A clear description of BECTS, or benign focal epilepsy of childhood with rolandic spikes, was given by Nayrac and Beaussart in 1958 [218]. Later, Beaussart [219] drew attention to the benign nature of the condition. This is a childhood seizure occurring between 3 and 13 years of age, at an average age of onset of 7, seen mostly during drowsiness and sleep. The clinical seizures are characterized by focal clonic facial seizures often preceded by perioral numbness. In many cases, the patients have generalized tonic-clonic seizures that appear to be secondary generalization. On occasion, there is speech arrest. Consciousness is preserved. The EEG shows centrotemporal or rolandic spikes or sharp waves (Fig. 44.7) with a typical morphology of a triphasic sharp wave of high amplitude localized to the centrotemporal region but sometimes spreading to the contralateral hemisphere. Epileptiform discharges sometimes may occur outside the centrotemporal region and show occipital paroxysms in children exhibiting symptoms similar to those noted in BECTS [53]. The activation of interictal epileptic discharges (IEDs) by NREM sleep is a well-known feature of benign epilepsy of childhood with rolandic spikes. Authors agree in reporting a marked increase of IEDs during NREM sleep and a substantial reduction during REM sleep [220–224]. IEDs tend to be present throughout the night in each consecutive NREM cycles, showing a higher correlation with spindle frequency activity with respect to slow-wave activity time course [224]. The same correlation with spindle frequency activity was found also in other epileptic syndromes of childhood characterized by a marked activation of interictal spikes during sleep as the benign epilepsy with occipital paroxysms, the Landau–Kleffner syndrome, and the continuous spike-and-wave discharges during slow-wave sleep syndrome [225–227]. These findings suggest that the thalamocortical network that plays an important role in pacing sleep spindles in the generalized 3-Hz spike-and-wave discharges may also modulate focal discharges [82].

Indeed, it has been observed that local recurrent oscillatory activity and localized spike-and-wave discharges can be confined within a regional thalamocortical circuitry involving a circumscribed pool of neurons within the cortex [228, 229]. The above data suggest that subtle age-dependent dysfunctions of the thalamocortical system may play an important role in the pathophysiology of epileptic syndromes of childhood characterized by a marked activation of interictal spikes during sleep.

The prognosis of BECTS is generally excellent, with cessation of seizures by the age of approximately 16 years, without any neurologic sequelae and satisfactory response to anticonvulsant therapy. However, in some patients, the marked activation of IEDs during sleep may induce neuropsychological disorders [230]. For these reasons, as proposed in the 2010 Report of the ILAE Commission on Classification and Terminology, it would be better to replace the term “benign” with the term “self-limited” [216].

JME, an electroclinical syndrome, was described by Janz and Mathes [231] and later published in detail by Janz and Christian [232]. The onset of the syndrome is usually between 13 and 19 years and is manifested by massive bilaterally synchronous myoclonic jerks, which are most commonly seen in the morning shortly after awakening [232, 233]. The EEG is characterized by generalized spike-and-wave and typically multiple spike-and-wave discharges (see Fig. 44.3), seen in a synchronous and symmetric manner. Photosensitivity [234] and the phenomenon of perioral myoclonia [235] of the lips, tongue, jaw, or throat (precipitated predominantly by talking) may occur in a large number of patients with JME. The excellent response to anticonvulsants makes this condition relatively benign, although not “self-limited” and easily distinguishable from the malignant syndrome of progressive myoclonus epilepsies.

Epileptic syndrome with generalized tonic-clonic seizure on awakening [233, 236] is manifested by the occurrence in the second decade of generalized tonic-clonic seizures on awakening from sleep. This is a rare syndrome, and clinically, there may be occasional absence or myoclonic manifestations and photosensitivity resembling JME. There is considerable overlap between generalized tonic-clonic seizures on awakening and JME. Patients with generalized tonic-clonic seizures on awakening should have had at least six generalized tonic-clonic seizures, and in JME patients, there are relatively frequent myoclonic jerks and infrequent generalized tonic-clonic seizures [94].

CSWS, formerly known as electrical status epilepticus during sleep (ESES), is a disease of childhood characterized by generalized continuous spike-and-wave EEG discharges during slow-wave sleep. All-night PSG study is necessary for diagnosis. The patients display progressive behavioral disturbances, although the seizures disappear within months or years. This entity is rare and found in children between 5 and 15 years of age. ESES was first described by Patry et al. [163] in 1971 in six children. Later, Tassinari and coworkers reviewed the literature and gave a comprehensive description of the entity [237–239]

Most of the patients had a prior history of epilepsy. The characteristic EEG finding consists of 2.0 to 2.5 cycles/sec generalized spike-and-wave discharges seen during at least 85 % of NREM sleep and suppressed during REM sleep (see Fig. 18.​22a–c). Occasional bursts of spike-and-wave discharges or focal frontal spikes were noted during REM sleep. There were a few bursts of generalized spike-and-wave discharges seen in the EEG during wakefulness. These EEG discharges disrupted the stages of NREM sleep. In particular, the vertex sharp waves, K complexes, and spindles could not be well recognized. However, the cyclic pattern of REM-NREM persisted normally. Generally, there were no sleep disturbances, but some children had difficulty awakening in the morning. CSWS is now considered an epileptic encephalopathy of childhood characterized by cognitive and motor impairment and epilepsy [238, 239]. The etiologic heterogeneity of CSWS has been emphasized by Veggiotti et al. [240] and supported by a recent report of its presence in Rett syndrome [241]. The EEG findings of continuous epileptic discharges generally disappear within 3 years of appearance [237–239]. Focal abnormalities, in the EEG, may persist, however. It is not clear whether CSWS is a focal epilepsy or a generalized epilepsy with heterogeneous presentation, and hence, it is classified under the category of undetermined epileptic syndromes. Seizures show a benign course and respond well to antiepileptic medications, with disappearance of seizures by the mid-teens. The psychological impairment, however, persists.

Landau–Kleffner syndrome (LKS) is an acquired aphasic syndrome occurring in a previously normal child and probably is a variant of CSWS [242]. The characteristic language dysfunction in LKS is an apparent “word deafness” or auditory verbal agnosia. There are many similarities between CSWS and LKS, and the type of neuropsychological dysfunction may depend on the location of the discharge (e.g., frontal in CSWS and temporal in LKS). Most CSWS patients have no evidence of language dysfunction. Approximately 70–80 % of children have seizures that are characterized by eyeblinking, head dropping, or minor automatisms with secondary generalization. These patients respond to antiepileptic medications and remain seizure-free by the mid-teens. The EEG pattern is similar to that noted in CSWS.

Nocturnal temporal lobe epilepsy (NTLE) has not been well characterized. It has been described by Bernasconi and coinvestigators [243] in a subgroup of 26 patients with refractory temporal lobe epilepsy without structural lesion, with more than 90 % of seizures occurring during sleep. Focal seizures with transient impairment of consciousness, staring, automatism, and experiential or other sensory components occurring predominantly during sleep characterize the clinical syndrome. These simple partial staring seizures are frequently followed by secondary generalization. The following features differentiate patients with NTLE from the typical nonlesional temporal lobe epilepsy patients with diurnal seizures: a rare family history of epilepsy, low prevalence of childhood febrile seizures, infrequent and nonclustered seizures, and favorable surgical outcome [94, 243].

In the early 1980s, Lugaresi and Cirignotta [244] and Lugaresi et al. [245] reported cases of paroxysmal attacks occurring during NREM sleep characterized by prominent motor behaviors in the form of dystonic posturing, tremors, and ballistic movements of the limbs, lasting 15 s to 2 min and not associated with epileptic abnormalities on the scalp EEG (Box 43.4). The attacks could respond to low doses of carbamazepine, posing the question of whether they represented epileptic seizures or sleep-related movement disorders. These attacks, termed nocturnal paroxysmal dystonia, were later demonstrated to represent a form of nocturnal frontal lobe epilepsy (NFLE) [245–248]. Nocturnal frontal lobe epilepsy (NFLE) is a syndrome of heterogeneous etiology as genetic, lesional, and cryptogenetic forms have been described. Although generally considered a benign clinical entity, severe, drug-resistant forms do exist [249].

The spectrum of frontal lobe epilepsy manifestations include the so-called paroxysmal arousals, characterized by abrupt arousals from NREM sleep often accompanied by asymmetric tonic or dystonic posturing or complex movements such as pelvic thrusting, pedaling, choreoathetoid and ballistic movements of the limbs (Videos 1, 2, 3), lasting less than 20 s [250–252] and the so-called epileptic nocturnal wanderings (Video 4) [253, 254], more complex events lasting 2–4 min and associated with agitated ambulation and jumping about. Notably, paroxysmal arousals, often recurring quasi-periodically every 20–40 s for long stretches during NREM sleep, are often associated with attacks of nocturnal paroxysmal dystonia and epileptic nocturnal wanderings in the same patient and can represent the initial manifestations of the more prolonged attacks [255]. The increasing complexity of NFLE ictal motor behaviors, from minor to major events, reflects a different duration and propagation of the discharge within the frontal lobe [255, 256].

The peculiar features of NFLE thus consist of its nocturnal recurrence, related to NREM sleep stages in over 80 % of the seizures [257], and of the characteristic motor pattern with truncal and bipedal gross and often violent movements with dyskinetic features; the latter have led to the definition of hypermotor or hyperkinetic seizures.

Functional brain imaging in frontal lobe seizures (nocturnal paroxysmal dystonia and paroxysmal arousals) indeed confirms that the peculiar motor patterns are related to involvement of mesial, especially cingulate, motor areas [258, 259]. Studies with intracerebral electrodes (stereo-EEG) conducted in drug-resistant patients with NFLE have shown that the seizure onset in patients with asymmetric tonic or dystonic posturing is generally localized in the posterior portion of the frontal cingulate gyrus and in the posterior mesial frontal cortex with a primary involvement of the supplementary motor area [56, 260]. In patients with seizures characterized by hyperkinetic automatisms and complex motor behaviors, the region of seizure onset may involve the dorsolateral and anterior frontal regions (frontopolar and frontal anteromesial regions) [56, 261, 262]. Seizure characterized by the association between fear and nocturnal wandering seem to involve a cerebral network including the anterior cingulate, orbitopolar, and temporal regions [260, 263–265]. Moreover, it has been shown that “hypermotor seizures” may also be found in seizures originating from the temporal lobe and the insula [263, 266–269]. In such cases, the hyperkinetic features appeared 8–40 s after the beginning of the discharge in the temporal or insular regions, when the discharge spreads to other cortical structures such as the cingulate, frontal, and parietal cortices (Video 5) [266, 269]. Such a possibility of an extra-frontal origin of seizures renders the term NFLE somewhat misleading. Recently, in order to improve the definition of the disorder and establish well-defined diagnostic criteria, a consensus conference was held. It was recommended to change the name into “Sleep related Hypermotor Epilepsy (SHE),” considering that seizures are associated with sleep rather than time of day, seizures may arise outside the frontal lobe and the hypermotor aspects if the attacks are characteristic [270]. The etiology of SHE may be genetics, due to a cortical abnormality or unknown. Diagnostic criteria were developed with three levels of certainty: witnessed (possible) SHE, video-documented (clinical) SHE, and video-EEG-documented (confirmed) SHE. Due to the recurrence of nocturnal motor events, several NFLE patients may complain of non-restorative sleep and of daytime sleepiness [249, 257, 261].

Scheffer et al. [271] described an autosomal dominant form of frontal lobe epilepsy in six families. Brief motor seizures usually occurred in clusters during sleep. The disorder usually started in childhood and persisted through adult life. Patients were of normal intellect and had normal neurologic examination and neuroimaging. Response to carbamazepine was excellent. In most cases, interictal EEGs were normal, although one family with daytime attacks had epileptiform discharges. Video telemetry during the attacks confirmed their epileptic nature. They called this condition autosomal dominant nocturnal frontal lobe epilepsy (ADNFLE). The clinical features of ADNFLE were later confirmed by Thomas et al. [272] in one family with very frequent seizures during infancy, in which carbamazepine therapy again was dramatically effective, and by Oldani et al. [273], who studied 33 patients and found similar results. In 1995, Phillips et al. [274] mapped a gene responsible for ADNFLE in a large Australian kindred to chromosome 20q13.2, and Steinlein et al. [275] demonstrated that epileptic nocturnal frontal lobe (ENFL) type 1 was due to mutations in CHRNA4, the gene encoding the A4 subunit of the acetylcholine (ACh) neural receptor. Another linkage site was later reported to chromosome 15q24 accounting for ENFL type [2, 276] and mutations in CHRNB2, the gene encoding for the B2 subunit of the ACh neural receptor localized on chromosome 1 as accounting for ENFL type 3 [277]. Another linkage locus to chromosome 8p12.3-8q12.3 and a missense mutation in the gene CHRNA2 encoding for the neural ACh receptor A2 subunit have been reported in familial seizures characterized by complex and finalized ictal behavior resembling epileptic nocturnal wanderings [264], as well as mutations in the corticotropin-releasing hormone gene [278].

Genetic findings thus implicate the nicotinic ACh receptors in ADNFLE. Mutations responsible for ADNFLE work by increasing the receptor sensitivity to ACh [279], indicating that a gain of function of the mutant receptors underlies the neuronal dysfunction responsible for the epileptic seizures. Mutated nicotinic receptors responsible for ADNFLE were also found to be more sensitive to carbamazepine, which works as a noncompetitive inhibitor of the nicotinic ACh receptors [280]. On the basis of the genetic findings and functional imaging data [281], the pathogenesis of ADNFLE has been attributed to dysfunction in the dorsal cholinergic ascending arousal system, and a common background with the arousal parasomnias has been hypothesized based on epidemiologic and clinical data [282, 283]. In animal models, it has been shown that a mutation of the nicotinic receptors may provoke an unbalanced excitation/inhibition circuitry within the GABAergic reticular thalamic neurons, thus favouring seizures through the synchronizing effect of spontaneous thalamocortical oscillations [96]. On the other hand, other experimental studies have reported an involvement of nAChR in the regulation of arousals, sleep stability, and the activity-rest pattern [95]. The observed genetic alterations in NFLE could create the conditions for both arousal instability and seizure generation [284].

However, ADNFLE is not only related to mutations in the cholinergic system. Since 2005, other genes, not belonging to the nACh receptor subunit family, have been identified. In particular, Combi et al. [278, 285] found mutations of the corticotropin-releasing hormone gene in sporadic and ADNFLE cases. The in vitro functional analysis of both variations demonstrated an altered level of protein expression suggesting an interrelation between CRH concentration and neuronal excitability with a possible effect on thalamocortical loop dysfunctions [278]. Very recently, a further CRH mutation has been found in the protein pro-sequence region of the CRH of two affected siblings of an Italian ADNFLE family [286]. In 2012, a further gene on chromosome 9 encoding the sodium-activated potassium channel subunit 1 (KCNT1) was associated with ADNFLE [287]. In particular, four variants in KCNT1 were identified in three families and in a sporadic ADNFLE case, all showing a severe ADNFLE phenotype with early onset, high prevalence of intellectual disabilities, and psychiatric or behavioral problems, including psychosis, catatonia, and aggression. Very recently, mutations in the DEPDC5 gene were reported as responsible for different types of focal epilepsies, including ADNFLE [288]. However, the prevalence of mutations in this gene remains to be assessed.

In summary, ADNFLE is a heterogeneous genetic syndrome that can be incidental to mutations in different genes. Mutations in these genes, however, account only for a minority of cases [289], and their mean penetrance ranges from 60 to 80 %. Hence, further studies are needed to better characterize this heterogeneous syndrome. Given the high intrafamilial variability and the overlapping features of the clinical manifestations, ADNFLE patients do not show a clear distinction from sporadic NFLE cases, except for certain ADNFLE mutations frequently associated with specific additional neurologic or psychiatric symptoms.

Patients with epilepsy may complain of excessive daytime sleepiness (EDS), insomnia (inability to fall or maintain sleep and early morning awakening), and adverse daytime consequences related to insomnia and EDS, as well as unusual movements and behaviors intruding into sleep. EDS in epileptics may result from clinical seizures, particularly nocturnal seizures and ictal or interictal EEG epileptiform discharges; comorbid conditions such as obstructive sleep apnea syndrome (OSAS; see later) and polycystic ovary syndrome (PCOS); and effects of antiepileptic medications (see later). Box 44.8 lists causes of EDS in epileptic patients. Maganti et al. [316] reported that EDS and sleep complaints are common among adults with epilepsy, and in some patients, these may be due to underlying sleep disorders such as sleep apnea. Khatami et al. [317] assessed sleep-wake habits and EDS using a standardized questionnaire in 100 consecutive outpatients with epilepsy and 90 controls. Sleep complaints were more common in epilepsy patients than in controls, and sleep maintenance insomnia and EDS were found frequently; they noted that loud snoring and restless legs symptoms are the only independent predictors of EDS in epilepsy patients. Jenssen et al. [318], based on a questionnaire and chart review of a tertiary referral center, noted subjective somnolence to be related mainly to depression rather than to obstructive sleep apnea (OSA) and other variables. A recent case–control study from Brazil [319] found that patients with epilepsy had more EDS, daytime dysfunction, and sleep disorders compared with a control group. In a systematic review of PubMed-cited articles from 2002 to 2012 (most studies were cross-sectional and questionnaire-based), Giorelli et al. [320] concluded that EDS was related more frequently to undiagnosed sleep disorders than to epilepsy-related factors.

Manni and Terzaghi [321] described two elderly men with late-onset sleep-related tonic-clonic seizures and RBD. The authors hypothesized that RBD may facilitate seizure occurrence. In a later study, Manni et al. [322] reported co-occurrence of epileptic seizures and RBD in six cases. The authors cautioned that further investigations of the occurrence of RBD episodes and epilepsy are needed to understand the neurobiological significance of this comorbidity.

The relationship between seizure type, severity of seizure, and extent of sleep deficits remains somewhat controversial, and the reports are contradictory. WASO, sleep stage shifts, and sleep fragmentation are found in all seizure types [157, 314, 323–326]. Reduction of REM sleep and an increase in NREM stages 1 and 2 are in part dependent on the type of epilepsy. Declerck et al. [314] found an increase of NREM stages 1 and 2 and a reduction of REM sleep in 258 patients with primary generalized or partial seizures with secondary generalization as compared with 223 nonepileptic subjects. Seizure occurrence during sleep accentuates sleep deficits, which are more marked in primary generalized and partial seizures with secondary generalization than in other types. In 25 % of epileptics, Declerck et al. [314] could not evaluate PSG recordings because of severe encephalopathies associated with seizures. Similar findings were obtained by Bessett [157]. Baldy-Moulinier [323] noted a decrease of REM sleep in patients with CPS occurring during sleep. However, Baldy-Moulinier found markedly reduced REM sleep in patients having only one attack of secondary generalized seizure during the night. It is interesting to note that Bessett [157] in human epileptics and Baldy-Moulinier [323] in temporal lobe epilepsy models found no rebound REM sleep in subsequent recordings after REM sleep loss, which is contrary to the usual findings of REM rebound after REM deprivation. In summary, WASO and sleep fragmentation are found in all types of epilepsy, and generalized seizures are associated with an increase of NREM stages 1 and 2 and a reduction of REM sleep. In CPS, there is often REM reduction only. Besset [157] could not discriminate NREM stages or REM sleep in the EEG because of disrupted sleep architecture due to the seizures (ictal and interictal).

Hoeppner et al. [327] studied self-reported sleep disorder symptoms in epilepsy. They gave a questionnaire relating to six aspects of sleep: delayed sleep onset, night awakenings, dreams, night terrors, sleepwalking, and fatigue on awakening. They evaluated four groups of subjects: (1) Four patients with simple partial seizures, (2) 18 patients with CPS, (3) Eight patients with generalized seizures, and (4) 23 controls (14 women and 9 men aged 16–53 years). They found significantly more sleep disorder symptoms (particularly frequent awakenings at night) in patients with simple partial seizures and CPS. The generalized group behaved like the control group. Patients with the most frequent seizures, irrespective of type, had the most sleep disturbances.

Roder-Wanner et al. [328] obtained polygraphic sleep recordings in 43 patients with different types of epilepsies. They found that patients with generalized epilepsy had a higher percentage of deeper stages of sleep (NREM 3 and 4) than patients with focal epilepsy. These observations are correlated with the factor of photosensitivity, which was noted in a subgroup of these patients. The authors concluded that there was no real relationship between sleep structure and the type of epilepsy. Thus, there is some controversy regarding the relationship between seizure type and sleep. It can be concluded, however, that the severity of sleep deficits is in part correlated with severity of the seizure disorder. Animal studies support such a conclusion [329, 330]. In previous studies, sleep structure abnormalities may have been related to clinical or subclinical seizure activity preceding the PSG investigation or to the medication received during the study.

There are contradictory reports regarding REM sleep disturbance [179]. On seizure-free nights, REM sleep is usually normal, but REM decrement is noted when there are primary or secondary generalized seizures during the night. There is no REM suppression during partial seizure without secondary generalization [157, 323]. Bowersox and Drucker-Colin [331] stated that increased cortical neuronal excitability and reduced seizure threshold may result from chronic REM sleep deprivation secondary to repeated and frequent nocturnal generalized seizures. In a recent review, Ng and Pavlova [332] hypothesized that desynchronized EEG with altered connectivity is responsible for seizure suppression in REM sleep. In fact, after reviewing several studies, Jaseja and Jaseja [333] observed a linear relationship between reduction of REM sleep and intractable or refractory epilepsy and postulated REM sleep duration as a biomarker for predicting intractability of seizure. The authors even made a novel suggestion of deep brain stimulation (specifically of pedunculopontine nucleus) as a novel therapeutic approach in intractable epilepsy to increase REM sleep percentage.

In a series of 15 patients with temporal lobe seizure disorders, Touchon et al. [334] found increased WASO, shifting of the sleep stages, and increases in NREM sleep stages 1 and 2. In a study of 23 patients with temporal lobe epilepsy, Kohsaka [335] found significantly decreased sleep efficiency and increased awakenings in both treated and untreated patients. He also noted increased NREM stage 4 in untreated patients compared to healthy controls. The site of the primary focus may determine the type of the sleep disturbances [18]. Foci in the amygdalohippocampal region may lead to increased WASO and decreased sleep efficiency. Frontal lobe epileptics, however, may show a specific reduction in stages 3 and 4 NREM sleep.

In a questionnaire-based study of 40 children with tuberous sclerosis, Hunt and Stores [336] found that concurrent epilepsy was significantly associated with sleep disturbances in these children. This observation was corroborated by Bruni et al. [337], who found a more disrupted sleep architecture in patients with tuberous sclerosis and epilepsy compared with seizure-free children. In large series of children with epilepsy, there was a correlation among seizure frequency, incidence of interictal epileptiform discharges, duration of seizure disorders, behavior problems, poor quality of sleep, and disturbed breathing during sleep [338, 339]. Becker et al. [340] noted that 80 % of 30 children complaining of sleep disturbance had PSG documentation of OSA, abnormal sleep architecture, and fragmentation.

de Weerd et al. [341] reported poor quality of life and sleep disturbances that were more common in adults with focal seizures with or without secondary generalization compared with controls. A variety of PSG-documented sleep abnormalities have been described in patients with JME [136, 342, 343] and focal temporal lobe epilepsy. Bazil et al. [136] documented reduction of REM sleep by seizures in temporal lobe epilepsy. Sleep abnormalities in absence seizures gave conflicting results [180]. In BECTS, no sleep architectural abnormalities are noted [42, 344].

Several authors have noted that sleep abnormalities are more common in patients with primary or secondary generalized seizures than in those with partial focal epilepsies [42, 344, 345], but other studies have shown that patients with severe or medically refractory temporal lobe seizures may have equally severe sleep structural abnormalities [343, 344]. Several investigators have reported severe sleep abnormalities in symptomatic epilepsies associated with neurologic deficits [166, 346–351]. In all of these patients, both sleep dysfunction and seizure disorders must be treated simultaneously to obtain best results.

In several reports, PSG findings and sleep abnormalities have been described in partial seizures, especially frontal lobe [352, 353] and temporal lobe [136, 353] seizures. Tachibana et al. [352] showed an improvement in sleep structure after treatment with appropriate antiepileptic drugs (AEDs) (Fig. 44.9).

It is generally thought that sleep deficits in seizure disorders are secondary to the severity of the seizure disorder and are a direct result of seizures during sleep. However, studies by Tanaka and Naquet [354] demonstrated progressive sleep deficits in amygdala kindling models. In addition, the sleep deficits persisted one month after discontinuation of kindling procedures.

Shouse and Sterman [329] produced amygdala kindling in 10 adult cats and studied their sleep and waking patterns chronically. They found a progressive sleep disturbance and retention of the deficit over a prolonged period after termination of amygdala stimulation. These findings suggest the “kindling” of a sleep disturbance in addition to a seizure disorder. The authors further stated that sleep abnormalities cannot be viewed as a simple or temporary side effect of epileptiform activity. It appears that a permanent change in sleep physiology occurs in epilepsy. These observations of Shouse and Sterman [329] partially answer the question posed by Passouant [1]: “Can epilepsy lead to a sleep disorder?” Effective treatment of epilepsy with anticonvulsant medications or surgical methods normalizes sleep disturbances in human epilepsy [325].

In 1995, Silvestri et al. [355] reported six patients who were diagnosed in childhood as having disorders of arousal and later developed epileptic seizures. The sleep disorders consisted of sleepwalking and night terrors, all confirmed by PSG studies. The seizures noted were complex partial in five and generalized tonic-clonic in one. Nocturnal monitoring confirmed the epileptic nature of these events. The authors hypothesized that because both disorders of arousal and epilepsy are related to sleep and share other common factors such as age of onset and precipitating factors, these disorders share common functional substrates, and it is possible that disorders of arousal may later turn into epileptic seizures. It should be noted, however, that sleepwalking and sleep terrors are frequently noted in children, and seizures may simply coexist with these NREM parasomnias. Most sleep specialists and epileptologists simply do not believe that such parasomnias can later turn into epileptic seizures. It should also be remembered that cases of typical disorders of arousal not associated with epileptic discharges in epileptic children have been described [14, 162, 356].

In most of the studies, phenobarbital is found to increase stage 2 NREM sleep and decrease sleep-onset latency, REM sleep, and WASO without any significant effect on slow-wave sleep [346, 373, 374]. Wolf et al. [373] reviewed the literature to assess the effect of barbiturates, phenytoin, carbamazepine, and valproic acid treatment on sleep. They noted significant reduction of REM sleep, a reduction in total awake time, and an increase in NREM stage 2 sleep as the short-term effects of barbiturates. The long-term effects of barbiturates are similar in general, but in some cases, the sleep pattern returned to the premedication level. Wolf et al. [373] performed a prospective polygraphic study of sleep in epileptic patients before and after medications using a crossover design. They studied phenobarbital, phenytoin, ethosuximide, valproic acid, and carbamazepine. The authors included 40 unmedicated patients to study the effect of phenobarbital and phenytoin. The short-term effects of phenobarbital included reduction of WASO and REM sleep and increase of stage 2 NREM sleep. There was no relationship with the serum drug levels.

Manni et al. [375] performed an objective and subjective assessment of daytime sleepiness using the multiple sleep latency test (MSLT), clinical, and psychometric data on 10 patients with generalized epilepsy treated chronically with phenobarbital, 10 patients with cryptogenic partial epilepsy treated with carbamazepine, and 10 healthy controls. These authors found that patients on phenobarbital had a greater daytime sleep tendency and performed worse on the digit symbol substitution test compared to the other two groups. In a similarly designed study [376], they noted a shorter mean sleep latency in patients on phenobarbital compared with patients on sodium valproate and controls. Psychomotor functioning was also poor in patients on phenobarbital compared to controls, whereas patients on valproate had some attentional impairment and a tendency toward longer motor movement time. However, they did not find a correlation between the assessed parameters and serum drug concentrations.

Phenytoin generally causes a reduction of sleep-onset latency, REM sleep, and slow-wave sleep as well as sleep efficiency, but causes increased stages 1 and 2 NREM sleep [351, 371]. Phenytoin also increases daytime sleepiness. The short-term effects of phenytoin included no change in the percentage of WASO, a decrease in NREM sleep stages 1 and 2, and an increase in sleep stages 3 and 4; there was no change in REM sleep and no relationship with the serum drug levels. Wolf et al. [373] studied the long-term effects of phenytoin in 12 patients. The long-term effects were in general a reversal of the short-term effects and consisted of an increase of NREM sleep stages 1 and 2 with a decrease of slow-wave sleep. REM sleep, however, remained unaltered.

Carbamazepine has been studied fairly extensively in various studies. This AED is found to increase sleep efficiency and slow-wave sleep but decrease REM sleep in healthy subjects [374] and transiently decrease REM in epileptics. Baldy-Moulinier [323] reported normalization of disturbed sleep pattern in temporal lobe epileptics after carbamazepine treatment. After acute carbamazepine administration in cats, Gigli et al. [377] reported an increase of NREM stage 1 sleep and total sleep time, a decrease of REM sleep, and reduced duration of awakenings. Some studies examined the effects of carbamazepine in treated versus untreated epileptic patients [334, 342, 343]. Studies in healthy normal subjects showed that carbamazepine can increase slow-wave sleep, decrease REM sleep, and consolidate sleep [378–380]. In some studies, the EEG effects in epileptics have been contradictory. For example, Legros and Bazil [371] failed to find any EEG effects during sleep in 10 epileptics treated with long-term carbamazepine, but Bell et al. [369] found increased slow-wave sleep and decreased stage 2 NREM sleep with carbamazepine monotherapy.

There have been very limited studies to describe effects of primidone on the sleep EEG. In one study using 30 healthy subjects, a single dose of primidone (250 mg) resulted in an increase of slow-wave sleep and a reduction in REM sleep [381]. Treating epileptic patients with 750 mg primidone daily for 3 months resulted in reduced sleep-onset latency and REM density but not percentage [382].

The effects of ethosuximide included disrupted sleep, increased sleep latency, increased stage 1 sleep, decreased slow-wave sleep, and increased REM sleep and awakenings [373, 383].

Valproic acid in general has minimal effects on sleep architecture in patients with epilepsy [368, 373–375, 384]. Findji and Catani [384] reported an improvement of sleep organization and increase of slow-wave sleep in epileptic children after treatment with valproic acid. At higher doses, however, Harding et al. [385] observed a decrease of delta and REM sleep.

The benzodiazepine group of drugs is generally used for status epilepticus (e.g., lorazepam, diazepam, and midazolam), but sometimes clonazepam is used in some drug-resistant seizures and certain types of seizures (e.g., myoclonic seizures and Lennox–Gastaut syndrome). These drugs generally cause decreased sleep efficiency, sleep-onset latency, stage 1 NREM and slow-wave sleep, and arousals [325, 368, 386].

Most of the traditional AEDs adversely affect nocturnal sleep, sleep architecture, and daytime vigilance, but newer AEDs in limited studies have shown minor and even positive effects [387].

Several newer AEDs have come onto the market to treat patients with seizure disorders; most of them have been indicated to use as add-on drugs, but some are being used as primary AEDs. These drugs have not been studied extensively to determine their effects on sleep architecture.

Felbamate is one of the earlier drugs in the newer generation but has largely been discontinued because of severe hepatotoxicity. This drug has been reported to cause insomnia in epileptic patients [363, 388].

Gabapentin was originally developed to treat seizure disorder, but later was found to be useful in many other conditions, such as neuropathic pain and restless legs syndrome/periodic limb movements in sleep. In healthy subjects, gabapentin increases slow-wave sleep [382, 389, 390]. Placidi et al. [391] studied the effects of long-term gabapentin treatment on nocturnal sleep in drug-resistant epileptics and observed an increase in slow-wave and REM sleep, and a reduction of arousals and stage 1 NREM sleep.

Lamotrigine has minimal effects on sleep in general [351, 371]. Lamotrigine may cause increased REM sleep and decreased slow-wave sleep [374]. Sadler [361] reported insomnia (difficulty initiating and maintaining sleep shortly after administration of the drug) requiring reduction in dosage or discontinuation of the drug in over 6 % of 109 patients treated with lamotrigine. This finding, however, was contradicted by Foldvary et al. [362], who did not find any insomnia in any of 10 adult patients with focal epilepsy on lamotrigine treatment.

Levetiracetam in general has minimal effect on sleep architecture in normal volunteers. Bell et al. [369] studied levetiracetam in normal volunteers and patients with epilepsy in a double-blind, placebo-controlled study. They found increased stage 2 NREM sleep in both epileptics and controls, and increased REM sleep latency only in the healthy subjects and decreased slow-wave sleep in patients. In a double-blind, crossover, placebo-controlled study in 14 healthy volunteers using PSG and the MSLT after oral administration of levetiracetam up to 2000 mg/day or placebo for 3 weeks, Cicolin et al. [392] found increased total sleep time and sleep efficiency and decreased WASO. MSLT findings did not differ between the two groups. The authors concluded that levetiracetam in healthy volunteers consolidated sleep without causing any daytime sleepiness.

Oxcarbazepine has not been adequately studied but has been noted to cause excessive sleepiness [346].

Pregabalin is a more recent AED and has been studied in a limited manner. Hindmarch et al. [393] studied 24 healthy volunteers and measured sleep objectively using PSG and subjectively using a questionnaire. Compared with the placebo, pregabalin significantly increased slow-wave sleep and reduced sleep-onset latency and REM sleep percentage. Subjective evaluation showed significant improvement in sleep quality, but ratings of behavior following awakening were impaired.

Tiagabine, a selective GABA reuptake inhibitor, has been used for partial and secondary generalized seizures. In healthy elderly subjects, tiagabine significantly increased slow-wave sleep and sleep efficiency [394, 395]. In patients with primary insomnia, tiagabine increased slow-wave sleep and reduced WASO in a dose-dependent manner [396, 397].

Topiramate, tiagabine, zonisamide, and vigabatrin, which are used in some partial secondary generalized seizures, especially in those not responding to other AEDs, have not been adequately tested to study the effects of these drugs on sleep architecture in epileptic patients. Bonanni et al. [370], following topiramate monotherapy in an open-label trial with 14 epileptics, found no difference in sleep architecture and daytime sleepiness as measured objectively by multiple sleep latency testing. Bonanni et al. [398] studied the effects of carbamazepine and vigabatrin on daytime sleepiness in patients with partial epilepsy by measuring with MSLT and overnight PSG. The results suggested that vigabatrin did not significantly affect sleep architecture in their patients with epilepsy. Vigabatrin treatment in medically refractory epilepsy, however, causes weight gain, making these patients susceptible to developing OSAS [399].

Lacosamide, a later generation novel AED used as an add-on therapy, has been shown in a small group of drug-resistant epileptics to have no detrimental effects on sleep quality and quantitative EEG characteristics [387]. The other more recently approved AEDs have not been studied to see the effects on sleep architecture.

Vagus nerve stimulation has been used with some success in patients with refractory or intractable seizure disorder. However, in addition to improving the seizure state and daytime alertness, vagus nerve stimulation caused sleep disordered breathing in some patients [400–403].