Distal Embolization in Saphenous Vein Graft Intervention

Although both vessels are composed of three layers - the intima, media and adventitia -the medial layer is much thicker in arteries. Remember that form follows function: the muscular medial portion of the arterial wall must accommodate the high systemic pressures generated by the left ventricle, with an average mean arterial pressure of 80-100 millimeters of mercury (mmHg). The thick medial layer also helps propel blood through the body.

In contrast, veins are thin-walled and designed to operate in a low-pressure system. They’re twenty times more distensible than arteries, and the highest pressure to which they’re exposed is 18-20 mm Hg at the level of the venuole, dropping to 4-5 mm Hg as blood returns to the right atrium.(3) Veins also have valves to prevent the backflow of blood, and they rely on large muscle contraction and the respiratory pump to assist in venous return.

When a vein is attached to the aorta and anastomosed to the native coronary artery, we’ve in effect taken something designed for a low-pressure system and attached it to a high-pressure system. Soon after implantation and exposure to this systemic pressure, SVGs undergo intensive intimal hyperplasia, followed by accelerated and progressive atherosclerosis.

Vein graft atherosclerosis (VGA) is usually not recognized before two to three years after CABG and doesn’t appear to cause measurable graft attrition before five years post-op. The increase in vein graft attrition seen more than five years post-surgery, however, appears to be in large part due to VGA, and the presence of late stenosis in vein grafts is a predictor of adverse clinical outcomes.(4)

There are important differences between SVG vessel disease and native vessel disease (see Fig. 1). Compared to native vessel disease, SVG plaques have a poorly defined cap. They’re bulkier, more diffuse, more friable, and have larger amounts of underlying thrombus (see Fig 2&3).(5) In addition, it’s thought that blood turbulence at the anastomosis site -- where the graft is sutured from the large aorta to a smaller coronary artery -- plays a role in SVG deterioration.(6)

Options for the patient with severe graft disease include reoperation, percutaneous coronary intervention (PCI) of the native artery, and intervention of the SVG itself. Bypass redo, however, carries a three to five times higher risk of mortality (8 to10%) and is often reserved for patients without a LIMA graft and diffuse SVG disease.

Unfortunately, many of these patients don’t have adequate distal targets or acceptable left ventricular (LV) function for bypass redo. They’re often advanced in age and have more co-morbidities than patients undergoing CABG for the first time.(7)

And while the best option in this era of drug eluting stents (DES) may be stenting of the native coronary, for many of these more fragile patients the only acceptable strategy is treatment of the vein graft.

Despite major advances in catheter-based therapy and adjunctive pharmacology, percutaneous revascularization of diseased SVGs remains a critical challenge for the interventional cardiologist and the cath lab team.

While SVG intervention accounts for 10 to 15% of PCIs in most centers, it’s complicated by high procedural, in-hospital and long-term event rates. SVG patients also have a greater atherosclerotic burden compared to native vessel patients.

Vein-graft PCI has been plagued by three problems:

• distal embolization and the resultant phenomenon of slow flow/no reflow.

• progressive graft disease outside the target lesion.

• restenosis, or recurrence of stenosis after ballooning or stenting.

Historically, SVG intervention has been characterized by high rates of no reflow (10 to 15% of procedures), post-procedural MI (17 to 20% of procedures) secondary to embolic debris, and 30-day major adverse cardiac events (MACE) rates of 16.9%.(8) SVG PCI patients also have a ten times higher incidence of in-hospital mortality.(9)

Because of the friability and non-encapsulated nature of SVG lesions, embolization of atherosclerotic debris is a major risk during PCI. It’s estimated that embolization occurs as much as 100% of the time, although angiographic evidence presents only 15 to 20% of the time because of the microscopic nature of much of the embolization. (10)

The effects of atheroembolization on the patient depend on the amount, size and composition of the atheroembolic debris as well as the patient’s left ventricular function and microcirculatory status.

The Slow Flow/No Reflow Phenomenon

Slow flow/no reflow is among the most serious atheroembolic effects. Patients who experience this complication have an increased risk of myocardial infarction (MI), morbidity and mortality during and after the procedure (see Fig. 4&5).

First described by Kloner and Jennings in 1974, slow flow/no reflow is defined as inadequate myocardial perfusion through a given segment of coronary circulation without angiographic evidence of mechanical vessel obstruction.

Hong et al. observed that distal embolization with associated slow flow/no reflow was an independent predictor of late mortality and in-hospital CK-MB elevation (>3-5X normal) in patients undergoing PCI for lesions in degenerated SVGs.(11)

The exact mechanism of slow flow/no reflow is not known but is thought to stem from dysfunction or obstruction of microcirculation at the level of the resistance arterioles. It may occur through spasms of distal microcirculation, platelet clumping and the distal embolization of pieces of friable lipid-rich plaque. The release of chemical mediators that may lead to constriction at the arteriole level has also been proposed as a mechanism of action. (12)

• Figure 1 Figure 1
• Figure 2 Figure 2
• Figure 3 Figure 3
• Figure 4 Figure 4
• Figure 5 Figure 5
•  

Embolic Protection Therapies

In 2002, the Saphenous Vein Graft Angioplasty Free of Emboli Randomized
(SAFER) trial transformed the standard of practice for SVG lesion intervention by demonstrating that PCI performed with embolic protection was associated with a lower incidence of no-reflow, peri-procedural MI and late adverse events. In addition, the SAFER trial showed a favorable cost-benefit profile.

The trial compared SVG intervention using the Medtronic GuardWire® Distal Occlusive Device (see Fig. 6) to intervention using a standard guide wire. The results were impressive. Thirty-day MACE rates fell from 16.9 to 9.6%, and no reflow fell from 8.8 to 3.3%: a nearly 50% reduction in event rate.(13) As a result, embolic protection for SVG intervention is now a Class 1 ACC/AHA/SCAI Practice Guideline and the recommended standard of care for SVG stenting.

Today there are many choices for embolic protection, with strategies falling into four broad categories:

• Distal occlusion

• Distal filters

• Proximal occlusion

• Pharmacologic strategies

Distal Occlusion

The strategy behind distal occlusion is to block the vessel being treated several centimeters beyond the target lesion so that plaque liberated from the lesion during angioplasty or stent placement remains suspended in the resulting stagnant column of blood. If that column of blood and the debris it contains can be aspirated completely before distal occlusion is relieved and antegrade flow is restored, distal embolization of debris will be prevented. (14)

Mechanism of operation The distal occlusion device consists of a 0.014-inch hypotube on which an inflatable occlusion balloon is mounted. The balloon tip is passed across the lesion in its deflated state before angioplasty or stenting takes place, and the hypotube shaft is used as the interventional guidewire throughout the procedure.

The distal balloon is inflated before intervention. It remains inflated until a distal aspiration catheter (PercuSurge Export, Medtronic) has been used to aspirate or lavage the stagnant blood column and its suspended debris (see Fig 7). At this point the occlusion balloon is deflated, flow is restored and the protective device is removed.

Occlusion time should not exceed seven minutes, and patients may complain of chest discomfort during occlusion time. Distal occlusion devices can be technically challenging, requiring interventionalists and cath lab teams to be proficient with the systems for optimum results.

Benefits of distal occlusion include the following:

• Distal embolization of debris can be prevented if the stagnant column of blood and the debris it contains can be aspirated before flow is restored. (15)

• Both small and large particles as well as soluble mediators can be trapped more completely than is the case with filters, which may allow some smaller particles and soluble mediators (thromboxane) to pass through.

Limitations include:

• Several minutes of end-organ ischemia occur because of occlusion throughout the intervention.

• There is limited contrast opacification of the target lesion during occlusion, which may make visualization for stent placement difficult.

• Systemic embolization of very proximal lesions is possible.

• Aspiration may fail to recover debris in stagnant pools near the fornices of the occlusion balloon or in the loosely adherent boundary layer near the stent surface.

• Interventionalists aren’t always able to tailor guidewire choice to other procedural requirements.

Device options for distal occlusion include:

• Medtronic AVE GuardWire Plus Temporary Occlusion and Aspiration System (Medtronic, Minneapolis, MN)

The Medtronic GuardWire System was the first distal protection system commercially available in the US and the device used in the SAFER trial.

Distal Filters

Distal filters operate on the principle that a deployed filter can allow ongoing perfusion and yet trap some, if not all, particulate debris. Continual flow can be more comfortable for the patient, and the physician has better visualization of the treatment area.

Mechanism of operation: Distal filters are advanced across the target lesion in their smaller collapsed state. The retaining sheath (the delivery catheter) is then withdrawn, allowing the filter to open and expand against the vessel wall. The filter remains in place to catch any liberated embolic material larger than the filter pore size (usually 100-120 microns) during intervention.

At the end of the intervention, the filter is collapsed using a sheath (the recovery catheter) and the captured embolic material is removed from the body.

Benefits of distal filters include the following:

• Athough distal filters might be expected to retrieve only larger particles, analysis shows nearly identical particle size distribution and aggregate volume of debris retrieved with either a distal filter (100 micron pore size) or a distal occlusion/aspiration system. A filter’s ability to trap particles far smaller than its nominal pore size may stem from the tendency of particles to clump or strand across filter pores, reducing the functional pore size.

• Even if distal filters do allow smaller debris particles to pass through, experimental data by Mauri et al. suggest that embolic particles <100 microns are tolerated in far larger number than are larger particles before interfering with microcirculatory function.
  Smaller particles are also less likely to cause end-organ damage.(16)

• Distal filters are technically less complex than distal occlusion devices and therefore easier to use.

Limitations include:

• Some patients requiring SVG stenting are not candidates for distal protection because of lesion location. In fact very distal lesions may preclude the use of any kind of distal protection device.

• Embolic debris may be released during initial lesion crossing with bulky delivery catheters without protection in place.

• Distal filters may become occluded before the intervention is complete.

• Filters do not prevent the transit of soluble mediators into the myocardium that may lead to slow flow/no reflow.

Device options Two distal embolic filter devices are currently approved by the US FDA for use in SVGs;

• FilterWire EZ® System (Boston Scientific, Natick, MA)

• Spider™ Embolic Protection Device (eV3, Plymouth, MN)

Both devices consist of a delivery catheter, the filter device, and a retrieval catheter. Both allow contrast imaging during the procedure as distal perfusion is maintained.

FilterWire EZ® Embolic Protection System

The FilterWire EZ System is available in two sizes: one for vessel diameters of 2.25 to 3.5 mm, and another for vessel diameters of 3.5 to 5.5 mm. (See Fig. 8)

The 0.014 in. FilterWire device serves as a steerable interventional wire for the procedure and is available in lengths of 190 and 300 cm. The filter, which looks like a butterfly net, is a perforated polyurethane sheet (110 microns) attached to a nitinol loop. The nitinol conforms to the vessel shape and keeps the filter in place. The distal 3 cm of the wire allow the operator to shape the tip to aid in crossing the lesion.(17)

Here are some hints for working with the Filterwire device:

Think black and white TV: the filter wire comes preloaded in a delivery catheter with a black tip. The recovery catheter has a white tip. A tear-away introducer sheath and torquer are also included in the kit.
Place the introducer sheath in the hemostatic valve and advance the Filterwire device as a unit, using the torquer to steer the tip past the lesion. Deploy the filter basket and remove the delivery catheter.
The basket should remain stationary and not drag along the inside of the vessel, which is counterproductive in preventing distal embolization. Because the filter basket wire is the interventional wire, any movement of the guidewire moves the basket up and down. Filter basket position and distal flow should be monitored frequently.
As the basket catches and fills with debris, distal perfusion may be diminished and the basket should be captured in the white tipped recovery sheath and removed.
The FilterWire device is single-use only: do not attempt to rinse the basket and replace it back in the vessel.

Spider™ Distal Embolic Protection System

The SpiderFX™ Embolic Protection Device consists of a capture wire and a dual-ended SpiderFX catheter. The device allows use of any 0.014 to 0.018 in. guidewire for initial lesion cross and delivery catheter placement. (See Fig. 9)

The capture wire is a nitinol braid filter mounted on a 190 cm or convertible 320/190 cm PTFE-coated 0.014 in. stainless steel wire. The capture wire is available in a broad range of filter diameters from 3 mm to 7 mm. It acts as the primary guidewire for other interventional devices compatible with a 0.014 in. wire. (18)

The dual-ended SpiderFX catheter is used to exchange the primary access guidewire with the capture wire, deploy the capture wire at the desired location, and recover it at the end of the procedure.
The delivery end of the catheter is green and has a crossing profile of 3.2 Fr. The recovery end is blue and has a crossing profile of 4.2 Fr. All components of the device are delivered in a single hoop, with the capture wire/nitinol braid filter pre-loaded in the delivery end of the catheter.

A note of caution about distal protection strategies: It’s important to note that both distal occlusion and distal filter protection strategies have limitations based on SVG lesion location. Some patients requiring SVG stenting are not suited to distal protection devices because of lesion location, and very distal lesions may preclude the use of any kind of distal protection device.

In addition, embolic debris may be released during initial lesion crossing with bulky delivery catheters without protection in place. Distal occlusion devices require several minutes of ischemia and must be placed and retrieved within seven minutes. Distal filters may become occluded before the intervention is complete, and filters do not prevent the transit of soluble mediators into the myocardium that may lead to slow flow/no reflow.

• Figure 6 Figure 6
• Figure 7 Figure 7
• Figure 8 Figure 8
• Figure 9 Figure 9
•  

Proximal occlusion

Proximal occlusion addresses the limitations of distal protection, in particular for the 40% of patients who are not candidates for distal protection devices by virtue of lesion location. Currently the Proxis Embolic Protection System (St. Jude Medical, Minneapolis, MN) is the only FDA-approved proximal occlusion system for use in SVG.

Mechanism of operation The Proxis system is a single-operator catheter that is deployed proximal to the target lesion before crossing. Inflation of the sealing balloon interrupts antegrade flow during the period of lesion intervention. Stagnated blood and emboli liberated during intervention are retrieved by gentle aspiration or ancillary flushing of the vessel.

The Proxis device is tracked through the guide catheter and into the target vessel proximal to the treatment area. The guidewire and interventional device are inserted through the Proxis device and may be staged proximal to the treatment area before balloon inflation.

Balloon inflation suspends blood flow, ensuring stagnation of blood and liberated embolic material during treatment of the lesion. During protection, a static column of contrast verifies adequate sealing and highlights the treatment area to facilitate interventional device placement.(19)

Benefits of proximal occlusion include:

• There is potentially complete recovery of particles of all sizes and humoral substances.

• Protection can be established before any device is passed across the target lesion.

• Very distal lesions can be treated, so vessels with distal lesions that are not candidates for distal protection devices can be protected.

Limitations include:

• The process depends on adequate collaterals for perfusion during occlusion and aspiration of suspended debris, although a distal infusion catheter can be used if spontaneous aspiration is inadequate.

• The internal working diameter of the short sheath is smaller, which may limit applicability in some complex lesions.

Device options As noted above, the most-studied and only FDA-approved proximal occlusion system is the Proxis Embolic Protection System (St. Jude Medical, Minneapolis, MN). It received FDA approval for use in SVG in 2007 based on results of the PROXIMAL (Proximal Protection During Saphenous Vein Graft Intervention Using the Proxis Embolic Protection System) Trial.(20)

Pharmacologic embolic protection

Pharmacologic embolic protection is the final strategy to consider. The important role of microvascular vasoconstriction has been demonstrated by the successful reversal of no-reflow events using a variety of microvasodilators (verapamil, diltiazem, adenosine, nitroprusside, nicardipine (Cardene®)).

In “Pharmacologic” Distal Protection Using Prophylactic, Intragraft Nicardipine to Prevent No-Reflow and Non-Q-Wave Myocardial Infarction during Elective Saphenous Vein Graft Intervention, Fischell et al demonstrated intragraft nicardipine administration without distal protection to be successful in preventing no-reflow.(21)

Nicardipine is a highly potent arterial vasodilator with attractive properties as an agent to prevent or reverse no-reflow. It has a longer duration of action than verapamil or diltiazem in intracoronary (IC) administration. It has relatively greater coronary vasoselectivity and greater microcirculatory vasodilating activity, and it is associated with minimal myocardial depression or atrioventricular (A-V) nodal disruption.

In the study, patients were medicated with 70 U/kg of intravenous unfractioned heparin (UFH) and weight-based eptifibatide. Activated clotting time (ACT) was monitored for a goal of 200 to 250 sec. Nicardipine IV (25mg in 10 ml vial) was mixed in normal saline to achieve a final concentration of 10 micrograms/ml. This is achieved by taking 2.5 mg (1cc) mixed in 250cc of normal saline.

Just prior to direct stenting of the SVG lesion, 300 micrograms of nicardipine were injected via the guiding catheter. Immediately after injection, a Cypher® drug eluting stent was advanced across the lesion and the balloon inflated for two minutes. Following pre-medication with a second dose of 300 micrograms of IC nicardipine, the stent was post-dilated. Final angiography demonstrated an excellent result with brisk antegrade flow.(22)

Average graft age was 11.9±6.6 years with thrombus in 26 of 83 vessels (31%). The primary adverse endpoint of total CPK >3X the upper limit of normal (ULN) or CK-MB> 3X the ULN was seen in 1 of 68 (1.5%) and 3 of 68 (4.4%) patients.

No-flow was observed transiently in two of 83 SVG interventions (2.4%). No patient had a Q-wave MI. In-hospital MACE (death, MI, repeat target vessel revascularization) were observed in only three of 68 patients. There were no additional MACE events (0/68) from hospital discharge to 30 days.(23)

Study results showed that prophylactic vasodilatation with intragraft nicardipine followed by direct stenting appeared to be a safe and cost-effective ($89/vial) means of performing elective SVG revascularization. This approach may provide a simple, time-and-cost effective alternative or adjunct to mechanical distal proximal protection for elective SVG intervention.

Important Take-Home Points

• Saphenous vein graft intervention remains a high-risk procedure. Even with embolic protection devices, it carries the highest risk of complications, both peri-procedural and 30-day MACE, with no reflow as an independent predictor of adverse long-term outcomes.

• It’s absolutely critical to be familiar with the drugs and dosages used to treat slow flow/no reflow events.

• All SVG lesions benefit from embolic protection. There’s no way to predict embolization: it’s present in all lesion lengths, with or without angiographic identifiable thrombus present, and with or without iib/iiia use or direct stenting.

• Never go into an SVG intervention without extensive preparation, even when a lesion looks “simple” and without diffuse disease. According to Baim, the greatest treatment effect (reduction in thirty-day MACE) post-intervention was observed in lesions with a stenosis of 25% or less.24 A “hit and run” approach can have devastating consequences for the patient during the procedure and far-reaching consequences in terms of later mortality.

• The what, when and where of embolic protection are multi-factorial. Vessel and lesion characteristics, location and thrombus burden are key to selecting the most appropriate device.

• Even with the technologies currently in place, SVG intervention remains a high-risk, challenging procedure for interventionalists and cath lab
  personnel.

• Intra Coronary Medications & Dosages Intra Coronary Medications & Dosages
• Comparison of SVG Embolic Protection Strategies Comparison of SVG Embolic Protection Strategies
•  

SVG Intervention Practice Pointers

Case Prep

• Know your equipment. Make sure you’re competent and confident in your ability to prepare, deliver, recover and trouble-shoot whatever device is in use in your institution.

• Know the technological advantages and disadvantages for different lesion types. A very distal lesion, for example, is not a candidate for distal protection in any form, occlusion or filter. This is the time to use the Proxis device or -- if it’s not available -- pharmacologic nicardipine. If the patient has one remaining graft that is supplying the entire myocardium, any form of occlusion device, proximal or distal, is not appropriate: the patient will rapidly become ischemic.

• When using filter devices during intervention, monitor flow by injecting contrast periodically. Remember that as the filter basket fills with liberated debris, flow can be diminished or absent, and the filter basket must be captured and removed.

• Always check distal flow after basket removal.

• Once in place, filter wires are the interventional wires and should be pinned in place securely during device exchanges.

• If the filter moves proximally and must be repositioned, the recovery catheter must be used to collapse the filter and reposition to the distal location. Never drag the open basket back into position, and never attempt to reuse the delivery catheter for that purpose.

Intervention with occlusion devices

• Know how to use the equipment. This is vital. “I think I know” isn’t sufficient: you must know. Occlusion devices are technically challenging, and interventionalists need a team that is confident in device operation.

• Once the occlusion balloon goes up, the clock is ticking: you have a maximum of seven minutes to perform PCI, aspirate debris and deflate the catheter.

• Monitor patient hemodynamics.

• Tell the patient to let you know immediately if s/he begins to experience chest discomfort.

• Keep an eye on the EKG ST segment: it may begin to elevate.

• Reassure and comfort the patient throughout the procedure.

Post-intervention

• Ensure stent apposition before retrieving the protection device.

• Use slow, methodical movements to deliver, deploy and retrieve embolic protection devices.

• Know the drugs, concentrations and formulas to prepare when slow flow/no reflow occurs.

• Know the pathophysiology of the no-reflow phenomenon. If you do, you’ll know not to offer the physician IC nitroglycerin (NTG) in slow flow/no flow, as it has little to no effect on microcirculation, although it’s the first choice in epicardial vessel spasm. Calcium channel blockers and adenosine are the drugs of choice for treating the slow flow/no reflow phenomenon.

• Expect the best results but prepare for the worst.

References

1. Hermiller, James MD. Commentary: Contemporary Management of Vein Graft Disease. Journal of Invasive Cardiology 2005; ISSN: 1042-3931: 17 (8): 399-400.

2. Fuster V, Alexander RW. Hurst’s The Heart 11th Ed. 2004; Ch 58: 1496.

3. Martini F. Fundamental of Anatomy & Physiology 7th Ed. Ch 20: 695.

4. Braunwald’s Heart Disease: A Textbook of Cardiovascular Medicine. 8th Ed. 2008: 492.

5. Schoenhagen P, Ziada K et al. Arterial remodeling and coronary artery disease: the concept of “dilated” versus “obstructive” coronary atherosclerosis. Jnl Am Coll Cardiol 2001; 38: 297-306.

6. www.medtronic.com Treatment: Distal Protection Systems. Why do saphenous vein grafts in the heart deteriorate? Are the forces the same as those causing deterioration in veins?

7. Fuster V, Alexander RW. Hurst’s The Heart. 11th Ed. 2004; Ch. 58: 1497.

8. de Feyter PJ. Percutaneous treatment of saphenous vein graft obstructions: A continuing obstinate problem. Circulation 2003; 107: 2284-2286.

9. Boston Scientific Corporation 2007. PowerPoint: Saphenous vein graft intervention.

10. Sugita J. CardioVascular Center, Frankfurt, Germany. PowerPoint: Embolic protection devices: what is available?

11. Hong MK, Popma JJ, Pichard AD et al. The clinical significance of distal embolization after transluminal extraction atherectomy in diffusely diseased saphenous vein grafts. Am Heart J 1994; 127: 1296-1303.

12. Salinas G, MD. UTMB Galveston, Department of Internal Medicine, Division of Cardiology. PowerPoint: No reflow phenomenon.

13. Baim DS, Wahr D, George B et al. Saphenous Vein Graft Angioplasty Free of Emboli Randomized (SAFER) Trial Investigators. Randomized trial of distal embolic protection device during percutaneous intervention of saphenous vein aorto-coronary bypass grafts. Circulation 2002; 105: 1285-1290.

14. www.medtronic.com Product Information: distal occlusion, GuardWire.

15. Carrozza JP, Mumma M, Breall J et al. Randomized evaluation of the TriActiv balloon-protection flush and extraction system for the treatment of saphenous vein graft disease. J Am Coll Cardio 2005; 46: 1677-1683.

16. Mauri L, Rogers C, Baim DS. Devices for distal protection during percutaneous coronary revascularization. Circulation 2006; 113: 2651-2656.

17. www.bostonscientific.com FilterWire EZ™ Embolic Protection System.

18. Curra F, Albertal M, O’Neill W et al. Distal protection during primary angioplasty: A feasibility and safety study utilizing a novel filter technology. J Invasive Cardiol 2006; 18(10): 442-445.

19. www.sjm.com Products: Proxis.

20. Mauri L, Cox D, Hermiller J et al. The PROXIMAL trial: Proximal protection during saphenous vein graft intervention using the Proxis embolic protection system. J Am Coll Cardiol 2007; 50: 1442-1449.

21. Fischell TA, Subraya RG, Ashraf K et al. “Pharmacologic” distal protection using prophylactic, intragraft nicardipine to prevent no-reflow and non-Q-wave myocardial infarction during elective saphenous vein graft intervention. J Invasive Cardiol 2007; 19: 58-62.

22. Fischell T, Haller S, Ashraf K. Intragraft nicardipine prophylaxis to prevent no-reflow in triple vessel saphenous vein graft intervention. J Invasive Cardiol 2005; 17 (6): 334-337.
    ..Ibid., 21.

23. Baim DS, Wahr D, George B et al. Saphenous Vein Graft Angioplasty Free of Emboli Randomized (SAFER) Trial Investigators. Randomized trial of distal embolic protection device during percutaneous intervention of saphenous vein aorto-coronary bypass grafts. Circulation 2002; 105: 1285-1290.