Secondary Epileptogenesis in Cerebral
Arteriovenous Malformations
Hwa-shain Yeh, MD, Michael D. Privitera, MD

\s=b\There is debate as to whether secondary
epileptogenesis occurs in humans. As part
of a series of patients with cerebral arteriovenous malformations and epilepsy, we identified two patients with probable secondary
epileptogenesis in the mesiotemporal regions, ipsilateral but anatomically distant
from an arteriovenous malformation causing
seizures. Both patients had arteriovenous
malformations outside the mesiotemporal
region and resection of the arteriovenous
malformation and epileptogenic areas identified by electrocorticography produced initial
freedom from seizures. Three to 6 months
later both patients developed a different seizure pattern that proved to be mesiotemporal
in origin by video/electroencephalogram
(one with depth electrodes) and both patients are seizure free after a second resection of anterotemporal and mesiotemporal

regions. Findings indicating a secondary
epileptogenic focus include (1) different seizure type by patient history, (2) second seizure type by ictal video/electroencephalographic recordings, and (3) lack of pathologic abnormalities in the resected mesiotemporal specimens.
(Arch Neurol. 1991 ;48:1122-1124)
T^ he incidence of patients with arterio¬
venous malformations (AVMs) pre¬
senting with epilepsy and with no clini¬
cal evidence of intracranial hemorrhage
varies between 17% and 40% in pub¬
lished series.1"3 The result of surgical
treatment for seizure control in patients
who have cerebral AVMs are not satisAccepted for publication May 17,1991.
From the Departments of Neurosurgery (Dr
Yeh) and Neurology (Dr Privitera), University of
Cincinnati (Ohio) College of Medicine; and Mayfield
Neurological Institute, Cincinnati, Ohio (Dr Yeh).
Reprint requests to the Department of Neurosurgery, University of Cincinnati College of Medi-

cine, 231 Bethesda Ave, Cincinnati, OH 45267-0515
(Dr Yeh).

factory4' Since 1982 we have per¬
formed a prospective study in 36 pa¬
tients who had supratentorial AVMs
who presented with epilepsy and with¬

out a history of intracranial hemor¬
rhage. All of the patients had been
treated medically before surgery. They

considered disabled because of
side effects or failure of anticonvulsant
therapy or the effect on their profes¬
sional lives of even infrequent seizures.
They underwent preoperative electro¬
encephalographic (EEG) evaluation in
addition to computed tomography scan¬
ning or magnetic resonance imaging and
cerebral angiography. Most of these op¬
erations were performed with the pa¬
tients under local anesthesia for an accu¬
rate intraoperative electrocorticographic recording. Cortical excision in
addition to removal of AVMs was per¬
formed in 19 patients if epileptic dis¬
charges were demonstrated adjacent to
or at a distance from the AVMs. Two
patients, one with a large AVM in the
right frontal region and the other one
with a thrombosed AVM in the right
were

posterior temporal region, presented
with complex partial seizures due to a
remote epileptic focus in the ipsilateral
mesial temporal region after the initial
procedure of excision of AVMs. Both
patients had a second surgery to remove
the remote seizure focus.

REPORT OF CASES

CASE 1.—A 30-year-old man presented
with a 9-year history of medically intractable
seizures. The seizures were not preceded by
an aura, but consisted of sudden onset of
unconsciousness, followed by generalized
tonic-clonic convulsions that lasted for 1 to 2
minutes. Frequency of seizures was about
one to two per week despite adequate anti¬
convulsant therapy. Computed tomography

Downloaded From: http://archneur.jamanetwork.com/ by a University of Pennsylvania User on 06/18/2015

showed a 2-cm calcified lesion located in the

right posterior temporal region (Fig 1). Ictal
EEG recordings showed seizure onset from
the right posterotemporal region.
He underwent a right frontotemporoparietal craniotomy under local anesthesia foiexcision of the thrombosed AVM from the
posterior portion ofthe right superior tempo¬
ral gyrus at University Hospital in Cincin¬
nati (Ohio) on February 13, 1986, and the

intraoperative

electrocorticography

re¬

vealed epileptiform activity from the inferior
portion of the right parietal lobe that was
resected. The patient has not experienced
generalized seizures since surgery.
Three months after surgery the patient
developed complex partial seizures, charac¬
terized by eye blinking, staring, and loss of
consciousness. There were no generalized
tonic-clonic seizures. He became disabled be¬
cause of increasing frequency of the attacks
despite trials of various antiepileptic drugs at
serum concentrations producing clinical

toxicity.

Sixteen months following the first proce¬
dure, video/EEG monitoring was performed
and showed five clinical complex partial sei¬
zures

with an initial blank stare, automa¬

tisms, and postictal confusion. Each seizure

rhythmic sharpdischarge from the right sphenoidal
electrode, starting simultaneously with the
clinical alteration in behavior. Intraopera¬
tive acute surface and depth electrode elec¬
trocorticography showed active epileptiform
discharges and polymorphic delta activity
from both the amygdala and hippocampus, as
well as from the anterior portions ofthe mid¬
dle and inferior temporal gyri. There was no
evidence of epileptiform activity or abnormal
slowing in the 3 to 4 cm of temporal lobe lying
between the previously resected AVM and
the electrophysiologically abnormal anterior
temporal lobe. Six centimeters of the right
anterotemporal lobe, including the amygdala
and the anterior 1.5 cm of pes hippocampus,
were resected en bloc. Neuropathologic anal¬
ysis of the antero/mesiotemporal lobe was
unremarkable. The patient has remained sei-

was associated with a clear

wave

zure

free for 3 years following the second

procedure.

CASE 2.—A 39-year-old right-handed man
presented with a 20-year history of general¬
ized seizures. He had denied warning symp¬
toms and had 4 to 6 generalized tonic-clonic
seizures per week despite treatment with
phenytoin and carbamazepine at serum con¬
centrations producing clinical toxicity. Cere¬
bral angiogram showed a large AVM in the
right anterofrontal region (Fig 2, left). An
electroencephalogram showed rhythmic
sharp and slow activity in the right antero¬
frontal region.
A right frontotemporal craniotomy was
performed at University Hospital in Cincin¬
nati on June 27, 1986. Intraoperative electro¬
corticography showed active spike and sharp
wave activity in the right frontal cortex pos¬
terior to the AVM. The AVM and the cere¬
bral cortex showing epileptiform activity
were both resected. Pathologic study re¬
vealed an AVM and gliosis of the removed

parenchyma. Postoperative angiogram
showed complete excision ofthe AVM (Fig 2,
right). The patient has not experienced sei¬

Fig 1.—Computed tomography without (left) and with (right) contrast showed a 2-cm calcified
thrombosed arteriovenous malformation in the right posterotemporal region.

zures after surgery.

During a follow-up visit 6 months after
surgery, he mentioned that he had experi¬

enced intermittent episodes of unpleasant
and strange smells. Despite serum concen¬
trations of phenytoin and carbamazepine
causing clinical toxicity, he continued to have
episodes of unpleasant smells followed by
complex partial seizures characterized by al¬
tered consciousness and automatisms. He
developed two episodes of generalized tonicclonic seizures preceded by lightheadedness
and a bad smell. Retrospectively, he admit¬
ted that he has occasionally experienced
transient episodes of a "rusty smell" before
surgery, but he did not think that was a
symptom of epilepsy.
Neither a magnetic resonance scan nor an¬
giography revealed abnormalities in the tem¬
poral lobe. Two years later the patient was
readmitted for video/EEG monitoring. Four
complex partial seizures were recorded
showing an initial aura of an unpleasant smell
followed by staring, alteration in conscious¬
ness, lip smacking, then secondary general¬
ization. Ictal EEG recording with scalp and
sphenoidal electrodes did not adequately lo¬
calize seizure onset, so the patient was read¬
mitted for chronic depth electrode record¬
ings that showed seizure onset from the right
hippocampus and amygdala. The patient un¬
derwent a right anterotemporal lobectomy
with resection of the amygdala and hippo¬
campus. He has been seizure free for 24
months. Pathologic evaluation of the right
temporal lobe was unremarkable.
COMMENT

Secondary epileptogenesis and kin¬
dling have been observed in experimen¬
tal animal studies, but are not well docu¬
mented in humans.M Kindling was first
observed in rats by Goddard1" as the
progressive enhancement of electrical
and behavioral response to a low-inten¬
sity initially ineffective electrical brain
stimulation. The limbic system struc¬
tures, especially the amygdala, are the

Fig 2.—Cerebral arteriogram (lateral view) disclosed a large arteriovenous malformation in the right
anterofrontal region (left). The postoperative study showed a complete excision of the arteriove¬
nous malformation (right).

most sensitive to electrical kindling."
The duration of stimulation required for
seizures varies in different species; an
average of 14 days is needed in rats, 25
days in cats, and 248 days in rhesus
monkeys.12 The interictal electrical dis¬
charges propagate and appear indepen¬
dently at distant sites for an indefinite
period of time in the animals with kin¬
dling in the amygdala. As in kindled
animals, a similar distant independent
epileptic discharge is believed to occur
in humans with partial seizures.13 Mech¬
anisms similar to those in animal kin¬
dling may cause some poorly controlled
seizures in humans. The mirror focus is
described as an independent secondary
epileptogenic lesion in the brain contralateral and homotopic to the primary
focus.14 The question of secondary epi¬
leptogenesis in humans, especially in
patients who had epilepsy caused by tu¬
mor ofthe brain, has been addressed.15'16
Twenty-one percent to 36% of patients
with epilepsy caused by tumors in var¬
ied locations of the human brain had
evidence of secondary epileptogenesis.
After surgical removal of the primary
lesion, in approximately one third ofthe

Downloaded From: http://archneur.jamanetwork.com/ by a University of Pennsylvania User on 06/18/2015

patients secondary foci persisted and
continued to give rise to clinical sei¬
zures.

The clinical course and EEG findings
in these two patients provide evidence
of a remote, independent epileptic focus
after removal of the primary lesion. In
both cases, an abnormality outside the
mesiotemporal region gave rise to sei¬
zures that were initially controlled by
resection of the AVM; however, an in¬
dependent seizure focus developed in
the mesiotemporal region and seizures

stopped by a second surgical re¬
section of anterior and mesiotemporal
structures. The change in the manifes¬
tations of the seizures by history, clini¬
cal observation on video, and ictal EEG
all support a region of epileptogenicity
anatomically distinct from the region of
the AVM capable of producing clinical
seizures. Both regions were capable of
producing clinically distinct seizures,
not merely interictal spikes. Further¬
more, pathologic analysis did not dem¬
onstrate any AVM or residual hemosiderin in the resected mesiotemporal
regions, indicating the epileptogenicity
was not due to a preexisting structural
were

abnormality of mesiotemporal struc¬
tures.

Secondary epileptogenesis presum¬
ably occurred prior to the surgical exci¬

sion of AVMs. The symptoms from the
secondary seizure focus may have exist¬
ed earlier, but were not recognized ini¬
tially in case 2. The first patient denied
complex partial seizures before the first
surgery. Perhaps the second seizure
pattern was clinically enhanced after
excision ofthe primary focus, which was
associated with more prominent and
disabling seizures.
Morrell1' postulated that "the density
ofthe excitatory synaptic connections of
the cells of the primary focus and the
excitability of the target cells appear to
exert a major influence on the likelihood
of development of secondary foci." In
the kindling model, the amygdala is the
most sensitive site for development of
seizures, and in humans, the mesiotem¬
site for
poral region is the most common
'"
complex partial seizures. In case 1, the
epileptic discharges may have reached
the mesiotemporal region via the fibers
of the inferior longitudinal fasciculus,
which extend from the occipital lobe,
pass through the length of the temporal
lobe, terminate in the temporal pole,
and reach the amygdala by the arcuate
fibers. In case 2, discharges may have
spread from the orbitofrontal region to
the rostral part ofthe ipsilateral tempo¬
ral lobe by the uncinate fasciculus.
Several investigators have reported
cases where ipsilateral secondary epi¬
leptogenesis was inferred but not prov¬
en. In the cases of Falconer et al19
removal of a lesion outside the mesio¬
temporal lobe produced gradual resolu¬
tion of presumed mesiotemporal onset
seizures. Falconer and coworkers' pa¬
tients may have had what Morrell" re¬
fers to as "reversible" secondary epilep¬
togenesis where removal ofthe primary
region produces freedom from seizures.
Our two cases may have reached the
irreversible stage where the secondary
focus continues to produce clinical sei¬
zures after removal of the primary
focus.
Sperling et al3' and Olivier et al21 pre¬
sented cases with predominantly mesio¬
temporal seizures and presumed prima¬
ry epileptogenic regions outside the
mesiotemporal region. In Sperling and
coworkers' case, a tumor involved lan¬
guage areas of the posterotemporal re¬
gion, but the patient became seizure
free after anterotemporal and mesio¬
temporal resection. Olivier and cowork¬
ers' patient had an initial seizure pat¬
tern indicating occipital lobe onset, but
after some years also developed inde¬
pendent temporal lobe complex partial
seizures. Depth electrode recording

demonstrated

independent epilepto¬

genic regions in the occipital and in the
mesiotemporal areas, and the patient
was almost completely seizure free af¬
ter only anterotemporal and mesiotem¬
poral resection. In these two cases, the

primary abnormality was not resected,
yet removal of a presumed secondary
epileptogenic lesion produced a good
outcome.
It has been postulated that the follow¬
ing mechanisms are active in the patho¬

physiology of epilepsy associated with
cerebral AVMs: (1) ischemia ofthe adja¬

cent cerebral cortex due to arteriove¬
nous shunting in the large AVMs; (2)

gliosis of the brain surrounding the
AVM because of previously subclinical
hemorrhage or hemosiderin deposits;
and (3) secondary epileptogenesis, es¬
pecially AVMs in the temporal lobe in
association with remote epileptogenic
foci.22 Therefore, thorough preoper¬
ative and intraoperative EEG studies
should be carried out, especially in pa¬

tients who have cerebral AVMs and
present with chronic epilepsy. Video/EEG monitoring of seizures is impor¬
tant to demonstrate whether indepen¬
dent regions of seizure onset exist. Ac¬
cording to our previous experience with
a favorable result of seizure control of
surgically treated patients associated
with cerebral AVMs, excision of the
AVM along with adjacent or remote epi¬
leptogenic cortex should be considered
to optimize seizure control after sur¬
gery23 As in Sperling and coworkers'2"
case, if a lesion is in eloquent cortex that
cannot be resected but seizures origi¬
nate in an accessible area, resection of
only the region producing seizures may
produce good seizure control. Continu¬
ous long-term follow-up and réévalua¬
tion of patients who still have seizures
after initial resection of cerebral AVM,
either the same type or a new type, are
important. As we have shown, some pa¬
tients may benefit from a second sur¬
gery to excise the secondary epilepto¬
genic focus to achieve good seizure
control.
Our cases are unique in that a focal
structural abnormality initially causing
seizures was removed, seizures im¬
proved, but then recurred with a differ¬
ent seizure pattern indicating a region
of secondary epileptogenesis; removal
of the pathologic normal secondary re¬
gion stopped the new seizure pattern.
References
1. Crawford PM, West CR, Chadwick DW,
Shaw MDM. Arteriovenous malformation of the
brain: natural history in unoperated patients.
J Neurol Neurosurg Psychiatry. 1986;49:1-10.
2. Moody RA, Poppen JL. Arteriovenous malformations. J Neurosurg. 1970;32:503-511.

Downloaded From: http://archneur.jamanetwork.com/ by a University of Pennsylvania User on 06/18/2015

3. Perret G, Nishioka H. Report of the cooperative study of intracranial aneurysms and subarachnoid hemorrhage, VI: arteriovenous malformation:
an analysis of 545 cases of cranio-cerebral arteriovenous malformations and fistulae reported to the
cooperative study. J Neurosurg. 1966;25:467-490.
4. Crawford PM, West CR, Shaw MDM, Chadwick DW. Cerebral arteriovenous malformation
and epilepsy: factor in the development of epilepsy.

Epilepsia. 1986;27:270-275.

5. Forster DMC, Steiner L, Hakanson S. Arteriovenous malformations of the brain: a long-term
clinical study. J Neurosurg. 1972;37:562-570.
6. Murphy MJ. Long-term follow-up of seizures
associated with cerebral arteriovenous malformations: results of therapy. Arch Neurol. 1985;42:477\x=req-\
479.
7. Parkinson D, Bachers G. Arteriovenous malformations: summary of 100 consecutive supratentorial cases. J Neurosurg. 1980;53:285-299.
8. Goldensohn ES. The relevance of secondary
epileptogenesis to the treatment of epilepsy: kindling and the mirror focus. Epilepsia. 1984;

25(suppl 2):S156-S168.

9. Morrell F, Wada J, Engel J. Potential relekindling and secondary epileptogenesis to
the consideration of surgical treatment for epilepsy. In: Engel J Jr, ed. Surgical Treatment of the
Epilepsies. New York, NY: Raven Press; 1987:701\x=req-\
707.
10. Goddard GV. Development of epileptic seizures through brain stimulation at low intensity.
Nature. 1967;214:1020-1021.
11. Goddard GV, McIntyre DC, Leech CK. A
permanent change in brain function resulting from
daily electrical stimulation. Exp Neurol. 1969;
25:294-330.
12. Wada JA. The clinical relevance of kindling:
species, brain sites and seizure susceptibility. In:
Livingston KE, Hornykiewicz D, eds. Limbic
Mechanisms. Orlando, Fla: Plenum Publishing
Corp; 1978:369-388.
13. Engel J Jr, Cahan L. The potential relevance
of kindling to human partial epilepsy. In: Wada JA,
ed. Kindling 3. New York, NY: Raven Press;
1986:37-54.
14. Morrell F. Physiology and histochemistry of
the mirror focus. In: Jasper HH, Ward AA Jr, Pope
A, eds. Basic Mechanisms of the Epilepsies. Boston, Mass: Little Brown & Co Inc; 1969:357-370.
15. Morrell F, Rasmussen T, de Toledo-Morrell
L, Quesney LF, Gloor P. Frontal lobe epilepsy of
neoplastic etiology: incidence of secondary epilepvance of

togenesis. Epilepsia. 1984;25:654-655.
16. Morrell F, Rasmussen T, Gloor P, de Toledo\x=req-\
Morrell L. Secondary epileptogenic foci in patients
with verified temporal lobe tumor. Electroencephalogr Clin Neurophysiol. 1983;54:26.
17. Morrell F. Varieties of human secondary epileptogenesis. J Clin Neurophysiol. 1989;6:227-275.
18. Williamson PD, Wieser HG, Delgado-Escueta AV. Clinical characteristics of partial seizures. In: Engel J Jr, ed. Surgical Treatment of the
Epilepsies. New York, NY; Raven Press; 1987:101\x=req-\

120.
19. Falconer MA, Driver MV, Serafetinides
EA. Temporal lobe epilepsy due to distant lesions:
two cases relieved by operation. Brain. 1962;
85:521-534.
20. Sperling MR, Cahan LD, Brown WJ. Relief
of seizures from a predominantly posterior temporal tumor with anterior temporal lobectomy. Epilepsia. 1989;30: 559-563.
21. Olivier A, Gloor P, Andermann F, Ives J.
Occipitotemporal epilepsy studied with stereotaxically implanted depth electrodes and successfully
treated by temporal resection. Ann Neurol.

1982;11:428-432.
22. Yeh H, Tew JM, Jr. Management of arterio-

the brain. Contemp Neurosurg. 1988;9:1-6.
23. Yeh H, Kashiwagi S, Tew JM Jr, Berger TB.
Surgical management of epilepsy associated with
cerebral arteriovenous malformations. J Neurosurg. 1990;72:216-223.
venous malformations of