Onyx versus nBCA and coils in the treatment of intracranial dural arteriovenous fistulas

Abstract

Background and purpose

Intracranial dural arteriovenous fistulas (DAVFs) with cortical venous drainage have significant morbidity and mortality. Complete closure of these lesions is necessary to reduce these risks. The purpose of our study was to compare the outcome of DAVFs treated with Onyx versus those treated with n-Butyl Cyanoacrylate (nBCA) and coil embolization in a case-control study. Compared with nBCA and coil embolization, we hypothesized that Onyx embolization for DAVF is safer and has a higher chance of complete obliteration, with no need for post-embolization surgery for the DAVF.

Materials and methods

From 1998 to 2015, 29 patients who had DAVFs were treated with endovascular embolization. Of these, 24 patients had imaging available for analysis. Successful closure rates, complications, and procedure time were compared between the embolization techniques.

Results

The chance of not requiring post-embolization surgery with Onyx (81.8%) was significantly higher (p = 0.005) than with nBCA (22.22%). The complication rate with Onyx (9.1%) tended to be lower compared with that of nBCA (22.22%; p = 0.37). Procedural time was not significantly different between Onyx (mean 267 minutes) and nBCA (mean 288 minutes) (p = 0.59). The odds ratio of a DAVF being treated with Onyx and then requiring no follow-up surgery was 17.5 (95% CI 1.97–155.4).

Conclusion

Our case-control study suggests that Onyx embolization is superior to nBCA and coil embolization in completely obliterating DAVFs, with higher odds of no post-embolization surgery. We also found that Onyx is safe for embolization of DAVFs, with no associated neurological mortality and morbidity.

Introduction

Intracranial dural arteriovenous fistulas (DAVFs) are abnormal connections between an artery and a venous sinus or leptomeningeal vein within the dura mater. True incidence of DAVFs is unknown,¹ but it is reported that they comprise 10–15% of all intracranial arteriovenous malformations.² The etiology of DAVFs is unclear, but proposed etiologies include trauma, venous thrombosis, transcranial surgery, otitis, sinusitis, tumor, and congenital causes.³

Classification of these lesions is according to the type of venous drainage. The classification schemes by Borden et al. and Cognard et al. are most commonly used.^4,5 Borden type I and Cognard types I and IIa DAVFs have a low risk of intracranial hemorrhage (ICH). Thus, treatment is only indicated if symptoms, such as tinnitus, are intolerable for the patient.⁶ In DAVFs classified as Borden types II and above or Cognard types IIb and above, there is cortical venous reflux (CVR). With CVR, annual mortality risk may be as high as 10.4%.⁶ If not treated, the annual risk for hemorrhage during follow-up for these lesions is 8.1%.⁶ The annual risk for non-hemorrhagic neurologic deficits during follow-up is 6.9%. In total, there is an annual event rate of 15% if not treated.⁶ Therefore, DAVFs with CVR require treatment with complete obliteration of the fistula. Additionally, Cognard types IV and V demonstrate venous ectasia. Sixty-five percent of these lesions present with ICH.⁵

Traditionally, DAVF embolization has been performed via transvenous coil embolization and/or trans-arterial injection of liquid embolic agents such as n-Butyl Cyanoacrylate (nBCA; Codman Neuro, Raynham, Bristol County, MA).⁷ The US Food and Drug Administration approved the use of Onyx, a non-adhesive liquid embolic agent (Medtronic, Irvine, CA), which is a mixture of ethylene vinyl alcohol, dimethyl sulfoxide (DMSO), and tantalum powder, for the embolization of intracranial AVMs in July 2005.⁸ Since then, embolization with Onyx has been a successful tool for obliterating intracranial DAVFs.^9–12

The purpose of our study was to compare the outcome of DAVFs treated with Onyx with those treated with nBCA and coil embolization in a case-control study. Our hypothesis was that compared with nBCA and coil embolization, Onyx embolization for DAVF is safer and has a higher chance of complete obliteration, lowering the need for post-embolization surgery for the DAVF.

Methods

The study was approved by our institutional research ethics board. A retrospective review of the institutional interventional neuroradiology database yielded 117 patients with endovascular embolization for intracranial vascular malformation from 1998 to July 2015. We included only patients with a diagnosis of DAVF who were treated with endovascular embolization and for whom we had imaging on our PACS. Patients with arteriovenous malformations or carotid cavernous fistulas were excluded. A total of 29 patients with DAVFs were treated with endovascular embolization during this time period. Of these, five patients did not having imaging available for evaluation and were excluded. Thus, 24/29 (83%) of the patients with DAVFs treated with endovascular embolization were included. The patients with whom Onyx was used were considered cases and patients treated with nBCA or coils were considered controls. Both cases and controls were classified according to location and Cognard and Borden types based on their diagnostic computed tomography, magnetic resonance, and digital subtraction angiography.

Patient selection and embolization

The choice of embolic agent was dependent on the availability of the agent and preference of the operator. The frequency of these DAVF embolizations has increased over time at our center. All nine patients before September 2010 were treated with glue or coil. Twelve of 15 patients after September 2010 were treated with Onyx. All Onyx embolizations were performed by a single operator (JJSS). The Onyx cases were defined according to the intention-to-treat principle: any time Onyx was used, it was classified as an Onyx case. Of the 24 DAVFs, Onyx was used in 12 cases. Nine patients were treated with only nBCA, and three were treated with only coils. All patients underwent general anesthesia for their embolization. Patients were heparinized throughout the procedure. The trans-arterial approach was used for Onyx and nBCA embolizations, whereas the transvenous approach was used for the coil embolizations.

For Onyx embolizations, access to the fistula was performed using DMSO-compatible micro-catheters and different micro-guidewires. The most commonly used micro-catheter was Marathon (Covidien, Irvine, CA) with the assistance of Mirage micro-guidewire (Covidien). The tip of the micro-catheter was advanced as close to the fistula point as possible. Initial patients (n = 4) were treated with Onyx 18 only. More recently, Onyx embolizations (n = 8) were started using Onyx 34 to make a proximal plug. The rest of the embolization was performed using Onyx 18 with standard plug and push technique. The total volume of Onyx ranged from 1–4 mL. Most patients were treated with a single sitting of embolization. Two Onyx patients needed a second sitting of embolization to achieve complete obliteration or a residual Borden type 1.

For nBCA embolization, the micro-catheter and concentration of nBCA used was based on the preference of the operator. For coil embolization, micro-catheter and coil selection was based on the preference of the operator and availability of these at the time of treatment.

The cases and controls were compared for demographic characteristics, presenting symptoms, obliteration rates, procedural complications, procedure time, and need for post-embolization surgery.

Statistical analysis

Frequency measures were used for descriptive statistics. For comparison of different parameters between cases and controls, odds ratio, Student’s t-test, and chi-square test were used, and 95% confidence intervals were calculated. A p-value of 0.05 was considered significant.

Results

Twenty-four patients (17 males; age 1.67–84 years, M = 56 ± 17.7 years) with DAVFs met our inclusion criteria. Table 1 shows the demographic characteristics, clinical presentation, angiographic characteristics, and treatment method for patients in our study in chronological order. Hemorrhage was the most frequent presenting symptom (7/24, 30%), followed by headache (5/24, 21%).

Table 1.

Demographics of patient population, location, and type of dural arterio-venous fistulas embolized.

Patient no.	Age (decade)	Presentation	Location	Borden	Cognard
1	80s	Visual disturbance	Occipital	III	IV
2	70s	Seizure	Occipital	III	III
3	60s	Headache	Parietal	III	III
4	30s	Headache	Occipital	III	III
5	60s	Hemorrhage	Temporal	III	III
6	60s	Headache	Temporal	III	IV
7	60s	Headache	Temporal	III	IV
8	0s	Macrocrania	Superior sagittal sinus	II	II a + b
9	70s	Hemorrhage	Frontal	III	IV
10	50s	Hemorrhage	Petrosal sinus	III	IV
11	60s	Tinnitus	Petrosal sinus	I	IIa
12	60s	Tinnitus	Sigmoid sinus	II	II a + b
13	40s	Headache	Frontal	III	IV
14	50s	Visual disturbance	Superior sagittal sinus	III	IV
15	50s	Tinnitus	Transverse sinus	III	III
16	60s	Hemorrhage	Posterior fossa	III	IV
17	50s	Tinnitus	Temporal	III	IIb
18	40s	Hemorrhage	Parietal	III	IV
19	20s	Hemorrhage	Transverse sinus	III	IV
20	70s	Visual disturbance	Transverse sinus	II	II a + b
21	60s	Hemorrhage	Transverse sinus	III	IV
22	40s	Visual disturbance	Transverse sinus	III	IV
23	40s	Paresthesia	Posterior fossa	III	V
24	70s	Unknown	Straight sinus	III	IV

The outcomes of treatment are shown in Table 2. The chances of not requiring post-embolization surgery with Onyx (81.8%) were significantly higher (p = 0.005) compared with nBCA (22.22%) and compared with non-Onyx (nBCA and coils) treatment (p = 0.013). The complication rate with Onyx (9.1%) was lower, but not significantly different, compared with that of nBCA (22.22%; p = 0.37) and non-Onyx treatment (p = 0.27). Importantly, Onyx embolization was associated with no neurological complications. Procedural time was not significantly different between Onyx (mean 267 minutes) and nBCA (mean 288 minutes; p = 0.59). Compared with treatment with nBCA, the odds ratio of a DAVF being treated with only Onyx and requiring no follow-up surgery was 17.5 (95% CI 1.97–155.4). Compared with treatment with nBCA and coils, the odds ratio of a DAVF being treated with only Onyx and requiring no follow-up surgery was 8 (95% CI 1.16–55.2).

Table 2.

Treatment outcomes, complications, and procedure time.

	Onyx	nBCA	Coils
Number	12	9	3
Complete obliteration	8	2	1
Residual Borden type 1	2	0	1
Perioperative morbidity	1 (arm vein thrombosis)	2 (hemiplegia, SAH)	1 (hemiplegia)
Death	0	0	0
Catheter retention/rupture	1	0	0
Required post-embolization surgery	2	7	1
Procedure time (minutes)	267	288	244

Mean clinical follow-up was 10.9 weeks (range 4–23.4 weeks) for Onyx and 17.8 weeks (range 1–52 weeks) for non-Onyx patients. Mean imaging follow-up was 11 weeks (range 1–52 weeks) for Onyx, 7.7 weeks (range 1–52 weeks) for nBCA, and 13 weeks (range 1–26 weeks) for coils. Ten of the 12 Onyx patients received short- to mid-term follow-up, which demonstrated no recurrence of the lesion. The majority of these follow-ups were performed between seven weeks and six months after embolization. The two patients who did not receive short-term follow-up had their embolization shortly before this retrospective review took place.

Discussion

Typically, endovascular embolization of DAVFs is attempted as the definitive treatment before open surgical intervention. Our case-control study, reviewing treatment of DAVF over the last 17 years in our institution, showed that the chances of open surgical intervention following Onyx embolization was significantly lower compared with embolization with nBCA or coils. The complication rate and procedure time between the cases and controls were not significantly different. The patients treated with Onyx embolization had no neurological complications.

In the past, endovascular embolization of DAVFs has been described with transvenous coil embolization and/or transvenous or trans-arterial injections of a liquid embolic agent such as n-nBCA.⁷ In 2006, Rezende et al. reported the first case of a DAVF embolized with Onyx.¹³ Since then, multiple other reports and trials have demonstrated positive results for DAVF embolization with Onyx.^9–12 The trans-arterial approach for Onyx DAVF embolization is safe and effective.^10,11,14 The properties of Onyx allow for the manipulation of the flow of the embolic material so that the path forming the embolus can be redirected, making it possible to embolize multiple arterial supplies before embolization of the venous component of the DAVF.

Although many studies demonstrated that Onyx was effective in embolization of DAVFs, fewer studies have compared DAVF occlusion rates with Onyx directly to those of nBCA. Notably, in 2013, Rabinov et al. published a retrospective study that demonstrated higher angiographic occlusion rates of DAVFs with Onyx versus nBCA.¹⁵ Our case-control study also found Onyx superior to non-Onyx treatment for DAVFs in terms of having a higher chance of complete obliteration and not requiring post-embolization surgery. Onyx embolization has become the first line of treatment for DAVFs with CVR in our institution. Both patients (patients 9 and 14) who had to undergo surgical intervention after Onyx embolization were treated early in our experience. At the time of Onyx embolization, patient 14 had residual DAVF with CVR after two prior embolizations and two prior open surgical interventions. The DAVF was finally clipped following our partial embolization with Onyx. The main access to the fistula was already lost due to the previous embolization and surgery. With experience, we became more aggressive with Onyx embolization. This approach is reflected in higher chances of complete obliteration with Onyx embolization without any neurological mortality and morbidity.

The non-adhesive nature of Onyx allows discontinuous and long injections facilitating embolization of multiple arterial and venous feeders. The discontinuous injection allows time to reassess the best method to achieve total fistulous obliteration that minimizes complications. The non-adhesive nature of Onyx also allows for venous filling to stop without concern of non-intended migration of Onyx into veins and other potential arterial collaterals. In 2010, Guedin et al. discouraged the use of Onyx partly due to the longer injection times compared with nBCA.¹⁶ Although Onyx does have a longer injection time, it may allow for the aforementioned benefits. Our study found no significant difference in the mean procedural duration for Onyx compared with nBCA or coil embolization for DAVFs.

One of the concerns regarding Onyx has been the stability of the embolization.^{5,7,11,17–19} In our study, 10/12 Onyx patients with short- to mid-term follow-up had no recurrence of the lesion, a result supported in another recent retrospective review by Rangel-Castilla et al.²⁰ The two patients who did not receive short-term follow-up had their embolization shortly before this retrospective review took place. In our experience, a solid Onyx cast is always occlusive and stable. For embolization of DAVFs, we always aim to get a solid Onyx cast in the draining vein of the fistula. The exact anatomical localization of the arterio-venous transition point using 3D spin angiogram was also key to successful embolization.

Limitations

A small sample size is an important limitation of our study. We reported all patients who presented with this rare disease and were treated in our institution. Of note, many of the DAVF series treated with Onyx in the literature have had similar sample sizes. We acknowledge that historical controls have a bias, as the technology, expertise, and our understanding of the disease change over time. These changes can contribute to the difference in outcomes between the two groups. We believe that all patients with DAVF treated with nBCA in our series could potentially be treated with Onyx. Given a superior outcome with Onyx embolization and the rarity of DAVFs, a randomized control trial between Onyx and nBCA embolization for DAVFs is not practical.

Conclusions

Our case-control study demonstrated Onyx embolization as superior to nBCA and coil embolization in completely obliterating DAVFs, with higher odds of no need for post-embolization surgery. We also found Onyx to be safe for embolization of DAVFs, with no associated neurological mortality and morbidity.

Declaration of conflicting interest

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 research, authorship, and/or publication of this article.