Strokes are frequently devastating events with potentially fatal complications and significant disability, resulting in a major impact on the affected individual’s quality of life. It is well accepted that the presence of carotid stenosis is responsible for 20–30% of strokes.1 Great effort has therefore been invested in the search for treatments to reduce the stroke risk attributed to it. Although the mechanism of stroke related to carotid atherosclerosis is not entirely clear, analogous to the occurrence of an acute coronary syndrome, plaque rupture with subsequent thrombus adhesion followed by distal embolisation of either thrombotic material or atherosclerotic debris2 is most commonly thought to be the cause. With the aim of plaque removal or stabilisation, medical and interventional treatment options have been explored. It is assumed that, by stent implantation, the plaque composition is modified such that the plaque becomes less vulnerable to rupture and/or embolisation. This brief review will focus on carotid artery stenting (CAS).
The discussion of the evolution of percutaneous treatment modalities would be incomplete without a brief summary of the extensive surgical experience prior to the advent of percutaneous treatment.
Evidence Following the Completion of Surgical Trials
Similar to the initial experience with percutaneous carotid intervention, the first carotid endarterectomy (CEA) in 19543 was followed by many years of mixed and controversial results4,5 until the performance of three important trials in the early 1990s: the North American Symptomatic Carotid Endarterectomy Trial (NASCET),6 the European Carotid Stenosis Trial (ECST)7 and the Veterans Affairs Cooperative trial.8 These trials unequivocally demonstrated a benefit of CEA for patients with symptomatic carotid stenosis greater than 50%.
In the presence of stenosis >69%, fewer than 10 patients needed to be treated with CEA to prevent one stroke, provided the surgical risk was 6% or less. This benefit included men and women, patients over 74 years of age and patients with contralateral carotid occlusion. As expected, in the following years the number of CEAs increased substantially, including in patients with asymptomatic carotid stenoses.
CEA for asymptomatic patients continued to be the subject of considerable debate until the completion of two major trials: the Asymptomatic Carotid Atherosclerosis Study (ACAS), published in 1995,9 and the Asymptomatic Carotid Surgery Trial (ACST), published in 2004.10 To summarise the results, it was associated with a lower stroke risk than the rudimentary medical therapy available at the time. Although the relative risk reduction was impressive (approximately 50%) and statistically significant, the absolute risk reduction was in the range of 1% annually and the number needed to treat therefore much higher than in patients with symptomatic carotid stenosis. In addition, several findings dampened the enthusiasm for surgery for asymptomatic patients. First, patients over 74 years of age did not benefit, which may in part be related to the small number of enrolled patients in this age category. Second, the benefit did not become apparent until five years after the operation. Most importantly, however, to obtain the benefit documented in the surgical trials, the peri-operative risk could not exceed 3% in this low-risk population.
The excitement of a percutaneous treatment option was fostered by one important aspect unique to all of the above surgical trials: the surgeries were performed by experienced, high-volume surgeons at centres of excellence. Surgical centres were allowed to participate in the trials only after demonstration of a low peri-operative complication rate. A substantial number of patients were excluded from the trials for anatomical reasons or due to significant co-morbidity. Hence, these excellent results frequently could not be reproduced in subsequent registries that included patients with higher surgical risk and/or less experienced operators.11,12 To illustrate patient selection, in ACAS, for randomisation of one patient, 25 needed to be screened. Therefore, although the efficacy of CEA is well established for low-risk patients when performed by experienced operators, the benefit in asymptomatic patients with high surgical risk remains less clear.
In high-risk patients particularly, percutaneous treatment has several hypothetical advantages. First, it does not require the performance of general anaesthesia, with the inherent risk of adverse cardiac events. Second, it is not associated with the risk of cranial nerve injury and other local side effects, such as haematomas and delayed wound healing. Third, in the conscious patient direct neurological feedback is available in case of cerebral emboli, which may allow prompt attention and potential percutaneous rescue therapy. Finally, high cervical lesion location and prior neck radiation or radical neck dissection are not contraindications for percutaneous treatment. These potential advantages led to the performance of a series of non-randomised and randomised trials as well as registries to assess the safety and efficacy of percutaneous carotid intervention.
Early Registries and Non-randomised Trials
After the advent of carotid angioplasty in 197913,14 and stenting in 1989,15 multiple observational studies have been published. The results of these studies are to be interpreted with caution due to the non-randomised nature of the design, the inconsistent use of neuroprotection, the limited and incomplete follow-up and the absence of an independent assessment of neurological events. Without being complete, the following paragraphs will review the findings of some of the larger studies.
Roubin et al. published the results of 604 CAS procedures in 2001 (the largest published study at the time) for asymptomatic and symptomatic patients without neuroprotection, with a reported 30-day peri-procedural event rate (stroke and death) of 7.4% (major stroke 1%, minor stroke 4.8%).16 Among patients who survived the 30-day peri-procedural period, the three-year stroke rate was approximately 5%. These event rates were comparable to those in the surgical trials despite the fact that a substantial number of symptomatic patients were considered NASCET-ineligible based on anatomical or clinical characteristics. Similar to later trials, the outcome for octogenarians was less promising, with a 16% incidence of 30-day non-fatal stroke and death rate. There also appeared to be a learning curve with significantly fewer events in the later years of the trial, underscoring the importance of equipment and operator experience.
In 2003 the results of the Global Carotid Artery Stent Registry (GCSR)17 and a pooled analysis of 26 observational studies including approximately 3,500 patients18 were published. For the GCSR, the participating 52 centres returned surveys regarding their CAS results. The clinical characteristics (asymptomatic versus symptomatic), technique (type of stent, use of distal protection) and peri-procedural event rates (stroke and death) were reported for a little over 12,000 CAS procedures. The 30-day peri-procedural stroke and death rate was 4.7%. Although initially predominantly balloon-expandable stents without neuroprotection were used, in later years there was a shift towards self-expanding stents and the use of neuroprotection. There was a significantly lower stroke risk in patients treated with distal protection and in asymptomatic compared with symptomatic patients. Without protection, stroke- and procedure-related death was 5.27% compared with 2.27% with protection. Likewise, in the pooled analysis of 26 observational studies, the combined peri-procedural stroke and death rate was significantly lower with the use of distal protection (1.8% with versus 5.5% without). The peri-procedural event rate was similar in the subsequently published Prospective Registry of Carotid Angioplasty and Stenting (Pro-CAS)19 and European Long-term Carotid Artery Stenting (ELOCAS)20 registries. In all above-described registries, the procedural success rate was well above 95%.
The Cartotid Revascularization using Endarterectomy or Stenting Systems (CaRESS) results were published in 200321 and 2005.22 CaRESS was a non-randomised, prospective, multicentre study including predominantly (85%) high-risk asymptomatic and symptomatic patients. The choice of procedure, CAS versus CEA, was left to physician and patient preference. The self-expanding Wallstent was used exclusively with the Guardwire Plus filter protection system. The combined end-point of stroke and death at 30 days was equivalent for both treatment arms (3.5% for CEA and 2.1% for CAS), with no significant difference at one-year follow-up (13.6% for CEA and 10% for CAS). Likewise, there was no difference in the secondary end-points of restenosis, residual stenosis or revascularisation.
Device-specific Multicentre Registries
The creation of several device-specific multicentre registries for the treatment of high-risk patients followed the publication of these observational studies. The purpose of most of these registries was to either obtain US Food and Drug Administration (FDA) or CE device approval, or to fulfil FDA requirements for post-marketing approval. Results were compared with a historical CEA control group.
Given the unclear role of CAS in low-risk patients, in most of the registries low-risk patients are excluded. To date, the results of several registries have been published or presented. Without detailed description of the various specific trials and devices, taking the available data – from the Acculink for Revascularization of Carotids in High-risk Patients (ARCHeR),23 Boston Scientific EPI: A Carotid Stent for High-risk Surgical Patients (BEACH),24 Carotid Revascularization with ev3 Arterial Technology Evolution (CREATE),25 Evaluation of the Medtronic AVE Self-exanding Carotid Stent System with Distal Protection in the Treatment of Carotid Stenosis (MAVeRIC), Carotid Artery Revascularization using Boston Scientific EPI Filterwire EX/EZ and the EndoTex NexSTent (CABERNET)26 and Carotid Rx Acculink/Accunet Post Approval Trial to Uncover Unanticipated or Rare Events (CAPTURE)27 registries – into account, the 30-day myocardial infarction, stroke or death rate generally ranges from 4 to 8%. Once again, in the above registries the procedural success rate is well above 95%. Based on the registry data, it is reasonable to state that the results are similar to those achieved with historical CEA trials.
Randomised Trials Comparing Carotid Artery Stenting with Carotid Endarterectomy
Early randomised trials were not completed due to an unacceptably high procedure-related risk. The first randomised trial (Leicester trial)28 of carotid stenting in symptomatic patients without distal protection was stopped after recruitment of only 17 patients (planned recruitment was 300 patients). This was due to the occurrence of strokes in five of the 11 patients randomised to carotid angioplasty. The percutaneous procedures were performed without prior aortic arch imaging and with single antiplatelet therapy, despite the widespread use of dual antiplatelet therapy in most centres at the time. Moreover, the operator had performed only eight prior carotid procedures, the majority outside an experienced unit. This was followed by the Wallstent trial (published as an abstract only in 2001),29 which randomised symptomatic patients with a high-grade (>60%) carotid stenosis to CEA versus CAS without distal embolic protection. After recruitment of 219 out of the intended 700 patients, the trial was stopped due to a high one-year event rate (ipsilateral stroke, procedure-related or vascular death of 12.1% in the CAS arm versus 3.6% in the CEA arm).
These rather disappointing trials were followed by the first completed randomised multicentre Carotid and Vertebral Artery Transluminal Angioplasty Study (CAVATAS) trial published in 2001.30 Five hundred and four patients with symptomatic carotid stenosis and relatively low risk were randomised to percutaneous intervention without embolic protection (conducted before the advent of neuroprotection devices) versus CEA. The majority (74%) of patients were treated with balloon angioplasty as dedicated carotid stents were only available towards the completion of the trial. The event rate at 30 days (any stroke or death) did not differ significantly from the surgical group (10 versus 9.9%). Further, it remained comparable after three years of follow-up (no significant difference in ipsilateral stroke rates). However, it was higher than expected in the surgical group compared with results in the previous surgical trials (NASCET,6 ECST7). The authors hypothesised that the trial included patients at somewhat higher risk than those in NASCET and ECST. Importantly, the results are to be interpreted with caution due to the high confidence intervals. It is noteworthy that restenosis was higher in the endovascular group (14 versus 4%) and that the occurrence of haematomas and cranial nerve palsies were, as expected, higher in the surgical group. A particular strength of the trial was the independent neurological review that, up to this point, had not consistently been performed in the available registries and non-randomised studies. Despite the above-described limitations, the publication of the CAVATAS data fostered enthusiasm for carotid stenting and the performance of more trials. In the same year, the results of the smaller single-centre Kentucky trial randomising patients with symptomatic carotid stenosis to CAS versus CEA were published.31 In this trial, 104 patients were randomised; stents were used in all patients but distal protection in none. The peri-procedural risk in both groups was exceedingly low (one patient died in the CEA group, related to a myocardial infarction immediately following the surgery, otherwise there were no procedure-related strokes or deaths in either group).
Subsequently, in 2004, the results of the Stenting and Angioplasty with Protection in Patients at High Risk for Endarterectomy (SAPPHIRE) trial were published.32 This was the first trial with mandatory neuroprotection for CAS. Included were 334 patients with both symptomatic carotid stenosis greater than 50% and asymptomatic stenosis greater than 80%. All patients met the criteria for high surgical risk, due to either co-morbidities or difficult anatomy. Embolic distal filter protection and stents were used in all patients in the CAS arm. Agreement was required prior to randomisation regarding the patients’ suitability for either treatment modality among the interventionalist, surgeon and neurologist. The trial aimed to demonstrate the non-inferiority of CAS compared with CEA. The primary end-point – stroke, death or myocardial infarction at 30 days – was equivalent for CAS compared with CEA (12.1 versus 20.1%). There was a significant trend towards fewer events in the CAS arm (p=0.054) due to a higher incidence of peri- operative myocardial infarction in the CEA arm. The equivalency remained at one year. In 2008 the three-year follow-up data were published, and the stroke rates remained no different between the two treatment groups (combined stroke, death and myocardial infarction: 24.6 versus 30.3%).33 The need for target vessel revascularisation was lower in the CAS group and the incidence of cranial nerve injury was, as expected, higher in the CEA group.
The trial was criticised due to the inclusion of a high number (greater than 20% in both treatment groups) of patients with recurrent carotid stenosis. This bias potentially favours the CAS group as the risk of repeat carotid surgery, particularly cranial nerve injury, is known to be high. Also, the embolic risk associated with carotid stenting is potentially lower given the plaque composition (predominantly intimal hyperplasia). Additional criticism regarding the higher frequency of myocardial infarction in the surgical group was related to the double antiplatelet therapy used in the CAS group, which could have reduced the rate of myocardial infarctions had it been used in the surgical group as well.
Most importantly, however, the overwhelming majority of patients treated with CEA or CAS were asymptomatic. Therefore, the peri- procedural (30-day) event rates (death, stroke or myocardial infarction) in both treatment arms (4.8% in the CAS group and 9.8% in the CEA group) were of concern. The rates are high if one takes into consideration that in the surgical trials for asymptomatic carotid stenosis the peri-procedural event rate needed to be 3% or less to confer a benefit compared with medical management. Nevertheless, based on these results it is probably reasonable to conclude that for the specific patient population studied (high-risk), CAS is likely to be at least equivalent to CEA.
It remains to be determined whether in asymptomatic high-risk patients any interventional treatment (surgical or percutaneous) is superior to medical management, as no medical arm was included in the study. The results of this trial led to the approval of the stent used in SAPPHIRE as well as the embolic protection device for the treatment of high-risk symptomatic patients. In the same year the results of a randomised single-centre study of 85 asymptomatic patients who underwent CAS without distal protection or CEA were published. No deaths or cerebral ischaemic events occurred in either group.34
The Stent-Supported Percutaneous Angioplasty versus Carotid Endarterectomy (SPACE) trial was designed to demonstrate non-inferiority of CAS in patients with high-grade symptomatic carotid stenosis (embolic protection optional and used in only about one-quarter of patients).35 Whereas the patients in SAPPHIRE were considered high-risk, in the SPACE trial high-risk patients were excluded. The enrolment goal was 1,900 patients. After interim analysis of 1,200 patients, however, the occurrence of the primary end-point – stroke and death – was similar (6.84% in the CAS group versus 6.34% in the CEA group) with a statistically insignificant trend towards a higher incidence of disabling ipsilateral strokes in the CAS group (4 versus 2.9%).
Taking these findings into account, it was determined that at least 2,500 patients would need to be enrolled to demonstrate non-inferiority of stenting. The trial was hence terminated prematurely due to insufficient funding for enrolment of the projected number of patients needed. Importantly, at medium-term follow-up the annual risk of ipsilateral stroke was 1% or less, irrespective of the treatment modality.36 This suggests that, similar to the SAPPHIRE trial and the Endarterectomy versus Angioplasty in Patients with Symptomatic Severe carotid Stenosis (EVA-3S) trial (described below), CAS was as durable as CEA.36 EVA-3Sl
Similar to SPACE, in EVA-3S low-risk symptomatic patients were randomised to CAS or CEA.37 It was terminated prematurely (after enrolment of 527 of the intended 872 patients) due to a high event rate in the CAS group (30-day stroke and death rate of 9.6% compared with 3.9% in the CEA group).
The trial was criticised for the following reasons:
- neuroprotection was not used uniformly (in only 92%);
- only single-agent antiplatelet therapy was administered in a significant number of patients (15%); and
- the operator experience was limited.
This last point was illustrated by the low number of patients enrolled per institution, even at the more active treatment sites (10 patients annually), the high number of procedural failures (5%) and the average procedural time (70 minutes). Moreover, only two procedures needed to be performed with a given device in order to be able to use the device for trial purposes. It is noteworthy that the procedural event rate (stroke and death) was much higher (25%) in the patients treated without embolic protection – similar to the findings of the Carotid Revascularization Endarterectomy Versus Stent Trial (CREST) described below. In a long-term follow-up excluding the peri-procedural (30-day) event rate, the annual stroke rate was low (well below 1%) and equivalent between CAS and CEA, once again suggesting that after the peri-operative period CAS offers the same stroke protection as CEA.
What Have We Learned from the Available Randomised Trials?
First, when performed by experienced operators similar results can be achieved with CAS with neuroprotection compared with CEA for high- risk surgical patients. Second, the procedural risk for both CAS and CEA in asymptomatic high-risk patients is slightly higher than anticipated from ACAS and ACST. The benefit of revascularisation in this patient population is not clear, especially considering the currently available medical treatment options. Third, the use of neuroprotection seems to be important for the reduction of peri-procedural complications. Fourth, operator experience appears to determine the success rate and incidence of complications in CAS. Finally, at medium-term (two- to three-year follow-up), excluding the peri-procedural risk, the annual stroke incidence after CAS is probably comparable to that of CEA.
How Are Ongoing Trials Designed?
There are five large ongoing trials, CREST, ICSS, TACIT, ACST-2 and ACT-1.
CREST will be the largest trial that compares CAS versus CEA in patients with symptomatic and asymptomatic carotid stenosis. It is sponsored by the National Institute of Neurological Disorders and Stroke and National Institutes of Health. In contrast to the SAPPHIRE trial, but similar to the two main North American CEA trials, NASCET and ACAS as well as the prematurely discontinued EVA-3S and SPACE, it specifically excludes patients considered to be at high risk.
The primary end-point will be peri-procedural (30-day) death, stroke or myocardial infarction and ipsilateral stroke at one to four years. The intention is to enrol 2,500 patients. Potential operators have gone through a rigorous credentialling process to qualify for trial participation. In the lead-in phase, the interventionalists underwent training on an animal model, attended a carotid stenting course, submitted procedural and follow-up documentation for peer review and underwent observation by a device company representative for the first case. Finally, before permission to enrol patients, 20 carotid stenting procedures were required in the lead-in phase, with demonstration of excellent technique. The follow-up is expected to be completed by late 2009. After data review from the lead-in phase, it has become apparent that the peri-procedural stroke and death rate is similar in men and women (4.1 and 4.5%, respectively).39 The peri-procedural complication rate in octogenarians was high (12%),40 and this group of patients was therefore excluded after the lead-in phase. Upon completion, more information will be available regarding equivalency of carotid stenting to CEA in a low-risk population.
Investigators of the Transatlantic Asymptomatic Carotid Intervention Trial (TACIT) have planned to randomise 2,400 asymptomatic patients to CAS with neuroprotection, CEA or medical management. Proposed exclusion criteria are less stringent than in the low-risk trials and will reflect practice in the real world. The primary end-point will be the combination of 30-day death and five-year all-stroke risk. In addition to the comparison of CAS versus CEA in this patient population, results may provide clarification on whether any revascularisation, surgical or percutaneous – in addition to optimal medical therapy with potent statins, antiplatelet therapy and antihypertensives (and established treatment goals) – provides any advantage over optimal medical therapy alone. Furthermore, it will be the first trial to include systematic follow-up examinations of neurocognitive function. It has been suggested that high-grade carotid stenoses may be associated with more rapid neurocognitive decline.41,42 Likewise, given the frequent peri-procedural observation of cerebral emboli (detected by either intracranial Doppler or imaging modalities such as diffusion- weighted MRI), it is conceivable that revascularisation by either modality may accelerate neurocognitive decline.43 The result of this trial will hopefully provide an answer to which strategy is least likely to affect neurocognitive function. If any of the invasive treatment strategies proves to be superior, the absolute benefit will likely be of low magnitude. It will therefore be even more important to identify patients at very low risk of stroke with medical treatment only, in which case withholding revascularisation may be acceptable.
Plaque composition will be analysed via ultrasound to provide useful future guidance for the detection of a high-risk profile for the occurrence of strokes with medical therapy as well as for the occurrence of strokes with revascularisation. Surgeons will have to demonstrate an event rate (stroke or death) in the year previous to the trial of less than 3% for CEA. Interventionalists must demonstrate an event rate of less than 5% with a documented performance of at least 25 CAS procedures. This will help to ensure adequate experience in the invasive treatment arms. Enrollment to this trial has not yet begun.
The Asymptomatic Carotid Surgery Trial (ACST-2) is sponsored primarily by the UK National Institute for Health Research and will randomise 5,000 asymptomatic patients to CAS or CEA. The primary end-point is stroke and death at 30 days and stroke or death at five years. There are few specific pre-defined exclusion criteria beyond physician judgement that the risk of either CAS or CEA would outweigh the potential benefit. This study has begun enrolment.
The Asymptomatic Carotid Trial (ACT-1) will randomise 1,540 asymptomatic patients at standard (low) risk to CAS versus CEA (3:1). The primary end-point is stroke, myocardial infarction or death at 30 days, as well as ipsilateral stroke between days 31 and 365. In addition, there will be a quality of life assessment and cost analysis. Enrolment has started but the results of the trial are not expected for the next five years or more. The trial is device-company-sponsored (Abbott Vascular) and one stent and embolic protection system are used exclusively.
The Asymptomatic Carotid Stenting Study (ICSS) is sponsored by the UK Stroke Association, Medical Research Council, Sanofi-Synthélabo and the EU. It compares CEA versus CAS with neuroprotection in patients with symptomatic carotid stenosis. All risk categories have been included. The device choice (stent and embolic protection) has been at the discretion of the operator to allow a tailored approach and reflect routine clinical practice. A total of 1,713 patients have been enrolled and recruitment was completed in October 2008. The primary end-point is stroke or death at 30 days and long-term outcome free of fatal or disabling stroke. In addition, a cost analysis will be performed and quality of life will be assessed at five-year follow-up.
Preliminary short-term safety data were presented at the European Stroke Conference in 2009. The reported 30-day stroke or death risk was significantly lower after CEA than after stenting (3.4 versus 7.4%). Likewise, in the MRI substudy, new ischaemia shortly after carotid revascularisation was more common after CAS than after CEA (50 versus 14%). Distal protection was used in approximately 80% of patients. Interestingly, the incidence of silent cerebral ischaemia detected by MRI was higher in the patients that underwent the procedure with the use of distal protection than in patients without. Further details are awaited.
Points of Discussion
The most pertinent problem to be addressed in the future is the almost invariable occurrence of distal embolisation.44–48 This phenomenon occurs with surgical revascularisation and with CAS.44,49,50 It can be readily demonstrated by non-invasive imaging44,45 and the consequences detected by diffusion-weighted MRI,46–48 as well as in some cases clinically, as transient ischaemic attack, strokes or neurocognitive impairment.43 Neurological event rates in most recent randomised trials and registries using distal protection appear to have decreased.17,18,51,52
Events still occur, however. This may in part be related to imperfect device designs. Filter pores are generally in the range of 100μm and there is ample evidence to support that brain injury can be the result of much smaller particle sizes.53–57 The filter apposition is not optimal, particularly in tortuous segments, and balloon occlusion devices may leave a zone that cannot be reached with suction before balloon deflation. Finally, embolisation occurs during catheter manipulation in the aortic arch prior to deployment of neuroprotective devices and has been described well beyond the completion of the procedure.58–60 This emphasises the need for better stent designs (e.g. more flexible and easily deliverable closed cell designs) and the importance of timely dual antiplatelet therapy. Although proximal protection devices with flow reversal have potential advantages, they are very bulky, not always tolerated and not yet rigorously tested in clinical trials.
The second problem, restenosis, which has been encountered relatively frequently with plain old balloon angioplasty,30 has become quite rare with stenting. Most recent trials and registries suggest severe restenosis rates well below 10%.61–65 The implication of this phenomenon will need further study as it is currently not clear if an in-stent stenosis of the same magnitude as a native carotid stenosis is associated with equivalent risks.66
The most controversial recent topic has been the safety of CAS in octogenarians. Rightfully, and not surprisingly, age has been identified as a significant risk factor in CAS.16,67 However, more recent, mainly non- randomised data suggest that in well-selected patients CAS can be performed with similar safety to CEA.32,68–70 It is likely that complex anatomy is more important than age itself. It is clear that CAS should be performed with significant reservation and by the most experienced operators in patients with complex anatomy (severe calcification, tortuosity, hostile arch). This should be the case until more friendly and effective equipment is available and has been tested in clinical trials in this subset of patients.
Finally, the utility of CAS (or CEA) in asymptomatic, particularly high-risk, patients has not been clarified in the presence of more effective medical therapy for stroke prevention. This group of patients should ideally be treated in the setting of a randomised clinical trial (e.g. TACIT) to guide the further management of these patients. Hopefully, future studies can identify the clinical parameters that are associated with a higher risk and take advantage of non-invasive imaging modalities to help risk-stratify asymptomatic patients.
In a perfect world, the event that we are trying to prevent with a treatment should not occur because of the treatment. Although a tremendous amount of progress has been made to reduce the occurrence of strokes as a result of carotid stenting, optimally this risk should be close to zero. In addition to improvements to the technology of the delivery devices (sheaths, catheters, wires, stents), proximal protection with flow reversal or the design of more effective filter systems (perhaps with smaller pore size and antithrombotic coating) should be the most important focus to minimise the stroke risk.
The impact of clinical and plaque characteristics on the stroke incidence of carotid atherosclerosis and stenting should be examined by the increasingly sophisticated imaging modalities (e.g. contrast ultrasound, MRI, CT angiography). This will allow guidance in the determination of which patients warrant treatment and with which method.