Cerebellar Hemorrhage after Spinal Surgery

Cerebellar Hemorrhage after Spinal
Surgery: Report of Two Cases and
Literature Review
Jonathan A. Friedman, M.D., Robert D. Ecker, M.D.,
David G. Piepgras, M.D., Derek A. Duke, M.D.
Department of Neurologic Surgery, Mayo Clinic, Rochester, Minnesota

OBJECTIVE AND IMPORTANCE: Cerebellar hemorrhage remote from the
site of surgery may complicate neurosurgical procedures. We describe
our experience with two cases of cerebellar hemorrhage after spinal
surgery and review the three cases previously reported in the literature to
determine whether these cases provide insight regarding the pathogenesis
of remote cerebellar hemorrhage.
CLINICAL PRESENTATION: One of our patients developed cerebellar hemorrhage in the vermis and right hemisphere after transpedicular removal
of a partially intradural T9–T10 herniated disc with the patient in the
prone position. The other patient developed cerebellar hemorrhage in the
vermis and bilateral hemispheres after L3–S1 decompression and instrumentation with the patient in the prone position, during which the dura
was inadvertently opened.
INTERVENTION: The first patient was treated conservatively and had mild
residual dysarthria and gait ataxia 2 months after surgery. The second
patient underwent exploration and revision of the lumbar wound with
primary dural repair. The cerebellar hemorrhage was treated conservatively, and the patient had mild dysarthria and ataxia 1 month after
surgery.
CONCLUSION: Cerebellar hemorrhage must be considered in patients with
unexplained neurological deterioration after spinal surgery. Dural opening with loss of cerebrospinal fluid has occurred in every reported case of
cerebellar hemorrhage complicating a spinal procedure, supporting the
hypothesis that loss of cerebrospinal fluid is central to the pathogenesis of
this condition. Because remote cerebellar hemorrhage can occur after
procedures with the patient in the supine, sitting, and prone positions,
patient positioning seems unlikely to play a causative role in its
occurrence. (Neurosurgery 50:1361–1364, 2002)
Key words: Cerebellum, Cerebrospinal fluid, Intracranial hemorrhage, Remote hemorrhage, Spine

C

erebellar hemorrhage remote from
the site of operation may complicate
intracranial surgery. The occurrence of
remote cerebellar hemorrhage (RCH) is
most frequently associated with frontotemporal craniotomy with the patient in
the supine position (1, 3, 5, 7, 10, 12).

Patients who undergo repair of an unruptured aneurysm or a temporal lobectomy seem particularly predisposed to
RCH (5).
Although the underlying pathophysiology of RCH has not been definitively
elucidated, recent reports suggest that
Neurosurgery, Vol. 50, No. 6, June 2002

1361

the condition is venous in origin (3, 5–7,
12). Cerebellar “sag” as a result of intraoperative cerebrospinal fluid (CSF) aspiration has been proposed to cause
stretching and occlusion of bridging cerebellar veins coursing in the cephalad
direction, leading to hemorrhagic venous infarction (5, 9, 12, 13). The relevance of patient positioning in the development of RCH is unclear.
If acute CSF hypovolemia does play a
causative role in the pathophysiology of
RCH, the phenomenon would be expected to occur after spinal procedures
during which CSF was lost. We report
our experience with two cases of cerebellar hemorrhage after spinal procedures and review the relevant literature
describing this occurrence.

CASE REPORTS
Patient 1
A 43-year-old man presented with
myelopathy and was found to have a
herniated T9–T10 disc. This patient was
briefly described previously (4). He had
no history of arterial hypertension,
trauma, hypercoagulable state, or recent
infection. The patient underwent removal of the disc via a posterolateral,
transpedicular approach in the prone
position. The disc had eroded through
the dura, and opening of the dura and
sectioning of the dentate ligament on
one side was necessary. The patient
awoke from surgery neurologically intact. Neither intraoperative nor postoperative blood pressures were substantially elevated. At 12 hours after
surgery, the patient developed some
dysarthria, followed by progressive
ataxia over the course of the next 24
hours. A computed tomographic (CT)
scan was obtained 48 hours after surgery, revealing cerebellar hemorrhage
in the vermis and right hemisphere (Fig.
1). Laboratory studies, including platelet count, prothrombin time, and partial
thromboplastin time, were within normal ranges. Bilateral vertebral angiography was normal. The patient was
treated conservatively, and his deficits
did not progress. At 2 months after sur-

1362

Friedman et al.
FIGURE 2. Patient 2. A, noncontrast CT of the head demonstrating acute cerebellar
hemorrhage within the superior vermis and both hemispheres; B, coronal fluidattenuated inversion recovery
magnetic resonance scan demonstrating early subacute
intraparenchymal cerebellar
hemorrhage.

FIGURE 1. Patient 1. CT scan of the
head without contrast obtained 2 days
after surgery, demonstrating acute hemorrhage in the cerebellar vermis and
superior aspect of the right cerebellar
hemisphere.
gery, he had slight residual dysarthria
and gait ataxia.

Patient 2
A 56-year-old woman underwent an
L3–S1 decompressive laminectomy and
pedicle screw fixation at an outside hospital for degenerative lumbar spondylolisthesis. The patient had no history
of arterial hypertension, trauma, hypercoagulable state, or recent infection. The
patient awoke from her initial surgery
neurologically intact. Neither intraoperative nor postoperative blood pressures
were noted to be elevated. On postoperative Day 2, the patient developed a
severe occipital headache and vomiting,
followed by the development of dysarthria and bilateral upper-extremity
ataxia. A head CT scan revealed a bilateral hemorrhage in the superior folia of
the cerebellar vermis and the cerebellar
hemispheres (Fig. 2A). The patient was
transferred to our institution.
On admission, the patient was alert
and complained of a severe occipital
headache. Neurological evaluation revealed dysarthria and bilateral ataxia. A
repeat CT scan was unchanged. Laboratory studies, including platelet count,
prothrombin time, and partial thromboplastin time, were within normal
ranges. A drainage catheter connected
to bulb suction remained in the subfascial space, and pink-tinged watery fluid
was aspirated. Magnetic resonance imaging revealed early subacute hemorrhage within the cerebellum without ad-

ditional abnormalities (Fig. 2B). The
patient’s headache continued to worsen,
but it was relieved by her lying supine.
A repeat CT scan demonstrated no enlargement of the hemorrhage or hydrocephalus. More than 500 ml of watery
fluid suggestive of CSF were drained
from the wound via the suction catheter.
The drain was removed on the evening
of postoperative Day 3. The next morning, the patient’s headache had resolved
completely, even when she was in the
upright position.
On postoperative Day 4, clear fluid
began to leak from the lumbar incision.
A CT myelogram was performed, which
revealed a posterior pseudomeningocele with contrast extravasation at both
the L3 and L5 levels. The patient was
returned to the operating room, and a
3-mm dural tear at the caudal aspect of
the previous laminectomy with copious
CSF leakage was identified and repaired. The patient had no further headache but had residual mild dysarthria
and ataxia at discharge 3 weeks after
surgery.

DISCUSSION
Cerebellar hemorrhage after a spinal
procedure has been reported previously
and should be considered as a potential
source of unexplained neurological decline after spinal surgery. The phenomenon was first described by Chadduck
(2) in 1981, after a cervical laminectomy
in which the dura was opened widely
and the dentate ligaments were sectioned while the patient was in a sitting
position. In 1994, Mikawa et al. (8) reported a cerebellar hemorrhage after
durotomy and revision of a cervical fusion. Most recently, Morandi et al. (9)
described a patient who developed cerebellar and supratentorial hemorrhage
Neurosurgery, Vol. 50, No. 6, June 2002

after resection of an intradural cervical
schwannoma.
An important common feature among
these three previously reported cases
and our two cases is the intradural nature of the procedures, with consequent
loss of CSF. In an analysis of RCH after
supratentorial craniotomy at our own
institution, almost all cases of RCH were
associated with procedures that involved opening the CSF cisterns or the
ventricular system (5). There is a growing consensus that hemorrhagic venous
infarction underlies most cases of RCH
(3, 5–7, 12). It is possible that downward
cerebellar displacement, or “sag,” resulting from intraoperative loss of CSF
causes transient stretch occlusion of superior cerebellar veins draining in the
cephalad direction into the deep venous
system (5, 6, 12, 13). This may cause
intracerebellar hemorrhage in patients
with insufficient venous collaterals.
The pattern of cerebellar hemorrhage
on the postoperative CT scans in our
two patients after spinal surgery is similar to that seen after supratentorial operations (1, 3, 5, 12). On the basis of our
two cases and the three previous reports, it would initially appear that cerebellar hemorrhage after spinal procedures
is
typically
symptomatic.
Postoperative imaging of the brain,
however, would be distinctly uncommon in an asymptomatic patient after a
spinal procedure. Thus, asymptomatic
cerebellar hemorrhages are unlikely to
be detected, and the true incidence of
this occurrence may be higher than suggested by existing reports. In contrast,
RCH after supratentorial surgery is frequently asymptomatic and detected incidentally on postoperative scanning.
Our patient with cerebellar hemorrhage after lumbar fusion with dural

Cerebellar Hemorrhage after Spinal Surgery
tear had a postural headache highly
suggestive of intracranial hypotension,
which resolved when suction drainage
of the wound was discontinued. This
case supports the association between
low CSF pressure and cerebellar hemorrhage remote from the operative site. In
patients with intracranial hypotension
caused by spontaneous CSF leakage
from spinal arachnoid diverticula, however, cerebellar hemorrhage has not
been described. This probably relates to
the chronicity of the disorder, during
which time collateral venous drainage
would be expected to develop. With the
more abrupt loss of CSF that may occur
during and immediately after spinal
procedures, such collateral venous
drainage would not have time to
develop.
The relevance of patient positioning
in the development of RCH after supratentorial surgery is unclear. Seoane and
Rhoton (11) have suggested that jugular
vein compression at the transverse process of C1 as a result of head positioning
may be a causative factor in RCH. If
RCH represents a hemorrhagic venous
infarction, it is plausible that venous
compression in the neck could cause or
exacerbate the process. The occurrence
of cerebellar hemorrhage after spinal
procedures with the patient in both the
prone and sitting positions, however,
suggests that position is not relevant to
the pathogenesis of this complication after spinal procedures and casts some
doubt on its role in the development of
RCH after supratentorial procedures.

CONCLUSIONS
Cerebellar hemorrhage must be considered in patients with unexplained
neurological deterioration after spinal
surgery. Dural opening is common to
every reported case of a spinal procedure complicated by cerebellar hemorrhage, supporting the hypothesis that
CSF loss is central to the pathogenesis of
this condition. Because RCH can occur
after procedures with the patient in the
supine, sitting, and prone positions, patient positioning seems unlikely to play
a causative role in its occurrence.

ACKNOWLEDGMENT
We are deeply indebted to Mary
Soper for expert assistance in manuscript preparation.
Received, August 3, 2001.
Accepted, September 22, 2001.
Reprint requests: Jonathan A. Friedman,
M.D., Department of Neurologic Surgery, Joseph 1-229, Saint Mary’s Hospital, 1216 Second Street SW, Rochester, MN 55905.
Email: friedman.jonathan@mayo.edu

REFERENCES
1. Brisman MH, Bederson JB, Sen CN, Germano
IM, Moore F, Post KD: Intracerebral hemorrhage occurring remote from the craniotomy
site. Neurosurgery 39:1114–1122, 1996.
2. Chadduck WM: Cerebellar hemorrhage complicating cervical laminectomy. Neurosurgery
9:185–189, 1981.
3. Cloft HJ, Matsumoto JA, Lanzino G, Cail WS: Posterior fossa hemorrhage after supratentorial surgery.
AJNR Am J Neuroradiol 18:1573–1580, 1997.
4. Friedman JA, Piepgras DG: Remote hemorrhage. J Neurosurg 95:279–280, 2001 (letter).
5. Friedman JA, Piepgras DG, Duke DA, McClelland RL, Bechtle PS, Maher CO, Morita A,
Perkins WJ, Parisi JE, Brown RD Jr: Remote
cerebellar hemorrhage after supratentorial surgery. Neurosurgery 49:1327–1340, 2001.
6. Konig A: Postoperative hemorrhage. J Neurosurg
86:916–917, 1997 (letter).
7. Konig A, Laas R, Herrmann HD: Cerebellar
hemorrhage as a complication after supratentorial craniotomy. Acta Neurochir (Wien) 88:104–
108, 1987.
8. Mikawa Y, Watanabe R, Hino Y, Ishii R, Hirano
K: Cerebellar hemorrhage complicating cervical
durotomy and revision C1–C2 fusion. Spine
19:1169–1171, 1994.
9. Morandi X, Riffaud L, Carsin-Nicol B, Guegan Y:
Intracerebral hemorrhage complicating cervical
“hourglass” schwannoma removal. J Neurosurg
94[Suppl 1]:150–153, 2001.
10. Papanastassiou V, Kerr R, Adams C: Contralateral cerebellar hemorrhagic infarction after
pterional craniotomy: Report of five cases and
review of the literature. Neurosurgery 39:841–
852, 1996.
11. Seoane E, Rhoton AL Jr: Compression of the
internal jugular vein by the transverse process
of the atlas as the cause of cerebellar hemorrhage after supratentorial craniotomy. Surg
Neurol 51:500–505, 1999.
12. Toczek MR, Morrell MJ, Silverberg GA, Lowe
GM: Cerebellar hemorrhage complicating temporal lobectomy. J Neurosurg 85:718–722, 1996.
13. Yoshida S, Yonekawa Y, Yamashita K, Ihara I,
Morooka Y: Cerebellar hemorrhage after supratentorial craniotomy: Report of three cases.
Neurol Med Chir (Tokyo) 30:738–743, 1990.

COMMENTS
Friedman et al. report two cases of patients who experienced unexplained cereNeurosurgery, Vol. 50, No. 6, June 2002

1363

bellar hemorrhage after spinal surgery. In
both patients, the spinal subarachnoid
space was entered. In the second patient,
cerebrospinal fluid (CSF) was actively aspirated from the wound by a bulb suction
drain. We do not know whether such a
drain was placed in the first patient. The
resulting cerebellar hemorrhages are diffuse and relatively widespread. This is
similar to the imaging picture seen with
remote cerebellar hemorrhage after supratentorial surgery.
The authors hypothesize that the hemorrhages are the result of venous stasis. I
am not convinced that these are venous
infarctions. The hemorrhages span more
than the territory of a single vein. The
computed tomographic images and magnetic resonance imaging scan do not demonstrate the low density frequently seen
surrounding a venous infarction. If the
hemorrhage arose from diffuse cerebellar
venous stasis, I would have expected to
see diffuse cerebellar swelling and obstructive hydrocephalus.
Remote cerebellar hemorrhages remain a perplexing problem. I do not
understand their pathophysiology or
any avoidable predisposing factor.
Allan H. Friedman
Durham, North Carolina
Friedman et al. provide us with two
interesting examples of cerebellar hemorrhage after spinal surgery. In both cases,
the dura was opened and CSF was lost.
Many readers are undoubtedly familiar
with the phenomenon of remote cerebellar hemorrhage in the setting of intracranial surgery. As the authors point out,
hemorrhagic venous infarction secondary
to obstruction of draining veins caused by
a “sagging cerebellum” has been the most
popular explanation for this phenomenon. In the two examples that I have personally observed, I was relatively convinced that this was indeed the
explanation. I believe that the two present
cases and the previously reported cases in
association with spinal surgery add credence to this theory. I disagree, however,
with the authors’ contention that position
plays no role in the phenomenon. I think
the authors would agree that although the
overall position is not relevant, the head
does need to be elevated above the site of
the CSF leak for the sagging of the cere-

1364

Friedman et al.

bellum to occur. One could argue that use
of a very slight degree of “head-down”
positioning when one contemplates opening the spinal dura might prevent this
complication. This, of course, should be
done only when intracranial hypertension
is not an issue.
Robert E. Breeze
Denver, Colorado
The authors describe two cases of patients presenting with a cerebellar hematoma after spine surgery and perform a
comprehensive literature review regarding this rare complication. In each patient,
the dura was opened, intentionally or not,
with an associated loss of CSF. In both
patients, initial postoperative evaluation
demonstrated no neurological deficits.
Symptoms were noted as early as 12
hours after surgery in one patient. Each

patient was managed conservatively regarding the cerebellar hemorrhage, without progression of symptoms. Both patients demonstrated mild residual deficits
consisting of dysarthria and gait ataxia at
the last follow-up evaluation. The authors’ literature review showed that only
three other cases of cerebellar hemorrhage
after spine surgery have been reported. In
all cases, the dura was opened, with loss
of CSF. On the basis of this association
and previous assumptions after intracranial surgery, the authors hypothesize that
cerebellar “sag” as a result of intracranial
hypovolemia is the primary event leading
to RCH. The downward displacement of
the cerebellum leads to stretch occlusion
of the bridging veins and venous
infarction.
This article provides a description of an
extremely rare complication associated

Neurosurgery, Vol. 50, No. 6, June 2002

with intradural spine surgery. As the authors correctly point out, the exact incidence of this complication is unknown,
and it is likely to remain so, because cranial imaging after spine surgery is not
routinely performed. Although definitive
conclusions are lacking because of the
small number of patients presenting with
this complication, the authors are able to
formulate several reasonable assumptions
on the basis of the associations observed.
Given the rarity of this complication, it is
virtually impossible to formulate preventive measures; however, the possibility
needs to be considered and aggressively
evaluated should unexplained neurological deterioration occur after any intradural spine procedure.
Paul C. McCormick
New York, New York