JLE

Epileptic Disorders

MENU

Surgical management of pediatric patients with encephalopathy due to electrical status epilepticus during sleep (ESES) Volume 22, numéro 1, February 2020

Electrical status epilepticus during sleep (ESES), also known as continuous spike-wave discharges during slow-wave sleep (CSWSS), is an age-dependent epileptic encephalopathy that often remits over time (Patry et al., 1971; Kramer et al., 2009; Rubboli and Tassinari, 2019). Defined by the Commission on Classification and Terminology of the International League Against Epilepsy 1989, it typically presents in children around five years of age, with clinical seizures which progress within two years to a severe epileptic encephalopathy. In addition to epileptic seizures, the onset of ESES is associated with a progressive encephalopathy characterized by developmental regression, neurobehavioral disorder and severe deficits in language, social skills, memory, global intellect and motor function (Galanopoulou et al., 2000; Tassinari et al., 2000; Veggiotti et al., 2001; Raha et al., 2012; Tassinari and Rubboli, 2019). Landau-Kleffner syndrome (LKS) is a closely related entity that involves compromised eloquent language development and typically presents with verbal language delay (Pal et al., 2016).

ESES limited to one cerebral hemisphere is referred to as hemi-ESES. Therapy is similarly directed towards eliminating the electrographic pattern in order to control seizures and halt neurocognitive decline. Medical management options include antiepileptic drugs (AEDs), high-dose corticosteroids, benzodiazepines, intravenous gamma globulins and the ketogenic diet (Aeby et al., 2005; Kramer et al., 2009; Caraballo et al., 2011; Fejerman et al., 2012; Veggiotti et al., 2012; Arzimanoglou and Cross, 2019; Jansen et al., 2019).

Surgery is a promising therapeutic option for pharmaco-resistant ESES, but is infrequently employed due to potential spontaneous remission of ESES, and concerns about surgical complications. Multiple subpial transections (MSTs) have been performed on patients with language-eloquent ESES in LKS, but the results are inconsistent. Focal resections and hemispherectomies are performed more rarely and yield variable results. Since the natural history of ESES and the predictors of surgical success are unclear, there is significant clinical equipoise on the optimal approach for children with this condition.

We report a cohort of children with seizures and electrographic findings of ESES who were evaluated and managed at a single tertiary center specializing in epilepsy. We herein describe the patient population including seizure semiology, radiographic findings, and neuropsychological assessment. Furthermore, we describe the factors influencing the decision to perform surgery, and report the safety and efficacy of surgical management. Lastly, we discuss the benefits of surgery compared to medical management, and describe patient factors predicting optimal outcome. Although other authors have reported small case series with resective procedures (table 1), to our knowledge, this is the largest case series with resective and disconnective surgical management for ESES reported to date.

Materials and methods

We performed a retrospective case review of consecutive patients undergoing surgery for ESES at Nicklaus (formerly known as Miami) Children's Hospital between January 2009 and July 2016. Patient demographic information, seizure semiology and frequency, pre-operative radiography, electrophysiology, and neuropsychological testing were evaluated. Information on the type of surgery, final surgical pathology, and any associated surgical complications were collected. At least two years of post-surgical follow-up was obtained in all patients, and post-operative follow-up was reviewed for seizure frequency, neurocognitive outcome, and academic performance. When available, a post-operative electroencephalogram (EEG) was also reviewed to determine resolution of ESES electrographic patterns.

Inclusion criteria

Between January 2009 and July 2016, the prospectively collected data for all patients undergoing video electroencephalography (VEEG) and epilepsy workup at Nicklaus Children's Hospital were reviewed. Patients were classified with ESES if they presented with the typical EEG findings of continuous spike-waves during at least 85% of slow wave (non-REM) sleep, usually noted after seizure onset and persisting on at least three or more recordings over a period of at least one month in association with epilepsy, neuropsychological and/or motor impairment. From this cohort, we identified patients with electrographic abnormalities that were either hemi-ESES, or asymmetrically prominent in one hemisphere. Patients in this cohort were boys and girls between 0-18 years of age, who ultimately underwent surgery at Nicklaus Children's Hospital.

Presurgical workup

Children with seizures were initially evaluated in the neurology clinic, where a detailed assessment of their seizure semiology, frequency, and age at seizure onset was performed. Developmental, past medical, surgical, and family histories were recorded, focusing on pertinent details of prematurity, history of infantile spasms, and central nervous system (CNS) insults (tumor, hemorrhage or hypoxia) and cortical malformation. Motor function, language competence, developmental milestones and visual fields were assessed.

EEG and VEEG monitoring captured waking and sleep interictal and ictal electrographic findings.

All patients underwent a comprehensive radiographic evaluation utilizing 3-Tesla magnetic resonance imaging (MRI). MRI diffusion tensor imaging (DTI) was occasionally used for tractography, and functional MRI (fMRI) was used to assess language dominance. Positron emission tomography (PET), ictal single-photon emission tomography (SPECT) and neuroelectromagnetic source imaging (NSI) were utilized selectively. Pertinent radiographic findings, including T2/FLAIR signal changes, polymicrogyria, focal cortical dysplasia, atrophy, infarction, and lesions, were noted. PET and SPECT findings of hypo- or hypermetabolic areas were also used to assess concordance with the seizure semiology and VEEG findings.

Nine children underwent formal pre-operative neuropsychological testing, which included assessment of overall intelligence, behavior, psychology and psychiatric co-morbidity. Intelligence was assessed with the Wechsler Abbreviated Scale of Intelligence (WASI) or Wechsler Intelligence Scale for Children (WISC-IV) for patients aged >six years old, or the Wechsler Preschool and Primary Scale of Intelligence™ - Third Edition (WPPSI™ - III) for patients aged 3-7 years old. Psychological screening was performed using two parental questionnaires, the Conner's Parent Report Scale-Revised (CPRS-R) and the Child Behavioral Checklist (CBCL). Academic achievement, school performance, and any learning disability was documented through clinical history taking. The Childhood Autism Rating Scale (CARS) was used to screen for pervasive development disorder. The Diagnostic and Statistical Manual of Mental Disorders (DSM IV) was used to assess psychiatric co-morbidity.

Children considered for surgery were presented at a weekly Epilepsy Surgery Case Conference. Based on seizure semiology, electrographic and MRI findings and functional imaging, a resective or disconnective procedure was proposed. Surgical intervention was offered to children who were unresponsive to AEDs and ancillary medications (benzodiazepines and corticosteroid therapy), and who evidenced neurocognitive decline coincident with ESES onset. In three cases, the children had decreased overall IQ scores and word processing speech on formal neuropsychological testing; in 10 cases, neurocognitive decline was reported by parents or documented academically through progressive academic delay. Children who met these criteria were assessed to be good surgical candidates, and underwent surgery at Nicklaus Children's Hospital by one of two neurosurgeons specialized in epilepsy.

Outcome

Post-operatively, children were re-evaluated by the neurology and neurosurgery services, and complications or unexpected deficits were recorded. Epilepsy outcome was assessed using the Engel classification at the time of last follow-up visit. Post-operative EEG or VEEG was performed to assess the degree of ESES resolution, and to evaluate interictal and ictal epileptiform activity during wakefulness. Postoperative neuropsychological outcome was obtained either through formal objective testing, or subjectively by parental questionnaires detailing the child's behavior, social interactions, verbal skills, and academic performance.

Results

Patient population

Of 9,502 pediatric patients who underwent VEEG evaluation at Nicklaus Children's Hospital between January 2009 and July 2016, 58 children met criteria for ESES. During the study period, 11 patients (19%) with an average age of 9.4 ± 3.4 years (range: 5-16 years) suffering from ESES underwent surgery at our institution (table 2).

Mean age at seizure onset was 3.2 ± 1.9 years (range: 0.003-6 years) and the mean age at formal ESES diagnosis was 7.1 ± 2.4 years (range: 4-11.5 years). The most common etiology was cortical malformation (MCD), consisting of polymicrogyria (PMG) or focal cortical dysplasia (FCD) in four children. Three children had thalamic or intraventricular hemorrhage (IVH) and subsequent encephalomalacia; two as a result of surgery for choroid plexus papilloma, and one with unknown etiology. One child had Rasmussen's encephalitis and another had perinatal middle cerebral artery (MCA) infarct leading to developmental porencephaly. Two children had gliosis of unknown etiology.

Two children experienced generalized seizures, including absence, atonic, and generalized convulsive semiologies; nine children experienced multiple focal seizure types. Secondary generalized seizures were the most common seizure semiology (eight patients), although one child experienced focal motor and non-motor seizures without generalization. Three patients experienced at least one documented episode of status epilepticus that required rescue medication and intensive care monitoring. Seizure frequency was daily (n=6), weekly (n=1), or monthly (n=4).

Preoperatively, nine children evidenced contralateral hemiparesis and three also had a homonymous hemianopsia. Seven children had language delay consistent with LKS. All patients were initially treated with AEDs and high-dose benzodiazepine (0.5-1.0 mg/kg diazepam at bedtime) as first-line therapy, followed by high-dose prednisone as second agent and administered over a period of one month. The ketogenic diet and vagus nerve stimulator (VNS) were also employed unsuccessfully in some patients. Five children were right-handed, and six were left-handed. Although handedness often correlated to ipsilateral ESES and contralateral paresis, three children had discordant handedness (Cases 4, 8, and 9).

Formal neuropsychological evaluation was performed at Nicklaus Children's Hospital on nine patients prior to surgery; the other children did not undergo formal evaluation due to either a language barrier or significant global delay (table 3). Neuropsychological impairments ranged widely, from global cognitive regression to more selective impairment of specific cognitive domains or expressive language. Overall intelligence was variable, with Intelligence Quotient (IQ) ranging from impaired to low average. Working memory index (WMI) and processing speed index (PSI) were impaired in the majority of children.

Pre-operatively, all patients had variable levels of developmental delay affecting different cognitive domains with documented regression. Furthermore, all patients experienced learning disability or poor academic performance. In addition, the majority of children displayed behavioral abnormalities, including impaired peer relations and social difficulties, aggression, hyperactivity, inattention, and emotional lability. Comorbidities included attention deficit hyperactivity disorder (ADHD), learning disorder, anxiety, and depression.

Radiography

MRI was performed for all children, and pertinent findings included hemispheric PMG, FCD, cerebral atrophy, focal sclerosis, and T2/FLAIR hyperintensity (table 2). Two patients had prior cerebral vascular insults. One child had hemispheric cystic encephalomalacia involving the thalamus (following prior tumor surgery and complicated by stroke) with additional contralateral white matter changes. Another patient had FCD and gliosis adjacent to a prior anterior temporal lobectomy (ATL) resection cavity. The child with prior MCA stroke had a large developmental porencephalic lesion and open-lip schizencephaly. In order to document thalamic involvement in ESES (Guzzetta et al., 2005), particular attention was paid to the appearance of the ipsilateral thalamus on radiography. Five children displayed thalamic changes on MRI (encephalomalacia, T2/FLAIR signal changes, and atrophy).

All children underwent either PET or SPECT, which revealed areas of hypo- or hyper-metabolism/perfusion concordant with the EEG findings, and provided additional lateralizing confirmation to support operative management.

Neurophysiology

All patients underwent VEEG monitoring preoperatively, with hemi-ESES isolated to the left hemisphere in eight children and right hemisphere in one child (table 2). Two children had bilateral ESES suggesting two-way (Pertola Type III) interhemispheric propagation. Electroclinical seizures during wakefulness were captured on VEEG in nine patients, showing epileptiform discharges from abnormal cerebral foci that were concordant with radiographic findings. In all cases, the ictal epileptic focus during wakefulness was located in the same hemisphere as the ESES electrographic findings.

Surgery

In this cohort, seven children underwent focal cortectomy and resection, and four children underwent hemispherectomy and disconnection (table 4). Surgery was undertaken in children following documentation of cognitive performance based on serial neuropsychological tests and deterioration in school performance. The period of time necessary to confirm these parameters was variable, but was never less than a year from the diagnosis of ESES. Children who responded to a period of conservative medical management with diazepam, prednisone and AEDs continued to be managed medically and followed in our Comprehensive Epilepsy Program.

Focal cortectomies were performed for children with VEEG and MRI findings that supported a localized epileptogenic focus, accompanied by a consistent seizure semiology and history. In some cases, operative adjuncts including subdural grid electrodes and intra-operative electrocorticography (ECoG) were used to guide the extent of resection. One child with left fronto-temporo-parietal and insular PMG, left hemisphere ESES, and a left temporo-parietal epileptic focus underwent subdural grid insertion followed by focal resection (Case 3). Another child with left ESES and MRI findings of basal ganglia and hemispheric signal abnormality, underwent ECoG-guided resection to achieve a maximally safe resection (Case 10).

Four children underwent functional peri-Sylvian hemispherectomies, one of whom required a subsequent anatomical hemispherectomy for persistent seizures (Case 6). This child had significant global delay due to daily seizures, and underwent surgical disconnection of multiple areas of neurologic injury and bilateral encephalomalacia, with a VEEG that revealed bilateral ESES and bilateral epileptogenic foci. Unfortunately, she experienced no seizure relief following the functional-, and later anatomical- hemispherectomies. Another patient with left hemispheric PMG and thalamic atrophy and hemispheric slowing and spike-wave discharges also underwent functional hemispherectomy (Case 2). A third patient underwent hemispherectomy for a large area of fronto-temporo-parietal PMG with multi-focal epileptiform discharges (Case 5). Lastly, a patient with left-sided developmental porencephaly underwent endoscopic disconnection of the posterior quadrant, essentially equivalent to a functional hemispherectomy (Case 11).

Pathological tissue analysis was unavailable for the patient undergoing endoscopic disconnection; abnormal pathological findings were observed in 10 specimens including FCD or PMG (n=7) and gliosis (n=3).

Clinical and neurophysiological outcome

There were no procedure-related complications, and all patients were discharged within a few days of surgery. Nine children experienced improvement in seizure frequency, and four children had complete seizure freedom (Engel Class I) at the time of the most recent follow-up visit.

Children with extensive cerebral injury and bilateral ESES activity pre-operatively experienced little or no seizure improvement (Engel Class III and IV). One child (Case 2) became seizure-free after a right hemispherectomy, but subsequently developed seizure recurrence 14 months post-operatively. Investigation with ictal SPECT revealed increased perfusion in the right basal ganglia and thalamus, and VEEG revealed nearly continuous independent epileptiform discharges in the right frontal and temporal regions. Another child with epilepsy due to an extensive CNS insult from choroid plexus papilloma and hemorrhage had no improvement at the last follow-up visit (Case 6). Unfortunately, she continued to experience intractable epilepsy despite extensive medical and surgical treatment. She was not assessed with post-operative VEEG.

Parents reported subjective neurocognitive, academic, social, and behavioral improvement in eight children, although all exhibited selective persistent deficits (table 4). Improved academic performance and learning included higher school achievement and improved cognition, processing speed, memory, verbal skills, and reading. Some children were able to return to regular school with good academic performance. Social and behavioral improvement included better interactions with family and schoolmates, increased social insight, and greater independence. Improved behavior and mood, including amelioration of anxiety and depressive symptoms, were also noted. Three children had significant developmental delay at baseline and subsequently did not experience neuropsychological improvement.

Ten patients underwent post-operative EEG or VEEG (table 4). Of those, two had persistent ESES and six had resolution of ESES with some persistent abnormal electrographic activity. Notably, two children had post-operative normal EEGs with no evidence of ictal or interictal abnormalities in either the waking or sleep states (Case 7 and 10). The resolution of VEEG findings was concordant with improved seizure and neurocognitive outcomes. One child (Case 7) improved significantly in both social and academic domains; he was more engaging with friends, learned new material faster, and displayed improved focus and attention. He had only occasional seizures. The other child (Case 10) had complete resolution of seizures and improved neurocognitive outcomes. Post-operatively, she was able to resume normal activities, return to regular school, travel, and exhibited improved mood and behavior. Both children with normal EEGs were weaned off their AEDs completely.

Discussion

ESES is an age-dependent disorder associated with epilepsy, cognitive decline and behavioral impairment. It is a relatively new entity, first described in 1971 in six children with continuous spike-wave discharges during non-REM sleep that subsided on awakening. The clinical presentation of children with ESES is variable, ranging from early-onset developmental delay to global cognitive regression with clinical seizures (van den Munckhof et al., 2015). The average age at ESES diagnosis is 6.9 years (range: 1-11.5 years), and most present with seizures, although it is absent in 20% of cases (van den Munckhof et al., 2015). The average age at ESES diagnosis in our cohort (7.1 years) is consistent with published studies, although our surgical series contains a disproportionate number of children with symptomatic seizures and neurocognitive decline.

As a relatively new entity, the pathophysiology and natural history of ESES is poorly understood. The occurrence of GRIN2A mutations in familial cases (Carvill et al., 2013; Lemke et al., 2013; Lesca et al., 2013; Lesca et al., 2019), underlying inflammatory disorders, or prior cerebral insults have been postulated as potential etiologies. However, the etiology was unknown in almost half of patients in a recent meta-analysis of 575 cases (van den Munckhof et al., 2015). Early recognition and effective therapy are crucial to improving the long-term neurocognitive prognosis. The natural history of ESES is thought to be a self-limiting condition, usually ceasing around puberty. However, long-term neuropsychological sequelae often remain, with permanent cognitive impairments estimated at 40%-60% (Kramer et al., 2009; Liukkonen et al., 2010a). Therefore, the major goal of ESES treatment is to resolve seizures and halt neurocognitive decline.

Medical therapy improves or abolishes ESES in only half of patients (Veggiotti et al., 2012). Furthermore, most symptomatic ESES non-responders to medical therapy develop cognitive decline (Caraballo et al., 2013) and less than half regain baseline cognitive status (Guerrini et al., 1998; Liukkonen et al., 2010b; Veggiotti et al., 2001, 2012; Smith and Hoeppner, 2003). Although neuropsychological and seizure outcomes respond to medical therapy in many cases, surgery remains an important option, especially for children with ESES secondary to a structural lesion. Specifically, based on the recent meta-analysis, surgery led to neurocognitive and epilepsy improvement in 80% of cases, exceeding the benefit of AEDs, benzodiazepines, and steroids (van den Munckhof et al., 2015). We acknowledge, however, that our cohort is still small and that our findings await confirmation from additional studies with larger samples.

Although other authors have reported small case series with resective procedures, or larger series with MST for children with LKS (table 1), to our knowledge, this is the largest case series with resective and disconnective surgical management for ESES reported to date. Heterogeneous pathologies including PMG, FCD, hypoxia, and hemorrhage were represented in our cases. All seven children who underwent focal cortical resections for hemi-ESES improved postoperatively. Their favorable outcomes are consistent with the hypothesis that their declines were secondary to epileptic encephalopathy and pathology-independent. Thus, similar to other patients with a focal seizure etiology, localized excision of the seizure-onset zone in children with ESES is also beneficial.

A less favorable outcome in two of the four children who underwent hemispherectomies was related to incomplete resection of the epileptogenic zone. Both patients evidenced bilateral ESES pre-operatively, and persistent electrographic ESES post-operatively. One patient made no recovery (Engel Class IV), while the other had minor improvement in seizure frequency (Engel Class III). There were no neuropsychological, behavioral or intellectual improvements in either child. Of note, the presence of bilateral ESES likely heralds a poorer epileptic prognosis when bilateral ESES is asynchronous with bi-directional interhemispheric propagation (Peltola type 3 pattern), suggesting incomplete removal of ESES-generating cortex with uni-hemispheric surgery.

Early surgical intervention, especially in medical non-responders, can halt cognitive decline and abolish epileptic EEG patterns. We noted similar results as cognitive regression was halted in the majority of our patients and gains were obtained in a small subset (table 4). However, some degree of cognitive and behavioral impairment persisted in all patients, suggesting that surgery cannot fully reverse pre-existing cognitive deficits. A similar outcome has been reported for conventional paediatric epilepsy surgery for other epileptic encephalopathies (Roulet-Perez et al., 2010).

Bartolini et al reported on a cohort of 27 patients with ESES in polymicrogyric cortex and noted that complete remission occurred by age 13 years (Bartolini et al., 2016). This long-term study confirmed that the natural evolution of ESES involves spontaneous remission in patients with this etiology. While this outcome might favor medical rather than surgical treatment, this cohort ultimately evidenced variable degrees of cognitive disability. It therefore remains unclear whether earlier surgical intervention might have prevented some of the permanent cognitive sequelae.

In this small case series of 11 children who underwent surgery for ESES, resection or disconnection improved seizure control and neuropsychological outcome in most patients. However, children with non-resected or bilateral ESES have persistent seizures and neuropsychological impairment. In ESES, the aim of treatment is to control both clinical seizures and paroxysmal EEG abnormalities, in order to improve motor, language, and cognitive functions. Although our cohort is too small to identify statistically significant risk factors, reduced epileptic and neurocognitive improvement occurred in patients who were more severely cognitively impaired and exhibited incompletely resected or bilateral epileptogenic and ESES foci, and non-resolution of postoperative ESES. Although the numbers are too small to be conclusive, these findings are reported in prior surgical series (table 1). Improved post-operative cognitive status is uncommon, and typically occurs in children with less severe developmental delay, preoperative IQ >75, and shorter time to treatment.

Although ESES is a relatively new entity with poorly understood etiology and natural history, increasing evidence advocates for early recognition and aggressive management. It would be reasonable to consider a six to 12-month course of escalating medical treatment, involving AEDs initially, and advancing to benzodiazepines, corticosteroids, and IVIG. If the child continues to experience persistent seizures or neuropsychological decline, surgical management in the form of resection or disconnection should be considered. As seen in this series and the recent literature, surgery is most favorable for developmentally normal children with early documented neuropsychological regression and be considered for patients with known ESES etiology, hemi-ESES findings, and focal epileptogenic focus and radiographic findings. Additional research on the optimal timing of surgery and the factors associated with improved outcome would be beneficial for the management of children with this new and rare entity.

Disclosures

None of the authors have any conflict of interest to declare.