Clinical genetic study of the epilepsy-aphasia spectrum

Authors

  • Meng-Han Tsai,

    1. Epilepsy Research Centre, Department of Medicine, Austin Health, University of Melbourne, Heidelberg, Victoria, Australia
    Search for more papers by this author
    • These authors contributed equally to the manuscript.

  • Danya F. Vears,

    1. Epilepsy Research Centre, Department of Medicine, Austin Health, University of Melbourne, Heidelberg, Victoria, Australia
    Search for more papers by this author
    • These authors contributed equally to the manuscript.

  • Samantha J. Turner,

    1. Department of Paediatrics, Royal Children's Hospital, University of Melbourne, Melbourne, Victoria, Australia
    Search for more papers by this author
  • Robert L. Smith,

    1. Newcastle Medical School, John Hunter Children's Hospital, Newcastle, New South Wales, Australia
    Search for more papers by this author
  • Samuel F. Berkovic,

    1. Epilepsy Research Centre, Department of Medicine, Austin Health, University of Melbourne, Heidelberg, Victoria, Australia
    Search for more papers by this author
  • Lynette G. Sadleir,

    1. Department of Paediatrics, School of Medicine and Health Sciences, University of Otago, Wellington, New Zealand
    Search for more papers by this author
  • Ingrid E. Scheffer

    Corresponding author
    1. Department of Paediatrics, Royal Children's Hospital, University of Melbourne, Melbourne, Victoria, Australia
    2. Florey Institute of Neuroscience, Heidelberg, Victoria, Australia
    • Epilepsy Research Centre, Department of Medicine, Austin Health, University of Melbourne, Heidelberg, Victoria, Australia
    Search for more papers by this author

Address correspondence to Ingrid E. Scheffer, Melbourne Brain Centre, Level 2, 245 Burgundy Street, Heidelberg, Victoria 3084, Australia. E-mail: scheffer@unimelb.edu.au

Summary

Purpose

To characterize the frequency and nature of the family history of seizures in probands with epilepsy falling within the epilepsy-aphasia spectrum (EAS) in order to understand the genetic architecture of this group of disorders.

Methods

Patients with epileptic encephalopathy with continuous spike-and-wave during sleep (ECSWS), Landau-Kleffner syndrome (LKS), atypical benign partial epilepsy (ABPE), and intermediate epilepsy-aphasia disorders (IEAD) were recruited. All affected and available unaffected relatives up to three degrees of relatedness underwent phenotyping using a validated seizure questionnaire. Pedigrees were constructed for all families. The proportion of affected relatives according to each degree of relatedness was calculated. The epilepsy phenotypes in close relatives were analyzed. The data were compared to the families of probands with benign childhood epilepsy with centrotemporal spikes (BECTS) using the same methodology.

Key Findings

Thirty-one probands, including five ECSWS, three LKS, one ABPE, and 22 IEAD were recruited. The mean age of seizure onset was 3.9 (range 0.5–7) years. A male predominance was seen (68%, 21/31) . Sixteen (51.6%) of 31 had a positive family history of seizures. Among 1,254 relatives, 30 (2.4%) had a history of seizures: 13 (10.2%) of 128 first-degree relatives, 5 (1.7%) of 291 second-degree relatives, and 12 (1.4%) of 835 third-degree relatives. Thirteen had febrile seizures, including two who had both febrile seizures and epilepsy. Of the 19 relatives with epilepsy, 4 had BECTS, 4 epilepsies with focal seizures of unknown cause, 3 IEAD, and 7 unclassified. One had genetic generalized epilepsy. In the families of the BECTS probands, 9.8% of first-degree, 3% of second-degree, and 1.5% of third-degree relatives had seizures, which was not significantly different from the EAS cohort families.

Significance

The frequencies of seizures in relatives of probands with EAS suggest that the underlying genetic influence of EAS is consistent with complex inheritance and similar to BECTS. The phenotypic pattern observed in the affected relatives comprised predominantly febrile seizures and focal seizures. These findings suggest that a shared genetic predisposition to focal epilepsies underpins the epilepsy-aphasia spectrum.

Landau-Kleffner syndrome (LKS) and epileptic encephalopathy with continuous spike-and-wave during sleep (ECSWS) are the archetypal epilepsy-aphasia syndromes in which an acquired aphasia or more pervasive cognitive regression occurs in the setting of electrical status epilepticus in sleep. The electroencephalography (EEG) correlate of these syndromes is continuous spike-and-wave during sleep (CSWS), which is defined as bilaterally synchronous discharges that comprise >85% of slow wave sleep. CSWS typically occurs at some time during LKS and is pathognomonic in ECSWS. Seizures occur in 70% patients with LKS and all with ECSWS. LKS and ECSWS are considered syndromes at the severe end of the epilepsy-aphasia spectrum (EAS).

At the mild end of the spectrum of epilepsy associated with language and speech disorders lies benign childhood epilepsy with centrotemporal spikes (BECTS), which is the most common focal epilepsy in children younger than 16 years of age (Shinnar et al., 1999; Dalla Bernardina et al., 2005). By definition, children with BECTS have normal cognitive function, yet in the past two decades, persistent or transient neuropsychological impairments in different domains, but especially language, have been described (Staden et al., 1998; Monjauze et al., 2005; Riva et al., 2007; Lillywhite et al., 2009). In rare instances, a much more severe seizure disorder may occur, termed atypical benign partial epilepsy of childhood (ABPE) in which epileptic negative myoclonus, atonic seizures, and typical rolandic seizures occur. The EEG ranges in severity from centrotemporal spikes (CTS) to CSWS, and the outcome is typically good (Aicardi & Chevrie, 1982; Hirano et al., 2009).

In clinical practice, some patients fall between the two ends of the epilepsy-aphasia spectrum. They show a combination of abnormal cognitive development or regression (predominantly affecting language function), with or without clinical seizures, yet do not fulfil the EEG criteria for CSWS. Their EEG shows sleep activation of focal epileptiform discharges (predominantly centrotemporal but other regions can be involved) either without bilaterally synchronous activity, or with synchronous activity occupying <85% of NREM (non–rapid eye movement) sleep. These cases were called intermediate epilepsy-aphasia disorders (IEADs).

The etiology of epilepsies within the epilepsy-aphasia spectrum is largely unknown. Although there are case reports describing familial occurrence, clinical genetic studies have not been performed. Herein we characterize the clinical features and study the families of 31 probands with EAS exploring the presence of seizures in first-, second-, and third-degree relatives. Our aim was to understand the underlying genetic architecture of these severe focal epilepsies of childhood.

Methods

Patients were recruited from three sources: (1) clinical practice of pediatric epileptologists (IES, LGS), (2) a community-based epilepsy twin study (SFB, Australian Twin Registry), and (3) referrals from neurologists or pediatricians in Australia and New Zealand to the Epilepsy Research Program. Four families were also included in our previous study of the family history of BECTS (Vears et al., 2012). Electroclinical phenotyping was performed by an epileptologist, and a validated seizure questionnaire was used (Reutens et al., 1992).

Probands

Patients were classified using the following criteria:

  1. ECSWS where probands had global or selective cognitive regression, seizures, and CSWS, defined as >85% of NREM sleep. Seizures could be focal, generalized, atonic, or negative myoclonic in type.
  2. LKS where probands had predominantly language regression on a background of normal development or isolated language delay, and bilaterally synchronous perisylvian discharges activated during sleep. Seizures were usually infrequent.
  3. ABPE where probands had epileptic negative myoclonic or atonic seizures in addition to rolandic seizures (RS). RS typically began with an aura of unilateral tongue or perioral paresthesia followed by facial tonic or clonic movement that could spread to the upper limb. EEG showed frequent CTS that were activated in sleep but did not fulfil the criteria for CSWS.
  4. IEAD where probands had cognitive regression, delay, or plateauing (predominantly in the language domain) with centrotemporal epileptiform discharges that showed sleep activation yet did not fulfil the criteria for CSWS. Seizures were common but not essential.

Exclusion criteria included:

  1. Tonic seizures
  2. No EEG abnormalities or where an EEG was not performed
  3. Children with classical BECTS (we recently reported a clinical genetics study of BECTS) (Vears et al., 2012)

The term EAS is used here to refer to all four groups.

Family study

Detailed pedigrees of each proband were constructed for up to three degrees of relatedness. Strenuous attempts were made to contact all affected relatives to invite them to participate in the study. Detailed interviews using the validated seizure questionnaire (Reutens et al., 1992) were conducted with affected relatives, and an eyewitness account of their seizures was obtained. Where possible, the mother of each affected individual was interviewed. In addition, a family history of speech and language impairments was noted. Medical records and all relevant investigations, including EEG and neuroimaging results, were obtained. Each individual was phenotyped according to the International League Against Epilepsy (ILAE) organization (Berg et al., 2010). Where insufficient data were available, the individual was labeled as unclassified. Asymptomatic individuals with CTS were regarded as unaffected.

Data analysis

For each degree of relatedness, the proportion of affected individuals was determined by calculating the number of affected individuals divided by the total number of individuals. We compared the proportion of affected relatives of probands with EAS with those of probands with BECTS (Vears et al., 2012). Chi-square test was used for comparison between different groups.

All participants, or their parent or legal guardian in the case of minors, provided written informed consent. This study was approved by the Austin Health Human Research Ethics Committee.

Results

Probands

Thirty-one unrelated probands were recruited: five had ECSWS, three had LKS, one had ABPE, and 22 had IEAD. A predominance of boys was observed (21/31, 68%). The mean age of seizure onset in the 26 cases with seizures was 3.9 years (range 6 months–7 years). The mean age of deterioration in speech in the remaining five cases was 3.6 years (range 1.5–6 years). The following seizure types occurred: febrile seizures (5 patients), rolandic seizures (12), focal dyscognitive seizures (11), tonic–clonic seizures (13), and negative myoclonic or atonic seizures (5). All probands had abnormal EEG studies with typical centrotemporal epileptiform discharges activated by sleep. In addition, seven had frequent multifocal sharp waves and eight had CSWS on EEG (clinically they had either ECSWS or LKS).

Cognitive regression, predominantly in language domains, occurred in 23 children, whereas 8 showed plateauing in their development. This occurred on a background of previous normal development in 16 and delayed development in 15 children. Improvement in cognitive function was observed with antiepileptic medication in 15 and steroids treatment in 9 patients. In terms of final outcome, intellectual disability was mild in seven, moderate in six, and severe in six children; the remaining 12 children were of normal intellect. Autistic features were noted in 10 probands and obsessive-compulsive traits in 6.

Brain MRI was performed on 29 probands; two could not lie still due to their intellectual disability. One patient had hippocampal sclerosis. Twenty-eight patients had either normal imaging or nonspecific changes such as small foci of high T2 signal in the white matter (two patients), Chiari I malformation (1), a congenital small temporal lobe on the side contralateral to that of predominant epileptiform activity (1), and mild bilateral gray–white matter junction blurring of the temporal poles (1).

Family study

For 31 probands, pedigrees were obtained describing up to three degrees of relatedness (Fig. 1). Overall, 16 (51.6%) of 31 probands had a positive family history of seizures. Three families showed bilineal inheritance of seizure disorders. Of the total 1,254 first-, second-, and third-degree relatives, 30 (2.4%) had seizures, including 19 (1.5%) with epilepsy and 13 (1%) with febrile seizures (Table 1). Two relatives had both febrile seizures and epilepsy, including IEAD in one and focal epilepsy in the other. Two IEAD and two ECSWS cases have family members with BECTS and so were previously included in our BECTS family study (Families A, J, L, M) (Vears et al., 2012).

Table 1. Family history of seizures in probands with EAS
ProbandsNFirst degree (%)p-ValueSecond degree (%)p-ValueThird degree (%)p-ValueTotal (%)p-Value
  1. NS, not significant.

EAS cohort3113/128 (10.2) 5/291 (1.7) 12/835 (1.4) 30/1,254 (2.4) 
BECTS cohort5321/214 (9.8)NS15/494 (3.0)NS21/1,377 (1.5)NS57/2,085 (2.7)NS
Figure 1.

Pedigrees of EAS probands with positive family history of seizures (n = 16). FS, febrile seizures; FS+, febrile seizures plus; BECTS, benign childhood epilepsy with centrotemporal spikes; ECSWS, epileptic encephalopathy with continuous spike-and-wave during sleep; IEAD, intermediate epilepsy-aphasia disorders; GGE, genetic generalized epilepsy; CTS, centrotemporal spikes.

Considering each degree of relatedness, 6 (4.7%) of 128 first-degree relatives, 3 (1%) of 291 second-degree relatives, and 10 (1.2%) of 835 third-degree relatives had epilepsy. Eight (6.3%) of 128 first-degree relatives, 2 (0.7%) of 291 second-degree relatives, and 3 (0.4%) of 835 third-degree relatives had febrile seizures (Table 2).

Table 2. Epilepsy syndromes in each degree relatives of EAS probands
SyndromesFirst-degree relatives (n = 128)Second-degree relatives (n = 291)Third-degree relatives (n = 835)Total (n = 1,254)
  1. a

    One first-degree relative had EAS + FS and one third-degree relative had focal epilepsy (occipital lobe epilepsy) +FS.

Febrile seizuresa82313
Epilepsy631019
BECTS2024
IEAD3003
Epilepsy with focal seizures of unknown cause1124
GGE0011
Unclassified epilepsy0257
Total1351230

With regard to speech and language disorders, in 7 (23%) of 31 probands a positive family history was reported. Among 1,254 first-, second-, and third-degree relatives, 9 had speech and language disorders: 4 (3.1%) of 128 first-degree, 2 (0.7%) of 291 second-degree, and 3 (0.4%) of 835 third-degree relatives. This was significantly higher than the frequencies of our previous BECTS cohort, where only 4 (0.2%) of 2,085 of first-, second-, and third-degree relatives reported speech and language disorders (p = 0.02, Vears DF, Tsai MH, Sadleir LG, Grinton BE, Lillywhite LM, Carney PW, Harvey AS, Berkovic SF, Scheffer IE, unpublished data).

Epileptic syndromes in relatives of probands with EAS

Of the six first-degree relatives with epilepsy, three had IEAD, including a monozygous twin pair (Fig. 1 Pedigree F) (Table 2). Another two had BECTS, and one had unclassified focal epilepsy. In the second- and third-degree relatives, there were three individuals with unclassified focal epilepsy, two with BECTS, and one with genetic generalized epilepsy (GGE). Two second- and five third-degree relatives were regarded as unclassified epilepsy, as insufficient information was available.

Comparison between EAS and BECTS cohorts

When we compared these data to our previous study of BECTS (Vears et al., 2012), we found that there was no significant difference in the proportion of probands with a positive family history of seizures (Table 1). The proportion of affected relatives in total, or for each degree of relatedness, showed no significant difference in terms of epilepsy or febrile seizures (p > 0.05). No differences emerged when the four families with both BECTS and EAS were excluded from the analysis (data not shown).

Discussion

We investigated the family history of seizure disorders in patients with epilepsies that fall along the epilepsy-aphasia spectrum (EAS), including LKS and ECSWS. We found that the proportion of affected individuals decreased rapidly from one degree to three degrees of relatedness. The seizure frequencies in the relatives and pattern of affected relatives were similar in both EAS and BECTS cohorts.

The epilepsy-aphasia spectrum

Two well-established syndromes, LKS and ECSWS, encompass language regression, sometimes in the setting of more global regression, together with electrical status epilepticus in sleep EEG recordings. In clinical practice, however, patients may not fit within the criteria accepted for these classical syndromes. They may have milder features in terms of seizures, EEG manifestations, and cognitive sequelae. The EEG studies in our IEAD patients showed CTS with sleep activation but do not fulfill the diagnostic EEG criterion for CSWS in which epileptiform activity occupies at least 85% of NREM sleep (Tassinari et al., 2000). Despite this well-accepted strict definition, some authors have adopted less restrictive criteria that suggest 25–60% epileptiform activity in sleep satisfies a diagnosis of CSWS (Beaumanoir, 1995; Inutsuka et al., 2006; Van Hirtum-Das et al., 2006; Scheltens-de Boer, 2009). Hence, this has resulted in confusion and debate as to whether the cases with epileptiform activity occupying <85% sleep are variants of the same disorder as ECSWS (McVicar & Shinnar, 2004; Nickels & Wirrell, 2008). Herein we regard the EAS as a spectrum that extends from BECTS at the mild end, to include ECSWS and LKS at the severe end (Fig. 2).

Figure 2.

The epilepsy-aphasia spectrum. The spectrum contains disorders ranging in severity in terms of epilepsy, EEG, and intellectual outcome. At the mild end of the spectrum is BECTS with mild epilepsy and minimal language dysfunction. The severe end includes Landau-Kleffner syndrome and epileptic encephalopathy with continuous spike-and-wave during sleep. Some patients have intermediate epilepsy-aphasia disorders.

Seizures often occur in patients with EAS, with rolandic seizures being the most common. In some cases, other seizure types including focal dyscognitive, negative myoclonic, atonic, and bilateral convulsive seizures are also observed. It is common for these patients to have been misdiagnosed as “symptomatic generalized epilepsy” or “Lennox-Gastaut syndrome” because of the presence of abundant bilateral epileptiform discharges in sleep on EEG; however, none had tonic seizures or generalized paroxysmal fast activity on EEG.

The hallmark of EAS is the association of cognitive regression or plateauing with sleep activation of epileptiform discharges. Typically, ECSWS presents with global regression which, in one third of cases occurs on a background of preexisting developmental delay (Tassinari et al., 2000; Nickels & Wirrell, 2008). In contrast, LKS is characterized by more restricted language regression, characteristically beginning with verbal agnosia. In our cohort of 31 EAS patients, preexisting developmental delay followed by cognitive regression or plateauing occurred in 15 cases. Improvement in cognition was observed in 24 (77%) of 31 patients with antiepileptic medications with or without steroids.

Previous studies of the genetics of EAS

Initially genetic factors were not considered relevant to ECSWS and LKS; however, more recently, the concept of shared underlying genetic factors for BECTS, ECSWS, and LKS has started to emerge (Beaumanoir, 1992; Tassinari et al., 2000; Doose et al., 2001; Nickels & Wirrell, 2008; Rudolf et al., 2009). Saltik et al. (2005) compared BECTS patients with and without later evolution to ECSWS and found a family history of epilepsy was three times higher in the ECSWS group, supporting the influence of genetic factors. Further evidence that genetic components have a more significant role in EAS than classic BECTS has been drawn from twin and family studies (Doose et al., 2001; Vadlamudi et al., 2006).

Only two monozygotic twin pairs with EAS have been reported. Of interest, one pair was concordant for ECSWS, whereas another with LKS was discordant with the unaffected twin having a normal EEG (Feekery et al., 1993; Blennow & Ors, 1995). Two sibling pairs with LKS, including one pair from the original paper by Landau and Kleffner, have also been reported (Landau & Kleffner, 1957; Nakano et al., 1989). In addition, in four families whose probands had ECSWS, two fathers had BECTS, one sibling had ABPE, and one sibling had developmental dysphasia with CSWS without seizures (De Tiege et al., 2006; Praline et al., 2006; Saadeldin & Al-Tala, 2011).

Anecdotal reports of “rolandic epilepsy” with chromosomal deletions (7q, 16p13, 1qter) have been described, although our criteria would suggest that these cases had EAS (Vaughn et al., 1996; Burke et al., 1997; Reutlinger et al., 2010). A mutation in SERPINI1 was found in one patient who had ECSWS (Coutelier et al., 2008). Rare families with dominant inheritance bearing phenotypic similarities to EAS have been reported, although no linkage or causative variants have been identified (Scheffer et al., 1995; Kugler et al., 2008; Michelucci et al., 2008). A mutation in the SRPX2 gene was identified in a French family where affected individuals had rolandic seizures, oral and speech dyspraxia, as well as intellectual disability (Roll et al., 2006).

Clinical genetics of EAS

The genetic architecture of EAS has not been systematically explored. This may be partially due to poor recognition of the EAS as well as the rarity of this group of disorders which are often regarded as “symptomatic” despite the majority having normal magnetic resonance imaging (MRI) studies.

We examined the pedigrees of 31 probands with EAS to three degrees of relatedness. Sixteen probands had a positive family history of epilepsy or febrile seizures. Previous estimates on reviewing the literature have suggested that a family history of seizures is present in 12% of patients with LKS and 15% with ECSWS, but the nature of the family history has not been well characterized (Beaumanoir, 1992; Tassinari et al., 2000).

Our approach differs from those of earlier studies in the systematic characterization of the seizure disorders in all relatives out to three degrees of relatedness. We found that 10.2% of first-degree relatives have seizures or epilepsy, which decreases to 1.7% and 1.3% in the second- and third-degree relatives, respectively. These frequencies are consistent with a complex inheritance pattern due to the involvement of multiple genes possibly with an environmental contribution. There was, however, one family (L) that could potentially have an autosomal dominant inheritance pattern, and three families showed bilineal inheritance of seizures. It is possible that, in some families, genes of major effect are responsible, whereas in others, multiple genes of lesser effect are involved.

We observed a lower frequency of febrile seizures in the second- and third-degree relatives in our families than in the general population. Given the fact that the second- and third-degree relatives were drawn predominantly from the older generations, this could be due to age-related recall bias where an accurate early history may not be available (Ottman et al., 1995). Although under-ascertainment of affected relatives may have occurred, these frequencies are similar to our previous BECTS study and the findings of focal epilepsy overall (Ottman, 1989; Abou-Khalil et al., 2007; Vears et al., 2012).

Of the epileptic syndromes in the relatives, the most common phenotype was FS in 12 (40%) of 30 relatives. Of interest, the second most common presentations in relatives were BECTS in four and epilepsy with focal seizures in another four. This was followed by three relatives with IEAD. Therefore, the picture emerging in families of individuals with EAS is of a predisposition to focal seizures as well as FS.

When we compared our EAS and BECTS families using the same methodology, we found no difference in the proportions of affected relatives with seizures between the two cohorts (Vears et al., 2012). Moreover, the patterns of epilepsies in relatives were similar, with a clear predilection for focal rather than generalized epilepsies. This suggests that the genetic architecture may be similar in both EAS and BECTS and that there are likely to be some shared genetic determinants.

The high proportion of speech and language disorders reported in the relatives of EAS probands is of interest. This is similar to the findings of Lesca et al. (2012) where they found that 45% of 61 patients with either ECSWS or LKS had a family history of febrile seizures, epilepsy including ECSWS or LKS, cognitive, speech, language, and reading impairments. Their key observation of copy number variants in 74% patients (46/61) highlights the likely genetic contribution to these complex and overlapping phenotypes. Moreover, many of the copy number variants included genes associated with autism spectrum disorders or specific language impairment. These observations support the hypothesis of a genetic predisposition affecting brain development that involves several networks (e.g., epilepsy, language, and autism) (Doose et al., 2000; Lesca et al., 2012). Further work examining the speech and language phenotypes in families with EAS is warranted.

EAS is a complex group of disorders for which little is known with regard to etiology and underlying mechanisms. It is likely that these represent network disorders incorporating speech and language networks influenced by frequent epileptiform activity. Aggressive early treatment may have the potential to improve long-term cognitive outcome. Our study demonstrates that the genetic architecture of EAS is likely to be complex. Identification of genes that contribute to these disorders will shed light on the underlying biologic mechanisms of the EAS.

Acknowledgment

The study was supported by National Health and Medical Research Council, Australia. The authors thank the patients and their families for participating in the research.

Disclosure

None of the authors has any conflicts of interest. 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