FULL-LENGTH ORIGINAL RESEARCH
Predictors for long-term seizure outcome in juvenile myoclonic epilepsy: 25–63 years of follow-up
Address correspondence to Felix Schneider, Department of Neurology, Epilepsy Center, University of Greifswald, Sauerbruchstrasse, 17489 Greifswald, Germany. E-mail: firstname.lastname@example.org
Purpose: The long-term seizure outcome of juvenile myoclonic epilepsy (JME) is still controversial; the value of factors that are potentially predictive for seizure outcome remains unclear. The aim of this study was both to investigate the long-term seizure outcome in patients with JME after a follow-up of at least 25 years and to identify factors that are predictive for the seizure outcome.
Methods: Data from 31 patients (19 women) with JME were studied. All of them had a follow-up of at least 25 years (mean 39.1 years) and were reevaluated with a review of their medical records and direct telephone or face-to-face interview.
Key Findings: Of 31 patients 21 (67.7%) became seizure-free; in six of them (28.6%) antiepileptic drug (AED) treatment was discontinued due to seizure freedom. The occurrence of generalized tonic–clonic seizures (GTCS) preceded by bilateral myoclonic seizures (BMS) (p = 0.03), a long duration of epilepsy with unsuccessful treatment (p = 0.022), and AED polytherapy (p = 0.023) were identified as significant predictors for a poor long-term seizure outcome, whereas complete remission of GTCS under AED significantly increased the chance for complete seizure freedom (p = 0.012). The occurrence of photoparoxysmal responses significantly increases the risk of seizure recurrence after AED discontinuation (p = 0.05).
Significance: This study shows conclusively that JME is a heterogeneous epilepsy syndrome. Life-long AED treatment is not necessarily required to maintain seizure freedom. Several long-term outcome predictors that can potentially increase the ability of clinicians and their confidence to recommend different treatment options to patients with JME were identified.
Juvenile myoclonic epilepsy (JME) is a well-known idiopathic generalized epilepsy syndrome with a prevalence of 5–11% among all patients with epilepsy (Janz & Christian, 1957; Janz, 1969, 1985, 1989; Panayiotopoulos et al., 1994). As a specific electroclinical syndrome it is characterized by mandatory or typical myoclonic seizures alone or combined with generalized tonic–clonic seizures (GTCS) or absence seizures (ABS) (Janz, 1985; Dreifuss, 1989) and generalized spikes ≥3 Hz on interictal electroencephalography (EEG) (Commission on Classification and Terminology of the International League Against Epilepsy, 1989).
Bilateral myoclonic seizures (BMS) typically occur on awakening (Janz & Christian, 1957); GTCS appear in the majority of the cases and ABS in only 15–30% (Asconapé & Penry, 1984; Delgado-Escueta & Enrile-Bacsal, 1984). Series of BMS often precede GTCS (Jain et al., 1997). Age of JME onset is usually in the adolescence with a female predominance (Genton et al., 2000). Interictal EEG in JME shows generalized interictal epileptiform discharges (Commission on Classification and Terminology of the International League Against Epilepsy, 1989) such as generalized spikes, polyspikes, spike-wave complexes, or combinations of these (Usui et al., 2005). An earlier study on childhood absence epilepsy (CAE) described a progression to JME in 44% of their patients who either continued antiepileptic drug (AED) treatment or had ongoing seizures at follow-up (Wirrell et al., 1996). However, some authors consider cases of CAE developing to JME as subtypes of JME with a different outcome (Camfield & Camfield, 2009). Several factors that potentially provoke epileptic seizures or epileptic discharges in patients with JME, such as higher mental activities like speaking, reading, writing, arithmetic calculation, and spatial construction (Matsuoka et al., 2000) or perioral reflex myoclonias induced by either reading or speaking (Mayer et al., 2006), have been identified. Furthermore, approximately 30% of the patients show photoconvulsive responses (Asconapé & Penry, 1984). Earlier studies reported JME to be a chronic disease that required lifelong AED treatment (Delgado-Escueta & Enrile-Bacsal, 1984; Martínez-Juárez et al., 2006). There are only few studies that focus on the long-term follow-up of JME. One recent study in 48 patients showed a remission of BMS in 26 of their patients (54.2%) after a median patient age of 32.9 years (Baykan et al., 2008). Another study with a mean follow-up of 25.8 years reported seizure freedom in 17% of their JME patients without AED; however, predictors for seizure freedom were not ascertained (Camfield & Camfield, 2009).
The aim of this study was both to investigate the long-term seizure outcome in patients with JME after a follow-up of at least 25 years and to identify factors that are predictive for seizure remission.
Materials and Methods
This study was approved by the University of Greifswald Institutional Review Board, including telephone consent. Written informed consent was received by every patient. The study was conducted among the inhabitants of the catchment area of the University Hospital of Greifswald (total population approximately 500,000) in the northeast of Germany. Data from all patients diagnosed with JME before January 1986 were retrospectively reviewed from the Epilepsy Center database. Inclusion criteria were (1) diagnosis of JME, (2) normal neurologic examination and overall intelligence, and (3) follow-up of at least 25 years. We excluded patients with a history of severe brain trauma and epilepsy syndromes other than JME. Diagnostic criteria of JME included the history of BMS (mandatory) with or without additional GTC and/or ABS. Unrecognized ABS before diagnosis were included. Brain imaging was not required. EEG studies at the time of diagnosis and during the course were performed in all patients using the international 10–20 system of electrode placement. Diagnosis of JME was made by experienced epileptologists (UR, RH) on the basis of the patients’ medical history and medical history by witnesses as well as the patients’ EEG studies. Discontinuation of AED treatment in seizure-free patients was attempted only if desired by the patient and was not common practice in this Epilepsy Center at the time. No regular follow-up program for patients with JME exists at our institution. However, all cases with JME were reevaluated by experienced epileptologists (JG, FS, UR) between May and September 2011 with a review of their medical records and direct telephone or face-to-face interview. Of 71 patients screened, 31 could be reached and were enrolled in the study. Of the remaining 40 patients 38 could not be contacted due to either unknown addresses or names of both the patients and the current treating physicians; the remaining two patients died in 2004 and 2006 for unknown reasons. Data about seizure types and age of onset of each seizure type, initial and current seizure frequency, detailed medical history, provoking factors, duration of epilepsy, family history, EEG results including photoparoxysmal responses as well as social aspects were collected by reviewing the medical record and during the interview at follow-up using a standardized questionnaire. The questionnaire included objective questions such as yes/no questions as well as open questions such as questions about the seizure frequency. Patients were accepted for this study only if they met the criteria for JME at the time of diagnosis. Both the follow-up interview and the collection of patients’ current medical history and medical history by witnesses were performed by experienced epileptologists (UR, FS, JG). EEG studies at the time of follow-up were not part of the study protocol; however, all included patients were in regular medical attendance and 23 (74.2%) of them had current EEG investigations. The authors reviewed hospital and physicians EEG records. Seizure-free outcome at follow-up was defined as a terminal seizure-free period up to enrollment of at least 5 years.
Photoparoxysmal response (PPR) is defined as the occurrence of spikes, spike waves, poly-spike waves, or repetitive spikes in response to intermittent photic stimulation (Wolf & Goosses, 1986; Doose & Waltz, 1993; Verrotti et al., 2005; Lu et al., 2008). Due to the lack of data on reflex epileptic traits other than PPRs in our patient group, such as higher mental activities or perioral reflex myoclonias induced by either reading or speaking, further subgroup analysis with regard to prognostic implications for the long-term seizure outcome was not part of our study protocol.
SPSS 20.0 (IBM Co., Armonk, NY, U.S.A.) was used for statistical processing of the data. The statistical methods were descriptive statistics with frequency analysis and cross-tab analysis, as well as mean and standard deviation calculation in parametric data. Sensitivity, specificity, and predictive values were calculated for each parameter to predict long-term seizure outcome. Statistical significance was assessed using Spearman correlation, Fisher’s exact test, Mann-Whitney-Wilcoxon test (MWW), and chi-square test, with a significance defined as a p-value of ≤0.05. Under consideration of the Bonferroni correction the p-value is ≤0.002.
Details of all patients included in the study are given in Table 1.
Table 1. Clinical data of all patients included in study
| 1||51||F||SF−||18||33||30||1||BMS||−||+||BMS + GTC||SD, ST, AL||+||+||−||−||−||−||0||1|
| 2||49||F||NSF+||19||30||1||1||BMS||−||+||BMS + GTC||SD, ST||+||+||+||−||−||−||1||2|
| 3||41||M||SF+||9||32||13||2||BMS||+||+||BMS + GTC||SD, AL||+||+||−||+||−||−||1||3|
| 9||65||F||SF+||16||49||17||1||GTC||+||−||BMS + GTC||SD||−||u||−||−||−||−||1||2|
|10||43||M||SF+||15||28||9||1||BMS||+||+||BMS + GTC||AL, ST||+||+||−||+||−||−||1||2|
|11||40||F||NSF+||8||32||0||3||ABS||−||u||BMS||SD, ST, AL||+||+||u||−||−||+||1||1|
|12||75||F||SF+||12||63||12||2||BMS||+||−||BMS + GTC||SD||−||−||+||−||−||−||2||5|
|14||60||M||SF−||11||49||21||1||GTC||−||−||BMS + GTC||SD, ST, AL||−||−||−||−||−||−||0||4|
|19||48||F||SF+||14||34||21||1||GTC||+||+||BMS + GTC||SD, ST||−||−||−||−||−||−||1||3|
|20||40||M||NSF+||15||25||2||1||BMS||−||−||BMS + GTC||No||−||−||−||−||−||−||2||3|
|24||68||F||SF+||15||53||16||1||GTC||−||−||BMS + GTC||SD, ST||−||−||−||−||−||−||1||3|
|25||46||F||NSF+||15||31||0||1||GTC||+||−||BMS + GTC||SD||+||+||+||+||−||−||1||2|
|27||73||F||SF+||14||59||5||1||GTC||+||−||BMS + GTC||SD, ST||−||−||−||−||−||−||2||2|
|28||44||F||SF+||12||32||16||2||GTC||−||−||BMS + GTC||SD||−||−||−||−||−||−||1||5|
We enrolled 31 patients (19 women) with a mean age of 52.2 years (standard deviation [SD] ± 12.27, range 29–76). The mean age of epilepsy onset (first recognized seizure) was 13.1 years (SD ± 4.73, range 2–23). The mean follow-up was 39.1 years (SD ± 11.9, range 25–63).
Forty (56.3%) of 71 screened patients (27 women) were lost to follow-up, their mean age was 47.1 years (SD ± 8.81, range 33–81), mean duration of epilepsy was 34.9 years (SD ± 8.70, range 25–69), and mean follow-up at the last follow-up was 9.2 years (range 2–17). Between the group of included patients and those who were lost to follow-up no significant difference was found with regard to age of epilepsy onset (MWW: p = 0.329), duration of epilepsy (MWW: p = 0.252), and types of occurring seizures (MWW: p = 0.934). However, PPRs were significantly more frequent among the lost patients (77.5%) compared to the study group (48.4%) (MWW: p = 0.011).
In our cohort of 31 patients, 16 (51.6%) had a history of BMS and GTCS, 11 (35.5%) additionally had ABS, whereas the remaining 4 (12.9%) had BMS and ABS only. In the majority of our patients (51.6%) the epilepsy started with GTCS (with BMS in 22.6%; with ABS in 25.8%). In four patients (12.9%) there was a history of CAE, which converted into JME. All patients were treated with AEDs; the mean number of AEDs used was 3.1 (SD ± 1.73, range 1–6). Nineteen patients (61.3%) received valproic acid (VPA), and 14 of them in monotherapy. Neither the total number of AED trials during the entire course (chi-square test: p = 0.34) nor the duration of AED treatment (chi-square test: p = 0.918) is significantly associated with the long-term outcome; however, a significant negative correlation was found between an increasing number of AEDs at follow-up and a seizure-free outcome (chi-square test: p = 0.023).
Seizure outcome at follow-up
Twenty-one patients (67.7%) became seizure-free under AED treatment. Mean duration of epilepsy was 34.2 years (SD ± 8.04, range 25–49). Discontinuation of AED treatment due to seizure freedom was attempted in 9 (42.9%) of them (by treating physician: n = 4, by the patient: n = 5). Six of these nine patients (66.7%; 28.6% of all seizure-free patients) remained seizure-free without AED for a mean duration of 19.2 years (SD ± 8.01, range 8–30) (mean duration of epilepsy until AED withdrawal was 15.2 years [SD ± 8.7, range 3–28]). Of these six patients, two became seizure-free with the first AED (ethosuximide, VPA), and two after the second AED trial (Table 2). In three (33.3%) of the nine patients seizures recurred (after 1, 10, and 38 years) and AED treatment was restarted, which led to seizure freedom in all of them (Table 1, Patients 2, 18, and 27); in contrast to the findings of Wolf et al. (2006), seizures recurred during AED reduction in one, and after complete AED discontinuation in two of our patients.
Table 2. Clinical data of all seizure-free patients without AED
|14||60||M||11||49||BMS/GTCS||GTCS||10 – 20||PRI||4||32||22||21||Patient||−||BMS/GTCS||−||−||−||−||−||−|
Including these three patients, 15 were seizure-free with AED at follow-up; they had AED treatment for at least 25 years with a mean seizure-free time of 17.1 years (SD ± 7.81, range 5–29). Eight of these patients had BMS and GTC, six additionally had ABS, and one had BMS and ABS. Three patients became seizure-free under the first AED (VPA) and another three under the second AED (two of them VPA). In the remaining nine patients three or more AEDs were tried. At follow-up 11 (73.3%) of the 15 patients were treated with monotherapy (nine VPA) and 4 received polytherapy (3 of them including VPA).
Ten patients (32.3%) continued to have seizures despite multiple different AEDs. The mean number of AEDs previously tried was 3.8 (SD ± 1.87, range 1–6). The mean number of AEDs at follow-up was 1.8 (SD ± 0.92, range 1–4).
Predictors of seizure outcome
Statistical values of predictors for the long-term seizure outcome of patients with JME are given in Table 3 (statistically significant factors are in bold). Details on several important aspects are described below.
Table 3. Diagnostic measures for predictors for the long-term seizure outcome of JME based on the outcome (statistically significant factors in bold)
|Nominally scaled data|| || || || || || |
| First recognized seizure type: BMS||0.82||0.813a||0.238||0.8||0.714||0.333|
| First recognized seizure type: GTCS||0.543||0.525a||0.632||0.5||0.75||0.364|
| Total number of GTCS (≤10 or >10)||0.093||0.087a||0.474||0.875||0.9||0.412|
| Age at first GTCS (≤12 or >12)||0.618||0.601a||0.211||0.875||0.8||0.318|
| Seizure-free patients with late-onset GTCS: AED discontinuation/seizure recurrence|| 0.011 || 0.013 a || 0.071 || 0.4 || 0.25 || 0.133 |
| AED treatment leads to remission of GTCS|| 0.023 || 0.012 a || 1 || 0.25 || 0.76 || 1 |
| GTCS preceded by BMS|| 0.03 || 0.03 a || 0.111 || 0.5 || 0.333 || 0.2 |
| Chronodependency of GTCS||0.834||0.826a||0.842||0.125||0.696||0.25|
| Chronodependency of BMS||0.885||0.88a||0.571||0.4||0.667||0.308|
| Occurrence of BMS in series||0.258||0.243a||0.2||0.6||0.5||0.273|
| JME-specific EEG discharges||0.21||0.197a||0.95||0.2||0.704||0.667|
| JME-specific EEG discharges||0.416||0.398a||0.75||0.4||0.714||0.444|
| Photosensitivity in SF patients: AED+/−|| 0.04 || 0.05 b || 0.667 || 0.833 || 0.909 || 0.5 |
| Family history of IGE||0.968||0.467a||0.095||0.9||0.667||0.321|
| Family history of JME||0.599||0.584a||0.3||0.6||0.6||0.3|
| JME developing from CAE||0.117||0.109a||0.143||0.6||0.429||0.25|
| No. of AEDs (at time of follow-up)|| 0.011 || 0.023 a || 0.81 || 0.6 || 0.81 || 0.6 |
|Metrically scaled data|| || || || || || |
| Age at epilepsy onset||0.982||0.659a|| || || || |
| Time from diagnosis until seizure freedom|| 0.003 || 0.022 a || || || || |
| Seizure types||0.439||0.671a|| || || || |
| No. of AEDs (in total)||0.118||0.34a|| || || || |
| Duration of AED treatment||0.854||0.918a|| || || || |
Generalized tonic–clonic seizures
Of 27 patients (87.1%) with GTCS, 10 patients (37.0%) had 1–10 GTCS, 7 (26.0%) had 11–50 GTCS, and 10 patients (37.0%) had >50 GTCS throughout their entire clinical course. Despite a high positive predictive value (PPV) of 90% (Table 3), no significant association between the total number of GTCS (≤10 or >10) and the long-term outcome was found (chi-square test: p = 0.087). In 22 patients (81.5%) the first recognized GTCS occurred at an age >12 years. Early or late GTCS onset (before or after the age of 12 years) is not significantly correlated with a seizure-free follow-up (chi-square test: p = 0.601); however, if AED treatment leads to seizure freedom in patients with late GTCS onset (age >12 years), AED continuation is required to maintain the patients’ seizure freedom (chi-square test: p = 0.013). Of 27 patients with GTCS, treatment with AEDs led to GTCS remission in 25 patients (92.6%); 19 of them became completely seizure-free. It is notable that remission of GTCS under AED is significantly associated with complete remission of BMS and/or ABS (chi-square test: p = 0.012). In six patients the individual GTCS were either occasionally or habitually preceded by BMS; only two patients became seizure-free, whereas 80.0% of the patients whose GTCS were never preceded by BMS became seizure-free. The occurrence of GTCS preceded by BMS is significantly associated with a lower likelihood of attaining seizure freedom (chi-square test: p = 0.03).
PPRs were found in 15 patients (48.4%) at least one time during the course of the patients’ epilepsy; 11 of these patients (73.3%) became seizure-free. Seizure-free rate in patients without PPR was 62.5%. No significant correlation between the occurrence of PPR and a seizure-free follow-up has been found (chi-square test: p = 0.519); however, in the group of seizure-free patients with PPR, AED withdrawal is significantly associated with seizure recurrence (Fisher’s test: p = 0.05; PPV 90.9%); only 9.1% of these patients no longer receive AEDs.
Duration of epilepsy
A shorter duration of epilepsy until seizure freedom is reached was significantly associated with a favorable long-term outcome (chi-square test: p = 0.022).
In this retrospective study, we investigated the long-term seizure outcome and predictors for seizure remission in patients diagnosed with JME after a unique mean follow-up of 39.1 years. To our knowledge, this is both the largest study with the longest follow-up to examine the long-term seizure outcome, and the first study to identify factors that are predictive for the seizure outcome in this most challenging group of patients.
The most outstanding finding from this work is that the majority of our patients (67.7%) became seizure-free; in nine of these patients AED treatment was stopped due to seizure freedom, and six patients remained seizure-free at follow-up. Due to the current opinion, that JME requires life-long AED therapy, AED discontinuation in seizure-free patients was not common practice in our epilepsy center and only attempted if desired by the patient. Consequently our numbers of attempts to discontinue AEDs are relatively small.
Our results also show that the time to reach seizure freedom is a significant predictor of the long-term seizure outcome (chi-square test: p = 0.022), suggesting that the chance of long-term seizure freedom decreases with the time of ineffective AED treatment. In a recent JME long-term outcome study, Camfield and Camfield (2009) demonstrated 26.1% of their patients to be seizure-free without AEDs, whereas 21.7% of their patients had seizure recurrences after AED withdrawal. Another prior study on 48 patients with JME showed 11 of 22 patients aged >40 years of age to be seizure-free (Baykan et al., 2008); only four of their patients had a remission of BMS and GTCS without AED. Taken together, in contrast to the current opinion, our results suggest that life-long AED therapy is not necessarily required to maintain seizure freedom in patients with JME; however, the decision to discontinue AEDs should depend on several long-term seizure outcome predictors identified in our study.
- 1 This is the first study showing that the occurrence of GTCS, which are preceded by BMS, is significantly associated with the long-term seizure outcome (chi-square test: p = 0.03). However, notwithstanding a high PPV, the total number of GTCS (PPV 90%) and the patient’s age at the time of the first recognized GTCS (PPV 80%) do not significantly influence the clinical course. That may give hope to patients with early GTCS onset and a high GTCS frequency in the beginning of their epilepsy.
- 2 A increased number of AEDs at the time of follow-up significantly associates with a worse seizure outcome (SC: p = 0.011). Due to legal restrictions on the availability of medications in former East Germany, including proper AEDs for JME treatment (such as VPA), physicians usually started treatment with easily accessible but often inadequate AEDs (such as carbamazepine, benzodiazepines), and if necessary, switched to other monotherapies or polytherapies to obtain seizure freedom. Consequently, our finding of a mean number of 3.1 AEDs used is relatively high. Despite that, the total number of AED trials is not significantly associated with the long-term outcome, though (chi-square test: p = 0.34). These findings are consistent with those of Martínez-Juárez et al. (2006) that more than two thirds of their seizure-free patients with JME were treated with AED in monotherapy. Confirmatory, our results show that an increase of AEDs in combination is associated with a worse long-term outcome, irrespective of the total number of AEDs tried in that patient. Conclusively the number of monotherapies tried has no predictive value for the long-term outcome.
- 3 No significant association was found between the different JME-specific seizure types and the long-term outcome (chi-square test: p = 0.671). In contrast to both our findings and those of Baykan et al. (2008), Gelisse et al. (2001) reported a significant association with drug resistance when all three seizure types are present. In addition, our study shows that complete remission of GTCS under AED treatment is significantly correlated with a complete seizure-free outcome (chi-square test: p = 0.012), suggesting that AED treatment should be adjusted to reach complete seizure freedom in these patients. Taken together, successful treatment of GTCS can be considered as a predictor for a seizure-free long-term outcome. In an earlier study, Jain et al. (1997) reported a benign subgroup of patients with JME with BMS only, which showed a better long-term prognosis. In our study the lack of those patients precludes any conclusions with regard to the seizure outcome in this special subgroup of patients with JME.
- 4 Consistent with prior findings of Baykan et al. (2008), our study shows no significant association between PPR and other interictal JME-specific EEG findings with a seizure-free long-term outcome. However, our study shows that in seizure-free patients, PPRs are significantly associated with seizure recurrence when AED treatment is discontinued (Fisher’s test: p = 0.05). Furthermore, 83.3% of our seizure-free patients without AED showed no PPRs. Taken together, it is likely that photosensitive JME can be considered as a JME subtype with a higher seizure risk, which makes AED treatment indispensable in these patients. However, PPRs were significantly more frequent among patients who were lost for follow-up compared to the study group. Assuming either favorable or poor seizure outcome in these patients as a reason for discontinuing follow-up with our epilepsy center, there is a potential bias toward a more select study group that has to be considered in the interpretation of our results.
Martínez-Juárez et al. (2006) focused on different JME subgroups and reported a seizure-free rate of 8.6% of their patients with CAE evolving to JME (all of them with AED). These data suggest that JME developing from CAE may represent a different JME subtype with a worse outcome, which was why prior studies on JME long-term outcome excluded these patients (Camfield & Camfield, 2009). In our study, a history of CAE was not significantly correlated with the long-term outcome; however, our small number of CAE patients precludes further subgroup analysis and needs to be considered when interpreting these results.
Several limitations of our single-center study need to be considered. First, given the relatively small group of 31 patients, the possibility of type 2 statistical errors that limit statistical validity should be considered. Secondly, we included patients with CAE, who might represent a different JME subtype with a worse outcome (Camfield & Camfield, 2009). Although only four of our patients had a history of CAE, this might have an effect on our findings and should be considered when interpreting these results. Thirdly, in 8 (38.1%) of 21 seizure-free patients determination of the outcome was based on the patients report without EEG confirmation. Lastly, 38 of 71 patients with JME could not be contacted due to either unknown addresses or names. Assuming either favorable or poor outcome in these patients as a reason for renunciation from our epilepsy center, a potential selection toward only easy-to-follow patients has to be considered (Callanan et al., 2001).
Despite the above limitations, our study shows that JME is likely a heterogeneous epilepsy syndrome. Our findings of long-term seizure freedom and the statistical validation of several outcome predictors can potentially increase the clinicians’ ability and confidence to recommend different treatment options to patients with JME. Conclusively our results provide patients with JME, especially the younger group, with a higher degree of confidence for their social and vocational planning. In contrast to the current opinion, life-long AED treatment is shown to be not necessarily required to maintain seizure freedom in these patients, and the decision should depend on several predictors identified in our study. Generalized tonic–clonic seizures preceded by BMS, a long duration until seizure freedom is reached, and AED polytherapy were identified as significant predictors of a poor long-term seizure outcome, whereas remission of GTCS under AED treatment is predictive of a long-term seizure-free outcome. The occurrence of PPRs significantly increases the chance of seizure recurrence after AED discontinuation. Future prospective studies will need to be undertaken in this challenging patient population to test the replicability and generalizability of our observations.
We gratefully acknowledge the assistance of Professor Dr. med. G. Rabending and Dr. rer. nat. P. Kolyschkow, University of Greifswald, Germany.
None of the authors has any conflict of interests to disclose. We confirm that we have read the Journal’s position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.