If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Pathophysiology explains not only epilepsy phenotype of Dravet syndrome but also non-epileptic features, and this condition should therefore be considered a disease.
Other genes than SCN1A produce the Dravet syndrome phenotypes but clinical reports do not provide data regarding non epileptic features.
The therapeutic strategy for Dravet syndrome should be modified, questioning chronic use of benzodiazepines.
To improve full-blown phenotype of Dravet disease requires targeting NaV1.1 channels modulation, including SCN1A gene therapy.
Dravet syndrome combines clonic generalized, focal or unilateral seizures, beginning within the first year of life, often triggered by hyperthermia whatever its cause, including pertussis vaccination. Long-lasting febrile seizures are frequent in infancy and repeat status epilepticus (SE) has negative prognostic value. Massive myoclonus, rare absences, complex partial seizures and generalized spikes may appear several years later. Myoclonic status may occur in childhood, but acute encephalopathy with febrile SE followed by ischemic lesions and psychomotor impairment, the most severe condition, occurs mainly within the first five years of life. Generalized tonic–clonic and tonic seizures in sleep predominate in adulthood. Non epileptic manifestations appear with age, including intellectual disability, ataxia and crouching gait. Incidence of SUDEP is high, whatever the age. SCN1A haploinsufficiency producing NaV1.1 dysfunction mainly affects GABAergic neurons. In cortical interneurons it explains epilepsy, in cerebellum the ataxia, in basal ganglia and motor neurons the crouching gait, in hypothalamus the thermodysregulation and sleep troubles, and dysfunction in all these structures contributes to psychomotor delay. Valproate, stiripentol, topiramate and bromide are the basis of antiepileptic treatment, whereas inhibitors of sodium channel worsen the condition. Benzodiazepines seem to facilitate acute encephalopathy when given chronically, and they should be restricted to SE. Ketogenic diet is useful in both chronic and acute conditions. Only targeting SCN1A haploinsufficiency and NaV1.1 dysfunction could improve non epileptic manifestations of this condition that deserves being considered as a disease, not only as an epilepsy syndrome.
Among rare and severe epilepsies, Dravet syndrome (DS) occupies a particular place that is well expressed by the names it was given along its short history. Initially “Severe myoclonic epilepsy of infancy” (SMEI) was distinguished from Lennox–Gastaut syndrome (LGS) which was then considered as the most common severe epilepsy of childhood, and with which it shared pharmacoresistance, drop attacks, episodes of status epilepticus (SE) and intellectual disability [
]. Charlotte Dravet noticed that some children wrongly labeled as LGS exhibited massive myoclonus with photosensitivity and had had febrile seizures (FS) from the first year of life, thus pointing to a previously overlooked condition. Later, when it appeared that myoclonus was missing in over half the cases, the condition received the eponym of DS. Since then, its molecular basis has been identified [
]. It is now increasingly clear that epilepsy is the tip of the iceberg and that cognitive delay is another expression of the mutation, not merely the consequence of seizures. DS also produces age-dependent movement disorders, to which central and peripheral nervous systems contribute. Furthermore, there is not only neurological but also cardiac impact. Should it therefore still be considered as an epilepsy syndrome or more likely as a disease?
]. The first, “febrile stage” is specific enough to permit high diagnostic probability within the first year of life. First seizure occurs before 12 months of age in over 90% of cases, usually between 4 and 8 months, in a child previously considered as developing normally and without neurological history. It is clonic, generalized, focal or unilateral, prolonged in one third of the cases, triggered by mild fever or hot bath. This complicated FS often follows pertussis vaccination. 2.5% of patients with seizures following vaccination develop DS. Many patients previously labeled “post-vaccination encephalopathy” carry SCN1A mutation, and disease course is similar to that of patients with DS whose first seizures was not triggered by vaccination [
]. Prior to repeat seizures, mild hyperthermia, 37.5–38.5 °C, is often not noticed by the surrounding and muscle activity including that due to the seizure itself may have contributed to rising temperature. 80% of patients have hemi-clonic and/or focal seizures, lasting over 15 min. Rectal/oral benzodiazepine (BZ) often fails to stop the seizure that requires admission to hospital. Seizure frequency is moderate at onset, once a month or even less until the end of the first year of life. Seizures may repeatedly affect the same body area for several months, misleadingly pointing to possible focal epilepsy but interictal electroencephalogram (EEG) shows no spikes for several years and theta activity becomes prominent in the second year of life [
]. In very rare instances, massive myoclonic jerks are the first seizure type, before clonic seizures occur.
Between 1 and 5 years of age, in the “worsening stage”, motor seizures become more frequent but shorter, although their severity is still linked to mild hyperthermia. They affect randomly various parts of the body, a sequence indicating that the whole motor strip is involved. A triggering factor can often be identified: mostly hyperthermia, but also physical exercise, emotion, whereas light is less frequently a trigger than used to be when the condition was first described. Some patients nevertheless exhibit self-stimulation, watching the sun and closing the eyes. Fever remains a major cause of anxiety for parents who each time fear a new seizure.
Several months after onset, additional seizure types appear (Fig. 1), in less than half the cases, namely myoclonus and absences [
]. Myoclonic jerks can be massive or erratic. Absences and complex partial seizures without any motor component are rare and their recognition requires ictal EEG. Generalized 2–3 Hz spike-waves can be seen on EEG after 3–5 years. Photosensitivity occurs in a small proportion of cases, namely in patients with massive myoclonus.
By the end of the first decade, the “stabilization stage” begins in which seizures are less frequent, are brief, occur in sleep, and are tonic in a small proportion of cases. It is only at the end of the first decade that in a small proportion of cases EEG shows bursts of high frequency generalized spikes distinct from slow spike waves of the LGS [
]. Cognitive troubles are now in the front scene. In milder forms seizures become rare, down to 1/year, but they rarely disappear completely. These cases exhibit better cognition with valuable speech.
Although conventional MRI identifies no abnormality in neocortex, basal ganglia and white matter, morphometry in the second decade showed global volume reduction of gray and white matters in brainstem, cerebellum, corpus callosum, corticospinal tracts and association fibers. Hippocampal atrophy following long lasting unilateral FS is a rare finding.
Various types of Status Epilepticus may occur: (1) in infancy, prolonged clonic FS, often unilateral, commonly last over 30 min; (2) between 4 and 8 years of age, obtundation (non-convulsive) status with confusion combined with erratic myoclonus of the extremities and around the mouth lasting several hours is associated with pseudo-rhythmic slow wave activity encroached with irregular spikes. (3) In addition to these transient events, severe febrile SE lasting several hours or days, usually quoted as “acute encephalopathy” remains an etiological challenge. It occurs with non-specific infection, usually within the first five years of age, several months or years after onset of epilepsy in patients chronically treated with BZ combined with valproate (VPA) [
]. To manage the SE, all patients received high doses of BZ and about half got barbiturates. Cortical, basal ganglia and white matter lesions, consistent with acute ischemia, sustain severe psychomotor sequelae with spastic tetraparesis and cognitive impairment.
At onset of first seizures, neither parents nor physician notice any delay, but already proper tests disclose visual troubles, namely acuity and fixation shift disorder. Along the second year of life psychomotor delay appears insidiously. Walking is delayed in 60% of cases, speech develops even slower, and attention deficit frequently appears with hyperkinesia. One quarter of patients have autistic behavior with poor eye contact and stereotypes. However, there is no clear loss of abilities, in contrast with West or Landau–Kleffner syndromes [
Walking is increasingly ataxic, and children tend to fall easily, without any jerks. Paroxysmal choreoathetosis may occur, with or without administration of phenytoin. In the second half of the first decade, it becomes apparent that instability is not the only reason for slowness, and that gait is increasingly crouching. Although pyramidal signs may develop early in the course of the disease, tendon reflexes are eventually difficult to elicit but there is no amyotrophy, and electromyography may show signs of both anterior horn and peripheral motor nerve involvement [
]. SE used to be a significant cause of death (36%), but has decreased with improved diagnosis and treatment, and SUDEP is presently the main cause (56%) with two peaks, at 1–3 and over 18 years, without evidence of worsening of epilepsy [
Factors that positively impact transfer to the adult health-care system are the quality of transition preparation, a longer duration of follow-up by the same child neurologist, a stable medical condition before transfer and transfer after the age of 18 [
]. Fever sensitivity persists in half the patients. Atypical absences, myoclonic or complex partial seizures tend to disappear. Hyperkinesia is replaced by slowness with crouching gait. Orthopedic troubles with kyphoscoliosis are frequent. Risk of SUDEP remains high. Most adults have severe intellectual disability. Autism spectrum disorder persists. Hallucinations and delusions may appear [
Better cognitive condition in adulthood is associated with having had less than three episodes of SE in infancy, whereas EEG abnormalities in the first year of life, motor disorders, tonic seizures and early appearance of myoclonus and absences are the main negative prognostic factors.
There is high incidence of FS and epilepsy in the family. The molecular genetics story started when DS was shown to be the most severe expression of the generalized epilepsy with febrile seizures plus (GEFS+) spectrum described in large Australian families that revealed dominantly inherited mutations in SCN1A [
]. De novo mutations in this gene were later confirmed to affect over 75–80% of patients with DS. More than 1250 different SCN1A mutations were identified, three quarters involving the paternal chromosome. There seems to be phenotype–genotype correlation [
]: approximately 52% are truncating mutations correlating with de novo cases of classical DS in 94% patients; missense mutations in the pore-forming part constitute about 27% and correspond to classical DS in 75%; and 12% are missense mutations in the voltage sensor part correlating with a clinical picture ranging from FS+ to DS.
Other mechanisms explain phenotypic variability. Mosaic mutation is associated with a milder phenotype than full heterozygous mutation but can be associated with more severe epilepsy phenotype in the offspring. Particularly severe DS (with small corpus callosum, cerebellar atrophy and delayed myelination), FS+ and FS+ followed by partial seizures were reported in consanguineous families related to SCN1A missense mutations inherited as homozygous traits. The function of the α-subunit of sodium channels may or may not be rescued by the β-subunit, explaining difference in severity, from DS to GEFS+. Association with CACNA1A gene polymorphic variants is correlated with earlier seizure onset, more frequent prolonged seizures in the first year of life and frequent absence seizures.
Microdeletion usually originates de novo from paternal chromosome. De novo balanced translocations were also reported. 10% of DS patients have deletion, duplication or amplification identified with multiplex ligation-dependent probe amplification. SCN1A deletion combined with SCN2A and SCN3A deletion produces a more severe phenotype than DS with earlier onset and progressive microcephaly, consistent with migrating partial seizures in infancy. On the other hand, association of SCN1A and SCN9A deletions produces milder DS phenotype.
In 10–20% of cases, the genetic cause remains unknown and additional genes are likely to be implicated, involving sodium channel subunits, but eventually the GABA neurotransmission. Less than 1% of DS patients have homozygous mutation in SCN1B, and very few have GABRG2 or SCN2A mutations. However, it is not clear from the published clinical descriptions whether patients with mutations in other genes than SCN1A develop only Dravet syndrome or the full-blown phenotype of Dravet disease including non-epileptic manifestations, particularly in the peripheral nervous system.
Mitochondriopathy has been suggested because of respiratory chain dysfunction findings, which raises the question of its mechanism and the risk of administration of VPA. However, the few reported heterozygous POLγ variants are unlikely to be pathogenic since these were recessive traits [
If there are three ILAE diagnostic features (normal development before seizure onset, exacerbation with hyperthermia, and the appearance of ataxia, pyramidal signs, or massive myoclonus), there is a 100% chance of having a positive test for SCN1A. However, for one given case, even truncating mutation does not predict the clinical severity. Given the price of the investigation, it should be performed if parents need genetic counselling for future pregnancy since although de novo mutation is the usual case, familial recurrence represents 7% of cases, and may affect the descent of non-symptomatic parents. Somatic and germline mosaicism may be the cause, concerning up to 10% of de novo mutations in DS children. Cases with moderate mosaicism (12–25%) with milder phenotype, namely focal seizures following FS carry a risk of more severe familial recurrence. Therefore, this small risk of severe recurrence should be balanced with the risk of preimplantation diagnosis.
4. Differential diagnosis
Hyperthermia-sensitive polymorphous clonic seizures in the first year of life with normal initial development and EEG offer little alternatives, except for FS+ which is not associated with cognitive delay, and mutation in PCDH19 in which there is no motor neuropathy [
]. In case of late diagnosis, the confusion may appear with malignant migrating partial seizures, severe forms of myoclono-atonic epilepsy or Lennox–Gastaut syndrome, although lack of migrating seizures, astatic and tonic seizures, as well as distinct EEG aspect, permit distinguishing the Dravet phenotype. Among emerging genes, epilepsy due to de novo GABRA1 and GABRG2 mutations [
] seems to give the most similar phenotype. De novo CHD2 mutations were shown to have variable expression, overlapping between myoclono-atonic epilepsy and Dravet-like phenotype. However, seizures onset is late and preceded in some patients by developmental delay, unlike DS [
]. This phenotype is rather similar to SCN2A encephalopathy, although the latter is more heterogeneous and begins earlier, usually in the neonatal period. Differential diagnosis is shown in Fig. 2.
SCN1A encodes the α-subunit of the main brain type-I voltage-gated sodium channel NaV1.1 which is specifically localized in neuronal cell body, namely the initial segment of GABAergic interneurons. NaV1.1, progressively replaces NaV1.3, the embryonic sodium channel subunit in the second week of postnatal life in rodents, which grossly corresponds to the first year of human life [
]. Then, at least 75% of the sodium current in GABAergic interneurons is conducted through NaV1.1 channels. Severe DS phenotype is usually caused by complete loss of SCN1A function resulting in reduction of NaV1.1-mediated sodium currents in GABAergic neurons. Rare cases with gain-of-function had similar phenotype.
GABAergic neurons are largely distributed in the central nervous system, including cortex, hippocampus, basal ganglia, hypothalamus, cerebellum and spinal cord, and they are mostly inhibitory after the neonatal period. An exception concerns the suprachiasmatic nucleus in which for a subpopulation of neurons, GABA may be either inhibitory or excitatory, depending of the enlightenment. Retino-suprachiasmatic fibers that belong to NMDA ganglion neurons of the retina activate Ca2+ channels in GABAergic neurons of the suprachiasmatic nucleus in light-time, transiently activating gene expression of the transmembrane NKCC1 co-transporter that accumulates chloride in neurons and is responsible of the excitatory character of these GABAergic neurons [
]. This generates day–night alternation within a suprachiasmatic nucleus GABAergic neurons’ subpopulation that ensures synchronization of the circadian rhythms with the environment.
Dysfunction of the suprachiasmatic nucleus could explain both sleep impairment and the slow EEG background rhythm from the second year of life. Scn1a heterozygous mice have a severe disruption of their circadian rhythms with longer circadian period and lack of light-induced phase shift, which is partially rescued by pharmacological enhancement of GABAergic transmission. Imbalance between excitatory neurons in the thalamic ventrobasal nucleus and inhibitory GABAergic neurons in the thalamic reticular nucleus could contribute to sleep troubles in DS patients including delayed sleep onset and difficulty maintaining it [
Scn1a heterozygous mice exhibit reduced sodium currents in GABAergic hippocampal and cortical interneurons, but not in pyramidal neurons, resulting in hyperexcitability and specific epilepsy phenotype, namely major sensitivity to triggering factors. It is now well established that mild impairment of NaV1.1 channels predisposes to FS, intermediate impairment leads to GEFS+, and severe or complete loss of function leads to the intractable seizures and non-epileptic manifestations of DS. GABA is an important neurotransmitter for thermoregulation in the preoptic area and anterior hypothalamus (PO/AH). It contributes to disinhibition of heat production or inhibition of heat loss under cold ambient temperature. Hyperthermia-induced seizure-susceptible (Hiss) rats carrying a missense Scn1a mutation (N1417H) showed significantly lower threshold than control animals in inducing epileptiform discharges in response to local stimulation of the hippocampus. Moreover, fever per se may worsen GABA channels dysfunction. It was shown that elevated temperature reversibly reduced GABA receptors by decreasing trafficking or the accelerated endocytosis [
This particularly severe SE appears in the context of high sensitivity to fever due to age, in children who have experienced repeat episodes of SE and were chronically treated with BZ that reduces the GABA response to benzodiazepines [
]. In the context of abnormal GABAergic transmission that characterizes Dravet syndrome, the consequences may be particularly devastating. Reported mitochondrial dysfunction associated with increased energy needs of defective GABAergic neurons could also contribute to the energy failure.
There is progressive slowdown of cognitive developmental to which several factors may contribute. Cognitive acquisition requires independence of cortical columns in day-time and proper slow sleep organization [
]. In DS, slow rhythmic activity awake impairs cognitive acquisition. Fast-spiking, parvalbumine positive GABAergic neurons which are critical to maintain precise temporal dynamics in cortical and subcortical networks, including the patterning and generation of oscillations such as theta and gamma rhythms in the hippocampus and cortex, are affected in DS [
]. Furthermore, programs developed by the cortex are stored in the basal ganglia, but this cannot take place in DS that comprises abnormal basal ganglia function. Attention deficit may be a consequence of reticular formation, frontal lobe and basal ganglia dysfunction. Scn1a+/− mice exhibit hyperactivity, stereotyped behavior, social interaction deficits and impaired context-dependent spatial memory.
There is no evident correlation between severity of SCN1A mutations and SUDEP. Scn1a haploinsufficient mice have QT prolongation, ventricular ectopic foci, idioventricular rhythms, beat-to-beat variability, ventricular fibrillation, and focal bradycardia. Although, AV blocks episodes are frequent, these abnormalities are not observed in mice with NaV1.1 channel deletion restricted to the heart, suggesting that these defects arise from outside the heart. Furthermore, selective inhibition of NaV1.1 in the heart with low concentrations of tetrodotoxin causes changes in intrinsic cardiac function that are not observed in DS mice, indicating that they are occluded by the more severe effects of nervous system deletion of NaV1.1. SUDEP in Dravet mice results from ictal and postictal dysregulation of the heart via hyperactivity of the parasympathetic nervous system, which leads to lethal bradycardia and electrical dysfunction of the ventricle. Heart rate variation is depressed in DS compared to other types of epilepsy, QT and P wave dispersions indicate major autonomic dysfunction with increased adrenergic tone [
]. Nevertheless, NaV1.1 is a major component of late sodium current in ventricular myocytes, particularly from failing heart. Therefore, SUDEP is more likely to be triggered by neurological than cardiac cause, but failing myocytes are likely to contribute.
5.5 Movement disorders
NaV1.1 channels are expressed in cell bodies of cerebellar GABAergic Purkinje neurons of ataxic mice that are unable to provide adequate inhibitory control of deep cerebellar nuclei neurons. This can explain ataxia observed from the second year of life in DS. Parkinsonism likely results from lenticular dysfunction since neurons in this structure are mainly GABAergic. NaV1.1 channels are expressed in nodes of Ranvier, including 80% of motor neurons [
]. This explains that patients with DS and SCN1A mutation, have distal motor deficit with gait disturbance and distal orthopaedic deformities although EMG is neurogenic with preserved nerve conduction. This is followed by proximal motor deficit leading to crouching. Interestingly, in ventral and dorsal roots no expression of NaV1.1 could be detected, explaining the absence of profound proprioceptive or sensitive troubles in DS patients.
Although seizure freedom and normal cognitive development is an obvious future aim, presently a reasonable objective is to prevent long lasting seizures, reduce seizure frequency to two a month or less, and improve cognitive and motor disorders. Following a first non-febrile or complicated FS before the diagnosis of DS is suspected, the administration of VPA is commonly agreed. It is important to avoid the use of sodium channel inhibitors: carbamazepine, oxcarbazepine, lamotrigine and phenytoin that may worsen the condition, phenytoin eventually triggering choreoathetosis. Vigabatrin can also worsen the condition, phenobarbital and rufinamide have poor effect. Repeat seizures require the addition of a second compound, either topiramate (TPM, 3–5 mg/kg/day), or stiripentol (STP, 50–80 mg/kg/day) with a reduction of the VPA dose to 20 mg/kg/day because STP is a cytochrome P450 inhibitor. Since STP was registered combined with VPA and clobazam when it was believed to act pharmacokinetically by increasing the effect of clobazam, STP was shown to have GABAergic effect per se [
], and to be efficient whatever the comedication (VPA/CLB). On the other hand, we showed that there are reasons to fear that chronic BZ administration could facilitate acute encephalopathy and insensitivity to BZ administered for SE. BZs also increase hypotonia, a risk in case of neuropathy. Therefore chronic use of BZ in DS may not be the best option, and should at least be questioned. The next step, in case seizures remain frequent or prolonged, is to replace TPM by STP or the reverse, then if necessary to replace them by bromide (30–50 mg/kg/day). The ketogenic diet can be an option for periods of frequent seizures (Fig. 3).
In case of a seizure lasting over 5 min buccal or rectal BZ is administered. In case of SE, the intravenous administration of a BZ remains the first step, as usual for the treatment of SE. Then, if seizures persist ketogenic diet should be preferred to barbiturates or phenytoin that may be dangerous [
Future tracks for treatment research need to take in account non epileptic troubles largely related to the gene dysfunction, independently of epilepsy. Several transgenic animal models, namely zebrafish, drosophila and mice, permit screening new compounds. Induced pluripotent stem cells (iPSCs) should permit drug screening and personalizing medication according to the gene anomaly carried by the patient. On the other hand, gene therapy would permit restoration of proper function of GABAergic neurons and therefore improve not only epilepsy but also non epileptic disorders, provided it is performed early enough in the course of the disease.
SCN1A positive DS is more than an epilepsy syndrome, it should be considered as a disease. Molecular genetics continues to identify new genes responsible of DS and other conditions with similar phenotypes. Neurophysiological studies on animal models played a crucial role explaining epileptic and non-epileptic manifestations, GABAergic neurons’ dysfunction underlying most components of this complex phenotype. Emerging data about BZ adverse effects should modify treatment algorithm restricting their administration to SE, in order to prevent acute encephalopathy. However antiepileptic treatment is no longer sufficient to improve outcome. SCN1A expression/NaV1.1 modulation should now become the therapeutic target.
Conflict of interest statement
The authors declare no conflict of interest.
Dravet syndrome (Severe myoclonic epilepsy in infancy).
in: Dulac O. Lassonde M. Sarnat H.B. Handbook of clinical neurology, vol. 111 (3rd series). Pediatric neurology part I. 2013: 627-633