Case reports
Spinal epidural hematoma without lesion of the
Report of 4 cases

L. C a l l i a u w * , M. D h a r a * , F. M a r t e n s * , and L. V a n n e r e m * *

Introduction
Spinal fracture or dislocation can provoke a spinal epidural hematoma by extravasation from
the bone. This type of case will not be discussed
in this paper. On the contrary spinal epidural
hematoma without a lesion of the spine occur
relatively rarely. Foo and Rossier reviewed the
world literature in 1981 and were able to collect
158 cases 1. The analysis of these cases enabled
them to find a correlation between the presurgical neurological status and the postoperative
outcome. Their correlation can be summarized
as follows: Postoperative return of motor function was noted in 95.3%, 87% and 45.3% of the
patients with incomplete sensorimotor, incomplete sensory but complete motor and complete
sensorimotor lesion respectively. Complete sensorimotor recovery occurred in 41.9%, 26.1%
and 11.3% of these 3 groups of patients, respectively. Since then other cases have been reported 2.
It is confirmed that pre-operative neurological status and delay of surgical treatment are
the principle factors in determining the outcome. There is still doubt about the source of
bleeding: whether the origin is arterial or venous, is a matter for debate.
In order to be able to participate in this discussion we present four cases of spinal epidural
hematoma.

Summary
The authors report 4 cases of spinal epidural
hematoma without a lesion of the spine. The
medical literature relating to this subject is
reviewed with special reference to the pathophysiology of the bleeding. There are arguments in favor of the venous origin as well the
arterial origin of the bleeding.
As far as recovery following surgical treatment
is concerned, the history of our cases confirms
the opinion that the prognosis depends on two
factors: the delay between onset of symptoms
and operation and the severity of the neurological deficit.
Key words: Spinal epidural hematoma, spontaneous hematoma.

Case reports
Case 1

A 13-year-old boy had a fall from a ladder of 5
meters; he was admitted to the hospital on the
same day. On admission he was conscious and
complaining of pain in the left wrist. H e was
treated for luxation of the left wrist.
Roentgenograms of the lumbar vertebra did
not show any evidence of fracture. The patient
was discharged on the same day. One week
later, at night, he awoke because of severe pain

* Department of Neurosurgery, University hospital, Ghent (Belgium), * * Department of Neurosurgery, St, Janshospitaal,
Brugge (Belgium).
Address for correspondence and reprint requests: L. Calliauw, Department of Neurosurgery, University Hospital, De Pintelaan
185, 9000 Ghent, Belgium.
Accepted 2.11.87

Clin Neurol Neurosurg 1988. Vol. 90-2.

131

in the cervico-dorsal region of the spine with
weakness of both lower limbs. This weakness
disappeared after 10 minutes. The following day
he also complained of backache aggravated by
walking; he was readmitted to the hospital.
Neurological examination revealed tenderness
over the lower cervical and T8-T9 thoracic vertebrals levels. Motor system examination of the
lower limbs showed normal power, tone and
bilateraly symmetrical knee and ankle reflexes.
Plantar reflexes were flexor. Motor system examination of the upper limbs was normal. Sensory examination revealed hyperpathia from T8
to T12 and below this level hypesthesia was
present. Micturition and defecation was normal. Bleeding and clotting time was normal.
Blood analysis ruled out bleeding diathesis.
No fractures were seen on spinal radiograms.
Myelography revealed widening of the ventral
epidural space from the lower cervical vertebrae
up to the level of T9; the spinal cord was pushed
backward. The cerebrospinal fluid analysis was
normal.
The computerized tomography suggested spinal epidural hematoma (Fig. 1). On 30.06.86 a
laminectomy from T4 to T7 was done. Ventral
to the dura an organized blood clot was found;
the blood clot was evacuated. No traumatic lesions were seen. The clot did not reveal any
abnormal vessels or neoplastic cells (histological
examination).
Postoperatively the patient was relieved of
pain and had improvement in his sensory deficits. At the time of discharge (12 days after
admission) there were no neurological deficits.
The patient is followed up at regular intervals;
he has no complains.
Case 2
A 20-year-old man was admitted on 01.09.86
following a motorcycle accident. There was a
short period of loss of consciousness. He was
treated for fracture of the left humerus. Roentgenograms of skull and spine did not reveal
abnormalities. He was discharged 8 days after
admission to the hospital. Three months later he
started to complain of weakness of the lower
limbs. He also complained of decreased sensation in his lower limbs and thoracic region.
Neurological examination at that time revealed paresis of the lower limbs with hypesthe132

Fig. 1.
Case 1. ComputerizedTomogramsuggestinga ventral epidural hematoma.

sia of the left lower limb. Knee reflex and ankle
reflex were bilaterally symmetrical. Babinski reflex had plantar response. Micturition and defecation were normal.
Cerebrospinal fluid analysis showed an elevated protein level (440 mg/100 ml). Electromyography showed diffuse peripheral neurogenic changes, considered then as characteristic
of a Guillain-Barr6 syndrome. The paresis of
the lower limb being progressive, he was treated
with methyl-prednisolone (40 mg 2 times a day)
and with vitamins. The patient was admitted to
our department one month later.
Neurological examination revealed paresis of
both lower limbs. The right lower limb was more
paretic than the left one. Exaggerated knee and
ankle reflexes were present bilaterally. Ankle
and patellar clonus was present. The Babinski
reflex on the right side was positive. The Babinski on the left side was indifferent.
Sensory examination revealed decreased epicritic, protopathic sensation at T5-T6 dermatome level. There was also decreased vibration
sensibility over the lower limbs. The position
sense was normal. In the upper limb the power,
tone and reflexes were bilaterally symmetrical
and normal. Micturition and defecation were
normal.
Blood analysis ruled out bleeding diathesis.
No fractures or other abnormalities were seen
on spinal radiograms. Myelography revealed a
complete stop at the level of T3 with compression of the spinal cord by an extradural space

occupying lesion up to C7.
A SSEP done on 10.12.86 by stimulating the
nervus fibularis and tibialis above the knee revealed absence of response on the dorsal medulla of the upper thoracic spinal cord.
A computerized tomography was suggestive
of spinal epidural hematoma. On 11.12.86 a
laminectomy from T1 to T4 was done. Dorsal to
the dura there was a xanthochromatic fluid collection with organized membranes. Evacuation
of the chronic epidural hematoma with the organized membranes was done. The dura was
opened at the level of T4 and revealed a normal
medulla. No traumatic lesions were seen. The
histological examination of the organized membranes showed epidural fat, few dilated blood
vessels; this was interpreted as fibrosis of the
dura.
Post-operatively there was improvement in
motor power of lower limbs with diminution of
exaggerated knee and ankle reflexes. Babinski
was positive on the right side. The sensory deficit remained the same.
A SSEP done 4 weeks after the operation when motor recuperation was already seen revealed answers over the right side of the dorsal medulla of the upper thoracic spinal cord but
on the left side there was no response.
Ct-scan (19.12.86) showed a backward shift of
the medulla from T1 to T4. An N M R scan done
on 27.01.87 showed an extradural fluid collection with atrophy of the spinal cord at the level

of T1 to T4 (Fig. 2a and 2b). The patient received intensive physiotherapy. A new SSEP
examination revealed reproducable responses
over right and left side of the dorsal medulla of
the upper thoracic spinal cord.
At that time (08.04.87) the neurological examination revealed normal motor power in both
lower limbs. The knee and ankle reflexes bilaterally were brisk. The ankle clonus was positive
bilaterally. The sensory examination revealed
improved proprioception and vibration sense at
T5-T6. The epicritic sensation remained the
same as before.
Case 3
A 39-year-old patient was admitted on 17.03.87
with a history of sudden onset of lower backache
with weakness of lower limbs. H e also complained of diffuse loss of sensation in both lower
limbs. There was a history of urinary retention
and constipation. The patient was known as
having hemophilia (factor VIII A) at an earlier
age.
There was history of occasional bleeding
tendencies. H e was admitted to our department
on 19.03.87. Neurological examination revealed
paraplegia with absence of knee and ankle reflexes. The plantar reflexes were flexor. T h e r e
was loss of proprioceptive, epicritic, thermoanalgesic, vibration and position senses below
the level of L2. There was loss of anal tone and
anal reflex was absent. There was loss of cre-

Fig. 2a and b.
Case 2. N M R scan which shows (a) an extradural fluid collection and (b) an atrophy of the spinal cord.

133

masteric and bulbocavernosus reflexes.
No abnormalities were seen on spinal radiograms.
A myelogram revealed a total stop at the level
of L4 with extradural compression of the cord
up to T8. On 19.03.87 a laminectomy from T8 to
L4 was done. Dorsal to the dura a large hematoma was evacuated. The antihaemophilic factor
was administered pre-, per- and post-operatively to stabilize factor 8.
14 days after the operation he was able to feel
some sensation in both lower limbs and showed
contraction in the right quadriceps. SSEP done
on 31.03.87 showed reproducable responses
over the dorsal medulla of the lower thoracic
spinal cord in the left side but no responses on
the right side. The patient received intensive
physiotherapy.
On 03.04.87 motor system examination of the
lower limbs showed normal tone, power grade
II, knee and ankle reflexes bilaterally symmetrical exaggerated. Plantar reflexes were normal.
There was improvement in epicritic, proprioceptive, thermo-analgesic, vibration and position sensation in both lower limbs. All modalities of sensation were absent at $1 to $5
dermatome level. Anal tone was lost and anal
reflex absent, bulbocavernosus reflex absent,
cremasteric reflex was positive on the right side
and absent on the left side.
The last follow-up done on 15.05.87 shows
normal tone in both lower limbs. The power is
grade IV. Bilaterally symmetrical exaggerated
knee and ankle reflexes. Sensory examination
revealed further improvement in all modalities
of sensation in both lower limbs. There is some
epicritic and proprioceptive sensation in $1 to
$5 region. Anal tone and anal reflexes absent.
Bulbocavernosus and cremasteric reflexes present. Patient needs intermittent catheterization
every 3 hours. During defecation he does not
feel the sensation of the stools.
Case 4
A 63-year-old man experienced lower lumbar
pain on 21.03.83. The following day he complained of weakness of the lower limbs with
retention of urine; he was admitted to our department. Neurological examination revealed
paralysis of the right lower limb and paresis of
the left lower limb. The knee and ankle reflexes

134

were absent bilaterally. The plantar reflex had a
flexor response bilaterally. Sensory examination revealed saddle anaesthesia and decrease
of all modalities of sensation in both lower
limbs. There was urinary retention. Blood analysis did not reveal any bleeding diathesis. Spinal
radiograms were normal.
A myelogram on 29.03.83 revealed a stop at
L1 due to compression by an epidural process
(Fig. 3).
A laminectomy T12 to L3 was done and an
epidural hematoma located dorsal to the dura
was evacuated. Histological examination revealed epidural fat and hematoma; there was no
evidence of abnormal vessels or neoplastic cells.
Post-operatively there was motor improvement
in both lower limbs. He was discharged from
hospital on 28.04.83. At the time of discharge he
had paresis of right foot due to involvement of
tibialis anterior. There was diffuse hypoesthesia
in both lower limbs. He was able to walk with
the help of a stick. He received physiotherapy.
A follow-up on 16.03.83 was done. Segmentary
muscle testing did not show any deficit.
Knee reflex was bilaterally symmetrical. Ankle reflex on the right side was absent. The
plantar reflex had a flexor response. Sensory
examination revealed absence of all modalities
of sensation in L5, S1 $2 of the right lower limb.
There was good control of micturition and defecation. Patient is followed up at regular intervals.
Fig. 3.
Case 4. Myelogram with a stop at L1 due to compression by
an epidural process.

Table 1: Summaryof the 4 cases
Case

Pre-operative
neurologicalstatus

Level of
laminectomy

Time between
onset of
symptoms and
operation

Results of surgical
treatment

Partial sensory
deficit below T8

T4-T7

2 days

complete recovery

Incomplete sensorymotor deficit

T1-T4

10 days

Complete sensorymotor deficit
Urinary retention

T8-L4

2 days

Paralysis right
lower limb
Paresis left lower
limb
Sensory deficit
Urinary retention

T12-L3

2 days

motor recovery
sensory deficit remains
Improvement of motor
and sensory function but
deficit remains. No
bladder control.
motor recovery
sensoryimprovement
spontaneous micturation

Discussion
The cause of these hematomas is often unknown; therefore some of them are considered
to be spontaneous. The term 'spontaneous'
means without an identified etiology. According to this definition, only our case 4 can be
called 'spontaneous'.
The other spinal epidural hemorrhages have
been associated with coagulopathy, A V M or
some iatrogenic causes (lumbar puncture, epidural anesthesia, surgical procedures).
The relationship to trauma is controversal except in cases with lesions on the spinal radiograms.
Some authors have included patients with minor trauma into the category of spontaneous
hematomas 3. As pointed out by G h a n e m and
Ivan it is difficult to draw the line between mild
trauma and the activities of daily living 4. The
same can be said for hematomas occurring after
lifting heavy weights 1, pulling a chair or in the
course of whooping coughS; is that spontaneous,
or is the increase of intrathoracic/intra-abdominal pressure associated to trauma? Whatever
the etiology may be, the mechanism of the
bleeding is also a mystery. Supporters of the
venous origin argue that the epidural veins belong to a primitive venous system which has no
venous valves. This valveless plexiform network
is connected above with the cranial dural sinus,
below with the pelvic veins and in the neck and

trunk with the azygos and caval systems. Hence
an abrupt pressure change in the visceral cavities can be transmitted directly to pre-existing
pathological epidural veins and provoke their
rupture. It is very probable that such an epidural
venous bleeding will be quickly tamponaded by
the expanded dural sac as the intrathecal pressure is higher than the pressure in the veins. To
explain a progression of the bleeding, leading to
compression of the cord, a low pressure in the
epidural space has to be supposed.
The venous origin is therefore problematic
although the clinical course of our case 1 and 2,
in which the clinical signs of cord compression
started respectively 9 days and 3 months after a
traumatic event, fall in favour of a venous origin
of the bleeding.
Those who favour the hypothesis that the
hemorrhage originates from an arterial source
stress the rapidity with which some patients become paraplegic after the onset of pain, a phenomenon which does not occur in the cases we
have outlined.
The arterial source has to be found in ascending and descending arteries running in the spinal
dura. Those arteries are fed by arteries that run
with the nerve sheath. According to Beatty and
Winston 3 some of these bridging-arteries - specialy in the cervical region - have a free portion.
Stretching of these arteries could result in hemorrhage. The above-mentioned authors bear
out their hypothesis by the fact that epidural
135

spinal hematomas frequently originate in the
cervical region and that, in the patients they
have operated, the bleeding was predominantly
lateral.
None of our 4 cases corresponds to these criteria. In addition it seems difficult to explain
why nobody has seen bleeding arteries after removing a clot, especially after an early operation. It is unlikely that all these arteries have
been thrombosed.
The origin of the hemorrhage remains a mystery, and the hypothesis of the 'cryptic vascular
anomalies' of Robertson is pure hypothesis 6, it
has never been proved.
Concerning the prognosis, the results of our 4
cases can only accord with the conclusions made
by Foo and Rossier 1.
- Operating quickly on an incomplete lesion
means a very good chance of complete recovery (case 1 and 4).
An incomplete lesion operated with delay
can recover but with greater chance of deficit
(case 2).

136

-Even
a complete sensory-motor lesion
should be operated on. The absence of sensorimotor functions before operation does not
necessarily indicate a poor prognosis, since
partial neurological return is noticed (case

3).
The SSEP seemed not to be of great value since
their return was always preceded by clinical improvement.
References

1

FOO DK, ROSSlER AB. Pre-operative Neurological S t a t u s i n
predicting surgical outcome of spinal epidural hematomas. Surgical Neurology 1981; 15:389-401.
2 KLOSSEKVH, HOLLERE. Zur Problematik der spontanen
spinal epidurale H~imatome. Zentralblatt fiir Neurochirurgie 1984; 45:116-123.
3
BEATrY RM, WINSTON KR. Spontaneous cervical epidural
hematoma. J Neurosurg 1984; 61:143-148.
4
GHANEM Q, IVAN LP. Spontaneous spinal epidural h e m a toma in an 8-year-old boy. Neurology 1978; 28:832-92.
5 JACKSONrE. Spontaneous spinal epidural hematoma coincident with whooping cough. J. Neurosurg 1963;
20:715-7.
6
ROBERTSONWC Jr, LEE YN, EDMONSON MB. Spontaneous
spinal epidural hematoma in the young. Neurology
1979; 29:120-2.