Article Text
PDF
Abstract
Background and purpose Transarterial Onyx embolization is an effective treatment for patients with intracranial dural arteriovenous fistula (DAVF). A study was performed to determine whether the clinical and radiological outcomes after transarterial Onyx treatment were affected by the type of cortical venous drainage (direct vs indirect) of high-grade DAVF.
Materials and methods Between May 2006 and December 2014, demographic data, clinical presentation, angiographic characteristics, and treatment-related outcomes were collected for 54 patients divided into two groups (intracranial DAVF with direct and indirect cortical venous drainage). Continuous variables were compared with the two-tailed t test and categorical variables with the χ2 test. Statistical significance was set at P<0.05.
Results Fifty-two patients (71% with direct and 29% with indirect cortical venous drainage) underwent Onyx embolization. Immediate complete occlusion after treatment was observed in about 55% of patients without between-group difference. During the long-term follow-up, complete angiographic occlusion was achieved in 83% of patients. Specifically, 15 additional patients (40%) in the direct cortical venous drainage group progressed to complete occlusion, but only one (6%) in the indirect cortical venous drainage group. Overall, the rate of complete occlusion was higher in patients with DAVF with direct cortical venous drainage (92%) than in those with DAVF with indirect cortical venous drainage (62.5%) (P=0.01). The rate of permanent treatment-related complications was 4%, mostly related to ischemic events. Overall, 80.5% of patients had a good neurological outcome (modified Rankin Scale score 0–2).
Conclusions Transarterial Onyx embolization of intracranial high-grade DAVF is safe and effective, particularly for lesions with direct cortical venous drainage.
- fistula
- liquid embolic material
- artery
- hemorrhage
Statistics from Altmetric.com
Introduction
Intracranial dural arteriovenous fistulas (DAVFs) account for 10–15% of all intracranial arteriovenous malformations. DAVFs are defined as abnormal vascular connections between dural arteries and dural venous sinuses or leptomeningeal veins, with the shunt usually located within the dural leaflets.1 The venous drainage pattern is the most important predictor of clinical symptoms, natural history, risk of hemorrhage, and possibility of treatment.2 3 DAVF management is based on the angioarchitecture and clinical symptoms, and includes conservative monitoring, endovascular embolization (transarterial or transvenous), surgical excision, and radiosurgery. Since the marketing of Onyx, most patients with intracranial DAVF can be satisfactorily treated through the transarterial route.4 5 Although the efficacy of Onyx is well established, few data are available on the influence of the fistula drainage pathway on the outcomes after such treatment. In this study we examined the clinical presentation and treatment-related outcomes after transarterial Onyx embolization of high-grade intracranial DAVFs with direct (DCVD) or indirect cortical venous drainage (ICVD).
Materials and methods
Patient population
This study analyzed data from consecutive patients with a diagnosis of intracranial DAVFs treated in our institution from May 2006 to December 2014. The study was approved by the local institutional review board. Data were extracted from a prospectively maintained database of consecutive patients with intracranial DAVF.
The medical and imaging records of 196 patients treated for intracranial DAVFs were reviewed and those with high-grade intracranial fistulas (Borden type II and III) and presence of DCVD or ICVD treated with transarterial Onyx embolization were included in the study (total n=54). Patients with carotido-cavernous fistula, DAVF without cortical venous drainage or pial arteriovenous fistula were excluded.
DCVD fistulas were defined as lesions with a direct connection between the feeding meningeal artery and a cortical vein (figure 1). ICVD fistulas represented a shunt between the meningeal arteries and dural sinus, with retrograde flow into the subarachnoid veins (figure 2).
Dural arteriovenous fistula (DAVF) with direct cortical venous drainage in patient 1 (a 57-year-old man presenting with hemiplegia secondary to a subarachnoid hemorrhage). (A) Right primitive carotid artery angiography showing the direct dural fistula with feeding vessels from the occipital and middle meningeal artery. The fistula drains into a superior hemispheric cerebellar vein (white arrow), which is a tributary of the lateral mesencephalic vein (white arrowheads) and the vein of Galen. (B) Final angiogram after two Onyx injections (black arrow showing the Onyx cast) with complete exclusion of the DAVF. (C) Digital subtraction angiogram follow-up at 3 months showing stable shunt occlusion.
Dural arteriovenous fistulas (DAVFs) with indirect cortical venous drainage. Upper panels: patient 2 (an 83-year-old woman with transient mild aphasia secondary to an indirect dural fistula of the left lateral sinus). (A) Left common carotid artery angiogram shows retrograde reflux inside an ‘isolated’ left lateral sinus (large black arrow) and into the contralateral sinus (black arrowhead) with cortical venous ectasia (thin black arrow). (B) Final control after Onyx injection and (C) digital subtraction angiography follow-up at 3 months showing absence of delayed occlusion. MR angiography follow-up at 2 years confirmed persistence of the DAVF. Lower panels: patient 3 (a 77-year-old woman presenting with cognitive disorder). (D) Initial angiogram showing a DAVF of the left ‘isolated’ lateral sinus (white arrow) and cortical venous reflux into the superior petrosal sinus and the vein of Trolard (white arrowheads). (E) Control after Onyx injection showing persistent shunt and cortical venous reflux. (F) Three-month control angiogram shows complete occlusion of the fistula.
Data collection
Based on the angiographic imaging results, patients were dichotomized into two groups: DAVFs with ICVD (ICVD group) and DAVFs with DCVD (DCVD group). DAVFs with ICVD were defined as lesions with the shunts between the meningeal artery and dural sinus, with retrograde flow into the subarachnoid veins. On the other hand, DAVFs with DCVD were defined as fistulas with direct drainage of the meningeal arteries into the subarachnoid veins or an ‘isolated’ thrombosed sinus segment. Demographic data, clinical presentation, DAVF features (location, main arterial feeders, venous drainage), treatment-related outcomes, and clinical and radiological follow-up data were collected. The Glasgow Coma Scale was used to assess the level of consciousness at diagnosis. The modified Rankin Scale (mRS) score was used for clinical outcome assessment. Clinical follow-up visits were at 3 months and 1 year after treatment. Radiological follow-up visits were scheduled at variable intervals on the basis of (1) the degree of angiographic occlusion after the first endovascular treatment, (2) age, (3) symptoms and neurological status. The first radiological follow-up examination was performed using catheter angiography whereas long-term radiological follow-up was performed with catheter angiography or magnetic resonance angiography (MRA), in function of the previous radiological results and clinical indications. All angiograms were reviewed by two senior neuroradiologists (DM and PHL) not involved in the treatment. Treatment outcome was classified as complete or incomplete occlusion, based on the remnant DAVF on angiographic images.
Treatment
All patients were treated by transarterial Onyx injection under general anesthesia in a biplane angiographic unit (Philips Allura, Best, The Netherlands). No additional surgical or radiosurgical treatment was performed. After transfemoral cannulation of a 6F or 8F catheter, the patient received 50 IU/kg heparin intravenously. A 6F or 7F guiding catheter was then navigated and superselective angiography was performed after catheterization of the target vessel with an intermediate 5F catheter and an Onyx-compatible microcatheter. The microcatheter was navigated into the DAVF feeding vessel as close as possible to the draining vein. Dimethyl sulfoxide was injected to fill the catheter dead space in order to prevent direct contact between Onyx and the bloodstream. DAVF embolization was then performed by slow Onyx injection under digital subtraction angiography (DSA) guidance. Onyx was injected through the middle meningeal artery or other branches of the external carotid artery according to the vascularization of the DAVFs.
Statistical analysis
For statistical analyses, patients were divided into two groups (DCVD vs ICVD). Statistical analyses were performed with the GraphPad QuickCalcs software. The two-tailed t test was used to compare continuous variables and the χ2 test for categorical variables. Statistical significance was set at P<0.05.
Results
Patient characteristics at diagnosis
For this study, 54 patients with high-grade (Borden II and III) intracranial DAVF (table 1) were subdivided into DCVD (n=38; 70%) and ICVD groups (n=16; 30%). The mean age and the percentage of men were slightly higher in the ICVD group than in the DCVD group (65.7 years vs 58.9 years, and 75% vs 58%, respectively). The most common clinical symptoms were intracranial hemorrhage (39%), tinnitus (18.5%), and headache (15%). Hemorrhagic presentation was more common in the DCVD than in the ICVD group (42% vs 31%). Overall, 15% of patients had impaired consciousness (score <15 on the Glasgow Coma Scale) at diagnosis. Patients with a Glasgow Coma Scale score <15 were more frequent in the ICVD group than in the DCVD group (25% vs 10.5%).
Patient characteristics and clinical presentation of high-grade intracranial dural arteriovenous fistulas
DAVF location
The most common DAVF locations were the lateral (31.5%), convexity/superior sagittal (31.5%), and tentorial sinuses (16%) (table 2). The convexity and superior sagittal sinuses were the most common locations in the DCVD group (42% of patients) and the lateral sinus in the ICVD group (81%). Multifocal arteriovenous shunts were observed in 55% of patients in the DCVD group and in 70% of patients in the ICVD group. Overall, pial supply to the DAVF was detected in three patients (5.5%), with a comparable proportion between groups (table 2).
Location of high-grade intracranial dural arteriovenous fistulas and proportion of lesions with pial supply
Treatment-related outcomes
Transarterial Onyx embolization could be performed in 52 patients (table 3). Due to the complex microcatheterization, endovascular treatment was not possible in two patients. The mean number of treatment sessions per patient was 1.5. Overall, complications during the first 48 hours after treatment occurred in 15% of patients. Two patients experienced permanent complications (3.8%): occipital hematoma during the first day after treatment in one patient from the DCVD group and ischemic lesion in the M2 territory in one patient from the ICVD group (see online supplementary table 1). The complication rate was slightly higher in the ICVD group than in the DCVD group (26% vs 11%), but the difference was not statistically significant (P=0.3). Hemorrhagic complications occurred in two patients (3.8%) and ischemic complications in six patients (11.5%), with no significant difference between groups. The hemorrhagic complications were secondary to venous infarction related to Onyx migration in cortical veins with secondary thrombosis (see online supplementary table 1). Sluggish flow within the venous system after embolization was managed by infusion heparin therapy for 1 week. However, in the case of hemorrhagic transformation, heparin was discontinued. With regard to the ischemic complications, in two cases they were related to arterial Onyx migration and concerned patients from the DCVD group (patients 2 and 3, online supplementary table 1). In one patient, dissection of the posterior inferior cerebellar artery occurred during catheterization, with secondary focal laterobulbar subarachnoid hemorrhage. The other ischemic events were related to lacunar strokes; most of them were transitory or asymptomatic and discovered during the early MRA follow-up. Neither complications related to unwanted Onyx migration into potentially dangerous anastomoses nor treatment-related mortality were reported.
Supplementary file 1
Radiological outcome, clinical outcome, and treatment-related complications after transarterial embolization of high-grade intracranial dural arteriovenous fistulas
Immediate complete occlusion after treatment was achieved in 55.7% of patients, with no difference between groups. During follow-up, DSA was performed in 42 patients (81%). Because of their comorbidities, three elderly patients (>80 years of age) were followed by MRA. Two other patients refused the control DSA and were monitored by MRA, but the immediate post-intervention MRA showed complete occlusion of the lesion. Five patients were lost during follow-up. Three of them were included in the group with incomplete occlusion because the fistula was incompletely occluded after treatment. During a mean follow-up of 34 months, the rate of complete occlusion increased to 83.3% and was significantly higher in the DCVD group (92%) than in the ICVD group (62.5%) (P=0.01). Complete cure was achieved after a single Onyx session in 62% of patients in the DCVD group and 53% of patients in the ICVD group (P=0.78). Overall, the rate of good neurological outcome (mRS score 0–2) was 80.5%.
Discussion
Our study of a large and selected cohort of patients with high-grade intracranial DAVF highlights a number of important findings on the clinical presentation and management of these lesions. In agreement with the more aggressive behavior of DAVFs with CVD, all patients were symptomatic and nearly 60% of them had aggressive symptoms (hemorrhage, seizure, focal neurological deficits) at the time of diagnosis. In our series, the most common fistula locations were the lateral sinus (80% of patients in the ICVD group) and convexity/superior sagittal sinus (42% of patients in the DCVD group). One of the major findings of our study is the high rate of long-term DAVF complete occlusion (83.3%) following transarterial Onyx injections. Moreover, this treatment led to better long-term angiographic results in the DCVD group than in the ICVD group (92% vs 62% complete occlusions, P=0.01). Finally, the incidence of permanent treatment-related complications was low in both groups, and 80% of patients reported a good neurological outcome during the follow-up. These findings suggest that DAVFs with cortical venous reflux can be treated by transarterial Onyx injection with a high success rate and limited treatment-related morbidity.
The venous drainage pattern of intracranial DAVFs is the most important predictor of clinical behavior, natural history, risk of hemorrhage, and possibility of treatment.2 3 Currently, DAVFs without CVD are considered ‘benign’ lesions, and treatment is recommended only in selected cases based on symptoms. Conversely, DAVFs with CVD have a high risk of aggressive symptoms related to intracerebral hemorrhage or venous hypertension, and urgent treatment is recommended in most patients. For lesions with cortical venous reflux, the annual rate of intracranial hemorrhage and non-hemorrhagic neurological deficits (NHND) is between 7% and 19%.6 In our series, the hemorrhage rate at diagnosis was close to 40%, and was higher in the DCVD group. In addition, about 30% of patients showed NHND related to venous hypertension. ‘Benign’ symptoms, such as tinnitus or headache, led to the diagnosis of DAVF in about 30% of our patients. According to historical studies, the rates of new bleeding and neurological events in untreated or partially treated patients with high-grade DAVF (Borden type II and III) are between 15% and 35%, with an annual mortality rate of 10%.7 8 In addition, the risk of new neurological events is higher in patients with DAVF, CVD and more aggressive presenting symptoms (hemorrhage or NHND) compared with patients with more ‘benign’ symptoms.9–11 Urgent treatment is recommended in most cases.
With the improvement of angiographic imaging, increased operators’ experience, and refinement of microcatheter and liquid embolic agent technologies, DAVF embolization has become the first-choice treatment and most patients with intracranial DAVF can be satisfactorily treated with an endovascular approach. For high-grade lesions, the treatment aim is the occlusion of the proximal segment of the venous drainage in order to eliminate the risk of new bleeding events or symptom worsening.12 With the introduction of Onyx, the transarterial endovascular approach has become an effective treatment with high curative rates. In recent series, the rate of DAVF occlusion after transarterial Onyx injection was between 60% and 100%.4 13–15 In our study the rate of immediate angiographic occlusion after treatment was close to 56%, without a significant difference between the DCVD and ICVD groups. Moreover, during a mean follow-up of 34 months, we observed the progressive thrombosis of the lesions with an angiographic complete occlusion rate close to 83%. However, it is important to point out that, during the follow-up, 15 patients (40%) in the DCVD group progressed to complete occlusion compared with only one (6%) in the ICVD group (table 3). Accordingly, the long-term angiographic complete occlusion rate was significantly higher in the DCVD group (92%) than in the ICVD group (62%). Very few studies have analyzed the role of venous drainage as a predictive factor of angiographic occlusion after DAVF embolization. In a series of 55 patients, Kim et al 16 reported favorable treatment outcomes among patients with non-sinus fistulas and DCVD (92% complete occlusion) compared with patients with sinus fistulas (63.3% complete occlusion). The better angiographic outcomes of DCVD fistulas after transarterial Onyx embolization are mainly explained by differences in the angioarchitecture of these lesion types. As DAVFs with ICVD usually have many small arteriovenous shunts, there is a higher probability that Onyx might not occlude all fistulating points. In addition, a larger volume of Onyx injection could be required, with the risk of Onyx reflux or migration in dangerous anastomoses. The progressive occlusion of DCVD fistulas can also be related to the higher penetration of Onyx into the fistulous sites and draining veins, with a higher rate of thrombosis progression during the follow-up. Conversely, in the case of ICVD lesions, the lower penetration of Onyx into the small shunts could lead to the development of new fistulous connections with lower rates of long-term occlusion.16–18
Our experience demonstrates that transarterial Onyx treatment in patients with high-grade DAVF is safe and effective. In the literature, the rate of treatment-related complications ranges from 8% to 15% in larger series of patients with intracranial DAVFs treated with Onyx.4 16 19 20 In our study, most of the reported complications were transient (11.5%) and were related to ischemic events. Interestingly, the incidence of complications after treatment was slightly higher in the ICVD group. This is likely due to the fact that more Onyx volume/injections are often required for the treatment of ICVD lesions, leading to a higher risk of migration of the embolic material into the parent artery or distal arteries.
Transvenous embolization of DAVFs with the occlusion of dural sinus is a possible strategy in selected cases. If the clinical presentation justifies the sinus sacrifice, this technique is safe and effective in patients with good collateral brain drainage. The main advantages of this approach are the easy access to the fistula, relatively safe approach, and the possibility of obliterating the fistula in one session. The angiographic obliteration rate ranges from 70% to 85% and the complication rate from 10% to 40%, with a frequency of permanent events between 4% and 7%.21 22 Based on our experience, the sacrifice of the venous sinus should be considered when Onyx embolization is not feasible or safe due to the inability to position the microcatheter close enough to the shunt and in the case of ICVD lesions with blinded sinus, torcular involvement or reflux in the superior petrosal sinus or posterior fossa veins. Such an approach could have been proposed to three (18.7%) of our 16 patients with ICVD.
Compared with other embolic materials such as cyanoacrylates, the main advantage of Onyx is the stability during injection that allows its progressive accumulation into the target point (the foot of the vein). In most of the recent published series, Onyx was more effective than n-butyl cyanoacrylate (n-BCA). Rabinov et al 23 reported 80% long-term occlusions after transarterial Onyx embolization of 31 high-grade DAVFs compared with 40% long-term occlusions after treatment with n-BCA of 21 high-grade lesions. Similarly, Choo et al 24 found better and more durable occlusion after Onyx injection compared with n-BCA, with higher odds of no post-embolization surgery in the first group. However, other studies highlighted the safety and effectiveness of n-BCA. Guedin et al 25 reported complete occlusion of 89.5% of Borden type II and III DAVFs treated with n-BCA with no permanent morbidity and mortality. On the other hand, the rate of progressive thrombosis of the residual shunt ranges between 12% and 16% after cyanoacrylate treatment.25 26 Kim et al 26 evaluated 121 patients with DAVFs treated with low concentration glue and distal catheterization, and found better angiographic results in patients with Borden type III fistulas (40% angiographic cure and 18% progressive thrombosis) than in those with Borden I and II lesions (26% angiographic cure and 15% progressive thrombosis). However, the use of n-BCA has several drawbacks: rapid rate of polymerization, higher risk of catheter retention, the injection must be fast and continuous, and the preparation requires experienced operators. Accordingly, in our experience, we consider transarterial Onyx injection as the treatment of choice for most DAVFs.
Study limitations
Our study has limitations intrinsic to single-center series and it is not a population-based study. In addition, although the data were prospectively collected, the analysis was retrospective. The radiological follow-up was largely based on the experience of the senior author, and there was a lack of standardization of the clinical and radiological patient follow-up. Finally, the DAVF classification and the imaging outcomes were assessed by the operators and not by independent researchers.
Overall, five patients had time-resolved MRA follow-up instead of DSA. Although DSA is still considered as the gold standard, MRA was previously used with good results.27
Conclusions
In our series, complete angiographic occlusion was observed in about 83% of patients after transarterial Onyx embolization, with a low incidence of permanent complications and a high rate of good neurological outcome during follow-up. Transarterial Onyx embolization seems to be more effective for the treatment of DAVFs with DCVD, with a complete occlusion rate close to 92%. Our study confirms that intracranial high-grade DAVFs, particularly those with DCVD, can be treated safely and effectively by transarterial Onyx injection.
References
Footnotes
Contributors DM, MLC, PHL, AB: Participated in the conception and design of the study. DM, PHL, MLC, FC, AB: Analyzed and interpreted the data. GG, OE, CR, CD, PM, VC: Collection, assembly, and possession of the raw data. CD, OE, VC, FC: Statistical expertise. DM, PHL, FC: Initial drafting of the article. PHL, FC, AB: Revision of article. DM, GG, OE, CR, CD, PM, AB: Revising the manuscript critically for important intellectual content.
Funding This research received no specific grant from any funding agency in the public, commercial or not- for-profit sectors.
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.