J Neurosurg Spine 20:322–326, 2014 ©AANS, 2014

Spinal dural arteriovenous shunt presenting with intramedullary hemorrhage Case report Ayumi Narisawa, M.D., Ph.D.,1 Toshiki Endo, M.D., Ph.D.,1 Kenichi Sato, M.D., Ph.D., 2 Mika Watanabe, M.D., Ph.D., 3 Akira Takahashi, M.D., Ph.D., 2 and Teiji Tominaga, M.D., Ph.D.1 Departments of 1Neurosurgery, 2Neuroendovascular Therapy, and 3Pathology, Tohoku University, Graduate School of Medicine, Sendai, Japan The authors report on a 49-year-old man with a thoracic spinal dural arteriovenous shunt (dAVS) in which rup­ ture of a varix caused intramedullary hemorrhage. In the literature, patients with a thoracic dAVS predominantly pre­ sent with congestive myelopathy; however, the patient featured in this report presented without increased deep tendon reflexes or muscle weakness, but instead with intermittent stabbing chest pain and paresthesia. Magnetic resonance images and angiograms demonstrated tortuous enlargement and the formation of a varix-like structure of the drain­ ing veins, features compatible with those of high-flow angiopathy. Recognition of this phenomenon is important in thoracic dAVS because intramedullary hemorrhage dramatically degrades outcome. A high index of clinical suspicion can prevent a similar case of thoracic dAVS from progressing to intramedullary hemorrhage. (http://thejns.org/doi/abs/10.3171/2013.12.SPINE12163)


Key Words      •      hematomyelia      •      intramedullary hemorrhage      •      myelopathy      • spinal dural arteriovenous shunt      •      spinal arteriovenous malformation      •      varix

dural arteriovenous shunt (dAVS) is the most common type of spinal vascular malformation, con­ stituting approximately 70% of all spinal arterio­ venous malformations.2,17 Clinical manifestations include pro­ gressive congestive myelopathy and subarachnoid hem­­or­rhage (SAH).9,11,16 Intramedullary hemorrhage in relation to spinal dAVS is extremely rare and has been reported only once in the English literature.14 Here, we report a case of tho­ racic dAVS initially presenting with intermittent, stab­ bing left-sided chest pain, as well as gradual worsening of paresthesia, and later complicated by the sudden onset of intramedullary hemorrhage. We evaluated the patient us­ ing MRI at different time points. Importantly, before the hemorrhage occurred, T2-weighted MRI demonstrated a high-intensity spot in the spinal cord around a wedged portion of draining veins. Given the clinical course and intraoperative findings in this case, the imaging finding was considered to indicate increased transluminal pres­ sure leading to rupture of the varix-like structure and in­ tramedullary hemorrhage. Clinical presentation and neurological examinations before manifestation of the hemorrhage were also distinct from those of congestive myelopathy. Meticulous preop­ erative evaluations with a high index of clinical suspicion may have prevented this unusual and devastating sequela. pinal

Abbreviations used in this paper: dAVS = dural arteriovenous shunt; SAH = subarachnoid hemorrhage.


Case Report History and Examination. A 49-year-old man was re­ ferred to our department with complaints of paresthesia of the chest, trunk, and bilateral lower limbs as well as difficulty in walking. He also had a history of intermittent stabbing left-sided chest pain, which was so intense that he had required admissions to another hospital to rule out angina pectoris. He needed a cane for walking long dis­ tances because of increased paresthesia over the preced­ ing few months. On admission, neurological examination revealed re­ duced superficial and deep sensation in areas below the T-6 dermatome. Muscle weakness was not apparent in the patient’s extremities. There was no Romberg sign and no pathological reflexes. The cremasteric reflex was present on the right side and absent on the left side. Patellar and Achilles tendon reflexes were slightly diminished bilater­ ally. Magnetic resonance imaging showed a flow-related signal void on the posterior surface of the spinal cord (Fig. 1). On T2-weighted MRI, hyperintense lesions were noted but limited to the T-6 level where a portion of the draining vein was noted to be tortuous and wedged into the dorsal surface of the spinal cord (Fig. 1C). Selective spinal angi­ This article contains some figures that are displayed in color on­line but in black-and-white in the print edition.

J Neurosurg: Spine / Volume 20 / March 2014

Spinal dAVS with intramedullary hemorrhage

Fig. 1.  Magnetic resonance images obtained before the intramedullary hemorrhage occurred. A T1-weighted MR image with Gd enhancement (A) and T2-weighted images (B and C) showing a dilated venous structure dorsal to the spinal cord. At the T-6 level (asterisks), a tangled venous complex appeared to adhere to the surface of the spinal cord (white arrows). In a magnified image (C), intramedullary T2 hyperintense areas were noted around the venous structure (black arrowheads) where a portion of the draining veins wedged into the spinal cord (black arrow).

ography detected a shunt at the left T-6 root sleeve (Fig. 2). A preoperative diagnosis of thoracic dAVS was made, and the patient was scheduled for elective surgery. While the patient waited for the surgical intervention, he experienced the sudden onset of back pain followed by exacerbation of paresthesia and muscle weakness in the left leg. He was not on anticoagulation or antiplate­ let therapy. At the ictus, he was lying and resting in bed. No fluctuation in systemic blood pressure was noted be­ fore the hemorrhagic event. He was unable to stand up and was admitted emergently to our hospital. Although he could move his right leg against resistance, muscle strength in his left leg was 2 or lower on manual muscle testing. His superficial and deep sensation was graded as 50% bilaterally below the T-6 dermatome. Surprisingly, MRI revealed an intramedullary hemorrhage at T-6. The wedged portion of the draining vein was now dilated to form a varix-like structure (Fig. 3). High-intensity areas in the spinal cord, cranial and caudal to the hematoma, were now noticeable on T2-weighted images. A rupture of the varix was considered to be a potential cause of intramed­ ullary hemorrhage and acute neurological deterioration. Operation. The patient underwent surgical interven­ tion on the following day after preoperative evaluations. With the patient prone, motor and sensory evoked po­ tentials were monitored as described previously.6 Par­ tial hemi-laminotomies of the left T-5 and T-6 laminae were performed for exposure. Following dural incision, a pedicle of the draining vein was recognized as aris­ ing from the left T-6 root sleeve (Fig. 4A and B). Soon J Neurosurg: Spine / Volume 20 / March 2014

after interruption of the draining vein, the arterialized vein turned purple. Indocyanine green videoangiography was used to confirm remission of the arterialized venous flow in the dilated veins. As we dissected and separated the veins from the dorsal spinal cord, a dilated varix-like portion was found protruding into the spinal cord paren­ chyma and facing a cavity in the hematoma. The varix was caught between thick pial membranes (Fig. 4C and D). We cut these membranous tissues and excised the varix for histological analysis. Careful inspection under an operative microscope did not reveal any other vascu­ lar malformation (Fig. 4E and F). Having considered that the location of this varix-like structure was identical to what preoperative MRI had indicated, we confirmed that the rupture of this varix-like structure was the source of intramedullary hemorrhage. No changes in the amplitude and latency of the motor and sensory evoked potentials were recorded during the operation. Pathological Findings. Histological examination re­ vealed hemosiderin deposition and fibrin attachment on the surface of the varix, which supported the contention that a rupture of the varix was the cause of the intramedul­ lary hemorrhage (Fig. 5). Postoperative Course. Postoperative MRI confirmed that the intramedullary hematoma was completely re­ moved. Flow-related signal voids of the dilated venous structures disappeared, and T2 hyperintensity lesions were remarkably diminished. Immediately after the operation, the patient’s chest pain subsided. His motor weakness showed gradual improvement. One month after surgery, 323

A. Narisawa et al.

Fig. 2.  Digital subtraction angiograms (A–C) of the left T-6 intercostal artery, anteroposterior view, and a reconstructed image (D). The arteriovenous shunt was localized (black arrow) at the left T-6 dural root sleeve. Ascending and descending draining veins were visualized. No other spinal vascular malformations were detected. A wedged portion of the draining vein (arrowheads) was noted as an oval-shaped structure.

muscle strength in his left leg was 4 on manual muscle testing, and he could walk with a cane. However, his post­ operative neurological recovery was limited thereafter. Two years after surgery, paresthesia and sensory distur­

bances remained in his trunk, back, and lower extremities. He still required a cane for ambulation because of left leg weakness, and he needed self-catheterization for voiding urine.

Fig. 3.  Magnetic resonance images obtained after the onset of intramedullary hemorrhage. A T2-weighted image (A) and T1-weighted image (B) showing the intramedullary hematoma at T-6 (asterisks). Hyperintense lesions were apparent along the spinal cord on the T2-weighted image. Gadolinium-enhanced T1-weighted image (C) revealing enlargement of the previously wedged portion of the draining vein protruding into the hematomyelia (arrow).


J Neurosurg: Spine / Volume 20 / March 2014

Spinal dAVS with intramedullary hemorrhage

Fig. 5. Histological examination of the varix using Masson’s trichrome stains. Internal elastic layers ended at the sites indicated by the arrows (left). The walls of the varix were thin and fragile. Hemosiderin depositions were detected at the tip of the varix at a higher magnification (arrowheads, right). Inset in the left image indicates the position of the image on the right.

Fig. 4. Intraoperative photographs including those taken with indocyanine green videoangiography. After dural incision, a pedicle of the draining vein was located at the left T-6 root sleeve (arrowheads in A and B). Following disconnection of the draining veins, the vessels turned purple (C and D). A varix-like structure was caught underneath the pial membranes (arrow in C). Following dissection of the membranes, the embedded varix was exposed (arrow in D) and excised from the hematoma cavity. Inside the cavity, there was no other source of bleeding (arrows in E and F). No arterial blood flow was detected inside of the cavity or in the draining veins according to indocyanine green videoangiography (F).


In the last 10 years, we have treated 42 cases of spinal dAVS using microsurgery or endovascular interventions. The currently reported case is the only one to present with intramedullary hemorrhage. Five cases have presented with SAH, and the shunts were localized in the cervical spine or craniocervical junction in all of these cases. In spinal dAVS, arteriovenous shunts often locate at the dural root sleeve.13 They were categorized as dorsal intradural arteriovenous fistulas in a system proposed by Spetzler et al.17 Increased intravascular pressure is trans­ mitted from radiculomeningeal arteries to the radicular veins and the venous system of the spinal cord.5 It causes stagnation of the venous outflow from the spinal cord, resulting in intramedullary venous hypertension and is­ chemic insult to the spinal cord.3 Thus, patients suffer progressive congestive myelopathy, which occurs most frequently in the thoracolumbar region.8,16 Subarachnoid hemorrhage is encountered as anoth­ er clinical presentation of this disease, especially when shunts are located in the craniocervical junction or the cervical spine.7,9,16 A pattern of venous outflow has been reported as significant in determining the clinical mani­ festation of the disease.7 More specifically, venous drain­ age ascending cranially may pose an increased risk of hemorrhage. In such cases, accelerated venous blood flow through the draining veins and focal dilation to form a venous sac4 or varix-like pouch were reported.9 J Neurosurg: Spine / Volume 20 / March 2014

Considering a case report of lumbar dAVS presenting with SAH, common features of hemorrhagic spinal dAVS may include an accelerated rate of arteriovenous circu­ lation and an altered configuration of the radiculospinal vein.10 In the presented case, successive MR images dem­ onstrated formation of the varix from a wedged portion of the draining vein. In this context, we speculated that increased venous hypertension had led to the formation of a varix. Furthermore, we confirmed, based on intra­ operative and histological findings, that the hemorrhage occurred where the varix was protruding into the spinal cord. The reason that the increased venous pressure was focused in the varix and caused an intramedullary hemor­ rhage but not SAH remains to be elucidated. As shown in Fig. 2C, venous drainage was equally distributed to cra­ nial and caudal directions. Thus, the same theory of SAH in a cervical or craniocervical junction dAVS cannot be applied in this case. Interestingly, it was intraoperatively confirmed that the varix was caught between pial membranes and pro­ truded into the spinal cord parenchyma. Given this unique configuration, we speculate that intraluminal pressure may have been acutely and heterogeneously increased inside the varix, resulting in a rupture. The walls of the varix, where the hemorrhage occurred, lacked elastic fibers and tended to be fragile. Dorsal spinal veins predominantly run in subarach­ noid spaces.12 However, our case is an extremely rare ex­ ample in which the draining veins of the dAVS ran under­ neath the pia mater, which may also explain the rarity of intramedullary hemorrhage in spinal dAVS. As demonstrated here, angioarchitecture of the drain­ ing veins, especially the relationship to the spinal cord pa­ renchyma, could be key in predicting the formation and rupture of a varix along the draining veins. One impor­ tant finding in preoperative MRI was the T2 hyperinten­ sity areas around the protruded portion of the perimed­ ullary vein in the spinal cord parenchyma (Fig. 1C). As reported previously, perifocal T2 hyperintensity indicated increased transluminal pressure in a case of intracranial arteriovenous malformation.1 In that report, intranidal an­ eurysm formation was apparent in the center of the T2 hy­ perintensity area on angiograms. Because of the increased risk of hemorrhage, urgent resection of the arteriovenous malformation as well as the intranidal aneurysm was per­ 325

A. Narisawa et al. formed, which resulted in satisfactory results. In our case, enlargement of the varix and hemorrhage occurred in the exact location where the abnormality appeared on the T2weighted MRI image. Thus, we speculate that intraparen­ chymal T2 hyperintensity around the varix wall may have indicated an increment of intraluminal pressure and pre­ dicted the hemorrhagic event. The clinical presentation of this case was also distinct from that of typical congestive myelopathy associated with midthoracic spinal dAVS. The patient complained of paresthesia and difficulty in walking. However, neurologi­ cal examination did not reveal motor weakness, increased deep tendon reflexes, or pathological reflexes. Given the lack of objective abnormality in the neurological exami­ nations, we did not consider the case to require urgent sur­ gical intervention, and the intramedullary hemorrhage oc­ curred before the treatment. However, when accompanied by clinical signs of spinal cord dysfunction, sudden and stabbing pain should be considered with caution, since it may indicate hemorrhage due to spinal arteriovenous mal­ formation.15 Given the limited number of similar reports, whether the stabbing pain preceded the intramedullary hemorrhage or the SAH remains to be determined. The patient suffered acute and severe neurological de­terioration and limited postoperative functional recov­ ery as a result of the hemorrhage. Given the devastating clinical course of our patient, it is important to recognize the possibility of such hemorrhagic events in patients with thoracic spinal dAVS.14


We describe a case of thoracic dAVS that presented with intramedullary hemorrhage with preceding intermit­ tent stabbing chest pains and paresthesia. A subpial wedge of the draining veins and intraparenchymal T2 hyperin­ tensity around the dilated venous structure could be a key finding in preoperative MRI to predict the formation and rupture of a varix. It is important to recognize that intra­ medullary hemorrhage, although it is rare, can occur in as­ sociation with thoracic dAVS, because it could dramatical­ ly degrade the clinical course and neurological outcomes of patients. Disclosure The authors report no conflict of interest concerning the mate­ rials or methods used in this study or the findings specified in this pa­per. Author contributions to the study and manuscript preparation in­clude the following. Conception and design: Endo, Narisawa, Tomi­ na­ga. Acquisition of data: Endo, Narisawa, Watanabe, Takahashi. Anal­ysis and interpretation of data: Endo, Sato. Drafting the article: En­­do, Narisawa, Sato. Critically revising the article: all authors. Re­viewed submitted version of manuscript: all authors. Approved the final version of the manuscript on behalf of all authors: Endo. Ad­­ min­is­trative/technical/material support: Takahashi, Tominaga. Study su­per­vision: Takahashi, Tominaga. References   1.  Akabane A, Jokura H, Ogasawara K, Takahashi K, Sugai K, Ogawa A, et al: Rapid development of an intranidal aneurysm


with perifocal brain edema in an unruptured cerebral arte­ riovenous malformation. Case report. J Neurosurg 97:1436– 1440, 2002  2. Aminoff MJ, Logue V: Clinical features of spinal vascular malformations. Brain 97:197–210, 1974   3.  Criscuolo GR, Oldfield EH, Doppman JL: Reversible acute and subacute myelopathy in patients with dural arteriovenous fis­ tulas. Foix-Alajouanine syndrome reconsidered. J Neurosurg 70:354–359, 1989   4.  Hashimoto H, Iida J, Shin Y, Hironaka Y, Sakaki T: Spinal du­ ral arteriovenous fistula with perimesencephalic subarachnoid haemorrhage. J Clin Neurosci 7:64–66, 2000   5.  Hurst RW, Kenyon LC, Lavi E, Raps EC, Marcotte P: Spinal dural arteriovenous fistula: the pathology of venous hyperten­ sive myelopathy. Neurology 45:1309–1313, 1995  6. Inoue T, Endo T, Nagamatsu K, Watanabe M, Tominaga T: 5-aminolevulinic acid fluorescence-guided resection of intra­ medullary ependymoma: report of 9 cases. Neurosurgery 72 (2 Suppl Operative):ons159–ons168, 2012   7.  Kai Y, Hamada J, Morioka M, Yano S, Mizuno T, Kuratsu J: Arteriovenous fistulas at the cervicomedullary junction pre­ senting with subarachnoid hemorrhage: six case reports with special reference to the angiographic pattern of venous drain­ age. AJNR Am J Neuroradiol 26:1949–1954, 2005  8. Kataoka H, Miyamoto S, Nagata I, Ueba T, Hashimoto N: Venous congestion is a major cause of neurological deteriora­ tion in spinal arteriovenous malformations. Neurosurgery 48:1224–1230, 2001   9.  Kinouchi H, Mizoi K, Takahashi A, Nagamine Y, Koshu K, Yoshimoto T: Dural arteriovenous shunts at the craniocervical junction. J Neurosurg 89:755–761, 1998 10.  Koch C, Gottschalk S, Giese A: Dural arteriovenous fistula of the lumbar spine presenting with subarachnoid hemorrhage. Case report and review of the literature. J Neurosurg 100 (4 Suppl Spine):385–391, 2004 11.  Krings T, Geibprasert S: Spinal dural arteriovenous fistulas. AJNR Am J Neuroradiol 30:639–648, 2009 12.  Lasjaunias P, Berenstein A, terBrugge KG: Surgical Neuroangiography: Clinical and Endovascular Treatment Aspects in Adults, ed 2. Berlin: Springer Verlag, 2006, Vol 2.2, pp 763–766 13.  McCutcheon IE, Doppman JL, Oldfield EH: Microvascular anatomy of dural arteriovenous abnormalities of the spine: a microangiographic study. J Neurosurg 84:215–220, 1996 14.  Minami M, Hanakita J, Takahashi T, Kitahama Y, Onoue S, Kino T, et al: Spinal dural arteriovenous fistula with hema­ tomyelia caused by intraparenchymal varix of draining vein. Spine J 9:e15–e19, 2009 15. Shephard RH: Spinal arteriovenous malformations and sub­ arachnoid haemorrhage. Br J Neurosurg 6:5–12, 1992 16.  Shinoyama M, Endo T, Takahash T, Shimizu H, Takahashi A, Suzuki M, et al: Long-term outcome of cervical and thoraco­ lumbar dural arteriovenous fistulas with emphasis on sensory disturbance and neuropathic pain. World Neurosurg 73:401– 408, 2010 17.  Spetzler RF, Detwiler PW, Riina HA, Porter RW: Modified classification of spinal cord vascular lesions. J Neurosurg 96 (2 Suppl):145–156, 2002

Manuscript submitted April 21, 2013. Accepted December 5, 2013. Please include this information when citing this paper: published online January 10, 2014; DOI: 10.3171/2013.12.SPINE12163. Address correspondence to: Toshiki Endo, M.D., Ph.D., Depart­ ment of Neurosurgery, Graduate School of Medicine, Tohoku Uni­ versity, 1-1 Seiryo Aoba, Sendai 980-8574, Japan. email: [email protected] nsg.med.tohoku.ac.jp.

J Neurosurg: Spine / Volume 20 / March 2014

Spinal dural arteriovenous shunt presenting with intramedullary hemorrhage: case report.

The authors report on a 49-year-old man with a thoracic spinal dural arteriovenous shunt (dAVS) in which rupture of a varix caused intramedullary hemo...
3MB Sizes 0 Downloads 0 Views