EEG Findings

The ictal EEG discharge is often a generalized spike wave (GSW) that may be slower than 3 Hz [Hirano et al., 2009]. Occasionally, the myoclonic seizures can be massive and result in a fall, or they may occur in a cluster. Some infants may exhibit photosensitivity at the onset of the syndrome, and the photic-induced myoclonic seizures are often more prominent than those that occur spontaneously [Capovilla et al., 2007]. Other infants may be sensitive to acoustic or tactile stimuli, leading some to argue for a distinctly separate epilepsy syndrome [Ricci et al., 1995]. We and others [Roger et al., 2002] do not believe this is a clinically distinct syndrome, but rather that the syndrome of BME may include both photosensitive and stimulus-provoked seizures in addition to spontaneous myoclonic seizures. In both groups, the waking EEG in between seizures is often normal but may contain rare generalized spike/wave discharges [Ricci et al., 1995; Capovilla et al., 2007], and the ictal EEG findings are indistinguishable.

Treatment and Outcome

The overall prognosis of myoclonic seizures is excellent, with complete resolution in almost all children usually within 1–2 years of diagnosis. In those who demonstrate photosensitivity, these seizures may be more difficult to control and may persist for a longer period of time [Roger et al., 2002]. The majority of children also have normal neurodevelopmental outcomes [Caraballo et al., 2009], although some studies have reported impaired psychomotor development and behavioral disturbances if the child is not treated or if onset of the syndrome is at less than 2 years of age [Mangano et al., 2005]. Those children with a prominent reflex component may have a more favorable neurodevelopmental outcome [Zuberi and O’Regan, 2006].

The medication of choice appears to be valproic acid (VPA), with 80–90 percent responding to VPA monotherapy, although high serum levels (>100 mg/L) may be necessary. When VPA monotherapy does not provide complete control, adjunctive use of a benzodiazepine, such as clonazepam or clobazam, can be helpful [Rossi et al., 1997].

Mutations in the SCN1A Gene

The importance of SCN1A gene mutations as an underlying etiology in epilepsy first became apparent when Scheffer and Berkovic reported missense mutations in two families with the syndrome of generalized epilepsy with febrile seizures plus (GEFS+) [Scheffer and Berkovic, 1997]. Due to the common occurrence of febrile seizures in both GEFS+ and Dravet’s syndrome, Claes et al. screened seven patients with Dravet’s syndrome and found de novo missense mutations in all seven patients [Claes et al., 2001]. Testing for mutations in the SCN1A gene is clinically available, and hundreds of patients with Dravet’s syndrome with SCN1A mutations have been reported [Depienne et al., 2009]. The main difference between individuals with a GEFS+ phenotype and a Dravet phenotype appears to be that the former often have a missense mutation with reduced penetrance, whereas patients with Dravet have mutations that occur in isolated patients (arise de novo) [Claes et al., 2001].

Mutations in the SCN1A gene are seen in approximately 70–85 percent of patients with Dravet’s syndrome [Harkin et al., 2007], and therefore it remains a clinical diagnosis based on age of onset and clinical phenotype. One should not exclude the diagnosis on the basis of a negative result on SCN1A mutation testing. Approximately 200 different point mutations in the SCN1A gene were identified in 271 probands with a strict clinical diagnosis of Dravet’s syndrome [Depienne et al., 2009]. Of those probands without a point mutation, the multiplex ligation-dependent probe amplification (MPLA) technique found an additional 14 micro-rearrangements of SCN1A [Depienne et al., 2009]. An SCN1A variant database was published in 2009, listing each type of mutation and the associated epilepsy syndrome [Claes et al., 2009]. As of 2009, 648 point mutations have been reported, of which 582 (90 percent) are in subjects with Dravet’s syndrome. An additional 67 genomic rearrangements were reported, with 46 (7 percent) of these occurring in patients with Dravet’s syndrome. The remaining clinical phenotypes with SCN1A mutations are wide and varied [Harkin et al., 2007]. In Dravet’s syndrome, the end result of each of these mutations appears to be a total loss of function of the mutated allele, most commonly caused by a missense mutation that leads to an abnormality in the pore-forming unit of the sodium channel [Claes et al., 2009]. While specific mutation genotype–phenotype correlations within Dravet’s syndrome are not yet available, further research in this area may lead to particular treatment strategies for specific genotypes.

Seizure Semiologies

Patients with Dravet’s syndrome have many different seizure semiologies, often occurring in the same patient. The most common presenting seizure type is a prolonged febrile convulsion lasting longer than 25 minutes, but afebrile seizures can be the presenting seizure type in as many as 35 percent [Roger et al., 2002]. Japanese authors have observed that hot water immersion can trigger seizures in these patients [Fujiwara et al., 1990], and this feature is one part of a clinical screening test aimed at predicting the diagnosis of Dravet’s syndrome [Hattori et al., 2008]. Other convulsive seizure types include generalized tonic-clonic seizures, hemiclonic seizures, and “falsely generalized,” seizures as described by Dravet [Roger et al., 2002]. The latter has a complex semiology, with a high degree of discrepancy between the clinical and EEG findings. Seizures often start in one part of the body, spread to another (on the opposite side), and then potentially return to the original side of origin, only to involve a new body part.

Nonconvulsive seizures are also common and include simple partial, complex partial, atypical absence, and myoclonic types. The specific semiologies of each of these seizure types are similar to that seen in other forms of childhood epilepsy. Tonic seizures are rare. A unique seizure type, termed “obtundation status,” is unique to the syndrome. The semiology of this seizure type consists of variable impairment in consciousness, and fragmentary and erratic segmental myoclonias involving the limbs and face. Patients may still be able to engage in simple activities, such as eating or playing with a toy. These episodes may last several hours to days and can either be initiated by, or conclude with, a convulsion [Roger et al., 2002]. The EEG during these episodes is not a pattern of classical atypical absence status, but rather is characterized by a diffuse dysrhythmia with focal or diffuse spikes [Roger et al., 2002]. We have observed one patient, who would begin with this type of clinical phenomenon, including prominent eyelid myoclonia that was dramatically accentuated with eye closure, and would remain in this state for 2–3 days, culminating in a generalized convulsion, after which her mental status cleared. She would remain clear for 2–3 weeks, and then repeat the same cycle over again.

EEG Findings

The EEG is invariably normal early in the course of the disease, and therefore, in contrast with other epilepsy syndromes, the diagnosis must be suspected on clinical grounds. As the disease progresses, the EEG background becomes slow but often continues to manifest normal sleep architecture [Roger et al., 2002]. Some authors have observed rhythmic 5–6-Hz centroparietal theta but this is not specific to the syndrome [Doose et al., 1998]. Over time, the interictal EEG can show both generalized and multifocal spikes, or may not show any of these features despite frequent seizures of multiple types, further highlighting the lack of specificity of EEG findings. Given this lack of specificity, a screening test for predicting the diagnosis of Dravet’s syndrome under 1 year of age has been developed, with a positive predictive value as high as 94 percent [Hattori et al., 2008]. This screening test uses simple clinical features, as shown in Table 56-2. A sum score of 6 or greater is associated with a high risk of having Dravet’s syndrome, and therefore SCN1A testing should be done in those cases. Even if the hot water-induced risk factor is excluded, a score of 5 or higher is still indicative of a high-risk patient.

Table 56-2 Clinical Screen for Dravet’s Syndrome

(Adapted from Hattori J et al. A screening test for the prediction of Dravet syndrome before one year of age. Epilepsia 2008;49(4):626–633.)

Imaging

Magnetic resonance imaging (MRI) studies are usually normal but may show nonspecific cerebral or cerebellar atrophy or isolated ventricular enlargement. Interestingly, these findings, specifically diffuse brain atrophy, appear to be more common in those children without associated SCN1A mutations. Furthermore, despite having frequent prolonged febrile and afebrile convulsions, mesial temporal sclerosis is rare, reported in only 1 of 58 patients [Striano et al., 2007].

Treatment

Until recently, treatment strategies were generally unsuccessful, with poor seizure control despite polytherapy. Many have observed that certain antiepileptic drugs (AEDs) exacerbate seizures, with the most notorious agents being carbamazepine (CBZ) and lamotrigine (LTG) [Wakai et al., 1996; Guerrini et al., 1998]. Some authors have recommended using CBZ as a means of “confirming” the diagnosis when there is a high index of clinical suspicion [Wakai et al., 1996]. Given the availability of genetic testing, we do not endorse this practice.

A study by Chiron et al. and a subsequent meta-analysis pooling these data with unpublished data showed that combination therapy with valproate (VPA) + stiripentol (STP) + clobazam (CLB) resulted in a 70 percent decrease in seizure frequency, with 43 percent of patients becoming seizure-free [Chiron et al., 2000; Kassai et al., 2008]. In this study, the STP dose began at 50 mg/kg/day divided in 2–3 doses, but could be titrated up to 100 mg/kg/day. The maximum recommended dose is 3500 mg/day [Chiron et al., 2000].

One proposed mechanism for the effect of STP is the potent inhibition of the p450 cytochrome system, resulting in higher levels of VPA and CLB. Plasma levels of CLB were significantly higher but levels of VPA were not [Chiron et al., 2000]. Interestingly, one other result of p450 inhibition is lower levels of metabolites, some of which are thought to explain some of the adverse toxic effects of AEDs. This finding perhaps explains why patients are able to tolerate the higher doses of certain AEDs [Chiron, 2005]. These observations, however, raise the possibility that STP has other mechanisms of action independent of its effects on the p450 system, including enhancement of gamma-aminobutyric acid (GABA)ergic transmission [Quilichini et al., 2006].

Of the newer agents, topiramate (TPM) appears to be very effective, especially if added to STP, resulting in 50 percent reduction of seizures in 78 percent of patients at a modest optimal dose of 3.2 mg/kg/day. Seventeen percent of patients remained seizure-free for at least 4 months [Kroll-Seger et al., 2006].

Bromide therapy has also shown promise in several studies [Oguni et al., 1994], specifically in reducing the number of convulsions. The ketogenic diet may be helpful in some patients [Caraballo et al., 1998].

Our strategy has been to start with VPA or TPM, followed by add-on therapy with a benzodiazepine in the form of clonazepam, clorazepate, or clobazam. When seizures prove to be refractory to this combination, STP is then added. If STP were more readily available in the United States and the cost were not prohibitive, we would very likely be able to utilize this medication earlier in the treatment algorithm.

Aggressive treatment of acute seizures and prevention of status epilepticus is critical, as some small studies have shown a trend towards improved neuropsychologic outcomes in those children with less frequent convulsive seizures [Chipaux et al., 2008]. Each patient should have a clearly defined “emergency plan,” and the clinician should ask each family if a particular treatment regimen has been particularly effective for their child [Nolan et al., 2006]. Parents should be educated about the inevitability of seizures, especially prolonged seizures in the setting of fever. Early use of high doses of benzodiazepines, either in the form of rectal diazepam (as high as 1 mg/kg per dose) or buccal midazolam, should be the first-line therapy. Some physicians have recommended insertion of a central venous port for those children who have proven to have difficult intravenous access and multiple episodes of status epilepticus [Dooley et al., 1995]. In these specific cases, families are able to cope much better, knowing that multiple intravenous attempts will not be necessary and appropriate treatment can be initiated to terminate the seizure faster.

Outcome

Long-term outcome with regard to seizure control and neuropsychological development is variable but generally poor. By definition, all children with Dravet’s syndrome have normal development at onset. Around 50 percent of children walk unsupported by a mean of 16 months of age; however, it is rare for children to utilize two-word sentences at the normal age of around 2 years [Wolff et al., 2006]. At around 2 years of age, there appears to be a gradual decline, and then there is a relative stabilization between the ages of 4 and 6 years [Wolff et al., 2006]. A trend was noted that fewer convulsive seizures often resulted in higher developmental quotients. Subsequent small cohorts have reported patients with normal or near-normal IQs and this favorable outcome is attributed to early diagnosis and appropriate management [Chipaux et al., 2008]. It remains to be seen in larger prospective studies whether early diagnosis and appropriate management with the medications outlined above will lead to improved seizure control and a subsequent improvement in neurodevelopmental outcome.

Etiology

Genetic factors clearly play a role in the pathogenesis of MAE, based on the high incidence of seizures (32 percent) in probands’ siblings and parents; the most common seizure type is absence seizures. Despite the initial discovery of the SCN1A gene in a family with GEFS+ that included a family member with MAE [Scheffer and Berkovic, 1997], only a small number of patients with mutations in the SCN1A gene have been identified in patients with MAE [Ebach et al., 2005; Harkin et al., 2007]. Further genotype–phenotype correlation studies are necessary to determine if there are underlying genetic factors that would be able to predict response to treatment and outcome.

Seizure Semiologies

Onset of MAE is often between the ages of 2 and 6 years, and children are usually developmentally normal [Kaminska et al., 1999]. These children may begin experiencing unexplained “jerks” and “falls,” which occur multiple times per day. These myoclonic seizures can take place in isolation, but often are followed by a brief period of atonia that results in a dramatic fall in which the child appears to be propelled to the ground. Polygraphic studies have documented that atonic falls can occur with or without the preceding myoclonus, and occur in about 64 percent of patients [Oguni et al., 2002]. Although atonic falls are not necessarily seen in all patients, more subtle myoclonic/atonic seizures are common. The only manifestation of these may be a brief head nod or “head drop,” with or without a myoclonic jerk of the upper extremities. When head drops occur in clusters with an associated alteration in level of consciousness, nonconvulsive status should be suspected. EEG findings during these episodes are always abnormal, with frequent runs of generalized spike/wave and slow spike-and-wave complexes. Absence seizures are quite common and may be similar to those seen in typical childhood absence epilepsy, but often are accompanied by more prominent myoclonus of the proximal upper extremities.

EEG Findings

The interictal EEG findings are varied and can be normal. The most common findings are generalized 3-Hz spike/wave discharges, as well as rhythmic 4–7-Hz biparietal theta activity. Some patients may also show occipital intermittent rhythmic delta activity (OIRDA) that characteristically attenuates with eye opening. Activation of generalized discharges during sleep is common, but generalized fast rhythms, such as generalized periodic fast activity (GPFA), are uncommon.

Treatment

Early recognition of the syndrome allows treatment to be focused on the broad-spectrum agents that have been shown to be particularly effective. The most commonly used and most effective AEDs have been VPA, LTG, levatiracetam, TPM and various benzodiazepines (clonazepam, clobazam, clorazepate). When absence seizures are frequent, ethosuximide (ETX) can be effective and has also been reported to decrease the frequency of myoclonic seizures; however, it is rarely used as monotherapy for this indication [Oguni et al., 2002]. The combination of VPA/ETX may be particularly effective, especially if high doses of VPA are used.

In those children who are refractory to medical management, a trial of the ketogenic diet should be considered early. The ketogenic diet has been shown to be as effective as other first-line agents in children with MAE [Kilaru and Bergqvist, 2007], and given this efficacy, we consider the ketogenic diet early in our treatment algorithm, especially if there is associated neurodevelopmental decline.

Outcome

The outcome with regard to cognitive development is highly variable but appears to be somewhat dependent on degree of seizure control. Some children may have fairly frequent seizures of multiple types and still have spontaneous remission, usually after approximately 3 years [Kaminska et al., 1999; Oguni et al., 2002]. In some children, the multiple seizure types remain intractable and these children have a higher risk of both cognitive and behavioral disturbances [Oguni et al., 2002]. The occurrence of tonic seizures has been described as a negative prognostic sign, although it remains unclear if these children may actually have an atypical form of Lennox–Gastaut syndrome rather than an atypical form of MAE [Kaminska et al., 1999].

Seizure Semiologies

In the classic syndrome of JME, myoclonic seizures may precede the first generalized tonic-clonic (GTC) seizure by 6–12 months, although GTCs occur as the first seizure type in approximately one-third of patients [Martinez-Juarez et al., 2006]. Some have proposed specific subgroups of JME, separating those patients who present with typical childhood absence or juvenile absence seizures, although this is much less common than the classical presentation of JME, accounting for only 10 percent of cases. The impact this distinction has on outcome will be discussed below. Photosensitivity is relatively common, occurring in approximately 30 percent of patients [Zifkin et al., 2005].

Myoclonic seizures often occur in the early morning hours shortly after awaking. Transcranial magnetic stimulation studies in patients with JME have demonstrated increased cortical excitability that is not present in other patients with focal epilepsy [Badawy et al., 2006]. Provocation of myoclonic seizures has been linked to higher cognitive tasks requiring higher-order thinking, such as writing or written calculation [Matsuoka et al., 2000].

Specific lifestyle features have been commonly associated with breakthrough seizures, including sleep deprivation, stress, alcohol, and menses [Martinez-Juarez et al., 2006].

EEG Findings

Patients with JME often have abnormal interictal EEGs, with the most common finding being generalized 4–6-Hz polyspike-and-wave (61 percent) or 3-Hz spike-and-wave discharges (14 percent) [Martinez-Juarez et al., 2006]. More prolonged (1–6 days; average 1–2 days) video EEG studies have shown EEG abnormalities of JME in 88 percent of patients [Park et al., 2009]. Other series have reported on the increased yield of an early morning EEG compared to an afternoon EEG (generalized epileptiform activity in 69 percent and 25 percent, respectively) [Labate et al., 2007]. A photoparoxysmal response in treatment-naive patients is also common and has been seen in as many as 35 percent of patients [Specchio et al., 2008].

Pathophysiology

It has long been accepted that the IGEs have a strong genetic component [Berkovic et al., 1998], and given the incidence of JME as a subtype of IGE, much research has been done to understand the underlying genetic basis of JME better. While JME appears to be a relatively homogeneous epilepsy syndrome, there are a number of studies implicating various genetic abnormalities. These include mutations in many different ion channels, including calcium, potassium, and chloride channels linked to at least seven different loci. There are also nonionic mechanisms that relate to neural migration during cortical development, and may explain some of the imaging and postmortem pathologic findings seen in these patients [de Nijs et al., 2009]. For further details about the specific genetic mutations and channels involved, please refer to Chapter 52.

Imaging

Imaging studies in patients with JME have indicated subtle structural changes in mesiofrontal cortex, with an increase in cortical gray matter [Woermann et al., 1999], although this has not been replicated in subsequent studies [Roebling et al., 2009]. These findings are interesting, given that the generalized discharges seen on routine EEGs often have a frontal predominance. This has been further investigated with dense array EEG, which further localized typical 4–6-Hz epileptiform activity to the orbitofrontal/medial frontopolar cortex, whereas some patients also demonstrated basal-medial-temporal sources [Holmes et al., 2010]. These electrophysiologic and anatomic abnormalities appear to be concordant with some of the neuropsychological deficits that are seen in patients with JME. Specific attention has been paid to executive dysfunction that has been demonstrated to correlate with smaller thalamic volumes and more frontal cerebrospinal fluid volumes in children with recent-onset JME when compared to children with benign rolandic epilepsy [Pulsipher et al., 2009].

Treatment

Valproate has long been considered the first-line agent, with reports of up to 86 percent of patients becoming seizure-free for at least 1 year [Penry et al., 1989]. Valproate has also been shown to be particularly effective in those patients with photosensitivity [Covanis et al., 2004]. Despite this efficacy, valproate has many side effects that may lead to significant comorbidities, including weight gain, hyperactivity, transaminitis, thrombocytopenia, pancreatitis. Women of child-bearing age must also consider the possible long-term cognitive effects on a fetus [Meador et al., 2009]. Due to these concerns, many newer AEDs are also commonly used, including lamotrigine, topiramate, zonisamide, and levetiracetam.

Lamotrigine is now accepted to be a broad-spectrum AED that is effective against both partial and generalized seizures [Biton et al., 2005]. Small studies have also shown it to be effective specifically in the management of patients with JME, both as monotherapy and polytherapy, resulting in seizure-free rates between 40 and 83 percent [Buchanan, 1996; Prasad et al., 2003]. LTG is also very well tolerated; however, it does have the potential for exacerbating myoclonic seizures [Prasad et al., 2003; Crespel et al., 2005]. The combination of LTG and VPA has been reported to be particularly effective in both partial and generalized epilepsy, including JME, and suggests the possibility of a synergistic effect [Brodie and Yuen, 1997].

Topiramate is also a broad-spectrum AED that has been shown to be effective in patients with JME, although there are no published studies on its use as monotherapy [Biton and Bourgeois, 2005]. It appears to be equally effective as LTG and VPA when used as polytherapy, although there is a suggestion that tolerability may be inferior [Prasad et al., 2003].

Zonisamide (ZNS) is another broad-spectrum AED with multiple proposed mechanisms of action. Small studies have shown ZNS to be broadly effective against all of the seizure types in JME, with 69, 62, and 38 percent of patients being free of GTC, myoclonic, and absence seizures, respectively [Kothare et al., 2004].

Levetiracetam is gaining a reputation as a broad-spectrum agent, as it has shown to be effective in patients with various forms of idiopathic generalized epilepsies including epilepsy with eyelid myoclonias and absence [Striano et al., 2009], as well as JME [Specchio et al., 2006; Noachtar et al., 2008]. In a study of 120 patients (93 percent of whom had JME) randomized to add-on levetiracetam or placebo, 25 percent were free of myoclonic seizures and 21.7 percent were free of all seizures, including GTCs, during the 12-week treatment phase [Noachtar et al., 2008]. In another open-label study of 48 patients with JME (10 newly diagnosed and 38 resistant to prior treatment), 73 percent were free of GTCs and 37.5 percent were free of myoclonic seizures over the study period, with a mean follow-up of 19 months [Specchio et al., 2006].

Benzodiazepines are widely accepted as having excellent antimyoclonic effects. This is also true of the myoclonic seizures seen in JME, with up to 88 percent of patients on clonazepam having complete control of myoclonic seizures but only 43 percent becoming free of GTCs [Obeid and Panayiotopoulos, 1989]. For this reason, benzodiazepines are not recommended as monotherapy for patients with JME, but can be extremely successful as add-on therapy when myoclonic seizures persist.

Outcomes

It has been widely accepted that JME is a lifelong condition with a high rate of relapse if weaned off AED therapy; however, some recent long-term follow-up studies have shown conflicting results [Martinez-Juarez et al., 2006; Baykan et al., 2008; Camfield and Camfield, 2009]. Furthermore, one must look at specific seizure types, as it appears the natural history of myoclonic seizures decreases after the fourth decade of life, independent of the on-going occurrence of the other seizure types [Baykan et al., 2008].

In the study by Martinez-Juarez et al., 257 patients were followed for a mean of 11± 6 years, and all patients were given a specific subtype of JME, including classic JME (72 percent), childhood absence epilepsy (CAE) evolving into JME (18 percent), JME with adolescent absence (7 percent), and JME with astatic seizures (3 percent). During a mean follow-up of 12.4 years in the classic JME group, 58 percent were free from all seizure types, with 91 percent of these patients remaining on AED treatment and only 9 percent seizure-free off medications. Treatment with VPA monotherapy (65 percent) was most common. The remaining 42 percent continued to have seizures despite AED treatment.

In contrast, patients in the CAE evolving into JME group fared much worse, with only 3 of 35 individuals achieving complete seizure freedom [Martinez-Juarez et al., 2006]. Although GTCs were controlled with medications in about 66 percent of patients, absence seizures persisted in 63 percent. These data support the conclusion that JME does appear to be a lifelong disease, specifically in the subgroup of patients with CAE evolving into JME. However, there is still a small subgroup (9 percent) of patients in whom the disease may not be lifelong, and based on these data, it is uncertain whether some of the patients who have been seizure-free on monotherapy for many years would be able to come off AED treatment successfully.

The study by Camfied and Camfield seems to suggest that this indeed may be a possibility. In 24 JME patients with a follow-up period of 25.8 years, 4 individuals (17 percent) were free of all seizures and off AEDs. The mean duration of therapy was 6.5 years before discontinuation. Two additional patients attempted to come off AEDs; they relapsed but later were able to be successfully weaned off medication and remain seizure-free. Another three patients (12.5 percent) had myoclonus only [Camfield and Camfield, 2009]. These results suggest that some patients with JME can successfully discontinue medications. Unfortunately, there do not appear to be clinical or EEG variables that can predict which patients will achieve this result. The decision to discontinue AEDs should be individualized for every patient, and a detailed discussion, with specific attention to the frequency of lifestyle provoking factors, is necessary in order to arrive at a safe and well-informed decision.