Benign Familial Neonatal Seizures

Benign familial neonatal seizures (BFNS), is a rare autosomal-dominant epilepsy of the newborn characterized by recurrent, mainly focal seizures that begin in the first few days of life and remit after a few weeks or months. Seizures start with a tonic posture accompanied by other symptoms such as apnea and other autonomic features. The seizures often progress to focal or bilateral clonic jerks. Generalized seizures have also been reported. Seizures are usually brief, lasting for approximately 1–2 minutes, but may occur as many as 20–30 times a day. The postictal state is short-lived, and the neonates look normal in the interictal period [Ronen et al., 1993]. The interictal electroencephalogram (EEG) pattern either is normal or shows minimal focal or multifocal abnormalities. EEG patterns suggestive of poor prognosis, such as a burst-suppression or inactive EEG, have not been reported. The original description of the disease dates back to the work of Rett and Teubel, who described a four-generation family with nine individuals affected [Rett and Teubel, 1964]. In their original work, the authors emphasized the heritability of the disease and the absence of acquired origin. The term “benign” was added only few years later, to highlight the fact that most affected individuals experienced normal psychomotor development. Nevertheless, 5 percent of BFNS patients present with febrile seizures and 11 percent develop epilepsy later in life [Plouin and Anderson, 2005]. In rare infants, the outcome may be poor [Dedek et al., 2003; Borgatti et al., 2004; Steinlein et al., 2007], and a single family with myokymia later in life has been described [Dedek et al., 2001].

BFNS is the first epilepsy syndrome for which a gene could be localized [Singh et al., 1998; Charlier et al., 1998]. This disorder has been linked to mutations in the neuronal voltage-gated potassium channel subunits, KCNQ2 and KCNQ3 [Singh et al., 1998; Charlier et al., 1998], located on chromosome 20 (20q13.3) and on chromosome 8 (8q24), respectively. KCNQ2 and KCNQ3 subunits mediate the M-current, a neuronal-specific K⁺ current, which plays a key role in controlling neuronal excitability [Biervert et al., 1998; Singh et al., 2003; Wang et al., 1998; Soldovieri et al., 2007]. The majority of cases, 56 percent, are associated with mutations in KCNQ2, with KCNQ3 mutations accounting for only 6.6 percent [Deprez et al., 2009]. More than 60 mutations of KCNQ2 have been reported so far, including missense, frameshift, nonsense, and splice-site mutations. Furthermore, submicroscopic deletions or duplications in KCNQ2 have been shown to account for a significant proportion of unsolved BFNS cases, in which mutations are not detected by direct sequencing [Heron et al., 2007a], and have been found in about one-sixth of BFNS families.

The age dependency of BFNS has been explained by the peculiar spatial and temporal expression pattern of KCNQ2 and KCNQ3 subunits in human development [Kanaumi et al., 2008]. A high degree of expression of KCNQ2 has been identified in the hippocampus, temporal cortex, cerebellar cortex, and medulla oblongata in fetal life. There is a decrease in expression of the gene after birth. The increased expression of KCNQ3 during fetal life compared with infancy has also been confirmed.

Benign Familial Neonatal-Infantile Seizures

Benign familial neonatal-infantile seizures (BFNIS) is an autosomal-dominant benign epilepsy of the early infancy with a high degree of penetrance. Age of onset ranges from 2 days to 6 months in different family members, with a mean age of onset of 3 months. Seizures have a focal onset, often with eye deviation, apnea, cyanosis, and staring. Secondary generalization may follow. Similarly to BFNS, seizures may occur in clusters. The interictal EEG is normal or shows occasional central or posterior focal spikes [Heron et al., 2002; Herlenius et al., 2007]. Seizures abate by 1 year of life, with a low risk of recurrence.

Unlike the previously discussed “benign familial neonatal seizures,” which are associated with potassium channel mutations, this syndrome has been linked to mutations in SCN2A [Heron et al., 2002], the alpha-2 subunit of the neuronal voltage-gated sodium channel [Berkovic et al., 2004a; Helbig et al., 2008]. Eight different mutations in ten families have been described so far [Heron et al., 2002; Berkovic et al., 2004b; Herlenius et al., 2007]. Mutations in SCN2A seem to be specific to this condition, with only very few cases being reported [Helbig et al., 2008]. The SCN2A mutations reported in association with BFNIS are all missense mutations. While the majority are predicted to be gain-of-function, more recently, Misra and colleagues reported three mutations that are a loss-of-function type [Misra et al., 2008]. An increase in the Na⁺ current caused by modifications of voltage-dependent gating, or a reduction in the current caused by reduced channel expression or impaired gating kinetics, has the ultimate effect of increasing neuronal excitability, thus favoring seizure generation or propagation [Avanzini et al., 2007; Misra et al., 2008].

Neonatal Epileptic Encephalopathies

The International League Against Epilepsy (ILAE) defines two further neonatal epilepsy syndromes, early myoclonic encephalopathy and Ohtahara syndrome (early infantile epileptic encephalopathy). Both syndromes have been defined as epileptic encephalopathies with neonatal onset and a typical suppression-burst EEG pattern [Aicardi and Ohtahara, 2005]. They are associated with severe psychomotor retardation and grave prognosis [Djukic et al., 2006]. A distinguishing feature is the main seizure type, which is myoclonic in the case of early myoclonic encephalopathy, and tonic in the case of early infantile epileptic encephalopathy. Ohtahara syndrome has been linked to a gene encoding syntaxin binding protein 1 (STXBP1), which encodes a vesicle release protein and has been implicated in 35.7 percent of cases [Saitsu et al., 2008; Deprez et al., 2009]. STXBP1 mutations have also been implicated in cases [Saitsu et al., 2008; Deprez et al., 2009] of mental retardation and nonsyndromic epilepsy [Hamdan et al., 2009]. Therefore, the role of STXBP1 in epilepsy is interesting on a couple of levels. First, it extends the genetic epilepsies to include molecules involved in neurotransmission through vesicle release, in addition to the many known ion channel mutations. Second, it redefines a syndrome as “genetic,” which might otherwise have been classified as “structural” or “metabolic.” Thus far, only one other protein involved in synaptic vesicle transmission, Synaspin-1, has been linked to epilepsy and learning difficulties, as reported in one family [Garcia et al., 2004]. Testing for STXBP1 may be useful in nonlesional cases of Ohtahara.

A second gene has recently been implicated in Ohtahara; an ARX protein truncation mutation has been found in one family [Fullston et al., 2010]. ARX is a homeobox gene located on the short arm of the X chromosome, and it has been implicated in a number of different syndromes including West’s syndrome (see below), lissencephaly, and nonsyndromic X-linked mental retardation and epilepsy [Strømme et al., 2002]. Limited human and extensive mouse gene-expression studies show high levels of ARX expression in the fetal brain, particularly in the neuronal precursors of the germinal matrix and the ventricular zone. High levels of expression are also observed in the subventricular zone, the caudate nucleus, putamen, substantia nigra, corpus callosum, amygdala, and hippocampus [Bienvenu et al., 2002; Kitamura et al., 2002; Colombo et al., 2004; Poirier et al., 2004]. This expression during early development and predilection for neuronal tissue suggests that ARX has a pivotal function in neurodevelopment. Recently, homozygous mutations on SLC25A22 have been described in case reports of consanguineous families with neonatal epileptic encephalopathies with suppression-burst [Molinari et al., 2005, 2009].

A final epileptic encephalopathy is characterized by early onset of a severe intractable seizure in girls with normal interictal EEG and severe hypotonia. The syndrome is associated with mutations in X-linked cyclin-dependent kinase-like 5 (CDKL5) [Bahi-Buisson et al., 2008a, b]. Mutations in the CDKL5 gene have also been shown to cause infantile spasms and Rett’s syndrome-like phenotype. Recent studies [Bahi-Buisson et al., 2008a, b] highlighted the key clinical features of this rare epileptic encephalopathy that should help establish a molecular diagnosis. In younger patients under 2 years of age, early-onset epilepsy is probably the most consistent sign in CDKL5 mutations. In all cases, epilepsy starts within 3 months, often in the neonatal period, with very frequent seizures and an interictal EEG that is normal or indicates background slowing. At this age, neurologic examination reveals severe hypotonia and poor eye contact. Subsequently, a large proportion of patients develop epileptic encephalopathy characterized by the occurrence of infantile spasms with hypsarrhythmia, and then multifocal epilepsy. While brain magnetic resonance imaging (MRI) might be normal at onset, it is usually found to be abnormal later in the course of the disease. Although it is nonspecific, most patients exhibit cortical atrophy combined with hyperintensities in the temporal lobe white matter that could be related to abnormal myelination. The evolution of the disorder includes the appearance of hand stereotypies from the age of 18 months to 3 years, hand apraxia, sleep disturbances and deceleration of head growth [Bahi-Buisson et al., 2008a]. All girls with CDKL5 mutations are severely delayed, with limited, if any, autonomy.

Benign Familial Infantile Seizures

Benign familial infantile seizures (BFIS) syndrome was first described in 1992 by Vigevano and co-workers, who reported on cases with focal seizures in infancy, a family history of convulsions, benign outcome, and autosomal-dominant inheritance [Vigevano et al., 1992]. This syndrome is now included in the most recent classification and terminology proposed by ILAE [Engel, 2001]. Age of onset ranges from 4 to 8 months, with a peak around 6 months. Seizures typically occur in clusters of brief recurrent episodes of up to ten a day, never reaching true status epilepticus. Seizures are often characterized by head and eye deviation that may alternate in laterality for different attacks; even in the same patient, EEG recordings corroborate independent right and left seizure onset, confirming the alternating clinical pattern [Vigevano, 2005].

While the overall prognosis remains excellent in terms of remission from seizures, there is a risk of paroxysmal dyskinesias in later childhood [Lee et al., 1998; Demir et al., 2005]. BFIS is the least well characterized genetically, with multiple loci implicated but few candidate genes. Multiple authors have linked the syndrome to the pericentromeric region of chromosome 16 [Lee et al., 1998; Demir et al., 2005]. There has also been one report linking the syndrome to chromosome 19 [Guipponi et al., 1997], and one linking it to chromosome 1 in a family with hemiplegic migraine and a mutation in the ATP1A2 gene, which encodes a sodium-potassium ATPase transporter [Vanmolkot et al., 2003]. In addition, mutations in SCN2A and KCNQ2 have been described [Striano et al., 2006; Zhou et al., 2006], suggesting a possible overlap of clinical and genetic characteristics of the three benign epilepsy syndromes occurring in the neonatal or neonatal-infantile period.

Dravet’s Syndrome

Dravet’s syndrome, previously known as severe myoclonic epilepsy in infancy (SMEI), was described by Charlotte Dravet in 1978 [Dravet, 1978]. This entity is an epileptic encephalopathy, a condition in which the epileptiform abnormalities themselves are believed to contribute to the progressive disturbance in cerebral function [Engel, 2001]. Onset is at approximately 6 months of age in previously healthy children. Dravet’s syndrome typically presents with prolonged hemiclonic or generalized febrile seizures, sometimes leading to status epilepticus. Myoclonic jerks, atypical absences, and complex partial seizures appear later in the course of the disease. Subsequently, children develop neurological deficits, such as ataxia and corticospinal tract dysfunction [Dravet et al., 1992], as well as psychomotor delay, behavioral disturbances, and significant learning problems [Wolff et al., 2006].

EEG findings may be normal at onset, and then show generalized, focal, and multifocal abnormalities, often associated with a strong photosensitivity. MRI is usually normal or, in a few cases, shows dilatation of the cisterna magna or slight diffuse atrophy [Dravet et al., 2005]. However, most recent neuroimaging techniques can identify hippocampal sclerosis in some patients during the course of the disease [Striano et al., 2007]. The onset during infancy and frequent association with prolonged febrile seizures help to distinguish the syndrome from epilepsy with myoclonic-astatic seizures, also known as Doose’s syndrome.

Major advances have occurred in our understanding of Dravet’s syndrome with elucidation of its association with SCN1A mutations. Since the report of Claes and colleagues in 2001 [Claes et al., 2001], the concept that SCN1A is the major gene responsible for Dravet’s syndrome has been established [Fujiwara et al., 2003; Wallace et al., 2003; Fukuma et al., 2004]. The gene encodes the neuronal sodium channel NaV1.1 alpha subunit, which has four homologous regions, each encoding six transmembrane domains and a region controlling interactions with the permeable ion. Mutations have been found throughout the gene [Gambardella and Marini, 2009], and a website has been created to track the increasing number of mutations at http://www.scn1a.info. There are currently more than 300 known SCN1A mutations, with mutational rates in patients with Dravet’s syndrome of 33–100 percent [Claes et al., 2003; Fukuma et al., 2004; Mulley et al., 2005; Sugawara et al., 2002; Nabbout et al., 2003; Wallace et al., 2003]. Mutations are commonly truncations, although missense mutations can occur [Claes et al., 2003; Kanai et al., 2004; Mulley et al., 2005]. Among SCN1A mutations, those associated with Dravet’s syndrome are quite variable, including frameshift, nonsense, deletion, amplification, and duplication. However, simple missense mutations still account for 40 percent of cases [Mullen and Scheffer, 2009].

Mutations are usually de novo, although cases of germline mosaicism have been reported and should be taken into account for genetic counseling [Gambardella and Marini, 2009]. Patients for whom standard polymerase chain reaction (PCR) fails to detect an SCN1A mutation may undergo additional testing with multiplex ligation-dependent probe amplification, which detects microchromosomal copy number variations, such as deletions or duplications, in an additional 10–25 percent of PCR-negative cases [Madia et al., 2006; Mulley et al., 2006; Suls et al., 2006; Marini et al., 2009]. In terms of treatment for Dravet’s syndrome, topiramate, stiripentol, valproate, clonazepam, and levetiracetam, as well as the ketogenic diet, have been effective. Lamotrigine and carbamazepine may induce seizure aggravation [Guerrini et al., 1998; Wakai et al., 1996].

The spectrum of SCN1A-related epilepsies has been extended in the past few years, and several subgroups related to Dravet’s syndrome have now been reported, such as the syndrome of borderline SMEI (SMEB), characterized by a lack of myoclonic seizures and generalized spike-wave discharges, and the intractable childhood epilepsy with generalized tonic-clonic seizures (ICE-GTC), in which tonic-clonic seizures predominate [Osaka et al., 2007; Fujiwara, 2006; Rhodes et al., 2005]. SCN1A has also been implicated in cases of familial migraine [Cestele et al., 2008; Vahedi et al., 2009] and in a variety of other epilepsy syndromes. While no association has been found with simple febrile seizures [Petrovski et al., 2009] in the 5 percent of Dravet cases found to have familial mutations, family members often have other phenotypes, such as generalized epilepsy with febrile seizures plus (GEFS+) [Wallace et al., 2001b; Sijben et al., 2009]. In a recent comprehensive study, Harkin and colleagues have analyzed a cohort of 188 patients with various epileptic encephalopathies and found SCN1A mutations in 48 percent of patients, mainly novel; 96 percent were de novo [Harkin et al., 2007]. No patients with West’s syndrome or progressive myoclonic epilepsy had an SCN1A mutation. However, mutations were not restricted to those with typical epileptic encephalopathies. Interestingly, a few patients with cryptogenic generalized or focal epilepsy, myoclonic-astatic epilepsy, and Lennox–Gastaut syndrome also carried mutations.

Because SCN1A is implicated in a variety of other epilepsy syndromes, some of which are more benign, there is new potential for confusion regarding prognosis for affected patients, and a new need for clinicians to understand the subtleties of genetic testing for SCN1A mutations. While many different SCN1A mutations have been identified in patients with Dravet’s syndrome, most of which are unique to individuals, several recurrent mutations have also been found [Mulley et al., 2005]. Mutations are spread throughout the gene and, therefore, have different predicted functional effects on the protein [Kanai et al., 2004]. However, it is apparent that wherever and whatever the functional effects of these mutations are, they all lead to a similar seizure phenotype. Collating the available functional data on such mutations does result in an obvious explanation of the shared epilepsy phenotype; however, mathematical modeling has predicted an increased excitability via augmented action potential firing [Spampanato et al., 2004]. Mutations in SCN1A are collectively the most frequent genetic cause of epilepsy. Hence, further efforts to clarify the precise pathophysiology are likely to be important to the fundamental understanding of epileptogenesis.

Generalized Epilepsy with Febrile Seizures Plus

Generalized epilepsy with febrile seizures plus (GEFS+), mentioned above, was first described by Sheffer and Berkovic, and is associated with SCN1A mutations in roughly 10 percent of cases [Scheffer and Berkovic, 1997]. Mutations are generally missense, and the syndrome is characterized by an autosomal mode of inheritance, a broad spectrum of phenotypes, and a penetrance of 60 percent. Cases are mostly considered to have generalized epilepsy. The most common phenotype is characterized by the association of febrile seizures (FS), or febrile seizures extending beyond 6 years of age (febrile seizures plus, FS+), with afebrile generalized tonic-clonic seizures. This condition is a heritable syndrome, so clinical characteristics may be present in multiple family members. Less common phenotypes show the association of FS+ with absences, or myoclonic or atonic seizures, while the most severe phenotype includes myoclonic-astatic epilepsy. Partial seizures have also been reported, though rarely [Baulac et al., 1999].

West’s Syndrome

The final syndrome to consider that has infantile onset and a genetic etiology, in a minority of cases, is West’s syndrome. While the majority of cases of infantile spasms have no genetic association, mutations in ARX, Aristaless-related homeobox gene, have been associated with cases of X-linked recessive infantile spasms or “ISSX” [Fullston et al., 2010]. Genetic testing may therefore be appropriate in cases in which multiple males are affected by infantile spasms and mental retardation; dystonia is also a feature [Poirier et al., 2008; Shinozaki et al., 2009]. A second, dominant form of ISSX has been associated with a balanced X-autosomal translocation that disrupts CDKL5 (see Neonatal Epileptic Encephalopathies). A final syndrome of epilepsy and mental retardation limited to females has also been described [Deprez et al., 2009; Hynes et al., 2009; Depienne et al., 2009; Dibbens et al., 2008]. This syndrome is often associated with autistic, obsessive, or aggressive traits, and has been linked to various mutation of PCDH19 (protocadherin 19). (For a broader review of infantile spasms, please refer to Chapter 56.)

Early- and Late-Onset Childhood Occipital Epilepsy

Early-onset benign childhood occipital epilepsy was first described by Panayiotopoulos in 1989 as an epilepsy syndrome characterized by the “ictal triad of nocturnal seizures, tonic deviation of the eyes and vomiting” [Panayiotopoulos, 1989]. The age of onset of Panayiotopoulos syndrome (PS) is typically between 3 and 5 years of age; it affects boys and girls equally and is characterized by a low seizure burden. However, seizures can be prolonged, are marked by autonomic features, and often involve alteration in consciousness. Seizures may march to involve clonic movement of the head and upper extremities or secondary convulsion. A minority of patients may develop seizures during the day. The syndrome is benign, with remission before age 12.

The EEG reveals normal background activity, but frequent surface-negative spike and slow-wave complexes over the occipital region, occurring singularly or repetitively with a frequency of 2–4 Hz. Discharges may be apparent bilaterally with a persistent voltage asymmetry or a shifting predominance, or they may be unilateral. The spike-wave complexes attenuate with eye-opening and fixation, and are induced in darkness or with eye closure. Generalized discharges may also be present, and were noted in 50 percent of the original series. Other focal discharges also occur. The occipital spike wave remains present despite treatment with antiepileptic medications, but tends to decrease with age [Capovilla et al., 2009].

The Gastaut type, also known as childhood epilepsy with occipital paroxysms (CEOP), was first described by Camfield in 1978 [Camfield et al., 1978] and elaborated by Gastaut in 1982 [Gastaut, 1982]. CEOP has a later onset, presenting between 3 and 16 years, and seizures are characterized by visual symptoms rather than gaze deviation. Visual symptoms may include amaurosis, phosphenes, illusions, or hallucinations, and may progress to involve automatisms or hemiclonic seizures. Events are followed by migraine and classically have a diurnal pattern. EEG abnormalities similarly involve occipital spike wave activated by eye closure. However, the prognosis is somewhat poorer in terms of chance for remission.

The Panayiotopoulos and Gastaut forms of CEOP are probably better viewed as a clinical spectrum. One-third of patients with CEOP appear to have a mixed syndrome, with features of both Panayiotopoulos and Gastaut forms [Taylor et al., 2008]. Similarly, while there is evidence for a genetic component to CEOP, with some reports of concordant monozygotic or dizygotic twins, the concordance does not appear to be higher in monozygotic twins, indicating that other epigenetic or environmental factors likely also play a role [Taylor et al., 2008]. Additionally, while a family history of epilepsy can be established in roughly 36 percent of patients with either form of childhood occipital epilepsy, affected family members have been found to have a variety of generalized or focal epilepsy syndromes rather than CEOP alone.

Benign Epilepsy with Centrotemporal Spikes

Benign epilepsy with centrotemporal spikes (BECTS) is the most common benign focal epilepsy of childhood, with an incidence of 10–20 per 100,000 in children under age 15. Onset occurs most commonly between 7 and 10 years, with a slight male predominance [Panayiotopoulos et al., 2008]. Seizures are focal and involve unilateral sensorimotor function of the face, speech arrest, and hypersalivation. Consciousness is generally maintained, although seizures may progress to become hemiconvulsive or generalized tonic-clonic.

The EEG reveals bilateral independent centrotemporal spikes reflecting a horizontal dipole with maximal negativity over the central region and maximal positivity over the frontal region. Epileptiform activity may be augmented in sleep and, rarely, other focal discharges or generalized spike wave has been reported. Three-quarters of rolandic seizures occur out of non-rapid eye movement (REM) sleep, often at sleep transitions.

Prognosis is excellent, with the majority of patients having fewer than ten seizures in total. Many will not require treatment with medications, and 10–20 percent will only have a single seizure. Remission occurs within 2–4 years of onset and prior to age 16 years. Rarely, and in the most severe 1 percent of cases, language and cognitive dysfunction can develop in association with significant activation of the EEG during sleep. Therefore, many view BECTS as being on a spectrum with other epilepsy syndromes, including Landau–Kleffner, atypical focal epilepsy of childhood, and epilepsy with continuous spike and wave during sleep (CSWS).

BECTS was classically believed to have a strong genetic component, probably based on the heritability of centrotemporal spikes, present in 11–48 percent of sibling pairs [Helbig et al., 2008], and early twin studies addressing overall heritability of the idiopathic epilepsy syndromes [Berkovic et al., 1998]. In addition, there have been case reports of mendelian inheritance patterns in families with variants of BECTS in which the disease is associated with dystonia [Guerrini et al., 1999] or with speech and language dysfunction [Hirsch et al., 2006; Roll et al., 2006; Scheffer et al., 1995b; Rudolf et al., 2009]. More recent twin studies, however, targeted to evaluate BECTS alone, have found relatively low concordance rates for the clinical syndrome, unlike the EEG trait of centrotemporal spikes, and suggest that nongenetic factors are important in the development of the phenotype [Vadlamudi et al., 2006]. While genetics is likely to play some role, given familial aggregation of the disease, the genetic risk appears to have been overestimated initially. Certainly, simple inheritance patterns only appear relevant in rare disease variants.

Childhood Absence Epilepsy

Childhood absence epilepsy (CAE) accounts for 5–15 percent of childhood epilepsy. It has a typical onset between 4 and 8 years of age, and there is a history of febrile seizures in 10–15 percent. Seizures are characterized by brief losses of awareness, lasting 10 seconds on average, and with a rapid return to baseline mental status within 2–3 seconds. Absences may be associated with a fixed blank stare, upward movement of the eyes, mild clonic, atonic, or tonic components, or quasi-purposeful movements. Tonic-clonic seizures occur in roughly 40 percent of patients but are infrequent, and absence status epilepticus occurs in as many as 10 percent. Myoclonic jerks are not typically seen.

Absences generally respond well to medical management, and remission is common. In the rare instance in which seizures persist into adulthood, they are often tonic-clonic in semiology, rather than absence. The EEG reveals a generalized 3-Hz spike wave that appears as an interictal trait, precipitated by hyperventilation, and in association with clinically apparent absence seizures.

CAE was previously categorized as an “idiopathic generalized epilepsy” (IGE) syndrome, along with juvenile myoclonic epilepsy (JME), juvenile absence epilepsy (JAE), and generalized epilepsy with tonic-clonic seizures. While the term “idiopathic” has now been abandoned, it may still be useful to think of these syndromes together in so far as their genetic underpinnings remain intertwined and without phenotypic specificity. Although rare families with mendelian inheritance patterns have been described for a subset of these syndromes [Cossette et al., 2002], polygenic inheritance, influenced by susceptibility genes, is widely accepted as accounting for the bulk of disease.

While the genetics for these previously categorized IGEs remains complex, a strong role for genetics has been well established through twin studies, demonstrating concordance for monozygotic twins of approximately 70 percent [Lennox, 1951; Vadlamudi et al., 2004]. In families without twins, however, the risk to siblings is roughly 6 percent, and as a result, only about one-third of families will report a family history. Within families with multiple affected members, there is often phenotypic heterogeneity, although there is some evidence for segregation of absence versus myoclonic phenotypes [Winawer et al., 2005].

A number of genes have been implicated in CAE. Mutations in GABRG2, which encodes the γ-2 subunit of the neuronal gamma-aminobutyric acid (GABA)_A receptor, have been implicated in the development of both CAE and febrile seizures [Wallace et al., 2001a; Kananura et al., 2002], as well as in GEFS+. Mutations in GABRA1, which encodes the alpha-1 subunit of the neuronal GABA_A receptor, have been associated with CAE in one series [Maljevic et al., 2006], and with JME in another [Cossette et al., 2002]. There is also evidence that the GABA_A receptor beta-3 subunit gene (GABRB3) may play a role in the development of CAE via reduced expression associated with mutation in its promoter [Urak et al., 2006]. In addition, malic enzyme 2, which plays a role in the synthesis of GABA, has been found to have a polymorphism associated with the adolescent-onset generalized epilepsies (JAE, JME, and generalized tonic-clonic seizures alone), conferring a sixfold increase in the odds of developing the disease [Greenberg et al., 2005].

Varied mutations in CACNA1H, the neuronal voltage-gated T-type calcium channel subunit, have also been implicated in CAE [Chen et al., 2003]. This channel is thought to play a key role in generating the 3-Hz spike wave that is characteristic of absence seizures. However, CACNA1H appears to function as a broader susceptibility gene and has also been shown to play a role in other generalized epilepsies, such as JME, JAE, and myoclonic astatic epilepsy, as well as in febrile seizures and even temporal lobe epilepsy [Heron et al., 2007b]. Similarly, CLCN2, a voltage-gated chloride channel, has also been implicated broadly in CAE, JME, JAE, and epilepsy with tonic-clonic seizures [D’Agostino et al., 2004]. SLC2A1, which encodes the GLUT1 glucose transporter, has recently been implicated in early-onset absence epilepsy before age 4 years [Suls et al., 2009].

The complex polygenic inheritance patterns involved in CAE and other genetic generalized epilepsy syndromes reveal that a single epilepsy phenotype can be generated by combinations of varied mutations, and that, conversely, a given mutation can be associated with more than one phenotype. With on-going research, understanding how such varied genetic factors can contribute to the development of seizures should tell us a great deal about the pathophysiology of epileptogenesis.

Juvenile Absence Epilepsy

Juvenile absence epilepsy (JAE) has an average age of onset of 12 years, representing a second peak in incidence after CAE with onset between 4 and 8 years. JAE is rarer than CAE, representing no more than 20 percent of absence epilepsy cases diagnosed. JAE is characterized by absence seizures, as described above for CAE, but there is a higher incidence of convulsive seizures, which occur in 80 percent, and myoclonus may occur in 15 percent, causing an area of overlap with JME.

The EEG reveals a generalized spike wave occurring ictally, interictally, and in response to hyperventilation. The spike wave is superimposed on a normal background and may be less regular or faster than is seen in CAE.

The course remains relatively benign, with few cases remitting but 80 percent responding well to first-line therapy. As noted above, there is evidence for polygenic inheritance, with many of the genes involved in CAE being also linked to JAE, and some evidence that there is more genetic overlap with CAE than with JME [Winawer and Shinnar, 2005].

Juvenile Myoclonic Epilepsy

Juvenile myoclonic epilepsy (JME) is believed to represent as much as 10 percent of all epilepsy cases. Onset occurs between ages 12 and 18 years. Seizures are often myoclonic, but generalized tonic-clonic occur in 95 percent. Events often begin just after awakening, and sleep deprivation and alcohol are commonly cited as triggers. Myoclonus most often involves the bilateral upper extremities, and as a result, it is common to elicit a history of difficulty eating breakfast, brushing teeth, or applying make-up in the morning. Generalized tonic-clonic seizures are often preceded by a period of increasing myoclonic jerks with preserved consciousness, allowing patients to predict their seizures on some occasions, despite the generalized onset. Absences are rare.

The EEG reveals generalized fast spike- and polyspike-wave activity that may have a frequency of 10–16 Hz during myoclonic jerks, or slow to 2–5 Hz with modest irregularity during more prolonged seizures. Generalized, fast-spike, and polyspike wave is also seen interictally superimposed on a normal background.

The prognosis for JME remains good, in so far as most cases are well controlled on medications, and as many as one-third may be able to cease medications during adulthood without relapse [Camfield and Camfield, 2009]. However, the disability associated with even rare convulsions should not be underestimated. In addition to the direct impact on morbidity and mortality, there are implications for driving privileges, and higher rates of unemployment have been reported [Camfield and Camfield, 2009].

As was discussed for CAE and JAE, JME has complex polygenic inheritance. There is a positive family history reported in 30–50 percent, and a number of common susceptibility genes have been identified to date. These genes include GABRA1, which encodes the alpha-1 subunit of the neuronal GABA_A receptor and has been associated with CAE and JME [Maljevic et al., 2006; Cossette et al., 2002], and malic enzyme 2, which is associated with adolescent-onset generalized epilepsies and plays a role in synthesis of GABA. In addition, CACNA1H and CLCN2, which encode a voltage-gated T-type calcium channel and a voltage-gated chloride channel, respectively, have been broadly implicated in primary generalized epilepsies and even in temporal lobe epilepsy [Heron et al., 2007; D’Agostino et al., 2004].

There are additional genes implicated specifically in the development of the JME phenotype. Several authors have reported gene mutations in EFHC1 associated with JME in a variety of affected families of differing origin [Suzuki et al., 2004; Annesi et al., 2007]. EFHC1 encodes a protein of unknown function with calcium-binding EF-hand motif, a helix-loop-helix structural domain found in a large family of calcium-binding proteins. There is evidence that the protein modulates apoptotic activity and R-type voltage-dependent calcium channel properties [Suzuki et al., 2004].

Epilepsy with Generalized Tonic-Clonic Seizures Alone

Previously known as epilepsy with generalized tonic-clonic seizures on awakening, this epilepsy syndrome was found to be associated with the adolescent onset of tonic-clonic seizures alone, without predilection for certain times of day or states [Reutens and Berkovic, 1995]. As a result, it is now referred to as epilepsy with generalized tonic-clonic seizures alone. The genetics of the syndrome are likely similar to those of CAE, JAE, and JME (see above), with complex polygenic inheritance of common susceptibility genes.

Autosomal-Dominant Nocturnal Frontal Lobe Epilepsy

Autosomal-dominant nocturnal frontal lobe epilepsy (ADNFLE) typically arises during the first (53 percent) or second (35 percent) decade of life, with a median age of onset of 8 years [Scheffer et al., 1995]. Clinically, the disease is characterized by nocturnal tonic or hypermotor activity, such as sitting up in bed, crawling into the bed, or being flung from bed with associated injury. Awareness is generally preserved, although there may be an inability to respond. Seizures occur in clusters and are most commonly reported soon after falling asleep or in the early hours of the morning. Clusters consist of a median of 6 attacks, but can range from 1 to 72, in extreme cases. Rarely, seizures are reported from naps or daytime wakefulness. Seizures are generally brief (median 60 seconds) and often accompanied by aura, which may awaken the subject from sleep prior to the first attack. The pattern and semiology of the seizures often lead to misdiagnosis with nocturnal parasomnia or other psychiatric or medical illness.

The EEG is notable for rare interictal epileptiform activity (16 percent) or slowing (22 percent), with the majority having no discernible interictal abnormalities (84 percent). The ictal EEG may show frontally predominant sharp- and slow-wave activity, or may fail to reveal a definite ictal pattern. Seizures most frequently arise from stage 2 sleep. Review of semiology often reveals prominent dystonic posturing.

Outcomes for ADNFLE are good, although the disease does not typically remit. Individuals demonstrate normal intellect and neurologic examination. The great majority requires antiepileptic medication, but less than one-third requires combination therapy. For example, in the description by Scheffer and coauthors [Scheffer et al., 1995], 32 percent of patients were well controlled on carbamazepine monotherapy, while 29 percent required more than one agent. There may be improvement in later life, with rare individuals successfully weaning medications after age 50 years.

The inheritance of ADNFLE is autosomal-dominant, as the name states, with a penetrance of 70–80 percent [Scheffer et al., 1995]. The phenotype has been associated with various mutations in genes encoding the alpha-4 (CHRNA4), beta-2 (CHRNB2), and alpha-2 (CHRNA2) subunits of the nicotinic acetylcholine receptor, as well as with a variant of the promoter for the corticotropin-releasing hormone (CRH). The acetylcholine receptor is a pentameric ion channel consisting of predominantly alpha-4 and beta-2 subunits. The second transmembrane domain of each subunit lines the ion channel pore and is the site for the majority of the mutations that have been described for CHRNA4 [Steinlein et al., 1995; Weiland and Steinlein, 1996] and CHRNB2 [Diaz-Otero et al., 2008; De Fusco et al., 2000; Phillips et al., 2001]. In fact, independent mutations in particular conserved amino acid residues in these domains have been demonstrated in families of differing ethnicity, suggesting that these sites play an important role in the development of the syndrome. Functionally, the mutations appear to alter the desensitization kinetics of the receptor, thereby contributing to hyperexcitability and seizure [Weiland et al., 1996]. In addition, in the case of CHRNB2, one of the mutations is associated with a more pronounced blockade of current by carbamazepine, suggesting a mechanism for treatment responsiveness.

Despite the evidence that the second transmembrane domain is important for the development of the disease, mutation of this region is not essential. One family has been described with ADNFLE and memory difficulties, associated with a mutation involving the third transmembrane region of CHRNB2 [Bertrand et al., 2005], and another family has been described with a CHRNA2 involving the first transmembrane domain of the receptor and affecting the receptor sensitivity to acetylcholine [Aridon et al., 2006]. Finally, despite an increasing number of nAChR mutations demonstrated in families with ADNFLE, the majority of cases remain without a putative gene, despite screening for nAChR mutations. Recently, a variation in the promoter of the corticotropin-releasing hormone gene (CRH) has been linked to the disease in one family [Combi et al., 2005], and it is possible that several additional loci will be found to explain the remaining cases, potentially helping our understanding of the various intersecting mechanisms at play in the development of the phenotype.

Autosomal-Dominant Partial Epilepsy with Auditory Features

Autosomal-dominant partial epilepsy with auditory features (ADPEAF), also known as autosomal-dominant lateral temporal lobe epilepsy (ADLTE), is a heritable partial epilepsy first described by Ottman and co-workers in 1995 [Ottman et al., 1995], with onset ranging from the second to fourth decades; it is estimated to make up 19 percent of genetic partial epilepsies [Michelucci et al., 2009]. To date, at least 35 families from Europe, the United States, Australia, and Japan have been reported, adding up to over 200 cases [Michelucci et al., 2003]. Seizures are focal, with a frequency from several per month to as few as twice a year. Secondary generalization occurs once or twice a year in 79–90 percent of cases. Auditory auras occur in 27–64 percent of cases, depending on case series, and may consist of poorly formed sounds, such as buzzing or ringing (74 percent); well-formed sounds, such as specific songs or voices (11 percent); or sound distortions, such as volume changes or muffling (28 percent). A variety of other auras can occur, including visual (17 percent), psychic (16 percent), autonomic (12 percent), vertiginous (95), and other sensory (13 percent) [Winawer et al., 2000]. Aphasia may also occur and is reported in 17 percent of cases.

The interictal EEG is often normal, but may show temporal slowing or epileptiform transients arising from the temporal region. Ictal EEG reveals seizure activity arising from the temporal region. Prognosis is good, in so far as seizures can be rapidly controlled with antiepileptic medications; however, there is a tendency for them to recur when patients are weaned off medications.

Roughly half of cases are associated with an autosomal-dominant mutation in LGI1, or leucine-rich glioma inactivated 1 protein [Ottman et al., 2004; Michelucci et al., 2003; Kalachikov et al., 2002]. Over twenty distinct mutations have been described, resulting in truncations, deletions, or single amino acid substitutions. Rarely, mutations are associated with idiopathic generalized epilepsy syndromes in some family members as well [Ottman et al., 2004]. Penetrance for the mutations ranges from 54 to 71 percent. Families with ADPEAF, for which no mutation has been identified, are more likely to have autonomic features associated with their epilepsy (56 percent vs. 16 percent) in some series, possibly indicating a more mesial onset.

It is of note, that LGI1 is one of few genes implicated in epilepsy that is not part of an ion channel, and its pathophysiologic role is under investigation. It has been shown to be downregulated in high-grade gliomas, although thus far there is no known connection between ADPEAF and increased risk for malignancy. In vitro studies show that the mutant form of LGI1 fails to be secreted [Michelucci et al., 2003], thereby inhibiting its ability to anchor to postsynaptic scaffolding proteins, where it plays a role in AMPA receptor-mediated synaptic transmission [Fukata et al., 2006]. Additional studies have found a role for the protein in presynaptic potassium channel assembly, with mutant forms of the protein resulting in changes in the inactivation kinetics of the channel [Schulte et al., 2006], and thereby potentially providing a mechanism for hyperexcitability.

Familial Partial Epilepsy with Variable Foci

Familial partial epilepsy with variable foci (FPEVF) is an autosomal-dominant focal epilepsy syndrome in which seizures may arise from different regions in various affected family members. The syndrome has been described in one Australian [Scheffer et al., 1998], one Dutch [Callenbach et al., 2003], one Spanish (Berkovic et al., 2004), and one large French–Canadian family [Xiong et al., 1999], with an age of onset ranging from the first to the third decade, median 7 years and mean 13 years. The syndrome is distinguished from the two other autosomal-dominant partial epilepsies (ADNFLE and ADPEAF) by the presence of daytime seizures, although nocturnal frontal lobe seizures often occur and may make identification of the syndrome more challenging. Seizures are most often preceded by autonomic, somatosensory, or sensory auras, may be followed by automatisms, and may be associated with hypermotoric, tonic, or tonic-clonic activity. Secondary generalization does occur. Auditory symptoms are not reported. While various family members may have different seizure symptoms, depending on the origin of their seizures, any given family member has one consistent seizure type throughout life.

With repeated EEGs, interictal abnormalities are found in roughly 50 percent of recordings and most often reveal epileptiform activity arising from the frontal or temporal regions. Ictal EEG shows focal onset to seizure activity. Imaging studies are generally normal. Prognosis is reasonable, in so far as patients maintain normal IQ and most achieve adequate control of their seizures; however, as many as half will require dual therapy with antiepileptic medications, and very few are able to be weaned off medications.

The syndrome has been linked to chromosome 22q in 3 of the 4 families [Callenbach et al., 2003; Xiong et al., 1999], with a penetrance of about 50 percent. In the remaining family, linkage was made to chromosome 2q [Sheffer et al., 1998], and some have suggested that this family may be clinically distinct due to a higher frequency of daytime seizures. A putative gene has not been identified for either linkage.

Other Mendelian Focal Epilepsies

While mesial temporal lobe epilepsy has classically been thought of as an acquired epilepsy syndrome, there is increasing appreciation for a genetic role in some variants. The genetic forms have been divided, based on whether there is associated hippocampal sclerosis or history of febrile seizures, and no putative genes have yet been identified, although linkage has been established in some cases.

In familial mesial temporal lobe epilepsy (FMTLE) without hippocampal sclerosis or febrile seizures, onset occurs in adolescence before age 18 years. Seizures are characterized by prominent psychic auras, such as déjà-vu, which often can be quite intense and frequent. Seizures with alteration in awareness, however, occur only infrequently, and secondary generalization is rare. EEG and MRI are normal and prognosis is good. A linkage has been found to chromosome 4q in one family [Hedera et al., 2007]. No gene has been identified, and the area of interest has not been found to contain any homologs of previously identified genes involved in heritable epilepsies. The linkage study suggests autosomal-dominant inheritance with variable penetrance.

In FMTLE with hippocampal sclerosis, the mean age of onset is around 10 years and seizures are more likely to involve alteration in consciousness and postictal confusion, as is seen in nonfamilial MTLE. Preceding febrile seizures are reported in about 10 percent and MRI reveals varying degrees of hippocampal sclerosis. While most affected individuals have a benign course and may remit, refractoriness is seen in as many as one-third of affected family members. Currently, there is no genetic linkage or candidate gene [Gambardella et al., 2009].

FMTLE with febrile seizures has been described in two large families with different linkage. Onset is in the first or second decade without hippocampal sclerosis on MRI, and the course is usually benign. In the first family, digenic inheritance was suggested, with linkage to 18qter and 1q25–31 [Baulac et al., 2001]. In the second, linkage was found to 12q22–23.3 [Claes et al., 2004]. Of course, febrile seizures also occur in other inherited epilepsies, such as GEFS+, and are certainly not specific to these rare cases; rather, febrile seizures probably represent a broad marker for increased susceptibility to seizure.

It should also be noted that temporal lobe epilepsy can occur as part of other inherited syndromes, such as FPEVF, discussed above. Temporal lobe seizures are also a feature of familial occipitotemporal lobe epilepsy, described in one Belgian family. This inherited epilepsy has a variable age of onset (mean 21 years), with migraine with visual aura occurring independently of seizures in 50 percent of affected individuals. Seizures are typically partial without alteration of consciousness and frequently consist of visual phenomena. EEG and MRI are normal, and prognosis is good. The syndrome has linkage to 9q21–q22, with a dominant mode of inheritance and a penetrance of 75 percent [Deprez et al., 2007].

The final mendelian focal epilepsy described to date is partial epilepsy with pericentral spikes. This syndrome was described in one large Brazilian family of Portuguese ancestry. Onset occurs in the first or second decade of life, and seizures may be hemiclonic, hemitonic, generalized tonic-clonic, or focal, with or without alteration in consciousness. EEG shows evidence of spikes or sharp waves in the pericentral region. The prognosis is good, with seizures remitting spontaneously or with a single antiepileptic agent. Linkage has been found to chromosome 4p15 [Kinton et al., 2002].