Commentary: Carotid Artery Revascularization for Stroke Prevention

Pharmacologic Therapy

Medical therapy reduces stroke risk in asymptomatic CS, but the residual risk remains substantial, particularly in patients with vascular comorbidities. The progress in pharmacologic prevention in cardiovascular medicine over the last 2 decades, including the use (and currently high penetration) of statins, ⁷ antiplatelet agents, ⁸ and angiotensin-converting enzyme inhibitors, ⁹ has led to a reduction in the statistical stroke risk in association with CS.^2,10

Unquestionably, all CS patients should receive optimized medical therapy (OMT). The symptomatic transformation of atherosclerotic CS, however, is far from being eradicated. Consequently, contemporary carotid revascularization studies^11–13 continue to include up to 55% symptomatic subjects in all-comer patient series. ¹¹ A significant proportion of these patients are already on antiplatelet and maximized statin treatment prior to the CS-associated stroke events,^11,14,15 consistent with the fact that medical therapy reduces but does not abolish the CS stroke risk.

CS and Stroke Risk

Recently, a series of influential communications have been addressed to the medical community with the message of a currently “low” (<1.0%^10,16 or ~0.5% ¹⁷ ) yearly incidence of stroke in relation to asymptomatic CS on medical management.^10,16,17 This is in contrast to contemporary data from vascular clinics that show a yearly stroke rate of ~2.0% to 2.5% in real-life cohorts including OMT patients.^15,18 This apparent difference in stroke risk between the general population and vascular clinic referrals is not surprising given that symptomatic disease in one vascular bed is well-known to be associated with future symptoms in another. ¹⁹ The differences in risk burden are likely to reflect fundamental variations in the populations evaluated, as well as the fact that (for a variety of reasons) studies generally include lower-than-average risk subjects.

Fundamental differences between CS patients in the general population and vascular clinic referrals include, for instance, the ~5 to 7-fold higher prevalence of diabetes (~5%–6% vs ~40%) and concomitant coronary heart disease (~10% vs ~60%) and the significantly higher prevalence of other risk factors of symptomatic vascular disease manifestation.^11,15,19,20 Thus, those “asymptomatic” patients who are referred for CS decision making are likely to have a higher stroke risk than the general population with CS for several reasons, including symptomatic vascular comorbidities and higher prevalence of diabetes and other risk factors known to increase the likelihood of symptomatic transformation of the carotid atherosclerotic plaque.^4,14,21 For a significant proportion of these patients, a “not-yet-symptomatic” or “at risk of symptomatic transformation” label would be more accurate than the “asymptomatic” label, but the field suffers from lack of effective risk discrimination tools. ¹⁹

While the need for revascularization in symptomatic patients is not controversial (albeit the treatment arrives too late to prevent brain damage!), it is surprising to the vascular community that, in the absence of any reliable evidence, vocal and powerful calls are being made to “stop revascularization in asymptomatic patients” ²² and “just stick to medical therapy.” ²³ Surprisingly, the notion that “asymptomatic CS is a relatively benign disease”^17,23 has found a particularly strong belief in the neurology community, where it has been largely internalized, resulting in a frequent bias toward routine no-revascularization and the “waiting for symptoms” strategy.^22–25 Large-scale contemporary neurologic patient series show that over 85% of strokes occur without a warning sign. ²⁶ At 5 years after stroke, survival is only ~40%. ²⁷ Of those who survive, about half are disabled and dependent, with 1 in 7 in permanent institutional care. ²⁷ In a substantial proportion of stroke survivors, the quality of life is miserable: 20% of stroke survivors at 5 years report a health-related quality of life score ≤0.1, where −0.04 indicates a state worse than death, 0.0 death-equivalent, and 1.0 full health. ²⁸ For CS patients with increased stroke risk in particular, this questions the validity of prescribing medications and “watchful waiting.” ²⁹

CS Stroke Risk and Evidence-Based Medicine

Misleading data reviews and meta-analyses are published that deliberately ¹⁶ exclude landmark, large-scale, level 1 evidence. The Asymptomatic Carotid Surgery Trial (ACST-1),^30,31 for example, randomized 3120 patients with asymptomatic CS (AHA/ASA definition ³² ) to immediate vs deferred carotid endarterectomy (CEA). The study found a profound 5.9% absolute risk reduction in nonperioperative stroke at 5 years (risk reduction from 10.0% to 4.1%) and 6.1% at 10 years (16.9% to 10.8%) with routine CEA. ³¹ The magnitude of the revascularization effect on stroke risk reduction was maintained in patients on lipid-lowering therapy (5.8% absolute stroke risk reduction at 10 years ³¹ ). It should be noted that such a significant treatment effect is hardly seen in recent clinical trials that usually need to point to a relative rather than absolute risk reduction to demonstrate a treatment effect (Note: in ACST-1, the relative stroke risk reduction with CEA was >50% at 5 years ³¹ ). When the stroke incidence evidence from 1560 ACST-1 patients randomized to deferred CEA (note the 10% 5-year risk of stroke in this large cohort^30,31) is discarded, ¹⁶ pilot data in small populations (in the context of epidemiologic studies ³³ ) become the pillar of “consistent evidence” for the benign nature of CS.^17,25,34

One example of the flagship evidence for the apparently benign nature of CS today is the 3-year follow-up of 101 patients with ≥50% asymptomatic carotid stenosis. ³³ The study registered a 5.9% rate of stroke or transient ischemic attack (TIA), with the majority of TIAs leading (rightly) to carotid revascularization. ³³ This resulted in a calculated annual ipsilateral TIA rate of 1.78% and stroke rate of 0.34%; the study conclusion was that “the risk of stroke on intensive contemporary medical treatment was low.” ³³ It is quite interesting that the second conclusion from this small observational study, ³³ calling for larger studies to determine the generalizability of this apparent improvement in prognosis, passes unnoticed when this work is further referenced.^17,25,34 The extremely high rate of death on medical treatment (29.13% at 3 years) in this work ³³ is not mentioned in the title, abstract, or indeed, when referencing it as apparent evidence that asymptomatic CS is a low-risk pathology.^17,25,34,35

Since the majority of stroke deaths occur out of hospital, ² a proportion of deaths in the “study of the prognosis of ≥50% asymptomatic carotid stenosis” ³³ could well result from stroke, changing the apparent message from that pilot work published in parallel with the (disregarded) large-scale multicenter ACST-1 data. ³¹ It is also quite interesting that the Marquardt study, ³³ with no CEA arm, becomes referenced in a recent “Best evidence for medical therapy for carotid artery stenosis” review ³⁴ as evidence that “if patients were prescribed best medical therapy ³³ then the rates of annual stroke were lower than in patients undergoing CEA ³³ (sic!).”

The rationale that “the worst nightmare for clinicians is that their medically treated asymptomatic CS patient experiences a disabling stroke” is contemporarily flagged as “faulty reasoning.” ²⁵ The nature of being a clinician is to treat real-life patients rather than books or statistical numbers. Very unfortunately, we ³⁶ and other reputable vascular centers^15,18 continue to see disabling strokes in asymptomatic CS patients on OMT who had not been referred for carotid revascularization. The question “why was I/my mother/father not referred to you earlier?” gets wrongly addressed to the operator rather than to those who decided on “restraint” ²⁵ on the basis of biased interpretation of selected data.^16,22,37 There is little doubt that with emerging access to low-risk carotid revascularization in particular ¹¹ those patients would have been better off with primary rather than secondary stroke prevention, particularly because the benefit of revascularization is uncertain in those who have already experienced a severely disabling stroke.

Routine limitation of asymptomatic CS patient management to medications only^16,22,37 is an experimental therapy against level 1 evidence ^30,31 ; it is unknown whether recommendations to manage patients against evidence-based medicine have been subjected to any ethical board reviews. Because there is evidence that faulty data overviews and meta-analyses can be the trigger for a significantly incorrect treatment in clinical practice, 38 the problem of biased data meta-analyses is increasingly appreciated in cardiovascular medicine.

Current Guidelines and CS Patient Management

There are currently at least 28 asymptomatic CS guideline recommendations produced by different international and national organizations. ²⁴ In the era of wide access to one common body of evidence, these guidelines provide conflicting recommendations that swing from CEA “yes” or carotid artery stenting (CAS) “yes” through CEA “maybe” or CAS “maybe” (with differential indication for CAS over CEA or CEA over CAS) to OMT alone. ²⁴ This encourages patient management choices according to the medical specialty of the deciding physician, his or her geographic region, guideline length or layout, or personal prejudice. There is also room for a random guideline choice–patient management choice, though this has nothing to do with conscious therapeutic decision making. The mess in recommendations ²⁴ is a sad reflection of the regrettable state of the field where the value of opinion(s) clouds the value of evidence.

Individual Patient Management Decision-Making Process

The ultimate responsibility for therapeutic decision-making with regard to a particular patient is in the hands of the particular physician (or, better, a multispecialty group of physicians similar to the cardiology-established Heart Teams for routine decision-making) rather than one or another guideline committee. The role of the guidelines is typically to indicate which therapeutic option(s) should be considered in a patient with a particular condition rather than to tell the physician what to do with the particular patient. Physicians need to incorporate in their decision-making process best evidence available in the public domain (and, if possible, evidence generated in patients similar to their patient) taken together with their experience, local expertise, and resources. Indeed, “relevant facts and patient’s best interests need to be distinguished from all else (distractions).” ²⁴

The proportion of asymptomatic CS patients who benefit from revascularization by avoiding ipsilateral stroke or stroke-related death will depend on characteristics of the particular cohort and the length and quality of observation. Unfortunately, there are fundamental problems with individualized risk stratification in asymptomatic CS. ³⁹ In particular, there is absence of prospectively validated risk stratification tools in CS. Nevertheless, several increased stroke risk factors have been identified, including progressive stenosis, irregular/ulcerated or thrombus-containing lesion, contralateral carotid artery occlusion or stroke, ipsilateral clinically silent brain infarct(s) on cerebral imaging, diabetes, and others. ³⁹ Indeed, until (and if ever) recalled on the basis of prospective validation, the above (and other) increased-risk features can be incorporated into a multispecialty neurovascular team (NVT) clinical decision-making algorithm. ¹¹

As with any interventional management, the risk of stroke death or severe disability, with extremely high personal and social costs,^27,28 needs to be weighed against the procedural risk(s) of carotid revascularization. ³ Neurology postulated that revascularization would be appropriate in low-stroke-risk CS patients only if it could be performed with a risk <1%. ⁴⁰ Emerging access to low-risk interventional management, even in symptomatic and increased-risk patients,^11,13 is likely to play an important part in establishing the balance between the individually unpredictable effect of “watchful waiting” on OMT vs OMT plus revascularization.

CAS Paradigm Evolution

The majority of evidence indicates that proximal neuroprotection, including ICA flow reversal, ⁵³ clamping^54,55 (Note: if needed, flow reversal can be routinely achieved with the endovascular clamping device ¹¹ ), or dynamic flow reversal ⁵⁶ can minimize the risk of intraprocedural cerebral embolization.^57–59 According to “tailored CAS” algorithms,^11,53,60 proximal neuroprotection is preferred in endovascular management of symptomatic and high-risk asymptomatic lesions. The use of ultra-closed-cell stent systems (achieved by covering the nitinol frame with a mesh made of different materials), on the other hand, not only further reduces the risk of intraprocedural neurologic complications but also, by preventing plaque protrusion through stent struts, eliminates postprocedural cerebral embolization as manifested on routine diffusion-weighted magnetic resonance imaging (DW-MRI). ⁶¹ This strategy has been termed intra- and postprocedural (sustained) “embolic prevention.” ¹¹ Sustained embolic prevention is thus complementary to the classic intraprocedural “embolic protection” using proximal (flow clamping or reversal) or distal (filter) temporary devices. Recent evidence indicates that incorporation of the sustained embolic prevention technology in otherwise routine CAS may achieve a CEA-like effect with no residual stenosis in all-comer symptomatic and increased-stroke-risk asymptomatic CS patients. ¹¹

Mesh-Covered Stent Family

Apart from the differences in the nitinol frame design [closed cell in RoadSaver/Casper and open cell in CGuard and the Gore hybrid stent (W.L Gore & Associates, Flagstaff, AZ, USA)], the 3 current double-layered carotid stent systems have other important design differences. First, there is the mesh material: braided nitinol in RoadSaver/Casper, polyethylene terephthalate (PET) single-fiber knitted MicroNet in CGuard, and interwoven polytetrafluoroethylene (PTFE) mesh in the Gore stent. Second, the mesh pore size differs: 500-µm-diameter in the Gore stent, 375–500 µm in RoadSaver/Casper, and only 150–180 µm in CGuard EPS. Finally, the position of the mesh in relation to the nitinol frame varies: outside the frame for the CGuard EPS and Gore stent and inside in the case of the RoadSaver/Casper.^61,62,65,67

The RoadSaver/Casper stent delivery system is low profile and very flexible. Similar to the Wallstent, it is re-sheathable up to 50% of the released stent length, which may offer a procedural advantage to some operators. However, RoadSaver/Casper exhibits very significant foreshortening on implantation, reaching 30% of the stent length. ⁶⁵ All 3 stents are commercially available in the monorail (rapid exchange) delivery system. Unsurprisingly, when minimizing the protective mesh cell size, the system delivery profile increases because the more protective microcells per stent frame area unit require more mesh “fibers.” So, as a general rule, the smaller the mesh cell size, the larger the delivery profile. The RoadSaver/Casper has the smallest delivery profile (5-F), with the Gore stent in the middle (5-F or 6-F depending on diameter). The CGuard EPS offers the largest open-cell nitinol frame size among the current carotid stent systems (21.7 mm² for the 8.0×40-mm stent size ¹¹ ) combined the smallest mesh microcell size (0.023–0.032 µm²), which corresponds to the pore size in distal protection filters. This makes the CGuard EPS (with 6-F outer diameter of the packed delivery system) the most “open” and, at the same time, the most “closed”-cell carotid stent among the current designs. ¹¹

Unsurprisingly, the tightly packed ultra-closed-cell mesh CGuard has a relatively high bending stiffness prior to delivery, ⁶² but after implantation it shows an extremely high flexibility and wall adaptation ⁶² exceeding that of RoadSaver/Casper. ⁶⁵ CGuard EPS has a high radial force that is similar to the Precise stent and is higher than that of RoadSaver/Casper.^62,65 In bench tests by Wissgott and colleagues, ⁶² CGuard showed both in a step model and in a curved model adaptation to vascular anatomies that was superior to RoadSaver/Casper.

When positioning the stent at the carotid bifurcation, some operators (in the absence of any systematic evidence for a clinical benefit) prefer a tapered design to avoid, theoretically at least, stent oversizing in the ICA and better fit to the particular carotid bifurcation anatomy. A tapered Gore stent is available in 3 diameter ranges (6–8, 7–9, and 8–10 mm). Because of its construction, the RoadSaver/Casper does not have a tapered version. On the other hand, the CGuard EPS, applying the SmartFit technology, has a property of self-tapering (“self-adaptation”) as confirmed in bench tests ⁶² and recently systematically demonstrated on independent angiographic core laboratory evaluation. ¹¹

Another CGuard EPS characteristic that may be important from the procedural standpoint is the lack of foreshortening on bench tests. ⁶² This CGuard EPS feature, which may offer an advantage particularly in anatomies that require precise stent placement, ¹¹ has been confirmed on independent systematic core laboratory analysis of over 100 CGuard EPS implants in consecutive CAS cases. ¹¹

MicroNet CGuard EPS Properties

Manufacturer-independent data on the CGuard EPS mechanical properties are summarized by “high radial force and strong support to expanded stenotic vessel sections, ease of implantation in absence of foreshortening, stent structure adaptation to changes in diameter and direction of tortuous vascular anatomies, and no measurable effect of the mounted MicroNet mesh on specific mechanical properties [of the device].” ⁶² These properties are critically important in understanding why the use of this system in CAS of all-comer lesions (including those that heretofore would have been managed with CEA rather than prior-generation carotid stents ¹¹ ) routinely confers a CEA-like effect on CAS, with no residual stenosis and reconstruction of the diseased carotid bifurcation fully respecting the native anatomy (no vessel straightening, kinking, etc ¹¹ ).

Another important contribution from Wissgott and colleagues ⁶² is their illustration of the feasibility and safety of exclusively primary (“direct”) CGuard EPS stenting, while in the 2 prior studies (CARENET ⁶¹ and PARADIGM ¹¹ ), the majority of carotid lesions were predilated. Taken together with the recent demonstration of CGuard’s compatibility with all types of temporary neuroprotection systems, ¹¹ the current data ⁶² inform the community on the feasibility and safety of different procedural strategies using CGuard EPS, allowing strategic tailoring to patient/lesion anatomy and symptomatology in a context that includes different temporary neuroprotection device use on the one hand and operator preference/experience on the other.

Routine Brain Imaging Evidence

With non-mesh-covered carotid stents, cerebral DW-MRI imaging clearly demonstrated ongoing postprocedural embolization. ⁴⁸ Today, DW-MRI is an important tool in evaluating minimization of intraprocedural embolization with improved temporary neuroprotection systems.^56.57,59 DW-MRI is similarly fundamental in evaluating the efficacy of different stent designs in reducing periprocedural and delayed cerebral embolism. ⁶⁸ With routine DW-MRI use, the CGuard EPS system was associated with >50% reduction in the incidence of periprocedural DW-MRI lesions and a >10-fold reduction in average lesion volume.^61,69 Furthermore, per-protocol DW-MRI imaging repeated at 30 days after the CGuard EPS implantation in the CARENET study ⁶¹ demonstrated evidence for near-elimination of new lesion occurrence during the stent healing period. Thus, systematic periprocedural and 30-day DW-MRI data on the 2 other mesh-covered stent systems (RoadSaver/Casper and the Gore hybrid stent) in cohorts of unselected patients are awaited to confirm the anti-embolic efficacy of these stents and to see whether there is any sign of a potential effect of design differences on embolic prevention efficacy. With the DW-MRI randomized data desired (but not unlikely to be performed on any large scale), any future comparison should correct for the proximal vs distal temporary protection type (which has a role in minimizing intraprocedural but not postprocedural embolization ⁵⁹ ) and the type of lesion(s)/patent symptomatic status.

Double-Layer Stents and Cerebral Protection

Because the mesh of the dual-layer stent design offers protection against plaque embolization starting only with the postdilation phase, it is certainly premature to call off the temporary brain protection devices in CAS using dual-layer stent systems. ¹³ This is particularly relevant in the absence of any criteria for the use of temporary neuroprotection devices (vs non-use) ¹³ and because the embolic material continues to be present, albeit at an apparently lower rate^11,60 in the temporary protection systems employed exclusively with the novel stent generation. ¹¹ Therefore, the concept on “no need for a neuroprotection device” with dual-layered stent use, ¹³ although definitely attractive to some operators, requires large scale, properly designed investigation with routine DW-MRI imaging. These studies potentially should be performed in certain patient/lesion subsets, such as treatment of carotid in-stent restenosis (note the need to differentiate between true restenosis and neo-atherosclerosis) or a fibrotic non-critical lesion (in clinically indicated revascularization).

Another important issue in the emerging field is to establish the role of transcarotid revascularization using surgical access and endovascular dynamic flow reversal, which is highly effective in minimizing intraprocedural cerebral embolization,^{12,56,57,58,66} in combination with the novel dual-layer carotid stent technology for sustained embolic prevention.

With a focus on ensuring CAS safety and durability, the vascular interventional community should not ignore a major issue that is more fundamental than the choice of revascularization mode and the importance of outcome monitoring as related to novel technologies. The technological progress that enables CAS safety contrasts with the lack of reproducible and clinically applicable stroke risk evaluation tools in CS compared with other fields in cardiovascular medicine. This is an important “first step” before the decision on the treatment mode (OMT + intervention vs OMT only) for CS patients, including indications for (any) intervention. For instance, reproducible risk stratification models for nonvalvular AF have not only been developed and validated but also effectively disseminated as a useful everyday clinical practice tool (eg, the CHA₂DS₂-VASc and HAS-BLED score decision-making model ⁷⁶ ), and the role of pharmacologic vs interventional stroke risk management in AF has been elucidated.^77,78 Urgent, multispecialty and large-scale endeavors, involving neurologists, vascular surgeons, radiologists, cardiologists, and angiologists are therefore greatly needed to close the embarrassing gap in prospectively validated evidence to determine the optimal management paths suitable for particular subsets of CS patients, particularly asymptomatic. ²⁹ Nevertheless, until (and if) prospectively validated algorithms similar to the CHA₂DS₂-VASc score become available for asymptomatic CS, neurologists and vascular specialists should endeavor to incorporate ³ in their decision making the already identified increased-risk criteria ³⁹ as per local feasibility. ¹¹

Sustained Embolic Prevention Evidence for CAS

Long-term data will be provided from the medium-sized studies that have already published their 30-day outcomes^11,13,72 and others that continue to recruit patients. ⁷⁹ Pilot evaluation of the anti-embolic efficacy of dynamic flow reversal using CGuard EPS in a series of consecutive symptomatic and increased-risk asymptomatic patients treated via the transcarotid route will be communicated at VEITH 2016, including routine DW-MRI cerebral imaging prior to CAS, at 24 to 48 hours, and 3 months. The Gore SCAFFOLD study of their double-layered stent in 312 patients in the US has recently crossed 50% recruitment and is expected to report its findings in 2017.

Recent communication of the first 12-month clinical and duplex ultrasound data ³⁶ for the CGuard EPS (CARENET ⁶¹ ) was reassuring in terms of the lack of any sign of an increased risk of in-stent restenosis or adverse neurologic events in midterm observation. However, larger-scale and longer-term clinical evidence is awaited by the endovascular community to firmly establish the role of dual layer anti-embolic stent technology. The 101 consecutive patient CGuard EPS PARADIGM study ¹¹ will report its 12-month clinical and duplex ultrasound data in fall 2016.

To answer the need for still-larger-scale data, a 2500-patient, investigator–initiated CGuard Euro-PARADIGM study protocol was finalized in 2015. Euro-PARADIGM follows the PARADIGM study concept ¹¹ in a multicenter multispecialty setting. Early in 2016, Terumo announced a 5000-patient European registry of RoadSaver/Casper. Data from these studies will fully determine the place of novel double-layered carotid stent technologies in large-scale clinical practice.

Large level 1 evidence for the efficacy of dual-layer carotid stent systems including clinical endpoints would be theoretically desired. Nevertheless, evidence from prior large-scale work shows that such a study would certainly suffer from the problems of effective recruitment volume, patient crossover, and funding.^43,80 Selection bias–free randomization would be either unethical (in the case of high-risk lesion randomization against prior-generation single-layer stent(s) with documented cerebral embolization significantly exceeding that of a mesh-covered stent system ^61,69 ) or the study would a priori likely be inconclusive (in case of low-risk lesion randomization). Therefore, cerebral DW-MRI imaging is going to continue to provide a sensitive surrogate in the evaluation of novel carotid revascularization technologies and treatment strategies.48,56,57,59,61,68,69 One obvious question would be whether different sizes of mesh pores are associated with any differences in embolic prevention efficacy as per the number and size of DW-MRI cerebral lesions. ⁶⁷ However, a number of other important factors of DW-MRI cerebral lesions, such as the plaque type and procedural strategy (eg, proximal vs distal neuroprotection) need to be taken into consideration in any direct or indirect comparative analyses of dual-layered stents.

Recently, OCT demonstrated that the RoadSaver/Casper stent was not completely free from plaque protrusion through the inner mesh layer into the lumen of the stented artery. ⁷² True prevalence of this phenomenon or its potential relation to cerebral embolization could not be determined due to non-consecutive patient enrollment and non-consecutive OCT evaluation, and there was no systematic DW-MRI imaging. However, no clinical manifestations of cerebral embolization were reported in 30-day observation.

The large clinical trials of carotid revascularization that are already ongoing will have to decide on whether (and how) to incorporate the novel stent technology so that the study results are not obsolete at the point of subject recruitment. Incorporation of novel carotid stent designs is feasible in the case of the Europe-based ACST-2 trial because 2 of the 3 dual-layer stent designs (CGuard EPS and RoadSaver/Casper) are registered for unrestricted use in Europe. In the US-based Carotid Revascularization Endarterectomy vs Stenting Trial 2 (CREST-2), the situation is more difficult because US-generated clinical evidence with a dual-layer carotid stent is likely to be required by the Food and Drug Administration. In the CREST-2 arm comparing contemporary OMT with CAS plus OMT, the ability to identify any relative outcome difference between the 2 treatment arms will depend on effective enrollment and randomization of a non-low-risk asymptomatic CS population, minimal patient crossover, and the ability to offer safe CAS with minimized post-CAS neurologic events.