Treatment of ruptured blood blister aneurysms using primary flow-diverter stenting with considerations for adjunctive coiling: A single-centre experience and literature review

Abstract

Objective

The objective of this article is to conduct a single-centre evaluation and quick literature review of the effectiveness of primary flow-diverter (FD) treatment of ruptured blood blister aneurysms (BBAs), with additional relevance of adjunctive coiling.

Methods

Patients presenting with subarachnoid haemorrhage (SAH) due to ruptured BBAs and subsequently treated with FDs were retrospectively selected from June 2010 to January 2017. Treatment techniques, angiographic data on occlusion rates and procedural success as well as clinical outcomes using the modified Rankin Scale (mRS) were collated. Cross-reference of results were made with available literature.

Results

Thirteen patients harbouring 14 BBAs were recruited. Of the 14 aneurysms, five (35.7%) showed immediate complete occlusion after the procedure (four of these five patients had adjunctive coiling). All of the aneurysms showed complete occlusion by the six- to nine-month control diagnostic angiogram. No rebleed or retreatment was experienced. Twelve of 13 (92%) patients had an mRS score of 0–1 at the last clinical follow-up. From the pooled data of the literature review, eventual aneurysm occlusion was achieved in 48/56 patients, with five patients requiring further endovascular treatment. In the clinical follow-up period, an mRS of 0–2 was recorded for 83.3% (45/54) of patients.

Conclusion

Endovascular reconstruction of BBAs using FD treatment is an effective method with good final clinical outcomes. Adjunctive use of coiling achieves higher incidence of immediate complete occlusion of BBAs.

Introduction

Blood blister aneurysms (BBAs) are small sessile lesions that typically occur at non-branching sites of the dorsal surface of the supraclinoid internal carotid artery (ICA).¹

These aneurysms are rare, contributing to less than 2% of all intracranial aneurysms.² Nonetheless, these account for 2.2% of all subarachnoid haemorrhages (SAH) from a ruptured ICA aneurysm. If left untreated, once ruptured, these demonstrate poor clinical outcomes.³

Histologically, BBAs are associated with dissections,² focal arterial wall loss of the internal elastic lamina and media, with a thin layer of fibrous tissue and/or thrombus covering the defect.⁴ Essentially, such lesions behave as pseudoaneurysms. BBAs pose significant challenges in diagnosis, typically presenting only when ruptured, and even then, may still require repeated angiographic studies before confirmation.⁵ Additionally, taking into account their fragile state and difficult morphology, such lesions prove challenging to manage either surgically or endovascularly once diagnosed.

Various surgical and non-flow diverter (FD) endovascular approaches have often been reported with mixed results.⁶ A recent systematic review by Szmuda et al. showed either endovascular (with limited FD case series) or surgical treatment for ruptured BBAs to be equally effective, though the authors mentioned stent-assisted coiling being the least beneficial due to increased need for repeat treatment.⁷ Collectively, this wide range of approaches highlights the persistent lack of unified treatment guidelines, with difficulty in organisation of a gold standard randomised trial due to the rarity of this condition.

Particularly in recent years, alternative treatment detailing reconstructive management using FDs have emerged from multiple centres, with encouraging results.^6,8–22

FD stents are unique devices designed for the treatment of complex intracranial aneurysms. These act by changing parent artery flow dynamics with diversion of blood from the aneurysm thereby inducing its thrombosis, as well as providing a scaffold for endothelisation to occur, resulting in eventual permanent aneurysm exclusion. Concerns for this treatment modality include the need for dual antiplatelets post-procedure and the associated risks in case of rebleed in a recently ruptured aneurysm, given that thrombosis of the aneurysm is gradual. Moreover, use of antiplatelet therapy may preclude or increase the risk of future surgical procedures which the patient requires.

This current series investigates our single-centre, retrospective experience and approach in primary FD treatment of ruptured BBAs, with mid- to long-term follow-up imaging and clinical outcomes. A quick literature review for the purpose of cross-reference with other centres’ approaches was also performed.

Materials and methods

All patients with intracranial BBAs presenting to our institution from June 2010 to January 2017 were retrospectively selected from a prospectively maintained database under the approval of our institutional review board.

Decision for an endovascular approach utilising FD treatment had been reached following consensus by the institution’s multidisciplinary team inclusive of neurosurgeons as well as interventional neuroradiologists. All selected lesions had an initial presentation of SAH, with characteristic small- and wide-necked configuration, arising from non-branching sites of the distal ICA. Information compiled included Fisher scale at onset, time to treatment, extraventricular drain (EVD) placement, treatment modality (number of FDs utilised/with or without adjunctive coiling), success of procedure and any complications (early or delayed) incurred.

In the clinical follow-up, the latest modified Rankin Score (mRS) was recorded as the final outcome. Regular post-procedural angiographic (three- then six-monthly if uneventful procedure) and computed tomography angiography (CTA)/magnetic resonance angiography (MRA) imaging (performed yearly after two initial diagnostic angiographic follow-ups were shown to be unremarkable with a treated aneurysm) were obtained. The degree of occlusion of the endovascularly treated aneurysm was graded on angiography using the Raymond-Roy classification, with complete obliteration, neck remnant and residual aneurysm, respectively. Any in-stent stenosis and patency of parent and branch arteries were also noted.

Technique

Before FD treatment, CT of the brain was performed. If an EVD was necessary, it was inserted before the endovascular procedure to reduce intracranial haemorrhagic (ICH) risk upon commencement of antiplatelet medications.

FD treatments were performed under general anaesthesia, with intravenous heparinisation to reach an activated clotting time (ACT) of ≥250 seconds. For patients with acute/subacutely ruptured aneurysms, systemic heparinisation was continued for three hours after the procedure, to allow overlap with commencement of effect of the antiplatelet therapy. Intravenous (IV) 500 mg aspirin was administered on an angiographic table just prior to the deployment of the FD, with 600 mg clopidogrel delivered via nasogastric tube (NGT) immediately upon procedure conclusion. Platelet function tests were routinely performed three hours post-procedure and repeated the next day (Multiplate Aggregometer; Dynabyte, Germany), to assess the effectiveness of the anti-aggregant therapy. All patients subsequently received double-antiplatelet therapy for three months post-treatment, and aspirin for a minimum of one year, though typically maintained on the latter indefinitely.

Utilising biplane angiography, coupled with three-dimensional (3D) rotational angiography, working projections, precise parent artery diameters and aneurysm sizes were obtained. Access was secured with a triaxial guide-catheter system including a 6Fr introducer sheath (Terumo Destination; Terumo, Tokyo, Japan); intermediate catheter (Fargomax 070; Balt, France); and 0.027-inch inner diameter microcatheter (XT-27 and Marksman; Covidien-ev3). The FDs were then deployed, and in general, only one device per patient was used in treatment. Additional FDs and whether adjunctive coiling were utilised was up to the operator’s discretion and individual characteristics of the aneurysms. After FD placement, stent positioning, wall apposition, patency of parent and branch arteries as well as the aneurysm were confirmed with an immediate control digital subtraction angiography (DSA) as well as on ConeBeam CT.

Results

Patients

A total of 14 ruptured BBAs from 13 patients were treated in the study (Table 1). There were three males and 10 females, with median age at treatment being 49.6 years, ranging from 33 to 69 years old. Eight patients (61.5%) had a poor initial clinical grade (Fisher scale ≥2) at presentation. Average height of the aneurysms at treatment was 3.0 mm (range of 1.4–5.7 mm), and all were confirmed with DSA utilising 3D reconstructions. Six patients were treated with a single FD, one patient with two overlapping FDs, five patients with a single FD plus coils, and one patient with initial surgical clipping, followed by single FD plus coils due to recurrence of aneurysm after clipping.

Table 1.

Summary of patients’ aneurysm characteristics (all supraclinoid ICA location), treatment technique, angiographic and clinical outcomes.

No.	Sex/ age	FS	A° size (mm) at treatment neck × height	Treatment post-bleed (days)	Treatment technique	RS immediately post-procedure	RS at two- to three-month DSA	RS at six- to nine-month DSA	Complication description	mRS at last follow-up
1	F/69	4	2.5 × 4.0	14	1 PED	3	2	1		1
2	F/44	1	3.0 × 3.0	18	1 PED	3	1	1		0
3	F/33	1	3.0 × 3.0	30	1 PED	3	NA	NA	Slow flow in left ophthalmic artery post-procedure. No follow-up imaging.	1
4	F/55	3	4.5 × 4.1	13	1 PED + coil	1	1	1		0
5	F/43	3	4.0 × 5.1	51	1 PED + coil	1	1	1		1
6	F/47	1	3.2 × 3.4	9	1 PED + coil	1	1	NA		0
7	F/35	2	2.3 × 2.0	17	1 PED	1	1	1	Stasis in left A1 due to PED. Eventual occlusion left A1 at six-month control DSA, with left A2 supplied via ACOM, and moderate stenosis at distal PED	1
8	F/54	4	2.9 × 2.0	2	2 PED	3	1	1	D6 post-bleed severe vasospasm: Left lower limb paresis, resolved with vasoactive treatment	0
9	M/59	4	3.4 × 2.9	2	1 PED + coil	2	1	NA	Inpatient ventriculitis	1
10	M/60	1	4.3 × 1.4	4	1 PED	3	1	1		0
11	F/52	3	1.4 × 1.5	3	1 PED	3	1	1		0
12	M/50	4	7.0 × 5.7	6	1 PED + coil	3	3 (increased size)	1	Worsening hydrocephalus at third month, with major symptomatic ICH on attempted initial shunt revision	4
13	F/44	4	2.2 × 1.8 (first); 2.6 × 2.5 (second)	5	1 PED + coiling of second A°	1 (second A°), 3 (first A°)	1 (second A°), 2 (first A°)	1 (both A°)	D8 post-bleed severe vasospasm: worsening aphasia/confusion, improved with vasoactive treatment. Subsequently inpatient meningitis	1

A°: aneurysm; ACA: anterior cerebral artery; ACOM: anterior communicating artery; DSA: diagnostic angiogram; F: female; FS: Fisher grade scale; ICA: internal carotid artery; ICH: intracranial haemorrhage; M: male; mRS: modified Rankin Scale; PED: Pipeline Embolization Device; RS: Raymond’s scale.

Time to definitive FD treatment from initial presentation ranged from 2 to 51 days, with an average of 13.4 days. The main outlier (Patient 5) with treatment at 51 days post-presentation was due to initial treatment with surgical clipping and subsequent delay after it recurred. Following treatment, none of the patients had an aneurysm rebleed or needed retreatment.

Periprocedural complications

During deployment of the FD, two patients demonstrated stasis of flow in the covered side branches (one in the ophthalmic artery, the other in the covered left A1 segment), which were otherwise of no clinical significance. No other major intraoperative complication was observed. In the immediate post-procedure period, two patients had clinically significant vasospasm (one presenting with left lower limb paresis and the other with worsening aphasia and increasingly confused state – both with initial Fisher scale of 4), requiring endovascular management with intra-arterial nimodipine, with subsequent good outcomes (mRS 0 and 1 respectively on discharge). Another two patients developed meningitis/ventriculitis secondary to EVD placement whilst inpatients, though again with good recovery.

Incidence of symptomatic ICH related to EVD or ventriculoperitoneal (VP) shunt placement

Five patients (38.5%) had hydrocephalus requiring EVD placement prior to treatment, with no ventriculostomy-related clinically significant sequelae/symptomatic ICH. Of these five patients, Patient 12 had an additional shunt revision of his EVD approximately three months into his rehabilitation due to worsening hydrocephalus and altered mental state. He consequently suffered a procedural-related ICH needing further symptom alleviating EVDs and another final successful shunt revision. Patient 13 had her right EVD converted to the left side 17 days after commencement of treatment/dual-antiplatelet therapy also due to drainage issues, with subsequent VP shunt revision a further three weeks later. In Patient 13’s case, no major symptomatic ICH was experienced, with the procedures performed under thrombocyte transfusion and temporary (one to two days’) cessation of antiplatelet therapy.

Present clinical and angiographic outcome

The status of one patient (Patient 3) was mRS 1 at the third-month post-treatment office visit, but was lost to follow-up thereafter. The average time of follow-up after PED placement for the remaining 12 patients was 22.0 months (range from six months to five years). Twelve of 13 (92%) patients had an mRS score of 0–1 on the last clinical follow-up, with Patient 12 having an mRS of 4.

Of the 14 aneurysms, five (35.7%) showed immediate complete occlusion (Raymond Class I) after the procedure (four of these five patients had adjunctive coiling with FD placement). One of 14 (7.1%) had residual aneurysmal neck (Raymond Class II) and eight of 14 (57.1%) had persistent aneurysm filling (Raymond Class III). Of the eight patients with Raymond Class III, seven demonstrated contrast stasis, with only one patient (Patient 12) showing persistent filling in the non-coiled segment with minimal contrast stagnation. Eleven patients (harbouring 12 aneurysms) had an early control angiogram (two to three months). Nine of these 12 aneurysms (75%) showed Raymond Class I, with two Raymond Class II and one Raymond Class III (Patient 12 – aneurysm increasing in size). By the intermediate (six- to nine-month) follow-up diagnostic angiography imaging period, 11/11 (100%) of the imaged aneurysms showed complete aneurysm occlusion (including Patient 12).

One of 14 patients demonstrated asymptomatic moderate in-stent stenosis in the follow-up DSAs, which did not warrant further intervention. There were no cases of in-stent thrombosis.

A summary of the clinical and angiographic results are demonstrated in Table 1.

Literature review

We performed a literature review from MEDLINE and EMBASE using the key terms “blood”, “blister”, “aneurysm”, “ruptured”, “flow diverter” in both “AND” and “OR” combinations. Available articles were further filtered from 1 January 2010 to 31 December 2016. To facilitate comparison with our study, only cases which detailed FD-treated, ruptured BBAs of a supraclinoid location were included.^8–24 Individual data collected from each of these studies included size of aneurysm, time taken to treatment, treatment technique, relevant periprocedural-related complications, degree of aneurysm occlusion post-FD placement, rebleed and/or recurrence/retreatment despite FD treatment and final recorded mRS score. As there was scanty information of whether and when EVDs were placed, such data were not included.

From a total of 17 studies, 56 patients which fitted the criteria were obtained (Tables 2 and 3). Twenty-three patients (41.1%) were treated in the acute phase of rupture, and a further six patients (10.7%) in the subacute phase (two to eight weeks post-bleed). The remaining 27 patients did not have a detailed period of onset of treatment. Thirty-four patients (60.7%) were initially treated with a single FD, 19 patients (33.9%) with ≥2 FDs and three (5.4%) were treated with a combination of FD and coils. Periprocedural-related complications were largely evenly distributed amongst the main treatment categories (Table 2). Excluding patients in whom degree of immediate post-procedural occlusion were not detailed, an immediate occlusion rate of 27.3% (6/22 patients) of single FDs, 18.8% (3/16) of telescoping FDs and 100% (1/1) of FDs + coils were noted. Eventual aneurysm occlusion was achieved in the majority, 48/56 patients (85.7%), with one additional patient showing persistent opacification at six-month last recorded follow-up (single PED), two patients with no available data provided and five deaths.

Table 2.

Literature overview of FD techniques used, aneurysm characteristics, occlusion details and relevant clinical data (summarizing Table 3 information).

Initial technique	No. of patients	Immediate obliteration					Final mRS
		Y	N	NA	Rebleed	Retreatment/Aneurysm size	0–2	3–5	6	NA	Relevant periprocedural complications
1 FD	34	6	16	12	Case 42: D14 Case 55: D9	Case 29: 1.8 × 1.5 mm Case 42: 3 × 3 × 1 mm, further increase of 2 mm at treatment Case 55: 4.3 × 2.2 mm	27	1	4	2	Case 9: distal ipsilateral haemorrhagic infarct/death Case 11: ipsilateral large MCA infarct Case 42: rupture during procedure, stopped with balloon inflation
≥2 FD	19	3	13	3	0	Case 46:2.5 × 2 × 3 mm Case 56:7 × 6.5 × 5 mm	15	3	1	0	Case 43: Type III endoleak Case 45: clot after second FD needing ReoPro Case 51: flow reduction in ICA/MCA Case 52: infarcts post-procedure in ant. choroidal, MCA/ACA territories
FD + coil(s)	3	1	0	2	0	0	3	0	0	0	Case 10: stent thrombosis/carotid occlusion Case 16: complete ipsilateral blindness
Total	56	10	29	17	2	5	45	4	5	2

ACA: anterior cerebral artery; FD: flow diverter; ICA: internal carotid artery; MCA: middle cerebral artery; mRS: modified Rankin Scale.

Table 3.

Overall literature review of FD techniques used, aneurysm characteristics, occlusion details and relevant clinical data.

Article	Size of aneurysm (mm)	Prior treatment	Time to FD treatment (days)	Treatment method	Relevant treatment-related complications		Immediate occlusion	Earliest occlusion (months)	Rebleed post-FD treatment	Recurrence/ retreatment	Latest mRS outcome
Princiotta et al. 20118	NA	Neuroform +  coils	10 weeks	1 SILK + coils	N		Y		N	N	1
Çinar et al. 20139	4 × 1.5		15	1 PED	N		NA	6	N	N	1
Çinar et al. 20139	3.5 × 2	D2 clip. D16 coil	26	2 PEDs	N		NA	6	N	N	4
Çinar et al. 20139	3.5 × 2		60	1 PED	N		NA	12	N	N	0
Çinar et al. 20139	4 × 1.5		10	2 PEDs	N		NA	6	N	N	2
Çinar et al. 20139	4 × 3		20	1 PED	N		NA	6	N	N	0
Çinar et al. 20139	4 × 2		35	1 PED	N		NA	6	N	N	3
Yoon et al. 201410	1.5 × 2		NA	2 PEDs	N		NA	4	N	N	0
Yoon et al. 201410	2 × 2		NA	1 PED	Distal haemorrhagic infarct, death		NA	NA (death)	N	N	6
Yoon et al. 201410	2.6 × 1.6		NA	1 PED + coil	Stent thrombosis +  carotid occlusion		NA	6	N	N	1
Yoon et al. 201410	2 × 2		NA	1 PED	Ipsilateral MCA infarct		NA	6 (persistent neck remnant)	N	N	1
Yoon et al. 201410	1.0 × 1.5		NA	1 PED	N		NA	NA	N	NA	NA
Yoon et al. 201410	2.5 × 3.2		NA	1 PED	N		NA	3	N	N	0
Yoon et al. 201410	2 × 3		NA	1 PED	N		NA	6	N	N	0
Yoon et al. 201410	3 × 2, 3 × 2		NA	1 PED	N		NA	3	N	N	1
Yoon et al. 201410	2.96 × 1.96		NA	2 PEDs + coil	Ipsilateral blindness		NA	18	N	N	2
Yoon et al. 201410	2.33 × 1.04		NA	1 PED	N		NA	9	N	N	0
Rasskazoff et al. 201011	2		NA	2 SILKs	N		N	18 days	N	N	0
Consoli et al. 201213	7 × 5	Enterprise stent + coil	∼2weeks	1 PED	N		N	3	N	N	NA
Aydin et al. 201514	2.5 × 2.5		6	1 SILK	N		N (Class III)	3			0
Aydin et al. 201514	2.5 × 2.0		5	2 SILKs	N		N (Class III)	NA (death sepsis)	NA	NA	6
Aydin et al. 201514	3.0 × 4.0		9	1 SILK + 1 enterprise	PAO at three months		N (Class III)	PAO at three months	N	N	2
Aydin et al. 201514	2.0 × 3.0		8	2 SILKs	N		N (Class III)	3	N	N	0
Aydin et al. 201514	2.0 × 2.5		12	1 SILK	N		N (Class II)	Short term <2 weeks	N	N	0
Aydin et al. 201514	2.0 × 3.0		14	1 SILK	N		N (Class III)	3	N	N	0
Aydin et al. 201514	2.0 × 3.0		10	1 SILK	N		N (Class III)	3	N	N	1
Aydin et al. 201514	1.5 × 2.5		14	1 SILK	N		N (Class III)	3	N	N	0
Aydin et al. 201514	2.5 × 6.0		5	2 SILKs	N		N (Class III)	3	N	N	0
Linfante et al. 201715	1.8 × 1.5		NA	3 PEDs over two procedures (1 + 2)	N		N (enlarging despite multiple PED)	NA (death)	N	Y	6
Linfante et al. 201715	1 × 1		NA	1 PED	N		Y		N	N	0
Linfante et al. 201715	1.5 × 2.8		NA	1 PED	N		Small filling	Complete occlusion	N	N	0
Linfante et al. 201715	0.75 x 1.5		NA	1 PED	N		Small filling	Complete occlusion	N	N	0
Linfante et al. 201715	1.5 × 1.3		NA	1 PED	N		Small filling	Complete occlusion	N	N	0
Linfante et al. 201715	1.3 × 1.2		NA	1 PED	N		Y		N	N	0
Linfante et al. 201715	1.5 × 1.2		NA	1 PED	N		Y		N	N	1
Linfante et al. 201715	1.5 × 1.3		NA	1 PED	N		Y		N	N	0
Linfante et al. 201715	2.1 × 2		NA	1 PED	N		Y		N	N	0
Linfante et al. 201715	1.6 × 1.3		NA	1 PED	N		Y		N	N	0
Hu et al. 201416	1.8		One week	2 PEDs	N		Contrast stagnation	Six weeks	N	N	0
Hu et al. 201416	1.2		NA	2 PEDs	N		Contrast stagnation	Seven weeks	N	N	0
Hu et al. 201416	1.5 (3.5 at treatment		∼2 weeks	2 PEDs	N		Y		N	N	0
Nerva et al. 201517	3 × 3 × 1 (increase height 2 mm at tx)		9	2 PEDs over two procedures	Rupture in procedure (stopped with balloon)		Rebleed D14, f urther increase size. No rebleed after 2 FDs	NA (death)	Y	Y	6
Nerva et al. 201517	2.8 × 2 × 1		5	2 PEDs	Type III endoleak		Contrast stagnation	3 weeks	N	N	0
Gonzalez et al. 201421	NA		NA	Multiple FD	N		Y		N	N	1
Gonzalez et al. 201421	1		NA	2 PEDs	Clot after second FD		Y		N	N	3
Gonzalez et al. 201421	2.5 × 2 × 3		NA	3 PEDs over two procedures (2 + 1)	N		No stasis. Third PED, complete occlusion		N	Y	1
Chalouhi et al. 201422	2		7	1 PED	N		N	3	N	N	0–2
Chalouhi et al. 201422	2		13	1 PED	N		NA	NA	N	NA	0–2
Chalouhi et al. 201422	3		10	1 PED	N		N	4	N	N	0–2
Chalouhi et al. 201422	3		2	1 PED	N		N	6; involution of aneurysm	N	N	0–2
Causin et al. 201119	3		NA	3 SILKs	Flow reduced initially		Contrast stagnation	7	N	N	0
Causin et al. 201119	<2		2	2 SILKs	Infarct in the Ach, M1 and ACA territories		Partial exclusion	7 days	N	N	Aphasia and right hemiplegia
Causin et al. 201119	5		1	3 SILKs	N		Significant reduction	8	N	N	0
Martin et al. 201220	4.8 × 3.0		4	2 PEDs	N		Persistent	9	N	N	Clinically well
Mazur et al. 201623	4.3 × 2.2		<5	2 PEDs over two procedures	N		N	9 days	Y	Y	6
Lang et al. 201724	7.0 × 6.5 × 5		NA	3 PEDs over two procedures (2 + 1)	N		N	NA	N	Y	0–1

ACA: anterior cerebral artery; FD: flow diverter; ICA: internal carotid artery; MCA: middle cerebral artery; mRS: modified Rankin Scale; N: no; PAO: parent artery occlusion; PED: Pipeline Embolization Device; tx: treatment; Y: yes.

Five patients received further endovascular treatment in the form of additional FDs, of whom three died. Two of the five patients (Cases 42 and 55) presented with rebleeds soon after placement of an initial single FD, with eventual deaths. Of the other three cases, two had an interval increase in aneurysm size (cases 29 and 56), whereas another was due to persistent opacification of the sac (Case 46) post-treatment.

In the clinical follow-up period, an overall good outcome (mRS 0–2) was recorded for 83.3% (45/54) of patients.

Discussion

BBAs typically present with SAH, generally afflicting a younger population than those harbouring saccular aneurysms.³ There is also significantly poor preoperative clinical condition (68% with a World Federation of Neurological Surgeons (WFNS) scale score greater than 3),²¹ with BBAs demonstrating rapid growth, high spontaneous rebleed rates²⁵ and typical radiologic evolution of a ruptured BBA to a saccular shape within days.²⁶

Ruptured BBAs also have an associated high rate of treatment-induced rebleed and death, regardless of most treatment modalities.^21,27–29 However, FD therapy for BBAs has emerged with promising results, gaining increasing popularity in recent years.^8–22 A meta-analysis by Rouchaud et al. has provided supportive data, showing improvement of FD treatment of BBAs over other endovascular reconstructive techniques (i.e. coiling, stenting or stent-assisted coiling), with lower retreatment rates, better mid-/long-term complete occlusion, with similar mid to long term good clinical outcomes.³⁰

Favourable results have also been demonstrated in our case series, with 92% of treated patients demonstrating an mRS score of 0–1 in the latest clinical follow-up. Eventual complete obliteration of the aneurysm was also achieved in 100% of patients.

Single/multiple FD versus FD with conjunctive coiling?

It has been postulated that the reconstructive aspects of FD therapy likely accounts for effective immediate prevention of rebleed/rerupture of a BBA,¹⁴ hence in theory, a single FD may be sufficient for treatment. Several reports also exist of success with stents alone for the treatment of small dissecting and BBAs, further supporting the notion of stent-induced remodelling of the parent vessel.³¹ However, whilst intraoperative contrast stasis within the aneurysm may suggest effective flow diversion, exactly when it predicts complete lesion thrombosis and neck endothelisation cannot be determined.³² Crucially, a reduction of intra-aneurysmal flow does not equate to a reduction of the pressures experienced, with a few cases illustrated by Cebral et al. using computational fluid dynamic models demonstrating why some aneurysms still rupture even after FD placement.³³

The review of other similar case reports/series (Table 2) showed a preferential use of single FD (60.7%), with rate of immediate complete occlusion of BBAs being 27.3% (single FD) and 18.8% (telescoping FD), respectively. The number of patients (three) receiving FD + coil as initial treatment was too small for significant analysis.

In our case series (Table 1), there was similar notable lack of immediate complete lesion thrombosis following deployment of a single or telescoping PED(s) (only one of eight – 12.5%). The natural course of BBAs are known to be malignant,^25,26 and patients being placed on dual-antiplatelet coverage despite a not entirely secured lesion is risky. Furthermore, any inclination for future coiling of recurring/expanding aneurysms which had already been jailed by an FD would have been impossible. Concerns are highlighted in case examples of persistent aneurysm flow/growth despite FD placement,^15,17,21,24 and more crucially, two cases^17,23 detailing a re-rupture soon after a single FD treatment, both eventually leading to death.

As such, we propose for lesions of suitable morphology demonstrating height of at least 2.5 mm or larger, a strategy of FD and concomitant coiling to promote earlier aneurysm complete occlusion. This has been shown by other authors to be a potentially successful and durable technique for BBAs.^8,10 Although of a small sample size, four of six (66.7%) BBAs treated in our institution in such a manner showed immediate complete occlusion, one of six in the early three-month angiographic follow-up, and the last patient showing increase in size (tight coil packing could not be achieved due to the parent vessel anatomy and awkward lesion angulation/morphology) on the early follow-up, with subsequent complete occlusion on the intermediate seven-month diagnostic angiogram. None of our patients had rebleed following treatment and results were overall positive. Potential drawbacks of this technique include increased vessel injury and intraprocedural prothrombosis from added device manipulation, especially in the setting of pre-existing vasospasm.¹⁰ Also, there is the risk of perforation and rehaemorrhage of the BBA whilst coiling.³⁴ Other options for lesions not showing immediate occlusion involve placing telescoping stents.^15,17,19,21 In theory, the increased metal coverage over the BBA would provide more significant flow diversion and increased support to the adjacent diseased wall. This may however expose the patient to increased risk of parent or branch artery occlusion, thromboembolic events and in-stent stenosis.^10,14,19,21 In addition, such a technique also does not guarantee immediate/permanent occlusion of the aneurysm, as illustrated in Cases 46 and 56 (Table 2), which subsequently required additional FD treatments.

Optimal time to treatment/initiation of dual antiplatelets

There has been so far no consensus regarding the timing of initiation of FD treatment (and hence commencement of dual-antiplatelet medications) in the setting of recently ruptured BBAs.

Some institutions have advocated delaying the procedure till the subacute phase to allow clot stabilisation/fibrin plug of the ruptured BBA to mature, thereby possibly being less susceptible to the effects of antiplatelet therapy.⁹

In our institution, we similarly adopted a watchful waiting approach (for about two weeks each after initial presentation) of our first few patients (Patients 1, 2 and 4), with commencement of dual antiplatelets only on table. Similarly, for Patients 5 (recurred lesion following initial clipping), 6 (delayed transfer/diagnosis from another hospital) and 3 and 7 (delayed diagnosis), treatment was not immediately provided following presentation.

During the observational period from initial diagnosis to treatment, we noted a majority of these aforementioned lesions showed subtle yet significant increase in size on repeat DSAs, keeping with the natural course of the disease.²⁶ Based on the increased risk of exposure to rupture with a longer waiting period, coupled with positive final outcomes in other studies with almost immediate treatment at presentation,^19,20,22 for subsequent patients (Table 1, Patients 8–13), we advocated treatment within one week of haemorrhage.

Between these two approaches toward treatment (though sample size is small), we did not experience any significant difference in terms of interval rebleeding (none), directly related periprocedural complications encountered or final mRS outcome (all patients were mRS 0–1, with the exception of Patient 12).

Another important factor for consideration when to treat involves vasospasm, which peaks particularly in the six- to eight-day post-bleed period. Severe vasospasm remains an important cause of morbidity and mortality, with delayed ischaemic neurologic deficits (symptomatic vasospasm) in 17%–40%.³⁵ Vasospasm could also pose significant challenges during device manipulation, with increased risk of vessel dissection or perforation in spastic vessels; accurate FD sizing may also be difficult. It may theoretically also promote thrombosis of the covered smaller branches, which are more sensitive to flow reduction.³⁶ In our case series, two patients with severe intraprocedural vasospasm (Patients 12 and 13) were treated on day 6 and day 5 of presentation respectively, both with initial modified Fisher scale of 4. No intraprocedural complications were encountered for either, though to minimise risk, treatment should preferably be undertaken prior to the vasospastic window.

Antiplatelet management in patients requiring EVDs/VP shunts

A main concern with the patient on dual antiplatelets is the risk of bleeding in those requiring invasive procedures such as EVDs or VP shunts, with as yet no established guidelines for the management of antiplatelet therapy in such patients.

In a study by Kung et al., EVD placement in patients undergoing stent-assisted coiling (with dual-antiplatelet coverage) for acutely ruptured aneurysms had an 8% rate of symptomatic haemorrhage, as compared to 0.9% for conventional coiling without concomitant dual-antiplatelet therapy. In these cases, the ventriculostomy was oftentimes placed prior to FD (and hence dual antiplatelet) onset.³⁷ A systemic literature review by Bodily et al.³⁸ showed clinically significant EVD-related bleeding complication rates in patients on antiplatelets who had undergone stent-assisted coiling in acutely ruptured aneurysms to be between 3% and 10%. Conversely, a small case series by Tumialán et al. revealed that of the two patients who received ventriculostomy placement after FD/antiplatelet commencement, both suffered intraventricular haemorrhage and death.³⁹

Premature discontinuation of antiplatelet therapy in favour of surgery, however, exposes the patient to risks of increased stent thrombosis, with raised platelet adhesiveness secondary to a rebound effect.⁴⁰ As primary studies into neurovascular stent/FD thrombosis prophylaxis protocols have been sparse, current data have largely been based on cardiac literature.⁴¹ Prior research on stopping clopidogrel for major surgery within the first three weeks after drug-eluting coronary artery stent placement increased mortality by 30% to 86%.⁴⁰ Patients undergoing non-cardiac surgery during the coronary stent endothelialisation window (estimated at three to four months for bare metal stents⁴²) are at 5 to 10 times greater risk of stent thrombosis as compared to matched patients for whom surgery was deferred.⁴³

In our institution, it is routine practice to assess and if required, place EVDs a few days prior to FD treatment (and hence dual-antiplatelet commencement). All five preprocedural EVDs performed did not suffer any related symptomatic intracranial haemorrhage. However, for patients who have already commenced dual antiplatelets needing additional intracranial invasive procedures, a very careful balance of the anti-aggregant medication must be sought.

Firstly, delay of any surgery (particularly if not urgent) for patients on dual antiplatelets is advocated,⁴⁴ till at least stent endothelisation has occurred. When delay is not an option, some guidelines recommend discontinuing aspirin and clopidogrel five to seven days prior to surgery and restarting antiplatelet therapy within 24 hours after operating.⁴⁵ Other regimens include discontinuing clopidogrel seven days prior to surgery while maintaining aspirin and possibly bridging therapy with eptifibatide and heparin.⁴⁶ An alternative, at least in the setting of VPS placement, would be performing the procedure even under dual-antiplatelet therapy, as published by Mahaney et al.⁴⁷ The authors had shown this to be feasible in a small subgroup, with low risk of symptomatic ICH. However, these studies mostly did not include patients with intracranial disease and given the lack of adequate prospective studies, each case must be managed on an individual basis.

For Patient 12 (Figure 1(a)–(e)), shunt revision was performed due to symptomatic progressive hydrocephalus, with increased drowsiness and fluctuating consciousness. Both antiplatelets, which had already been administered for approximately three months, were stopped one week prior to shunt placement. This was with the dual intention of hopefully promoting further thrombosis of the BBA (which was enlarging despite prior coils and an FD), as well as decreasing the risk of iatrogenic ICH. A regimen of low molecular-weight heparin (LMWH) was simultaneously started. LMWH was chosen, as it induces rapid inhibition of thrombin, a key component of both the blood coagulation and platelet aggregation cascade. It would therefore be effective in the prevention and treatment of arterial thrombosis, as evidenced from results of prior clinical trials.⁴⁸ In addition, unlike antiplatelets, its onset of action is almost immediate, has a much shorter half-life, and most important, its actions are reversible, which would be beneficial in case of severe haemorrhage. Despite medication modification, Patient 12 still had a moderate VPS-related intraventricular haemorrhage, requiring further symptom-alleviating EVDs and another final successful shunt revision. On subsequent discharge with rehabilitation, he had re-attained his preadmission baseline mRS of 4, and was continued only on a therapeutic dose of LMWH. An intermediate seven-month (i.e. four months after cessation of antiplatelets for shunt revision) follow-up diagnostic angiogram was then performed. This showed a patent FD, with interval complete occlusion of the aneurysm. No additional thromboembolic events, stent stenosis or occlusion was noted. Patient 12 was a clear outlier in terms of his protracted clinical course. We attribute his poor outcome to the severity of his original SAH and subsequent vasospasm, complicated by the late haemorrhagic events related to his hydrocephalus and subsequent shunt operations. We do not think there was any causative relationship between the pre-treatment aneurysm size and his clinical outcome.

Figure 1.

Illustration of Patient 12’s right supraclinoid internal carotid artery blood blister aneurysm. (a) Pre-treatment angiogram performed on first day of admission demonstrating a blood blister aneurysm measuring 4.2 × 2.5 mm. (b) Treatment performed on day 6 of admission shows interval increase in size of the blood blister aneurysm, now measuring 7.0 × 5.7mm. Significant vasospasm was also noted in the proximal segments of the right anterior and middle cerebral arteries, (c) Control angiogram immediately post single FD deployment with adjunctive coiling. There is a still patent component at the antero-medial aspect of the aneurysm which could not be adequately coiled due to morphology, (d) Control angiogram three months post treatment, with interval increase in blood blister aneurysm size and compaction of the coil mass, (e) Control angiogram seven months post treatment, with no residual aneurysm detected, with a well patent FD and parent artery.

Conclusion

In this case series, FD implantation appears to be a safe and effective treatment modality for ruptured BBAs. Additional coiling for aneurysms larger than 2.5 mm may be considered, as these appear to result in faster lesion occlusion, with no significantly increased complications. We have also illustrated the significant challenges in managing patients requiring invasive intracranial procedures post-commencement of dual-antiplatelet therapy, highlighting the need for relevant guidelines and future research.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported in part by a grant (KTIA_NAP_13-1-2013-0001) from the National Brain Research Program in Hungary.