Stent placement in patients with acute subarachnoid haemorrhage: when is it justified?

Coil embolisation has become the standard, evidence-based treatment for ruptured brain artery aneurysms since the landmark ISAT trial [1]. Subsequent technical developments have included balloon-assisted coiling for broad-based aneurysms [2] and endovascular stenting [3]. Balloon-assisted coiling is routinely performed in patients with ruptured aneurysms. However, endovascular stents are thrombogenic and require concomitant antiplatelet medication. This presents two problems in patients with acute intracranial bleeds; firstly, it increases the risk of bleeding complications, and secondly, the delayed onset of action of antiplatelet medication makes early stent thrombosis more likely. Therefore, stents are predominantly used for the treatment of unruptured aneurysms [4]. Despite the risks, certain situations may lead to the placement of a stent after subarachnoid haemorrhage; for example, if an aneurysm is too broad-based to safely contain a coil, if there is a vessel dissection or if there is the need to preserve flow when coil embolisation is complicated by coil migration or thrombus formation.

Adequate planning and patient consent require knowledge of the complication rates of acute stenting, especially where the use of a stent is considered prior to the procedure. It is also of great importance to determine the optimal antiplatelet medication for acute stenting and to establish the optimal device for this indication.

This study was conducted with appropriate institutional approval to audit against good practice and published treatment results. The prospectively kept central intracranial stent registry of the department was reviewed to identify all stenting procedures in patients with acute subarachnoid haemorrhage from 2008 to 2016. Acute stenting procedures were defined as those occurring within 30 days of presentation at hospital with acute ruptured subarachnoid haemorrhage. Relevant imaging studies were reviewed on the institutional PACS and the patients’ notes were retrieved to assess clinical outcomes.

All procedures were carried out under general anaesthetics by various operators. Aneurysm anatomy was assessed before the procedure by CTA according to the institutional standard protocol and/or by 3D DSA. For most patients, a standard 6F guide catheter was used. All patients received a bolus of heparin.

Due to the 9-year timespan of the review, various devices were used, reflecting the changing operator preferences and technical developments. As a result, a variety of delivery systems were also used: Prowler™ 21 for Enterprise™ (Codman Neuro, Raynham MA) and LVIS™ stents (Microvention Aliso Viejo CA), Vasco 21™ for Leo™ and SILK™ (BALT Extrusion, Montmorency, France), Marksman™ for Pipeline embolization device (PED) ™ (Medtronic, Minneapolis, MN), Excelsior SL10™ and Echelon 10™ (Stryker, Kalamazoo MI) for LVIS Jr. and Leo⁺ baby. Early Neuroform™ stents were preloaded on a dedicated delivery system. Later variants of the Neuroform stent were delivered through the Excelsior XT 27™, Rebar 27™ (Neuroform EZ™) or the Excelsior Sl 10™ (Neuroform Atlas™).

Spontaneous acute subarachnoid haemorrhage is a life-threatening condition. While the rupture of an intracranial saccular aneurysm is the most common underlying condition, other causes include dissections of cerebral arteries and fusiform aneurysms. The patients presented here comprise saccular aneurysms, very broad-based aneurysms, intracranial dissections and fusiform aneurysms.

Endovascular placement of platinum coils was shown to be superior to neurosurgical clipping in a prospective randomised trial several years ago [1]. Patients regarded as suitable for both treatment modalities were enrolled into the trial, which showed better functional outcomes and a marginally higher rebleeding rate after coiling. Although only a proportion of patients presenting with subarachnoid haemorrhage were randomised, endovascular treatment has widely replaced clipping as the standard of care for ruptured aneurysms. Coiling has also become the first-line treatment for vascular pathologies outside the ISAT trial, and for which there is no equally robust outcome data.

Endovascular treatment of aneurysms has been aided by subsequent technical developments such as balloons and stents [3]. While the use of balloon-assisted coiling for the treatment of ruptured broad-based saccular aneurysms is now regarded as standard of care [5], the use of stents in the acute situation remains problematic. Thromboembolic complications during balloon-assisted coiling can be reduced by peri-procedural heparin injection. In contrast, stents require platelet function inhibition until they are endothelialised. Depending on stent type, adequate levels of platelet inhibition may require the use of two synergistic drugs [3]. Aspirin, the most commonly used antiplatelet agent, can be injected intravenously and has an immediate onset of action. However, other commonly used drugs such as clopidogrel have a slow onset of action and cannot be administered intravenously (newer oral and intravenous antiplatelet agents were not available at the time of the treatments). For this reason, elective patients are given clopidogrel several days before the procedure, or several hours before the intervention as a loading dose. In acute subarachnoid haemorrhage, where this is not an option, stents are considered only if other, less risky treatment modalities are not available. Such situations include particularly adverse anatomy or an unforeseen procedural problem.

Fifty-one stents of varying types were implanted in 49 patients with acute subarachnoid haemorrhage in our institution during a 9-year period. In this time interval, a total of around 1400 acutely ruptured aneurysms were embolised and more than 300 intracranial stents were deployed (elective and acute cases). Owing to the retrospective nature of our review, and the time interval, a variety of devices were employed. Stent selection was based on device availability, operator experience and underlying vascular pathology. During the initial part of the study period, early versions of Neuroform stents prevailed, supplanted by Enterprise and later the Neuroform EZ and Neuroform Atlas. These devices were used for all indications including dissecting and blister aneurysms. Later, flow-diverting SILK and Pipeline devices were used, primarily in the setting of ICA blister aneurysms [6, 7]. The braided stents LVIS and Leo [8] were also used in this cohort. Until recently, LVIS Jr and LEO⁺ baby were the smallest stents available, and both are compatible with standard 0.017″ coiling microcatheters such as Excelsior SL10 and Echelon 10. This allowed the use of two microcatheters through one 6F guide catheter and ‘bail out’ stenting through a coiling catheter. Older stents require larger microcatheters (Prowler select 0.021″, Excelsior XT 27). Old versions of the Neuroform were preloaded on an ‘over the wire’ delivery system and had to be advanced over an exchange wire. This created additional problems in emergency situations.

Patients with blister aneurysms were treated with single or double stenting at the start of the study period, using conventional stents [9]. Later, these patients were treated with flow-diverting stents without additional coil placement [10, 11].

Braided stents and flow diverters are likely to be more thrombogenic than laser-cut stents. This may be reflected in a slightly higher stroke rate for the LVIS stents, the first braided stent used for acute stenting in our institution. The comparatively lower stroke rates for the Leo stents and flow diverters may reflect increasing experience in managing the antiplatelet medication for these devices.

Four of 13 patients treated with Neuroform and three of 19 of patients treated with Enterprise stents were placed on aspirin alone as maintenance treatment. There was one ischaemic event attributable to these procedures. However, the number of patients is too small to adopt routine single antiplatelet medication after placement of laser-cut stents. While there were two rebleeds from the treated aneurysms, we did not encounter other intracranial haemorrhages such as from the placement of ventricular catheters. A low risk of haemorrhage after application of antiplatelet medication in the setting of acute SAH is consistent with recent literature [12, 13].

Acute clot formation was observed during 11 procedures. Standard treatment was intra-arterial injection of the glycoprotein IIb/IIIa antagonist abciximab [14] which has been long established in our unit. This resulted in at least partial clot resolution in nine patients. Four patients suffered a stroke. There were no clinically significant bleeding events following Reopro administration. ReoPro was only used in one of six LVIS procedures despite a high stroke rate. This may reflect the developing experience with using braided stents, which later lead to more aggressive treatment of clot formation. Treatment with abciximab appears most effective if it is initiated early, upon first signs of clot formation [15]. Higher doses are required if a thrombotic stent occlusion has already occurred. Tirofiban has also been used for this purpose [16] and with similar results.

The death rate was 13% for patients with saccular aneurysms and 10% overall. This appears consistent with other case series of stenting in acute SAH [17, 18] [19] but is significantly higher than among all patients treated for ruptured aneurysms [3].

Stenting in patients with acute subarachnoid haemorrhage carries a higher complication rate than coiling alone. For difficult aneurysms/dealing with complications, it may be appropriate and should be considered on an individual basis.