Epilepsia, 47(1):176180, 2006

Blackwell Publishing, Inc.

C 2006 International League Against Epilepsy

Efficacy of Dexamathasone on Cerebral Swelling and Seizures

during Subdural Grid EEG Recording in Children
Takashi Araki, Hiroshi Otsubo, Yuko Makino, Irene Elliott, Koji Iida, Ayako Ochi,
Shelly K. Weiss, Sylvester H. Chuang, James T. Rutka, and O. Carter Snead III
Divisions of Neurosurgery, Neurology, and Diagnostic Imaging; Department of Pediatrics, The Hospital for Sick Children and
University of Toronto, Toronto, Ontario, Canada

Summary: Purpose: To evaluate the impact of steroid treatment on cerebral swelling and seizures during subdural grid EEG
(SGEEG) monitoring.
Methods: We reviewed data from 37 pediatric patients with
intractable epilepsy who underwent SGEEG monitoring and divided the patients into those who received dexamethasone and
those who did not. We then correlated administration of steroids
to incidence of cerebral swelling on computed tomography (CT)
scans and to frequency of seizures during SGEEG.
Results: Twenty-three patients received dexamethasone prophylactically every 6 hours (dosage range, 17.5 mg; mean, 3.2
mg) from the first day of SGEEG placement (group A); 14 patients received no dexamethasone (group B). Eight (21.6%) of 37
patients experienced cerebral swelling on CT: two (9%) were in
group A, and six (42.9%) were in group B (p < 0.05). SGEEG

monitoring time for recording habitual seizures that localized

cortical areas for surgical excision was longer in group A (1
6 days; mean, 3.0) than it was in group B (13 days; mean,
2.2), (p < 0.05). Habitual seizures were recorded in 36 patients.
One group A patient experienced obtundation due to cerebral
swelling, and monitoring in this patient was discontinued.
Conclusions: The prophylactic administration of steroids
to pediatric patients during SGEEG monitoring is efficacious for reducing cerebral swelling. Although it decreases
the frequency of habitual seizures and increases seizuremonitoring time, dexamethasone reduces the risk of complications from cerebral swelling during the SGEEG procedure.
Key Words: DexamethasoneSubdural grid EEGCerebral
swellingChildrenSeizure.

In the early 1960s, French and Galicich (1) introduced

steroids as treatments for cerebral edema. Glucocorticoids
decreased the edema that was associated with brain tumors
(2), and in the treatment of epilepsy, adrenocorticotropic
hormone (ACTH) relieved infantile spasms in patients
with West syndrome (35). Dexamethasone, a synthetic
glucocorticoid, was frequently administered to decrease
cerebral swelling; however, a meta-analysis of head-injury
trials reported that steroid treatment had no overall benefit
in outcome (6).
The role of steroids for treating brain edema in patients
undergoing invasive video-EEG is controversial (7). Dexamethasone has a neuroprotective effect during subdural
grid EEG (SGEEG) monitoring in patients with refractory
epilepsy. However, the steroid also produces counterproductive effects during invasive monitoring for localizing
the epileptogenic zone. Although dexamethasone prevents

cerebral swelling, it also suppresses the seizures needed

for localization.
We previously reported on the complications of SGEEG
monitoring in children with intractable epilepsy (8). We
hypothesized that steroids reduced brain swelling and decreased seizure frequency during SGEEG monitoring. To
evaluate the impact of steroid treatment, we correlated
use of dexamethasone to incidence of cerebral swelling
and seizure frequency. We retrospectively compared computed tomography (CT) findings, seizure profiles, electrode types, surgical procedures, monitoring durations,
and ages of pediatric patients who received steroids during
SGEEG with the same parameters of patients who were
not given steroids.
METHODS
Patient population
We reviewed the hospital charts of 37 patients who underwent SGEEG monitoring between May 1995 and October 2002 at The Hospital for Sick Children. Multiple
anticonvulsant therapies for localization-related neocortical epilepsy failed for all patients, and they had SGEEG

Accepted August 6, 2005.

Address correspondence and reprint requests to Dr. H. Otsubo at Division of Neurology, Department of Pediatrics, The Hospital for Sick
Children, 555 University Avenue, Toronto, Ontario, Canada, M5G 1X8.
E-mail: hiroshi.otsubo@sickkids.ca

176

DEXAMATHASONE AND CEREBRAL SWELLING AND SEIZURES

monitoring to localize the epileptogenic zone for surgical excision. All patients or their parents gave informed
consent.
Subdural grid placement
We made individualized subdural grid arrays for each
patient by using data from seizure semiology, interictal
and ictal scalp recordings from EEG, and interictal spike
sources and somatosensory evoked fields from magnetoencepholography (MEG). We used 5-mm diameter electrodes embedded in a silicone rubber (Silastic) sheet. Interelectrode distance ranged from 10 to 13 mm.
For grid placement, we performed a craniotomy with
the patient under general anesthesia. After removing the
bone flap, we made a large M-shaped dural opening based
on the location of the sagittal sinus, placed the subdural
grid array on the exposed cortical surface of the epileptic hemisphere, and sutured the grid array to the dura to
prevent displacement. We closed the dura with a large
subgaleal membrane graft. We tunneled individual cables
separately through the scalp remote from the incision and
secured them in place with purse-string sutures. In addition to placing the grid on the epileptogenic hemisphere,
we used depth electrodes in the mesial temporal regions
and subdural strip electrodes for the margin of the grid or
the opposite hemisphere (Ad-Tech, Racine WI, U.S.A.).
SGEEG monitoring
We discontinued routine anticonvulsants from patients
before their admission to the Pediatric Intensive Care
Unit (PICU), where we performed SGEEG monitoring
and functional mapping. Immediately after subdural grid
implantation, we verified the location of electrodes from
plain scalp x-ray films and started dexamethasone treatment. Patients also received prophylactic antibiotics (a
combination of cefotaxime and vancomycin) throughout
the monitoring period. We also gave ranitidine to prevent
gastrointestinal bleeding. During monitoring, we evaluated patients white blood cell (WBC) counts, glucose
levels, and physical findings (meningeal signs, fever, and
systolic blood pressure). Patients received brain CT scans
on the day after implantation and whenever any neurologic deficits occurred. We conducted functional cortical
mapping on the third day of monitoring.
For localization of the epileptogenic zone, we planned
to capture at least three habitual seizures. When seizures
became longer than 5 minutes, we gave the patient 0.1
to 0.3 mg/kg of intravenous diazepam. After a sufficient
number of seizures had been recorded, we removed the
grid and performed cortical excision or lobectomy of the
epileptogenic zone or both with the patient under general
anesthesia.
Data analysis
We retrospectively analyzed the 37 patients by first classifying them as to whether they received dexamethasone

177

before SGEEG. Prior to 1998, we routinely gave dexamethasone to patients at the beginning of SGEEG; we
classified these patients as group A. However, because
these patients with intractable and frequent seizures who
received dexamethasone experienced silent days without
habitual seizures during SGEEG monitoring, in 1998, we
discontinued prophylactic dexamethasone if patients had
no brain swelling at grid placement; we classified these patients as group B. Group A consisted of 23 patients who
received dexamethasone [dosage range, 17.5 mg every
6 hours (q6h); mean, 3.2 mg q6h] prophylactically from
the start of recording, and group B contained 14 patients
who did not receive the steroid. Thirteen (57%) of the
group A patients and seven (50%) of the group B patients had lesions on magnetic resonance imaging (MRI).
A neuroradiologist (S.H.C.), blind to dexamethasone administration, analyzed each patients CT scans for cerebral swelling as evidenced by (a) midline shift (>5 mm);
(b) sulcus effacement; (c) shape of the ventricle; and (d)
compression of basal cistern. To determine monitoring durations, we used two parameters from the patients seizure
profile: (a) the day of initial seizure onset after grid placement, and (b) the number of days between the day of initial seizure onset and the day when we recorded sufficient
seizures for detecting the epileptogenic zone.
We analyzed continuous variables by using the nonparametric MannWhitney test and dichotomous variables
using the Fishers exact test. We used univariate analysis to test correlations between incidences of cerebral
swelling and (a) the total number of electrodes on subdural and other depth electrodes; (b) the combined number
of grids, strip electrodes, and depth electrodes used; (c)
hemispheres and locations of grid arrays; and (d) patients
ages.
RESULTS
Ages of the 37 patients ranged from 2 to 19 years (mean,
11.9 years). Sixteen were male, and 21 were female patients. Subdural grids were placed over the left hemisphere
in 19 (51%) patients and the right hemisphere in 18 (49%)
patients. Thirty-three (86%) patients had strip electrodes,
and 20 (54%) had depth electrodes in addition to a grid
array (range, 80153 electrodes; mean, 109 electrodes).
Eight (22%) of 37 patients had cerebral swelling on
CT scans during SGEEG monitoring. Three (8%) patients
had complications during the monitoring session that were
not directly related to cerebral swelling: two (5%) had
wound infections, and in one (3%), a prominent subgaleal
CSF fluid accumulation developed that required lumbar
drainage for 5 days.
Cerebral swelling
Six (43%) patients in group B experienced cerebral
swelling during SGEEG, whereas only two (9%) in group
A had swelling. The difference in incidence of cerebral
Epilepsia, Vol. 47, No. 1, 2006

178

T. ARAKI ET AL.
TABLE 1. Profiles and CT findings for eight patients with cerebral swelling

Patient
No. of Electrodes
Midline
Sulcal
Compression Compression of
No.
Group Age Sex Hemisphere
in grid
structure shift effacement of ventricle
basal cistern
1
2
3
4
5
6
7
8

A
A
B
B
B
B
B
B

7
13
10
11
13
18
18
18

M
F
F
F
M
F
M
F

R
R
L
L
L
L
R
R

91
94
120
111
66
89
104
116

>1 cm

<5 mm

<5 mm

>1 cm
>5 mm

+
+
+
+
+
+
+

+
+
+

+
+
+

Neurological
findings
Obtundation

Right motor weakness

F, female; L, left; M, male; R, right.

swelling between group A and group B was significant

(p < 0.05). Table 1 summarizes profiles and CT findings
for these eight patients.
Cerebral swelling was identified from CT as a midline
shift in one group A and four group B patients, as sulcal
effacement in two group A and five group B patients, as
compression of the ventricle in one group A and three
group B patients, and as compression of basal cistern in
one group A and three group B patients.

FIG. 1. Patient 1 from group A (see Table 1) had obtundation and

subsequent unresponsiveness on the third day after subdural grid
placement in the right hemisphere. On the fourth day, computed
tomography showed compression of the right ventricle and midline
shift. The grid was removed without cortical excision.

Epilepsia, Vol. 47, No. 1, 2006

Patient 1 from group A experienced obtundation on the

third day after right hemispheric grid implantation. CT
showed extensive cerebral swelling on day 4, consisting
of compression of the right ventricle and midline shift
with progressive unresponsiveness (Fig. 1). We removed
the grid on day 4. Because only subclinical seizures had
been recorded, we performed no cortical excision. Patient 1 never experienced neurologic sequelae, temporary
or permanent, yet continued to have seizures after grid
removal.
Patient 2, also from group A, experienced no neurologic
sequelae from the cerebral swelling. We gave no treatment
and kept her under observation.
For five (patients 37) of the six group B patients with
cerebral swelling, we administered intravenous dexamethasone as treatment; none needed surgical intervention.
Patient 6 showed right motor weakness 2 days after left
hemispheric grid placement. During the perioperative period, this patient received 1-deamino-8-D-arginine vasopressin (DDAVP) prophylactically for abnormal platelet
function. CT showed compression of the left ventricle and
cistern (Fig. 2). We temporarily removed the bone flap and
continued SGEEG monitoring. On the sixth day of monitoring, we removed the subdural grid, and 4 days after
grid removal, we performed a lesionectomy with additional cortical excision. We covered the bone flap 2 weeks
after cortical excision. The patient experienced impaired
speech, hemianopia, and mild motor weakness of the upper extremity that persisted for 8 weeks. The patient was
seizure free 24 months after surgery.
The number of grid electrodes in the eight patients
with cerebral swelling ranged from 66 to 120 (mean, 97),
whereas the number in the 29 patients without cerebral
swelling ranged from 40 to 125 (mean, 95).
Seizure frequency
Of 37 patients, we recorded habitual seizures in 22 of 23
group A and in all 14 group B patients. The total number of
seizures ranged from 2 to 70 (mean, 18.3 seizures); group
A had a range of two to 48 seizures (mean, 18.7); group B,
five to 70 (mean, 17.6). Time of first seizure occurrence
after subdural grid placement for all patients ranged from

DEXAMATHASONE AND CEREBRAL SWELLING AND SEIZURES

179

10.8 4.3 years; p = 0.3) did not correlate with cerebral

swelling.
DISCUSSION

FIG. 2. Patient 6 from group B (see Table 1) experienced prominent right motor weakness 2 days after subdural grid placement in
the left hemisphere. Computed tomography showed compression
of ventricle and cistern. The bone flap was temporarily removed,
and monitoring was continued. The subdural grid was removed
on the sixth day of monitoring. Lesionectomy and additional cortical excision were performed 4 days after subdural grid removal.
Two weeks after cortical excision, the bone flap was again covered. The patients impaired speech, hemianopia, and mild motor
weakness of the upper extremity persisted for 8 weeks. The patient was seizure free 24 months after surgery.

day 1 to day 5 (mean, 2.1 days), with similar times for

group A (day 1 to day 5; mean, 2.0 days) and group B
(day 1 to day 5; mean, 2.3 days).
Total monitoring time, including time for functional
mapping that was performed after at least three seizures
occurred, ranged from 2 to 10 days (mean, 3.9 days). For
group A, 1 to 6 days (mean, 3.0 days) of recording were required to capture three habitual seizures for localizing the
epileptogenic zone; whereas in group B, 1 to 3 days (mean,
2.2 days) were needed. The difference between group A
and group B in number of days required for seizure recording was significant (p < 0.05).
Other factors
In univariate analysis, the total number of electrodes on
subdural grids (group A, 107 14; group B, 110 18;
p = 0.9), the combined number of grids, strip electrodes,
and depth electrodes (group A, 4 2; group B, 4 2;
p = 0.6), hemispheres and locations of grid arrays (group
A: right, 11; left, 12; bilateral, 2; group B: right, 6; left, 8),
and patients ages (group A, 13.8 4.3 years; group B,

Cerebral swelling is a serious complication of SGEEG

monitoring in children because swelling interferes with
localization and resection of the epileptogenic zone and
identification of functional cortex. Moreover, cerebral
swelling that is refractory to intervention can provoke
cerebral herniation and result in death. Morrison et al.
(9) reported that four of 48 pediatric epilepsy patients
with cerebral swelling required a second-stage operation
sooner than the planned surgery; one of the four patients
died of malignant cerebral swelling. Pilcher et al. (10) also
reported that a pathologic elevation of intracranial pressure during SGEEG in one patient resulted in death. Lee
et al. (11) reported that the overall morbidity rate during invasive monitoring in adult patients was 16% (eight of 49
patients), including one case who had cerebral swelling
without steroids. Hamer et al. (7) reported that increased
intracranial pressure occurred in 2.5% or five of 187 adult
patients during monitoring sessions that included the administration of dexamethasone. They also noted that more
grids, longer monitoring times, and left-side grid insertion significantly correlated to more complications. We
saw cerebral swelling on CT scans after grid implantation
in eight (22%) of 37 patients. In two patients, we needed
temporarily to remove the bone flap; one of these patients
did not successfully complete SGEEG monitoring. None
of our patients had lethal complications, however.
Silberbush et al. (12) reported that MRI after grid implantation reduced local artifacts and demonstrated complications better than CT did. But in our pediatric patients
who required extensive hemispheric grids with more than
100 electrodes and bundles of cables, MRI did not clarify
intracranial complications.
Because the subdural grid is a massive foreign body
implanted in the patient with neocortical epilepsy, many
factors can contribute to the pathogenesis of cerebral
swelling. The hemispheric grid not only can provoke the
mass effect, but also can compress extensive cortical veins
to interrupt venous circulation. Klatzo and Seitelberger
(13) have categorized cerebral edema as vasogenic or cytotoxic. Vasogenic edema, commonly seen as peritumoral
edema, is highly responsive to corticosteroids (2). However, cytotoxic edema due to cerebral infarction or hemorrhage in patients with stroke does not respond to steroids
(1416). In patients who have extensive cortical areas of
venous congestion due to subdural grids, the mechanism
of cerebral swelling may be vasogenic, and steroids can be
effective in treating the cerebral swelling. In our patients,
we saw no signs of infarction on CT, and clinical deterioration after steroid administration was reversible, indicating
that the changes were not those of cytotoxic damage from
Epilepsia, Vol. 47, No. 1, 2006

180

T. ARAKI ET AL.

an infarction. To avoid mechanical compression of veins

by grids, we ensured that the edges of large grids did not
compress or impede the outflow of major draining veins,
such as the vein of Labbe or Trolard (8).
For adult patients with subdural grid electrodes, Hamer
et al. (7) reported that administration of steroids decreased the complication rate. Our results also indicated
that steroids are efficacious in decreasing the incidence
of cerebral swelling during SGEEG monitoring in children. For 22 of 23 patients receiving steroids, we were
able to complete SGEEG monitoring and capture a sufficient number of habitual seizures to localize the resection area. In seven of the eight patients who had cerebral
swelling and were treated with intravenous dexamethasone, we were able to record habitual seizures and perform surgical resections; these patients had no permanent
deficits.
For children with intractable epilepsy, steroids affect
seizure frequency during SGEEG. Dexamethasone has
been used as therapy for pediatric epilepsy (17,18), and
ACTH is an accepted therapy for infantile spasms (35).
ACTH reduces neuronal excitability in infantile spasms
by inducing steroid release and by influencing receptors
in the central nervous system (3,5). Because our group A
patients required a longer recording time for localization
of the cortical excision area, dexamethasone may reduce
ictogenesis and the production of clinical seizures during
SGEEG monitoring, even in patients who have frequent
daily seizures. Nonetheless, we succeeded in localizing
the epileptogenic zone for cortical resection in most patients who were given steroids. When steroids are administered to avoid the incidence of cerebral swelling during
SGEEG, we recommend close observation of the patients
neurologic, thrombocytic, and neuroimaging findings.
We conclude that prophylactic dexamethasone may potentially decrease the incidence of cerebral swelling and
safely treat swelling in some children with intractable
neocortical epilepsy during SGEEG monitoring with
extensive grids and strips. Although it decreases seizure
frequency and increases recording periods for capturing
habitual seizures, dexamethasone administration is efficacious in reducing the risk of complications due to cerebral
swelling during SGEEG monitoring. We suggest a ran-

Epilepsia, Vol. 47, No. 1, 2006

domized clinical trial with a larger number of patients to

study further the use of dexamethasone before SGEEG.
Acknowledgment: We thank Mrs. Carol L. Squires for her
editorial assistance.

REFERENCES
1. French LA, Galicich JH. The use of steroids for control of cerebral
edema. Clin Neurosurg 1964;10:212223.
2. Leenders KL, Beaney RP, Brooks DJ, et al. Dexamethasone treatment of brain tumor patients. Neurology 1985;35:16101616.
3. Snead OC III. How does ACTH work against infantile spasms?
Bedside to bench. Ann Neurol 2001;49:288289.
4. Mackay MT, Weiss SK, Adams-Webber T, et al. Practice parameter: medical treatment of infantile spasms: report of the American
Academy of Neurology and the Child Neurology Society. Neurology
2004;62:16681681.
5. Brunson KL, Avishai-Eliner S, Baram TZ. ACTH treatment of infantile spasm: mechanisms of its effects in modulation of neuronal
excitability. Int Rev Neurobiol 2002;49:185197.
6. The Brain Trauma Foundation, The American Association of Neurological Surgeons, the Joint Section on Neurotrauma and Critical
Care. Role of steroids. J Neurotrauma 2000;17:531535.
7. Hamer HM, Morris HH, Mascha EJ, et al. Complications of invasive
video-EEG monitoring with subdural grid electrodes. Neurology
2002;58:97103.
8. Onal C, Otsubo H, Araki T, et al. Complications of invasive
subdural grid monitoring in children with epilepsy. J Neurosurg
2003;98:10171026.
9. Morrison G, Duchowny M, Resnick T, et al. Epilepsy surgery
in childhood: a case report of 79 patients. Pediatr Neurosurg
1992;18:291297.
10. Pilcher WH, Rusyniak WG. Complication of epilepsy surgery. Neurol Clin North Am 1993;4:311325.
11. Lee W, Lee J, Lee S, et al. Complications and results of subdural grid electrode implantation in epilepsy surgery. Surg Neurol
2000;54:346351.
12. Silberbusch MA, Rothman MI, Bergey GK, et al. Subdural grid implantation for intracranial EEG recording: CT and MR appearance.
AJNR Am J Neuroradiol 1998;19:10891093.
13. Klatzo I, Seitelberger F. Brain Edema. Vienna: Springer Verlag,
1967.
14. Poungvarin N. Steroids have no value in stroke therapy. Stroke
2004;35:229230.
15. Kumar N, Jain S, Maheshwari MC. Role of dexamethasone in the
outcome from acute stroke. J Assoc Physicians India 1989;37:315
317.
16. Ogun SA, Odusote KA. Effectiveness of high dose dexamethasone
in the treatment of acute stroke. West Afr J Med 2001;20:16.
17. Yoshikawa H, Yamazaki S, Abe T, Oda Y. Liposteroid therapy
for refractory seizures in children. J Child Neurol 2000;15:702
704.
18. Sinclair DB. Prednisone therapy in pediatric epilepsy. Pediatr Neurol 2003;28:194198.