Treatment of carotid cavernous sinus fistulas with flow diverters. A case report and systematic review

Abstract

Introduction

Carotid cavernous fistulas (CCFs) are rare, usually follow head trauma or aneurysmal rupture. Recent treatment options include endovascular techniques such as flow diversion devices (FDDs).

Objective

To present our case treated with FDD application with transarterial cavernous-sinus coiling and present a systematic review on the use and effectiveness of FDDs in CCF treatment.

Materials and methods

We present our case of CCF treatment with FDD. A search was also conducted in PubMed, EMBASE and Cochrane until November 2020. Reference lists were also cross-checked.

Results

Including our case, thirty-eight patients were identified with a CCF that was treated with FDDs in sixteen studies. Twenty-two patients were females, nine were males and the rest unidentified. The mean age was 52,6 years (range 17–86, SD± 19.28). Thirty-six patients suffered from direct and two from indirect CCFs. Single FDD was used in four cases, single FDD with embolic materials in eleven cases, multiple overlapping FDDs were used in six cases and multiple overlapping FDDs with embolic materials were used in seventeen cases. Thirty-five patients (92,1%) had clinical improvement, immediate angiographic occlusion was seen in 44,7% of the cases, while long-term occlusion rate was 100% but with variable follow-up periods. One patient (2,6%) presented with a neurological deficit related to FDD deployment.

Conclusion

Targeted treatment of CCFs with single or overlapping FDDs with or without adjunct embolic agents offers a high success rate, both clinically and long-term angiographically compared to other endovascular methods alone. However, further research with multi-center prospective trials is warranted.

Introduction

Carotid cavernous fistulas (CCFs) are abnormal vessel connections between the internal carotid artery (ICA) or external carotid artery and the cavernous sinus (CS). ¹ According to the Barrow classification, CCFs can be grouped into four types, A to D. Type A represents high-flow, direct CCFs and while B, C and D are subtypes of indirect, low-flow ones. ² CCFs can follow head trauma, ICA aneurysm rupture, brain surgery near the CS or connective tissue disease. Blood congestion due to irregular high pressures in the CS and the surrounding venous plexuses presents with a characteristic symptomatology including chemosis, pulsatile exophthalmos, ocular bruit, cranial nerve palsies, and progressive, potentially irreversible visual loss. ¹

Therapeutic measures aim to achieve CCF occlusion, preserving patency of adjacent arterial branches and integrity of the nerves crossing the CS by preventing future complications. ³ Nowadays, endovascular techniques are mostly used: balloon- or stent-assisted, transarterial or transvenous coiling, liquid embolic agents or even ICA occlusion.^3–5 Flow diverter devices (FDDs) are flexible, self-expandable stent-like devices that have a dense, yet porous wall. ⁶ Apart from FDDs’ main use in therapy of particular internal carotid artery aneurysms, recent reports describe their off-label use for CCF treatment. ⁷ , ⁸

In this paper, we present a case of a 62-year-old woman with a type A CCF treated with a combination of FDD and transarterial coil application. Moreover, we systematically review the relevant literature on the safety and efficacy of FDDs in the treatment of CCFs.

Case presentation

A 62-year-old female presented in our department with six-days symptoms of diplopia and right visual difficulty. During examination, there was right eye congestion with exophthalmos, diplopia and significant visual loss (visual acuity 20/100). The rest of the neurological examination was unremarkable. Notably, the patient had a history of mild head injury following a bicycle fall with an uneventful recovery a few months earlier. No medication or drug use was reported. Imaging studies included a brain Magnetic Resonance Imaging (MRI), Magnetic Resonance Angiography (MRA) which revealed a right, direct high-flow CCF and a Digital Subtraction Angiography (DSA) which confirmed it (Figure 1(a) to (f)). ²

Figure 1.

Presentation of our patient's angiographic images, before and after coiling and eventually after FDD application, showing total occlusion of the CCF. (a) to (c) Lateral angiographic views of the right-sided CCF before intervention (a), after coiling application via the transarterial route into the cavernous sinuses with residual flow into the contralateral cavernous sinus (b) and after the FDD application with immediate total angiographic occlusion of the CCF (c). As above with AP views demonstrating eventual total occlusion of the CCF (d) to (f).

Our team decided to proceed with the endovascular treatment of the CCF using a single transfemoral arterial approach. Preoperatively, the patient received aspirin 100 mg and clopidogrel 75 mg for 10 days. Intraoperatively, titanium coils were initially placed into the CS via the transarterial route at the anterior portion of the CCF (Figure 1(b)). Angiography demonstrated reduced but residual blood flow into the CS, indicating partial occlusion of the CCF. Based on our previous experience on the use of FDDs on ICA aneurysms, we decided to deploy a Surpass Streamline FDD into the cavernous segment of the right ICA. Subsequently, immediate angiographic elimination of the CCF was demonstrated (Figure 1(f)).

Postoperatively, the patient’s symptomatology gradually resolved with decreased congestion of the right eye and improving visual acuity. Five days later, she was discharged from our unit with a good neurological outcome. She received dual antiplatelet therapy of aspirin 100 mg and clopidogrel 75 mg daily for 3 months and then aspirin 100 mg alone for three more months. At the 6-month follow-up, DSA confirmed total obliteration of the CCF with an excellent neurological outcome.

Materials and methods

The Centre for Review and Dissemination and Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines were followed (Figure 2). ⁹ The search strategy was applied in PubMed, EMBASE and Cochrane Library databases from their introduction up until November 2020. The search strategy included Medical Subject Heading (MeSH) terms and word variants for CCFs and FDDs (Table 1). Articles in the reference lists of included studies were hand searched separately in case any relevant papers were missed.

Figure 2.

The PRISMA flow diagram of this study.

Table 1.

Search strategy of the systematic review on the use of different types of flow diverter devices on the treatment of carotid-cavernous fistulas.

Group A	and	Group B
Carotid cavernous sinus fistul* Or		Flow diverter or
Carotid-cavernous sinus fistul*Or		Flow diversionor
CCF		FDD
		or Pipeline embolization device or Pipeline flex flow diverter or
		Surpass flow diverterorFRED flow diverter

fistul*: includes fistoula, fistulae, fistulous.

Due to the rarity of this condition and relative recent arrival of FDDs in clinical practice, only case reports and case series were found. All English articles involving high and low flow CCFs in adult patients were included, irrespective of CCFs’ etiology. Studies were included whether FDDs were initially planned for the CCF occlusion or when CCF appeared after FDD application for an ICA aneurysm and additional FDDs were deployed. Articles where CCFs were occluded using another technique and then FDDs were deployed for various reasons were excluded. Studies with insufficient data and pediatric patients were excluded. ¹⁰ , ¹¹

Guidelines suggest a two-step process in study selection for inclusion in a systematic review. This process involved two researchers [TS and EA] analyzing each title and abstract independently and then meeting to discuss any differences until consensus was achieved. A data extraction form was used to record relevant information regarding study design, sample size, demographics, CCFs’ etiology and type, pre-operative management, therapeutic intervention, outcome measures, results and follow-up findings. All post-operative outcomes were recorded, but only adverse events related to FDD use were analyzed.

All studies included were case reports and case series. The methodological quality of each study was evaluated using an evaluation tool designed for systematic reviews of case series and case reports. ¹² Total validity scores are unreliable and not recommended in a systematic review; thus, they were not calculated. Each component of the quality assessment was analyzed in supplementary material and was conducted by both researchers independently, who met to discuss any discrepancies. One study was excluded from the review as it included limited preoperative and postoperative information, which are analyzed in this study. ¹¹

Results

A total of 839 records were identified. After duplicates were removed, 84 publications were eventually eligible for full-text inspection. Thorough hand searching of the references in each study was also conducted. Our case excluded, thirty-seven patients were identified in sixteen studies that were eligible for this review (Figure 2). All studies classified their cases according to Barrow classification. Only Wendl et al. classified CCFs according to Rooij classification.^1,2,13 Information regarding demographics, type of CCF, therapeutic intervention, postoperative clinical and angiographic outcome, home antiplatelet therapy and follow up were extracted for this analysis.

Our case is not included in the literature results, however we incorporated it into the analysis to reach a total of 38 patients (Table 2 – Case #15). In total, twenty-two patients were females (57,8%), nine were males (23,6%) and seven (18,4%) were of unknown gender. Median age was 58 years and mean age was 52,6 years (range 17 – 86, SD± 19.28); age was not provided in seven patients. Direct Barrow type A CCFs were thirty-six (94,7%), while only two of type B. ²

Table 2.

Reported cases of direct cavernous sinus carotid fistulas treated with flow diverters.

Study	Case	Age, gender	Type, etiology	Treatment	Angiograpgic outcome	Clinical outcome	Post-op therapy	Maximum follow-upFinal outcome
Single FDD
Wendl et al. 13	#1	72, male	Low-flow type A, iatrogenic	1 FDD	Complete occlusion	Improvement of symptoms	DAT: aspirin 100 mg and clopidogrel 75 mg	18 months, complete occlusion, resolution of symptoms
	#2	58, female	Low-flow type A, iatrogenic	1 FDD	Complete occlusion	Improvement of symptoms		25 months, complete occlusion, resolution of symptoms
Castano et al. 22	#3	65, female	Low-flow type B, CCF fed by the meningohypophyseal trunk	1 FDD	Complete occlusion	Improvement of symptoms	N/A	1 month, complete occlusion, resolution of symptoms
	#4	77, female	Low-flow type B, small branches of the contralateral ICA	1 FDD	Complete occlusion	Improvement of symptoms	N/A	1 month, complete occlusion, resolution of symptoms
Single FDD with additional embolic agents
Mustafa et al. 20	#5	39, female	High-flow type A, ICA aneurysm rupture two weeks after FFD treatment	– 1 FDD – 5 coils transvenous	Incomplete occlusion	Improvement of symptoms	–DAT: aspirin and clopidogrel for two months -Aspirin	2 months, complete occlusion, resolution of symptoms
Lin et al. 21	#6	Middle-aged, unknown gender	High-flow type A, ICA aneurysm rupture six weeks after FFD treatment	1 FDD, 19x coils transvenous	Complete occlusion	Improvement of symptoms	–DAT: aspirin 325 mg and clopidogrel 75 mg for six months –325 mg aspirin after 6 months	6 months, complete occlusion, resolution of symptoms
	#7	Middle-aged, unknown gender	High-flow type A, ICA aneurysm rupture one month after FFD treatment	1 FDD, 4x coils transvenous	Complete occlusion	Improvement of symptoms	–DAT for 1 month -100 mg aspirin for 3 months	6 months, complete occlusion, resolution of symptoms
Chan et al. 14	#8	33, male	High-flow type A, post-traumatic (two months)	1 FDD, 8x coils transarterial	Complete occlusion	Clinically stable	–Post-op: 6000UI i.v. heparin twice daily,–DAT: for 2 weeks, –Aspirin 100 mg home therapy	4 months complete occlusion, neurologic deficits due to head trauma
Iancu et al. 27	#9	48, female	High-flow type A, iatrogenic	1 FDD, coils transarterial	Incomplete occlusion	Improvement of symptoms	clopidogrel for 3 months	12 months, complete occlusion
Ogilvy et al. 24	#10	35, male	High-flow type A, ICA traumatic aneurysm rupture	1 FDD, coils, Onyx	Complete occlusion	Improvement of symptoms	DAT: aspirin 325 mg and clopidogrel 75 mg for at least 3 months	6.5 months, complete occlusion, resolution of symptoms
Yoon et al. 15	#11	42, female	High-flow type A, trauma – rupture point of a persistent trigeminal artery	1 FDD, 1 coil transarterial	Incomplete occlusion	Improvement of symptoms	N/A	6 months, complete occlusion, resolution of symptoms
Baranoski et al. 18	#12	–, unknown gender	High-flow type A, spontaneous ICA aneurysm rupture	1 FDD, 8 coils transarterial	Incomplete occlusion	Improvement of symptoms	DAT for 6 months	55 months, complete occlusion, resolution of symptoms
	#13	–, unknown gender	High-flow type A, spontaneous	–1 FDD, transvenous coils, transvenous Onyx 18 and Onyx 34 with arterial balloon inflation -Transvenous Onyx 18 with arterial balloon inflation (POD 1)	–Incomplete occlusion –Incomplete occlusion	Improvement of symptoms	DAT for 6 months	2 months, complete occlusion, resolution of symptoms
	#14	–, unknown gender	High-flow type A, trauma	–1 FDD (PED) -Transvenous 21coils -(POD 7) –Transvenous nBCA and –3 coils (POD 10)	–Incomplete occlusion –Incomplete occlusion	Improvement of symptoms	DAT for 6 months	40 months, complete occlusion, resolution of symptoms
Stamatopoulos et al.	#15	62, female	High-flow type A, trauma	– 1 FDD and coils transarterial	– Complete occlusion	Improvement of symptoms	–DAT for 3 months: aspirin 100 mg and clopidogrel 75 mg. -Aspirin 100 mg for 3 months	6 months, complete occlusion, resolution of symptoms
Multiple FDDs
Hauck et al. 16	#16	64, female	High-flow type A, iatrogenic – acute ICA aneurysm rupture intraoperatively	6 FDD	Complete occlusion	Neurologic deficits due to intraoperative manipulations (mRS: 1)	DAT: aspirin 325 mg and clopidogrel 75 mg	N/A
Nadarajah et al. 23	#17	19, female	High-flow type A, trauma	4 FDD	Incomplete occlusion	Improvement of symptoms	-i.v. heparin infusion for 2d post-op, aPTT: 50–90 s.–DAT: aspirin 100 mg and clopidogrel 75 mg	1 month, complete occlusion, resolution of symptoms
Ogilvy et al. 24	#18	32, male	High-flow type A, trauma	2 FDD	Complete occlusion	Improvement of symptoms	– DAT: aspirin 325 mg and clopidogrel 75 mg for at least 3 months	2,5 months, complete occlusion, resolution of symptoms
Wendl et al. 13	#19	60, male	Low-flow type A, iatrogenic	2 FDD	Complete occlusion	Improvement of symptoms	DAT: aspirin 100 mg and clopidogrel 75 mg	17 months, complete occlusion, resolution of symptoms
	#20	50, female	Intermediate flow type A, trauma	–3 FDD -3 FDD (POD 20)	–Incomplete occlusion -Incomplete occlusion	Improvement of symptoms		18 months, complete occlusion, resolution of symptoms
Amuluru et al. 25	#21	65, female	High-flow type A, ICA aneurysm rupture during FDD application	– 5 FDD	Incomplete occlusion	Improvement of symptoms	DAT: aspirin and ticagrelor	12 months, complete occlusion, resolution of symptoms
Multiple FDDs with additional embolic agents
Nossek et al. 26	#22	59, female	High-flow type A, spontaneous rupture of ICA aneurysm	3 FDD, multiple coils into the aneurysmal fundus and the CS transarterial	Incomplete occlusion	Improvement of symptoms	–DAT: aspirin 325 mg and clopidogrel 75 mg for 6 months –81 mg aspirin after 6 months	48 months, complete occlusion, resolution of symptoms
Pradeep et al. 17	#23	28, male	High-flow type A, trauma	2 FDD, 3 coils and Onyx embolization transvenous	Complete occlusion	VI nerve palsy, improvement of vision	N/A	3 months, complete occlusion, resolution of symptoms
	#24	23, male	High-flow type A, trauma-headshot	–1 FDD, Onyx transvenous –1 FDD (after 2 months) –1 FDD (after 4 months)	–Incomplete occlusion -Incomplete occlusion –Incomplete occlusion	Improvement of symptoms	N/A	13 months, complete occlusion, resolution of symptoms
Ogilvy et al. 24	#25	78, female	High-flow type A, trauma	– 2x FFD and coils	Complete occlusion	Improvement of symptoms	DAT: aspirin 325 mg and clopidogrel 75 mg >3 months	4 months, complete occlusion, resolution of symptoms
Wendl et al. 13	#26	74, male	High-flow type A, Iatrogenic	–2x SG, 4 FDD	Incomplete occlusion	Improvement of symptoms	DAT: aspirin 100 mg and clopidogrel 75 mg	15 months, complete occlusion, ICA occlusion due to interruption of medication intake
	#27	76, female	intermediate flow type A, iatrogenic	–2 FDD –12 coils transvenous (POD 70)	–Incomplete occlusion -Complete occlusion	Improvement of symptoms		40 months, complete occlusion, resolution of symptoms
	#28	17, female	Intermediate flow type A, trauma	– 6 FDD, 24 coils transarterial -4 FDD (POD 5)	–Incomplete occlusion -Incomplete occlusion	Improvement of symptoms		8 months, complete occlusion, resolution of symptoms
	#29	66, female	Intermediate flow type A, trauma	–2 FDD, 65 coils transvenous -1 FDD, 9 coils transvenous (POD 63)	–Incomplete occlusion -Incomplete occlusion	Improvement of symptoms		32 months, complete occlusion, resolution of symptoms
	#30	44, female	Low-flow type A, Iatrogenic	–2 FDD, 1xSG	Incomplete occlusion	Slight disability due to cerebral ischemia – bilateral chronic ICA occlusion		31 months, complete occlusion, resolution of symptoms
	#31	74, female	High-flow type A, spontaneous	–65 coils transvenous –8 FDD (POD 2) -20 coils transvenous (POD 26)	–Incomplete occlusion –Incomplete occlusion –Incomplete occlusion	Improvement of symptoms		17 months, complete occlusion, resolution of symptoms
	#32	64, female	Low-flow type A, trauma	–1 FDD, 3 coils transarterial	Complete occlusion	Improvement of symptoms		22 months, complete occlusion, resolution of symptoms
	#33	54, female	High-flow type A, trauma	–6 FDD, 85 coils transarterial	Incomplete occlusion	Improvement of symptoms		0.6 months, complete occlusion, ICA occlusion due to interruption of medication intake, mild disability due to the head trauma
	#34	86, female	Intermediate flow type A, spontaneous cavernous ICA aneurysm rupture	–8 FDD, 35 coils, transvenous	Complete occlusion	Improvement of symptoms		5 months, complete occlusion, slight disability, resolution of symptoms
	#35	46, female	Low-flow type A, iatrogenic	–3 FDD -2 FDD (POD 6)	–Incomplete occlusion -Incomplete occlusion	Improvement of symptoms		13 months, complete occlusion, resolution of symptoms
Baranoski et al. 18	#36	–, unknown gender	High-flow type A, trauma	– 2 FDD, 19 coils	Complete occlusion	Improvement of symptoms	DAT for 6 months	Lost at follow-up
	#37	–, unknown gender	High-flow type A, trauma	–2 FDD, 14 coils transvenous -Onyx transvenous -(POD 3)	–Incomplete occlusion -Incomplete occlusion	Improvement of symptoms		8 months, complete occlusion, resolution of symptoms
Huseyinoglu et al. 19	#38	22, male	High-flow type A, trauma	–FDD x 5, coils 24 transarterial	Incomplete occlusion	Improvement of symptoms	N/A	3 months, complete occlusion, residual neurologic deficit due to severe head trauma

CCF: carotid cavernous fistula; ICA: internal carotid artery; ACT: activated clotting time; DAT: dual antiplatelet therapy; FDD: flow diverter devices; POD: post-operative day; i.v.: intravenous; nBCA: n-butyl-2-cyanoacrylate; SG: stent graft; intra-op: intraoperatively; post-op: postoperatively; UI: units; s: seconds; h: hour; aPTT: activated partial thromboplastin time; mRS: modified Rankin Score; N/A: not available.

Cases are classified based on the means which were used in the process of carotid cavernous sinus fistula occlusion.

Depending on the etiology of type A CCFs, we classified them into the following subgroups: associated with ICA aneurysm rupture, iatrogenic, related to trauma or spontaneous (Table 3). Among direct CCFs, nine (25%) cases were associated with ICA aneurysmal rupture: two cases presented acutely after FDD application, while three were associated with a delayed aneurysmal rupture after FDD application. Spontaneous ICA aneurysm rupture happened at three cases and one case was traumatic aneurysmal rupture. Seventeen (47,2%) cases were traumatic CCFs, whereas eight (22,2%) cases where iatrogenic (seven cases were caused during ICA angioplasty and one case presented after transsphenoidal surgery for pituitary adenoma). Two (5,5%) cases of direct CCFs were spontaneous.

Table 3.

Table showing the treatment used in each category of the carotid cavernous sinus fistulae in the literature.

Etiology – type of CCF	Single FDD	Single FDD with embolic agents	Multiple FDD	Multiple FDDs with embolic agents	Total
Type B CCF	2	–	–	–	2
Iatrogenic	2	1	1	4	8
ICA aneurysm rupture	–	5	2	2	9
Trauma	–	4	3	10	17
Spontaneous	–	1	–	1	2
Total	4	11	6	17	38

FDD: flow diverter devices; CCF: carotid cavernous fistula; ICA: internal carotid artery.

Regarding post-operative outcomes, seventeen (44,7%) patients had complete immediate CCF obliteration, while twenty-one (55,3%) patients had almost complete CCF closure with some residual flow. Among all cases, 10 (26,3%) needed reoperation. Thirty-four patients (89,4%) improved clinically postoperatively, and one presented with a neurological deficit due to intraoperative manipulations (complication rate 2,6%). Thirty-two (84,2%) patients received dual antiplatelet therapy (DAT) postoperatively. Follow-up periods varied widely, however, all CCFs were eventually reported as occluded.

Regarding intervention, there was some heterogeneity among the studies and we decided to further classify the patients into four different categories: single FDD, single FDD with adjunct embolic agents (coils, ethylene vinyl alcohol copolymer/Onyx, stent grafts), multiple FDDs and multiple FDDs with adjunct embolic agents (Table 2).

A single FDD was used in two cases of Barrow type A CCFs and two cases of low flow type B CCFs. All four CCFs were occluded intraoperatively (immediate occlusion rate 100%), that lasted up until the follow-up (long-term occlusion rate 100%). All patients improved clinically and DAT was used as home therapy.

A single FDD along with additional embolic agents was used in eleven cases, all type A CCFs. Five of them were occluded immediately, six had residual flow postoperatively and two of them needed re-operation. Eventually, all of them were occluded during the follow-up (immediate occlusion rate 45,4%, reoperation rate 18,1%, long-term occlusion rate 100%). Ten (90,9%) had clinical improvement postoperatively and resolution of symptoms at the last follow up. One patient (#8, Chan et al.) who suffered from severe head trauma was clinically stable post-operatively and had neurological deficits at the last follow-up. ¹⁴ Nine patients received DAT as home therapy, one (#9, Iancu et al.) received clopidogrel alone, while no information was available for another (#11, Yoon et al.). ¹⁵

Multiple FDDs were used in six patients with type A CCFs. Three were occluded intraoperatively, three had some residual flow and two of them, eventually needed reoperation (immediate occlusion rate 50%, reoperation rate 33,3%). Long-term follow up was missing for case #16 (Hauck et al.) which was reported with neurological deficits due to intraoperative manipulations. ¹⁶ From the rest, all five were reported to be closed during follow up and have long-term clinical improvement (100% long-term occlusion rate and clinical improvement). ¹⁶ DAT was used as home therapy for all patients.

Lastly, multiple FDDs with additional embolic agents were used in seventeen cases. Five had immediate CCF closure, twelve had residual flow and seven eventually needed reoperation (immediate closure rate 29,4%, reoperation rate 41,1%). Sixteen patients (94,1%) had clinical improvement, patient #23 (Pradeep at al.) remained with a VI nerve palsy which was related to pre-operative ICA aneurysm rupture and patient #34 remained with neurologic deficits which existed prior the procedure. ¹³ , ¹⁷ In all seventeen patients, but patient #36 (Baranoski et al.) who was lost in follow- up, CCFs were occluded (long-term occlusion rate 100%). ¹⁸ DAT was used as home therapy for fourteen patients; three patients had no information on the home antiplatelet therapy used. ¹⁷ , ¹⁹

Discussion

Although CCFs can cause serious consequences, limited data exist on the most optimal treatment method. ¹ , ³ Specific anatomical difficulties necessitating the occlusion of the fistula while protecting the vascular and nervous components crossing the CS, make the therapeutic intervention challenging. In this paper, we present the case of a 62-year-old woman with a post-traumatic CCF who was treated with the application of a single FDD following transarterial CS coiling. Additionally, we conducted a systematic review on the use of FDDs in the treatment of CCFs. The search of existing literature indicated that FDDs have been used for the treatment of type A as well as type B CCFs with excellent clinical and long-term angiographic results. ² , ⁷ Depending on the case, a single or multiple overlying FDDs were deployed with or without adjunct embolizing materials (titanium coils or liquid embolic agents). Postoperatively, most patients presented with clinical improvement of the CCF-related symptoms and in all cases, but two that had no relevant data available, long-term follow up revealed complete occlusion of the CCF.

FDDs have been approved and initially used for the treatment of unruptured intracranial aneurysms of the internal carotid artery. ⁷ However, based on the better understanding of their mechanism of action, properties and postoperative outcomes from published cases, FDDs are gradually being used for other neurovascular pathologies as well, such as CCFs. ⁷ Whether a single FDD or multiple FDDs with or without additional embolic agents are used depends on the type of CCF, the experience of the neuro-interventionalist and the result of immediate angiographic occlusion or residual flow.

The first reported case of CCF treatment with FDD use was a rather complicated case presented by Hauck et al. where CCF was actually caused intraoperatively by a misplaced FDD during treatment of an giant ICA aneurysm and five more PEDs subsequently being used to successfully obliterate the aneurysm and fistula. ¹⁶ Mustafa et al. also described a case where CCF-related symptoms presented in a patient two weeks after a single FDD deployment for an intracavernous ICA aneurysm. At this case, five coils were placed into the aneurysmal sack via a transvenous approach in order to occlude the fistulous point. The patient was discharged with asymptomatic partial occlusion of the CCF, while total obliteration was confirmed by MRA two months postoperatively. This is a situation where an FDD was evidently not used to treat the CCF, but its presence may have offered the environment for a controlled delayed rupture of the aneurysm and successive long-term complete obliteration and excellent neurological outcome following coiling. ²⁰ Yet another study by Lin et al. reported two patients with high-flow CCFs due to rupture of a cavernous ICA aneurysm, which was previously treated with one FDD each. The fistula was occluded with trans-cavernous coiling with intraoperative no contrast leaking into the CS and complete resolution of symptoms. ²¹

One single FDD as a treatment option seems to be suitable for specific low-flow type A CCFs due to low pressure gradient and type B CCFs.^1,2 Four cases have been reported in the literature. Castano et al. presented two cases of Barrow type B CCFs, which were treated with one FDD each without any additional embolization treatment. ² , ²² Additionally, Wendl et al. described two cases with direct, low-flow type A CCFs, which were treated with a single FDD. ² , ¹³ The low flow gradient of this type of CCFs seems to be sufficiently further reduced by a sole FDD. Authors believe that this reduction of the flow leads to initial CCF thrombosis and then endothelization of the FFD seems to provide a permanent closure of the CCF.

The sole use of multiple overlapping FDDs has also been described in six cases in the literture. ¹³ , ¹⁶ ,^23–25 Since the FDD offer some porosity and variable metal coverage, application of multiple FDDs could efficiently stop high-flow fistulas and, more importantly preserve the lumen of the ICA with parent vessel reconstruction. Owing to the fact that no placement of embolic agents into the CS is being used, there is no risk of direct injury of cranial nerves passing through the CS, as previously described in other studies. ⁵ Amuluru et al. reported a case where an ICA aneurysm treated with FDD ruptured intraoperatively, causing a high-flow CCF, but was completely occluded following delivery of a second FDD. ²⁵

FDDs were designed with specific properties, so that occlusion of side arterial branches nearby an ICA aneurysm, such as the ophthalmic artery, won’t happen. Theoretically, the pressure gradient in a high-flow CCF and the porosity of the FDD’s mesh make a single FDD inefficient as a sole treatment option. Indeed, most high-flow type A CCFs seem to need a more individualized approach. In 73,6% of the cases reported in the literature, transarterial and/or transvenous coiling, stent-grafts or Onyx were additionally used with FDD deployment. ¹³ , ¹⁴ ,^17–21, ²⁴ , ²⁶ , ²⁷ Pradeep et al. reported two patients with post-traumatic high-flow CCFs. In the first case, the fistula was occluded using transvenous coiling and Onyx injection along with three FDDs, whereas in the second case, transvenous Onyx with three FDDs were needed. In both cases, the CCF was not totally closed immediately, but it was found occluded at three and seven months follow-up, respectively. ¹⁷ Similarly, Ogilvy et al. presented two cases where coils and Onyx embolization of CS were combined with FDDs. ²⁴ Interestingly, no neurologuc deficits related to coiling or Onyx were reported.

Additionally, there are some FDDs’ properties that make them useful in challenging cases. FDDs can protect the ICA from coil herniation back into the lumen and the device itself can be used to facilitate coiling. ¹⁵ , ¹⁸ , ²⁶ FDDs’ reconstructive nature has been utilized for the treatment of more complicated cases where CCF treatment can be combined with reconstruction of a torn ICA. The neointima which is being grown seems to offer important mechanical coverage to the lacerated arterial segment. ⁸ Chan et al. described a case of a post-traumatic CCF which coexisted with a ruptured ICA pseudoaneurysm. ¹⁴ Both pseudoaneurysm and CCF were embolized using coils with subsequent application of a single FDD achieving complete occlusion and maintaining the patency of ICA despite the multiple lacerations. Oligvy et al. also described a case of post-traumatic CCF co-existing with an ICA pseudoaneurysm treated with FDDs. ²⁴ Even when the CCF is caused after FDD application for pseudoaneurysm treatment, additional CS coiling as described by Park et al. can be enough to stop the fistula. ¹¹ FDDs have also been reported in treatment of a postoperative CCF after transsphenoidal surgery for pituitary adenoma. ¹³ Although small leakage was noted from ICA to CS, patient’s symptoms immediately improved. At one month follow-up, CCF was completely occluded. ²⁷ As already mentioned, in cases where an FDD already exists, it can be used as a scaffold for further treatment either using additional FDDs or embolic agents. ¹⁶ , ²⁰ , ²¹ , ²⁸

Whether high-flow CCFs eventually need multiple overlapping FDDs with coils or Onyx cannot be answered by this study, as the relatively small number of patients drawn from case studies and case series offer evidence of low quality. This remains an individualized approach based on the characteristics of the fistula, the immediate angiographic result following each single intervention and the interventionalist’s experience. The ideal target remains the immediate complete occlusion or at least minimal residual flow through the CCF. Among all cases in this review, 89,4% of patients had clinical improvement and a long-term occlusion rate of 100%. Interestingly, despite the fact that the immediate occlusion rate ranged from 29,4 to 100% among the four combinations of treatment used for different types of CCFs, long-term complete occlusion was achieved in all cases. Residual flow postoperatively seems to be not only expected given the nature of FDDs attributes and mechanism of action, but also acceptable considering long-term outcomes. ¹³ , ¹⁵ ,^17–20, ²⁶ , ²⁷ This latency period is partly explained by the endothelialization which is achieved throughout. ²⁹ The authors of this paper believe that six months is the ideal period of time for the last follow-up.

One of the important aspects of post-operative guidance is the antiplatelet therapy used for FDD thrombosis prevention. ³⁰ Thirty-one patients received DAT as home therapy and all had an uneventful recovery and complete occlusion of the CCF. The authors agree with the opinion stated by Wendl et al., that an unreasoned stop of the antiplatelet therapy shortly after surgery can result in FDD thrombosis, as ICA occlusion has been reported at three patients who stopped antiplatelet medication intake. ¹³ Given the high percentage of CCFs which are related to head trauma, the administration of antiplatelet therapy pre- and postoperatively can be a deterrent factor in the choice of FDDs for many clinicians, however in this study, no hemorrhagic complications were observed but in one patient with delayed ICA aneurysm rupture (2,6%). ¹

After direct comparison with other techniques widely used in literature for CCF occlusion, FDDs in combination with other embolizing materials offer a lower immediate occlusion rate, but comparable if not better long-term occlusion rates and clinical improvement.^31–35 Arifin et al. presented a case series with 100% immediate occlusion rate of CS coiling, while embolization using detachable balloons was 40%. ⁴ Additionally, Prasad et al. reported a 100% occlusion rate at 6 month follow-up in a study where detachable balloons, detachable coils, both detachable balloons plus coils and both detachable plus push coils were used. ⁵ He et al. reported that in 105 cases of CCFs treated with detachable balloon or polyvinyl alcohol copolymer, occlusion rates ware 96,5% and 84,2% respectively, but with some major complications, including death. ³⁵ Moreover, Ertl et al. using a combination of endovascular techniques without FDD, reported a complication rate of 8% and long-term follow up occlusion rate of 96%. ³⁴

Although not initially constructed and researched for CCF closure, FDDs seem to be an efficient and safe choice and can be successfully combined with additional endovascular techniques. In total, immediate closure of CCF was 44,7% and long-term occlusion rate is 100%. Authors of this study believe that the reoperation rate of FDDs is acceptable, as they are do not imply life-long neurological deficits nor death for the patients and they have a very high long-term occlusion rate. FDDs can easily take the shape of ICA providing broad coverage at the fistulous point even in cases of trauma, pseudoaneurysm or ruptured ICA aneurysm. More importantly, their nature protects side-branches of ICA from occlusion and a dense structure to support intraoperative maneuvers, preventing coils from protruding into the ICA lumen. Also, specifically for low-flow type B CCFs, a single FDD alone may be enough for immediate and long-term obliteration of the fistula. Major drawbacks of FDDs are their high cost compared with other techniques and the relatively high reoperation rate.

Given the nature of this study, it faces some limitations. The relatively small number of patients described in the literature and the heterogeneity on the intervention, deter the extraction of safe results on the patient selection and the therapy of choice. We met a wide spectrum of therapeutic combinations described in the literature, ranging from a single FDD application to multiple overlapping FDDs with stent-grafts, embolic agents delivery into the CS, coiling and Onyx. Thus, no direct comparisons with other techniques such as sole coiling can be made without caution. ⁴ , ⁵ The surgeon’s experience on FDD use as well as patient unique characteristics predetermine the therapeutic choice.

Conclusion

This systematic review showed that depending on the case and using a tailored approach for each patient, FDDs with or without adjunct embolization should be considered in the treatment of CCFs with a favorable safety profile. The choice of a single FDDs seems to be reasonable only for low-flow type B CCFs. However, the evidence is scarce and the decision of broad FDD usage in CCFs has to be made upon thorough evaluation. Higher quality prospective studies are needed. Patient selection, the adjunct use of additional techniques along with FDDs and the optimal postoperative antiplatelet medication scheme have to be included in the planning of any future studies.

Highlights

• Carotid Cavernous Fistulas (CCF) are uncommon entities classically causing pulsatile exophthalmos, ophthalmoplegia and progressive visual loss

• Flow Diversion Devices (FDDs) have been approved for treatment of specific-type intracranial aneurysms

• Our case included, FDDs have been used as an alternative to treat CCFs with promising results.

• Depending on the patient, CCFs can be treated with a single or overlapping FDDs, combined with coiling and/or embolic agent application.