What this paper adds
- • Normal sleep EEG does not exclude Landau–Kleffner syndrome.
Aim We report three cases of Landau–Kleffner syndrome (LKS) in children (two females, one male) in whom diagnosis was delayed because the sleep electroencephalography (EEG) was initially normal.
Method Case histories including EEG, positron emission tomography findings, and long-term outcome were reviewed.
Results Auditory agnosia occurred between the age of 2 years and 3 years 6 months, after a period of normal language development. Initial awake and sleep EEG, recorded weeks to months after the onset of language regression, during a nap period in two cases and during a full night of sleep in the third case, was normal. Repeat EEG between 2 months and 2 years later showed epileptiform discharges during wakefulness and strongly activated by sleep, with a pattern of continuous spike-waves during slow-wave sleep in two patients. Patients were diagnosed with LKS and treated with various antiepileptic regimens, including corticosteroids. One patient in whom EEG became normal on hydrocortisone is making significant recovery. The other two patients did not exhibit a sustained response to treatment and remained severely impaired.
Interpretation Sleep EEG may be normal in the early phase of acquired auditory agnosia. EEG should be repeated frequently in individuals in whom a firm clinical diagnosis is made to facilitate early treatment.
Continuous spike–waves during slow-wave sleep
Interictal epileptiform discharge
Positron emission tomography
Landau–Kleffner syndrome (LKS) is an epileptic syndrome of childhood recognized by the International League Against Epilepsy, and is synonymous with acquired epileptic aphasia.1 The syndrome was first described by Landau and Kleffner in six children with normal early language development who became aphasic after the onset of subtle seizures.2 The aphasia was further characterized as auditory verbal agnosia, i.e. the inability to decode phonemes despite intact peripheral hearing mechanisms, leading to a severe receptive and expressive verbal deficit.3 Further observations showed that language regression may occur in the absence of clinical seizures.4–7
In their original paper, Landau and Kleffner reported that affected individuals exhibited bilateral epileptiform discharges on electroencephalography (EEG) that were prominent in the temporal regions.2 In 1986, Aicardi8 stated that EEG abnormalities are present, by definition, in all cases, but may be highly variable from one recording to another: patients have interictal epileptiform discharges (IEDs) that are generally multifocal and bilateral, often with a predominance over the temporal and parietal regions. This opinion was repeated in other review articles,6,9,10 and by the International League Against Epilepsy, which defined LKS as a childhood disorder characterized by acquired aphasia, multifocal spikes, and spike and wave discharges.1 Non-rapid eye movement sleep activates IEDs in LKS. A pattern of continuous spike–waves during slow-wave sleep (CSWS; i.e. diffuse spike–waves occurring during at least 85% of slow sleep6) was found in the course of the disease in all patients reported in some series.4,11–13 In a longitudinal EEG analysis, it was found that at epilepsy onset, but before language regression, patients exhibited bilateral synchronous IEDs occupying from 30 to 90% of slow-wave sleep, and subsequently showed the CSWS pattern at some time during the follow-up, with marked fluctuations.14
LKS is now classified among the epileptic encephalopathies, i.e. conditions in which epileptiform abnormalities may contribute to progressive cognitive dysfunction.15 it is considered part of the spectrum of idiopathic focal epilepsies with CSWS, having in common cognitive regression and strong activation and spread of focal IEDs during sleep.6,9,10 However, the presence of CSWS is not a prerequisite for a diagnosis of LKS. For instance, in one series of 25 patients with LKS, the CSWS pattern was present in only nine patients.16 We report here three children in whom sleep EEG, performed several months after onset of language regression, showed normal findings. A diagnosis of LKS was made when subsequent sleep EEG showed IEDs. These cases highlight that normal sleep EEG does not rule out LKS, and reactivate the debate about the pathophysiology of language regression in this condition.
Patients were retrospectively selected from the databases of the departments of Paediatric Neurology of Erasme University Hospital of Brussels (patient 1: male, age at onset of language regression 3y 6mo; patient 2: female, age at onset 2y 6mo) and the Children’s University Hospital Temple Street, Dublin (patient 3: female, age at onset 2y 3mo), according to the following criteria: (1) regression of language in childhood over a period of a few weeks leading to auditory agnosia, (2) normal structural cerebral magnetic resonance imaging, (3) normal peripheral hearing as assessed by auditory evoked potentials, (4) first sleep EEG showing the absence of IEDs, and (5) subsequent EEG showing the presence of IEDs, leading to a diagnosis of LKS.
Only three patients fulfilled these criteria over the last 20 years. Parents signed informed consent for the publication of their children’s data. Table I summarizes the relevant clinical and EEG data of the three patients.
|1 (Male)||2 (Female)||3 (Female)|
|Age at onset of language regression||3y 6mo||2y 6mo||2y 3mo|
|Age at first sleep EEG (duration: results)||4y (overnight: normal)||3y 9mo (nap: normal)||4y 3mo (nap: normal)|
|Non-verbal IQ (Leiter International Performance Scale), age, result||4y 8mo, 93||3y 9mo, 110||5y 5mo, 111|
|Age at first sleep EEG showing IEDs (results)||4y 9mo (R and L rolandic IEDs awake, CSWS)||6y (R temporal IEDs awake, CSWS)||4y 5mo (L temporal and R occipital, no CSWS)|
|Drug trials||Hydrocortisone (21m cure)||Dexamethasone (1m cure), and various conventional AED||Corticosteroids (6wks cure) and various conventional AED|
|EEG evolution under treatment||Complete and sustained normalization after 1m treatment||No significant changes||Initial normalization, then reappearance of IEDs and CSWS pattern (5y 2mo)|
|Epileptic seizures||None||A few GTCS between ages 6 and 7y||Three focal seizures between ages 7 and 9y|
|Long-term outcome of auditory agnosia||Partial recovery||No recovery||No recovery|
The first normal sleep EEG was recorded either overnight (patient 1, illustrated in Fig. 1a) or during a nap in which only sleep stages 1 and 2 were attained (patients 2 and 3) and the findings were analysed by one of the co-authors (JD for patient 1, CW for patient 2, and MDK for patient 3).
The delay between the first normal sleep EEG and a second EEG recording showing IEDs varied from 2 months to 2 years 3 months. Notably, a CSWS pattern on the second EEG was recorded in only two of the three patients (see Fig. 1b for patient 1), although a CSWS pattern was documented later in the evolution of the disease in the third patient.
In all three cases, personal histories were unremarkable before the onset of language regression. In particular, epileptic seizures were not documented and there was normal acquisition of verbal skills. There was no family history of epilepsy.
In all cases, an investigation of regional glucose distribution was performed using positron emission tomography (PET) with 18-F-fluorodeoxyglucose (FDG). In patient 1, PET was carried out at age 4 years 8 months, in a period where CSWS were present, and showed regional hypermetabolism in the left upper temporal neocortical area and hypometabolism in mesial temporal regions and in the cingulate cortex. In patients 2 and 3, PET investigations took place during a period without CSWS on EEG (at age 3y 9mo for patient 2 and 6y 9mo for patient 3) and showed either right temporal hypometabolism (patient 2) or a normal pattern (patient 3).
All three patients received corticosteroids during the course of their disease. They induced sustained sleep EEG normalization only in patient 1, who was treated with a prolonged course of hydrocortisone started at 5mg/kg/day, following the regimen published by Buzatu et al.17 Patients 2 and 3 did not show a sustained EEG response to either corticosteroids or conventional antiepileptic drugs; however, there was a long interval between onset of aphasia and administration of corticosteroids, which were given for short periods and poorly documented for patient 2. Moreover, fluctuations between a normal EEG and a pattern of CSWS were observed later in the course of the disease in patient 3. Indeed, this patient was assessed at 6 years 9 months for possible subpial transections because of the absence of language recovery and CSWS on previously recorded EEG, but two attempts to elicit the CSWS pattern under anaesthetic in order to perform the amytal study were unsuccessful, and EEG recording during natural sleep was normal at that time.
Outcome for language was noticeably different in the three patients. Patient 1 showed gradual, albeit incomplete, language recovery. At the age of 6 years 8 months, recognition of non-verbal sounds and auditory attention continue to improve, but comprehension of spoken language remains inferior to that of signed language. The patient is able to repeat phonemes and simple syllables and, expressively, he communicates by combining signs, gestures, imitation, intonation, and utterances. On the other hand, patients 2 and 3 have persisting severe auditory verbal agnosia and have learned sign language. At 25 years of age, patient 2 communicates by signing or by using pictograms. She is able to identify non-verbal sounds but is unable to understand verbal messages. Learning abilities and level of signing remain weak. She is socially well integrated but is unable to live independently. Patient 3 experienced some return of limited verbal understanding at age 15 years and was transferred to a mainstream school. At 22 years of age she is able to live independently.
The three patients reported here presented with a history of language regression and verbal auditory agnosia at the age of 2 to 3 years, consistent with LKS, although earlier than the peak age at onset of between 5 and 7 years.6 There was a delay in diagnosis because the first sleep EEG, performed several months after onset of regression when the patients were profoundly aphasic, was normal. The finding of IEDs on subsequent EEG (including a pattern of CSWS in two cases) led to the diagnosis of LKS.
Evolution was variable among these three patients. Patient 1 showed rapid, complete, and sustained remission of EEG abnormalities after a prolonged course of hydrocortisone commenced more than 1 year after onset of regression, and is gradually recovering language. The other two patients remained severely aphasic despite several trials of antiepileptic drugs and short courses of corticosteroids.
The initial normal EEG in the three patients with LKS may have two explanations. The first is that EEG abnormalities may appear long after the onset of language regression in LKS and are thus an epiphenomenon, as proposed by some authors.18 The second explanation is that these cases are extreme examples of the well-known phenomenon of fluctuating EEG abnormalities previously described in LKS patients,8,13 and that IEDs were not recorded because of sampling limitations. It could be argued that, as only sleep stages 1 and 2 were recorded during a nap in the first EEG of two of our three patients, an overnight recording would have captured IEDs. However, this assumption remains purely speculative, as nap-time recordings were the same as those obtained during full nocturnal sleep in one LKS series.13 We consider fluctuation of the EEG abnormalities as the most likely explanation for the following reasons. First, fluctuation between normal EEG and CSWS was documented in patient 3 not only at onset but also later in the course of the disease when the patient was evaluated for surgery. Secondly, the patient who had the most sustained EEG response to treatment (patient 1) showed the greatest recovery in language. This observation supports the hypothesis that the epileptic discharges activated by sleep play a major role in the language, cognitive, and behavioural deficits seen in these children, a finding previously reported.4,6,9,12 The reason for the day-to-day variation in frequency of IEDs remains unknown.
These three patients had FDG-PET investigations showing heterogeneous results: focal temporal hypermetabolism (patient 1), focal temporal hypometabolism (patient 2), and absence of any regional metabolic change (patient 3). Such heterogeneity in PET findings is consistent with previous reports on FDG-PET studies in LKS and other idiopathic focal epilepsies with CSWS.19,20 Focal hypermetabolism supports the epileptic hypothesis of language regression. Indeed, both animal models and previous PET studies performed in children with CSWS found that focal cortical hypermetabolism is related to intense interictal spiking and induces inhibition of remote connected areas which become hypometabolic.20,21 As PET studies were performed during a period when sleep EEG was relatively free of IEDs in two patients, we propose that the absence of focal hypermetabolism was related to the absence of intense interictal spiking, and that a normal metabolic pattern was consistent with a state of recovery from CSWS, as already described in other patients.22
Another important issue raised by these case reports is whether pure auditory agnosia acquired during childhood may occur in conditions other than LKS. Auditory agnosia is believed to result from bilateral dysfunction of primary auditory cortical areas. When acquired, it could theoretically result from vascular, traumatic, or infectious insults. The bilateral opercular syndrome with pseudobulbar palsy following encephalitis is well described.23 One case of verbal auditory agnosia was reported to be a sequel to herpetic encephalitis, but this child had bilateral upper temporal lesions on imaging.24 In a series of 149 children who underwent long-term EEG for language regression of unknown aetiology (i.e. without brain lesion or encephalitis, etc.), 15 patients had language regression without autistic features and EEG findings were normal.16 The authors did not describe the type of language deficit in detail or the final diagnosis in this subgroup of patients, and so it is possible that some had LKS, undiagnosed because of the normal EEG. Stefanatos et al.25 reported on a 6-year-old male with language regression at age 22 months and a picture of auditory verbal agnosia associated with autistic-like behaviour. A 24-hour EEG recording was normal but single-photon emission computed tomography showed bilateral temporal hypoperfusion, suggesting the possibility of LKS. Treatment with steroids resulted in marked improvement. It is possible that such cases are more common than recognized, particularly if the patients do not develop seizures or are lost to neurological follow-up.
In conclusion, we propose that LKS is the most likely diagnosis when auditory agnosia occurs in a child following a period of normal language development, and that an initial normal sleep EEG does not exclude this diagnosis. Therefore, we recommend a repeat EEG, preferably a full overnight sleep recording, within a few weeks if the first EEG is normal. If the clinical features are very suggestive of LKS, the diagnosis should be considered even in the absence of an abnormal EEG. The same approach should be applied when relapse occurs in a child already diagnosed with LKS.