Guidelines for EEG in encephalopathy related to ESES/CSWS in children


Address correspondence to Marjan Scheltens-de Boer, MD, Department of Clinical Neurophysiology, room BA 400, Erasmus Medical Centre’s Gravendijkwal 230, 3015 CE Rotterdam, The Netherlands. E-mail:


Electrical status epilepticus during slow sleep (ESES) or continuous spikes and waves during slow sleep (CSWS) is a phenomenon characterized by strong activation of epileptiform activity in the electroencephalogram (EEG) during sleep. The literature contains several small series of patients and many case reports. Large prospective studies are lacking. Definitions of the syndromes and EEG criteria and methods vary, as does their classification. The fluctuating clinical course and EEG findings complicate the diagnostic process and evaluation of effect of therapy. Studies describing quantitative aspects of the epileptiform abnormalities in EEG are overrepresented in literature, whereas qualitative aspects are relatively undervalued. Guidelines for evaluation of the EEG in these syndromes, which focus on both aspects, are presented.

Several encephalopathy syndromes in children exist, in which the electroencephalogram (EEG) during sleep mostly shows a strong activation of epileptiform activity. Symptoms vary: deterioration of one or more cognitive functions with or without motor, behavioral, and/or psychomotor decline. Mostly these children have or develop epileptic seizures.

Tassinari et al. (1977) introduced the term electrical status epilepticus during slow sleep (ESES) for this EEG phenomenon, which was converted to status epilepticus during sleep (SES) (Tassinari et al., 2000). In 1989 the Commission on Classification and Terminology (CCT) of the International League Against Epilepsy (ILAE) introduced the more descriptive term continuous spikes and waves during slow sleep (CSWS). In this report I will focus on the EEG aspects and present guidelines for evaluation of the EEG in patients with ESES/CSWS-related syndromes.

Review of the Literature

In the literature many small series and case reports of patients with ESES/CSWS-related syndromes have been described, but large prospective studies are lacking. In the description of these patients, inconsistent criteria for the EEG parameters and clinical syndromes and other diagnostic tests are used, which hamper objectivity, comparison, and generalization.

The EEG parameters, which are used in the literature dealing with ESES/CSWS, roughly can be divided into quantitative and qualitative parameters.

Quantitative parameters

The percentage of epileptiform activity during sleep can be expressed as spike-wave index (SWI) (Tassinari et al., 2000), originally described as the percentage of (diffuse) spikes and waves during slow-wave sleep. Various criteria are used for ESES/CSWS: an SWI of at least 85% (Tassinari et al., 2000), 50% (Beaumanoir, 1995; Beaumanoir et al., 1995), 90% (Rossi et al., 1999), 60% (Inutsuka et al., 2006), and 25% (Van Hirtum-Das et al., 2006). Others merely mention a strong activation of epileptiform activity during sleep (CCT of the ILAE, 1989). By some authors no significant drop in performance is thought to occur when the SWI is lower than 85% (Beaumanoir, 1995; Beaumanoir et al., 1995;Guzzetta et al., 2005). The fluctuating clinical course and EEG findings complicate the diagnostic process and therapy evaluation (Stroink et al., 1997; Tassinari et al., 2000).

The methods to determine the SWI vary: considering the percentage of diffuse spikes-and-waves during the whole-night non-REM (rapid eye movement) sleep (Tassinari et al., 1977), percentage of spikes-and-waves during at least 15 min slow wave sleep (Lewine et al., 1999), percentage spikes-and-waves during the total duration of each cycle of slow wave sleep (Massa et al., 2000), the percentage of seconds with ≥1 spike-wave complex during the first 30 minutes non-REM sleep of the first and last sleep cycles (Aeby et al., 2005), SWI of at least one sleep–wake cycle (Saltik et al., 2005), SWI during the whole-night, first non-REM cycle or nap EEG (Inutsuka et al., 2006).

Although the importance of a high SWI during sleep is agreed upon (Morrell et al., 1995; Roulet Perez, 1995; Rossi et al., 1999; Eriksson et al., 2003; Smith & Hoeppner, 2003; Holmes & Lenck-Santini, 2006), relatively little attention is paid to the SWI during the awake stage. Some authors claim that epileptic activity itself hampers cognitive function (Seri et al., 1998; Gordon, 2000; Pearl et al., 2001; Holmes & Lenck-Santini, 2006; Praline et al., 2006; Tassinari & Rubboli, 2006).

Several authors developed grading scales to quantify the amount of epileptiform activity during sleep: Mikati et al. (2002) used a gradation of 0–4 [no spike-waves (SW); 0–25% SW; 25–50% SW; 50–75% SW; 75–100% SW, respectively]. Beaumanoir (1995) and Beaumanoir et al. (1995) distinguished four categories: CSWS** (>85%), CSWS* (50–80%), CSWS- (<50%) and CSWS0 (no epileptiform activity). Aeby et al. (2005) developed a grading system of four items containing epileptiform and nonepileptiform parameters.

The described ESES/CSWS-associated sleep stages also vary. Patry et al. (1971) only consider slow wave sleep as does the CCT of the ILAE (1989). Galanopoulou et al. (2000) stress sleep stages 3 and 4, Rossi et al. (1999) stages 1 and 2, and Genton et al. (1995) non-REM and REM sleep.

A variation of the percentage epileptiform activity during the night is mentioned sporadically (Galanopoulou et al., 2000; Tassinari et al., 2000), with predominance of epileptiform activity during the first part of the night.

Only incidentally the amplitude of epileptiform activity is considered (Veggiotti et al., 2001; Aeby et al., 2005).

Qualitative parameters

Morphology and frequency of epileptiform activity and discharges are mentioned infrequently (Bureau, 1995; Morikawa et al., 1995; Lewine et al., 1999; Tassinari et al., 2000; Guzzetta et al., 2005; Popović Miocinović et al., 2005, Saltik et al., 2005). The presence of dipoles is mentioned by Morrell et al. (1995), Galanopoulou et al. (2000), and Smith and Hoeppner (2003).

Distribution of epileptiform activity during sleep mostly is described as secondary bilateral synchronized (Kobayashi et al., 1994; Morrell et al., 1995; Galanopoulou et al., 2000; Tassinari et al., 2000; Aeby et al., 2005; Guzzetta et al., 2005); other descriptions are less frequent: hemi-ESES (Hirsch et al., 1995; Galanopoulou et al., 2000; Tassinari et al., 2000; Irwin et al., 2001; Veggiotti et al., 2001; Aeby et al., 2005; Guzzetta et al., 2005; Saltik et al., 2005); asymmetric ESES (Aeby et al., 2005; Saltik et al., 2005); bitemporal ESES or BTESES in Landau-Kleffner syndrome (Rossi et al., 1999); focal ESES or FES (Genton et al., 1995;Tassinari et al., 2000; Ballaban-Gil & Tuchman, 2000; Teixeira et al., 2007); and multiple foci (Aeby et al., 2005; Saltik et al., 2005).

Epileptiform foci during the awake stage are generally thought to be associated with symptomatology: For instance, receptive aphasia is related to the localization of spikes and waves in the temporal region, whereas behavioral disturbances and more global cognitive decline seem to be correlated with a frontal localization of epileptiform activity (Beaumanoir, 1995; Beaumanoir et al., 1995; Tassinari et al., 2000). Saltik et al. (2005) use an elaborate description of epileptiform activity according to localization.

The global picture of the epileptiform activity of the EEG during sleep sometimes is described, that is, fragmented (Veggiotti et al., 2001), continuous (CCT of the ILAE 1989), subcontinuous, and periodic.

Nonepileptiform abnormalities in the description of ESES/CSWS-related EEGs are mentioned less frequently: Background pattern (Rossi et al., 1999; Tassinari et al., 2000; Yan Liu & Wong, 2000; Irwin et al., 2001; Aeby et al., 2005), slow wave foci (Tassinari et al., 2000; Irwin et al., 2001; Aeby et al., 2005; Popović Miocinović et al., 2005; Saltik et al., 2005), (presence of) sleep stages and sleep phenomena (Nobili et al., 2001; Guzzetta et al., 2005; Saltik et al., 2005), and evolution of epileptic activity during sleep (Tassinari et al., 2000). Inclusion of these parameters probably yields a better correlation with clinical symptoms (Aeby et al., 2005).

Sleep architecture

Sleep architecture is disturbed, which by itself might have an adverse effect on cognitive function (Hirsch et al., 1995; Mikati et al., 2002).

Certain sleep parameters such as delta waves and sleep spindles might play a role in the activation of epileptic activity in sleep (Yung et al., 2000; Nobili et al., 2001; Guzzetta et al., 2005). Beaumanoir (1995), Beaumanoir et al. (1995) found a decrease in sleep spindles in patients with CSWS.

EEG methods

Several ways to register epileptiform EEG activity during wakefulness and/or sleep relating to ESES/CSWS are reported: ambulatory 24-h EEG (Yan Liu & Wong, 2000), polysomnography with or without video (García-Peñas, 2005; Guzzetta et al., 2005; Popović Miocinović et al., 2005; Inutsuka et al., 2006; Teixeira et al., 2007). Sleep EEGs (whole night registration or nap EEGs) with (Teixeira et al., 2007) or without sleep medication and with (Tachikawa et al., 2001; Eriksson et al., 2003), or without video registration and conventional awake EEG (Teixeira et al., 2007).

Clinical syndromes

Table 1 shows a summary of the ESES/CSWS-related syndromes and their most important EEG aspects (Gordon, 1997; Galanopoulou et al., 2000; Tassinari et al., 2000; Nickels & Wirrell, 2008).

Table 1.   Clinical syndromes
SyndromeFeaturesEpileptic focusSleep stage with high SWI and %
  1. CSWS/ESES, continuous spikes-and-waves during slow sleep/electrical status epilepticus during slow sleep; BECTS, benign epilepsy with centrotemporal spikes; REM, rapid eye movement; BTESES, bitemporal ESES.

CSWS/ESES syndromeGlobal cognitive decline, (atonic) seizures, motor disturbancesFrontocentral/frontotemporal Centrotemporal/FrontalNon-REM >85% or >50%, most diffuse
Landau-KleffnerReceptive/mixed aphasia, verbal agnosia, infrequent seizures(Centro)temporal/ posterior temporal/ parietooccipital, (vertical dipole)(Non-)REM Any % BTESES/unilateral/diffuse
Atypical BECTSMost nocturnal (partial) seizures with cognitive and behavioral disturbancesCentro(temporal),(horizontal dipole)(Non-)REM <50 or >50%, focal or diffuse
Opercular syndromeResembling BECTSCentrotemporal (bilateral)Non-REM (REM?) Any %

Generally autistic epileptiform regression (Tuchman & Rapin, 1997), Lennox-Gastaut syndrome (Galanopoulou et al., 2000; Tassinari et al., 2000), benign occipital epilepsy (Galanopoulou et al., 2000; Tassinari et al., 2000), and myoclonic-astatic epilepsy or Doose’s syndrome (Galanopoulou et al., 2000) are headed under differential diagnosis.

It is debated whether the epileptiform activity during ESES/CSWS in the absence of evident clinical seizures has an epileptic significance (Kallen, 2001). Opinions regarding the temporal relationship between the clinical course and EEG findings vary from a strict one, a delayed one, to no relationship at all (Paquier et al., 1992; Lagae et al., 1998; Rossi et al., 1999; Galanopoulou et al., 2000; Massa et al., 2000; Tassinari et al., 2000; Ducuing et al., 2004).

Proposed New Guidelines

Based on the preceding the following new guidelines may be formulated.

Strong activation (an SWI of at least 50%) of epileptiform activity during non-REM, and also sometimes REM-sleep, should trigger the possibility of an ESES/CSWS-related syndrome.

The distribution of epileptiform activity in the EEG during the awake stage and during sleep in patients with an ESES/CSWS-related syndrome can be focal, multifocal, unilateral, asymmetric bilateral, symmetric bilateral, diffuse, or more restricted. The ESES/CSWS pattern may be (sub)continuous, fragmented, or periodic.

Not only epileptiform activity during sleep and the awake stage but also background pattern, focal slow activity, sleep characteristics of the EEG, and sleep architecture probably play an important role in the determination of symptoms, their severity, and prognosis of the ESES/CSWS-related syndromes and should be described.

For scientific purposes, the SWI and discharges in a preferentially ambulatory 24-h EEG if possible with video registration should be described during the awake stage, per sleep stage, and during the course of the night (stable, decreasing, or increasing). For scientific purposes, extra sensors can be added [electro-oculography (EOG), pulse oximeter]; hypnogram can be made afterwards. In our center I developed the following scale for the SWI: 0 (no SW); 1 (0–20% SW); 2 (20–50% SW); 3 (50–85% SW); 4 (>85% SW) to facilitate comparison with literature. For a more detailed description it seems better to use the terms range, mean, and most encountered percentage epileptiform activity to differentiate a monotonous EEG pattern from a more fragmented or periodic one. In addition, epileptic discharges have to be described separately with accompanying clinical events if possible. This method favors selection of the most representative parts of the EEG in the future.

For clinical purposes, a nap EEG after sleep deprivation without sleep medication combined with an EEG during the awake stage probably will suffice. If there is a high suspicion of CSWS, one could expand the registration to a longer period, preferentially a 24-h EEG.

In long sleep registrations a selection of the first and last sleep cycle or part of the EEG probably will suffice to reliably quantify the epileptiform activity. The criterion of at least 50% epileptiform activity during non-REM and/or REM sleep seems most adequate, especially if the clinical picture fits a CSWS/ESES-related syndrome.

EEG follow-up: For clinical purposes, only on clinical suspicion of relapse or when doubt exists regarding the cause or clinical changes. For scientific purposes, frequent (initially, for instance, biweekly, and later on quarterly), EEGs are recommended in order to correlate EEG parameters with the clinical course.

Large prospective multicenter studies with clear uniform clinical and EEG criteria, such as formulated in the preceding text, have to be performed in the future for determination of the optimal SWI to define the syndromes and identification of possible other clinical important parameters in the EEG.


Drs G.H. Visser and J.C. Perumpillichira and Prof W.F.M. Arts are gratefully acknowledged for their help.

I confirm that I have read the Journal’s position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.

Disclosure: I have no conflicts of interest to disclose.