Insights from Carotid Artery Stenting Registries
Registries tend to have less stringent inclusion and exclusion criteria than randomised trials and provide real-world data on a broad variety of patients treated by a range of interventionists of varying experience. Registries serve to provide first-time evaluation of new devices, help set up learning curve pearls, can be hypothesis-generating and can also help to validate results from randomised controlled trials in a broader spectrum of patients.
The 30-day major adverse cardiovascular event (MACE) rate of several US carotid artery stenting (CAS) registries (see Figure 1) shows a progressive improvement, which is largely attributable to an improved learning curve. In addition, the improved results are secondary to refined case selection, avoidance of anatomically adverse carotid patients and improvements in device profile and performance. Most of the device innovation has focused on refinements and redesigns to distal protection devices. Most manufacturers have improved their carotid filters. From Accunet II, Angioguard Xp and Filterwire EZ Spider to Emboshield NAV 6, the device profile has improved and transition zones have been smoothed, allowing easy trackability of devices and shorter device landing zone requirements. More recently, the FiberNet 3D filter development has used 3D fibres in a short landing zone, providing another option. All of these iterations have increased ease of use, safety and predictability of performance. Similar enhancements to the carotid stent designs have led to a reduction in profile from 9 to 5Fr.
Two large-scale post-market studies, Emboshield® and Xact® Post approval carotid stent trial (EXACT) (n=2,145) and Carotid ACCULINK/ACCUNET post approval trial to uncover rare events 2 (CAPTURE 2) (n=4,175), by 672 operators at 280 sites in the US show remarkably low complication rates in non-octogenarians meeting the American Heart Association (AHA)/American College of Cardiology (ACC) outcomes thresholds for carotid endarterectomy (CEA: symptomatic 5.3% and asymptomatic 2.9%).1 Both studies had independent neurological evaluation and audit. However, octogenarians had unacceptable outcomes in both the symptomatic (MACE rate 10.5%) and asymptomatic groups (MACE rate 4.4%). Schreiber reported similar data in the Carotid Artery Stenting with Emboli Protection Surveillance-Post-Marketing Study (CASES PMS)2 (n=1,492), with a MACE rate of 5% at 30 days.
These results are comparable to those of the Stenting and angioplasty with protection in patients at high risk for endarterectomy (SAPPHIRE) trial, suggesting that clinical trial results can be reproduced in the real world by physicians with varying levels of training. The three-year Carotid artery revascularization using the Boston Scientific FilterWire EX/EZ and the EndoTex NexStent (CABERNET) study3 showed a remarkably low stroke rate in stented patients at follow-up (7.2%), with very low restenosis rates. Similarly, the Carotid revascularization using endarterectomy or stenting systems (CARESS) real-world parallel registry study4 showed a low 30-day MACE rate (under 2.5% for both treatments (CEA and CAS) and at four-year follow-up reaffirmed low stroke rates (9.6% for CEA and 8.6% for CAS).
Using propensity-matched analysis, Bangalore et al.5 reported no difference between CAS (n=836) and CEA (n=836) patients in the Reduction of atherthrombosis for continued health (REACH) registry in terms of death, myocardial infarction (MI) and stroke at two years.
Several European investigators have shown consistently better results with proximal protection compared with distal filters, having had access to such devices earlier than their US counterparts. For example, Stabile et al.6 reported an impressive 30-day MACE rate of 3.04% in symptomatic and 0.82% in asymptomatic patients. These results were duplicated in the US in the Proximal protection with the MO.MA device during carotid stenting (ARMOUR) study7 (MO.MA device) and the Embolic protection with reverse flow (EMPIRE) study (Gore Neuro protection),8 using proximal protection from two different devices in high-risk patients. The stroke rates in symptomatic patients are the lowest ever achieved,8 perhaps raising an argument in favour of proximal over distal protection in symptomatic groups. A new 3D filter (FiberNet) was used by our group in the Emboli protection device in carotid artery stenting (EPIC) trial9 in 237 patients at 26 centres, with a 30-day MACE rate of 3% in an all-comer high-risk group using any commercially available stent. Similar impressive data were also provided in the Protected carotid artery stenting in subjects at high risk for carotid endarterectomy (PROTECT) trial by Gray, using the low-profile, new-generation Emboshield Pro filter.10
Negative Signal in Octogenarians
Data from several registries and the Carotid revascularization endarterectomy versus stent trial (CREST) lead-in study (see Figure 2) showed that high-risk octogenarian patients treated with distal protection devices have unacceptably high complication rates. While the reasons for this are multifactorial, including elevated risk due to age, cardiac comorbidity and arch atheroma, the presence of carotid anatomical adversity11 is probably the biggest risk factor. Much of this is due to significant carotid tortuosity, the presence of type III arch and sharp entry and exit angles at the lesion. It is also possible that octogenarians have deranged cerebral reserve and autoregulatory mechanisms predisposing them to hyperperfusion syndrome, cerebral oedema and intracranial haemorrhage. Some of this elevated risk in early studies represents the learning curve (see Figure 3), a lack of physician awareness of anatomical adversity and lack of knowledge of bailout techniques. More recent data suggest that in experienced hands, octogenarians do not necessarily carry a prohibitive risk. Chiam et al.12 reported a major stroke or death rate of 2% in 153 patients, with 0% in symptomatic patients (n=39) and 2.6% in asymptomatic patients (n=114). In another small series reported by Bacharach et al.,13 78 patients (3.8%) compared favourably with 157 non-octogenarians (2.5%; p<0.426).
Insights from Personal Case Series
Yadav et al.14 have provided valuable insights from early case series. They refined the carotid stenting technique, bringing in coronary technology, using stratified risk categories for access based on external carotid artery (ECA) disease and popularising the no-touch access technique and periprocedural management including dual antiplatelet therapy. The current authors also learned a great deal from Mathias, Theron and Kachel on carotid anatomy and haemodynamic monitoring.
Since 1996, the current authors have been collecting data on aortic arch patterns (see Figure 4)15–18 and ease of access from the femoral approach. We classified aortic arch types into types I to III based on the elongation of the aortic arch and its relation to the junction of the descending aorta and the distal end of the arch. As the arch sinks deeper, it takes with it the origins of the great vessels, rendering the femoral approach increasingly complex. We have devised techniques for access and equipment choice based on aortic arch type that are now used extensively.
More recently, we began noticing carotid lesion adversity as the most important critical factor affecting ease of access to the distal internal carotid artery (ICA) for distal protection devices. In addition to subtotal occlusion, the presence or absence of sharp lesion angulations at entry and exit (see Figures 5 and 6) is the single major risk factor for complications during use of carotid filters. Sharp entry and exit angles increase friction when the filter is being advanced and adversely create an opportunity for plaque prolapse into the stent. Sharp angles prevent easy advancement of filter-retrieval catheters through the stent in the bends and can in fact promote filter migration into a more proximal position and risk filter stent entanglement (see Figure 7). Strategies to deal with sharp angles include adjusting guide catheters to a more proximal position to relax the central carotid artery (CCA) and using buddy wires, buddy catheters, Venture catheters, pre-filter balloon dilatation catheters, Spider and Emboshield pro NAV 6 filters and proximal protection.
Interestingly, most of the valuable information on CAS became available in personal case series, with registries and randomised trials providing validation and refinement of such techniques to a larger audience.
Insights from Randomised Carotid Artery Stenting Trials
The early randomised Carotid and vertebral artery transluminal angioplasty study (CAVATAS) trial compared carotid angioplasty alone versus CEA. The acute and long-term data suggest comparable results (10.6% stroke or death rate in both groups of symptomatic patients), with a higher restenosis rate in the angioplasty alone group than in the stent group. The first randomised trial in high-risk groups with distal protection and independent neurological audit was the SAPPHIRE trial, whose unique design with parallel registries has been described in the literature, as have the acute 30-day and three-year follow-ups.14 In the group of eligible patients randomised to either CAS or CEA, no between-group differences in the composite end-points (stroke, MI and death) were observed. There was no difference in the major stroke or death rate, although there were numerically more minor strokes in the stent group and more MIs in the CEA group. The same theme can be observed in the much larger CREST trial, which is discussed below. There has been repeated criticism by some19 about the inclusion of MI in the composite end-point; therefore, it is important to remember the adverse consequence of perioperative MI after vascular surgery.20,21 The Mayo Carotid surgery trial was aborted early in the US owing to unacceptably high rates of perioperative MI and death.22
European Randomised Carotid Artery Stenting Trials and Their Impact
Endarterectomy versus angioplasty in patients with severe symptomatic carotid stenosis (EVA 3S) and Stent-supported percutaneous angioplasty of the carotid artery versus endarterectomy (SPACE)23,24 are two large-scale randomised trials from France and the EU, respectively. Both trials were conducted on symptomatic patients by clinicians with excellent surgical expertise. The relative lack of expertise by the CAS operators compared to the surgeons has been widely criticised, particularly in the EVA 3S study, where mentored proctoring and participation was allowed in at least 50% of the procedures. Both trials by design allowed optional use of distal protection; this was later changed in EVA 3S to be mandatory, after unacceptably high rates of stroke were noted in the unprotected group. EVA 3S produced some of the best surgical results published so far. EVA 3S had inferior results for CAS with or without protection compared with CEA. The SPACE study used a distal protection device in only 27% of the patients; this study was terminated for financial reasons but not because of any safety signal. The published results show no difference in outcomes between the stent and surgery groups. The more recent International Carotid stenting study (ICSS)25 showed inferior results for symptomatic patients treated with CAS and distal protection compared with those treated with CEA. All of these studies together had a negative impact (see Table 1) on the widespread use of CAS and enrolment in ongoing clinical trials.
It is important to remember that no single trial provides all the answers. Some are designed to provide answers based on theories and assumptions that were prevalent at the time of the study design. Many of these trials had no mandated CAS device training or an adequate roll-in period to make sure the learning curve had peaked by the start of the study and distal protection was not mandated with a uniform device system. Sometimes, the medical therapy was not contemporary; for example, dual antiplatelet agents were not uniformly used.
It is axiomatic that the more procedures we perform, the better we perform them; we do not need randomised studies to prove this point. It is essential that the operator is trained and his or her results vetted by an independent interventional committee before the operator is allowed to enter the roll-in phase prior to randomisation.
The operator must have adequate expertise in the device being tested. With that in mind, the requirements for endovascular experience in the three European trials are as follows:
- SPACE: Minimum of 25 cases, but proctored training allowed for those who had performed only 10 lifetime cases. This did not have to be recent experience or concurrent.
- EVA 3S: At least 12 procedures or at least 30 supra-aortic (sub-clavian) procedures. Tutor-assisted procedures allowed at certain centres not fulfilling the minimum requirements. A significant number of procedures in the trial were the first ever procedures performed by trainees.
- ICSS: At least 50 stenting procedures, 10 of which had to be carotid stenting. Tutor-assisted procedures allowed for sites with limited experience.
Contrast this with the CREST trial enrolment requirements. An independent interventional committee evaluated each operator’s experience. The results were reviewed and a minimum of 20 carotid stent procedures were required with acceptable stroke and death rates meeting AHA criteria. CREST had an elaborate roll-in programme, with almost 1,400 patients enrolled in the roll-in phase – more than the number of stent patients enrolled in any of the European trials. The roll-in phase shows clearly the role of training and outcomes when individual outcomes are compared within the specialties, suggesting that cardiologists with early adoption of embolic protection have learned how to troubleshoot far more quickly than non-cardiologists in asymptomatic patients stratified by age. Roffi et al. elegantly highlight the conundrum by stating that tutoring and randomised trials should be considered mutually exclusive.26 No surgeon would be allowed to enrol in a randomised trial if he or she had performed only 10 endarterectomy procedures.
The Carotid Revascularization Endarterectomy versus Stent Trial Trial and its Interpretation
CREST is the largest randomised trial27 comparing CAS with routine embolic protection versus CEA. The CREST trial included multiple disciplines and had surgical and interventional management committees that used stringent criteria for investigator selection. The trial required a lead-in phase prior to allowing investigators to perform carotid stenting in the randomised phase.
Routine embolic protection was required in the study, with a single device system used. CREST allowed inclusion of both symptomatic and asymptomatic patients. CREST showed no difference in MACE outcomes with a composite end-point of all-cause stroke, MI or death between CAS and CEA (CAS versus CEA 5.2 versus 4.5%, hazard ratio [HR] 1.18, 95% confidence interval [CI] 0.82–1.68; p=0.38). There was no difference in the incidence of major stroke (CAS versus CEA: all stroke 4.1 versus 2.3%; p=0.01; major stroke 0.9 versus 0.7%; p<0.52). When separating the end-point for the type of stroke, there was an absolute increase in minor strokes with stenting, which was balanced or cancelled out in the composite end-point by an absolute increase in MIs in the CEA group (CAS versus CEA: stroke 4.1 versus 2.3%; p<0.01; MI 1.1 versus 2.3%; p=0.03). This has created the controversy of whether it is unreasonable to include MI in the composite end-point of a study procedure designed to prevent strokes. However, if CEA prevents strokes at the cost of additional heart attacks, which can lead to adverse consequences, it is reasonable to consider this as a cost of protection. Rates of ipsilateral stroke during a mean follow-up of 2.5 years were equal between groups, at 2% for stenting and 2.4% for surgery.
The CREST Quality short form (SF) sub-study (see Figures 8 and 9) attempted to study the patient perception of the CREST adverse event outcomes on their lives. It claims to show that patients perceived minor strokes more negatively than MI while admitting that those who had cranial nerve palsies felt better than those who did not after CEA. This raises doubts about the accuracy and predictive value of SF study. It is hard for patients to perceive an MI that is silent. It is also possible that those with cranial nerve palsies had more focussed follow-ups and felt better cared for. A growing concern among interventionists is the habit of singling out cranial nerve palsies separately from CEA and not including them in the neurological events as a better catch-all than stroke alone. After all, if a patient had hypoglossal palsy from cranial nerve injury or minor stroke, the impact on the patient would not be that different since both are eventually resolved within several months. The incidence of cranial nerve palsies at 5–7% has remained steady through decades of mature carotid surgery from the North American symptomatic carotid endarterectomy trial (NASCET), the European carotid surgery trial (ECST), the SAPPHIRE and CREST trials. This is an important comorbidity that should be considered in the global cost of revascularisation as composite neurological complications. This brings an important doctrine; to reach clinical equipoise between treatment arms one needs to appreciate complications. It would seem hypocritical to portray perioperative MI and cranial nerve palsies as inconsequential and maintain that CEA is superior to stenting by minor stroke margin alone.
The CREST trial has contributed significantly to our understanding of the relative roles of CEA and CAS, which do not necessarily compete and may in fact be complementary. For the older person, CEA appears to have an advantage in terms of stroke protection, but at the cost of an increased risk of MI. In the younger patient, CAS appears to have an advantage over surgery (see Figure 10). Consequently, patient selection is important in choosing the right modality.
The Asymptomatic carotid trial (ACT) has neared enrolment of asymptomatic patients in an asymmetrical design between CAS and stenting using the XACT stent and Emboshield Filter, sponsored by Abott vascular. This trial will provide important answers in this group of patients and may find that both revascularisation methods have their place in appropriately selected patients and in experienced hands.
To quote Vladimir Hutchinsky: “For medical knowledge to progress we need both the optimist and the pessimist. For without the optimist we have no progress, while without the pessimist we have no proof.” CAS has come a long way and the state-of-the-art has been redefined with the recent availability of new-generation distal protection devices and proximal protection systems. Proximal protection could in fact be the game changer for both symptomatic and older patients compared with distal protection alone. Both the highly anticipated ACT I trial and proximal protection might ultimately prove the pessimists wrong.