Dravet syndrome: New potential genetic modifiers, imaging abnormalities, and ictal findings


  • Eija Gaily,

    Corresponding author
    1. Department of Pediatric Neurology, Helsinki University Central Hospital, Helsinki, Finland
    • Address correspondence to Eija Gaily, Department of Pediatric Neurology, Helsinki University Central Hospital, PO Box 280, Helsinki 00029 HUS, Finland. E-mail: eija.gaily@hus.fi

    Search for more papers by this author
  • Anna-Kaisa Anttonen,

    1. Folkhälsan Institute of Genetics, Helsinki, Finland
    2. Department of Medical Genetics, Haartman Institute and Research Program's Unit, Molecular Neurology and Neuroscience Center, University of Helsinki, Helsinki, Finland
    3. Department of Clinical Genetics, Helsinki University Central Hospital, Helsinki, Finland
    Search for more papers by this author
  • Leena Valanne,

    1. Helsinki Medical Imaging Center, Helsinki University Central Hospital, Helsinki, Finland
    Search for more papers by this author
  • Elina Liukkonen,

    1. Department of Pediatric Neurology, Helsinki University Central Hospital, Helsinki, Finland
    Search for more papers by this author
  • Ann-Liz Träskelin,

    1. Folkhälsan Institute of Genetics, Helsinki, Finland
    Search for more papers by this author
  • Anne Polvi,

    1. Folkhälsan Institute of Genetics, Helsinki, Finland
    2. Department of Medical Genetics, Haartman Institute and Research Program's Unit, Molecular Neurology and Neuroscience Center, University of Helsinki, Helsinki, Finland
    Search for more papers by this author
  • Markus Lommi,

    1. Department of Pediatric Neurology, Helsinki University Central Hospital, Helsinki, Finland
    2. Folkhälsan Institute of Genetics, Helsinki, Finland
    Search for more papers by this author
  • Mikko Muona,

    1. Folkhälsan Institute of Genetics, Helsinki, Finland
    2. Institute for Molecular Medicine Finland (FIMM), University of Helsinki, Helsinki, Finland
    3. National Institute for Health and Welfare, Public Health Genomics Unit, Helsinki, Finland
    Search for more papers by this author
  • Kai Eriksson,

    1. Department of Pediatrics, Tampere University Hospital, Tampere, Finland
    Search for more papers by this author
  • Anna-Elina Lehesjoki

    1. Folkhälsan Institute of Genetics, Helsinki, Finland
    2. Department of Medical Genetics, Haartman Institute and Research Program's Unit, Molecular Neurology and Neuroscience Center, University of Helsinki, Helsinki, Finland
    Search for more papers by this author



Dravet syndrome is an autosomal dominant epileptic encephalopathy of childhood, which is caused mainly by SCN1A and PCHD19 mutations. Although Dravet syndrome is well recognized, the causes of acute encephalopathy are still elusive, and reported data on ictal electroencephalography (EEG) and structural brain abnormalities are scarce.


We studied 30 children who fulfilled the clinical criteria for Dravet syndrome. All patients were screened for SCN1A mutations and 25 for POLG mutations with bidirectional sequencing. Clinical data, including etiologic studies done as part of the clinical workup, were collected from hospital charts. Ictal video-EEG recordings and magnetic resonance (MR) images were reanalyzed by the authors.

Key Findings

SCN1A mutations were found in 25 patients (83%). Two SCN1A mutation–negative patients had chromosomal translocations involving chromosomes 9 and X, and one had a mutation in PCDH19. Prolonged seizures were associated with acute encephalopathy in three SCN1A mutation–positive patients. One showed evidence of a significant hypoxic–ischemic event during status epilepticus. The other two demonstrated new persistent neurologic deficits postictally; they both carried heterozygous POLG variants (p.Trp748Ser or p.Gly517Val). Hippocampal sclerosis or loss of gray–white matter definition in the temporal lobe was observed in 7 of 18 patients who had MRI after age 3 years (39%). Motor seizures were recorded on video-EEG for 15 patients, of whom 12 were younger than 6 years at recording; 11 patients (73%) showed posterior onsets.


Our data imply that a heterozygous X;9 translocation and rare POLG variants may modify the clinical features of Dravet syndrome. The latter may increase susceptibility for acute encephalopathy. Temporal lobe abnormalities are common in patients imaged after 3 years of age. Focal seizures seem to localize predominantly in the posterior regions in young children with Dravet syndrome.

Dravet syndrome (DS), previously also known as severe myoclonic epilepsy of infancy (SMEI), is an epilepsy syndrome with onset in the first year of life; it is drug-resistant, and often characterized by prolonged tonic–clonic seizures typically provoked by fever and infections, and cognitive decline (Dravet et al., 2005; Guerrini & Oguni, 2011). Myoclonias and frequent episodes of status epilepticus are common but are not present in all patients.

Although recovery from status epilepticus occurs in most patients with DS within 24 h, several case reports have recently described acute encephalopathy with severe neurologic sequelae (Sakakibara et al., 2009; Chipaux et al., 2010; Nishri et al., 2010; Takayanagi et al., 2010; Tang et al., 2011; Okumura et al., 2012). Some of these patients have shown muscle biopsy findings suggesting mitochondrial disorders (Castro-Gago et al., 1995; Nishri et al., 2010), but no mutations in POLG or other genes affecting mitochondrial function have been found. The only previously described patient with SCN1A and POLG mutations did not present with acute encephalopathy (Bolszak et al., 2009).

Nearly 80% of patients with DS harbor a mutation in the SCN1A gene encoding the neuronal voltage-gated sodium channel subunit Nav.1.1 (Marini et al., 2011). Some patients who remain negative in sequence analysis have deletions or chromosomal rearrangements involving SCN1A. SCN1A mutations arise usually de novo, but 5–10% are inherited from a healthy or mildly affected parent. Mutations in other genes have also rarely been associated with DS, especially PCDH19 mutations, carried by a small percentage of female patients (Depienne et al., 2009a; Marini et al., 2011).

There are two previous studies reporting magnetic resonance imaging (MRI) abnormalities after reevaluation by neuroradiologists in patients with DS at age 4 years or older; these have provided contradicting data. Siegler et al. (2005) reported hippocampal sclerosis in 10 (71%) of 14 patients, but Striano et al. (2007) observed hippocampal sclerosis in only one of 58 patients, whereas other abnormalities such as atrophy were found in 12 patients (21%). Hippocampal sclerosis was also rare in two recent outcome studies (Akiyama et al., 2010; Ragona et al., 2010) and in patients who were imaged in adulthood (Catarino et al., 2011).

Major motor seizures in DS are typically described as unilateral or bilateral tonic–clonic, psychomotor (complex partial), and focal clonic (Ohki et al., 1997; Oguni et al., 2001; Ragona et al., 2010). In 25 seizures recorded in eight patients (age at recording not given, but most were probably children), ictal onsets occurred in the frontal area in six seizures, temporal in nine, and occipital in six (Ohki et al., 1997). In adult patients, focal electroencephalography (EEG) seizure onsets have been localized mainly to the frontal regions (Akiyama et al., 2010; Catarino et al., 2011).

The aim of this retrospective study was to determine the molecular genetic background of Finnish patients with DS and to further delineate the clinical spectrum by investigating the occurrence of acute encephalopathy, structural brain abnormalities, and focal seizures recorded on video-EEG.


Clinical information of all 114 patients for whom we had performed SCN1A mutation analysis in our laboratory was reviewed by one of the authors (EG) to define epilepsy syndrome and etiology. Patients with seizure onset in the first year, intractable epileptic seizures triggered by infections and increased temperature, normal development in the first year, and no evidence of structural-metabolic etiology at seizure onset were diagnosed as having DS (Dravet et al., 2005). Based on these criteria, 31 patients with DS were identified. One patient was excluded from further analysis because his case report has been published previously (Bolszak et al., 2009). The remaining 30 patients (born in 1988–2008) were included in the present study and the results of their genetic studies were reviewed. All seizures recorded on video-EEG and all MR images performed at our hospital (19 patients) or elsewhere (11 patients) were reanalyzed (EG and LV). The authors had personally seen 25 (83%) of the patients.

For the 30 included patients, the mean age at seizure onset was 6.2 months (range, 2.4–12.0 months), and at latest follow-up was 11.1 years (range, 4.3–22.1 years). Interictal EEG in the first year was normal in 28 patients. One patient (S056) had posterior spikes between seizures during a seizure cluster but normal EEG recordings between clusters. Another patient (S066) had sparse multifocal spikes during wakefulness 1 month after seizure onset at 6 months. Five patients (S017, S079, S117, S121, and S183 in Table S1) were never observed to have myoclonic seizures and would earlier have been classified as borderline SMEI (Guerrini & Oguni, 2011). The majority of patients had mild to severe mental deficiency at latest follow-up; only one patient older than 10 years (S025) had preserved normal intelligence.

For SCN1A analysis, total genomic DNA was isolated from blood cells using PUREGENE DNA Purification Kit (Gentra Systems Inc., Minneapolis, MN, U.S.A.). The 26 coding exons of SCN1A (AB093548) were amplified and sequenced with intronic primers (sequences available on request) using BigDye Terminator v3.1 Cycle Sequencing Kit and an ABI 3730 DNA Analyzer (Perkin Elmer, Foster City, CA, U.S.A.). The available parental samples and at least 94 unrelated Finnish controls were analyzed for the identified mutations by sequencing. The clinical relevance of the variants was further assessed from databases: the SCN1A infobase (http://www.scn1a.info/), the SCN1A variant database (http://www.molgen.vib-ua.be/SCN1AMutations/), the NHLBI Exome Sequencing Project (ESP) Exome Variant Server (http://evs.gs.washington.edu/EVS/) and the dbSNP Short Genetic Variations database (http://www.ncbi.nlm.nih.gov/SNP/).

To detect copy number changes in SCN1A, multiplex ligation-dependent probe amplification (MLPA) analysis was performed in those mutation-negative patients whose samples were available or as part of their routine work-up in diagnostic laboratory using SALSA MLPA P137 SCN1A kit (MRC-Holland, Amsterdam, The Netherlands) according to manufacturer's instructions and analyzed as described by Mulley et al. (2006).

The POLG gene mutation analysis had been done based on clinical suspicion of a mitochondrial disorder in four patients with DS. Because one of them turned out to have a heterozygous mutation (see 'Results'), we proceeded to look for POLG mutations in all patients whose DNA was still available. We amplified and sequenced as described above the 22 coding exons and exon–intron boundaries of POLG (X98093) from an additional 21 patients. The phase I release of the 1000 genomes project (1000 Genomes Project Consortium et al., 2012) containing genome-wide variant frequency information for 93 Finnish healthy individuals (http://browser.1000genomes.org, date accessed 15 March 2013) and a local database of aggregate variant data from 500 Finnish exomes sequenced from the FINRISK population cohort were used to estimate POLG variant frequencies in the Finnish population.

Fluorescence in situ hybridization (FISH) with telomeric DNA probes was performed on chromosome preparations from peripheral blood cells in a routine diagnostic laboratory. Either Chromoprobe Multiprobe-T kit (Cytocell, Cambridge, United Kingdom) or Vysis ToTelVysion kit (Abbott Laboratories, Abbott Park, IL, U.S.A.) was used according to the supplier's instructions. The kits carry subtelomere specific probes for both the p-arm and the q-arm of each chromosome, except for the acrocentric chromosomes, where the probes are only for the q-arm sub-telomeres.

The ethics committee of the Helsinki University Central Hospital approved this study. All participating families provided written informed consent for the study.


Genetic data

A mutation in SCN1A was found in 25 (83%) of 30 patients (Table S1). The mutation arose de novo in 22 patients and was inherited from a parent in 2 patients (S131, K903). In both cases, the parent carrying the mutation was affected by both febrile and afebrile seizures. Inheritance was unknown for one patient (S032) for whom the paternal sample was not available.

In one female patient (S056), a de novo missense mutation in PCDH19, c.695A>G (p.Asn232Ser) was detected in a clinical diagnostic laboratory. It was predicted by PolyPhen2 as probably damaging, and it was absent in databases. The seizures of the patient with PCDH19 mutation started at 2 months, 24 h after she had received the pertussis vaccination. Her seizures were strongly provoked by fever and infection and occurred in clusters. In infancy, they typically started with hypomotor–apneic semiology often followed by tonic–clonic activity and posterior discharges on ictal EEG. Later, she also presented clusters of tonic–clonic and hypermotor seizures with an aura of abdominal or lower extremity pain, and myoclonias.

Two patients (S066, S160) without SCN1A abnormality had a chromosomal rearrangement involving chromosomes 9 and X. Both had three signals of the subtelometric marker 9ptel: two on both chromosomes 9 and one in the middle of the long arm on chromosome X (Xq). Comparative genomic hybridization was normal in both patients. Clinical onset was with tonic–clonic seizures provoked by fever at 6 (S066) and 8 (S160) months of age. Seizures recurred monthly despite drug treatment, and the patients also showed afebrile seizures, ataxia, myoclonias, and moderate (S160) or severe (S066) mental deficiency. Patient S066 had several episodes of convulsive status but became seizure-free at 11 years on valproate and topiramate. Patient S160 continued to have seizures in the second decade.

The mothers of both patients carried the chromosomal abnormality seen in the child. The mother of patient S066 was clinically unaffected. The mother of S160 had febrile seizures, epilepsy, and late onset ataxia, but a carrier sister was healthy, whereas a noncarrier sister had febrile seizures and developmental delay.

The remaining two patients without SCN1A abnormality had seizure onsets at 4–9 months of age, provocation by fever, and normal early development, EEG, and MRI. Both patients continued to be drug resistant, with febrile and afebrile bilateral and focal tonic–clonic and psychomotor seizures, myoclonias, and later mental deficiency.

The POLG gene was analyzed in 25 patients (Table S1). One patient (S200) carried two variants in cis (c.2243G>C; c.3428A>G) (p.Trp748Ser; p.Glu1143Gly), of which the first is a recessive mutation and the latter a polymorphism (Hakonen et al., 2005). Another patient (S077) carried a rare variant c.1550G>T (p.Gly517Val). The carrier frequencies of these variants in 500 Finnish exome controls were 0.40% (p.Trp748Ser) and 1.20% (p.Gly517Val). In addition, neither of the two variants is present in the 93 Finnish samples of the phase I release of the 1000 genomes project.

Status epilepticus and acute encephalopathy

One or more episodes of convulsive status epilepticus (SE) lasting for more than 30 min had occurred in 25 patients (83%). The median (mean) age at the first episode was 9.1 (11.6) (range 4.0–49.6) months. In most cases, recovery occurred within 24 h to the prestatus clinical condition. Three patients (S176, S200, S077) had significant neurologic sequelae after SE (summarized in Table 1).

Table 1. Patients with acute encephalopathy and neurologic sequelae after status epilepticus
Patient SCN1A POLG SE description (age, months)AED before SE onset, mg/kg/dayTreatment of SENeurologic sequelaeMRI in the acute phaseOutcome at 12 months post SE
  1. SE, status epilepticus; LEV levetiracetam; VPA, valproate; PB, phenobarbitone; CLB, clobazam; STP, stiripentol; DZP, diazepam; MDZ, midazolam; phosPHT, phosphenytoin; THP, thiopental; LZP, lorazepam,

  2. a

    0.7 mg/kg prior to hospital admission.

S200, malec.3452C>G, p.Ser1151* de novoc.2243G>C ,p.Trp748Ser, heterozygous, paternal inheritanceTonic–clonic, duration 12 h (26), in coma for 5 daysLEV 50DZP rectal and iv, MDZ buccal, phosPHT iv, THP infusion for 10 hSevere regression of all skillsFocal hyperintensity in the left frontal lobe at day 12 (Fig. 1B–C)Almost complete recovery within 2 months, mild residual clumsiness of the right hand
S077, femalec.3655_3685delinsATA, p.Trp1219Ilefs*42 de novoc.1550G>T, p.Gly517Val heterozygous, parents not testedLeft-sided tonic–clonic, duration over 30 min (6.5)None (first seizure)DZP rectalNew left-sided hemiparesisNot doneHemiparesis resolved within a year
S176, femalec.4300T>C, p.Trp1434Arg, de novoNo mutationsTonic–clonic, duration 1.5 h (19), in coma for 9 days

VPA 36

PB 2.7

CLB 0.9

STP 45

MDZa, LZP iv, THP bolus, propofol infusion for 10 hSevere regression of all skillsEdema of gray and white matter, thalami and basal ganglia at day 7No significant recovery, severe motor and mental disability

Seizure onset of patient S200 was at 8 months with bilateral tonic–clonic semiology, followed by frequent febrile and afebrile right-sided or bilateral tonic–clonic seizures and myoclonias. MRI (Fig. 1A) and EEG were normal. His development was normal until the first SE occurred, with fever at the age of 2.2 years. The first symptoms suggested partial onset with eyes and head turning to the left, followed by tonicclonic activity. EEG at 16 h after onset showed no ictal activity, but short seizures recurred on the second day. On day 11 after onset EEG showed left-sided background attenuation. MRI on day 12 showed hyperintensity of the left frontal lobe in diffusion weighted images, suggesting cytotoxic edema (Fig. 1B–C). At 3 weeks after status, the patient was able to stand supported for a few minutes and had severe visual dysfunction. Gradual recovery of neurologic function occurred within 5–6 weeks after onset. Follow-up 3T MRI at age 2.8 years showed nearly complete resolution of the left frontal abnormalities (Fig. 1D). At 4 years—on levetiracetam, stiripentol, clobazam, and modified Atkins diet—the patient continued to have 1–2 motor seizures per week. Speech and social skills were delayed.

Figure 1.

MRI (1.5 T) of patient S200. (A) (FLAIR): Normal finding at age 1.3 years. (B) (FLAIR) and (C) (diffusion-weighted image): Hypersignal (arrow) in the left frontal lobe anterior to the motor cortex 12 days after onset of acute encephalopathy, age 2.3 years. (D) (FLAIR): Almost complete resolution of the abnormality, 5 months after acute encephalopathy, age 2.7 years.

Patient S077 had her first seizure at the age of 6.5 months. The exact duration of the seizure was not noted in the hospital files, but it had continued until the ambulance arrived. Left-sided hemiparesis, affecting both extremities and lower face, was observed after the seizure, and persisted for more than 6 months. Later MRI studies at 11 months and 4.8 years of age were normal. Ataxia and myoclonias occurred intermittently. Occasional elevations of the alanine aminotransferase levels up to 182 unit/L (normal <40) were observed during valproate and ethosuximide treatment. At 15 years of age, she had severe mental deficiency and tonic–clonic seizures up to one to two times per week.

Patient S176 had her first seizure at the age of 2.5 months. Convulsive SE episodes with fever recurred approximately once a month, but neurodevelopment was normal until the complicated SE with fever and bilateral tonic–clonic semiology occurring at 19 months. On arrival at the emergency room at 70 min after SE onset, she continued seizing, was acidotic, and had a high temperature (40°C). EEG recorded at 6 h after seizure onset showed no ictal activity. Alanine aminotransferase levels in the plasma increased up to 3,500 unit/L (normal <40), aspartate aminotransferase to 13,141 unit/L (normal <50), and blood ammonium to 97 micromol/L (normal <50) on the second day after onset. There was no permanent liver damage. At 5 years, she had severe mental and motor disability. Tonic–clonic seizures occur one to two times per month.


All patients had at least one high field MRI. At our own hospital we use an epilepsy-dedicated imaging protocol. This consists of axial T2-weighted 3-mm and fluid-attenuated inversion recovery (FLAIR) 4-mm slices, T2-weighted 3-mm coronal slices through the head and 2-mm images through the temporal lobes angled perpendicular to the hippocampi, as well as T1-weighted three-dimensional (3D) images with 1-mm isotropic voxel in sagittal plane. The first MRI available for reanalysis was performed at a median age of 1.1 (range 0.2–13.0) years and was normal in 26 patients (87%). Abnormalities were found in four children: one cortical atrophy at 3.1 years and three temporal lobe abnormalities. Nineteen patients with initially normal MRI had at least one subsequent MRI available for reanalysis; five of them showed new temporal lobe abnormalities, two had developed cortical or cerebellar atrophy at 2.2–2.4 years, and another two had acute changes associated with SE (shown in Table 1 and Fig. 1B–C). The frequency of temporal lobe abnormalities was 7 (39%) of 18 in patients imaged after the age of 3 years. Either hippocampal sclerosis or loss of gray–white matter definition in the anterior temporal lobe (Mitchell et al., 2003) was observed. An example of the latter is shown in Fig. 2. The temporal lobe findings are summarized in Table 2.

Table 2. Temporal lobe abnormalities detected on MRI in patients with Dravet syndrome
PatientMRI before 3 years of age (age at imaging in years)MRI at 3 or more years of age (age at imaging in years)
  1. a

    Imaged at another hospital using a similar protocol.

S056Hypersignal, right hippocampus (0.4)Normal (8.4)
S127Normal (1.9)Hippocampal sclerosis, bilateral, and cortical atrophy (7.6)
S128Normal (1.0)Hippocampal sclerosis, right (3.0)
S002NoneLoss of gray–white matter definition, bilateral (3.7 and 5.1)a
S033NoneLoss of gray–white matter definition, bilateral (4.1)
S170Normal (1.8)Loss of gray–white matter definition, left (5.9)
S030Normal (0.6)Loss of gray–white matter definition, bilateral (4.1)
S160Normal (2.0)Loss of gray–white matter definition, bilateral (5.2)
Figure 2.

MRI (1.5 T, T2-weighted images). (A) Patient S030 shows bilateral anterior temporal blurring of the gray–white junction (dysmyelination), age 4 years. (B) Normal temporal lobes in a 4-year-old child for comparison.

Focal seizures recorded on video-EEG

One or more video-EEG recordings (from a few hours to 4 days) were available for review for 25 patients. The recordings contained no ictal activity in five patients and only myoclonias with or without absences in another five. Video and ictal EEG of 49 motor seizures (other than myoclonias) of 15 patients were reanalyzed (Table 3). The durations of the focal seizures varied from <1 to 18 min.

Table 3. Motor seizuresa recorded on video-EEG
PatientAge(s) at recording (years)N motor (focal) szsIctal discharge localizationSeizure behaviors
  1. PO, posterior; C, central; T, temporal; F, frontal; FC, frontocentral; R, right; L, left; tc, tonic–clonic; psm, psychomotor; szs, seizures; N, number.

  2. a

    Excluding myoclonias.

  3. b

    One ictal discharge in the right frontal region without clinical symptoms.

S0560.4 and 0.611 (3)PO right, C leftHypomotor with apnea
S1240.5 and 1.73 (3)PO right, C left and rightVersion left, clonic left and right
S1311.91 (1)PO rightMyoclonias→version left→left clonic
S1762.45 (1)PO leftVersion right, tonic
S0332.61 (1)PO rightMyoclonias→psychomotor
S1283.02 (1)PO left and bilateralMyoclonias→version right, left
S0163.22 (2)PO right and leftVersion left and psm, psm
S1603.51 (1)PO rightVersion right and psm
K9034.02 (1)PO right→left→rightVersion left→right→left, bilateral tc
S0064.46 (6)PO right and left, FC right, F bilateralMyoclonias→version, unilateral clonic, bilateral tc
S1705.96 (3)T left, F rightPsm, bilateral tc
S16912.31 (2b)PO left, F rightPsm, subclinical
S0302.82 (0)Not localizedBilateral tc
S14718.02 (0)Not localizedBilateral tc
S00213.24 (0)Not localizedBilateral tc

Focal ictal discharges were observed in 25 (51%) of 49 seizures (Table 3). The most common localization was unilateral posterior (11 of 15 patients, 73%). Most posterior discharges were preceded by bilateral spike-waves, which lasted for a few seconds and were often associated with myoclonic jerks. An example of a posterior seizure is shown in Fig. S1A–C. The most common seizure behavior during the posterior discharges was turning of the head and eyes to the contralateral side with impaired responsiveness. Eye blinking was also common. Ictal vomiting occurred in one patient.


Genetic findings

The proportion of SCN1A mutation–positive patients (83%), as well as the different types of identified mutations in our study, was similar to earlier reports in DS (Marini et al., 2011). More than half of the mutations we identified (13/25, 52%) have previously been published, which highlights the importance of recurrent mutations in SCN1A. Two patients with DS had inherited SCN1A mutations from a mildly affected parent. The more severe phenotype may be explained by other unknown genetic factors modifying the symptoms and, in one of these children (K903), perhaps also by severe perinatal asphyxia. We did not find any copy number changes in SCN1A. Because partial or whole SCN1A deletions or duplications are present in around 13% of SCN1A mutation–negative patients with DS (Depienne et al., 2009b), the absence of these could thus be due to chance. A mutation in PCDH19 was identified in one female patient, compatible with the previously estimated proportion of 5% of PCDH19 mutations in DS.

A chromosomal translocation X;9, which was present in two patients (one female, one male) and their mothers has not been described previously as etiology for DS. Given that subtelomeric FISH implied existence of chromosome 9–derived material in the middle of the X chromosome, it is tempting to speculate that the translocation has disrupted the PCDH19 gene, which is located on Xq22.1. However, this was not experimentally verified. Moreover, there was a history of both epilepsy and ataxia in one family, but these symptoms did not segregate with the translocation, suggesting that in this family, it acts as a modifier rather than is a primary cause for DS. Whether the translocation breakpoint on Xq disrupts PCDH19 or other genes modifying the disease phenotype remains unknown. Mosaic deletion of PCDH19 region has been described as a cause of DS in one male patient (Depienne et al., 2009a).

Two (8%) of the 25 tested patients were found to be heterozygous for one of two rare POLG variants (p.Gly517Val or p.Trp748Ser), which occur at the frequency of around 1% in the Finnish control population. Because both of these patients had uncommon neurologic sequelae after status epilepticus, it may be that carriership for a rare POLG variant predisposes to encephalopathy albeit not being deleterious alone (see below).

Acute encephalopathy

The causes of acute encephalopathy in patients with Dravet syndrome have remained elusive. Barbiturates, unidentified viruses, mitochondrial dysfunction, and inappropriate antiepileptic drug treatment have been implied (Chipaux et al., 2010; Nishri et al., 2010; Takayanagi et al., 2010; Tang et al., 2011) as possible causative or contributing agents. Patient S176 with extensive bilateral cortical injury, reversible hepatic failure, and catastrophic neurologic outcome associated with status epilepticus resembles the case report by Nishri et al. (2010), although we found no evidence of predisposing mitochondrial dysfunction. She probably had a significant hypoxic–ischemic event occurring prior to hospital admission. This could have been caused by respiratory depression associated with the relatively large dose of midazolam prior to admission, especially considering the inhibitory action of stiripentol on midazolam metabolism.

Another patient (S200) with encephalopathy showed a focal lesion on MRI and less catastrophic although severe neurologic sequelae. He was shown to be heterozygous for the POLG mutation p.Trp748Ser, which has not been described previously in patients with DS. Homozygosity or compound heterozygosity of this mutation leads to the mitochondrial recessive ataxia syndrome or MIRAS (Hakonen et al., 2005). Carriers of the heterozygous p.Trp748Ser mutation have been shown to have decreased cerebral glucose metabolism without overt neurologic symptoms (Rantamäki et al., 2007).

The third patient (S077) with signs of brain injury (persistent hemiparesis) after a prolonged tonic–clonic seizure in infancy was heterozygous for p.Gly517Val in POLG. The heterozygous p.Gly517Val variant has been associated with a variety of mitochondrial disorders and other neurologic symptoms including seizures and metabolic strokes (Horvath et al., 2006; Hopkins et al., 2010; Kasiviswanathan & Copeland, 2011; Staropoli et al., 2012). There is one previous case report of concomitant heterozygous p.Gly517Val variant and SCN1A mutation in a child whose clinical presentation was DS without episodes of acute encephalopathy (Bolszak et al., 2009). That child also had another POLG mutation (p.Arg722His) and enlarged mitochondria without biochemical evidence of mitochondrial dysfunction in muscle biopsy. Our patient has not had muscle biopsy. Biochemical analysis of the p.Gly517Val variant in vitro has revealed only marginally compromised functional properties (Kasiviswanathan & Copeland, 2011). However, these may be augmented in situations in vivo when the requirements for optimal protein function are high.

We postulate that rare POLG variants could act as modifiers in DS by compromising cortical metabolism during the high energy demands associated with status epilepticus. Seizure semiology in both our patients suggested that they had prolonged focal or lateralized ictal discharges, which may have caused local cortical failure of energy metabolism (Mueller et al., 2001) and subsequent structural damage (Christiaens et al., 2003).

Temporal lobe abnormalities

In line with Siegler et al. (2005) but contrasting with other previous studies (Striano et al., 2007; Akiyama et al., 2010; Ragona et al., 2010; Catarino et al., 2011), we found temporal lobe abnormalities to be very common in patients with DS, provided that they had been imaged after the age of 3 years. In addition to hippocampal sclerosis, we observed abnormally high white matter signal in anterior temporal lobes. This results in the loss of anterior temporal lobe gray–white matter definition, which has been described previously by Mitchell et al. (2003) in children with hippocampal sclerosis, associated with early seizure onset (under 2 years). As in several of our patients, it may be present also without visible hippocampal sclerosis, but probably represents the same spectrum of temporal lobe abnormalities. These abnormalities become visible with time, as the white matter in the anterior temporal region is not myelinated before 3 years (by that age, the white matter is visualized as darker than the cortical gray on T2-weighted images), and are usually missed in those patients who are imaged only during infancy. They may represent focal cortical dysplasia type III, which has been defined recently by the International League Against Epilepsy (ILAE) Diagnostic Methods Commission (Blümcke et al., 2011), reorganization of the cortical cytoarchitecture as a response to an injury of the cortex. To our knowledge, our study is the first to report loss of anterior temporal lobe gray–white matter definition in patients with DS.

Analysis of surgical specimens has shown that the hippocampus is often affected in intractable epilepsy even if the primary seizure focus is extratemporal (Riney et al., 2006). Evidence from case reports shows that injury caused by prolonged febrile seizures is probably one mechanism leading to hippocampal sclerosis (Merkenschlager et al., 2009). If the temporal lobe abnormalities observed on MRI are a consequence of seizure-related damage, neuropathologic studies would be expected to confirm the existence of temporal lobe injury in patients with DS. So far, scarce data are available. An 11-year-old child with DS who died of sudden unexpected death in epilepsy (SUDEP) exhibited multifocal micronodular dysplasia of the left temporal cortex and bilateral gliosis of the hippocampal CA4 region (Le Gal et al., 2010). In contrast, three adult patients with DS showed no neuronal loss in the hippocampal regions (Catarino et al., 2011).

Focal seizures

We found posterior ictal onsets to be common in young children (up to 4 years) with DS. This was different from the almost exclusive frontal or frontocentral localization of focal seizures recorded in adults with DS (Akiyama et al. 2010; Catarino et al., 2011). The predominance of posterior seizures in the youngest age group and a shift to more anterior localizations with increasing age has been described earlier in children with partial epilepsy (Nordli et al., 2001). As suggested by Nordli et al. (2001), this is probably an age-dependent phenomenon associated with the orderly progression of brain maturation from posterior to anterior parts. Perhaps early posterior focal seizures are the first indication of epileptogenicity in the occipital cortex, which is later manifested as light and contrast sensitivity in many patients with DS.

Seizure semiology in our patients with frequent head and eye turning to the contralateral side and clonic activity was similar to what has been described before in complex partial seizures in DS (Ohki et al., 1997) and in posterior seizures (Montassir et al., 2010). When motor signs are subtle and no ictal EEG is available, posterior focal seizures may be mistaken for atypical absences. Ictal vomiting, which was observed in one of our patients, has been reported before in symptomatic occipital epilepsy in childhood (Montassir et al., 2010).


We provide further evidence for temporal lobe injury being common in patients with DS, and for focal seizures predominating in the posterior regions in young patients with DS. We propose that rare heterozygous POLG variants may increase susceptibility to focal brain injury during prolonged seizures in patients with DS, and that the identified X;9 translocations may act as modifiers of DS. The role of POLG variants as a modifying factor in DS and possibly other epilepsy phenotypes should be explored in future studies.


We thank the participating patients and their parents. We thank Eila Herrgård MD, Maria Arvio MD, Pirkko Karttunen MD, Reija Alen MD, Marja Hietala MD, Leila Pajunen MD, and Liisa Metsähonkala MD for referring patients to our study. Hanna Hellgren and Mira Aronen are acknowledged for laboratory analyses, Sinikka Lindh RN for coordinating patient sampling, and medical student Sini Vanhanen for help with the collecting of clinical data. Pirjo Isohanni MD and Anu Suomalainen-Wartiovaara MD are thanked for expert advice concerning POLG studies, and the EuroEPINOMICS Rare Epilepsy Syndromes consortium for carrying out exome sequencing and sequence analysis of patient S200. We thank The Sequencing Initiative Suomi (The SISu Project) for providing access to the aggregate variant data from exomes of the FINRISK cohort (exome sequence data produced by support from the Wellcome Trust grant number WT089062, the Academy of Finland grants 200923 and 00213, and the ENGAGE project, grant agreement HEALTH-F4-2007-201413). This study was funded by the Folkhälsan Research Foundation, Academy of Finland Center of Excellence of Complex Disease Genetics (grant no 213506), Foundation for Pediatric Research, and Arvo and Lea Ylppö Foundation.


Eija Gaily has participated as a paid consultant in an Eisai advisory board meeting. Kai Eriksson has received lecture fees from Eisai and UCB Pharma Ltd. The other authors have no conflicts of interest to disclose. We confirm that we have read the Journal's position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.