Clin Neuroradiol DOI 10.1007/s00062-014-0362-y
O r i g i n a l A rt i c l e
Endovascular Treatment of Intracranial Dural Arteriovenous Fistulas Presenting with Intracranial Hemorrhage in 46 Consecutive Patients: With Emphasis on Transarterial Embolization with Onyx C. Li · X. Yang · Y. Li · C. Jiang · Z. Wu
Received: 23 September 2014 / Accepted: 18 November 2014 © Springer-Verlag Berlin Heidelberg 2014
Abstract Purpose The purpose of this study was to evaluate the effectiveness and safety as well as the clinical and angiographic results of endovascular treatment (EVT) for patients with hemorrhagic dural arteriovenous fistulas (DAVFs). Methods From April 2009 to November 2013, 46 consecutive patients (7 women, 39 men; mean age, 46.7 years) diagnosed with hemorrhagic intracranial DAVFs at our department were enrolled in this study. Clinical and angiographic data were reviewed and evaluated. Results Two fistulas were cured by transvenous approach, and all other fistulas were embolized through transarterial route. After treatment with last embolization, a residual shunt was observed in 15 patients, including near-total obliteration of the fistula in 6 patients (13.0 %) and partial obliteration of the fistula in 9 patients (19.6 %). Seven of them underwent supplementary Gamma knife surgery. Complications occurred during or after EVT in six patients. One patient died 10 days after EVT because of complications of Onyx embolus into the posterior inferior cerebellar artery. Other complications included microcatheter entrapment in one case, trigeminocardiac reflex in one, facial nerve paresis in two, and occulomotor nerve paresis in one. The facial nerve paresis in one patient markedly improved but did not completely resolve during follow-up period. The other four patients recovered well, without symptoms at clinical follow-up. Angiographic follow-up result was obtained in
Z. Wu, MD () · C. Li, MS · X. Yang, MD · Y. Li, MD · C. Jiang, MD Beijing Neurosurgical Institute and Beijing Tiantan Hospital, Capital Medical University, No. 6, Tiantan Xili, Dongcheng District, 100050 Beijing, People’s Republic of China e-mail: [email protected]
35 cases, with complete obliteration in 27 of them (77.1 %). Clinical follow-up outcomes were good, with 38 patients (82.6 %) reporting modified Rankin Scale of either 0 or 1. No patient suffered recurrent intracranial hemorrhage during follow-up. Conclusions EVT was effective and safe in the modern management of ruptured intracranial DAVFs, with complete cure in most lesions. Clinical outcomes were good despite patients presenting with intracranial hemorrhage. Keywords Dural arteriovenous fistula · Intracranial hemorrhage · Endovascular treatment · Transarterial embolization · Onyx Introduction Intracranial dural arteriovenous fistulas (DAVFs) are pathologic shunts between dural arteries and dural venous sinuses, meningeal veins, or cortical veins, which account for 10–15 % of intracranial arteriovenous malformations [1]. Intracranial hemorrhage is one of the most serious clinical manifestations of intracranial DAVFs [2–4]. Intracranial DAVFs presenting with hemorrhage require early treatment aimed at complete and definitive fistula obliteration, considering the substantial risks of rebleeding after the first hemorrhage [5, 6]. Most previous studies regarding hemorrhagic intracranial DAVFs focused only on risk factors for intracranial hemorrhage and found that the existence of cortical venous reflux (CVR) increases the risk of intracranial hemorrhage for DAVFs [6–9]. Currently, large case series specifically evaluating the clinical and angiographic outcomes of patients diagnosed with ruptured DAVFs are rare [3]. During the past 2 decades, endovascular embolization has become a first-line treatment for DAVFs [4, 10–12].
13
2
The advent of Onyx (EV3, Irvine, CA), a new nonadhesive liquid embolic agent, has changed the treatment strategy for DAVF because of its advantages over other embolic agents [11, 12]. In this retrospective study, we present angiographic and clinical outcomes of 46 consecutive patients with hemorrhagic DAVFs, which were treated with endovascular treatment (EVT). The purpose of this study was to evaluate the effectiveness and safety of EVT for these lesions, with emphasis on transarterial embolization using Onyx. Materials and Methods Institutional review board approval was obtained for this retrospective study. Patients or their family members have given written informed consent to publish relevant materials. Patients and Data Collection We reviewed our patient database from April 2009 to November 2013 and identified 46 consecutive patients diagnosed with hemorrhagic DAVFs managed with EVT. Every patient underwent preoperative planar images examination, including head computed tomography (CT) scan or magnetic resonance (MR) imaging. Evidence of intracranial hemorrhage was confirmed with signs of intracranial blood on CT scan or MR imaging, or sometimes in the cerebrospinal fluid, by lumbar puncture. Diagnosis of DAVF was confirmed by a six-vessel digital subtraction angiography (DSA) including external carotid injection. The intracranial hemorrhage was attributed to the DAVF if there was no other explanation. Information collected and analyzed included demographics (age and sex), clinical presentation and extension of hemorrhage (intraparenchymal, subdural, subarachnoid, or intraventricular), location and angiographic features of the fistulas, EVT modality and endovascular materials selection, treatment complications, and angiographic and clinical follow-up outcomes. The diagnostic, immediately postoperative, and follow-up DSA findings were evaluated by at least two experienced neurointerventionalists to identify lesion locations, angiographic characteristics of the fistulas, and treatment results. Treatment Protocol and EVT Techniques In this series, every treatment decision-making was discussed in a multidisciplinary team consisting of neurologists, neurosurgeons, and neurointerventionalists. Every treatment was tailored to the individual patient on a case-bycase basis, considering various clinical and imaging parameters, including clinical presentation, patient age, the results
13
C. Li et al.
of imaging (location, arterial feeders, and venous drainage of the fistula), patient preference, and so on. All patients in this series received EVT as initial treatment to manage the fistulas. All procedures were performed with the patients under general anesthesia. Anticoagulation during the treatment process included heparin used in the flushing system of the guiding catheter (3000 IU/l of saline), without intravenous bolus of heparin given at the beginning of each procedure. Before the treatment procedure, cerebral angiogram was performed in all cases to confirm the diagnosis and angiographic characteristics, and to assist treatment planning. When transarterial approach was selected, vascular access was obtained via a transfemoral approach using a 6F guiding catheter. Sometimes when the lesion was fed by bilateral arteries or both the anterior circulation and posterior circulation, the contralateral transfemoral approach was also prepared to perform control angiograms to evaluate the degree of fistula obliteration during procedure. Using roadmapping technique, a microcatheter was coaxially navigated through the guiding catheter into a selected feeding artery. Progression of the microcatheter (Enchelon-10, EV3 or Marathon, EV3) was aided by the use of a microguidewire (Silverspeed-10/14, EV3; Mirage 008, EV3; Traxcess-14, Microvention) to reach to or as close as possible to the fistula. The choice of the microcatheter was at the discretion of the operator, and the most commonly used microcatheter in our center was Marathon microcatheter (EV3, Irvine, CA). When a safe embolization position was achieved, superselective angiography was performed to reveal the focal angioarchitecture and dynamic flow characteristics of the fistula. The vascular anatomy was carefully considered before embolization began for potential dangerous anastomoses and the cranial nerve supply. The decision regarding which endovascular material (Onyx/Nbutyl-2-cyanoacrylate, NBCA) and which concentration of the embolized material to use was subjective and was based on the distance from the fistula, flow velocity of the shunt, and risk of venous migration, often at the discretion of the operator. Sometimes blood flow of the shunt was so fast that we had to place appropriate coil to decrease the flow velocity at the fistula, and make the subsequently injected liquid embolic agent stable. The aim of glue embolization was to push the glue from the distal arterial feeder segment into the most proximal venous segment. As to Onyx injection, “plug and push” technique was often used to enable its penetration into the fistular compartments. Once Onyx advanced, it was intermittently injected until the proximal draining veins were completely filled, thereby completely obliterating the fistula. Some lesions needed multiple microcatheter embolization or multiple treatment sessions to obtain complete obliteration of the fistula or satisfactory flow reduction with suppression of CVR, until there was no appropriate route for
Endovascular Treatment of Intracranial Dural Arteriovenous Fistulas Presenting with Intracranial
further embolization. When the satisfactory embolization was unavailable using transarterial route or the supplying arteries were deemed as too tortuous to obtain satisfactory embolization position using transarterial approach, transvenous embolization was attempted with detachable coils and/or in combination with Onyx. The balloon-assisted flow control technique, which has been recently recommended as effective methods for high-flow DAVFs, was applied in selected cases, as described by Albuquerque et al. [10] and Shi et al. [12]. Angiographic follow-up was generally recommended at 3–9 months after treatment, preferably using DSA. CT or MR angiography was regarded as an alternative in cases when DSA was not available. If persistent CVR was noted after the embolization procedure, the patient was recommended for microsurgical disconnection or stereotactic radiosurgery (SRS) to achieve complete disconnection of the CVR (in this series, no patient agreed to accept adjunctive surgery). Clinical follow-up information was acquired by means of outpatient visits and/or phone contact and evaluated according to modified Rankin Scale (mRS). Results There were 39 men and 7 women, with a mean age of 46.7 years (range, 23–68 years). Tentorial fistulas (n = 21, 45.7 %) were the most common location of DAVFs followed by the anterior cranial fossa (n = 9, 19.6 %). All fistulas had direct cortical venous drainage or retrograde flow into cortical veins (Cognard type II b or greater). A total of 34 cases (73.9 %) were associated with venous varices (Cognard type IV). These patients experienced 51 intracranial hemorrhages in all; 41 patients experienced one intracranial hemorrhage each. Five patients suffered from rebleeding before treatment. Intraparenchymal hemorrhage, which occurred in 28 cases (60.9 %), was the most common bleeding pattern, and the site of intraparenchymal hemorrhage was most often near venous varix of the fistula. Other hemorrhage extensions included subarachnoid (n = 11, 23.9 %), intraparenchymal and intraventricular (n = 7, 15.2 %), subdural (n = 2, 4.3 %), and intraventricular (n = 1, 2.2 %). One patient presenting with intraparenchymal hemorrhage underwent surgical evacuation of the intraparenchymal hematoma. The fistula was not disconnected during operation in this patient. Four patients with intraventricular hemorrhage underwent extraventricular drainage. The demographics and clinical data of these patients are summarized in Table 1. All patients received EVT as initial treatment to manage their fistulas. A patient with right transverse sinus DAVF (Cognard type II b) underwent right transverse sinus packing initially using coils in association with Onyx, and the fistulas were completely obliterated. All fistulas in the other
3
Table 1 Demographics, location of fistula, extension of hemorrhage, clinical presentation, and Cognard classification Parameter Value Number of patients 46 Mean age (years) 46.7 Gender Male 39 (84.8) Female 7 (15.2) Location of fistula Tentorium 21 (45.7) Anterior cranial fossa 9 (19.6) Transverse-sigmoid sinus 6 (13.0) Cerebral convexity 6 (13.0) Foramen magnum 2 (4.3) Middle cranial fossa 2 (4.3) Intracranial hemorrhage extensions Intraparenchymal 28 (60.9) Subarachnoid 11 (23.9) Intraparenchymal and intraventricular 7 (15.2) Subdural 2 (4.3) Intraventricular 1 (2.2) Clinical presentation Before hemorrhage Asymptomatic 35 (76.1) Headache, bruit, and tinnitus 7 (15.2) Nonhemorrhagic neurological deficit 3 (6.5) Seizure 1 (2.2) After hemorrhage Headache, without neurological deficit 27 (58.7) Neurological deficit 14 (30.4) Coma 9 (19.6) Seizure 6 (13.0) Cognard classification Type V 2 (4.3) Type IV 34 (73.9) Type III 6 (13.0) Type IIa + b 3 (6.5) Type IIb 1 (2.2) Numbers in parentheses are percentages
patients were embolized with transarterial route as initial treatment. A patient with a fistula at the anterior cranial fossa was initially embolized through transarterial route. However, the supplying arteries were too tortuous to obtain a satisfactory embolization position, so transvenous coil embolization via the superior sagittal sinus was finally performed to completely obliterate the fistula. For the remaining 44 patients receiving transarterial embolization, a total of 57 embolization procedures were performed, with two or more procedures in 10 of them. Of these patients, 24 (24 of 44, 54.5 %) were cured with a single microcatheter embolization. Six of them required a second microcatheterization for complete occlusion. One patient required a third microcatheterization for an embolic cure. After treatment with last embolization, a residual shunt was observed in 15 patients (near-total obliteration, n = 6,
13
4
13.0 %; partial obliteration, n = 9, 19.6 %). Seven of them accepted Gamma knife surgery as supplementary treatment. No patient agreed to accept microsurgical intervention as adjunctive treatment in this series. Angiographic follow-up result was obtained in 35 cases (76.1 %), with a mean length of follow-up of 9.9 months. Of these patients, 27 (27 of 35, 77.1 %) had complete obliteration of their fistula. Angiographic follow-up results was available in four of the seven patients accepting supplementary Gamma knife surgery, and the results were as follows: two near-totally obliterated fistulas were transformed into totally obliterated; another near-totally obliterated fistula was still near-totally obliterated, but suppression of CVR was observed compared with immediate angiographic outcomes after EVT, and one partially obliterated fistula was transformed into near-totally obliterated. Complications occurred during or after EVT in six (13.0 %) cases. The microcatheter was stretched and broken during retrieval in one case, without any clinical symptoms. One patient encountered migration of the Onyx embolus into the posterior inferior cerebellar artery (PICA) and died of multiple organ failure10 days after EVT. During Onyx injection, bradycardia caused by trigeminocardiac reflex (TCR) occurred in one case, which was effectively extinguished after administering atropine. Other complications included facial nerve paresis in two cases and occulomotor nerve paresis in one. One patient presenting with facial nerve paresis still suffered from slight facial nerve paresis at 22-month clinical follow-up. The other two patients with cranial nerve paresis recovered well during the follow-up period, without symptom at clinical follow-up. Clinical follow-up outcome was obtained in 44 cases (one patient was lost). All the 44 patients recovered from their baseline condition during the follow-up period. Clinical outcomes were good in this series, with 38 patients (82.6 %) reporting mRS of either 0 or 1 (no symptoms or minimal residual symptoms). No patient suffered recurrent intracranial hemorrhage during the follow-up. The data of treatment, angiographic, and clinical outcomes are summarized in Table 2. Illustrative cases are presented in Figs. 1, 2, and 3. Discussion The clinical presentation of DAVFs varies among individuals, which could be “asymptomatic” type, “benign” type presenting with pulsatile tinnitus, proptosis, and other slight symptoms, or “aggressive” type presenting with intracranial hemorrhage or nonhemorrhagic neurologic deficits [4]. Despite the recognition of hemorrhage as one of the most serious sequelae of intracranial DAVFs and the treatment urgency of hemorrhagic DAVFs, to our knowledge, there is
13
C. Li et al. Table 2 Treatment, angiographic, and clinical outcomes Parameter Value Treatment TAE 37 (80.4) TAE and GKS 7 (15.2) TVE 2 (4.3) Endovascular material Onyx 37 (80.4) NBCA 2 (4.3) Onyx and NBCA 3 (6.5) Onyx and Coils 3 (6.5) Coils 1 (2.2) Endovascular complication Cranial nerve paresis 3 (6.5) Microcatheter entrapment 1 (2.2) Trigeminocardiac reflex 1 (2.2) Migration of the Onyx embolus 1 (2.2) Angiographic outcomes after EVT Complete obliteration 31 (67.4) Near-total obliteration 6 (13.0) Partial obliteration 9 (19.6) Angiographic follow-up 35 (76.1) Mean time (months) 9.9 Modalities DSA 26 (56.5) MRA or CTA 9 (19.6) Outcomes Complete obliteration 27 (58.7) Near-total obliteration 5 (10.9) Partial obliteration 3 (6.5) Clinical follow-up 44 (95.7) Mean time (months) 20.1 Outcomes (mRS) 0–1 38 (82.6) 2–3 4 (8.7) 4 2 (4.3) TAE transarterial embolization, GKS Gamma knife surgery, TVE transvenous embolization, NBCA N-butyl-2-cyanoacrylate, EVT endovascular treatment, DSA digital subtraction angiography, MRA magnetic resonance angiography, CTA computed tomography angiography, mRS modified Rankin Scale Numbers in parentheses are percentages
only one large series report specifically focusing on clinical characteristics and outcome of patients with hemorrhagic DAVFs [3]. In that article, Daniels et al. reported clinical outcomes of 28 patients with DAVFs who presented with intracranial hemorrhage, which were treated with EVT, surgery, SRS, or a combination of multiple modalities. In their case series, the majority of fistulas (75 %) were completely obliterated, and 71 % of patients experienced excellent clinical outcome (mRS of 0 or 1) [3]. In the present study, we analyzed the angiographic and clinical outcomes of 46 consecutive patients with hemorrhagic DAVFs to increase the experience for EVT of these hemorrhagic lesions. We also demonstrated a high rate of complete obliteration and safety
Endovascular Treatment of Intracranial Dural Arteriovenous Fistulas Presenting with Intracranial
5
Fig. 1 A quadragenarian suddenly presented with headache and speech arrest. Head computed tomographic scan shows intraparenchymal hemorrhage (a, arrow). Angiogram of left external carotid artery shows a fistula at the left middle cranial fossa feeding by the left middle meningeal artery (MMA), left accessory meningeal artery (AMA), and left sphenopalatine artery (b). The venous drainage was into the left vein of Labbe that had a venous aneurysm (b, arrow) and drained into the left transverse sinus. Three microcatheterizations with Onyx injection was performed through the left MMA and the left AMA during two treatment sessions to obtain an embolic cure. Cast of Onyx after the final embolization (c, arrow). Follow-up digital subtraction angiography 6 months later confirmed complete obliteration of the lesion (d). The patient made a very good clinical recovery (modified Rankin Scale = 0)
of endovascular embolization for hemorrhagic DAVFs. The study further demonstrates that EVT was effective in management of hemorrhagic DAVFs, with good clinical outcomes despite patients presenting with intracranial hemorrhage. Currently, there are mainly three alternatives for the treatment of DAVFs, including EVT, microsurgical disconnection of fistulas, and SRS [2, 4, 10–14]. Microsurgery is a traditionally recommended treatment approach [13]. However, due to the risks inherent in craniotomy and the efficacy of EVT, surgery is currently only indicated in cases where endovascular approaches have failed or are not feasible [4]. For SRS, the risk of hemorrhage still remains during the latency period ranging from several months to years, until vessel thrombosis and fistula closure happen, so it is inappropriate as the primary treatment in hemorrhagic DAVFs considering the high risk of rebleeding [14]. This technique should be reserved for carefully selected DAVFs for which endovascular and surgical options have been exhausted. Currently, endovascular embolization using transarterial or transvenous approach has become a first-line treatment for DAVFs [2, 4, 10–12]. With modern endovascular tech-
niques such as flow control techniques using balloon, better understanding of the pathophysiology and anatomy of intracranial DAVFs can be achieved, with EVTs resulting in high rates of complete obliteration with a low risk to the patient. Treatment of DAVFs is aimed at complete cure of the arteriovenous fistulas—incomplete treatment allows recruitment of collateral vessels and persistent risk of hemorrhage. NBCA has been used as an embolic agent with relatively good results, but it requires quick and continuous injection technique to avoid unintended catheter tip adherence to the vessel wall due to its high thrombogenicity. The advent of new nonadhesive liquid embolic agent, Onyx (EV3, Irvine, CA), has changed the treatment strategy for DAVFs because of its greater safety and efficacy compared with NBCA in endovascular embolization, especially when distal enough navigation of the microcatheter was difficult [11, 12, 15]. The increased control of Onyx made it possible to deal with DAVFs with multiple arterial supplies, allowing the Onyx to penetrate slowly and precisely into the afferent arteries well until the proximal draining veins are affected. In this series, a fistula (Fig. 2) “recurred” on follow-up angiogram although “complete obliteration” of the fistula was obtained
13
6
C. Li et al.
Fig. 2 A sexagenarian suddenly presented with alexia, followed by seizure and coma. Head computed tomographic scan shows intraparenchymal hemorrhage (a, arrow). Angiogram of the left common carotid artery shows a fistula at the left occipital lobe with feeders from posterior branches of the left middle meningeal artery (MMA) and meningeal branches of the left occipital artery (OA), drained into cortical venous vein with venous varix (b, arrow). Angiogram obtained after two microcatheterization with Onyx embolization through the routes of the left MMA and the left OA shows “complete obliteration” of the fistula (c). However, digital subtraction angiography (DSA) 6 months later revealed recurrence of this “cured” fistula (d, arrow). The patient recovered well, and no rebleeding occurred during 9 months of clinical follow-up (modified Rankin Scale = 1). This patient underwent Gamma knife surgery as supplementary treatment after the follow-up DSA examination and was waiting for DSA follow-up again
via transarterial embolization with Onyx. In fact, the fistula in this case was not really cured because the proximal venous segment was not affected, thus fistula recanalization may occur with filling from newly recruited arterial feeders. Adamczyk et al. [16] have reported that despite adequate Onyx penetration into the fistula and draining vein, DAVF recanalization may also appear. However, no such case was observed in this series. Although this kind of case was uncommon and the definite underlying mechanism for this recurrence remains unclear, the possibility for DAVF recanalization should highlight the need for performing followup angiography after EVT. In this series, tentorial fistulas (45.7 %) were the most common location of DAVFs followed by the anterior cranial fossa DAVF (19.6 %). DAVFs in these two locations differ from other subtypes because they usually have direct drainage into cortical veins, often with formation of venous ectasia (Cognard type IV) [17–19]. These characteristics make the lesions prone to intracranial hemorrhage. Many previous reports have described surgical treatment for DAVFs located at the anterior cranial fossa and the tentorium with promising result. Agid et al. [19] have reported open surgical dis-
13
connection was effective in obliterating the fistula in all 15 patients with anterior cranial fossa DAVFs, with complication of brain edema and confusion after fistula disconnection in only one of them. The surgical treatment for tentorial DAVFs is more difficult than DAVFs in the anterior cranial fossa because of the deep location and the surrounding vital structures, such as the deep veins. Lawton et al. [20] have described the results of microsurgery for treatment of 31 tentorial fistulas during a 9-year period, with complete obliteration rate of 94 % and transient neurological morbidity rate of 13 %. Although open surgery is effective for complete disconnection of fistulas in the anterior cranial fossa and the tentorium with a low risk of severe complication, due to the inherent invasiveness of craniotomy, a less invasive endovascular approach may be still preferred. For tentorial DAVFs, the posterior branch of the MMA was often the most common feeder and also the most common pedicle catheterized. The posterior branch of the MMA is usually enlarged and relatively straight when feeding a DAVF, allowing for distal catheterization adjacent to the point of the fistula and forward progression of the embolic material, thus increasing the likelihood of penetration of the embolic agent well into
Endovascular Treatment of Intracranial Dural Arteriovenous Fistulas Presenting with Intracranial
7
Fig. 3 A sexagenarian suddenly presented with headache. Head computed tomographic scan shows left frontal intraparenchymal hemorrhage (a, arrow). Lateral cerebral angiogram of the left internal carotid artery shows a fistula at the anterior cranial fossa feeding by the left ophthalmic artery, with drainage into the left inferior frontal cortical vein with venous varices (b, arrow). Superselective angiograms after catheterization of the ethmoidal feeder (c). Note the absence of a retinal choroidal blush and prominent opacification of the dural arteriovenous fistula. Immediately after Onyx embolization, control angiogram demonstrated complete closure of the fistula (d). The patient made a good clinical recovery after procedures and was waiting for follow-up angiograms (modified Rankin Scale = 1)
the proximal draining vein. Arterial supply of anterior cranial fossa DAVF is fed preferentially by ethmoidal branches of the ophthalmic artery (OphA) [18, 19]. Of nine anterior cranial fossa DAVFs of this series, eight were embolized by Onyx through transarterial approach, mostly often via the OphA route. The vascular anatomy must be carefully considered before embolized material injections for potential dangerous anastomoses and the cranial nerve supply to avoid perioperative complications. When embolizing from the posterior MMA branch, it is critical to avoid any reflux into the petrosal branch, which supplies the facial nerve. For anterior cranial fossa DAVFs using transarterial route of OphA, the origin of central retinal artery should be kept in mind, and close attention should be paid during Onyx injection to ensure that Onyx reflux does not occlude the central retinal artery [18, 19]. A relatively high complication rate of 13 % is observed in this series during or after treatment, including microcatheter entrapment, TCR, cranial nerve paresis, and migration of the Onyx embolus into the PICA, partially because of the lack of experience in the early period of using Onyx. Among them, the complication of Onyx embolus migrat-
ing into the PICA is depressing and fatal eventually. The other five patients with complications recovered well during follow-up period, and were asymptomatic or mildly symptomatic when clinical follow-up was performed. We should respect the anatomic limitations to avoid perioperative complications when using EVT for these lesions. As mentioned earlier in the text, the vascular anatomy must be carefully considered before embolized material injections for potential dangerous anastomoses and the cranial nerve supply. Complete and definitive angiographic fistula obliteration was the goal of the treatment of DAVFs, especially in patients presenting with intracranial hemorrhage. However, currently, not all the DAVFs could be cured even by multidisciplinary treatment. In this series, 14 patients (30.4 %) had residual fistulas at the end of their final treatment. The role of incomplete embolization is unclear. In our series, no rebleeding at follow-up was observed in the patients who had incomplete obliteration of the fistula after EVT. However, the length of follow-up is relatively short, so the present result may not be able to reflect the accurate prognosis of these patients, and longer follow-up is necessary to evaluate the long-term efficacy of incomplete embolization of these lesions.
13
8
Several limitations of the present study should be noted. Firstly, our study is a single-center, retrospective study that carries the inherent risk of patient selection bias. Secondly, the number of patients enrolled in this study was small, and the follow-up period in some cases was relatively short. This prevents us from generalizing the results precisely. Moreover, this analysis only focuses on EVT as treatment modality for patients with hemorrhagic DAVFs and generally represents only the main observational clinical and angiographic outcomes. Thus, we could not evaluate the superiority or shortcoming of EVT for these lesions because of the lack of control with other therapeutic approaches. In fact, many previous reports have described surgical treatment for DAVFs with promising results, particularly the lesions located at the anterior cranial fossa and the tentorium. Nevertheless, considering rarity of hemorrhagic DAVFs, the results of this study may provide some experience for better management of these lesions. Conclusions EVT was effective and safe in the modern management of ruptured intracranial DAVFs, with complete cure in most lesions. Clinical outcomes were good despite patients presenting with intracranial hemorrhage. Acknowledgments This work was supported by the National Science Foundation of China (grant nos. 81220108007, 81171079 and 81371315) and Special Research Project for Capital Health Development (grant no. 2014-1-1071). Conflict of Interest None
References 1. Kwon BJ, Han MH, Kang HS, Chang KH. MR imaging findings of intracranial dural arteriovenous fistulas: relations with venous drainage patterns. AJNR Am J Neuroradiol. 2005;26:2500–7. 2. Piippo A, Niemelä M, van Popta J, Kangasniemi M, Rinne J, Jääskeläinen JE, Hernesniemi J. Characteristics and long-term outcome of 251 patients with dural arteriovenous fistulas in a defined population. J Neurosurg. 2013;118:923–34. 3. Daniels DJ, Vellimana AK, Zipfel GJ, Lanzino G. Intracranial hemorrhage from dural arteriovenous fistulas: clinical features and outcome. Neurosurg Focus. 2013;34:E15. 4. Gandhi D, Chen J, Pearl M, Huang J, Gemmete JJ, Kathuria S. Intracranial dural arteriovenous fistulas: classification, imaging findings, and treatment. AJNR Am J Neuroradiol. 2012;33:1007–13.
13
C. Li et al. 5. Satoh K, Satomi J, Nakajima N, Matsubara S, Nagahiro S. Cerebellar hemorrhage caused by dural arteriovenous fistula: a review of five cases. J Neurosurg. 2001;94:422–6. 6. Duffau H, Lopes M, Janosevic V, Sichez JP, Faillot T, Capelle L, Ismaïl M, Bitar A, Arthuis F, Fohanno D. Early rebleeding from intracranial dural arteriovenous fistulas: report of 20 cases and review of the literature. J Neurosurg. 1999;90:78–84. 7. Borden JA, Wu JK, Shucart WA. A proposed classification for spinal and cranial dural arteriovenous fistulous malformations and implications for treatment. J Neurosurg. 1995;82:166–79. 8. Cognard C, Gobin YP, Pierot L, Bailly AL, Houdart E, Casasco A, Chiras J, Merland JJ. Cerebral dural arteriovenous fistulas: clinical and angiographic correlation with a revised classification of venous drainage. Radiology. 1995;194:671–80. 9. Singh V, Smith WS, Lawton MT, Halbach VV, Young WL. Risk factors for hemorrhagic presentation in patients with dural arteriovenous fistulae. Neurosurgery. 2008;62:628–35. 10. Albuquerque FC, Ducruet AF, Crowley RW, Bristol RE, Ahmed A, McDougall CG. Transvenous to arterial Onyx embolization. J Neurointerv Surg. 2014;6:281–5. 11. Cognard C, Januel AC, Silva NA, Jr., Tall P. Endovascular treatment of intracranial dural arteriovenous fistulas with cortical venous drainage: new management using Onyx. AJNR Am J Neuroradiol. 2008;29:235–41. 12. Shi ZS, Loh Y, Duckwiler GR, Jahan R, Viñuela F. Balloon-assisted transarterial embolization of intracranial dural arteriovenous fistulas. J Neurosurg. 2009;110:921–8. 13. Hoh BL, Choudhri TF, Connolly ES, Jr., Solomon RA. Surgical management of high-grade intracranial dural arteriovenous fistulas: leptomeningeal venous disruption without nidus excision. Neurosurgery. 1998;42:796–805. 14. Guo WY, Pan DH, Wu HM, Chung WY, Shiau CY, Wang LW, Chiou HJ, Yen MY, Teng MM. Radiosurgery as a treatment alternative for dural arteriovenous fistulas of the cavernous sinus. AJNR Am J Neuroradiol. 1998;19:1081–7. 15. Abud TG, Nguyen A, Saint-Maurice JP, Abud DG, Bresson D, Chiumarulo L, Enesi E, Houdart E. The use of Onyx in different types of intracranial dural arteriovenous fistula. AJNR Am J Neuroradiol. 2011;32:2185–91. 16. Adamczyk P, Amar AP, Mack WJ, Larsen DW. Recurrence of “cured” dural arteriovenous fistulas after Onyx embolization. Neurosurg Focus. 2012;32:E12. 17. Puffer RC, Daniels DJ, Kallmes DF, Cloft HJ, Lanzino G. Curative Onyx embolization of tentorial dural arteriovenous fistulas. Neurosurg Focus. 2012;32:E4. 18. Li C, Wu Z, Yang X, Li Y, Jiang C, He H. Transarterial treatment with Onyx of Cognard type IV anterior cranial fossa dural arteriovenous fistulas. J Neurointerv Surg. 2014;6:115–20. 19. Agid R, Terbrugge K, Rodesch G, Andersson T, Söderman M. Management strategies for anterior cranial fossa (ethmoidal) dural arteriovenous fistulas with an emphasis on endovascular treatment. J Neurosurg. 2009;110:79–84. 20. Lawton MT, Sanchez-Mejia RO, Pham D, Tan J, Halbach VV. Tentorial dural arteriovenous fistulae: operative strategies and microsurgical results for six types. Neurosurgery. 2008;62:110–25.
O r i g i n a l A rt i c l e
Endovascular Treatment of Intracranial Dural Arteriovenous Fistulas Presenting with Intracranial Hemorrhage in 46 Consecutive Patients: With Emphasis on Transarterial Embolization with Onyx C. Li · X. Yang · Y. Li · C. Jiang · Z. Wu
Received: 23 September 2014 / Accepted: 18 November 2014 © Springer-Verlag Berlin Heidelberg 2014
Abstract Purpose The purpose of this study was to evaluate the effectiveness and safety as well as the clinical and angiographic results of endovascular treatment (EVT) for patients with hemorrhagic dural arteriovenous fistulas (DAVFs). Methods From April 2009 to November 2013, 46 consecutive patients (7 women, 39 men; mean age, 46.7 years) diagnosed with hemorrhagic intracranial DAVFs at our department were enrolled in this study. Clinical and angiographic data were reviewed and evaluated. Results Two fistulas were cured by transvenous approach, and all other fistulas were embolized through transarterial route. After treatment with last embolization, a residual shunt was observed in 15 patients, including near-total obliteration of the fistula in 6 patients (13.0 %) and partial obliteration of the fistula in 9 patients (19.6 %). Seven of them underwent supplementary Gamma knife surgery. Complications occurred during or after EVT in six patients. One patient died 10 days after EVT because of complications of Onyx embolus into the posterior inferior cerebellar artery. Other complications included microcatheter entrapment in one case, trigeminocardiac reflex in one, facial nerve paresis in two, and occulomotor nerve paresis in one. The facial nerve paresis in one patient markedly improved but did not completely resolve during follow-up period. The other four patients recovered well, without symptoms at clinical follow-up. Angiographic follow-up result was obtained in
Z. Wu, MD () · C. Li, MS · X. Yang, MD · Y. Li, MD · C. Jiang, MD Beijing Neurosurgical Institute and Beijing Tiantan Hospital, Capital Medical University, No. 6, Tiantan Xili, Dongcheng District, 100050 Beijing, People’s Republic of China e-mail: [email protected]
35 cases, with complete obliteration in 27 of them (77.1 %). Clinical follow-up outcomes were good, with 38 patients (82.6 %) reporting modified Rankin Scale of either 0 or 1. No patient suffered recurrent intracranial hemorrhage during follow-up. Conclusions EVT was effective and safe in the modern management of ruptured intracranial DAVFs, with complete cure in most lesions. Clinical outcomes were good despite patients presenting with intracranial hemorrhage. Keywords Dural arteriovenous fistula · Intracranial hemorrhage · Endovascular treatment · Transarterial embolization · Onyx Introduction Intracranial dural arteriovenous fistulas (DAVFs) are pathologic shunts between dural arteries and dural venous sinuses, meningeal veins, or cortical veins, which account for 10–15 % of intracranial arteriovenous malformations [1]. Intracranial hemorrhage is one of the most serious clinical manifestations of intracranial DAVFs [2–4]. Intracranial DAVFs presenting with hemorrhage require early treatment aimed at complete and definitive fistula obliteration, considering the substantial risks of rebleeding after the first hemorrhage [5, 6]. Most previous studies regarding hemorrhagic intracranial DAVFs focused only on risk factors for intracranial hemorrhage and found that the existence of cortical venous reflux (CVR) increases the risk of intracranial hemorrhage for DAVFs [6–9]. Currently, large case series specifically evaluating the clinical and angiographic outcomes of patients diagnosed with ruptured DAVFs are rare [3]. During the past 2 decades, endovascular embolization has become a first-line treatment for DAVFs [4, 10–12].
13
2
The advent of Onyx (EV3, Irvine, CA), a new nonadhesive liquid embolic agent, has changed the treatment strategy for DAVF because of its advantages over other embolic agents [11, 12]. In this retrospective study, we present angiographic and clinical outcomes of 46 consecutive patients with hemorrhagic DAVFs, which were treated with endovascular treatment (EVT). The purpose of this study was to evaluate the effectiveness and safety of EVT for these lesions, with emphasis on transarterial embolization using Onyx. Materials and Methods Institutional review board approval was obtained for this retrospective study. Patients or their family members have given written informed consent to publish relevant materials. Patients and Data Collection We reviewed our patient database from April 2009 to November 2013 and identified 46 consecutive patients diagnosed with hemorrhagic DAVFs managed with EVT. Every patient underwent preoperative planar images examination, including head computed tomography (CT) scan or magnetic resonance (MR) imaging. Evidence of intracranial hemorrhage was confirmed with signs of intracranial blood on CT scan or MR imaging, or sometimes in the cerebrospinal fluid, by lumbar puncture. Diagnosis of DAVF was confirmed by a six-vessel digital subtraction angiography (DSA) including external carotid injection. The intracranial hemorrhage was attributed to the DAVF if there was no other explanation. Information collected and analyzed included demographics (age and sex), clinical presentation and extension of hemorrhage (intraparenchymal, subdural, subarachnoid, or intraventricular), location and angiographic features of the fistulas, EVT modality and endovascular materials selection, treatment complications, and angiographic and clinical follow-up outcomes. The diagnostic, immediately postoperative, and follow-up DSA findings were evaluated by at least two experienced neurointerventionalists to identify lesion locations, angiographic characteristics of the fistulas, and treatment results. Treatment Protocol and EVT Techniques In this series, every treatment decision-making was discussed in a multidisciplinary team consisting of neurologists, neurosurgeons, and neurointerventionalists. Every treatment was tailored to the individual patient on a case-bycase basis, considering various clinical and imaging parameters, including clinical presentation, patient age, the results
13
C. Li et al.
of imaging (location, arterial feeders, and venous drainage of the fistula), patient preference, and so on. All patients in this series received EVT as initial treatment to manage the fistulas. All procedures were performed with the patients under general anesthesia. Anticoagulation during the treatment process included heparin used in the flushing system of the guiding catheter (3000 IU/l of saline), without intravenous bolus of heparin given at the beginning of each procedure. Before the treatment procedure, cerebral angiogram was performed in all cases to confirm the diagnosis and angiographic characteristics, and to assist treatment planning. When transarterial approach was selected, vascular access was obtained via a transfemoral approach using a 6F guiding catheter. Sometimes when the lesion was fed by bilateral arteries or both the anterior circulation and posterior circulation, the contralateral transfemoral approach was also prepared to perform control angiograms to evaluate the degree of fistula obliteration during procedure. Using roadmapping technique, a microcatheter was coaxially navigated through the guiding catheter into a selected feeding artery. Progression of the microcatheter (Enchelon-10, EV3 or Marathon, EV3) was aided by the use of a microguidewire (Silverspeed-10/14, EV3; Mirage 008, EV3; Traxcess-14, Microvention) to reach to or as close as possible to the fistula. The choice of the microcatheter was at the discretion of the operator, and the most commonly used microcatheter in our center was Marathon microcatheter (EV3, Irvine, CA). When a safe embolization position was achieved, superselective angiography was performed to reveal the focal angioarchitecture and dynamic flow characteristics of the fistula. The vascular anatomy was carefully considered before embolization began for potential dangerous anastomoses and the cranial nerve supply. The decision regarding which endovascular material (Onyx/Nbutyl-2-cyanoacrylate, NBCA) and which concentration of the embolized material to use was subjective and was based on the distance from the fistula, flow velocity of the shunt, and risk of venous migration, often at the discretion of the operator. Sometimes blood flow of the shunt was so fast that we had to place appropriate coil to decrease the flow velocity at the fistula, and make the subsequently injected liquid embolic agent stable. The aim of glue embolization was to push the glue from the distal arterial feeder segment into the most proximal venous segment. As to Onyx injection, “plug and push” technique was often used to enable its penetration into the fistular compartments. Once Onyx advanced, it was intermittently injected until the proximal draining veins were completely filled, thereby completely obliterating the fistula. Some lesions needed multiple microcatheter embolization or multiple treatment sessions to obtain complete obliteration of the fistula or satisfactory flow reduction with suppression of CVR, until there was no appropriate route for
Endovascular Treatment of Intracranial Dural Arteriovenous Fistulas Presenting with Intracranial
further embolization. When the satisfactory embolization was unavailable using transarterial route or the supplying arteries were deemed as too tortuous to obtain satisfactory embolization position using transarterial approach, transvenous embolization was attempted with detachable coils and/or in combination with Onyx. The balloon-assisted flow control technique, which has been recently recommended as effective methods for high-flow DAVFs, was applied in selected cases, as described by Albuquerque et al. [10] and Shi et al. [12]. Angiographic follow-up was generally recommended at 3–9 months after treatment, preferably using DSA. CT or MR angiography was regarded as an alternative in cases when DSA was not available. If persistent CVR was noted after the embolization procedure, the patient was recommended for microsurgical disconnection or stereotactic radiosurgery (SRS) to achieve complete disconnection of the CVR (in this series, no patient agreed to accept adjunctive surgery). Clinical follow-up information was acquired by means of outpatient visits and/or phone contact and evaluated according to modified Rankin Scale (mRS). Results There were 39 men and 7 women, with a mean age of 46.7 years (range, 23–68 years). Tentorial fistulas (n = 21, 45.7 %) were the most common location of DAVFs followed by the anterior cranial fossa (n = 9, 19.6 %). All fistulas had direct cortical venous drainage or retrograde flow into cortical veins (Cognard type II b or greater). A total of 34 cases (73.9 %) were associated with venous varices (Cognard type IV). These patients experienced 51 intracranial hemorrhages in all; 41 patients experienced one intracranial hemorrhage each. Five patients suffered from rebleeding before treatment. Intraparenchymal hemorrhage, which occurred in 28 cases (60.9 %), was the most common bleeding pattern, and the site of intraparenchymal hemorrhage was most often near venous varix of the fistula. Other hemorrhage extensions included subarachnoid (n = 11, 23.9 %), intraparenchymal and intraventricular (n = 7, 15.2 %), subdural (n = 2, 4.3 %), and intraventricular (n = 1, 2.2 %). One patient presenting with intraparenchymal hemorrhage underwent surgical evacuation of the intraparenchymal hematoma. The fistula was not disconnected during operation in this patient. Four patients with intraventricular hemorrhage underwent extraventricular drainage. The demographics and clinical data of these patients are summarized in Table 1. All patients received EVT as initial treatment to manage their fistulas. A patient with right transverse sinus DAVF (Cognard type II b) underwent right transverse sinus packing initially using coils in association with Onyx, and the fistulas were completely obliterated. All fistulas in the other
3
Table 1 Demographics, location of fistula, extension of hemorrhage, clinical presentation, and Cognard classification Parameter Value Number of patients 46 Mean age (years) 46.7 Gender Male 39 (84.8) Female 7 (15.2) Location of fistula Tentorium 21 (45.7) Anterior cranial fossa 9 (19.6) Transverse-sigmoid sinus 6 (13.0) Cerebral convexity 6 (13.0) Foramen magnum 2 (4.3) Middle cranial fossa 2 (4.3) Intracranial hemorrhage extensions Intraparenchymal 28 (60.9) Subarachnoid 11 (23.9) Intraparenchymal and intraventricular 7 (15.2) Subdural 2 (4.3) Intraventricular 1 (2.2) Clinical presentation Before hemorrhage Asymptomatic 35 (76.1) Headache, bruit, and tinnitus 7 (15.2) Nonhemorrhagic neurological deficit 3 (6.5) Seizure 1 (2.2) After hemorrhage Headache, without neurological deficit 27 (58.7) Neurological deficit 14 (30.4) Coma 9 (19.6) Seizure 6 (13.0) Cognard classification Type V 2 (4.3) Type IV 34 (73.9) Type III 6 (13.0) Type IIa + b 3 (6.5) Type IIb 1 (2.2) Numbers in parentheses are percentages
patients were embolized with transarterial route as initial treatment. A patient with a fistula at the anterior cranial fossa was initially embolized through transarterial route. However, the supplying arteries were too tortuous to obtain a satisfactory embolization position, so transvenous coil embolization via the superior sagittal sinus was finally performed to completely obliterate the fistula. For the remaining 44 patients receiving transarterial embolization, a total of 57 embolization procedures were performed, with two or more procedures in 10 of them. Of these patients, 24 (24 of 44, 54.5 %) were cured with a single microcatheter embolization. Six of them required a second microcatheterization for complete occlusion. One patient required a third microcatheterization for an embolic cure. After treatment with last embolization, a residual shunt was observed in 15 patients (near-total obliteration, n = 6,
13
4
13.0 %; partial obliteration, n = 9, 19.6 %). Seven of them accepted Gamma knife surgery as supplementary treatment. No patient agreed to accept microsurgical intervention as adjunctive treatment in this series. Angiographic follow-up result was obtained in 35 cases (76.1 %), with a mean length of follow-up of 9.9 months. Of these patients, 27 (27 of 35, 77.1 %) had complete obliteration of their fistula. Angiographic follow-up results was available in four of the seven patients accepting supplementary Gamma knife surgery, and the results were as follows: two near-totally obliterated fistulas were transformed into totally obliterated; another near-totally obliterated fistula was still near-totally obliterated, but suppression of CVR was observed compared with immediate angiographic outcomes after EVT, and one partially obliterated fistula was transformed into near-totally obliterated. Complications occurred during or after EVT in six (13.0 %) cases. The microcatheter was stretched and broken during retrieval in one case, without any clinical symptoms. One patient encountered migration of the Onyx embolus into the posterior inferior cerebellar artery (PICA) and died of multiple organ failure10 days after EVT. During Onyx injection, bradycardia caused by trigeminocardiac reflex (TCR) occurred in one case, which was effectively extinguished after administering atropine. Other complications included facial nerve paresis in two cases and occulomotor nerve paresis in one. One patient presenting with facial nerve paresis still suffered from slight facial nerve paresis at 22-month clinical follow-up. The other two patients with cranial nerve paresis recovered well during the follow-up period, without symptom at clinical follow-up. Clinical follow-up outcome was obtained in 44 cases (one patient was lost). All the 44 patients recovered from their baseline condition during the follow-up period. Clinical outcomes were good in this series, with 38 patients (82.6 %) reporting mRS of either 0 or 1 (no symptoms or minimal residual symptoms). No patient suffered recurrent intracranial hemorrhage during the follow-up. The data of treatment, angiographic, and clinical outcomes are summarized in Table 2. Illustrative cases are presented in Figs. 1, 2, and 3. Discussion The clinical presentation of DAVFs varies among individuals, which could be “asymptomatic” type, “benign” type presenting with pulsatile tinnitus, proptosis, and other slight symptoms, or “aggressive” type presenting with intracranial hemorrhage or nonhemorrhagic neurologic deficits [4]. Despite the recognition of hemorrhage as one of the most serious sequelae of intracranial DAVFs and the treatment urgency of hemorrhagic DAVFs, to our knowledge, there is
13
C. Li et al. Table 2 Treatment, angiographic, and clinical outcomes Parameter Value Treatment TAE 37 (80.4) TAE and GKS 7 (15.2) TVE 2 (4.3) Endovascular material Onyx 37 (80.4) NBCA 2 (4.3) Onyx and NBCA 3 (6.5) Onyx and Coils 3 (6.5) Coils 1 (2.2) Endovascular complication Cranial nerve paresis 3 (6.5) Microcatheter entrapment 1 (2.2) Trigeminocardiac reflex 1 (2.2) Migration of the Onyx embolus 1 (2.2) Angiographic outcomes after EVT Complete obliteration 31 (67.4) Near-total obliteration 6 (13.0) Partial obliteration 9 (19.6) Angiographic follow-up 35 (76.1) Mean time (months) 9.9 Modalities DSA 26 (56.5) MRA or CTA 9 (19.6) Outcomes Complete obliteration 27 (58.7) Near-total obliteration 5 (10.9) Partial obliteration 3 (6.5) Clinical follow-up 44 (95.7) Mean time (months) 20.1 Outcomes (mRS) 0–1 38 (82.6) 2–3 4 (8.7) 4 2 (4.3) TAE transarterial embolization, GKS Gamma knife surgery, TVE transvenous embolization, NBCA N-butyl-2-cyanoacrylate, EVT endovascular treatment, DSA digital subtraction angiography, MRA magnetic resonance angiography, CTA computed tomography angiography, mRS modified Rankin Scale Numbers in parentheses are percentages
only one large series report specifically focusing on clinical characteristics and outcome of patients with hemorrhagic DAVFs [3]. In that article, Daniels et al. reported clinical outcomes of 28 patients with DAVFs who presented with intracranial hemorrhage, which were treated with EVT, surgery, SRS, or a combination of multiple modalities. In their case series, the majority of fistulas (75 %) were completely obliterated, and 71 % of patients experienced excellent clinical outcome (mRS of 0 or 1) [3]. In the present study, we analyzed the angiographic and clinical outcomes of 46 consecutive patients with hemorrhagic DAVFs to increase the experience for EVT of these hemorrhagic lesions. We also demonstrated a high rate of complete obliteration and safety
Endovascular Treatment of Intracranial Dural Arteriovenous Fistulas Presenting with Intracranial
5
Fig. 1 A quadragenarian suddenly presented with headache and speech arrest. Head computed tomographic scan shows intraparenchymal hemorrhage (a, arrow). Angiogram of left external carotid artery shows a fistula at the left middle cranial fossa feeding by the left middle meningeal artery (MMA), left accessory meningeal artery (AMA), and left sphenopalatine artery (b). The venous drainage was into the left vein of Labbe that had a venous aneurysm (b, arrow) and drained into the left transverse sinus. Three microcatheterizations with Onyx injection was performed through the left MMA and the left AMA during two treatment sessions to obtain an embolic cure. Cast of Onyx after the final embolization (c, arrow). Follow-up digital subtraction angiography 6 months later confirmed complete obliteration of the lesion (d). The patient made a very good clinical recovery (modified Rankin Scale = 0)
of endovascular embolization for hemorrhagic DAVFs. The study further demonstrates that EVT was effective in management of hemorrhagic DAVFs, with good clinical outcomes despite patients presenting with intracranial hemorrhage. Currently, there are mainly three alternatives for the treatment of DAVFs, including EVT, microsurgical disconnection of fistulas, and SRS [2, 4, 10–14]. Microsurgery is a traditionally recommended treatment approach [13]. However, due to the risks inherent in craniotomy and the efficacy of EVT, surgery is currently only indicated in cases where endovascular approaches have failed or are not feasible [4]. For SRS, the risk of hemorrhage still remains during the latency period ranging from several months to years, until vessel thrombosis and fistula closure happen, so it is inappropriate as the primary treatment in hemorrhagic DAVFs considering the high risk of rebleeding [14]. This technique should be reserved for carefully selected DAVFs for which endovascular and surgical options have been exhausted. Currently, endovascular embolization using transarterial or transvenous approach has become a first-line treatment for DAVFs [2, 4, 10–12]. With modern endovascular tech-
niques such as flow control techniques using balloon, better understanding of the pathophysiology and anatomy of intracranial DAVFs can be achieved, with EVTs resulting in high rates of complete obliteration with a low risk to the patient. Treatment of DAVFs is aimed at complete cure of the arteriovenous fistulas—incomplete treatment allows recruitment of collateral vessels and persistent risk of hemorrhage. NBCA has been used as an embolic agent with relatively good results, but it requires quick and continuous injection technique to avoid unintended catheter tip adherence to the vessel wall due to its high thrombogenicity. The advent of new nonadhesive liquid embolic agent, Onyx (EV3, Irvine, CA), has changed the treatment strategy for DAVFs because of its greater safety and efficacy compared with NBCA in endovascular embolization, especially when distal enough navigation of the microcatheter was difficult [11, 12, 15]. The increased control of Onyx made it possible to deal with DAVFs with multiple arterial supplies, allowing the Onyx to penetrate slowly and precisely into the afferent arteries well until the proximal draining veins are affected. In this series, a fistula (Fig. 2) “recurred” on follow-up angiogram although “complete obliteration” of the fistula was obtained
13
6
C. Li et al.
Fig. 2 A sexagenarian suddenly presented with alexia, followed by seizure and coma. Head computed tomographic scan shows intraparenchymal hemorrhage (a, arrow). Angiogram of the left common carotid artery shows a fistula at the left occipital lobe with feeders from posterior branches of the left middle meningeal artery (MMA) and meningeal branches of the left occipital artery (OA), drained into cortical venous vein with venous varix (b, arrow). Angiogram obtained after two microcatheterization with Onyx embolization through the routes of the left MMA and the left OA shows “complete obliteration” of the fistula (c). However, digital subtraction angiography (DSA) 6 months later revealed recurrence of this “cured” fistula (d, arrow). The patient recovered well, and no rebleeding occurred during 9 months of clinical follow-up (modified Rankin Scale = 1). This patient underwent Gamma knife surgery as supplementary treatment after the follow-up DSA examination and was waiting for DSA follow-up again
via transarterial embolization with Onyx. In fact, the fistula in this case was not really cured because the proximal venous segment was not affected, thus fistula recanalization may occur with filling from newly recruited arterial feeders. Adamczyk et al. [16] have reported that despite adequate Onyx penetration into the fistula and draining vein, DAVF recanalization may also appear. However, no such case was observed in this series. Although this kind of case was uncommon and the definite underlying mechanism for this recurrence remains unclear, the possibility for DAVF recanalization should highlight the need for performing followup angiography after EVT. In this series, tentorial fistulas (45.7 %) were the most common location of DAVFs followed by the anterior cranial fossa DAVF (19.6 %). DAVFs in these two locations differ from other subtypes because they usually have direct drainage into cortical veins, often with formation of venous ectasia (Cognard type IV) [17–19]. These characteristics make the lesions prone to intracranial hemorrhage. Many previous reports have described surgical treatment for DAVFs located at the anterior cranial fossa and the tentorium with promising result. Agid et al. [19] have reported open surgical dis-
13
connection was effective in obliterating the fistula in all 15 patients with anterior cranial fossa DAVFs, with complication of brain edema and confusion after fistula disconnection in only one of them. The surgical treatment for tentorial DAVFs is more difficult than DAVFs in the anterior cranial fossa because of the deep location and the surrounding vital structures, such as the deep veins. Lawton et al. [20] have described the results of microsurgery for treatment of 31 tentorial fistulas during a 9-year period, with complete obliteration rate of 94 % and transient neurological morbidity rate of 13 %. Although open surgery is effective for complete disconnection of fistulas in the anterior cranial fossa and the tentorium with a low risk of severe complication, due to the inherent invasiveness of craniotomy, a less invasive endovascular approach may be still preferred. For tentorial DAVFs, the posterior branch of the MMA was often the most common feeder and also the most common pedicle catheterized. The posterior branch of the MMA is usually enlarged and relatively straight when feeding a DAVF, allowing for distal catheterization adjacent to the point of the fistula and forward progression of the embolic material, thus increasing the likelihood of penetration of the embolic agent well into
Endovascular Treatment of Intracranial Dural Arteriovenous Fistulas Presenting with Intracranial
7
Fig. 3 A sexagenarian suddenly presented with headache. Head computed tomographic scan shows left frontal intraparenchymal hemorrhage (a, arrow). Lateral cerebral angiogram of the left internal carotid artery shows a fistula at the anterior cranial fossa feeding by the left ophthalmic artery, with drainage into the left inferior frontal cortical vein with venous varices (b, arrow). Superselective angiograms after catheterization of the ethmoidal feeder (c). Note the absence of a retinal choroidal blush and prominent opacification of the dural arteriovenous fistula. Immediately after Onyx embolization, control angiogram demonstrated complete closure of the fistula (d). The patient made a good clinical recovery after procedures and was waiting for follow-up angiograms (modified Rankin Scale = 1)
the proximal draining vein. Arterial supply of anterior cranial fossa DAVF is fed preferentially by ethmoidal branches of the ophthalmic artery (OphA) [18, 19]. Of nine anterior cranial fossa DAVFs of this series, eight were embolized by Onyx through transarterial approach, mostly often via the OphA route. The vascular anatomy must be carefully considered before embolized material injections for potential dangerous anastomoses and the cranial nerve supply to avoid perioperative complications. When embolizing from the posterior MMA branch, it is critical to avoid any reflux into the petrosal branch, which supplies the facial nerve. For anterior cranial fossa DAVFs using transarterial route of OphA, the origin of central retinal artery should be kept in mind, and close attention should be paid during Onyx injection to ensure that Onyx reflux does not occlude the central retinal artery [18, 19]. A relatively high complication rate of 13 % is observed in this series during or after treatment, including microcatheter entrapment, TCR, cranial nerve paresis, and migration of the Onyx embolus into the PICA, partially because of the lack of experience in the early period of using Onyx. Among them, the complication of Onyx embolus migrat-
ing into the PICA is depressing and fatal eventually. The other five patients with complications recovered well during follow-up period, and were asymptomatic or mildly symptomatic when clinical follow-up was performed. We should respect the anatomic limitations to avoid perioperative complications when using EVT for these lesions. As mentioned earlier in the text, the vascular anatomy must be carefully considered before embolized material injections for potential dangerous anastomoses and the cranial nerve supply. Complete and definitive angiographic fistula obliteration was the goal of the treatment of DAVFs, especially in patients presenting with intracranial hemorrhage. However, currently, not all the DAVFs could be cured even by multidisciplinary treatment. In this series, 14 patients (30.4 %) had residual fistulas at the end of their final treatment. The role of incomplete embolization is unclear. In our series, no rebleeding at follow-up was observed in the patients who had incomplete obliteration of the fistula after EVT. However, the length of follow-up is relatively short, so the present result may not be able to reflect the accurate prognosis of these patients, and longer follow-up is necessary to evaluate the long-term efficacy of incomplete embolization of these lesions.
13
8
Several limitations of the present study should be noted. Firstly, our study is a single-center, retrospective study that carries the inherent risk of patient selection bias. Secondly, the number of patients enrolled in this study was small, and the follow-up period in some cases was relatively short. This prevents us from generalizing the results precisely. Moreover, this analysis only focuses on EVT as treatment modality for patients with hemorrhagic DAVFs and generally represents only the main observational clinical and angiographic outcomes. Thus, we could not evaluate the superiority or shortcoming of EVT for these lesions because of the lack of control with other therapeutic approaches. In fact, many previous reports have described surgical treatment for DAVFs with promising results, particularly the lesions located at the anterior cranial fossa and the tentorium. Nevertheless, considering rarity of hemorrhagic DAVFs, the results of this study may provide some experience for better management of these lesions. Conclusions EVT was effective and safe in the modern management of ruptured intracranial DAVFs, with complete cure in most lesions. Clinical outcomes were good despite patients presenting with intracranial hemorrhage. Acknowledgments This work was supported by the National Science Foundation of China (grant nos. 81220108007, 81171079 and 81371315) and Special Research Project for Capital Health Development (grant no. 2014-1-1071). Conflict of Interest None
References 1. Kwon BJ, Han MH, Kang HS, Chang KH. MR imaging findings of intracranial dural arteriovenous fistulas: relations with venous drainage patterns. AJNR Am J Neuroradiol. 2005;26:2500–7. 2. Piippo A, Niemelä M, van Popta J, Kangasniemi M, Rinne J, Jääskeläinen JE, Hernesniemi J. Characteristics and long-term outcome of 251 patients with dural arteriovenous fistulas in a defined population. J Neurosurg. 2013;118:923–34. 3. Daniels DJ, Vellimana AK, Zipfel GJ, Lanzino G. Intracranial hemorrhage from dural arteriovenous fistulas: clinical features and outcome. Neurosurg Focus. 2013;34:E15. 4. Gandhi D, Chen J, Pearl M, Huang J, Gemmete JJ, Kathuria S. Intracranial dural arteriovenous fistulas: classification, imaging findings, and treatment. AJNR Am J Neuroradiol. 2012;33:1007–13.
13
C. Li et al. 5. Satoh K, Satomi J, Nakajima N, Matsubara S, Nagahiro S. Cerebellar hemorrhage caused by dural arteriovenous fistula: a review of five cases. J Neurosurg. 2001;94:422–6. 6. Duffau H, Lopes M, Janosevic V, Sichez JP, Faillot T, Capelle L, Ismaïl M, Bitar A, Arthuis F, Fohanno D. Early rebleeding from intracranial dural arteriovenous fistulas: report of 20 cases and review of the literature. J Neurosurg. 1999;90:78–84. 7. Borden JA, Wu JK, Shucart WA. A proposed classification for spinal and cranial dural arteriovenous fistulous malformations and implications for treatment. J Neurosurg. 1995;82:166–79. 8. Cognard C, Gobin YP, Pierot L, Bailly AL, Houdart E, Casasco A, Chiras J, Merland JJ. Cerebral dural arteriovenous fistulas: clinical and angiographic correlation with a revised classification of venous drainage. Radiology. 1995;194:671–80. 9. Singh V, Smith WS, Lawton MT, Halbach VV, Young WL. Risk factors for hemorrhagic presentation in patients with dural arteriovenous fistulae. Neurosurgery. 2008;62:628–35. 10. Albuquerque FC, Ducruet AF, Crowley RW, Bristol RE, Ahmed A, McDougall CG. Transvenous to arterial Onyx embolization. J Neurointerv Surg. 2014;6:281–5. 11. Cognard C, Januel AC, Silva NA, Jr., Tall P. Endovascular treatment of intracranial dural arteriovenous fistulas with cortical venous drainage: new management using Onyx. AJNR Am J Neuroradiol. 2008;29:235–41. 12. Shi ZS, Loh Y, Duckwiler GR, Jahan R, Viñuela F. Balloon-assisted transarterial embolization of intracranial dural arteriovenous fistulas. J Neurosurg. 2009;110:921–8. 13. Hoh BL, Choudhri TF, Connolly ES, Jr., Solomon RA. Surgical management of high-grade intracranial dural arteriovenous fistulas: leptomeningeal venous disruption without nidus excision. Neurosurgery. 1998;42:796–805. 14. Guo WY, Pan DH, Wu HM, Chung WY, Shiau CY, Wang LW, Chiou HJ, Yen MY, Teng MM. Radiosurgery as a treatment alternative for dural arteriovenous fistulas of the cavernous sinus. AJNR Am J Neuroradiol. 1998;19:1081–7. 15. Abud TG, Nguyen A, Saint-Maurice JP, Abud DG, Bresson D, Chiumarulo L, Enesi E, Houdart E. The use of Onyx in different types of intracranial dural arteriovenous fistula. AJNR Am J Neuroradiol. 2011;32:2185–91. 16. Adamczyk P, Amar AP, Mack WJ, Larsen DW. Recurrence of “cured” dural arteriovenous fistulas after Onyx embolization. Neurosurg Focus. 2012;32:E12. 17. Puffer RC, Daniels DJ, Kallmes DF, Cloft HJ, Lanzino G. Curative Onyx embolization of tentorial dural arteriovenous fistulas. Neurosurg Focus. 2012;32:E4. 18. Li C, Wu Z, Yang X, Li Y, Jiang C, He H. Transarterial treatment with Onyx of Cognard type IV anterior cranial fossa dural arteriovenous fistulas. J Neurointerv Surg. 2014;6:115–20. 19. Agid R, Terbrugge K, Rodesch G, Andersson T, Söderman M. Management strategies for anterior cranial fossa (ethmoidal) dural arteriovenous fistulas with an emphasis on endovascular treatment. J Neurosurg. 2009;110:79–84. 20. Lawton MT, Sanchez-Mejia RO, Pham D, Tan J, Halbach VV. Tentorial dural arteriovenous fistulae: operative strategies and microsurgical results for six types. Neurosurgery. 2008;62:110–25.
