Sleep and Sleep Disorders in Epilepsy

Introduction

Sleep problems and primary sleep disorders are far more likely to be found in children with epilepsy (CWE) than the general pediatric population.1–14 The etiology of sleep disruption in children with epilepsy may be multifactorial including: epilepsy per se, frequent nocturnal seizures disrupting nocturnal sleep organization, effects of antiepileptic medications on daytime alertness and nighttime sleep, and treatable primary sleep disorders. Comorbidities such as physical disability,⁸ intellectual disability,^2,15,16 neurodevelopmental syndromes,^17,18 autism spectrum disorder,¹⁹ and behavioral disorders^1,7–9,20 may add to the likelihood of sleep disorders in a child with epilepsy. In CWE, they are associated with negative effects on daytime behavior and academic performance.^21–23 Recognizing this has led to increasing numbers of children or adolescents referred to sleep specialists to evaluate whether undiagnosed sleep disorders are contributing to their seizures.

Questionnaire-Based Studies on the Prevalence of Sleep Disorders in Children with Epilepsy

Children with epilepsy (CWE) are much more likely to have sleep problems than the general pediatric population.1–10,24 Sleep problems in CWE may be due to varying combinations of: epilepsy per se, nocturnal seizures disrupting nocturnal sleep organization, effects of antiepileptic medications on daytime alertness and nighttime sleep, and treatable primary sleep disorders. Worse yet, comorbidities such as physical disability,⁸ intellectual disability,^2,15,16 neurodevelopmental syndromes,^17,18 autism spectrum disorder,¹⁹ and behavioral disorders^1,7–9,20 often increase the likelihood of sleep disorders in CWE.

A recently published study prospectively explored the relationships between the effect of pediatric epilepsy on child sleep, parental sleep and fatigue, and household sleeping arrangements in 105 households with a child with epilepsy and 79 controls.²⁵ Using multiple different validated pediatric sleep questionnaires, the authors found: (1) increased rates of both parent–child room sharing and cosleeping compared to controls; (2) CWE had significantly more sleep disturbances, especially for parasomnias, nocturnal awakenings, sleep duration, daytime sleepiness, sleep onset delay and bedtime resistance than controls. Parents of CWE were more likely to report fatigue and sleep dysfunction. Severity of a child’s epilepsy correlated positively with the degree of child and parent sleep dysfunction and parental fatigue. Antiepileptic drug polytherapy predicted that greater childhood sleep disturbances would be reported. Child sleep problems were associated with room sharing and cosleeping. Sixty-two percent of parents described decreased quantity and/or quality of sleep when cosleeping. Forty-four percent of parents reported rarely or never feeling rested because they were concerned about their children having seizures during sleep.

An earlier prospective study found 89 children with idiopathic partial or generalized epilepsy (in whom seizures were more often well controlled and not associated with comorbidities) were significantly more likely to have more sleep problems than their 49 siblings or 321 age-matched healthy controls.¹ Post-hoc comparisons found the CWE had significantly more excessive daytime sleepiness (EDS), bedtime difficulties, sleep fragmentation, and parasomnias than their siblings or controls. Multiple regression analysis showed that sleep complaints, longer sleep latencies, and shorter sleep times were much more likely to be found in the children whose seizures were poorly controlled. Furthermore, daytime seizures and high nighttime interictal epileptiform discharge (IED) rates predicted daytime drowsiness, and explained 15% of its variance. Sleep problems in these CWE greatly increased the likelihood they would have far more behavior problems (inattention, hyperactivity, impulsivity, oppositional defiant disorder) than their siblings or controls. Three variables (age, higher rates of IEDs during sleep, and length of freedom from seizures) accounted for 24% of the variance of which CWE were at greatest risk.

Juvenile myoclonic epilepsy (JME) is one of the most common adolescent-onset epilepsies. Patients with JME typically have generalized convulsive motor seizures heralded by myoclonic jerks which typically occur soon after awakening. Seizures are often triggered by sleep deprivation with or without alcohol (‘activation by celebration’). Krishnan et al. prospectively studied the effect of epilepsy on sleep in patients with JME, comparing them with age- and gender-matched controls using detailed clinical assessment, EEG, neuroimaging and multiple different sleep questionnaires.²⁶ They found that patients with JME had significant sleep disturbances characterized by EDS and disturbed nocturnal sleep despite adequate AED therapy and good seizure control.

Another case-control study found sleep problems were twofold higher (mean 4 ± 3 vs. 2 ± 2 ) in 79 CWE (mean age 10 ± 3 years) compared with 73 age- and gender-matched controls. Other questionnaire-based studies have demonstrated: (1) symptoms of OSA were 15 times more likely to be reported by the parents of 26 children with epilepsy (mean age 15 years) than a similar number of healthy controls (65% vs. 4%); ((2) children with epilepsy compared with control subjects had more daytime sleepiness, less on-task behavior, and less attention;¹⁶ and (3) children with benign rolandic epilepsy had significantly shorter sleep duration, more frequent parasomnias, and daytime sleepiness than a reference sample of children.^5,24

Poor sleep hygiene may contribute to sleep problems in CWE. A prospective case-control study found 121 CWE needed to be put to bed by their parents, take an afternoon nap, awaken during the night, take >30 minutes to fall asleep, express fear of the dark, awake complaining of a distressing dream or worry, call for parent(s) during the night and/or visit the parental bed.² These studies have further found that: (1) CWE whose seizures are poorly controlled are more likely to have poor sleep habits compared to those whose seizures are controlled; and (2) CWE whose seizures occur primarily during sleep are significantly more likely to have difficulty falling asleep, afternoon napping, nocturnal awakenings, awakening with fear or a dream, calling out for their parents at night, or visiting the parental bed when compared to CWE whose seizures occur primary when awake.

Is Sleep Architecture Abnormal in Children with Epilepsy?

Pereira et al. recorded a single night of polysomnography (PSG) in 31 CWE who had drug-resistant epilepsies comparing sleep architecture in them with a group of normal age-matched healthy controls and a group of children with benign rolandic epilepsy.²⁷ Compared to normal controls, the children with drug-resistant epilepsy showed a significant reduction of time in bed, total sleep time, rapid eye movement (REM) sleep, sleep stage NREM 3, and sleep efficiency, and a significant increase in wake after sleep onset. Compared to children with benign rolandic epilepsy, those with drug-resistant epilepsy had greater reduction in NREM 3, REM sleep and sleep efficiency.

Other small case-control studies 28–32 recording in-laboratory PSG have shown that abnormalities in sleep architecture are significantly more likely to be found in CWE compared to healthy age-matched controls. Findings among the case-control studies were: (1) children who had partial seizure(s) during a comprehensive in-laboratory overnight PSG with 24-channel EEG had significantly less time in bed and sleep time compared to CWE who did not have seizures during the study or normal controls;³¹ (2) children with primary generalized epilepsy whose seizures were well controlled still had significantly more NREM 1 and longer REM sleep latency than normal controls;³² (3) compared to controls, children with Lennox–Gastaut syndrome (a severe childhood-onset epileptic encephalopathy with medically refractory seizures) had reduced percentages of REM, NREM 2 and NREM 3 sleep;³³ (4) children with benign rolandic epilepsy (who often have few or no seizures) still had reduced total sleep time, sleep efficiency, and percent REM sleep compared to controls;³⁰ (5) compared to controls, children with intellectual disability and epilepsy (mean age 13 ± 4 years) had longer sleep latency, higher percentage of wake after sleep onset and NREM 3 sleep, lower sleep efficiency, more awakenings and stage shifts, a higher CAP rate, increased A1 index, long and less numerous CAP sequences;³⁴ and (6) abnormalities in the density and frequency of sleep spindles were observed in children with untreated epilepsy compared to CWE whose seizures were treated or healthy controls.³⁵ However, two studies found normal sleep architecture in overnight PSG in children with absence epilepsy (and whose petit mal seizures typically occur when awake).^36,37

Another small case-control study compared PSG findings in 10 children (mean age 11) with tuberous sclerosis complex (TSC) and 10 healthy controls.³⁸ They found the children with TSC more often had lower sleep efficiency (ranging from 60% to 88%), increased WASO (>10%), more nocturnal awakenings and stage shifts, poorly organized sleep cycles, increased NREM 1, and decreased REM sleep compared with controls. Furthermore, sleep architecture was significantly more disrupted in the three children who had at least one seizure recorded during their PSG (sleep efficiency 69%, WASO 24%, mean awakenings 16/h) compared to the TS children who did not (sleep efficiency 88%, WASO 5%, and mean awakenings 3/h).

Prevalence of Sleep Apnea on Sleep Studies in Children with Epilepsy

A retrospective case-control study compared PSG findings in 40 CWE referred for suspected obstructive sleep apnea (OSA) with 11 children who had moderate pediatric OSA (pediatric obstructive apnea–hypopnea indexes (PAOHI)) of 5–10/h of sleep.²⁸ Twenty percent of the CWE had OSA (PAOHI >1/h), 33% obstructive hypoventilation, 8% upper airway resistance syndrome, 18% primary snoring, and 10% periodic limb movements (>5/h). These findings are not particularly revelatory, given the children were symptomatic. The CWE and OSA compared with those with uncomplicated moderate OSA had a higher body mass index (BMI), were more likely to be obese (BMI >95th percentile in 62% vs. 18%), had longer sleep latencies (51 vs. 16 min), higher arousal indexes, and lower nadir SpO₂ (86% vs. 90%) even though their mean PAOHI was only 3/h compared to 7/h in the children with uncomplicated moderate OSA. The investigators further found CWE whose seizures were poorly controlled had significantly lower sleep efficiency, a higher arousal index, and a higher percentage of REM sleep compared with children who were seizure-free or exhibited good seizure control.

A recently published small prospective study found OSA was more likely to be found in CWE who were on multiple antiepileptic medications and whose seizures were poorly controlled.³⁹ Uncontrolled epilepsy was a risk factor for OSA (80%) compared with primary snoring (47%, P = 0.02). The obstructive apnea–hypopnea index increased with increasing number of AEDs. Based on these findings, the investigators argue that children with uncontrolled seizures on multiple AEDs should be routinely screened for obstructive sleep apnea.

Impact of Treating Obstructive Sleep Apnea in Children with Epilepsy

Few studies have examined the effects of treating OSA in children with epilepsy. A recent retrospective analysis by Segal et al. evaluated the effects of adenotonsillectomy on seizure frequency in 27 CWE (median age 5 years).⁴⁰ Three months after surgery they found median reduction in seizure frequency for the group was 53%; 10 (37%) patients became seizure-free, 3 (11%) had a >50% reduction in the frequency of their seizures, 6 (22%) showed lesser degrees of improvement. Seizure frequency was unchanged in 2 (7%) and worse in 6 (22%). Multivariate analysis demonstrated a trend toward seizure freedom with each percentile increase in BMI and early age of surgery. A 45% or greater reduction in seizure frequency was found in three of six adults with epilepsy who used CPAP, and ≥60% in one of three children who tolerated it.⁴¹

Sleep–Wake and Circadian Patterns and Types of Seizures in Children

Three retrospective studies from the same research group have recently been published investigating sleep–wake, day/night and 24-h periodicity of different types of seizures in children with epilepsy.42–44 Among these papers, the investigators reported that: (1) tonic seizures were more frequently seen in sleep; (2) clonic seizures awake between 6 to 9 a.m. and 12 to 3 p.m.; (3) absence seizures 9 a.m. to 12 p.m. and 6 p.m. to 12 a.m. primarily awake; (4) atonic seizures between 12 to 6 p.m.; (5) myoclonic seizures occurred in wakefulness (6 a.m. to noon);⁴² (6) primarily generalized tonic–clonic seizures 9 a.m. and 12 p.m.; (7) focal-onset generalized convulsions emanating from the temporal lobe were more likely to occur when awake, extratemporal during sleep;⁴³ and (8) infantile (epileptic) spasms in children less than age 3 between 9 a.m. and 12 p.m. and 3 to 6 p.m., 6 a.m. to 9 a.m. in children older than age 3.⁴⁴

Nrem Parasomnias are More Common in Patients with Nocturnal Frontal Lobe Epilepsy

Nocturnal frontal lobe seizures in children are often initially misdiagnosed as sleepwalking, sleep terrors, nightmares or even a psychiatric problem.45,46 A case series found frontal lobe seizures in 22 children occurred almost exclusively during sleep in 77%, were brief (30 seconds to 2 minutes), frequent (3–22 per night) and characterized by a sudden arousal from NREM 2 sleep accompanied by screaming, agitation, dystonic posturing, kicking or bicycling of the legs, and/or urinary incontinence.⁴⁵ Interictal EEGs were normal in 86% of the children, although frontal or bifrontal electrographic seizure activity accompanying some of the frontal lobe seizures was seen in 95%. Given these clinical and EEG features, it is understandable why they are often initially misdiagnosed as sleepwalking, sleep terrors, nightmares or even a psychiatric problem.^45,46 Adding to the diagnostic challenge, NREM arousal disorders and sleep bruxism are significantly more common in patients and their relatives with NFLE.

A prospective case-control study found a higher incidence of parasomnias among 89 children with idiopathic epilepsy compared with 49 siblings and 321 healthy control children using parental sleep questionnaires. A recent retrospective study found the lifetime risk for NREM arousal disorders was sixfold greater in individuals with NFLE and fivefold higher for sleep bruxism compared to controls.⁴⁷ Among relatives of those with NFLE, the lifetime prevalence of a NREM arousal disorder was 4.7 times greater and nightmares 2.6 times higher compared to the relatives of control subject. NREM arousal disorders which are frequent (≥2–3 per week) warrant overnight PSG, which often identify another primary sleep disorder (most often OSA, less often RLS or PLMD). If OSA is found with PSG, tonsillectomy often eliminates both sleep disorders.⁴⁸

Indications for Video-Polysomnography in Children with Suspected or Known Epilepsy

Comprehensive video-PSG is most often done in CWE for suspected OSA or to identify primary sleep disorders contributing to complaints of sleepiness, fragmented sleep and/or poorly controlled seizures. When symptoms suggest this, a PSG is warranted.⁴⁹ Sometimes, we are asked to confirm whether paroxysmal nocturnal behaviors are epileptic or not in children with or without known epilepsy. The majority of children referred for parasomnias have NREM sleep disorders of arousal. This is not surprising since a recently published longitudinal study of child development reported an overall prevalence of 39.8% for sleep terrors and 14.5% for sleepwalking in children 6 years or younger.⁵⁰

The American Academy of Sleep Medicine (AASM) practice parameters for PSG suggest a PSG is unnecessary if the nocturnal behavior events are typical, non-injurious, infrequent, and not disruptive to the child or family.⁵¹ Common, uncomplicated, non-injurious parasomnias (such as typical disorders of arousal, nightmares, enuresis, sleeptalking and bruxism) can usually be diagnosed by a clinical history. One caveat, though: unusually frequent sleep terrors or sleepwalking events (more than 2–3 times per week) warrant PSG to identify another sleep disorder precipitating them (most often OSA, occasionally PLMD). OSA was found on overnight PSG in 58% of 84 prepubertal children who had sleep terrors and/or sleepwalking.⁴⁸ OSA and parasomnias were eliminated in 43 who had tonsillectomy. Two had restless legs syndrome and treatment of it with pramipexole eliminated the confusional arousals, restless legs and PLMS.

A PSG is indicated if the paroxysmal nocturnal events are atypical, frequent, potentially injurious, and/or disruptive to patient or family.⁵¹ The AASM recently published practice parameters⁵² for non-respiratory indications for PSG in children, accompanied by an evidence-based review justifying these.⁵³ These recommend a PSG may help differentiate atypical paroxysmal nocturnal behaviors from nocturnal seizures or to identify when sleep-disordered breathing or other sleep disorders contribute to frequent parasomnias, enuresis or affect control of seizures. Box 44.1 summarizes the differential diagnosis of nocturnal behavioral events in children. Box 44.2 provides a summary of the clinical features which are typical for a NREM arousal disorder. Box 44.3 summarizes red flags for atypical parasomnias which warrant consideration of video-PSG.

Box 44-1 Differential Diagnosis of Nocturnal Behavioral Events in a Child