eMedicine Specialties > Neurology > Pediatric Neurology

Epileptic and Epileptiform Encephalopathies

Author: Stavros M Hadjiloizou, MD, Instructor, Department of Neurology, Division of Epilepsy and Clinical Neurophysiology, Children's Hospital, Harvard University Medical School
Coauthor(s): James J Riviello Jr, MD, Professor of Pediatrics, Division of Neurology, Baylor College of Medicine; Chief of Neurophysiology, Texas Children's Hospital
Contributor Information and Disclosures

Updated: May 15, 2006

Introduction

Background

Epilepsy, one of the most common neurologic diagnoses, is defined as a chronic condition characterized by spontaneous, recurrent seizures. An epileptic seizure is defined as a clinical event associated with a transient, hypersynchronous neuronal discharge and represents only the symptom of a potential underlying brain pathology and not the actual disease. Epileptic is a descriptive term used to denote the presence of an epileptic mechanism. EEG typically is used to help in the diagnosis and classification of epileptic disorders. Encephalopathy refers to central nervous system dysfunction of any cause.

Epilepsy is a heterogeneous condition that has many, often completely unrelated, underlying causes. Similarly, the cognitive profiles of patients with epilepsy vary significantly and may be related to different factors including etiology, age of onset, seizure control, and adverse effects related to treatment.

Age of seizure onset is a determining factor for the cognitive impact of epilepsy—it also is strongly associated with its etiology—and is responsible for the many differences between childhood and adult-onset epilepsies. The continuous evolution and development of the child's brain alters the symptoms of epilepsy in the same patient and may often lead to the remission of seizures. On the other hand, childhood is unique in regard to learning and psychosocial development, and epilepsy may have a long-lasting impact for the patient's life, especially if interferes with them. Epileptic or epileptiform encephalopathies are characterized by slowing or regression of development that is attributed to seizuresor interictal, often sleep activated, abnormalities on the EEG, or both (Nabout and Dulac, 2003). Hence, regression in intellectual abilities or behavioral deterioration should raise the suspicion of these disorders.

Recently, the ILAE Task Force on Classification and Terminology (Engel, 2001) proposed a modified diagnostic scheme for epileptic seizures and epilepsy that for the first time recognizes a distinct category for epileptic encephalopathies in which the epileptiform abnormalities may contribute to progressive dysfunction. This category includes the following: early myoclonic encephalopathy, Ohtahara syndrome, West syndrome, Dravet syndrome, myoclonic status in nonprogressive encephalopathies, Lennox-Gastaut syndrome, Landau-Kleffner syndrome, and epilepsy with continuous spikes during slow wave sleep (CSWS).

Chatrian et al, in their 1974 glossary of EEG terms, described "epileptiform" as an interpretive term used in electroencephalography that applies to distinctive waves or complexes distinguishable from the background activity, which resemble the waveforms recorded in a proportion of human subjects suffering from an epileptic disorder. Epileptiform patterns include spike and sharp waves, alone or accompanied by slow waves, occurring singly or in bursts lasting at most a few seconds. The term epileptiform typically refers to interictal paroxysmal activity and not to the EEG activity seen during an actual seizure, which is called an electrographic seizure. The probability of the association of epileptiform activity with an actual epileptic disorder varies.

The typical example of an epileptic encephalopathy is the Lennox-Gastaut syndrome, consisting of a mixed seizure disorder, with tonic, atonic, and myoclonic seizures, mental retardation, and a slow spike and wave pattern on the EEG. In fact, this syndrome has been called "epileptic encephalopathy" or "epileptic encephalopathy with diffuse spike and wave discharges." The majority of children with Lennox-Gastaut syndrome have intractable seizures occurring on a daily basis, some with as many as hundreds of seizures per day. Children with multiple seizures per day may have cognitive dysfunction resulting from the large number of seizures, the side effects of multiple antiepileptic drugs typically required in Lennox-Gastaut syndrome, the underlying cause itself, or a combination.

As noted above, other epileptic encephalopathies include early – neonatal-myoclonic encephalopathy and early infantile epileptic encephalopathy (also called Ohtahara syndrome), both of which are associated with frequent epileptiform activity on the EEG and a poor prognosis. Typical EEG patterns in these epileptic encephalopathies include slow spike waves, as in Lennox-Gastaut syndrome; an electrodecremental event, typically seen with the spasm in infantile spasms; the burst-suppression pattern, consisting of discontinuity in the EEG background; and the "beta burst," consisting of runs of fast activity, also called a "malignant burst," which may be associated with a tonic seizure (see Image 1).

Other epilepsies called catastrophic epilepsies or malignant epileptic encephalopathies include West syndrome, or infantile spasm syndrome; severe myoclonic epilepsy of infancy and childhood; myoclonic-astatic epilepsy (Doose syndrome); the progressive myoclonic epilepsies; and Rasmussen syndrome.

Unlike the epileptic encephalopathies, in which frequent, obvious seizures occur, the epileptiform encephalopathies refer to disorders with epileptiform activity on the EEG without marked clinical seizure activity. Epileptiform features (ie, spikes, sharp waves, and spike and wave discharges) in these disorders indicate cortical irritability but do not necessarily mean that the patient has epilepsy. However, when epileptiform activity is present, and especially when repetitive, the question of whether anticonvulsant therapy would be helpful is raised.

Our warning of "treat the patient, not the EEG" means that we will not treat with anticonvulsants if we do not think that the patient has actual clinical seizures. This particular therapeutic dilemma is faced frequently not only with the epileptiform encephalopathies but also in patients with either learning or psychiatric disorders; when EEGs are done to "rule out epilepsy"; when no overt clinical seizure activity is apparent; or, in the past, when EEGs were done in children with migraines. Treating epileptiform abnormalities in the absence of overt seizures has been called EEG cosmetics by Besag, who notes that "the extent to which epileptiform discharges cause temporary or permanent cognitive impairment, profoundly influences decisions on whether to treat with antiepileptic medication or surgery."

Epileptiform activity that occurs in certain developmental disorders and whether this activity warrants a therapeutic trial with anticonvulsants or even corticosteroid treatment has been a topic of recent interest. Although this scenario has been applied especially to the Landau-Kleffner syndrome (LKS; ie, acquired epileptiform aphasia) and its differentiation from autism or pervasive developmental disorder (PDD), especially when regression has occurred (called autistic epileptiform regression), this also occurs with disintegrative epileptiform regression, the congenital aphasias, learning disabilities, and the epileptic syndrome called continuous spikes and waves during slow sleep (CSWS) or electrical status epilepticus of sleep (ESES; see Image 2).

Although the term Landau-Kleffner syndrome typically is applied only to language deterioration, Hirsch et al have proposed that the eponym "LKS" should be applied to the acquired deterioration of any higher cortical function associated with seizures or sleep-activated epileptiform activity on EEG.

Status epilepticus (SE) also must be included as an epileptic encephalopathy, not the generalized tonic-clonic type that is easy to recognize, but the nonconvulsive type usually consisting only of altered awareness with either generalized or focal electrographic SE demonstrated on EEG. Although some patients may have subtle motor manifestations, many have no outward motor manifestations; therefore, EEG is crucial for its detection. Nonconvulsive SE may be prevalent in intensive care units, even in patients with medical illnesses and without apparent clinical seizure activity.

Mikati et al reported a protracted epileptiform encephalopathy in a child with an acute encephalopathy with seizures, language regression, and behavioral abnormalities. In order to detect these patients, the index of suspicion must be high. However, this article focuses on disorders characterized by abnormalities in higher cortical function related to epileptic or epileptiform activity rather than disorders with actual clinical seizures or SE.

The following are the definitions of certain terms:

Epilepsy: This is a chronic condition with spontaneous, recurrent seizures; a seizure is defined as a clinical event associated with a transient, hypersynchronous neuronal discharge.

Epileptic: This term denotes the presence of epilepsy.

Encephalopathy: This refers to a disturbance in brain functioning, particularly in intellectual activity or higher cortical functioning as used in this review.

Epileptiform: This refers to spike waves, sharp waves, spike and wave activity, or other rhythmic waveforms that imply epilepsy or may be associated with epilepsy. However, epileptiform activity alone does not confirm a diagnosis of epilepsy.

Dementia, epileptic dementia, and epileptiform dementia: Dementia is defined as a loss of previously learned intellectual ability that interferes with functioning; this can be a global or isolated phenomenon. Epileptic dementia refers to this loss of higher cortical function(s) resulting from actual clinical seizures, whereas epileptiform dementia refers to the loss of higher cortical function(s) associated with epileptiform activity on EEG.

Epileptic aphasia and epileptiform aphasia: Although the term epileptic aphasia has been used for LKS, epileptic aphasia by its strict definition refers to an aphasia caused by an actual seizure, or in other words, an ictal aphasia. Epileptiform aphasia refers to a language disorder, expressive, receptive, or mixed, associated with epileptiform features on EEG. The terms congenital aphasia, developmental aphasia, or acquired aphasia are used with this to describe whether the condition was present early or acquired. Acquired aphasia implies previously normal language development with subsequent regression. Note that regression might occur even in developmental language disorders or the congenital aphasias.

Epiphenomenon: This refers to associated disorders without a direct cause and effect relationship. For example, the epileptiform activity in most cases of LKS is postulated to be an epiphenomenon, occurring because of dysfunction in the language areas, and the language disorder itself to occur because of dysfunction in that location but to not be caused directly by actual ictal activity or an epileptic mechanism.

Developmental language disorder and developmental language delay: Developmental language disorder and developmental language delay (DLD) refers to a delay in the normal sequence of language acquisition—a child typically starts talking at age 12 months, speaks in phrases by age 24 months, and speaks in sentences by age 36 months. Delay is observed in both expressive and receptive language. The term DLD is used when no cause, such as an overt neurological problem, mental retardation, autism or PDD, or hearing loss, is apparent.

Children with severe epilepsy or specific malignant epileptic syndromes commonly have mental retardation, PDD, or other cognitive and learning disorders. For example, Boyer and Deschatrette reported 9 patients with primary autism who had Lennox-Gastaut syndrome; Riikonen and Amnell reported psychiatric disorders in 53 of 192 children with infantile spasms in Finland, and Chugani et al reported a high incidence of autism (10 of 14) in children with infantile spasms and bitemporal abnormalities on positron emission tomography (PET) scan. In addition, Taylor et al reported a 10% incidence of autism in children with refractory epilepsy evaluated for possible epilepsy surgery. Many children with autism or PDD have difficulty in control of seizures. Therefore, disorders that could be "epileptic" and those that could be "epileptiform" may occur in the same person.

LKS is discussed first, since this is the prototype of an epileptiform encephalopathy, and then other disorders that relate to LKS, have epileptiform features on EEG, and have defects in higher cortical functioning, such as dementia and regression, are discussed. These other disorders include the following:

• LKS variants
• Continuous spikes and waves during sleep
• Autistic spectrum disorders or pervasive developmental disorders (PDD) with special emphasis on PDD with regression (autistic epileptiform regression)
• Disintegrative epileptiform regression
• Dysphasia and aphasia, congenital developmental language disorder
• Learning problems with epileptiform features on EEG (transient cognitive impairment)

Pathophysiology

The underlying mechanisms of these disorders are complex and far from being elucidated. It is presumed that the encephalopathic picture is, at least partially, the result of seizures, the epileptiform electrical activity, or both. Landau and Kleffner, in 1957 suggested that "persistent convulsive discharges in brain tissue largely concerned with language communication" may be responsible for the patient's deficits. Many believe that CSWS and LKS may be "the two sides of the same coin" (Kallerman, 1978; Tassinari, 1992).

The role of sleep activation, and particularly the extreme phenomenon of electrical status epilepticus of sleep, offers an appealing and challenging paradigm that could lead to better understanding of the pathophysiologic basis of these conditions. Two crucial questions still await an answer: (1) what are the mechanisms involved in the generation of such significant, interictal, sleep activation; and (2) what are the mechanisms involved in the cognitive/developmental regression that accompanies these conditions.

Although currently no clear answers to these questions are available, evidence suggests that defective mechanisms of synaptogenesis and thalamocortical circuit formation during a critical period may be involved in the generation of CSWS.

Secondary bilateral synchrony, that is facilitated by the corpus callosum and that may involve the thalamocortical connections, was hypothesized as the possible mechanism for the generation of ESES and LKS discharges (Morrell, 1985; Morrell, 1995; Kobayashi, 1992; Monteiro, 2001).

Recently, Guzzetta et al reported the EEG abnormalities in 32 children with early thalamic injury, primarily due to vascular mechanisms, and their association with CSWS (Guzzetta, 2005). Notably, the EEGs of 29 out of the 32 patients showed major sleep activation. Among these 29 patients, 2 different groups were distinguished: the first included the more or less typical CSWS (12 cases), generally with symmetry of spike and waves (SWs) and often with no spindle at all. The other cases had an usual asymmetry of SWs and the presence or reduction of spindles, plus other atypical features concerning synchronism and morphology of SWs. Behavioral disorders were significantly more present in patients with a true CSWS; their improvement paralleled the disappearance of CSWS. The generally predominant injury was in the lateral aspect of the thalamus including reticular the nucleus and ventral nuclei. The authors speculated that an imbalance between GABA-B- and GABA-A receptors may be a cofactor predisposing to

thissymptomatic CSWS.

Also increasing evidence suggests that interictal EEG abnormalities can produce transient cognitive impairment (Shewmon and Erwin, 1988; Kasteleijn-Nolst Trenite, 1988; Aarts, 1984; Binnie, 2003; Binnie, 1992; Binnie, 1991). Furthermore, benign rolandic epilepsy (benign epilepsy with centrotemporal spikes) may be not so benign since the interictal discharges may have a substantial effect on cognitive function (Beleasel and Bye, 2004; Massa, 2001). Additionally, the presence of continuously abnormal discharges during sleep may cause disruption of hippocampal function and interfere with the consolidation of memory (Moruzzi and Magoun, 1995; Lorincz and Buzsaki, 2000; Louie and Wilson, 2001).

Furthermore, autoimmune mechanisms are also considered in the pathogenesis of developmental disorders and epilepsy in children. Connolly et al reported higher autoantibodies to several neural antigens in these children compared with controls. The presence of both brain-derived neurotrophic factor (BDNF) autoantibodies and elevated BDNF levels in some children with autism and childhood disintegrative disorder (CDD) suggests a previously unrecognized interaction between the immune system and BDNF.

Frequency

International

In a recent epidemiologic study of childhood epilepsy, in a 20-year cohort from Tel Aviv, Kramer et al reported that West syndrome occurred in 9%, myoclonic seizures in 2.2%, Lennox-Gastaut syndrome in 1.5%, and LKS, Ohtahara syndrome, myoclonic astatic epilepsy, and ESES in 0.2% each.

Autism occurs in approximately 4 per 10,000 children in the general population.

Mortality/Morbidity

Mortality and morbidity relate to the underlying causes of these disorders, the treatment, or both.

Sex

Developmental disorders have a higher frequency in males. McLaren and Bryson reported that the male-to-female ratio ranges from 1.3 to 1.9:1.

Age

These disorders may involve all ages, although they present in childhood.

Clinical

History

• Landau-Kleffner syndrome (LKS) or acquired epileptiform aphasia
  • In 1957, Landau and Kleffner reported 6 children who presented with aphasia after apparently normal acquisition of language. Since then, LKS was recognized as an epileptic syndrome characterized by language regression, an abnormal EEG and absence of specific underlying brain pathology.
  • The classic features of LKS include a developmentally normal child and normal language acquisition that is followed by a verbal auditory agnosia (ie, word deafness), language regression, seizures, and an epileptiform EEG. LKS usually presents in children aged 3-8 years. An important corollary is intact peripheral hearing. Patients without all classic features have been referred to as having LKS variants.
  • Both seizures and behavioral disturbances, particularly hyperactivity, occur in approximately two thirds of children with LKS.
  • By 1992, only 198 additional cases had been reported. Although considered rare, uncommon is a better descriptive term, since it has been underdiagnosed in the past.
  • LKS is still classified by the International League against Epilepsy (ILAE - 1989) under "epilepsies and syndromes undetermined as to whether they are focal or generalized"; it is currently proposed (Engel 2001) to move under the newly described category of epileptic encephalopathies (in which epileptiform abnormalities may contribute to progressive dysfunction). Most cases reported to date have been classified as idiopathic, although specific pathologic disorders have been identified, including infectious disorders such as cysticercosis and toxoplasmosis, inflammatory disorders such as CNS vasculitis and demyelinating disease, tumors such as temporal lobe astrocytomas and dysembryoplastic neuroepithelial tumors (DNET), and other structural abnormalities such as a left sylvian arachnoid cyst.
  • The EEG in LKS shows bilateral, multifocal spikes and spike and wave discharges, usually occurring in the posterior regions—bitemporally (50%) or parieto-occipitally (30%) (see Image 3) and associated with marked activation during sleep. However, discharges occur in many locations and may even be generalized. Some centers require the presence of ESES in order to make the diagnosis. LKS is associated with ESES, but SWI may be only 50%.
  • The EEG may change over time, either spontaneously or with treatment, and EEG abnormalities are more likely to be present at the time of the language regression. Therefore, a critical time period could be present in which the EEG is abnormal. This has important clinical impact since, in the past, obtaining an EEG at the time of language regression was unusual; EEGs typically are performed after the regression had occurred. In the first author's series of the first 20 patients diagnosed with LKS or LKS variants at Children's Hospital, reviewed by Bolanos et al, EEG became normal in 4 patients within a year.
  • The actual cause of the language regression and aphasia seen in LKS is not known. Moreover, whether the language regression is due to actual epileptic activity or ictal activity directly affecting the function of language areas or the epileptiform abnormalities are a result of damage or dysfunction in these anatomic areas, or in other words, an epiphenomenon, is questionable. Holmes et al suggested that EEG abnormalities are an epiphenomenon and not the cause of the aphasia.
  • Pathologic specimens obtained after multiple subpial transection (MST), a surgical procedure designed by Morrell to eliminate an epileptic focus in eloquent cortex, done under EEG guidance, have demonstrated subcortical astrocytosis, perivascular lymphocytosis, and microglial nodule formation (ie, findings that suggest inflammation), a cryptic arteriovenous malformation, or excessive ectopic neurons. These findings demonstrate a pathologic process in these anatomic regions. MST typically is done in patients with an idiopathic/cryptogenic disorder, since lesionectomy would be done in symptomatic patients.
  • Since MST disrupts an epileptic focus, preventing epileptic activity (MST interrupts horizontal connecting cortical fibers while leaving vertical cortical connections intact), the resulting improvement indicates that the epileptiform activity in these specific cases must have caused the aphasia. If epileptiform activity can cause language dysfunction, then whether LKS must always be acquired or could, in a critical area, prevent the normal acquisition of language, such as in a congenital aphasia, is a debated issue. The presence of a cortical dysplasia in tissue obtained from MST suggests this possibility.
  • The prognosis for LKS has varied, depending on the series. Montovani and Landau followed up the original patients reported by Landau and Kleffner in 1980. Of 9 patients, with follow-up that varied from 10-28 years, 4 had full recovery, one had a mild language disability, and 4 had moderate disability. Later papers have not reported favorable outcomes.
  • Bishop did a literature review of patients with LKS in 1985 and identified 45 cases. Age of onset seemed related to the outcome the outcome was less favorable if the onset occurred before age 4 years. Deonna et al reported that only 1 of 7 adult patients had normal language, the other 6 demonstrating varying degrees of language deficits, some with complete absence of language. In a recent paper on the neuropsychologic follow-up of 12 patients, Soprano et al reported that 9 of the 12 patients had persistent language deficits of different degrees.
  • Praline et al reported their experience with the seizure and cognitive outcome of two young adults who were diagnosed with LKS in childhood and compared with 5 young adults who were diagnosed with CSWS in childhood. They concluded that, although the epilepsy associated with these syndromes has a good prognosis, the neuropsychological disorders particular to each syndrome persisted. The intellectual functions of the 2 patients with LKS were normal; however, their everyday lives were disrupted by severe, disabling language disturbances.
  • Rossi et al described 11 patients with LKS treated with various AEDs and a mean follow-up of 9 years and 8 months. At the last observation, only 18.2% of cases presented a complete language recovery, while mental retardation was evident in 63.6%.
• Landau-Kleffner syndrome variants
  • Children with language regression without the classic features of LKS have been referred to as having LKS variants. Originally, children without clinical seizures were considered as having variants because all the children reported by Landau and Kleffner had clinical seizures. However, clinical seizures occur in only two thirds of the patients with this syndrome. Epileptiform features must be present on EEG to make this diagnosis.
  • The variants include abnormal EEG without clinical seizures; involvement of more anterior language areas with dysfunction characterized by oral-motor apraxia, sialorrhea, more expressive rather that receptive aphasia, seizures, and an abnormal EEG (centrotemporal spikes similar to those seen with benign focal epilepsy); language regression and abnormal EEGs with PDD; and even congenital aphasias, also called developmental language disorders, with epileptiform EEGs.
  • Additionally, the criteria for LKS include normal acquisition of language followed by regression. However, language regression also can occur in children with previously abnormal language development such as in PDD, developmental language disorders, and CSWS, and these disorders may overlap. Some children with PDD or language delay and epileptiform EEGs, who have never had regression, have been regarded as having "developmental LKS" or an LKS variant because of the epileptiform activity. In the series of autistic and dysphasic children reported by Tuchman, Rapin, and Shinnar, language regression was seen in 5% (11 of 229) of children with a dysphasia.
• The syndrome of electrical status epilepticus of sleep (ESES) or epilepsy with continuous spikes during slow wave sleep (CSWS)
  • In 1971, Pantry et al reported 6 children with significant cognitive regression and "subclinical electrical status epilepticus induced by sleep". All 6 children had abnormal EEGs characterized by markedly sleep-activated pattern with paroxysmal features covering at least 85% of slow wave sleep.
  • CSWS was later expanded by Tassinarri et al (Tassinarri, 1992) to represent an epileptic syndrome characterized by various seizure types and continuous spike and waves during slow wave sleep associated with broader range of neuropsychological deterioration. CSWS presents in children aged 3-14 years (peak age, 8 y), and the onset of seizures usually precedes the onset of neuropsychological deterioration. CSWS may be further divided into symptomatic and cryptogenic, depending on whether normal neurological or psychomotor development was present before its onset.
  • Like LKS, "epilepsy with CSWS" is still classified by the International League against Epilepsy (ILAE - 1989) under the category of "epilepsies and syndromes undetermined as to whether they are focal or generalized"; the 2001 proposal of ILAE Task Force on Classification and Terminology (Engel et al) suggests that this syndrome should be described under the newly created category of epileptic encephalopathies (in which epileptiform abnormalities may contribute to progressive dysfunction), like LKS. The ILAE definition suggests that CSWS "results from the association of various seizure types, partial or generalized, occurring during sleep, and atypical absences when awake. Tonic seizures do not occur. The characteristic EEG pattern consists of continuous diffuse spike-waves during slow wave sleep, which is noted after the onset of seizures. Duration varies from months to years. Despite the usually benign evolution of seizures, prognosis is guarded because of the appearance ofneuropsychologicaldisorders."
  • Interestingly, the ILAE definition of the syndrome does not involve specific definition of the electrographic abnormalities and does not require a particular percentage of sleep activation as a criterion of the syndrome. This is indicative of the controversies in these terms and suggestive that a continuum of activation may evolve over time (Galanopoulou, 2000). Consequently, 85% of sleep activation should not be considered an absolute figure and should not be a requirement for this syndrome. In fact, this degree of activation may represent only a fraction of the patients with this disorder.
  • These children have seizures, which may not be frequent and usually respond well to AED treatment, although the regression in language, mental ability, and behavior may remain resistant to such treatment. Tassinari reported a series of 29 children with CSWS. All children except one had seizures, and 26 had had either several seizures or frequent seizures for more then 3 months preceding CSWS; 18 had normal psychomotor development prior to CSWS; and 11 had abnormal psychomotor development prior to the onset of CSWS.
    • The 18 patients with normal development all had a severe decrease in intelligence quotient (IQ) score; behavioral disturbances defined as decreased attention span, hyperactivity, aggression, difficulties with interaction; and inhibition after the development of CSWS. Two patients developed a psychotic state.
    • Ten children had a marked impairment of temporospatial orientation.
    • In the 11 patients with abnormal psychomotor development, mental deterioration was noted in all, 3 developed marked hyperactivity, and one showed "massive regression" including language and loss of interest in all activities.
  • Eriksson et al described a child with occipitotemporal CSWS and visual agnosia who demonstrated major deficits in visual perception, especially in object recognition, impaired shape discrimination and detection, and poor copying skills.
  • The EEG pattern of CSWS can occur in LKS, and some criteria require CSWS to confirm a diagnosis of LKS. Guilhoto and Morrell reported that language regression was the most prominent symptom in LKS, whereas generalized neurobehavioral problems were the predominant symptoms in generalized CSWS. Therefore, LKS and CSWS may have similar clinical and electrographic features.
  • Veggiotti et al recently emphasized the difference between the EEG pattern of CSWS and the epileptic syndrome of CSWS. Not all patients with a sleep-activated pattern consistent with CSWS have the age-related epileptic syndrome of CSWS. In their series of 32 patients with CSWS, only 10 (34%) had features of the CSWS syndrome, whereas in the remainder, 4 had LKS, 3 had the acquired opercular syndrome, and 15 had symptomatic epilepsy.
  • Differential diagnosis of CSWS, especially when considering the EEG, includes LKS, Lennox-Gastaut syndrome (LGS), benign epilepsy with centrotemporal spikes, and atypical benign partial epilepsy. LKS and CSWS are differentiated primarily by the clinical manifestations, with LKS featuring primarily language regression, while CSWS features primarily neurobehavioral regression. LGS can have similar seizures, but tonic seizures are the hallmark of LGS and do not occur in CSWS. EEG in LGS has fast bursts, also called beta bursts or malignant bursts. Benign focal epilepsy of childhood with centrotemporal spikes, also referred to as benign Rolandic epilepsy or benign epilepsy with centrotemporal spikes (BECTS), may have a sleep-activated EEG but should not reach the 85% needed for the diagnosis of CSWS. Aicardi and Chevrie described atypical benign partial epilepsy, which may have similar seizures and EEG patterns but does not cause intellectual regression.
  • Praline et al reported their experience with the seizure and cognitive outcome in 5 young adults who were diagnosed with CSWS in childhood. Although they confirmed the good outcome associated with epilepsy in CSWS, they found persistence of the neuropsychological deficits. Three of the 5 patients with a CSWS syndrome during childhood remained globally and nonselectively mentally deficient.
  • Outcome of CSWS: Scholtes et al
• Pervasive developmental disorder, pervasive developmental disorder with regression
  • The terms of pervasive developmental disorders (PDD) and autistic spectrum disorders (ASD) are equivalent terms and refer to the same spectrum of life-long developmental challenges. PDD is a clinical syndrome characterized by defects in language and social interaction and stereotyped repetitive behaviors. Children with autism were thought to be abnormal from very early in life, but several studies have reported regression in children with PDD. Tuchman and Rapin reported language regression in 30% of 582 children with PDD, and Kurita reported speech loss in 37% of 261 children with autism.
  • Wilson et al reported their experience with 196 consecutive children (143 males and 53 females) who presented with language regression or plateau. Mean age at regression was 21.2 months, and the mean interval to referral was 34.8 months. Seventy percent of the children became nonverbal, and 75% were cognitively impaired. Language regression was associated with a more global autistic regression in 93% of children. Fifteen percent of the children had a history of seizures. Some recovery occurred in 61%, but only one child recovered fully. Improvement was more likely in the 49% who were entirely developmentally normal before the regression.
  • PDDs are associated with increased, although variable, risk of developing epilepsy depending on age, cognitive impairment, and type of language disorder. It is suspected that between 10 and 40% of children with PDD will develop epilepsy by adolescence or early adulthood. PDD may be also associated with abnormal epileptiform EEG abnormalities without clinical seizures.
  • In the authors' experience, children with PDD (ie, autism) and language regression and abnormal EEGs make up the largest number of the variants, and in fact, the largest number of the patients referred for "rule out LKS."
  • Perhaps most controversial is the possible relationship between LKS and autism (PDD). The differentiation of LKS from PDD with regression can be difficult and perhaps even impossible, since patients with language regression at a younger age regress in social skills and eye contact, and may develop self-stimulatory behaviors, therefore fulfilling diagnostic criteria for PDD. However, the "pure" or "typical" child with LKS still has an ability to relate to others, share emotions, and use nonverbal communication. Therefore, the relative preservation of nonverbal skills is an important differentiating factor.
  • In contrast, the essential characteristics of autistic disorder are "markedly abnormal social interaction and communication and a markedly restricted repertoire of activity and interest" (Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition). The specific diagnostic features of autism include (1) marked abnormality in the quality of social-relating interaction (such as eye contact and other nonverbal behavior, sharing emotions, reciprocal relating to peers), (2) impairment in communication (eg, delay in development of speech and lack of compensating nonverbal communication, or stereotyped and otherwise abnormal speech), and (3) restricted and stereotyped behaviors (such as extreme preoccupation with one's interest and/or routine and self-stimulation). At least some of these developmental abnormalities have to be seen before age 3 years.
  • While some children with autism initially have some language acquisition, they stop using it between ages 1-3 years, and language usually is quite limited.
  • Children with autism may have a delay in the development of spoken language, and a large number of children with autism do not develop any useful spoken language. A smaller number of children with autism may have had normal development, usually until age 18 months to 3 years, and then have had a period of regression. The best study to date on this topic comes from Tuchman and Rapin. In their database of 533 children with PDD, regression occurred in 29%; 64 of these children had regression before age 2 years and, in this group, 20% had abnormal EEGs.
  • As noted above, both EEG abnormalities and epilepsy occur in children with autism, so their presence alone does not help to differentiate autism from LKS. The differentiation of children with LKS from those with autism may be especially difficult in a child with autism with language regression and either seizures or EEG abnormalities. Whereas a child with LKS and an epileptiform EEG can be treated with either antiepileptic drugs or steroids, steroid therapy has no proven role in children with autism.
  • In 1975, Small reported paroxysmal EEG abnormalities, including focal spikes, paroxysmal spike and wave, and multifocal spike activity, in 65% of 147 children with autism. Detection of EEG abnormalities depends on 2 factors: (1) number of EEGs recorded and (2) obtaining the EEG in all states—wakefulness, drowsiness, and sleep. When only one EEG was recorded, the chance of having an abnormal EEG was only 40%; it increased to 60% for 2 EEGs, and 80% for 3 or more recordings. More EEG abnormalities were detected when all 3 states were recorded.
  • Minshew recently reviewed the neurophysiologic abnormalities in autism. The recent studies have involved predominantly single EEGs and have reported abnormalities in 32-43%. These included diffuse or focal spikes or slow waves, paroxysmal spike and wave activity, with a mixed discharge being the most common abnormality. These tended to be bilateral, and even when unilateral did not show a predilection for a single cortical area. She notes that Small's study reflects a tendency to do repeat EEGs in the more severely affected patients and that this study included both primary and secondary cases of autism (eg, acquired brain damage or a genetic syndrome such as Tay-Sachs syndrome).
  • Tonic-clonic seizures are the most common type of seizures in children with PDD and occur more frequently in those with more severe impairment. The onset is usually in early childhood, with a second peak during adolescence. The best longitudinal studies are documented by Lockyer and Rutter, and Gillberg and Steffenburg. The cumulative incidence of seizures in these studies was 25% and 33%, respectively. Lockyer and Rutter also reported serious neurological deterioration during adolescence, accompanied by seizures, in 12% of subjects.
• Childhood disintegrative disorder
  • This is a subgroup of the PDD spectrum of children with normal development for at least the first 2 years after birth, with a clinically significant regression after age 2 years but before age 10 years. This also has been called Heller syndrome, dementia infantilis, or disintegrative psychosis, and it usually is associated with severe mental retardation. Affected children undergo autistic regression later and have higher risk of developing epilepsy than more typical cases of PDD.
  • Disintegrative disorder may be classified as disintegrative epileptiform regression when the EEG has epileptiform activity. In data from Tuchman and Rapin, only 5% (n=9) of the children had regression after age 3 years. In these 9 children, one had epilepsy and one had an epileptiform EEG. Tuchman reports that the epileptiform activity is more global in disintegrative disorder. Mouridsen et al compared seizures in disintegrative psychosis and in infantile autism and found that the incidence was 77% in disintegrative psychosis and 33% in infantile autism. The peak period of onset was before puberty in both groups.
• Developmental language delay, developmental language disorder, congenital aphasia
  • These terms refer to a delay in the normal sequence of language acquisition in the absence of an apparent cause. As in LKS, this diagnosis requires normal peripheral hearing and no delay in motor or intellectual milestones. A developmental language disorder may be expressive, receptive, or mixed, and verbal auditory agnosia or verbal apraxia may be observed.
  • Epileptiform EEGs have been reported in these patients. Sato and Dreifuss reported a 13-year-old boy with speech delay and EEG showing runs of bitemporal, rhythmic, continuous spike and sharp waves at 5-6 Hz without any clinical manifestations of seizure activity. Maccario et al reported 7 children with language delay and dysphasic errors with EEG abnormalities consisting of focal and generalized spike-wave and sharp-wave discharges, which did not respond to anticonvulsants.
  • In a larger study, Echenne et al evaluated EEG findings in 32 children with congenital aphasias. Nine of these patients had occasional seizures and 4 were epileptic; 22 of 32 patients had normal EEGs and 10 showed interictal epileptiform discharges. During a prolonged EEG after sleep deprivation, 30 of 32 had epileptiform activity, 4 with ESES. Tuchman et al reported regression even in this group; if the EEG were epileptiform, distinguishing this clinically from LKS would be difficult except by the normal development prior to regression in LKS.
  • No randomized study has been done in children with this disorder. Some series of patients with LKS have included those with speech delay, most notably that of Soprano et al, in which 9 of 12 patients with LKS had some language disturbance prior to the acquired aphasia. The author has treated 2 children with congenital aphasia and marked sleep-activation on EEG, who had clinical improvement on valproic acid (VPA); lower levels of VPA were associated with regression in their abilities. These cases suggest that treating the interictal activity in a case of a congenital aphasia might be useful.
• Learning problems or cognitive impairment with epileptiform features on EEG, also referred to as transient cognitive impairment
  • Generalized tonic-clonic convulsive (GTC) seizures are well known to have a potential negative impact on cognitive functioning; therefore, cognitive testing preferably should be deferred for several days after a GTC seizure. Also, either untreated or uncontrolled absence epilepsy with frequent absence seizures may interfere with neuropsychological testing. Therefore, the patient preferably should have the best seizure control possible before doing this type of testing. However, subclinical seizures (ie, epileptiform discharges not associated with clinical signs) may have a negative effect on performance. This is referred to as transient cognitive impairment (TCI) by Kasteleijn-Nolst Trenite et al.
  • Detection of TCI by generalized spike and wave discharges is dependent on the duration of the discharge and the complexity of the task. The chance of cognitive impairment increases when the discharges are longer than 3 seconds in duration; choice reaction time and short-term memory tests are more sensitive than simple motor tasks.
  • Gordon et al reported an instructive case of a 7-year-old boy who presented with poor progress in school and an EEG that showed very active, independent frontal spike discharges. He did not have clinical seizures. He was randomized to receive VPA, 125 mg bid (4 periods), or matching placebo (4 periods) over 8 weeks, with primary end-points being the score on the Wechsler Intelligence Scale for Children-Revised (WISC-R) and a handwriting sample; behavior was monitored using teacher and parents' Connors questionnaires. While on VPA, the boy's WISC-R score significantly improved, and the EEG went from a pretreatment frequency of 28 spike discharges per minute to normal. The children with PDD treated with VPA by Gillberg and Schaumann and also by Plioplys had no clinical seizures, yet they had marked improvements.
  • Studying 91 patients with epilepsy, Binnie et al found TCI in approximately 50% of patients during subclinical generalized and even focal discharges. Specifically, right-sided discharges were associated with defects in spatial memory and left-sided discharges were associated with verbal errors. A relationship between the timing of the discharge and the stimulus was noted—dysfunction was greater if the discharge occurred either during, or even more significantly, within 2 seconds before, the stimulus or in the time between the stimulus and the response.
  • Kasteleijn-Nolst Trenite reported similar findings in a different study in which 36% of children had defects in either visual spatial or verbal tasks, or both; he also reported that right-sided discharges had a greater effect on visual spatial tasks and left-sided discharges had a greater effect on verbal tasks. Both Kasteleijn-Nolst Trenite et al and Marston et al have demonstrated improvement in school performance in children treated with VPA for subclinical epileptiform discharges.
  • Binnie studied TCI in 10 children with rolandic spikes, using the Corsi short-term memory test performed during continuous EEG monitoring; 5 of 10 children showed significant TCI. Kasteleijn-Nolst Trenite recommended suppressing epileptiform discharges when TCI has been demonstrated with EEG testing.
  • These cases raise questions, therefore, about the usefulness of EEG in identifying epileptiform activity in children with learning difficulties and about what these discharges mean and what type of testing is needed to determine their significance in children with definite epilepsy. These cases raise the question of the need for this type of testing routinely in patients with epilepsy. In addition, these clinical histories suggest that epileptiform activity alone may be sufficient to cause an impairment in cortical functioning in some children.

• Search for CME/CE on This Topic »

• Brain and Nervous System Center
• Epilepsy Overview
• Epilepsy Causes
• Epilepsy Symptoms
• Epilepsy Treatment