Bypass Graft Interventions

Introduction

Two types of grafts are currently used for coronary artery bypass graft surgery (CABG): saphenous vein grafts (SVGs) and arterial grafts (internal mammary artery grafts [IMAs], radial artery grafts, and gastroepiploic grafts). IMA grafts have the best long-term patency rates, but in most cases, SVGs are still being used to bypass all coronary arteries that need revascularization. SVGs have high rates of failure, which increase with increasing time post-CABG ( Figure 11-1 ).

Epidemiology

Angina in patients with prior CABG can be due to native coronary artery disease progression, bypass graft disease, or proximal subclavian artery stenosis development ( Figure 11-1 ). In an analysis from the National Cardiovascular Data Registry (NCDR), between 2004 and 2009, percutaneous coronary intervention (PCI) in prior CABG patients represented 17.5% of the total PCI volume (300,902 of 1,721,046). The PCI target vessel was a native coronary artery in 62.5% and a bypass graft in 37.5% of cases. Bypass graft PCI represented 6.6% of all PCI and consisted mainly of SVG PCI (6.1% of all PCI), followed by arterial graft PCI (0.44% of all PCI), and, rarely, both arterial graft and SVG PCI (0.04% of all PCI). The proportion of SVGs as PCI target vessels increased after 5 years and even more so after 10 years from CABG ( Figure 11-2 ). From January 1, 2010, through June 2011, bypass graft and SVG PCI constituted 6.0% and 5.5%, respectively, of the total NCDR PCI volume.

Indications for Bypass Graft Interventions

Patients who present with bypass failure can be treated with redo CABG, medical therapy, PCI of a native coronary artery, or bypass graft PCI.

Redo CABG is infrequently performed due to high morbidity and mortality. Redo CABG can also cause injury of patent grafts, which is especially concerning for IMA grafts. In the Angina With Extremely Serious Operative Mortality Evaluation (AWESOME) trial, the risk for subsequent clinical events was similar after redo CABG and after PCI. According to the 2011 American College of Cardiology/American Heart Association (ACC/AHA) PCI guidelines, redo CABG is favored in patients with vessels unsuitable for PCI, multiple diseased bypass grafts, availability of the IMA for grafting chronically occluded coronary arteries, and good distal targets for bypass graft placement. Factors favoring PCI over CABG include limited areas of ischemia causing symptoms, suitable PCI targets, a patent graft to the left anterior descending artery, poor CABG targets, and comorbid conditions.

In patients with SVG lesions, native coronary artery PCI is preferred over SVG PCI, supplying the same territory because of better short- and long-term outcomes, especially in diffusely diseased and degenerated SVGs. Indeed, a native coronary artery was the target vessel in the majority of prior CABG patients undergoing PCI in NCDR. However, native coronary arteries supplied by a failing SVG may be chronic total occlusions (CTOs) that can be challenging to recanalize, although the SVG can, at times, be used as a conduit for retrograde native vessel revascularization.

SVG PCI has two major limitations: high rates of distal embolization and in-stent restenosis ( Figure 11-3 ), which are discussed in subsequent sections. Patients undergoing SVG PCI are usually older and have more comorbidities compared with patients undergoing native coronary artery PCI, and as a result, they are at high risk for subsequent events, both cardiovascular and noncardiovascular.

Distal Embolization and Embolic Protection Devices

SVG lesions can have complex morphology ( Figure 11-4 ) and friable atheromas that can result in distal embolization during PCI ( Figure 11-5 and Videos 11-1 and 11-2 ). Distal embolization can cause no reflow and acute ST-segment elevation or present as asymptomatic cardiac biomarker elevation. Cardiac biomarker elevation post-SVG PCI (especially CK-MB increase >5× upper limit of normal) has been associated with increased mortality; hence, it is important to prevent distal embolization or promptly treat it if it occurs.

Use of an embolic protection device (EPD) is the only proven strategy for preventing distal embolization during SVG PCI ( Figure 11-6 and Videos 11-3 and 11-4 ). EPDs capture debris liberated during PCI before it enters the coronary microcirculation causing injury. As of January 2015, three EPDs are available in the United States ( Figure 11-7 , Table 11-1 ): the FilterWire (Boston Scientific, Natick, Massachusetts), the Spider (Covidien, Mansfield, Massachusetts), and the GuardWire (Medtronic Vascular, Santa Rosa, California). The first two EPDs are filters, whereas the GuardWire is a 0.014 inch guidewire with a distal balloon that when inflated stops antegrade flow; after completion of PCI, any column of blood within the SVG is aspirated with a thrombectomy catheter before restoring antegrade flow ( Figure 11-8 ). The GuardWire allows “complete” protection, that is, capture of all released particles and humoral factors, in contrast to filters that only capture larger size particles. Moreover, it has a lower crossing profile and requires a shorter landing zone (20 mm vs. 25-50 mm for filters). However, the GuardWire can be cumbersome to use and cessation of blood flow may be poorly tolerated by some patients, especially those in whom the SVG supplies a large area of myocardium.

The Saphenous vein graft Angioplasty Free of Emboli Randomized (SAFER) trial randomized 801 patients undergoing SVG PCI to GuardWire or stenting over a standard guidewire. The study's primary endpoint (composite of death, myocardial infarction, emergency CABG, or target lesion revascularization by 30 days) occurred in 65 patients (16.5%) assigned to control versus 39 patients (9.6%) assigned to the GuardWire (p = 0.004). This significant 42% relative reduction in the primary endpoint was driven by a reduction in the incidence of myocardial infarction (8.6% vs. 14.7%, p = 0.008). No reflow was also less common in the EPD group (3% vs. 9%, p = 0.02). Given the significant clinical benefit with EPD use, subsequent SVG PCI studies used a noninferiority design to compare one EPD to another, as summarized in Table 11-2 .

Choosing an EPD for a specific SVG lesion is based on several factors, such as lesion location, device availability, local expertise in EPD use, and the potential hemodynamic consequences of SVG flow cessation ( Figure 11-9 ). SVG body lesions can be protected with any EPD, as long as there is an adequate landing zone. Ostial SVG lesions should only be protected with a FilterWire or Spider, since use of the GuardWire could result in debris embolization in the aorta from the stagnant column of blood in the SVG. Although ostial SVG lesions were excluded from the pivotal SVG PCI trials, a recent study showed high success rates with EPD use in ostial lesions at the cost of difficulty retrieving the filter in 11% of the lesions; one of these patients developed acute stent thrombosis causing cardiac arrest ( Figure 11-10 ). Moreover, ostial and distal anastomotic lesions are more likely to consist of fibrous tissue and less likely to contain lipid core plaque compared with SVG shaft lesions, and hence may be less likely to embolize. Distal anastomotic lesions ( Figure 11-9 ) cannot be protected with any of the currently available EPDs (manufacturing of proximal embolic protection devices stopped in 2012). Distal anastomotic lesions constitute approximately 19% of SVG lesions undergoing PCI. Routine use of EPDs in SVG in lesions due to in-stent restenosis may be unnecessary because these lesions are usually caused by neointimal proliferation, making distal embolization unlikely. Similarly, EPD use may not be necessary for recently implanted (<2 years old) SVGs that have not had enough time to develop significant degeneration predisposing to embolization.

The 2011 ACC/AHA PCI guidelines state that, “EPDs should be used during SVG PCI when technically feasible” (Class I indication, level of evidence B). ¹¹ Yet, EPDs were only used in 23% of SVG PCI between 2004 and 2009 in the NCDR registry, although some studies suggest that up to 77% of SVG lesions are eligible. Potential explanations for EPD underutilization include the following: lack of reimbursement for EPDs; technical difficulties and lack of familiarity with EPD use; fear of EPD-related complications, such as device entrapment and acute vessel occlusion; and uncertainty regarding the magnitude of clinical benefit with EPD use. The SAFER trial was performed before the era of potent ADP P2Y12-receptor inhibitors. In the “Is Drug-Eluting-Stenting Associated with Improved Results in Coronary Artery Bypass Grafts?” (ISAR-CABG) trial in which all patients were pretreated with 600 mg clopidogrel before PCI, despite very infrequent EPD use (in <1% of SVG PCIs), the incidence of myocardial infarction was 6%, which is lower than the incidence of myocardial infarction reported in the control arm of the SAFER trial (14.7%).

When use of an EPD is not feasible (for example, in distal anastomotic lesions, lesions without an adequate landing zone, tight lesions that cannot be crossed with an EPD, or thrombotic lesions in which EPD insertion may cause embolization by itself) ( Figure 11-9 ), alternative interventions to reduce distal embolization (or obviate its adverse consequences) include the following: (a) intragraft vasodilator administration (such as adenosine, nitroprusside, nicardipine, and verapamil ), (b) use of an excimer laser, (c) implantation of slightly undersized stents (which did not result in higher restenosis rates in one retrospective study ³⁹ ), (d) direct stenting without predilation, or (e) use of micromesh-covered stents, which are currently not approved for clinical use in the United States.

SVG Stenting

Drug-eluting stents (DES) are currently preferred for SVG PCI to reduce the risk for in-stent restenosis ( Figure 11-2 ) based on three randomized controlled clinical trials ( Table 11-3 ).

The Reduction of Restenosis In Saphenous vein grafts with Cypher sirolimus-eluting stent trial (RRISC) compared the sirolimus-eluting stent (SES) Cypher (Cordis, Warren, New Jersey), with a bare-metal stent (BMS) in 75 patients and reported less angiographic restenosis and target lesion revascularization at 6 months. During long-term follow-up, the SES group had higher mortality (29% vs. 0%, p = 0.001) rates, and target vessel revascularization was similar in both groups; however, most patients died because of noncardiac causes or cardiac causes unrelated to the target SVG.

The Stenting Of Saphenous Vein Grafts trial (SOS) trial compared a paclitaxel-eluting stent (PES), Taxus (Boston Scientific, Natick, Massachusetts), to a similar BMS in 80 patients and demonstrated less angiographic restenosis and lower incidence of clinical events (both repeat revascularization and myocardial infarction) with PES.

The ISAR-CABG study randomized 610 patients to a first-generation DES (SES, PES, or sirolimus-eluting ISAR stent) or a BMS. At 12 months, the DES group had a significantly lower incidence of target lesion revascularization (7% vs. 13%, p = 0.01) compared to BMS, with a similar incidence of all-cause death (5% vs. 5%, p = 0.83), myocardial infarction (5% vs. 6%, p = 0.27), and definite or probable stent thrombosis (1% vs. 1%, p = 0.99). The incidence of occlusive restenosis was 6% with DES and 12% with BMS (p = 0.008). ⁵³ Stent failure in SVG lesions frequently presents with an acute coronary syndrome.

The 2011 ACC/AHA PCI guidelines state, “DES are generally preferred over BMS” for SVG lesions. ¹¹ There is limited data on whether second-generation DES further improve outcomes compared to first-generation DES. In a pilot study use of the second-generation everolimus-eluting stent was associated with high rates of stent strut coverage but also high malapposition rates at 12 months postimplantation. Three retrospective studies have been published, with one showing better (less target vessel revascularization) and two showing similar outcomes with second-generation DES. DES with bioabsorbable polymer showed encouraging outcomes in the NOBORI 2 registry, although the risk of death, myocardial infarction, and target vessel revascularization were higher than the corresponding rates of patients undergoing treatment of non-SVG lesions. Preliminary reports have shown promising results with bioabsorbable stents in SVGs, but further study is needed.

BMS implantation should be avoided in distal anastomotic graft lesions due to high rates of in-stent restenosis; DES implantation or balloon angioplasty alone is preferred, and both of these strategies appeared to provide similarly good outcomes in a pilot study. SVG shaft lesions have been associated with a higher risk of distal embolization and long-term adverse clinical events compared to proximal and distal anastomotic lesions.