Spinal Hodgkin’s Disease
torts the thoracic aorta is not much different in Figure 4A than it appears in Figure
2A.
Mark N. Hadley
Birmingham, Alabama
This case report by Citow et al. details
the very unusual incidence of primary extranodal Hodgkin’s disease of the spine
presenting as spinal cord compression.
These authors describe a case of initial
misdiagnosis in today’s medicolegal climate. The initial radiological and clinical

presentations were certainly consistent
with tuberculosis of the spine and warranted presumptive treatment. The authors were very astute in pursuing the
correct diagnosis after clinical progression
despite appropriate medical treatment.
My initial surgical procedure would
have been an anterior transthoracic approach. The posterior parascapular technique is certainly an option but results in
a longer fixation segment (nine levels in
this case) and probably would be associated with a higher risk of neurological
complications in this patient. In general

Conus Perimedullary Arteriovenous
Fistula with Intracranial Drainage:
Case Report
G. Edward Vates, M.D., Ph.D., Alfredo Quiñones-Hinojosa, M.D.,
Van V. Halbach, M.D., Michael T. Lawton, M.D.
Departments of Neurological Surgery (GEV, AQ-H, MTL) and Interventional Neuroradiology
(VVH), University of California, San Francisco, San Francisco, California

OBJECTIVE AND IMPORTANCE: Perimedullary arteriovenous fistulae (AVFs)
do not commonly present with subarachnoid hemorrhage or intracranial
venous drainage causing neurological symptoms. We present a case with
both of these features. The patient was inadvertently treated for an unruptured intracranial aneurysm before his true problem was recognized.
CLINICAL PRESENTATION: A 65-year-old man presented with sudden-onset
lower-extremity weakness, diplopia, nausea, and dysarthria on the day of
admission. A lumbar puncture documented subarachnoid hemorrhage,
and imaging studies revealed a left middle cerebral artery aneurysm. It
was noted during surgery that this aneurysm was unruptured, and the
patient did not exhibit improvement after surgery.
INTERVENTION: Spinal angiography demonstrated a spinal perimedullary
AVF feeding from the left T12 radicular artery; venous drainage extended
rostrally into the posterior fossa venous system. The AVF was surgically
occluded via a posterior laminectomy at the level of the AVF. After
surgery, the patient’s symptoms began to abate.
CONCLUSION: Conus perimedullary AVFs can have venous drainage that
extends as far as intracranial veins, which can lead to confusing clinical
findings because the symptoms may suggest an intracranial process,
although the lesion is in the spine. Surgeons must be aware of this
confusing presentation. (Neurosurgery 49:457–462, 2001)
Key words: Perimedullary arteriovenous fistula, Petrosal sinus, Subarachnoid hemorrhage, Veins

Neurosurgery, Vol. 49, No. 2, August 2001

457

terms, I believe that an anterior spinal
lesion resulting in pathological focal kyphosis and anterior spinal cord compression should be corrected surgically via an
anterior approach. This approach allows
more effective, biomechanically sound
correction of the sagittal plane deformity,
in addition to being a more direct and
probably safer approach, than the approach the authors used to treat the spinal
cord compression.
Stephen Papadopoulos
Ann Arbor, Michigan

I

ntradural spinal perimedullary arteriovenous fistulae (AVFs) were first
described by Djindjian et al. (6) and, by
most accounts, are the second most
common type of spinal arteriovenous
malformation (AVM) (5, 7, 23). As with
most other kinds of spinal AVMs, perimedullary AVFs can cause symptoms
through steal, venous congestion, or,
less commonly, hemorrhage or mass effect and usually present with either
chronic, insidious, progressive myelopathy or sudden, catastrophic, subarachnoid or intramedullary hemorrhage (1,
4, 5, 10, 18, 21, 23). It is unusual, however, for patients to present with symptoms from both categories. It is even
more unusual for a conus perimedullary
AVF to drain intracranially and cause
symptoms of brainstem dysfunction attributable to venous congestion. We report the first case in the literature of a
patient with a conus perimedullary AVF
with intracranial venous drainage that
caused subarachnoid hemorrhage and
bulbar signs; the patient also harbored
an incidental intracranial aneurysm,
and he was inadvertently treated for the
aneurysm before the true cause of his
symptoms (his perimedullary AVF) was
discovered and obliterated.

CASE REPORT
A 65-year-old right-handed man presented to another hospital with suddenonset lower-extremity weakness, diplopia, nausea, and dysarthria on the day
of admission. The patient had complained of intermittent headaches for 2

458

Vates et al.

weeks before admission; the initial evaluation indicated paranasal sinus disease, and the patient was treated by his
primary care physician with nasal decongestants and azithromycin. On the
day of admission, the patient had been
moving about in his house when he
suddenly felt his legs become weak and
he fell to the ground. His sudden weakness was not associated with headache,
back pain, loss of consciousness, or seizure, but the patient did lose control of
his bladder and bowels. He was brought
to the emergency room, and the evaluation there revealed marked lowerextremity weakness and hyporeflexia,
without associated sensory deficits. The
patient also had right abducens palsy,
mild right facial weakness, and dysarthria, without tongue or palatal asymmetry. With further questioning, the patient reported 2 months of increased
difficulty in walking. He had previously
been able to easily walk 6 to 8 blocks;
however, his endurance had decreased
in the previous 2 months and he could
then walk only 1 to 2 blocks before
needing rest. This change was not associated with back pain, numbness, paresthesia, or changes in bowel or bladder
function. The past history of the patient
was significant for hypertension and excision of a malignant melanoma 3 years
before admission, without any evidence
of metastatic disease.
At admission, the patient underwent
head computed tomography, which
demonstrated no intracranial hemorrhage or brain abnormalities; paranasal
sinus opacification was evident, consistent with his previous history of sinus
disease. Cervical and lumbar magnetic
resonance imaging (MRI) scans revealed
degenerative joint disease and mild/
moderate cervical and lumbar canal stenosis, to a degree insufficient to explain
the profound deficits of the patient.
There were prominent flow voids, suggesting engorged perimedullary vessels, but these did not attract attention at
the time the MRI scans were obtained
(Fig. 1). Brain MRI scans and magnetic
resonance angiograms were obtained at
the same time as the spinal MRI scans,
and these demonstrated a left middle
cerebral artery (MCA) aneurysm (not
shown) and T2-weighted hyperintensity
in the pons and medulla (Fig. 1). A lum-

FIGURE 1. T2-weighted MRI scans of the thoracic (A) and cervical (B and C)
spine, showing tortuous extramedullary flow voids (white arrows) consistent with
engorged perimedullary veins extending the entire length of the spinal canal. B,
sagittal MRI scan also demonstrating increased T2-weighted signal intensity within
the pons and medulla (black arrow). C, axial T2-weighted scan of the caudal cervical spine, showing an enlarged extramedullary flow void on the anterior aspect of
the spinal cord, consistent with an engorged perimedullary vein.
bar puncture was performed, and evaluation of the cerebrospinal fluid revealed 540,000 red blood cells, 1,200
white blood cells, and normal protein
and glucose levels. Another lumbar
puncture performed 2 days later demonstrated similar cell counts, and neither procedure was considered traumatic. Four-vessel cerebral angiography
confirmed the presence of the left MCA
aneurysm; no vasospasm or other cerebrovascular abnormality was noted. Although the findings were not classic indications of intracranial aneurysm
rupture, the physicians caring for the
patient at that time recommended surgical clipping of his aneurysm. The patient was taken to the operating room on
his third hospital day, where the surgeon performed a left pterional craniotomy and clipped the left MCA aneurysm without difficulty. The surgeon
noted that the aneurysm did not seem to
have ruptured previously. After surgery, the patient developed no complications but he exhibited no improvement in his lower-extremity weakness

Neurosurgery, Vol. 49, No. 2, August 2001

and he lost joint position sense in his
right leg. Further review of the spinal
MRI scans indicated a spinal vascular
abnormality, and the patient was referred to our cerebrovascular service for
spinal angiography and treatment.
At the time of arrival at our facility,
the patient demonstrated persistent left
facial weakness, dysarthria, and lowerextremity weakness. Cerebral and spinal angiograms were obtained. These
demonstrated successful clipping of the
left MCA aneurysm, with no residual
lesion and no vasospasm. However, the
spinal angiogram demonstrated a perimedullary AVF centered at L2, with
supply at T12 from the left radicular
artery contribution to the anterior spinal
artery and additional posterior spinal artery supply at L1 and T10 from the
right and left radicular arteries, respectively (Fig. 2). Interestingly, the venous
drainage coursed cephalad, through
perimedullary veins, throughout the entire length of the spinal canal to the
cervicomedullary junction and finally
emptied, through intracranial perimed-

Perimedullary Arteriovenous Fistula with Intracranial Drainage
FIGURE 2. A, digitally subtracted,
anteroposterior, spinal angiogram, showing injection of contrast material through
the left T12 radicular artery contribution
to the anterior spinal artery, as well as
abnormally early emergence of contrast
material into a perimedullary draining
vein. B, slightly later phase of the same
injection, at higher magnification, showing an abnormally early flow of contrast
material from the anterior spinal artery
into a perimedullary vein, ascending rostrally within the draining vein. C, anteroposterior fluoroscopic view of the thoracic region, to provide bony landmarks
for orientation in the digitally subtracted
image in D; this shows how contrast
material injected through the left T12
radicular artery, as in A, leads to early
abnormal emergence of contrast material
in an ascending perimedullary vein that
drains throughout the extent of the thoracic spine. E, lateral fluoroscopic view of
the craniocervical region, to provide bony
landmarks for orientation in the digitally
subtracted image in F; this shows how
contrast material injected through the left
T12 radicular artery leads to emergence
of contrast material in the perimedullary
and perimesencephalic veins in the posterior fossa. The black arrow in F indicates
the superior petrosal sinus, through which
contrast material empties into the
transverse-sigmoid sinus system on the
left (anteroposterior view not shown).
ullary and perimesencephalic veins,
into the petrosal sinus (Fig. 2).
Because of the progressive nature of
the deficits of the patient and the clear
evidence that this perimedullary AVF
had both ruptured and caused venous
congestion affecting his pons and mesencephalon, we recommended surgical
obliteration of the fistula. The patient
agreed, and he was taken to the operating room, where wide laminectomies at
L1 and L2 were performed. After opening of the dura, the nerve roots of the
conus were carefully separated, and the
conus medullaris and filum terminale
were identified. Using the large draining vein of the fistula as a map, we
identified the fistula site on the anterior
surface of the conus (Fig. 3). From this
location, we were also able to identify
and isolate the small posterolateral spi-

nal artery supply, which was coagulated and divided. The conus was rotated slightly, to yield a good view of
the anterior spinal artery as it fed into
the fistula, and the more proximal anterior spinal artery was identified and
carefully dissected from the conus. The
isolated segment of the anterior spinal
artery feeding into the fistula was then
occluded with a 5-mm AVM clip, and
the fistula was noted to darken considerably. As an extra precaution, the
draining vein was occluded close to the
fistula with another 5-mm AVM clip.
This completely isolated the fistula and
occluded both its inflow and outflow.
We then coagulated the fistula with bipolar cautery, to shrink it and prevent
additional refilling.
The patient tolerated the procedure
well, and postoperative angiograms

Neurosurgery, Vol. 49, No. 2, August 2001

459

FIGURE 3. A, intraoperative view of
the surgical field, showing the engorged
draining vein of the perimedullary AVF,
with arterialized blood (white arrow).
B, higher-magnification view, showing
the arteriovenous connection (white
arrow). C, same view as in B, showing
the aneurysm clip applied to the fistulous connection (white arrow). D,
lower-magnification view of the surgical field, showing the same aneurysm
clip as in C, applied to the arteriovenous connection, as well as a smaller
clip applied to the draining vein more
caudally (black arrow).
FIGURE 4. Digitally
subtracted spinal
angiogram with
injection of contrast material
through the left
T12 radicular
artery, showing
that, after surgery,
the fistulous connection was obliterated and no
abnormal venous
drainage persisted.
confirmed that the AVF had been obliterated (Fig. 4). By the time of discharge,
the patient demonstrated improved
strength and sensation in both legs and
was able to walk cautiously with a
walker. His dysarthria and facial weakness both resolved by the end of his
hospitalization. He was transferred to a
rehabilitation center 7 days after surgery and, at the follow-up examination
6 weeks after surgery, was walking with
the assistance of only a cane.

460

Vates et al.

DISCUSSION
This case is unusual because the patient’s complaints and the diagnostic
test results did not agree with the classic
description of either intracranial subarachnoid hemorrhage attributable to
an aneurysm or a spinal cord perimedullary AVF. In retrospect, this disagreement was attributable to two factors,
i.e., the aneurysm was not the cause of
this patient’s symptoms and the perimedullary AVF had a very unusual
drainage pattern, which made localization of the symptoms very difficult.
Spinal cord AVMs are a heterogeneous group of vascular anomalies, including both “true” AVMs and AVFs. In
the past 15 years, recognition of differences in pathogenesis, anatomic features, pathophysiological characteristics, radiographic appearance, clinical
manifestations, and optimal treatment
for these various disorders has led to the
construction of a classification system,
as follows (4, 7, 8, 11, 23). Type I spinal
AVMs are spinal dural AVFs with a direct connection between the dural
branch of the spinal ramus of a radicular
artery and an intradural medullary
vein. Type II and III spinal AVMs are
true glomus AVMs; type II (glomus)
spinal AVMs are completely contained
within the substance of the spinal cord,
whereas type III (juvenile) spinal AVMs
extend to involve surrounding paraspinous tissues. Type IV spinal AVMs are
intradural leptomeningeal fistulous connections between an extramedullary artery and a dilated perimedullary vein
and are commonly called perimedullary
spinal cord AVFs (12). Subsequent work
has attempted to subclassify perimedullary spinal AVFs on the basis of architecture and flow (2, 16–18), but that is
not the focus of this report. The cause of
perimedullary AVFs remains a mystery;
most are considered congenital lesions
(5), but cases resulting from traumatic
injury have been reported (4, 9). These
lesions become clinically manifest in the
third to sixth decades of life and demonstrate no sex preference (4, 9, 18, 23).
They are most commonly observed in
the conus medullaris or cauda equina (4,
5, 9, 18, 23), although cervical and thoracic perimedullary AVFs have also
been reported (13, 15, 18, 19).

As noted by Wong and Awad (23), spinal vascular malformations can mimic a
diverse group of diseases affecting the
central and peripheral nervous systems,
and the syndromes associated with spinal
vascular malformations are rarely stereotypical. This was particularly true in this
case. The patient’s cranial nerve deficits
and the brain MRI results suggesting ischemia in the pons and medulla indicated
an intracranial disorder, but the patient’s
leg weakness and hyporeflexia indicated
a spinal cord lesion. In retrospect, spinal
MRI was a correct first step in diagnosis,
but the evidence for a spinal vascular
anomaly was overlooked, perhaps because an intracranial lesion (the left MCA
aneurysm) was discovered simultaneously. Nevertheless, the confusing nature of this patient’s presentation is now
clear; he had a spinal vascular lesion that,
because of the extremely unusual pattern
of venous outflow, was causing symptoms in two widely separated regions of
the central nervous system. Ricolfi et al.
(21) recently reviewed their experience
with giant spinal perimedullary AVFs
and did present the case of a 5-year-old
boy with subarachnoid hemorrhage and
progressive paraplegia secondary to a T12
perimedullary AVF of the conus with intracranial venous drainage. As we can
best tell from the scant primary data provided in their report, however, their patient did not present with the same constellation of bulbar signs and spinal cord
dysfunction. In addition, their case involved a giant spinal perimedullary AVF,
and our case demonstrates that dramatic
symptoms involving the entire neuraxis
can occur even with a smaller lesion.
Accurate diagnoses and anatomic localization are needed for the appropriate treatment of spinal perimedullary
AVFs. Myelography was historically
used to identify and localize spinal perimedullary AVFs, but MRI and spinal
angiography are now most commonly
used to identify these lesions (4, 5, 7, 11).
MRI can indicate spinal perimedullary
AVFs by documenting serpentine flow
voids in the subarachnoid extramedullary space; in addition, MRI can demonstrate subarachnoid hemorrhage or
changes in the brain and spinal cord
that suggest hemorrhage or venous congestion, as in this case. However, MRI
does not adequately delineate the ana-

Neurosurgery, Vol. 49, No. 2, August 2001

tomic features of the arterial feeders and
venous drainage of these lesions. Spinal
angiography is required to define the
arterial and venous anatomic features of
the fistulae sufficiently to guide
therapy.
Spinal perimedullary AVFs are associated with grim prognoses if left untreated. Patients experience either progressive symptoms of spinal cord
dysfunction attributable to presumed
venous congestion or steal or a stuttering course punctuated by episodes of
sudden neurological impairment attributable to rupture (1, 2, 5, 7, 8, 11, 12, 18,
23). Spinal perimedullary AVFs can be
successfully treated by using either microsurgical occlusion or endovascular
obliteration, and the relative merits of
these two approaches are better described elsewhere (2–5, 10, 11, 13, 14, 17,
18, 20, 22, 23). We chose surgical occlusion because the major supply originated from the anterior spinal artery.
Occlusion of this supply via the endovascular route can be treacherous, because it is difficult to control the deposit
of embolic material; this puts the dominant arterial supply to the spinal cord at
risk. In addition, the fistula was located
in an area that is easily accessed via a
standard surgical approach, with little
risk of surgery-related morbidity.

CONCLUSIONS
Spinal perimedullary AVFs may have
venous drainage that extends as far as
intracranial veins; this can lead to confusing clinical findings because the
symptoms may suggest an intracranial
process, although the lesion is in the
spine. This case emphasizes the need for
awareness of this potentially confusing
presentation. With the knowledge gathered from this case, surgeons should be
better prepared to diagnose spinal perimedullary AVFs and to plan treatments
for patients with these lesions.

ACKNOWLEDGMENT
We thank Ty Thaiyananthan for help
in preparing this report.

Perimedullary Arteriovenous Fistula with Intracranial Drainage
Received, January 25, 2001.
Accepted, March 28, 2001.
Reprint requests: Michael T. Lawton, M.D.,
Department of Neurological Surgery, University of California, San Francisco, 505 Parnassus Avenue, M779, San Francisco, CA
94143. Email: lawtonm@neurosurg.ucsf.edu

REFERENCES
1. Aminoff MJ, Gutin PH, Norman D: Unusual
type of spinal arteriovenous malformation.
Neurosurgery 22:589–591, 1988.
2. Anson JA, Spetzler RF: Classification of spinal
arteriovenous malformations and implications
for treatment. BNI Q 8:2–8, 1992.
3. Barrow DL, Alleyne CH: Surgical approaches:
Spinal lesions, in Jafar JJ, Awad IA, Rosenwasser RH (eds): Vascular Malformations of the Central Nervous System. Philadelphia, Lippincott
Williams & Wilkins, 1999, pp 339–345.
4. Barrow DL, Colohan AR, Dawson R: Intradural
perimedullary arteriovenous fistulas (type IV
spinal cord arteriovenous malformations).
J Neurosurg 81:221–229, 1994.
5. Cawley CM, Barrow DL: Intradural perimedullary spinal cord arteriovenous fistulas, in Barrow DL, Awad IA (eds): Spinal Cord Arteriovenous Malformations. Park Ridge, IL, AANS,
1999, pp 147–160.
6. Djindjian M, Djindjian R, Rey A, Hurth M,
Houdart R: Intradural extramedullary spinal
arterio-venous malformations fed by the anterior spinal artery. Surg Neurol 8:85–93, 1977.
7. Grote EH, Bien S: Arteriovenous malformations
of the spinal cord, in Youmans JR (ed): Neurological Surgery. Philadelphia, W.B. Saunders
Co., 1996, pp 1511–1530.
8. Grote EH, Voigt K: Clinical syndromes, natural
history, and pathophysiology of vascular lesions of the spinal cord. Neurosurg Clin North
Am 10:17–45, 1999.
9. Gueguen B, Merland JJ, Riche MC, Rey A: Vascular malformations of the spinal cord: Intrathecal perimedullary arteriovenous fistulas fed
by medullary arteries. Neurology 37:969–979,
1987.
10. Halbach VV, Higashida RT, Dowd CF, Fraser
KW, Edwards MS, Barnwell SL: Treatment of
giant intradural (perimedullary) arteriovenous
fistulas. Neurosurgery 33:972–980, 1993.
11. Heros RC: Arteriovenous malformations of the
spinal cord, in Ojemann RG, Heros RC, Crowell
RM (eds): Surgical Management of Neurovascular
Disease. Baltimore, Williams & Wilkins, 1995, pp
519–537.
12. Heros RC, Debrun GM, Ojemann RG,
Lasjaunias PL, Naessens PJ: Direct spinal arteriovenous fistula: A new type of spinal AVM—
Case report. J Neurosurg 64:134–139, 1986.
13. Hida K, Iwasaki Y, Goto K, Miyasaka K, Abe H:
Results of the surgical treatment of perimedullary arteriovenous fistulas with special reference to embolization. J Neurosurg 90[Suppl
4]:198–205, 1999.

14. Khayata M, McKenzie J, Vishteh AG, Dean B,
Spetzler RF: Endovascular approach to spinal
vascular lesions, in Maciunas RJ (ed):
Endovascular Neurological Intervention. Park
Ridge, AANS Publications Committee, 1995, pp
247–261.
15. Markert JM, Chandler WF, Deveikis JP, Ross D:
Use of the extreme lateral approach in the surgical treatment of an intradural ventral cervical
spinal cord vascular malformation: Technical
case report. Neurosurgery 38:412–415, 1996.
16. Merland JJ, Reizine D: Treatment of arteriovenous spinal cord malformations. Semin
Intervent Radiol 4:281–290, 1987.
17. Merland JJ, Reizine D: Embolization techniques
in the spinal cord, in Dondelinger RF, Rossi P,
Kurdziel JC (eds): Interventional Radiology. New
York, Thieme, 1990, pp 433–442.
18. Mourier KL, Gobin YP, George B, Lot G,
Merland JJ: Intradural perimedullary arteriovenous fistulae: Results of surgical and
endovascular treatment in a series of 35 cases.
Neurosurgery 32:885–891, 1993.
19. Nagashima C, Miyoshi A, Nagashima R,
Ogawa M, Enomoto K, Watabe T: Spinal giant
intradural perimedullary arteriovenous fistula:
Clinical and neuroradiological study in one
case with review of literature. Surg Neurol 45:
524–532, 1996.
20. Nelson PK, Bose A: Spinal vascular malformations:
Angiographic diagnosis and neuroendovascular
treatment, in Jafar JJ, Awad IA, Rosenwasser RH
(eds): Vascular Malformations of the Central Nervous
System. Philadelphia, Lippincott Williams &
Wilkins, 1999, pp 413–435.
21. Ricolfi F, Gobin PY, Aymard A, Brunelle F,
Gaston A, Merland JJ: Giant perimedullary arteriovenous fistulas of the spine: Clinical and
radiologic features and endovascular treatment.
AJNR Am J Neuroradiol 18:677–687, 1997.
22. Ricolfi F, Manelfe C, Meder JF, Arrue P, Decq P,
Brugieres P, Cognard C, Gaston A: Intracranial
dural arteriovenous fistulae with perimedullary
venous drainage: Anatomical, clinical and therapeutic
considerations.
Neuroradiology
41:803–812, 1999.
23. Wong JH, Awad IA: Clinical syndromes: Spinal
vascular malformations, in Jafar JJ, Awad IA,
Rosenwasser RH (eds): Vascular Malformations
of the Central Nervous System. Philadelphia, Lippincott Williams & Wilkins, 1999, pp 155–160,
1999.

COMMENTS
This is an interesting case report that
emphasizes referred neurological symptoms attributable to spinal cord arteriovenous malformations (AVMs). Although these classic data are well
known (1, 2), they point to the importance of the venous drainage of the
AVM and the absence of correlation between the location of the shunt (whatever its type or size) and the level of
symptoms observed.
In this case report, the clinical history
was typical of neither subarachnoid

Neurosurgery, Vol. 49, No. 2, August 2001

461

hemorrhage nor a supratentorial lesion.
Sudden weakness of both legs, with diplopia, nausea, and dysarthria, should
have indicated a lesion in the pons and
not a left middle cerebral artery aneurysm, especially in the absence of sudden intense headache. With modern radiological tools, more incidental lesions
are now being discovered. It may be
difficult to avoid acting on imaging
findings that have no link to the patient
history. After intraoperatively observing that the aneurysm was unruptured,
the authors searched elsewhere for the
cause of the subarachnoid bleeding diagnosed during the lumbar puncture,
and they discovered a spinal perimedullary arteriovenous fistula (AVF). As
stated above, this finding is classic, although rare. Spinal AVFs can exhibit
venous drainage extending as far as intracranial veins. This case report should
remind us to pay close attention to clinical signs and anamnesis.
Georges Rodesch
Jacques Brotchi
Brussels, Belgium
1. Di Chiro G, Doppman JL: Endocranial drainage
of spinal cord veins. Radiology 95:555-560, 1970.
2. Djindjian R, Hurth M, Thurel CL: Cervicocranial
phlebography of angiomas of the spinal cord.
Neuroradiology 1:42-46, 1970.

This is a report of a rare and very
interesting case of a spinal perimedullary fistula in the low thoracic region
that drained upward and secondarily
arterialized pial veins in the posterior
fossa, causing both hemorrhage and
brainstem dysfunction. This case reminds me of one I reported with my
colleagues in 1986 (1). Our patient almost underwent proton-beam irradiation to treat what was initially interpreted as a posterior fossa AVM. As we
studied the films in more detail, we
were struck by the fact that we observed
only arterialized veins in the posterior
fossa, without any real arterial input.
Consideration of this finding in conjunction with the fact that the patient had
exhibited slowly progressing paraparesis, which had been erroneously attributed to his “posterior fossa AVM,” led
us to speculate that perhaps there was a
vascular lesion in the spinal cord that
was draining upward and secondarily

462

Vates et al.

arterializing the veins of the posterior
fossa. In fact, with spinal angiography,
the patient was proven to have a huge,
upward-draining, perimedullary fistula
in the low thoracic region, arterializing
the entire venous system of the posterior
fossa. Because the fistula was located anteriorly and was fed by the anterior spinal
artery, I chose to treat it via a thoracotomy
and a two-level corpectomy, with intradural direct closure of the fistula. The case
presented by the authors is more complicated, because the patient had an incidental aneurysm and, in addition, was simultaneously exhibiting symptoms resulting
from subarachnoid hemorrhage, from
brainstem dysfunction, and apparently
also from spinal cord involvement.
The reverse of this situation is also
known to occur, and we have had experience with two cases in which the patient
presented with myelopathy resulting
from an intracranial dural fistula in the
posterior fossa that drained downward,
secondarily arterializing the perimedullary veins and causing progressive spinal

cord myelopathy resulting from venous
congestion. The lesson from these cases is
that, when intracranial arterialized veins
are observed without an obvious intracranial AVM or dural fistula, one must look
for the vascular abnormality in the spinal
cord; conversely, when a patient presents
with myelopathy and arterialized veins in
the spinal cord without an obvious spinal
AVM or spinal dural fistula, one must
look for an intracranial dural fistula that
drains downward.
Roberto C. Heros
Miami, Florida

1. Heros RC, Debrun GM, Ojemann RG, Lasjaunias
PL, Naessens PJ: Direct spinal arteriovenous fistula: A new type of spinal AVM—Case report.
J Neurosurg 64:134–139, 1986.

The authors present the interesting case
of a patient with a conus perimedullary
AVF with intracranial drainage and bulbar signs. They describe a lesion in the
region of the conus medullaris, with com-

ponents of both an AVF and an AVM.
These lesions can have extensive venous
drainage, and it is not surprising that they
can extend as far as the intracranial venous system. We are in the process of
describing a modified classification system for spinal cord vascular malformations and have considered it necessary to
create a new subcategory, i.e., the conus
AVM, given the unique nature of lesions
in this region. Conus AVMs are locationspecific lesions and, as mentioned above,
exhibit features of both AVFs and AVMs
of the spinal cord. In our experience, they
can be both perimedullary and intramedullary. Additionally, they manifest with
progressive myeloradiculopathy and subarachnoid hemorrhage. As in the case described by the authors, we have observed
very complex patterns of venous drainage
in association with these lesions.
Howard A. Riina
Robert F. Spetzler
Phoenix, Arizona