Introduction

Epilepsy and sleep are related in several aspects, from physiology and pathophysiology to clinical appearance and neurophysiological diagnosis to treatment. It has been known since the days of Aristotle that epileptic seizures may occur exclusively during sleep.¹ Later, several authors observed that certain epileptic seizures may show a tendency to occur during sleep or waking periods and recognized the relation to the sleep–wake cycle.²^–⁸ Also, epileptiform discharges occur more frequently during non-REM sleep than in REM sleep and waking periods,⁹^–¹¹ and arousals and transitions periods between sleep stages are considered to facilitate epileptiform discharges.¹²^–¹⁵ Antiepileptic drugs have effects on sleep, and it has even been speculated that some of their antiepileptic effect is due to their effect on sleep.¹⁶

In this chapter, we will discuss the relationships between epilepsy and sleep, the use and ambiguity of sleep in the diagnosis of epilepsy, and the effects of treatment on epilepsy and sleep.

Physiology and Diagnostic Studies of Sleep

Sleep is a complex behavioral trait, in which consciousness and responsiveness to environmental stimuli are reduced. There is an ongoing discussion as to why we sleep, but some physiologic functions, such as memory consolidation,¹⁷ have been shown to be dependent on sleep. The pressure on the organism to initiate sleep depends largely on circadian and homeostatic influences.¹⁸ Several brain regions are actively involved in the sleep process. Commonly, the cortical activity, and with it the sleep electroencephalography (EEG), is described as “synchronous” in a periodical fashion. The synchronicity refers in particular to the cortical neurons.¹⁹ Deeper brain structures such as the midbrain and brainstem involved in this active process do not necessarily reflect the same level of synchronicity as the cortex. There are two general different states of sleep, rapid eye movement sleep (REM sleep) and non–rapid eye movement sleep (NREM sleep), with the latter being divided further into four stages.²⁰ Over the course of a night, NREM and REM sleep alternate with an approximately 90-minute rhythmicity, creating four to six sleep cycles. The NREM stages differ in their amount of slow-wave sleep and special defining patterns like K-complexes and sleep spindles. Electrophysiologically, the dynamic of NREM sleep can be understood as the neurobehavioral attempt to reach a high level of synchrony, expressed as slow-wave sleep and a deafferentiation of sensory input.²¹ REM sleep is defined by fast eye movements, loss of muscle tone, and an EEG desynchrony. These sleep oscillations are mainly generated by the corticothalamic system, the reticular thalamic and thalamocortical circuit, and thalamically projecting brainstem structures.²² This difference in cellular and systemic neuronal activity is important for the difference in occurrence of seizures and interictal epileptiform discharges during sleep.

During sleep, external (for example, environmental noises) and internal stimuli (for example, sleep apnea episodes) may evoke arousal reactions. In the arousals, an adequate reaction to the stimulus and the attempt to stay asleep are competing impulses. A night with many arousals will be not as restful as a night with few arousals. In the classic sleep evaluation according to the criteria of Rechtschaffen and Kales ²⁰ an arousal is an event of fast, desynchronized EEG discharges. Terzano and colleagues²³ described a 20- to 40-second rhythmicity of changes in the EEG as the cyclic alternating pattern (CAP), defined by an alternation of phasic EEG events and background activity. Now CAP is considered an expression of different arousal mechanisms that are expressed differentially in the various phases of sleep and in sleep stage transitions.²⁴ During arousals and during stage shifts, interictal epileptic discharges and seizures are more likely to develop, especially in generalized epilepsies.^13,^25,²⁶

In evaluating and scoring sleep, polysomnography is the specific electrophysiological diagnostic procedure. The EEG, an electrooculogram (EOG), and a chin electromyograph (EMG) are needed for sleep-stage scoring. For more sleep specific diagnostics, respiratory parameters, ECG, and extremity EMG are derived. For polysomnography, an EEG derivation with the full 10 to 20 placed electrodes is usually only performed for research purposes or in the differential diagnosis of epilepsy and sleep disorders. The conventional EEG scoring requires only one EEG channel (electrodes C3–A2 or C4–A1).

Despite not incorporating all the derivations needed for polysomnography, routine EEG or EEG-video monitoring will already enable the diagnosis of all sleep stages, although with some ambiguity. The prefrontal leads (Fp1 and Fp2) and the lateral temporal leads (FT9 and FT10 or T7 and T8) can be used to interpret eye movements. An epilepsy-focused EEG will allow the evaluation of sleepiness (eye rolling in the frontal leads, temporal theta of drowsiness), falling asleep (change of background, vertex waves, and posterior sharp transients of sleep [POSTS]), and sleep stages 2, 3, and 4 (K-complexes, sleep spindles in stage 2, increase of delta activity in stages 3 and 4). REM sleep can be detected by fast eye movements and the low-amplitude beta and gamma activity in the EEG, but its exact beginning may be obscured without an EMG recording of the chin muscles. The formal scoring of conventional sleep stages requires 30-second epochs starting with the lights-off signal.²⁰

Excessive daytime sleepiness can be assessed by the multiple sleep latency test (MLST), which provides more objective data than questionnaires.²⁷ The MLST is usually performed on a day following a polysomnography to provide accurate documentation of the preceding night’s sleep.

Effects of Epilepsy on Sleep

Epilepsy patients typically report poorer sleep than healthy controls,²⁸ and it seems that sleep disturbance affects their quality of life.^29,³⁰ The prevalence is higher in patients with comorbid anxiety and depression.³⁰

SLEEP IN PATIENTS WITH GENERALIZED EPILEPSIES

Janz ¹⁶ described that patients with awakening epilepsy and those with sleep epilepsy had distinct sleep patterns and sleep habits. According to that study, patients with generalized epilepsy like to stay up late in the evening and often have difficulty falling asleep. Their sleep appeared to be disrupted. In the morning they feel drowsy and unrefreshed, and they prefer to get up late if they can.^6,¹⁶ Early polygraphic studies seemed to support this concept,^31,³² but the sleep EEG recordings were discontinuous in one study,³² which does not allow interpretation of the sleep structure. Other polysomnographic sleep studies in patients with idiopathic generalized epilepsy differentiated more and found normal sleep patterns unless seizures occurred during the night.³³^–³⁵ These studies were performed either on patients with chronic antiepileptic medication or on patients in whom the drugs (usually phenobarbital and phenytoin) were discontinued a few days prior to the sleep investigations. Thus, chronic drug effects or rebound effects after discontinuation, which may influence the results, cannot be excluded. Only one study³⁶ investigated the night sleep of unmedicated epilepsy patients. Adaptation nights to the sleep lab were also included in this study³⁶ to avoid “first-night” effects.³⁷ Photosensitive patients with generalized epilepsy had significantly more deep sleep (sleep stages 3 and 4) and less light sleep (sleep stages 1 and 2) than the other patients with generalized epilepsy.³⁶ However, this effect was no more significant if the patients were age matched, as the photosensitive patients were younger, and younger patients tend to have more deep sleep. Thus, the systematic polysomnographic evaluation of unmedicated patients with generalized epilepsy³⁶ did not reveal distinct sleep patterns in patients with “awakening” and “sleep” epilepsy, as hypothesized by Janz⁶ based on unstructured interviews.

SLEEP IN PATIENTS WITH FOCAL EPILEPSIES

Patients with focal epilepsy have poorer sleep as compared to controls in many categories: epilepsy patients sleep less, have a poorer quality of sleep, and report a disturbed sleep.³⁰ This was found not only for patients with catastrophic epilepsy, but also already for patients who are relatively well controlled on 1 or 2 AED.²⁹ Deep sleep is reduced and there are more arousals and more stage shifts,³⁸ especially in nights with seizures, but also in seizure-free nights. REM sleep is decreased if seizures occur during the night as well as during the preceding day. The earlier the seizures appear during the night, the shorter the total REM time,³⁹ but the amount of REM sleep increases with better seizure control.⁴⁰ Interictal epileptiform discharges (IEDs) alone seem to have little effect on the sleep architecture.⁴¹

Effects of Sleep and Sleep Deprivation on Epilepsy

Langdon-Down and Brain ⁵ were the first to subdivide epilepsy patients according to the occurrence of their seizures. They described (1) a “diurnal type,” whose seizure occurred predominantly during the day with a maximum following morning awakening and two smaller peaks in the afternoon; (2) a “nocturnal type,” in whom seizures occurred during the night with maxima shortly after falling asleep and in the early morning hours; and (3) a group without any discernible pattern (“diffuse type”).⁵ Relations of epileptic seizures to the sleep–wake cycle were evident from the observation that the times at which the seizures occurred changed when the sleep regimen was altered.^3,⁴²

The circadian sleep–wake cycle also significantly influences the occurrence of epileptiform discharges.¹¹ During non-REM sleep, generalized epileptiform discharges are more frequent than during waking.^9,^10,⁴³ For this reason, sleep EEGs are performed on patients in whom the diagnosis of epilepsy is not established or the epilepsy syndrome is unclear and the waking EEG is unrevealing.

RELATION OF EPILEPTIFORM DISCHARGES TO THE SLEEP-WAKE CYCLE AND SLEEP STAGES IN GENERALIZED EPILEPSY

Generalized epileptiform discharges gradually increase with deepening of non-REM sleep ^34,⁴⁴ In patients with absence epilepsy, the lowest discharge rates were found during REM sleep.³⁴ Because deep sleep stages are most pronounced during the first sleep cycle, it is not surprising that the highest rate of interictal epileptiform discharges were found during the first sleep cycle.³⁴ The morphology of generalized spike-wave complexes is more irregular during non-REM sleep and is similar to waking and REM sleep.^34,^35,⁴⁴

The early hypothesis ⁸ that transitional states between wake and sleep and vice versa may be epileptogenic in selected patients was supported by polygraphic studies. Several authors demonstrated the facilitating effects of transitional states of sleep such as sleep onset, changes between sleep stages, and arousals on the occurrence of epileptiform discharges in patients with absence epilepsy.^12,^33,³⁵ Niedermeyer⁴⁵ emphasized that an abnormal paroxysmal response to arousal and the influx of upward traveling stimuli seem to be the most important epileptogenic mechanism in primary generalized epilepsy. The concept of CAP, which is based on these observations, provides a new approach and supports the idea that transitional cyclic sleep patterns activate epileptiform discharges.¹³

RELATIONS OF EPILEPTIC SEIZURES AND SLEEP IN GENERALIZED EPILEPSIES

Generalized tonic-clonic seizures show a clearer relation to the sleep–wake cycle than “minor seizures.”⁴⁶ However, myoclonic seizures tend to occur predominantly in the early morning hours shortly after awakening from night sleep in juvenile myoclonic epilepsy.^14,⁴⁷ Based on the patient’s recollection of seizures and the patient’s history, Janz⁴⁶ classified epilepsies according to the time of occurrence of the “grand mal” seizures. He coined the term awakening epilepsy for patients whose generalized tonic-clonic seizures predominantly occur in the first 2 hours after awakening, with a second peak in the afternoon.⁴⁶ Janz⁶ reported earlier that in 45% of his 2110 patients with grand mal epilepsy, seizures occurred predominantly during sleep, in 34% during the first 2 hours after awakening from night sleep, and in 21% no relation to the sleep–wake cycle was found. During the course of the epilepsy, the pattern may change as less patients have an awakening predominance (31%) and more patients show a diffuse pattern (26%).¹⁶ The awakening type seemed to be associated with idiopathic generalized epilepsy.⁴⁶ Half of Janz’s patients with hereditary idiopathic epilepsy belonged to the awakening group, whereas most patients with focal epilepsy secondary to brain tumors (53%) had a diffuse pattern of seizure occurrence, and only 8% showed an awakening pattern.⁴⁶ The minor seizures in patients with epilepsy on awaking comprised predominantly absence seizures (94%) and myoclonic seizures (96%) and less frequently psychomotor seizures (16%) or Jacksonian seizures (9%).⁴⁶ A major bias of the study of is the observation that the recollection of epileptic seizures by the patients as well as caregivers is unreliable.⁴⁸ Also, the etiological diagnoses are subject to uncertainties in the study of Janz⁴⁶ because at that time no modern imaging studies were available. Few recent studies dealt with the relation of the occurrence of generalized tonic-clonic seizures to the sleep–wake cycle.^49,⁵⁰ The distribution of generalized tonic-clonic seizures was as follows in 77 of 141 patients with generalized epilepsy, in whom generalized tonic-clonic were observed:⁴⁹ (1) 16.8% had seizures after awakening, (2) 36.3% had seizures during waking hours of the day, (3) 28.5% had seizures during night sleep, and (4) 18.1% had seizures both during night sleep and during waking periods of the day. Morning and nocturnal awakenings accounted for 36 of 51 seizures in 33 patients with juvenile myoclonic epilepsy, whereas the evening relaxation period (n = 6), sleep onset (n = 3), and sleep (n = 6) were less frequently associated with seizures.⁵⁰