Coronary artery bypass grafting (CABG) is a highly effective therapy in ischemic heart disease. However, its success is affected by the limited life expectancy of saphenous vein bypass grafts (SVGs), which are still used in almost all interventions.1,2 Most recent studies report an estimated patency rate of 50–60% for SVGs after 10 years, compared to 85–95% for arterial conduits.3 Three main mechanisms are described for vein graft failure. In the early postoperative period (first month), acute thrombosis due to technical and anatomical factors is the dominant etiology. In the subacute period (second to twelfth month) the major role is played by intimal hyperplasia, resulting from the graft's adaptation to higher arterial pressures with consequent loss of endothelial inhibition and smooth muscle cell proliferation. During the late period (after the first year), accelerated atherosclerosis becomes the most important mechanism for graft stenosis and occlusion. Compared to native vessels, vein graft atheromas are more diffuse and concentric and less calcified and have poorly developed or absent fibrous caps, making them especially prone to rupture with consequent thrombotic occlusion of the graft.1,4

Stenosed or occluded SVGs can be treated by percutaneous coronary intervention (PCI) or repeat CABG. Compared to first CABG, repeat CABG is technically more difficult, carries a higher risk of death and provides less symptomatic improvement.2 As a result, PCI is currently the preferred treatment for SVG lesions. However, two major questions concerning percutaneous revascularization of CABG patients remain unanswered. The first is where to intervene: in the venous graft itself or in the native coronary artery. Percutaneous treatment of soft and friable, degenerated vein graft lesions is associated with less favorable immediate results and more adverse in-hospital events, and poses unique procedural challenges due to the frequent presence of thrombi superimposed on critical graft stenosis with a consequent tendency for distal embolization and periprocedural myocardial infarction. In the long term, results are disappointing, with high recurrence rates due to restenosis and emergence of new lesions resulting in target vessel failure (TVF).4–7 Nonetheless, only a few small retrospective studies comparing the two strategies have been published, with inconclusive results, and the general practice is to decide on an individual basis.8,9 The second question concerns the most suitable stent for PCI of SVG lesions. Although superior to balloon angioplasty,10–13 results of stent implantation in SVG lesions are not as good as in native coronary arteries.13 Initial evidence retrospectively comparing bare-metal (BMS) with drug-eluting (DES) stents tended to favor DES, with a lower incidence of major adverse cardiac events mostly due to a reduction in target vessel revascularization (TVR).14–17 However, the recent publication of the only two randomized studies in this area has questioned the superiority of DES. The RRISC (Reduction of Restenosis In Saphenous vein grafts with Cypher stent) trial, comparing sirolimus-eluting stents (SES) with BMS, showed SES to be superior in target lesion and target vessel revascularization at 6 months, with similar mortality and myocardial infarction (MI) rates.18 However, the DELAYED RRISC (Death and Events at Long-term follow-up AnalYsis: Extended Duration of the Reduction of Restenosis In Saphenous vein grafts with Cypher stent) trial showed loss of repeated revascularization benefits and significantly higher mortality in SES compared with BMS at 32 months.19 The SOS (Stenting Of Saphenous vein grafts) trial, comparing paclitaxel-eluting stent (PES) with BMS, also showed disappointing results, with less TVR and TVF in the PES group but similar MI and mortality rates at a mean follow-up of 1.5 years.20 This lack of consistent long-term benefit of DES over BMS in this context21–24 led to the recent publication of two meta-analyses, also with controversial results: one, including 19 studies with 3420 patients, reported a benefit for DES vs. BMS only in TVR and MI25; the second, including 22 studies and 5543 patients, showed significantly better results for DES in TVR, TLR and total and cardiovascular mortality.26 Although revitalized, this question seems far from settled.

The aim of our study was to compare long-term angiographic and clinical outcomes of different percutaneous revascularization options in patients with SVG lesions, with regard to intervention strategy – PCI in SVG or native coronary artery – and type of stent to implant – BMS or DES.

Coronary artery bypass surgery remains the most effective treatment for complex multivessel disease, but its long-term results are affected by the short life expectancy of SVGs.1,2 Nearly half of SVGs are occluded 10 years after implantation,3 making percutaneous revascularization of CABG patients a frequent procedure that represents 10–15% of interventions in most catheterization labs.27 Our study was aimed to clarify different therapeutic options regarding PCI of SVGs only, vs. PCI of native vessels only or a combined approach.

Concerning the intervention strategy and the question of where to intervene, in our study there were no differences in mortality, MI or TVF between patients revascularized in the saphenous graft, in the native coronary artery or in both territories. These results are in agreement with other published studies.8 Angioplasty of the SVG instead of the native vessel is feasible and safe and should be considered a valid option in selected cases, despite the expected higher difficulty of interventional maneuvers inside a vein graft and the higher risk of complications such as vein wall perforation and dislodgment and distal embolization of atherosclerotic and thrombotic material. The key element must be careful and individualized patient selection for each strategy, guided by angiographic indications of ischemia. In our population, operator-based decisions led preferentially to native vessel PCI (74.4%) rather than other strategies. This decision appeared not to be influenced by the patient's clinical status, since baseline characteristics were similar in the two subgroups (Table 2). Combined revascularization in both SVG and native vessel proved to be a good option in complex cases, with similar results despite the presence of more serious coronary disease and more extensive intervention.

Regarding the second issue, the most suitable type of stent, although PCI of coronary lesions with DES has been shown to be superior in several clinical scenarios, the pivotal trials which established the anti-restenotic effect of DES specifically excluded patients with vein graft disease due to its distinct and mostly unknown physiological environment.28 Studies on the effectiveness of DES in SVG disease are limited in number, population size, duration of follow-up, and, with only two exceptions – the RRISC and SOS trials, respectively with sirolimus and paclitaxel-eluting stents – are all retrospective. Most importantly, published results are conflicting, target vessel revascularization being the only endpoint for which DES consistently showed better results than BMS in SVG,5,26 and even this is unsure in the long term, as seen in the DELAYED-RRISC trial.19

Our analysis showed different results in PCI performed in native vessels and in SVGs. In native coronary arteries DES showed significantly lower mortality (p<0.001) and a tendency for lower TVF at 27-month mean follow-up. These results underline the well-documented impact of antiproliferative agents in the prevention of stent restenosis, with consequent benefits in mortality and morbidity. However, our data suggest that in the setting of SVG angioplasty DES do not offer a significant clinical advantage over BMS. In SVG interventions BMS show similar long-term performance with regard to mortality and a tendency for better results in MI and TVF. These results mirror the findings of other published comparison data with long-term follow-up. Possible reasons for the non-inferiority of BMS in this setting are the slightly larger diameters of implanted BMS, the highly thrombotic milieu of saphenous grafts, with potential for a more pronounced inflammatory and thrombotic reaction after DES deployment, and the progressive nature of SVG disease, with a large proportion of events attributable to progression of disease in untreated segments rather than to restenosis.29 The need for long-term follow-up has become more evident in view of the recent results of the DELAYED RRISC trial, which reported a trend for more repeated revascularization procedures in the DES group after the first 6 months, probably reflecting a “late catch-up” phenomenon resulting in more frequent late and very late thrombosis and late restenosis than in native coronary arteries.19 The use of BMS in SVG PCI may therefore be preferable in this cohort of patients.

In patients with saphenous graft disease there were no significant long-term follow-up differences between intervention in SVGs and in native vessels. The choice between drug-eluting and bare-metal stents is related to intervention strategy: while in native vessel angioplasty DES should remain the first choice due to their survival advantage, in SVG intervention the use of BMS was not inferior.

The authors have no conflicts of interest to declare.