CASE REPORT

Vertebral artery dissection as an extremely rare causeof spinal epidural hematoma: case report and reviewof the literature

Zsolt Kulcsar & Zsolt Berentei & Miklos Marosfoi &Istvan Nyary & Istvan Szikora

Received: 30 July 2008 /Accepted: 12 February 2009 /Published online: 28 February 2009# Springer-Verlag 2009

AbstractObjective To present a patient with a cervico-thoracicventrally located epidural hematoma caused by dissectionand subsequent bleeding of the cervical portion of thevertebral artery.Summary and background data Non traumatic epiduralhematoma is a rare entity. The etiology usually is notclarified: a venous origin is usually suspected although anarterial source is also possible.Clinical report A 32-year-old woman presented with aventrally located cervico-thoracic epidural hematomacaused by non traumatic dissection and dissecting aneu-rysm rupture of the cervical portion of the vertebral artery.The dissection was demonstrated by magnetic resonanceimaging and digital subtraction angiography. The patienthad no neurological symptoms and was treated byconservative methods. Follow up imaging showed healingof the vertebral artery and resorption of the epiduralhematoma.Conclusion Dissection of the cervical portion of thevertebral artery with subsequent perivascular bleeding isnot well recognized as a possible cause of a spinal epiduralhematoma. Even though this entity and the underlyingcause may be rare, we suggest a vigilant search forvertebral artery injury in cases of ventrally located cervicaland upper thoracic epidural hematoma.

Keywords Spinal epidural hematoma . Vertebral artery .

Dissection . Dissecting aneurysm . Conservative therapy

Introduction

Non traumatic spinal epidural hematoma is a well recog-nized but uncommon disease. Causative factors are identi-fied only in some patients with the etiology remainingunclarified in the vast majority. The most commonlyaccepted concept is venous bleeding from spinal epiduralveins, but the possibility of an arterial source of hemor-rhage was raised by some authors. [4, 10, 33]. We report anextremely rare example of a spinal epidural hematoma withproven dissection of the vertebral artery and perivascularbleeding as the cause of the hematoma formation.

Clinical report

A 32 year old female presented with sudden onset of a stiffneck and chest pain radiating towards the neck. Atadmission she complained of progressive headache, nauseaand vomiting. She had experienced similar pain 2 weekspreviously, associated with fever, which was considered tobe a viral infection. There was not a history of injury beforeeither episode. Examination showed neck stiffness and apositive Kernig’s sign with no other neurological symp-toms. Subarachnoid hemorrhage (SAH) was suspected butwas excluded by emergency computed tomography (CT)scan of the head. The patient was transferred to aneurosurgical department. Magnetic resonance angiography(MRA) of the circle of Willis did not show any evidence ofan intracranial aneurysm or vascular malformation. Herbackache persisted, so a contrast-enhanced MR scan of the

Acta Neurochir (2009) 151:1319–1323DOI 10.1007/s00701-009-0223-7

Z. Kulcsar : Z. Berentei :M. Marosfoi : I. Nyary : I. SzikoraNational Institute of Neurosurgery,Budapest, Hungary

Z. Kulcsar (*)Klinik Hirslanden, Neurozentrum,Witellikerstrasse 40,8032 Zurich, Switzerlande-mail: kulcsarzsolt22@gmail.com

thoracic and cervical spine was performed 2 and 8 daysafter the onset of her symptoms and showed a spinalepidural hematoma extending from the cranio-cervicaljunction down to the level of the fifth thoracic vertebra(T5) in the ventral epidural space (Fig. 1).

Twelve days after the initial presentation, the patient wastransferred to our hospital for angiographic evaluation andpotential treatment. Digital subtraction angiography (DSA)of the entire neural axis was performed, it excluded spinaland intracranial vascular malformations but showed twodissecting aneurysms of the right vertebral artery, at thelevels of C5 and C2, with irregular stenosis of the segmentin between (Fig. 2). Retrospectively review of the initialimaging studies identified in the previous acute CT-scan aventrally-located hyperdensity representing an acute hema-toma at the level of the foramen magnum and alsosusceptibility artifacts in the wall of the right vertebralartery in the MR studies (Fig. 3). Based on these findings,dissection of the right vertebral artery with dissectinganeurysm formation and bleeding was proposed as themost probable cause of the spinal epidural hematoma. Inview of the patient having had two episodes consistent withvertebral dissection, an endovascular occlusion of the rightVA was considered. However, as no neurological deficitdeveloped and the patient’s meningeal signs graduallyimproved, no invasive treatment was carried out. Antico-agulant or antiplatelet treatment was not proposed as therisk of rebleeding was considered to be elevated, and thepatient was discharged. When reviewed 1 month later, shestill had some neck stiffness but had no neurologicalsymptoms. Repeat DSA at this time showed improvementin the right VA dissection. Complete resolution of stiff neckand pain followed and the patient made a completerecovery. Three months after the event, MR scanning ofthe cervical and thoracic spine with 3D Time-of Flight

MRA showed complete resolution of the spinal epiduralhematoma and healing of the dissected right vertebral artery(Fig. 4).

Discussion

Spinal epidural hematoma is a rare entity that needs urgentdiagnostic evaluation and may necessitate immediatedecompressive spinal surgery if associated with progressiveneurological symptoms [23, 25]. The outcome after surgicalevacuation seems to be inversely related to the timing of

Fig. 1 Sagittal T2-weightedMR scan shows low signalintensity in the ventral epiduralspace from the cranio-cervicaljunction to the level of T5consistent with acute hemor-rhage (a), and most porminent atthe level of T4 resulting in cordcompression (b), demonstratedby the axial GRE T2-weightedimage

Fig. 2 DSA of the right vertebral artery demonstrates two dissectinganeurysm like dilatations at the level of C5 and C2 (black arrows),with luminal narrowing of the segment between

1320 Z. Kulcsar et al.

surgery and the severity of symptoms before decompression[23, 25]. For patients presenting with no or mild neurolog-ical symptoms and for those rapidly improving onobservation, conservative management is also an acceptableoption as a good clinical outcome has been observed inthese cases [5, 7, 8, 12, 14, 19].

The etiology of non traumatic SEH is usually presumedto be associated with coagulopathies [20] including druginduced coagulation disorders [17, 22, 24, 28–30], epiduraltumors [35], underlying spinal vascular malformations [26,27, 32] and also some kind of mechanical rupture of spinalvessels secondary to a straining effort [11, 25]. In themajority of patients, however the cause remains unclear[15]. The most widely accepted hypothesis of the mecha-nism of hematoma formation is bleeding from epiduralveins. Groen et al. found spinal epidural hematoma in theposterior epidural space in 99% of the examples reviewedsuggesting a venous origin for the bleed [15]. This couldexplain the usually slow progression of spinal cordcompression symptoms. Primitive epidural veins do notcontain valves [36] and any sudden rise in pressure in thecentral venous system, eg. during weight lifting or cough-ing—may be transmitted directly to the epidural plexus thuscausing vessel rupture and hemorrhage. In contast, otherauthors have argued against a venous source of thebleeding, noting that the pressure in the epidural veins ismuch less than the intrathecal pressure and concluding thata severe spinal cord compression can not result fromvenous bleeding [4]. As an alternative, they suggest anarterial source of hemorrhage, although true spinal arterio-venous vascular malformations are rarely found at angiog-raphy or surgery as a source of a spinal epidural haematoma[4, 15, 25].

We believe that the coexisting dissection of the vertebralartery and the subsequent development of a dissectinganeurysm was the most likely cause of the epiduralhematoma. Vertebral artery dissections are often preceded

by some form of injury to the neck, however the majority ofthe reported cases are spontaneous in origin [18]. Dis-sections are thought to arise from an intimal tear andsubsequent penetration of circulating blood into the vesselwall leading to formation of an intramural haematoma.Subintimal dissection tends to cause luminal narrowing orocclusion, whereas subadventitial dissection may cause adissecting aneurysm [16]. Combined forms with stenosisand aneurismal dilatation have also been described [2].Dissection of the cervical vertebral artery typically presentswith posterior neck pain ipsilateral to the dissection andheadache. It may be followed by posterior circulation

Fig. 3 Acute brain CT imagedemonstrates a thin hyperden-sity in the anterior aspect of theforamen magnum consistentwith epidural hemorrhage (a).The axial GRE T2-weightedimage at the level of C4 (b)shows abnormal low signal in-tensity consistent with intramu-ral hematoma involving the rightvertebral artery (white arrow)

Fig. 4 Sagittal T2-weighted MR image performed 3 months afterpresentation reveals the spontaneous resorption of the epiduralhematoma (a), and the 3D TOF MRA demonstrates healing of theright vertebral artery with no narrowing of the lumen and normal flowsignal (b)

Vertebral artery dissection: cause of spinal epidural hematoma 1321

transient ischaemic attacks or ischemic stroke, tinnitus, andsensorimotor cervical radiculopathy but may also remainsilent [3, 9, 34]. Most often, the V2 and V3 portion of theartery is affected [3, 13]. The initial diagnosis of a VAdissection is best made with a combination of mural andluminal imaging, such as MRI and MRA or by CT andCTA, although the gold standard for demonstrating luminalirregularities is still conventional DSA [13]. The treatmentof vertebral artery dissections is usually conservative bymeans of anticoagulation or antiplatelet therapy. Endovas-cular therapy is proposed when medical therapy fails orwhen the dissection impairs intracranial perfusion [1, 3, 13,21]. Therapeutic decision making can be difficult whenthromboembolic and hemorrhagic risks are both present.However permanent occlusion of the vertebral artery mayeliminate the risk of rebleeding.

Anatomical considerations

The VA is divided anatomically into four segments: (1) theV1 segment, also called the ostial, extraosseus or proximalsegment, extends from the subclavian artery to the C6transverse process; (2) the V2, called foraminal or trans-versary segment, runs in and between transverse processesfrom C6 to C2; (3) the V3, extraspinal or suboccipitalsegment, begins at the C2 transverse foramen and extendsup to the dura mater of the foramen magnum; (4) the V4segment, or intracranial segment, is intradural and forms thebasilar artery after joining the contralateral segment.Passing through one transverse process to another, the V2segment of the VA is running free and is enclosed with theperivertebral venous plexus into a thin continuous sheathmade up of the extension of the transverse processesperiosteum. Between two transverse processes, the VAcrosses the cervical nerve root anteriorly, these two beingseparated only by this thin layer of periosteum [6]. Weassume that a transmural dissection could eventually breakthe thin anatomical layer separating the VA from thevertebral foramen and the ventro-lateral epidural space, sothat bleeding could penetrate into the spinal epidural space.

In our patient the DSA and MR-findings clearlydemonstrated the signs of dissection of the arterial wallwith aneurysm formation and the hematoma in the epiduralspace. We believe that this forms an anatomical basis toexplain also the ventral location of hematoma, with theepidural hyperdensity of blood seen on the initial CT scanat the level of the foramen magnum which was continuousfrom the level of C2 to T4.

A combined search with the specific keywords in theMedline database yielded two foreign language articleswhich raised the possibility of a connection betweenvertebral artery dissections and cervical epidural hematoma,although the relationship was not confirmed [31, 37].

However reviewing the relevant literature we found onearticle presenting a similar case of dissecting V3 segmentaneurysm associated with epidural hematoma [9]. In thiscase the hematoma was also ventrally located, and thepatient recovered without any specific treatment. Themechanism of epidural hematoma formation was notdiscussed by the authors of the article.

Conclusions

We believe that our patient is the second to be reported withdocumented dissection and extraluminal rupture of thevertebral artery as the most probable cause of a nontraumatic ventrally located cervical epidural hematoma.We think that the etiology of this pathology is not wellunderstood and dissection of the vertebral artery withperivascular bleeding might be underdiagnosed. Rare asthis entity and the underlying cause may be, we suggest amore vigilant investigation of the vertebral arteries in apatient with a ventrally located cervical and upper thoracicepidural hematoma. We do not suggest performing spinaland vertebral DSA in all patients with a spinal epiduralhaematoma; nevertheless CT- or MR-angiography of thecervical arteries should always be considered in thediagnostic workup.

References

1. Ansari SA, Thompson BG, Gemmete JJ, Gandhi D (2008)Endovascular treatment of distal cervical and intracranial dis-sections with the neuroform stent. Neurosurgery 62:636–646discussion 636–646. doi:10.1227/01.NEU.0000311350.25281.6B

2. Arnold M, Bousser MG (2005) Carotid and vertebral arterydissection. Pract Neurol 5:100–109. doi:10.1111/j.1474-7766.2005.t01-1-00292.x

3. Arnold M, Bousser MG, Fahrni G, Fischer U, Georgiadis D,Gandjour J, Benninger D, Sturzenegger M, Mattle HP, Baumgart-ner RW (2006) Vertebral artery dissection: presenting findings andpredictors of outcome. Stroke 37:2499–2503. doi:10.1161/01.STR.0000240493.88473.39

4. Beatty RM, Winston KR (1984) Spontaneous cervical epiduralhematoma. A consideration of etiology. J Neurosurg 61:143–148

5. Binder DK, Sonne DC, Lawton MT (2004) Spinal epidural hema-toma. Neurosurg Q 14:51–59. doi:10.1097/00013414-200403000-00009

6. Bruneau M, Cornelius JF, George B (2005) Anterolateralapproach to the V2 segment of the vertebral artery. Neurosurgery57:262–267 discussion 262–267. doi:10.1227/01.NEU.0000176414.58086.2B

7. Clarke CE, Hall DJ, Zagnoon A, Loizou LA (1993) Successfulconservative management of spontaneous spinal extraduralhaematoma. Br J Neurosurg 7:209–211. doi:10.3109/02688699309103482

8. Clarke DB, Bertrand G, Tampieri D (1992) Spontaneous spinalepidural hematoma causing paraplegia: resolution and recovery

1322 Z. Kulcsar et al.

without surgical decompression. Neurosurgery 30:108–111.doi:10.1097/00006123-199201000-00019

9. Crum B, Mokri B, Fulgham J (2000) Spinal manifestations ofvertebral artery dissection. Neurology 55:304–306

10. Cywinski JB, Parker BM, Lozada LJ (2004) Spontaneous spinalepidural hematoma in a pregnant patient. J Clin Anesth 16:371–375. doi:10.1016/j.jclinane.2003.08.007

11. David S, Salluzzo RF, Bartfield JM, Dickinson ET (1997)Spontaneous cervicothoracic epidural hematoma following pro-longed valsalva secondary to trumpet playing. Am J Emerg Med15:73–75. doi:10.1016/S0735-6757(97)90054-1

12. Duffill J, Sparrow OC, Millar J, Barker CS (2000) Canspontaneous spinal epidural haematoma be managed safelywithout operation? A report of four cases. J Neurol NeurosurgPsychiatry 69:816–819. doi:10.1136/jnnp.69.6.816

13. Flis CM, Jager HR, Sidhu PS (2007) Carotid and vertebral arterydissections: clinical aspects, imaging features and endovasculartreatment. Eur Radiol 17:820–834. doi:10.1007/s00330-006-0346-7

14. Groen RJ (2004) Non-operative treatment of spontaneous spinalepidural hematomas: a review of the literature and a comparisonwith operative cases. Acta Neurochir (Wien) 146:103–110.doi:10.1007/s00701-003-0160-9

15. Groen RJ, Ponssen H (1990) The spontaneous spinal epiduralhematoma. A study of the etiology. J Neurol Sci 98:121–138.doi:10.1016/0022-510X(90)90253-J

16. Guillon B, Brunereau L, Biousse V, Djouhri H, Levy C, BousserMG (1999) Long-term follow-up of aneurysms developed duringextracranial internal carotid artery dissection. Neurology 53:117–122

17. Han YM, Kwak HS, Jin GY, Chung GH, Song KJ (2006) Spinalepidural hematoma after thrombolysis for deep vein thrombosiswith subsequent pulmonary thromboembolism: a case report.Cardiovasc Intervent Radiol 29:450–453. doi:10.1007/s00270-004-0133-2

18. Haneline MT, Lewkovich GN (2005) An analysis of the etiologyof cervical artery dissections: 1994 to 2003. J ManipulativePhysiol Ther 28:617–622. doi:10.1016/j.jmpt.2005.08.016

19. Hentschel SJ, Woolfenden AR, Fairholm DJ (2001) Resolution ofspontaneous spinal epidural hematoma without surgery: report of twocases. Spine 26:E525–E527. doi:10.1097/00007632-200111150-00025

20. Hutt PJ, Herold ED, Koenig BM, Gilchrist GS (1996) Spinalextradural hematoma in an infant with hemophilia A: an unusualpresentation of a rare complication. J Pediatr 128:704–706. doi:10.1016/S0022-3476(96)80141-7

21. Kadkhodayan Y, Jeck DT, Moran CJ, Derdeyn CP, Cross DT 3rd(2005) Angioplasty and stenting in carotid dissection with orwithout associated pseudoaneurysm. AJNR Am J Neuroradiol26:2328–2335

22. Kirazli Y, Akkoc Y, Kanyilmaz S (2004) Spinal epidural hematomaassociated with oral anticoagulation therapy. Am J Phys MedRehabil 83:220–223. doi:10.1097/01.PHM.0000107498.91919.44

23. Lawton MT, Porter RW, Heiserman JE, Jacobowitz R, SonntagVK, Dickman CA (1995) Surgical management of spinal epiduralhematoma: relationship between surgical timing and neurologicaloutcome. J Neurosurg 83:1–7

24. Lederle FA, Cundy KV, Farinha P, McCormick DP (1996) Spinalepidural hematoma associated with warfarin therapy. Am J Med100:237–238. doi:10.1016/S0002-9343(97)89465-X

25. Liao CC, Lee ST, Hsu WC, Chen LR, Lui TN, Lee SC (2004)Experience in the surgical management of spontaneous spinalepidural hematoma. J Neurosurg 100:38–45

26. Miyagi Y, Miyazono M, Kamikaseda K (1998) Spinal epiduralvascular malformation presenting in association with a spontane-

ously resolved acute epidural hematoma. Case report. J Neurosurg88:909–911

27. Muller H, Schramm J, Roggendorf W, Brock M (1982) Vascularmalformations as a cause of spontaneous spinal epidural haema-toma. Acta Neurochir (Wien) 62:297–305. doi:10.1007/BF01403637

28. Ozgocmen S, Yoldas T, Kocakoc E, Ozkurt-Zengin F, ArdicogluO (2004) Spinal epidural hematoma associated with streptokinasetreatment for myocardial infarction. Spinal Cord 42:374–377.doi:10.1038/sj.sc.3101604

29. Sawin PD, Traynelis VC, Follett KA (1995) Spinal epiduralhematoma following coronary thrombolysis with tissue plasmin-ogen activator. Report of two cases. J Neurosurg 83:350–353

30. Seet RC, Lim EC, Wilder-Smith EP, Ong BK (2005) Spontaneousepidural haematoma presenting as cord compression in a patientreceiving clopidogrel. Eur J Neurol 12:811–812. doi:10.1111/j.1468-1331.2005.01057.x

31. So G, Debata A, Baba S, Tsunoda K, Suyama K, Nagata I (2008)Spontaneous cervical epidural hematoma presenting with hemi-paresis following neck extension: a case report. No Shinkei Geka36:731–734

32. Solero CL, Fornari M, Savoiardo M (1980) Spontaneous spinalepidural haematoma arising from ruptured vascular malformation.Case report. Acta Neurochir (Wien) 53:169–174. doi:10.1007/BF02074789

33. Spengos K, Sameli S, Tsivgoulis G, Vassilopoulou S, Vemmos K,Zakopoulos N (2005) Spontaneous spinal epidural hematoma inan untreated hypertensive patient. Eur J Intern Med 16:451–453.doi:10.1016/j.ejim.2005.01.019

34. Tabatabai G, Schober W, Ernemann U, Weller M, Kruger R(2008) Vertebral artery dissection presenting with ispilateral acuteC5 and C6 sensorimotor radiculopathy: a case report. Cases J1:139. doi:10.1186/1757-1626-1-139

35. Tashjian RZ, Bradley MP, Lucas PR (2005) Spinal epiduralhematoma after a pathologic compression fracture: an unusualpresentation of multiple myeloma. Spine J 5:454–456.doi:10.1016/j.spinee.2005.03.006

36. van der Kuip M, Hoogland PV, Groen RJ (1999) Human radicularveins: regulation of venous reflux in the absence of valves. AnatRec 254:173–180. doi:10.1002/(SICI)1097-0185(19990201)254:2<173::AID-AR3>3.0.CO;2-B

37. Yasaka M, Arihiro S, Hasegawa Y, Minematsu K (2005) A case ofacute cervical epidural hematoma mimicking medial medullaryinfarction complicated by arterial dissection. Rinsho Shinkeigaku45:652–656

Comment

The authors present an interesting case report, and a very surprisingfinding in a patient presenting with chest pain. However, this is not anew observation. Previous reports describe vertebral artery dissectionpresenting with chest or back pain in the absence of apparent trauma.The authors did well in re-creating the patient’s story and the chain ofevents. MRI and angiogram images depict the situation well. Theliterature agrees with nonoperative management in patients withoutneurological compromise. Although complete proof of cause andeffect is not possible, it seems very probable that the spinal EDH iscaused by vertebral artery dissection.

John ZhangJan EckermannUSA

Vertebral artery dissection: cause of spinal epidural hematoma 1323