Secondary Logo

Journal Logo

Original Contribution

Successful Treatment of Six Cases of Indirect Carotid-Cavernous Fistula with Ethylene Vinyl Alcohol Copolymer (Onyx) Transvenous Embolization

Bhatia, Kartik D MBBS; Wang, Lily MBBS; Parkinson, Richard J MBBS, FRACS; Wenderoth, Jason D BSc, MBBS, FRANZCR

Author Information
Journal of Neuro-Ophthalmology: March 2009 - Volume 29 - Issue 1 - p 3-8
doi: 10.1097/WNO.0b013e318199c85c
  • Free


Carotid-cavernous fistulas (CCFs) are uncommon but clinically significant vascular anomalies that may be associated with serious neurologic or ophthalmic morbidity. The most commonly used classification system of CCFs is that proposed in 1985 by Barrow et al (1), consisting of four different grades (A-D) based on angiographic identification of the arterial feeding sources of the fistula (1,2). All 6 cases of the indirect CCFs in this series were Barrow type D fistulas, defined as having arterial feeding sources from both the internal carotid artery (ICA) and external carotid artery (ECA) (1-3).

Endovascular treatments have become the mainstay in the management of indirect CCFs over the last 20 years (2-7). The occlusion of the complex vessel network via endovascular approaches using standard balloon and coil techniques has been difficult, however, because the honeycomb morphology of the cavernous sinus is not amenable to these large and stiff devices, and the result sometimes is incomplete closure of the fistula with resultant worsening in morbidity associated with shunting into orbital or cortical venous systems (2,4,6,8-10). In addition, coil and balloon techniques are associated with a high rate of mechanical cranial nerve injury (4,5).

Liquid embolic systems such as pure ethanol and n-butyl cyanoacrylate (nBCA) are increasingly being used as adjuncts or alternatives to balloon and coil systems owing to the liquid's ability to penetrate and occlude vessels of small caliber or with difficult access (11-13). However, the liquid poses a small but significant risk of venous and arterial thromboembolism via spontaneous droplet movement (11,13).

The ethylene vinyl alcohol copolymer (EVOH, Onyx) liquid embolization system (MicroTherapeutics Incorporated, Irvine, CA) uses a polymer that precipitates into an artificial embolus upon contact with blood or bodily fluids (12,14,15). This system possesses all of the advantages of other liquid embolic systems but allows greater control of polymer distribution, thus potentially reducing the risk of spontaneous droplet movement. It can fill cavities in a manner similar to that of foam-based embolization products (12,14,15). Unlike nBCA, Onyx also precipitates in a coherent fashion rather than having a tendency to fragment. It can be delivered over much more prolonged periods because of its nonadhesive nature (14,15,16). The relative simplicity of CCF occlusion with Onyx compared with standard coil occlusion techniques has the potential to reduce angiographic screening times and thereby patient radiation dose.

We present 6 type D CCFs in 5 patients treated only with the Onyx system via a transvenous approach.


Five consecutive patients with indirect type D CCFs were treated from December 2005 until May 2007 at the Prince of Wales Hospital, Sydney, Australia, using the Onyx system via a transvenous approach. Three of the patients were women, aged 44, 54, and 64 years at the time of treatment, and two were men, aged 63 and 64 years at the time of treatment. All endovascular procedures were undertaken by experienced neurovascular interventionalists (JW and RP).

Onyx was chosen as the sole embolization agent for these patients because of its theoretical advantages in reducing screening times, its ease and flexibility of use, and the theoretical reduction in risk of incomplete closure, cranial nerve injury, adhesive complications, and thromboembolism. In none of the 6 patients was there a need to use alternative embolic products.

The Onyx system consists of a copolymer that is dissolved in a dimethyl sulfoxide (DMSO) solvent (15). Within this mixture there is a suspended micronized tantalum powder that allows fluoroscopic visualization. When this liquid mixture is delivered via a catheter to the desired site, it makes contact with blood that causes the DMSO solvent to diffuse away rapidly, resulting in precipitation of the EVOH polymer into a putty-like mass with radiopaque properties (12,15). The nature of the precipitation process is such that the layer of Onyx in contact with blood forms a skin, the center of the precipitating mass remaining in its dissolved state (15). As pressure is applied to the delivery syringe, the liquid Onyx ruptures the overlying skin at its weakest points much as lava flows occur on the ocean floor. The specific formulation used in this series was Onyx-34, which has a nominal liquid viscosity of 34 centistokes (15). This formulation has relatively high viscosity so as to reduce the risk of retrograde penetration into the ICA.

In our patients, the liquid was delivered by an MTI Rebar-10 microcatheter in Case 1 and by an MTI Echelon-10 microcatheter in all other cases in accordance with the manufacturer's instructions (15).

All 5 patients had pretreatment imaging of the fistulas via time-resolved magnetic resonance angiography (TR-MRA) and digital subtraction angiography (DSA). All catheterization was completed under full anticoagulation using heparin to achieve an activated clotting time of 250-300 seconds. All procedures were performed under general anesthesia. DSA was used to obtain views of right and left internal carotid, external carotid, and vertebral arteries before treatment via catheterization of the femoral artery.

A catheter was maintained in the arterial system at all times to allow visualization of the arterial side of the fistula. The diagnostic angiograms were closely studied for supply of the fistulas by inferolateral or meningohypophyseal branches of the ICA. Such supply would be a relative contraindication to the use of Onyx or would necessitate the use of balloon protection in the ICA during Onyx injection.

The cavernous sinus was catheterized via the femoral vein using an inferior petrosal approach in 5 procedures. In 1 patient, in whom there were bilateral CCFs, a second treatment was required, necessitating access to the opposite cavernous sinus via the superior ophthalmic vein, which was cannulated under transorbital ultrasound guidance.

The initial targeted point of embolization in all cases was the venous point that filled first on contrast injection from the arterial side of the fistula, but this was not always necessary. Onyx-34 was progressively delivered into the cavernous sinus until complete obliteration of the sinus and the fistula was achieved.

The average angiographic screening time for the 5 procedures was 36.58 minutes. In 1 procedure, the screening time could not be obtained because of equipment failure. Confirmatory post-Onyx DSA films were obtained from the arterial supply to demonstrate definitive closure of the fistula. All patients were examined after the procedure to assess the efficacy of the treatment by imaging and clinical criteria. Follow-up included MRI and ophthalmologic evaluation at 3 months after treatment.


Table 1 shows clinical and imaging features of our 5 patients.

Clinical and imaging features of our 5 patients

Case 1

A 63-year-old man had a 2-week history of redness of the right eye. Four days before presentation, he had noted double vision on looking to the right and then began to have a progressive reduction in the vision of his right eye. He was found to have ptosis, exophthalmos, and chemosis of the right eye. There was a significant reduction in visual acuity in the right eye. Diplopia was present on right, upward, and downward gaze, and there was an obvious abduction deficit of the right eye. There were no other pertinent clinical abnormalities.

TR-MRA showed a right-sided Barrow type D CCF fed by right ICA and ECA branches (Fig. 1A-B). Given the rapid progression in symptoms and the risk of blindness to the right eye, he underwent endovascular management of the lesion with the Onyx system the next day.

FIG. 1:
Digital subtraction angiography of Case 1. Midarterial (A) and late arterial (B) phase lateral right internal carotid arteriograms show fistulous opacification of the right cavernous sinus (arrow) and superior ophthalmic vein (arrowhead). Lateral right internal carotid arteriogram (C) after the deployment of Onyx into the cavernous sinus via a transvenous approach shows total occlusion of the fistula.

Transvenous embolization was achieved using the Onyx-34 Liquid Embolization System with an MTI Rebar-10 catheter via an inferior petrosal approach from the right femoral vein. The cavernous sinus was progressively and completely obliterated with Onyx. Post-Onyx DSA views demonstrated definitive closure of the fistula (Fig. 1C). Total screening time was 37.6 minutes. There was an immediate dramatic reversal in the proptosis and chemosis of the right eye. Postoperatively the patient was noted to have a right sixth cranial nerve palsy, unchanged when compared with his preprocedural status. However, on follow-up examination at 3 months, the palsy had completely resolved. Successful closure of the fistula was confirmed on follow-up MRI examination.

Case 2

A 54-year-old woman presented with a 3-week history of blurred vision and painful movements of the left eye. For 10 days before presentation, she had had double vision upon looking to the left and a persistent whooshing sound in both ears. She was found on examination to have reduced visual acuity, conjunctival congestion, and mild proptosis of the left eye. There was also reduced abduction of the left eye. Noncontrast CT imaging revealed dilated superior ophthalmic veins bilaterally.

DSA demonstrated a complex left type D CCF with arterial supply bilaterally from the ICA and ECA (Fig. 2A-B). There was also a right type D CCF. The left cavernous sinus was accessed via an inferior petrosal approach from the right femoral vein and was successfully embolized using the Onyx-34 Liquid Embolization System with an MTI Echelon-10 microcatheter. Post-Onyx views revealed closure of the complex left CCF (Fig. 2C). Screening time of this procedure could not be obtained because of imaging equipment failure at the end of the procedure. The patient had made a full clinical recovery by the time of discharge. The right fistula (Fig. 2D) was closed 3 months later using the same Onyx-34 technology. Because of the known occlusion of the inferior petrosal sinus, the right cavernous sinus was accessed via the right superior ophthalmic vein (SOV) using transorbital ultrasound guidance for catheterization of this vessel (Fig. 2E). The right cavernous sinus was closed using Onyx-34. Post-Onyx DSA views demonstrated successful closure of the fistula (Fig. 2F). Screening time was 23.7 minutes. There was some moderate postoperative conjunctival edema of the right eye, which was probably a result of the saline flushing of the right SOV. This resolved within 36 hours. The patient made a full recovery before discharge. Successful closure was confirmed on follow-up ophthalmic and MRI examination 3 months later.

FIG. 2:
Digital subtraction angiography of Case 2. Lateral (A) and anteroposterior (B) late phase arteriograms show a complex left indirect carotid-cavernous fistula (CCF) (arrow) fed from the left internal and external carotid arteries via petrosal and left inferior maxillary arterial branches. C. After Onyx occlusion of the left CCF (arrow), a right CCF is exposed (arrowhead). D. Arteriogram shows filling of the right CCF via branches of the right external carotid artery. E. The right cavernous sinus is accessed via a catheter in the superior ophthalmic vein. F. Lateral arteriogram shows complete closure of the right CCF by Onyx cast (arrow).

Case 3

A 44-year-old woman presented with a 2-month history of severe headaches associated with nausea and vomiting in the morning. Pulsatile tinnitus, blurred vision of the left eye, and double vision on left gaze developed soon afterward. There was proptosis and conjunctival injection of the left eye, with diplopia on abduction. TR-MRA and DSA demonstrated a left CCF. The left cavernous sinus was accessed by advancing the Echelon-10 microcatheter through the right inferior petrosal sinus (IPS) and right cavernous and circular sinuses, and the fistula was successfully occluded using Onyx-34 (Fig. 3). Post-Onyx DSA views confirmed closure of the fistula. Screening time was 60.0 minutes. Ophthalmic and MRI review at 3 months after the procedure confirmed complete imaging closure and clinical recovery.

FIG. 3:
Digital subtraction angiography of Case 3. A. Lateral midphase right carotid arteriogram shows a complex left indirect CCF (arrow). B. An Echelon-10 microcatheter (arrow) deploys Onyx (arrowhead) into the left cavernous sinus via the right inferior petrosal sinus and right cavernous sinus. C. Lateral left carotid arteriogram shows complete closure of the complex left CCF (arrowhead).

Case 4

A 64-year-old woman presented with a 6-month history of double vision on looking to the right. The right eye had also developed progressive reddening over the previous 4 months. There was conjunctival injection of the right eye without associated proptosis and diplopia on right gaze. MRA demonstrated a right type D CCF. It was successfully closed using Onyx-34 via the right IPS with an Echelon-10 microcatheter. Screening time was 39.1 minutes. The patient demonstrated complete reversal of symptoms and was discharged home the next day. The success of the treatment was confirmed on MRI and ophthalmic assessment 3 months later.

Case 5

A 64-year-old man presented with double vision on looking to the right. Arterialized conjunctival vessels were present in the right eye. DSA confirmed the TR-MRA evidence of a right type D CCF. The fistula was successfully closed after embolization of the right cavernous sinus with Onyx-34 delivered via a right IPS approach with an Echelon-10 microcatheter. Screening time was 22.5 minutes. There was complete reversal in the patient's symptoms postoperatively. Follow-up at 3 months with MRI and ophthalmic examination confirmed complete closure of the fistula and no return of ophthalmic symptoms.

The only possible complication among these patients was a temporary sixth nerve palsy in Case 1, which had resolved completely at the time of follow-up 3 months later.


We have reported successful closure without enduring complications of six Barrow type D CCFs in 5 patients using the Onyx Liquid Embolization System via a transvenous approach.

The major advantage of liquid systems such as the one we used is their ability to conform to irregular and complex vascular structures (11,13,14). However, these agents have historically been difficult to control owing to their liquid nature, which can render them dangerous because of the risks of occlusion of important vascular structures (11,13). The use of pure ethanol as a liquid embolization agent for endovascular procedures was first described by Yakes et al (17) in a series of 17 patients with cerebral vascular malformations. That series demonstrated the efficacy of the agent to cause vessel sclerosis but at the cost of a high complication rate, with 8 of 17 patients having significant neurologic deficits (11,17). Such high complication rates are in part a result of ethanol's immediacy and reliability as a sclerosing agent, such that unwanted distribution of the agent resulted in cerebral infarction (11,17).

The next developmental stage involved the use of nBCA. This agent allows greater user-based control and thus has a reduced risk of cerebral infarction (12,13,18). Wakhloo et al (13) demonstrated this advantage in a series of 14 indirect CCFs in which angiographic and clinical cure was achieved with nBCA alone (n = 6) or in combination with coil embolization (n = 7) or polyvinyl alcohol (n = 1). Only one patient had spillage of an nBCA droplet into a cerebral vessel, a substantial improvement on the results of ethanol-based procedures (13). However, perforation of the inferior petrosal sinus during microcatheter placement was encountered in one patient as the result of catheter adhesion (12,13). The adhesion of microcatheters to the vessel wall with nBCA was also described by Debrun et al (19), who encountered 29 episodes of catheter adhesion in a series involving brain arteriovenous malformation embolizations.

The Onyx Liquid Embolization System has the ability to penetrate the complex network of vessels typical of indirect type D CCFs. In addition, it also has several properties that render it superior to other liquid-based systems such as pure ethanol and nBCA. Onyx is nonadhesive, maintains excellent penetrative ability for complex vascular structures, allows greater user-based control to reduce the risk of accidental embolization, and can be delivered via a single slow injection (11,12,13,16,20,21).

The use of Onyx to treat indirect CCFs was first described in a single indirect CCF by Arat et al (16) in 2002. This same team later demonstrated its successful use in treatment of a superior sagittal sinus dural arteriovenous fistula (dAVF) (14). Suzuki et al (12) demonstrated the safe and effective use of Onyx in combination with detachable coils for the treatment of 3 indirect CCFs. Baccin et al (20) also described the use of Onyx in association with coils for the treatment of a single direct CCF. Before this report, no series of patients had been published in which the Onyx system was used as a single treatment technique for indirect CCFs, despite increasing theoretical and clinical evidence demonstrating its advantages over other liquid embolic systems.

The major drawback associated with the use of Onyx is the risk of retrograde penetration into the feeding arterial vessels of the fistula. This risk is particularly important in indirect CCFs when there is significant supply of the fistula by meningeal branches of the ICA, such as the inferolateral or meningohypophyseal trunks. Retrograde arterial penetration has been reported previously by Nogueira et al (22) in 12 intracranial dural dAVFs treated with Onyx as a single treatment technique. They noted that such retrograde penetration into arterial feeding systems had the potential to result in cerebral infarction, although this did not occur. Van Rooij et al (23) reported major cerebral infarctions as a result of such retrograde penetration in 2 of 44 patients with brain arteriovenous malformations (AVMs). Difficulty in visualizing the tiny arterial feeding vessels as they enter the cavernous sinus in indirect CCFs may allow accidental penetration of Onyx into the ICA, with the potential for cerebral artery embolization (22,23). The risk of this complication can be reduced by pausing the delivery of Onyx for several minutes when such reflux is visualized on fluoroscopy. The use of biplane fluoroscopy during Onyx injection will aid such visualization (22,23). In this series, we used Onyx-34 exclusively because of its higher viscosity to reduce the risk of such retrograde penetration. In addition, the diagnostic angiograms were closely studied beforehand for significant supply to the fistulas by inferolateral or meningohypophyseal trunks.

In this series we also described the use of a superior ophthalmic vein approach to the fistula (Case 2). This approach (2,8,21) is often necessary because there is a significant incidence of inferior petrosal sinus thrombosis, which interferes with access by that route.

This case series is limited by its low patient numbers, retrospective approach, and lack of a comparative cohort. As such, it is not possible to comment definitively on the overall safety and efficacy of the Onyx system in treatment of indirect CCFs via the transvenous approach. However, the excellent outcomes for patients in this series, increased control of agent delivery, reduced screening times, and lack of adhesive complications all indicate that there is great potential for the use of Onyx in the management of this condition.


1. Barrow DL, Spector RH, Braun IF, et al. Classification and treatment of spontaneous carotid cavernous sinus fistulas. J Neurosurg 1985;62:248-56.
2. Shownkeen H, Bova D, Origitano TC, et al. Carotid-cavernous fistulas: pathogenesis and routes of approach to endovascular treatment. Skull Base 2001;11:207-18.
3. Cognard C, Gobin YP, Pierot L, et al. Cerebral dural arteriovenous fistulas: clinical and angiographic correlation with a revised classification of venous drainage. Radiology 1995;194:671-80.
4. Zhang YJ, Dion JE, Barrow DL, et al. Endovascular therapy for cavernous sinus vascular lesions. Tech Neurosurg 2003;8:238-50.
5. Albuquerque FC, Heinz GW, McDougall CG. Reversal of blindness after transvenous embolization of a carotid-cavernous fistula: case report. Neurosurgery 2003;52:233-7.
6. Clatterbuck RE, Spetzler RF. Surgical treatment of intracavernous vascular lesion: indications and results. Tech Neurosurg 2003;8:230-7.
7. Kai Y, Hamada J, Morioka M, et al. Treatment of cavernous sinus dural arteriovenous fistulae by external manual carotid compression. Neurosurgery 2007;60:253-7.
8. Kirsch M, Henkes H, Liebig T, et al. Endovascular management of dural carotid-cavernous fistulas in 141 patients. Neuroradiology 2006;48:486-90.
9. Cheng KM, Chan CM, Cheung YL. Transvenous embolisation of dural carotid-cavernous fistulas by multiple venous routes: a series of 27 cases. Acta Neurochir (Wien) 2003;145:17-29.
10. Hara T, Hamada J, Kai Y, et al. Surgical transvenous embolization of a carotid-cavernous dural fistula with cortical drainage via a petrosal vein: two technical case reports. Neurosurgery 2002;50:1380-4.
11. Koebbe CJ, Horowitz M, Jungreis C, et al. Alcohol embolization of carotid-cavernous indirect fistulae. Neurosurgery 2003;52:1111-6.
12. Suzuki S, Lee DW, Jahan R, et al. Transvenous treatment of spontaneous dural carotid-cavernous fistulas using a combination of detachable coils and Onyx. AJNR Am J Neuroradiol 2006;27:1346-9.
13. Wakhloo AK, Perlow A, Linfante I, et al. Transvenous n-butyl-cyanoacrylate infusion for complex dural carotid cavernous fistulas: technical considerations and clinical outcome. AJNR Am J Neuroradiol 2005;26:1888-97.
14. Arat A, Cekirge S, Saatci I, et al. Transvenous injection of Onyx for casting of the cavernous sinus for the treatment of a carotid-cavernous fistula. Neuroradiology 2004;46:1012-5.
15. Onyx Liquid Embolic System-Instructions for Use (product information). Irvine, CA: MicroTherapeutics Incorporated; 2003.
16. Arat A, Inci S. Treatment of a superior sagittal sinus dural arteriovenous fistula with Onyx: technical case report. Neurosurgery 2006;59(Suppl 1):ONSE169-70.
17. Yakes WF, Krauth L, Ecklund J, et al. Ethanol endovascular management of brain arteriovenous malformations: initial results. Neurosurgery 1997;40:1145-53.
18. Luo CB, Teng MM, Chang FC, et al. Transarterial balloon-assisted n-butyl-2-cyanoacrylate embolization of direct carotid cavernous fistulas. AJNR Am J Neuroradiol 2006;27:1535-40.
19. Debrun GM, Aletich VA, Shownkeen H, et al. Glued catheters during embolization of brain AVMs with acrylic glue. Interv Neuroradiol 1997;3:13-9.
20. Baccin CE, Campos CM, Abicalaf R, et al. Traumatic carotid-cavernous fistula: endovascular treatment with Onyx and coils. Interv Neuroradiol 2005;11:363-7.
21. Baldauf J, Spuler A, Hoch HH, et al. Embolization of indirect carotid-cavernous sinus fistulas using the superior ophthalmic vein approach. Acta Neurol Scand 2004;110:200-4.
22. Nogueira RD, Dabus G, Rabinov JD, et al. Preliminary experience with Onyx embolization for the treatment of intracranial dural arteriovenous fistulas. AJNR Am J Neuroradiol 2008;29:91-7.
23. van Rooij WJ, Sluzewski M, Beute GN. Brain AVM embolization with Onyx. AJNR Am J Neuroradiol 2007;28:172-7.
© 2009 Lippincott Williams & Wilkins, Inc.