Acta Neurochir DOI 10.1007/s00701-014-2225-3
CLINICAL ARTICLE - VASCULAR
Frequency, risk of hemorrhage and treatment considerations for cerebral arteriovenous malformations with associated aneurysms Johannes Platz & Joachim Berkefeld & Oliver C. Singer & Robert Wolff & Volker Seifert & Jürgen Konczalla & Erdem Güresir
Received: 16 July 2014 / Accepted: 1 September 2014 # Springer-Verlag Wien 2014
Abstract Background Data on arteriovenous malformations (AVMs) of the brain with AVM-associated aneurysms (AAA) are scarce. This study addresses the incidence, rate of hemorrhage, treatment strategies and stability during follow-up in a neurovascular center. Methods We retrospectively reviewed patients harboring an AVM with at least one AAA treated at our neurovascular center between 2002 and 2013. Results Of 216 patients, 59 (27.3 %) had at least one AAA (n=92 aneurysms total). Compared to patients without AAA, hemorrhagic presentation occurred more frequently (61.0 % versus 43.9 %, p=0.025), and the rate of infratentorial AVMs was higher (37.3 % versus 16.6 %, p=0.001). The aneurysm was the origin of the bleeding in most cases, most often Parts of this study were presented as an abstract at the annual meeting of the DGNC in 2013. Electronic supplementary material The online version of this article (doi:10.1007/s00701-014-2225-3) contains supplementary material, which is available to authorized users. J. Platz (*) : V. Seifert : J. Konczalla : E. Güresir Johann Wolfgang Goethe-University, Department of Neurosurgery, Schleusenweg 2-16 D, 60528 Frankfurt am Main, Germany e-mail: [email protected] J. Berkefeld Johann Wolfgang Goethe-University, Department of Neuroradiology, Frankfurt am Main, Germany O. C. Singer Johann Wolfgang Goethe-University, Department of Neurology, Frankfurt am Main, Germany R. Wolff Gamma Knife Center, Johann Wolfgang Goethe-University, Frankfurt am Main, Germany
categorized as a feeding artery aneurysm. Overall, the first and recurrent hemorrhage were associated with a high mortality and morbidity (15.3 % and 39 %, respectively). Aneurysms were treated by coiling (n=21), surgery (n=18), or embolizaton with liquid embolization agents (n=11). All aneurysms treated by embolization and surgery remained occluded during follow-up (mean follow-up 39.0±45.0 months). However, in incomplete AVM obliteration, significant recurrence of the treated aneurysm was noted after endovascular coiling (37.5 %), which may be related to the persistence of pathological blood flow. Conclusion In our series, AAA was a significant risk factor for hemorrhage and was associated with a poor outcome. It seems worthwhile to consider whether the aneurysm itself is a risk factor or only an epiphenomenon of severely altered hemodynamics induced by these special AVMs and therefore only the most common site of rupture. As the complication rate was low for aneurysm occlusion, we recommend treating these aneurysms whenever possible. Furthermore, obliteration of the AVM should be strived for as this subtype may be associated with an increased risk of hemorrhage.
Keywords Aneurysm treatment . AVM treatment . Cerebral arteriovenous malformation . Intracranial aneurysm . Intracranial hemorrhage
Abbreviations AAA AVM-associated aneurysm AVM Arteriovenous malformation DSA Digital subtraction angiography LEA Liquid embolic agents mRS Modified Rankin Scale
Acta Neurochir
Background Cerebral AVMs are rare congenital lesions associated with a risk of hemorrhage of 1 to more than 4 % per year [1–4]. This rate varies according to various features of the AVM such as the location, draining pattern, or history of previous hemorrhage. Since the publication of the long-anticipated randomized trial on unruptured brain AVMs (ARUBA), many issues and questions have arisen [5]. One of the most important issues may become the selection of AVM patients considered for treatment. Therefore, risk factors of AVMs need to be identified clearly. The presence of AVM-associated aneurysms (AAAs) is thought to influence the hemorrhage rate. AAAs can be classified as feeding artery aneurysms (type 1), nidal aneurysms (type 2), and aneurysms not associated with the AVM (type 3, Fig. 1) [6]. Type 1 aneurysms can be further divided into proximal (type 1a, aneurysm closer to the circle of Willis than to the AVM) and distal (type 1b, aneurysm closer to the AVM). Due to limitations and biases of case series, data concerning the bleeding risk of AAAs are inconsistent. Publications dealing with the question of whether AAAs present an independent risk for hemorrhage yielded contradictory results [7–20]. We reviewed our series of AVM patients to determine the frequency of associated aneurysms, treatment strategies and clinical results.
Patients and methods We retrospectively identified all patients with cerebral AVMs who presented at our center between February 2002 and March 2013. This study was approved by the local ethics committee. Initially, each case was evaluated on the basis of digital subtraction angiography (DSA) by a neurovascular team consisting of neurosurgeons, neuroradiologists, neurologists and radiosurgeons. DSA consisted of a four-vessel angiogram in all patients, which was carefully reviewed for the presence of AAA. Only in selected patients with a poor clinical status after hemorrhage were CT angiography or MR angiography accepted as sufficient if the AVM and aneurysm could be clearly identified. Patients with at least one type 1 or type 3 aneurysm were included. AAAs were defined as arterial saccular dilations of the lumen with at least the double diameter of the vessel carrying it. Cases with insufficient workup or with infundibula, arterial ectasias and venous aneurysmal ectasias were excluded. In our institution, we do not aggressively search for intranidal (type 2) AAAs, which may easily be missed without selective catheterization of the different feeding arteries. Therefore, those AAAs were not inclusion criteria. According to the patients’ records and imaging findings, we determined the rate of ruptured and unruptured aneurysms. Hemorrhage was defined as a symptomatic event with blood
on imaging or in the cerebrospinal fluid. To define the origin of the hemorrhage, multimodal imaging was used (including multiplanar reconstructions and angiography techniques in CT, DSA or MRI) as well as analysis of the blood distribution, anatomic relationship of the AVM and the aneurysm(s) to the hemorrhage and, if available, intraoperative findings. In an interdisciplinary consensus, the origin of the hemorrhage could be distinguished as either aneurysmal or AVM-related in most cases (Fig. 2). Five patients were partially treated at other institutions: one surgical aneurysm occlusion, one partial aneurysm embolization, one partial aneurysm and AVM embolization (all because of hemorrhage prior to presentation to our center), one surgical AVM excision and one combined treatment. Aneurysm treatment The indication for aneurysm treatment was determined by interdisciplinary consensus between the neurovascular specialists based on the patient’s clinical condition and imaging findings. The best treatment option was chosen interdisciplinarily between endovascular and microneurosurgical techniques. In ruptured aneurysms, we usually focused on the aneurysm without striving for definite treatment of the AVM, aiming for aneurysm occlusion within 24 h. In unruptured aneurysms, we developed an individualized treatment strategy independently from the AVM assuming that the risk of AAA rupture might be at least as high as in the ISUIA study [21]: in case of surgical excision of the AVM, the aneurysm was treated in the same procedure when technically feasible. For all others, more remote AAAs, the treatment was planned individually, bearing in mind that there are reports of aneurysm shrinkage after successful AVM treatment [6, 20, 22–24]. Endovascular aneurysm treatment Endovascular treatment was performed under general anesthesia after administration of 5,000 IU of heparin and included aneurysm coiling or embolization with liquid embolization agents (LEA), sometimes including parent artery occlusion. Various types of platinum coils were used during the treatment period as well as different LEAs (Onyx and Histoacryl). Aneurysms were packed with coils as densely as possible. In selected patients, the balloon remodeling technique was applied and/or an endovascular stent was placed. Heparinization was routinely continued for 24 to 48 h afterwards (goal: activated clotting time 1.5× normal). Microsurgical aneurysm treatment Clipping of the aneurysm was the primary goal using standard microneurosurgical methods [25, 26]. For confirmation of
Acta Neurochir Fig. 1 Types of aneurysms associated with AVMs. (a) Type 1a: “Proximal feeding artery aneurysm,” meaning that the aneurysm is located closer to the circle of Willis on the feeding artery than to the AVM nidus. (b) Type 1b: “Distal feeding artery aneurysm,” meaning that the aneurysm is located closer to the AVM nidus than to the circle of Willis on the feeding artery (white arrow). Note the additional type 1a aneurysm in this patient (black arrow). (c) Type 1b: Multiple aneurysms. (d) Type 3: Aneurysm not associated with the AVM (white arrow). In this case, the AVM was supplied by vessels of the MCA and ACA only; therefore, the ruptured basilar tip aneurysm was categorized as type 3
complete aneurysm occlusion, all patients underwent DSA about 1 week after surgery. AVM treatment AVM treatment was always planned by the interdisciplinary neurovascular team. After AVM rupture, treatment of the AVM was deferred for 4 to 6 weeks if there was no lifethreatening condition. AVM treatment was chosen among neurosurgical excision with or without preoperative partial Fig. 2 Determining the origin of hemorrhage. Using multidirectional reconstruction, the origin of the hemorrhage could be detected in most cases as shown here: In this patient, clearly the aneurysm (white arrow) is the source of the intraventricular hemorrhage (*), while the nidus of the AVM is located remotely (black arrow)
embolization, embolization alone or radiosurgery (see Supplemental Figure 2 for radiosurgery details). In case of AVM rupture, radiosurgery was only chosen if the location or angioarchitecture of the AVM prevented definitive neurosurgical or endovascular AVM treatment. Follow-up Follow-up was obtained during outpatient visits using the modified Rankin cale (mRS). Alternatively, a telephone interview
Acta Neurochir
was performed. Outcome was dichotomized into favorable (mRS 0 to 2) and poor (mRS 3 to 6). Neurologic deficits were classified as hemorrhage and/or treatment related. As occlusion of all AAAs and AVMs could not always be achieved safely, patients with selective treatment of AAAs and/or the AVM were included. In case of incomplete treatment, serial MR imaging studies were acquired using intervals of 1 to 2 years.
Overall, 39 aneurysm-related hemorrhages were documented in this series: additionally to the 33 initial ruptures, 6 delayed ruptures occurred (Fig. 3). Ruptured aneurysms were larger in diameter (6.25±4.2 vs. 4.17±2.6 mm, p
CLINICAL ARTICLE - VASCULAR
Frequency, risk of hemorrhage and treatment considerations for cerebral arteriovenous malformations with associated aneurysms Johannes Platz & Joachim Berkefeld & Oliver C. Singer & Robert Wolff & Volker Seifert & Jürgen Konczalla & Erdem Güresir
Received: 16 July 2014 / Accepted: 1 September 2014 # Springer-Verlag Wien 2014
Abstract Background Data on arteriovenous malformations (AVMs) of the brain with AVM-associated aneurysms (AAA) are scarce. This study addresses the incidence, rate of hemorrhage, treatment strategies and stability during follow-up in a neurovascular center. Methods We retrospectively reviewed patients harboring an AVM with at least one AAA treated at our neurovascular center between 2002 and 2013. Results Of 216 patients, 59 (27.3 %) had at least one AAA (n=92 aneurysms total). Compared to patients without AAA, hemorrhagic presentation occurred more frequently (61.0 % versus 43.9 %, p=0.025), and the rate of infratentorial AVMs was higher (37.3 % versus 16.6 %, p=0.001). The aneurysm was the origin of the bleeding in most cases, most often Parts of this study were presented as an abstract at the annual meeting of the DGNC in 2013. Electronic supplementary material The online version of this article (doi:10.1007/s00701-014-2225-3) contains supplementary material, which is available to authorized users. J. Platz (*) : V. Seifert : J. Konczalla : E. Güresir Johann Wolfgang Goethe-University, Department of Neurosurgery, Schleusenweg 2-16 D, 60528 Frankfurt am Main, Germany e-mail: [email protected] J. Berkefeld Johann Wolfgang Goethe-University, Department of Neuroradiology, Frankfurt am Main, Germany O. C. Singer Johann Wolfgang Goethe-University, Department of Neurology, Frankfurt am Main, Germany R. Wolff Gamma Knife Center, Johann Wolfgang Goethe-University, Frankfurt am Main, Germany
categorized as a feeding artery aneurysm. Overall, the first and recurrent hemorrhage were associated with a high mortality and morbidity (15.3 % and 39 %, respectively). Aneurysms were treated by coiling (n=21), surgery (n=18), or embolizaton with liquid embolization agents (n=11). All aneurysms treated by embolization and surgery remained occluded during follow-up (mean follow-up 39.0±45.0 months). However, in incomplete AVM obliteration, significant recurrence of the treated aneurysm was noted after endovascular coiling (37.5 %), which may be related to the persistence of pathological blood flow. Conclusion In our series, AAA was a significant risk factor for hemorrhage and was associated with a poor outcome. It seems worthwhile to consider whether the aneurysm itself is a risk factor or only an epiphenomenon of severely altered hemodynamics induced by these special AVMs and therefore only the most common site of rupture. As the complication rate was low for aneurysm occlusion, we recommend treating these aneurysms whenever possible. Furthermore, obliteration of the AVM should be strived for as this subtype may be associated with an increased risk of hemorrhage.
Keywords Aneurysm treatment . AVM treatment . Cerebral arteriovenous malformation . Intracranial aneurysm . Intracranial hemorrhage
Abbreviations AAA AVM-associated aneurysm AVM Arteriovenous malformation DSA Digital subtraction angiography LEA Liquid embolic agents mRS Modified Rankin Scale
Acta Neurochir
Background Cerebral AVMs are rare congenital lesions associated with a risk of hemorrhage of 1 to more than 4 % per year [1–4]. This rate varies according to various features of the AVM such as the location, draining pattern, or history of previous hemorrhage. Since the publication of the long-anticipated randomized trial on unruptured brain AVMs (ARUBA), many issues and questions have arisen [5]. One of the most important issues may become the selection of AVM patients considered for treatment. Therefore, risk factors of AVMs need to be identified clearly. The presence of AVM-associated aneurysms (AAAs) is thought to influence the hemorrhage rate. AAAs can be classified as feeding artery aneurysms (type 1), nidal aneurysms (type 2), and aneurysms not associated with the AVM (type 3, Fig. 1) [6]. Type 1 aneurysms can be further divided into proximal (type 1a, aneurysm closer to the circle of Willis than to the AVM) and distal (type 1b, aneurysm closer to the AVM). Due to limitations and biases of case series, data concerning the bleeding risk of AAAs are inconsistent. Publications dealing with the question of whether AAAs present an independent risk for hemorrhage yielded contradictory results [7–20]. We reviewed our series of AVM patients to determine the frequency of associated aneurysms, treatment strategies and clinical results.
Patients and methods We retrospectively identified all patients with cerebral AVMs who presented at our center between February 2002 and March 2013. This study was approved by the local ethics committee. Initially, each case was evaluated on the basis of digital subtraction angiography (DSA) by a neurovascular team consisting of neurosurgeons, neuroradiologists, neurologists and radiosurgeons. DSA consisted of a four-vessel angiogram in all patients, which was carefully reviewed for the presence of AAA. Only in selected patients with a poor clinical status after hemorrhage were CT angiography or MR angiography accepted as sufficient if the AVM and aneurysm could be clearly identified. Patients with at least one type 1 or type 3 aneurysm were included. AAAs were defined as arterial saccular dilations of the lumen with at least the double diameter of the vessel carrying it. Cases with insufficient workup or with infundibula, arterial ectasias and venous aneurysmal ectasias were excluded. In our institution, we do not aggressively search for intranidal (type 2) AAAs, which may easily be missed without selective catheterization of the different feeding arteries. Therefore, those AAAs were not inclusion criteria. According to the patients’ records and imaging findings, we determined the rate of ruptured and unruptured aneurysms. Hemorrhage was defined as a symptomatic event with blood
on imaging or in the cerebrospinal fluid. To define the origin of the hemorrhage, multimodal imaging was used (including multiplanar reconstructions and angiography techniques in CT, DSA or MRI) as well as analysis of the blood distribution, anatomic relationship of the AVM and the aneurysm(s) to the hemorrhage and, if available, intraoperative findings. In an interdisciplinary consensus, the origin of the hemorrhage could be distinguished as either aneurysmal or AVM-related in most cases (Fig. 2). Five patients were partially treated at other institutions: one surgical aneurysm occlusion, one partial aneurysm embolization, one partial aneurysm and AVM embolization (all because of hemorrhage prior to presentation to our center), one surgical AVM excision and one combined treatment. Aneurysm treatment The indication for aneurysm treatment was determined by interdisciplinary consensus between the neurovascular specialists based on the patient’s clinical condition and imaging findings. The best treatment option was chosen interdisciplinarily between endovascular and microneurosurgical techniques. In ruptured aneurysms, we usually focused on the aneurysm without striving for definite treatment of the AVM, aiming for aneurysm occlusion within 24 h. In unruptured aneurysms, we developed an individualized treatment strategy independently from the AVM assuming that the risk of AAA rupture might be at least as high as in the ISUIA study [21]: in case of surgical excision of the AVM, the aneurysm was treated in the same procedure when technically feasible. For all others, more remote AAAs, the treatment was planned individually, bearing in mind that there are reports of aneurysm shrinkage after successful AVM treatment [6, 20, 22–24]. Endovascular aneurysm treatment Endovascular treatment was performed under general anesthesia after administration of 5,000 IU of heparin and included aneurysm coiling or embolization with liquid embolization agents (LEA), sometimes including parent artery occlusion. Various types of platinum coils were used during the treatment period as well as different LEAs (Onyx and Histoacryl). Aneurysms were packed with coils as densely as possible. In selected patients, the balloon remodeling technique was applied and/or an endovascular stent was placed. Heparinization was routinely continued for 24 to 48 h afterwards (goal: activated clotting time 1.5× normal). Microsurgical aneurysm treatment Clipping of the aneurysm was the primary goal using standard microneurosurgical methods [25, 26]. For confirmation of
Acta Neurochir Fig. 1 Types of aneurysms associated with AVMs. (a) Type 1a: “Proximal feeding artery aneurysm,” meaning that the aneurysm is located closer to the circle of Willis on the feeding artery than to the AVM nidus. (b) Type 1b: “Distal feeding artery aneurysm,” meaning that the aneurysm is located closer to the AVM nidus than to the circle of Willis on the feeding artery (white arrow). Note the additional type 1a aneurysm in this patient (black arrow). (c) Type 1b: Multiple aneurysms. (d) Type 3: Aneurysm not associated with the AVM (white arrow). In this case, the AVM was supplied by vessels of the MCA and ACA only; therefore, the ruptured basilar tip aneurysm was categorized as type 3
complete aneurysm occlusion, all patients underwent DSA about 1 week after surgery. AVM treatment AVM treatment was always planned by the interdisciplinary neurovascular team. After AVM rupture, treatment of the AVM was deferred for 4 to 6 weeks if there was no lifethreatening condition. AVM treatment was chosen among neurosurgical excision with or without preoperative partial Fig. 2 Determining the origin of hemorrhage. Using multidirectional reconstruction, the origin of the hemorrhage could be detected in most cases as shown here: In this patient, clearly the aneurysm (white arrow) is the source of the intraventricular hemorrhage (*), while the nidus of the AVM is located remotely (black arrow)
embolization, embolization alone or radiosurgery (see Supplemental Figure 2 for radiosurgery details). In case of AVM rupture, radiosurgery was only chosen if the location or angioarchitecture of the AVM prevented definitive neurosurgical or endovascular AVM treatment. Follow-up Follow-up was obtained during outpatient visits using the modified Rankin cale (mRS). Alternatively, a telephone interview
Acta Neurochir
was performed. Outcome was dichotomized into favorable (mRS 0 to 2) and poor (mRS 3 to 6). Neurologic deficits were classified as hemorrhage and/or treatment related. As occlusion of all AAAs and AVMs could not always be achieved safely, patients with selective treatment of AAAs and/or the AVM were included. In case of incomplete treatment, serial MR imaging studies were acquired using intervals of 1 to 2 years.
Overall, 39 aneurysm-related hemorrhages were documented in this series: additionally to the 33 initial ruptures, 6 delayed ruptures occurred (Fig. 3). Ruptured aneurysms were larger in diameter (6.25±4.2 vs. 4.17±2.6 mm, p
