With medical therapy the risk of stroke declines fairly rapidly, so that the benefit of endarterectomy is highest early. In NASCET, the risk of patients randomized to medical therapy declined quickly, becoming equal to that of surgical therapy in 18 months in patients with moderate stenosis, and after 30 months in patients with severe stenosis [77]. As medical therapy has improved, this time has been compressing, so that patients benefit most from surgery done within 2 weeks or less, and the benefit is marginal if done more than 3 months after the sentinel stroke or TIA [81].

Carotid stenting involves placement of a metal sheath into the narrow segment of the artery by catheterization (often from the femoral artery). Sometimes the stent is dilated by balloon angioplasty; an alternative is self-expanding stents. This approach seems intuitively attractive, but has important problems that result from the difficulty of passing a catheter through a tortuous craggy artery. Table 14.2 summarizes trials comparing carotid endarterectomy to stenting.

In contrast to coronary intervention, the purpose of carotid endarterectomy or stenting is not to increase blood flow to the brain, but to prevent embolization of atheromatous debris into the brain (Fig. 14.1). Furthermore, the purpose of intervention is to prevent ipsilateral stroke due to large artery disease, not to prevent contralateral stroke, lacunar infarction, or myocardial infarction. In our personal opinion, physicians should not be distracted by the finding in CREST [82] that stenting carried a lower risk of myocardial infarction, nor by the lower risk with stenting in patients at high risk for surgery [83], most of whom were not good candidates for carotid revascularization [84].

Carotid revascularization with stenting is often, even usually, complicated by procedure-related embolization. This problem is illustrated in Fig. 14.2. The Calgary stroke group [85] performed MRI scans after deployment of stents, with transcranial Doppler embolus detection during the stenting procedure, in 30 patients, 23 % of whom were asymptomatic. They divided emboli into malignant and nonmalignant emboli according to signal intensity. The median embolic signal count was 212.5. The embolic signal count was highest during stent deployment, followed by deployment of the protection device. New diffusion-weighted imaging (DWI) lesions were found in 80 % of the patients after stenting, with a median of four new DWI lesions (interquartile range 7). Two of 30 (6.7 %) had new or worsening clinical deficits post-CAS.

As a result, with carotid stenting the risk of stroke is approximately double that with carotid endarterectomy. The myocardial infarctions in CREST, many of which were “biochemical” myocardial infarctions identified by elevations of troponin levels, impaired quality of life much less than did strokes. A major stroke impaired quality of life three times more than did a minor stroke, and a myocardial infarction only reduced quality of life 2/3 as much as did a minor stroke [82]. Furthermore, a year after the intervention, an endpoint stroke had large adverse effects on physical function, role functioning, and vitality (energy/fatigue); endpoint myocardial infarctions and cranial nerve palsies had quantitatively smaller and not statistically significant effects [86, 87].

In the USA, more than 90 % of carotid interventions are for asymptomatic stenosis [92], even though the evidence discussed below now indicates that 90 % of patients with carotid stenosis would be at lower risk with medical therapy alone [93]. This is being justified by historical data that no longer pertain (Table 14.3). When the Asymptomatic Carotid Artery Surgery trial (ACAS) [94] was carried out there were essentially no patients on statins; during the later years of the European Asymptomatic Carotid Surgery Trial (ACST) [95] there were some patients on low-dose statins, but very few (if any) on high-dose statins. It is therefore not legitimate to justify carotid stenting on the basis of the Carotid Revascularization with Endarterectomy or Stenting Trial (CREST) [82], which had no concurrent medical arm [93]. Fortunately the CREST 2 trial and the recently initiated European carotid stenosis trial will compare endarterectomy, stenting, and best medical therapy, so this issue should be sorted out within a few years.

In CREST, the procedural (30-day) risk of stroke or death for asymptomatic patients was 2.5 % for stenting and 1.4 % for endarterectomy; the 4-year risk was 4.5 % with stenting and 2.7 % with endarterectomy [82]. The 2011 report of Wang et al. [92] documents in Medicare patients a 1-year risk of stroke or death of 16.7 % for stenting and 11 % for endarterectomy. It is now clear that stenting carries a higher risk of stroke than does CEA and a higher risk than that with intensive medical therapy [96].

Intervention with CEA or CAS in asymptomatic stenosis is based on the historical risk of stroke in patients randomized to medical therapy in ACAS and ACST, which were approximately 2 % per year. However, as shown in Fig. 14.3, the annual risk has declined since 2005 to approximately 0.5 %. This was shown in a population-based study in the UK [5], a stroke prevention clinic population in Canada [6], and in meta-analyses [4]. The consequence of this change is that most patients with asymptomatic stenosis (approximately 90 %) would be better off with medical therapy.

Patients with symptomatic carotid stenosis are also at lower risk now than at the time the randomized trials were carried out. Surgical risk has been declining, as has the risk with more intensive medical therapy. Some have objected that the risk with medical therapy is higher in patients with more severe stenosis, but Naylor has recently shown that this is not the case [97].

Women, patients with less severe stenosis [81], and patients with only retinal TIAs [98] benefit less from endarterectomy. Patients with chronic ocular ischemia, who may benefit from revascularization, represent a special case.

That there are lower risk patients with symptomatic stenosis, and that risks are declining both with medical and surgical therapy, has led to calls for randomized trials of medical vs. surgical therapy, not only for asymptomatic carotid stenosis, but also for symptomatic stenosis [7].

The challenge, therefore, is how to identify the few who might benefit from CEA or CAS.

Approaches to identifying patients with high-risk carotid stenosis were reviewed in 2012 [99]. These include evaluation of clinical characteristics [100], transcranial Doppler (TCD) microemboli [42, 101], ulceration on 3D ultrasound [102], plaque composition/texture on ultrasound [103–105], neovascularization on ultrasound [106], intraplaque hemorrhage on MRI [107], juxtaluminal black plaque [108, 109], and plaque inflammation on PET/CT [110–112]. Plaque roughness on ultrasound and ulcer volume on 3D ultrasound are in development.