- Email: [email protected]
Spontaneous resolution of an idiopathic cervical direct vertebral arteriovenous fistula after partial coil embolization in a patient presenting with myeloradiculopathy
Clinical Neurology and Neurosurgery 104 (2002) 146– 151 www.elsevier.com/locate/clineuro
Case report
Spontaneous resolution of an idiopathic cervical direct vertebral arteriovenous fistula after partial coil embolization in a patient presenting with myeloradiculopathy Jun-ichiro Asai a,*, Takaki Hayashi b, Ryuta Suzuki a, Tsukasa Fujimoto a, Goro Nagashima a a
Department of Neurosurgery, Fujigaoka Hospital, School of Medicine, Showa Uni6ersity, 1 -30 Fujigaoka, Aoba-ku, Yokohama, Kanagawa 227 -8501, Japan b Department of Radiation, Fujigaoka Hospital, School of Medicine, Showa Uni6ersity, 1 -30 Fujigaoka, Aoba-ku, Yokohama, Kanagawa 227 -8501, Japan Received 21 May 2001; received in revised form 1 November 2001; accepted 23 November 2001
Abstract A 53-year old female presented with paresis of the left upper extremity. Magnetic resonance imaging (MRI) and magnetic resonance angiography (MRA) disclosed a single high-flow vertebral arteriovenous fistula (AVF) with vertebral artery (VA) transection. The AVF was also fed by steal flow from the contralateral VA. The left posterior inferior cerebellar artery (PICA) branched just distal to the fistula. The fistula drained into the neighboring paravertebral veins and refluxed into the intradural venous systems. The dilated drainers compressed the spinal cord. Embolization was attempted at the drainer just behind the fistula orifice using platinum coils. The fistula was still fed slightly by right VA after the embolization, but spontaneous complete obliteration was achieved after one week. The clinical symptoms and signs disappeared. Although, detachable balloon embolization is the quickest and most effective procedure to obliterate a fistula, stepwise embolization using GDC can be considered, and may avoid the normal pressure perfusion break-through phenomenon. Spontaneous obliteration of the fistula after partial embolization in our case may result from intravenous embolization just behind the fistula orifice. It may therefore be a useful approach to the embolization of an AVF to begin the embolization at the venous side of the fistula. © 2002 Elsevier Science B.V. All rights reserved. Keywords: AVF; Spinal disease; Cervical spine; Embolization; Platinum coil
1. Introduction
2. Case report
Cervical direct vertebral arteriovenous fistula (AVF) without dural components is a different clinical entity from dural AVF and rare vascular disease [1 – 3]. This report discusses a treatment for vertebral AVF in the context of a case with spontaneous occlusion of a single high-flow vertebral AVF after platinum coil embolization.
A 53-year old female suffered from paresis of her left upper extremity. Magnetic resonance imaging (MRI), magnetic resonance angiography (MRA), and computer tomography (CT) myelography revealed a vascular lesion spreading over the intradural extramedullary, extradural, and paravertebral regions. No abnormal intensity was detected in the spinal cord, but the dilated vessels compressed the spinal cord (Fig. 1). Left vertebral angiography (VAG) disclosed a single, direct, and very high-flow AVF at C1-2 with vertebral artery (VA) transection. The AVF was also fed by steal flow from
* Corresponding author. Tel.: +81-45-971-1151; fax: + 81-45-9444701. E-mail address: [email protected] (J.-i. Asai).
0303-8467/02/$ - see front matter © 2002 Elsevier Science B.V. All rights reserved. PII: S 0 3 0 3 - 8 4 6 7 ( 0 1 ) 0 0 1 9 4 - 9
J.-i. Asai et al. / Clinical Neurology and Neurosurgery 104 (2002) 146–151
147
the right VA and ascending cervical arteries. The left posterior inferior cerebellar artery (PICA) branched just distal to the fistula. The fistula drained into the neighboring paravetebral veins and refluxed into the radiculomedullary and anterior median veins, all of which were significantly dilated around the spinal canal (Fig. 2). A microcatheter was inserted via the left VA until it reached the drainer just behind the orifice of the fistula. Embolization was attempted at this site using Guglielmi detachable coils (GDCs) (Boston Scientific, Natik, MA, USA), Vortex coils (Boston Scientific, Natik, MA, USA), and Diamond coils (Boston Scientific, Natik, MA, USA). The fistula and left VA were finally occluded, but the fistula was still fed slightly by backward flow from right VA after the embolization (Fig. 3). One week later, spontaneous complete obliteration of the fistula with the preservation of the left PICA was achieved (Fig. 4), and the patient’s symptoms disappeared.
3. Conclusions
Fig. 1. (a and b) Axial MRI (TE 22 ms, TR 600 ms, FOV 16*18, thickness 5 mm, 256*192 Matrix, 1 NEX) at C1 –C2. (c and d) Sagittal MRI (TE 25 ms, TR 500 ms, FOV 24*12, thickness 4 mm, 256*254 Matrix, 4 NEX). (e and f) Cervical MRA (2D-fast gradient echo, TE 1.6 ms, TR 6.4 ms, FOV 32*24, thickness 30 mm, 256*160/1 NEX). (g and h) CT myelogram. A vascular lesion spreading over the spinal canal and the paravertebral soft tissue was revealed (arrow). The lesion compressed the spinal cord. MRA disclosed a large vascular lesion around the spinal cord.
The paresis of the left arm in this case was attributed to the compression of the intradural nerve roots, spinal cord, and/or brachial plexus by the intradural and extradural dilated and congestive drainers. The origin of the spinal cord myelopathy from spinal venous congestion and/or hypertension was not theoretically ruled out, but there was at least no abnormal signal on spinal cord MRI. Endovascular treatment is now the simplest, quickest, most reliable, non-invasive, and cost-effective method of treating vertebral fistulas. Beauieux et al. and Halbach et al. reported in their large series that their best treatment results for vertebral fistulas were achieved using detachable balloons [2,4]. Coil embolization through the ipsilateral and/or contralateral VA has also been described [2,3,5,6]. We used GDC to embolize the fistula in our case, because obtaining the detachable balloon system in Japan is difficult at this time. Abrupt closure of a fistula may give rise to the ‘normal perfusion break through’ characterized by postoperative swelling or hemorrhage, which is thought to be due to a loss of autoregulation of cerebral blood flow surrounding the AVM [2,7]. We believe that if a fistula is fed only by the ipsilateral VA with no transection, showing antegrade flow into distal VA (Fig. 5a), the antegrade approach should be selected and only the fistula should be embolized. If the VA needs to be occluded, occlusion of only the proximal VA should be avoided, and the fistula flow should be decreased in comparison with the preoperative stage to avoid increasing the steal flow. When the fistula is fed by the ipsilateral VA and by
148
J.-i. Asai et al. / Clinical Neurology and Neurosurgery 104 (2002) 146–151
Fig. 2. Left VAG (a and b: Anterior – posterior (A –P) view; c and d: lateral view) revealed the direct high-flow AVF at the level of the atlas with a transection of the left VA. Right VAG (e and f: A –P view; g and h: lateral view; and i and j: right anterior oblique (RAO) view) disclosed that the AVF was also fed by backward flow from the right VA. The left PICA branched just distal to the fistula (arrow). The fistula drained into the neighboring paravertebral veins and refluxed into the radiculomedullary and anterior median veins (arrowhead), all of which were significantly dilated around the spinal canal. The AVF was also fed by the left ascending cervical artery (k).
backward steal flow from the contralateral VA with transection, that is showing no antegrade flow into VA distal to the fistula (Fig. 5b), and if the microcatheter can reach the VA distal to the fistula, then the embolization should begin at the fistula and gradually occlude the distal VA. If the distal VA cannot be accessed via the antegrade route, the fistula should be approached from both the antegrade and retrograde directions, and gradual or stepwise occlusion of the fistula and/or of the VA distal and the proximal to the fistula should be performed.
In our case, spontaneous occlusion of the fistula was thought to be the result of the GDC embolization that was begun at the venous side of the fistula. It is therefore a useful technique to begin the embolization of an AVF from the venous side just behind the fistula.
Acknowledgements The authors thank Kay Asai for the preparation of the manuscript and references.
J.-i. Asai et al. / Clinical Neurology and Neurosurgery 104 (2002) 146–151
149
Fig. 3. The fistula flow from the left VA was completely closed (a: lateral view; and b: A – P view), but a slight fistula flow from the right VA was still detected (c and d: A –P view; e and f: RAO view).
150
J.-i. Asai et al. / Clinical Neurology and Neurosurgery 104 (2002) 146–151
Fig. 4. One week after the embolization, the fistula was entirely obliterated (a and b: A – P view; c and d: RAO view), preserving the left PICA (arrow).
Fig. 5. Schematic drawing of the hemodynamics in VA –AVF with or without VA transection. Arrows indicate the directions of blood flow. LVA: left vertebral artery; RVA: right vertebral artery. (a) Without VA transection, the ipsilateral VA feeds the fistula and the distal VA; that is, the fistula is fed only by the ipsilateral VA. (b) With VA transection, the ipsilateral VA feeds only the fistula with no antegrade filling of the distal VA, and the fistula is also fed by backward flow from the contralateral VA.
References [1] Nagashima C, Iwasaki T, Kawanuma S, Sakaguchi A, Kamisasa A, Suzuki K. Traumatic arteriovenous fistula of the vertebral artery with spinal cord symptoms. Case report. J Neurosurg
1977;46:681 – 7. [2] Halbach VV, Higashida RT, Hieshima GB. Treatment of vertebral arteriovenous fistulas. AJR Am J Roentgenol 1988;150:405 – 12. [3] Yoshida S, Nakazawa K, Oda Y. Spontaneous vertebral arteri-
J.-i. Asai et al. / Clinical Neurology and Neurosurgery 104 (2002) 146–151 ovenous fistula—case report. Neurol Med Chir (Tokyo) 2000;40:211 – 5. [4] Beaujeux RL, Reizine DC, Casasco A, Aymard A, Rufenacht D, Khayata MH, Riche MC, Merland JJ. Endovascular treatment of vertebral arteriovenous fistula. Radiology 1992;183:361 – 7. [5] Roper PR, Guinto FC Jr., Wolma FJ. Posttraumatic vertebral artery aneurysm and arteriovenous fistula: a case report.
151
Surgery 1984;96:556 – 9. [6] Freitag HJ, Grzyska U, Zeumer H. The use of the ‘cross over technique’ in the management of a traumatic vertebro –vertebral fistula. Neuroradiology 1989;31:174 – 6. [7] Young WL, Kader A, Prohovnik I, Ornstein E, Fleischer LH, Ostapkovich N, Jackson LD, Stein BM. Pressure autoregulation is intact after arteriovenous malformation resection. Neurosurgery 1993;32:491 – 7.
Case report
Spontaneous resolution of an idiopathic cervical direct vertebral arteriovenous fistula after partial coil embolization in a patient presenting with myeloradiculopathy Jun-ichiro Asai a,*, Takaki Hayashi b, Ryuta Suzuki a, Tsukasa Fujimoto a, Goro Nagashima a a
Department of Neurosurgery, Fujigaoka Hospital, School of Medicine, Showa Uni6ersity, 1 -30 Fujigaoka, Aoba-ku, Yokohama, Kanagawa 227 -8501, Japan b Department of Radiation, Fujigaoka Hospital, School of Medicine, Showa Uni6ersity, 1 -30 Fujigaoka, Aoba-ku, Yokohama, Kanagawa 227 -8501, Japan Received 21 May 2001; received in revised form 1 November 2001; accepted 23 November 2001
Abstract A 53-year old female presented with paresis of the left upper extremity. Magnetic resonance imaging (MRI) and magnetic resonance angiography (MRA) disclosed a single high-flow vertebral arteriovenous fistula (AVF) with vertebral artery (VA) transection. The AVF was also fed by steal flow from the contralateral VA. The left posterior inferior cerebellar artery (PICA) branched just distal to the fistula. The fistula drained into the neighboring paravertebral veins and refluxed into the intradural venous systems. The dilated drainers compressed the spinal cord. Embolization was attempted at the drainer just behind the fistula orifice using platinum coils. The fistula was still fed slightly by right VA after the embolization, but spontaneous complete obliteration was achieved after one week. The clinical symptoms and signs disappeared. Although, detachable balloon embolization is the quickest and most effective procedure to obliterate a fistula, stepwise embolization using GDC can be considered, and may avoid the normal pressure perfusion break-through phenomenon. Spontaneous obliteration of the fistula after partial embolization in our case may result from intravenous embolization just behind the fistula orifice. It may therefore be a useful approach to the embolization of an AVF to begin the embolization at the venous side of the fistula. © 2002 Elsevier Science B.V. All rights reserved. Keywords: AVF; Spinal disease; Cervical spine; Embolization; Platinum coil
1. Introduction
2. Case report
Cervical direct vertebral arteriovenous fistula (AVF) without dural components is a different clinical entity from dural AVF and rare vascular disease [1 – 3]. This report discusses a treatment for vertebral AVF in the context of a case with spontaneous occlusion of a single high-flow vertebral AVF after platinum coil embolization.
A 53-year old female suffered from paresis of her left upper extremity. Magnetic resonance imaging (MRI), magnetic resonance angiography (MRA), and computer tomography (CT) myelography revealed a vascular lesion spreading over the intradural extramedullary, extradural, and paravertebral regions. No abnormal intensity was detected in the spinal cord, but the dilated vessels compressed the spinal cord (Fig. 1). Left vertebral angiography (VAG) disclosed a single, direct, and very high-flow AVF at C1-2 with vertebral artery (VA) transection. The AVF was also fed by steal flow from
* Corresponding author. Tel.: +81-45-971-1151; fax: + 81-45-9444701. E-mail address: [email protected] (J.-i. Asai).
0303-8467/02/$ - see front matter © 2002 Elsevier Science B.V. All rights reserved. PII: S 0 3 0 3 - 8 4 6 7 ( 0 1 ) 0 0 1 9 4 - 9
J.-i. Asai et al. / Clinical Neurology and Neurosurgery 104 (2002) 146–151
147
the right VA and ascending cervical arteries. The left posterior inferior cerebellar artery (PICA) branched just distal to the fistula. The fistula drained into the neighboring paravetebral veins and refluxed into the radiculomedullary and anterior median veins, all of which were significantly dilated around the spinal canal (Fig. 2). A microcatheter was inserted via the left VA until it reached the drainer just behind the orifice of the fistula. Embolization was attempted at this site using Guglielmi detachable coils (GDCs) (Boston Scientific, Natik, MA, USA), Vortex coils (Boston Scientific, Natik, MA, USA), and Diamond coils (Boston Scientific, Natik, MA, USA). The fistula and left VA were finally occluded, but the fistula was still fed slightly by backward flow from right VA after the embolization (Fig. 3). One week later, spontaneous complete obliteration of the fistula with the preservation of the left PICA was achieved (Fig. 4), and the patient’s symptoms disappeared.
3. Conclusions
Fig. 1. (a and b) Axial MRI (TE 22 ms, TR 600 ms, FOV 16*18, thickness 5 mm, 256*192 Matrix, 1 NEX) at C1 –C2. (c and d) Sagittal MRI (TE 25 ms, TR 500 ms, FOV 24*12, thickness 4 mm, 256*254 Matrix, 4 NEX). (e and f) Cervical MRA (2D-fast gradient echo, TE 1.6 ms, TR 6.4 ms, FOV 32*24, thickness 30 mm, 256*160/1 NEX). (g and h) CT myelogram. A vascular lesion spreading over the spinal canal and the paravertebral soft tissue was revealed (arrow). The lesion compressed the spinal cord. MRA disclosed a large vascular lesion around the spinal cord.
The paresis of the left arm in this case was attributed to the compression of the intradural nerve roots, spinal cord, and/or brachial plexus by the intradural and extradural dilated and congestive drainers. The origin of the spinal cord myelopathy from spinal venous congestion and/or hypertension was not theoretically ruled out, but there was at least no abnormal signal on spinal cord MRI. Endovascular treatment is now the simplest, quickest, most reliable, non-invasive, and cost-effective method of treating vertebral fistulas. Beauieux et al. and Halbach et al. reported in their large series that their best treatment results for vertebral fistulas were achieved using detachable balloons [2,4]. Coil embolization through the ipsilateral and/or contralateral VA has also been described [2,3,5,6]. We used GDC to embolize the fistula in our case, because obtaining the detachable balloon system in Japan is difficult at this time. Abrupt closure of a fistula may give rise to the ‘normal perfusion break through’ characterized by postoperative swelling or hemorrhage, which is thought to be due to a loss of autoregulation of cerebral blood flow surrounding the AVM [2,7]. We believe that if a fistula is fed only by the ipsilateral VA with no transection, showing antegrade flow into distal VA (Fig. 5a), the antegrade approach should be selected and only the fistula should be embolized. If the VA needs to be occluded, occlusion of only the proximal VA should be avoided, and the fistula flow should be decreased in comparison with the preoperative stage to avoid increasing the steal flow. When the fistula is fed by the ipsilateral VA and by
148
J.-i. Asai et al. / Clinical Neurology and Neurosurgery 104 (2002) 146–151
Fig. 2. Left VAG (a and b: Anterior – posterior (A –P) view; c and d: lateral view) revealed the direct high-flow AVF at the level of the atlas with a transection of the left VA. Right VAG (e and f: A –P view; g and h: lateral view; and i and j: right anterior oblique (RAO) view) disclosed that the AVF was also fed by backward flow from the right VA. The left PICA branched just distal to the fistula (arrow). The fistula drained into the neighboring paravertebral veins and refluxed into the radiculomedullary and anterior median veins (arrowhead), all of which were significantly dilated around the spinal canal. The AVF was also fed by the left ascending cervical artery (k).
backward steal flow from the contralateral VA with transection, that is showing no antegrade flow into VA distal to the fistula (Fig. 5b), and if the microcatheter can reach the VA distal to the fistula, then the embolization should begin at the fistula and gradually occlude the distal VA. If the distal VA cannot be accessed via the antegrade route, the fistula should be approached from both the antegrade and retrograde directions, and gradual or stepwise occlusion of the fistula and/or of the VA distal and the proximal to the fistula should be performed.
In our case, spontaneous occlusion of the fistula was thought to be the result of the GDC embolization that was begun at the venous side of the fistula. It is therefore a useful technique to begin the embolization of an AVF from the venous side just behind the fistula.
Acknowledgements The authors thank Kay Asai for the preparation of the manuscript and references.
J.-i. Asai et al. / Clinical Neurology and Neurosurgery 104 (2002) 146–151
149
Fig. 3. The fistula flow from the left VA was completely closed (a: lateral view; and b: A – P view), but a slight fistula flow from the right VA was still detected (c and d: A –P view; e and f: RAO view).
150
J.-i. Asai et al. / Clinical Neurology and Neurosurgery 104 (2002) 146–151
Fig. 4. One week after the embolization, the fistula was entirely obliterated (a and b: A – P view; c and d: RAO view), preserving the left PICA (arrow).
Fig. 5. Schematic drawing of the hemodynamics in VA –AVF with or without VA transection. Arrows indicate the directions of blood flow. LVA: left vertebral artery; RVA: right vertebral artery. (a) Without VA transection, the ipsilateral VA feeds the fistula and the distal VA; that is, the fistula is fed only by the ipsilateral VA. (b) With VA transection, the ipsilateral VA feeds only the fistula with no antegrade filling of the distal VA, and the fistula is also fed by backward flow from the contralateral VA.
References [1] Nagashima C, Iwasaki T, Kawanuma S, Sakaguchi A, Kamisasa A, Suzuki K. Traumatic arteriovenous fistula of the vertebral artery with spinal cord symptoms. Case report. J Neurosurg
1977;46:681 – 7. [2] Halbach VV, Higashida RT, Hieshima GB. Treatment of vertebral arteriovenous fistulas. AJR Am J Roentgenol 1988;150:405 – 12. [3] Yoshida S, Nakazawa K, Oda Y. Spontaneous vertebral arteri-
J.-i. Asai et al. / Clinical Neurology and Neurosurgery 104 (2002) 146–151 ovenous fistula—case report. Neurol Med Chir (Tokyo) 2000;40:211 – 5. [4] Beaujeux RL, Reizine DC, Casasco A, Aymard A, Rufenacht D, Khayata MH, Riche MC, Merland JJ. Endovascular treatment of vertebral arteriovenous fistula. Radiology 1992;183:361 – 7. [5] Roper PR, Guinto FC Jr., Wolma FJ. Posttraumatic vertebral artery aneurysm and arteriovenous fistula: a case report.
151
Surgery 1984;96:556 – 9. [6] Freitag HJ, Grzyska U, Zeumer H. The use of the ‘cross over technique’ in the management of a traumatic vertebro –vertebral fistula. Neuroradiology 1989;31:174 – 6. [7] Young WL, Kader A, Prohovnik I, Ornstein E, Fleischer LH, Ostapkovich N, Jackson LD, Stein BM. Pressure autoregulation is intact after arteriovenous malformation resection. Neurosurgery 1993;32:491 – 7.