Carotid Endarterectomy

Surgical treatment for carotid artery stenoses was introduced in the early 1950s.² Although early observational data suggested benefit for surgery over medical therapy, large prospective randomized trials investigating the beneficial effect of CEA on stroke reduction were not initiated until the late 1980s and early 1990s. Landmark studies including the North American Symptomatic Carotid Endarterectomy Trial (NASCET),^3–6 European Carotid Surgery Trial (ECST),⁷ Asymptomatic Carotid Atherosclerosis Study (ACAS),⁸ and Asymptomatic Carotid Surgery Trial (ACST)⁹ confirmed the benefits of surgery over best available medical treatment for reducing the risk of stroke in both symptomatic (NASCET, ECST) and asymptomatic patients (ACAS, ACST). Carotid endarterectomy surgery is comprehensively discussed in Chapter 33.

Patients included in these surgical studies were carefully selected; specifically, subjects who were considered high CEA risk were excluded from participation. Thus, octogenarians, patients with recurrent stenosis following prior ipsilateral endarterectomy, intracranial stenosis that was more severe than the surgically accessible lesion in the neck, unstable angina pectoris, recent myocardial infarction (MI), contralateral CEA, patients on long-term anticoagulation therapy, and surgically inaccessible lesions were all excluded from these trials.³,⁸

Endovascular Approaches to Treat Carotid Stenosis

The mission to develop safer percutaneous endovascular solutions to treat arterial stenosis was pioneered by Dotter¹⁰ and Gruntzig and Hopff¹¹ in the 1960s and 1970s. In 1977, Klaus Mathias, an interventional radiologist, described a catheter system that could be used for performing balloon angioplasty of cervical carotid stenosis,¹² and this was followed by a few case reports of successful carotid angioplasty performed in the surgical suite.¹³,¹⁴ In 1984, Vitek and his neuroradiology colleagues from the University of Alabama at Birmingham (UAB)¹⁵ reported angioplasty of the innominate artery aided by balloon occlusion protection of the common carotid artery (CCA). This early report represents the first percutaneous intervention performed with the benefit of distal embolic protection. During the 1980s, clinical reports of carotid angioplasty were sporadic and limited to small single-center series of patients.¹⁶ Kachel et al. summarized the results of carotid angioplasty published in the literature through 1995 and noted that 503 of the 523 (96%) procedures were technically successful. There were no deaths, major strokes occurred in 2.1%, and minor complications were in the single digits (6.3%).¹⁷ In 1985, Rabkin and Germashev began using early-generation nitinol (an alloy of nickel and titanium) stents as an endovascular prosthesis and subsequently reported their 5-year experience.¹⁸

Resistance to widespread acceptance and the slow progress of angioplasty involving the supraaortic vessels was largely due to two major concerns: (1) local vessel injury related to balloon inflation causing a flow-limiting dissection with a risk of acute vessel closure (prestent era), and (2) the risk of distal embolization. In later years, these key limitations would be overcome by the introduction and widespread adoption of stents and EPDs.

In March 1994, Iyer, Vitek, and Roubin initiated the carotid angioplasty program at UAB under carefully scrutinized institutional protocols.¹⁹ Initial interventions were performed using stand-alone balloon angioplasty (no stents). To maximize the luminal result, a long inflation was performed using a 5-mm over-the-wire balloon; the center port of this balloon could accommodate a 0.035-inch guidewire. Once the balloon was in place, the wire was withdrawn, and oxygenated arterial blood withdrawn from the femoral artery was infused through the center port of the balloon with the help of a special pump device, permitting a long 10-minute balloon inflation. The first four patients were treated without complications. Patient #5, a woman with contralateral carotid occlusion, presented with a transient ischemic attack (TIA) related to a high-grade stenosis in the index carotid artery and underwent an uncomplicated balloon angioplasty procedure. Despite a perfectly acceptable angiographic result, approximately an hour after the procedure, there was acute closure of the angioplastied carotid artery. Although a technically successful, urgent reintervention with recanalization and stenting of the occluded vessel was performed, the patient did not recover from the major stroke related to the acute closure and subsequently expired. This case triggered the decision by the UAB group to perform elective carotid stenting—irrespective of the angiographic results of balloon angioplasty—and primary stenting became the intervention of choice for treatment of cervical carotid stenosis. The subsequent rapid adoption of this approach by interventional cardiologists in particular, and the endovascular interventional community in general, heralded the modern era of endovascular treatment for extracranial carotid bifurcation disease.^19–22

Although balloon expandable stents were used in the first 100 patients, by the summer of 1995 (when the initial patients returned for their follow-up angiograms) it became clear that these stents were prone to deformation (stent crush) because of the superficial location of the carotid artery and the associated movements of the neck.²³ The Alabama group were the first to report this complication, seen in approximately 15% of patients at 6-month follow-up.²³ Fortunately, stent deformation was largely a cosmetic issue, with only one patient presenting with symptoms in this series. This observation, as well as the recognition that chances for regulatory approval for balloon expandable stents for treating extracranial carotid stenosis were slim, led to the rapid introduction, testing, and adoption of self-expanding stents. Stents have all but abolished acute carotid vessel closure, and in contemporary practice, primary carotid stenting is the norm. The reader should note that unlike in coronary arteries, the risk of acute stent thrombosis and instant restenosis, two major limitations of coronary stents, are nonissues when stents are deployed in the extracranial carotid location.

In 1996, Theron et al.²⁴ reported results from his seminal work using his triple coaxial catheter that incorporated distal balloon occlusion for providing embolic protection during carotid bifurcation angioplasty. Unfortunately, this early-generation distal protection balloon could only be used with balloon angioplasty (and not with stents). By 2000, the first investigational distal balloon occlusion EPD, the Percusurge Guardwire (Medtronic, Minneapolis, Minn.) was introduced into clinical trials in the United States. This was soon followed by a number of clinical trials, all of which included a filter as the distal EPD. Unlike Theron’s distal occlusion balloon, the Percusurge balloon—as well as all such filters—can be used with both over-the-wire and monorail stent delivery systems. As increasing clinical data became available, use of distal protection devices was recognized and accepted by many (but not all²⁵) as an integral if not mandatory part of carotid artery dilation and stenting.^26–31 In our opinion, EPDs, when selected and used appropriately, improve procedure safety by significantly reducing the risk of procedure related embolic major and fatal strokes.

The multicenter Carotid and Vertebral Artery Transluminal Angioplasty Study (CAVATAS-I)⁹ was conducted between 1992 and 1997 in the United Kingdom during an era when stents were neither widely available nor perceived as integral. This prospective randomized trial compared outcomes of balloon angioplasty and CEA in 504 patients; only 55 patients (26%) within the group assigned to endovascular treatment received stents. Initially, stents were used only as a bailout treatment, with increased elective use toward the end of the study; distal protection devices were not used. Major event rates within 30 days after treatment did not differ significantly between endovascular treatment and surgery: disabling stroke or death (6.4% vs. 5.9%) and any major stroke lasting more than 7 days or death (10.0% vs. 9.9%). This study also demonstrated that endovascular techniques were superior to surgery when considering other risks related to the incision in the neck and use of general anesthesia. Cranial nerve injury was reported in 8.7% of surgical patients; no events occurred in patients undergoing endovascular procedures (P < 0.0001). Major groin or neck hematomas occurred less often after endovascular treatment than after surgery (1.2% vs. 6.7%, P < 0.0015). The results of this early clinical trial set the stage for investigation of carotid stenting.

Symptomatic Patients

Symptomatic carotid stenosis refers to ischemia or infarction in the distribution of the internal carotid artery (ICA) causing neurological abnormalities that include but are not limited to contralateral motor and/or sensory events, speech, and/or visual problems (monocular blindness, field defects). Amaurosis fugax refers to transient monocular visual loss, typically described by the patient as a shade being drawn down or across the eye (amaurosis, Greek for “darkening,” and fugax, Latin for “fleeting”). Dizziness and problems with balance are symptoms that typically result from ischemia or infarction in the vertebrobasilar system, and the presence of a carotid artery stenosis in a patient presenting with dizziness is almost always incidental i.e., the carotid stenosis is most often asymptomatic and NOT causally related to the symptoms. The culprit stenosis is considered symptomatic for 6 months beyond the event. Additionally, the risk of recurrent stroke is lower in patients who present with amaurosis as the sole symptom in comparison to patients who present with a hemispheric TIA.

It is well accepted that revascularization should be offered to all symptomatic patients if the diameter of the ICA is reduced more than 70% as documented by noninvasive imaging, or more than 50% as documented by catheter angiography (Table 32-1). There is, however, one important caveat: the periprocedural risk of stroke or death related to the revascularization procedure (CEA or CAS) should be under 6%.³² In symptomatic patients, there is a well-established correlation between increasing stenosis severity and stroke risk. The NASCET study³ demonstrated the benefit of CEA over medical treatment for reducing the risk of future stroke in symptomatic patients with carotid stenosis between 70% and 99%. The NASCET results also showed that symptomatic patients with a lesser degree of stenosis (between 50% and 70%) benefit less. Revascularization is typically recommended in this group if there are additional unfavorable angiographic features (e.g., ulceration or other features associated with increased risk of vessel-to-vessel embolization).

^{Table 32-1} Modified American Heart Association Recommendations for Carotid Artery Revascularization

CABG, coronary artery bypass graft surgery.

^* Lesion severity is determined according to the North American Symptomatic Carotid Endarterectomy Trial (NASCET) methodology (i.e., the ratio between lumen diameter at the point of maximal stenosis and the lumen diameter of the non tapered segment of the distal internal carotid artery).¹⁵³

Modified from Brott TG, Halperin JL, Abbara S, et al: ASA/ACCF/AHA/AANN/AANS/ACR/ASNR/CNS/SAIP/SCAI/SIR/SNIS/SVM/SVS guideline on the management of patients with extracranial carotid and vertebral artery disease: executive summary. A Report of the American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines, and the American Stroke Association, American Association of Neuroscience Nurses, American Association of Neurological Surgeons, American College of Radiology, American Society of Neuroradiology, Congress of Neurological Surgeons, Society of Atherosclerosis Imaging and Prevention, Society for Cardiovascular Angiography and Interventions, Society of Interventional Radiology, Society of NeuroInterventional Surgery, Society for Vascular Medicine, and Society for Vascular Surgery, developed in collaboration with the American Academy of Neurology and Society of Cardiovascular Computed Tomography. J Am Coll Cardiol 57:1002–1044, 2011; and Roubin GS, Iyer S, Halkin A, et al. Realizing the potential of carotid artery stenting: proposed paradigms for patient selection and procedural technique. Circulation 113:2021–2030, 2006.

An important, albeit controversial and unsettled, issue in the treatment of symptomatic patients relates to the timing of the revascularization procedure after the index symptomatic event.^33–37 Risk of a recurrent neurological event after a TIA or stroke is estimated to be between 15% and 20%, and this elevated risk persists for approximately 6 months after the initial event and underlies the rationale for the 6-month threshold for defining symptomatic patients. Proponents of early intervention i.e., within a few days of the symptomatic event. Argue that the highest risk of a recurrent event is during this early period and any delay in treatment will significantly diminish its therapeutic value, since a substantial portion of these patients would have already experienced a neurological event during the waiting period. A key reason underlying the reluctance of operators to perform revascularization (CEA or CAS) soon after a stroke (less so after a TIA) is the concern that early treatment increases the risk of hemorrhagic transformation of the culprit, (nonhemorrhagic) infarct. Although the increased risk of intracranial hemorrhage following early intervention has been challenged,³³ a recent retrospective analysis³⁸ of a large national inpatient database involving more than 57 million in-hospital admissions, conducted to determine the prevalence and risk factors of intracranial hemorrhage among patients undergoing CEA (N = 215,012) and CAS (N = 13,884), arrived at a different conclusion. Symptomatic presentations represented the minority of indications for CEA (n = 10,049 [5%]), as well as CAS (n = 1251 [10%]). Intracranial hemorrhage occurred significantly more frequently after CAS than CEA in both symptomatic (4.4% vs. 0.8%; P <0.0001) and asymptomatic presentations (0.5% vs. 0.06%; P <0.0001). Multivariate regression suggested that symptomatic presentations (vs. asymptomatic) and CAS procedures (vs. CEA) were both independently predictive of six- to sevenfold increases in the frequency of postoperative intracranial hemorrhage. The observations from this retrospective population-based analysis are at variance with observations from recent large prospective randomized trials involving symptomatic and asymptomatic patients, wherein the risk of mortality and major stroke was uniformly low in both CEA and CAS arms, and the higher incidence of neurological events in the CAS arm was a result of minor strokes (ischemic rather than hemorrhagic).^39–41 Pending resolution of this issue by future studies, the current approach of waiting a minimum of 3 weeks following the index event (longer for larger strokes) is likely to continue.

Asymptomatic Patients

Asymptomatic carotid disease refers to the presence of a stenosis resulting in a 60% or greater reduction of the luminal diameter of the extracranial ICA without symptoms of ipsilateral stroke, TIA, or amaurosis fugax. Treatment of patients with asymptomatic carotid artery stenosis has become extremely controversial, with two main issues fuelling this ongoing debate^42–44:

Procedure-Related Risks

Over the past decade, one of the most important advances in the field of carotid stenting relates to our understanding of what constitutes “high stent risk.” It is important to remember that CEA was first performed in the 1950s, and over the course of the next several years, surgeons identified both anatomical features and comorbidities that would increase the risk of endarterectomy (high–CEA risk group). Furthermore, these high–CEA risk patients were excluded from participating in the major randomized CEA trials.

Standard stent risk

Although carotid stenting outcomes are not influenced by gender, age is a very important determinant. The concept of brain reserve is akin to cardiac reserve—a patient with poor left ventricular (LV) function is more likely to manifest and experience complications related to a percutaneous coronary intervention (PCI) or CABG procedure. Similarly, a patient with compromised brain function (diminished brain reserve) is more likely to clinically manifest neurological events related to periprocedural embolization. Embolization is a universal occurrence with all CAS procedures and happens despite the use of EPDs. Patients with prior large strokes, multiple small strokes, or lacunar infarcts and those with dementia are examples of patients with compromised brain reserve. Dementia in particular is a problem. Despite having a perfectly acceptable angiographic and clinical result (i.e., no procedure-related events) in the follow-up period, anecdotal reports suggest a marked deterioration in memory and other cognitive functions. The reason for this is unclear, but dementia should be considered at least a relative contraindication for CAS.

Close attention should be paid to the end-diastolic ultrasound flow velocity. If this value exceeds 100 cm/s (especially >120 cm/s), the angiographic stenosis severity will exceed 80% (as defined by the NASCET criteria; Fig. 32-1) and will meet the treatment threshold for treating asymptomatic lesions.

Figure 32-1 Ideal lesion and vessel morphology for carotid stenting.

Ideal lesion (A) has high-grade (>80%) internal carotid artery (ICA) stenosis (enddiastolic velocity 124 cm/s). Note that ICA/external carotid artery angle is acute, and the artery cephalad to stenosis is free of significant bends. B, Same lesion following carotid stenting using a distal embolic protection device (EPD [filter]).

Figure 32-2 illustrates lesion and vessel features that are ideal for stenting using distal embolic protection. These features include:

• A narrow acute angle between the ICA and external carotid artery (ECA). The wider this bifurcation (i.e., the angle approaches 90 degrees or is frankly obtuse), the greater the technical difficulty in advancing a distal embolic protection filter device with a fixed-wire system (Fig. 32-3). The technical difficulty imparted by an open ICA/ECA angle is compounded if there is additional tortuosity in the ICA distal to the stenosis (Fig. 32-4).

• Minimal calcification and no ulceration. Some degree of calcification is nearly ubiquitous in a diseased carotid bifurcation, but heavy concentric calcification in association with a severe stenosis is a major problem. Although the demonstration of carotid calcification is straightforward and requires only fluoroscopy (Fig. 32-5), the distinction between deep vessel wall calcium and superficial calcium encroaching on the vessel lumen may be difficult, and the decision to declare the case unsuitable for CAS is largely subjective. We arbitrarily define heavy calcification as calcification 3 mm or more in width, with concentricity defined by imaging in two orthogonal views. The unyielding nature of calcium, along with the stiffness it imparts to the involved vessel segment, makes it difficult to predilate and advance the EPD and stent delivery system through the lesion (especially the stent). Forcing these devices in an attempt to cross the stenosis not only increases the chances of prolapsing the sheath out of the CCA, it also increases the risk of embolization, spasm, and dissection. Inability to completely dilate and expand the deployed stent despite using larger and/or high-pressure balloons (resulting in a stent with an hourglass appearance) is an intraprocedural nightmare.

• The artery, especially cephalad to the stenosis, is free of any significant kinks or bends. Presence or absence of this key unfavorable feature is extremely important to note on preprocedure magnetic resonance angiography (MRA), computed tomographic angiography (CTA), or invasive angiography, since it increases the degree of difficulty when attempting to place a distal filter EPD. Excessive vascular tortuosity is defined as two or more bend points that are 90 degrees or greater (see Fig. 32-4). At times the tortuosity can be extreme and may impart a hairpin bend to the ICA (see Fig. 32-4). Worsening grades of tortuosity increase the difficulty when attempting to cross the stenosis and may make device delivery difficult or impossible. Straightening of the tortuous vessel segment by stiff wires or devices may result in vessel spasm and reduced antegrade flow. Thus, despite filter placement, the patient does not receive the benefit of brisk antegrade flow and may manifest ischemic symptoms in the absence of adequate collaterals. Additionally, slow flow increases the risk of fibrin deposition within the filter. The longer the dwell time of the EPD, the higher the risk of an iatrogenic thrombus. Iatrogenic tortuosity can also be introduced by placement of the sheath in a redundant carotid artery, so tortuosity should be assessed after the sheath is in place below the carotid bifurcation.

Figure 32-2 This lesion is unsuitable for carotid artery stenting.

The high grade of stenosis, 90-degree angle between the internal carotid artery and external carotid artery, and proximal calcification make passage of wires and equipment hazardous.

Figure 32-3 These lesions are unsuitable for carotid artery stenting using a distal embolic protection device.

This is due to the high degree of tortuosity of the internal carotid artery distal to the lesion, preventing safe and effective positioning of a filter device.

Figure 32-4 Extensive circumferential calcification of the internal carotid artery makes this lesion unsuitable for carotid artery stenting.

Subtraction imaging without contrast shows calcification of both the internal carotid artery and external carotid artery.

Figure 32-5 This vessel demonstrates a string sign with high-grade stenosis and reduced cranial flow and internal carotid artery filling (arrow).

Note the relatively complete filling of the external carotid artery vessels in comparison.

Stenosis severity and eccentricity/concentricity are not problems as long as the flow in the vessel is normal (Thrombolysis in Myocardial Infarction [TIMI] grade III). A severe stenosis in association with less than TIMI III flow (string sign, Fig. 32-6) and an occluded artery are contraindications (Fig. 32-7) for CAS. Ulceration is often noted, even on angiograms from asymptomatic patients, and although not a contraindication, operators should be aware that the risk of embolization might be higher, particularly during the phase of poststent balloon dilation. Angiographic filling defects that are consistent with a thrombus are a contraindication to CAS (Fig. 32-8). Note that both calcium and thrombus may appear as filling defects, and the differentiation is based on the clinical presentation. Whereas a filling defect in a symptomatic patient should be presumed to be thrombus (see Fig. 32-7), filling defects in asymptomatic patients are frequently a result of calcium encroaching on the vessel lumen (Fig. 32-9).

Figure 32-6 An occluded carotid artery is an absolute contraindication to stenting.

Figure 32-7 Thrombus located in proximal internal carotid artery in a patient with symptomatic carotid artery disease.

This is identified by the hazy appearance and is only visible following contrast injection.

Figure 32-8 This patient demonstrates eccentric calcification, which may at times closely resemble thrombus.

Fluoroscopy without contrast injection will typically reveal calcification, as also shown in Figure 32-5.

Figure 32-9 Unfavorable vessel morphology for carotid stenting.

A, Unfavorable lesion. Note obtuse internal carotid artery/external carotid artery (ICA/ECA) angle, high-grade eccentric stenosis immediately distal to bifurcation, and ulcer proximal to stenosis near carotid bulb. Vessel distal to stenosis is straight. B, Result can be seen after treatment using an Emboshield (Abbott Vascular, Santa Clara, Calif.) filter. Wire is independent of filter, and negotiating the unfavorable bifurcation and severe eccentric stenosis is far easier with a wire uncoupled from the filter element. Pre-predilation may be needed. Open-cell stent was used to treat the lesion on the bend; this stent design does not introduce any additional bends in ICA post stenting. Care should be taken to place proximal end of stent flush with origin of ICA. If it hangs between ICA origin and common carotid artery (CCA), stent edge can cause problems in advancing postdilatation balloon as well as filter retrieval catheter. Note that ulcer is excluded, not obliterated, and no attempt should be made to obliterate ulcer by using larger balloons. Flow to ulcer crater will seal off in time.

As a rule, unfavorable anatomical features (Figs. 32-10 and 32-11; also see Figs. 32-3 through 32-9) should be considered contraindications for CAS. Although special techniques (e.g., use of a heavy-gauge buddy wire to straighten tortuous vessel segments, use of cutting balloons to dilate unyielding lesions) may result in a satisfactory angiographic outcome, the risk of a procedure-related neurological event should be presumed to breach the accepted periprocedural complication threshold.

Figure 32-10 Unfavorable vessel morphology for carotid stenting.

A, Complex lesion. There is an obtuse internal carotid artery/external carotid artery (ICA/ECA) angle as well extreme tortuosity of the ICA distal to the stenosis. Note the carotid stent in the contralateral carotid artery. B, Final result after carotid artery stenting. This lesion was treated using a Percusurge GuardWire (Medtronic, Minneapolis, Minn.) distal occlusion balloon for embolic protection and an open-cell stent. Distal filters are contraindicated. Proximal flow reversal is an option, provided the arch anatomy is favorable for placing larger French-size catheters in the carotid artery. This type of vessel morphology will be technically challenging and should be a contraindication for the beginner and low- and medium-volume operators.

Figure 32-11 Classification of the aortic arch.

In the frontal projection, a horizontal line is drawn across the origin of the left subclavian artery. Type I: all the great vessels originate at same level and meet this line. Access to left carotid and innominate arteries is easiest with this aortic arch configuration. Type II and type III: as the aorta becomes more unfolded and elongated (a function of increasing age and hypertension), origin of great vessels becomes displaced more posteriorly, and on the frontal projection, origins are progressively displaced inferior to the horizontal line referenced above. Access becomes increasingly difficult because a catheter approaching from the descending aorta tends to prolapse into the ascending aorta.

Initial Evaluation

Preprocedure Issues

Patients who consent to brachiocephalic angiography and possible stent placement should undergo a comprehensive clinical cardiovascular evaluation to assess for presence of coronary artery disease(CAD), aortic stenosis, and LV dysfunction. The presence of these comorbidities will impact a patient’s tolerance of the hemodynamic effects of CAS. All patients should undergo evaluation of the preprocedure neurological status, and baseline NIH, Rankin, and Barthel Stroke Scales should be documented. All symptomatic patients, those with a prior history of stroke, and those with an abnormal neurological examination should have a CT or MRI scan of the brain to document baseline status.

Antiplatelet Agents and Anticoagulants

In the clinical protocol, great emphasis is placed on dual antiplatelet therapy before and after carotid stenting. The non-event of stent thrombosis and the low rates of peri- and postprocedural embolic events are predicated upon administration of the correct doses of adjunctive antiplatelet therapy. All patients should receive aspirin, 81 to 325 mg daily, and clopidogrel, 75 mg daily, prior to the procedure and for a minimum of 30 days after the procedure.³² If a patient has not received both aspirin and clopidogrel on a daily basis, we suggest they receive a 600-mg loading dose of clopidogrel at least 4 hours prior to the procedure. If this is not possible, the procedure should be rescheduled. There is no experience using prasugrel in patients undergoing carotid stenting.

The approach to patients who are on chronic anticoagulation with warfarin should be individualized, with the acknowledgement that triple therapy with aspirin, clopidogrel, and warfarin increases the risk of bleeding. Discontinuing warfarin while the patient is on dual antiplatelet treatment may be acceptable in patients at low risk for systemic embolism. If the patient requires anticoagulant therapy because of a high risk of thromboembolism, such as a prosthetic mechanical valve, it is appropriate to discontinue warfarin for approximately 4 days prior to the scheduled invasive procedure, “bridge” the patient with heparin if appropriate, and then restart warfarin on the evening of the carotid stent procedure. In patients requiring warfarin in the poststent period, dual antiplatelet therapy should include 81 mg of aspirin together with 75 mg of clopidogrel. Dual antiplatelet medications maintained for 6 to 8 weeks after carotid stenting is optimal.

Antihypertensive and β-Blocker Medication

Blood pressure and/or heart rate–lowering medications are typically withheld the day of the procedure to avoid excessive bradycardia and hypotension resulting from procedure-related carotid baroreceptor stimulation. Postprocedure, blood pressure and heart rate should be followed closely and medications reintroduced as soon as the clinical situation permits. In patients with restenosis following prior CEA (denervated carotid bulb) or in cases where the location of the stenosis is such that balloon inflations and stent deployment are clearly cephalad to the carotid bifurcation, there may be no need to discontinue these medications. In these patients, postprocedural blood pressure requires careful management to minimize the risk and/or consequences of cerebral hyperperfusion syndrome (discussed later).

Vascular Access

Femoral artery access is the preferred and recommended approach. Carotid interventions via brachial or radial artery approach have been described in patients with so-called hostile anatomy of the aortic arch. Direct percutaneous puncture of the carotid artery as a method of vascular access has, for the most part, been abandoned. The frequent need for general anesthesia, proximity of the access site to the site of the lesion, problems related to local hematoma including the risk of airway compromise, and difficulty in compressing a superficial stented vessel for securing hemostasis are some of the reasons why direct carotid artery catheter insertion is no longer used. Femoral venous access is unnecessary unless a reliable peripheral venous access is unavailable. Routine prophylactic placement of a temporary venous pacemaker is no longer recommended but should be readily available.

Diagnostic Angiographic Evaluation

It is mandatory to have a high-quality complete diagnostic cerebral and extracranial carotid angiogram prior to initiating the stenting procedure. Imaging of the aortic arch by angiography, MRA, or CTA may be helpful to define the arch type and anomalous origins of the vessels. The most common anomaly, seen in approximately 7% of patients, is independent origin of the left vertebral artery from the arch and origin of the left carotid artery from the innominate. Figure 32-12 shows classification of the aortic arch.

Figure 32-12 Examples of commonly used catheters for cervico-cerebral angiography: the double-curved Vitek and Simmons catheters, JR4, Berenstein and Headhunter catheter.

A complete cerebral angiogram requires anatomical definition of both intracranial and extracranial carotid arteries as well as the dominant vertebral artery. The decision to perform selective cannulation and angiography of the vertebral artery should be individualized. The vertebral arteries frequently have a tortuous course, and the vessel is prone to spasm—features that predispose to dissection with catastrophic sequelae. It is important for the operator to understand the collateral circulation to the brain hemisphere of interest. Clear definition of the collateral circulation is especially important if a flow-arresting balloon occlusion–type EPD is being considered. Interrupting antegrade flow in the absence of good collateral circulation may cause the patient to have a seizure and require rapid premature deflation of the occlusion balloon. Interruption or even termination of the procedure under these circumstances increases the risk of periprocedural complications.

Catheter selection for diagnostic angiography

A variety of catheters are available for diagnostic cerebral angiography, and selection is tied to operator familiarity and experience. Examples of diagnostic catheters are shown in Figure 32-13. It should be understood that catheters that require additional manipulations to reshape them within the ascending aorta increase the risk of embolization. Use of such catheters should be reserved for negotiating the difficult aortic arch anatomy (e.g., patients with extended, stiff, calcified aortas [see Fig. 32-12, type III arch]) when use of alternative preshaped catheters that require less manipulation have either failed or are expected to fail.

Figure 32-13 North American Symptomatic Carotid Endarterectomy Trial (NASCET) criteria for determining degree of carotid artery stenosis.

Luminal diameter at site of greatest narrowing is recorded in three planes and used as the numerator (a). A reference diameter is taken across a plaque-free section of internal carotid artery distal to stenosis (b) and is used as the denominator. A percentage stenosis is then calculated.

(From North American Symptomatic Carotid Endarterectomy Trial. Methods, patient characteristics, and progress. Stroke 22:711–720, 1991.)

Diagnostic angiography involves injection of 2 to 3 mL of nonionic contrast diluted with an equal amount of saline. Immediately prior to acquisition of the subtraction angiogram, patients are asked not to breathe, move, or swallow to minimize motion artifact. They are also warned that they may experience a funny taste and may see flashing or multicolored lights in the ipsilateral eye.

Diagnostic angiography consists of visualization of the origins of the innominate and left common carotid arteries from the aortic arch (by selective injections), both carotid bifurcations in orthogonal projections, and both vertebral arteries (usually by nonselective injections). Intracranial images of both carotid arteries are routinely acquired, and occasionally selective injection of one or both vertebral arteries is also performed. Brachiocephalic angiography has several advantages:

1. It is a reliable, reproducible method for precisely measuring the degree of carotid artery stenosis (see Fig. 32-1).

2. It demonstrates anatomical conditions that can be unfavorable for carotid stenting. Examples include dilated/extended aortic arch (see Fig. 32-12), marked vessel tortuosity, heavily calcified stenosis, and lesions with obvious filling defects (see Figs. 32-3 through 32-11).

3. It helps define the status of collateral circulation to the ipsilateral cerebral hemisphere (i.e., the one supplied by the stenotic carotid artery being evaluated for treatment). Knowledge of contralateral carotid stenosis or occlusion and status of the collateral supply influences the stenting technique: shorter balloon inflations, for example, and choice of protection device—flow interrupting (occlusion balloon) vs. flow preserving (filter devices). The term isolated hemisphere describes the anatomical situation where the cerebral hemisphere of interest is entirely dependent on the ipsilateral ICA for its blood supply, owing to absence of the anterior and posterior communicating arteries.

4. It reliably demonstrates significant flow-limiting stenosis distal to the carotid bifurcation. Although the bifurcation stenosis may be treatable, the ultimate benefit of stroke reduction may not accrue to the patient because of additional cephalad disease.

5. In the event there is an intraprocedural neurological event, the postevent intracranial angiograms can be compared with the baseline preprocedure pictures.

The main risks of invasive cerebral angiography relate to the use of contrast and the possibility of a neurological event. The typical sequence of acquisition and the usual angiographic views are listed in Box 32-3.

Box 32-3 Overview of Carotid Angiography Acquisition Views

Left Subclavian Angiogram (PA View)

The ostium of the left vertebral artery is usually seen well in the frontal projection; if not, the RAO-cranial projection should be tried.

If a selective angiogram of the left vertebral artery is to be acquired, a roadmap to facilitate entry and selective placement of the catheter is recommended.

Selective or nonselective intracranial views of the vertebrobasilar system are acquired in the lateral and steep AP cranial views.

Selective Left Carotid Angiograms

The bifurcation is imaged in LAO 45-degree as well as lateral projections. If the bifurcation is “overrotated,” an AP caudal view usually separates the external and internal carotid arteries very well.

Intracranial images of the left carotid artery are acquired in the lateral and AP cranial (15- to 30-degree) views.

Innominate Angiogram in the RAO Caudal View

Best for separating the carotid and the right subclavian arteries and hence very helpful for defining lesions involving the origin of the right subclavian artery.

Selective Right Carotid Angiograms

Bifurcation is imaged in the RAO 45-degree as well as lateral projections.

Intracranial images of the right carotid artery are acquired in the lateral and AP cranial (15- to 30-degree) views.

AP, anteroposterior; LAO, left anterior oblique; PA, posteroanterior; RAO, right anterior oblique.

Carotid Sheath Placement

Once the diagnostic study is completed, the ICA with the target stenosis is identified, and there are no anatomical contraindications for stenting, a 90-cm long, 6 F sheath is advanced into the CCA using one of two techniques shown in Figure 32-14.

Figure 32-14 Carotid sheath placement.

A, In the two-step approach using a diagnostic catheter, a Glidewire (Boston Scientific, Watertown, Mass.) is inserted into external carotid artery (ECA). Catheter is advanced to the ECA, and the Glidewire exchanged for a support wire such as a Supra Core (Abbott Vascular, Santa Clara, Calif.) wire. Catheter is then removed, and sheath with introducer is advanced over support wire to common carotid artery (CCA). Finally the introducer and wire are removed. B, In the single-step or telescoping approach, sheath is placed in descending aorta. A catheter is used to place a Glidewire in the ECA, and the catheter is advanced to the CCA. Glidewire and catheter are fixed in place and the sheath advanced over the two. With sheath in place, Glidewire and catheter are then removed.

Embolic Protection Devices

Although not all steps of the carotid intervention can be “emboli protected” because placement of the sheath occurs prior to deployment of the EPD. It is important to recognize, however, that the risk of embolization is highest during stent deployment and balloon dilation.⁶⁴

A number of EPDs are available and can be broadly divided into flow-interrupting (occlusive) and flow-preserving (nonocclusive) devices (Fig. 32-15). Occlusion EPDs can be subdivided into distal occlusion and proximal occlusion types. An example of a flow-interrupting EPD that is deployed distal to the stenosis is the Percusurge GuardWire (Medtronic, Minneapolis, Minn.). Proximal occlusion devices include those developed by Gore (Flagstaff, Ariz.) and the MoMa device (Medtronic, Minneapolis, Minn.). Although flow-interrupting occlusive-type EPDs are intuitively appealing (no blood flow = no emboli), some issues with their use have arisen. For example, to permit occlusion and interruption of antegrade flow in the ipsilateral carotid artery, it is mandatory to demonstrate robust collateral circulation to the hemisphere being treated. Infrequent use of the GuardWire device, as well as occasional device failure (inadequate balloon inflation and suboptimal seal of the carotid artery and/or premature deflation), has resulted in virtual abandonment of this device for carotid intervention. The rationale for use of proximal occlusion devices is based not only on interruption of antegrade flow but also flow reversal in the ipsilateral ICA. This combination is achieved by balloon occlusion of both the CCA and ECA. Although procedural outcomes using these devices are favorable,⁶⁵,⁶⁶ they are more in the category of a niche device, since the catheters are substantially larger. They are particularly useful in cases with tortuous distal ICA anatomies or lesions with filling defects that may embolize during wire passage prior to deployment of the distal EPD.

Figure 32-15 Correct placement of embolic protection device.

Device is deployed cephalad to stenosis in a straight segment of the internal carotid artery (ICA) to ensure all antegrade flow passes through, not around, the filter. It must also be sufficiently distal to lesion to enable stent deployment and expansion. Filters must not be placed on curves of the ICA.

The nonocclusive flow-preserving devices are the filters. Depending on construction, the filter body can be supported with a nitinol frame or be unsupported and resemble a windsock. The underlying principles and rationale for use are the same for all filter EPDs: preserving antegrade flow while preventing passage of embolic debris. As may be expected, each of these EPDs has certain unique features, and familiarity with their construction and attributes will allow judicious selection of the device most appropriate for use with a particular anatomy.

Embolic protection device deployment

The device has to be placed approximately 2 cm cephalad to the stenosis to provide adequate space (the “landing zone” for the EPD) to accommodate the tip of the stent delivery system and allow satisfactory coverage of the lesion with the stent (Fig. 32-16).

Figure 32-16 Examples of embolic protection devices.

Clockwise from Top Left, MoMa proximal occlusion device (Medtronic, Minneapolis, Minn.) and Gore flow-reversal system (Gore, Flagstaff, Ariz.). Nonocclusive filters are the Accunet and Emboshield (both from Abbott Vascular, Santa Clara, Calif.).

For severe subocclusive stenosis, “pre-predilation” using a low-profile coronary angioplasty balloon (2 or 2.5 mm diameter) is very helpful. Although this type of dilation is not protected, the risk of this approach should be counterbalanced with the problems of trying to force an EPD through a very severe stenosis. The tip of the wire relative to the EPD should be angiographically visible throughout all steps of the procedure, and the operator should ensure that the wire and EPD are stable with minimal or no movement, since excessive movement can lead to spasm and slow flow.

Lesion Predilation

Lesion predilation should be considered the default strategy because:

Use of a 3 or 3.5 mm × 30 mm coronary balloon inflated to nominal pressure is recommended for predilation. The markers on the ends of the 30-mm-long balloon help the operator select the length of the stent (30 or 40 mm). Predilation is generally brief, and gradual deflation is recommended. On rare occasion when treating a heavily calcified lesion, the stent may not pass easily through the stenosis, despite adequate predilation. In this setting, a larger 4- or 5-mm balloon may be needed for additional predilations. Following predilation and immediately prior to stenting, an arteriogram is performed to once again establish the relationship of the stenosis to the bony landmarks.

Stents

To eliminate the risk of deformation and crushing²³ seen with balloon expandable stents, self-expanding stents are exclusively used for carotid stenting, with the following notable exceptions:

Although the Elgiloy (a cobalt chromium alloy) tracheobronchial Wallstent, (Boston Scientific, Natick, Mass.) was initially used, most operators currently use self-expanding nitinol stents. The Wallstent, an example of a closed-cell stent (see later discussion), fell out of favor largely because of its unpredictable foreshortening. This feature makes precise positioning of the stent very difficult. Despite structural and technical differences between the Elgiloy Wallstent and nitinol stents, there was no significant difference in carotid stenting outcomes using the Wallstent approved for use in the carotid circulation on the basis of the Boston Scientific EPI: A Carotid Stenting Trial for High-Risk Surgical Patients (BEACH) trial.⁶⁸

The unconstrained diameter of the self-expanding stent should be at least 1 or 2 mm larger than the largest-diameter segment intended to be covered by the stent, almost always the CCA. For example, our preference in most cases is to use an 8- to 10-mm tapered stent; an 8-mm-diameter stent segment is deployed in the ICA, while the wider 10-mm end of the stent extends into the CCA. The stent is deployed slowly to minimize its tendency to jump forward. It is important to ensure that the caudal end of the stent is firmly anchored in the CCA. If not, the angulated proximal edge of the stent can cause difficulties in recrossing the stent with either the post-dilation balloon or the EPD retrieval sheath.

Tips in stent selection, deployment, and positioning include:

Figure 32-17 Comparison of open- and closed-cell stent designs.

Operators should have one of each type available and be familiar with its use, enabling the correct choice to be made in each individual case.

Figure 32-18 Angulation of cranial segment of internal carotid artery (ICA) due to catheter introduction and stent implantation.

First panel demonstrates acute angulation of ICA distal to stenosis, with a 5 F diagnostic catheter in position. Introduction of a 6 F guide sheath further exacerbates this and induces a stenosis at the bend. Following stent implantation (open-cell) in third panel and almost complete withdrawal of sheath, stenosis and angulation significantly resolve. This lesion was not treated. Panel 4 shows same lesion at angiography 3 months later when patient returned for contralateral carotid artery stenting.

Post-dilation

The size of the post-dilation balloon is matched to the diameter of the ICA at the site of the stenosis. Typically, the self-expanding stent is post-dilated with a 5-mm, 0.014-inch wire-compatible balloon. On occasion, high-pressure balloons may be needed to postdilate a stent deployed within a heavily calcified stenosis.

Post-dilation of the stent is a critical step:

Final Angiographic Assessment

Embolic Protection Device and Sheath Removal and Access Site Hemostasis

To be effective as an emboli trapping device, the size of the pores in the fabric of the filter have to be microns in diameter. Passage of blood through these small pores stimulates the deposition of fibrin within the filter, providing the perfect conditions for formation of a thrombus within the filter. The longer the dwell time, the greater the chances of formation of an iatrogenic thrombus. When using filter EPDs, it is important to minimize both the number of contrast injections and the in vivo dwell time; median time should not exceed 7 minutes, and consistent breach of this time limit is a quality concern. If the proximal edge of the stent is not well opposed to the wall of the CCA, either as a result of stent undersizing or improper positioning, it may be difficult to advance the EPD retrieval sheath (or postdilation balloons) through the stent. The best ways to overcome this difficulty are to (1) ask the patient to turn the head to one side or (2) advance the sheath into the stent. Some have recommended that only the proximal part of the filter should be captured, since pulling the filter completely into the capture sheath may squeeze out the emboli. Following withdrawal of the EPD, the Tuohy-Bourst adaptor should be fully opened and allowed to bleed back generously in case particles and atheromatous debris have been released into the guide catheter during withdrawal of the EPD device.

At the completion of the procedure, the sheath should be gently pulled back and out of the carotid artery and exchanged for a short sheath, which can be removed when hemostasis can be safely achieved. The vagal response to manual sheath removal and compression can compound the baroreflex effect of carotid stenting and lead to profound hypotension and bradycardia. In most patients, access-site hemostasis can be obtained using closure devices, although no improvement in outcomes has been demonstrated.⁷⁴ Finally, although low blood pressure is not unusual in the immediate postprocedure phase, other causes (e.g., retroperitoneal hemorrhage related to access-site bleeding) should be considered as reasons for unexplained persistent hypotension.

Patients are monitored in a telemetry bed overnight, and more than 95% will be ready for discharge approximately 24 hours following the procedure. All patients should have a postprocedural neurological exam and assessment of the NIH and Rankin Stroke Scales. Once patients are able to ambulate without difficulty, they can be discharged with instructions to return for follow-up in 4 weeks (Box 32-4).

Management of Hemodynamics

The carotid sinus baroreceptor, innervated by the glossopharyngeal nerve (cranial nerve IX), is located in the adventitia of the carotid artery and is activated by stretching of the blood vessel wall. The physiological role of this mechanoreceptor is to detect, respond, and regulate the systemic blood pressure. Stimulation of the carotid sinus baroreceptor (via stretching) stimulates the glossopharyngeal nerve, which in turn stimulates the vasomotor center in the medulla via the tractus solitarius, which then modulates the activity of the autonomic nervous system (suppresses sympathetic and stimulates parasympathetic output). The net effect of stimulation of the carotid baroreceptor is bradycardia (negative chronotropy), as well as some reduction in cardiac contractility. Additionally, the decreased sympathetic output results in vasodilation, causing hypotension. As might be expected, carotid sinus stimulation and the resultant hemodynamic perturbations are most profound when the lesion and the stretch from the postdilation balloon involve the carotid bifurcation, less so if the stenosis is cephalad to the bifurcation. These changes are not seen (or occur with much less intensity and frequency) in the postendarterectomy patient whose carotid bulb is denervated. The bradycardia and hypotension can be profound and are worse in younger patients and when treating highly calcified stenosis (where higher pressures are used for postdilating the stent). In anticipation of these hemodynamic changes:

Close attention should be paid to both intra- and post-procedure blood pressure management. The hemodynamic effect of the atropine and α-agonists used during the procedure is short-lived. In the postprocedure phase, hypotension with systolic blood pressures as low as 70 to 80 mmHg is usually asymptomatic and may be managed expectantly. The systolic blood pressure will typically increase by 15 to 30 mmHg by next morning. Hypotension usually presents a problem in two settings: (1) patients with a periprocedural embolic event or a severe contralateral carotid stenosis can become symptomatic, and (2) patients with baseline renal insufficiency may have worsening kidney function as a result of the combination of low blood pressure and contrast exposure.

The risk of hyperperfusion syndrome (discussed later) is increased if blood pressure remains elevated following dilation (e.g., in the post-endarterectomy patient with severe post-CEA restenosis).

Cerebral Hyperperfusion Syndrome

Cerebral hyperperfusion syndrome (CHS) is a rare but serious complication of carotid revascularization. Following successful carotid revascularization, there is an increase in ipsilateral blood flow to the affected cerebral hemisphere. Most often the patient remains asymptomatic⁷⁵; however, in some patients this increase can overwhelm normal compensatory mechanisms, resulting in a marked increase in flow. Cerebral hyperperfusion is defined as an increase in blood flow of greater than 100% from baseline, and CHS is hyperperfusion associated with neurological deficit. Estimated incidence varies according to methodology and definitions, but most published studies report a rate less than 3%.⁷⁶

The pathophysiology of CHS is not completely understood but is most likely due to a combination of factors. Patients with chronic cerebral hypoperfusion have maximal dilation of the intracranial arterioles, and normal autoregulation may not be restored for several days or weeks following revascularization. Impaired autoregulation refers to failure of the brain at the microcirculatory level to modulate blood flow and blood pressure such that sudden increase in flow and pressure is not transmitted to small blood vessels. Rather, pressure and flow are maintained within a narrow range. Impaired autoregulation of small-vessel cerebral blood flow (CBF) has been implicated in CHS in experimental models. This is either due to endothelial dysfunction resulting from free radical accumulation,⁷⁷ or neurogenic failure of smooth-muscle regulation.⁷⁸ Failure of the normal baroreceptor reflex is another possibility, with uncompensated postprocedural hypertension contributing to hyperperfusion.⁷⁹ Impairment of the trigeminovascular reflex may also contribute because it plays a role in maintaining normal cerebrovascular tone.⁸⁰ Hyperperfusion results in fluid entering the interstitial space and causing edema, predominantly in the posterior circulation.

Those with severe (>90%) subocclusive stenosis, with limited collateral supply (isolated hemisphere), are most at risk for hyperperfusion.⁸¹ Baseline severe hypertension and persistence of increased blood pressure in the postprocedure period (not uncommon in patients postendarterectomy) can result in hyperperfusion lasting several days following a procedure.⁸¹

The most common clinical symptom of cerebral hyperperfusion is headache. Not uncommonly, patients report this symptom on the table soon after the procedure is completed. The headache may last a few days, is typically unilateral, and is associated with a nonfocal neurological exam. A more serious presentation, possibly related to cerebral edema, is seizure. Patients with CHS can present with a variety of neurological deficits, including isolated speech disturbance. In patients who develop neurological symptoms following carotid revascularization, other etiologies must be considered in the differential diagnosis. Differentiation must be made between procedure-related embolic complications with an ischemic infarct, manifestation of a low-flow state, and a high-flow state characteristic of CHS. The diagnosis of CHS is best confirmed with MRI.⁷⁶

Treatment involves lowering blood pressure, reduction in cerebral edema, and anticonvulsant therapy. Because vasorelaxation may further increase cerebral blood flow, calcium antagonists and nitrates are contraindicated in the treatment of CHS. Hence, the operator must resist the urge to administer nitrates to relieve spasm (e.g., at the site of deployment of the EPD noted on the postprocedure angiogram); it will resolve gradually with time. Although nitrates will rapidly resolve spasm, they can also predispose to development of hyperperfusion. Owing to their ability to reduce cerebral perfusion pressure, labetalol and clonidine⁷⁶,⁸² are the preferred agents. Duration of therapy is not well defined; treatment may be needed for several days or months post procedure.⁸³ There is insufficient evidence to support the treatment of cerebral edema, although corticosteroids and barbiturates have been used to good effect. Prophylactic anticonvulsant treatment is not recommended but should be introduced if seizures occur.⁷⁶

The relative infrequency of this condition and the heterogeneity of presentations make it hard to generalize outcomes following CHS. However, the limited data from the CEA literature suggest that a third of patients may remain disabled following CHS,⁸⁴ with others reporting mortality rates of up to 50%.⁸⁵ Therefore, although rare, the diagnosis should be considered in all patients who develop neurological symptoms following carotid revascularization.

Post-approval Registries

As part of the condition for marketing approval of its devices (stents and EPDs), the sponsor, Abbott Vascular, agreed to perform FDA-mandated post-approval studies to assess the occurrence of rare and unanticipated device-related events. These post-approval studies have also provided an opportunity to assess the outcomes of carotid stenting in high–CEA risk patients in the non-trial setting (real world). In 2009, Gray et al.⁷³ reported the outcomes of two such prospective multicenter (280 U.S. sites, 672 operators) post-market surveillance studies involving CAS with distal embolic protection using filters in high–CEA risk patients. These studies had pre- and postprocedure neurological evaluation and independent adjudication of neurological events. Results of these two studies, as well as the outcomes of an earlier large (n = 4225) post-approval study (CAPTURE),¹⁰² are summarized in Table 32-6.

In summary, key outcomes from the CAPTURE 2 and EXACT data sets⁷³ are:

Operator Experience

In a recent publication, the CAPTURE 2 study investigators¹⁰⁵ analyzed the carotid stenting outcomes of 3388 patients (representing 64% of the total number of patients) to determine whether physician or site-related variables affected outcomes of CAS. Symptomatic patients and patients older than 80 years of age (two known predictors of adverse outcomes) were excluded. During a 3-year interval between March 2006 and January 2009, 459 operators treated the study population in 180 U.S. hospitals. The composite rates of death, stroke, and MI, and the composite rates of death and stroke at 30 days were 3.5% and 3.3%, respectively, for the full CAPTURE 2 study cohort and 2.9% and 2.7%, respectively, for the asymptomatic nonoctogenarian subgroup.

Two thirds of the sites (118 of 180 [66%]) had no death or stroke events. Within the remaining sites, an inverse relationship between adverse event rates and hospital patient volume as well as individual operator volume was observed. The death and stroke rates trended lower for interventional cardiologists compared with other specialties. Similar conclusions were drawn from a German registry analysis¹⁰⁶ and a recent meta-analysis of published studies.¹⁰⁷ The CAPTURE 2 study authors concluded that both site and operator volume were the most important determinants of perioperative events, and defined a threshold of 72 cases for consistently achieving a 30-day death and stroke rate of less than 3%. This 72 case number is higher than the entry threshold for operators participating in randomized trials like CREST or ACT I (Carotid Stenting vs. Surgery of Severe Carotid Artery Disease and Stroke Prevention in Asymptomatic Patients) trials.

Analysis and Critique

The analysis and critique that follows should be interpreted bearing in mind the limitations imposed by subset analysis and small numbers of patients within these subsets.

With the exception of SAPPHIRE, all other studies dealing with carotid stenting in high–CEA risk patients were registry studies. Despite the registry label, it should be noted that the high-risk CAS registries were performed under FDA-scrutinized clinical protocols, with prospective data collection and event adjudication. The importance of prospective independent clinical review was demonstrated by Rothwell and Warlow¹⁰⁸ investigating adverse event rates following CEA. They found a threefold difference in neurological events between operator self-reported and independent neurologist-assessed events. Both NASCET and ACAS used such prospective independent evaluations.

Individually and collectively, a large number of patients, mainly asymptomatic, have been studied within the context of these registry trials and provide a robust data set for analysis of high–CEA risk patients undergoing carotid stenting in the United States. Cumulatively, a total of more than 10,000 patients were included and analyzed in the three postmarketing studies (90% asymptomatic), and analysis of the data has helped provide answers to important questions concerning carotid stenting in a real-world setting.

Carotid stenting outcomes have shown a steady and continuous improvement since the initial introduction of these devices in U.S. trials (Fig. 32-19). For example, compared with outcomes in the ARCHeR and the original CAPTURE registries, the combined results of the CAPTURE 2 and EXACT studies showed improvement (30-day death and stroke: 6.9% vs. 5.7% vs. 3.6%, respectively), and these improved outcomes meet AHA guidelines in the younger than 80 years age group. As will be discussed later, in the CREST study, the majority of neurological events were minor strokes, with substantial if not complete resolution of the neurological deficits during the follow-up period.

Figure 32-19 Published adverse event rates for carotid artery stent studies and registries over the past 10 years.

Note progressive decline in events over time. The year in brackets is the time of last patient enrollment.

Several factors contributed to the improvement in the results of carotid stenting. First is recognition of the clinical features that define the high–stent risk patient. This was not appreciated at the time when these registry studies were initiated. As operators became more experienced, these high–stent risk patients were excluded from studies initiated later in the decade with corresponding better outcomes.

Second, the pool of qualified experienced operators has expanded with time, with corresponding improvement in outcomes. Moreover, there have been improvements in the devices. For example, the profile of the EPDs has become smaller, and a variety of nitinol stents specifically for use at the carotid bifurcation have been developed. Improvements in technology, along with minor adjustments in the procedure protocol (e.g., limiting post inflation to a single inflation), have likely contributed to improved outcomes.

As a result, CAS outcomes in non-octogenarian, high–surgical risk patients, both symptomatic and asymptomatic, are within the acceptable thresholds of AHA standards. Additionally, in patients deemed high CEA risk due to unfavorable anatomical features, the outcomes meet AHA guidelines in both symptomatic and asymptomatic patients irrespective of age.

The higher event rates in the early (pre-2005) high risk CEA carotid stent registries (i.e., event rates that breached the AHA thresholds) is one likely explanation for lack of Centers for Medicare and Medicaid Services (CMS) reimbursement for asymptomatic patients undergoing carotid stenting, despite FDA approval for this indication in 2004. The outcomes of the CAPTURE 2 and the EXACT post-marketing registries are compelling because at least in the non-octogenarian high–CEA risk population, both symptomatic and asymptomatic (Table 32-7), the death and stroke outcomes were within the AHA guidelines.

The octogenarian and older population continues to be a challenge, and the decision to recommend and perform carotid stenting, especially in the asymptomatic patient older than 80 years of age, has to be individualized.

Analysis and Critique

Minor Strokes

All interventions for extracranial carotid artery disease, CEA or CAS, are prophylactic, and the specific goal of the procedure is to reduce the patient’s future risk of a stroke. Although survival free of a major stroke is the principal goal of the therapeutic intervention, minor strokes cannot be ignored. In fact, especially when treating asymptomatic patients, operators should have an extremely low tolerance for any procedure-related neurological event, be it major or minor stroke or cranial nerve injury. In the CREST study, an increased incidence of minor strokes contributed to the excess stroke hazard in the CAS arm. Analysis of the NIH Stroke Scale data of patients suffering a peri-procedural minor stroke reveals that although residual defects were disproportionately higher in the CAS arm at 1 month (1.10% vs. 0.60%), this difference was no longer evident at the 6-month time point (0.6% vs. 0.6%). A similar trend to equalization was noted when objective classification of the residual deficits was performed using the Rankin Scale. Importantly, the occurrence of a minor stroke did not negatively impact the patient’s long-term survival. Thus, although there were more minor strokes in the CAS arm, the neurological impairment was minimal, and by 6 months both groups had similar outcomes.

Myocardial Infarction

In CREST, MI was defined by biomarker elevation (creatine kinase [CK]-MB or troponin > twice the upper limit of normal) plus either chest pain or electrocardiographic (ECG) evidence of ischemia (>1 mm ST elevation or depression in two contiguous leads). An additional pre-specified category included biomarker elevation without chest pain or ECG abnormality (biomarker positive only). When compared to patients without biomarker elevation, mortality was higher over 4 years for those with MI (HR 3.40; 95% CI, 1.67-6.92) or biomarker positive only (HR 3.57; 95% CI, 1.46-8.68). After adjustment of baseline risk factors, the occurrence of MI or elevation of biomarkers only remained independently associated with increased mortality.¹¹⁷ In EVA-3-S, SPACE, and ICSS, cardiac biomarkers were not measured as part of the protocol.

The negative impact on survival in patients experiencing a periprocedural MI is consistent with observations from other cardiac and non-cardiovascular procedures. It has repeatedly been shown that small peri-procedural elevations of cardiac enzymes were associated with increased future mortality.^118–121 How should one interpret these findings? The occurrence of a peri-procedural MI or biomarker elevation likely serves as a marker for more extensive underlying atherosclerotic disease. None of the other studies referenced in this section, including CREST, was able to positively conclude whether or not the occurrence of this procedure-related ischemic event in some way further adds to the baseline increased mortality risk. Thus, physicians have to factor in the occurrence as well as the consequences of MI when making treatment recommendations for an individual patient—especially the asymptomatic octogenarian.

Cranial Nerve Injuries and Their Sequelae

Cranial nerve injuries, a consistent complication of CEA (≈︀5%; Table 32-10) are rarely considered when discussing outcomes. Some of these deficits can be permanent and on occasion can be the source of major morbidity. For example, in the ICSS study, two of the cranial nerve injuries (out of 857 endarterectomies) were classified as disabling—both patients required gastrostomies. In the CREST study, more than 80% of the cranial nerve injuries were motor deficits, and 2% of the cranial nerve injuries were unresolved at 6-month follow-up. In future trials, consideration should be given to the inclusion of cranial nerve injuries as part of a composite primary endpoint together with death, stroke, and MI.

^{Table 32-10} Published Cranial Nerve Injury Rates from Large Randomized Trials*

CAS, carotid artery stenting; CEA, carotid endarterectomy.

^* The 1.1% rate seen in EVA-3 S was due to three events, two from patients who crossed over to CEA (intent-to-treat [ITT] analysis).

Secondary Endpoints

Pre-specified interaction analysis showed symptom status and gender did not modify the treatment effect. However, age at treatment did have an influence, and the crossover was approximately 70 years. The impact of age on treatment selection is discussed later.

Clinical Durability of Carotid Stenting

Knowledge of the risk of an ipsilateral stroke after CAS during the follow-up period is vital to demonstrate the clinical durability of CAS. Clinical durability of CEA in both symptomatic and asymptomatic patients with extracranial carotid artery disease was established by NASCET³ and ACAS.⁸ The rates of ipsilateral stroke during the 4-year CREST follow-up period were low and similar in both treatment arms—2.0% for CAS and 2.4% for CEA. These results are similar to the rates in the EVA-3-S¹¹⁰ and SPACE trials,¹²² suggesting excellent durability for up to 4 years. The durable benefits of carotid stenting (i.e., freedom from ipsilateral stroke) in these large randomized studies reinforce the results published by the authors almost a decade ago.²² To further investigate the long-term outcomes following carotid intervention, subject follow-up in the CREST study has been extended to 10 years and will provide additional information on this issue.

Restenosis Following Carotid Stenting

Restenosis rates have been documented following CAS in the large randomized trials and are approximately 10% at 2 years,¹²²,¹²³ as measured by duplex ultrasonography. Although duplex ultrasound is an excellent method of following these patients, velocity criteria conventionally applied to diagnose stenosis in nonstented arteries may require modification because of mechanical changes in the carotid artery following stenting. It has been suggested that a peak systolic velocity greater than 300 cm/s be used to define a stenosis of 70%^124–126 in the stented patient. Other modalities also present challenges. Computed tomography angiography may be used as a complementary form of imaging,¹²⁷ but it requires contrast and radiation exposure. Magnetic resonance angiography is not useful in a stented patient.

Evaluating restenosis by tracking the need for target lesion revascularization is a commonly accepted method to define clinically relevant in-stent restenosis. At 1-year follow-up, both CAS and CEA had close to 99% freedom from target lesion revascularization,³⁹ with only 0.8% of patients requiring repeat revascularization. Furthermore, even in patients with evidence of in-stent restenosis, the event rate is low, suggesting restenosis is generally a benign condition.¹²³ These results, together with previously reported very low rates of carotid stent restenosis, support durability of the self-expanding stent for this indication. Since the stent is implanted within a superficial artery that is subject to a variety of torsional forces associated with neck movement, questions of stent fracture have been raised and are being addressed by follow-up fluoroscopy in some ongoing trials.¹²⁸ In the authors’ experience, stent thrombosis, stent fractures, and in-stent restenosis are rare issues for stents placed in the extracranial cervical carotid location. The rate of stent fracture is currently under study.

Summary of CREST Results

Carotid Stenosis Following Cervical Irradiation

Carotid artery stenosis is a delayed complication of neck irradiation (radiation arteritis).¹³⁰ Although stenosis following prior radiation is often seen at the carotid bifurcation, patients with prior neck radiation frequently have lesions that are high (at or above the C2 vertebra) or low (at or below the level of the clavicle). Stenosis at multiple locations within the same artery is not unusual. The typical interval between the radiation treatment and detection of stenosis is at least 10 years. In patients who had neck irradiation but also have multiple risk factors for atherosclerotic vascular disease and who present with an isolated bifurcation stenosis, it may be very difficult to determine the etiology of the carotid disease.

Surgical treatment of these patients is complicated by a number of factors. Many patients have had radical neck dissections in association with radiation therapy, resulting in fibrosis and scarring. There is an increased risk of stenosis of the CCA, its low position making surgical access more difficult. There is also an increase in postoperative necrosis, infections, wound breakdown, and cranial nerve palsies that have been observed post CEA in this population,¹³¹ compared to patients without prior radiation. In contrast, none of these difficulties are faced by operators performing CAS, making it an appealing technique for this indication.

Owing to the infrequent nature of this presentation, there are no randomized studies; however, several case series have demonstrated that CAS is safe and efficacious in this setting.¹³²,¹³³ Reported peri-procedural stroke and death rates have been low and have not exceeded those observed in general patient cohorts undergoing CAS. The only limitation of stenting in this group is the possible increased risk of in-stent restenosis following the procedure. Restenosis rates from 5%¹³² to 41%¹³³ have been reported at intervals from 12 to 24 months. Although higher than that seen in patients without prior radiation, these events have not been associated with symptoms and restenotic lesions are amenable to repeat percutaneous intervention.

Despite the absence of randomized trial data apart from those patients enrolled in the SAPPHIRE study, the low peri-procedural complication rates with CAS makes this the preferred approach for this patient group and the authors support the level IIa recommendation of the multispecialty guidelines³² in recommending the use of CAS for carotid stenosis following cervical radiotherapy. Patients should be closely followed by serial duplex ultrasound examinations after the procedure to monitor for restenosis. Because of problems of delayed healing and incomplete endothelialization of the stent struts, patients with prior neck radiation who undergo stenting are prescribed lifelong dual antiplatelet treatment.

Post–Carotid Endarterectomy Restenosis

Restenosis after CEA is a well-recognized phenomenon. Restenosis that occurs within the first 2 years after CEA has been attributed to intimal hyperplasia, whereas after 2 years, progressive atherosclerosis is thought to be responsible. Restenosis rates vary depending on the study and methodology and are reported from 5% at 1 year¹²³ to 36% during long-term follow-up.¹³⁴ Revision CEA is technically challenging and associated with higher complication rates than primary surgery, with increased rates of stroke (4.8% vs. 0.8%), TIA (4.0% vs. 1.1%), and cranial nerve injury (17.0% vs. 5.3%).¹³⁵ Carotid artery stenting is therefore an appealing technique in this setting.

There have been no randomized trials to assess which method is superior. Single-center studies evaluating the experience of CAS following prior CEA have demonstrated that low complication rates can be achieved. Thirty-day death/stroke rates ranging from 1%¹³⁶ to 4%¹³⁷ have been published and have compared favorably to endarterectomy.¹³⁶ Of note, these published figures are slightly superior to those seen in randomized trials of symptomatic patients.³⁹ The durability of CAS after previous CEA has not been well established; however, the limited data suggest that stent restenosis is not significantly higher than that seen in primary CAS.¹³⁷

Given that the current literature demonstrates no particular increase in CAS risk following prior CEA but a significant increase in complications during revision CEA, CAS is the recommended approach in this setting.

Carotid Stenting in Women

Some authors have postulated a variance in outcomes dependent on the sex of the patient. Women typically have smaller-caliber arteries and aortic arches, potentially leading to increased technical difficulties and higher complication rates. Although individual trials have not had sufficient power to detect such differences, the Carotid Stenting Trialists Collaboration (CSTC) published a meta-analysis of three randomized controlled trials. This study revealed no significant difference in outcomes at 4 years between men and women, although there was a trend toward higher adverse event rates in men.¹³⁸ A separate review of the CREST study data found that women are more likely to have complications following CAS, and men more likely after CEA.¹³⁹ The authors suggested that the treatment strategy for carotid stenoses could be customized, with a preference for CEA in women and vice versa. If the data from this study are further pooled with the data from CSTC, there is no significant difference in outcomes.¹⁴⁰ Given the conflicting study data and absence of clear evidence, there is currently no basis for tailoring treatment approaches on the basis of gender.

Carotid Stenting in the Elderly

Elderly patients (generally interpreted to mean patients > 80 years) with carotid stenosis pose several challenges for carotid revascularization. The association between older age and increased risk of adverse events after CAS was seen in the CREST lead-in cohort,¹⁴¹ the SPACE trial,⁴⁰ and the ICSS.¹⁰⁹ Indeed, octogenarian patients are more likely to be excluded because they have one or more features deemed high risk for carotid stenting, as shown in Box 32-1. By carefully excluding patients with adverse features—decreased cerebral reserve, excessive tortuosity (2 bends >90 degrees after the origin of the CCA from the aortic arch or ICA from the carotid bifurcation), or heavy concentric calcification—it has been shown that adverse event rates comparable to those seen following CEA¹⁴² can be achieved.

In the CREST study, an interaction between age and treatment efficacy was detected with a crossover at an age of approximately 70 years. This led the CREST investigators to conclude that “carotid-artery stenting tended to show greater efficacy at younger ages, and CEA at older ages.”³⁹ Hence, treatment recommendation in the elderly patient with carotid stenosis must be highly individualized and take into account high-risk features that will make stenting or CEA (or both options) unsuitable.