Do seizures and epileptic activity worsen epilepsy and deteriorate cognitive function?



Relevant to the definition of epileptic encephalopathy (EE) is the concept that the epileptic activity itself may contribute to bad outcomes, both in terms of epilepsy and cognition, above and beyond what might be expected from the underlying pathology alone, and that these can worsen over time. The review of the clinical and experimental evidence that seizures or interictal electroencephalography (EEG) discharges themselves can induce a progression toward more severe epilepsy and a regression of brain function leads to the following conclusions:

  1. The possibility of seizure-dependent worsening is by no means a general one but is limited to some types of epilepsy, namely mesial temporal lobe epilepsy (MTLE) and EEs.
  2. Clinical and experimental data concur in indicating that prolonged seizures/status epilepticus (SE) are a risky initial event that can set in motion an epileptogenic process leading to persistent, possibly drug-refractory epilepsies.
  3. The mechanisms for SE-related epileptogenic process are incompletely known; they seem to involve inflammation and/or glutamatergic transmission.
  4. The evidence of the role of recurrent individual seizures in sustaining epilepsy progression is ambiguous. The correlation between high seizure frequency and bad outcome does not necessarily demonstrate a cause–effect relationship, rather high seizure frequency and bad outcome can both depend on a particularly aggressive epileptogenic process.
  5. The results of EE studies challenge the idea of a common seizure-dependent mechanism for epilepsy progression/intellectual deterioration.

Epileptic encephalopathies (EEs) are defined as conditions in which cognitive, sensory, and/or motor functions deteriorate as a consequence of epileptic activity, which consists of frequent seizures and/or interictal paroxysmal activity (Dulac, 2001). The definition embodies the notion that the epileptic activity itself may contribute to severe cognitive and behavioral impairments above and beyond what might be expected from the underlying pathology alone (e.g., cortical malformation), and that these can worsen over time (Berg et al., 2010). In this section we review the clinical and experimental evidence that epileptic activity, either ictal or interictal, can per se induce a progressive worsening of epilepsies: both seizure frequency/severity and regression of brain function.

Clinical Evidence of Epileptic Activity–Related Deterioration

It is worth noting that most patients with epilepsy do not have a progressive disorder, as in both seizure worsening and intellectual decline. Indeed, many types of human epilepsies are known that do not progress despite seizure repetition: benign rolandic and occipital epilepsies, absence epilepsies, juvenile myoclonic epilepsy, benign familial neonatal and infantile seizures, and autosomal dominant nocturnal frontal lobe epilepsy, among the most common ones. Therefore, the hypothesis of a progressive worsening induced by seizures or interictal EEG discharges applies to only a limited number of human epilepsies: some types of focal epilepsy (and namely, mesial temporal lobe epilepsy; MTLE) and EEs.

Epidemiologic studies have assessed the risk of recurrence after a first unprovoked seizure (Haut et al., 2004). Patients with febrile seizures, postcraniotomy seizures, or posttraumatic seizures, treated and not treated after the acute event, show equal incidence of later epilepsy. Furthermore, studies from developing countries have found a high rate of spontaneous remission (about 50%) in individuals with untreated epilepsies (Osuntokun et al., 1987; Placencia et al., 1992).

Mesial temporal lobe epilepsy

Particular emphasis has been placed on the role of an episode of prolonged seizures/status epilepticus (SE) as the initiator of a process leading to chronic epilepsy. In several instances, the natural history of epilepsy indicates an initial precipitating event as an underlying cause of slowly progressive epileptic syndromes (Mathern et al., 1996). In some case (trauma, infection, autoimmune process) the initial event is associated with repeated seizures, often taking the shape of SE, which have been proposed to set into motion a sequence of slowly evolving cellular events, which presumably alter neural circuits and eventually render these circuits vulnerable to spontaneous synchronization and to the expression of seizures (see Sutula, 2004 for a review). This hypothesis is supported by many experimental data, which are reviewed in the next section of this article. It must be said, however, that, with the notable exception of MTLE (see below), the role of the epileptic activity, which is sometimes associated with the initial event, is not clearly demonstrable in a clinical setting. Several studies (Hauser et al., 1990; Shinnar et al., 1996; Kho et al., 2006) have shown that only symptomatic SEs correlate with brain damage and late epilepsy development, which raises the question of the respective role of epileptic activity versus the SE causative lesion in the initiation and maintenance of the epileptogenic process (Haut et al., 2004).

Based on the correlation of the results of experimental and clinicopathologic studies, MTLE is considered the prototype of epilepsy, in which epileptic activity plays a crucial role both in the initial event (e.g., febrile seizures/status) and in the pathogenesis of the ensuing process leading to chronic epilepsy. A number of retrospective studies have shown that a significant proportion of patients with mesial temporal sclerosis and MTLE had antecedents of complex febrile seizures in early childhood (Cendes, 2004). The experimental evidence is reviewed in the next section of this article. From a clinical standpoint, the evidence of MTLE as the result of a process sustained or facilitated by the epileptic activity is ambiguous. An analysis of the natural history of MTLE cannot answer the question of whether the unfavorable outcome results from the persistence of the epileptic activity, which is usually undetectable in the latent period between the initial event and the chronic phase, or is intrinsically related to the nature of the underlying epileptogenic process set in motion by the initial event. Previous longitudinal analyses have been inconclusive (Scott et al., 2003); the results of the ongoing FEBSTAT (Consequences of Prolonged Febrile Seizures) prospective study (Nordli, 2012) may clarify the issue. Similar considerations apply to other retrospective studies of MTLE aimed at demonstrating its progressive course and the role of seizure recurrence in sustaining the epileptogenic process. Several lines of evidence are consistent with possible acute seizure-associated changes (see review in Briellmann et al., 2005), but a subsequent evolution in chronic MTLE with hippocampal sclerosis has not been unequivocally demonstrated. A prospective analysis of 103 patients with newly diagnosed focal epilepsy (Salmenpera & Duncan, 2005) showed a decrease in hippocampal volume after 1–3 years in 13% of the patients. Again, the coexistence of uncontrolled seizures and hippocampal atrophy in a limited subset of patients does not unequivocally prove a cause–effect relationship. Sporadic reports on acute, possibly inflammatory, damage of the hippocampus (localized limbic encephalitis; Bien et al., 2007) and febrile SE (Scott et al., 2003) suggest that acute seizures and hippocampal damage develop within weeks or months and result in MTLE (Kaiboriboon & Hogan, 2002; Sokol et al., 2003) that shows little, if any, imaging sign of further progression of structural damage, despite persistence of seizure activity (Lewis et al., 2002).

Epileptic encephalopathies

Clinical evidence of epileptic activity–related deterioration in EEs can be evaluated as both drug sensitivity of seizures and intellectual disability. EEs included in the epilepsy classification are the following: early myoclonic encephalopathy (EME), Ohtahara syndrome, epilepsy of infancy with migrating focal seizures, West syndrome (WS), Dravet syndrome (DS), myoclonic encephalopathy in nonprogressive disorders, Lennox-Gastaut syndrome (LGS), epileptic encephalopathy with continuous spike-and-wave sleep/encephalopathy status epilepticus sleep (CSWDS/ESES), and Landau-Kleffner syndrome (LKS). Because all of these epilepsies are age related and occur at different stages of postnatal development, a first question is whether and how the underlying etiologic factors can interfere with the brain development, thus accounting for the slowing or regression of brain functions.

The above question is particularly relevant for the severe myoclonic epilepsy of infancy (SMEI) also known as DS for which a genetic etiology, that is, SCNA1A gene mutation has been demonstrated by Claes et al. (2001). It is in fact logical to assume that the loss of Na+-channel function resulting from the mutation is potentially harmful for the brain development. This of course does not rule out the possibility that seizures and EEG discharges also contribute to the deterioration. At present, however, there is no sound clinical evidence that this would be the case. Rather, by longitudinally analyzing the cognitive profile of a group of children with DS, Ragona et al. (2010, 2011) concluded that the apparent intellectual worsening is not a consequence of deterioration, but rather reflects an arrest of cognitive development occurring during the early stages of the disease (Fig. 1). Ongoing studies will clarify the issue. The available clinical data, however, do not provide clear evidence that the recurrence of seizures is responsible for a progressive course of the disease and the evidence even tends to dispute the classification of DS as an EE according to the EE accepted definition.

Figure 1.

Cognitive development of a group of patients with Dravet syndrome followed for at least 30 months (from Ragona et al., 2011, with permission).

Another dreadful EE is WS, which is characterized by infantile spasms and hypsarrhythmia, and often by psychomotor regression, including autistic behaviors. The latter often appear later and may indicate a progression of the syndrome. The syndrome is a relatively rare manifestation of common insults, including genetic mutations, brain dysplasias, perinatal asphyxia, and other perinatal traumatic events. Progression can therefore be viewed as the result of a continuing interaction of underlying systems and epigenetic influence. What is the role of epileptic activity? The answer is complicated by the difficulty of defining the electric and clinical manifestations as interictal or ictal phenomena in WS. It has been proposed that hypsarrhythmia may represent ongoing seizure activity, and that infantile spasms and electrodecremental events result from activation of subcortical circuits attempting to control cortical seizure activity (see Lado & Moshé, 2002). According to this interpretation, hypsarrhythmia could be viewed as an electric status epilepticus, and the spasm would express the activation of an intrinsic antagonistic mechanism. Alternatively, the spasm could be interpreted as an ictal phenomenon occurring over a diffusely altered cortical electrogenesis. Whichever is the interpretation, the typical electroclinical picture requires the active participation of a pathologic system, in which different brain areas (the cortex, thalamic nuclei, and brainstem) either work or do not work together. Also accepted is the idea that a dysfunction of the widespread system involved in WS epileptogenesis easily accounts for the failure to acquire new developmental milestones, which results in regression involving impairment of widespread networks organized as a system (see Capovilla et al., 2013). Recent EEG–functional magnetic resonance imaging (fMRI) studies of WS have demonstrated that epileptiform discharges in hypsarrhythmia are associated with hemodynamic and metabolic changes in the cerebral cortex, and that high-voltage slow waves correlate with blood oxygen level–dependent (BOLD) MR changes in cortical and subcortical structures (Siniatchkin et al., 2007). Notably, it has been demonstrated that the longer the active WS, the worst is the cognitive outcome, which makes mandatory an aggressive therapeutic attitude aimed at discontinuing the electrical and clinical manifestations as soon as possible. A better understanding of the pathogenesis of WS could contribute substantially to advancing our knowledge of the influence of epileptic activity in sustaining epilepsy progression.

In the above considerations we made reference to epileptic activity without discussing separately the relative role of clinical seizures versus EEG discharges. The group of EEs with EEG activation during sleep, encephalopathy with status epilepticus during sleep (ESES; Patry et al., 1971, otherwise labeled as continuous spike and wave discharges during sleep (CSWS) and Landau-Kleffner syndrome (LKS; Landau & Kleffner, 1957), raises the possibility that EEG discharges, apparently unrelated to any clinical manifestations, are correlated with neuropsychological deterioration, as first proposed by Tassinari et al. (1977). ESES is an EE, the main features of which are neurologic deterioration in different domains (cognitive, motor, and behavioral) and various seizure types related to the appearance of a peculiar EEG pattern characterized by epileptic activity significantly activated during slow sleep (i.e., from 85% to 100% of slow wave sleep)—that is, a condition of continuous spikes and waves, or status epilepticus during sleep (SES). LKS is an acquired epileptic aphasia or auditory agnosia, occurring in a previously normal child who has already developed age-appropriate speech, associated with epileptic EEG abnormalities particularly prominent during sleep, with or without apparent clinical seizures. Imaging examination is usually negative, and the neurologic examination is normal except for the neuropsychological deficit. Recently, LKS has been proposed to be part of the spectrum of ESES syndrome (Tassinari et al., 2012).

The current interpretation is that the intense epileptic activity as detected on EEG disrupts the normal sleep architecture, thereby interfering with memory consolidation taking place during slow wave sleep. A role of sleep in neuroplastic remodeling of neural networks mediating cognitive performance and behavior, particularly in children, is supported by a growing body of experimental data that demonstrate, for instance, the crucial role of sleep in learning and memory consolidation, whereas, sleep deprivation has an opposite, deleterious effect on these processes.

A recent hypothesis—the synaptic homeostasis hypothesis (Tononi & Cirelli, 2006)—suggests that plastic changes occurring during wakefulness result in a progressive increase in synaptic strength in many brain circuits (see also Lado et al., 2013). A progressive increase in synaptic strength is unsustainable, because stronger synapses consume more energy, occupy more space (they are more numerous and/or larger), and saturate the ability to learn. Therefore, sleep may serve an essential function by promoting a homeostatic reduction in synaptic strength, through the proportional reduction (downscaling) of synaptic weights after the synaptic potentiation gathered during wakefulness. EEG slow wave activity (SWA) is the marker of the synaptic normalization occurring during sleep: in fact, a local SWA increment during sleep after a learning task can be observed in the cortical regions involved in the performance of that task (Huber et al., 2004). This local increment of SWA results in significant improvement in the task performance after sleep. The opposite effect has been reported in subjects with prolonged immobilization of an arm in whom a local decrement of EEG SWA during sleep occurred in the corresponding cortical arm presentation (Huber et al., 2006).

Considering these data, it can be postulated that prolonged focal epileptic activity during sleep (as it occurs in ESES) interferes with SWA-generating mechanisms at the site of the epileptic focus, thereby impairing the neural processes and, possibly, the local plastic changes associated with learning and other cognitive functions (Tassinari & Rubboli, 2006). In this respect, ESES might represent a model of the clinical effects of a localized disruption of EEG activity during sleep caused by long-lasting sleep-related focal epileptic activity, adding further evidence to support the crucial role of sleep in neuroplasticity mechanisms underlying higher cognitive functions (Tassinari et al., 2009). The plausibility of this hypothesis has been recently demonstrated by Bölsterli et al. (2011), who analyzed the time course of the slope of EEG slow waves, a sensitive measure of cortical synaptic strength (Vyazovskiy et al., 2008) in children with ESES and control subjects. As expected, controls showed a decrease of the slope of slow waves from the first to the last hour of sleep, whereas ESES patients did not show any significant change in slope across the night, and this effect was more pronounced in the hemisphere with more spike-wave activity. Although this study is preliminary, owing to the small number of subjects, it is conceivable that these kinds of experiments may open a promising line of research in the field of electroclinical correlations, providing interesting insight into the pathophysiology of ESES and possibly of other childhood epileptic conditions with enhanced activation of epileptic paroxysms during sleep (Cantalupo et al., 2011).

Experimental Evidence of Seizure Activity–Related Progression of Epilepsy

In principle, experimental findings could help to clarify the issue of seizure-related progression of epileptogenesis, even though currently available data are still contradictory. The ability of simple seizures to promote secondary epileptogenesis is based on the assumption, first proposed by William Gowers (Eadie, 2011), that “seizures beget seizures.” This attractive but highly speculative hypothesis was formulated in 1881 on the basis of the idea that the molecular neuronal changes persisted after the end of a seizure discharge and predisposed to the occurrence of further seizures. The concept that recurrent seizures in an active epileptic condition promote further seizures is mentioned in several studies, even though the rational evidence that it is correct is still ambiguous and not conclusive. Strong and sustained epileptiform discharges, such as those occurring during SE, but not “simple seizures,” may beget further seizures by means of brain damage and the ensuing postreactive plasticity that forms novel synapses/networks (Ben-Ari et al., 2008). A second argument that is supposedly supporting the concept that seizures beget seizures is the demonstration of the occurrence of mirror foci (Khudoerkov, 1977). An active epileptogenic region acutely induced by local pharmacologic manipulations both in vivo (Sherwin, 1984; Szente & Boda, 1994) and in vitro (Khalilov et al., 2003) may induce activation at sites contralateral to the original focus. These experimental studies demonstrated that repetition of 10–20 ictal discharges is sufficient to induce secondary seizures or epileptic-like discharges in homologous cortical regions contralateral to the primary focus that within minutes or hours become independent (mirror foci). The chronic persistence of these acutely induced mirror foci has not been properly investigated. The original studies on epileptic mirror foci demonstrated that chronic foci generated in the rat cerebral cortex after implantation of cobalt-gelatinous rods into the contralateral hemisphere developed were associated with degeneration of callosal fibers/synapses and reversible retrograde alterations of neurons; it was not reported if such changes were related to seizure activity in the primary focus (Bogolepov & Pushkin, 1975). Therefore, the idea that a focus of epileptiform activity can generate remote excitability changes that sustain independent chronic foci in target areas is not definitively demonstrated by experimental studies.

The kindling phenomenon

A strong validation to the concept of epileptogenesis as a self-sustained process that involves seizure activity derives from the kindling phenomenon. It has been demonstrated that repeated activation of neural circuits by patterned electrical stimulation (or repetitive convulsant applications in the pharmacologic kindling) induces progressive and permanent increases in seizure susceptibility to the inducing agent and eventually results in spontaneous seizures (Goddard et al., 1969; Morimoto et al., 2004). Therefore the kindling phenomenon provides a valuable paradigm to evaluate the progressive nature of focal epilepsy (Sutula, 2004). In the fully kindled state, repetitive trains of stimuli evoke after discharges initiated by and over lasting the duration of the stimulation. Spontaneous seizures may occur in fully kindled animals and do not vary in frequency with time in the fully kindled state (Morimoto et al., 2004), suggesting that when stimulation is stopped the progression of epileptogenesis is also suspended. The progressive, stimulus-induced development of the kindling phenomenon has been associated with the progressive and cumulative stimulus-related damage characterized by neuronal loss, dentate gyrus neurogenesis, gliosis, and sprouting of the mossy fibers. It is suggested that the electrical stimulation and/or the associated afterdischarge result in an excitotoxic injury. To support this hypothesis, it was demonstrated that neuronal apoptosis occurs in the rat dentate gyrus 5 h after one single hippocampal kindling stimulation that generates an afterdischarge (Bengzon et al., 1997). As for other models of focal epilepsy (see subsequent text), the specificity of damage versus epileptiform activity in determining the progression of epileptogenesis during the kindling phenomenon cannot be disambiguated and remains an open issue.

Kindling-induced focal epilepsy was originally aimed at reproducing features of specific epileptic syndromes, namely MTLE (Cavazos et al., 1994); however, it is also accepted as a model of complex partial epilepsy with secondarily generalized seizures (McNamara, 1984).

Animal models of MTLE

Extensive literature is available on animal models that reproduce the primary features of human MTLE (see Pitkänen et al., 2006). The most used MTLE models result from chemically induced SE. In most models the systemic applications of drugs such as pilocarpine or kainate induce an intense convulsive SE that requires the use of a benzodiazepine to be interrupted to avoid death of the animals. After a latent period, a chronic epileptic condition that mimics relevant structural features of MTLE with recurrent convulsive generalized seizures has been described with extensive bilateral brain damages (Turski et al., 1983; Ben-Ari et al., 1979; Sloviter, 2005). Of interest, not all treatments that induce a sustained SE promote secondary epileptogenesis and result in chronic focal epilepsy. Treatments with γ-aminobutyric acid (GABA)–receptor antagonists, such as pentylenetetrazole and bicuculline, induce SE but do not promote brain damage; neither chronic epilepsy in the pilocarpine model seizures or brain damage progressively worsen after SE. It is not clear in these models if the two phenomena occur in parallel or if seizures precede structural alterations. In SE-induced MTLE models, seizure frequency in the full chronic condition shows high variability, and several reports have demonstrated that after an initial increase, seizure frequency reaches a plateau and show a late decrease after several months. In rat electrically induced and kainic acid–induced SE, several authors have shown that seizure duration, behavior, and frequency all intensified during early stages, but after 8–12 weeks were characterized by a plateau for all measures (Bertram & Cornett, 1994; Hellier et al., 1998). Moreover, in the MTLE model induced by electrical stimulation, ongoing hippocampal damage measured by MR imaging was associated with the epileptogenic insult (i.e., the initial SE) rather than with the recurrence of spontaneous seizures (Pitkänen & Sutula, 2002) correct. When the drug is applied locally (i.e., kainate or tetanus toxin in the hippocampus or the amygdala), the initial SE is more focal, with less generalized convulsions, does not require any interrupting treatment, and results in limited unilateral cell loss (Bouilleret et al., 1999; Bragin et al., 1999; Riban et al., 2002; Carriero et al., 2012). In particular, when a small dose of kainate is injected unilaterally in the dorsal hippocampus in mice, SE is associated with an early onset of excitotoxic damage (i.e., cell loss in CA1, CA3, and hilus during the first 24 h; Suzuki et al., 1995) restricted to the injected hippocampus, whereas no cell loss is observed in the contralateral hippocampus despite the existence of sustained epileptic discharges in this structure during the SE. In this model, a progressive aggravation of focal epileptiform discharges has been reported only during the first 3 weeks, which parallels the dispersion of the dentate gyrus (Heinrich et al., 2011). Then spontaneous focal seizures arise regularly, with occasional generalizations, and remain stable in frequency and duration for months after the SE (Langlois et al., 2010). The dispersion of the dentate gyrus reaches a plateau with a similar time course (Suzuki et al., 2005). Similarly, in the guinea pig, this protocol induces a strictly unilateral damage, despite the fact that SE involves both hemispheres (Carriero et al., 2012). This suggests that, in both models, intense seizure activity per se does not induce brain damage. Finally, when tetanus toxin is injected either in the hippocampus or in the neocortex, seizures and occasional acute SE are followed by chronic seizures that appear at 1–2 weeks and usually remit within 4–6 months. Of interest, minimal cell loss and brain damage and gliosis are observed in this model (Mellanby et al., 1977; Jefferys, 1992). If the same protocol is performed in neonate rats, it induces brain damage (in DG after hippocampal injection) and chronic seizures may persist without remission (Lee et al., 1995).

Altogether, studies on animal models of MTLE suggest that SE induced by several convulsive drugs that activate directly (e.g., kainate) or indirectly (e.g., pilocarpine) the glutamatergic neurotransmission lead to epileptogenic changes that may chronically generate further recurrent seizures. On the contrary, there is no experimental evidence in chronic MTLE models that the recurrent simple focal seizures are sufficient to promote further epileptogenesis and worsen brain damage. SE-induced recurrent seizures depend on the drug- and/or SE-induced initial brain damage and generally remit if damage is not established.

Animal models of traumatic brain injury

Models of traumatic brain injury (TBI) also reproduce a condition in which an acute event is followed by chronic focal epilepsy that develops slowly. As for the MTLE models, TBI also induces brain damage, as evaluated by MRI, that starts 3 h after traumatic injury and continues to progress for up to 6 months and reaches a plateau over a period of months (Kharatishvili et al., 2007; Immonen et al., 2009).

A larger number of epileptiform discharges was reported in different brain areas after TBI (D'Ambrosio et al., 2005). Seizure-like events increase with time after TBI in the limbic regions, and these changes parallel structural alterations, in particular gliosis (D'Ambrosio & Perucca, 2004). Seizures in the parietal cortex close to TBI showed an initial peak at 2 weeks after trauma and decreased despite the persistence of a marked damage in this region. As for the MTLE models, the extension of damage also correlated with the expression of epileptiform events in these experiments, but a causal relationship between the latter and the progression of TBI has not been proven.

Animal models of epileptic encephalopathies

Studies on models of progressive epileptic encephalopathies are limited to rodent models of West syndrome (WS) and Dravet syndrome (DS). In DS mouse models, the epileptic phenotype is reproduced by heterozygous expression of mutations in the voltage-gated sodium channel NaV1.1 identified in DS patients (Yu et al., 2006). Homozygous mice show a more severe phenotype and do not survive after 1 week of age. No correlation between the frequency and/or duration of seizures and brain damage progression has been shown. Hyperthermia-induced seizures were shown to induce high mortality, and ataxia has been reported in heterozygous Nav1.1 knockout mice, whereas no clear indication of a progression of damage-dependent on seizure occurrence was demonstrated in DS models (Kalume et al., 2007; Oakley et al., 2009; Mantegazza, 2011).

Infantile spasms and hypsarrhythmia represent the main clinical features of WS that can be induced in rat pups, either by cortical infusion of tetrodotoxin, a fast sodium conductance blocker (Lee et al., 2008), or by cortical coinjection of lipopolysaccharide and doxorubicin (Scantlebury et al., 2010) and N-methyl-d-aspartate (Velísek et al., 2007). Development of mice to reproduce human mutations associated with a malignant form of pediatric epilepsy, characterized by infantile spasms, lissencephaly and developmental abnormalities (Kitamura et al., 1997) by conditional deletion of the developmental active X-linked aristaless-related homeobox gene (Arx gene) are associated with IS and a defect of migration of interneurons from the ganglionic eminence (Kitamura et al., 2002; Marsh et al., 2009). In addition, in these models that reproduce a symptom of WS, no indication of a correlation between a progression of brain damage and seizure recurrence was demonstrated.


From the above-reviewed data we can conclude that:

  •  Most clinical studies are retrospective; therefore, separation of seizure-induced pathology from the cause of the seizure is difficult.
  •  The positive correlation between high seizure frequency and bad outcome does not necessarily demonstrate that the outcome is the consequence of seizures, rather high seizure frequency and bad outcome can both depend on a particularly aggressive epileptogenic process (risk of circular argument).
  •  Clinical and experimental data concur in indicating prolonged seizures/SE as a risky initial event, but the role of subsequent seizures in sustaining progression is much less clear and probably limited to some special situation.
  •  Cognitive deterioration in EEs may depend on mechanisms that differ from one type to another (e.g., sleep disruption in CSWS and LKS, cortical hypometabolism in WS), which raises the question of whether it is appropriate to hypothesize common seizure dependent mechanisms for epilepsy progression/intellectual deterioration.


None of the authors has any conflict of interest 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.