Clinical manifestations

Presentation and course

Seizure onset usually occurs in the second decade of life with a mean age of onset of 11 to 14 years, but with a very wide range of onset from 2 months to 64 years. The only mandatory core diagnostic criteria for sleep-related hypermotor (hyperkinetic) epilepsy is brief focal motor seizures with hyperkinetic or asymmetric tonic/dystonic features occurring predominantly from sleep. A mean of 20 seizures per month is reported, with 1 to 20 seizures occurring per night. Most individuals have no identifiable trigger for their seizures. Approximately one third of patients have occasional daytime seizures, and for a similar proportion, focal seizures evolve to bilateral tonic-clonic seizures, both usually occurring during periods of poor seizure control. A significant number of individuals with sleep-related hypermotor epilepsy also report a personal or family history of events that are clinically consistent with benign parasomnias such as sleepwalking and sleep terrors. Some patients with sleep-related hypermotor epilepsy may complain only of sleep disturbance, and others may be unaware of their events, with the seizures only being reported by relatives. In some cases of “sporadic sleep-related hypermotor epilepsy,” a family history of nocturnal events may be suggestive of a diagnosis of ADSHE if family members are also experiencing focal seizures that may not have been recognized.

Exclusionary criteria for the diagnosis of sleep-related hypermotor epilepsy include seizures only during wakefulness, generalized onset seizures, and age of onset younger than 2 months or over 64 years of age.

Focal motor seizures with vigorous hyperkinetic or asymmetric tonic/dystonic features are seen, usually with autonomic signs (tachycardia, tachypnea, irregular respiratory rhythm), vocalization, and negative emotional expression, such as fear. There may be head and eye deviation. Hyperkinetic movements involve proximal limb or axial muscles, producing irregular large-amplitude movements, such as pedaling, pelvic thrusting, jumping, thrashing, or rocking movements. Focal motor seizures may be clinically subtle (previously termed "paroxysmal arousals") or may have longer duration and greater complexity (such as "epileptic wandering"). Patients may describe a focal aware sensory or cognitive seizure before the motor features commence. Focal to bilateral tonic–clonic seizures can occur. Although occurrence of seizures from sleep is characteristic of this syndrome, seizures from the awake state occur in 27% to 45% of patients at some time in their life.

Neurologic examination is usually normal. Neuroimaging studies may be normal or show a focal cortical dysplasia or acquired structural abnormality. The interictal EEG tends to show a normal background although focal (usually frontal) epileptiform abnormalities can be seen. When patients with sleep-related hypermotor epilepsy are studied using video EEG monitoring, by far the most common seizure type recorded is the paroxysmal arousal; paroxysmal nocturnal dystonia is seen less frequently, with episodic nocturnal wanderings recorded infrequently. Most patients have at least two of these seizure types identifiable on monitoring, and autonomic features such as tachycardia and irregular respiration are also prominent.

Prognosis and complications

The prognosis of sleep-related hypermotor epilepsy is predominantly related to the underlying etiology. Most patients have normal intellect and normal imaging and respond to first-line antiseizure medications. However, intellectual disability and behavior disorders have been reported in sleep-related hypermotor epilepsy, especially in the familial form related to mutations in KCNT1. Patients with intellectual disability, neurologic or imaging abnormality, or seizures in wakefulness are less likely to achieve sustained seizure remission. Epilepsy surgery may be effective in selected etiologies. The best surgical outcome is seen when the etiology is a well-defined structural pathology, especially focal cortical dysplasia type IIb.

However, it is increasingly recognized that in some cases, sleep-related hypermotor epilepsy may have a more severe presentation. Some individuals have highly refractory epilepsy throughout their lives, with periods of status epilepticus. They may also have significant psychiatric and cognitive morbidity (34; 16). Moreover, evidence is accumulating that even in relatively mildly affected individuals, neuropsychological abnormalities of frontal lobe function may be identified on formal testing (56; 81). In a cohort of 60 individuals with sleep-related hypermotor epilepsy, 15% had intellectual disability or cognitive deterioration and nearly half had deficits in at least one test of cognitive function (30).

As discussed earlier, the clinical features of sleep-related hypermotor epilepsy in most families with different mutations are indistinguishable. There is some evidence, however, to suggest that phenotypic features such as severity of epilepsy, and cognitive and psychiatric morbidity, may be influenced by the specific underlying genetic mutation. Steinlein and colleagues analyzed the clinical features of 150 affected individuals from 19 families in 12 countries and found that certain nAChR mutations appeared to confer a higher risk for mental retardation, schizophrenia-like symptoms, and marked cognitive deficits (69). For example, only 2 of 67 identified individuals with the CHRNA4-S248F mutation had major psychiatric or neurologic features, and only minor neuropsychological deficits were found on formal testing. In contrast, 11 of the 19 known individuals with the CHRNA4-S252L mutation had intellectual capacities in the low normal or moderately intellectually disabled range, and most individuals developed epilepsy at a younger age (6 months to 2 years). However, the small numbers involved in such studies, along with multiple other potential confounding factors, make definitive conclusions difficult.

Clinical vignette

A 33-year-old man presented with nocturnal attacks from the age of 10 years. The events were characterized by waking from sleep with a sensation of his breath “locking in his throat,” feeling unable to breathe. The episodes would usually happen soon after he had fallen asleep (within 30 minutes to an hour) and would often occur 5 to 10 times in close succession on the same night over a few hours. He felt he had partial or full recollection of all the events. He remembered sitting forward during the attacks and sometimes shouting “help.” His mother described how he would suddenly sit forward, with a distressed facial expression and apparently hyperventilating. His legs would kick and his arms would be held stiffly out to his sides, and he would shout. The attacks usually lasted 30 seconds to a minute, but occasionally they would last up to 2 minutes. They occurred exclusively from sleep.

The patient had a strong family history of similar sleep-related events; his father and brother had been fully investigated and diagnosed with nocturnal epilepsy on the basis of video-EEG monitoring. His father had suffered frequent and troublesome nocturnal events, which started in early childhood, and which were diagnosed as night terrors and psychogenic attacks for many years. His grandmother also had a history of troublesome sleepwalking in her early life, which had settled spontaneously soon after the birth of her first child.

Standard investigations, including MRI and EEG (during sleep and wakefulness), were normal. Video-EEG monitoring captured several events consistent with frontal lobe seizures, characterized by sitting forward with dystonic posturing of the upper limbs and hypermotor bipedal automatisms of the legs. EEG during these episodes was significantly marred by movement artefact, but bifrontal rhythmic sharp and slow activity was clearly present during some of the attacks.

He commenced on carbamazepine, which reduced his seizure frequency, but the episodes still occurred during periods of sleep deprivation or stress. However, during his twenties, the episodes became gradually less frequent and eventually stopped without further changes to his medication regimen at 23 years. He remained seizure-free on carbamazepine monotherapy.

Biological basis

Etiology and pathogenesis

The etiology of sleep-related hypermotor (hyperkinetic) epilepsy may be genetic, genetic–structural, or acquired. In a majority of patients, the etiology is unknown. No specific clinical features distinguish etiologies.

Familial sleep-related hypermotor epilepsy is usually inherited in an autosomal dominant fashion (autosomal dominant sleep-related hypermotor epilepsy, ADSHE), with a penetrance of approximately 70%. A pathogenic gene variant is found in approximately 19% of ADSHE and in 7% of sporadic sleep-related hypermotor epilepsy. Genetic causes of ADSHE include pathogenic variants in GATOR1 complex genes (DEPDC5, less frequently NPRL2 or NPRL3), in acetylcholine receptor subunit genes (CHRNA4, less frequently CHRNB2 or CHRNA2), and in the sodium-activated potassium channel gene KCNT1. Individuals with GATOR complex pathogenic gene variants may have focal cortical dysplasias, with implications for epilepsy surgery. Individuals with KCNT1 pathogenic variants have a more severe form of sleep-related hypermotor epilepsy, with intellectual disability, psychosis, and sometimes regression, and higher penetrance in families. Rare families with autosomal recessive sleep-related hypermotor epilepsy are described, and pathogenic variants in PRIMA1 have been identified in one family.

nAChR gene mutations. Mutations in CHRNA4, CHRNB2, and CHRNA2 can cause ADSHE (70; 51; 01) and are estimated to account for 10% to 20% of families in some populations (48). Penetrance is estimated at 70% in these families. Functional studies of the nicotinic receptor mutations associated with ADSHE demonstrate that result is gain-of-function, sometimes with altered response to antiepileptic medications and acetylcholine (10; 23; 24; 69). The importance of cholinergic signaling in sleep-related hypermotor epilepsy has also been emphasized by the finding of homozygous mutation in the gene PRIMA1 (which encodes a protein that anchors acetylcholinesterase to the neuronal membrane) in a family with autosomal recessive sleep-related hypermotor epilepsy (22).

Nevertheless, the mechanism by which mutations in the nAChR subunit genes result in epilepsy is not clear. The nAChRs are important ligand-gated ion channels that are distributed widely throughout the central nervous system, although the function of these receptors is only partially understood. There is good evidence that they are involved in presynaptic modulation of neurotransmitter release in a number of systems, enhancing the release of norepinephrine, dopamine, GABA, serotonin, and acetylcholine (61). In addition, there may be direct nicotinic synaptic neurotransmission in the CNS (although there is relatively little evidence of this), and the nAChR is believed to play a role in the regulation of gene expression and neuronal pathfinding during development (61). Positron emission tomography (PET) studies have suggested that this developmental function may be important in ADSHE, possibly via influences on dopamine receptor expression (19) or through changes in regional nAChR density (55).

KCNT1 mutations. The discovery of missense mutations in the sodium-gated potassium channel gene KCNT1 in families with severe ADSHE (21), and a sporadic case with a de novo mutation, demonstrated that other genetic mechanisms can also cause ADSHE. Mutations in this gene appear to be associated with a fully penetrant, often more severe form of ADSHE associated with behavioral and psychiatric problems and intellectual disability, but with otherwise typical seizure semiology. Psychiatric features include aggression and psychosis; severe developmental regression is also observed (16). Intriguingly, de novo KCNT1 mutations have also been identified in the syndrome of epilepsy of infancy with migrating focal seizures (EIMFS), a severe epileptic encephalopathy associated with migrating focal seizures and profound developmental impairment (04). A family was comprised of four children with two sets of half-siblings, two half-siblings with ADSHE and two half-siblings with EIMFS inherited from their mildly affected mother with ADSHE (39). Functional studies of the mutations causing EIMFS are consistent with a gain of function resulting in hyperexcitability (04), though there is a single report of a child with EIMFS and a KCNT1 loss of function mutation with decreased membrane expression (18). KCNT1 mutations causing ADSHE show gain of function but less than that observed in EIMFS (38). The precise function of KCNT1 and how mutations in this gene result in these distinct epilepsy phenotypes is, however, poorly understood.

mTOR pathway gene mutations. In 2013, the gene underlying familial focal epilepsy with variable foci (FFEVF) was discovered to be DEPDC5 (67; 25; 17). FFEVF is characterized by autosomal dominant inheritance of focal epilepsies in which different family members have focal epilepsies emanating from different cortical regions (67). DEPDC5 mutations were identified in seven of the eight large families reported (82; 11; 09; 42; 17). When small families with two or more individuals with focal epilepsy were studied for DEPDC5 mutations, 12% (10 out of 82) had mutations. Some families only contained individuals with sleep-related hypermotor epilepsy and could, therefore, have been diagnosed with ADSHE. These families effectively represent a subset of FFEVF in which some families only express a sleep-related hypermotor epilepsy phenotype. Many of the DEPDC5 mutations produce truncation of the protein and consequent haploinsufficiency.

DEPDC5 is a subunit of the GATOR1 complex, which exhibits regulatory control over mammalian target of rapamycin (mTOR), a protein complex involved in regulation of cell proliferation, among other functions (05). The other two components of GATOR1 are NPRL2 and NPRL3. Mutations of the genes encoding these proteins (NPRL2 and NPRL3) have also been identified in a smaller fraction of families with FFEVF, with individuals having sleep-related hypermotor, temporal lobe seizures, or other focal epilepsies (60; 78). Rare ADSHE families have been identified with mutations of GATOR1 genes (17; 26; 60).

An ADSHE family was described with bottom of the sulcus focal cortical dysplasia in two individuals and nonlesional sleep-related hypermotor epilepsy in the remainder of the family who were well characterized (66). Exome sequencing identified a truncation mutation of DEPDC5 in this family, highlighting that this gene could cause lesional and nonlesional epilepsy in a family. This observation has subsequently been confirmed in other families with focal cortical dysplasia and GATOR1 mutations (06; 60; 78). The mechanism of action by which mutations in GATOR1 genes result in epilepsy may relate to small, radiologically unidentifiable, focal cortical dysplasias in those with nonlesional epilepsy. In those with overt focal cortical dysplasia, a second hit such as a mutation in a gene within the same pathway may result in a developmental lesion (63). That mutation may be restricted to the lesional region or germline in origin.

In addition to MTOR regulator genes, a mother-daughter pair, both with nonlesional sleep-related hypermotor epilepsy, were identified and found to both carry a possibly pathogenic variant in the MTOR gene itself (40). Further study is necessary to clarify whether MTOR mutations are a significant cause of ADSHE.

CRH mutations. Missense mutations in the promoter and coding regions of CRH, encoding corticotropin-releasing hormone, have also been implicated in ADSHE (13; 62). The phenotype in these cases appears to be relatively mild, with no reports of the intellectual or psychiatric comorbidities. The mechanism by which altered CRH expression or function leads to epilepsy remains unclear; however, animal studies have demonstrated a proconvulsive effect of corticotropin-releasing hormone in the developing brain (03).

ADSHE can be caused by mutations in different genes and likely arises through different etiologic mechanisms, including channelopathies and focal cortical dysplasia. Even within the same family, variable presentations can be seen, demonstrating that ADSHE exhibits both genetic and phenotypic heterogeneity. The term “phenotypic heterogeneity” refers to the phenomenon of a single genetic mutation producing a phenotype of widely differing severity; this variability is usual within families with ADSHE (65). “Genetic heterogeneity,” on the other hand, describes different mutations causing a similar phenotype and takes two forms: allelic heterogeneity, referring to different mutations at the same locus, and locus heterogeneity, referring to mutations at different loci. Both forms of genetic heterogeneity are seen in ADSHE.

CABP4. A large family with ADSHE was reported in which all affected individuals had a variant in CABP4, the gene encoding calcium-binding protein 4 (12). The significance of CABP4 in ADSHE is unclear at this time, as the familial variant in this report is also present 19 times in the Genome Aggregation Database, of individuals without severe childhood-onset disease.

Mutations in known genes still only account for approximately 20% to 30% of ADSHE cases; thus, the remainder must occur via other mechanisms. These include as yet unrecognized genes acting in a monogenic fashion, though some cases may occur as a result of complex or polygenic inheritance. Epigenetic phenomena may also play an important role in some families.

Epidemiology

An Italian study estimated prevalence of sleep-related hypermotor (hyperkinetic) epilepsy at 1.8 to 1.9 per 100,000, based on identification of a total of 14 patients (7 women) in the regions of Modena and Bologna (76). None of the patients in that study had a family history of sleep-related hypermotor epilepsy, suggesting ADSHE is rare. Sleep-related hypermotor (hyperkinetic) epilepsy may occur more often in men than women, with the ratio estimated at seven men to three women (57), though the limited available data suggest sex prevalence is likely equal in familial cases.

Sudden unexpected death in epilepsy (SUDEP) occurs in sleep-related hypermotor epilepsy, with an estimated incidence of 0.36 per 1000 person-years (44). The incidence of SUDEP in ADSHE is not known, though at least one case has been reported, involving a family with a KCNT1 mutation (39), and SUDEP may be associated with DEPDC5 mutations (02).

Prevention

No recognized measures can be taken to prevent the development of epilepsy in ADSHE. However, prompt diagnosis and appropriate management have a key role in minimizing the psychosocial sequelae of this condition. In many individuals, the burden of their epilepsy is compounded by delayed diagnosis, with patients often misdiagnosed for many years with sleep disorders or non-epileptic psychogenic seizures. Appropriate diagnosis, investigation, and management can help to minimize the harm that such misdiagnoses can cause. Appropriate management of seizures results in improved sleep, which can improve on learning and daily function.

Differential diagnosis

The principal differential diagnosis of nocturnal frontal lobe epilepsy is benign sleep disorders, primarily the NREM arousal parasomnias (15). However, a number of sleep disorders can cause potential confusion with sleep-related hypermotor (hyperkinetic) epilepsy, including the following:

• NREM arousal parasomnias (sleepwalking, confusional arousals, sleep terrors) • REM sleep behavior disorder • Rhythmic movement disorder of sleep (jactatio capitis nocturna) • Catathrenia (nocturnal groaning) • Sleep starts • Nocturnal panic attacks • Psychogenic non-epileptic attacks • Sleep-related breathing disorders

Despite the relatively long list of differential diagnoses, by far the biggest practical problem in clinical setting is the distinction of sleep-related hypermotor epilepsy from the NREM arousal parasomnias such as sleep terrors and somnambulism. A careful history, however, will usually be sufficient to distinguish these disorders (14). In the patient with paroxysmal nocturnal events, the most important historical features indicating sleep-related hypermotor epilepsy as opposed to benign parasomnias are the timing of events (in sleep-related hypermotor epilepsy this is often within 30 minutes of sleep onset, whereas parasomnias typically occur between 1 and 2 hours after sleep onset); the number of events per night (in sleep-related hypermotor epilepsy, there are often multiple events in a single night, whereas parasomnias rarely occur more than once or twice per night); the duration of events, which may be brief or prolonged in parasomnias, but in sleep-related hypermotor epilepsy are almost invariably brief (usually less than 1 minute); and the presence of an aura (very common in at least a proportion of seizures in individuals with sleep-related hypermotor epilepsy, but almost never a feature of parasomnias). Extensive wandering is rare (though occasional) in sleep-related hypermotor epilepsy but is common in parasomnias; conversely, lucid recall for events is common in at least a proportion of seizures in some patients with sleep-related hypermotor epilepsy but is very uncommon in parasomnias (although some vague recollection is not uncommon).

Importantly, few individual features are exclusive to either condition, and the diagnosis must be made by taking into account all pertinent aspects of the history. This can be problematic, even for clinicians with considerable experience with these disorders. The diagnostic process may be improved by use of the Frontal Lobe Epilepsy and Parasomnias (FLEP) scale. This is a brief, validated clinical questionnaire that has been shown to generate reliable diagnoses on the basis of the history in most cases in this setting (14).

Although the history is paramount in this setting, in some cases video-EEG monitoring is required to make a diagnosis. However, this is only practical when events are happening on a frequent basis. Moreover, events in sleep-related hypermotor epilepsy are often associated with normal or non-specific EEG features, and in some cases diagnosis may be difficult even if events are successfully recorded. The key to diagnosis is the stereotyped nature of the attacks, which can be identified on video-EEG studies.

Once sleep-related hypermotor epilepsy has been diagnosed, a careful family history is essential to detect ADSHE. As penetrance is 70%, the diagnosis may not be readily apparent, especially as there is great variation in disease severity between individuals in any given family. Many affected individuals may have only very mild events, and adults in a family may have had no events since childhood or adolescence. A superficially taken family history may often miss affected relatives who are not diagnosed as having epilepsy.

Diagnostic workup

The diagnostic workup in ADSHE is the same as for any patient with epilepsy. It is important to appreciate that the history of the events and family history are paramount in arriving at the correct diagnosis, and other investigations may be noncontributory in many cases.

Routine EEG in wakefulness is normal in up to 90% of individuals with ADSHE, but sleep EEG shows frontally dominant epileptiform discharges in up to 50% (47).

Video-EEG monitoring may be essential in some patients to make the diagnosis, but diagnostic uncertainty may occasionally remain even after monitoring. Ictal patterns include evolving sharp or spike-and-wave discharges, rhythmic slow activity, or diffuse background flattening over frontal areas. Postictal focal slowing may be seen. Intracranial EEG recordings (eg, stereo-EEG) have demonstrated that ictal discharges may start in various extrafrontal areas (insulo-opercular, temporal, and parietal cortices). The key to the diagnosis remains in reviewing the events and confirming the stereotyped semiology, especially in cases without clear electrographic correlate on the scalp EEG. Electrographically, seizures tend to occur during NREM sleep, with about 70% occurring in stage 1 or 2.

Neuroimaging is usually normal. Occasionally, a structural brain abnormality is found, most commonly focal cortical dysplasia but also, less commonly, an acquired structural pathology.

At present, there is a limited role for diagnostic genetic testing in sleep-related hypermotor epilepsy. The diagnosis is made on the clinical presentation and family history in ADSHE. Where a patient has no known family history and no lesion is found on MRI, mutational analysis may uncover a de novo mutation in some cases with significant genetic counselling implications (52; 21). Multi-gene panels and next-generation sequencing techniques are becoming increasingly popular and affordable modalities; however, the yield in cases of sleep-related hypermotor epilepsy may be low given the relatively low frequency of mutations in known genes. Once a genetic mutation is found, it carries considerable implications in terms of selection of medications, recognition of comorbidities, and genetic counseling.

Management

Seizures in sleep-related hypermotor (hyperkinetic) epilepsy are usually responsive to treatment with antiseizure medications, with carbamazepine being particularly effective (65; 47; 71; 54; 20). Both acetazolamide and lacosamide have been reported as effective in resistant cases of SHE/ADSHE (75; 31). A case series of 24 individuals with ADSHE treated with topiramate reported seizure freedom, or a 50% reduction of seizures, in almost all cases (87.5%) (46). Perampanel has been studied as an add-on agent.

An open-label trial of fenofibrate as add-on therapy in 11 adults with sleep-related hypermotor epilepsy, five of whom had ADSHE, found that seizure frequency decreased and quality of life was improved (58). In a single patient with a known alpha4 nicotinic acetylcholine receptor subunit gene mutation, nicotine (administered via patch) was found to result in a significant reduction in seizure frequency (80); however, nicotine has not been used in larger trials in ADSHE.

In cases with KCNT1 mutations, the KCNT1 gain of function shows dose-dependent reversibility with quinidine, a broad-spectrum potassium channel blocker (38). Although case reports describe some benefit of quinidine in terms of seizure reduction and mild developmental gains in children with EIMFS, a single case of KCNT1 focal epilepsy with regression did not derive any benefit (07; 37). A placebo-controlled randomized trial of quinidine for sleep-related hypermotor epilepsy with KCNT1 mutations also did not show a therapeutic benefit, and several patients developed prolonged PT intervals even at low quinidine doses (45).

Outcomes

Outcome is primarily a function of the underlying etiology. A study with 139 patients with sleep-related hypermotor epilepsy showed total remission in only 22.3% of the cohort (29). This included sporadic and familial cases. Any underlying brain disorder (any combination of intellectual disability, perinatal insult, pathologic neurologic examination, and brain structural abnormalities) and seizures in wakefulness were more frequent among the group with refractory seizures.

Intellect is normal in most cases but over half of sleep-related hypermotor epilepsy patients will have some degree of neuropsychological deficit on formal testing (30). Those with gene mutations have greater risk of intellectual disability (30); in particular, individuals with KCNT1 mutations typically have intellectual disability and severe psychiatric and behavioral comorbidities.

Epilepsy surgery may be effective in some children with focal cortical dysplasia. Seizure outcomes may be compromised by extensive epileptogenic zones.

Special considerations

Pregnancy

There are no specific reports about the effect of pregnancy on seizures in ADSHE, although some women report exacerbation during pregnancy (65). Women with ADSHE should be treated in the same manner as other women with epilepsy before, during, and after pregnancy.

Anesthesia

There are no specific data about the use of anesthesia in ADSHE.