Predictive factors for a good prognosis following surgery for temporal lobe epilepsy: a cohort study in Spain.

Abstract

To investigate the outcome of temporal lobe epilepsy surgery and identify the variables which predict a good prognosis with respect to seizures in postoperative follow-up after two and four years. This retrospective study included 115 selected patients who underwent surgery for temporal lobe epilepsy between 1996 and 2007. In the second year after surgery 86.1% of patients had a good prognosis for seizure control (73.9% Engel class I and 12.2% Engel class II) and 89.2% (76.3% Engel class I and 12.9% Engel class II) in the fourth year. Sixty-four of 93 (68.8%) patients were free of disabling seizures (Engel class I) during the entire period and 78 (83.8%) had good prognosis (Engel class I and II). For the second year, logistic regression analysis revealed the following variables to be independently predictive of good seizure control: absence of two or more seizure episodes in the first year after surgery, normal postoperative video-EEG, and age at surgery of less than 35 years. In the fourth year, mesial temporal sclerosis, female sex and normal postoperative video-EEG were the predictive factors. For the group with a good prognosis in both the second and the fourth year, the predictive variables were: absence of two or more seizure episodes in the first year after surgery (OR: 13.762, CI 95%: 2.566-73.808, p<0.002) and normal postoperative video-EEG (OR: 16.301, CI 95%: 3.704-71.740, p<0.001). This study illustrates the sustained benefit of temporal lobe epilepsy surgery. The multivariate logistic regression analysis failed to identify a good predictive model composed of preoperative variables alone, although it was possible to build such a model with either pre- and postoperative variables or only postoperative variables.

Full text

Page 1

doi:10.1684/epd.2011.0413
36
Epileptic Disord, Vol. 13, No. 1, March 2011
Correspondence:
M. I. Forcadas Berdusán
Department of Neurology,
Hospital de Cruces,
Plaza de Cruces s/n. ES,
48903 Baracaldo, Spain
<MISABEL.FORCADASBERDUSAN@
osakidetza.net>
Original article
Epileptic Disord 2011; 13 (1): 36-46
Predictive factors for a good
prognosis following surgery
for temporal lobe epilepsy:
a cohort study in Spain
María Isabel Forcadas-Berdusán1,2,
José Luis Bustos-Sánchez1,2, Elena Valle-Quevedo1,4,
Jon Aurrecoechea Obieta1,5, Beatriz Mateos Go˜ ni1,6,
Lorea Martinez-Indart3, Ana Molano Salazar1,7,
Juan Carlos Gomez-Esteban2, I˜ nigo Garamendi-Ruíz1,2
1Medical-Surgical Epilepsy Unit
2Department of Neurology
3Clinical Epidemiology Unit
4Neurophysiology Service
5Department of Neurosurgery
6Department of Neuroradiology
7Department of Neuropsychology, Hospital de Cruces, Baracaldo, Spain
Received July 15, 2010; Accepted November 29, 2010
ABSTRACT – Objective. To investigate the outcome of temporal lobe epilepsy
surgery and identify the variables which predict a good prognosis with respect
to seizures in postoperative follow-up after two and four years. Methods. This
retrospective study included 115 selected patients who underwent surgery for
temporallobeepilepsybetween1996and2007.Results. Inthesecondyearafter
surgery 86.1% of patients had a good prognosis for seizure control (73.9% Engel
class I and 12.2% Engel class II) and 89.2% (76.3% Engel class I and 12.9% Engel
class II) in the fourth year. Sixty-four of 93 (68.8%) patients were free of dis-
abling seizures (Engel class I) during the entire period and 78 (83.8%) had good
prognosis (Engel class I and II). For the second year, logistic regression analysis
revealedthefollowingvariablestobeindependentlypredictiveofgoodseizure
control: absence of two or more seizure episodes in the first year after surgery,
normal postoperative video-EEG, and age at surgery of less than 35 years. In
the fourth year, mesial temporal sclerosis, female sex and normal postoperative
video-EEG were the predictive factors. For the group with a good prognosis in
both the second and the fourth year, the predictive variables were: absence
of two or more seizure episodes in the first year after surgery (OR: 13.762, CI
95%: 2.566-73.808, p<0.002) and normal postoperative video-EEG (OR: 16.301,
CI 95%: 3.704-71.740, p<0.001). Discussion. This study illustrates the sustained
benefit of temporal lobe epilepsy surgery. The multivariate logistic regression
analysis failed to identify a good predictive model composed of preoperative
variables alone, although it was possible to build such a model with either pre-
and postoperative variables or only postoperative variables.
Key words: seizures, mesial temporal sclerosis, epilepsy surgery, prognosis,
predictors, temporal lobe

Page 2

Epileptic Disord, Vol. 13, No. 1, March 2011
37
Predictive factors for TLE surgical outcome
The incidence of epilepsy in industrialised countries
ranges from 24 to 53 cases per 100,000 people per
year, while developing countries report 77 to 114 cases
per 100,000 people per year (Hauser, 1997). Despite an
optimal therapeutic regimen with antiepileptic drugs
(AEDs), between 30 and 40% of epileptic patients
will be seizure-free for no more than five years;
thesepatientshaveso-called“drug-resistant”epilepsy
(Aicardi and Shorvon, 1997; Hauser and Hesdorffer,
2001).Between10and15%ofthesepatientsarepoten-
tial candidates for surgical treatment (Guberman and
Bruni, 1999).
Resective epilepsy surgery is mostly carried out for
cases of symptomatic focal epilepsy and 70 to 90%
of patients who have undergone surgery suffer from
mesial temporal lobe epilepsy (Jallon and Loiseau,
2001). Of the patients with temporal lobe epilepsy, 49
to 80% of adults are drug-resistant and mesial tempo-
ralsclerosisistheprimarycause,beingresponsiblefor
between 65% and 70% of cases (Zumsteg et al., 2006).
Depending on when and how measurements were
performed, as well as the surgical procedure car-
ried out, seizure outcome after temporal lobe
epilepsy surgery varies between authors. The study by
McIntosh et al., (2001), reported that 64% of patients
did not experience seizures causing impairment of
consciousness one year after surgery and, further-
more, 42% were seizure-free. Other authors (Foldvary
etal.,2000)obtainedsatisfactoryresultsforthecontrol
of seizures (Engel class I and II) in 80% of patients in
the second year of follow-up.
Recently, an increasing number of studies have
attempted to identify predictive variables which will
guarantee appropriate selection of candidates for
surgery to obtain the best postoperative results. How-
ever, there is so far no agreement on the key variables.
In Spain, Villanueva et al., (2004) reported the out-
comes of 41 patients, based on the prognostic factor
analysis according to Engel and ILAE (International
LeagueAgainstEpilepsy)postoperativeseizurecontrol
classifications. In contrast to other series, one study
(Solaetal.,2005)reportedalowrateoftemporalmesial
sclerosis cases (14%).
The objective of this study was to investigate the out-
come of temporal lobe epilepsy surgery carried out in
ourhospitalandtoidentifyvariableswhicharepredic-
tive of good seizure control in the second and fourth
year of follow-up.
Methods
We studied retrospectively a cohort comprised of
patients suffering from drug-resistant temporal lobe
epilepsy, who underwent surgery at the Epilepsy
Unit of Cruces Hospital between January 1996 and
December 2007. Epilepsy was considered to be drug-
resistant when at least two antiepileptic drugs used
alone, and in combination, had failed to control
seizures (table 1).
Preoperative evaluation included: neurological, psy-
chiatric and neuropsychological assessment, epilepsy
protocolmagneticresonanceimaging(MRI)usinga1.5
Tesla resonator, computer-assisted campimetry, and
long-term video-electroencephalography (video-EEG)
which included recording of at least two seizures.
If results were not conclusive, additional tests were
carried out such as monitoring using foramen ovale
electrodes, subdural grids and strip electrodes, MRI
with a 3 Tesla machine, as well as positron emis-
sion tomography (PET) or single photon emission
computed tomography (SPECT) in critical and/or inter-
critical phases. Finally, an interdisciplinary commit-
tee (composed of neurologists, neuropaediatricians,
neurophysiologists,neurosurgeons,
gists, a psychiatrist and a neuropsychologist) decided
whether the patients were appropriate candidates for
surgery.
Possiblesurgicalapproachesincludedanteriortempo-
ral lobectomy (sparing the superior temporal gyrus)
with the removal of the amygdala and hippocampus,
or in cases of vascular malformations, tumour or dys-
plastictissue,lesionectomywithorwithoutremovalof
the amygdala and hippocampus.
Postoperative follow-up was carried out after one,
three and six months and then yearly. MRI monitoring,
24-hour video-EEG, computer-assisted campimetry,
and neuropsychological assessment were performed
preferentially within the first year after surgery.
Patients were excluded from the study if they pre-
sented extratemporal lesions on neuroimaging, if they
were not monitored during the first two years after
surgery and if they underwent surgery without follow-
ing the preoperative protocol.
The studied variables were: sex, age at onset of
seizures, age at the time of surgery, duration of
epilepsy, physical examination, history of febrile con-
vulsions (Freeman, 1980) and central nervous system
infections, MRI findings of the brain, surgical tech-
nique, laterality of surgery, year of surgery, average
number of seizures per month in the year prior to
surgery, presence of typical ictal semiology of mesial
temporal epilepsy (Wieser, 2004), presence of aura
(Blume et al., 2001), pre- and postoperative intelli-
gence quotient (IQ), neurophysiological parameters
based on video-EEG, location of intercritical epilep-
tiform activity, frequency, location and spread of the
initial ictal pattern, result of the postoperative video-
EEG, invasive and semi-invasive monitoring, surgical
re-intervention for epilepsy, presence of two or more
seizures in the first year after surgery (excluding
neuroradiolo-

Page 3

38
Epileptic Disord, Vol. 13, No. 1, March 2011
M.I. Forcadas-Berdusán, et al.
Table 1. Characteristics of the cohort of patients who underwent surgery for temporal lobe epilepsy betweeen
1996 and 2007 (n=115).
Variable Number of cases (%)Average (DS)
Sex Female
Male
68 (59.1)
47 (40.9)
Age at onset of epilepsy Median: 9 yearsIR: 14 years
Age at surgery 35.60 (11.58) years
Duration of epilepsy 23.77 (11.56) years
Medical historyAbnormal neurological
CNS Infection
Febrile convulsion
4 (3.5)
15 (13)
35 (30.4)
Laterality of surgery Right
Left
53 (46.1)
62 (53.9)
MRI findingsNormal
Mesial Sclerosis
DNET
Other tumours
VM (vascular malformation)
Cortical dysplasia
Dual pathology
Non-conclusive
7 (6.1)
77 (67)
5 (4.3)
5 (4.3)
3 (2.6)
6 (5.2)
2 (1.7)
10 (8.7)
Surgical technique Anterior temporal lobectomy
with amygdalo-
hippocampectomy
Lesionectomy with/without
removal of amygdala and
hippocampus
105 (91.3)
10 (8.7)
Preoperative IQGlobal
Verbal
Manipulative
97.07 (16.73)
96.80 (16.67)
96.48 (16.65)
Postoperative IQ Global
Verbal
Manipulative
98.90 (16.27)
97.96 (16.83)
99.56 (15.34)
Average number of
seizures/month
Median: 6IR: 8
Typical semiology of
mesial TLE
58 (50.4)
Aura 66 (57.4)
Video-EEG location of
interictal discharges
Unitemporal
Right
Left
Bitemporal (independent)
Multifocal
Normal
74 (64.3)
32
42
34 (29.6)
5 (4.3)
2 (1.7)

Page 4

Epileptic Disord, Vol. 13, No. 1, March 2011
39
Predictive factors for TLE surgical outcome
Table 1. (Continued)
Variable Number of cases (%)Average (DS)
EEG frequency of
initial ictal pattern
<5Hz
5-8Hz in the first 30 seconds
5-8Hz after 30 seconds
>8Hz
Irregular
15 (13)
88 (76.5)
7 (6.1)
4 (3.5)
1 (0.9)
Video-EEG location
of ictal pattern
Unitemporal
Right
Left
Synchronous bitemporal
Not temporal location
No lateral discharges noted
Diffuse
103 (89.6)
48
55
3 (2.6)
5 (4.3)
2 (1.7)
2 (1.7)
Video-EEG spread
of ictal pattern
58 (50.4)
Monitoring with
additional electrodes
Foramen ovale electrodes
Subdural strips
13 (11.3)
3 (2.6)
Surgical re-intervention
for epilepsy
6 (5.2)
Postoperative video-EEG Normal
Abnormal
71 (61.7)
44 (38.3)
Two or more seizure
episodes in the first year
after surgery
31 (27)
Good prognosis at year 2
after surgery
99/115 (86.1)
Good prognosis at year 4
after surgery
83/93 (89.2)
seizures within the first month), results of the post-
operative classification in the second and fourth year
after surgery, and presence of complications and/or
sequelae.
More specifically, the duration of epilepsy was mea-
sured from the first non-febrile seizure until surgery.
Based on MRI, mesial sclerosis was defined as the
presence of at least two of the following features:
hippocampal formation volume loss, temporal lobe
atrophy with involvement of the parahippocampal
gyrus, volume loss in three-dimensional volumetry
of the hippocampus, and T2 signal hyperintensity.
The category “other tumours” corresponded to low
grade tumours, all of which had no substantial growth
effects and no evidence of recurrence at consecu-
tive follow-up visits. The variable “typical semiology
of temporal mesial epilepsy” included the four follow-
ing phenomena: aura, arrest of activity, automatisms,
and impairment of consciousness. IQ was established
using the Wechsler Intelligence Scale. The location of
interictal epileptiform activity was considered to be
“independentbitemporal”whenthepresenceofsuch
activitywasdetectedinbothtemporallobesregardless
of the relative involvement of each lobe. By “surgical
re-intervention”, we refer to those cases in which the
resection was enlarged in a second intervention due
to failure to control seizures, attributed to an initial
incomplete resection of the hippocampus. In post-
operative follow-up, seizure status was assessed using
the Engel classification system (Engel et al., 1993).
ThecasesweredrawnfromthearchivesoftheEpilepsy
Unit and corresponding data extracted from clinical

Page 5

40
Epileptic Disord, Vol. 13, No. 1, March 2011
M.I. Forcadas-Berdusán, et al.
histories, after obtaining verbal and written informed
consent. Data analysis was carried out using SPSS 16.0
statistical software.
For the bivariate analysis, the dependant variable
“prognosis” was categorised in the following way:
“Good prognosis” for Engel class I and II patients, and
“poor prognosis” for Engel class III and IV patients.
In addition, other data were converted to dichotomic
variables: the “NMR findings” variable was trans-
formed by contrasting mesial temporal sclerosis with
other cases; pre- and postoperative IQ by compa-
ring scores which were either less than, greater than
or equal to 90; “location of interictal discharges and
location of the initial ictal pattern in the video-EEG”
by comparing cases of unilateral temporal focus with
the other cases; “frequency of the initial ictal pattern
in the video-EEG” by comparing relative frequencies
of 5-8 Mhz; “year of surgery” by comparing before or
during the year 2000 with later times points; “age at
onset of epilepsy” by comparing patients under the
age of 11 years old with those aged 11 and older; “age
at surgery” by comparing those aged 35 years and
under with those over the age of 35; and “duration
of epilepsy” by comparing duration of less than 23
years with greater than or equal to 23 years. For the
last three variables the average was used as the cut-off
point.
Categorical variables were characterised using fre-
quencydistributionsandpercentages,whilemeasures
of central tendency, such as the mean and median,
and variability, such as standard deviation (SD) or
inter-quartile range (IR), were used for quantita-
tive variables. The mean and standard deviation
was calculated for less asymmetric distributions and
the median and inter-quartile range were calcu-
lated for asymmetric distributions. The comparison
of proportions between categorical variables was per-
formed using Pearson’s Chi-square coefficient and
corresponding corrections, such as Fisher’s correc-
tion, when the expected frequencies were less than
five.
Univariate logistic regression analysis was undertaken
toidentifywhichvariableswereassociatedwithagood
prognosis in the postoperative follow-up of seizure
status. The variables found to be associated with a
good prognosis (p<0.200) were included in a non-
automaticstepwisemultivariatelogisticregression.Of
allthesevariables,theonewiththehighestpvaluewas
removed from the model, and the logistic regression
was performed again using the remaining variables.
This process was repeated, each time removing the
variable with the highest p value. The process was
complete when all the variables used in the model
were significant (p<0.05). Results are expressed as OR
(oddsratio)and95%CI(95%confidenceintervals).The
robustness of the model was assessed on the basis of
the area under the ROC curve and the model assump-
tions checked using the residuals.
The study was approved by the institution’s Ethics
Committee.
Results
The cohort from the Epilepsy Unit of Cruces Hospi-
tal comprised 115 patients who underwent resective
surgery for the treatment of temporal lobe epilepsy
between 1996 and 2007 (inclusive). All patients met the
presurgicalselectioncriteria(asexplainedabove).The
patient characteristics are shown in table 1.
The cases of dual pathology included those with
mesial sclerosis and cortical dysplasia, both involving
the temporal lobe. The most commonly used surgi-
cal technique was anterior temporal lobectomy with
removal of the amygdala and hippocampus. Lesionec-
tomy was carried out in three cases and lesionectomy
with removal of the amygdala and hippocampus in
seven. In the three cases where additional monitor-
ing with subdural strips was required, foramen ovale
electrodes were also used.
Using the Engel classification to assess the control of
seizures,thefollowingwasidentified:attwoyearsafter
surgery (n=115), 85 (73.9%), 14 (12.2%), 14 (12.2%) and
2 (1.7%) patients were classified as Engel Classes I, II,
III and IV, respectively and at four years after surgery
(n=93), 71 (76.3%), 12 (12.9%), 8 (8.6%) and 2 (2.2%)
patients were classified as Engel Classes I, II, III and
IV, respectively. Two patients were lost in the follow-
up by the end of the fourth year due to death (in one
case from a neck neoplasm and the other from brain
haemorrhage due to head trauma caused by seizures).
Twentypatientsdidnotcompletethefour-yearfollow-
up as they underwent surgery after 2005; of these
patients, 16 were classified as Engel class I, two as class
II, two as class III and no patients as class IV, after two
years.
Of the 93 patients studied at both follow-up points, 64
(68,8%) were classified as Engel class I for the entire
period and 78 (83,8%) had good prognosis (Engel class
IandII).Thepercentageofpatientswithgoodprogno-
sis for postoperative seizure control according to the
MRIfindingsisshownintable2.Thegroupsofpatients
withtheworstresultsincludedpatientssufferingfrom
focal cortical dysplasia and those with non-conclusive
recordedevents.Logisticregressionanalysisusingthe
variables associated with good postoperative progno-
sis was performed for the second and fourth years, as
well as for the group with good prognosis.
Thevariablesassociatedwithagoodprognosisforcon-
trolofseizuresinthesecondpostoperativeyear,witha
valueofp<0.05,were“absenceoftwoormoreseizure
episodes in the first year after surgery” and “normal

Page 6

Epileptic Disord, Vol. 13, No. 1, March 2011
41
Predictive factors for TLE surgical outcome
Table 2. Percentage of cases with a good prognosis in postoperative follow-up.
Second yearFourth Year Maintained across both years
MRI findingsn=115n=93 n=93
Normal85.7%100% 85.7%
Mesial temporal sclerosis89.6% 95.2% 87.3%
DNET100%100%100%
Other tumours80%75%75%
Vascular malformations100% 100% 100%
Cortical dysplasias50%60%50%
Dual pathology100%100%100%
Non-conclusive70%60%60%
Table 3. Logistic regression with results from the second year after surgery.
Variables associated with a good postoperative prognosis.
Univariate analysisa
Multivariate analysisb,c
OR CI 95%p OR CI 95%p
Absence of two or more seizure episodes in
the first year after surgery
33.765 7.018-162.452
<0.001 39.627 5.683-276.317
<0.001
Unilateral ictal temporal location-- 0.052---
Age at surgery >35 years-- 0.1865.931 1.006-34.9820.049
Normal postoperative video-EEG36.207 4.569-286.9330.001 31.553 3.072-324.0570.004
Reintervention surgery-- 0.180---
Surgery after year 2000-- 0.107---
Unilateral interictal temporal location--0.071---
Temporal mesial sclerosis from MRI-- 0.127---
aValues with p<0.200 are shown. The odds ratio and the confidence interval of 95% are only given for variables with p<0.05.
bOnly variables from the final multivariate logistic regression model are shown.cArea under the ROC curve: 0.940.
postoperative video-EEG” in the univariate analysis.
The other variables were not statistically significant.
However, in addition to the two variables mentioned
above, the variable “age at surgery over 35 years old”
was also statistically significant in the multivariate ana-
lysis (table 3).
With respect to the fourth year of follow-up, the uni-
variateanalysisshowedfourvariablesassociatedwitha
good prognosis with statistical significance: “absence
of two or more seizure episodes in the first year after
surgery”, “unilateral ictal temporal location”, “mesial
temporalsclerosis”,and“normalpostoperativevideo-
EEG”. In the multivariate analysis, the first two of these
variables were not included but another, the variable
“female”, was included (table 4).
The univariate analysis showed that in the group with
good prognosis in both the second and fourth year
aftersurgery,thefollowingvariableshadstatisticalsig-
nificance: “absence of two or more seizure episodes
in the first year after surgery”, “unilateral ictal and

Page 7

42
Epileptic Disord, Vol. 13, No. 1, March 2011
M.I. Forcadas-Berdusán, et al.
Table 4. Logistic regression with results from the fourth year after surgery.
Variables associated with a good postoperative prognosis.
Univariate analysisa
Multivariate analysisb,c
ORCI 95%P ORCI 95%p
Absence of two or more seizure episodes in
the first year after surgeryd
7.860 1.851-33.3790.005---
Mesial temporal Sclerosis5.736 1.368-24.0490.017 15.163 2.335-98.4500.004
Unilateral ictal temporal location5.500 1.125-26.8970.035---
Female-- 0.177 6.4151.025-40.135 0.047
Presence of Aura-- 0.124---
Normal postoperative video-EEG7.857 1.563-39.501 0.01232.1563.661-282.432 0.002
Age at onset of the first seizure epiosode
>10years
-- 0.181---
Surgery after year 2000-- 0.155---
aValues with p<0.200 are shown. The odds ratio and the confidence interval of 95% are only given for variables with p<0.05.
bOnly variables from the final multivariate logistic regression model are shown.cArea under the ROC curve: 0.854.
dRemoved from the multivariate analysis in the last step with p=0.066.
Table 5. Logistic regression for the subgroup with a good prognosis
at both the second and fourth year after surgery.
Univariate analysisa
Multivariate analysisb,c
OR CI 95%p OR CI 95%p
Absence of two or more seizure episodes in
the first year after surgery
19.911 5.070-78.188
<0.00116.301 3.704-71.740
<0.001
Unilateral ictal temporal location4.554 1.078-19.2370.039---
Intercritical unilateral temporal location 4.5111.509-13.486 0.007---
Normal postoperative Video-EEG17.187 3.643-81.089
<0.00113.7622.566-73.808 0.002
Surgery after year 2000-- 0.153---
Mesial Temporal Sclerosis-- 0.081---
aValues with p<0.200 are shown. The odds ratio and the confidence interval of 95% are only given for variables with p<0.05.bOnly
variables from the final multivariate logistic regression model are shown.cArea under the ROC curve: 0.912.
interictal temporal location”, and “normal postopera-
tive video-EEG”. In the multivariate analysis, only the
first and last of these variables demonstrated p<0.05
(table 5).
The values of the area under the ROC curve of the
logistic regression models in the second and fourth
year, and in the group which maintained a good
postoperative prognosis, were 0.940, 0.854 and 0.912
respectively.
In relation to the sequelae and complications after
surgery, contralateral homonym quadrantanopsia was
the most common (43 patients), followed by paresis of

Page 8

Epileptic Disord, Vol. 13, No. 1, March 2011
43
Predictive factors for TLE surgical outcome
Table 6. Sequelae and/or complications associated
with surgery for temporal lobe epilepsy.
Sequelae and/or complicationsNumber of cases
Quadrantanopsia 43
Verbal memory impairment3
CSF fistula3
CN III paresis 16
CN VI paresis2
CN VII paresis5
Persistent tinnitus1
Haemorrhagic CVD2
Ischaemic CVD4
Dysphasia 10
Infection following craniectomy1
Postoperative cavity3
Anxiety1
Psychosis2
Depression3
Personality disorders3
Temporary hemiparesis with no
evidence of ischaemic lesion on
neuroimaging
2
Epidural haematoma requiring drainage 1
CSF: cerebrospinal fluid; CN: cranial nerve;
CVD: Cerebrovascular disease/
cranial nerves (23 cases) and dysphasia (10 patients).
Ischaemic cerebrovascular (four patients) and haem-
orrhagic events including epidural haematoma (three
cases) were considered to be severe complications.
However, no deaths were associated with surgery.
Moderate to severe persistent neurological deficits
were seen in three patients, two with hemiparesis and
one with dysphasia. All deficits, including those which
were psychiatric, were seen before the second year
(table 6).
Discussion
The general characteristics of this cohort are similar to
those of others (Salanova et al., 2002; Villanueva et al.,
2004; Elsharkawy et al., 2009; Jaramillo-Betancur et al.,
2009). For temporal lobe epilepsy surgery, we believe
therestillexistsasignificantdelaybeforesurgery,thus,
all efforts should be made to warrant early identifica-
tion of patients with potential drug-resistant epilepsy.
This would allow appropriate referral to specialized
epilepsy units in order to manage refractory cases cor-
rectly and without delay, and enable candidates who
would benefit from surgical treatment to be selected.
In addition, this study illustrates the high prevalence
ofmesialtemporalsclerosisaspreoperativepathology
amongpatientswithrefractorytemporallobeepilepsy.
The postoperative outcome of seizure control is con-
sistent with, or marginally better than, the outcome
reported by other authors (Tellez-Zenteno et al., 2005;
Janszky et al., 2005). Our data are also similar to very
recent reports (Elsharkawy et al., 2009), demonstrating
continued benefit, two and four years after surgery. In
our series, more than two thirds of patients were free
of disabling seizures during the entire study period,
and almost 84% were classified in the group with good
prognosis (Engel class I and II). The percentage of
patients with good prognosis at year four may even
havebeenhigher,sincedataarecurrentlylackingfrom
20 patients who have not yet completed the four-year
follow-up, 90% of whom had a good prognosis at the
secondyear.Ourpositiveresultsregardingthecontrol
of seizures after surgery are directly related to strict
selection criteria and correlation between video-EEG
monitoring and ictal semiology. It may be surprising
that more patients had a good prognosis after four
years than after two years. We assume that this is
because we established the postoperative prognosis
using the Engel classification, which is dynamic and
based on the number of seizures in one period, and
so may vary in the same patient from one period to
another. We believe the cause may be multifactorial;
possible reasons for this include antiepileptic drug
adjustment and the “running-down” phenomenon.
Most publications regarding seizure prognosis after
temporal lobe surgery have proposed measures of
associationwithpre-andpostoperativevariables.Find-
ings vary depending on the surgical protocols of the
respective institution, the diagnostic tests available at
the time, research design, and data analysis. Taken
together, a fair comparison between different series
is difficult.
In our series, we found an association between a good
prognosis and the following variables, for at least one
of the time periods analysed: absence of two or more
seizure episodes in the first year after surgery, normal
postoperative video-EEG, temporal unilateral inter-
critallocalization,unilateraltemporalictallocalization,
female gender, and age 35 years or older. The first
five of these variables have already been mentioned

Page 9

44
Epileptic Disord, Vol. 13, No. 1, March 2011
M.I. Forcadas-Berdusán, et al.
in other series (Armon et al., 1996; Radhakrishnan
et al., 1998; Park et al., 2002: Di Gennaro et al., 2004;
Janszky et al., 2005; Cohen-Gadol et al., 2006; Jeha
et al., 2006). However, the last of these variables has
not been previously described, rather, other reports
have suggested a worse prognosis with increasing age
at surgery (Hennessy et al., 2001; Janszky et al., 2005).
We have not been able to explain this finding. More-
over, this variable was identified from data from the
secondyear,andnotfromthefourthyearorthegroup
of patients with good prognosis in both years, thus we
areunabletoprovideanexplanationforthisidentified
association.
While the information relating to associations derived
from the bivariate analysis is in itself useful, it would
bemoreinformativetoproduceapredictivemodelfor
postoperative results and this is why a logistic regres-
sion analysis was undertaken.
Inourstudy,themultivariatelogisticregressionmodel
could not be based on preoperative variables alone,
for any of the cut-off points. The same difficulty was
encountered in other reports (Uijl et al., 2008) which,
using only preoperative variables, did not succeed
in establishing a good predictive model for postop-
erative seizure control. Furthermore, in our results,
for the group with a good prognosis in the second
postoperative year, the model only included postop-
erative variables, i.e. absence of two or more seizure
episodes in the first year after surgery and normal
postoperative video-EEG. However, the model still
produced a high value for the area under the ROC
curve (0.912), as did the models for the second (0.94)
and fourth years (0.854), although confidence inter-
vals were wide. Only the variable relating to normal
postoperative video-EEG was present in all models,
but the absence of two or more seizure episodes in
the first year after surgery was also present in two
models (in the four-year postsurgical follow-up it was
removed from the multivariate analysis in the last
step; p=0.066), and more importantly, in the model
of good prognosis at both the second and fourth
years. Therefore, we suppose that it may be an impor-
tant variable associated with good prognosis. One
reason that may explain why we were unable to iden-
tify a model based on preoperative variables may be
due to the fact that the presurgery study was not
completely homogeneous for all patients, since, for
some patients, specific studies such as deep elec-
trodes were carried out (as described in the methods).
However, postsurgical follow-up was the same for all
patients (24-hour video-EEG monitoring and control
of seizures reported in consecutive visits). Another
reason may be the limits of the preoperative
basic studies, i.e. semiology, neuroimaging with
a 1.5 Tesla resonator and long-term video-EEG
monitoring.
Thesamplesizeofourstudymayhavelimitedthepos-
sibilitiesoffindinganassociationbetweensomeother
variables and a good prognosis. Because the majo-
rity of patients who undergo temporal lobe epilepsy
have good prognosis (as in other series), the group of
patientswithpooroutcomeissmall,henceconfidence
intervalsarewide.Alargersamplemayhaveproduced
abetterfitfortheconfidenceintervalsofourpredictor
variables. Nevertheless, the area under the ROC curve
is high in all the models.
We therefore believe that we have developed a reli-
able predictive model for sustained good prognosis
over a period of three years. Although not useful
as a preoperative tool since it is comprised of post-
operative variables, this may be used as a predictive
model for patient monitoring, as both variables may
be measured in the early postoperative period (one
year after surgery). In this way, the results of this
modelcouldbeconsideredwhendecidingwhetherto
reduce or withdraw antiepileptic drugs and even with
respecttothedailylifeofpatients(suchasconsidering
whethertheyshoulddriveavehicle).Furthermore,the
variables included in our models may be considered
in studies attempting to establish a predictive model
for the postoperative control of seizures, including a
larger number of patients to overcome the limitation
imposed by the sample size.
Finally, we must consider surgical re-intervention,
sequelae and/or complications. Re-intervention was
deemed appropriate when failure to control seizures
was associated with an incomplete resection of the
hippocampus. In relation to this, the surgeon’s expe-
rience plays a crucial role since the majority of
these cases occurred during the initial part of the
surgeon’s learning curve. After re-intervention, all
patients were seizure-free on follow-up. All the side
effectsappearedinthefirsttwoyearsofthefollow-up.
Haemorrhagiccerebrovasculareventscanbeconside-
red to have occurred by chance, as there was no
anatomical correlation between the haemorrhage and
the epileptogenic lesion site. There was spontaneous
remission of paresis of cranial nerves. In our series,
moderate or severe neurological persistent deficits
only occurred in three patients (two with hemiparesis
and one with dysphasia). Psychiatric sequelae were, in
allcases,controlledwithmedication.Personalitydisor-
dersandmemoryandlanguageimpairmentareamong
the risks to be weighed up for each patient (Hermann
et al., 1991; Behrens et al., 1997; Blumer et al., 1998;
Salanova et al., 2002; Tanriverdi, 2009). ?

Page 10

Epileptic Disord, Vol. 13, No. 1, March 2011
45
Predictive factors for TLE surgical outcome
Acknowledgments.
The work has been carried out in memory of Dr. P. Madoz (Head
of the Neurophysiology Service at Cruces Hospital).
WeexpressourthankstoDrs.J.JZarranzImirizaldu,J.PratsVi˜ nas
and J. Garibi Undabarrena, Heads of Neurology, Neuropaedi-
atrics and Neurosurgery, for their support in the establishment
and running of this service, the nursing team without whose
excellent technical and person-to-person assistance this publi-
cation would not have been possible, and to all those colleagues
who referred their patients to us.
Disclosure.
Dr. Jose Luis Bustos-Sanchez was sponsored by a grant from the
Fundación Carolina-BBVA.
None of the authors has any conflict of interest to declare.
References
Aicardi J, Shorvon SD. Intratable Epilepsy. In: EngelJ,
PedleyTA, editors. Epilepsy: a comprehensive textbook.
Philadelphia:Lippincott-Raven, 1997, p. 1325-31.
Armon C, Radtke RA, Friedman AH, Dawson DV. Predictors
of outcome of epilepsy surgery: multivariate analysis with
validation. Epilepsia 1996;37:814-21.
Behrens E, Schramm J, Zentner J, König T. Surgical and neu-
rological complications in series of 708 epilepsy surgery
procedures. Neurosurgery 1997;4:1-9.
Blume WT, Lüders HO, Mazrahi E, Tassinari C, van Emde
W, Engel Jr. J. Glossary of descriptive terminology for ictal
semiology: report of the ILAE task force on classification and
terminology. Epilepsia 2001;42:1212-8.
Blumer D, Wakhlu S, Davies K, Hermann B. Psychiatric
outcome of temporal lobectomy for epilepsy: incidence
and treatment ofpsychiatric
1998;39:478-86.
complications. Epilepsia
Cohen-Gadol AA, Wilhelmi BG, Collignon F, et al. Long-term
outcome of epilepsy surgery among 399 patients with nonle-
sional seizure foci including mesial temporal lobe sclerosis.
J Neurosurg 2006;104:513-24.
Di Gennaro G, Quarato PP, Sebastiano F, et al. Postoperative
EEG and seizure outcome in temporal lobe epilepsy surgery.
Clin Neurophysiol 2004;115:1212-9.
Elsharkawy AE, Alabbasi AH, Pannek H, et al. Long-term out-
comeaftertemporallobeepilepsysurgeryin434consecutive
adult patients. J Neurosurg 2009;110:1135-46.
Engel J. Jr, Van Ness P, Rasmussen T, Ojeman LM. Outcome
with respect to epileptic seizures. In: Engel J Jr. Surgical
treatment of the epilepsie. New York: Raven Press, 1993:
609-21.
FoldvaryN,NasholdB,MaschaE,et al.Seizureoutcomeafter
temporal lobectomy for temporal lobe epilepsy: a Kaplan-
Meier survival analysis. Neurology 2000;54:630-4.
Freeman JM. Febrile Seizures: a consensus on their signifi-
cance, evaluation and treatment. Pediatrics 1980;66:1009-12.
Guberman A, Bruni J. Surgical Treatment. In: Guberman A,
Bruni J. Essentials of Clinical Epilepsy. Woburn: Butterworth
Heinemann, 1999: 181-8.
Hauser
PedleyTA.Epilepsy:acomprehensivetextbook.Philadelphia:
Lippincott-Raven, 1997: 47-57.
WA. IncidenceandPrevalence. In:EngelJ,
Hauser WA, Hesdorffer DC. Remission, intratability, mortal-
ity and comorbility of seizures. In Wyllie E. The treatment of
epilepsy.Philadelphia:LippinconttWilliamns&Wilkins,2001:
139-45.
Hennessy MJ, Elwes RD, Honovar M, Rabe-Hesketh S,
Binnie CD, Polkey CE. Predictors of outcome and patholog-
ical considerations in the surgical treatment of intractable
epilepsyassociatedwithtemporallobelesions.JNeurolNeu-
rosurg Psychiatry 2001;70:450-8.
HermannBP,WylerbAR,SomesG.Languagefunctionfollow-
ing anterior temporal lobectomy. J Neurosurg 1991;74:560-6.
Jallon P, Loiseau J. Newly diagnosed unprovoked epileptic
seizures: presentation at diagnosis in CAROLE study. Epilep-
sia 2001;42:464-75.
Janszky J, Janszky I, Schulz R, et al. Temporal lobe epilepsy
with hippocampal sclerosis: predictor for long-term surgical
outcome. Brain 2005;128:395-404.
Jaramillo-BetancurH,JiménezME,Massaro-CeballosM,et al.
Risk factor for seizures recurrence and short term outcome
afterepilepsysurgeryformesialtemporalsclerosis.RevNeu-
rol 2009;49:175-80.
Jeha LE, Najm IM, Bingaman WE, et al. Predictors of outcome
after temporal lobectomy for the treatment of intractable
epilepsy. Neurology 2006;66:1938-40.
McIntosh AM, Wilson SJ, Berkovic SF. Seizure outcome after
temporal lobectomy: current research practice and findings.
Epilepsia 2001;42:1288-307.
Park K, Buchhalter J, McClellarnd R, Raffel C. Frecuency
and significance of acute postoperative seizures follow-
ing epilepsy surgery in children and adolescents. Epilepsia
2002;43:874-81.
Radhakrishnan K, So EL, Silbert PL, et al. Predictors of
outcome of anterior temporal lobectomy for intractable
epilepsy: A multivariate study. Neurology 1998;51:465-71.
Salanova V, Markand O, Worth R. Temporal Lobe Epilepsy
Surgery: Outcome, Complications, and Late Mortality Rate
in 215 Patients. Epilepsia 2002;43:170-4.
Sola RG, Hernando-Requejo V, Pastor J, et al. Pharmacoresis-
tant temporal-lobe epilepsy. Exploration with foramen oval
electrodes and surgical outcomes. Rev Neurol 2005;41:4-16.
TanriverdiT,AjlanA,PoulinN,OlivierA.Morbidityinepilepsy
surgery:anexperiencebasedon2449epilepsysurgeryproce-
duresfromasingleinstitution.JNeurosurg2009;110:1111-23.
Tellez-Zenteno JF, Dhar R, Wiebe S. Long-term seizure out-
comes following epilepsy surgery: a systematic review and
meta-analysis. Brain 2005;128:1188-98.

Page 11

46
Epileptic Disord, Vol. 13, No. 1, March 2011
M.I. Forcadas-Berdusán, et al.
UijlSG,LeijtenFS,ArendsJB,ParraJ,VanHuffelenAC,Moons
KG.Prognosisaftertemporallobeepilepsysurgery:thevalue
of combining predictors. Epilepsia 2008;49:1317-23.
VillanuevaV,PeralE,AlbisuaJ,deFelipeJ,SerratosaJM.Prog-
nostic factors in temporal lobe epilepsy surgery. Neurología
2004;19:92-8.
Wieser HG. ILAE Commission on Neurosurgery of Epilepsy.
ILAE Commission Report. Mesial Temporal Lobe Epilepsy
with Hippocampal Sclerosis. Epilepsia 2004;45:695-714.
Zumsteg D, Andrade DM, Wennberg RA. Clinical electro-
physiology factors indicative of intractability in patients with
temporal lobe epilepsy. Adv Neurol 2006;97:45-62.