guided SVG PCI between 2002 and 2005.42 All of the patients underwent DES implantation (153 SES and 56 PES). By dividing the population into three groups according to post-PCI ratio stent diameter/IVUS average reference lumen diameter (group I <0.89, group II 0.9–1, group III >1), the authors showed no significant differences in stent malapposition and TLR, but a significant difference in post-PCI CK-MB elevation between the three groups (>3 x Upper Limit of Normal (ULN): group I 6%, group II 9%, group III 19%; p=0.025). The result of this study is consistent with a previous study by Iakovou et al. showing that an aggressive stent expansion resulted in greater MI rates, without any advantage in terms of TVR.43
Percutaneous Coronary Intervention of Moderate Stenoses
Rodes-Cabau et al.44 reported that about a half of patients with mild to moderately degenerated SVGs show significant disease progression after 15 months despite low-density lipoprotein (LDL) levels <90mg/dl. Ellis et al.45 first reported the important prognostic implications of moderate SVG lesions (cardiac event rate at three-year follow-up 45 versus 2% in patients with and without moderate lesions, respectively). At the 2009 edition of the ACC Scientific Session/i2 Summit the The Moderate VEin Graft LEsion Stenting With the Taxus Stent and Intravascular Ultrasound Pilot Trial (VELETI) trial was presented. This trial compared an aggressive strategy (PCI with PES) versus medical therapy of 70 moderate non-significant SVG lesions evaluated at 12 months with angiography and IVUS. The authors showed that ‘sealing’ moderate SVG lesions with PES significantly reduced SVG disease progression, and this resulted in a trend towards lower MACE at one year (3% in DES group versus 19% with medical therapy; p=0.09).
Theoretically, the use of covered stents in SVG intervention could be beneficial. The opportunity of these stents to entrap atherosclerotic plaque and superimposed thrombotic debris underneath the device struts could prevent distal microembolisation. Some preliminary data in terms of the use of such devices come from A Prospective Randomized Trial Evaluating the Symbiot Covered Stent System in Saphenous Vein Grafts (SYMBIOT) III trial, in which the Symbiot self-expanding covered stent (Boston Scientific®) was used, and the trial showed interesting acute results with a trend towards the reduction of post-PCI CK-release. It was associated with a poor long-term outcome with a higher rates of restenosis and need of TVR and TLR compared with BMS.46 The MGuard stent, a recently developed mesh-covered stent (InspireMD) has been studied in the INSPIRE trial, a registry enrolling 30 patients affected by SVG lesions and native coronary stenosis at high risk of embolisation that was recently presented at EuroPCR 2010.47 Almost 53% of the enrolled patients had a SVG lesion treated with MGuard-covered stent. In this cohort, the use of MGuard stent can assure a good acute result with a TIMI III flow in 100% of the population. One-year results showed a total MACE rate of 23.3% (TVR 20%), but systematic use of this device continues to be prevented by the unavailability of a comparison trial with BMS.
SVG intervention is an attractive therapeutic alternative to re-operation in cases of SVG failure, but is associated with poor acute and long-term results.
Distal embolisation is the main factor influencing immediate procedural success rate. This phenomenon can be significantly reduced by a large variety of EPDs, each of whom may decrease the incidence of periprocedural MI and no-reflow phenomenon. There is no evidence of clear-cut superiority of any single device over the others. Direct stenting and avoiding stent overexpansion are technical glitches that seem to reduce CK-MB release and periprocedural MI. Intra-coronary vasodilators can be of help in preventing and treating no-reflow phenomenon, whereas no benefits accrued from use of IIb/IIIa inhibitors (see Table 1).
The potential role of DES in improving long-terms result of SVG intervention is still under debate. To date, the systematic use of DES in SVG PCI is not based on evidence. Well-designed, large randomised trials powered for hard end-points are highly desirable.
Some emerging techniques and devices such as undersized DES implantation and covered stent represent interesting hypotheses that need to be adequately tested. Ôûá