The first description of “subclinical electrical status epilepticus induced by sleep in children” dates back to 1971, when Patry, Lyagoubi, and Tassinari (Patry et al., 1971) described a peculiar electroencephalography (EEG) pattern occurring almost continuously during sleep in six children and characterized by apparent “subclinical” spike and wave complexes lasting from months to years. Five of them had mental retardation and two had no acquired language. Subsequently, the term “SES” or “ESES” (electrical status epilepticus during sleep) was introduced, and a connection between electroencephalographic pattern and clinical signs of cognitive impairment was suggested (Dalla Bernardina et al., 1978). The expression “continuous spikes and waves during slow sleep” (CSWS), dates back to 1989 when CSWS was included in the group of syndromes of undetermined origin (focal or generalized) by the Commission on Classification and Terminology of the International League Against Epilepsy (1989).
ESES is an age-related, self-limited disorder characterized by the following:
- Neuropsychological impairment in the form of global (Morikawa et al., 1995; Tassinari et al., 2000; Veggiotti et al., 2001) or selective cognitive or language regression (Beaumanoir et al., 1995; Deonna & Roulet, 1995; Debiais et al., 2007);
- Motor deterioration including ataxia, dyspraxia, dystonia (Neville et al., 1998), or unilateral deficits (Veggiotti et al., 1999, 2005);
- Epilepsy, with focal or generalized seizures (unilateral or bilateral clonic; tonic–clonic; absence; focal-motor; complex-partial; epileptic negative myoclonus);
- Typical EEG findings characterized by continuous epileptiform activity occupying more than 85% of non–rapid eye movement (NREM) sleep and persisting in at least three EEG recordings over a period of at least 1 month (Morikawa et al., 1989; Beaumanoir, 1995; Rossi et al., 1999; Tassinari & Rubboli, 2006). To date, a significant increase in EEG abnormalities during sleep with the presence of neurocognitive impairment is considered a hallmark of ESES.
Several studies have investigated the long-term evolution of CSWS (Veggiotti et al., 2002; Praline et al., 2003; Liukkonen et al., 2010; Seegmüller et al., 2012) and identified prognostic factors that affect cognitive outcome. Our study reports on the long-term follow-up of 25 subjects sharing the same EEG pattern, and describes significant differences in terms of cognitive outcome.
- Top of page
All 25 patients were assessed with standardized Wechsler scales in different moments of their clinical history. Twenty-one (84%) showed worsening of general cognitive abilities with a mean decrease of 14.00 points (range −41 to −1) in global IQ; 12.15 points in verbal tasks (range −35 to −2); and 13.30 points in performances items (range −40 to −2). The remaining four patients (16%) presented a severe brain damage associated with a complex neurologic condition prior to CSWS onset. In these patients, modifications in the global IQ were minimal during ESES and always below the normal range (range 40–70).
We stratified our sample in five subgroups in relation to the final outcome.
Group 1 (cases 1–7; see Fig. 1) included seven patients (28%) without cerebral lesions, who manifested a decline in language and performance tasks during the CSWS pattern, and an improvement during the regression of EEG abnormalities. This group shows a typical evolution of the neuropsychological profile with deterioration of cognitive abilities during the florid phase of ESES, and improvement during the regression of epileptic abnormalities. During the florid phase of CSWS, the global IQ declined below the normal average in four patients (57%). The mean increase was 22.00 points in global IQ (range 8–44), 14.70 in verbal IQ (range 3–35), and 26.00 in performance IQ (range 3–50). After CSWS remission all patients reached a global IQ within the normal range (87–128). Group 1 final outcome was interpreted as positive.
Figure 1. (A) Group 1 IQ (“typical evolution”). (B) Group 1 evolution of patient 1. QIT1, global IQ before ESES onset; QIT2, global IQ during ESES; QIT3, global IQ after ESES; QIT N, normal global IQ; QIT, global IQ; QIV, verbal IQ; QIP, performance IQ.
Download figure to PowerPoint
In group 2 (cases 8–10; Fig. 2), the electroencephalographic pattern of CSWS was associated with a mainly motor clinical impairment. This group includes three patients (12%) without cerebral lesions. We observed only minimal variation in the global IQ, which ranged between 80 and 101 before CSWS onset, and between 81 and 94 after CSWS disappearance. The mean decrease scores were 7.00 points in global IQ (range −16 to +1); 9.00 in verbal IQ (range −17 to −2); 4.00 in performance IQ (range −26 to +10). Two patients showed a decline in the global IQ below the normal average during the CSWS florid phase (global IQ, respectively, of 67 and 68); however, all patients reached a final IQ (after remission of the EEG pattern) within the normal average (range 81–94).
Figure 2. (A) Group 2 IQ (“motor” ESES). (B) Group 2 evolution of patient 9. QIT1, global IQ before ESES onset; QIT2, global IQ during ESES; QIT3, global IQ after ESES; QIT N, normal global IQ; QIT, global IQ; QIV, verbal IQ; QIP, performance IQ.
Download figure to PowerPoint
In group 3 (cases 11–17; see Fig. 3) cognitive deficit persisted despite the normalization of the EEG pattern. This group consisted of seven patients (28%); in five of them the CSWS pattern lasted more than 2 years; two of them had more than three EEG relapses treated with cycles of corticosteroids. Two patients showed a cognitive deficit before the CSWS onset (global IQ, respectively, of 68 and 64), all patients presented a decline in global IQ below the normal average (range 41–78), and none of them reached a global IQ within the normal range (35–78) at ESES regression. All patients showed a decline in language and performance tasks during and after the ESES pattern. Group 3 final outcome was negative.
Figure 3. (A) Group 3 IQ (negative evolution). (B) Group 3 evolution of patient 14. QIT1, global IQ before ESES onset; QIT2, global IQ during ESES; QIT3, global IQ after ESES; QIT N, normal global IQ; QIT, global IQ; QIV, verbal IQ; QIP, performance IQ.
Download figure to PowerPoint
Group 4 (cases 18–23; see Fig. 4) includes patients with a cerebral lesion prior to ESES onset associated with a complex neurologic clinical picture. The global IQ showed minimal improvement at the last follow-up; however, it remained constantly below the normal range, with values of 30–69 before ESES onset and 38–70 after ESES remission.
Figure 4. (A) Group 4 IQ (severe cerebral lesion). (B) Group 4 evolution of patient 18. QIT1, global IQ before ESES onset; QIT2, global IQ during ESES; QIT3, global IQ after ESES; QIT N, normal global IQ; QIT, global IQ; QIV, verbal IQ; QIP, performance IQ.
Download figure to PowerPoint
Group 5 (cases 24–25; see Fig. 5) includes two patients (8%) with a progressive deterioration in neuropsychological tests without temporal correlation with ESES duration (less than 7 months), and no association with clinical and electroencephalographic relapses. No metabolic or structural lesions were detected and the global IQ was, respectively, 71 and 106 at ESES onset. During the florid phase of ESES the IQ dropped to 70 and 102, respectively; however, upon ESES remission a further decrease was observed (64 and 79, respectively). The mean reduction in global IQ was 7.00 points (range −7), 15.50 points in verbal IQ (range −20 to −11), and 5.00 points in performance IQ (range −9 to −1). ESES onset was, respectively, at 10.20 years and 4.60 years, with ESES duration of 6 and 7 months (Table 1).
Table 1. Summary of clinical data before ESES onset
|Cases/sex||Etiology (MRI findings)||Age at onset of epilepsy||Epilepsy||Type of seizures||Neurologic problem before ESES|
|GROUP 1 (patients with “typical evolution”)|| || || || || |
|1/M||Unknown||3 year||LKS||Focal motor||Dysphasia – ADHD|
|2/F||Unknown||4 year||AR||Complex partial||Delayed speech|
|3/M||Unknown||5 year||AR||Complex partial||Normal|
|4/M||Unknown||3 year 4 month||AR||Focal motor||Normal|
|5/F||Unknown||4 year 8 month||LKS||Focal motor||Aphasia|
|7/F||Unknown||4 year 2 month||AR||Complex partial||Normal|
|GROUP 2 (patients with “motor ESES”)|| || || || || |
|9/F||Unknown||3 year 5 month||LKS||ENM||Normal|
|GROUP 3 (patients with “negative evolution”)|| || || || || |
|11/M||Prenatal Chlamydia||5 year||AR||Complex partial, SGTCS||Delayed motor development|
|12/M||Unknown||4 year||LKS||Complex partial, AA||Normal|
|14/M||Pre/perinatal vascular||5 year||AR||Complex partial, AA||Normal|
|15/M||Perinatal vascular (PLV thalamus sx)||7 year||LKS||GTCS||Global delay, aphasia|
|16/M||Pre/perinatal vascular (PVL)||5 year 6 month||AR||Focal motor||ADHD|
|17/F||Unknown||2 year||AR||Complex partial||Normal|
|GROUP 4 (patients with other neurologic conditions)|| || || || || |
|18/F||Neoplasia – ganglioglioma||5 year||LKS||Focal motor||Dysphasia|
|19/F||Perinatal CMV||1 year||Structural||Focal motor||Global delay|
|20/F||Hydrocephaly||6 year||Structural||Complex partial||Global delay|
|21/F||Perinatal vascular (PVL sx)||1 year||Structural||Focal motor, SGTCS||Hemiplegia|
|22/M||Malformation (Bilateral dysplasia)||8 year 9 month||Structural||GTCS||Global delay|
|23/M||Pre/perinatal vascular (PLV thalamus dx)||7 year||Structural||GTCS||Spastic tetraparesis|
|GROUP 5 (patients with inexplicable evolution)|| || || || || |
|24/F||Unknown||4 year 6 month||AR||SGTCS||Normal|
|25/F||Unknown||9 year||AR||Focal motor, SGTCS||Normal|
Figure 5. (A) Group 5 IQ (inexplicable ESES). (B) Group 5 evolution of patient 24. QIT1, global IQ before ESES onset; QIT2, global IQ during ESES; QIT3, global IQ after ESES; QIT N, normal global IQ; QIT, global IQ; QIV, verbal IQ; QI, performance IQ.
Download figure to PowerPoint
The ESES duration is significantly different among the five groups (F = 6.82, p < 0.001); post hoc tests with Bonferroni method further confirm that ESES duration in group 3 (negative outcome) (mean 28.14) is longer than those of group 1 (positive outcome) (mean difference 15.00, standard error [SE] 3.57, p < 0.001), group 2 (mean difference 21.48, SE 4.61, p < 0.001), group 4 (mean difference 13.48, SE 3.72, p < 0.01), and group 5 (mean difference 21.64, SE 5.35, p < 0.01).
- Top of page
In follow-up studies of patients with ESES, indicators of long-term evolution emerge. Veggiotti et al. (2002) described the long-term neuropsychological follow-up of five patients with ESES but with three different types of outcome: acquired frontal dementia, language deficits, and normal evolution. Four patients initially diagnosed with Landau-Kleffner syndrome showed clearly different outcomes in terms of severity and type of impairment. The authors concluded that the initial diagnosis does not have any predictive significance and focused their attention on some prognostic factors that are more useful in identifying the long-term outcome: length and age of onset of CSWS, location of epileptiform activity, and individual neuropsychological profile.
Praline et al. (2003) studied seven young adults, five of them with CSWS syndrome and two with Landau-Kleffner syndrome during childhood. After assessing IQ, spoken and written language, and executive functions, they concluded that although the epilepsy associated with these syndromes has a good prognosis, neuropsychological disorders persisted. In fact, only two patients attended mainstream school, three of five patients with CSWS syndrome presented some learning disabilities, and the two patients with Landau-Kleffner syndrome continued to show severe and disabling language disturbances. The authors identified some prognostic factors such as the location of interictal electric focus and age of ESES onset.
Liukkonen et al. (2010) evaluated 32 children diagnosed and prospectively followed for at least 3 years (neuropsychological follow-up for at least 5 years was available in 18 patients): Six children had atypical rolandic (AR) epilepsy, nine Landau-Kleffner syndrome, and 17 symptomatic epilepsy. Before ESES onset, 20 children had normal IQ. A regression of ESES pattern was observed in 16 patients after pharmacologic treatment. Ten children (31%) (four [67%] with AR, three (33%) with Landau-Kleffner syndrome, and three (19%) with symptomatic etiology), including nine with treatment response, regained pre-ESES cognitive level. The authors concluded that most children with ESES experience permanent cognitive impairment, and the unfavorable cognitive outcome is predicted by younger age at ESES diagnosis, lower IQ at the time of the diagnosis, and no response to drug treatment.
Seegmüller et al. (2012) reported on the long-term follow-up (15.6 years) of 10 adolescents and young adults with nonlesional epilepsy with CSWS. None of the patients fully recovered after ESES, yet four regained borderline to normal intelligence and were almost independent. Patients with prolonged global intellectual regression had the worst outcome. The authors suggested that cognitive recovery after CSWS regression depends on the severity and duration of the initial impairment. They concluded that early recognition of epilepsy with CSWS and rapid introduction of effective therapy are crucial for obtaining the best possible outcomes.
All patients in our sample experienced different levels of cognitive regression during ESES. We can, therefore, put forward the hypothesis of a temporal correlation between CSWS pattern and decline of cognitive functions; CSWS syndrome seems a “real” epileptic encephalopathy. Despite sharing the same ESES EEG pattern, the evolution in our patients is not always consistent with the pattern that would be typically expected.
Group 1 patients present the “typical evolution” described in the literature: IQ within the normal range before CSWS onset, deterioration in cognitive ability during the CSWS EEG pattern, and progressive improvement with rescue of cognitive functions after CSWS remission. These patients do not have detectable brain damage, and ESES onset is between 4 and 13.2 years (mean 6.53 years); all patients in this group show a good response to antiepileptic medications without relapses. Mean total duration of ESES was 13.1 months (range between 7 and 24 months).
Patients in group 2 show prevalent motor deficits and ESES onset from 2.2 to 8.2 years (mean 5.5 years). Considering the clinical manifestations of this group, it seems possible to hypothesize that the “encephalopathy” mainly affects nonassociative areas and/or site of cognitive functions. Mean ESES duration was 6.6 months (range from 6 to 8 months).
Patients in group 3 present cognitive impairment due to the epileptic encephalopathy: rather than the symptomatic etiology, this outcome seems associated to the long duration of the EEG pattern and/or to nonresponsiveness to the drug treatment. The mean ESES duration was 28.1 months (range from 24 to 36 months) with a statistically significant difference (15 months) compared to the first group (t = 5.34; p < 0.001).
Group 4 is characterized by an evident brain lesion that interferes with cognitive functions. In this group ESES pattern only marginally affects the outcome. ESES onset was between 5 and 9.3 years (mean 7.4 years). The mean duration of ESES (14.6 months) is between the first and third group.
Group 5 displays the most “atypical evolution.” The two affected subjects show a progressive decline in cognitive functions, also after the regression of the EEG pattern. They do not share any of the previously cited negative prognostic factors. The onset is between 4.6 and 10.2 years. The mean duration of ESES was 6.5 months (range from 6 to 7 months), which is lower than in group 1 (Table 2).
Table 2. Summary of CSWS period and pharmacologic data
|Case/sex||Age of CSWS onset||CSWS duration||Type of treatment||Age treatment began||Treatment duration||CSWS response (improvement of EEG pattern and/or neurocognition)|
|GROUP 1 (patients with “typical evolution”)|| || || || || || |
|1/M||4 year||2 year||VPA+LTG, CORT, LEV+ESM||6 year 11 month||8 year||CORT (EEG pattern)|
|2/F||8 year 2 month||1 year||OSP||8 year 2 month||2 year||OSP (EEG pattern)|
|3/M||7 year 5 month||9 month||VPA+CLB+LTG+ACTH, VPA+ESM+OSP+CORT, VPA+ESM||7 year 5 month||2 year 10 month||VPA+ESM+OSP+CORT (EEG pattern /neurocognition)|
|4/M||4 year 11 month||11 month||CZP, VPA+ESM, VPA+LTG||4 year 11 month||7 year 5 month||VPA+ESM (EEG pattern/neurocognition)|
|5/F||4 year 5 month||7 month||ACTH, CORT, VPA+CZP+FBM, VPA+CORT+CLB, VPA+CZP+ACTH, CZP||4 year 5 month||3 year||VPA+CORT+CLB (EEG pattern/neurocognition)|
|6/F||13 year 2 month||1 year 6 month||VPA, VPA+CLB||13 year 2 month||2 year 7 month||VPA+CLB (EEG pattern/neurocognition)|
|7/F||4 year 2 month||11 month||CLB||5 year 1 month||1 year||CLB (EEG pattern /neurocognition)|
|GROUP 2 (patients with “motor ESES”)|| || || || || || |
|8/F||2 year 2 month||6 month||ESM, CBZ, CORT, VPA||2 year 2 month||4 year 4 month||VPA + ESM (EEG pattern)|
|9/F||8 year 2 month||8 month||CBZ, VPA+CLB+OSP, VPA+OSP, VPA+LEV, VPA+ESM+CORT, OSP||8 year 2 month||5 year 11 month||VPA+ESM+CORT (EEG pattern/neurocognition)|
|10/F||6 year 3 month||6 month||VPA+CLB, OSP+VPA, VPA+CLB+OSP||6 year 3 month||3 year||VPA+CLB+OSP (EEG pattern/neurocognition)|
|GROUP 3 (patients with “negative evolution”)|| || || || || || |
|11/M||6 year 8 month||2 year||VPA+CZP, VPA+CORT||6 year 8 month||3 year 2 month||VPA+CORT (EEG pattern/neurocognition)|
|12/M||6 year||2 year 6 month||PB+VPA, BDZ+CORT, VPA+CBZ+BDZ, ESM+VPA||6 year||12 year||CORT (EEG pattern/neurocognition)|
|13/F||2 year 7 month||2 year 4 month||CLZ+PB, VPA+CLB+CBZ, VPA+CLZ+OXC, VPA+LEV+CLZ||2 year 7 month||8 year 1 month||VPA+LEV+CLZ (EEG pattern)|
|14/M||11 year||2 year||CLB,VPA,LEV, ESM,LTG,TOP, CORT,OSP||5 year||10 year 5 month||CORT+CLB (EEG pattern)|
|15/M||10 year 10 month||3 year||VPA, VPA+CLB, VPA+CORT||8 year||7 year||VPA+CORT (EEG pattern/neurocognition)|
|16/M||6 year 5 month||2 year||VPA+ESM, LEV+ESM, VPA+LEV, OSP+ESM, VPA||6 year 5 month||14 year 2 month||VPA+LEV (EEG pattern)|
|17/F||9 year||2 year 7 month||VPA+LEV, VPA+ESM, OSP+LEV, OSP, VPA+OSP+LEV||9 year||6 year 10 month||LEV+OSP (EEG pattern)|
|GROUP 4 (patients with other neurologic conditions)|| || || || || || |
|18/F||5 year||2 year 9 month||VPA+CLB, LEV+ESM, VPA+ESM, OSP, CORT||5 year 3 month||7 year 8 month||CORT (EEG pattern/neurocognition)|
|19/F||7 year 9 month||1 year 6 month||VPA,OSP,TOP, CLB,RUF||7 year 9 month||2 year 3 month||VPA+TOP+OSP+CLB (EEG pattern /neurocognition)|
|20/F||7 year 11 month||3 month||PB+VPA+CBZ+ CZP, LEV+OSP, LEV+CZP, OSP||6 year||7 year 3 month||LEV+OSP (EEG pattern/neurocognition)|
|21/F||9 year 3 month||1 year 4 month|| ||9 year 9 month||8 year||VPA+CLB/surgery (EEG pattern /neurocognition)|
|22/M||9 year||7 month||VPA, VPA+CLB, LEV+CORT, LEV+ESM+CORT, LEV+ESM, TOP+ESM, VPA+LEV||9 year||3 year 5 month||VPA+CORT (EEG pattern/neurocognition)|
|23/M||6 year 4 month||11 month||CLB, VPA, VPA+CLB, VPA+LEV+CORT, VPA+LEV||6 year 6 month||4 year 4 month||VPA+LEV+CORT (EEG pattern)|
|GROUP 5 (patients with inexplicable evolution)|| || || || || || |
|24/F||4 year 6 month||7 month||CLB||5 year 1 month||6 month||CLB (EEG pattern /neurocognition)|
|25/F||10 year 2 month||6 month||VPA+LEV, VPA+OSP||10 year 2 month||4 year||VPA+OSP (EEG pattern/neurocognition)|
The first factor to take into account is the etiology, as it seems to affect cognitive outcome. Where a brain damage or malformation is present, the “encephalopathy effect” is less detectable. In addition, our study suggests that ESES duration and poor response to the pharmacologic treatment, in terms of number of electroclinical and clinical relapses, may be considered as possible indicators of long-term evolution. The analysis of variance (ANOVA) results in fact indicate a significant difference in terms of ESES duration. The response to the treatment is also different in the two groups, with an initial modification of the EEG pattern followed by its reappearance in group 3 and absence of any relapse in group 1.
In group 2 the impairment is mainly motor and negative myoclonus is present. Here the “encephalopathy effect” does not seem evident at a cognitive level. Unfortunately, it was not possible to perform functional tests; however, from the clinical manifestation of these patients it is plausible to hypothesize that the most extensively involved area is the motor area with no involvement of associative areas (“motor ESES”). This finding would corroborate our earlier suggestions regarding the role of the site of epileptic activity (Veggiotti et al., 2002).
Data from group 5 escape any straightforward interpretation because cognitive decline seems to proceed without any clear correlation with the EEG pattern. A possible interpretation is the recognition of an underlying disease. However, this finding may also reflect a limitation of our study, which did not use the new techniques of cortical EEG sources (Brazzo et al., 2012).
Does ESES affect clinical outcome always in the same way? Findings from our study support a negative answer. The total duration of the EEG pattern, the unresponsiveness to pharmacologic treatment, and the site of epileptic activity seem important predictors of long-term outcome. ESES is not a “real” syndrome; rather, it is an EEG pattern that can modify the clinical evolution of the underlying disease and patients’ neuropsychological profile while not uniformly affecting the clinical outcome. It is thus important to adopt an aggressive therapeutic approach to immediately stop a status condition and prevent cognitive and neuropsychological deterioration secondary to long-term EEG abnormalities. Based on the literature and on our experience, the use of several AED therapies does not lead to the disappearance of the EEG pattern and, thus, the early use of corticosteroids is crucial (Veggiotti et al., 2012). Because our study is retrospective, the limitation of not using the new techniques for EEG signal analysis is worth noting. As suggested by Bölsterli et al., (2011) and Cantalupo et al., (2011), the normal overnight decrease of incidence of slow waves during NREM sleep and the decline of slow-wave slope in the course of the night are absent in patients with ESES (Seri et al., 2009). This observation supports the hypothesis that spikes, which are restricted to NREM sleep, interact with the restorative processes during slow-wave sleep. The overnight decrease of EEG slow-wave slopes was linked to slow-wave sleep downscaling. ESES is associated with cognitive regressions, and the downscaling process is thought to be a plasticity process important for learning and memory. This lack of downscaling in children with ESES may explain cognitive deterioration. These findings may reflect disruption of sleep-dependant synaptic homeostasis that could interfere with cortical plasticity. Further studies are needed to analyze together sleep physiology and epilepsy. These might provide electrophysiologic evidence that any function related to slow waves is impaired in patients with ESES, and may suggest a new mechanism leading to the global or specific cognitive and behavioral regression observed in children with ESES.
Prospective studies with temporally related overnight sleep EEG and detailed neuropsychological assessments are needed to further investigate the interaction between continuous spike waves and sleep functions.