predicting seizure freedom after two or more chronic invasive evaluations in patients with...

W

PEER-REVIEW REPORTS

Predicting Seizure Freedom After Two or More Chronic Invasive Evaluations in Patientswith Intractable Epilepsy
Sean J. Nagel1, Lara E. Jehi2, Rebecca O’Dwyer4, Dani Bidros5, Girish K. Hiremath6, William E. Bingaman3
rnmvdp

lraglqa

tsyTismptoctsa

msrfw

�

INTRODUCTION

Epilepsy remains a major health problem indeveloped and developing countries. In theUnited States alone, 0.5% to 0.9% of thepopulation is diagnosed with epilepsy (5).Although the majority of patients withnewly diagnosed epilepsy will eventuallyachieve remission with antiepileptic drugs,nearly one in five patients will continue tohave chronic recurrent seizures despite op-timal treatment with antiepileptic drugs(6). A conservative estimate is that one-halfof patients with medically refractory sei-zures are potential candidates for epilepsysurgery. The results of epilepsy surgeryhave gradually improved in the past decade,approaching 60% to 90% seizure-free out-come in temporal lobe epilepsy and 45%to 66% in extratemporal lobe epilepsy (2, 11,30). Improvements in seizure-free out-

Key words� Epilepsy� Epilepsy surgery� Outcomes� Reoperation� Subdural grids

Abbreviations and AcronymsAPOS: Acute postoperative seizureEEG: ElectroencephalographyMRI: Magnetic resonance imagingSDE: Subdural electrodes

From the 1Department of Neuroscience,MetroHealth Medical Center, Cleveland, Ohio;

2Epilepsy Center, and 3Department of Neurosurgery,Neurological Institute, Cleveland Clinic, Cleveland, Ohio;4Department of Neurology, Georgetown University Hospital,

ashington, D.C.; 5Princeton Brain and Spine Care,Princeton, New Jersey; and 6Neurological Associates,Columbus, Ohio, USA

To whom correspondence should be addressed:Sean J. Nagel, M.D. [E-mail: [email protected]]

Citation: World Neurosurg. (2012) 77, 3/4:548-555.DOI: 10.1016/j.wneu.2011.06.037

Journal homepage: www.WORLDNEUROSURGERY.org

Available online: www.sciencedirect.com

1878-8750/$ - see front matter © 2012 Elsevier Inc.All rights reserved.

comes are likely related to advances in neu- (

548 www.SCIENCEDIRECT.com

oimaging along with the expanded use ofoninvasive video-electroencephalographyonitoring, and the willingness to use in-

asive monitoring techniques (i.e., sub-ural grids, strips, and depth electrodes) inoorly localized cases.

Failure to localize and/or resect the epi-eptic focus after an invasive monitoring pe-iod is multifactorial. Reasons include in-dequate electrode coverage, multifocal oreneralized epilepsy, and proximity of epi-

eptogenic zone to eloquent cortex. The fre-uency of reoperation reported in the liter-ture is variable, ranging from 3% to 14%

� OBJECTIVE: In patients with intracresect the epileptic focus after invasithese patients return for a second invaclearly characterized. This study aimssecond invasive electroencephalograpredictors.

� METHODS: We retrospectively reviore invasive evaluations at Clevelan

urgical, imaging, and pathology inforegression model. A favorable outcreedom, allowing for auras and seiz

eek.

RESULTS: Ten patients (33%) becam(mean follow-up duration 3.13 years [rawas more likely in patients reporting(mean 19 seizures/mo), and in thosescalp EEG ictal pattern (32% seizureseizure frequency (mean 70 seizures/mseizure free; P � 0.04). There waspostoperative seizures and failure ofpatients, at least one complication waevaluation compared with a complicevaluation.

� CONCLUSIONS: A second invasiveone-third of patients. However, thiscomplication rate with reoperation.

23). In rare instances, patients with uncon- i

WORLD NEUROSURGE

rolled seizures or seizure recurrence after aeizure-free period may return months toears later for a second invasive evaluation.he literature on this population of patients

s sparse. The aim of this report is to de-cribe the characteristics, risks, and deter-inants of seizure outcomes in a group of

atients who have had repeat subdural elec-rode evaluation. A previous report fromur institution reviewed the long-term out-ome of patients who underwent reopera-ion after failed epilepsy surgery (14). Ab-ent from this article was the outcomesnalysis of patients who had invasive mon-

epilepsy, failure to localize and/oronitoring is multifactorial. Rarely doevaluation, and their outcome is not

etermine the seizure outcome after a(EEG) evaluation, and its possible

d 30 patients who underwent two orinic between 1980 and 2007. Clinical,on was analyzed using a multivariate

was defined as complete seizurerestricted to the first postoperative

izure free after the second operation6 months–17 years]). Seizure freedomer seizures per month preoperativelya predominantly regional/lateralized), as opposed to those with higher; P � 0.02) or diffuse ictal EEGs (0%

gnificant association between acuteat surgery (P � 0.0083). In 17 of 30orted (57%) after the second invasiverate of 23% after the first invasive

uation may lead to seizure freedom int be weighed against the increased

tableve msiveto dphic

ewed Clmatiomeures

e sengefew

withfreeonth

a sirepe

s repation

evalmus

toring on two or more occasions. Should

RY, DOI:10.1016/j.wneu.2011.06.037

mailto:[email protected]

http://www.worldneurosurgery.org

http://www.sciencedirect.com

treiehctTtl

zarfostizr

rtMaasvna

fccvt

s

goamdeaf

balsttwrup(

rwrw

o(tcget

SDumffsttWtpats

PEER-REVIEW REPORTS

SEAN J. NAGEL ET AL. INVASIVE REEVALUATION FOR EPILEPSY

the trend toward early surgery continue inthe management of epilepsy, this analysiscould aid in the decision-making process inpatients whose seizures return after inva-sive monitoring. In an effort to improvepatient selection, we attempted to identifyfactors that are predictive of success afterredo-subdural monitoring in our series ofpatients at the Cleveland Clinic.

METHODS

We retrospectively reviewed 30 patients,identified using the Cleveland Clinic epi-lepsy patient database, who were invasivelymonitored on two or more occasions be-ween 1980 and 2007. Patients monito-ed with subdural electrodes, epidurallectrodes, or depth electrodes were

ncluded. Patients whose initial invasivevaluation was performed at an outsideospital were included. Patients were ex-luded from analysis if follow-up was lesshan 6 months after the second operation.he protocol was reviewed and approved by

he institutional review board at the Cleve-and Clinic.

Age of seizure onset, frequency of sei-ures preoperatively and postoperativelyfter each invasive surgery, time to recur-ence, and seizure freedom at most recentollow-up were collected from the chartsf each patient. For the purposes of thistudy, seizure frequency was recorded ashe total number of seizures per month,ncluding patients with two or more sei-ure types as documented in the medicalecord.

Neuroimaging, specifically the magneticesonance imaging (MRI) report was usedo identify patients with lesional (abnormal

RI) or nonlesional (normal MRI) epilepsyt baseline when this information was avail-ble. The most recent MRI report before theecond invasive evaluation was used to di-ide patients into groups with residual ab-ormal tissue and those with changes onlyfter surgery.

The etiology of the seizures was culledrom the neuropathology reports or fromlinic notes. The etiology was classified asryptogenic, malformation of cortical de-elopment, mesial temporal sclerosis, orumor.

The surgical data analyzed included the
ide of surgery, age at first and second sur- p
WORLD NEUROSURGERY 77 [3/4]: 548-5

ery, location of surgery (Cleveland Clinicr outside hospital), type of invasive evalu-tion, and the lobe resected. The invasiveonitoring techniques consisted of sub-

ural electrodes (strips or grids), epidurallectrodes, depth electrodes, or a combinedpproach. The lobe resected was defined asrontal, temporal, parietal, or multilobar.

The scalp electroencephalography (EEG)efore the second invasive monitoringnd the invasive EEG characteristics col-ected from the second evaluation weretudied. This included elements of bothhe interictal and ictal patterns. The ana-omic margins of the surgical resectionere compared with the imaging and neu-

ophysiologic data from the invasive eval-ation to determine whether they encom-assed the ictal and interictal activitycomplete or incomplete resection).

The preoperative and postoperative neu-ological examination after each surgeryas reviewed to identify postoperative neu-

ological deficits. New neurological deficitsere divided into expected (in context of

Table 1. Summary of Demographic and CSecond Invasive Monitoring and SubsequFreedom

Al(

Female (%)

Left hemisphere surgery (%)

Mean age seizure onset (years)

Range (years)

Mean age at first surgery (years)

Range (years)

Mean preoperative seizure frequency/month

Range

APOS (%) (n � 29)*

Yes

No

Seizure etiology

Cryptogenic

Malformation of cortical development

Mesial temporal sclerosis

Tumor

APOS, acute postoperative seizure.*APOS data was not available on one patient.

replanned resection of functional cortex to 5

55, MARCH/APRIL 2012 ww

ptimize seizure freedom) or unexpectedsuch as secondary to hemorrhages, infec-ions). Complications after surgery werelassified into the following categories: sur-ical site infection, hemorrhage, cerebraldema, sinus thrombosis, deep venoushrombosis, hydrocephalous, and death.

tatistical Analysisescriptive statistical methods were firstsed to summarize the data, reportingeans, medians, and standard deviations

or continuous variables and frequenciesor categorical variables. Univariate analy-is was initially performed to comparehose patients who were seizure free andhose who had seizure recurrence, using

ilcoxon rank sum, �2, and Fisher’s exactests. This analysis was used to identifyotential prognostic factors and those vari-bles with a significance of 10% were fur-her tested in a multivariate linear regres-ion model. Variables with a significance of

al Characteristics in Patients After theResection According to Seizure

ents30)

Seizure Free(n � 10)

Recurred(n � 20) P Value

3) 3 (33) 10 (50) 0.29

0) 4 (40) 11 (55) 0.44

13.5 7.3 0.11

0–37 0–22

27.3 27.4 0.98

.8 8.6–47.9 10.3–57.8

19 70 0.02

0 2–60 2–300

0.0083

1) 1 (8) 11 (92)

9) 9 (53) 8 (47)

0.14

8) 5 (36) 9 (64)

8) 3 (21) 11 (79)

) 1 (100) 0 (0)

) 1 (100) 0 (0)

linicent

l Patin �

13 (4

15 (5

9.9

0–37

27.4

8.6–57

53

2–30

12 (4

17 (5

14 (4

14 (4

1 (3

1 (3

% or less were considered significant. This

w.WORLDNEUROSURGERY.org 549

m5df

EMcmsto

MBatIe

bapassf

ONai(u1p

tt1tste2A(ttslmvlwf

IETfitcpp0ibb

ord fro

PEER-REVIEW REPORTS


method allowed the interaction of multiplevariables and their cumulative effect on theoutcome to be considered.

RESULTS

Summary of Clinical andDemographic DataTable 1 summarizes the clinical and demo-graphic data of the 30 patients included inthe analysis. Seventeen patients were men(57%) and 15 were left-handed (50%). Themean age at seizure onset was 9.9 (� stan-dard deviation [SD] 9.3) years and the meanat age at the first evaluation was 27.4 years(range 8.6 –57.8 years). The mean durationbetween first and second invasive evalua-tion was 6.1 years (� SD 5.2 years) and the

ean baseline seizure frequency was3.1 (� SD 76.8) seizures/mo. Follow-upata were available after the second surgeryor 3.13 years (range 6 months–17 years).

tiologyalformation of cortical development and

ryptogenic epilepsy were the most com-only reported etiologies in the patients

tudied (Table 1). Only one patient with aumor warranted invasive evaluation on twoccasions.

agnetic Resonance Imagingaseline brain MRI information was avail-ble in 23 patients (Table 2). Sixteen ofhese patients had normal imaging (69%).n patients presenting for a second invasive

Table 2. Summary of Magnetic Resonancthe Second Invasive Monitoring Accordin

All Patients(n � 30)

Baseline MRI (n � 23)

Normal 16 (69)

Abnormal 7 (30)

Preoperative MRI (n � 29)*

Postsurgical changes 22 (76)

Residual abnormalities 7 (24)

MRI, magnetic resonance imaging.*No imaging information was available in the medical rec

valuation, the preoperative MRI of the


rain showed residual abnormalities in 7nd expected changes after surgery in 22atients. Imaging information was unavail-ble in one patient before the second inva-ive evaluation. There was no statisticallyignificant correlation between seizurereedom and a baseline abnormal MRI.

perative Resultsineteen patients were initially operated on

t the Cleveland Clinic (Table 3). In the firstnvasive evaluation, subdural electrodesSDEs) were used in 24 patients, depth eval-ations in 4 patients, epidural electrodes inpatient, and a combined approach in 1

atient.

aging Findings at Baseline and Prior toSeizure Freedom



0.24

6 (38) 10 (63)

1 (14) 6 (86)

0.87

7 (32) 15 (68)

2 (29) 5 (71)

m one patient before the second invasive evaluation.

Table 3. Summary of Operative Details inEvaluation According to Seizure Freedom

Al(

Location of first surgery (%) (n � 30)

Cleveland Clinic

Other

Lobe resected first surgery (%) (n � 30)

None

Temporal

Frontal

Parietal

Multilobar

Lobe resected second surgery (%) (n � 30)

Temporal

Frontal

Multilobar 8 (27

WORLD NEUROSURGE

After the first invasive evaluation 6 pa-ients did not have a resection (20%), 8 pa-ients (27%) had a temporal lobe resection,1 patients (37%) had a frontal lobe resec-ion, 2 patients (7%) had a parietal lobe re-ection, and 3 patients (10%) had mul-ilobar resections. The second invasivevaluation consisted of SDE evaluation in3 patients and depth and SDE in 7 patients.fter the second evaluation, 9 patients

30%) had temporal lobe resections or ex-ension of a previous temporal lobe resec-ion, 13 patients (43%) had frontal lobe re-ections or extension of a previous frontalobe resection and 8 patients (27%) had

ultilobar resections or extension of a pre-ious multilobar resection. In 3 patients theobe resected during the second surgery

as remote from the site of the resectionrom the first surgery.

nvasive and Noninvasivelectroencephalographyable 4 summarizes the noninvasive EEGndings before the second invasive moni-

oring. Regionalized, interictal EEG did notorrelate with seizure freedom (P � 0.46). Aredominant ictal regional and lateralizedattern predicted seizure freedom (P �.04). The EEG findings from the second

nvasive monitoring are summarized in Ta-le 5. There was a significant associationetween focal interictal activity (P � 0.04)

ents Requiring a Second Invasive

ents0)



0.06

) 5 (26) 14 (74)

) 6 (55) 5 (45)

0.58

) 2 (33) 4 (67)

) 4 (50) 4 (50)

) 3 (27) 8 (73)

0 (0) 2 (100)

) 1 (33) 2 (67)

0.81

) 2 (22) 7 (78)

) 5 (38) 8 (62)

e Img to

Pati

l Patin � 3

19 (63

11 (37

6 (20

8 (27

11 (37

2 (7)

3 (10

9 (30

13 (43

) 3 (38) 5 (63)

RY, DOI:10.1016/j.wneu.2011.06.037

ctpmftsl

(cz0pat9

CNC(TciwttwoabcnhtOtr

rpwid

D

Wwsskrmiltetofopssud1ftpgp

PEER-REVIEW REPORTS


alone and seizure freedom. When interictalspiking and ictal onset were positively cor-related, there was a significant increase inseizure recurrence (P � 0.0005). No signif-icant association was identified between fo-cal ictal activity and seizure freedom (P �1.0). Complete resection of the ictal onsetzone did not significantly improve seizurecontrol (P � 0.31).

OutcomesTen patients (33%) became seizure free af-ter the second operation at the time of mostrecent follow-up. In 20 patients (67%) withpersistent seizures, there was no improve-ment in overall seizure frequency (mean70.2 � 19.9 seizures/mo preoperatively vs.67.1 � 40.6 seizures/mo postoperatively;P � 0.99). In these patients, seizures re-

urred within 1 year of surgery. Twelve pa-ients (60%) had seizures immediatelyostoperatively, 5 patients (25%) failed by 6onths, and the other 3 patients (15%) be-

ore 1 year postoperatively. In patients withemporal lobe epilepsy, 4/7 patients wereeizure free. In patients with extratemporalobe epilepsy, 6/23 had a favorable out-

come (P � 0.13). Patients reporting fewerseizures per month preoperatively weremore likely to be seizure free after a sec-

Table 4. Summary of Noninvasive ElectroSecond Invasive Monitoring According to

All(

Interictal EEG (%) (n � 24)

Regional

Multiregional

Ictal EEG Regional/lateralized (%) (n � 24)

Yes

No

Ictal EEG Nonlocalizable (%) (n � 24)

Yes

No

Predominant ictal pattern (%) (n � 24)

Regional/lateralized

Nonlocalizable

EEG, electroencephalography.

ond invasive evaluation and resection c


P � 0.02). There was a significant asso-iation between acute postoperative sei-ures and failure of repeat surgery (P �.0083). This was the only independentredictor of outcome using a multivari-ble regression model that remained sta-istically significant (odds ratio � 12.4,5% confidence interval 1.8 –252.7).

omplications and Neweurological Deficitsomplications were reported in 7 patients

23%) after the first invasive evaluation.wo of the patients had more than oneomplication. Brain edema was describedn three patients, postoperative infections

ere identified in two patients, deep veinhrombosis was identified in two pa-ients, and mild aphasia and hematomaere documented in one patient. Theverall complication rate was increasedfter the second invasive evaluation (Ta-le 6). There was a postoperative compli-ation in 17 patients (57%), includingew neurological deficits. One patientad more than one complication. Infec-

ion was reported in four patients (13%).ne infection was superficial. Two pa-

ients developed meningitis. One patientequired bone flap removal and delayed

phalographic Findings Before theure Freedom

nts0)



0.46

) 5 (45) 6 (55)

) 4 (30) 9 (69)

0.15

) 8 (35) 15 (65)

0 (0) 1 (100)

0.74

) 3 (33) 6 (67)

) 6 (40) 9 (60)

0.04

) 7 (32) 15 (68)

0 (0) 2 (100)

ranioplasty. One patient died. New neu-


ological deficits were expected in eightatients. Of those eight patients, threeent on to become seizure free. These

ncluded new hemiparesis, hemisensoryeficit, aphasia, or visual field cut.

ISCUSSION

e evaluated our experience in patientsith intractable epilepsy who were inva-

ively monitored on two or more occa-ions during a 27-year period. To ournowledge, this is the largest series toeview outcomes in patients who have had

ore than one intracranial monitoringnvestigation. Reoperation for failed epi-epsy surgery in general, has concludedhat half of the patients undergoing reop-ration will benefit (3, 13-15, 24, 33). Ofhe 30 patients included in our study, onlyne-third were seizure free after a meanollow-up of about 3 years. Our lower ratef seizure freedom reflects the diffuse oroorly localized epilepsy in patients con-idered for invasive monitoring. In fact,everal studies have already identified these of SDE as a poor seizure outcome in-icator after resective epilepsy surgery (1,6). Of note, the 30% chance of seizurereedom found in our study correspondso patients who have already failed onerevious invasive EEG evaluation. Theeneral favorable seizure outcome ex-ected after any reoperation (with first

surgery guided by invasive or noninvasivedata) is higher (25). As such, invasive EEGas a reoperation technique might improvethe localization of an epileptogenic focusthat was solely defined based on noninva-sive methods before the first surgery, butis not expected to provide much addi-tional useful information if already per-formed with the initial resection.

Yet, although only one-third of our pa-tients were seizure free after a second inva-sive evaluation, given their poor seizureprognosis, such an outcome may still justifythe continued use of repeat surgery in selectcases of intractable epilepsy.

In the first reported series that studiedoutcomes after a second invasive evalua-tion published in 2000, Siegel et al. (28)analyzed nine patients after failed local-ization because of presumed sampling er-ror. In seven of these patients, a seizurefocus was localized and was eventually

enceSeiz

Patien � 3

11 (46

13 (54

23 (96

1 (4)

9 (38

15 (63

22 (92

2 (8)

resected in six. Five of these patients were


uslpeotep

pfifiMAc(ssAtt(wgiflp

gp(darnlttiro

artpbusIzwrftide

fwmpcw

PEER-REVIEW REPORTS


seizure free after the second operation.The sixth patient had a third intracranialevaluation and eventually was seizure freeafter a limited extension of the resection(28). They concluded that a second reeval-

ation should be considered when aampling error is the reason for failedocalization. Interestingly, although thereoperative data favored temporal lobepilepsy, in six of seven patients, seizurenset originated in an extratemporal loca-

ion, suggesting the usefulness of a repeatvaluation when evaluating patients withseudotemporal lobe epilepsy (28).

Risk Factors for FailureThe most significant risk factor that pre-dicted seizure recurrence after the second

Table 5. Summary of the Ictal and IntericInvasive Monitoring According to Seizure

All(n

Interictal EEG (n � 24)

Focal (%)

Yes

No

Ictal EEG (n � 24)

Focal (%)

Yes

No

Ictal EEG features (%) (n � 24)

Repetitive spiking

Paroxysmal fast

Diffuse suppression

Correlation of ictal onset and interictalspiking (%) (n � 24)

Yes

No

Resection of interictal spiking region (%)(n � 21)

Complete

Incomplete

Resection of ictal onset zone (%) (n � 19)

Complete

Incomplete

EEG, electroencephalography.

invasive monitoring session was an acute b


ostoperative seizure (APOS). This is de-ned as a seizure occurring during therst 7 to 10 days after surgical resection.ani et al. (21) recently reported that

POS predicts poor long-term seizureontrol. In that report, 26% of children�18 years) who underwent either hemi-pherectomy or extratemporal resectionuffered from APOS. The incidence ofPOS was higher in patients that had ex-

ratemporal resections compared withhose who underwent hemispherectomy21). Patients who suffered from APOSere 87% less likely to have achieved En-el class I outcome at the 12- to 24-month

nterval after surgery (21). Malla et al. (20)ound that patients with APOS are lessikely to be seizure free after anterior tem-oral lobectomy when the APOS resem-

indings Recorded from the Seconddom

nts0)



0.04

) 9 (45) 11 (55)

) 0 (0) 4 (100)

1.00

) 6 (38) 10 (63)

) 3 (38) 5 (63)

0.54

) 2 (40) 3 (60)

) 3 (30) 7 (70)

) 4 (45) 5 (55)

0.0005

) 4 (21) 15 (79)

) 5 (100) 0 (0)

0.70

) 3 (38) 5 (63)

) 6 (65) 7 (54)

0.31

) 4 (33) 8 (67)

) 4 (57) 3 (43)

les the patient’s habitual seizures. Other d

WORLD NEUROSURGE

roups have reported similar findings inatients undergoing hemispherectomy19). However, in a small study (31), APOSid not warn of future seizures in patientsfter surgery for frontal lobe epilepsy. In aecent study (17), APOS were a poor prog-ostic marker in patients with frontal

obe epilepsy. A possible explanation ishat APOS reflects residual active epilep-ogenic tissue, and their occurrence in themmediate postoperative period is a sur-ogate marker of an incomplete resectionf the currently active epileptogenic zone.

To reduce the need for reoperation aftern invasive evaluation, Bauman et al. (4)ecently described a novel approach to mul-istage epilepsy surgery in a select group ofediatric patients suffering from intracta-le epilepsy. They evaluated 15 patients whonderwent two intracranial monitoringessions during the same hospitalization.n this highly select group of patients, sei-ure localization by intracranial monitoringas inconclusive. Therefore, the grids were

eplaced after an initial resection was per-ormed to monitor for residual seizure ac-ivity, improve localization, address the ep-leptogenic potential of multiple foci, orefine the proximity of a seizure focus toloquent cortex (4).

This multistage surgery can be safely per-ormed with a complication rate that is on parith the traditional two-stage technique. Theultistage approach may be of benefit in our

atients with APOS, who then went on to re-ur. Future clinical studies that include adultsill validate the utility of this technique.High preoperative seizure frequency pre-

Table 6. Summary of PostoperativeComplications and NeurologicalDeficits After a Second InvasiveMonitoring and Surgical Resection

Complications (%) Number of Patients

Infection 5 (17)

DVT 2 (7)

Sinus thrombosis 1 (3)

Hydrocephalous 1 (3)

Death 1 (3)

Neurological deficit 8 (27)

DVT, deep venous thrombosis.

tal FFree

Patie� 3

20 (83

4 (17

16 (67

8 (33

5 (21

10 (42

9 (38

19 (79

5 (21

8 (38

13 (62

12 (63

7 (37

icted failure in this study. Foldvary et al.

RY, DOI:10.1016/j.wneu.2011.06.037

Mccstimft

PEER-REVIEW REPORTS


(12) also found that a high seizure burdenpreoperatively is related to an increased rateof seizure recurrence postoperatively in pa-tients undergoing temporal lobectomy. In astudy (17) analyzing 371 temporal lobecto-mies performed at the Cleveland Clinic,similar results were found.

Correlation With Preoperative Imagingand Disease EtiologyIn 23 of 30 patients, the baseline MRI reportwas available. Sixteen patients (69%) had anormal MRI, although this did not correlatewith postoperative seizure freedom. Therehave been numerous reports that have com-pared seizure outcomes based on the pre-operative MRI findings (7, 9, 10, 22, 26, 27,29). In the 1990s, Berkovic et al. (7) reportedon the recurrence of seizures in 135 patientsafter temporal lobectomy. They found that,although patients with lesional MRI find-ings had a postoperative seizure freedomrate in the 50% to 60% range, only 21% ofpatients with normal MRIs were seizurefree at the end of a 5-year follow-up period.Other reports corroborated these find-ings (26).

Other studies have reported that pa-tients with a normal MRI had a similarrate of postoperative seizure freedomwhen compared with patients who havelesional MRIs. Siegel et al. evaluated 43patients with normal MRIs who under-went intracranial monitoring. Focal sei-zure onset was detected in 58% of thesepatients. In those patients who subse-quently underwent resection, 83% had asatisfactory postoperative seizure-freeoutcome (27). In our previously pub-lished series of 24 patients with a normalMRI, 37% were seizure free after surgicalresection of the seizure focus (8). Thislower rate of seizure freedom may be re-lated to cortical dysplasia, which is oftennot evident on imaging. Therefore, sei-zure freedom was similar to other serieswhere a sizeable proportion of the groupconsisted of patients harboring corticaldysplasia (18, 20). In our study, 47% ofthe patients had cortical dysplasia. This isconsistent with a previous study from ourinstitution, where, 42% were found tohave cortical dysplasia (8). In patientswith a lesional MRI, the etiology of sei-zures, often neoplasia or hippocampalsclerosis, is favorable. In contrast, corti-
cal dysplasia, frequently undetectable on

RI, is less amenable to complete surgi-al resection. Removing the patients withortical dysplasia as the pathologic sub-trate causing intractable epilepsy may,herefore, improve seizure-free outcomesn this study, as 11/14 of our patients with

alformation of cortical developmentailed reoperation as opposed to 9/16 ofhose with other etiologies (P � 0.09).

Therefore, in patients without a clear eti-ology for seizures, a normal MRI evalua-tion, suspicion for cortical dysplasia, andconcordance of other tests, intracranialmonitoring should be considered.

Electroencephalographic Findings asPredictors of OutcomeFocal interictal EEG activity independentlycorrelated with postoperative seizure free-dom (P � 0.04). Surprisingly, when interic-tal and ictal onset were positively corre-lated, seizure recurrence was significantlyincreased (P � 0.0005). There was also nosignificant association identified betweenfocal ictal activity and seizure freedom (P �1.0). Complete resection of the ictal onsetzone also did not significantly improve sei-zure control (P � 0.31). Explaining thesecounterintuitive findings is challenging,but may be possible if we remember that thepatient cohort under study here is a veryspecific one: a group that already failed aresection guided by a previous invasive EEGrather than “naïve” intractable epilepsy pa-tients presenting for their first resection.Seizure recurrence after the first invasiveEEG may have been due to an incompletecoverage and delineation of the epileptic fo-cus—in which case a repeat evaluation withbetter coverage might improve the out-come— but may also reflect a more diffuseepileptic pathology (e.g., microscopic dys-plasia) or a true underlying epileptogenicityin a remote functionally connected elementof the network in question. As such, evidenceof an adequate “ictal onset zone” definitionmay not necessarily have the same expectedstrong correlation with “epileptogenic zone”definition. For example, the positive correla-tion between interictal and ictal onset zones,as defined by our repeat evaluation, may actu-ally reflect a consistent spread pattern from aremote uninvestigated and as such missed ep-ileptogenic focus. Alternatively, these EEGfindings may not be a valid independent out-come predictor, as they were not confirmed by
the regression analysis. The relatively small

number of available EEGs for review limitedthe number of useful conclusions from theEEG data.

ComplicationsThe number of complications increasedafter the second evaluation. In approxi-mately 57% (17/30) of patients, a compli-cation was reported. A new neurologicaldeficit was the most common complica-tion (27%). A neurological deficit seemedacceptable to many patients in exchangefor lasting seizure freedom. In fact, in theeight patients with a postoperative neuro-logical deficit, it was expected. However,these patients overall faired only slightlybetter. Other series of intracranial moni-toring, which only include one invasivestudy, reported lower complication rates,approximately 6% to 10% in recent series(32). In the series described by Bauman etal. (4), a very low rate of complications—which only included those patients need-ing blood transfusions (13%)—wasnoted. They did not report any permanentneurological deficits, wound infection,cerebral hemorrhage, or edema. Infec-tions occurred in 13% of patients who hadtwo or more invasive evaluations. Since1995, we have modified our surgical tech-nique. We currently tunnel all wires andmaintain patients on antibiotics duringtheir evaluation. A single grid is sent tomicrobiology after removal for culture.

The one death reported in our series oc-curred in a 59-year-old man. In 1992, he hada standard right temporal lobectomy. Fouryears later his resection was extended to in-clude Heschl’s area. He continued to haveone to five generalized seizures per month.In 2003, four subdural grids were im-planted over the lateral frontal, parietal,temporal, and basal occipital cortex. Inva-sive monitoring indicated that the seizureswere originating from an extensive malfor-mation of cortical development located atthe temporoparietal junction. This area wasresected without complication and he wasdischarged 5 days later on phenytoin, leve-tiracetam, and oxcarbazepine. One monthlater the patient died after a generalized sei-zure. The patient’s family reported that hehad fewer auras since surgery and the dura-tion of his generalized seizures had de-creased.

It is important to counsel patients and
their families about the high complication

1

1

1

1

1

1

1

1

1

2

2

2

2

2

2

2

2

2

2

3

PEER-REVIEW REPORTS


rate with a repeat invasive evaluation. Fur-thermore, before invasive monitoring, theyshould understand that lasting seizure free-dom might only be attainable after the re-moval of eloquent cortex. Oftentimes, theextent of resection is limited by the proxim-ity to functional cortex.

Study LimitationsThere are several recognized limitations inthis study. All retrospective studies arehampered by the quality of the medical re-cords. Furthermore, we included data fromoutside institutions. Taken together, thisprecluded the useful analysis of select riskfactors and confounded the interpretationof others. Although this is the largest studyto date, repeating an invasive epilepsy eval-uation is rare. Consequently, the total num-ber of patients included in our analysis issmall (N � 30), requiring caution when in-terpreting these results. Furthermore, weelected to include patients with follow-up asshort as 6 months because we had so fewpatients to evaluate.

CONCLUSIONS

In select cases, patients who fail a singlechronic invasive monitoring investigationshould be referred for additional invasivestudies. In many cases, new localizing in-formation may be uncovered. The increasedrisk inherent in reoperative neurosurgerytogether with the relatively low success rateshould be carefully considered in these pa-tients. Not infrequently, resection of an el-oquent area of the brain is needed to preventseizure recurrence.

REFERENCES

1. Armon C, Radtke RA, Friedman AH, Dawson DV:Predictors of outcome of epilepsy surgery: multivar-iate analysis with validation. Epilepsia 37:814-821,1996.

2. Arruda F, Cendes F, Andermann F, Dubeau F, Vill-emure JG, Jones-Gotman M, Poulin N, Arnold DL,Olivier A: Mesial atrophy and outcome afteramygdalohippocampectomy or temporal lobe re-moval. Ann Neurol 40:446-450, 1996.

3. Awad IA, Nayel MH, Lüders H: Second operation
after the failure of previous resection for epilepsy.Neurosurgery 28:510-518, 1991.
1


4. Bauman JA, Feoli E, Romanelli P, Doyle WK, Devin-sky O, Weiner HL: Multistage epilepsy surgery:safety, efficacy, and utility of a novel approach inpediatric extratemporal epilepsy. Neurosurgery 62Suppl 2:489-505, 2008.

5. Begley CE, Famulari M, Annegers JF: The cost ofepilepsy in the United States: an estimate from pop-ulation-based clinical and survey data. Epilepsia 41:342-351, 2000.

6. Berg AT: Understanding the delay before epilepsysurgery: who develops intractable focal epilepsy andwhen? CNS Spectr 9:136-144, 2004.

7. Berkovic SF, McIntosh AM, Kalnins RM, JacksonGD, Fabinyi GC, Brazenor GA, Bladin PF, HopperJL: Preoperative MRI predicts outcome of temporallobectomy: an actuarial analysis. Neurology 45:1358-1363, 1995.

8. Chapman K, Wyllie E, Najm I, Ruggieri P,Bingaman W, Lüders J, Kotagal P, Lachhwani D,Dinner D, Lüders HO: Seizure outcome after epi-lepsy surgery in patients with normal preoperativeMRI. J Neurol Neurosurg Psychiatry 76:710-713,2005.

9. Cukiert A, Buratini JA, Machado E, Sousa A, VieiraJO, Argentoni M, Forster C, Baldauf C: Results ofsurgery in patients with refractory extratemporal ep-ilepsy with normal or nonlocalizing magnetic reso-nance findings investigated with subdural grids.Epilepsia 42:889-894, 2001.

0. Edwards JC, Wyllie E, Ruggeri PM, Bingaman W,Lüders H, Kotagal P, Dinner DS, Morris HH, Pray-son RA, Comair YG: Seizure outcome after surgeryfor epilepsy due to malformation of cortical develop-ment. Neurology 55:111-104, 2000.

1. Engel J Jr, Van Ness PC, Rasmussen TB, Ojemann L:Outcome with respect to epileptic seizures. In: EngelJ Jr, ed. Surgical Treatment of the Epilepsies, ed 2.New York: Raven Press, 1993:609-621.

2. Foldvary N, Nashold B, Mascha E, Thompson EA,Lee N, McNamara JO, Lewis DV, Luther JS, Fried-man AH, Radtke RA: Seizure outcome after tempo-ral lobectomy for temporal lobe epilepsy: a Kaplan-Meier survival analysis. Neurology 854:630-634,2000.

3. Germano IM, Poulin N, Olivier A: Reoperation forrecurrent temporal lobe epilepsy. J Neurosurg 81:31-36, 1994.

4. Gonzalez-Martinez JA, Srikijvilaikul T, Nair D,Bingaman WE: Long-term seizure outcome in reop-eration after failure of epilepsy surgery. Neurosur-gery 60:873-880, 2007.

5. Holmes MD, Wilensky AJ, Ojemann LM, OjemannGA: Predicting outcome following reoperation formedically intractable epilepsy. Seizure 8:103-106,1999.

6. Janszky J, Jokeit H, Schulz R, Hoppe M, Ebner A:EEG predicts surgical outcome in lesional frontallobe epilepsy. Neurology 54:1470-1476, 2000.

7. Jeha LE, Najm I, Bingaman W, Dinner D, Widdess-Walsh P, Lüders H: Surgical outcome and prognos-

WORLD NEUROSURGE

tic factors of frontal lobe epilepsy surgery. Brain130(Pt 2):574-584, 2007.

8. Kloss S, Pieper T, Pannek H, Holthausen H, Tux-horn I: Epilepsy surgery in children with focal corti-cal dysplasia (FCD): results of long-term seizureoutcome. Neuropediatrics 33:21-26, 2002.

9. Koh S, Nguyen S, Asarnow RF, LoPresti C, YudovinS, Shields WD, Vinters HV, Mathern GW: Five ormore acute postoperative seizures predict hospitalcourse and long-term seizure control after hemi-spherectomy. Epilepsia 45:527-533, 2004.

0. Malla BR, O’Brien TJ, Cascino GD, So EL, Rad-hakrishnan K, Silbert P, Marsh WR: Acute postop-erative seizures following anterior temporal lobec-tomy for intractable partial epilepsy. J Neurosurg89:177-182, 1998.

1. Mani J, Gupta A, Mascha E, Lachhwani D, PrakashK, Bingaman W, Wyllie E: Postoperative seizuresafter extratemporal resections and hemispherec-tomy in pediatric epilepsy. Neurology 66:1038-1043,2006.

2. Paolicchi JM, Jayakar P, Dean P, Yaylali I, MorrisonG, Prats A, Resnick T, Alvarez L, Duchowny M: Pre-dictors of outcome in pediatric epilepsy surgery.Neurology 54:642-647, 2000.

3. Polkey CE, Awad IA, Tanaka T, Wyler AR: The placeof reoperation. In: Engel J Jr, ed. Surgical Treatmentof the Epilepsies, ed 2. New York: Raven Press, 1993:663-667.

4. Salanova V, Quesney LF, Rasmussen T, AndermannF, Olivier A: Reevaluation of surgical failures and therole of reoperation in 39 patients with frontal lobeepilepsy. Epilepsia 35:70-80, 1994.

5. Schwartz TH, Spencer DD: Strategies for reopera-tion after comprehensive epilepsy surgery. J Neuro-surg 95:615-623, 2001.

6. Scott CA, Fish DR, Smith SJ, Free SL, Stevens JM,Thompson PJ, Duncan JS, Shorvon SD, HarknessWF: Presurgical evaluation of patients with epilepsyand normal MRI: role of scalp video-EEG telemetry.J Neurol Neurosurg Psychiatry 66:69-71, 1999.

7. Siegel AM, Jobst BC, Thadani VM, Rhodes CH,Lewis PJ, Roberts DW, Williamson PD: Medicallyintractable, localization-related epilepsy withnormal MRI: presurgical evaluation and surgicaloutcome in 43 patients. Epilepsia 42:883-888,2001.

8. Siegel AM, Roberts DW, Thadani VM, McInerney J,Jobst BC, Williamson PD: The role of intracranialelectrode reevaluation in epilepsy patients afterfailed initial invasive monitoring. Epilepsia 41:571-580, 2000.

9. Smith JR, Lee MR, King DW, Murro AM, Park YD,Lee GP, Loring DW, Meador KJ, Harp R: Resultsof lesional vs. nonlesional frontal lobe epilepsysurgery. Stereotact Funct Neurosurg 69:202-209,1997.

0. Spencer SS, Berg AT, Vickrey BG, Sperling MR, Ba-
zil CW, Shinnar S, Langfitt JT, Walczak TS, Pacia SV:Predicting long-term seizure outcome after resec-
RY, DOI:10.1016/j.wneu.2011.06.037

3

3

3

C

cc

r

CD

J

A

1

PEER-REVIEW REPORTS


tive epilepsy surgery: the multicenter study. Neu-rology 65:912-918, 2005.

1. Tigaran S, Cascino GD, McClelland RL, So EL,Marsh WR: Acute postoperative seizures after fron-tal lobe cortical resection for intractable partial epi-lepsy. Epilepsia 44:831-835, 2003.

2. Van Gompel JJ, Worrell GA, Bell ML, Patrick TA,
Cascino GD, Raffel C, Marsh WR, Meyer FB: Intra-cranial electroencephalography with subdural grid a
Leonardo Rangel-Castilla, Sean Barber, Y

few reports that evaluate the role of ETV in

smns73tarH

wpopnp

�


electrodes: techniques, complications, and out-comes. Neurosurgery 63:498-505, 2008.

3. Wyler AR, Hermann BP, Richey ET: Results of reop-eration for failed epilepsy surgery. J Neurosurg 71:815-819, 1989.

onflict of interest statement: The authors declare that the
rticle content was composed in the absence of any A
i Jonathan Zhang

CP.

M

Ttsrsas1


ommercial or financial relationships that could beonstrued as a potential conflict of interest.

eceived 20 October 2010; accepted 23 June 2011

itation: World Neurosurg. (2012) 77, 3/4:548-555.OI: 10.1016/j.wneu.2011.06.037

ournal homepage: www.WORLDNEUROSURGERY.org

vailable online: www.sciencedirect.com

878-8750/$ - see front matter © 2012 Elsevier Inc.
ll rights reserved.
The Role of Endoscopic Third Ventriculostomy in the Treatment ofCommunicating Hydrocephalus

INTRODUCTION

Endoscopic third ventriculostomy (ETV)has become the treatment of choice in themanagement of obstructive hydrocepha-lus (HCP) due to its low risk and highsuccess rates (70%– 80%) (13, 14, 16, 21,23–26, 31, 36, 38). Endoscopic third ven-triculostomy bypasses the site of obstruc-tion in obstructive HCP and restores nor-mal physiologic cerebrospinal fluid (CSF)circulation. The role of ETV in communi-cating HCP is still unknown. There are

Key words� Communicating hydrocephalus� Endoscopic third ventriculostomy� Endoscopy

Abbreviations and AcronymsCSF: Cerebrospinal fluidETV: Endoscopic third ventriculostomyFIESTA: Fast imaging employing steady stateacquisitionHCP: HydrocephalusINPH: Idiopathic normal pressure hydrocephalusMRI: Magnetic resonance imagingNPH: Normal pressure hydrocephalusVP: Ventriculoperitoneal

Department of Neurosurgery, The MethodistNeurological Institute, Houston, Texas, USA

To whom correspondence should be addressed:Leonardo Rangel-Castilla, M.D.[E-mail: [email protected]]

Citation: World Neurosurg. (2012) 77, 3/4:555-560.DOI: 10.1016/j.wneu.2011.06.038

Journal homepage: www.WORLDNEUROSURGERY.org

Available online: www.sciencedirect.com

1878-8750/$ - see front matter © 2012 Elsevier Inc.All rights reserved.

econdary communicating HCP (post-eningitis and posthemorrhagic) and

ormal pressure HCP (NPH), reporting auccessful rate of 50% to 70% and 45% to2%, respectively (1, 5, 7, 11, 18, 20, 21,2). In the present study, we retrospec-ively reviewed the indications for ETVnd outcome after the procedure in a se-ies of 36 patients with communicating

� OBJECTIVE: To elucidate the role ofpatients with secondary and idiopathi

� METHODS: A series of 36 patientsomen) were treated by ETV betwee

atients age ranged from 19 to 81 yearf 6 to 36 months (mean 9.2 months). Thatients with secondary communicatormal pressure HCP. Sixteen (44.4%)eritoneal shunt placement that prese

RESULTS: The etiology of secondameningitis, trauma, neoplasm, and othein some patients. The outcome of Epatients (75%). A Kaplan-Meier analysof ETVs in secondary communicating H0.83, 0.8, and 0.77, respectively; in thewas 0.83 initially and became stablesuccessful proportion of ETV in commuof follow-up was 0.97, 0.83, 0.78, and

� CONCLUSIONS: ETV is a good ocommunicating HCP, normal pressuretriculoperitoneal shunts. The indicatcommunicating HCP needs further stu

ETHODS

hirty-six patients underwent ETV for thereatment of communicating HCP at our in-titution between November 2007 and Feb-uary 2010. The patient population con-isted of 15 women and 21 men with a meange of 52 years (range 19 – 81 years). Pre-enting symptoms included headaches in

oscopic third ventriculostomy (ETV) inmmunicating hydrocephalus (HCP).

communicating HCP (21 men and 15vember 2007 and February 2010. The(mean 52 years), and had a follow-uptients were divided into a group of 29

HCP and a group of 7 patients withe patients had a previous ventriculo-with shunt malfunction.

CP was subarachnoid hemorrhage,tiology was not possible to determine

was considered successful in 27/36vealed that the successful proportiont 0.5, 1, and 3 months of follow-up waspathic normal pressure HCP group it.66 after the first month. Overall, theting HCP was at 0, 0.5, 1, and 3 months

in the management of secondary, and replacing malfunctioning ven-of ETV as a first-line treatment inowever, results are promising.

endc co

withn Nos olde pa

ingof thnted

ry Hrs. ETVis reCP aidioat 0nica0.75.

ptionHCP

ionsdy; h

6 patients, gait or balance difficulty in 14




mailto:[email protected]



predicting seizure freedom after two or more chronic invasive evaluations in patients with...

Documents