Epilepsy Research 122 (2016) 3743

Contents lists available at www.sciencedirect.com

Epilepsy Research
journal homepage: www.elsevier.com/locate/epilepsyres

Resective focal epilepsy surgery Has selection

of candidates changed? A systematic review
Churl-Su Kwon a , Jonathan Neal a , Jose Tellz-Zenteno b , Amy Metcalfe c ,
Kathryn Fitzgerald d , Lizbeth Hernandez-Ronquillo b , Walter Hader e , Samuel Wiebe e,f ,
Nathalie Jett e,f, , for the CASES Investigators
a

Department of Neurosurgery, Massachusetts General Hospital, Boston, MA, USA

Department of Neurology, University of Saskatchewan, Saskatoon, Canada
Department of Obstetrics and Gynecology, Alberta Childrens Hospital Research Institute, University of Calgary, Calgary, Alberta, Canada
d
Harvard School of Public Health, Boston, MA, USA
e
Department of Clinical Neurosciences, Hotchkiss Brain Institute, University of Calgary, Alberta, Canada
f
Department of Community Health Sciences, OBrien Institute for Public Health, University of Calgary, Alberta, Canada
b
c

a r t i c l e

i n f o

Article history:
Received 14 May 2015
Received in revised form 13 January 2016
Accepted 11 February 2016
Available online 12 February 2016
Keywords:
Epilepsy surgery
Temporal lobe epilepsy
Extra-temporal lobe epilepsy
Systematic review

a b s t r a c t
Objective: No standard, widely accepted criteria exist to determine who should be referred for an epilepsy
surgical evaluation. As a result, indications for epilepsy surgery evaluation vary significantly between
centers. We review the literature to assess what criteria have been used to select patients for resective
epilepsy surgery and examine whether these have changed since the publication of the first epilepsy
surgery randomized controlled trial in 2001.
Methods: A systematic review was conducted using PubMed and EMBASE, bibliographies of reviews and
book chapters identifying focal epilepsy resective series. Abstract, full text review and data abstraction (i.e.
indications for surgery) were performed independently by two reviewers. Descriptive historical analysis
was done to examine indications over time.
Results: Out of 5061 articles related to epilepsy surgery, 384 articles met all eligibility criteria. Most common criteria for selecting patients for evaluation for resective surgery were: AED resistance (n = 303, most
commonly >2 AEDs = 46), epilepsy duration (n = 53, most commonly >1 year = 42) and seizure frequency
(most commonly at least one seizure/month, n = 29). Out of the prospective studies the most notable
change over time (pre-2000 vs. post-2000) was failure of 2 AEDs (8% vs. 43% respectively, p < 0.001).
Conclusions: Important variations between studies make it difficult to identify consistent criteria to guide
surgical candidacy or changes in indications over time. With increasing evidence that earlier surgery
is associated with better outcomes, it is recommended that patients be evaluated as soon as they have
failed two AEDs, consistent with the new definition of drug resistant epilepsy. Furthermore, low seizure
frequency should not be a barrier to epilepsy surgery. Anyone with drug resistant epilepsy should be
promptly evaluated for possible surgery, regardless of seizure frequency.
2016 Elsevier B.V. All rights reserved.

1. Introduction

Corresponding author at: Department of Clinical Neurosciences, Foothills Medical Centre, 1403-29 St NW, Calgary, Alberta, Canada.
E-mail addresses: churlsu.kwon@gmail.com (C.-S. Kwon), jonbneal@gmail.com
(J. Neal), jose.tellez@usask.ca (J. Tellz-Zenteno),
amy.metcalfe@albertahealthservices.ca (A. Metcalfe), katefitzgrld@gmail.com
(K. Fitzgerald), liztellez33@hotmail.com (L. Hernandez-Ronquillo),
walter.hader@albertahealthservices.ca (W. Hader), swiebe@ucalgary.ca (S. Wiebe),
Nathalie.jette@albertahealthservices.ca (N. Jett).
http://dx.doi.org/10.1016/j.eplepsyres.2016.02.007
0920-1211/ 2016 Elsevier B.V. All rights reserved.

No standard, widely accepted criteria exist to determine who

should be referred for an epilepsy surgical evaluation. As a
result, indications for epilepsy surgery evaluation vary significantly
between centers.
In 2003 the International League Against Epilepsy (ILAE) Pediatric Epilepsy Surgery Subcommission proposed that the following
individuals be considered possible surgical candidates: (1) those
who have failed two appropriate antiepileptic drugs (AEDs); (2)
those who are disabled by medication side effects; (3) those with

38

C.-S. Kwon et al. / Epilepsy Research 122 (2016) 3743

seizures that do not fit into a definite electroclinical epilepsy syndrome based on the ILAE classification; and (4) those that have
an identified lesion seen on neuroimaging (Cross et al., 2006). In
infants and children with frequent seizures, rapid surgical evaluation may be necessary to prevent deleterious effects on brain
development (Engel and Shewmon, 1993). In 2009 the American
Epilepsy Society consensus conference also underlined the importance of integration of data and care pathways in selecting patients
for epilepsy surgery, utilizing a well-functioning multidisciplinary
team with a systematic approach to investigations and thus conveying realistic expectations for surgical outcomes to patients and their
families (Duncan, 2011). Other recommendations have been published to guide selection of surgical candidates for epilepsy surgery
but the recommendations are still very general (Engel et al., 2003;
Labiner et al., 2010). Most agree that for a patient to be considered
for surgery, the seizures should be disabling; however, this concept is not easily defined. It has been proposed that rather than
considering seizure frequency, a physician should consider how
significantly the seizures interfere with a patients quality-of-life
(Jones and Andermann, 2000). More recently there is increasing
consensus that anyone who meets the new drug resistant epilepsy
definition (failed 2 AEDs) should be considered for a surgical evaluation (Kwan et al., 2010).
The objectives of this study were to provide estimates of the
standard criteria that are used for recruiting patients for epilepsy
surgery and to identify sources of heterogeneity between studies.
We hypothesized that criteria for selection changed over time and
that failure of fewer AED failures would be noted as a criterion in
more recent cohorts.

2. Material and methods

2.1. Literature search strategy
A systematic review was conducted as part of a larger project
(CASES, the Canadian Appropriateness and necessity Study of
Epilepsy Surgery) (www.epilepsycases.com) on the development
of an appropriateness and necessity rating tool to identify patients
with focal epilepsy who should be referred for an epilepsy surgery
evaluation (Jett et al., 2012). An extensive search using both
PubMed and EMBASE databases, bibliographies of reviews, original articles and book chapters were completed to identify English
language articles published between January 1965 and June 2008
inclusive (see Appendix A for full search strategy). The aim of the
search was to identify all of the focal resective surgical series published since 1965. A separate broader search was also carried out
to identify all review articles about epilepsy surgery (including systematic reviews) published for hand searching.

as bibliographies of book chapters and published reviews about

epilepsy surgery.
2.3. Data extraction and analysis
Two authors independently extracted data from the included
studies using a standardized form piloted by the authors on an
initial sample of articles. Data extracted included the following:
year of study, study type, country, sample size, age, and whether
the following eligibility criteria (or any additional ones) were used
as selection for surgery evaluation, and how they were defined:
epilepsy duration, frequency of seizures, disabling seizures, resistance to AEDs, duration of AED failure, number of AEDs failed,
intolerable side effects to AED, absence of developmental delay,
absence of psychiatric conditions, absence of serious medical conditions, absence of progressive neurological conditions, prior surgery,
EEG criteria, neuroimaging criteria.
To further characterize pre-surgical criteria and to examine
groupings for these criteria, the studies were divided into prospective (Supplementary Table 1) and retrospective (Supplementary
Table 2) series based on when exposure status and outcomes
were ascertained. Check marks as seen in Supplementary Tables
1 and 2 indicate the criteria each study used. The aforementioned
categories were modified into the following columns: epilepsy
duration, seizure frequency, disabling seizures, AED resistance (not
further specified), AED trials completed, one AED failed, two or
more AEDs failed, duration of AED failure, side effects of medications, EEG, and MRI. If none of these criteria were utilized for
the study, the other criteria only column has been marked. If
drug resistant epilepsy was documented, the study was placed
into the AED resistance (not further specified) column. If AED trials were explicitly mentioned, but the required number of failed
drugs not specified, the study was placed in the AED trials completed column. The data were further examined pre- and post-year
2000 within the prospective data to see whether criteria to select
patients for surgery evolved over time (Table 1). The year 2000
was selected as it was around the time when results of the first
randomized controlled trial of focal resective epilepsy surgery
began to be presented at international meetings (Wiebe et al.,
2001). The years of data collection were used rather than year of
publication.
2.4. Statistical analysis
Descriptive statistics and statistical analysis of proportions were
performed. A p-value of less than 0.05 was pre-defined to indicate
a statistically significant result. A meta-analysis was not conducted
due to heterogeneity between studies.
3. Results

2.2. Study selection

3.1. Literature search
Abstracts were screened independently by two reviewers.
Inclusion criteria for the initial search were English language,
focal epilepsy or epilepsy with partial seizures (lesionectomy,
lobectomy, corticectomy, selective amygdalo-hippocampectomy).
Initial exclusion criteria were neonatal studies, non-resective
surgery, palliative procedures, stimulation studies and studies with
<20 patients. We also required that articles stated clear inclusion
or exclusion criteria in the study methods. All abstracts meeting
the above eligibility criteria were then reviewed as full text articles
again by two independent reviewers, to determine final eligibility
for data extraction. Bibliographies of all full text articles meeting
final eligibility criteria for data extraction were screened by two
reviewers, to make sure no additional studies were missed, as well

Of the 5061 articles identified from the CASES literature review,

572 articles were either surgical series or review articles on indications or contraindications for surgery evaluation selection. Review
articles without original data were excluded, leaving 489 surgical series reviewed in full text and 384 articles meeting all
eligibility criteria for data abstraction (Supplementary references)
(Fig. 1).
3.2. Study design and population
Study design types were as follows: 73 (19.0%) were prospective, 311 (81.0%) were retrospective and 2 (0.5%) were randomized

39

C.-S. Kwon et al. / Epilepsy Research 122 (2016) 3743

Table 1
Pre- and post-year 2000 prospective and retrospective data examining whether criteria to select patients for surgery evolved over time.
Prospective

Seizure frequency

One seizure per month

Five seizure per year
Frequent seizures
Not specified

Total
Epilepsy duration

>1 year
>1 year to 2 years
>2 years
Longstanding epilepsy
Not specified

Total
Duration of AED Failure

Used as criteria
Not specified

Total
Number of AEDs failed

1 AED
2 AEDs
3 AEDs
Multiple trials
Refractory
No mention

Total
AED resistance

Resistance
No mention

Total
Other criteria

Neuroimaging
EEG criteria
Prior surgery
Disabling seizures
Intolerable side effects
Psychiatric conditions
Developmental delay
Medical condition
Progressive neurological

Retrospective

Pre-2000, n (%)

Post-2000, n (%)

n (%)

5 (10)
1 (2)
0 (0)
46 (88)

8 (38)
0 (0)
1 (5)
12 (57)

16 (5)
3 (1)
6 (2)
286 (92)

52 (100)

21 (100)

311 (100)

0 (0)
3 (6)
5 (10)
0 (0)
44 (84)

2 (10)
1 (5)
3 (14)
0 (0)
15 (71)

2 (1)
12 (4)
14 (5)
11 (4)
272 (86)

52 (100)

21 (100)

311 (100)

8 (16)
44 (84)

6 (29)
15 (71)

5 (2)
306 (98)

52 (100)

21 (100)

311 (100)

0 (0)
4 (8)
6 (12)
1 (2)
34 (65)
7 (13)

0 (0)
9 (43)
1 (5)
0 (0)
7 (33)
4 (19)

1 (0)
19 (6)
7 (2)
11 (4)
202 (65)
71 (23)

52 (100)

21 (100)

311 (100)

45 (87)
7 (13)

17 (81)
4 (19)

240 (77)
71 (23)

52 (100)

21 (100)

311 (100)

14 (27)
16 (31)
2 (4)
3 (6)
4 (8)
8 (15)
14 (27)
0 (0)
0 (0)

4 (19)
7 (33)
5 (24)
0 (0)
0 (0)
2 (10)
8 (38)
2 (10)
2 (10)

99 (32)
69 (22)
28 (9)
21 (7)
3 (1)
17 (5)
36 (12)
3 (1)
6 (2)

Note: The values in the table reflect numbers (percentages) of studies in which selection criteria were stated. This does not reflect the proportion of surgical patients fulfilling
selection criteria, as these values were not able to be determined from the data.

controlled trials. Of all the studies included, 159 (41.4%) included

only adults, 59 (15.4%) included only children and 149 (38.8%)
included all age groups. 17 studies (4.4%) did not specify age
group.
3.3. Selection criteria
The following variables were used in the studies as selection
criteria for surgery.
3.3.1. Seizure frequency
Forty (10.4%) studies required a specific seizure frequency as an
inclusion criterion. Twenty-nine (7.6%) studies used 1 seizure per
month as an inclusion criterion. Thirteen out of the twenty-nine
were prospective studies. Four studies (1.0%) cited a minimum of
5 seizures/year, of which 1 was prospective, while 7 studies (1.8%)
(including 1 prospective study) stated frequent seizures as a criterion for surgery but did not stipulate a specific seizure frequency.
Overall, 344 (89.6%) of studies did not specify seizure frequency as
a selection criterion. Looking purely at the prospective studies that
mentioned a specific seizure frequency as selection criteria, 6 studies (11.5%) were pre-year 2000 and 9 (42.9%) were post-year 2000
(11.5% vs. 42.9%, p = 0.003).

3.3.2. Epilepsy duration

Four studies (1.0%) (2 prospective) specified an epilepsy duration of >1 year; 16 (4.2%) studies (4 prospective) used >1 year to 2
years; and 22 (5.7%) studies (8 prospective) used >2 years. Eleven
retrospective studies (2.9%) said longstanding epilepsy but without
specifying an actual duration. The majority of studies (331 studies;
86.2%) did not use epilepsy duration as selection criteria with 59
of 331 being prospective. Looking purely at the prospective studies
that mentioned a specific epilepsy duration as selection criteria,
8 studies (15.4%) were pre-year 2000 and 6 studies (28.6%) were
post-year 2000 (15.4% vs. 28.6%, p = 0.20).
3.3.3. Duration of AED failure
Duration of AED failure was used as a selection criterion in nineteen (4.9%) studies whereas the vast majority 365 (95.1%) did not
specify this criterion in patient selection. Fourteen out of the nineteen studies were prospective of which 8 studies (15.4%) were
pre-year 2000 and 6 studies (28.6%) were post-year 2000 (15.4%
vs. 28.6%, p = 0.20).
3.3.4. Number of AEDs failed
A single retrospective study (0.3%) with study cohort pre2000 required that patients had failed 1 AED. Thirty-two (8.3%)

40

C.-S. Kwon et al. / Epilepsy Research 122 (2016) 3743

Fig. 1. Literature search.

studies (13 prospective) required failure of 2 AEDs. Fourteen

(3.6%) studies (7 prospective) required failure of 3 AEDs. Twelve
(3.1%) studies (1 prospective) simply required failure of multiple
AEDs or an adequate trial of AEDs without further specification.
Most studies (243 (63.3%), 41 prospective) did not use these criteria
in their selection process and only mentioned that the candidates
had to be refractory to medication, without being more specific.
Thus there were 302 (78.6%) studies that mentioned drug resistance as an indication for surgery in their study. Eight-two studies
(11 prospective) did not mention AEDs at all (21.4%). Looking purely
at the prospective studies that mentioned a specific number of AEDs
failed as selection criteria, 10 (19.2%) were pre-year 2000 and 10
(47.6%) were post-year 2000 (19.2% vs. 47.6%, p = 0.01). The most
notable change over time (pre-2000 vs. post-2000) was failure of
2 AEDs (8% vs. 43% respectively, p < 0.001). Sixty percent of preyear 2000 prospective studies that mentioned a specific number of
AEDs failed required candidates to have failed 3 AEDs whereas
90% of studies post year 2000 required failure of only 2 or more
AEDs (Table 1).

3.3.5. Other criteria

Other criteria used included a particular neuroimaging finding
(or non-lesional) in 119 (31.0%) studies, particular EEG criteria in 92
(24.0%) studies. Disabling seizures (24 studies, 6.3%) and intolerable
side effects to medications (6 studies, 1.7%) were listed as indications for selection for evaluation for surgery in some studies, while
the following were listed as contraindications for an evaluation for
surgery in other studies: psychiatric conditions (27 studies, 7.0%),
developmental delay (58 studies, 15.1%), serious medical conditions (5 studies, 1.3%), and progressive neurological conditions (8
studies, 2.1%). Comparing the pre- vs. post-year 2000 prospective
studies, no significance differences were observed in the use the
specific inclusion/exclusion criteria: neuroimaging (27% vs. 19%,
p = 0.56), EEG criteria (31% vs. 33%, p = 0.83), disabling seizures (6%
vs. 0%, p = 0.55), intolerable SE (8% vs. 0%, p = 0.32), psychiatric conditions (15% vs. 10%, p = 0.71), developmental delay (27% vs. 38%,

p = 0.71), medical conditions (0% vs. 10%, p = 0.08), progressive neurological conditions (0% vs. 10%, p = 0.08). Prior surgery was more
likely to be an exclusion criteria post-2000 (4% vs. 24%, p = 0.02)
(Table 1).
3.4. Selection criteria grouping
Examining inclusion criteria across studies we see that in the
prospective studies 37 out of the 73 (51%) had just one pre-surgical
criterion vs. 170 out of the 311 (55%) in the retrospective studies
(p = 0.63). Comparing how many criteria were used in prospective
vs. retrospective studies we can see that there were more prospective studies that included 3 criteria: prospective studies 22/73
(30%) vs. retrospective studies 48/311 (15%) (p < 0.01).
Supplementary Table 3 presents a detailed account of the
prospective studies and their inclusion criteria. As there were only
two randomized trials at the time of our review, the criteria utilized for these studies were specifically examined (Lutz et al., 2004;
Wiebe et al., 2001). Lutz et al. (2004) required only the following indication for inclusion (based on our indications of interest):
MRI data (presence of hippocampal sclerosis or gliosis) (Lutz et al.,
2004). They also listed the following criteria: age >16, IQ > 60, and
availability of complete pre- and post-operative neuropsychological results. The more comprehensive paper by Wiebe et al. (2001)
included the following as selection for evaluation: epilepsy duration (more than one year), seizure frequency (monthly seizures on
average during the preceding year), failure of two or more AEDs
(one of which was phenytoin, carbamazepine or valproic acid), EEG,
and MRI (along with age >16) (Wiebe et al., 2001). Supplementary
Table 4 presents a detailed account of excluded articles that did not
mention inclusion and exclusion criteria.
4. Discussion
In this systematic review of published surgical series, we found
that the two criteria used most commonly to define surgical

C.-S. Kwon et al. / Epilepsy Research 122 (2016) 3743

candidacy are (1) seizure frequency and (2) failure of AED treatment. Unfortunately, the degree of heterogeneity among published
epilepsy surgical series demonstrates existence of diverse benchmarks, poverty in defining eligibility criteria and dissimilar patient
selection for surgery across centers. Taken together, it is difficult to compare results and develop recommendations. This in
part is a reflection of the fact that 80% of the surgical series
are retrospective in their design but also that criteria used to
defined surgical eligibility often lack standardized widely accepted
definitions.
To see whether criteria to select patients for surgery have
evolved over time, we examined studies pre-year 2000 vs. postyear 2000. Concentrating on the prospective data we found that
a higher proportion of centers over time reported criteria of
seizure frequency (pre-year 2000: 11.5% vs. post-year 2000: 42.9%,
p = 0.003) and number of AEDs failed (pre-year 2000: 19.2% vs.
post-year 2000: 47.6%, p = 0.01). The most important and rather disquieting outcome of the analysis was that, for the vast majority of
studies inclusion criteria were incompletely documented for presurgical evaluation. However, it was encouraging to see that of all
the studies documenting a specific number of AEDs that had to be
failed to be considered for surgery, 90% of prospective studies after
2000 only required failure of 2 or more AEDs, consistent with the
new definition of drug resistant epilepsy (Kwan et al., 2010). This
is in contrast to studies before 2000 where 60% studies required
failure of 3 or more AEDs when a number of AED was specified as
a criterion.
The notion of failure of AEDs seems self-explanatory, but as
seen in this systematic review, it remains relatively unintuitive,
and a pragmatic operational definition of AED failure was lacking
until recently (Kwan et al., 2010). Fortunately the ILAE Commission
on Classification and Terminology published in 2009 a consensus
definition of drug resistant epilepsy with the aim of articulating
a mutual language among practitioners: drug resistant epilepsy
may be defined as failure of adequate trials of two tolerated
and appropriately chosen and used AED schedules (whether as
monotherapies or in combination) to achieve sustained seizure
freedom (Kwan et al., 2010). This common clinical framework was
developed in order to aid physicians to recognize as well as facilitate prompt referral to epilepsy specialists for those who satisfy
the definition. While this consensus definition may still need to
be refined, it provides a suitable platform to be assessed in future
prospective research.
When identifying possible epilepsy surgery candidates, it is
important not only to keep in mind the concept of drug resistance, but to also understand the course of epilepsy. It is well
known when treating those with epilepsy, that approximately
one third of patients remain drug resistant (Kwan and Brodie,
2000); however, this proportion is much higher in those with
focal epilepsy, in whom more than 70% may be drug resistant
(Gilioli et al., 2012; Stephen et al., 2001). It has been seen in
well-conducted studies that those who go into remission with
seizure medication often later relapse. A recent study in drug
resistant children found that 57% of these children go through
at least one 12-month remission interval, with 68% eventually
relapsing; repeated remissions and relapses were a common occurrence (Berg et al., 2009). A recent study in adults reported a
yearly 5% cumulative probability of entering seizure remission in
drug resistant individuals. However the risk of relapse in these
individuals was 71.2% (Callaghan et al., 2011). As such, drug
resistant individuals are not likely to become seizure-free from
future drug trials and should have a prompt surgical evaluation.
The ideal approach to the management of epilepsy is to have
well-controlled seizures without new or disabling side effects. Few

41

disabling seizures despite adequate AED treatment would justify

a patient to be evaluated for surgery; however, this study found
that most studies had no documentation of disabling seizures.
Only 24 (6.3%) studies mentioned this criterion, of which 21 were
retrospective and the three prospective studies were all published prior to 2000. Treatment initiation should be performed
with haste before unfavorable outcomes from the result of continued seizures ensue. Medical therapy however can often be
accompanied by disabling side effects, which can reduce an individuals quality of life (QOL) (Gilliam, 2002). Surgery should always
be considered a possible option when adequate medication trials do not stop seizures, particularly disabling seizures or if the
epilepsy is disabling and should also be considered if medication side effects are significant (Engel et al., 2003). Once again,
standardized widely accepted definitions for the terms disabling
seizures or disabling epilepsy do not exist and these terms
are often used interchangeably when defining surgical candidacy.
In the surgical series reviewed for this systematic review, disabling epilepsy was generally defined as: drug resistant epilepsy
that significantly interferes with the patients ability to function,
significantly impairs QOL, or threatens the future cognitive or
physical state of the patient or inadequate control by seizure medication. Successful surgery can eradicate the disabling aspects of
seizures and epilepsy in the majority of carefully selected individuals with surgically remediable epilepsy syndromes (Engel,
1996).
Our systematic analysis shows that duration of epilepsy was
only documented in 53 of 384 studies, of which only 14 were
prospective studies. Further to what we have already witnessed, a
disquieting facet of the analysis demonstrates again the paucity of
documentation. Selection for evaluation for epilepsy surgery such
as epilepsy duration or seizure frequency (e.g. monthly seizures)
are no longer appropriate for epilepsy surgery, as anyone with
ongoing seizures should be promptly referred to an epilepsy program to investigate whether they are suitable candidates for
epilepsy surgery (Haneef et al., 2010). Indeed, early surgical therapy
for drug resistant temporal lobe epilepsy was recently found to be
superior to ongoing medical management with none of the participants in the medical group vs. 73% of participants in the surgical
group seizure free during the two year of follow up (Engel et al.,
2012).
Past research has cited uncertainties in eligibility criteria as
a reason for the under-utilization of epilepsy surgery (Kumlien
and Mattsson, 2010). A web-based clinical decision tool (www.
epilepsycases.com) was recently developed using the RAND/UCLA
appropriateness methodology, to guide physicians in identifying
patients who are candidates for an epilepsy surgery evaluation
(Jett et al., 2012). This tool was developed based on a minimum set
of variables which should be available as part of the basic epilepsy
workup (seizure frequency, epilepsy duration, seizure type, number of AEDs failed, AED side effects, EEG and MRI results). However,
the authors of the tool still recommend that all those with drug
resistant epilepsy, regardless of the tool, be referred for an epilepsy
surgery evaluation.
The most significant limitation to this systematic review is
the extensive heterogeneity noted across different study populations and about how each eligibility criterion was defined making
any sophisticated analysis (e.g. meta-analyses) impossible. Most
of the studies in this review were retrospective and thus ascertainment of well-defined criteria are challenging. Additionally, the
inclusion and exclusion criteria engaged in most retrospective
studies many years earlier may not be recalled accurately by the
authors of series reports, and thus will be influenced by their current practices. The studies we identified mostly included patients
who only proceeded to surgery. A major limitation is the lack of

42

C.-S. Kwon et al. / Epilepsy Research 122 (2016) 3743

information on those who were deemed not to be candidates due

to high risk of surgery (e.g. a cavernoma in the central sulcus in
a young patient who works as a carpenter or a lesion in one of
the occipital lobes in a teacher, etc.). More studies are required
examining surgical candidacy for all those evaluated and not just
those who proceed to surgery after a long decision making process. Furthermore, of the 386 final included articles, there were
only two randomized trials, with only one being a randomized
double blind controlled trial with clearly defined inclusion and
exclusion criteria (Wiebe et al., 2001). Finally, the original search
included only articles in English although the two main databases
(EMBASE and PubMed) were used, and book chapters about
epilepsy surgery as well as published reviews on the topic were
also hand searched to ensure no additional surgical series were
missed.

5. Conclusions
The two main criteria that were used to define surgical
candidacy in the published series reviewed are: seizure frequency and failure to respond to AEDs. However, the terms
are loosely defined and display significant intra-study variation.
Overall, having one seizure per month and failure of two or
more AEDs were the commonest specific criteria. Yet, seizure
frequency should not be one of the sole determinants of surgical candidacy, as infrequent seizures may still pose significant
risks and impair quality-of-life. Important variations between
studies make it difficult to identify consistent objective criteria
to determine surgical candidacy. Moreover, this review provides convincing data on the often incomplete data on criteria
for pre-surgical evaluation. Development of accepted standard
criteria for surgical candidacy including clear definitions of eligibility criteria are needed in order to ensure timely referrals to
epilepsy centers and to allow for comparison of outcomes between
centers.

Acknowledgements
Dr. Nathalie Jett holds a salary award from Alberta Innovates
Health Solutions (AIHS) and a Canada Research Chair in Neurological Health Services Research. Dr. Jose F. Tellz-Zenteno holds
salary award from the Royal University Hospital Foundation in
Saskatoon trough the Mudjadik Thyssen Mining Professorship in
Neurosciences and also receives operating funds from the University of Saskatchewan. Dr. Samuel Wiebe holds the Hopewell
Professorship of Clinical Neurosciences Research at the University of Calgary. This study was funded by AIHS and the Canadian
Institutes of Health Research.

Appendix A. Full search strategy

1. epilep*.mp.
2. (surger* or surgi*).mp. [mp = title, abstract, subject headings,
heading word, drug trade name, original title, device manufacturer, drug manufacturer name]
3. (epilepsy and surgery).mp. [mp = title, abstract, subject headings,
heading word, drug trade name, original title, device manufacturer, drug manufacturer name]
4. (incidence or mortality or follow-up studies or prognosis or
prognos* or predict* or course or outcome or psychology or quality of life or memory or survival analysis or seizure* or utilization
or cost or efficacy or complications or effectiveness or sudep).mp.
[mp = title, abstract, subject headings, heading word, drug trade

5.
6.
7.

7
8.

name, original title, device manufacturer, drug manufacturer

name]
(1 and 2) or 3
4 and 5
(randomized controlled trial or random* or (double and blind)
or placebo or drug therapy or therapeutic or cohort studies or
risk or (odds and ratio) or (relative and risk) or case control or
casecontrol studies or clinical trial or random allocation or case
series or decision analysis or economic).mp. [mp = title, abstract,
subject headings, heading word, drug trade name, original title,
device manufacturer, drug manufacturer name]
6 and 7
limit 8 to (human and English)

Appendix B. Supplementary data

Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.eplepsyres.2016.
02.007.
References
Berg, A.T., Levy, S.R., Testa, F.M., DSouza, R., 2009. Remission of epilepsy after two
drug failures in children: a prospective study. Ann. Neurol. 65,
510519.
Callaghan, B., Schlesinger, M., Rodemer, W., Pollard, J., Hesdorffer, D., Hauser, W.A.,
French, J., 2011. Remission and relapse in a drug-resistant epilepsy population
followed prospectively. Epilepsia 52, 619626.
Cross, J.H., Jayakar, P., Nordli, D., Delalande, O., Duchowny, M., Wieser, H.G.,
Guerrini, R., Mathern, G.W., International League against Epilepsy,
Subcommission for Paediatric Epilepsy Surgery, Commissions of Neurosurgery
and Paediatrics, 2006. Proposed criteria for referral and evaluation of children
for epilepsy surgery: recommendations of the Subcommission for Pediatric
Epilepsy Surgery. Epilepsia 47, 952959.
Duncan, J.S., 2011. Selecting patients for epilepsy surgery: synthesis of data.
Epilepsy Behav. 20, 230232.
Engel Jr., J., Shewmon, D., 1993. Who should be considered a surgical candidate?
In: Engel Jr., J. (Ed.), Surgical Treatment of the Epilepsies. , 2nd ed. Raven Press,
New York, pp. 2334.
Engel Jr., J., 1996. Current concepts: surgery for seizures. N. Engl. J. Med. 334,
647652.
Engel Jr., J., Wiebe, S., French, J., Sperling, M., Williamson, P., Spencer, D., Gumnit,
R., Zahn, C., Westbrook, E., Enos, B., Quality Standards Subcommittee of the
American Academy of Neurology, American Epilepsy Society, American
Association of Neurological Surgeons, 2003. Practice parameter: temporal lobe
and localized neocortical resections for epilepsy: report of the Quality
Standards Subcommittee of the American Academy of Neurology, in
association with the American Epilepsy Society and the American Association
of Neurological Surgeons. Neurology 60, 538547.
Engel Jr., J., McDermott, M.P., Wiebe, S., Langfitt, J.T., Stern, J.M., Dewar, S., Sperling,
M.R., Gardiner, I., Erba, G., Fried, I., Jacobs, M., Vinters, H.V., Mintzer, S.,
Kieburtz, K., Early Randomized Surgical Epilepsy Trial (ERSET) Study Group,
2012. Early surgical therapy for drug-resistant temporal lobe epilepsy: a
randomized trial. JAMA 307, 922930.
Gilioli, I., Vignoli, A., Visani, E., Casazza, M., Canafoglia, L., Chiesa, V., Gardella, E., La
Briola, F., Panzica, F., Avanzini, G., Canevini, M.P., Franceschetti, S., Binelli, S.,
2012. Focal epilepsies in adult patients attending two epilepsy centers:
classification of drug-resistance, assessment of risk factors, and usefulness of
new antiepileptic drugs. Epilepsia 53, 733740.
Gilliam, F., 2002. Optimizing health outcomes in active epilepsy. Neurology 58,
S9S20.
Haneef, Z., Stern, J., Dewar, S., Engel, J., 2010. Referral pattern for epilepsy surgery
after evidence-based recommendations. A retrospective study. Neurology 75,
699704.
Jett, N., Quan, H., Tellez-Zenteno, J.F., Macrodimitris, S., Hader, W.J., Sherman,
E.M.S., Hamiwka, L.D., Wirrell, E.C., Burneo, J.G., Metcalfe, A., Faris, P.D.,
Hernandez-Ronquillo, L., Kwon, C.S., Kirk, A., Wiebe, S., the CASES expert
panelists, 2012. Development of an online tool to determine the
appropriateness for an epilepsy surgery evaluation. Neurology 79, 10841093.
Jones, M.W., Andermann, F., 2000. Temporal lobe epilepsy surgery: definition of
candidacy. Can. J. Neurol. Sci. 27, S11S13.
Kumlien, E., Mattsson, P., 2010. Attitudes towards epilepsy surgery: a nationwide
survey among Swedish neurologists. Seizure 19, 253255.
Kwan, P., Brodie, M.J., 2000. Early identification of refractory epilepsy. N. Engl. J.
Med. 342, 314319.
Kwan, P., Arzimanoglou, A., Berg, A.T., Brodie, M.J., Hauser, W.A., Mathern, G.,
Mosh, S.L., Perucca, E., Wiebe, S., French, J., 2010. Definition of drug resistant
epilepsy: consensus proposal by the ad hoc Task Force of the ILAE Commission
on Therapeutic Strategies. Epilepsia 51, 10691077.

C.-S. Kwon et al. / Epilepsy Research 122 (2016) 3743

Labiner, D.M., Bagic, A.I., Herman, S.T., Fountain, N.B., Walczak, T.S., Gumnit, R.J.,
National Association of Epilepsy Centers, 2010. Essential services, personnel,
and facilities in specialized epilepsy centers revised 2010 guidelines.
Epilepsia 51, 23222333.
Lutz, M.T., Clusmann, H., Elger, C.E., Schramm, J., Helmstaedter, C., 2004.
Neuropsychological outcome after selective amygdalohippocampectomy with
transsylvian versus transcortical approach: a randomized prospective clinical
trial of surgery for temporal lobe epilepsy. Epilepsia 45, 809816.

Stephen, L.J., Kwan, P., Brodie, M.J., 2001. Does the cause of localisation-related
epilepsy influence the response to antiepileptic. Epilepsia 42,
357362.
Wiebe, S., Blume, W.T., Girvin, J.P., Eliasziw, M., 2001. A randomized, controlled
trial of surgery for temporal-lobe epilepsy. N. Engl. J. Med. 345,
311318.

43