Content uploaded by Istvan Bodi
Author content
All content in this area was uploaded by Istvan Bodi on Nov 03, 2014
Content may be subject to copyright.
ORIGINAL ARTICLE
Spontaneous epidural spinal haematoma in children caused
by vascular malformations
Ramanan Sivakumaran •Andrew King •
Istvan Bodi •Christopher L. Chandler •
Daniel C. Walsh
Received: 1 October 2013 / Revised: 5 October 2014 / Accepted: 5 October 2014
ÓSpringer-Verlag Berlin Heidelberg 2014
Abstract
Purpose The occurrence of spinal epidural haematoma of
‘spontaneous’ origin in adults is a well-documented entity,
though it is rare in children. In the literature to date, there
are few cases of this kind of spontaneous haematoma
proven to be due to an underlying vascular abnormality.
Method Retrospective review of two cases of children
under 15 years of age with spontaneous epidural spinal
haematoma.
Results Underlying arteriovenous malformations were
identified in both cases. Intra-operative photographs and
histological sections of these anomalies are presented.
Conclusion These are the first two such cases described
with clinico-pathological correlation.
Keywords Spinal Epidural Haematoma
Arteriovenous Malformation Children
Introduction
Spontaneous epidural spinal haematoma (SESH) is a rare
cause of spinal cord compression in children [1]. Although
well recognised in adults, reports of this entity in the
paediatric population are few and are ascribed to idiopathic
causation. We present two children with their surgical
management, outcomes and pathological features.
Case 1
A previously healthy 8-year-old girl developed severe neck
pain while dancing. She also reported feeling a ‘pop’ as the
pain developed. Over a 6-week period, she became
increasingly lethargic with progressively worsening
mobility. After a further week, she attended the casualty
department of the local district general hospital unable to
walk.
Clinical examination on admission revealed MRC grade
3/5 spastic tetraparesis and urinary retention. Spinal mag-
netic resonance imaging (MRI) revealed an epidural lesion
compressing the theca on the right from C4/5 to C7/T1
levels (Fig. 1). The lesion was centrally bright on T1-
weighted imaging with a peripheral lower signal intensity
rim. The signal characteristics of this lesion were consistent
with an acute or subacute epidural haematoma. No other
abnormality was seen in the neuroaxis. All haematological
investigations on admission were within normal limits.
The girl was taken for an urgent C5–C7 laminoplasty
and evacuation of haematoma. Significant bleeding was
encountered from abnormally dilated arterialised epidural
veins, which was quickly controlled by diathermy. These
veins extended caudal to the laminotomy (Fig. 2), but as no
vascular imaging was available to provide architectural
R. Sivakumaran C. L. Chandler D. C. Walsh
Department of Neurosurgery, King’s College Hospital, London,
UK
R. Sivakumaran (&)
Academic Neurosurgery Unit, St George’s University of
London, Cranmer Terrace, London SW17 0RE, UK
e-mail: ramsivakumaran@doctors.org.uk
A. King I. Bodi
Department of Neuropathology, King’s College Hospital,
London, UK
D. C. Walsh
Institute of Psychiatry, King’s College University, London, UK
123
Eur Spine J
DOI 10.1007/s00586-014-3616-0
information (on what we believed to be an arteriovenous
malformation), we elected not to extend the exposure
within the immature spine. With the existing exposure, the
abnormal vessels were removed from the dural sac and
excellent haemostasis achieved. Once the veins at the level
of exposure had been excised, the bleeding from the
remaining veins was evidently no longer arterial.
Post-operatively, the child awoke immediately to make
a good neurological recovery over the subsequent 3 weeks.
At discharge, she was ambulant independently with normal
bladder function. Time-resolved imaging of contrast
kinetics (TRICKS) sequence MRI scan performed post-
operatively did not show any residual haematoma or
abnormal vasculature.
Histology revealed blood clot and a complex arrange-
ment of vessels with varying wall thickness, including
venous like structures but with fibrous thickening and some
arterialized vessels with thinner walls (Fig. 3). The features
were those of a vascular malformation, consistent with an
AVM.
Case 2
An otherwise healthy 13-year-old boy developed sudden-
onset interscapular pain with no obvious precipitating
cause. The following day, he developed progressively
severe lower limb weakness and urinary retention. He was
admitted to a local children’s hospital where neurological
examination revealed paraplegia (MRC grade 0/5) with
areflexia, loss of anal tone and a T4 sensory level. MRI of
the thoracic spine (Fig. 4) demonstrated a mixed-signal
epidural lesion located posteriorly and to the left extending
between C7/T1 and T4/5. The lesion displaced the spinal
cord anteriorly and its signal characteristics were most in
keeping with an acute or subacute epidural haematoma.
Fig. 1 T2-weighted image of cervico-thoracic spine illustrating
dorsal epidural lesion
Fig. 2 Intra-operative photograph (a) demonstrating residual arterio-
venous malformation underneath cervical lamina (arrow) with
accompanying illustration (b) depicting arteriovenous malformation
(AVM) under cut edge of lamina, with dura and overlying haematoma
exposed
Fig. 3 Histology showing vessel walls with varying thickness
including a thin arterialized vessel wall (A) and a thick fibrous
walled vessel (V) (elastic Van Gieson stain, 920)
Eur Spine J
123
He underwent emergency T1–T5 laminoplasty and
evacuation of an epidural haematoma and resection of an
AVM. Histological analysis revealed fragments of haem-
orrhagic tissue containing vascular channels with abnor-
mally thick and irregular walls, which were also, in places,
very thin (Fig. 5). These findings were consistent with the
presence of an arteriovenous malformation.
Post-operatively, this boy’s neurological deficit had
improved to MRC grade 4/5. At the time of discharge from
hospital, he still experienced intermittent urinary inconti-
nence but was able to walk without assistance. By
3 months post-operatively, he was independently ambulant
with very mild right lower extremity weakness on close
scrutiny of his gait, though his intermittent urinary conti-
nence persisted. MRI of his thoracic spine performed at
that time demonstrated no flow voids or abnormal contrast
enhancement pattern to indicate residual vascular
malformation.
Discussion
Spontaneous epidural spinal haematoma (SESH) is an
uncommon cause of spinal cord compression in children,
with 37 cases previously reported in English literature to
date. The clinical presentation may be variable or even
non-specific making it difficult to diagnose. The primary
complaint is usually spinal pain at the level of haematoma,
which may or may not be accompanied by radicular pain
appropriate to that level [2,3]. In some cases, irritability
may also be the only early presenting feature of SESH.
Cakir et al. [4] describe a case of a 9-year-old girl with a
spontaneous thoracic epidural haematoma initially mis-
diagnosed as Guillain–Barre
´syndrome. Intravenous
immunoglobulin therapy did not improve her symptoms,
prompting MR imaging which revealed a thoracic epidural
haematoma. This case illustrated the importance of ade-
quate, timely imaging in supporting the diagnosis.
Magnetic resonance imaging is considered to be the gold
standard in diagnosing spinal epidural haematomas [3,5].
Not only does MRI help to identify haematoma, but it also
helps to estimate its age and may help to identify the
presence of an underlying vascular malformation. Fur-
thermore, TRICKS MRI provides high-resolution vascular
imaging of the spine and is particularly invaluable when
studying arteriovenous malformations. TRICKS MRI thus
obviates the need for digital subtraction angiography,
which is considerably more invasive and technically
demanding.
In our first case, the child had a 5-week interval between
the onset of symptoms and the development of a neuro-
logical deficit. Patel et al. [2] describe patients primarily
complaining of pain ranging from several hours to several
days prior to the development of neurological signs.
Interestingly, Matsumae et al. [6] describe a case of a
13-year-old boy who developed spastic paraparesis which
was managed non-operatively as plain C-spine radiography
and CSF analysis were normal. Symptoms completely
resolved within a fortnight only to recur over 3 years later,
when myelography was performed revealing an epidural
mass lesion extending from C4 to C7 vertebral levels.
Subsequently laminectomy was performed to remove the
haematoma and intra-operative inspection did not reveal
the presence of any abnormal vessels or neoplastic process.
Fig. 4 Sagittal (a) and axial (b) T2-weighted images demonstrating
epidural lesion displacing spinal cord anteriorly
Eur Spine J
123
Abram et al. [7] report a 10-year-old girl with three
discrete separate episodes of paraparesis, with MR imaging
interpreted as unremarkable after the first two occasions.
After the third episode, MRI displayed a cervico-thoracic
junction epidural haematoma. Emergency decompressive
laminectomy and clot evacuation was performed with no
AVM identified. Interestingly, spinal angiography on that
admission was also interpreted as unremarkable and she
was discharged. Follow-up spinal angiography at 2 years
was suggestive of a vascular abnormality. However,
exploratory laminectomy did not reveal evidence of an
AVM.
The site of epidural haematoma in children is most
commonly dorsal [8] and occurs largely between C5 and
T1 vertebral levels. The underlying abnormality or pre-
cipitating cause of the haematoma is found in only
50–60 % of cases [9], though the bleeding is generally
thought to be venous in origin. Normal activities of daily
living may be responsible and cases of SESH have been
reported after activities such as sneezing, coughing and
bending over. These activities may cause changes in intra-
abdominal or intra-thoracic pressure altering the dynamics
of blood flow within the veins of the valveless epidural
venous plexus [2,8]. Other possible aetiologies include
coagulopathies, AVMs or an underlying neoplasm. In
many cases, an underlying abnormality, whether anatomi-
cal or physiological, may never be discovered [10].
The presence of abnormal epidural veins identified intra-
operatively was a feature our first case. Akutsu et al. [11]
describe a similar occurrence in a 27-year-old man with a
SESH whereby, after evacuation of the haematoma, the
presence of abnormally dilated epidural veins was dis-
covered. These veins were found to form an epidural
plexus and histological assessment showed varying calibre
of the vessels with muscular hyperplasia and collagenous
tissue contributing to their thickening. Some of the veins
even had an internal elastic lamina like an artery, but no
arterio-venous anastomosis was identified.
The presence of an AVM identified as the underlying
cause of SESH is rarer still, comprising 2 % of all spinal
lesions in children below 10 years of age [12]. Chen et al.
describe a case of an 8-year-old boy who developed back
pain radiating to the right anterior chest wall and became
paraplegic. Investigations showed an epidural lesion and
the operative finding was an AVM fed by a radicular artery
from the right T1–T2 root entry zone. However, in most
cases, an underlying vascular malformation as the cause for
SESH may never be found. Tewari et al. [9] reports a case
series of 3 children (aged 11, 8 and 5 years) each with
spontaneous epidural haematomas where no underlying
vascular malformation was identified.
The role of spinal digital subtraction angiography (DSA)
in the investigation of spinal epidural haematomas is
reserved for those cases where vascular malformations are
suggested on MR imaging [1,13]. Interestingly, Abram
et al. [7] report a case in a 10-year-old girl where DSA was
performed 8 days post-operatively and revealed no spinal
vascular abnormality. However, a repeat angiogram was
performed at 2 years was suggestive of ‘an unusual vas-
cular abnormality’ in the right-sided extradural space at C7
level. The patient then underwent subsequent exploratory
laminectomy, but operative findings revealed no vascular
abnormality.
In the case of a 14-year-old girl described by Rosenberg
[8], spinal DSA was performed at 2 weeks after the patient
was discharged from hospital. No spinal vascular
Fig. 5 Ruptured arteriovenous malformation surrounded by extensive haemorrhages. Thick, irregular blood vessel (V) and another vascular
channel revealed segmentally thinned elastic laminae (arrow), showing transition between artery and vein (elastic Van Gieson stain, 920)
Eur Spine J
123
abnormality was identified in this study. Other than the two
children mentioned, no cases of spontaneous spinal epi-
dural haematoma in the available literature underwent
spinal angiography.
The presence of a vascular malformation might be
anticipated in a child presenting with a spontaneous epi-
dural haematoma even without overt evidence of flow
voids on MRI scan. There is a potential for significant
arterial blood loss and the evacuation of such a haematoma
will be safest in the hands of a suitably experienced sur-
geon. Thirty-seven cases of SESH have been reported in
the world literature. These two cases are the first to present
with clinico-pathological correlation. The optimal clinical
follow-up for such cases remains to be defined as does the
natural history of asymptomatic lesions.
Acknowledgments The authors acknowledge the contribution of
Dr. Timothy Hampton, Consultant Radiologist at King’s College
Hospital, in providing radiological reports of the magnetic resonance
images in both cases.
Conflict of interest None.
References
1. van Heesewijk JP, Casparie JW (2000) Acute spontaneous spinal
epidural haematoma in a child. Eur Radiol 10:1874–1876
2. Patel H, Boaz JC, Phillips JP, Garg BP (1998) Spontaneous spinal
epidural hematoma in children. Pediatr Neurol 19:302–307
3. Pai SB, Maiya PP (2006) Spontaneous spinal epidural hematoma
in a toddler–a case report. Childs Nerv Syst 22:526–529. doi:10.
1007/s00381-005-0002-6
4. Cakir E, Karaarslan G, Usul H, Baykal S, Kuzeyli K, Mungan I,
Yazar U, Peksoylu B, Aynaci M, Cakir F (2004) Clinical course
of spontaneous spinal epidural haematoma mimicking Guillain–
Barre syndrome in a child: a case report and literature review.
Dev Med Child Neurol 46:838–842
5. Nagel MA, Taff IP, Cantos EL, Patel MP, Maytal J, Berman D
(1989) Spontaneous spinal epidural hematoma in a 7-year-old
girl. Diagnostic value of magnetic resonance imaging. Clin
Neurol Neurosurg 91:157–160
6. Matsumae M, Shimoda M, Shibuya N, Ueda M, Yamamoto I,
Sato O (1987) Spontaneous cervical epidural hematoma. Surg
Neurol 28:381–384
7. Abram HS, DeLaHunt MJ, Merinbaum DJ, Hammond DN (2007)
Recurrent spontaneous spinal epidural hematoma in a child: first
case report. Pediatr Neurol 36:177–180. doi:10.1016/j.pedia
trneurol.2006.09.009
8. Rosenberg O, Itshayek E, Israel Z (2003) Spontaneous spinal
epidural hematoma in a 14-year-old girl. Case report and review
of the literature. Pediatr Neurosurg 38:216–218. doi:10.1159/
000069091
9. Tewari MK, Tripathi LN, Mathuriya SN, Khandelwal N, Kak VK
(1992) Spontaneous spinal extradural hematoma in children.
Report of three cases and a review of the literature. Childs Nerv
Syst 8:53–55
10. Wittebol MC, van Veelen CW (1984) Spontaneous spinal epi-
dural haematoma. Etiological considerations. Clin Neurol Neu-
rosurg 86:265–270
11. Akutsu H, Sugita K, Sonobe M, Matsumura A (2003) A case of
nontraumatic spinal epidural hematoma caused by extradural
varix: consideration of etiology. Spine J 3:534–538
S1529943003001530 [pii]
12. Chen CC, Wang CM, Chu NK, Wu KP, Tang SF, Wong AM
(2008) Spinal cord arteriovenous malformation presenting as
chest pain in a child. Spinal Cord 46:456–458. doi:10.1038/sj.sc.
3102133
13. Fountas KN, Kapsalaki EZ, Robinson JS (2006) Cervical epidural
hematoma in children: a rare clinical entity. Case report and
review of the literature. Neurosurg Focus 20:E6
Eur Spine J
123
