History

The carotid artery is unique in the human body. It is the only artery that bifurcates into a bulb larger than itself. It is subjected to unique hemodynamic and pathophysiologic forces and has been a source of fascination for centuries. In 438 BC, the Greeks depicted manual neck compression on the Parthenon as a battlefield technique to stun, stupefy or plunge an opponent into a deep sleep (karos or karotides). In the same century, Hippocrates recognized that lesions of one neck vessel resulted in contralateral hemiplegia and coined the term apoplexy (to strike down) to describe stroke. Rufus of Ephesus in 100 AD is credited with officially naming these vessels carotid ^{Reference Robicsek and Robicsek10,Reference Gurdjian and Gurdjian11} and Vesalius in 1543 was the first to provide accurate illustrations and to recognize their significance for cerebral circulation. ^{Reference Robicsek and Robicsek10–Reference Barnett12} Wepfer in 1658 was the first to provide detailed descriptions of the cerebral and extracranial vasculature, and to describe extracranial carotid thrombosis in association with apoplexy. ^{Reference Robicsek and Robicsek10,Reference Gurdjian and Gurdjian11} Willis in 1664 is credited with recognizing the physiologic and pathologic importance of his eponymous Circle at the base of the brain, including the potential for collateral circulation in proximal occlusion, a principle dramatically proven by John Hunter in the 1700s.

Pathophysiology

In 1905, Chiari provided autopsy confirmation of atheromatous plaque in the cervical vessels as the source of cerebral emboli in stroke, a condition he termed “endarteritis chronica deformans.” ^{Reference Chiari13} Chiari in 1906 and Hunt in 1914 both re-emphasized the importance of the extracranial carotid arteries, ^{Reference Hunt14} which were first imaged by Moniz in 1927 with direct puncture cerebral angiography. ^{Reference Moniz15} It was not until the 1940s that detailed clinico-pathological association between carotid atherosclerosis and stroke was observed, a correlation then firmly established by C Miller Fisher’s seminal reports from the 1950s. ^{Reference Fisher16,Reference Fisher17} His descriptions of carotid plaque associated thrombus and intraplaque hemorrhage, and the observation that most symptomatic patients had at least a 75% stenosis, or luminal diameter of 1 millimeter (mm) or less, form the basis for our understanding of extracranial disease and cerebral ischemia. ^{Reference Fisher, Gore, Okabe and White18,Reference Fisher and Ojemann19} External to internal carotid artery (ICA) collaterals were first described in 1911, and leptomeningeal collaterals, initially postulated by Heubner in 1874, in 1953. ^{Reference Vander Eecken and Adams20}

Carotid surgery

The earliest reports of surgery on the carotid artery appeared in 1793 and were usually related to trauma and hemorrhage. John Abernethy, a protégé of John Hunter demonstrated that ligation of the carotid artery was survivable because of collateral flow in 1803. ^{Reference Thompson21} C Miller Fisher was an early proponent of revascularization surgery for extracranial atherosclerosis, although initially for restoration of normal blood flow rather than prevention of cerebral emboli. ^{Reference Fisher17}

The first successful surgical reconstruction of the carotid artery was performed in Buenos Aires in 1951, ^{Reference Thompson21} and the first published report of successful CEA was by Eastcott et al in 1954; ^{Reference Eastcott, Pickering and Rob22} an earlier claim by DeBakey from 1953 was not reported until 1975. Interestingly, Eastcott used systemic hypothermia of 28C during 28 minutes of carotid occlusion. The first use of a shunt during CEA, a technique still debated today, was by Cooley in 1956. ^{Reference Thompson21} Surgical pioneers such as Debakey and Cooley initially believed that the problem of carotid stenosis was due to reduced cerebral blood flow, and that stenosed vessels should be opened to improve flow, not prevent emboli. ^{Reference Robertson23} It was Francis Murphey in 1973 ^{Reference Murphey24} who championed the embolic theory of stroke, and Wesley Moore in 1978 who documented the significance of ulcerated carotid plaques. ^{Reference Moore, Boren and Malone25}

Over the next several decades, the numbers of CEA procedures worldwide grew rapidly, with essentially no guidelines on who would benefit from the procedure, or what would be acceptable complication rates. In the United States, surgical mortality in over 2,400 cases performed between 1961 and 1968 was approximately 4.5%. ^{Reference Easton26} By 1976, over 34,000 procedures were being performed annually, with persistent high morbidity and mortality rates, up to 21% in one community-based series. ^{Reference Barnett, Plum and Walton27} By 1985, over 100,000 CEAs were done annually, the third most common operation in the United States (US), with morbidity and mortality estimated at 10%. ^{Reference Easton26,Reference Barnett, Plum and Walton27} General unhappiness with both medical and surgical treatments of stroke led to the Joint Study on Extracranial Occlusive Disease in the US between 1959 and 1976, which added a large amount of surgical information to the debate but showed no significant outcome difference between medical and surgical groups. ^{Reference Barnett12,Reference Fields, Maslenikov, Meyer, Hass, Remington and Macdonald28} Heterogeneity and unreliability of accumulated clinical data prompted calls for multicentre randomized controlled trials (RCTs). ^{Reference Barnett, Plum and Walton27}

Initial RCTs of CEA vs. medical therapy were negative for surgical benefit. ^{Reference Fields, Maslenikov, Meyer, Hass, Remington and Macdonald28–Reference Shaw, Venables, Cartlidge, Bates and Dickinson30} It was not until the landmark RCTs, the European Carotid Surgery Trial (ECST) and the North American Symptomatic Carotid Endarterectomy Trial (NASCET), published in 1991, that CEA was clearly shown to effectively prevent strokes in symptomatic patients with carotid stenosis > 70%. ^31,32 In NASCET, perioperative stroke and mortality rates of 5.5% and 1.2% were achieved and generally accepted as the standard. Despite organizational and methodological differences, both studies showed significant benefit of CEA over medical therapy, with NASCET showing an absolute risk reduction of 16%, and the number needed to treat of 6 patients to prevent one stroke over 5 years in patients with 70%–95% stenosis. Both men and women with 70–99% stenosis had significant benefit, and men with 50%–69% stenosis had some benefit with CEA performed within 2–3 weeks of their ischemic event. A combined ECST-NASCET analysis also confirmed that the benefit of surgery was muted in patients with lesser degrees of stenosis (50%–69%) and those with near occlusions (95%–99%), and that there was harm from surgical intervention in those with < 50% stenosis. ^{Reference Barnett12}

Concurrently, investigators were studying patients with asymptomatic carotid stenosis. The Veterans Affairs study of 1993 ^{Reference Hobson, Weiss and Field33} showed no benefit of CEA over medical therapy. The Asymptomatic Carotid Atherosclerosis Study (ACAS) in 1995 showed marginal benefit from CEA in patients with severe stenosis, with an annual risk reduction of 1% and a number needed to treat of > 65 to prevent one stroke over 2 years. ³⁴ Only men aged 75–80 with 60%–99% stenosis were found to benefit from CEA. These studies did not recognize or analyze for near occlusions as did NASCET. The inclusion of transient ischemic attack as an endpoint in this study distinguished it from ECST and NASCET, weakening the conclusions. The Asymptomatic Carotid Surgery Trial (ACST), published in 2004 showed similar results with marginal benefit, but only if perioperative stroke and death did not exceed 3%. ³⁵ These studies and others have justified intervention on vast numbers of asymptomatic patients despite impressive improvements in medical therapy ^{Reference Hackam36} and evidence that complication rates in the community often exceed 3%. ^{Reference Kresowik, Bratzler and Karp37} Approximately 90% of carotid revascularization procedures are performed in asymptomatic patients in the USA. ^{Reference Golnari, Nazari and Hurley38,Reference Otite, Khandelwal, Malik and Chaturvedi39} The rates of intervention elsewhere are much lower, estimated at 7% in the UK. ^{Reference Naylor40} Contemporary RCTs of intervention vs. medical therapy for asymptomatic disease such as ACST-2, ECST-2 and Stent Protected Angioplasty versus Carotid Endarterectomy (SPACE-2) have shown inconclusive findings, and results of the most comprehensive trial, Carotid Revascularization Endarterectomy vs. Stenting Trial (CREST-2) are pending. ^{Reference Rieff, Eckstein and Mansmann41,Reference Nederkoorn, Donners, van Velzen, Cheng and Gregson42}

Measurement of carotid stenosis

The measurement of carotid stenosis, historically so essential to clinical decision-making, has always been controversial. Absolute luminal diameter, cross sectional area, ratios, % reduction and “eyeball” techniques using a variety of modalities, beginning with angiography by direct puncture and then catheters, have all been used with varying degrees of validity. It is relatively easy to over-estimate stenosis severity, and arbitrary criteria, with stenoses as low as 30% ^{Reference Bauer, Boulos and Meyer43} were used to justify treatment. The Joint Study in the US was one of the first to rigorously define stenosis comparing the narrowest luminal diameter to the internal carotid artery beyond the bulb (28), but it was not until NASCET and ECST that validated measurement schemes were widely accepted.

NASCET used a ratio of the ICA diameter at the site of maximal stenosis to the ICA diameter well beyond the bulb where the walls are parallel ^{Reference Fox44} whereas ECST used a ratio of the maximal ICA stenosis to the carotid bulb, unseen on a catheter angiogram. There are pitfalls with both systems. For NASCET, interpretation of partial or complete collapse beyond a proximal stenosis, ^{Reference Fox, Eliasziw, Rothwell, Schmidt, Warlow and Barnett45} resulting in “near occlusion” will produce otherwise fallacious ratios of percent stenosis. Near occlusions are common when recognized correctly. ^{Reference Johansson and Fox46} NASCET percentage calculation uses a denominator of a normal ICA diameter well beyond the bulb where the walls are parallel, not commonly complied with and located outside the range of ultrasound windows. The ECST pitfall is to measure an imagined bulb diameter. These pitfalls lead to unique prognostic and therapeutic distortions.

There was initial confusion when comparing outcome results using NASCET and ECST systems, with a study in 1994 demonstrating their essential incompatibility, i.e., a 70% ECST stenosis was calculated as suggesting only 40% by NASCET, 75% ECST suggesting 50% NASCET, and 85% ECST suggesting 70% NASCET. ^{Reference Rothwell, Gibson, Slattery, Sellar and Warlow47} The NASCET method became the most durable and widely accepted for research and outside the bounds of clinical trials because endarterectomy clearly corrected the very serious stroke risk for the most severe stenoses. It was always relatively easy, however, to “fudge” the percent stenosis to exaggerate severity, thereby giving the appearance of surgical eligibility for patients with a stenosis that didn't meet the rigorous NASCET criteria. The denominator of the commonly used ratios has always been a source of ambiguity and unreliability. ^{Reference Fox44}

The replacement of instrumented angiography and its intrinsic risks with less-invasive modalities such as duplex ultrasound, CTA and MRA permits much more accurate and detailed evaluation of carotid disease. Direct measurement in mms at the site of greatest stenosis is the key parameter, accurate from CTA without the stroke risk of catheter angiography, and nearly so with MRA. NASCET percentages can be extrapolated from CTA images, ^{Reference Bartlett, Walters, Symons and Fox48} but overestimation of percentage stenosis is common ^{Reference Horev, Honig and Goldbart49,Reference Jareczek, Farrell and Lehman50} and the vagaries of percent stenosis continue. The real comparative measurement is the stenosis diameter without a ratio, as originally proposed by Miller Fisher. ^{Reference Fisher and Ojemann19}

Ultrasound and MRI now provide detailed analysis of plaque morphology and composition, allowing demonstration of plaque progression, regression and vulnerability. ^{Reference Brinjikji, Huston, Rabinstein, Kim, Lerman and Lanzino51} More sophisticated, contemporary criteria are becoming available to determine the need for carotid intervention. ^{Reference Saba, Moody and Saam52,Reference Saba, Cau and Murgia53} Maximum wall thickness, lipid-rich necrotic core, intraplaque hemorrhage, fibrous cap rupture, plaque inflammation and neovascularity are all associated with increased stroke risk. The recently introduced Carotid Plaque-RADS scale ^{Reference Saba, Cau and Murgia53} is a multi-modality scoring system that incorporates these variables, as well as plaque burden, stenosis progression and calcification to produce an overall assessment of plaque vulnerability and stroke risk.

It is now time for stroke researchers to abandon ratios and percent stenosis, debated principles now decades old. NASCET outcomes were based on stenosis degree, and this remains valid. New studies in both symptomatic and asymptomatic patients, however, must work not only from stenosis assessed at the narrowest diameter in mms but should also incorporate current determinants of plaque vulnerability and stroke risk. ^{Reference Saba, Cau and Murgia53}

Carotid angioplasty and stenting

The earliest report of percutaneous transluminal balloon angioplasty of the carotid artery was in 1981. ^{Reference Mathias54} The procedure was initially performed primarily by cardiologists and general interventional radiologists, ^{Reference Yadav, Roubin and Iyer55} who soon realized that the carotid was not just another peripheral vessel, but one in which recoil, dissection and generation of emboli from friable atherosclerotic plaque following angioplasty could have major clinical consequences, eventually progressing to the routine use of stents and embolic protection devices (EPDs). CAS became a popular alternative to CEA, perceived as less invasive, particularly in high surgical risk patients. ^{Reference Bettmann, Katzen and Whisnant56}

RCTs of CEA vs. CAS soon followed, all defining primary endpoints as stroke or death and two including myocardial infarction. The Carotid and Vertebral Artery Transluminal Angioplasty Study (CAVATAS) in 2001 ⁵⁷ showed no difference in outcomes between groups, but relatively high major adverse effects in both (9%–10%). The Stenting and Angioplasty with Protection in Patients at High risk for Endarterectomy (SAPPHIRE) in 2004 studied high surgical risk patients, 71% of whom were asymptomatic, and found that CAS was safer, mainly due to the lower incidence of myocardial infarction. ^{Reference Yadav, Wholey and Kuntz58} This study had a major influence on health care policy in the US, leading to regulatory approval of CAS as a valid alternative to CEA in this group of patients. SPACE and Endarterectomy versus Stenting in patients with Symptomatic Severe Carotid Stenosis (EVA 3-S) in 2006 ^{59,Reference Mas, Chatellier and Beyssen60} showed conflicting outcomes, the former demonstrating non-inferiority, the latter a higher risk of stroke and death at 30 days with CAS. The International Carotid Stenting Study (ICSS) in 2014 reported equivalent 5-year stroke and death rates, but twice as many non-disabling strokes with CAS. ^{Reference Bonati, Dobson and Featherstone61} CREST, published in 2010, is considered to be the most informative comparison study, despite almost half of the 2,502 patients being asymptomatic. There was no significant difference in periprocedural and 4-year stroke and death rates, with more strokes in the CAS group and more myocardial infarcts in the CEA group. ^{Reference Brott, Hobson and Howard62} This study solidified the belief that CAS is equivalent to CEA in average surgical risk patients as well. However, there is evidence that stroke has a greater adverse effect on quality of life, so this represents a questionable equivalence. ^{Reference Cohen, Stolker and Wang63} CEA was safer than CAS in older patients (>75 yrs.) and those with multiple medical co-morbidities. This may seem paradoxical but is explained by the risk of emboli during balloon angioplasty and stenting of rigid, friable atherosclerotic arteries. ^{Reference Spence64} Up to 80% of patients undergoing CAS can show new diffusion weighted imaging lesions on MRI, with almost 7% worsening clinically in one series. ^{Reference Almekhlafi, Demchuk and Mishra65}

There are now multiple RCTs containing CAS arms specifically assessing asymptomatic patients. SPACE-2 and ECST-2 showed no difference between medical management and CAS, ^{Reference Rieff, Eckstein and Mansmann41,Reference Nederkoorn, Donners, van Velzen, Cheng and Gregson42} and ACST-2 was reported as showing equivalent complication rates of CAS and CEA in these patients. ^{Reference Halliday, Bulbulia and Bonati66} However, the periprocedural risk of stroke or death was 3.8% with CAS vs. 2.7% with CEA, and the 5-year risk was 5.5% with CAS vs. 3.6% with CEA.

Medical management

In the late 1800s, Sir William Osler and others were aware of the association of carotid disease and stroke, but it was not until the work of Miller Fisher and Adams in the 1950s and 1960s that the connection of atherosclerotic plaque as a source of embolism and therefore stroke was established, rather than vasospasm and hemodynamic alterations. ^{Reference Grotta67,Reference Fisher and Adams68} The first use of anticoagulant therapy for threatened stroke was reported in 1955 ^{Reference Millikan, Siekert and Schick69} but was eventually shown to be ineffective. ^70,71 The first successful trials, not for treatment but for prevention of stroke, occurred in the 1970s with the use of antiplatelet agents such as aspirin, dipyridamole, ticlopidine and clopidogrel, often with the active collaboration of vascular surgeons. ^{Reference Fields, Lemak, Frankowski and Hardy72–Reference Haas, Easton and Adams75} Stroke risk was reduced by up to 20% in these studies. Statins have also figured prominently in stroke prevention in both asymptomatic and symptomatic patients since the 1990s. ^{Reference Baigent, Blackwell and Emberson76–79} As imaging modalities became more sophisticated, the ability to determine features of plaque vulnerability dramatically improved. The American Heart Association developed well validated criteria of histologic classification in 1995 (80,51). Plaque echogenicity, neovascularity, ulceration, intraplaque hemorrhage, fibrous cap rupture and lipid-rich necrotic core are all associated with increased stroke risk. In combination with the presence of micro-emboli on transcranial Doppler, the ability to predict and monitor response to medical therapy has expanded the options available for the treatment of carotid disease. ^{Reference Kamtchum-Tatuene, Noubiap, Wilman, Saqqur, Shuaib and Jickling81–Reference Spence, Coates and Li83}

An approach to treatment of atherosclerosis based on measurement of carotid plaque implemented in 2003, ^{Reference Spence and Hackam84} was associated with a > 80% reduction of risk among patients with asymptomatic carotid stenosis: the 2-year risk of stroke declined from 8.8% to 1% (i.e., 0.5%/year), and the 2-year risk of myocardial infarction declined from 6.7% to 1%. ^{Reference Spence, Coates and Li83} Meta-analyses reported more than 10 years ago that with more intensive medical therapy, the risk of stroke or death was well below the risk of CAS or CEA. ^{Reference Naylor85} With truly optimal medical management including smoking cessation and a Mediterranean diet, most patients with asymptomatic carotid stenosis would not benefit from intervention. The risk of stroke from a severe, asymptomatic carotid stenosis is approximately 1% annually, or 4.7% over 5 years ^{Reference Chang, Tucker and Rothenberg86} and it seems likely that < 10% of asymptomatic patients require intervention. ^{Reference Hackam36,Reference Paraskevas, Veith and Spence82}