Stroke accounts for over 150,000 deaths yearly in the United States, and there are an additional 500,000 non-fatal strokes, many of which are permanent and debilitating. It is now known that emboli arising from the carotid artery are a significant cause of stroke, but the link between carotid disease and cerebrovascular events was not clarified until the 1950’s. Dr. C. Miller Fisher, a neurologist at the Massachusetts General Hospital, recalled the work of Chiari, who in 1905 had made the link between carotid disease and neurologic events. In 1951, Fisher described tight lesions of the carotid artery with ulceration and hemorrhage in patients with hemiplegia. He noted that some carotid occlusions were asymptomatic while others caused strokes, sometimes preceded by prodromal symptoms that we now recognize as transient ischemic attacks or amaurosis fugax (temporary monocular blindness). Fisher famously hypothesized that “it is even conceivable that some day vascular surgery will find a way to bypass the occluded portion of the artery during the period of ominous fleeting symptoms.”
Direct evidence for embolization from the carotid system came in 1958, when Hollenhorst described cholesterol emboli in the retinal artery (now known as Hollenhorst plaques), which are refractile crystals, often lodging at arteriole bifurcations, and which frequently dislodge over time. This observation further confirmed that emboli from the carotid artery could travel to the cerebrovascular circulation, in this case going to the retina by way of the ophthalmic artery - the first branch of the internal carotid artery. (Although indicative of embolization, Hollenhorst plaques are common in the elderly and are not associated with an increased risk of stroke.)
Removal of carotid plaque eliminates the source of emboli, which arise from advanced lesions in the intima. Plaques typically occur in the proximal internal carotid artery opposite of the flow divider, where there is abnormally low and oscillating shear due to flow separation from the wall (Figure 1). This is where lipids infiltrate and are engulfed by macrophages. These early lesions grow as more lipid and cholesterol is deposited and the inflammatory response recruits macrophages and proliferating mesenchymal cells. As the wall thickens beyond its blood supply, it is invaded by thin-walled, delicate neovasculature, which is prone to rupture and hemorrhage, further enlarging the plaque. Eventually, necrotic cores develop, composed of dead inflammatory and mesenchymal cells, decaying red blood cells, calcium deposits, and cholesterol clefts. A fibrous cap consisting mainly of myofibroblasts and matrix covers these necrotic cores. Inflammation, partially in the shoulder regions, contributes to cap thinning and eventual rupture, spewing emboli into the circulation and exposing a highly thrombogenic surface, which serves as a source of further thromboembolism. While small emboli such as these do not cause significant downstream damage in other arterial beds, they can cause major strokes when released into the carotid circulation, where they may block blood flow to the retina (causing amaurosis fugax), or to the portions of the brain supplied by the anterior and middle cerebral arteries (causing hemispheric, focal, neurologic deficits such as numbness, paralysis, or aphasia).
Figure 1. Flow patterns in the carotid artery bifurcation.
The Rise of Carotid Endarterectomy
Fisher’s hypothesis proved correct when carotid endarterectomy was developed to eliminate the source of emboli by removing plaque from the bifurcation. The procedure carries a low risk of stroke or death (just a few percent). The common carotid artery and its internal and external branches are controlled with clamps, the artery is opened, and the offending plaque removed. This is made possible by the structure of the artery wall and the nature of carotid atherosclerosis. Plaque resides in the intima, sparing the media, which is a layer of concentric laminae of smooth muscle cells, aligned circumferentially and separated by distinct layers of elastin (Figure 2).
Figure 2. Structure of the arterial wall. From Clark JM, Glagov S: Transmural organization of the arterial media: The lamellar unit revisited. Arteriosclerosis 5:19–34, 1985.
There is a natural cleavage plane between these layers, which allows the surgeon to gently separate the inner lining of the wall containing plaque from the outer wall leaving a thin layer of media and the surrounding adventitia intact. The remaining wall may be paper thin, but the adventitia is rich in collagen and remarkably resistant to stretching over time, preventing later carotid distention or aneurysm formation. Plaque tends to be confined to the carotid bifurcation, making it possible to remove it all the way to its most distal extent in the internal carotid artery. Ulceration and thrombosis are often evident on gross inspection of the removed specimen (Figure 3). Cooley performed the first successful carotid endarterectomy in 1953 on a 53-year-old patient with transient ischemic attacks. The operation was done without the benefit of carotid imaging, since none was available at that time.
Figure 3. Carotid plaque specimen opened to show a large thrombus within the lumen.
Ultrasound Diagnosis Becomes the Standard
Early clinical trials of carotid endarterectomy were performed when angiography was still used as the standard preoperative diagnostic test. This procedure is invasive and carries a small risk of stroke from catheter-related thromboembolism. Later trials relied on duplex ultrasound for the diagnosis of the extent of carotid narrowing, since this noninvasive test was accurate enough to eliminate the need for angiography. Catheter angiography has since been replaced largely by angiography with computed tomography or magnetic resonance imaging using appropriate contrast agents given intravenously. These techniques have the advantage of giving more detailed anatomic information in addition to imaging of the brain, but they are considerably more costly than ultrasound and are unnecessary prior to carotid endarterectomy when the duplex scan is of high quality and there are no concerns about intracranial pathology.
Duplex scanning classifies the extent of carotid narrowing into groups defined by the degree of luminal narrowing as a percentage of diameter reduction. It does so by measuring the velocities where they are the greatest, generally at or slightly beyond the point of maximum stenosis - usually in the very proximal internal carotid. This method is now well established and when performed properly has an accuracy of 90% in determining if there is a 70% degree of stenosis. The test is highly dependent on the skill of the examiner and the methods used. Concern was raised about the accuracy of the test during early clinical trials of carotid endarterectomy for symptomatic patients (particularly NASCET – see below). These trials did not use standardized equipment or methods. In 2003 a consensus panel with representatives from radiology and vascular surgery was convened to develop standards for reliable carotid ultrasound examinations. The panel made the following recommendations:
• Examinations should use gray-scale, color Doppler, and velocity waveform analysis.
• The degree of stenosis should be stratified into categories (see Table below).
• Grading of internal carotid artery (ICA) stenosis should be based on peak systolic velocity (PSV) and presence of plaque on imaging.
• ICA-to-common carotid artery PSV ratio and ICA end-diastolic velocity can be used when results are in doubt.
Figure 4. Table of Consensus Panel Criteria for Carotid Diagnosis. From: Grant EG, Benson CB, Moneta GL, et al. Carotid artery stenosis: gray-scale and Doppler US diagnosis--Society of Radiologists in Ultrasound Consensus Conference. Radiology. 2003 Nov;229(2):340-6.
Establishment of Carotid Endarterectomy for Symptomatic Patients
In the 1960’s and 1970’s carotid endarterectomy was used increasingly for stroke prevention, peaking at an estimated 150,000 procedures in 1987. In 1984 Dr. H.J.M. Barnett, a Canadian neurologist, and others raised concern about the growing number of carotid endarterectomies being performed and by reports of high rates of complications. Barnett became the principle investigator of the North American Symptomatic Carotid Surgery Trial (NASCET), which randomized symptomatic patients with greater than seventy percent carotid stenosis to medical or surgical treatment. All 659 patients received aspirin. Only centers with demonstrated low surgical morbidity and mortality participated. The trial was designed for a five-year follow-up, but was halted by the safety monitoring committee after only two years when the benefit of surgery was statistically significant for the primary end point of ipsilateral stroke (26% versus 9%), as well as all other endpoints:
|Two-Year NASCET Results (NEJM 1991;325(7):445-453)
||Medical Rate (%)
||Surgical Rate (%)
| Ipsilateral stroke
| Any stroke
| Andy stroke or death
| Major or fatal ipsilateral stroke
| Any major or fatal stroke
| Any major stroke or death
These highly significant results were corroborated by the European Carotid Surgery Trial (ECST), which also released their findings in 1991. The efficacy of endarterectomy in stroke prevention was thus established for symptomatic patients with high-grade stenosis, however there are still concerns about patient selection. The risk reduction of 17% means that 83% of patients who undergo the procedure do not benefit from it. In other words, five procedures are required prevent one stroke (this is referred to as the number needed to treat). Attention has thus been directed at identifying the subgroup of patients most likely to benefit from surgery. NASCET data suggest that age is a factor: older patients benefit more from surgery (especially the > 75 year-olds). Other factors that favored intervention include male gender, stenosis of greater than 90% (but interestingly, not very high-grade lesions with incomplete filling of the distal ICA, so called “string sign”), irregular plaques, contralateral occlusion, recurrent symptoms, and hemispheric rather than retinal symptoms.
Most strokes occur within weeks of the presenting symptom, before unstable plaques reorganize to a nonthrombogenic state. Thus, surgery is most effective in stroke reduction during this time. In the past there was concern for conversion of bland to hemorrhagic infarct if surgery was performed within four weeks of stroke, but it has been shown that surgery need not be delayed if strokes are non-disabling. Subgroup analysis of NASCT data showed maximum benefit of endarterectomy if performed within 2 weeks of the presenting neurologic event. Six months following symptoms, the risk of future stroke is no more than that of a patient who never had a symptom, thus patients are considered to be asymptomatic if they have been symptom free for six months.
Trial results were not as clear for symptomatic patients with lesser degrees of stenosis. ECST included patients with less than 30% stenosis (NASCET did not) and found surgery to be harmful in this group. From NASCET, there were too few patients in the 30-69% stenosis group to establish a benefit at two years. Thus trials were extended beyond two years for patients with lesser degrees of stenosis. Surgery was either ineffective or harmful for patients with less than 50% stenosis. For moderate stenosis (50-69%) there was modest but significant benefit in the NASCET group (15.7% versus 22.2% at five years, number needed to treat 15). Patients were most likely to benefit if they were male and if they had hemispheric symptoms rather than amaurosis fugax.
The standard of care in most institutions now favors carotid endarterectomy in symptomatic patients with greater than 70% stenosis who are deemed good candidates for surgery. High risk may be related to physiologic factors such as symptomatic coronary disease, uncontrolled congestive failure, severe pulmonary disease, and uncontrolled diabetes. It may also be due to anatomic factors such as prior neck radiation, surgically inaccessible lesion, contralateral occlusion, laryngeal palsy, and presence of a tracheostomy. Surgery may be indicated in patients with greater than 50% stenosis if no other source of emboli can be found, especially if the lesion is particularly irregular and ulcerated. High-risk patients may be treated with stenting using emboli protection. Stenting may also be used in the context of clinical trials. It is contraindicated in patients with significant circumferential carotid calcification or unfavorable aortic arch anatomy that makes catheterization of the carotid artery both difficult and hazardous. Patients over the age of 75 tend to do better with surgery. See more on stenting below.
Improved methods of predicting stroke are being sought to reduce the number needed to treat. The nature of rupture-prone (or “vulnerable”) plaques is being examined along with ways to identify them biochemically or with imaging. One possibility is high-definition magnetic resonance scanning, which can identify many different plaque components. Such studies have found, for example, that intraplaque hemorrhage is associated with an increased risk of stroke. Ultrasound imaging has also been evaluated. Heterogeneous plaques tend to behave worse than homogeneous plaques, but the resolution and ability to discern different plaque components is relatively poor compared to magnetic resonance imaging.
Questionable Benefit of Carotid Endarterectomy for Asymptomatic Patients
Clinical trials of carotid endarterectomy for asymptomatic patients appeared a few years after NASCET and ECST. The Asymptomatic Carotid Surgery Trial was published in 1995 (JAMA 1995;273(18):1421-8). The study included over 1600 patients with greater than 60% stenosis followed for five years. All patients received aspirin. Those randomized to surgery underwent confirmation of disease severity with angiography, which was the standard of care at the time, followed by endarterectomy if stenosis was confirmed. At five years there was a 47% relative risk reduction in the surgery group (11% stroke rate in the medical arm versus 5.1% for surgery). Of note, 1.2% of the strokes in the surgery group were caused by angiography, which is no longer used in care of these patients. Similar results were reported from the multinational Asymptomatic Carotid Surgery Trial (Lancet 2004;364(9432):416), which randomized over 3000 asymptomatic patients to observation versus surgery. At five years the stroke rate was 11.8% in the medical group and 6.4% in the surgery group. The Veterans Affairs Cooperative Study Group published similar results in the same year. This trial included 444 asymptomatic patients with greater than 50% stenosis. The trial was halted when results of the larger trials demonstrated benefit of intervention. At the conclusion of this trial, there was a significant reduction in all neurologic events (TIA, amaurosis fugax, and stroke) in the endarterectomy group (8.0% versus 20.6%), but the study was underpowered to show a significant benefit in ipsilateral stroke alone (4.7% versus 9.4%). Three of the ten stroke in the surgery group were caused by preoperative angiography.
Although these randomized clinical trials documented a statistically significant stroke risk reduction in patients undergoing endarterectomy, the clinical significance has been intensely contested. While the overall relative reduction in stroke rate is fifty percent, the absolute stroke reduction is only about one percent per year, and the number needed to treat is a costly 40. The risk reduction with surgery is so small that any increase in the trials’ excellent perioperative stroke or death rate would eliminate the benefit. Reports from databases that include a broad range of providers have suggested that surgical complication rates outside of study groups are too high to justify the procedure in asymptomatic patients. Also, women in general tend to have a higher complication rate with carotid endarterectomy than men, and there have been too few women in the cohorts studied to determine if this group benefits from surgery.
Results of these trials were followed by an increase in the number of procedures in the US, where surgeons are rewarded economically for doing procedures and society is willing to pay a high cost for any reduction in the number of strokes, and a decrease in the UK, where the National Health System found the cost to outweigh the minimal benefit.
As stenting became established therapy for stenotic lesions in other arterial beds, it was a natural extension to use it for carotid stenosis. In peripheral arteries, stenosis causes symptoms due to flow reduction, and stenting is an effective way to restore flow. However, in the cerebral circulation, flow reduction usually is not the issue since collateral circulation is so robust. Here the problem is thromboembolization from plaques causing distal, cerebral infarction. Thus the role of treatment is to stabilize plaque, and it is not obvious that the open-structure of standard, wire mesh stents would achieve this goal.
Early experience with stenting of carotid arteries was not favorable due to strokes caused either by manipulation of catheters through diseased arteries or disruption of plaque with release of emboli at the time of balloon angioplasty. These early, unfavorable results subsequently have been reduced by experience and improved technology. Because the results of surgery are so good, most trials of stenting are designed to show noninferiority rather than superiority, with the assumption that patients would prefer the seemingly less invasive procedure to surgery if the two had comparable results. Cost does not necessarily favor stenting, which is surprisingly expensive.
In 2008 the industry-supported Stenting and Angioplasty with Protection in Patients at High Risk for Endarterectomy Trial (SAPPHIRE), reported three-year results. The trial enrolled high risk, symptomatic and asymptomatic patients randomized to stenting with emboli protection versus endarterectomy in a noninferiority design. There was no significant difference in outcomes. In the same year, a similar study with similar results in1200 patients followed for 24 months was reported from the Stent-Protected Angioplasty versus Carotid Endarterectomy trial (SPACE). Two years later, four-year results of the smaller Endarterectomy Versus Angioplasty in Patients with Symptomatic Severe Carotid Stenosis (EVA-3S) trial reported comparable results. The International Carotid Stenting Study (ICSS) reported similar long-term functional outcomes for stenting and endarterectomy at five years, although the stenting group had a higher rate of any stroke (15.2% versus 9.4%). The Asymptomatic Carotid Trial (ACT I), published in 2016, enrolled over 1400 asymptomatic patients who were not high risk and were randomized to stenting with embolic protection or carotid endarterectomy. The study found that during five-year follow-up stenting was not inferior to surgery for post-procedure stroke or mortality.
Results of these various studies are mirrored in the large Carotid Revascularization Endarterectomy versus Stenting (CREST ) Trial. This trial included 2500 patients, both symptomatic and asymptomatic, with prolonged follow-up. To achieve this level of enrollment, the trial used 117 centers, many years for recruitment, and it broadened the initial inclusion criteria to include both symptomatic and asymptomatic patients with high-grade carotid stenosis as long as they were candidates for either treatment. As in prior reports, early results favored surgery for its lower stroke risk and stenting for fewer cardiac events. Patients over the age of 75 did better with surgery, possibly due to challenges of angiography in the setting of more advanced, generalized arterial disease that includes the aortic arch. Long-term results were similar for the two treatment arms. Results are summarized in the table below.
|Long-term CREST results
| 30-d Stroke
| 30-d MI
| 30-d Death
||9 (0.7 %)
| 10-yr stroke
| 10-year Stroke/MI/death
(N Engl J Med. 016 Mar 17;374(11):1021-3)
Pre-procedure imaging with MR or CT angiography is required to select appropriate candidates for stenting. These studies determine if there are anatomic limitations such as unfavorable arch anatomy, peripheral arterial obstructive disease, or severe bifurcation calcification. Current practice is dictated by reimbursement guidelines by the Centers for Medicare and Medicaid Services (CMS). Stenting with embolic protection devices is reimbursed for symptomatic patient with 70% stenosis who are high-risk for surgery either due to anatomic or physiologic factors. Otherwise it is reserved for use in approved clinical trials involving high-risk patients with 50%-70% stenosis and symptoms or > 80% stenosis and no symptoms.
CREST-2: Does stenting or endarterectomy reduce stroke rates in asymptomatic patients in the modern era?
Current interventions for carotid stenosis have improved over time. Better methods of emboli protection, which is now standard of care, reduce the rate of peri-procedure stroke with stenting. These include methods to reverse flow in the internal carotid artery during the procedure and better devices to capture released emboli. Long-term results may improve with mesh-covered stents. Surgery results are improved by better perioperative cardiac risk management, intraoperative anesthesia, and postoperative care. At the same time, modern medical management has reduced stroke rates to near those achieved by surgery in prior randomized trials. Thus, there is equipoise regarding intervention for asymptomatic carotid stenosis: it is not known if intervention is better than intensive medical management alone.
In 2014 the CREST trialists have embarked on a second trial, CREST-2, to determine if either stenting or endarterectomy benefit asymptomatic patients. There are separate arms for stenting and endarterectomy, each of which is being compared to maximal medical management, but not to each other. To insure that the study was ethical, the NIH required that patients not having carotid intervention receive the best possible treatment of their atherosclerosis. This involves very close medical follow-up with intensive management of risk factors as outlined in the table below.
Aspirin for most patients
| Lifestyle modifications
|| Smoking cessation
|| Coaching Specially trained coaches
Every 2 weeks for 3 months
Every 3 months for 9 months
Every 6 months thereafter
The plan is to enroll asymptomatic patients over the age of 35 with at least 70% stenosis as determined by angiography or the duplex criteria described above and who are candidates for either surgery or stenting. They will be followed for up to four years for stroke or death. As with the first CREST study, enrollment has been slow because the difference between intervention and non-intervention is significant and participants often have strong preconceived bias toward one or the other. A detailed discussion with each potential subject emphasizes the fact that it is not known if intervention is beneficial. Further, all subjects benefit from intensive medical treatment, which is known to reduce complications from atherosclerosis, both cardiac and cerebrovascular. A similar trial by the ECST group also is underway. It compares revascularization with medical treatment alone for asymptomatic patients. It is enrolling symptomatic patients who are not high-risk. Results of these studies will guide future management of asymptomatic patients. Until results are known, asymptomatic patients with high-grade stenosis (> 70%) may be treated with medical management only, carotid endarterectomy if they are not high-risk, or stenting if in the context of clinical trials.
TransCarotid Artery Revascularization (TCAR)
As noted above, the main problem with stenting is stroke caused by embolization at the time of the procedure. For surgery the main problems are myocardial infarction and cranial nerve injury. The later occurs at a rate of 5-15%, primarily involving the vagus, hypoglossal, or facial nerve (fortunately, most recover). TCAR is a novel approach to carotid revascularization that addresses all of these issues. The common carotid artery is accessed through a small, supraclavicular incision. This surgery can often be done using local anesthesia and involves no dissection around cranial nerves other than the vagus, which is large and easily identified in the carotid sheath. A catheter device is placed into the common carotid artery well below the stenotic lesion. This device is designed to provide access to the carotid bifurcation and also to allow reverse flow in the internal carotid artery during stenting. To reverse flow, the inflow is blocked and back bleeding from the internal carotid artery flows through the catheter, which is connected to the femoral vein for return of the blood. A filter removes any embolic debris from the blood before it is returned to the venous circulation.
Preliminary studies with TCAR in 141 high-risk patients were favorable with a 1.4% 30-day stroke rate, a 0.5% cranial nerve injury rate (hoarseness from vagus nerve manipulation, all recovered), and a 3.7% combined 30-day stroke and death rate. Nearly half were accomplished with local anesthesia. The rate of clinically silent microembolization detected by diffusion-weighted magnetic imaging, was also favorable, occurring in 23% of TCAR patients as compared to 70% with standard, transfemoral stenting with embolic protection devices (data from ICSS), and 17% with endarterectomy. TCAR has been approved for use in high-risk symptomatic patients with > 70% stenosis. Trials are currently underway for high-risk symptomatic patients with 50-70% stenosis and high-risk asymptomatic patients with > 80% stenosis.
Embolization from carotid artery bifurcation disease was identified as a significant cause of stroke in the 1950’s. Carotid endarterectomy was developed during this time and was used increasingly until the 1980’s when concern was raised about its effectiveness. Several randomized clinical trials subsequently proved the efficacy of this procedure in reducing stroke in symptomatic patients with high-grade carotid stenosis (greater than 70%). These studies were terminated at two years at which time intervention was clearly superior to medical treatment alone. Still significant, but less benefit was found at five years in patients with moderate stenosis (50% to 70%). In asymptomatic patients, at five years, surgery provided a smaller but significant stroke reduction, 11% versus 5%, or about 1% per year. Since then results of medical therapy have improved to rival those of intervention. For this reason, clinical trials are underway to reassess the benefit of surgery in asymptomatic patients. Results of stenting are improving with development of better stents and emboli protection devices. Carotid stenting with embolic protection devices is now approved for high-risk symptomatic patients with > 70% stenosis. It is being evaluated in clinical trials for treatment of symptomatic patients with 50% to 70% stenosis and asymptomatic patients. Also in evolution are imaging methods that will allow better identification of patients who benefit from carotid intervention since numbers needed to treat are high: even symptomatic patients with high-grade stenosis have only a 20% five-year risk of stroke. Currently, duplex ultrasound is the most cost effective screening method for carotid diagnosis. When results are clear, it is sufficient for patients prior to carotid endarterectomy. When stenting is considered, either CT or MR angiography is required to fully visualize the anatomy from the aortic arch to the brain. Treatment of carotid stenosis has evolved significantly since the 1950’s and continues to do so with development of improved interventions and better patient selection.
Muller MD, von Felten S, Algra A, et al. Immediate and Delayed Procedural Stroke or Death in Stenting Versus Endarterectomy for Symptomatic Carotid Stenosis. Stroke 2018 Nov;49(11):2715-2722.
Brott TG, Howard G, Roubin GS, et al. Long-Term Results of Stenting versus Endarterectomy for Carotid-Artery Stenosis. N Engl J Med 2016 Mar 17;374(11):1021-31.
Grant EG, Benson CB, Moneta GL, et al. Carotid artery stenosis: gray-scale and Doppler US diagnosis--Society of Radiologists in Ultrasound Consensus Conference. Radiolody. 2003 Nov;229(2):340-6.
Fisher CM: Occlusion of the internal carotid artery. Arch Neurol Psych 1951;65:346-377.
North American Symptomatic Carotid Endarterectomy Trial Collaborators : Beneficial effect of carotid endarterectomy in symptomatic patients with high-grade carotid stenosis. N Engl J Med 1991;325:445-453.
European Carotid Surgery Trialists' Collaborative Group : European carotid surgery trial: Interim results for symptomatic patients with severe (70-99%) or with mild (0-29%) carotid stenosis. Lancet 1991;337:1235-1243.
Kashyap VS, King AH, Foteh MI, et al. A Multi-Institutional Analysis of Contemporary Outcomes After Transcarotid Artery Revascularization versus Carotid Endarterectomy. J Vasc Surg. 2018 Jun;67(6), e191–e192.