Cardiac Catheterization after Bypass Surgery

Abstract

After coronary artery bypass graft (CABG) surgery, the typical patient will have progression of the original native coronary disease as well as atherosclerosis of the bypass grafts. When this leads to angina or myocardial infarction, repeat cardiac catheterization may be necessary. However, the risks of catheterization in post-CABG patients are higher than in non-CABG patients, and the benefits are smaller, so optimal medical therapy should be employed and clear indications should be present before post-CABG catheterization is undertaken. In the past decade, two advancements have been made in strategies for post-CABG catheterization. First, for patients with a left internal mammary artery graft, left radial access should be routinely used and is safer than femoral access. Second, diseased saphenous vein bypass grafts may offer a retrograde approach to chronic total occlusions of the native artery. When successful, retrograde stenting of the bypassed native coronary artery is more durable than interventions on the saphenous vein graft supplying it. This chapter summarizes indications, techniques, and tricks of catheterization and strategies for coronary intervention in patients with prior CABG.

1. Introduction

The two main methods of revascularization in coronary artery disease are percutaneous coronary intervention (PCI) and coronary artery bypass surgery (CABG). In modern medicine, coronary artery bypass surgery is mostly reserved for the most severe or complex coronary artery disease. Patients who are status post-CABG can develop further coronary disease and myocardial ischemia in the years following surgery. As in any other patient who is suspected of having coronary artery disease, cardiac catheterization provides the definitive test (angiography) and is often the treatment modality of choice (PCI) in patients with prior CABG. This chapter aims to highlight the most important aspects of cardiac catheterization, coronary angiography, bypass graft angiography, and percutaneous coronary intervention in patients who are status post coronary artery bypass surgery.

1.1 Types of bypass grafts

The left internal mammary artery (LIMA) graft to the left anterior descending (LAD) coronary artery provides CABG with its primary benefit over PCI in multi-vessel disease. The LIMA is a branch of the left subclavian artery, which itself branches from the aortic arch. The LIMA arises from the inferior-anterior aspect of the subclavian artery and courses caudally down the left chest. This graft is generally used as an in situ graft with its free end anastomosed to a coronary artery (usually the LAD).

Other than the LIMA, other bypass graft options include the right internal mammary artery (RIMA), radial artery, and saphenous veins. Most often grafts to arteries other than the LAD utilize saphenous veins. These are harvested from the legs and an anastomosis is created most often from the ascending aorta to the target coronary artery. Rarely an in situ gastroepiploic artery is anastomosed to the right coronary artery or the inferior epigastric artery is harvested and used as a free graft anastomosed to the aorta.

Free arterial grafts are superior to saphenous vein grafts (SVGs) [1, 2] however their use is limited by several factors. Radial artery grafts must meet stringent requirements before harvesting for use in CABG. Rarely radial arteries cannot be used because they are too small, previously traumatized (i.e. prior transradial catheterizations), or supply all blood flow to the hand. The RIMA can also be used, either in situ or as a free graft, but the use of both the LIMA and the RIMA is associated with an increased risk of sternal wound infections [2]. For these reasons, SVGs remain the most frequently used graft other than the LIMA.

1.2 Configurations of bypass grafts

Most commonly grafts have a single origin and single terminal anastomosis. However, several variations are used by surgeons:

• “Jump” or sequential grafts: Often a LIMA or SVG will be anastomosed side-to-side to an artery or branch and then the distal graft will be anastomosed end-to-side to a second artery or branch. For the LIMA it is typical to anastomose to a LAD-diagonal branch and terminate at the distal LAD. SVGs are often jumped from a first to a second obtuse marginal or from a right coronary artery (RCA) posterior descending branch to an RCA-posterolateral branch.

• “Snake” or long circular grafts: In a technique that has fallen out of favor but still may be found occasionally, a single long saphenous vein graft is anastomosed to the aorta and then anastomosed side-to-side to the LAD and/or branches then to the circumflex branches and finally anastomosed end-to-side to the RCA.

• Under unusual circumstances (e.g., a third CABG surgery) a surgeon may anastomose an SVG from the descending aorta to a coronary artery (usually the circumflex).

• An in situ RIMA may be anastomosed to the circumflex, right coronary artery, or the right coronary posterior descending branch.

• An in situ gastroepiploic artery may be anastomosed to the right coronary artery. This can be easily cannulated using a Judkins right (JR4) catheter to identify the hepato-splenic trunk, advancing the catheter over the wire into the hepatic artery and then the gastro-epiploic artery.

• Y grafts: Surgeons may anastomose a free radial artery to a LIMA with the radial graft going to a diagonal branch of the left anterior descending and the LIMA ending at the LAD. Rarely surgeons will anastomose a SVG segment to an SVG in a similar fashion.

• Common aortic “hoods” or “buttons”. Occasionally surgeons will anastomose two SVGs to a single spot on the aorta so that both arise from the same point.

1.3 Natural history of bypass grafts

Arterial grafts are more durable than venous grafts. When grafted to the LAD, the LIMA graft has a five-year patency rate of 91%, whereas vein grafts had a five-year patency rate of 78% [3, 4]. In patients who underwent CABG between 1995 and 2010, at a 7-year follow-up the patency of the LIMA was 87%, the patency of a radial artery graft to the RCA or LCx was 82%, and the patency of saphenous vein grafts was 58% [5].

Three processes lead to SVG failure, and the mechanism of failure can be predicted by the timing of failure. A useful rule of thumb is that about 10% of grafts occlude in under 1 month due to thrombosis or surgical issues, about 10% occlude between 1 month and 1 year due to intimal proliferation, in about 2–3% more occlude per year due to accelerated atherosclerosis. Within the first month after CABG, thrombosis (i.e. due to hypercoagulability) and technical failure (i.e. damage to or defects of the graft) are the predominant mechanisms. From the first month to the first year after CABG intimal hyperplasia is the predominant mechanism, a process in which smooth muscle cells proliferate and fibroblasts lay down extracellular matrix (also known as “arterialization” of the graft) in response to exposure to arterial pressures. And beyond the first year of CABG atherosclerosis is the predominant mechanism, a process that is accelerated in SVGs as compared to native arteries and in which unstable plaques often form [6].

1.4 Indications for cardiac catheterization after CABG

The 2012 Appropriate Use Criteria for Diagnostic Catheterization provide indications for cardiac catheterization in patients with prior CABG [7]. Common indications include acute coronary syndromes or electrical instability. Emergent coronary angiography may be indicated for postoperative CABG patients who have clear signs of ischemia, unexplained hemodynamic instability, low cardiac output syndrome, electrical instability, diffuse electrocardiogram changes, new ischemic wall motion abnormalities, or very large troponin elevations after CABG. Troponin elevations of >10x the upper limits of normal qualify as type 5 myocardial infarction (MI) in the Fourth Universal Definition of Myocardial Infarction [6, 8, 9].

For stable patients, the indications are more limited. In general, asymptomatic patients should not undergo catheterization unless there is other evidence of extensive ischemia. Specifically, a small or even moderate-sized ischemic abnormality on stress testing would not warrant catheterization in a patient with no symptoms or atypical symptoms. The indication for catheterization strengthens as symptoms increase despite guideline-directed medical therapy or as the evidence for extensive ischemia increases. Consideration of catheterization in patients after CABG must balance the risks of catheterization (which are about twice those of diagnostic coronary arteriography in non-CABG patients) and the risks of subsequent PCI against the benefits of symptom relief or of diagnosing atypical symptoms. To our knowledge, no study has demonstrated improved survival from repeat PCI or CABG in any sub-group of post-CABG patients.

2. Approach to cardiac catheterization and bypass graft angiography

The approach to cardiac catheterization in a patient with prior CABG is the same as the approach to cardiac catheterization in patients without CABG for a right heart catheterization, left heart catheterization, and native coronary angiography. Graft arterography includes finding and selectively engaging each graft, usually one LIMA graft and one or more grafts arising from the ascending aorta.

2.1 Pre-catheterization preparation

It is critically important for the operator to know the details of the CABG surgery before starting catheterization, in order to plan access. For example, the best access for a patient with LIMA and RIMA grafts, or with the left radial used for CABG, may be femoral access. It is critically important for the operator to review the operative report because this is the only reliable roadmap to finding grafts. Downstream descriptions of the surgery become progressively unreliable. Specifically, the discharge summary is usually written by an advanced practice provider who may misinterpret the operative report, and subsequent summaries by cardiologists or primary care providers are routinely misleading. For example, a LIMA to the LAD with radial Y-graft to the diagonal and an SVG jumping from the second obtuse marginal to the RCA postero-lateral branch will be recorded in subsequent clinic notes as a 4-vessel CABG. But without details, the operator will not know how many anastomoses from the aorta to look for, or whether a graft will be arising from the right side of the aorta as is typical of grafts to the RCA. When the allowable contrast dose is limited by kidney disease it is particularly important to know details of coronary anatomy to prevent excessive test injections while searching for grafts.

When details of the surgery are unavailable, patients are usually reliable sources of the number of distal anastomoses. Usually, when patients are told the results of their surgery by the surgical team, they are told the number of distal anastomoses, which may exceed the number of proximal anastomoses. The wise operator will make sure all distal anastomoses are accounted for before ending a procedure.

2.2 Vascular access

Radial access decreases vascular complications compared to femoral access in patients without prior CABG. The same is true for patients after CABG, but left radial artery access is preferred since it offers easy access to the origin of the LIMA. In patients with the left radial artery harvested for use as a bypass graft, femoral access is usually used although experienced operators can non-selectively (and occasionally selectively) cannulate the LIMA using right radial access. With left radial access, the left arm can be pulled across the abdomen so the operator does not have to reach across the table. The use of the distal radial access site (“snuffbox approach”) can bring the access point even closer to the operator standing on the right side of the table. The RADIAL-CABG randomized trial compared femoral access to left radial access at a single center and demonstrated higher radiation doses, contrast volumes, and longer procedure times with left radial access as compared to femoral access; though radial access was associated with higher patient satisfaction. The crossover rate was higher (17%) in the transradial group compared to the transfemoral group [0%] [10]. A meta-analysis found fewer vascular complications with radial access [11].

2.3 Graft markers

Graft markers are used or not used variably by cardiac surgeons. Common varieties include a small disk usually placed above the aortic anastomosis, a horseshoe or wire ring around the proximal part of the graft, or occasionally just a clip by the aortic anastomosis. Often SVGs or in situ LIMA grafts will have clips where side branch veins were cut; these can lead like breadcrumbs along the course of the graft and give a hint as to the location of its terminus.

2.4 Catheter selection and angiographic views

A typical patient will have a LIMA graft arising from the left subclavian anastomosing distally to the LAD and two or three free grafts, usually SVGs, with anastomoses from the aorta to the target vessel in the LCX system, RCA system, or a diagonal branch of the LAD. Our general approach is described in Table 1.

Graft	LIMA to LAD	SVG/radial to RCA	SVG/radial to LCX	SVG/radial to Diagonal	RIMA to LCX	In-situ GEA
Catheters	1: JR4; 2: IMA; 3: VB1	1: Multi A; 2: AL1/2; 3: BG right	1: JR4; 2: AL2; 3: Multi; 4: BG left	1: JR4; 2: AL2; 3: Multi; 4: BG left	1: JR4; 2: IMA; 3: VB1	JR4 engages through the hepato-splenic artery
View	AP cranial; RAO; Left lateral	LAO; RAO; RAO cranial; LAO cranial	LAO; RAO; AP caudal	LAO cranial LAO; RAO	LAO; RAO	(Depends on anastomosed artery)

Table 1.

An approach to bypass graft angiography.

The LIMA is engaged by finding its ostium in the subclavian artery. It may arise on the more proximal vertical section or on the more distal horizontal section of the subclavian. We use the anterior–posterior view although occasionally the right anterior oblique view will better separate the proximal LIMA from the subclavian. From left radial access, the JR4 catheter is advanced over a wire retrograde in the left subclavian to the LIMA ostium. From femoral access, the JR is advanced retrograde through the transverse aorta. Counter-clockwise rotation allows the operator to place the catheter sequentially in the right innominate, then the left carotid, and finally into the left subclavian. The JR4 catheter can be advanced over a wire distally into the subclavian. From either access point, the JR4 can be gently maneuvered proximally in the subclavian with gentle counter-clockwise rotation and test injections. If the origin of the LIMA is acute the JR4 can be exchanged over a wire for an IMA catheter and maneuvered similarly. For a severely angulated LIMA origin, a VB-1 or similar catheter with a pigtail-like curve can be positioned beyond the ostium and pulled back to engage the LIMA ostium (Table 2).

Name	Type	Study
GuardWire	Distal balloon	SAFER demonstrated improved rates of periprocedural MI and no-reflow as compared to usual therapy
TriActiv	Distal balloon	PRIDE demonstrated noninferiority to the GuardWire and FilterWire
FilterWire*	Distal filter	FIRE demonstrated noninferiorty to the GuardWire
SpideRx*	Distal filter	SPIDER demonstrated noninferiority to the GuardWire and FilterWire
CardioShield	Distal filter	CAPTIVE failed to demonstrate noninferiority to the GuardWire
Proxis	Proximal balloon	PROXIMAL demonstrated noninferiority to the GuardWire and FilterWire

Table 2.

Embolic protection devices.

*

Currently available in the USA.

Source: Lee et al. [12].

Free grafts to the other coronary arteries (i.e. SVGs or radial grafts) are found in the proximal ascending aorta. The grafts are found by selecting a catheter and searching the aorta above the level of the coronary arteries. Right coronary artery grafts will be located on the right side of the aorta whereas left circumflex and diagonal grafts will be located on the left or posterior aspects of the aorta. Generally, grafts are arranged in the following ascending position in the aorta: RCA grafts lowest in the aorta, followed by LAD grafts (if there is SVG to LAD) located a little higher, followed by diagonal branch, then left circumflex first obtuse marginal, second obtuse marginal, and circumflex posterolateral grafts highest in the aorta. We favor multi-purpose shapes (or right bypass graft shape) for grafts to the RCA (which usually have a downward takeoff). Grafts to diagonal branches or circumflex branches may be cannulated with the JR or multi-shaped catheter, but if necessary Amplatz-shaped catheters or left bypass graft catheters can be used. For all of these, we use a clockwise rotation of the catheter with frequent test injections to engage grafts.

On occasion, it can be hard to find all of the grafts. When searching for grafts, start with a specific catheter for the suspected graft as described above. A proximally occluded graft may be demonstrated by test injections showing a short stump in a side view or a circle in an end-on view. Occasionally grafts are flush occluded at the aorta and cannot be identified. For RCA grafts it is important to point the catheter downward in the graft using a slight counter-clockwise torque since injection in the proximal graft orthogonal to its direction can mimic a total occlusion. Consider that a graft may arise from an unusual location on the ascending aorta or even from the descending aorta [13], or that a RIMA or gastroepiploic artery may have been used. When all else fails, non-selective aortography can be performed although it does not reliably demonstrate all patent grafts. The last option for finding a graft would be a CT or MRI angiogram.

It may be helpful to identify native vessels that appear to have been grafted. Occasionally the stump of the graft where it is terminally anastomosed to the vessel may be seen. In other cases where the graft has flush-occluded, a characteristic upward omega-bend of the native vessel caused by scarring/retraction of the graft after surgery may reveal where the graft was anastomosed to the native vessel. Occasionally a segment of a jump graft between two native branches will remain patent even after the graft from the aorta to the first anastomosis has occluded.

3. Bypass graft PCI

PCI in patients who are post-CABG is common. Data published from the NCDR CathPCI registry in 2011 show that PCI in prior CABG patients represents 17.5% of all PCIs. Native arteries were targeted alone in 62.5% of PCI in prior CABG patients, saphenous vein grafts were the target in 34.9%, and arterial grafts were the target in 2.5% [14]. A similar observational analysis from VA medical centers in 2016 showed overall similar data (73.4% of PCI was in a native artery, 25.0% in an SVG, and 1.5% in an arterial graft). The VA analysis demonstrated that procedure-related complications were more frequent in bypass PCI patients compared to those without, including in-hospital mortality, procedural complications, peri-procedural MI, no-reflow, and dissection. The patients who received PCI to graft lesions were also noted to have higher mortality, MI, and revascularization at 1 and 5 years of follow-up [15].

Indications for PCI in post-CABG patients are similar to those without prior CABG. Graft lesions causing acute coronary syndromes may undergo PCI or may be used as conduits for retrograde PCI of the native vessel to which they anastomose. In stable patients, PCI is generally not indicated for asymptomatic patients. The strength of indication for PCI increases as the severity of symptoms despite guideline medical therapy increases.

3.1 Approach to SVG PCI

There are several issues with intervention on SVGs, and as such, the operator must carefully consider their options before embarking on SVG intervention. SVG intervention carries a high risk of distal embolization, no-reflow, and peri-procedural MI. Degenerated vein grafts are noted in both the ACC/AHA and SCAI classification schemes to be high-risk lesions and to have worse outcomes as compared to low-to-intermediate risk native vessel lesions [16]. Several principles affect decisions regarding SVG intervention.

A first principle of vein graft intervention is that PCI in vein grafts is less reliable than PCI of native coronary arteries. Observational data suggest that PCI to SVGs is associated with worse outcomes than PCI to native coronary arteries [15, 17, 18]. For this reason, when reasonable, restoration of blood flow by performing PCI to the native vessel is preferred to PCI of the SVG. Preferencing PCI to the native artery where possible is given a Class 2a recommendation in the updated 2021 ACC/AHA Coronary Artery Revascularization guidelines [19]. It should be noted that this strategy is complicated by the high rate of CTOs in bypassed native arteries, and referral to a physician with experience in complex coronary disease and CTO may be necessary [20, 21]. A strategy of PCI to the SVG followed by staged PCI to the native artery, especially in the setting of acute MI, may be useful [22]. Intentional iatrogenic occlusion of the SVG after native vessel PCI may be beneficial to reduce competitive flow [23, 24].

A second principle is that intermediate lesions should in general be treated medically. Two trials, VELETI and VELETI II studied the utility of stenting intermediate SVG lesions. While there was a trend in the VELETI pilot study towards improved outcomes with stenting, the larger VELETI II study showed no benefit [25, 26, 27]. Additionally, the use of FFR has been studied in intermediate lesions. While there may be benefit to the use of FFR in arterial grafts, no benefit was seen in SVG lesions and should probably not be used in this setting [28].

A third principle is that PCI to CTOs of SVGs is not of benefit and should not be performed. Chronic total occlusions of SVGs were studied in a retrospective study published in 2010 that found success rate of PCI of SVG CTO was 68%. In the successful PCI group, the ISR rate was 68% and TVR rate was 61% with a median follow-up of 18 months [29]. Due to the low success rates and high rate of revascularization, current guidelines give PCI of SVG CTOs a Class 3: No Benefit designation [19].

3.2 Balloons and stents

Bare-metal stenting was clearly an improvement over balloon angioplasty for SVG lesions. The Saphenous Vein in De Novo (SAVED) trial compared bare-metal stents to balloon angioplasty for focal, de-novo SVGs lesions. Stenting increased the procedural success, demonstrating 92% success with BMS versus 69% for angioplasty [30]. This benefit of BMS as compared to balloon angioplasty alone was reinforced with data from the Venestent trial [31].

Several studies have examined the use of bare-metal versus drug-eluting stents in SVG PCI. The RRISC trial initially demonstrated improved outcomes of DES as compared to BMS [32], however the DELAYED RRISC study (a post hoc analysis of the RRISC trial) appeared to support increased mortality of patients treated with DES as compared to BMS [33]. Subsequent randomized controlled trials and meta-analyses have however demonstrated the safety of DES in SVGs [34, 35]. In addition to some smaller trials, two larger RCTs compared DES to BMS: ISAR-CABG and DIVA. While ISAR-CABG did demonstrate lower target lesion revascularization with DES as compared to BMS at 12 months [36], by follow-up at 5 years no difference between DES and BMS was observed [34]. The DIVA trial showed no difference at 12 months between DES and BMS [37]. A meta-analysis of the available RCTs done in 2018 showed no difference between DES and BMS [35]. Of note, in the ISAR-CABG trial, most stents were first-generation, while in the DIVA trial most stents were second-generation indicating that neither first nor second-generation DES stents are an improvement over BMS [35]. Two retrospective studies have found no difference between first- and second-generation DES [38, 39].

Directly stenting SVG lesions (as opposed to performing pre-dilation) might prevent distal embolization. One observational study done in 2003 indicated that direct stenting decreased post-procedural MB-CK elevation, and the one-year composite endpoint of death, Q-wave MI, and target lesion revascularization [40].

Under-sizing stents may improve outcomes in SVG PCI. Hong et al. in 2010 examined a series of patients who underwent SVG PCI with IVUS. They compared patients based on the ratio of stent diameter to vessel diameter and found that patients with relatively under-sized stents had fewer post-procedural CK-MB elevations without worse outcomes at 1 year [41].

3.3 Embolic protection devices

SAFER was a trial in which a distal balloon device called the GuardWire demonstrated a significant decrease in peri-procedural MI and a decrease in no-reflow [42]. The GuardWire is a distal balloon embolic protection device wherein the balloon is inflated distal to the PCI target. The operator then stents the lesion and aspirates the blood containing post-PCI embolic debris out of the vessel before deflating the balloon [42]. The FIRE trial compared a device called the FilterWire, a distal filter-based device, against the GuardWire and showed non-inferiority [43]. Numerous other trials have been investigated (see table below), but all of these trials were in some way compared their device to the GuardWire to show non-inferiority as opposed to a comparison against usual therapy. The TRAP trial would have been a second RCT but was ended due to lack of enrollment and was therefore under-powered; the trend however was of findings consistent with SAFER (decreased peri-procedural MI) [44].

There have been multiple analyses since these trials in the early 2000s looking at EPDs. Iqbal et al. examined the British Columbia Cardiac Registry and showed that patients undergoing SVG PCI had improved post-procedural TIMI flow after EPD use, however had no difference in TVR or mortality at 2 years [45]. Brennan et al. examined the Cath PCI database and showed no difference in rates of death, MI, or TVR with the use of EPDs but did show increased rates of no-reflow, vessel dissection, perforation, and periprocedural MI with the use of EPDs [4]. Paul et al. performed a meta-analysis and review in 2017, which suggested no benefit to EPD use in SVG intervention [46].

The 2011 ACC/AHA guidelines on PCI gave the use of embolic protection devices (EPDs) a Class I recommendation based upon strong randomized control trial evidence from the SAFER trial. However, with the subsequent data described above, current guidelines downgrade the recommendation for use of EPDs from Class I (in 2011) to Class IIa (in 2021) [19, 47]. Despite the data supporting EDP use, estimates of usage rates in SVG lesions based on large registry data range from 14–22% [48, 49]. EPD use may be discouraged by the technical difficulty of using these somewhat bulky devices [49].

In summary, the only randomized trial data available shows the benefit to use of EPD. Multiple other EPDs have shown non-inferiority to the GuardWire. EPDs can be difficult to use which significantly limits their use in clinical practice. And while significant observational data have called into question the findings of the SAFER trial, guideline recommendations are unlikely to change significantly until further RCTs are performed.

3.4 Pharmacology of SVG intervention

In general, antiplatelet drugs are used in the same way post SVG PCI as they would be used post native vessel PCI. The PLATO trial demonstrated the efficacy of ticagrelor over clopidogrel in ACS patients. A post hoc analysis of PLATO showed that ticagrelor was as effective for post-CABG patients as it was for no-CABG patients [50]. In addition, SVG lesions are high-risk lesions and may benefit from more intensive antiplatelet therapy than some native vessel lesions. The DAPT trial showed that in patients who had SVG PCI, there was less stent thrombosis with 30 months of DAPT as compared to 12 months of DAPT [51]. An analysis of the DAPT study developed and validated a prediction rule intended to determine patients who would benefit most from prolonged DAPT. In the generated scoring system, the presence of a vein graft stent was one of the strongest predictors of deriving benefit from prolonged DAPT [24].

The use of GP IIb/IIIa inhibitors does not appear to be of benefit. A meta-analysis of five randomized trials published in 2002 showed that the use of GP IIb/IIIa inhibitors in graft interventions provided no benefit and had an association with worse outcomes [52].

The use of anticoagulants is similar in SVG PCI as in native-vessel PCI. Heparin is the dominant drug used, however, bivalirudin has been shown to be safe and effective [53].

Vasodilator drugs may decrease the rate of no-reflow in SVG PCI. Adenosine, nitroprusside, and the calcium channel blockers verapamil and nicardipine have been investigated. Overall, the quality of the evidence is low however all the studies show some degree of improvement in no-reflow, post-procedural CK-MB elevation, or both in association with the use of vasodilators [54, 55, 56, 57]. Nicardipine is often preferred as it causes less hypotension and a longer duration of action [58].

3.5 Other therapeutic options and techniques for SVG

The CORAL trial examined the use of excimer laser coronary atherectomy before stenting. The study failed to enroll enough patients and so they compared laser atherectomy with a stent to the SAFER data (control and EPD groups). The rate of MACE, driven by peri-procedural MI, was lower in the SAFER GuardWire group [59]. One case–control registry indicated that ELCA showed better angiographic outcomes and lower rates of Type IVa MI as compared to distal embolic protection devices [60].

The VeGAS 2 trial compared the AngioJet rheolytic thrombectomy device to urokinase infusion for SVG thrombus. The AngioJet creates a local vacuum using high-velocity water jets, with the intention of sucking thrombus into the catheter for degradation and removal. AngioJet did show some improvements over urokinase infusion, especially in the rates of procedural success, non-Q-wave MI, and vascular complications [61].

3.6 Arterial graft PCI

Arterial grafts are significantly more durable and significantly fewer in number than venous grafts, and they are therefore significantly less likely to be the targets of PCI. PCI in arterial grafts is generally more successful and with lower complication rates than in PCI of vein grafts [14, 15].

The IMA is the most important arterial graft, and there are a few relevant points regarding PCI in these arteries. The risk of complication is not negligible. The most common cause of unsuccessful PCI in an IMA graft is excessive vessel tortuosity. Straightening a tortuous LIMA can cause pseudolesions which may cause ischemia; this effect must be distinguished from vasospasm (as it will not improve with vasodilators) and dissection. Removal of the guidewire should resolve a pseudolesion [58]. Tortuous subclavian arteries may be an issue as well – ipsilateral (usually meaning left) radial access can help in this case. On occasion, coronary ischemia in the distribution of the IMA can be caused by a stenosis of the subclavian artery proximal to the IMA graft, and PCI of the subclavian artery (by an experienced peripheral operator) can relieve the ischemia [62].

Ostial dissections can occur in IMA PCI and therefore the ostium should be evaluated at the end of an IMA PCI procedure. PCI of distal anastomotic IMA lesions has been shown to have better outcomes (less restenosis) with balloon angioplasty as compared to stenting; stents are typically used in lesions of the ostium and the body of IMA grafts [63, 64].

4. Summary

Indications for catheterization and PCI in post-CABG patients are similar to those for patients without CABG. Graft anatomy (taken from the source CABG operative report) should be known before starting a diagnostic procedure. Diagnostic procedures involving grafts are more difficult, require more time, contrast, and catheters, and produce more complications than procedures in patients without prior CABG. A set of unique “tricks” is required to selectively cannulate all grafts known to be present. PCI of grafts, particularly of SVGs, produces frequent complications and is often followed by restenosis. PCI of the native vessel supplying the grafted territory, either antegrade or retrograde, which may be preferred over graft arteriography. As the incidence of CABG is decreasing over recent decades, the number of post-CABG patients undergoing catheterization is decreasing. However, the ability to perform angiography of post-CABG patients will continue as a required skill of invasive interventional cardiologists.