Since the publication of the Endarterectomy versus Angioplasty in Patients with Symptomatic Severe Carotid Stenosis (EVA-3S) and Stent-supported Percutaneous Angioplasty of the Carotid Artery versus Endarterectomy (SPACE) studies, doubts have been raised regarding the safety of carotid artery stenting (CAS) as an alternative carotid intervention to carotid endarterectomy (CEA). A recent metaanalysis¹ recommended the cautious use of CAS, and concluded that CAS for low-operative-risk patients should be subjected to randomised controlled trials. It might be said that the once optimistic view of CAS has softened, and has been replaced by a healthy scepticism. This accepted, it is clear that CAS is preferable when the patient is deemed to be at high surgical risk and, as meaningful subset analyses emerge from the ongoing trials, it is likely that some patients will respond well to CEA and some with CAS, and the remainder should be equally well treated by either.

The purpose of this article is to try to put forward the author’s view on cerebral protection during CAS, and to highlight Södersjukhuset’s experience with the flow reversal system, Gore Neuro Protection System (NPS) (WL Gore, Flagstaff, Arizona).

Before embolic risk and the reduction of this risk by the deployment of cerebral protection systems are discussed, we must remember that embolic risk is not the only source of procedural neurological hazard. There are, of course, numerous other causes of procedural stroke, including haemodynamic injury, acute carotid occlusions and the uncommon contrast encephalopathy.² Discussion of these is beyond the scope of this article.

Can Vulnerable Plaque Be Identified Prior to Intervention?
Biasi et al.³ described the importance of plaque characteristics in determining outcome after CAS. The term greyscale median (GSM) describes an objective parameter of plaque echolucency, and a GSM score <25 predicts a worse outcome than for the group of patients with a GSM score >25. The only other predictor of a worse outcome is the degree of stenosis: the tighter the stenosis, the higher the risk. Of course, the stenosis measurement is simply a surrogate marker of emboligenic risk. Other authors have not reproduced these findings,⁴ and there is disagreement about the importance of plaque echolucency in identifying patients at risk of peri-procedural neurological events.

The Influence of Patient Anatomy on Outcome
Tortuous or variant anatomy can increase both the technical complexity of CAS (leading to increased procedural stroke) and the technical failure rate.⁵ It is crucial to know the anatomical challenges of a case for planning purposes, and this usually mandates ‘overview’ imaging such as magnetic resonance angiography (MRA), computed tomography angiography (CTA) or catheter arch aortography. Patients should be graded according to their anatomical complexity and chosen in a way that suits the experience of the physician performing the CAS procedure.

Neuroprotection During Carotid Artery Stenting
Do we need to protect the brain during CAS? We do not know for certain. There is no first-class evidence that supports the theory that patients undergoing protected CAS do better than those undergoing unprotected CAS. What we know is that debris is released from the atherosclerotic plaque during manipulation with guidewires, filters, stent deployment and even lesion crossing, pre- and post-dilatation, with a filter and filter deployment and retrieval stages. Moreover, catheter manipulation in the aortic arch and selective catheterisation of the supra-aortic vessel, a prerequisite for CAS from a femoral or brachial approach, is emboligenic. We also know that up to eight times more microembolic signals are detected during unprotected carotid angioplasty than during CEA on procedural transcranial Doppler (TCD).⁶ More recent work on new white lesions detected with diffusionweighted magnetic resonance imaging (MRI) confirms the higher microembolic load in filter-protected CAS compared with CEA.⁷ Kastrup et al.⁸ performed a systematic review of 839 patients treated with protected CAS (largely of the filter type) and 2,357 without protection. This indicated that cerebral protection significantly reduced embolic complications (1.8% in the protected group versus 5.5% in the unprotected population), although many of the studies included employed historical controls and were self-audited and of small sample size. There are no randomised trials to guide us. However, it is clear that all available protection systems can and do trap macroemboli that would otherwise pose a significant threat of major stroke. A recent study highlights that in 279 patients treated with filter-protected CAS, visible debris was found in 169 filters (60%).⁹ Therefore, it would seem wise to use some kind of cerebral protection device (CPD), analogous to the use of a parachute when skydiving. A randomised trial is not necessary to prove a parachute’s value under these circumstances.

Since 1990, when Théron et al.¹⁰ first described cerebral protection during CAS using an occlusion balloon to avoid distal embolisation, a number of dedicated devices have been developed. Thanks to new materials and technical improvements, we now have a range of products that can be used in order to protect the brain from macroand microembolism during CAS.

Products available operate in different ways, and can be grouped in order of how they work, i.e. distal filters, distal balloon occlusion, proximal balloon occlusion and flow reversal. The remainder of this article will focus on filters, balloon occlusion systems (proximal and distal) and flow reversal.

Filters
Filters are the most frequently used type of CPD. These are easy to use and the pore size varies between 60 and 140 microns. There are two main designs: a basket of Nitinol mesh (e.g. SpideRx™ filter, ev3, Inc., Plymouth, Minnesota), or a perforated polyurethane membrane (e.g. FilterWire EZ, Boston Scientific, Natick, Massachusetts). Filters allow constant perfusion of the brain, which could be important in patients with hypoperfusion syndrome or in patients with an incomplete circle of Willis. Particles smaller than the pore size probably pass unhindered to the brain, and some may pass beside the filter in case of poor wall apposition. In order to establish protection using a filter, the undeployed filter has to cross the lesion while flow in the internal carotid artery is antegrade. This is clearly an unprotected and potentially emboligenic stage of the procedure. However, these devices, like every other protection system, are capable of trapping macroemboli.

There are trade-offs. There is potential for intimal damage in the distal internal carotid artery (ICA) caused by the filter.¹¹ If the ICA above the lesion is tortuous, there might not be enough ‘landing zone’ for safe placement of the filter.

Distal and Proximal Ballon Occlusion These devices create a no-flow situation in the ICA beyond the treated segment, and it must be remembered to aspirate debris after completion of procedural steps. One example of distal balloon occlusion is the PercuSurge GuardWire (Medtronic, Santa Rosa, California). The MO.MA device, (Invatec, Roncadelle, Italy) effects proximal balloon occlusion of the common and external carotid arteries (CCA and ECA). The ECA balloon in this system is fixed to the sheath at a pre-determined length, and delivery of guidewires and stents is performed through a sidehole of the sheath. Schmidt et al.¹² compared the MO.MA device with the EPI Filterwire (Boston Scientific Corp., Santa Clara, California). Microembolic signals (MES) were monitored during five stages of the CAS procedure (see Figure 1):

• placement of the protection device;
• passage of the stenosis;
• stent deployment;
• balloon dilatation; and
• retrieval of the protection device.

The authors reported a significant reduction (but not an abolition) of MES counts during phases II–IV, and in total when using the MO.MA device.