Vascular Interventions

Fundamentals

Abnormalities of the blood vessel lumen, whether resulting in a lumen that is too small (stenosis) or too big (aneurysm), are readily treated with percutaneous techniques. We will first focus on stenotic or occluded blood vessels. A variety of strategies have been devised to restore patency of both arteries and veins, but all have certain elements in common ( Box 4-1 ). This section focuses on these fundamentals, with specific interventions discussed in subsequent sections. Interventions should be performed only after a complete diagnostic evaluation in appropriately symptomatic patients. The lesion responsible for the symptoms should be identified and characterized, and both the inflow and outflow evaluated. Lesion morphology is very important in guiding intervention and predicting outcome ( Fig. 4-2 ). For example, short concentric noncalcified atherosclerotic lesions usually respond well to simple angioplasty, whereas long, calcified occlusions may be impossible to cross or dilate. Careful review of the appearance of the lesion, sometimes in several obliquities, may be necessary.

Once the decision to intervene is made, a vascular sheath should be inserted. If a large sheath will be required, consider partially deploying a suture-mediated closure device before the sheath is inserted (see Box 2-10 ). Medications commonly used during interventions should be prepared in advance ( Table 4-2 ). The patient should be anticoagulated with unfractionated heparin or bivalirudin just before or immediately after crossing the lesion.

Crossing the lesion is essential to completing a revascularization procedure. Stenoses are crossed with an atraumatic torqueable selective guidewire, often directed by a selective catheter ( Box 4-2 ). The use of digital “roadmap” fluoroscopy is extremely helpful, especially for complex and long lesions. Once the guidewire is across the lesion, the catheter is advanced over the guidewire, and the guidewire removed. Aspiration of blood followed by gentle injection of contrast confirms that the lumen has been reached on the other side. Static flow indicates either occlusion of the stenosis by the catheter or subintimal injection ( Fig. 4-3 ). Distinguishing between the two is essential; filling of normal caliber lumen and branches indicates successful crossing of the lesion, whereas focal pooling of contrast or tracking around a central unopacified lumen indicates a subintimal location. A pressure gradient can be measured at this time if necessary.

Crossing an occlusion is more difficult than crossing a stenosis. The longer and more calcified the occlusion, the less probable the successful negotiation from one side to the other. There are two basic approaches: intraluminal and intentional subintimal. With intraluminal revascularization, the obliterated vascular lumen is recanalized. The procedure begins with injection of contrast close to the site of occlusion. This may unveil a tapered lumen that will direct the wire through the occlusion. Often, a smooth fibrotic cap adjacent to the last collateral around the lesion is present. To gain access to the occluded lumen, an angled or straight catheter is advanced up to the cap and a guidewire is then used to probe for a “soft spot.” The choice of initial guidewire varies with the operator, the vessel, and the anatomy, but often a straight hydrophilic guidewire stays within the lumen once it gets through the cap. When the initial attempts to cross the occlusion fail, different shaped catheters and guidewires with gradually increased stiffness are used. Mechanical devices that burrow, vibrate, drill, or melt their way through the obliterated lumen are available ( Fig. 4-4 ). If all else fails, the back end of a hydrophilic guidewire can be used to puncture the cap, but this risks puncture of the artery wall as well.

Subintimal revascularization is an intentional dissection in the media around the occluded lumen and intimal disease. The proposed benefits of this approach are the absence of atheromatous disease in the media and the ability to recanalize extremely long distances, such as groin to ankle. An angled catheter is used to direct a guidewire (usually angled hydrophilic) into the subintimal space above the obstruction ( Fig. 4-5 ). A loop of guidewire is formed in the subintimal space and used to dissect around the lesion. The guidewire spontaneously reenters the true lumen distally at the interface of the plaque and normal intima in approximately two thirds of the procedures. This may occur well distal to the actual obstruction. When spontaneous reentry does not occur, specialized devices can puncture back into the lumen ( Fig. 4-6 ). At centers experienced with the technique, failure to successfully reenter the true lumen is reported in about 10% of cases.

After confirming successful navigation across the lesion by aspirating blood or injecting contrast, a working guidewire is advanced through the catheter to provide stability for the intervention. In general, relatively stiff guidewires are used that allow devices to be advanced into position without losing access across the lesion. These are available in a range of lengths and diameters, including as small as 0.010-inch diameter. Hydrophilic coated guidewires are usually not a good choice for working wires, because they easily slip out of place during catheter exchanges.

Accurate device sizing (e.g., for balloons, stents, or stent-grafts) requires consideration of several factors ( Box 4-3 ). Oversizing of device diameters by 5%-10% greater than the diameter of the normal lumen is the general rule. Very tight or calcified lesions may require progressive dilation with sequentially larger balloons. The final desired diameter of a blood vessel is determined from an adjacent normal segment of vessel, the same vessel on the other side of the body in the case of bilateral structures or the known average size of the vessel (i.e., “rule of thumb” technique) ( Fig. 4-7 ). When measuring directly from images, it is helpful to calibrate to a known internal standard such as a marker catheter or guidewire, but this method is not always accurate. Many digital units include electronic measuring software. Intravascular ultrasound can be used to obtain direct measurements. The vessel and the type of lesion also impact sizing, in that veins are usually more compliant, whereas heavily calcified arteries may fracture when overdilated.

The length of the device should be sufficient to treat the diseased area, with minimal trauma to adjacent normal or slightly diseased vessels. When the area of disease to be treated extends up to or across a bifurcation into a smaller diameter vessel, the device should be sized or delivered in a manner that avoids trauma to the smaller or normal vessel. (See Balloon Angioplasty.)

The optimal clinical outcome of revascularization procedures is durable improvement of the patient’s symptoms. Because the clinical outcome cannot always be measured during a procedure, most interventionalists use technical endpoints such as a residual luminal stenosis of less than 20% or reduction of the pressure gradient across the lesion to a predetermined level. Pressure gradients across lesions are a very useful means of deciding when to perform and when to stop a procedure. Reliance upon angiographic appearance as the sole indication for intervention leads to treating the image not the patient. Pressures are measured intravascularly, both proximal and distal to the lesion ( Table 4-3 ). The most accurate systems obtain simultaneous pressures using two end-hole catheters or an end-hole catheter and a sheath ( Fig. 4-8 ). A quick approach is to withdraw a catheter through the lesion while continuously recording pressures, but the patient’s baseline pressure could change during this process. In many instances a catheter must be through the lesion of interest in order to obtain a distal pressure, such as in the renal artery. If the lesion is tight, the catheter itself accentuates the severity of the gradient by partially obstructing the lumen. Specialized pressure-sensing guidewires may be useful in these cases, but in general most interventionalists believe that any symptomatic lesion tight enough to be partially obstructed by an angiographic catheter requires treatment. When in doubt, injection of 200-300 μg of nitroglycerin or another vasodilator through a catheter distal to the lesion may induce hyperemia and unmask a gradient. Technical success and clinical success do not always coincide; pushing an intervention to an extreme to obtain an ideal image or perfect gradient risks harming the patient (see Fig. 4-8 ).

The guidewire should remain across the lesion until success has been documented. When appropriate, the status of the vessels distal to or proximal to the intervention should be reevaluated as well. Clinical follow-up tailored to the intervention should be performed by the interventional radiologist.

Procedural complications associated with opening of blood vessels are related to the patient’s general status, the difficulty of the procedure, the size and type of the device, the underlying condition of the vessels, and the intensity of anticoagulation ( Box 4-4 ). Older patients with acute illnesses, diffuse vascular disease, and concurrent major illnesses are most likely to experience a complication. Large complex devices have more complications than small simple devices.

Failure (i.e., return of symptoms) of revascularization procedures occurs for different reasons at different times ( Table 4-4 ). Most early failures are due to technical issues such as an occlusive dissection adjacent to the intervention site that impedes flow, elastic recoil of a fibrotic lesion, or perhaps a missed lesion that continues to impair flow. Hypercoagulable syndromes, episodes of hypotension, or other low-flow states can result in acute thrombosis at arterial or venous intervention sites at any time. After approximately 3 months, and up to 1 year from the intervention, intimal hyperplasia is the main cause of failure. The degree of intimal hyperplasia that occurs after an intervention is dependent upon the biology of the vessel and the extent of the trauma to the endothelium (see Fig. 1-8 ). In general, the more extensive the area that is treated, the higher the likelihood that intimal hyperplasia will be a problem. Restenosis occurs in or adjacent to the original lesion. Drug-eluting balloons, drug-eluting stents, and stent-grafts show promise for improved long-term clinical success rates in various vascular applications. There is evidence suggesting that antiplatelet and statin therapies may reduce restenosis rates following peripheral arterial interventions. After 1 year, failure of a revascularization procedure is more likely to be due to progression of the original disease in the inflow or outflow vessels. Factors such as smoking, diet, hyperlipidemia, and homocysteinemia contribute to late failure of arterial interventions.

Balloon Angioplasty

The primary mechanism of arterial balloon angioplasty is controlled fracture of the obstructing plaque ( Fig. 4-9 ). This results in formation of fissures in the plaque itself, and tearing of the edges of the plaque away from the adjacent normal intima. With proper oversizing of the balloon, the muscular media is stretched as well. Plaque is not remodeled, redistributed, or vaporized by the balloon. Distal embolization of microscopic and, occasionally, macroscopic debris does occur but is usually asymptomatic. Visualization of “cracks” or small dissections in lesions following angioplasty is a normal finding at angiography ( Fig. 4-10 ). Over time these areas may remodel and the lumen resume a more normal appearance. Venous lesions, which are usually fibrotic, are primarily stretched during angioplasty.

Andreas Gruentzig performed the first successful balloon angioplasty in 1977. Since then, the technology of balloons has become very complex, but in practical terms it can be divided into compliant and noncompliant devices ( Fig. 4-11 ). Compliant balloons are constructed from material that continues to stretch as pressure is applied, allowing the balloon to expand until the point of rupture. Compliant balloons conform to the vessel walls rather than dilate. This type of balloon is most commonly used to temporarily occlude flow or sweep away thrombus during a balloon thrombectomy.

Noncompliant balloons reach a nominal predetermined diameter during inflation and remain close to that diameter as pressure is increased to the bursting point (many of these balloons actually increase slightly in diameter as pressure increases). During inflation, a waist is visualized in the balloon at the site of maximal stenosis ( Fig. 4-12 ). Continued inflation of the balloon ultimately obliterates the waist, often suddenly. Noncompliant balloons are desirable for angioplasty; otherwise the balloon expands on either side of the waist without dilating the stenosis. Very high pressure noncompliant balloons (up to 30 atm) are used in venous and nonvascular applications, whereas lower pressure balloons are used in arterial lesions. However, some lesions cannot be dilated no matter how much pressure is applied. In this situation, balloon rupture may occur. Burst pressures are included on the packaging of all balloons.

The balloon material and construction determine the burst pressure. A balloon that ruptures before the lesion is fully dilated is of little value, but a balloon that is so strong that the vessel ruptures first is potentially dangerous. Balloons are designed to split longitudinally to minimize damage to the vessel wall and facilitate removal ( Fig. 4-13 ).

Balloon diameters range from a millimeters to more than 4 cm, and lengths range from less than 1 cm to more than 20 cm. The vascular bed and lesion length influence the balloon choice ( Table 4-5 ). When in doubt, undersize. An initially undersized balloon rarely causes a complication, and a larger balloon can always be used if the result is unsatisfactory. The reverse is not true for an initially oversized balloon. Because of the physics of balloons (tension = pressure × radius), the larger the balloon, the lower the necessary inflation pressure.

The balloon material is not radiopaque. To aid in positioning of the balloon, metallic rings are placed on the catheter at both ends of the balloon (see Fig. 4-12 , Box 4-5 ). Dilute contrast (1 part contrast to 2-3 parts flush solution) is used for inflation to allow visualization of the balloon–lesion interaction and facilitate rapid deflation. Centering the balloon on the lesion maximizes stability during inflation and force transmitted to the lesion. During inflation, the ends of the balloon inflate first, so-called “dog-boning,” followed by “popping” the waist (see Fig. 4-12 ). Soft lesions dilate with relatively little pressure. Calcified or fibrotic lesions may require substantial pressure and may never fully dilate. Inflation should be gradual, in that rapid, explosive dilation results in greater trauma to the adjacent normal vessel. In the hands of most humans, a 10-mL Luer-lock syringe can maximally generate 10-12 atm, which is sufficient to dilate the majority of arterial lesions. A mechanical inflation device or a smaller syringe (5 or 3 mL) can be used to generate the higher pressures sometimes needed for fibrotic or venous lesions. Balloons are kept inflated for 20-120 seconds depending on the lesion, the vessel, and the degree of anticoagulation. The patient should be asked about discomfort during inflation. Total lack of sensation on the patient’s part implies an undersized balloon (provided there is normal vascular innervation). Mild pressure or pain indicates stretching of the adventitia and a properly oversized balloon. Intense, severe, or sharp pain suggests overdilation with possible dissection or rupture of the blood vessel. To minimize trauma, balloons should not be moved back and forth through the lesion while inflated. Short balloons have a tendency squirt out of a lesion during inflation (similar to pinching a wet watermelon seed). This can be prevented by stabilizing the balloon shaft at the diaphragm of the sheath and using a longer balloon. Deflation of the balloon with a 20-mL syringe is quicker and more complete than with a 10-mL syringe. Rapid deflation is desirable to restore flow as quickly as possible.

Angioplasty balloons are mounted on angiographic catheters, usually with two lumens: one for a guidewire and one for balloon inflation/deflation. The back end of the standard over-the-wire balloon catheter usually has two separate Luer-lock hubs, one for the central guidewire lumen and one for balloon inflation. Balloons mounted on 3- or 4-French catheters have the smallest overall insertion profiles, require 0.018 inch or smaller guidewires, and can be inserted through 4- or 5-French sheaths. Rapid exchange balloon catheters have two lumens in the distal portion of the catheter, with the shorter guidewire lumen exiting the side of the catheter ( Fig. 4-14 ). The balloon lumen continues for the entire length of the catheter, ending in single Luer hub. This allows for lower catheter profiles and insertion of a long catheter over a standard length guidewire, because the majority of the catheter slides alongside the guidewire, not over it.

When removing a balloon through a sheath, resistance may be encountered because the deflated balloon does not return to its original low profile. Continued aspiration and counterclockwise rotation as the catheter is withdrawn into the sheath rewraps the balloon around the shaft and facilitates removal.

Several enhanced angioplasty balloons are available. Cutting and sculpting balloons use small longitudinal blades on the balloon surface or a metallic cage on the outside of the balloon to create controlled splitting rather than random plaque fractures ( Fig. 4-15 ). This may enable successful angioplasty of tough, fibrotic lesions. Cryoplasty, in which liquid nitrogen is used to inflate a specially designed balloon, results in apoptosis of smooth muscle cells in the media. Angioplasty balloons are increasingly used to deliver drugs or gene therapy vectors directly to lesions.

Angioplasty of lesions that involve or are close to a vascular bifurcation can be challenging because of the size discrepancy of the lumens proximal and distal to the branch point. Simultaneous inflation of two balloons across the branch point, each sized to the diameter of the vessel distal to the branch, prevents complications at the vessel origins ( Fig. 4-16 ). This is known as the “kissing balloon” technique and is also used with stent deployment. In some cases the lesion in the branch is close to, but does not involve, the vessel origin. Rather than insert a second balloon, a “safety wire” in the uninvolved vessel can be used to preserve access in case a complication (e.g., origin dissection) occurs ( Fig. 4-17 ).

Specific complication rates for angioplasty of each anatomic area will be found in later chapters. One of the most potentially devastating generic complications is vascular rupture ( Fig. 4-18 ). Patients complain of severe pain persisting after deflation of the balloon, which may be accompanied by hypotension and tachycardia. Immediate reinflation of the balloon across or proximal to the lesion is a life-saving maneuver ( Box 4-6 ). Prolonged balloon inflation, reversal of anticoagulation, stent-graft placement, or urgent surgical repair (see later) may be required to stop the bleeding. The risk of vessel rupture during revascularization procedures, however, is much less than 1%.

Embolic Protection Devices

Particulate embolization likely occurs with all revascularization procedures, but it is clinically silent in most cases. In some vascular beds, such as the kidneys, the manifestations of particulate embolization may be subtle, such as a blunted or paradoxical clinical response to the intervention. The risk of clinically apparent embolization is highest with interventions in diseased arteries. Embolization of plaque elements is termed “cholesterol” embolization ( Box 4-7 ). This can occur as chunks of plaque (macroembolization) or cholesterol crystals (microembolization). Macroembolization results in acute occlusion of branch arteries with acute ischemia. Cholesterol crystals are mobilized during procedures by unroofing lipid-rich plaques and can continue to shower for days to weeks. Crystalline emboli lodged in arterioles are not angiographically visible, cause localized ischemia, and incite a painful inflammatory reaction in addition to the occlusion that can ultimately lead to tissue ischemia and necrosis (see Fig. 1-34 ). The overall incidence of cholesterol embolization is less than 1%, with patients with extensive atheromatous disease or friable plaques at greatest risk. Amputation, permanent renal failure, bowel ischemia, stroke, and death are all potential outcomes of cholesterol embolization.

Protection devices have been developed to minimize the risk of particulate embolization ( Fig. 4-19 ). The clinical drivers for these devices were originally coronary saphenous vein bypass and carotid artery interventions, but application to renal and other peripheral arterial interventions are being evaluated. There are three basic types: distal filters, distal occlusion balloons, and proximal occlusion balloons with or without flow reversal. Filter and distal occlusion balloons are advanced through the lesion on a guidewire before the intervention and then must be recovered afterward. The filtration devices have pore sizes from 80 to about 200 μm and permit continued antegrade flow of blood during the procedure unless a large volume of debris is collected. One limitation of filters is that small debris can pass through the filter. Distal occlusion balloons prevent all emboli while inflated, but do not allow any antegrade flow during the procedure and rely on complete aspiration of debris to prevent emboli after deflation. Proximal occlusion devices also prevent antegrade flow, but usually rely on reversal of flow rather than aspiration to remove debris. The efficacy of these devices in peripheral interventions is under evaluation.

Stents

Stents provide an intravascular scaffold for the vessel lumen. The mechanism of action of stents is very different from angioplasty, in that the plaque and vessel wall are literally pushed aside by the stent to enlarge the lumen ( Figs. 4-8 and 4-20 ). Many different stents are available, but not all are approved for vascular use in the United States ( Figs. 4-21 and 4-22 ). Many stents are approved and labeled as biliary or tracheal stents. The off-label use of stents in blood vessels is ubiquitous but not officially promoted by manufacturers.

The metal used to make the stent as well as the manufacturing technique have a major impact on the clinical performance of the device ( Table 4-6 ). The typical metals used are stainless steel (iron and chromium), nitinol (nickel and titanium), and Elgiloy (cobalt, chromium, nickel, and molybdenum). The latter two metals are not ferromagnetic and therefore do not cause artifacts on magnetic resonance imaging. Stents are categorized as either balloon-expandable (deployed by inflating an angioplasty balloon) or self-expanding (deployed by releasing a constraining mechanism). Stents are further categorized as closed or open-celled . Open-celled stents are the most flexible, but can “shingle” or “fish scale” when deployed in tight curves or over irregular lesions. Most stents are manufactured from small diameter tubes of metal that are cut in a proprietary pattern with a laser to create an expandable device, with the type depending on the properties of the metal. Self-expanding stents may also be made from wire woven or sewn together. Nitinol has thermal memory, so that it is soft and flexible at room temperature but becomes rigid and resumes a predetermined shape and size at body temperature. This property is useful for self-expanding intravascular devices.

Stent design greatly influences clinical use. Balloon-expandable stents are deployed by applying force from within (usually with an angioplasty balloon), and they resist vessel wall elastic recoil. However, they will not reexpand spontaneously if the resistance is temporarily overcome (i.e., the stent remains collapsed). Self-expanding stents attempt to reach a predetermined diameter unless continuously and externally constrained. Although not as resistant to elastic recoil as balloon-expandable stents, self-expanding stents reexpand spontaneously if compressed. This is desirable in superficial locations, such as the cervical carotid artery and superficial femoral artery.

Drug-eluting or releasing stents are a class of devices that combine the mechanical properties of a stent with the ability to locally deliver drugs to prevent restenosis. There are several strategies for bonding drugs to stents, including applying polymer coatings on the surface or microscopic “wells” in the stent metal. These devices were used initially in coronary arteries but are now being applied in the peripheral arteries. Using polymer and metal stents that dissolve over time is another approach to reducing restenosis.

The indications for stent placement depend on patient- and lesion-specific factors as well as the preference of the operator ( Box 4-8 ). In practice, the tendency is to place a stent as the first intervention without trying to optimize the result of angioplasty (“primary stenting”). Although expensive (stents can cost more than $1000 each), a good initial result is obtained quickly and with greater certainty. Nevertheless, in many instances it is wise to attempt angioplasty first, reserving stents for failed or recurrent lesions. This is especially true in heavily calcified lesions that may not respond to angioplasty, let alone stent placement. The long-term results of stents in most anatomic locations are better than angioplasty alone, but sometimes not by a great amount. In certain conditions, such as fibromuscular dysplasia, stents do not seem to offer any advantage over angioplasty, and therefore should probably be reserved to salvage failed angioplasty. Stents should not be placed at sites of anticipated surgical anastomoses, because the presence of the device may complicate surgery or render it impossible. Anticoagulation (e.g., heparin or bivalirudin bolus) should be used during stent placement. A 1- to 3-month course of antiplatelet therapy (aspirin, clopidogrel, or both) is often prescribed following stent placement although efficacy in preventing restenosis in peripheral arteries has not been studied thoroughly.

The techniques for stent placement vary depending upon the device, but certain broad principles can be followed. Delivery over a guidewire is essential to maintain access through the stent after deployment. Predilation of very tight lesions with an angioplasty balloon ensures that the lesion is stretchable (i.e., can be treated with a stent), and makes positioning the device easier. However, primary stent placement reduces overall manipulation of the lesion with potentially fewer distal embolic complications. All stents should be long enough to cover the lesion, with minimal extension into normal areas of the vessel.

Balloon-expandable stents deploy from both ends toward the middle (see Fig. 4-22 ). These stents are sized in the same manner as angioplasty balloons, up to 5%-10% greater in diameter than the measured normal lumen. Balloon-expandable stents shorten slightly during expansion in proportion to the final diameter. This shortening can be dramatic if the stent is markedly overdilated. When mounting a stent on a balloon by hand, careful crimping in a radial fashion (not twisting) is necessary to prevent dislodgment of the balloon during delivery. The stent should not be longer than the working surface of the balloon (i.e., the distance between the radiopaque marker bands on the catheter shaft inside the balloon). When placing a balloon-expandable stent, it may be necessary to first advance a long sheath or guiding catheter across the lesion so that the stent can be positioned without catching on plaque. Premounted balloon-expandable stents with smooth polished edges that are firmly situated on a balloon can often be “bare-backed” through lesions that are not too tight or irregular. Small-diameter stents mounted on small-shafted balloons are surprisingly flexible and can negotiate tortuous vessels.

Self-expanding stents typically deploy from distal to proximal relative to the operator. The unconstrained diameter of the stent should be 10%-20% larger than the normal diameter of the target vessel because fixation depends on secure apposition of the stent to the vessel wall. Self-expanding stents are usually constrained by an outer sleeve or membrane (see Fig. 4-21 ). These devices do not necessarily require the protection of a long sheath or guiding catheter to advance through a lesion, but predilation of the lesion is useful. Radiopacity of the stent is crucial to aid in correct placement and complete expansion. Most nitinol-based stents have highly radiopaque markers at the ends. Laser-cut nitinol-based stents have the least shortening during delivery, but are not reconstrainable like some woven stents. However, woven stents may shorten markedly because they are made small in diameter for insertion by being stretched out on the delivery catheter. Laser-cut self-expanding nitinol stents cannot be dilated beyond their resting maximum diameter. Woven stents have some capability to be dilated beyond their nominal resting diameter, but will shorten.

In general, a stent cannot do anything that an inflated balloon cannot (see Fig. 4-8 ). In other words, if the lesion cannot be dilated with a balloon, then a stent will not provide any additional benefit. When the primary abnormality is chronic extrinsic compression of the lumen by an extravascular structure, placement of a stent without first relieving the compression may result in stent fracture ( Fig. 4-23 ).

Stent placement originally had a higher overall complication rate than angioplasty alone owing to the larger size of the first devices and more complex delivery. Complication rates have decreased dramatically as experience has increased and devices refined. In most cases the complication rates are probably lower than aggressive angioplasty, in that acceptable results are easily obtained with a stent whereas multiple balloon inflations would have been required in the past. The types of procedural complications are largely the same as for angioplasty. However, there are several that are unique to stent placement ( Table 4-7 ). Early thrombosis of fully deployed stents is unusual unless runoff or inflow is compromised (see Fig. 4-3 ). Stent fracture can result when there is repetitive compression, torsion, or flexion of a device (see Fig. 4-23 ). Fractures are thought to be associated with restenosis. Infection is rare but usually results in pseudoaneurysm formation around the stent.

Stent-Grafts

A stent-graft is a device constructed from a stent and a fabric that is inserted from a remote access using catheter techniques and image guidance. The stent anchors the graft in the blood vessel lumen and in most cases also provides structural support for the graft material. The graft material provides a conduit for blood flow. The first clinically successful stent-grafts were handmade by physicians by combining available stents and surgical vascular grafts. There are now a variety of specially designed and commercially manufactured stent-grafts ( Fig. 4-24 ). These devices employ a range of metals, designs, and graft materials. Nitinol, stainless steel, and Elgiloy are commonly used to make the stents. The stents may be rigid or flexible, continuous or interrupted, and located inside, outside, or in a sandwich of graft material. The graft material may be synthetic, such as woven polyester or expanded polytetrafluoroethylene, or biological.

The function of a stent-graft varies with the application. In occlusive disease the stent-graft not only props the vessel open but also forms a physical barrier between the diseased intima and the lumen. When used to treat an aneurysm, the stent-graft provides a new flow channel and excludes the sac of the aneurysm from the circulation ( Fig. 4-25 ). In the treatment of vascular injuries such as an acute arteriovenous fistula or pseudoaneurysm, the stent-graft seals over the hole in the wall of the vessel. In each of these cases, the basic principle is diversion of blood flow through the stent-graft. Therefore, in order to function, the stent-graft must fully appose the inner walls of the vessel at the attachment sites. Otherwise, blood will leak between the stent-graft and the intimal surface. This is a fundamental difference between stent-grafts and surgical grafts, which are sewn to the vessel wall.

The indications for stent-grafts are evolving as new devices and data become available ( Box 4-9 ). Stent-grafts are considered a standard treatment for arterial aneurysms, particularly of the abdominal and thoracic aorta. The results of the transjugular intrahepatic portosystemic shunt procedure have been improved dramatically by stent-grafts. Recanalization of arterial occlusions, kissing common iliac artery stents, and long segment superficial femoral arterial disease are some of the peripheral arterial applications of stent-grafts.