Introduction

Flow diversion represents a major milestone in the treatment of cerebrovascular aneurysms. Flow diverters (FDs) were initially indicated only for specific challenging aneurysm morphologies previously difficult to treat.1 However, clinical studies have expanded indications and ‘off-label’ use has dramatically expanded at many centers.2–4 Flow diversion requires a large metal surface area to be effective. Consequently, FDs are thrombogenic, leading to ischemic complications and necessitating dual antiplatelet therapy (DAPT). Even with effective DAPT, ischemic events represent the majority of complications following FD treatment.5 6 Suppression of platelet function increases the risk of serious bleeding,7 and yet it does not always prevent thromboembolic complications.8 9 An FD that was safe and effective with single antiplatelet therapy (SAPT), preferably aspirin alone, would be of significant clinical value. Although a variety of solutions have been considered including bioabsorbable materials for stent struts,10 FD surface modification represents one of the attractive strategies for reducing thromboembolic ischemic complications and avoiding DAPT after implantation.

Several potential modifications have previously shown good safety profiles but variable efficacy when studied in coronary stents. Over time, interest in surface modification faded as drug-eluting stents, particularly sirolimus, demonstrated high efficacy in reducing restenosis.11 Even with a robust body of literature surrounding many of these surface modifications for coronary applications, further work is needed to apply these concepts to FDs. Treating coronary atherosclerotic disease, which with exposed plaque components is thrombogenic by itself, is different from treating aneurysmal disease of the neurovasculature. Furthermore, the coronary circulation is driven by diastolic flow through larger vessels creating relatively low shear stress compared with the systolic flow through the neurovasculature.12–14 Although stent thrombosis has devastating consequences in both locations, there are unique considerations. Very small thrombi forming at jailed side branches and perforators are a specific concern after flow diversion, while stent restenosis is more problematic following coronary intervention.5 11 15 Finally, the eloquence of the two organs involved means that the neurovascular circulation is far less forgiving. These distinctions and the design differences between coronary stents and FDs demonstrate the need for further investigation of surface modifications for FDs.

While there are conceptually different approaches to reducing DAPT-associated complications (figure 1, table 1), both strategies will also reduce thromboembolic complications. The first and most obvious is to directly decrease device thrombogenicity using mechanisms that reduce or eliminate the need for systemic therapies. An alternative strategy is to promote rapid device surface endothelialization, as normal endothelium is the best antithrombotic coating. While this may not eliminate DAPT, it may significantly decrease the duration of therapy. These two approaches conflict somewhat. Coatings that prevent the attachment of platelets may also delay endothelialization, while coatings designed to improve endothelial growth and adhesion may be a similarly attractive surface for platelet adhesion and activation, thereby potentially increasing the risk of thrombosis.

Figure 1

Chart showing the approaches taken by various surface modifications to reduce or eliminate the need for dual antiplatelet therapy (DAPT). NiTi, nitinol; PMEA, poly(2-methoxyethyl acrylate).

Table 1

Summary of different surface modifications/coatings for flow diversion and intracranial stents

Currently, only the p48 and p64 devices with a hydrophilic polymer coating (HPC; Phenox, Bochum, Germany) come with instructions for SAPT in some regions when DAPT is contraindicated; however, there are several different surface modifications designed to achieve a similar goal. This review will focus on currently available surface-modified FDs, concepts in translational stages showing great promise, and some research supporting their design and use.

Surface modificationsBiopassive coatings

When exposed to the circulation, the surface of bare metal devices such as FDs is coated with proteins that can trigger undesirable reactions including inflammation, contact activation of the intrinsic coagulation cascade, and platelet activation.16 17 One way to avoid these reactions and their associated sequelae is to mask the device surface from circulating substrates such as platelets, fibrinogen, clotting factors, and leukocytes, among others. This can be accomplished by mimicking natural hydrophilic circulatory system components that do not induce these deleterious processes, or with non-interacting polymers that interrupt natural device surface interactions.18 Either strategy can impart reduced thrombogenicity while still allowing for normal neointima growth and proliferation.

Phosphorylcholine

Phosphorylcholine is a component of phosphatidyl choline headgroup and is the major constituent of all eukaryotic cell membranes, including red blood cells. Phosphatidyl choline headgroups do not induce platelet activation or allow for effective adhesion. It creates a hydrophilic boundary layer, discouraging protein conformational changes required for intrinsic pathway activation.19 It was perhaps the original coating proposed for vascular stents.20 It has been well studied on devices in the coronary vasculature; however, the vastly different cerebral vascular environment necessitates additional research and consideration. Medtronic (Irvine, California, USA) has built on coronary stent coating experiences to design a phosphorylcholine surface modification (Shield) for their Pipeline and Vantage FDs. It is a 3 nm thick chemically modified surface created from phosphorylcholine covalently linked to the devices’ braided wires.

Early in vitro and ex vivo work showed that this surface modification resulted in significantly less prothrombin proteolytic activation and platelet deposition compared with uncoated devices. In vascular flow loops, the results were approximately the same as in the control group with no device.21–23 However, these model systems only test the intrinsic pathway for contact activation of blood products. In vivo modelling is required to study both the intrinsic and extrinsic pathways. More recently, rabbit model data showed that Shield-coated devices had significantly less thrombus formation than uncoated devices, as measured with intravascular optical coherence tomography, even in animals not treated with DAPT.24 25 It is worth noting that coated devices also had less side branch thrombus formation, an important consideration for implant sites in the distal cerebral vasculature.24 The coating does not negatively affect aneurysm occlusion rates and reduces neointimal hyperplasia, with earlier and more uniform neointimal growth compared with uncoated devices.26 27 In a rabbit model, the Shield modification did not interfere with longitudinal healing, which is a concern with many surface modification strategies.28

With the efficacy and antithrombotic properties of the coating so well established, the next step is to evaluate whether this modification allows for SAPT, a particularly important question for patients with a ruptured aneurysm for whom DAPT carries significant risks.29 Early case reports demonstrate conflicting results,30 31 but some recent literature is promising.32–34 The largest multicenter retrospective study reported success in treating 14 ruptured aneurysms with Shield-coated devices and aspirin monotherapy with few ischemic and hemorrhagic complications.35 Continued work with larger patient populations and more standardized antiplatelet regimens is needed to completely understand the safety of this practice. The regulatory challenge has been to find the indication in which the risk–benefit ratio for SAPT would be acceptable. The ELEVATE trial (clinicaltrials.gov, NCT04391803) will test short-duration DAPT (48 hours) for acutely ruptured shallow intradural aneurysms treated with FDs having Shield technology.

Glycans

Glycans are another ubiquitous constituent of biological systems. Designed to mimic the glycocalyx and reduce hydrophobic interactions, HPC (Phenox) is a multilayer surface made of glycans covalently bonded to braids of the p48 and p64 FDs.

Initial experiments in vitro showed that, based on immunohistochemistry, there is significantly less platelet adherence and activation with HPC-coated FD wires compared with similar uncoated wires.36 37 Given that the p48 device is designed for more distal circulation,38 it is important to consider side branch clot formation. In a porcine model investigating coated and uncoated devices with various antiplatelet regimens, both the HPC coating and systemic administration of aspirin were shown to significantly reduce thrombus formation over side branches, and the combination of the two eliminated 99.8% of side branch clots.15

Although this is the only surface modification on the market that comes with instructions for SAPT after implantation in certain regions and when DAPT is contraindicated, results in the literature are mixed. A prospective study investigating these devices in the treatment of distal intracranial aneurysms with aspirin alone was stopped after eight patients were treated due to a very high rate of ischemic complications.39 Similar findings have also been reported in retrospective studies treating patients with ruptured aneurysms.40–42 Thromboembolic complications all occurred under SAPT with aspirin, and there is evidence that HPC-coated devices may be safe with SAPT when P2Y12 inhibitors are used.43–46 The COATING trial (clinicaltrials.gov, NCT04870047) is the first randomized clinical trial designed to test whether SAPT with prasugrel is as safe and effective as DAPT in the treatment of aneurysms with p64 HPC FDs. However, it is possible that the omission of aspirin, the weaker antiplatelet agent in the pairing, may limit the benefit of SAPT.

Poly(2-methoxyethyl acrylate) (PMEA)

PMEA belongs to a family of synthetic flexible biocompatible polymers including poly(ethylene glycol), poly(2-ethyl-2-oxazoline), poly(2-methoxyethyl acrylate) and others that disrupt normal interaction with circulating proteins and cells.47 48 PMEA is a thromboresistant polymer with years of medical device application evaluation, including extracorporeal membrane oxygenation machines. When any hydrophilic polymer coating is exposed to an aqueous environment it hydrates, forming a protective layer. PMEA attracts intermediate water molecules, which interact only weakly with the methoxy group of PMEA. When crystallizing and weakly associated, the dynamics of intermediate water appear to be involved in disrupting the interactions between the polymer and various proteins (eg, fibrinogen and factor XII). As a result, the proteins do not undergo conformational changes which would activate them and initiate the intrinsic pathway. Therefore, platelets are less likely to activate or adhere to the surface.49 Importantly, this coating does not seem to alter other binding interactions, such as those with endothelial and smooth muscle cells, because the cells use multiple simultaneous interactions between cell adhesion molecules and the surface-adsorbed proteins. This allows for normal endothelial growth over the device surface.

X-technology covalently links PMEA to all surfaces of the dual-layer FRED device (FRED-X; Microvention, Aliso Viejo, California, USA) to impart thromboresistance. In vitro research shows superior hemocompatibility by reducing contact activation. Reported platelet counts, β-thromboglobulin (β-TG), and thrombin–antithrombin complex (TAT) levels were similar between Shield and FRED-X devices, and negative controls with no devices. Additionally, the β-TG and TAT levels of uncoated devices were significantly increased while platelet count was significantly decreased.50 On SEM analysis of the devices, the fewest platelet deposits were noted on the FRED-X device and immunohistochemistry findings suggest that the coating did not impede endothelial cell attachment.50 The multicenter FRESH study supported in vitro findings with results from 161 patients treated with FRED-X under several DAPT regimens depending on institutional standards. The study reported a low rate of thrombotic complications (4.3%) and major adverse events (3.1%), while 66% of the aneurysms were completely occluded at the study endpoint.51 These positive findings raise the possibility that the antithrombotic properties of the coating could make these devices safe with SAPT, but clinical validation is still required.

Drug bound devices

An alternative strategy for creating antithrombotic device coatings involves binding active drugs to the surface to reduce or eliminate the need for systemic drug therapy. In these examples, the drug is immobilized permanently to the surface of the device rather than eluted over time. The advantage of this approach is that it addresses the presence of the foreign material and the underlying damage to the intima during device deployment, which activates the extrinsic coagulation pathway. Data have shown that simple microcatheterization leads to endothelium denudation,52 which exposes thrombogenic subendothelial structures (figure 2). An active drug coating could theoretically limit thrombosis in this setting, or potentially lyse clots as they form.

Figure 2

Scanning electron microscope image from a cross section of a canine anterior spinal artery after a microcatheter was passed through. Denudation of the endothelium on the vascular wall along the path of the microcatheter can be seen (arrowheads), as well as the path of the microwire (arrow).

Immobilized heparin

Heparin coating to reduce device thrombogenicity has been explored since the 1960s.53 54 A staple anticoagulant for 100 years, heparin enhances the natural antithrombotic mechanisms of the coagulation cascade. Specifically, it activates antithrombin III, which primarily inactivates factor V and thrombin, as well as other serine proteases.55 This, in turn, also inhibits platelet adhesion and activation.56 The decreased adhesion is likely due to reduced fibrinogen deposition, while the reduced activation is due to a reduction in platelet-stimulating factors like thrombin.54 57 These additional properties highlight the promise of using a heparin-coated neurovascular device without DAPT.

Heparin coating was an exciting approach to prevent stent thrombosis in the coronary circulation before the ubiquity of drug-eluting designs. Although the larger registry studies are inconclusive,58 some have suggested heparin coatings to be safe and effective at reducing subacute stent thrombosis.59 Additionally, the HOPE trial demonstrated a 30-day stent thrombosis rate of 1% in 200 treated coronary arteries under aspirin monotherapy.60

To build on this proven technology, Stryker recently entered a partnership with Carmeda (a wholly owned subsidiary of W.L. Gore Associates, Newark, Delaware, USA), seeking to improve their Evolve FD with a surface of covalently bonded heparin molecules (figure 3). Some early evidence supports the safety and efficacy of this surface modification in the neurovasculature,61 but it has yet to be applied to specific neurovascular devices. This coating has shown promise recently in a canine model. Both coated and uncoated FDs were implanted in the basilar artery. Animals in the coated stent group had fewer magnetic susceptibility artifacts on susceptibility weighted imaging than those in the bare stent group after 1 week, implying less thromboembolism. The coated devices also showed a trend to less acute thrombus formation on optical coherence tomography immediately following implantation compared with controls. These results suggest that the coating confers a reduced risk of thromboembolic events without a difference in efficacy.62 63 Future studies will look to build on these preliminary results.

Figure 3

(A) The anticoagulant effects of freely soluble heparin systemically administered, which is dependent on concentration and molecular weight. (B) Heparin attached and immobilized on the surface of an implantable device which inhibits clotting factors by antithrombin at the material interface. (C) heparan sulfate proteoglycan of the endothelium which naturally binds antithrombin and inhibits clotting factors. (Figure from Advanced Drug Delivery Reviews 2017;112:12–23).54

Fibrin nano-coating with bound heparin

Fibrin-heparin coatings are being investigated to both decrease device thrombogenicity and improve endothelial ingrowth. Activated platelets aggregate and begin clot formation by binding fibrinogen through GPIIb/IIIa.55 Platelets also aggregate through GPIb and integrin-mediated interactions with fibrin. Unlike typical aggregation mechanisms, this alternative pathway depends on activated thrombin, the coagulation cascade’s final product.64 Heparin, which induces thrombin inactivation, can therefore be covalently bound to a fibrin coating, creating a surface that is highly conducive to endothelial growth and inert to coagulation.65 In vitro, foil sheets with this coating have significantly more bound endothelial cells compared with uncoated foil after incubation with a medium containing isolated endothelial cells. The fibrin-heparin-coated sheets were also found to be significantly less thrombogenic than plain fibrin-coated and bare metal sheets. This was reflected by platelet counts and SEM imaging after incubation with whole blood.66

Acandis (Pforsheim, Delaware, USA) has used these principles to create a fibrin-heparin nanocoating for their DERIVO 2Heal embolization device. In a rabbit model, the device showed similar biocompatibility to bare stents with significantly less fibrin and platelet deposition. The endothelial growth over the device did not appear to be different between the coated and uncoated stents.67 To further investigate the long-term angiographic outcomes and safety of this coated device, Acandis is sponsoring the REheal observational trial (clinicaltrials.gov, NCT05543447), which will soon be enrolling patients. SAPT with this device appears promising,68 but larger studies are needed to corroborate the preliminary data.

Immobilized FXa-GPIIb/IIIa inhibitors

Another compelling drug choice is EP224238, which has the actions of both the antiplatelet drug tirofiban and the anticoagulant drug idraparinux. The combined antiplatelet and anticoagulant activity may reduce stent thrombosis substantially without systemic agents. Coated and uncoated discs showed similar results on platelet adhesion assays; however, qualitative platelet morphology analysis on SEM did demonstrate reduced platelet activation. Furthermore, drug-coated discs had good anti-FXa activity after exposure to whole blood, and human endothelial cell adhesion was unaffected.69 Theoretically, targeting FXa activation may also have advantages over heparin in certain patient populations as it does not require native antithrombin III activity. While we are not aware of any FD currently being developed with this coating, this concept is theoretically compelling.

Modifications to improve endothelial growth

While the previous surface modifications focused on reducing device thrombogenicity, another strategy for decreasing the requirement for DAPT is to increase the rate of implant endothelialization. Theoretically, DAPT should only be required for as long as the device is exposed to circulating blood.70 Accelerating endothelial ingrowth could theoretically decrease the amount of time patients spend on DAPT and the risk of thromboembolic complications.

Anti-CD34 antibody

Circulating CD34+ endothelial progenitor cells were shown to result in endothelization of implanted devices.71 Surface modification of FDs with covalently bound anti-CD34 antibodies should more efficiently bind circulating progenitor cells and may decrease the vasculature exposure time to a device. OrbusNeich has combined this antibody coating with a more traditional unidirectional drug-eluting polymer coating in their COMBO Plus coronary stent.

Early work showed that the addition of the antibody surface modification allowed for the maintenance of the antiproliferative effects of sirolimus while also enhancing the endothelial growth over the device.72 The results of the REMEDEE registry further support this technology in the coronary circulation. Of the 1000 patients enrolled over the 5-year study, only 0.9% experienced stent thrombosis, with no cases reported after 3 years.73 These results are promising, and it may be worthwhile to consider the development of this technology on devices designed for the cerebral vasculature.

UV-irradiated nitinol

UV irradiation has been shown to alter the wettability of titanium dioxide surfaces.74 Building on these findings, another possible strategy for improving endothelial growth is through UV irradiation of nitinol. Electron transfer reactions via UV irradiation result in the conversion of Ti4+ to Ti3+, which has affinity for dissociative water adsorption.75 The UV-modified device surface led to increased albumin absorption, which is known to be inversely correlated with thrombogenicity. Endothelial cells were also shown to migrate over the surface more easily than on non-irradiated metal.75

CD31 analogs

CD31, also known as PECAM-1, is a transmembrane glycoprotein selectively expressed on platelets, leukocytes, and endothelial cells (figure 4).76 The molecule decreases the activation of leukocytes and platelets and may improve the healing of the vascular endothelium.77 Given this role, surface modification with an analog could conceivably improve device incorporation into the vessel wall, while also decreasing implant site inflammation and thrombosis.

Figure 4

(A) Principal functions of CD31. CD31 is highly expressed at the surface of endothelial cells, platelets, and leukocytes. It plays key regulatory functions for circulation homeostasis on intercellular trans-homophilic binding. (B) Schematic of the P8RI peptide grafted onto the flow diverter strut. P8RI acts as a CD31 agonist by binding to a juxtamembrane CD31 domain. It aims at favoring endothelialization over the flow diverter while preventing thrombosis and inflammation. (Figure courtesy of Professors Aymeric Rouchaud and Jonathan Cortese).

It has been shown that the P8RI peptide, an analog of the extramembrane portion of CD31, can be linked to the surface of Silk Vista Baby devices. In vitro, endothelial adhesion was improved and reaction with blood components was significantly decreased.78 In a rabbit model, endothelial ingrowth was improved and platelet activation decreased. Importantly, even though the coating stimulated neointimal growth, the vessels remained patent with adequate caliber.78 Surface modification with a CD31 analog is a promising approach to promote rapid endothelial coverage.