With the exception of near-occlusion, CEA is of overall benefit for selected patients with recent symptomatic carotid stenosis =50% (NASCET method), provided surgical stroke/death risk is low. The benefit is greater with greater stenosis, men, the elderly (aged =75y), most recent ischaemic event within 2w, irregular plaque surface, and impaired cerebral perfusion reserve. Patients with recent symptomatic carotid territory ischaemic events should be screened by Doppler ultrasonography, MRA, or CTA, confirming substantial stenosis with a second non-invasive investigation. Catheter angiography may be required to confirm uncertain results. The surgical peri-operative stroke and death rate (7% in RCTs) is higher in women, hypertension, peripheral arterial disease, and occlusion of the contralateral ICA or ipsilateral ECA. The experience of the surgeon and hospital are crucial, and audited peri-operative complication rates should be publically available. Carotid stenting is less invasive than CEA and causes fewer local complications (cranial neuropathy and neck haematoma), but carries a higher procedural risk of stroke. Stenting should be considered in younger patients, or those at increased risk from CEA. While stenting is of high risk for intracranial vertebral artery stenosis, risk is low for extracranial stenosis and should be considered for recurrent symptoms despite optimal medical therapy.
This chapter will cover only the revascularization treatment of patients with recently symptomatic carotid stenosis, except in the section on carotid artery stenting (CAS) where some of the relevant studies have mixed symptomatic and asymptomatic patients.
The fact that moderate or severe stenosis is responsible for most cases of ipsilateral stroke was demonstrated by randomized trials that showed that removing the extracranial internal carotid artery (ICA) stenosis at its origin, by means of carotid endarterectomy (CEA), significantly reduces the risk of subsequent ipsilateral carotid territory ischaemic stroke (European Carotid Surgery Trial [ECST] Collaborative Group, 1991; North American Symptomatic Carotid Endarterectomy Trial [NASCET] Collaborators, 1991a,b).
Observational studies in white populations done 30 years ago suggested that about one-quarter of all first-ever ischaemic strokes and transient ischaemic attacks (TIAs) were caused by atherothromboembolism from the origin of the extracranial ICA (Sandercock et al., 1989). More recent studies report an around 8% prevalence of stenosis measuring >50% thought to be causative of ipsilateral ischaemic stroke, with an additional 11% of patients found to have contralateral asymptomatic stenosis (Cheng et al., 2019).
The mechanisms by which extracranial ICA atherosclerosis causes ischaemic stroke are listed in Table 20.1.
|‘Vulnerable’ atherosclerotic plaque lesions usually situated close to the carotid bifurcation undergo endothelial erosion, fissuring, or rupture, which exposes the subendothelial tissue to circulating blood, resulting in local thrombus formation. This can cause infarction in the territory of the brain supplied by the ICA by the following mechanisms (Torvik et al., 1989; Ogata et al., 1990):|
|1. Plaque debris or thrombus may embolize and block a more distal vessel, commonly the middle cerebral artery or its branches, occasionally the anterior cerebral, and rarely the ipsilateral posterior cerebral artery in patients with a widely patent posterior communicating artery.|
|2. Atherosclerotic plaque and/or thrombus may encroach upon the lumen of the ICA sufficiently to cause severe stenosis or occlusion, which may lead to hypoperfusion of distal brain regions, particularly in arterial border zones, and thus to border zone infarction (also known as watershed infarction).|
|3. Thrombosis at the site of stenosis may propagate up the ICA to occlude its distal branches.|
|ICA: internal carotid artery.|
The risk of stroke ipsilateral to a carotid stenosis increases with the degree of symptomatic carotid stenosis until the artery distal to the stenosis begins to collapse (ECST Collaborative Group, 1991; NASCET Collaborators, 1991; Rothwell and Warlow, 2000; Rothwell et al., 2000). A group of symptomatic patients with ‘collapse’ or ‘abnormal post-stenotic narrowing’ of the ICA resulting in near-occlusion of the artery was separately identified in NASCET. This was because it was not possible to measure the degree of stenosis using the NASCET method when the post-stenotic ICA was severely narrowed as a result of markedly reduced post-stenotic blood flow (Morgenstern et al., 1997). In NASCET, such patients were therefore arbitrarily allocated a stenosis of 95%. The ECST subsequently also separately analysed such patients, and both trials reported consistent results that these patients had a paradoxically low risk of stroke on best medical treatment alone to the extent that they did not benefit from CEA (Morgenstern et al., 1997; Rothwell and Warlow, 2000; Rothwell et al., 2000). The low risk of stroke in patients with collapse of the ICA distal to a severe stenosis, selected for the trials because of recent TIA or minor stroke, was most probably due to the presence of a good collateral circulation, which is visible on angiography in the vast majority of such patients. Without a good collateral circulation, these patients would most likely have had a major stroke and would not have been eligible for the trials.
The risk of stroke ipsilateral to a carotid stenosis is greater in patients with recent neurological symptoms from carotid territory ischaemia than in those with more distant symptoms (Lovett et al., 2003, 2004; Coull et al., 2004). The risk is time dependent, being highest in the few weeks after the presenting event, reasonably high for the first year, and falling quickly over the next 2 years to that of neurologically asymptomatic carotid stenosis (ECST Collaborative Group, 1991; NASCET Collaborators, 1991; Rothwell et al., 2000; Coull et al., 2004). The high early risk of recurrence probably represents the presence of an active unstable atherosclerotic plaque, and the rapid decline in risk over the subsequent year most likely reflects ‘healing’ of the unstable atheromatous plaque or possibly in some cases an increase in collateral blood flow to the symptomatic hemisphere (ECST Collaborative Group, 1998; NASCET Collaborators, 1998; Rothwell and Warlow, 2000; Rothwell et al., 2000a).
In addition to the severity of carotid stenosis, the presence of neurological symptoms, and the time since symptom onset, a number of other factors have been associated with an increased risk of stroke in the presence of carotid stenosis. These include increasing age, stroke as opposed to a single TIA, multiple TIAs (suggesting an active, unstable plaque), an irregular and ulcerated plaque surface morphology (which is pathologically unstable), absence of angiographic collateral flow, impaired cerebral reactivity, a high frequency of transcranial Doppler (TCD)-detected emboli to the brain, hypertension, and coronary heart disease (Table 20.2) (Rothwell and Warlow, 1999; Rothwell et al., 2000, 2005). Of all the possible symptoms associated with carotid stenosis, amaurosis fugax (transient monocular blindness) is associated with the lowest risk of recurrence.
|Risk factor||Hazard ratio||95% CI||P-value|
|Stenosis (per 10% increase)||1.18||1.10–1.25||<0.0001|
|Time since last event (per 7 days)||0.96||0.93–0.99||0.004|
ECST: European Carotid Surgery Trial. CI: confidence interval. TIA: transient ischaemic attack. MI: myocardial infarction.
More recent research has concentrated on the use of imaging to identify ‘vulnerable’ plaque and predict those at increased risk of stroke from carotid stenosis. These plaque imaging techniques include ultrasound imaging, single-photon emission computed tomography (SPECT) scanning, and magnetic resonance imaging (MRI) (Liem et al., 2017). Although there is good evidence that plaque characteristics, such as the presence of plaque haemorrhage identified by MRI, identify patients at higher risk of recurrence, these techniques have yet to be shown to be useful in clinical practice and have not entered routine clinical practice.
The purpose of revascularization of a symptomatic extracranial ICA stenosis is to reduce the risk of recurrent ipsilateral carotid territory ischaemic stroke by removing the source of carotid occlusion or thromboembolism. The two most commonly used strategies are CEA and CAS.
There have been five randomized controlled trials (RCTs) comparing the effect of endarterectomy for symptomatic carotid stenosis combined with best medical therapy with the effect of best medical therapy alone (Fields et al., 1970; Shaw et al., 1984; Mayberg et al., 1991; ECST Collaborative Group, 1998; NASCET Collaborators, 1998).
The first two trials were small and are now largely ignored as not reflecting later standards of practice (Fields et al., 1970; Shaw et al., 1984). The larger Veterans Administration (VA#309) trial (Mayberg et al., 1991) showed a non-significant trend in favour of surgery, but was stopped prematurely when the initial results of the two largest trials were published (ECST Collaborative Group, 1991; NASCET Collaborators, 1991). The analyses of these trials were stratified by the severity of stenosis of the symptomatic carotid artery, but the degree of stenosis on pre-randomization angiograms was measured using different methods. It is notable that these trials were started more than 30 years ago at a time when medical therapy was much less intensive than contemporary ‘best’ medical therapy. Nevertheless, their results continue to provide the main basis for current practice in the treatment of carotid stenosis.
The NASCET method underestimated stenosis compared with the ECST method. Stenoses reported to be 70–99% in the NASCET were equivalent to 82–99% by the ECST method, and stenoses reported to be 70–99% by the ECST method were 55–99% by the NASCET method (Rothwell et al., 1994).
Analysis of ECST Using the NASCET Method of Measurement of Carotid Stenosis
The ECST group re-measured the degree of carotid stenosis on their original angiograms using the method adopted by the NASCET group, and also redefined their outcome events to match those of NASCET, for comparability (Rothwell et al., 2003a).
Re-analysis of the ECST using NASCET measurements of stenosis showed that endarterectomy reduced the 5-year risk of any stroke or surgical death by 5.7% (95% confidence interval [CI]: 0–11.6) in patients with 50–69% stenosis (n = 646, p = 0.05) and by 21.2% (95% CI: 12.9–29.4) in patients with 70–99% stenosis without ‘near occlusion’ (n = 429, p < 0.0001). Surgery was harmful in patients with <30% stenosis (n = 1321, p = 0.007) and of no benefit in patients with 30–49% stenosis or near-occlusion (n = 478, p = 0.6 and n = 78, p = 0.7, respectively). These results of ECST, when analysed in the same way as NASCET, were consistent with the NASCET results (Rothwell et al., 2003a).
A pooled analysis of data from the ECST, NASCET and VA#309 trials, which included over 95% of patients with symptomatic carotid stenosis ever randomized to endarterectomy vs medical treatment, showed that there was no statistically significant heterogeneity between the trials in the effect of the randomized treatment allocation on the relative risks (RRs) of any of the main outcomes in any of the stenosis groups (Rothwell et al., 2003b). Data were therefore merged on 6092 patients with 35,000 patient-years of follow-up.
It is ironic that the precise purpose of CEA is to prevent stroke (and death), and paradoxically its major potential complication is peri-operative stroke (and death).
In the three trials of endarterectomy for symptomatic carotid stenosis, where postoperative complications were systematically reviewed, the overall pooled operative mortality was 1.1% (95% CI: 0.8–1.5), and the operative risk of stroke and death within 30 days of surgery was 7.1% (95% CI: 6.3–8.1) (Rothwell et al., 2003b).
A systematic review of 57 surgical case series, involving a total of 13,285 CEAs, indicated that the reported peri-operative risk of stroke and death varied widely from <1% to more than 30%, but was usually about 3–8%, and was on average 5.1% (95% CI: 4.6–5.6%) (Rothwell et al., 1996a,b; Goldstein et al., 1997). The risk was higher in surgical case series in which patients were assessed postoperatively by a neurologist (7.7% [95% CI: 5.0–10.2%], odds ratio [OR]: 1.62 [95% CI: 1.45–1.81]) (Rothwell, 1996a).
Based on the results from the trials, about one in 14 patients undergoing CEA for symptomatic carotid stenosis experiences a peri-operative stroke or dies in the peri-operative period. However, this risk cannot be generalized to one’s own institution, surgeons, or patients. In particular, personal experience, and recent case series, suggest that operative rates of CEA for symptomatic patients have declined considerably since NASCET and ECST were published (Matsen et al., 2006). A recent local prospective audit of a large number of patients undergoing CEA by each surgeon and centre is therefore required, and referring doctors (and patients) should have access to the peri-operative stroke and death rate of their prospective surgeon(s) and centres, derived from such an independent and rigorous audit (Goldstein et al., 1997). However, interpretation of unusually high or low operative risks must take into account the effects of chance and case mix. Otherwise, over-simplistic interpretation of crude results may lead to unjustified criticism of individual surgeons, and not to improvements in patient care.
There are several other important prognostic factors for stroke, and peri-operative stroke and death associated with CEA which may influence the decision to operate.
Prognostic Factors for Peri-operative Stroke or Death
Patient factors The risk of peri-operative stroke is greater in patients with symptomatic carotid stenosis than asymptomatic carotid stenosis. However, patients with symptoms of minor stroke lasting less than 7 days have a lower risk than those with stroke lasting longer, while patients with ocular ischaemia due to carotid stenosis (e.g. amaurosis fugax) have the lowest risk of peri-operative stroke or death (Rothwell et al., 1996b).
Other patient factors which have been associated with an increased risk of peri-operative stroke and death in trials of CEA for symptomatic carotid stenosis (Table 20.3) are as follows:
1. Female gender (perhaps because of smaller carotid arteries, perhaps more difficult to operate on).
2. Systolic blood pressure (SBP) >180 mm Hg (perhaps increased risk of reperfusion injury and cerebral haemorrhage).
3. Peripheral arterial disease (a marker of atherosclerotic plaque burden).
4. Occlusion of the contralateral ICA (indicates poor collateral cerebral circulation).
5. Stenosis of the ipsilateral external carotid artery (poor collateral circulation) (Rothwell et al., 1997). The risk of peri-operative stroke or death associated with CEA does not appear to be related to the degree of ipsilateral internal carotid stenosis (Table 20.4).
Surgical factors The surgical factors which may be associated with an increased risk of peri-operative stroke or death are: (a) inexperience due to low surgeon and hospital case volumes (Feasby et al., 2002); and (b) undertaking CEA in the very acute phase of stroke in evolution and crescendo TIAs, and during coronary artery surgery for patients with angina whose carotid stenosis was discovered during preparation for coronary artery surgery (Bond et al., 2003). For neurologically stable patients, such as those enrolled into the trials, there was no evidence of any increase in operative risk in patients operated on within 2 weeks of their last event (Rothwell et al., 2004). Moreover, in a systematic review of surgical case series, early surgery in neurologically stable patients was not associated with an increased operative risk (Bond et al., 2003). These data have led to pressure on surgeons to operate very soon after the onset of symptoms. However, operating on patients with unstable plaque might be hazardous and recent registry data have suggested that the risk is significantly increased in patients operated on within 2 days of symptoms (Strömberg et al., 2012). Thus, the optimal time to operate is probably on the third day after minor symptoms, which allows time for antiplatelet therapy to become active.
Local Adverse Effects of CEA
Other specific adverse effects of CEA, besides those inherent in any operation, include:
Prognostic Factors for Peri-operative Cranial Nerve Palsy
Patients undergoing CEA for recurrent carotid stenosis have an increased risk of cranial nerve injury and wound haematoma (Bond et al., 2003).
Because CEA carries a significant risk of peri-operative stroke and death, CEA is only superior to medical therapy alone when the risks of stroke over time on medical therapy are significantly higher than the combined peri-operative risk from CEA plus any stroke that occurred thereafter despite CEA. As discussed above, these risks vary from one individual to another. The peri-operative risks of surgery after randomization in the RCTs for recently symptomatic carotid stenosis initially exceeded the risk of stroke on medical treatment for a period of time and it is only after the risks of the two treatments cross over for a similar period of time that surgery shows a net benefit over medical therapy. The way in which the long-term risk of events on medical treatment alone varies with stenosis severity, while the long-term risk of events after surgery remains fairly constant with stenosis severity, is illustrated in Figure 20.1.
Figure 20.1 Effects of carotid endarterectomy (thick line) vs medical treatment alone (thin line) over 10-year follow-up at different degrees of symptomatic carotid stenosis analysed by intention-to-treat using data pooled from ECST, NASCET, and VA309.
It is a striking feature of patients with severe (70–99%) recently symptomatic carotid stenosis that although the risk of recurrent stroke is higher than attributable to lesser degrees of stenosis, the long-term risk of stroke on medical treatment alone flattens off over time, being much higher in the first 2 years after the index symptom than it is in subsequent years (Figure 20.1). This effect is less pronounced in patients with <70% stenosis. Thus, the net benefit of surgery is greatest in patients with severe stenosis (excluding near-occlusion) but the absolute risk reduction does not increase beyond 3 years of follow-up (Rothwell et al., 2003b). In contrast, the absolute risk reduction achieved by surgery for moderate degrees of stenosis (50 to 79%) increases over time up to at least 8 years after randomization (Figure 20.2).
Figure 20.2 Net benefit of carotid endarterectomy treatment above that of medical treatment alone shown as absolute risk reduction for various trial outcomes at different degrees of symptomatic carotid stenosis and follow-up times, analysed by intention-to-treat using data pooled from ECST, NASCET, and VA309. Near-occl = near-occlusion.
Evidence from the Cochrane Library
The most recent systematic review for the Cochrane Library was published in 2017 and included data from the 6343 patients enrolled in the NASCET, ECST, and VA trials. No new completed studies were identified. The authors performed a pooled analysis of 6092 participants in whom individual data were available using a re-analysis of all the baseline angiograms for stenosis severity using the NASCET method and a uniform definition of outcome events (Orrapin and Rerkasem, 2017).
Mild carotid stenosis (NASCET <50%) For patients with very mild stenosis measuring <30%, surgery increased the 5-year risk of any stroke or operative death with a risk ratio (RR) of 1.25 (95% CI: 0.99–2.01). For patients with mild stenosis measuring 30–49%, there was no evidence of benefit or harm, with a risk ratio of 0.97 (95% CI: 0.79–1.19). The assumed event rates for this degree of stenosis were 21% with best medical treatment alone and 20% with additional carotid surgery (Figure 20.3).
Moderate carotid stenosis (NASCET: 50–69%) For patients with moderate stenosis, surgery significantly reduced the risk of any stroke or operative death by 23% (RR 0.77, 95% CI: 0.63–0.94). This corresponded to a reduction from an assumed 5-year risk of 23% to 18% with surgery and a number needed to treat (NNT) of around 20 to prevent one event (Figure 20.3).
Figure 20.3 Forest plots showing the effects of CEA plus best medical therapy (surgery) vs best medical therapy alone (no surgery) in patients with recently symptomatic carotid stenosis in subgroups of severity of carotid stenosis (1, near occlusion; 2, 70–99%; 3, 50–69%; and 4, 30–49%).
Severe carotid stenosis (NASCET 70–99%) For patients with severe stenosis, excluding those with near-occlusion, surgery reduced the risk of any stroke or operative death by 47% (RR 0.53, 95% CI: 0.42–0.67%). This corresponds to a reduction from an assumed 5-year risk of 29% to 15% with surgery and a NNT of around 7 to prevent one event (Figure 20.3).
Near-occlusion In patients with near occlusion (defined as very severe stenosis associated with distal luminal collapse of the ICA), there was no evidence of overall benefit or harm from surgery (RR 0.95, 95% CI: 0.59–1.53).
Other outcome measures The Cochrane review also analysed several other outcomes, including ipsilateral carotid territory stroke or operative stroke or death, and disabling or fatal ipsilateral ischaemic stroke or operative stroke or death. The conclusions in relation to the benefit of surgery at different degrees of stenosis were very similar to those discussed above in relation to any stroke or operative death.
Although endarterectomy reduces the RR of stroke or operative death by 47% over the next 5 years in patients with a recently symptomatic severe (70–99%) carotid stenosis, the absolute risk reduction (ARR) is relatively small (14%), because only 29% of these patients have a stroke on medical treatment alone (and 15% have a stroke after CEA).
The operation is therefore of no value for at least two-thirds (71%) of patients who, despite having a severe (70–99%), recently symptomatic carotid stenosis, are destined to remain stroke free for the next 5 years with best medical therapy alone (i.e. without surgery) and can only be harmed by surgery.
It would, therefore, be useful to be able to identify in advance, and operate on, only the 14% of patients (one in seven) with severe symptomatic carotid stenosis who are going to benefit over the next 5 years.
We therefore need to be able to more precisely identify those patients with severe symptomatic carotid stenosis who have a very high risk of stroke on medical treatment alone, coupled with a relatively low operative risk. Simply relying on the presence or absence of recent carotid territory ischaemic symptoms and the degree of carotid stenosis is not good enough (although a good start).
When a systematic review of RCTs reveals a substantially significant overall treatment effect of several standard deviations, as is the case of CEA for severe symptomatic carotid, there is sufficient statistical power to perform subgroup analyses with some confidence in the results (other than purely hypothesis-generating ‘data dredging’) (see ‘Subgroup analyses’, Chapter 2, page 17).
Subgroup analyses of pooled data from ECST and NASCET revealed that the effectiveness of CEA in patients with severe symptomatic carotid stenosis was modified significantly by three clinical variables: the patients’ sex (P = 0.003), age (p = 0.03), and time from the last symptomatic event to randomization (p = 0.009) (Rothwell et al., 2004). Benefit from surgery was greatest in men, patients aged ≥75 years, and patients randomized within 2 weeks after their last ischaemic event, and fell rapidly with increasing delay. For patients with ≥50% stenosis, the number of patients needed to undergo surgery (NNT) to prevent one ipsilateral stroke in 5 years was 9 for men versus 36 for women, 5 for age ≥75 versus 18 for age <65 years, and 5 for patients randomized within 2 weeks after their last ischaemic event versus 125 for patients randomized at >12 weeks (Rothwell, 2005). These observations were consistent across the 50–69% and ≥70% stenosis groups and similar trends were present in both ECST and NASCET (Figure 20.4).
Figure 20.4 Numbers of events and absolute risk reduction (ARR) in 5-year actuarial risk of ipsilateral carotid territory ischaemic stroke or any stroke or death within 30 days after trial surgery according to pre-defined subgroups in all patients, patients with 50–69% stenosis, and those with ≥70% stenosis.
Women had a lower risk of ipsilateral ischaemic stroke on medical treatment and a higher operative risk in comparison to men. For recently symptomatic carotid stenosis, surgery is very clearly beneficial in women with ≥70% stenosis, but not in women with 50–69% stenosis (Figure 20.5). In contrast, surgery reduced the 5-year absolute risk of stroke by 8.0% (3.4–12.5) in men with 50–69% stenosis. This sex difference was statistically significant even when the analysis of the interaction was confined to the 50–69% stenosis group.
Figure 20.5 Absolute risk reduction (ARR) in 5-year cumulative risk of ipsilateral carotid territory ischaemic stroke or any stroke or death within 30 days after surgery plotted separately for patients with 70–99% stenosis (excluding near-occlusion) and those with 50–69% stenosis stratified by the time from last symptomatic event to randomization. Numbers above bars indicate actual absolute risk reduction. Vertical bars are 95% CIs.
Benefit from CEA increased with age in patients with recently symptomatic stenosis, particularly in patients aged over 75 years because of their high risk of stroke on medical treatment without a substantially increased risk of peri-operative stroke (Figure 20.4). These findings are consistent with a systematic review of all published surgical case series which reported no increase in the operative risk of stroke and death in older age groups (Rothwell, 2005).
There is therefore no justification for withholding CEA in patients aged over 75 years who are deemed to be medically fit to undergo surgery.
Timing of CEA
Given the high early risk of stroke on medical treatment alone after a TIA or minor stroke in patients with carotid disease (Lovett et al., 2003, 2004; Coull et al., 2004) and the lack of an increased operative risk in neurologically stable patients (see above), early surgery is likely to be most effective. The pooled analysis of data from the trials shows that the benefit from endarterectomy is greatest in patients randomized within 2 weeks of their last event (Figures 20.4 and 20.5). This was particularly important in patients with 50–69% stenosis, where the reduction in the 5-year risk of stroke with surgery was considerable in those who were randomized within 2 weeks of their last event (14.8%, 95% CI: 6.2–23.4), but minimal in patients randomized later (Figure 20.5). It should be borne in mind that surgery was often delayed by up to a week or two after patients were randomized in the RCTs and therefore the benefit of surgery might extend for a week or two longer between symptoms and the date of surgery.
Benefit from surgery is probably also greatest in patients presenting with stroke, intermediate in those with cerebral TIA, and lowest in those with retinal events (see Figure 20.4). There was also a trend in the trials towards greater benefit in patients with irregular plaque than a smooth plaque.
The outcome within the ECST of individuals with different characteristics that influence future risk of stroke have been incorporated into a model that predicts the 5-year ipsilateral risk of stroke on medical treatment alone in patients with recently symptomatic carotid stenosis. Although this model was derived from the ECST, it was validated externally using data from NASCET after the ECST angiograms and outcome event definitions were re-analysed to match NASCET (Rothwell and Warlow, 1999; Rothwell et al., 2005). The ECST model showed an excellent fit with the NASCET data (Figure 20.6).
Figure 20.6 Reliability of the ECST risk model in predicting the 5-year rate of ipsilateral ischaemic stroke on medical treatment alone in patients with 50–99% stenosis included in NASCET (filled squares). The 30-day risk of operative stroke in patients randomized to surgery in NASCET stratified by predicted risk on medical treatment is shown to demonstrate that patients at high risk when treated medically did not have a high risk from surgery.
Rothwell and colleagues (2005) provided colour-coded tables to allow clinicians to quickly assess an individual patient’s 5-year risk of ipsilateral stroke treated medically from five clinically important variables (Figure 20.7). It should be noted that because the data used to calculate the risk of stroke plotted in Figures 20.6 and 20.7 was derived from ECST, the results are likely to overestimate the current risks of stroke on modern medical therapy in patients being considered for surgery in the current era. Various stands of evidence suggest that current rates of stroke in patients with carotid stenosis treated optimally are now at least half those recorded in ECST and NASCCET (Cheng and Brown, 2017).
Figure 20.7 Tables showing predicted absolute risk of ipsilateral ischaemic stroke at 5 years in patients with recently symptomatic carotid stenosis treated with medical therapy alone according to five clinically important patient characteristics in (A) men and (B) women. This table was derived from ECST data using a Cox regression model. The terms Stroke, TIA, and Ocular in the tables refer to the most severe symptomatic ipsilateral ischaemic event in the past 6 months.
Types of CEA
Although CEA dramatically reduces the risk of stroke in patients with recently symptomatic severe ICA stenosis, the benefit is dependent on maintaining a lower peri-operative stroke risk. For many years, the optimal anaesthetic technique during CEA was controversial. Those favouring local or regional anaesthesia argued that it was safer in terms of medical complications and allowed the patient to be monitored for neurological deficits during surgery. The issue was resolved by the results of a large randomized trial (General Anaesthetic vs Local Anaesthetic, GALA), which showed no significant difference in major outcomes between patients allocated to either anaesthetic technique (GALA Trial Collaborative Group, 2008). A slight reduction in stroke with local anaesthesia was offset by an increase in myocardial infarction. Thus, the choice between the two techniques is a matter for the individual surgeon taking into account their own and the patient’s preference.
CEA is conventionally undertaken by a longitudinal arteriotomy. Eversion CEA, which employs a transverse arteriotomy and reimplantation of the carotid artery, has been reported to be associated with low peri-operative stroke and restenosis rates but an increased risk of complications associated with a distal intimal flap.
A Cochrane review of 5 RCTs evaluating whether eversion CEA was safe and more effective than conventional CEA in a total of 2465 patients reported no significant differences between eversion and conventional CEA techniques in the rate of peri-operative stroke and/or death (1.7% eversion CEA vs 2.6% conventional CEA, OR 0.44, 95% CI: 0.10–1.82), and stroke during follow-up (1.4% vs 1.7%; OR 0.84, 95% CI: 0.43–1.64) (Cao et al., 2001).
Although eversion CEA was associated with a significantly lower rate of restenosis >50% during follow-up than conventional CEA (2.5% eversion vs 5.2% conventional, Peto OR 0.48, 95% CI: 0.32–0.72), there was no evidence that the eversion technique for CEA was associated with a lower rate of neurological events when compared to conventional CEA.
There were no statistically significant differences in local complications between the eversion and conventional group.
A more recent systemic review came to very similar conclusions from the 5 available RCTs, but also included published case series comparing outcomes of eversion with conventional CEA. Among the case series (N = 20,270 patients), eversion CEA was associated with significantly lower rates of peri-operative 30-day rates of stroke or death (2.26% vs 4.32%, OR 0.52, 95% CI: 0.44–0.61) and late restenosis >50% (2.34% vs 4.68%, OR 0.49, 95% CI: 0.25–0.94) (Paraskevas et al., 2018). There were no significant differences in cranial nerve injury or neck haematoma rates.
Conventional CEA included patients undergoing either primary closure or patch closure of the arteriotomy after removal of the plaque. In a separate analysis of both the RCTs and the case series, the authors found no difference between outcome event rates when eversion CEA was compared with the conventional CEA patients who had patch closure (30-day stroke or death rate: 2.35% vs 3.31%, OR 0.64, 95% CI: 0.35–1.18) (Paraskevas et al., 2018).
Interpretation of the Evidence
There is insufficient evidence from randomized trials to reliably determine the RRs and benefits of eversion and conventional CEA. It is possible that carotid eversion is associated with a lower risk of long-term carotid occlusion and restenosis, but it is still unclear whether this is associated with a lower rate of subsequent neurological events. The recent meta-analysis of case series suggests that eversion CEA is safer and more durable than conventional CEA with primary closure, but eversion CEA has similar outcomes to conventional patched CEA. Patching is discussed in more detail below.
Implications for Practice
Until more reliable evidence is available, the choice of the surgical technique for CEA should depend on the experience and preference of the operating surgeon.
Implications for Research
Further randomized trials are needed to more precisely define the relative and absolute benefits and risks of eversion and conventional CEA, and establish the clinically relevance (or not) of restenosis of the carotid artery (that was previously operated on) as a cause of subsequent stroke. Studies analysing the costs of eversion and conventional CEA are also needed.
Routine or Selective Carotid Artery Shunting for CEA (and different methods of monitoring in selective shunting)
One of the limitations of CEA is the risk of peri-operative stroke, usually ipsilateral carotid territory ischaemic stroke. Peri-operative stroke can result from temporary interruption of cerebral blood flow while the carotid artery is clamped during CEA if collateral flow from the contralateral carotid artery or via the vertebrobasilar circulation is compromised. The duration of interrupted blood flow to the brain can be minimized by bridging the clamped section of the artery with a shunt. Potential disadvantages of shunting include complications such as air and plaque embolism and carotid artery dissection, and an increased risk of local complications such as nerve injury, haematoma, infection, and long-term restenosis. However, reliable data on these risks and the potential benefits are limited. Consequently, some surgeons advocate routine shunting, whereas others prefer to use shunts selectively or avoid them altogether.
A Cochrane systematic review of the effect of routine vs selective, or never, shunting during CEA identified six trials involving 1270 patients (Chongruksut et al., 2014). Three trials with 686 patients compared routine shunting with no shunting, one trial with 200 patients compared routine shunting with selective shunting, one trial with 253 patients compared selective shunting with and without near-infrared refractory spectroscopy monitoring, and the other trial with 131 patients compared shunting with a combination of electroencephalographic and carotid pressure measurement with shunting by carotid pressure measurement alone. In general, the authors considered that the trial quality was poor.
For routine versus no shunting, there was no significant difference in the rate of all stroke, ipsilateral stroke or death up to 30 days after surgery, although data were limited.
No significant difference was found between selective shunting with and without near-infrared refractory spectroscopy monitoring, although again the numbers were small.
There was no significant difference in the risk of ipsilateral stroke in patients selected for shunting with the combination of electroencephalogram (EEG) and carotid pressure assessment compared with pressure assessment alone, although again the data were limited.
Implications for Practice
There is insufficient evidence from RCTs to support or refute the use of routine or selective shunting during CEA and there is little evidence to support the use of one form of monitoring over another in selecting patients requiring a shunt. The use of EEG monitoring combined with carotid pressure assessment may reduce the number of shunts required without increasing the stroke rate, but more data are required to prove this.
Implications for Research
A large, multicentre randomized trial is required to assess whether shunting reduces the risk of peri-operative and long-term death and stroke. Even a modest 25% reduction in the RR of peri-operative stroke or death would result in approximately 15 fewer strokes and deaths per 1000 patients undergoing endarterectomy. However, to detect this reliably (80% power, 5% significance level) would require between 3000 and 5000 patients.
Two policies could be considered: routine shunting for all patients undergoing CEA or selective shunting in those at high risk of intra-operative cerebral ischaemia. The trial needs to be truly randomized, have long-term follow-up (several years), and have blinded outcome assessment, preferably by neurologists. Patients should be stratified by age, sex, degree of ipsilateral and contralateral internal carotid stenosis, the experience of the surgeon, the use of patching, and, in selective shunting, the method of monitoring of cerebral ischaemia.
As regards the method of monitoring in selective shunting, until the efficacy of shunting has been demonstrated, further trials of the method of monitoring are probably not merited. However, a systematic review of the sensitivity and specificity of the various methods of monitoring for cerebral ischaemia would be worthwhile to identify the best method of monitoring to be used in any trial of selective shunting.
When undertaking a CEA, many surgeons use a patch of autologous vein, or synthetic material, to close the artery, to enlarge the lumen, and, so they hope, to reduce the risk of restenosis and stroke. However, it remains uncertain whether carotid patch angioplasty (with either a venous or a synthetic patch) reduces the risk of carotid artery restenosis and subsequent ischaemic stroke compared with primary closure of the ICA.
A Cochrane systematic review of RCTs assessing the safety and efficacy of routine or selective carotid patch angioplasty compared with CEA with primary closure identified 7 trials involving 1967 patients undergoing 2157 operations (Rerkasem and Rothwell, 2009). Follow-up varied from hospital discharge to 5 years. The quality of trials included in the review was judged to be generally poor. Nevertheless, use of a carotid patch was associated with a reduction in the risk of ipsilateral stroke both during the peri-operative period (OR 0.31, 95% CI: 0.15–0.63) and over long-term follow-up (OR 0.32, 95% CI: 0.16–0.63). Patching was also associated with a lower risk of peri-operative carotid occlusion (OR 0.18, 95% CI: 0.08–0.41), and decreased restenosis during long-term follow-up (OR 0.24, 95% CI: 0.17–0.34).
Very few arterial complications, including haemorrhage, infection, cranial nerve palsies and pseudo-aneurysm formation, were recorded with either patch or primary closure.
No significant correlation was found between use of patch angioplasty and the risk of either peri-operative or long-term all-cause death rates.
Interpretation of the Evidence
Limited evidence suggests that carotid patch angioplasty might reduce the risk of peri-operative arterial occlusion and restenosis, and longer-term stroke or death.
Implications for Practice
Although some vascular surgeons do not routinely use patching in patients undergoing CEA, the present data from RCTs (albeit based on small numbers) appear to support a recommendation in favour of routine patching.
The use of selective patching (e.g. for very narrow arteries) has not been studied in RCTs, so no evidence-based recommendations can be made.
Implications for Research
The potential benefit of routine patching could be clinically important (up to 40 strokes prevented per 1000 patients treated) but in order to have reliable evidence on the risks and benefits of patching compared to primary closure, a large multicentre RCT will be required.
This trial should concentrate on clinical outcomes (deaths, all strokes, particularly fatal or disabling strokes, and ipsilateral strokes) as opposed to restenosis and have long-term follow-up (perhaps 5 years).
Assuming a 30-day risk of stroke or death of 5%, the trial would need to recruit about 3000 patients to have an 90% chance of detecting a reduction in the absolute risk of death or stroke to 2.5% (this number would also give a >90% chance of detecting a reduction in the risk of stroke or death at 5 years from 25% to 20%). Such a trial should use a secure method of randomization and be performed on a truly intention-to-treat basis with complete follow-up of all patients. Patients rather than arteries should be randomized so that the number of deaths and strokes is reported on a patient basis rather than an artery basis. Clinical follow-up should be blinded with independent assessment of strokes, preferably by neurologists. The results should be analysed according to the degree of narrowing of the artery and whether the patient had had a previous stroke or TIA or not. It would be possible to use a factorial design for such a trial so that some other procedure could be tested simultaneously, such as routine shunting. Until the benefit of carotid patching in terms of clinical outcomes for the patient is established, any future trials should include a control group of primary closure.
Although CEA is effective in patients with symptomatic carotid stenosis, it is not clear whether different surgical techniques affect the outcome, and whether the use of carotid patch angioplasty is superior to primary closure in reducing the risk of restenosis and improving both short- and long-term clinical outcome. Consequently, many vascular surgeons use carotid patching either routinely or selectively. Among those who do use carotid patching, however, there is debate over the choice of patch material. Vein patching (usually harvested from the saphenous vein and sometimes from the jugular vein) is favoured by some because it is easily available, easy to handle and possibly has a greater resistance to infection. Synthetic material, such as Dacron or polytetrafluoroethylene (PTFE), is favoured by others who feel that it offers a lower risk of patch rupture and aneurysmal dilatation, and also that it spares the morbidity associated with saphenous vein harvesting and leaves the vein which may be required for coronary bypass grafting at a later date. It is also possible that one type of synthetic material is better than the other. Furthermore, there are less commonly used materials such as bovine pericardium which have yet to be widely accepted.
A systematic review of 13 RCTs of the safety and efficacy of different materials for carotid patch angioplasty in 2083 operations identified seven trials comparing vein closure with PTFE closure, and six comparing Dacron grafts with other synthetic materials (Rerkasem and Rothwell, 2010). In most of the trials a patient could be randomized twice so that each carotid artery was randomized to a different treatment group.
There were no significant differences in the outcomes between vein patches and synthetic materials, except that there were fewer pseudoaneurysms associated with synthetic patches than vein patches (OR 0.09, 95% CI: 0.02–0.49). However, the numbers were small and the clinical significance of this finding uncertain.
Compared to other synthetic patches, Dacron was associated with a higher risk of peri-operative stroke or TIA (p = 0.03); restenosis at 30 days (p = 0.004); peri-operative stroke (p = 0.07); and peri-operative carotid thrombosis (p = 0.1). During follow-up for more than 1 year, there were also significantly more strokes (p = 0.03) and arterial restenoses (p < 0.0001) with Dacron, but again the numbers of outcomes were small and the significance uncertain.
Interpretation of the Evidence
It is likely that the differences between different types of patch material are very small. Consequently, more data than are currently available will be required to establish whether any differences do exist.
Some evidence exists that PTFE patches may be superior to collagen impregnated Dacron grafts in terms of peri-operative stroke rates and restenosis. However, the evidence is limited and more studies that compare different types of synthetic graft are required to make firm conclusions.
Pseudo-aneurysm formation might be more common after use of a vein patch compared with a synthetic patch.
Implications for Practice
There is no evidence to support the use of vein over synthetic patch material in CEA. The decision of which type of patch to use, if any, remains a matter of individual preference. However, if synthetic material is used, the currently available (limited) evidence from a single trial appears to show benefits from PTFE as opposed to Dacron material.
Antiplatelet drugs are effective and safe for TIA and ischaemic stroke patients in preventing recurrent vascular events, and for patients undergoing vascular surgical procedures in reducing the risk of graft or native vessel occlusion. However, this does not necessarily mean they are effective and safe after CEA.
Antiplatelet vs Control
A systematic review of RCTs evaluating whether antiplatelet agents are safe and beneficial after CEA (Engelter and Lyrer, 2003, 2004) identified six trials of antiplatelet therapy administered for at least 30 days after CEA in a total of 907 patients who were followed up for at least 3 months. Most trials used high dose aspirin alone or in combination with other antiplatelet drugs.
Antiplatelet therapy was associated with a significant reduction in the odds of stroke of any type during follow-up (OR 0.58, 95% CI: 0.34–0.98; p = 0.04) but not death (OR 0.77, 95% CI: 0.48–1.24), intracranial haemorrhage (OR 1.71, 95% CI: 0.73–4.03), or other serious vascular events.
Given the high risk of recurrence in patients with recent symptoms despite aspirin therapy, it is logical to try to improve medical therapy in patients waiting for CEA. One such approach is to use combined antiplatelet therapy. In one study, 100 patients were assigned aspirin 150 mg per day for 4 weeks prior to CEA and then, on the night before CEA, they were randomized to placebo plus long-term aspirin 150 mg per day (n = 54) or clopidogrel 75 mg per day plus long-term aspirin 150 mg per day (n = 46) (Payne et al., 2004). Compared with placebo plus aspirin, assignment to clopidogrel plus aspirin was associated with a small (8.8%) but significant reduction in platelet response to adenosine diphosphate (ADP) (p = 0.05), a 10-fold reduction in the RR of patients having >20 emboli detected by TCD within 3 hours of CEA (OR 10.23, 95% CI: 1.3–83.3; P = 0.01), and a significantly increased time from flow restoration to skin closure (an indirect marker of haemostasis) (P = 0.04). Surgeons often report increased bleeding in patients taking combined aspirin and clopidogrel, but in this study there was no increase in bleeding complications or blood transfusions.
These results are supported by the Clopidogrel and Aspirin for Reduction of Emboli in Symptomatic carotid Stenosis (CARESS) trial, which has shown that among patients with symptomatic carotid stenosis (before, not after, CEA), that the addition of clopidogrel to aspirin reduces the incidence of cerebral microemboli before (not after) CEA. The CARESS study was a randomized, double-blind, controlled trial comparing clopidogrel plus aspirin versus aspirin in 107 patients with recently (past 3 months) symptomatic carotid artery stenosis >50% and at least one characteristic microembolic signal (MES) detected during a 1-hour TCD recording of the ipsilateral middle cerebral artery (MCA) before randomization. At randomization, patients were assigned either a 300 mg loading dose of clopidogrel (four tablets), followed by 75 mg clopidogrel once daily from day 2 to day 7 ± 1, or a matching placebo loading dose, and once daily placebo from day 2 to day 7 ± 1. Patients in both arms also received 75 mg aspirin (ASA) once daily from day 1 to day 7 ± 1 (on top of clopidogrel or placebo). All study drugs were administered orally. The primary efficacy endpoint was the occurrence of ≥ 1 MES versus none (MES positive or negative patient), detected by off-line analysis (by central reading centre) of the 1-hour recording carried out on day 7 ± 1. Intention-to-treat analysis revealed a significant reduction in the primary endpoint: 43.8% of dual-therapy patients were MES positive on day 7, as compared with 72.7% of monotherapy patients (relative risk reduction 39.8%, 95% CI: 13.8–58.0; p = 0.0046) (Markus et al., 2005).
Interpretation of the Evidence
Antiplatelet drugs reduce the odds of stroke after CEA, but not other major outcomes. However, a statistically significant hazardous effect of antiplatelet therapy may have been missed because of the relatively small sample size.
There is some evidence that the combination of clopidogrel and aspirin might be more effective than aspirin alone in reducing post-operative thromboemboli, as detected by TCD (Payne et al., 2004). Data from several studies suggest that the presence of asymptomatic embolization, as measured by TCD evidence of MESs, might predict future strokes or TIAs in patients with symptomatic stenosis; the risk is reported to be 8- to 31-fold higher in patients who are MES positive (or who have ≥1 MES per TCD recording) than patients who have MES negative ICA stenosis (Valton et al., 1998; Molloy and Markus, 1999).
Implications for Practice
Antiplatelet therapy should be prescribed for all patients after CEA. Aspirin should probably be the first-line agent. Other antiplatelet agents such as clopidogrel and extended-release dipyridamole plus aspirin have also been shown to be effective in patients with TIA and ischaemic stroke not undergoing CEA (Chapter 9), and are likely (based on reasoning [Chapter 2], but not evidence) to also be effective after CEA. Many physicians and surgeons routinely prescribe the combination of aspirin and clopidogrel after TIA or stroke up to the time of CEA and then continue single therapy after surgery.
Implications for Research
Further research should focus on the effectiveness and safety of combinations of aspirin with other antiplatelet drugs.
There is preliminary evidence to suggest that a short period of treatment with aspirin and clopidogrel, for perhaps 1 month, might be more effective than aspirin during the acute phase of the event when patients with symptomatic carotid stenosis are at high risk of recurrent ischaemic events (Payne et al., 2004; Markus et al., 2005). This requires confirming in an adequately powered, dedicated study in a large number of patients with the occurrence of recurrent stroke, other serious vascular events, and bleeding complications as the primary outcome measure.
There is no evidence to support the use of anticoagulation in patients with recently symptomatic carotid stenosis who are in sinus rhythm (Chapter 10). Warfarin with a target international normalized ratio (INR) of 3–4.5 was harmful in the Stroke Prevention in Reversible Ischaemia Trial (SPIRIT) Study Group (1997), and there was no additional benefit over aspirin from warfarin at a mean INR of 1.8 (target INR 1.4–2.8) in the WARSS trial (Warfarin Aspirin Recurrent Stroke Study Group, 2001) (Chapter 19).
No trials have evaluated warfarin vs aspirin specifically in patients with carotid disease (carotid stenosis [>50%] was an exclusion criterion in the WARSS trial).
Implications for Practice
As the effect of warfarin in patients with symptomatic carotid stenosis undergoing CEA is likely to be qualitatively similar to that seen in other patients with TIA and ischaemic stroke due to arterial disease (based on reasoning [Chapter 18], not evidence), there is no indication to use oral anticoagulation after CEA. However, an exception can be patients with TIA or ischaemic stroke who have both an apparently symptomatic carotid stenosis and atrial fibrillation.
Warfarin is effective and indicated in patients with TIA or ischaemic stroke who are in atrial fibrillation (Chapter 18). The decision to recommend anticoagulation and/or endarterectomy in these patients depends to some extent on whether the recent TIA or stroke is thought to be cardioembolic or due to carotid thromboembolism. This is can be difficult, if not impossible, to determine sometimes. But if the computed tomography (CT) or magnetic resonance imaging (MRI) (diffusion-weighted imaging, DWI) brain scan shows multiple recent infarcts in multiple vascular territories, or echocardiography reveals apical thrombus or atrial enlargement, the source is likely to be the heart and anticoagulation is probably indicated. Alternatively, if echocardiography is normal and perhaps if any ischaemic lesions on brain imaging are confined to the ipsilateral carotid territory, it may be more appropriate to recommend endarterectomy and antiplatelet therapy.
Lowering blood pressure effectively reduces the risk of recurrent stroke and other serious vascular events (Chapter 15) (PROGRESS Collaborative Group, 2001), but the effect in different aetiological subtypes of ischaemic stroke, such as symptomatic carotid stenosis, is unknown.
Among patients with bilateral severe, flow-limiting (≥70%) carotid stenosis who were randomized to best medical treatment alone in ECST and NASCET, the risk of subsequent stroke was significantly increased if their SBP at randomization was <130 mm Hg (hazard ratio [HR] 6.0, 95% CI: 2.4–14.7) or 130–149 mm Hg (HR 2.5, 95% CI: 1.5–4.4) compared with patients with bilateral non-flow-limiting carotid stenosis <70% (Rothwell et al., 2003c). There was no increase in stroke risk with higher SBP of 150 mm Hg or more. The 5-year risk of stroke in patients with bilateral ≥70% stenosis was 64.3% in those with SBP <150 mm Hg (median value) versus 24.2% in those with higher blood pressures (p = 0.002).
However, among patients with bilateral severe, flow-limiting (≥70%) carotid stenosis who were randomized to CEA (plus best medical treatment) in ECST and NASCET, the 5-year risk of stroke in patients with bilateral ≥70% stenosis was not significantly different in those with SBP <150 mm Hg (median value) at randomization compared with higher blood pressures (13.4% vs 18.3%, P = 0.6).