Trials and Frontiers in Carotid Endarterectomy and Stenting

I. Carotid Revascularization: Trials and Guidelines

Based on the strength of randomized trials from the 1990s, major societal guidelines recommend carotid endarterectomy (CEA) for severe (≥ 70%), symptomatic carotid stenosis if an operative stroke/death rate of <6% can be maintained (“History and major trials in carotid revascularization” are summarized in the Supplemental File, please see http://stroke.ahajournals.org),^1-4 Though the benefit is less evident, most guidelines also recommend CEA be considered for 50-69% symptomatic stenosis.^2-4

There are subtle differences in recommendations regarding carotid artery stenting (CAS) in symptomatic patients. Some guidelines stipulate that CEA should be preferred over CAS in patients with severe (≥70%) symptomatic carotid stenosis^{2, 5}, especially if > 70 years old⁴, whereas others position CAS as an “alternative”.^{1, 3} Though the risk of operative stroke/death is higher with CAS, major RCTs report event rates under the recommended 6% cutoff for both treatment modalities.

Regarding asymptomatic disease, CEA is recommended for patients with stenosis ≥ 60-70% in “highly selected” patients as long as operative stroke/death rates <3% can be maintained.¹ A predicted life expectancy of at least 3-5 years has also been suggested.² The 3% threshold has been easily met by CEA cohorts in the Carotid Revascularization Endarterectomy versus Stenting Trial (CREST) (1.4%)⁶ and the Asymptomatic Carotid Trial (ACT1) (1.7%)⁷, suggesting that an even lower threshold may be appropriate.

Controversially, some guidelines have recommended that CAS can be considered in “highly selected” patients with asymptomatic carotid stenosis ≥ 60-70%^{1, 4, 8}, whereas others argue that the evidence remains insufficient.² The lack of consensus in the management of asymptomatic carotid stenosis is reflective of an ongoing need for high quality randomized clinical trial data to guide practice.

II. Operative Risk

a. CEA Operative Stroke Risk and “High Risk” Designation

Most clinical trials in carotid revascularization have focused on “average” operative risk patients, excluding patients with anatomic risk factors, major comorbidities or advanced age. Many patients have also been excluded from these studies if they had neurologic dysfunction or dementia that would limit stroke assessment, or other common conditions that carry stroke risk such as atrial fibrillation.^{7, 9} However, clinicians frequently must make treatment decisions for patients who would have been excluded from these trials due to presumed excessive operative risk, thereby limiting the generalizability of RCT findings.

The “high risk” designation (Table 1) was adopted by CMS from CAS trials such as the Stenting and Angioplasty with Protection in Patients at High Risk for Endarterectomy (SAPPHIRE) trial, attempting to define a population at increased risk for complications with CEA who might, therefore, be better suited for CAS. A review of the Society for Vascular Surgery’s Vascular Registry showed that the “high risk” classification is successful in identifying a group of CEA patients at increased risk of operative stroke/death/MI in symptomatic (7.3% in “high risk” vs 4.6% in normal risk, P<0.01) and asymptomatic patients (5.0% in “high risk” vs 2.2% in normal risk, P<0.001).¹⁰

Table 1.

Centers for Medicare and Medicaid Services “High Risk” for CEA criteria^*

Anatomic High Risk

Tandem stenosis > 70%

Bilateral carotid stenosis

Contralateral carotid occlusion

Recurrent carotid stenosis

Prior radiation therapy or radical neck dissection

High (above C2) or low (below clavicle) carotid bifurcation

Physiologic High Risk

Age ≥ 75 years

NYHA CHF Class III/IV

Left ventricular ejection fraction (LVEF) < 30%

Coronary artery disease involving ≥ 2 vessels

Unstable angina

Recent MI (within 6 weeks)

Severe pulmonary disease

Chronic renal insufficiency

^*Including “other conditions that were used to determine patients at high risk for CEA in prior carotid artery stenting trials and studies”

However, it has not been demonstrated that “high risk” patients do better with CAS than CEA. Registry and institutional data suggest that in asymptomatic “high risk” patients, CEA and CAS have similar stroke/death/MI rates and in symptomatic “high risk” patients, CAS carries higher risk.^{11, 12}

Apart from “high risk” designation, reviews of various databases have led to identification of various risk factors that predict poor outcomes after CEA.^13-15 Despite this work, there are no reliable, validated prediction models for determining high operative risk that can currently assist in choosing between revascularization strategies. In general, high anatomic risk patients may be more appropriate for CAS² as are patients with severe comorbid cardiac disease, based on excess operative cardiac risk that may be mitigated with CAS.¹⁶ Furthermore, true high risk patients (such as the largely asymptomatic cohort of high-risk patients studied in SAPPHIRE) may do better with intensive medical therapy alone, especially if operative stroke/death rates are anticipated to be higher than recommended safety thresholds (6% for symptomatic patients and 3% for asymptomatic patients).

b. CAS Operative Stroke Risk

i. High risk anatomy

To tailor treatment strategy to individual patients, major interest has developed in identifying whether any carotid lesion characteristics place a patient at high risk for operative stroke with CAS. Various lesion-related and procedure-related risk factors have been described which may increase the CAS-related risk of operative stroke (Table 2, Figure 1), many of which have been identified on secondary analyses of major CAS trials.^17-19 It is likely that increased prevalence of these high risk features is at least partially responsible for worse CAS outcomes in elderly patients (Table 3). Patients ≥ 80 years are more likely to have several of the lesion- and procedure-related anatomic characteristics which make CAS more difficult and/or are associated with higher stroke risk.^20-23

Table 2.

Factors potentially increasing risk of stroke with CAS

Access-related

Arch calcification

Arch type II or III

Tandem lesion in CCA or innominate

ICA-CCA angulation ≥ 60°

Lesion-related

Increasing stenosis severity

Circumferential calcification

Ulcerated lesion

Ostial lesion

Lesion length > 10-15mm

Multiple stent use

Sequential lesions

Echolucent plaque (on ultrasound)

Distal ICA

Tortuosity

Diffuse atherosclerosis

Tandem lesion

Thrombus

Small caliber

Operator Characteristics

Inexperience

Lack of pre-procedure CTA/MRA

Aortic arch injection

Failure to use EPD

Predilation prior to EPD

Failure to choose correct EPD for anatomy

Figure. Favorable and unfavorable anatomy for carotid stenting.

RIGHT (unfavorable): (A) Innominate takeoff from ascending aorta, Proximal CCA tandem lesion, Tortuous CCA; (B) High-grade, calcified ICA stenosis, ICA thrombus, Angulated ICA takeoff; (C) Occluded ECA; (D) Small diameter distal ICA with disease

LEFT (favorable): (a) Left CCA takeoff from top of aortic arch, CCA free of disease and straight; (b) Moderate-grade, smooth ICA stenosis with minimal calcification, Straight ICA takeoff; (c) Patent ECA; (d) Moderate to large diameter distal ICA with no disease.

Adapted with permission from: Schermerhorn M. Mastery of Surgery. 5^th ed. Philadelphia: Wolters Kluwer Health/Lippincott Williams & Wilkins; c2007. Chapter 187, Carotid Artery Stenting.

Table 3.

High CAS risk anatomic features associated with age ≥ 80 years

Access-related

Aortic arch/supra-aortic vessel calcification

Aortic arch elongation (type II/III arch)

Tandem lesion in CCA or innominate

ICA-CCA tortuosity

Lesion-related

Severe stenosis > 85%

Heavily calcified stenosis

Ulcerated lesion

In order to recognize these features and guide treatment choice, preoperative CT angiography of the neck (including the aortic arch) is recommended for patients being considered for CAS. This helps to determine the best means of accessing the lesion, any disadvantageous lesion characteristics, distal ICA anatomy and embolic protection device (EPD) selection. MR angiography would be an acceptable alternative although visualization of the aortic arch is not routinely provided and calcium identification is more limited.

ii. Embolic Protection

Reviews have consistently suggested that use of EPDs at the time of CAS reduces the risk of operative stroke,¹⁹ though data from randomized trials are lacking. Accordingly, EPD use is mandated in major randomized clinical trials (CREST, CREST-2[NCT02089217]) and is required for reimbursement by CMS.

However, catheter manipulation in a highly calcified aortic arch can lead to cerebral embolization prior the deployment of an EPD, which may be a major cause of stroke associated with CAS.^24-26 Even if the arch is safely navigated, deployment of a distal EPD first requires passing the device across the lesion in an unprotected fashion, a step which also incurs risk of embolic stroke.

Trans-Carotid Artery Revascularization (TCAR) was developed whereby direct surgical exposure of the common carotid artery allows for CCA access and initiation of embolic protection via flow-reversal, with blood return via the femoral vein.²⁷ This technique avoids risks associated with catheterization of the aortic arch and proximal CCA and does not necessitate crossing of the stenosis prior to initiation of embolic protection. In a mixed cohort of symptomatic (25%) and asymptomatic (75%) patients, a single-arm trial of TCAR was associated with an operative stroke/death rate of 2.8%.²⁷

iii. Stent Design

The contribution of stent design to operative stroke risk has been controversial. Some have argued that a closed cell architecture is more likely to act as a scaffold for unstable plaque and prevent embolization, while others prefer the flexibility of open cell designs for navigating severely stenotic and tortuous arteries. Analyses of large CAS databases and CAS trials have suggested that closed-cell stents incur lower operative TIA/stroke rates, compared to open-cell stents.^28-30 As a result, newer carotid stent designs emphasize small free cell areas, in some cases utilizing multi-layer or micromesh coverage.

iv. Operator Experience

For CAS, operator experience is critically important. A pooled analysis of early carotid stent trials for symptomatic carotid stenosis showed that operators with low (mean ≤ 3.2 procedures/year) or intermediate (mean 3.2-5.6 procedures/year) in-trial case volume had 10.1% and 8.4% risk of operative stroke/death, respectively. High volume operators (>5.6 procedures/year) had the lowest operative stroke/death rate at 5.1%.³¹ In CREST, physicians underwent rigorous credentialing with case review and participation in a lead-in phase prior to enrolling patients to mitigate the effect of operator experience.³²

Furthermore, there is reason to believe that CAS outcomes are not as good in the community where operator volume is lower. In a study of > 20,000 Medicare patient undergoing CAS, the median annual operator volume was only 3.0/year.³³ Patients treated by operators with <6 procedures/year were found to have an elevated risk of 30-day mortality [OR 1.9 [95% CI 1.3-1.9]) compared to patients treated by operators with ≥ 24 procedures/year. While an analysis of the CAPTURE 2 CAS registry has suggested that it may take up to 72 cases for an operator to achieve an operative stroke/death rate <3% in asymptomatic patient,³⁴ decreasing case volumes may make this goal unrealistic.

c. Myocardial Infarction and Cranial Nerve Injury

In initial studies comparing CEA to CAS, rates of MI were <1% for both procedures, likely because cardiac biomarkers were not measured routinely.^35-37 The high-risk SAPPHIRE trial, which systematically collected cardiac biomarkers, was the exception, reporting MI rates of 5.9% for CEA and 2.4% for CAS.³⁸ In the average-risk group studied in CREST, MI occurred in 2.3% of patients undergoing CEA and 1.1% undergoing CAS (P=0.03).^{16, 39} In CREST, the MI endpoint required biomarker elevation plus either chest pain or ECG evidence of ischemia (biomarkers were routinely collected, but isolated biomarker elevations were not considered as MIs). As a result, the higher MI rates seen in CREST may be partially due to the detection of some asymptomatic MIs. This has led to considerable controversy regarding the inclusion of MI in the primary composite endpoint, essentially equating the clinical importance of perioperative myocardial ischemia to stroke or death. Patients who suffered an MI (as defined in CREST) had increased risk of 4-year mortality (HR 3.4 [95% CI 1.7-6.9]) which was comparable to the increased risk of 4-year mortality associated with operative stroke (HR 2.8 [95% CI 1.6-4.8]), illustrating the importance of these events.^{16, 40} However, a subsequent quality of life study showed that, in CREST, operative stroke had a greater and more sustained impact on quality of life than MI.⁴¹ Of major ongoing trials, CREST-2 and the European Carotid Surgery Trial 2 (ECST-2, ISRCTN 97744893) will not include MI as a component of their composite primary endpoints.

Cranial nerve injury (CNI) may also have meaningful clinical consequences in patients undergoing CEA. In CREST, the rate of CNI for CEA was 4.6%.⁴² However, 34% of deficits had resolved at 1-month follow-up and 81% resolved by 1 year. No difference in quality of life associated with CNI were detected at 1-year follow-up. Registry studies have confirmed that, in most cases, CNI-related symptoms resolve in follow-up.⁴³

III. Developments and Controversies in Symptomatic Carotid Stenosis

a. Timing and type of revascularization after onset of symptoms

Data from randomized trials suggests that the attributable risk reduction of stroke/death associated with revascularization is highest within 14 days of symptom onset, and diminishes thereafter.⁴⁴ As a result, most guidelines recommend revascularization within this 14 day period.^2-4 Revascularization within 48 hours of symptom onset has been associated with higher risk of neurologic complications (including hemorrhagic conversion) and is typically avoided.⁴

Interestingly, randomized trials and large database reviews have shown that CAS performed within the first 7-14 days after symptom onset is associated with high stroke/death rates, particularly when compared to CEA.^45-47 CEA is therefore preferred over CAS when revascularization is performed within 14 days of symptom onset.⁴

b. Revascularization after thrombolysis or endovascular intracranial intervention

With increasing utilization of systemic thrombolysis and endovascular intracranial interventions, the risk of revascularization may be affected. Though limited, current data suggest that CEA can be safely performed within 14 days of thrombolysis but should be avoided within the first 72 hours due to risk of intracranial hemorrhage.⁴⁸ It has been recommended that patients only be considered for early CEA after thrombolysis if 50-99% ipsilateral ICA stenosis is present and the following criteria are met: (1) rapid neurologic recovery, (2) infarct <1/3 middle cerebral artery territory, (3) previously occluded MCA mainstem has recanalized, and (4) no parenchymal hemorrhage or brain edema.⁴

There are little data to guide decisions related to revascularization after endovascular intracranial interventions (thrombectomy, intra-arterial thrombolysis) and the practice of concurrent CAS at the time of intracranial intervention are variable.

c. Role of dual antiplatelet therapy

There are compelling data that dual antiplatelet therapy (DAPT), when initiated early after symptom onset, can reduce recurrence of neurologic events after noncardioembolic TIA/Stroke.⁴⁹ Specific to carotid disease, it has been shown that initiation of DAPT in patients with recently symptomatic carotid stenosis leads to decreased transcranial Doppler-detected microembolization (which are associated with TIA/Stroke risk)⁵⁰, and recurrent neurological events.⁵¹ Early initiation of DAPT should be considered after symptom onset, though this must be weighed against bleeding risk of any planned revascularization.^{4, 52}

IV. Developments and Controversies in Asymptomatic Carotid Stenosis

a. Identifying Asymptomatic Patients with High Long-Term Stroke Risk

The most widely used estimator of long-term stroke risk in asymptomatic patients is severity of stenosis.^{1, 2, 8} Additionally, stenosis progression occurs in approximately 5% of patients with asymptomatic carotid stenosis annually and leads many to consider close ultrasound follow-up or revascularization due to perception of high associated risk.^{53, 54} However, these methods in isolation are imperfect predictors of stroke risk, which has prompted research into other ways of identifying asymptomatic patients with high long-term stroke risk who might benefit most from revascularization.⁵⁵

The Asymptomatic Carotid Stenosis and Risk of Stroke (ACSRS) study determined predictors of ipsilateral TIA/Stroke in asymptomatic patients on medical therapy, incorporating plaque morphology characteristics from ultrasound.⁵⁶ The addition of clinical and ultrasound-detected plaque features to stenosis severity improved the ability to predict stroke: the highest risk subgroup was found to have a 5-year stroke rate of 10-20%.

The effect of plaque morphology on subsequent stroke risk has also been studied using MRI. A meta-analysis showed that MRI-detected intraplaque hemorrhage (HR 3.7) and lipid-rich necrotic core (HR 5.7) are both significant predictors of TIA/stroke in patients with asymptomatic carotid disease.⁵⁷

Transcranial Doppler of the middle cerebral artery can detect microembolization from a proximal carotid stenosis. The Asymptomatic Carotid Emboli Study (ACES) prospectively studied patients with ≥70% asymptomatic stenosis. TCD surveillance detected microemboli in 16%, with significant hazard of subsequent ipsilateral TIA/stroke (HR 2.54).⁵⁸

TCD can also be used to quantify cerebrovascular reserve (CVR), though other methods are also used. Patients with normal CVR will have increased flow in the middle cerebral artery following administration of a vasodilator; patients with impaired CVR will not show this typical response. A meta-analysis of multiple methods for measuring CVR showed a significant association between impaired CVR and subsequent TIA/stroke in asymptomatic patients (OR 4.7).⁵⁹

These methods are promising and may guide selection of asymptomatic patients who will benefit most from revascularization. In fact, the European Society for Vascular Surgery (ESVS) guidelines recommend consideration for revascularization of asymptomatic carotid stenosis ≥60% only if one of these high risk imaging characteristics is also seen (Supplemental Table I, please see http://stroke.ahajournals.org).⁴

b. Best Therapy for Asymptomatic Patients: A Moving Target

i. Improving Stroke Risk with Intensive Medical Therapy

In major trials comparing CEA to best medical therapy in asymptomatic patients, statin medications were not widely used. It was only in the latter years of ACST that lipid-lowering medications were implemented.⁶⁰ Even within these trials, rate of any stroke on BMT improved from 3.5%/year in ACAS to 1.4%/year in the latter half of ACST as medical management improved.⁶¹ It should be noted that the gradual implementation of lipid-lowering therapy in ACST did not negate the beneficial effect of CEA, though it did decrease the absolute benefit of revascularization.⁶⁰

Very aggressive lipid lowering has become a recent area of interest although this approach has not yet been tested in the setting of carotid artery disease. Injectable PCSK9 inhibitors (evolocumab, alirocumab) lead to reduction of low-density lipoprotein cholesterol (LDL-C) even in patients already on statins. Though these agents are not yet widely available, their use is associated with reduction in cardiovascular events.^{62, 63}

Improvements in antiplatelet therapy for carotid disease may also be demonstrated in the coming years. New P2Y12 antagonists include clopidogrel, ticagrelor, prasugrel, ticlopidine and cangrelor. PAR-1 inhibitors, such as vorapaxar, have also been developed. For ticagrelor, a subgroup analysis of the SOCRATES trial⁶⁴ demonstrated that in patients with potentially symptomatic atherosclerotic disease (including some with ICA stenosis), ticagrelor treatment was associated with 90-day stroke, MI or death rate of 6.7% compared to 9.6% in patients on aspirin (P=0.003).⁶⁵

In the context of improvements in medical therapy, reviews of randomized and nonrandomized studies have shown that annual risk of stroke with medical therapy for asymptomatic carotid stenosis has consistently improved since the publication of ACAS.^{61, 66}

ii. Improving Operative Risks with Revascularization and Intensive Medical Therapy

Operative stroke/death rates have continued to improve for CEA based on randomized trial evidence and clinical registries (Supplemental Table II, please see http://stroke.ahajournals.org). From 1991-2010, published data have shown a 6% annual reduction in operative stroke/death.⁶⁷ These trends coincide with the precipitous drop in cigarette smoking, adoption of routine statin use, and the increasing use of dual antiplatelet therapy, which may be associated with improved operative outcomes and long-term outcomes following revascularization.⁶⁸ Accordingly, there have been increasing calls to lower the acceptable stroke/death thresholds set by many guidelines (<6% symptomatic, <3% asymptomatic), especially given improvements in medical therapy.

Complication rates have also improved with CAS (Supplemental Table II, please see http://stroke.ahajournals.org). In CREST and ACT1, which mandated EPD use, 30-day stroke/death rates of 2.5% and 2.9% were reported, respectively,^{6, 7} and are acceptable based on current guidelines.

The successful implementation of intensive medical therapy outside of the trial setting is another area of uncertainty. Current evidence suggests that IMT regimens should include anti-thrombotic therapy, aggressive hyperlipidemia treatment with high-intensity statins when tolerated (regardless of LDL level), anti-hypertensive medications, aggressive control of hyperglycemia in diabetics to A1C<7%, smoking cessation and lifestyle modification (exercise and diet counseling).^{8, 69} Future trials will yield important information as to how frequently the varied goals of IMT are met but the generalizability of these regimens remains unclear.

V. The Need for Additional Trial Data

Ongoing uncertainty regarding indications for carotid revascularization has led to wide international variation in practice patterns, particularly for asymptomatic disease. In the United States, registry-based data suggest that roughly 60% of carotid revascularizations are performed in asymptomatic patients.⁷⁰ Internationally, rates of asymptomatic carotid revascularization range from 0% in Denmark to 73% in Italy.⁷⁰ This variability is reflective of ongoing uncertainty. The results from major ongoing randomized controlled trials, such as CREST-2 and ECST-2 will guide management in the years to come.

Carotid endarterectomy and stenting are among the most studied surgical procedures in history, with more than 20,000 patients in randomized clinical trials. Both forms of revascularization are proven to be safe, when performed by experienced practitioners in properly selected patients. Future efforts must focus on determining which patient populations truly benefit from these sophisticated techniques, as every treatment decision carries complex risk implications.