Anatomic and Technical Considerations of CAS
Patient selection, choice of devices, and guidelines for avoiding complications in successful carotid stent placement.
Patient selection in managing carotid artery occlusive disease by an endovascular stenting procedure is the single most important consideration for a successful outcome. Operator experience is the second most important factor. Combined, these criteria will certainly assist in minimizing periprocedural strokes. Strokes occurring in the early stages of a carotid program can be destructive in terms of the procedure's acceptability as an alternative to carotid endarterectomy (CEA). There are several additional stroke predictors including length of lesion, presence or absence of ulceration, clot versus plaque at the target lesion site, lesion position within the cervical segment of the carotid, tandem lesions of the higher cervical carotid segment, or coexisting occlusive disease at the ostium of the great vessels off the aortic arch.
In a series of 364 patients who had undergone carotid stenting with a follow-up of 2 years or more, the single most important stroke risk predictor was length of lesion. We noted a 15% periprocedural stroke, death, or myocardial infarction rate in those patients whose length of lesion was >2 cm. Eighty-five percent of the lesions were less than 1.5 cm in length and those patients had only a 2% (stroke death and myocardial infarction) periprocedural event.
The status of the external carotid artery is also quite important in patients with basilar artery insufficiency and vertebral artery occlusive disease when the external occipital from the external carotid is responsible for reconstitution of the vertebral artery and subsequently antegrade flow in the basilar artery. Occlusion of the external carotid in these situations will result in basilar artery infarction. The presence or absence of functioning anterior or posterior communicating vessels to reconstruct the circle of Willis is also a risk stratification factor in the evaluation of patients undergoing carotid stenting. Understanding the neurovascular anatomy in cerebral hemispheres and the posterior fossa circulation is critical in assessing risk stratification. For these reasons, I developed an anatomic classification of target lesion characteristics: type A, type B, and type C.
TARGET LESION CHARACTERISTICS
The type A lesion represents patients with ³80% stenosis and with concentric stenosis, the length of which is not greater than 1.5 cm. The proximal internal carotid is involved but spares the ostium and is nonulcerative. We recommended that when beginning carotid stenting, the operator should initially choose 15 to 20 patients with type A lesions (Figure 1). Unfortunately, even type A lesions cannot be taken for granted. Carotid plaque composition is still an unsettled issue and, in fact, could represent the same type of vulnerability that exists in the coronary circulation. In order to better understand the plaque composition, we have now incorporated into our routine the use of intravascular ultrasound (IVUS) with virtual histology (VH). The VH provides a quantitative analysis of plaque characteristics, including fibrous, necrotic, calcific, and lipid components. From this analysis, it is conceivable that we could determine from these characteristics which patients may not be ideally suited for endovascular stenting and may be more appropriately managed with CEA (Figure 2).
After experience with 15 or 20 type A lesions, an operator could proceed to type B lesions. The type B lesion is somewhat more complex, has ³85% stenosis at the target lesion site, is longer than 1.5 cm, may be ulcerative and/or eccentric, and involves the bifurcation including the distal common carotid artery and the ostium.
Type C lesions are found in patients with lesions longer than 2 cm and >85% stenosis. Tandem lesions may be present with common carotid ostium involvement or extensive involvement at the bifurcation and the proximal internal carotid. Major ulceration and suspicion of clot may be present. Occasionally, the dystrophic calcification is in a globular configuration at the bifurcation. When the presence of plaque versus clot cannot be differentiated angiographically or with IVUS and VH, those patients should categorically be referred for CEA and endovascular stenting should not be attempted. An additional absolute contraindication is the polypoid rigidly calcified defect that will not respond to stenting. If the carotid lesion cannot be dilated with the conventional predilatation balloon, there is unreasonable risk in attempting to stent those lesions. The anatomic risk predictors that basically represent the type C lesion are best avoided. In addition, aneurysmal involvement, dissection, and the complex aortic arch require special attention (Figure 3).
LIMITATIONS OF COMPLEX ARCHES
In the early stages of carotid endovascular stenting, the aortic arch was not considered a major procedural limitation. More recently, however, the complex aortic arch has become the single most limiting factor in successful carotid stenting. Regardless of the complexity of the internal carotid lesion or the degree of stenosis, we invariably are technically successful in accessing those lesions with the appropriate filter and endovascular stent choice. The arch, however, is a different situation. The recent classification of the aortic arch into type 1, type 2, and type 3 has been helpful in our approach to the design of catheter configurations, allowing easier access to the great vessels off the arch. We have recently added a type 4 configuration. Type 1 represents a relatively level aortic arch, making accessibility with conventional linear-directed catheters (eg, H1[Cook Incorporated, Indianapolis, IN], JR4 [Cordis Corporation, a Johnson & Johnson Company, Miami, FL]) less difficult. At least 85% of our procedures are done with the use of an H1 diagnostic catheter. Angulation of the aortic arch is classified as type 2 and type 3, with both representing increasing degrees of angulation, and with the innominate and left common carotid rising essentially from an ascending aortic configuration. When the plane of the aortic arch is angled inferiorly, the left common carotid and the innominate artery are rising at acute angles and require complex curves for access to the great vessels. In these situations, we will use the Simmons 1 or Simmons 2 catheter (Cook Incorporated) configurations for access. The type 4 aortic arch may be the most difficult, with not only angulation but the length of the arch and transverse diameters increased, which is associated with redundancy of the common carotid artery. It is more frequently noted on the right, where the tortuosity and redundant loop exist in the proximal segment. In these situations, it is extremely difficult to track either the sheath or guiding catheter to an appropriate position for stenting. This configuration may require the use of a coronary-configured guide, AR2 or AL3 (Cordis Corporation), to be positioned at the ostium of the innominate artery or the left common carotid utilizing a two-wire technique to maintain stability of the guide. A .014-inch wire is positioned in the external carotid and the filter wire is obviously positioned in the internal carotid beyond the target lesion, allowing deployment of the stent with a guide at the ostium and passage to the target lesion. The .014-inch wire in the external carotid provides the necessary stability. After the stent is deployed, the supplementary wire in the external carotid can be removed. This procedure is recommended only in situations in which there is an absolute surgical contraindication. The risk in the absence of the additional support wire in the external carotid artery is that if the guide should be displaced to the aorta during the procedure, it may be impossible to reposition the guide to an appropriate position because the .014-inch filter wire will not support that degree of tracking.
Most technical failures in carotid stenting are related to a complex aortic arch. Excessive manipulation in the aortic arch is also to be avoided considering that, in those situations, shower emboli can occur, which can cause either contralateral or ipsilateral hemispheric strokes during the procedure. It is worth remembering that a technical failure is acceptable, but a technical failure with a stroke is not. Having difficulty with the bovine origin of the left common carotid or the deep-seeded innominate artery and persisting more than 5 minutes with any single catheter is not worth the risk. Most of the successful catheter configurations have anterior and superior directions at the aortic arch and do not require an unreasonable amount of manipulation. The brachial artery is also an option but rarely a necessary site for access. Whether you choose to use a guide or a sheath is a matter of physician preference. We have become accustomed to the guides, considering that the softer tip might result in less dissection and is less likely to kink. In addition, the guide has torque and steerable features. The multiple configurations also allow the ability, when necessary, to successfully choose a guide based on the anatomy. The disadvantage is that the guide requires an 8-F sheath versus a 6-F sheath. In our experience, this has not been an important issue.
In choosing the stent and filter, we would like to match the most appropriate design for the anatomic characteristics of the lesion, although currently, we have not had this luxury because only a few stents have received FDA approval. For example, short focal segments may be best managed with self-expanding nitinol stent technology. This may be either open- or closed-cell geometry. Lengthy lesions with ulceration extending across the bifurcation might be best served with the braided monofilament alloys. The short focal segment that is rigidly calcified may require support from the reinforced closed-cell design. Open-cell configurations across complex bifurcations unfortunately have a "fish scale" effect as the stent is deployed, resulting in the open-cell strut impinging on the intima (Figure 4). Actual extravasation of contrast media has been noted to extend to the adventitia. This also results in the possibility of restenosis at this site in addition to a stress point that may result in stent fracture. We have observed one patient with stent fracture restenosis on follow-up examination. This was effectively managed with adjunctive stenting. More recently, there has been interest in the development of the helically designed stent, with improved flexibility and controllable cell size, both of which allow improved trackability while eliminating plaque prolapse.
RISKS AND BENEFITS OF EMBOLIC PROTECTION
Computational fluid model analyses of filters have shown considerable variation in their efficacy. Ideally, we prefer the filter best designed for the lesion and for efficacious particle capture. The pore size of most distal protection filters varies from 80 µm to 120 µm. Obviously, the larger the pore size is, the less effective the particle capture will be. On the other hand, the smaller the pore size is, the more likely that flow limitation will occur during the procedure. Proximal balloon control with flow reversal embolic protection systems has become an integral part of procedures, but are not yet available in the US. We have noted on several occasions that simple aspiration of the sheath or guiding catheter during or at the completion of the procedure has revealed significantly visible macroparticles coming from either the stent surface or the adjacent carotid. At this point in time, flow reversal systems cannot be ignored.
The limitation of the flow reversal system is the need to clearly understand the intracranial circle of Willis collateralization across either the anterior communicating or posterior communicating arteries in allowing efficient collateralization during occlusion. One advantage of the embolic filter device is the ability to provide flow during the procedure while fully expanded. Variations in designs may result in poor apposition in tortuous vessels, allowing the microemboli to pass through these sites. There is also the risk of inducing spasm or damage to the vessel with dissection by design limitations of the distal wire or unnecessarily rigid struts. Filter detachment with embolization and filter clotting have been described. Furthermore, there is always the possibility of trapped embolic particles being squeezed out of the filter as it is retracted or collapsed. Filters have reduced the periprocedural stroke event rate by 50% in essentially all registries. Their value has been established, but they all are not created equal, and experience and caution are necessary to avoid complications. Finally, when a neurologic event does occur, one should have a contingency plan and the necessary equipment for a neurointervention.
Few procedures in the history of medicine have had such resistance as carotid stenting as an alternative to CEA. Certainly in the high-risk surgical subsets of patients, carotid stenting has been at least the equivalent and, in several parameters, more favorable than CEA. However, there are a few unanswered issues, including the moderate- to low-risk subset because essentially all of the data that have been validated have been in the high-risk subset. The octogenarian subset also needs clarification because the early reports suggest that symptomatic as well as asymptomatic octogenarians have periprocedural stoke event rates varying from 8% to 16%. This is clearly unacceptable, and I am quite certain that those numbers can be improved upon with better technology. Major strokes are also an issue, and although the incidence is only 2% or less in essentially all of the high-risk registries and trials, the problem is the uncertainty of who is at risk for a major stroke during a procedure. This "Russian roulette" needs clarification. Hopefully with the introduction of IVUS and the VH component for plaque characteristic analysis, it may be possible to separate patients who are better suited for CEA than stenting.
A solution to the shower embolic events that may be occurring from the aortic arch or excessive manipulation within the common carotid artery during the stenting procedure has yet to be found. Progress has, however, been impressive, and it is only a matter of time until we have the solutions for the octogenarians, the complex aortic arch, more efficient distal protection systems, and newly designed stent configurations.
Mark H. Wholey, MD, is a Clinical Professor of Radiology at the University of Pittsburgh School of Medicine, UPMC Shadyside Hospital, and Chairman of the Pittsburgh Vascular Institute in Pennsylvania. He has disclosed that he is a paid consultant to Cordis Corporation. Dr. Wholey may be reached at (412) 623-2083; firstname.lastname@example.org.