Endovascular treatment of dural arteriovenous fistulas with sinus drainage: Do we really need to protect the sinus?

Abstract

Background and purpose

Dural arteriovenous fistulas (DAVFs) with direct antegrade sinus drainage have a benign natural history but bruit can be disabling. Disconnection of the draining sinus is considered curative. We present the treatment results of 14 patients with a dural arteriovenous fistula with antegrade sinus return with emphasis on functionality of the involved sinus and the need for sinus patency.

Materials and methods

Between January 2009 and January 2018, 14 patients with a DAVF with direct antegrade sinus drainage were treated in our institution. There were two men and 12 women (1: 6) with a mean age of 61 years (median 61, range 43–78). Clinical presentation was unbearable pulsatile bruit in all cases. Thirteen were draining in the sigmoid or transverse sinus and one drained into the inferior petrosal sinus.

Results

Twelve of 14 (86%) patients had a draining sinus non-functional for the brain. In ten of these 12 patients the sinus was occluded with liquid embolic or coils. In the two patients with a functional sinus, the fistula was successfully occluded without occluding the sinus. One patient with inferior petrosal sinus drainage required transvenous embolization. There were no permanent adverse events of treatment.

Conclusions

Patients with DAVFs with direct antegrade sinus drainage were cured using a strategy of endovascular transarterial and transvenous embolization. Most patients had a non-functional sinus that could be occluded for cure.

Introduction

Dural arteriovenous fistulas (DAVFs) are uncommon vascular disorders and account for 10–15% of all intracranial vascular shunts.^1,2 DAVFs are acquired through multiple hole fistulas connecting dural arteries with venous or sinusal drainage of the brain. The clinical course of DAVFs is variable depending on the pattern of venous drainage.³ Symptoms range from asymptomatic to an aggressive clinical presentation consisting of haemorrhage, venous hypertension and neurological deficit.

DAVFs consist of multiple arterial dural feeders draining into a single dural sinus or cortical vein. To eliminate the fistula the operator has to disconnect the draining vein from the supplying dural arteries. Incomplete treatment allows recruitment of collateral vessels and persistent fistulous connections. DAVFs with direct sinus return are a special subgroup. These fistulas have a benign natural history and never present with haemorrhage. However, these fistulas can present with debilitating tinnitus and therefore may require treatment. They can be classified according to Cognard et al. as type I and IIa DAVFs.³

Treatment of DAVFs has traditionally been by surgical disconnection of the draining vein, but in recent decades endovascular transarterial or transvenous embolization using coils or liquid embolics has become first-line treatment. Careful consideration of transarterial, transvenous and combined approaches should be given in each case before proceeding with embolization.

Recently, transvenous balloon-assisted sinus protection during transarterial embolization has been reported to be a useful adjunct to the endovascular treatment of DAVFs.⁴ Transvenous balloon protection of the recipient sinus has been associated with preservation of sinus patency, facilitation of occlusion of abnormal arteriovenous connections within the sinus wall and separate venous channels, and increased penetration of embolic material by retrograde reflux into other dural feeders of the fistula network.⁵

In this study we evaluate the clinical presentation, vascular anatomy and treatment considerations of a single-centre cohort of 14 DAVFs with direct antegrade sinus drainage. In particular, functionality of the draining sinus for normal brain venous return is evaluated with respect to therapeutic preservation or occlusion of the involved sinus.

Materials and methods

Patient cohort

Between January 2009 and January 2018, 14 patients with a DAVF with direct sinus drainage were treated in our institution. There were two men and 12 women (1: 6) with a mean age of 61 years (median 61, range 43–78). Clinical presentation was devastating pulsatile bruit in all cases. Thirteen DAVFs drained in the sigmoid- or transverse sinus. One DAVF drained in the inferior petrosal sinus with arterial feeders from the petrosal branch of the middle meningeal artery and the stylomastoid branch from the posterior auricular artery. Of the patients with sigmoid- or transverse sinus drainage, the main feeders were the middle meningeal artery and the occipital artery in all cases. Treatment strategy was decided in a multidisciplinary meeting with neuroradiologists, neurosurgeons and neurologists. Endovascular transarterial approach without sinus balloon protection was the primary treatment in all cases but one. One patient with inferior petrosal sinus drainage required transvenous embolization (see Figure 3).

Figure 3.

(a) Lateral view of an external carotid artery injection showing a DAVF involving the middle meningeal artery, draining into the inferior petrosal sinus. (b) Lateral view of a distal middle meningeal artery injection showing the DAVF and venous drainage via the inferior petrosal sinus in the jugular vein and emissary veins. (c) Lateral view demonstrating a second transvenous catheter with coils in the venous receptacle. (d) Lateral view demonstrating trans venous coiling of the inferior petrosal sinus. (e, f) Lateral view of a subtracted (e) and non-subtract (f) common carotid injection demonstrating obliteration of the fistula.

Sinus evaluation

A 5F microcatheter was positioned in the internal carotid artery. Multiple angiographic runs of all cerebral vessels were performed to evaluate the functionality of the ipsilateral fistulous sinus in draining normal brain. Whenever the draining sinus opacified in the early arterial fistulous phase and not during the late venous phase of an angiographic run, it was considered not functional for draining normal brain. The critical consideration is the patency of the superficial draining vein of Labbé into the sigmoid sinus. In situations where the vein of Labbé is patent in the venous phase with normal antegrade drainage into the jugular vein, the sinus is considered functional, even though there is no contrast filling of the transverse sinus (see Figure 1).

Figure 1.

(a–c) AP view of a right internal carotid injection demonstrating normal arterial (a), parenchymal (b) and venous phase (c). (d–f) AP view of a left internal carotid artery injection demonstrating a parenchymal and venous phase of a functional sigmoid sinus on the left. Normal drainage via the vein of Labbé and the sigmoid sinus. (g, i) AP view of a left internal carotid artery injection demonstrating a non-functional sigmoid sinus on the left. Drainage is via the superficial middle cerebral vein to the cavernous sinus.

Embolization procedure

Endovascular treatment was performed with the patient under general anaesthesia on a biplane angiographic unit (Integris Allura Neuro, Philips Healthcare, Best, The Netherlands). A 5F microcatheter was positioned in the internal carotid artery, and multiple angiographic runs of all cerebral vessels were performed. Next, a 5F catheter was positioned in the external carotid or vertebral artery, and an Onyx-compatible microcatheter (Rebar 105-5081-153, Medtronic, Minneapolis, MN or Scepter micro balloon, MicroVention, Aliso Viejo, CA) was coaxially advanced in a dural feeding artery as close as possible to the fistula site. Whenever possible, the middle meningeal artery was used for access to the fistula, since this dural artery usually has a straight course allowing for easy distal access. Once the micro catheter was in place, Onyx 18 (Medtronic) or PHIL (MicroVention) was slowly injected. When reflux occurred, the injection was paused for a few minutes. In most cases reflux occurred several times before penetration of Onyx through the fistula site into the sinus wall and retrograde into arterial feeders. With distal middle meningeal artery injection, more than 5 cm reflux was accepted. Once Onyx advanced, it was intermittently injected until all feeders were filled retrograde, thereby completely occluding the fistula. In case of a non-functional sinus, the injection was continued when liquid embolic entered the draining sinus. In recent years, the Scepter micro catheter was used. This catheter has a micro balloon just proximal to the tip. When inflated, the balloon prevents reflux in the feeding artery.

Immediate angiographic cure was defined as no residual arteriovenous shunting. Post-procedural pain was managed with medication according to local protocol, in some cases including intravenous morphine or fentanyl. Follow-up angiography was performed after 6–12 weeks. Illustrative cases are provided in Figures 1–5.

Figure 2.

(a) AP view of an left internal carotid artery injection demonstrating a non-functioning sigmoid sinus on the left. (b) Lateral view of an external carotid artery injection demonstrating a DAVF draining to the sigmoid sinus on the left. (c) Lateral view demonstrating distal access with a Scepter catheter (Microvention, Tustin, CA) in a feeding branch of the middle meningeal artery. (d, e) Progressive injection of liquid embolic (PHIL, Microvention, Tustin, CA) entering and occluding the sigmoid sinus, including retrogradely occluding arterial feeders. (f) Lateral view of a common carotid artery injection showing obliteration of the fistula.

Figure 4.

(a) AP view of a left internal carotid artery injection demonstrating a non-functioning sigmoid sinus on the left. (b) Lateral view of a common carotid artery injection demonstrating a DAVF draining to the sigmoid sinus on the left. (c) Lateral view of a middle meningeal artery injection demonstrating direct sinus drainage. (d) Lateral view demonstrating distal access with a Scepter catheter (Microvention, Tustin, CA) in a feeding branch of the middle meningeal artery. (d, e) Progressive injection of liquid embolic (PHIL, Microvention, Tustin, CA) entering and occluding the sigmoid sinus, including retrogradely occluding arterial feeders. (f) Lateral view of a common carotid artery injection showing obliteration of the fistula.

Figure 5.

(a) Lateral view of a left internal carotid artery injection demonstrating a Type 1 DAVF draining via the left sigmoid sinus. (b) AP view in the venous phase demonstrating a functioning transverse and sigmod sinus. The contrast opacification is diluted on the left, because of the continuous inflow of the DAVF. (c, d) Lateral view demonstrating distal access with a Scepter catheter (Microvention, Tustin, CA) in a feeding branch of the middle meningeal artery. Continuous injection of liquid embolic (PHIL, Microvention, Tustin, CA) and final cast. (e) Lateral view of a common carotid artery injection demonstrating obliteration of the fistula. (f) Lateral view in venous phase demonstrating both sigmoid sinus remaining open.

Results

All patients were treated endovascular only; 12 via arterial approach, one using a combined arterial and venous approach and one patient transvenous only.

In 12 of 14 patients (86%) the fistula was obliterated in one treatment session. Two patients had partial treatment with relief of symptoms. Both were retreated 1 year later because of aggravated symptoms. After the second treatment the fistula was obliterated in both. Twelve of 14 (86%) patients had a draining sinus non-functional for the brain. In ten of these 12 patients the sinus was occluded with liquid embolic or coils. In the two patients with a functional sinus, the fistula was successfully occluded without occluding the sinus. One patient with inferior petrosal sinus drainage required transvenous embolization.

There were no permanent adverse events. Five patients experienced post-procedural pain after embolization, probably because of dural ischemia. All patients were managed according to local analgaesic protocol. At the time of follow-up angiography, all patients were pain-free without medication. Post-procedural radiation induced alopecia did not occur.

Discussion

In our small patient cohort with selected patients with DAVFs with sinus drainage treated because of unbearable bruit, the majority had a non-functional drainining sinus for normal brain. The sinus in these patients was only used to drain the fistula, sometimes even with additional drainage to the opposite transverse and sigmoid sinus interfering with normal brain drainage. In the patients with a non-functional sinus, this sinus could be safely occluded with coils or liquid embolics to cure the fistula, with durable results. In the two patients with functional sinus, the fistula could be completely occluded with liquid embolics through the arterial route without balloon protection of the involved sinus.

In our series of treated type 1 DAVFs there was a striking preponderance for females (6: 1). This contradicts the results in our previous series of DAVFs with direct cortical venous drainage where there was a striking male preponderance.⁷

Although sinus protection treatments are increasingly performed, until now a structured evaluation of benefit is lacking. Sinus protection strategies might be valuable when the draining sinus is still used for drainage of normal brain parenchyma, especially when a vein of Labbé is present and draining into the involved sinus. In cases with non-functional sinus, preservation of the sinus might not lead to functional recovery. After all, the brain has found different venous drainage routes and there is no direct need to recruite the preserved sinus again for venous drainage. One might even argue that conflicting flow patterns may arise, in turn increasing the risk for sinus thrombosis. In addition, when injecting liquid embolics through arterial feeders while occluding the sinus with a long balloon, excessive reflux and penetration of the embolic agent might cause occlusion of small arteries. When this occurs, there is an increased risk of occluding normal small arteries suppling nerves and skin with possible neurological deficit and skin necrosis.

Cerebral dural fistulas with direct antegrade sinus drainage are rare acquired lesions. Given the benign natural history of these types of DAVFs, the primary goal of treatment is relief of symptoms instead of completely occluding the fistula. At the same time, it is essential to limit the risk for the patient. For example, in our cohort one patient was primarily treated transvenous to obliterate the draining inferior petrosal sinus, because the feeding arteries were the petrosal branch of the middle meningeal artery and the stylomastoid branch of the posterior auricular artery. These two branches form part of the of the facial arcade supplying the facial nerve.⁶ To exclude the risk of damaging the facial nerve a transvenous approach was performed.

In our cohort, as well as in the literature, the most favourable artery for transarterial access is the middle meningeal artery, even if this artery is not the dominant supply. This has several reasons. The middle meningeal artery usually has a straight course because of its location between the internal tabula of the skull and the periosteal layer of the dura. This straight course allows for easy distal access in the proximity of the draining vein. Another advantage of using the middle meningeal artery is the absence of dangerous extracranial to intracranial anastomoses in its distal part. These potential dangerous orbital, cavernous and petrosal anastomoses are located proximal and anterior at the skull base.⁶

Conclusion

In conclusion, endovascular treatment and management of patients with DAVFs with antegrade sinus drainage led to resolution of symptoms in all patients. The vast majority of patients could be treated with a liquid embolic via the arterial route in a single session. Most patients had a non-functional sinus that could be safely occluded to cure the fistula.

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 received no financial support for the researcg, authorship, and/or publication of this article.