Long-term course of Dravet syndrome: A study from an epilepsy center in Japan

Authors


Summary

Objective

This study attempted to clarify the long-term course of Dravet syndrome (DS).

Methods

Sixty-four patients diagnosed with DS (44 with typical DS, and 20 with atypical DS) were studied. The long-term outcomes of clinical seizures, electroencephalographic findings, neuropsychological findings, and social situation were analyzed. The follow-up period ranged from 11 to 34 years 5 months (median 24 years).

Results

At the last visit, the ages ranged from 19 years to 45 years (median 30 years). Fifty-nine patients continued to have generalized tonic–clonic seizures (GTCS). Status epilepticus and unilateral seizures were not observed and myoclonic seizures, atypical absence seizures, and photosensitive seizures were resolved in most patients. The frequency of complex partial seizures was equally low, with five patients at presentation and six patients at the last visit, respectively. Five patients achieved seizure remission (seizure-free for 1 year or longer). Only 1 of 44 patients with typical DS had seizure remission, whereas 4 of 20 patients with atypical DS remitted, with a statistically significant difference between the two phenotypes (p = 0.03). Intellectual disability was found in all patients; especially, severe intellectual disability was prevalent. Patients with atypical DS tended to have milder intellectual disability compared to those with typical DS (p = 0.0283). Occipital alpha rhythm in the basic activity was associated with milder intellectual disability (p = 0.0085). The freedom from seizures correlated with appearance of occipital alpha rhythms (p = 0.0008) and disappearance of epileptic discharges (p = 0.0004). Two patients with GTCS died. Mutations of the neuronal voltage-gated sodium channel alpha subunit type 1 gene were detected at a high frequency (33 of 36 patients examined). Seizure remission was found only in the missense mutation group.

Significance

The long-term seizure and intellectual outcomes are extremely poor in patients with typical DS compared to those with atypical DS. Epilepsy phenotype may influence long-term course of DS.

image

Dr. Rumiko Takayama, pediatric neurologist, Shizuoka Institute of Epilepsy and Neurological Disorders, Japan.

Dravet syndrome (DS), otherwise known as severe myoclonic epilepsy in infancy (SMEI), is a rare epileptic syndrome estimated to affect one in 40,000 children.[1] A recent report from the United Kingdom estimated the incidence of mutation-positive DS to be one in 40,900 births.[2] Although DS is representative of epileptic encephalopathy,[3] in which seizures remain refractory to all conventional antiepileptic drugs (AEDs), the long-term clinical course of this epileptic syndrome is largely unknown.

DS presents in the first year of life in an otherwise normal infant in the form of febrile or afebrile generalized or unilateral clonic seizures, or generalized tonic–clonic seizures (GTCS), later on with additional myoclonic seizures, atypical absence seizures, or partial seizures. The seizures are refractory to AEDs, and mental retardation appears from the second year,[4] but may occur later up to the fourth year.[5] The term “borderline SMEI” (SMEB) has been used to designate a subset of patients with similar clinical symptoms but who do not manifest minor seizures such as myoclonic seizures and atypical absence seizures. We advocated the term “intractable childhood epilepsy with generalized tonic–clonic seizure” (ICEGTC) for a group of patients who manifest refractory grand mal seizures with/without complex partial seizures (CPS).[6, 7] ICEGTC overlaps with SMEB in many aspects.

Because the older terminologies such as SMEI, SMEB, and ICEGTC are somewhat confusing, in this article we use the term “typical DS” for core SMEI and “atypical DS” for ICEGTC and SMEB. Atypical DS refers to patients without minor generalized seizures.

Mutations of the neuronal voltage-gated sodium channel alpha subunit type 1 gene (SCN1A) were first reported in core SMEI (typical DS) patients.[8] The majority of patients with DS have mutations in the SCN1A.[9-11] Currently, these related epilepsies are known to have high frequencies of SCNIA mutations; they are positioned as the SMIE spectrum, and included in DS.[11, 12]

DS is one form of epileptic encephalopathy, but the long-term clinical course is not well studied. Only four series included patients older than 20 years were published.[13-16] Although reports of various investigators concur on the point that the course of disease is marked by uncontrolled seizures, it remains unclear whether seizure control is ultimately achieved in the long term, whether there is a relationship between seizure severity and intellectual outcome, and what is the social and survival outcome of these patients. Fujiwara et al.[17] studied the clinical course of 29 patients with DS aged from 6 years 7 months to 17 years 10 months and found no seizure remission in all the patients, regardless of phenotype. The present study is an extension of the previous study, with a larger number of cases and prolonged observation period, aiming to clarify the long-term outcomes in these patients.

Methods and Subjects

We performed a retrospective review of clinical records to identify patients with DS who presented to our epilepsy center before 1998 and were followed for at least 10 years and who were older than 19 years at the last visit. In this study, patients with a diagnosis of SMEI or ICEGTC who had the following characteristics were selected. (1) Before onset, the infant had no history of brain damage and no developmental problem. (2) Around 6 months after birth, fever or bathing triggered the first episode of unilateral or bilateral clonic or tonic–clonic seizure. (3) The first seizure may manifest as convulsive status. (4) Unilaterally dominant seizure may involve the left side or right side alternatively. (5) Uncontrolled convulsive seizures persisted, with additional appearance of myoclonic seizures or atypical absence seizures after 1 year of age. (6) Cases without the above minor generalized seizures were also included. (7) Complex partial seizures may coexist. The above characteristics match the description of DS in a recent review.[18] Seizure remission was defined as seizure free for 1 year or longer at last visit. The Fisher's exact test and chi-square were used for statistical analysis. A p-value of <0.05 was regarded as significant.

A total of 64 patients (30 male and 34 female) were identified based on the above criteria. Data of clinical seizures, neuropsychiatric disorders, and social outcome were extracted. All the patients had experienced hospitalization for treatment, and thereafter were followed and treated at the outpatient clinic. All the patients underwent electroencephalography (EEG) examination at the last visit. During follow-up, EEG including awake and sleep states was examined at least once a year in almost all cases. Intellectual level was determined by clinical assessment. SCN1A analysis was conducted in 36 patients, and gene abnormalities were detected in 33 patients. The methods of gene analysis were described elsewhere.[9, 11] Written informed consent was obtained from the parents or responsible adults where necessary, and the study protocol was approved by the ethical committees of Shizuoka Institute of Epilepsy and Neurological Disorders and of the Institutional Review Board of RIKEN-BSI.

Results

Patients

The ages of epilepsy onset ranged from 2 to 11 months (median 5 months). The ages of initial presentation to our epilepsy center ranged from 1 year to 16 years and 8 months (median 5 years and 5 months); more than half of the patients (33/64, 51.6%) were aged 5 years or below. The ages at the last visit ranged from 19 years to 45 years (median 30 years). The follow-up period ranged from 11 years to 34 years and 5 months (median 24 years). A family history of febrile convulsion was found in 29 patients, and a family history of epilepsy was found in 8 patients. All patients were taking multiple AEDs including carbamazepine in three patients, but no new drugs such as stiripentol, topiramate, or lamotrigine. Discontinuation of AEDs was not attempted in any of the patients.

The 64 patients were divided into two groups according to the presence or absence of “combined” minor generalized seizures (myoclonic seizures or atypical absence seizures) other than GTCS throughout the course of epilepsy. Forty-four patients who had minor generalized seizures met the criteria of typical DS according to the proposal of the Commission on Classification and Terminology of the International League Against Epilepsy (ILAE).[19] Twenty patients were classified as atypical DS.

Seizure outcome

From onset to presentation

In all patients, prolonged generalized clonic seizures, GTCS, or unilateral seizures (US) were often induced by fever or bathing. These seizures occurred in clusters and often evolved to status epilepticus (SE). Despite treatment with AEDs, these seizures, whether febrile or afebrile, persisted uncontrolled, and convulsive SE occurred repeatedly in 39 patients. Myoclonic seizures coexisted in 40 patients, atypical absence seizures in 24 patients, and CPS in 18 patients (some patients had multiple seizure types). Photo-sensitive and/or pattern-sensitive seizures were found in seven patients.

At presentation and last visit

At presentation to our epilepsy center, all patients had GTCS or US. Eleven patients had recurrent SE, 24 had myoclonic seizures, 19 had atypical absence seizures, and five had CPS (some patients had multiple seizure types). Photo-sensitive and/or pattern-sensitive seizures were present in seven patients.

At the last visit, none of the patients had US and SE no longer occurred; 59 patients had GTCS, 5 had myoclonic seizures, 1 had atypical absence seizures, and 6 had CPS. Photo-sensitive and/or pattern-sensitive seizures were observed in no patients. In the majority of the patients, GTCS persisted, but the seizure frequency tended to decrease over time. Low-grade fever became less provocative. In most cases, frequency of GTCS was between weekly and monthly occurrence. Seizures occurred mostly during nocturnal sleep. There was an increase in the proportion of patients with GTCS reduced to yearly occurrence. In five patients, GTCS were controlled at ages ranging from 15 to 28 years (median 19 years). All five cases were terminal remission cases and were free from all types of seizure for 5–22 years at the time of the last visit (Figs. 1 and 2).

Figure 1.

Evolution of various seizure types in patients with DS. At the last visit, none of the patients had US, and none had SE. MS and AAS decreased with time. The frequency of patients manifesting CPS was unchanged. In the majority of the patients, GTCS persisted. GTCS, generalized tonic–clonic seizures; SE, status epilepticus; MS, myoclonic seizures; AAS, atypical absence seizures; CPS, complex partial seizures; US, unilateral seizures; DS, Dravet syndrome.

Figure 2.

Evolution of GTCS frequency. The seizure frequency tended to decrease over time. There was a marked increase in the proportion of patients with GTCS reduced to yearly occurrence. In five patients, GTCS were completely controlled. GTCS, generalized tonic–clonic seizures.

When examining the seizure remission rate according to phenotype, only one (2.3%) of 44 patients with typical DS had seizure remission, whereas 4 (20.0%) of 20 patients with atypical DS remitted, with a significant difference between the two phenotypes, indicating an association between seizure remission rate and phenotype in patients with DS (p = 0.03).

EEG

At presentation to our epilepsy center, 48 patients (75.0%) had epileptic discharges of various qualities. The epileptic discharges consisted of diffuse spike-wave complex or diffuse polyspike-wave complex in 14 patients (21.9%); diffuse spike-wave complex, diffuse polyspike-wave complex and focal spike and sharp waves in 14 patients (21.9%); focal spike and sharp waves in 20 patients (31.2%; frontal in three patients, temporal in 5 patients, and multifocal in 12 patients). When the patients were divided into those with diffuse epileptic discharges and those with focal epileptic discharges, the distribution was similar, with 28 patients (43.8%) in the former group and 34 patients (53.1%) in the latter group (some patients had both types). Epileptic discharges were not observed in 16 patients (25.0%) (Fig. 3).

Figure 3.

Evolution of EEG findings. Focal discharges became dominant during follow-up.

At the last visit, epileptic discharges were detected in 49 patients (76.6%). The epileptic discharges consisted of diffuse spike-wave complex or diffuse polyspike-wave complex in 5 patients (7.8%); diffuse spike-wave complex, diffuse polyspike-wave complex and focal spike and sharp waves in 1 patient (1.6%); and focal spike and sharp waves in 43 patients (67.2%; frontal in 21 patients, temporal in 3 patients, central in 1 patient, parietal in 1 patient, occipital in 1 patient, and multifocal in 16 patients). When the patients were divided into those with diffuse epileptic discharges and those with focal epileptic discharges, there were 6 patients (9.4%) in the former group and 44 patients (68.8%) in the latter group, showing a marked dominance of focal epileptic discharges (some patients had both types). Epileptic discharges were not observed in 15 patients (23.4%). Of these, five patients achieved freedom from seizures. On the other hand, among 49 patients who had epileptic discharges at the last visit, no patients achieved freedom from seizures. The freedom from seizures significantly correlated with disappearance of epileptic discharges (p = 0.0004). Among the 16 patients who had no epileptic discharges at presentation (from 1 to 16 years of age), 11 patients had epileptic discharges at the last visit. Of the remaining five patients with no epileptic discharges at the last visit, four patients had rare epileptic discharges sometimes during follow-up. In one patient there were no epileptic discharges at all despite numerous EEG recordings during follow-up (7 years 10 months to 34 years of age).

Diffuse high potential slow wave activity with central and parietal dominance was observed in 49 patients at the time of presentation to our epilepsy center, but the number was reduced to 16 at their last visit. At the last visit, occipital alpha rhythms were found in their basic EEG activity in 17 patients. Among them, 3 patients had slower posterior dominant rhythms (PDRs) at 8 Hz and 14 patients had normal PDRs ranging from 9 to 11 Hz. In five patients who were taking benzodiazepine, β waves were found accompanied by PDRs. Among 17 patients who had occipital alpha rhythms, 5 patients achieved freedom from seizures. On the other hand, among 47 patients who had no occipital alpha rhythms no patients achieved freedom from seizures. The freedom from seizures significantly correlated with appearance of occipital alpha rhythms in the background activity (p = 0.0008).

Regarding correlation between occipital alpha rhythms and epileptic discharges at the last follow-up EEG, 17 patients had occipital alpha rhythms; of these, 7 had epileptic discharges and 10 had no epileptic discharges. Forty-seven patients had no occipital alpha rhythms; of these 40 had epileptic discharges and 7 had no epileptic discharges. The appearance of occipital alpha rhythms correlated significantly with the disappearance of epileptic discharges (p = 0.001; Table 1).

Table 1. Occipital alpha rhythms and outcome
Occipital alpha rhythms at last visit (n)Patients with seizure freedom (n)aPatients without epileptic discharges (n)bPatients with mild intellectual disability (n)c
  1. n, number of patients.

  2. a

    The freedom from seizures correlated significantly with the appearance of occipital alpha rhythms (p = 0.0008).

  3. b

    The appearance of occipital alpha rhythms correlated significantly with disappearance of epileptic discharges (p = 0.001).

  4. c

    Patients with occipital alpha rhythms tended to have milder intellectual disability compared to those without (p = 0.0085).

+ (17)5/1710/173/17
− (47)0/477/470/47

Neuropsychological outcome at the last visit

Four patients were bedridden; whereas the remaining 60 patients were capable of walking, over half of them had ataxic gait. Four patients had hemiplegia, all of whom had permanent hemiplegia as a sequel of prolonged febrile SE, which occurred between 1 year and 1 month and 4 years of age.

Intellectual disability was observed in all patients: severe in 49 patients (76.5%), moderate in 12 patients (18.8%), and mild in 3 patients (4.7%), with most patients having severe intellectual disability. Seventeen patients (26.6%) were admitted to institution, 45 patients (70.3%) were attending sheltered workshops, and only two patients (3.1%) were living independently.

Table 2 shows the relationship between the severity of intellectual disability and the frequency of GTCS at the last visit. Among those with severe intellectual disability, 26 patients had daily to weekly GTCS, 21 patients had monthly to yearly GTCS, and 2 patients were seizure-free. Among those with moderate intellectual disability, four patients had daily to weekly GTCS, seven patients had monthly to yearly GTCS, and one patient was seizure-free. Among those with mild intellectual disability, no patient had daily to weekly GTCS, one patient had monthly to yearly GTCS, and two patients were seizure-free. These data showed a tendency of more severe intellectual disability in patients with higher GTCS frequency. A significant difference was observed between the severity of intellectual disability and the GTCS frequency at the last visit (p = 0.0019).

Table 2. Intellectual outcome and frequency of generalized tonic–clonic seizures at the last visit
Intellectual disabilityDaily to weekly seizures (n)Monthly to yearly seizures (n)Remitted (n)
  1. n, number of patients.

  2. Higher GTCS frequency correlates with more severe intellectual disability (p = 0.0019).

Mild012
Moderate471
Severe26212

Among 17 patients showing occipital alpha rhythms, intellectual disability was mild in 3 patients, moderate in 4 patients, and severe in 10 patients. On the other hand, among 47 patients showing no occipital alpha rhythms, no patient had mild intellectual disability, whereas 8 patients had moderate and 39 patients had severe intellectual disability. Patients with occipital alpha rhythms tended to have milder intellectual disability compared to those without (p = 0.0085; Table 1).

When examining the cognitive outcome according to phenotype, among 44 patients with typical DS, intellectual disability was moderate in 8 patients and severe in 36 patients. No patient had mild intellectual disability. On the other hand, among 20 patients with atypical DS, 3 patients had mild, 4 patients had moderate, and 13 patients had severe intellectual disability. Patients with atypical DS tended to have milder intellectual disability compared to those with typical DS (p = 0.0283; Fig. 4).

Figure 4.

Phenotypes and outcomes of intellectual disability. Patients with atypical DS tended to have milder intellectual disability compared to those with typical DS (p = 0.0283).

Mortality

Among 64 patients, 2 patients (one male and one female) died. The causes of death were febrile illness–related sudden death in one patient and drowning in another. Both patients had GTCS, and one patient had myoclonic seizures and atypical absences. There was no significant correlation between phenotype and mortality.

Intellectual disability was severe in both patients. Both patients had persistent GTCS, and seizure frequency did not change before death.

SCN1A mutation and clinical course

Mutational analysis was performed on all coding exons and splice sites of SCN1A in 36 patients, and 32 heterozygous mutations were found in 33 patients (91.7%), consisting of 7 frameshift mutations in 7 patients, 7 nonsense mutations in 8 patients, and 18 missense mutations in 18 patients, which were not found in the control population (n = 93–111). No mutation was detected in the coding region of SCN1A in three patients (no. 34, 35, and 36). The mutations were all different except in one pair of monozygotic twins (Table 3).

Table 3. Genotype-phenotype correlation
PatientGenderDS phenotypeLast visit SCN1A
Age (years)Seizure typeIntellectual disabilityMotor disabilitySocial outcomeAlpha rhythmsMutationDNA
Protein
  1. GTCS, generalized tonic–clonic seizures; MS, myoclonic seizures; CPS, complex partial seizures; SE, status epilepticus.

  2. GTCS frequency: y, yearly; m, monthly; w, weekly; d, daily.

  3. Group A: missense mutation (patients 1–18). Group B: nonsense mutation (patients 19–26), frameshift (patients 27–33).

  4. SCN1A mutation listed for the SCN1A variant database.[40]

1[11]FTypical21GTCS(w)SevereAtaxic gaitInstitutionalizedMissensec.5459T>C
F1820S
2[11]FTypical23GTCS(m)SevereUnremarkableSheltered workshopMissensec.2845A>G
M949V
3[11]MTypical26GTCS(w), MSSevereAtaxic gaitInstitutionalizedMissensec.5021C>A
A1674D
4[11]FTypical32GTCS(m)SevereUnremarkableInstitutionalizedMissensec.2921A>T
N974I
5[11]MTypical33GTCS(w), CPSSevereAtaxic gaitSheltered workshopMissensec.1028G>A
G343E
6[11]MTypical34GTCS(m)SevereAtaxic gaitInstitutionalized+Missense 
W1801G
7[11]FTypical34RemittedSevereAtaxic gaitSheltered workshop+Missensec.3756C>G
F1252L
8[11]FTypical35GTCS(m)SevereAtaxic gaitSheltered workshop+Missensec.307A>G
S103G
9[35]MTypical27GTCS(m)SevereUnremarkableSheltered workshopMissensec.2957T>C
L986P
10FTypical32GTCS(w)SevereAtaxic gaitSheltered workshopMissensec.840G>C
W280C
11MTypical37GTCS(m)SevereRight hemiplegia following febrile SE at 3 years 3 months oldSheltered workshop+Missensec.1055T>G
V352G
12[11]MAtypical27GTCS(m)SevereAtaxic gaitInstitutionalizedMissensec.2902G>A
G968R
13[36]FAtypical28RemittedMildUnremarkableSheltered workshop+Missensec.4096G>A
V1366I
14[11]MAtypical29GTCS(d)SevereAtaxic gaitSheltered workshopMissensec.2915T>C
V972A
15[11]MAtypical33GTCS(m)SevereAtaxic gaitSheltered workshopMissensec.5093C>T
T1698I
16[11]FAtypical33RemittedSevereUnremarkableSheltered workshop+Missensec.5389T>C
F1797L
17[37]FAtypical34GTCS(m)MildUnremarkableLiving independently+Missensec.1177C>T
R393C
18FAtypical20GTCS(w), CPSSevereUnremarkableSheltered workshopMissensec.719T>C
L240P
19[11, 38]MTypical19GTCS(w)SevereAtaxic gaitInstitutionalizedNonsensec.3604C>T
R1202X
20[11, 38]MTypical19GTCS(w)SevereAtaxic gaitInstitutionalizedNonsensec.3604C>T
R1202X
21[9]FTypical29GTCS(w)SevereAtaxic gaitSheltered workshopNonsensec.4514C>A
S1505X
22[11]MTypical29GTCS(d)SevereAtaxic gaitSheltered workshop+Nonsensec.3819G>A
W1273X
23[9]MTypical32GTCS(y)SevereAtaxic gaitSheltered workshopNonsensec.2101C>T
R701X
24[39]FTypical24GTCS(d)SevereRight hemiplegia following febrile SE at 2 years oldInstitutionalizedNonsensec.1834>T
R612X
25[35]FAtypical26GTCS(m)SevereAtaxic gaitSheltered workshop+Nonsensec.1738C>T
R580X
26FAtypical45GTCS(y)SevereLeft hemiplegia following febrile SE at 1 year 1 month oldInstitutionalizedNonsensec.4219C>T
R1407X
27[9]MTypical19GTCS(d)SevereAtaxic gaitSheltered workshopFrameshiftc.5678insACTA
Leu1659fsX1667
28[9]MTypical32GTCS(w)SevereAtaxic gaitInstitutionalizedFrameshiftc.2841delC
Asp947fsX962
29MTypical19GTCS(d), CPSSevereBedriddenInstitutionalizedFrameshiftc.5059delT
F1687fsX1714
30FTypical36GTCS(m), CPSModerateUnremarkableSheltered workshopFrameshiftc.4265-8insGGGT
Y1422fsX1431
31FTypical28GTCS(m)ModerateAtaxic gaitSheltered workshopFrameshiftc.5131delG
F1710fsX1714
32FTypical37GTCS(m), CPSSevereBedriddenSheltered workshopFrameshiftc.5540insAAAC
L1847fsX1861
33FAtypical35GTCS(w)ModerateUnremarkableSheltered workshopFrameshiftc.5108insACAT
M1704fsX1709
34[11]MTypical32GTCS(y)SevereUnremarkableSheltered workshop+ Not detected
35MTypical28GTCS(w), MSModerateAtaxic gaitSheltered workshop Not detected
36[11]FAtypical19GTCS(d)SevereBedriddenSheltered workshop+ Not detected

Mutations of the SC1NA were detected at a high frequency both in typical DS and in atypical DS (23/25; 92.0% vs. 10/11; 90.9%). Although statistically not significant, truncating mutations were more common in typical DS than in atypical DS (12/23; 52.2% vs. 3/10; 30.0%); missense mutations were less common in typical DS than in atypical DS (11/23; 47.8% vs. 7/10; 70.0%).

Regarding types of mutations, the patients were divided into two groups: 18 patients with missense mutation as group A, and 15 patients with truncating mutation (by grouping nonsense and frameshift mutations together) as group B. The two groups were compared with respect to clinical parameters and long-term outcomes. Group A consisted of 11 patients with typical DS and 7 patients with atypical DS, whereas group B consisted of 12 patients with typical DS and 3 patients with atypical DS (Fig. 5). Group A consisted of 8 male and 10 female patients, whereas group B consisted of seven male and eight female patients. The ages of onset ranged from 2 to 11 months (median 5 months) in group A, and 4 to 11 months (median 6 months) in group B. The ages at presentation to our center ranged from 1 year to 11 years and 2 months (median 4 years 5 months) in group A, and 1 year 8 months to 16 years 8 months (median 4 years 1 month) in group B. The ages of last visit ranged from 21 to 37 years (median 32 years) in group A, and 19 to 45 years (median 29 years) in group B.

Figure 5.

Types of SCN1A mutations and phenotypes. Typical DS tended to be more common in group B, although this relation did not reach statistical significance (p = 0.2828).

When the clinical subtypes were compared, typical DS tended to be more common in group B, although the difference was not significant (p = 0.2828; Fig. 5). No differences in gender, and in ages at onset, at presentation, and at the last visit were found between group A and group B. The outcome of various seizure types was also not different between the two groups (Fig. 6). Three patients in group A were seizure-free for 1 year or longer. Complete seizure control was observed only in group A. The phenotypes in the remitted cases were typical DS in one patient; and atypical DS in two patients. At the last visit, intellectual disability was severe in 16 patients in group A and 12 patients in group B, moderate in no patients in group A and 3 patients in group B, and mild in 2 patients in group A and no patient in group B (Fig. 7).

Figure 6.

Types of SCN1A mutations and seizure outcome. The outcomes of various seizure types were not different between the two groups. Patients with seizure freedom for 1 year or longer were found only in group A.

Figure 7.

Types of SCN1A mutations and outcomes of intellectual disability. Patients with mild intellectual disability were found only in group A.

Discussion

Compared to other epileptic encephalopathies such as West syndrome and Lennox-Gastaut syndrome, the concept of DS as an epileptic encephalopathy was established at a later date, which partially accounts for the paucity of reports on the long-term course of this epilepsy syndrome. Our center, the first epilepsy center in Japan, was inaugurated in 1975. Because of the concentration of refractory epilepsy cases referred from all over Japan, we have taken an interest in a group of fever-induced intractable grand mal epilepsies (nowadays referred to as DS) from the early days. Therefore, we were in a favorable position to study the long-term course of these epilepsies in a retrospective manner. Another feature of the present study is that we were able to observe the long-term course of a cohort longitudinally at a single epilepsy center.

The long-term course of DS was studied by investigating the courses of three outcome measures: death, refractory seizures, and remission. This study is characterized by a long follow-up period of 11 years to 34 years and 5 months (median 24 years). In terms of the long-term outcome of epileptic seizures, myoclonic seizures and atypical absence seizures decreased with time, and photosensitive seizures were also reduced markedly. On the other hand, only GTCS persisted in all patients. However, the frequency of GTCS also decreased over time, and SE no longer occurred during follow-up. This trend is consistent with our previous report[17] and those of other investigators.[13, 20, 21] Although few in number, some patients did achieve remission of seizures. Remission was seen after 15 years of age, with the latest at 28 years. These findings indicate that epileptogenicity attenuates with age.

In the present study, the frequency of patients manifesting CPS was equally low at presentation and at the last visit. Akiyama et al.[15] reported an increase in partial seizures over time. In our study, however, we observed no increase in partial seizures. Genton et al.[16] also reported that only one of 24 patients continued to have partial seizures at the last visit. Although Genton et al.[16] reported persistence of US, we found that US disappeared in all the cases. The present finding that GTCS continued for a long period is consistent with previous reports.

Photosensitive seizures disappeared in all patients during follow-up. Although photosensitivity seen in idiopathic epilepsy is common during adolescence,[22] photosensitivity in DS is marked during infancy and is strongly manifested. The photoparoxysmal response (PPR) in patients with idiopathic general epilepsy is wave-length dependent.[23] In contrast, DS patients exhibit quantity-of-light dependent PPR.[24] The age dependence of photosensitivity in DS patients may reflect the difference in mechanism of PPR expression.

Regarding the evolution of EEG findings, focal discharges became dominant during follow-up. Moreover, localization of spike focus in the frontal lobe became more common. However, no increase in patients with partial seizures and no evolution to frontal lobe epilepsy were observed. There was no correlation between the location of EEG focus and clinical seizures. Bureau and Dalla Bernardina[25] proposed that the interictal focal spikes lack epileptogenicity and represent fragmentary expression of diffuse discharges.

In the long-term course, there was an increase in number of patients showing reduced diffuse high potential slow waves and increase in occipital alpha rhythms. In all five patients who achieved seizure remission, occipital alpha rhythms appeared and epileptic discharges disappeared.

Because paroxysmal EEG abnormalities in DS patients can fluctuate according to the different conditions, a precise evaluation of the true incidence and frequency of the paroxysmal EEG abnormalities in DS patients is difficult.[25] In this series, five patients had no epileptic discharges both at the first and the last visit. Probably the lack of EEG recordings in young age can explain why those patients showed no or rare epileptic discharges during observation. Wide range of ages at presentation is one of the limitations of this study.

Intellectual outcome was very poor. Most patients had moderate to severe intellectual disability. The intellectual disability of the five patients with seizure remission was severe in two, moderate in one, and mild in two. Higher GTCS frequency correlated with more severe intellectual disability (Table 2).

Despite decrease of seizure severity with age, the neuropsychological outcome was poor. Akiyama et al.[15] reported a correlation between the presence of occipital alpha rhythms and mild intellectual disability. We also observed a trend of milder intellectual disability in patients with occipital alpha rhythms compared with those without.

In DS, SE has been reported to cause severe neuropsychological impairments and acute encephalopathy.[26] In our series also, permanent hemiplegia and developmental regression developed after febrile SE in four cases. Attention has to be given to severe neurologic sequelae that occur occasionally after SE in DS.

Two patients died. Mutational analysis was not performed for them. According to Sakauchi et al.,[27] sudden death peaked at 1–3 and 18 years of age, acute encephalopathy-related death peaked at 6 years, and death from drowning occurred at 7 years and above. In the present study, the ages at death were 19 years from drowning and 24 years from febrile illness–related sudden death, similar to those reported by Sakauchi et al.[27] Both cases had had GTCS, and their seizure frequency had not changed before death.

Since it became known that SC1NA mutation is the main cause of DS,[8-10] the relationship between genotype and phenotype has been a topic of much discussion. The phenotype-genotype correlation is complex. Zuberi et al.[28] examined >800 SCN1A mutations and conducted a detailed analysis on the relationship between the mutation genotype (truncating mutation or missense mutation) and phenotype. They reported that the ages of seizure onset (months after birth) for prolonged seizures, myoclonic seizures, and atypical absence seizures were earlier in patients with truncating mutation than in patients with missense mutation, suggesting that truncating mutation has greater contribution to severe phenotype than missense mutation. When the cohort was further followed and the factors associated with 5-year outcome were studied, the factors associated with poor outcome were frequent SE, EEG abnormality before 1 year of age, and motor disorder, whereas mutation class (truncating vs. missense) had no relation with outcome.[2]

There are few studies that verify the relationship between genotype and long-term course, especially the outcome in adult patients. Jansen et al.[13] reported that any particular SCN1A mutation type did not affect seizure outcome, but the number of cases was small. Akiyama et al.[15] investigated the relationship between long-term outcome and genotype in 25 adult patients with DS who harbored SCN1A mutations, and found that the mutation class had no effect on the long-term seizure and intellectual outcomes. Catarino et al.[29] reported the relationship between genotype and phenotype in adults with DS, including autopsy cases. Although all four patients with sudden death occurring during infancy had truncating mutations, two patients that died during adulthood had no mutation (one case) or missense mutation (one case) but no truncating mutation. Because missense mutation was found in milder phenotype capable of survival until adulthood, whereas truncating mutation was present in severe phenotype causing death during childhood, the potential effect of genotype class on phenotype may be speculated.

Epilepsies associated with SCN1A abnormalities distribute in a continuous spectrum of severity, with DS being the most severe and generalized epilepsy with febrile seizure plus (GEFS+) the least severe. DS consists of a typical type and an atypical clinical subtype. Whether seizure outcome differs between the two types of DS is not certain; we encountered no remitted cases in both types in our previous study.[17] In the present study, however, long-term follow-up of patients until adulthood revealed a difference in seizure outcome between the two phenotypes. Seizure remission is extremely rare in patients with typical DS, even after reaching adulthood.

Recently a study of cognitive development in DS was reported.[31] In this series, early appearance of myoclonus and absences was associated with the worst cognitive outcome. The authors suggest that the epilepsy phenotype has a prognostic value. In our series we found significant difference in intellectual outcome between typical DS and atypical DS (p = 0.0283). Patients with atypical DS tended to have milder intellectual disability. Our results show that epilepsy phenotype influences not only seizure outcome but also intellectual outcome.

The long-term seizure and intellectual outcomes are affected by various factors interacting in a complex manner, and identification of single attributing factors is difficult. Our study is a retrospective and descriptive one. Much more research and long-term prospective studies of large numbers of patients are needed to examine the roles of various factors. The fact that the extent of damage from convulsion plays an important role in outcome has been elucidated in previous studies as well as the present results. Further research about the possible role played by different kinds of epileptic seizures on development is needed to clarify the contribution of epilepsy-related factors to the cognitive phenotype.[5]

The present study reconfirms that the long-term intellectual and seizure outcomes of DS are extremely poor except a few cases. Although DS is refractory to all the conventional AEDs, recent studies have reported the effectiveness of new drugs such as stiripentol[31, 32] and topiramate. However, the patients in this series were not treated with the new drugs. It is important to avoid drugs with aggravating effects such as carbamazepine and lamotorigine.[4] Identification of the DS in the early stage would be of the utmost important to start appropriate treatment and avoid aggressive heavy medication. Simple screening test could help to predict DS.[34]

We must acknowledge that our series consisted of severe cases diagnosed a few decades ago. Current earlier recognition of DS and early initiation of suitable new drugs and further development of new effective therapies are likely to lead better outcomes.

Disclosure

None of the authors has any 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.

Ancillary