Management of antiepileptic drugs following epilepsy surgery: A meta-analysis

Epilepsy Research (2014) 108, 765—774

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

Management of antiepileptic drugs following epilepsy surgery: A meta-analysis Lady D. Ladino a,b, Lizbeth Hernández-Ronquillo a, José F. Téllez-Zenteno a,∗ a b

Department of Medicine, Division of Neurology, University of Saskatchewan, Saskatoon, SK, Canada Neurology Section, Hospital Pablo Tobón Uribe, Medellín, Colombia

Received 17 November 2013; accepted 26 January 2014 Available online 15 February 2014

KEYWORDS Epilepsy surgery; Discontinuation; Drug-resistant; AEDs; Seizure recurrence; Withdrawal

Summary Objective: No consensus exists regarding the management of antiepileptic drugs (AEDs) after successful epilepsy surgery (ES). We performed a meta-analysis with the most relevant evidence in this topic. Our aim was to provide evidence-based estimates of results on AEDs discontinuation after ES. Methods: We searched MEDLINE and Embase using Medical Subject Headings and keywords related to AEDs discontinuation after ES. Two reviewers independently applied the following inclusion criteria: original published research that directly compared seizure outcomes in patients having or not AEDs discontinuation after ES. Two investigators independently extracted data, resolving disagreements through discussion. A random and fixed-effect model was used to derive a pooled odds ratio (OR) for either seizure recurrence in both groups. Results: Of 257 abstracts initially identified by the search, 57 were reviewed as full text. Sixteen articles fulfilled eligibility criteria and described outcomes in 1456 patients with AEDs discontinuation and 685 patients with no discontinuation. The odds of having seizure recurrence after AEDs discontinuation was 0.39 times lower in patients with attempted discontinuation after surgery (OR 0.39, CI 95% 0.300—0.507, p < 0.001). Most likely the difference is related with a selected population where discontinuation was attempted. Significance: Seizure recurrence was higher for patients without AED modification than for the withdrawal group. Patients with seizure recurrence after discontinuation can be managed easily after re-start of medications. The discontinuation of medications should be done in good candidates and the decision should be individualized taking into account clinical, electrographical, imaging and histopathological variables. © 2014 Elsevier B.V. All rights reserved.

∗ Corresponding author at: Department of Medicine, Division of Neurology Royal University Hospital, Saskatoon, SK S7N 0W8, Canada. Tel.: +1 306 9668011. E-mail addresses: [email protected] (L.D. Ladino), [email protected] (L. Hernández-Ronquillo), [email protected], [email protected] (J.F. Téllez-Zenteno).

0920-1211/$ — see front matter © 2014 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.eplepsyres.2014.01.024

766

Introduction The safety and efficacy of epilepsy surgery (ES) for temporal lobe epilepsy (TLE) has been well established in two randomized clinical trials (RCTs) (Engel et al., 2012; Wiebe et al., 2001), as well as their sustained beneficial effects in the long-term, such as prolonged seizure freedom (Téllez-Zenteno et al., 2005), lower mortality, improved psychosocial and memory outcomes (Téllez-Zenteno et al., 2007), and improved quality of life (Mohammed et al., 2012). Potential reduction and eventual withdrawal of antiepileptic drugs (AEDs) is one of the most difficult therapeutic challenges after successful ES. There are legitimate reasons to stop AEDs after ES, including avoiding undesirable long-term toxicity, to reduce cognitive adverse effects of AEDs, to reduce costs and to remove daily treatment that serves as a major affirmation of the sick role in patients (Cole and Wiebe, 2008). Compared with other surgical outcomes such as the seizure outcome, few studies have been published regarding the management of AEDs after ES and current recommendations are based largely on local experience at different centers. The selection of candidates for AEDs withdrawal is complex and we do not have uniform criteria across epileptologists and epilepsy centers. According to three medical surveys (Berg et al., 2007; Swisher and Sinha, 2013; Téllez-Zenteno et al., 2012) of clinical practice regarding AEDs withdrawal, the majority of US and Canadian epileptologists and neurologists prefer to wait between six months to two years before any change in medication. AEDs levels, electroencephalogram (EEG), and a magnetic resonance imaging (MRI), are typically done before stopping AEDs (Berg et al., 2007; Swisher and Sinha, 2013; Téllez-Zenteno et al., 2012). The most important factors considered by epileptologists in North America were the following: focal pathology, complete postoperative seizure freedom, complete resection of a well-defined epileptic lesion, lack of postoperative epileptiform discharges on EEG, a patient’s desire to stop medications and a temporal localization for the surgery (Berg et al., 2007; Téllez-Zenteno et al., 2012). Although successful AEDs withdrawal has been associated with improvements in scales of general satisfaction and quality of life, some studies have shown controversial information. We performed a systematic review and meta-analysis of the evidence comparing seizure recurrence in patients with and without AEDs discontinuation after successful ES. Our aim was to provide clinicians with a scientifically valid and coherent summary of the best current evidence, and to provide a best estimate of seizure recurrence rates in patients with and without AEDs discontinuation after ES.

Methods Data source A medical librarian performed a comprehensive literature search of the Medline® , Embase® , Index Medicus® , and Cochrane databases from January 1980 to July 2013 that incorporated Medical Subject Headings and text

L.D. Ladino et al. words for literature on the management of AEDs after ES (Literature search strategy in Appendix A). We also searched bibliographies of reviews, original articles and book chapters, and consulted experts about other studies. We included studies if they contained original research involving patients, irrespective of age and regardless of language or country of origin.

Study selection and classification Two reviewers independently applied the following study inclusion criteria: (a) Original published research with ≥30 patients of any age undergoing resective ES; (b) reports that directly compared seizure outcome in patients having or not AEDs discontinuation after successful ES; (c) description of number of patients having each intervention; (d) quantitative description of seizure outcomes in patients with and without AED discontinuation, thus allowing for direct comparisons, and (e) seizure outcomes reported after at least one year of follow-up since surgery. We considered all outcomes in children and adults. Children were considered less than 16 years old. We excluded studies with not consistent data or any overlapping patient populations from the same center. We selected the most recent publication for inclusion if studies were duplicate reports from the same population. For this study we used the term ‘‘controlled studies’’ to describe studies where a comparison was done among patients where discontinuation was attempted after ES vs. not attempted. The outcomes explored included: postoperative seizure outcome using the Engel’s classification; time to start the discontinuation of AEDs; time to achieve discontinuation; seizure recurrence in both groups, and seizure freedom rate after the re-start of medications.

Data gathering and extraction We reviewed full texts in duplicate and selected those that met our inclusion criteria for meta-analysis. Data extracted from eligible studies included year of publication, country in which the study was conducted, study type (e.g., controlled/non-controlled), study design (e.g., case and control, cohort, etc.), number of participants and timing and method of discontinuation. Demographic data included sex distribution, mean age at seizure onset, mean age at ES, epilepsy etiology, time interval from surgery to start of AED discontinuation (ten months was selected as a cutoff for ‘‘early’’ vs. ‘‘late’’ drug tapering), time to achieve discontinuation and duration of total follow-up. Surgical data included Engel’s classification of seizure outcome, surgical resection type, and area of resection. Two reviewers independently abstracted all data, resolving disagreements through discussion and included a senior author where necessary. An attempt was made to assess risk factors for seizure recurrence in some studies (Al-Kaylani et al., 2007; Berg et al., 2006; Boshuisen et al., 2012; Hoppe et al., 2006; Kuzniecky et al., 1992; Lachhwani et al., 2008; Lee et al., 2008; Menon et al., 2012; Murro et al., 1991; Park et al., 2010; Rathore et al., 2011; Schiller et al., 2000).

Management of antiepileptic drugs following epilepsy surgery Articles identified through Medline and Embase N = 257

Eligible articles N = 57 (22%) Full text review

Excluded articles N=32 (56%) Other outcome Review articles Duplicate publication Case report Editorials Basic science study

13 (41%) 5 (16%) 4 (12%) 4 (12%) 4 (12%) 2 (6%)

Included articles N=25 (44%) Controlled studies Non-controlled studies

16 (64%) 9 (36%)

Figure 1 Flow diagram of the literature search and study identification for the meta-analysis.

Analysis Studies were the unit of analysis. We used Comprehensive Meta Analysis V2® for all meta-analyses, and Statistics with Confidence® to obtain exact 95% confidence intervals (95% CI) around proportions. We assessed the data for heterogeneity by visual analysis of the forest plots, and quantitatively using the Q statistic. The proportions of seizure recurrence in each category were pooled using the random and fixed effects model. We used odds ratio (OR) to compare pooled proportions of seizure recurrence in patients with AEDs discontinuation vs. not.

Results Evidence base The literature search yielded 257 citations, of which 57 (22%) were potentially eligible and were examined in full text independently by two reviewers (Fig. 1). Twenty-five studies met our inclusion criteria for the meta-analysis. Twelve (48%) studies described TLE surgery patients, 12 (48%) temporal and extratemporal surgical cases and one (4%) described only extratemporal lobe epilepsy surgical patients. Overall, 90% (95% CI 89.6, 90.6) of the surgical cases were TLE cases. Ten studies (40%) were in adults, six (24%) in children, and nine (36%) in mixed population. These 25 studies included a total of 3586 patients, 2901 (81%) patients had AED discontinuation after ES and 685 (19%) patients did not have any change on AEDs after ES.

767 Overall, 52% of patients were males. Age at operation ranged from 7 to 49 years. Sixteen studies documented the preoperative duration of epilepsy, which ranged from 5 to 27 years. Fifteen studies documented the results of MRI (used to define the presence of a lesion), and 85% of patients had lesional epilepsy demonstrated by imaging. Five studies evaluated the extent of the surgical resection extension by MRI, 81% of patients had a complete resection. Only eleven studies documented histopathology results, of these 32% reported hippocampal sclerosis (HS). We identified 16 (64%) controlled studies (Table 1), and nine (36%) non-controlled studies (Table 2).

Controlled studies characteristics Five (31%) of the 16 controlled studies collected data prospectively; the rest of studies collected data retrospectively (69%). There was one prospective RCT of AED (Kuzniecky et al., 1992). All studies were limited to singlecenter institutions an only one study combined data from seven centers in the United States (Berg et al., 2006). Overall eight countries from four continents were represented. These controlled studies included 2141 patients, 1456 (68%) patients discontinued AEDs and 685 (32%) patients did not have changes in AEDs after surgery.

Discontinuation method The timing of medication withdrawal was mainly influenced by patient preference according the authors of articles. Commonly AEDs were tapered using an individualized schedule for each patient. In only eight (32%) of 25 studies a specific protocol for AED reduction was specified (Al-Kaylani et al., 2007; Kerling et al., 2009; Kuzniecky et al., 1992; Maehara and Ohno, 2011; McIntosh et al., 2004; Menon et al., 2012; Murro et al., 1991; Rathore et al., 2011). In general, three principles regarding tapering methods were described. First of all, for patients receiving multiple medications, each one was individually tapered completely prior to the subsequent medication. Second, the reduction in dose was done in small tapering steps at every two to three months except in patients taking three or more AEDs, who underwent a more rapid withdrawal of the third AED and finally, one of the primary drug (typically carbamazepine or phenytoin) was always withdrawn in the end. Finally the interaction between parents and doctor influenced the decision to start medication withdrawal and the time interval at which to do so. Twenty-two studies reported the time interval between surgery and AED withdrawal; the average was 14 (3—32) months. Fourteen studies reported the time to achieve withdrawal; the average was 30 (2—60) months. The general postoperative follow-up of all studies was 65 (12—148) months.

Seizure outcomes Overall, 73% of patients in all studies had discontinuation of AEDs (95% CI 70.9, 74.5) and the postoperative Engel class I was 98% (95% CI 97.6, 98.4). In the 16 controlled studies the seizure recurrence was 18% (95% CI 16, 19.6)

768 Table 1 Seizure outcome in controlled studies. In 16 studies with a total of 2141 patients, 98% had class I Engel after surgery. Overall in all the studies 73% of patients discontinued AEDs and 27% did not. Eighteen percent of patients from the discontinuation groups had seizure recurrence compared with 76% of patients without discontinuation. Study

Author/year

Kuzniecky et al. (1992) Schiller et al. (2000) Van Veelen et al. (2001) Griffin et al. (2004) McIntosh et al. (2004) Kim et al. (2005) Berg et al. (2006) Lachhwani et al. (2008) Tanriverdi et al. (2008) Kerling et al. (2009) Boshuisen et al. (2009) Park et al. (2010) Rathore et al. (2011) Das et al. (2012) Menon et al. (2012) Pimentel et al. (2012) Total

Total patients

Surgery cases

40 210 143 30 157 66 291 97 63 60 109 223 310 169 106 67 2141

Seizure recurrence comparative results

Population

Surgery location

POP Engel class

Mix Mix Mix Adult Adult Mix Mix Pediatric Adult Adult Pediatric Adult Adult Adult Mix Adult

Temporal Mix Mix Temporal Temporal Temporal Temporal Mix Temporal Mix Mix Mix Temporal Mix Extratemporal Temporal

30 (75) 210 (100) 143 (100) 27 (90) 122 (78) 66 (100) 291 (100) 97 (100) 52 (82) 60 (100) 85 (78) 121 (54) 258 (83) — 63 (59) 67 (100) 1692 (98)

AEDs discontinuation group

AEDs no discontinuation group

Patients with discontinuation N (%)

Patients with no discontinuation N (%)

20 180 108 22 83 60 129 68 20 34 84 147 258 111 94 38

Seizure recurrence N (%)

20 30 35 8 74

Seizure recurrence N (%)

(50) (86) (76) (73) (53) (91) (44) (70) (32) (57) (77) (66) (83) (66) (89) (57)

4 (20) 35 (19) 32 (30) 6 (27) 13 (16) 20 (33) 44 (32) 11 (16) 0 (0) 8 (24) 5 (6) 68 (46) 64 (25) 11 (10) 44 (47) 13 (34)

(50) (14) (24) (27) (47) 6 (9) 162 (56) 29 (30) 43 (68) 26 (43) 25 (23) 76 (34) 52 (17) 58 (34) 12 (11) 29 (43)

6 (30) 2 (7) 6 (17) 6 (75) 24 (32) 1 (17) 73 (45) 1 (3) 12 (28) 10 (39) 19 (76) 68 (83) 52 (100) 28 (48) 12 (100) 9 (31)

1456 (73)

378 (18)

685 (27)

329 (76)

L.D. Ladino et al.

Management of antiepileptic drugs following epilepsy surgery

769

Table 2 Seizure outcome in non-controlled studies. Nine studies with a total of 1445 patients, 98% had Engel class I after surgery and 100% underwent AED discontinuation, from these patients 18% suffered seizure recurrence, there was no a clear control group. Study

Patients features

Author/year

Surgery cases

Murro et al. (1991) Khan and Onar (2006) Hoppe et al. (2006) Al-Kaylani et al. (2007) Sinclair et al. (2007) Lee et al. (2008) Maehara and Ohno, 2011 Boshuisen et al. (2012) Zeng et al. (2012) Total

77 62 102 57 80 169 33 766 99

AEDs discontinuation

Population

Surgery location

Engel class I N (%)

Mix Pediatric Pediatric Adult Pediatric Adult Mix Pediatric Mix

Temporal NS Mix Temporal Mix Temporal Temporal Mix Mix

77 (100) 62 (100) 91 (89) 57 (100) 80 (100) 89 (52) 33 (100) 766 (100) 99 (75)

1445

in patients with AED discontinuation and 76% (95% CI 70.7, 74.5) in patients with no discontinuation (p < 0.001). From the controlled studies, only five (Kim et al., 2005; Lachhwani et al., 2008; Pimentel et al., 2012; Schiller et al., 2000; Van Veelen et al., 2001) reported a higher seizure recurrence in the withdrawal group, 24% (95% CI 17.7, 30.4) vs. 7% (95% CI 1.02, 12.4). In the nine non-controlled

1445 (98)

Discontinuation group N (%) 77 62 102 57 80 169 33 766 99

Seizure recurrence N (%)

(100) (100) (100) (100) (100) (100) (100) (100) (100)

1445 (100)

34 17 11 18 35 93

(44) (27) (11) (32) (44) (55) 3 (9) 93 (12) 26 (26) 330 (18)

studies the rate of seizure recurrence in the discontinuation group was 18% (95% CI 16, 19.7). Additionally, 75% (95% CI 72.4, 79) of patients who initially had AED withdrawal and seizure recurrence were seizure free again after AEDs were restarted. Thirty nine percent (95% CI 29.8, 48.7) of patients who never attempted AED withdrawal and had seizure recurrence regained seizure freedom. In general (in

Table 3 Seizure free status after restarts AEDs in both groups of patients, n = 21 studies. Seventy six percent of patients with AEDs discontinuation regained seizure freedom after restarted AEDs compared with 39% in the group without discontinuation. In general, 77% of patients regained seizure freedom after restarted AEDs. There is no complete information about each group in some studies. Study

Seizure outcome in both groups

Author/year

Discontinuation groupN (%)

Kuzniecky et al. (1992) Schiller et al. (2000) Van Veelen et al. (2001) Griffin (2004) Kim et al. (2005) Berg et al. (2006) Lachhwani et al. (2008) Kerling et al. (2009) Boshuisen et al. (2009) Park et al. (2010) Rathore et al. (2011) Menon et al. (2012) Pimentel et al. (2012) Khan and Onar (2006) Hoppe et al. (2006) Al-Kaylani et al. (2007) Sinclair et al. (2007) Lee et al. (2008) Maehara and Ohno (2011) Boshuisen et al. (2012) Zeng et al. (2012) Total

No discontinuation groupN (%)

General information (%)

— — 18 (56) — 14 (70) 26 (63) 7 (64) 4 (50) 2 (40) — 56 (88) 30 (68) — 15 (88) 7 (64) 14 (78) 33 (94) 20 (33) 3 (100) 61 (70) 22 (85)

2 (33) — — — — 37 (51) — 1 (10) — — — — — — — — — — — — —

2 (33) 20 (90) 18 (56) 3 (50) 14 (70) 63 (55) 7 (64) 5 (28) 2 (40) 36 (44) 56 (88) 30 (68) 12 (100) 15 (88) 7 (64) 14 (78) 33 (94) 20 (33) 3 (100) 61 (70) 22 (85)

332 (76)

40 (39)

443 (77)

770 Table 4

L.D. Ladino et al. Identified risk factors in studies for seizure recurrence after AED discontinuation, n = 15 studies.

Factor

Studies

Author/year

Seizure recurrence before withdrawal

5

Non-lesional epilepsy

4

Early withdrawal (<10 months)

4

Longer duration of epilepsy (≥11 years)

4

Older age at surgery (> 30 years)

4

Abnormal postoperative EEG

3

Absence of abnormal pathology or poorly circumscribed pathologies as gliosis Incomplete resection Frequent preoperative generalized seizures (more than two per year) Extratemporal surgery Multifocal MRI lesions

3 2 1

Berg et al. (2006), Hoppe et al. (2006), Park et al. (2010), Rathore et al. (2011), and Menon et al. (2012) Kuzniecky et al. (1992), Schiller et al. (2000), McIntosh et al. (2004), and Park et al. (2010) Lee et al. (2008), Lachhwani et al. (2008), Park et al. (2010), and Boshuisen et al. (2012) Murro et al. (1991), Park et al. (2010), Rathore et al. (2011), and Menon et al. (2012) Al-Kaylani et al. (2007), Lee et al. (2008), Rathore et al. (2011), and Menon et al. (2012) Rathore et al. (2011), Boshuisen et al. (2012), and Menon et al. (2012) McIntosh et al. (2004), Rathore et al. (2011), and Menon et al. (2012) Boshuisen et al. (2009) and Boshuisen et al. (2012) McIntosh et al. (2004)

1 1

Sinclair et al. (2007) Boshuisen et al. (2012)

controlled and non-controlled studies) 77% (95% CI 74.6, 79.7) of patients with or without discontinuation regained seizure freedom after AEDs were re-started, see Table 3. In general, the studies had limitations in the assessment of risk factors associated with seizure recurrence, however there were consistent observations (Table 4).

Meta-analysis of seizure recurrence Sixteen controlled studies comprise a total of 2141 participants that were included in the final meta-analysis (Fig. 2). We ran two meta-analysis due to the presence of a RCT (Kuzniecky et al., 1992). The second plot does not have this study (Fig. 3). Visual analysis of forest plots revealed clinical homogeneity. The point estimate of all but five studies was in the same direction. The Q statistic for all metaanalysis revealed no statistically significant heterogeneity. Both meta-analyses showed a protective odd ratio (OR 0.39, CI 95% 0.30—0.50, p < 0.001 and OR 0.38, CI 95% 0.29—0.50, p < 0.001) indicating a lower seizure recurrence in the group of patients where discontinuation was attempted.

Discussion Once seizure freedom is achieved with surgery, patients and clinicians must carefully ponder whether to taper or discontinue AEDs after successful ES. Unfortunately, the evidence is not robust and there is no consensus about management of AEDs in patients after ES (Jehi, 2013). Seizures recur whether patients are taking AEDs or not (Cole and Wiebe, 2008), and it is difficult to estimate the weight of each variable such as the clinical characteristics, EEG and MRI findings, type of surgery, among others. Some published studies suggest a risk of recurrence ranging from 20 to 50% over ten years in patients who are initially seizure free, and

suggest that the chances of remaining seizure free are not significantly affected by whether or not AEDs are discontinued (Berg et al., 2006; Kerling et al., 2009; McIntosh et al., 2004; Tanriverdi et al., 2008). In fact less than 20% of all postoperative seizure recurrences start in the setting of AED withdrawal (Jehi et al., 2010). An earlier non-systematic review performed by Schmidt et al. (2004), described that after planned discontinuation of AEDs in adult patients who rendered seizure-free after ES, the mean recurrence rate was 34%. This meta-analysis compared patients with and without AEDs discontinuation after ES (most often in TLE surgery), and provided a pooled estimate of the probability of having seizure recurrence. The seizure relapse rate in patients with planned discontinuation of AEDs in this dataset is lower than patients with no changes in the AEDs after ES (18% vs. 76%, p < 0.001). Most likely this difference is related with a highly selected population with a low risk profile for seizure recurrence after discontinuation of AEDs. It is hard to estimate the weight of the different variables, but the majority of patients in these studies had TLE (90%), with a focal lesion on MRI (85%), a complete surgical resection (81%) and almost all studies only included patients that were completely seizure free one year after surgery (98% Engel class I). AEDs were not reduced in patients who asked an increase of AEDs to control simple partial seizures, in patients with persistent seizures or fear of relapse, in patients that had undergone previous ES and in patients that reject changes in the medical treatment. Therefore, our meta-analysis demonstrates a selected sample of patients that could explain the lower recurrence in the group where discontinuation was attempted. Our findings are consistent with a recent study published by Hemb et al. (2013). In this study the long-term clinical outcomes of ES for mesial TLE and unilateral HS were evaluated. Patients were followed for a minimum of eight years to a maximum of 18 years, being one of the longest follow-up of patients

Management of antiepileptic drugs following epilepsy surgery

771

Figure 2 Meta-analysis of seizure recurrence between AED discontinuation vs. no discontinuation. The results of a fixed and random effect model, shows an OR of 0.39 (CI 95% 0.300—0.507). The seizure recurrence was lower in 11 out 16 studies in the group of patients were discontinuation was attempted.

after ES reported. They found that patients in whom significant or complete reduction of AEDs was achieved over the years had greater chances of remaining seizure free (55% vs. 21%, p < 0.03) (Hemb et al., 2013). It is unclear what the minimum time should be waited before AEDs can be safely withdrawn. Twenty years ago the discontinuation timing was shorter, some studies reported conversion of patients to monotherapy in the immediate postoperative period (Kuzniecky et al., 1992). Other studies showed that if patients were on monotherapy preoperatively, a lowering of dose was initiated within the first week after surgery and prior to discharge from hospital (McLachlan and Maher, 2000) and once initiated the reduction AEDs were quickly tapered over a 3-week to 3-month

period (Murro et al., 1991). This practice has changed over the years, currently 50% of the physicians would wait one year after surgery before reducing medications and 10% of them would wait at least two years to reduce medications according to a survey of Canadian epileptologists (TéllezZenteno et al., 2012). In a survey of clinical practice from US epileptologists the discontinuation timing varied according the number of medications, for patients on ≥2 AEDs who underwent successful ES, the majority of the epileptologists preferred to initiate AED withdrawal after 6—12 months (67%). In patients on monotherapy after ES the epileptologists preferred to wait longer to initiate tapering (12 months, 28%; 2 years, 31%; >2 years, 34.5%) (Swisher and Sinha, 2013). Another survey from US epileptologists,

Figure 3 Meta-analysis of seizure recurrence between AED discontinuation vs. no discontinuation in studies reporting medical treatment after epilepsy surgery. The results of a fixed and random effect model, shows an OR of 0.38 (CI 95% 0.295—0.502). This plot does not include the randomized clinical trial of Kuzniecky et al. (1992).

772

L.D. Ladino et al.

Table 5 Ideal candidates for AED discontinuation before and after successful epilepsy surgery according our review. 1 2 3 4 5 6 7

Young patients at surgery time (less than 30 years old) Short duration of epilepsy history (less than 11 years) Temporal lobe epilepsy clearly demonstrated by clinical features and EEG findings Lesional epilepsy by MRI Seizure freedom after surgery Normal postoperative EEG (no interictal spikes) Hippocampal sclerosis in pathology

showed more conservative results, 62% of physicians would wait at least two years before any change in AED dose (Berg et al., 2007). Our meta-analysis also showed a conservative management in the majority of studies. Overall the average time for starting AED withdrawal was 14 months. A period of less than 10 months between surgery and AED discontinuation is a common reported risk factor for recurrence (Boshuisen et al., 2012; Lachhwani et al., 2008; Lee et al., 2008; Park et al., 2010). The European TimeToStop observational study analyzed the seizure recurrence after AEDs discontinuation in a pediatric cohort of 766 children. The median time for initiation of AED taper was 12.5 months; they found that although shorter time to withdraw predicted a higher seizure recurrence, it did not predict the long-term seizure outcome (Boshuisen et al., 2012). Ziemba et al. (2011) reported that early tapering of AEDs (established as less than 9 months in their study) is a predictor of seizure recurrence in patients with nonlesional preoperative MRIs, but not in those with identified structural lesions. Additionally, most recurrences after pediatric (Boshuisen et al., 2009) and adult (Radhakrishnan et al., 2003; Jehi et al., 2010) ES occur during the first postoperative year; and having one seizure the first year had a six fold increase in likelihood of suffering continued seizure recurrences (Ramesha et al., 2011). Therefore, the available surveys of practice and our meta-analysis consider one year as the minimum time for beginning the AED discontinuation in carefully chosen patients. Another relevant information derived from our metaanalysis is the potential definition of a good candidate for AED discontinuation before and after surgery (Table 5). The most common risk factors reported in our meta-analysis for determining relapse after AEDs discontinuation were older age at surgery (>30 years) (Al-Kaylani et al., 2007; Lee et al., 2008; Menon et al., 2012; Rathore et al., 2011), longer duration of epilepsy (>11 years) (Menon et al., 2012; Murro et al., 1991; Park et al., 2010; Rathore et al., 2011) and non-lesional epilepsy (Kuzniecky et al., 1992; McIntosh et al., 2004; Park et al., 2010; Schiller et al., 2000) as presurgical variables. The main surgical variable was other procedure that a standard temporal resection (Sinclair et al., 2007). And the main postsurgical factors were having seizures before the withdrawal (Berg et al., 2006; Hoppe et al., 2006; Menon et al., 2012; Park et al., 2010; Rathore et al., 2011), an early withdrawal (before than 10 months) (Boshuisen et al., 2012; Lachhwani et al., 2008; Lee et al., 2008; Park et al., 2010) an abnormal EEG after the surgery

(Boshuisen et al., 2012; Menon et al., 2012; Rathore et al., 2011), incomplete resection (Boshuisen et al., 2009, 2012) (determined mainly by MRI or Electrocorticography) and the results of pathology (McIntosh et al., 2004; Menon et al., 2012; Rathore et al., 2011). The absence of definite HS on pathology (Rathore et al., 2011), the absence of abnormal pathology (McIntosh et al., 2004; Yoon et al., 2003) and poorly circumscribed pathologies like gliosis and focal cortical dysplasia (FCD) (Menon et al., 2012) were associated with higher risk of seizure recurrence. One recent study reported that patients having FCD and other malformations had significantly earlier relapses than did those with HS (de Tisi et al., 2011). Therefore, if the surgery removes the seizureproducing cortex in its totality, and if patients do not have an inherent tendency to have seizures, or surgery changed the brain in such a way that tendency was removed (Cole and Wiebe, 2008) is highly possible that the AEDs withdrawal will have a successful outcome. It is possible also that AED discontinuation may not increase the risk of seizure recurrence and other underlying epileptogenic features may be much more relevant for the probability of achieving and maintaining seizure freedom (Hemb et al., 2013). Finally, one of the most relevant observations of this meta-analysis is the demonstration that the great majority of patients where medications were tapered and relapsed became seizure free after re-starting therapy (76%), in contrast with patients who relapsed due to causes other than AED tapering whom were less likely to regain seizure freedom (39%). The recurrence of seizures after drug withdrawal has in general a good prognosis, since a readjustment of medical therapy is usually effective in preventing further seizures to occur. In our analysis 77% of patients re-gain seizure free status after the medications were re-started. A retrospective study showed a higher tendency to regain seizure free status in patients where the recurrences occurred after complete AED discontinuation vs. patients with breakthrough seizures followed by a simple initial dose reduction (80% vs. 57%, p = 0.06) (Jehi et al., 2010). This result suggests that if you cannot get a complete withdrawal there is an underlying tendency to continue having seizures (Jehi, 2013). The reasons why patients do not re-gain seizure free status after re-onset of treatment could be related with the following aspects: incomplete resection of epileptogenic tissue, dual pathology, bitemporal involvement since the onset, extratemporal impairment as temporal plus epilepsy, and a potential activation of a secondary epileptic focus (Schmidt and Löscher, 2003). Our meta-analysis has strengths. This study reviews the best available evidence delivering precise estimates of a group of studies rather than individual studies. With the analysis of a relative homogeneous group of studies we detected a strong selection sample bias that is common in all the studies. Probably the most difficult aspect in our meta-analysis is the interpretation of the protective odds ratio (OR) identified in the analysis. The inverse of this OR (2.5), suggest that the seizure recurrence can be as 2.5 times higher in patients where discontinuation was not attempted vs. attempted. We believe that we cannot recommend epileptologists to not withdraw medications due to a potential high seizure recurrence. Due to the selection bias in the studies we believe that the discontinuation of medications should be individualized favoring

Management of antiepileptic drugs following epilepsy surgery patients with the best profile, as we describe in previous sections. The main limitation of our study pertains to the quality of the evidence due to the retrospective nature of the majority of studies. The absence of RCTs brings into question the comparability of patients having each clinical intervention, but also the influence of confounders on observed outcomes. Sources of heterogeneity were identified in studies such as study design, patient selection, demographics, epilepsy etiology and type of ES. Of importance, the AED withdrawal series published to date have mostly combined seizure recurrences in the context of medication reduction with those occurring after complete AED discontinuation. The two situations may not be equivalent. Seizures occurring during AED reduction are more difficult to control than those occurring after AED discontinuation. Lack of standardized AED discontinuation method may also pose problems.

773 Third step: ‘‘Withdraw’’ [TW] OR ‘‘Withdrawal’’ [TW] OR ‘‘Discontinue’’[TW] OR ‘‘Discontinuation’’[TW] OR ‘‘stopping’’ [TW] OR ‘‘tapering’’ [TW] OR ‘‘reduction’’ [TW] OR ‘‘policy’’ [MeSH] OR ‘‘protocol’’ [TW] OR ‘‘Management’’ [TW] Fourth step: ‘‘Recurrence’’ [MeSH] OR ‘‘Remission’’ [MeSH] OR ‘‘relapse’’ [TW] OR ‘‘Treatment outcome’’ [MeSH] OR ‘‘efficacy’’ [TW] OR ‘‘curative’’ [TW] OR ‘‘cohort studies’’ [MESH] OR ‘‘risk’’ [MESH] OR (‘‘odds’’ [TW] AND ‘‘ratio*’’ [TW]) OR (‘‘relative’’ [TW] AND ‘‘risk’’ [TW]) OR ‘‘case—control*’’ [TW] OR ‘‘case—control studies’’ [MESH] Fifth step: Limits Publication Year 1980—2013 Human Journal article

Conclusion Our study delivers important information; a minority of patients (one in five) postoperatively seizure-free has a seizure recurrence after AED discontinuation. Patients with seizure recurrence after discontinuation can be managed easily after the re-star of medications. The discontinuation of medications should be performed in good candidates and the decision should be individualized taking into account clinical, electrographical, imaging and histopathological variables. Future research in the form of well-designed RCTs in large samples is needed to further delineate the best timing and method of AEDs modification after ES as well as to identify patients with greater risk for recurrence.

Disclosure Dr. Tellez receives grants from the University of Saskatchewan, UCB Canada and the Royal University Hospital Foundation, Saskatoon, Saskatchewan, through the Mudjadik Thyssen Mining Professorship in Neurosciences. Dr. Ladino and Dr. Hernandez-Ronquillo have nothing to disclose.

Acknowledgment We would like to thank Richard Zhu for his help in the literature search.

Appendix A. Literature search strategy First step: ‘‘Epilepsy/surgery’’ [MeSH] OR (‘‘epilepsy’’ [TW] AND ‘‘surgery’’ [TW]) OR ‘‘Anterior temporal lobectomy’’ [MeSH] OR ‘‘postsurgical’’ [TW] OR ‘‘Postoperative period’’ [MeSH] OR ‘‘postoperative’’ [TW] Second step: ‘‘Anticonvulsant(s)’’ [MESH] OR ‘‘Drug therapy’’ [MESH] OR (‘‘Antiepileptic(s)’’ [TW] AND ‘‘drugs’’ [TW]) OR ‘‘Treatment’’ [TW] OR ‘‘AED(s)’’ [TW] OR ‘‘therapeutic use’’ [MeSH]

References Al-Kaylani, M., Konrad, P., Lazenby, B., Blumenkopf, B., AbouKhalil, B., 2007. Seizure freedom off antiepileptic drugs after temporal lobe epilepsy surgery. Seizure 16 (2), 95—98. Berg, A.T., Langfitt, J.T., Spencer, S.S., Vickrey, B.G., 2007. Stopping antiepileptic drugs after epilepsy surgery: a survey of US epilepsy center neurologists. Epilepsy Behav. 10 (2), 219—222. Berg, A.T., Vickrey, B.G., Langfitt, J.T., Sperling, M.R., Shinnar, S., Bazil, C., Walczak, T., Spencer, S.S., Multicenter Study of Epilepsy Surgery, 2006. Reduction of AEDs in postsurgical patients who attain remission. Epilepsia 47 (1), 64—71. Boshuisen, K., Arzimanoglou, A., Cross, J.H., Uiterwaal, C.S., Polster, T., van Nieuwenhuizen, O., Braun, K.P., TimeToStop Study Group, 2012. Timing of antiepileptic drug withdrawal and long-term seizure outcome after paediatric epilepsy surgery (TimeToStop): a retrospective observational study. Lancet Neurol. 11 (9), 784—791. Boshuisen, K., Braams, O., Jennekens-Schinkel, A., Braun, K.P., Jansen, F.E., van Rijen, P.C., van Nieuwenhuizen, O., 2009. Medication policy after epilepsy surgery. Pediatr. Neurol. 41 (5), 332—338. Cole, A.J., Wiebe, S., 2008. Debate: should antiepileptic drugs be stopped after successful epilepsy surgery? Epilepsia 49 (9), 29—34. Das, R.R., Artsy, E., Hurwitz, S., Wen, P.Y., Black, P., Golby, A., Dworetzky, B., Lee, J.W., 2012. Outcomes after discontinuation of antiepileptic drugs after surgery in patients with low grade brain tumors and meningiomas. J. Neurooncol. 107 (3), 565—570. de Tisi, J., Bell, G.S., Peacock, J.L., McEvoy, A.W., Harkness, W.F., Sander, J.W., Duncan, J.S., 2011. The long-term outcome of adult epilepsy surgery, patterns of seizure remission, and relapse: a cohort study. Lancet 378 (9800), 1388—1395. 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 (9), 922—930. Griffin, C.T., Abastillas, M.E., Armon, C., Lacanlale, J., Liwnicz, B.H., Kaptan, G., Dayes, L.A., 2004. Early antiepileptic drug reduction following anterior temporal lobectomy for medically intractable complex partial epilepsy. Seizure 13, 434—437. Hemb, M., Palmini, A., Paglioli, E., Paglioli, E.B., Costa da Costa, J., Azambuja, N., Portuguez, M., Viuniski, V., Booij, L., Nunes, M.L., 2013. An 18-year follow-up of seizure outcome after surgery for

774 temporal lobe epilepsy and hippocampal sclerosis. J. Neurol. Neurosurg. Psychiatry 84 (7), 800—805. Hoppe, C., Poepel, A., Sassen, R., Elger, C.E., 2006. Discontinuation of anticonvulsant medication after epilepsy surgery in children. Epilepsia 47 (3), 580—583. Jehi, L., 2013. Medication management after epilepsy surgery: opinions versus facts. Epilepsy Curr. 13 (4), 166—168. Jehi, L., Sarkis, R., Bingaman, W., Kotagal, P., Najm, I., 2010. When is a postoperative seizure equivalent to epilepsy recurrence after epilepsy surgery? Epilepsia 51 (6), 994—1003. Kerling, F., Pauli, E., Lorber, B., Blümcke, I., Buchfelder, M., Stefan, H., 2009. Drug withdrawal after successful epilepsy surgery: how safe is it? Epilepsy Behav. 15 (4), 476—480. Khan, R.B., Onar, A., 2006. Seizure recurrence and risk factors after antiepilepsy drug withdrawal in children with brain tumors. Epilepsia 47 (2), 375—379. Kim, Y.D., Heo, K., Park, S.C., Huh, K., Chang, J.W., Choi, J.U., Chung, S.S., Lee, B.I., 2005. Antiepileptic drug withdrawal after successful surgery for intractable temporal lobe epilepsy. Epilepsia 46 (2), 251—257. Kuzniecky, R., Rubin, Z.K., Faught, E., Morawetz, R., 1992. Antiepileptic drug treatment after temporal lobe epilepsy surgery: a randomized study comparing carbamazepine and polytherapy. Epilepsia 33 (5), 908—912. Lachhwani, D.K., Loddenkemper, T., Holland, K.D., Kotagal, P., Mascha, E., Bingaman, W., Wyllie, E., 2008. Discontinuation of medications after successful epilepsy surgery in children. Pediatr. Neurol. 38 (5), 340—344. Lee, S.Y., Lee, J.Y., Kim, D.W., Lee, S.K., Chung, C.K., 2008. Factors related to successful antiepileptic drug withdrawal after anterior temporal lobectomy for medial temporal lobe epilepsy. Seizure 17 (1), 11—18. Maehara, T., Ohno, K., 2011. Preoperative factors associated with antiepileptic drug withdrawal following surgery for intractable temporal lobe epilepsy. Neurol. Med. Chir. (Tokyo) 51 (5), 344—348. McIntosh, A.M., Kalnins, R.M., Mitchell, L.A., Fabinyi, G.C., Briellmann, R.S., Berkovic, S.F., 2004. Temporal lobectomy: long-term seizure outcome, late recurrence and risks for seizure recurrence. Brain 127 (9), 2018—2030. McLachlan, R.S., Maher, J., 2000. Management of antiepileptic drugs following epilepsy surgery: a review. Can. J. Neurol. Sci. 27 (1), 106—110. Menon, R., Rathore, C., Sarma, S.P., Radhakrishnan, K., 2012. Feasibility of antiepileptic drug withdrawal following extratemporal resective epilepsy surgery. Neurology 79 (8), 770—776. Mohammed, H.S., Kaufman, C.B., Limbrick, D.D., Steger-May, K., Grubb Jr., R.L., Rothman, S.M., Weisenberg, J.L., Munro, R., Smyth, M.D., 2012. Impact of epilepsy surgery on seizure control and quality of life: a 26-year follow-up study. Epilepsia 53 (4), 712—720. Murro, A.M., King, D.W., Smith, J.R., Gallagher, B.B., Flanigin, H.F., Loring, D.W., Lee, G.P., Campbell, R., Meador, K., Revolinski, P.H., 1991. Withdrawal of antiepileptic medications following epilepsy surgery. J. Epilepsy 4, 39—42. Park, K.I., Lee, S.K., Chu, K., Jung, K.H., Bae, E.K., Kim, J.S., Lee, J.J., Lee, S.Y., Chung, C.K., 2010. Withdrawal of antiepileptic drugs after neocortical epilepsy surgery. Ann. Neurol. 67 (2), 230—238. Pimentel, J., Peralta, A.R., Campos, A., Bentes, C., Ferreira, A.G., 2012. Antiepileptic drugs management and long-term seizure outcome in post surgical mesial temporal lobe epilepsy with hippocampal sclerosis. Epilepsy Res. 100 (1/2), 55—58. Radhakrishnan, K., So, E.L., Silbert, P.L., Cascino, G.D., Marsh, W.R., Cha, R.H., O’Brien, P.C., 2003. Prognostic implications

L.D. Ladino et al. of seizure recurrence in the first year after anterior temporal lobectomy. Epilepsia 44 (1), 77—80. Ramesha, K.N., Mooney, T., Sarma, P.S., Radhakrishnan, K., 2011. Long-term seizure outcome and its predictors in patients with recurrent seizures during the first year after temporal lobe resective epilepsy surgery. Epilepsia 52 (5), 917—924. Rathore, C., Panda, S., Sarma, P.S., Radhakrishnan, K., 2011. How safe is it to withdraw antiepileptic drugs following successful surgery for mesial temporal lobe epilepsy? Epilepsia 52 (3), 627—635. Schiller, Y., Cascino, G.D., So, E.L., Marsh, W.R., 2000. Discontinuation of antiepileptic drugs after successful epilepsy surgery. Neurology 54 (2), 346—349. Schmidt, D., Baumgartner, C., Löscher, W., 2004. Seizure recurrence after planned discontinuation of antiepileptic drugs in seizurefree patients after epilepsy surgery: a review of current clinical experience. Epilepsia 45 (2), 179—186. Schmidt, D., Löscher, W., 2003. How effective is surgery to cure seizures in drug-resistant temporal lobe epilepsy? Epilepsy Res. 56 (2/3), 85—91. Sinclair, D.B., Jurasek, L., Wheatley, M., Datta, A., Gross, D., Ahmed, N., Quigley, D., Snyder, T., Aronyk, K., McKean, J., 2007. Discontinuation of antiepileptic drugs after pediatric epilepsy surgery. Pediatr. Neurol. 37 (3), 200—202. Swisher, C.B., Sinha, S.R., 2013. Survey of current practices among US epileptologists of antiepileptic drug withdrawal after epilepsy surgery. Epilepsy Behav. 26 (2), 203—206. Tanriverdi, T., Olivier, A., Poulin, N., Andermann, F., Dubeau, F., 2008. Long-term seizure outcome after mesial temporal lobe epilepsy surgery: corticalamygdalohippocampectomy versus selective amygdalohippocampectomy. J. Neurosurg. 108 (3), 517—524. Téllez-Zenteno, J.F., Dhar, R., Hernandez-Ronquillo, L., Wiebe, S., 2007. Long-term outcomes in epilepsy surgery: antiepileptic drugs, mortality, cognitive and psychosocial aspects. Brain 130 (2), 334—345. Téllez-Zenteno, J.F., Dhar, R., Wiebe, S., 2005. Long-term seizure outcomes following epilepsy surgery: a systematic review and meta-analysis. Brain 128 (5), 1188—1198. Téllez-Zenteno, J.F., Ronquillo, L.H., Jette, N., Burneo, J.G., Nguyen, D.K., Donner, E.J., Sadler, M., Javidan, M.M., Gross, D.W., Wiebe, S., Canadian Epilepsy Surgery Study Group, 2012. Discontinuation of antiepileptic drugs after successful epilepsy surgery. A Canadian survey. Epilepsy Res. 102 (1/2), 23—33. Van Veelen, C.W., van Rijen, P.C., Debets, R.M., van Wijk-Leenaars, P.W., van Emde Boas, W., 2001. Dutch Collaborative Epilepsy Surgery Program: reduction of seizures, operative complications and tapering of medication in 338 patients, 1973—1998. Ned. Tijdschr. Geneeskd. 145 (46), 2223—2228 (in Dutch). Wiebe, S., Blume, W.T., Girvin, J.P., Eliasziw, M., Effectiveness and Efficiency of Surgery for Temporal Lobe Epilepsy Study Group, 2001. A randomized, controlled trial of surgery for temporal lobe epilepsy. N. Engl. J. Med. 345 (5), 311—318. Yoon, H.H., Kwon, H.L., Mattson, R.H., Spencer, D.D., Spencer, S.S., 2003. Long-term seizure outcome in patients initially seizurefree after resective epilepsy surgery. Neurology 61 (4), 445—450. Zeng, T.F., An, D.M., Li, J.M., Li, Y.H., Chen, L., Hong, Z., Lei, D., Zhou, D., 2012. Evaluation of different antiepileptic drug strategies in medically refractory epilepsy patients following epilepsy surgery. Epilepsy Res. 101 (1/2), 14—21. Ziemba, K.S., Wellik, K.E., Hoffman-Snyder, C., Noe, K.H., Demaerschalk, B.M., Wingerchuk, D.M., 2011. Timing of antiepileptic drug withdrawal in adult epilepsy patients after neocortical surgical resection: a critically appraised topic. Neurologist 17 (3), 176—178.