Review Article

Surgical Treatment of Epilepsy in Children
Elaine Wyllie, MD
In carefully selected children, as well as adults, intractable seizures may be eliminated or greatly reduced by
cortical resection or hemispherectomy. Critical features
of surgical candidacy include intractable disabling epilepsy, a localized epileptogenic zone, and a low risk of new
postoperative deficits; however, the analysis may be
complicated in children. Compared with adults, pediatric
patients are especially likely to present with poorly localizing electroencephalographic features because of their
high incidence of extratemporal localization and developmental pathology. Maturation factors may result in unusual epilepsy manifestations, for example, infantile
spasms and hypsarrhythmia caused by a focal cortical
lesion. The cognitive and psychosocial costs of continued
frequent seizures during infancy and childhood must be
assessed differently from those in adults and may include
stagnation of developmental progression. The risk for
new postoperative deficits may be modified if surgery is
performed during stages of active brain maturation with
developmental plasticity. For each individual child, the
potential risk/benefit ratio for surgery must be carefully
weighed on the basis of results of an extensive preoperative evaluation. Results from pediatric epilepsy surgery
series are encouraging, with percentages of seizure-free
patients similar to those in adult series. In some series,
delaying surgery for childhood-onset epilepsy into adulthood was associated with greater permanent psychosocial, behavioral, and educational problems. The available
data suggest that children should be considered for
surgical evaluation at whatever age they present with
severe intractable localization-related epilepsy. Complicated cases warrant referral to specialized centers with
extensive pediatric epilepsy surgery experience. © 1998
by Elsevier Science Inc. All rights reserved.
Wyllie E. Surgical treatment of epilepsy in children.
Pediatr Neurol 1998;19:179-188.

From the Pediatric Epilepsy Program; The Cleveland Clinic
Foundation; Cleveland, Ohio.

© 1998 by Elsevier Science Inc. All rights reserved.
PII S0887-8994(98)00062-9 ● 0887-8994/98/$19.00

Introduction
Several series have shown that intractable seizures may
be eliminated or greatly reduced in children, as well as
adults, after cortical resection [1-13] or hemispherectomy
[14-17]. Small series have also reported gratifying outcomes after epilepsy surgery in infants [18-23]. Critical
features of surgical candidacy include intractable epilepsy
interfering with quality of life (QOL) or development, a
localized epileptogenic zone, and a low risk of new
postoperative neurologic deficits. At first glance, this
formula appears clear-cut, but extensive clinical judgment
enters into every case. How many and which drugs must
fail before the epilepsy is declared intractable? How
frequent must the seizures be to warrant surgery? What
will be the psychosocial and developmental costs of
continued seizures if surgery is not performed? What are
the risks of new postoperative deficits if surgery is
performed during infancy or early childhood when active
brain maturation allows developmental plasticity? These
issues complicate the identification of appropriate pediatric candidates for epilepsy surgery.
An added concern is that compared with adult patients,
pediatric epilepsy surgery candidates are especially likely
to present with poorly localizing electroencephalographic
(EEG) features because of their high incidence of extratemporal localization and developmental abnormalities.
Maturation factors may also be important, as suggested by
age-specific manifestations of focal cortical lesions, such
as infantile spasms and hypsarrhythmia [19,20,23-26].
Clearly, the potential risk/benefit ratio for surgery must
be carefully weighed for each child on the basis of the
results of an extensive preoperative evaluation. Factors
weighing into these analyses are addressed here in the
context of the experience with pediatric epilepsy surgery
at The Cleveland Clinic Foundation (CCF) [13] and other
centers.

Communications should be addressed to:
Dr. Wyllie; Pediatric Epilepsy Program; The Cleveland Clinic
Foundation; 9500 Euclid Avenue; Cleveland, OH 44195.
Received January 22, 1998; accepted May 8, 1998.

Wyllie: Epilepsy Surgery 179

Temporal Lobe Resection
Temporal lobe resection is the most common epilepsy
surgery performed in adults, usually for hippocampal
sclerosis [27]. This epileptogenic lesion can now be
reliably identified in vivo with high-resolution magnetic
resonance imaging (MRI), including thin cuts through the
temporal lobes [28-34], and further confirmed by glucose
hypometabolism in the temporal lobe on positron emission
tomography (PET) scan [35]. Patients typically have
complex partial seizures with automatisms, with ictal and
interictal epileptiform discharges localized to the anterotemporal region on EEG [36]. In the setting of unilateral
hippocampal sclerosis, anteromesial temporal resection
offers freedom from seizures for approximately 70-85% of
patients [34,37-39].
Although hippocampal sclerosis may also occur in
children, it is uncommon in pediatric epilepsy surgery
candidates less than 12 years old. For example, Duchowny
et al. [3] reported hippocampal sclerosis in only two (13%)
of 16 preadolescent children who had temporal lobe
resection. In the series [13] of patients who had temporal
resection at the CCF between January 1990 and June 1996,
hippocampal sclerosis was the etiology for six (19%) of 31
children 3 months to 12 years of age and 12 (29%) of 41
adolescents 13-20 years of age. Overall, hippocampal
sclerosis was the etiology for only six (10%) of 62
children and 12 (16%) of 74 adolescents who had epilepsy
surgery of any type [13]. By comparison, in the predominantly adult multicenter data compiled by Engel [27],
73% of 5,446 epilepsy surgeries (excluding corpus callosotomies) were anterotemporal resections or amygdalohippocampectomies performed for nonlesional epilepsy including hippocampal sclerosis.
More commonly, pediatric candidates for temporal lobe
resection have focal cortical dysplasia or a low-grade
tumor (astrocytoma, ganglioglioma, or dysembryoplastic
neuroepithelial tumor) [3,11-13]. In such cases the EEG
may sometimes be poorly localizing [11,25]. For epilepsy
surgery in general, congruence between EEG and neuroimaging is an important factor for favorable long-term
seizure outcome [40], but exceptions may occur. A small
number of reported children—limited almost exclusively
to patients with tumors— have had seizure-free postoperative outcomes despite poorly localized EEG features
[11,25]. By contrast, the few pediatric surgical cases
reported so far with hippocampal sclerosis have each had
EEG findings similar to those in adults, with predominantly unilateral temporal ictal and interictal epileptiform
discharges on the side of the MRI abnormality [3,11,41].
High-resolution MRI and PET scans now routinely
permit detection of previously occult epileptogenic lesions
[19,28-33,42-45], and this detection has led to a shift in
relative weighting of preoperative EEG results. However,
EEG, seizure semiology, and clinical course remain critical factors to be carefully considered in every case. For
example, not every child with a potentially epileptogenic

180

PEDIATRIC NEUROLOGY

Vol. 19 No. 3

lesion on MRI has intractable epilepsy. Focal cortical
dysplasia and hippocampal sclerosis were first identified
in patients with intractable epilepsy, and many patients
have had good outcomes after resection of such lesions.
However, as new neuroimaging techniques are applied to
broader patient groups, it has become clear that these
lesions may occasionally be identified in normal control
patients or patients with only rare seizures or wellcontrolled epilepsy [44,46]. In most cases the presence of
a focal cortical lesion or hippocampal sclerosis should not
lead to surgery in the absence of poorly controlled seizures
confirmed by ictal video EEG. An exception may be the
child with infrequent seizures resulting from a low-grade
tumor in a surgically accessible area for whom resection
would permit histopathologic diagnosis and eliminate
concerns about potential invasiveness or malignancy.
Seizure outcome after temporal lobe resection appears
to be favorable as often for children as adults. Few data are
available from preadolescent children, but Duchowny et
al. [3] reported seizure-free outcomes for 12 (75%) of 16
patients who had temporal lobe resection before 12 years
of age. Similarly in the CCF series [13], 23 children (74%)
were free of seizures after temporal lobe resection and an
additional five children (13%) had only rare seizures.
Results were similar in adolescent patients, with seizurefree outcomes achieved for 33 (80%) of 41 patients, and in
both age groups the results far surpassed those achieved
during controlled trials of new antiepileptic drugs [47].
Waiting to perform temporal lobe resection later in life
does not appear to have advantages. Mizrahi et al. [7]
reported seizure-free outcomes for 18 (82%) of 22 predominantly adult patients who had temporal lobe resection
for intractable childhood-onset epilepsy after a mean
interval to surgery of 15 years and noted that later surgery
was associated with greater permanent psychosocial, behavioral, and educational problems.
The most frequent complication of temporal lobe resection is an asymptomatic homonymous superior quadrantanopia [48]. However, memory and language problems
have been reported after temporal lobe resection in adults,
with risk factors including higher preoperative functioning
or left resection [49]. Few systematic data are available
from preadolescent children because of the difficulties in
performing neuropsychologic tests in young children, the
lag in development of appropriate testing instruments for
use at young ages, and the low numbers of children who
have undergone the operation. However, preliminary results suggest that the neurocognitive risks of temporal lobe
resection may be similar in children and adults, with IQ
remaining stable [6,50-52] and a possible decrease in
memory function [50,52]. Risk factors for a decline in
postoperative verbal memory scores after temporal resection in childhood may include higher preoperative performance and left resection [52].
Language outcome has not yet been systematically
studied in preadolescent children who underwent left
temporal lobe resection, but preliminary results indicate

that developmental plasticity may play a role in permitting
safe resection of congenital left posterior temporal tumors.
When a low-grade left posterior temporal tumor grows
during the early stages of brain maturation, language
function may develop elsewhere in the right hemisphere or
in unaffected areas of the left hemisphere [53]. Language
mapping with cortical stimulation of subdural electrodes
may be critical to define safe margins for resection, with
extraoperative techniques usually more tolerable to young
children than intraoperative stimulation under local anesthesia.
Extratemporal and Multilobar Resection
Extratemporal or multilobar resections and hemispherectomies comprise a large percentage of pediatric
epilepsy surgeries, especially at younger ages. In the CCF
series [13], these procedures comprised 50% of the epilepsy surgeries performed by 12 years of age, 68% of
those performed by 6 years of age, and 90% of those
performed by 2 years of age.
Among focal extratemporal surgeries, frontal resections
predominate in children and adults [13], although in
infants, temporoparietal-occipital resections are especially
common [19]. In pediatric candidates for extratemporal, as
well as for temporal, resection the most common etiologies of the epilepsy are focal cortical dysplasia and
low-grade tumor [13].
Fewer patients are nearly or completely seizure-free
after extratemporal resection than after temporal resection
[13,27,54]. Sparse pediatric data are available, but Fish et
al. [55] reported rare or no postoperative seizures in only
12 (27%) of 45 children and adolescents who had frontal
lobe resection for nontumoral epilepsy between 1940 and
1980. In the CCF series [13], however, 26 (54%) of 48
children and adolescents were seizure-free after extratemporal or multilobar resection, and another 19% had only
rare seizures. These more favorable results may reflect the
more recent nature of the CCF pediatric series [13],
limited to the experience after 1990 when modern MRI
techniques were available to reveal cortical dysplasia,
tumor, or other focal lesions in 85% of cases. In a
predominantly adult series, Zentner et al. [56] reported
results similar to those in the CCF series [13], with
freedom from seizures after extratemporal resection for
54% of 60 patients, most of whom had lesional epilepsy.
The presence of a focal epileptogenic lesion on preoperative MRI appears to be the key to a better prognosis for
extratemporal resection.
Etiology also plays a role, with seizure-free outcomes
more frequent after temporal or extratemporal resection
for patients with tumor than for patients with cortical
dysplasia. In the CCF series [13], 36 (82%) of 44 patients
with tumor had seizure-free outcomes compared with 16
(52%) of 31 patients with dysplasia. Dysplasia was more
common among extratemporal (22 of 48, 46%) than
temporal (9 of 72, 12%) cases [13], perhaps influencing to

some degree the lower frequency of seizure-free outcomes
obtained after extratemporal resection.
A surgical approach to extratemporal and multilobar
cases is illustrated by the cases presented below. The
strategy at CCF, especially in very young patients, is to
define the extent of resection as much as possible with
preoperative video EEG, MRI, and PET, plus intraoperative stereotactic MRI, electrocorticography, cortical stimulation, and evoked potentials when appropriate. Newer
neuroimaging techniques, including magnetic resonance
spectroscopy [57-59] or ictal single-photon emission computed tomography [60-62], may also be helpful. Alternatively, Duchowny et al. [21] found extraoperative subdural
electrode studies to be useful for surgical planning in five
infants, using a special cortical stimulation paradigm [63].
It has also been found at CCF that extraoperative studies
with subdural electrodes may be helpful in selected cases
when resection is planned near eloquent cortex, especially
in older children. Chronically implanted subdural electrodes permit functional mapping of language, motor, and
sensory regions, as well as definition of the epileptogenic
zone by electrocorticography of spontaneous seizures
[8,64]. Language and somatosensory mapping with extraoperative cortical stimulation may be accomplished in
verbal children as young as 7 or 8 years old, but in younger
children functional localization is usually limited to results
from stimulation of primary motor areas and cortical
recording of somatosensory evoked potentials. Both of
these procedures, as well as interictal electrocorticography, can also be accomplished intraoperatively under light
general anesthesia. For this reason, chronically implanted
subdural electrodes are infrequently used at CCF for
evaluation of infants and young children.
Because seizure-free outcome is less common after
extratemporal than temporal lobe resection an obvious
question arises. Is the surgery a failure if seizures are
significantly reduced but not completely stopped? Sperling et al. [65] confirmed that driving and full employment
require complete freedom from seizures after epilepsy
surgery for adults. However, these are clearly not the
immediate goals for surgery in an infant or young child
with daily seizures and developmental stagnation. Anecdotal experience suggests that significant lightening of a
severe seizure burden in childhood may have important
positive effects on psychosocial and cognitive development and QOL, but systematic data are not yet available as
proof. Ongoing prospective pediatric studies will be important to better clarify these issues. Standardized healthrelated QOL assessments may be helpful in this analysis,
although validated instruments have only recently become
available for children [66].
Case Reports
Case 1. A male presented for preoperative evaluation at 3 months of
age. Perinatal course was unremarkable, but seizures began at 6 weeks of
age and occurred at a rate of 30-100/day despite multiple trials of antiepi-

Wyllie: Epilepsy Surgery 181

Figure 1. Case 1. Axial (TR: 4,700 ms, TE: 96 ms) and
sagittal (TR: 3,640 ms, TE: 96 ms) MRI scans revealing a
lesion in the right parietal lobe.

leptic medication. Seizures involved sustained, forceful bilateral eye
deviation to the right with chewing movements, restless limb movements,
and apparent partial preservation of awareness, punctuated by brief
epileptic spasms with abrupt extension and stiffening of the arms.
Neurologic examination was normal except for decreased visual attentiveness. MRI at 3 months of age revealed a tumor in the right parietal
lobe posterior to the postcentral gyrus (Fig 1). 18-Fluorodeoxyglucose
PET scan demonstrated intense focal hypermetabolism in the region of
the lesion. Interictal EEG revealed continuous slowing and spikes
maximum in the right parieto-occipital region (Fig 2). Ictal EEG
consisted of rhythmic spiking in the right parieto-occipital region (Fig
3A) interrupted by brief diffuse electrodecrements during spasms (Fig
3B, arrow). The boy underwent resection of the lesion at 3 months of age
(Fig 4), with localization aided by intraoperative MRI and electrocorticography. Histopathologic analysis of resected tissue revealed a gangliocytoma. The child has had no further seizures during the 12 months after
surgery, and antiepileptic medications have been withdrawn. Postoperative neurologic examination, developmental progress, and visual attentiveness were normal at follow-up 7 months after surgery.
Case 2. Another boy presented for preoperative evaluation at 2 years
6 months of age. Perinatal course was unremarkable. Seizures began at 1
month of age but were infrequent until 16 months of age, when they
became very frequent, intractable, and disabling. The seizures occurred
in clusters, 10-100/day, 2 or 3 days/week, and involved bilateral eye
blinking, version of eyes toward the left, lip smacking, arching of the

trunk, and random movements of all extremities for 45-60 seconds.
Developmental progress stagnated after 16 months of age, and at 2 years
6 months he took only a few steps with support and spoke only three
words. Neurologic examination revealed dense left homonymous hemianopia without hemiparesis.
MRI revealed evidence of cortical malformation in the right temporooccipital region (Fig 5), and 18-fluorodeoxyglucose PET scan demonstrated hypermetabolism in the right occipital lobe, maximum mesially.
Interictal EEG revealed continual periodic sharp waves, maximum in the
right occipital region (Fig 6), as well as decreased posterior background
and intermittent slowing in the same area. EEG seizures were maximum
in the right occipito-parietal region (Fig 7). Ictal single-photon emission
computed tomography revealed increased perfusion in the right occipital
and temporal regions. Magnetic resonance spectroscopy demonstrated
decreased n-acetylaspartate levels in the region of abnormality on MRI.
Right temporo-occipital resection was performed at 2 years 6 months
of age (Fig 8), with localization aided by intraoperative stereotactic MRI
and electrocorticography. Postoperative recovery was complicated by
aseptic meningitis with fever for several days. A hygroma was aspirated
2 months after surgery. Seven months after surgery, the boy has had no
seizures while taking reduced antiepileptic medications. He sustained no
new neurologic deficits after surgery. Developmental progression resumed but remains significantly delayed.

Hemispherectomy

Figure 2. Case 1. Interictal EEG demonstrating continuous slowing
and spikes maximum in the right parieto-occipital region.

182

PEDIATRIC NEUROLOGY

Vol. 19 No. 3

Hemispherectomy results in near or complete seizure
freedom for 60-85% of patients [16,17,27,67,68]. The
experience [13] at CCF has been similar, with 11 (69%) of
16 patients seizure-free. The procedure may be indicated
for patients with extensive unilateral hemispheric damage,
hemiparesis without fine finger movements, and intractable seizures arising from multiple or widespread regions of
the damaged hemisphere [67]. Candidates typically have
intractable epilepsy because of hemispheric syndromes
such as Sturge-Weber syndrome, unilateral perinatal middle cerebral artery infarction, hemimegalencephaly, or
Rasmussen chronic focal encephalitis [13,67,68].
Within the first year of life, hemiparesis may be difficult
to appreciate because of immature development of corticospinal tracts. The most reliable clue is decreased spontaneous limb movement on one side compared with the

Figure 3. Case 1. (A) Ictal EEG at clinical onset revealing rhythmic
spiking in the right parieto-occipital region. (B) Ictal EEG 20 seconds
after onset demonstrating interruption of the right parieto-occipital
spiking by a brief diffuse electrodecrement during a spasm (arrow).

other. Later in childhood, it is easier to appreciate the
absence of fine finger movements. Hemispherectomy does
not significantly worsen the hemiparesis in patients with
extensive hemispheric injury or malformation [16,69], and
patients retain proximal limb movement and walking.
Postoperative motor deficits may be less pronounced in
patients who have hemispherectomy during the first few
years of life because of developmental plasticity [70].
Many hemispherectomy candidates have a pre-existing
homonymous hemianopia so that hemispherectomy does
not alter visual function. In patients without a pre-existing
hemianopia, Tinuper et al. [16] found that the visual field
deficit after hemispherectomy did not result in functional
deterioration or interference with activities of daily living.
However, serious concerns about visual outcome after
hemispherectomy should be raised for a child without
hemianopia who has other significant pre-existing visual
problems, for example, amblyopia or retinopathy of prematurity. In addition, it must be recognized that a com-

plete homonymous hemianopia is exclusionary for driving
privileges in many states.
When left hemisphere damage and right hemiparesis
occur before language development, up to the age of 5 or
6 years, there is little or no change in language function
after left hemispherectomy [67,69,71,72]. Language that
has shifted to the right hemisphere at an early age is rarely
completely normal [73,74], although earlier lesions produce less pronounced deficits. Transferred right hemisphere language function may be confirmed preoperatively in children old enough to cooperate with the
intracarotid amobarbital procedure [75], with continued
verbal responses after anesthesia of the damaged left
hemisphere.
The optimum surgical procedure for hemispheric ablation has been debated. A variety of procedures are performed at different centers [68]. Serious, potentially fatal
late neurologic deterioration occurred in some early patients after complete anatomic hemispherectomy leading
to modifications of the procedure [14]. One alternative is
the functional hemispherectomy of Rasmussen and Villemure [14,16,76], with resection of central cortex and
transection of all interhemispheric connections and all
white matter tracts in the remaining frontal, parietal,
occipital, and temporal regions. This procedure leaves
most of the hemisphere vascularized and anatomically
intact but functionally disconnected from the rest of the
brain. Postoperative EEG reveals ongoing EEG seizures,
spikes, and periodic lateralized epileptiform discharges
from the disconnected cortex [15], but the abnormalities
are of no clinical significance because they cannot spread
to activate other functioning brain regions. Compared with
complete anatomic hemispherectomy, functional hemispherectomy appears to offer a similar rate of seizure-free
outcomes with a lower rate of serious late neurologic complications [14]. Other surgical techniques more conservative
than the classic anatomic hemispherectomy include hemispheric deafferentation [77] and hemicorticectomy [78].

Epilepsy Surgery in Infants
Although many of the principles described above apply
at all ages, some considerations are specific to infants. The
approach to epilepsy surgery in the youngest age groups
has undergone significant recent evolution.
The earliest reports of epilepsy surgery in infants
included small numbers of patients who had good outcomes after hemispherectomy for Sturge-Weber syndrome
[22] or hemimegalencephaly [79-81]. These syndromes
were easily appreciated by physical examination and
computed tomography, and the affected patients had
devastating epilepsy. In 1990, Duchowny et al. [21]
reported three infants who were free of seizures and two
who had reduced seizures after focal cortical resection in
the first year of life, with surgery performed on the basis
of discovery of a tumor or hamartoma on MRI or a

Wyllie: Epilepsy Surgery 183

localized epileptogenic zone on subdural EEG. Wyllie et
al. [23] also reported favorable seizure outcomes with rare
or no seizures for nine (75%) of 12 infants with catastrophic epilepsy who had cortical resection or functional hemispherectomy at 6 weeks to 29 months of age

Figure 4. Case 1. Sagittal MRI (TR: 500 ms, TE: 12 ms) performed 1
day after complete resection of the tumor at 3 months of age. The air and
excess cerebrospinal fluid were resorbed within the next few days.

Figure 5. Case 2. Axial MRI (TR: 300 ms, TE: 4.4 ms) depicting right
temporo-occipital macrogyria (arrows) with indistinct gray-white junction and poor white matter arborization.

184

PEDIATRIC NEUROLOGY

Vol. 19 No. 3

(mean 5 15 months). Each of the patients had a
localized cortical lesion identified by MRI or PET scan,
including focal cortical dysgenesis, Sturge-Weber syndrome, low-grade tumor, or hemimegalencephaly [23].
An important benefit in several cases was the resumption of developmental progression.
A significant shift in the approach to identification of
infant epilepsy surgery candidates occurred after recognition that infantile spasms and hypsarrhythmia may occasionally be caused by a localized cortical lesion [19,20,
23-26,82,83]. In most cases, infantile spasms and
hypsarrhythmia are the result of diffuse brain injury or
abnormality, but in some cases, focal cortical lesions have
been identified by structural or functional neuroimaging.
Focal cortical dysgenesis is the most commonly reported
lesion associated with infantile spasms [19,20], but other
types of lesions have also been noted, including hemimegalencephaly, low-grade tumors, unilateral cerebral
infarction, and Sturge-Weber syndrome. Lesion location
has varied but has most often been temporoparietal or
occipital [19,26].
A surgical approach to patients with localization-related
infantile spasms was pioneered by Chugani et al. [18-20]
at the University of California at Los Angeles. Favorable
results have since been achieved for similar patients at
other centers [23,25]. In small series, resection of the
cortical lesion has resulted in cessation of the spasms and
resolution of the hypsarrhythmia, with good outcome rates
similar to those achieved in other clinical settings at older
ages. Chugani et al. [19] reported seizure freedom or at
least 90% seizure reduction for 18 (78%) of 23 patients
who initially presented with infantile spasms and underwent focal cortical resection or hemispherectomy at 5
months to 3 years 8 months of age (mean 5 18 months).
MRI and PET were critical in identifying the lesions for
resection, but other clues to localized cortical dysfunction
included a past history of partial seizures, focal abnormality on neurologic examination (hemiparesis or hemi-

Figure 6. Case 2. Interictal EEG revealing continual periodic sharp
waves from the right occipital region, with lesser involvement of the left
occipital region because of volume conduction.

Figure 8. Case 2. Axial MRI (TR: 3,640 ms, TE: 96 ms) performed 24
hours after right temporo-occipital resection, with blood layered in the
resection bed.

Figure 7. Case 2. (A) Ictal EEG at seizure onset demonstrating
cessation of the continual periodic sharp waves with evolution of a theta
rhythm maximum in the right occipital region. (B) Ictal EEG 25 seconds
after onset depicting repetitive spiking in the right hemisphere, maximum
in temporal-parietal-occipital regions.

anopia), or focal EEG features to the hypsarrhythmia.
Although patients did not have localized EEG seizure
patterns during spasms, other “softer” EEG findings of
focality were often present, including sharp waves predominantly over one cortical area; regional slowing, decreased background, or absent sleep spindles over one
cortical region or hemisphere; and unilateral electrodecremental events [70]. One of the strongest EEG clues to a
focal epileptogenic lesion was a history of earlier partial
seizures with localized EEG onset before the onset of
infantile spasms and hypsarrhythmia. Video EEG soon
after the onset of seizures, therefore, may provide critical
information for use later after the seizure semeiology and
ictal EEG are no longer localizing.
The mechanism by which focal cortical lesions may
manifest as infantile spasms and hypsarrhythmia is unknown. However, a clue may be the relationship between
age of onset of spasms and location of the lesion. Koo and
Hwang [26] found significant differences, with onset of

spasms earliest in patients with occipital lesions
(mean age 5 3 months), later in patients with centraltemporal-parietal lesions (mean age 5 6 months), and
latest in patients with frontal lesions (mean age 5 10
months). This timing coincides with the timing of most
active maturation in those regions, with rapid increases in
synaptic density and myelination proceeding sequentially
from the back to the front of the brain [26]. During this
period with dramatic increases in connectivity, aberrant
circuits may develop between the cortical lesion and other
cortical or subcortical regions. Infantile spasms appear to
result from a unique pathologic interaction between a focal
cortical lesion and normal developmental processes.
The application of epilepsy surgery to patients with
infantile spasms and hypsarrhythmia is revolutionary because these patients lack the hallmark of localizationrelated epilepsy later in life (i.e., partial seizures with
localized EEG onset). In these infants the strategy for
surgery is resection of the zone of cortical abnormality
[70] instead of the zone of EEG seizure onset. These
patients provide the most extreme example of poorly
localized EEG findings compatible with successful seizure
outcome after resection of a lesion detected by neuroimaging. It is not yet clear whether these observations from
a specific maturation epoch will be applicable to any other
clinical setting later in life.
Chugani et al. [19] noted no new postoperative deficits
after surgery in infancy, although a perioperative death

Wyllie: Epilepsy Surgery 185

was reported in an earlier series [18]. A perioperative
death also occurred in the series of infants reported by
Wyllie et al. [23], and two infant deaths were reported by
Vining et al. [84] in their hemispherectomy series. These
results emphasize the need for surgery to be offered only
to those infants with truly catastrophic epilepsy. Surgical
mortality is higher in infancy than later in life but can be
reduced by a dedicated team of pediatric neurosurgeons,
anesthesiologists, and intensivists. Surgical mortality at
any age, however, must be balanced against the mortality
of uncontrolled seizures treated medically, with an incidence estimated to be on the order of 1:295/year for
children and adolescents with epilepsy and learning disabilities [85].
Does early epilepsy surgery result in improved longterm developmental outcome? The answer to this important question is not yet known, but preliminary observations [19,23] suggest that early relief from catastrophic
epilepsy may allow resumption of developmental progression during critical stages of brain maturation. In addition,
Asarnow et al. [86] studied developmental outcome in 24
children with infantile spasms who had resective surgery
at a mean age of 21 months and found a significant
increase at 2 years after surgery compared with preoperative levels, although only four children had a normal rate
of development. Prospective studies are in progress to
further clarify the developmental effects of epilepsy surgery in infancy.
Weighing the Risk/Benefit Ratio for Surgery
Certain difficult questions arise in every case during the
evaluation for pediatric epilepsy surgery. No pat answers
are available, but one approach is presented here.
How many drugs must fail before surgery should be
considered? It appears that the answer differs between
patients on the basis of etiology, location of the epileptogenic zone, and severity of the epilepsy. Some children may be appropriate surgical candidates before
many drugs have been tried, for example when the
cause is a tumor in a safely resectable region such as the
anterior temporal lobe. Other children warrant a thorough exploration of medical options, for example, in the
setting of nonlesional temporal lobe epilepsy, lesional
extratemporal epilepsy, or a hemispheric syndrome. The
most conservative approach is appropriate for children
with complicated or incongruent findings on EEG and
neuroimaging or nonlesional extratemporal epilepsy.
Most of these children are not favorable candidates for
epilepsy surgery, and they deserve aggressive trials of
all reasonable medical options (including the ketogenic
diet if possible).
Decisions about surgery are strongly affected by the
severity of the epilepsy. For the infant with dozens of
seizures every day because of focal cortical dysplasia, it
does not require longer than a few months to explore the
reasonable medical options before surgery. For the child

186

PEDIATRIC NEUROLOGY

Vol. 19 No. 3

with infrequent seizures caused by nonlesional temporal
lobe epilepsy, it may be appropriate to steer the family
away from epilepsy surgery and work with medical
options for several years. However, the same seizure
frequency may later become intolerable in adolescence as
driving, employment, and independence emerge as major
issues.
How should age be weighed in the decision about
timing of surgery? The results so far from pediatric
epilepsy surgery series indicate that results with respect to
seizures are similar throughout infancy, childhood, and
adolescence [13]. However, age at surgery may be an issue
in relation to developmental or psychosocial outcome,
with earlier relief of seizures potentially providing some
benefit. Age may also be a factor in relation to operative
complications. Surgery in infancy may afford the potential
for greater neurologic recovery because of developmental
plasticity, but it also entails a higher risk of mortality than
does surgery later in childhood.
Conclusions
Epilepsy surgery is a well-established treatment for
intractable localization-related epilepsy in adolescents and
adults [27], but it is less often considered for children. The
usual delay from onset of intractability to surgery remains
in the range of 12-15 years at most centers, reflecting a
reluctance to consider epilepsy surgery during childhood.
Results from pediatric surgical series so far do not justify
this reluctance but instead suggest that children should be
referred for surgical evaluation at whatever age they
manifest with severe intractable localization-related epilepsy. The potential risk/benefit ratio for surgery must then
be carefully weighed for each child in light of the many
complex age-related issues. Complicated cases warrant
referral to specialized centers with extensive pediatric
epilepsy surgery experience.
Case surgery performed by Dr. William Bingaman.

References
[1] Polkey CE. Selection of patients with intractable epilepsy for
resective surgery. Arch Dis Child 1980;55:841-4.
[2] Davidson S, Falconer M. Outcome of surgery in 40 children with
temporal-lobe epilepsy. Lancet 1975;2:1260-3.
[3] Duchowny M, Levin B, Jayakar P, et al. Temporal lobectomy in
early childhood. Epilepsia 1992;33:298-303.
[4] Falconer M. Significance of surgery for temporal epilepsy in
childhood and adolescence. J Neurosurg 1970;33:233-52.
[5] Green JR. Surgical treatment of epilepsy during childhood and
adolescence. Surg Neurol 1977;8:71-80.
[6] Meyer FB, Marsh WR, Laws ER, Sharbrough FW. Temporal
lobectomy in children with epilepsy. J Neurosurg 1986;64:371-6.
[7] Mizrahi EM, Kellaway P, Grossman RG, et al. Anterior temporal lobectomy and medically refractory temporal lobe epilepsy of
childhood. Epilepsia 1990;31:302-12.
[8] Nespeca M, Wyllie E, Lüders H, et al. Subdural electrodes in
infants and young children. J Epilepsy 1991;3 (Suppl):107-24.

[9] Rasmussen TB. Surgical aspects. In: Wise G, ed. Topics in child
neurology. Englewood Cliffs, NJ: Spectrum Publications, 1977:143-53.
[10] Whittle IR, Ellis HJ, Simpson DA. The surgical treatment of
intractable childhood and adolescent epilepsy. Aust N Z J Surg 1981;51:
190-6.
[11] Wyllie E, Chee M, Granstrom ML, et al. Temporal lobe
epilepsy in childhood. Epilepsia 1993;34:859-68.
[12] Adelson PD, Peacock WJ, Chugani HT, et al. Temporal and
extended temporal resections for the treatment of intractable seizures in
early childhood. Pediatr Neurosurg 1992;18:169-78.
[13] Wyllie E, Comair YG, Kotagal P, Bulacio J, Bingaman W,
Ruggieri P. Seizure outcome after epilepsy surgery in children and
adolescents. Ann Neurol 1998;44.
[14] Rasmussen T. Hemispherectomy for seizures revisited. J Can
Sci Neurol 1983;10:71-8.
[15] Smith SJM, Andermann F, Villemure JG, Rasmussen TB,
Quesney LF. Functional hemispherectomy: EEG findings, spiking from
isolated brain postoperatively, and prediction of outcome. Neurology
1991;41:1790-4.
[16] Tinuper P, Andermann F, Villemure JG, Rasmussen TB,
Quesney LF. Functional hemispherectomy for treatment of epilepsy
associated with hemiplegia: Rationale, indications, results, and comparison with callosotomy. Ann Neurol 1988;24:27-34.
[17] Peacock WJ, Wehby-Grant MC, Shields WD, et al. Hemispherectomy for intractable seizures in children: A report of 58 cases.
Childs Nerv Syst 1996;12:376-84.
[18] Chugani HT, Shewmon DA, Peacock WJ, Shields WD, Mazziotta JC, Phelps ME. Surgical treatment of intractable neonatal-onset
seizures: The role of positron emission tomography. Neurology 1988;38:
1178-88.
[19] Chugani HT, Shewmon DA, Shields WD, et al. Surgery for
intractable infantile spasms: Neuroimaging perspectives. Epilepsia 1993;
34:764-71.
[20] Chugani HT, Shields WD, Shewmon DA, Olson DM, Phelps
ME, Peacock WJ. Infantile spasms: I. PET identifies focal cortical
dysplasia in cryptogenic cases for surgical treatment. Ann Neurol
1990;27:406-13.
[21] Duchowny MS, Resnick TJ, Alvarez LA, Morrison G. Focal
resection for malignant partial seizures in children. Neurology 1990;40:
980-4.
[22] Hoffman HH, Hendrick EB, Dennis M, Armstrong D. Hemispherectomy for Sturge-Weber syndrome. Childs Brain 1979;5:233-48.
[23] Wyllie E, Comair YG, Kotagal P, Raja S, Ruggieri P. Epilepsy
surgery in infants. Epilepsia 1996;37:625-37.
[24] Wyllie E, Comair Y, Ruggieri P, Raja S. Epilepsy surgery in the
setting of periventricular leukomalacia and focal cortical dysplasia.
Neurology 1996;46:839-41.
[25] Brockhaus A, Elger CE. Complex partial seizures of temporal
lobe origin in children of different age groups. Epilepsia 1995;36:117381.
[26] Koo B, Hwang P. Localization of focal cortical lesions influences age of onset of infantile spasms. Epilepsia 1996;37:1068-71.
[27] Engel JJ. Surgery for seizures. N Engl J Med 1996;334:647-52.
[28] Jack CRJ. MRI-based hippocampal volume measurements in
epilepsy. Epilepsia 1994;35 (Suppl 6):S21-9.
[29] Berkovic SF, Andermann F, Olivier A, et al. Hippocampal
sclerosis in temporal lobe epilepsy demonstrated by magnetic resonance
imaging. Ann Neurol 1991;29:175-82.
[30] Cascino GD, Jack CRJ, Parisi JE, et al. Magnetic resonance
imaging-based volume studies in temporal lobe epilepsy: Pathological
correlations. Ann Neurol 1991;30:31-6.
[31] Cendes F, Andermann F, Dubeau F, et al. Early childhood
prolonged febrile convulsions, atrophy and sclerosis of mesial structures,
and temporal lobe epilepsy: MRI volumetric study. Neurology 1993;43:
1083-7.
[32] Cendes F, Andermann F, Gloor P, et al. MRI volumetric
measurement of amygdala and hippocampus in temporal lobe epilepsy.
Neurology 1993;43:719-24.

[33] Jack CRJ, Sharbrough FW, Cascino GD, Hirschorn KA,
O’Brien PC, Marsh WR. Magnetic resonance image-based hippocampal
volumetry: Correlation with outcome after temporal lobectomy. Ann
Neurol 1992;31:138-46.
[34] Berkovic SF, McIntosh AM, Kalnins RM, et al. Preoperative
MRI predicts outcome of temporal lobectomy: An actuarial analysis.
Neurology 1995;45:1358-63.
[35] Engel JJ, Henry TR, Risinger MW, et al. Presurgical evaluation
for partial epilepsy: Relative contributions of chronic depth-electrode
recordings versus FDG-PET and scalp-sphenoidal ictal EEG. Neurology
1990;40:1670-7.
[36] Wieser H, Engel JJ, Williamson PD, Babb TL, Gloor P.
Surgically remediable temporal lobe syndromes. In: Engel JJ, ed.
Surgical treatment of the epilepsies. Philadelphia: Lippincott-Raven,
1993:49-63.
[37] Patrick S, Berg A, Spencer SS. EEG and seizure outcome after
epilepsy surgery. Epilepsia 1995;36:236-40.
[38] Salanova V, Markand ON, Worth R. Clinical characteristics
and predictive factors in 98 patients with complex partial seizures treated
with temporal resection. Arch Neurol 1994;51:1008-13.
[39] Arruda F, Cendes F, Andermann F, et al. Mesial atrophy and
outcome after amygdalohippocampectomy or temporal lobe removal.
Ann Neurol 1996;40:446-50.
[40] Armon C, Radtke RA, Friedman AH, Dawson DV. Predictors
of outcome of epilepsy surgery: Multivariate analysis with validation.
Epilepsia 1996;37:814-21.
[41] Pelaez JM, Wyllie E, Bulacio J, et al. Epilepsy surgery for
hippocampal sclerosis in preadolescent children (abstract). Epilepsia
1997;38 (Suppl 8):72.
[42] Wyllie E. Clinical applications of neuroimaging: Surgical
planning in pediatrics. In: Cascino GD, Jack CRJ, eds. Neuroimaging in
Epilepsy. Stoneham, MA: Butterworths Heinemann, 1996:261-72.
[43] Palmini A, Andermann F, Olivier A, et al. Focal neuronal
migration disorders and intractable partial epilepsy: A study of 30
patients. Ann Neurol 1991;30:741-9.
[44] Raymond AA, Fish DR, Sisodiya SM, Alsanjari N, Stevens
JM, Shorvon SD. Abnormalities of gyration, heterotopias, tuberous
sclerosis, focal cortical dysplasia, microdysgenesis, dysembryoplastic
neuroepithelial tumour and dysgenesis of the archicortex in epilepsy:
Clinical, EEG and neuroimaging features in 100 adult patients. Brain
1995;118:629-60.
[45] Wyllie E, Baumgartner C, Prayson R, et al. The clinical
spectrum of focal cortical dysplasia and epilepsy. J Epilepsy 1994;7:30312.
[46] Ambrosetto G. Treatable partial epilepsy and unilateral opercular neuronal migration disorder. Epilepsia 1993;34:604-8.
[47] Marson AG, Kadir ZA, Chadwick DW. New antiepileptic
drugs: A systematic review of their efficacy and tolerability. BMJ
1996;313:1169-74.
[48] Katz A, Awad IA, Kong AK, et al. Extent of resection in
temporal lobectomy for epilepsy. II. Memory changes and neurologic
complications. Epilepsia 1989;30:763-71.
[49] Chelune GJ, Naugle RI, Lüders H, Awad IA. Prediction of
cognitive change as a function of preoperative ability status among
temporal lobectomy patients seen at 6-month follow-up. Neurology
1991;41:399-404.
[50] Adams CBT, Beardsworth ED, Oxbury SM, Oxbury JM,
Fenwick PBC. Temporal lobectomy in 44 children: Outcome and
neuropsychological follow-up. J Epilepsy 1990;3:157-68.
[51] Harbord MG, Manson JI. Temporal lobe epilepsy in childhood: Reappraisal of etiology and outcome. Pediatr Neurol 1987;3:
263-8.
[52] Szabó CA, Wyllie E, Stanford LD, et al. Neuropsychological
outcome of temporal lobe resection in children with epilepsy. Epilepsia
1998;39:814-9.
[53] DeVos K, Wyllie E, Geckler C, Kotagal P, Comair Y. Language
dominance in patients with early childhood tumors near left hemisphere
language areas. Neurology 1995;45:349-56.

Wyllie: Epilepsy Surgery 187

[54] Spencer SS. Long-term outcome after epilepsy surgery. Epilepsia 1996;37:807-13.
[55] Fish DR, Smith SJ, Quesney LF, Andermann F, Rasmussen T.
Surgical treatment of children with medically intractable frontal or
temporal lobe epilepsy: Results and highlights of 40 years’ experience.
Epilepsia 1993;34:244-7.
[56] Zentner J, Hufnagel A, Ostertun B, et al. Surgical treatment of
extratemporal epilepsy: Clinical, radiologic, and histopathologic findings
in 60 patients. Epilepsia 1996;37:1072-80.
[57] Laxer K, Garcia PA, Ng T. 1H magnetic resonance spectroscopy. In: Wyllie E, ed. The treatment of epilepsy: Principles and practice.
Baltimore: Williams & Wilkins, 1997:1030-5.
[58] Connelly A, Jackson GD, Duncan JS, King MD, Gadian DG.
Magnetic resonance spectroscopy in temporal lobe epilepsy. Neurology
1994;44:1411-7.
[59] Cross JH, Connelly A, Jackson GD, Johnson CL, Neville BGR,
Gadian DG. Proton magnetic resonance spectroscopy in children with
temporal lobe epilepsy. Ann Neurol 1996;39:107-13.
[60] Marks DA, Katz A, Hoffer P, Spencer SS. Localization of
extratemporal epileptic foci during ictal single photon emission computed tomography. Ann Neurol 1992;31:250-5.
[61] Harvey AS, Bowe JM, Hopkins IJ, Shield LK, Cook DJ,
Berkovic SF. Ictal 99mTc-HMPAO single photon emission computed
tomography in children with temporal lobe epilepsy. Epilepsia 1993;34:
869-77.
[62] Harvey AS, Hopkins IJ, Bowe JM, Cook DJ, Shield LK,
Berkovic SF. Frontal lobe epilepsy: Clinical seizure characteristics and
localization with ictal 99mTc-HMPAO SPECT. Neurology 1993;43:
1966-80.
[63] Jayakar P, Alvarez LA, Duchowny MS, Resnick TJ. A safe and
effective paradigm to functionally map the cortex in childhood. J Clin
Neurophysiol 1992;9:288-93.
[64] Lesser RP, Gordon B, Fisher R, Hart J, Uematsu S. Subdural
grid electrodes in surgery of epilepsy. In: Lüders H, ed. Epilepsy surgery.
Philadelphia: Lippincott-Raven, 1991:399-408.
[65] Sperling MR, Saykin AJ, Roberts FD, French JA, O’Connor
MJ. Occupational outcome after temporal lobectomy for refractory
epilepsy. Neurology 1995;45:970-7.
[66] Gilliam F, Wyllie E, Kashden J, et al. Epilepsy surgery
outcome: Comprehensive assessment in children. Neurology 1997;48:
1368-74.
[67] Andermann F. Functional hemispherectomy: Clinical indications and outcome. In: Wyllie E, ed. The treatment of epilepsy: Principles
and practice. Baltimore: Williams & Wilkins, 1997:1074-80.
[68] Holthausen H, May TW, Adams CTB, et al. Seizures post
hemispherectomy. In: Tuxhorn I, Holthansen H, Boenigk HE, eds.
Pediatric epilepsy syndromes and their surgical treatment. London: John
Libbey, 1997:749-73.

188

PEDIATRIC NEUROLOGY

Vol. 19 No. 3

[69] Wilson PJE. Cerebral hemispherectomy for infantile hemiplegia: A report of 50 cases. Brain 1970;93:147-80.
[70] Peacock WJ, Comair Y, Chugani HT, Shewmon DA, Shields
WD. Epilepsy surgery in childhood. In: Lüders HO, ed. Epilepsy surgery.
Philadelphia: Lippincott-Raven, 1991:589-98.
[71] Basser LS. Hemiplegia of early onset and the faculty of speech
with special reference to the effects of hemispherectomy. Brain 1962;
85:427-60.
[72] McFie J. The effects of hemispherectomy on intellectual functioning in cases of infantile hemiplegia. J Neurol Neurosurg Psychiatry
1961;24:240-9.
[73] Dennis M. Capacity and strategy for syntactic comprehension
after left or right hemidecortication. Brain Lang 1980;10:287-317.
[74] Rankin JM, Aram DM, Horwitz SJ. Language ability in right
and left hemiplegic children. Brain Lang 1981;14:292-306.
[75] Szabó CA, Wyllie E. Intracarotid amobarbital testing for
language and memory dominance in children. Epilepsy Res 1993;15:23946.
[76] Villemure JG. Hemispherectomy: Techniques and complications. In: Wyllie E, ed. The treatment of epilepsy: Principles and practice.
Baltimore: Williams & Wilkins, 1997:1081-6.
[77] Schramm J, Behrens E, Entzian W. Hemispherical deafferentation: An alternative to functional hemispherectomy. Neurosurgery
1995;36:509-16.
[78] Winston KR, Welch K, Adler JR, Erba G. Cerebral hemicorticectomy for epilepsy. J Neurosurg 1992;77:889-95.
[79] King M, Stephenson JBP, Ziervogel M, Doyle D, Galbraith S.
Hemimegalencephaly—A case for hemispherectomy? Neuropediatrics
1985;16:46-55.
[80] Vigevano F, Bertini E, Boldrini R, et al. Hemimegalencephaly
and intractable epilepsy: Benefits of hemispherectomy. Epilepsia 1989;
30:833-43.
[81] Vigevano F, DiRocco C. Effectiveness of hemispherectomy in
hemimegalencephaly with intractable seizures. Neuropediatrics 1990;21:
222-3.
[82] Asanuma H, Wakai S, Tanaka T, Chiba S. Brain tumors
associated with infantile spasms. Pediatr Neurol 1995;12:361-4.
[83] Alvarez LA, Shinnar S, Moshe SL. Infantile spasms due to
unilateral cerebral infarcts. Pediatrics 1987;79:1024-6.
[84] Vining EPG, Freeman JM, Pillas DJ, et al. Why would you remove
half a brain? The outcome of 58 children after hemispherectomy—The
Johns Hopkins experience: 1968-1996. Pediatrics 1997;100:163-71.
[85] Nashef L, Fish DR, Garner S, Sander JWAS, Shorvon SD.
Sudden death in epilepsy: A study of incidence in a young cohort with
epilepsy and learning difficulty. Epilepsia 1995;36:1187-93.
[86] Asarnow RF, LoPresti C, Guthrie D, et al. Developmental
outcomes in children receiving resection surgery for medically intractable infantile spasms. Dev Med Child Neurol 1997;39:430-40.