Neuroradiology (2014) 56:763–769 DOI 10.1007/s00234-014-1395-2
INTERVENTIONAL NEURORADIOLOGY
Ruptured brain arteriovenous malformations associated with aneurysms: safety and efficacy of selective embolization in the acute phase of hemorrhage Francesco Signorelli & Benjamin Gory & Isabelle Pelissou-Guyotat & Jacques Guyotat & Roberto Riva & Frédéric Dailler & Francis Turjman
Received: 1 February 2014 / Accepted: 17 June 2014 / Published online: 28 June 2014 # Springer-Verlag Berlin Heidelberg 2014
Abstract Introduction The impact of targeted embolization of ruptured cerebral arteriovenous malformation (AVM)-associated arterial aneurysms in the acute phase of bleeding is not well known. The objective of our study was to analyze the safety and efficacy of this treatment strategy with special emphasis on its protective effect against rebleeding. Methods From a prospective database, all patients presenting between December 2005 and March 2012 with a ruptured cerebral AVM associated with arterial aneurysms contiguous to the hemorrhage were selected. Hemorrhagic stroke severity and clinical outcome were measured using, respectively, Glasgow Coma Scale (GCS) and modified Rankin Scale (mRS) scores. Results Twenty-five patients with ruptured cerebral AVMs associated with 31 arterial aneurysms contiguous with the hemorrhage were included. In the first group of 17 patients, the mean time to treatment was 30.1 h, while in the second group with 1 patient in poor initial clinical conditions and 7 admitted 7 to 28 days after bleeding, it was 17 days. All F. Signorelli (*) : I. Pelissou-Guyotat : J. Guyotat Department of Neurosurgery, Hospices Civils de Lyon, Université Claude Bernard, “Pierre Wertheimer”, 59, Boulevard Pinel, 69500 Lyon, France e-mail: [email protected] B. Gory : R. Riva : F. Turjman Department of Interventional Neuroradiology, Hospices Civils de Lyon, Université Claude Bernard, Lyon, France F. Dailler Department of Intensive Care and Critical Care Unit, Hospices Civils de Lyon, Université Claude Bernard, Lyon, France F. Signorelli Department of Experimental and Clinical Medicine, University “Magna Graecia”, Catanzaro, Italy
arterial aneurysms were totally occluded. Four patients presented rebleeding before treatment (mean interval 9.6 days) and four had rebleeding after treatment (mean interval 8.25 months). At a mean follow-up of 56 months, 21 patients were alive (19 mRS≤2, 2 mRS>2). Three patients died 2 days, 9 days, and 2 months after bleeding and another died of rebleeding of her residual AVM 19 months after treatment. Treatment-related morbidity and mortality were, respectively, 4 and 0 %. Conclusion Selective embolization of arterial aneurysms associated with ruptured AVMs is a safe procedure and could lower the immediate risk of rebleeding. Keywords Arteriovenous malformation . Aneurysm . Intracranial hemorrhage . Rebleeding . Embolization
Introduction Intracranial hemorrhage (ICH) is the most important factor affecting morbidity and mortality in patients with brain arteriovenous malformations (AVMs) [1–3]. Several authors have reported a stringent association between hemorrhagic AVM presentation, rebleeding, and presence of prenidal, intranidal, and postnidal aneurysms [4–20]. However, although plausible, the protective role of embolization of arterial aneurysms associated with ruptured brain AVMs is not demonstrated yet. In this article, we describe our treatment paradigm, conceived several years ago and maintained to date, that guided the treatment of patients presenting with ruptured AVMs associated with one or multiple arterial aneurysms considered to be the AVM’s weak point and deemed to require urgent occlusion. The objective of our study was to verify whether selective embolization of these aneurysms was safe and could be
764
proposed to induce a protective effect against AVM rebleeding.
Methods Study population From a prospectively collected database including all patients with AVMs admitted to our institution, we selected all patients treated for a ruptured cerebral AVM associated with one or more arterial aneurysms between December 2005 and March 2012. Intracranial hemorrhages related to the AVM were demonstrated on either computer tomography (CT) or magnetic resonance imaging (MRI). All AVMs were diagnosed based on preoperative four-vessel digital subtraction angiography (DSA). Arterial aneurysms were defined on imaging as saccular dilatations with a diameter equal or larger than the diameter of the parent vessel, thought to be a weak point in the angioarchitecture of the AVM. We took into consideration flow-related and dissecting aneurysms located on the arteries feeding the AVM [18] (types II and III of the Yu classification [10]) and intranidal aneurysms [18] (type IV of the Yu classification [10]), providing that they were contiguous with the hemorrhage, as seen on the preoperative imaging workup (Figs. 1 and 2). Unruptured brain AVMs and ruptured brain AVMs without arterial aneurysms or presenting with an aneurysm on the venous side or not related to the hemorrhage on the preoperative cerebral angiogram and CT and/or MR images and other types of vascular malformations such as dural fistulas or vein of Galen malformations were excluded from the analysis. Demographic, clinical, and morphological treatment and follow-up data not specified into the database were collected reviewing the clinical chart of each patient. Hemorrhagic stroke severity and clinical outcome after hemorrhagic presentation and during follow-up were measured using, respectively, the Glasgow Coma Scale (GCS) and the modified Rankin Scale (mRS). Treatment strategy for each case was discussed by a multidisciplinary team including interventional neuroradiologists, neurosurgeons, and neurointensivists and was aimed at occluding the arterial aneurysm as soon as possible and the AVM and the associated intracranial hypertension accordingly. All patients had an immediate postoperative cerebral CT scan and DSA, and all surviving patients were followed-up with at least 1-year postoperative clinical evaluation and mRS determination, brain CT and/or MR imaging, and cerebral DSA. Embolization procedure All endovascular procedures were performed under general anesthesia at a biplane angiography unit (Neurostar and Axiom Artis, Siemens, Erlangen, Germany). Vascular access was
Neuroradiology (2014) 56:763–769
obtained with a 6F guiding catheter (Envoy; Cordis, Miami Lakes, FL, USA). Superselective angiography was carried out with an over-the-wire microcatheter and with a DMSOcompatible detachable tip microcatheter (Sonic; Balt extrusion, Montmorency, France). After the microcatheter tip was placed at a perinidal position, superselective angiography was performed to confirm the aneurysms and its characteristics and relationships with the parent vessel. Aneurysms were embolized with N-butyl cyanoacrylate (n-BCA; Cordis Microvascular, Inc., New Brunswick, NJ), Onyx (Covidien), or bare platinum coils (Trufill from Cordis, Miami Lakes, Florida; target from Stryker Neurovascular). When deemed safe, the associated AVM was treated either completely or partially immediately thereafter.
Results In the reviewed period, 25 patients presented to our hospital with ruptured cerebral AVMs associated with an aneurysm as a suspected source of bleeding. They were 16 males and 9 females, with a mean age of 42.2 years (range 10–79 years) (see Table 1). Clinical presentation and type of intracranial bleeding The overwhelming majority of patients (21/25, 84 %) presented with sudden violent headache, which was associated with meningismus in 5. Six patients (24 %) had a focal or generalized seizure and 15 (60 %) displayed a fixed neurological deficit. The mean GCS was 11.7, with seven patients (28 %) being comatous (GCS≤8) and four (16 %) having a GCS of 3. The hemorrhage was both intraparenchimal and subarachnoid in 19 out of 25 cases (76 %). Of those, three (12 %) had an intraventricular extension. Three patients (12 %) presented with a pure intraparenchimal bleeding. A pure subarachnoid or a pure intraventricular bleeding occurred in one patient each (4 %), while one patient (4 %) had a subarachnoid bleeding with intraventricular extension. Angioarchitecture and site of aneurysm There were 3 Spetzler-Martin [21] grade 1, 11 grade 2, 7 grade 3, and 4 grade 4 AVMs. In all cases, the initial angiogram demonstrated an aneurysm contiguous to the intracranial hemorrhage, which was considered the source of bleeding (Figs. 1 and 2). A total of 31 aneurysms were visualized. The ruptured aneurysm was located inside the nidus in 15 cases (48.4 %) and on the arterial feeders in 16 cases (51.6 %) (Figs. 1 and 2), four of them (12.9 %) being ruptured flow-related aneurysms [18], type II of Yu classification [10] (Fig. 2). Six patients (24 %) displayed multiple ruptured aneurysms (Fig. 2): three
Neuroradiology (2014) 56:763–769
765
Fig. 1 A 45-year-old man with sudden headache, impaired consciousness, and cerebellar ataxia. a Contrast enhanced T1weighted MR imaging demonstrates a cerebellar hematoma with an aneurysm on an AVM feeder indicating the source of bleeding (arrow). b Lateral view of vertebral angiography shows a supracerebellar AVM with distal arterial feeder aneurysm as the source of hemorrhage (arrow). c Lateral view of microcatheter angiography during NCBA embolization of the superior cerebellar artery aneurysm (arrow). d Complete obliteration of the AVM after subsequent surgical removal
patients had two intranidal aneurysms each, two patients had two aneurysms on AVM arterial feeders, and one patient had one aneurysm on arterial feeder and one intranidal. Timing of treatment Two groups of patients were identified according to the mean time to endovascular treatment (EVT): for the first group of 17 patients, the mean time to EVT was 30.1 h, while for the second group of 8 patients, it was 17 days. EVT of the aneurysm was delayed for one patient in poor initial clinical condition and for seven patients admitted to our hospital between 7 and 28 days after bleeding. Two of those patients were admitted after AVM rebleeding respectively 1 and 4 weeks after the first hemorrhage. All six patients having multiple aneurysms underwent single-stage embolization of all lesions because, in all cases, the preoperative imaging workup showed no distinctive pattern of the hemorrhage neither aneurysms’ distinctive features of rupture allowing to identify with certainty the bleeding lesion (Fig. 2). Type of treatment and angiographic results Twenty-four (77.4 %) aneurysms were embolized with nBCA (Figs. 1 and 2) and four (12.9 %) with Onyx (Fig. 2), while coils were used for three (9.7 %) aneurysms that were localized proximally on arterial feeders (Fig. 2). All
aneurysms were completely excluded on postoperative angiograms. The goal of the treatment by glue or Onyx was to obtain a segmental occlusion of the vessel along with the aneurysm, the migration of the embolic liquid inside the sac being considered a proof of cure, but not a treatment requirement. A segmental occlusion was performed based on the assumption that cortical anastomoses will take care of the distal vascularization. A nonadhesive liquid embolic material such as Onyx was preferred to glue when complete or partial treatment of associated nidus was planned in the same setting as the embolization of ruptured aneurysm and the chance of jeopardizing vascularization of normal brain with a reflux of 15 to 20 mm was small. Seventeen patients (68 %) underwent EVT of the sole aneurysm. The other eight patients (32 %) underwent additional occlusion of the AVM in the same setting as the embolization of the aneurysms, which was complete in five (20 %). Treatment of residual AVM was completed by subsequent embolization in seven patients (28 %), with a mean delay of 4.2 months (range 1–10 months), once blood clot is resolved or patient has recovered from initial bleeding. Residual AVM was completely treated by embolization and surgery or radiosurgery in two (8 %) patients each, by solesurgery or radiosurgeryin four (16 %) and one patient (4 %), respectively, while four other patients (16 %), in poor clinical condition (GCS≤6) at admission, had only embolization of the aneurysm as treatment strategy (Table 2).
766
Neuroradiology (2014) 56:763–769 Table 1 Baseline characteristics of 25 patients with AVM-associated ruptured arterial aneurysms Age, mean (range)
42.2 years (10–79 years)
Sex Clinical presentation Headache Meningismus
16 males, 9 females
Seizures Fixed neurodeficit GCS 15–13 12–9 ≤8 Spetzler-Martin grade Grade I Grade II Grade III Grade IV Location of AVM Lobar Deep Cerebellar Brain Stem Arterial feeders Pial
Fig. 2 A 63-year-old woman with sudden headache and cerebellar ataxia. a CT scan shows a vermian hemorrhage. b Lateral view of vertebral angiography shows a supracerebellar AVM with two aneurysms (arrows) on the telovelotonsillary segment of L PICA as the source of hemorrhage. c Lateral view of vertebral angiography shows coil embolization of the distal aneurysm (arrow). d Lateral view of vertebral angiography shows subsequent NCBA occlusion of the proximal aneurysm (arrow) with segmental occlusion of the parent arterial branch. e Six months later, CT scan shows rebleeding of the AVM. f Lateral view of vertebral angiography shows a de novo aneurysm on the distal segment of L PICA as the source of rebleeding (arrow). g Nonsubtracted image of Onyx cast after embolization, with complete filling of the sac (short arrow) and occlusion of the parent branch. Note the coils (long arrow) occluding one of the two aneurysms source of the first bleeding. h Angiography demonstrating complete obliteration of the de novo aneurysm
Perforators Mixed ≤2 >2 Venous drainage Superficial Deep Mixed Associated aneurysms Proximal on arterial feeder Distal on arterial feeder Intranidal First hemorrhage site Intraparenchimal only Subarachnoid only Intraventricular only Mixed Rebleeding Before treatment After treatment
21 (84 %) 5 (20 %) 6 (24 %) 15 (60 %) 12 (48 %) 6 (24 %) 7 (28 %) 3 (12 %) 11 (44 %) 7 (28 %) 4 (16 %) 17 (68 %) 3 (12 %) 4 (16 %) 1 (4 %) 15 (60 %) 3 (12 %) 7 (28 %) 20 (80 %) 5 (20 %) 13 (52 %) 1 (4 %) 11 (44 %) 31 (100 %) 4 ( 12.9 %) 12 (38.7 %) 15 (48.4 %) 3 (12 %) 1 (4 %) 1 (4 %) 20 (80 %) 8 (32 %) 4 (16 %) 4 (16 %)
GCS Glasgow Coma Scale, AVM arteriovenous malformation
Neuroradiology (2014) 56:763–769
767
Table 2 Management categories and outcome in AVMs (n=25) after embolisation of associated ruptured aneurysms Type of management of AVM
Number of cases (%)
mRS 0–1
mRS 2
mRS 3–4
Death
Embolisation only
12 (48 %)
7 (28 %)
2 (8 %) due to: 2 gravity of 1st bleed
0
Surgery only
4 (16 %)
3 (12 %)
0
0
Radiosurgery only
1 (4 %)
0
0
0
Multimodality treatment
4 comprising: 2 embolisation and surgery 2 embolisation and RxS 4 comprising: 3 embolisation of aneurysm only 1 embolisation of aneurysm and partial embolisation of AVM
4
3 (12 %) due to: 1 EVT complication 2 gravity of 1st bleed 1 (4 %) due to: 1 AVM rebleeding 1 due to: 1 gravity of 1st bleed 0
0
0
0
0
0
4 due to: 3 gravity of 1st bleeding 1 AVM rebleeding
Conservative management
EVT endovascular treatment, RxS radiosurgery, mRS modified Rankin Scale, AVM arteriovenous malformation
Rebleeding During the mean prospective follow-up period of 56 months (range 2–108 months), eight rebleedings occurred in 116.7 prospective years of observation. Four patients (16 % of patients, 12.9 % of aneurysms) presented rebleeding of their AVM before treatment of their aneurysm with a mean interval of 9.6 days (1–28 days) (Fig. 2). All these four patients had emergent embolization of their aneurysm, within 24 h after rebleeding. For one patient, rebleeding did not lead to appreciable worsening of her condition, while in three other patients, it resulted in sudden impairment of consciousness and new neurological deficits related to enlargement of existing intracerebral hematoma. One patient held a postoperative left side hemiplegia and gradually improved to a mRS of 2 at 9-year follow-up. The other patient displayed aggravation of his ataxia postoperatively, that improved subsequently and had a mRS of 1 at 20-month follow-up. The last patient died as a consequence of rebleeding 2 months thereafter. Four other patients (16 % of patients, 12.9 % of aneurysms) had rebleeding of their residual AVM after treatment of the aneurysm (two patients) or partial embolization of the AVM (two patients) at 3, 5, 6, and 19 months (mean 8.25 months). The first three patients did not show any clinical worsening after rebleeding. In all of them, repeated DSA showed a “de novo” false aneurysm, intranidal in two cases and on the distal segment of an arterial feeder in one (Fig. 2), absent on DSA performed at the time of first bleed, that colocalized with the rebleeding. These lesions were
considered the weak point of the AVM and were treated in emergency respectively at 16, 24, and 36 h after rebleeding. The last patient died of rebleeding before repeating DSA. The four rebleedings after treatment in 116.7 prospective years of observation equaled to a 3.4 % annual rehemorrhage rate. Clinical and radiological follow-up One patient experienced immediate postoperative embolic occlusion of the distal left ACA, indirect feeder of the AVM, due to accidental migration of n-BCA into the artery, with worsening of his right lower limb paresis and aphasia that improved in 12 months to a mRS of 2. At a mean follow-up of 56 months, 21 patients (84 %) were alive. Nineteen patients (76 %) were independent, with mRS≤2. Of these, three presented with GCS
INTERVENTIONAL NEURORADIOLOGY
Ruptured brain arteriovenous malformations associated with aneurysms: safety and efficacy of selective embolization in the acute phase of hemorrhage Francesco Signorelli & Benjamin Gory & Isabelle Pelissou-Guyotat & Jacques Guyotat & Roberto Riva & Frédéric Dailler & Francis Turjman
Received: 1 February 2014 / Accepted: 17 June 2014 / Published online: 28 June 2014 # Springer-Verlag Berlin Heidelberg 2014
Abstract Introduction The impact of targeted embolization of ruptured cerebral arteriovenous malformation (AVM)-associated arterial aneurysms in the acute phase of bleeding is not well known. The objective of our study was to analyze the safety and efficacy of this treatment strategy with special emphasis on its protective effect against rebleeding. Methods From a prospective database, all patients presenting between December 2005 and March 2012 with a ruptured cerebral AVM associated with arterial aneurysms contiguous to the hemorrhage were selected. Hemorrhagic stroke severity and clinical outcome were measured using, respectively, Glasgow Coma Scale (GCS) and modified Rankin Scale (mRS) scores. Results Twenty-five patients with ruptured cerebral AVMs associated with 31 arterial aneurysms contiguous with the hemorrhage were included. In the first group of 17 patients, the mean time to treatment was 30.1 h, while in the second group with 1 patient in poor initial clinical conditions and 7 admitted 7 to 28 days after bleeding, it was 17 days. All F. Signorelli (*) : I. Pelissou-Guyotat : J. Guyotat Department of Neurosurgery, Hospices Civils de Lyon, Université Claude Bernard, “Pierre Wertheimer”, 59, Boulevard Pinel, 69500 Lyon, France e-mail: [email protected] B. Gory : R. Riva : F. Turjman Department of Interventional Neuroradiology, Hospices Civils de Lyon, Université Claude Bernard, Lyon, France F. Dailler Department of Intensive Care and Critical Care Unit, Hospices Civils de Lyon, Université Claude Bernard, Lyon, France F. Signorelli Department of Experimental and Clinical Medicine, University “Magna Graecia”, Catanzaro, Italy
arterial aneurysms were totally occluded. Four patients presented rebleeding before treatment (mean interval 9.6 days) and four had rebleeding after treatment (mean interval 8.25 months). At a mean follow-up of 56 months, 21 patients were alive (19 mRS≤2, 2 mRS>2). Three patients died 2 days, 9 days, and 2 months after bleeding and another died of rebleeding of her residual AVM 19 months after treatment. Treatment-related morbidity and mortality were, respectively, 4 and 0 %. Conclusion Selective embolization of arterial aneurysms associated with ruptured AVMs is a safe procedure and could lower the immediate risk of rebleeding. Keywords Arteriovenous malformation . Aneurysm . Intracranial hemorrhage . Rebleeding . Embolization
Introduction Intracranial hemorrhage (ICH) is the most important factor affecting morbidity and mortality in patients with brain arteriovenous malformations (AVMs) [1–3]. Several authors have reported a stringent association between hemorrhagic AVM presentation, rebleeding, and presence of prenidal, intranidal, and postnidal aneurysms [4–20]. However, although plausible, the protective role of embolization of arterial aneurysms associated with ruptured brain AVMs is not demonstrated yet. In this article, we describe our treatment paradigm, conceived several years ago and maintained to date, that guided the treatment of patients presenting with ruptured AVMs associated with one or multiple arterial aneurysms considered to be the AVM’s weak point and deemed to require urgent occlusion. The objective of our study was to verify whether selective embolization of these aneurysms was safe and could be
764
proposed to induce a protective effect against AVM rebleeding.
Methods Study population From a prospectively collected database including all patients with AVMs admitted to our institution, we selected all patients treated for a ruptured cerebral AVM associated with one or more arterial aneurysms between December 2005 and March 2012. Intracranial hemorrhages related to the AVM were demonstrated on either computer tomography (CT) or magnetic resonance imaging (MRI). All AVMs were diagnosed based on preoperative four-vessel digital subtraction angiography (DSA). Arterial aneurysms were defined on imaging as saccular dilatations with a diameter equal or larger than the diameter of the parent vessel, thought to be a weak point in the angioarchitecture of the AVM. We took into consideration flow-related and dissecting aneurysms located on the arteries feeding the AVM [18] (types II and III of the Yu classification [10]) and intranidal aneurysms [18] (type IV of the Yu classification [10]), providing that they were contiguous with the hemorrhage, as seen on the preoperative imaging workup (Figs. 1 and 2). Unruptured brain AVMs and ruptured brain AVMs without arterial aneurysms or presenting with an aneurysm on the venous side or not related to the hemorrhage on the preoperative cerebral angiogram and CT and/or MR images and other types of vascular malformations such as dural fistulas or vein of Galen malformations were excluded from the analysis. Demographic, clinical, and morphological treatment and follow-up data not specified into the database were collected reviewing the clinical chart of each patient. Hemorrhagic stroke severity and clinical outcome after hemorrhagic presentation and during follow-up were measured using, respectively, the Glasgow Coma Scale (GCS) and the modified Rankin Scale (mRS). Treatment strategy for each case was discussed by a multidisciplinary team including interventional neuroradiologists, neurosurgeons, and neurointensivists and was aimed at occluding the arterial aneurysm as soon as possible and the AVM and the associated intracranial hypertension accordingly. All patients had an immediate postoperative cerebral CT scan and DSA, and all surviving patients were followed-up with at least 1-year postoperative clinical evaluation and mRS determination, brain CT and/or MR imaging, and cerebral DSA. Embolization procedure All endovascular procedures were performed under general anesthesia at a biplane angiography unit (Neurostar and Axiom Artis, Siemens, Erlangen, Germany). Vascular access was
Neuroradiology (2014) 56:763–769
obtained with a 6F guiding catheter (Envoy; Cordis, Miami Lakes, FL, USA). Superselective angiography was carried out with an over-the-wire microcatheter and with a DMSOcompatible detachable tip microcatheter (Sonic; Balt extrusion, Montmorency, France). After the microcatheter tip was placed at a perinidal position, superselective angiography was performed to confirm the aneurysms and its characteristics and relationships with the parent vessel. Aneurysms were embolized with N-butyl cyanoacrylate (n-BCA; Cordis Microvascular, Inc., New Brunswick, NJ), Onyx (Covidien), or bare platinum coils (Trufill from Cordis, Miami Lakes, Florida; target from Stryker Neurovascular). When deemed safe, the associated AVM was treated either completely or partially immediately thereafter.
Results In the reviewed period, 25 patients presented to our hospital with ruptured cerebral AVMs associated with an aneurysm as a suspected source of bleeding. They were 16 males and 9 females, with a mean age of 42.2 years (range 10–79 years) (see Table 1). Clinical presentation and type of intracranial bleeding The overwhelming majority of patients (21/25, 84 %) presented with sudden violent headache, which was associated with meningismus in 5. Six patients (24 %) had a focal or generalized seizure and 15 (60 %) displayed a fixed neurological deficit. The mean GCS was 11.7, with seven patients (28 %) being comatous (GCS≤8) and four (16 %) having a GCS of 3. The hemorrhage was both intraparenchimal and subarachnoid in 19 out of 25 cases (76 %). Of those, three (12 %) had an intraventricular extension. Three patients (12 %) presented with a pure intraparenchimal bleeding. A pure subarachnoid or a pure intraventricular bleeding occurred in one patient each (4 %), while one patient (4 %) had a subarachnoid bleeding with intraventricular extension. Angioarchitecture and site of aneurysm There were 3 Spetzler-Martin [21] grade 1, 11 grade 2, 7 grade 3, and 4 grade 4 AVMs. In all cases, the initial angiogram demonstrated an aneurysm contiguous to the intracranial hemorrhage, which was considered the source of bleeding (Figs. 1 and 2). A total of 31 aneurysms were visualized. The ruptured aneurysm was located inside the nidus in 15 cases (48.4 %) and on the arterial feeders in 16 cases (51.6 %) (Figs. 1 and 2), four of them (12.9 %) being ruptured flow-related aneurysms [18], type II of Yu classification [10] (Fig. 2). Six patients (24 %) displayed multiple ruptured aneurysms (Fig. 2): three
Neuroradiology (2014) 56:763–769
765
Fig. 1 A 45-year-old man with sudden headache, impaired consciousness, and cerebellar ataxia. a Contrast enhanced T1weighted MR imaging demonstrates a cerebellar hematoma with an aneurysm on an AVM feeder indicating the source of bleeding (arrow). b Lateral view of vertebral angiography shows a supracerebellar AVM with distal arterial feeder aneurysm as the source of hemorrhage (arrow). c Lateral view of microcatheter angiography during NCBA embolization of the superior cerebellar artery aneurysm (arrow). d Complete obliteration of the AVM after subsequent surgical removal
patients had two intranidal aneurysms each, two patients had two aneurysms on AVM arterial feeders, and one patient had one aneurysm on arterial feeder and one intranidal. Timing of treatment Two groups of patients were identified according to the mean time to endovascular treatment (EVT): for the first group of 17 patients, the mean time to EVT was 30.1 h, while for the second group of 8 patients, it was 17 days. EVT of the aneurysm was delayed for one patient in poor initial clinical condition and for seven patients admitted to our hospital between 7 and 28 days after bleeding. Two of those patients were admitted after AVM rebleeding respectively 1 and 4 weeks after the first hemorrhage. All six patients having multiple aneurysms underwent single-stage embolization of all lesions because, in all cases, the preoperative imaging workup showed no distinctive pattern of the hemorrhage neither aneurysms’ distinctive features of rupture allowing to identify with certainty the bleeding lesion (Fig. 2). Type of treatment and angiographic results Twenty-four (77.4 %) aneurysms were embolized with nBCA (Figs. 1 and 2) and four (12.9 %) with Onyx (Fig. 2), while coils were used for three (9.7 %) aneurysms that were localized proximally on arterial feeders (Fig. 2). All
aneurysms were completely excluded on postoperative angiograms. The goal of the treatment by glue or Onyx was to obtain a segmental occlusion of the vessel along with the aneurysm, the migration of the embolic liquid inside the sac being considered a proof of cure, but not a treatment requirement. A segmental occlusion was performed based on the assumption that cortical anastomoses will take care of the distal vascularization. A nonadhesive liquid embolic material such as Onyx was preferred to glue when complete or partial treatment of associated nidus was planned in the same setting as the embolization of ruptured aneurysm and the chance of jeopardizing vascularization of normal brain with a reflux of 15 to 20 mm was small. Seventeen patients (68 %) underwent EVT of the sole aneurysm. The other eight patients (32 %) underwent additional occlusion of the AVM in the same setting as the embolization of the aneurysms, which was complete in five (20 %). Treatment of residual AVM was completed by subsequent embolization in seven patients (28 %), with a mean delay of 4.2 months (range 1–10 months), once blood clot is resolved or patient has recovered from initial bleeding. Residual AVM was completely treated by embolization and surgery or radiosurgery in two (8 %) patients each, by solesurgery or radiosurgeryin four (16 %) and one patient (4 %), respectively, while four other patients (16 %), in poor clinical condition (GCS≤6) at admission, had only embolization of the aneurysm as treatment strategy (Table 2).
766
Neuroradiology (2014) 56:763–769 Table 1 Baseline characteristics of 25 patients with AVM-associated ruptured arterial aneurysms Age, mean (range)
42.2 years (10–79 years)
Sex Clinical presentation Headache Meningismus
16 males, 9 females
Seizures Fixed neurodeficit GCS 15–13 12–9 ≤8 Spetzler-Martin grade Grade I Grade II Grade III Grade IV Location of AVM Lobar Deep Cerebellar Brain Stem Arterial feeders Pial
Fig. 2 A 63-year-old woman with sudden headache and cerebellar ataxia. a CT scan shows a vermian hemorrhage. b Lateral view of vertebral angiography shows a supracerebellar AVM with two aneurysms (arrows) on the telovelotonsillary segment of L PICA as the source of hemorrhage. c Lateral view of vertebral angiography shows coil embolization of the distal aneurysm (arrow). d Lateral view of vertebral angiography shows subsequent NCBA occlusion of the proximal aneurysm (arrow) with segmental occlusion of the parent arterial branch. e Six months later, CT scan shows rebleeding of the AVM. f Lateral view of vertebral angiography shows a de novo aneurysm on the distal segment of L PICA as the source of rebleeding (arrow). g Nonsubtracted image of Onyx cast after embolization, with complete filling of the sac (short arrow) and occlusion of the parent branch. Note the coils (long arrow) occluding one of the two aneurysms source of the first bleeding. h Angiography demonstrating complete obliteration of the de novo aneurysm
Perforators Mixed ≤2 >2 Venous drainage Superficial Deep Mixed Associated aneurysms Proximal on arterial feeder Distal on arterial feeder Intranidal First hemorrhage site Intraparenchimal only Subarachnoid only Intraventricular only Mixed Rebleeding Before treatment After treatment
21 (84 %) 5 (20 %) 6 (24 %) 15 (60 %) 12 (48 %) 6 (24 %) 7 (28 %) 3 (12 %) 11 (44 %) 7 (28 %) 4 (16 %) 17 (68 %) 3 (12 %) 4 (16 %) 1 (4 %) 15 (60 %) 3 (12 %) 7 (28 %) 20 (80 %) 5 (20 %) 13 (52 %) 1 (4 %) 11 (44 %) 31 (100 %) 4 ( 12.9 %) 12 (38.7 %) 15 (48.4 %) 3 (12 %) 1 (4 %) 1 (4 %) 20 (80 %) 8 (32 %) 4 (16 %) 4 (16 %)
GCS Glasgow Coma Scale, AVM arteriovenous malformation
Neuroradiology (2014) 56:763–769
767
Table 2 Management categories and outcome in AVMs (n=25) after embolisation of associated ruptured aneurysms Type of management of AVM
Number of cases (%)
mRS 0–1
mRS 2
mRS 3–4
Death
Embolisation only
12 (48 %)
7 (28 %)
2 (8 %) due to: 2 gravity of 1st bleed
0
Surgery only
4 (16 %)
3 (12 %)
0
0
Radiosurgery only
1 (4 %)
0
0
0
Multimodality treatment
4 comprising: 2 embolisation and surgery 2 embolisation and RxS 4 comprising: 3 embolisation of aneurysm only 1 embolisation of aneurysm and partial embolisation of AVM
4
3 (12 %) due to: 1 EVT complication 2 gravity of 1st bleed 1 (4 %) due to: 1 AVM rebleeding 1 due to: 1 gravity of 1st bleed 0
0
0
0
0
0
4 due to: 3 gravity of 1st bleeding 1 AVM rebleeding
Conservative management
EVT endovascular treatment, RxS radiosurgery, mRS modified Rankin Scale, AVM arteriovenous malformation
Rebleeding During the mean prospective follow-up period of 56 months (range 2–108 months), eight rebleedings occurred in 116.7 prospective years of observation. Four patients (16 % of patients, 12.9 % of aneurysms) presented rebleeding of their AVM before treatment of their aneurysm with a mean interval of 9.6 days (1–28 days) (Fig. 2). All these four patients had emergent embolization of their aneurysm, within 24 h after rebleeding. For one patient, rebleeding did not lead to appreciable worsening of her condition, while in three other patients, it resulted in sudden impairment of consciousness and new neurological deficits related to enlargement of existing intracerebral hematoma. One patient held a postoperative left side hemiplegia and gradually improved to a mRS of 2 at 9-year follow-up. The other patient displayed aggravation of his ataxia postoperatively, that improved subsequently and had a mRS of 1 at 20-month follow-up. The last patient died as a consequence of rebleeding 2 months thereafter. Four other patients (16 % of patients, 12.9 % of aneurysms) had rebleeding of their residual AVM after treatment of the aneurysm (two patients) or partial embolization of the AVM (two patients) at 3, 5, 6, and 19 months (mean 8.25 months). The first three patients did not show any clinical worsening after rebleeding. In all of them, repeated DSA showed a “de novo” false aneurysm, intranidal in two cases and on the distal segment of an arterial feeder in one (Fig. 2), absent on DSA performed at the time of first bleed, that colocalized with the rebleeding. These lesions were
considered the weak point of the AVM and were treated in emergency respectively at 16, 24, and 36 h after rebleeding. The last patient died of rebleeding before repeating DSA. The four rebleedings after treatment in 116.7 prospective years of observation equaled to a 3.4 % annual rehemorrhage rate. Clinical and radiological follow-up One patient experienced immediate postoperative embolic occlusion of the distal left ACA, indirect feeder of the AVM, due to accidental migration of n-BCA into the artery, with worsening of his right lower limb paresis and aphasia that improved in 12 months to a mRS of 2. At a mean follow-up of 56 months, 21 patients (84 %) were alive. Nineteen patients (76 %) were independent, with mRS≤2. Of these, three presented with GCS
