Clinical manifestations

Presentation and course

Other clinical features consist of marked autonomic activation, including flushing of the skin, diaphoresis, and mydriasis. The heart rate rises; tachypnea and increased respiratory amplitude are common (50; 02; 39).

The patient may feel anxious, fearful, or terrified and is generally inconsolable until the intense agitation dissipates. A sense of respiratory oppression or of doom or impending death may accompany the episode. As the patient calms down at the end of the episode, sleep follows easily, and the episodes usually do not occur more than once per night (50). Sleep-related violence during disorders of arousal episodes or potentially harmful behaviors, which are rare in childhood, could appear in adults (31).

Recall of the episode either immediately after or the next morning is usually absent or limited to a single, unpleasant frightening visual scene, and patients, especially children, usually do not remember them in the morning (05). In contrast with REM sleep behavior disorder, disorders of arousal, including sleep terrors are often considered as nondreaming states as they emerge from N3 sleep; however, recent studies have challenged this view (48; 19). If children usually could not recall any mental activity associated with their episodes, many adults recalled at least one mental experience. The content of the collected reports was dominated by dynamic actions acted out from a self-perspective, often with apprehension and in response to misfortune and danger, in a home-setting environment (09).

Sleep terror episodes can be precipitated by fever, sleep deprivation, unusual circadian rhythms, or the use of central nervous system depressant medications (39; 16). Attacks can often be triggered by forced arousals in slow-wave sleep. Young children with poorly controlled sleep disorders (eg, obstructive sleep apnea) or medical conditions, such as esophageal reflux or asthma, had more sleep disruptions, including more frequent sleep terrors, than children without (38; 16).

Prognosis and complications

In children, sleep terrors are usually benign and become less frequent with age, tending to resolve spontaneously during adolescence, perhaps because the amount of time spent in delta-wave sleep decreases with age (39). When sleep terror episodes include sleepwalking activity, the risk of injuries to oneself or to others is increased due to the lack of control on the behavior and on the real environment and transient indifference to pain (19).

In children and adolescents with disorders of arousal, behavioral and emotional problems are surprisingly common, and the severity of emotional or behavioral problems is positively correlated with the severity of the nocturnal episodes (12).

A significant bidirectional association between depressive symptoms and disturbed sleep in children and adolescents was found (37). If disturbed sleep is associated with the consolidation of depressive symptoms starting in childhood, which, in turn, is associated with ongoing sleep problems, timely and appropriate interventions could prevent spiraling effects on both domains.

Clinical vignette

A 7-year-old child presented with a 3-year history of sudden awakenings, usually 1 hour after sleep onset. The child’s medical history was unremarkable. Family history revealed that a maternal cousin had experienced sleep terror episodes in the past. During the episodes the patient sat up with a fearful expression and glassy eyes, vocalized, screamed, and shook; his body seemed fighting against an unknown danger. He was usually inconsolable, unresponsive to external stimuli, and was difficult to waken and calm. He had tachycardia, tachypnea, flushing of the skin, and mydriasis. The patient was confused and disoriented if awakened. Often the child might be able to provide only an indication of fear, but return to sleep might occur without achieving full waking consciousness, and morning amnesia for the whole event was the rule. The episodes occurred twice monthly, lasting 3 to 5 minutes; during febrile illness, they were more frequent and prolonged. Neurologic examination was negative. Based on the characteristic history, he was diagnosed with sleep terrors. No pharmacologic treatment was begun because the attacks did not produce harm or injury; parents and child were reassured that the episodes are generally self-limiting, decreasing in frequency, and cease during adolescence or young adulthood.

Biological basis

Etiology and pathogenesis

A number of genetic, developmental, psychological, and organic risk factors have been identified for sleep terrors.

Genetic factors. There is clear evidence that many patients with sleep terrors have a family history of sleep terrors, sleepwalking, or both, supporting the hypothesis that sleepwalking and night terrors share a common strong genetic predisposition (41; 33). Still, no genes have been identified in family pedigrees, and the most probable mode of inheritance is considered to be multifactorial (33). Studies of twins strongly support the heritability of sleep terrors, with a higher correlation between monozygotic rather than dizygotic twins of patients with sleep terrors. In addition, it has been proposed that the human leukocyte antigen (HLA) DQB1*04 and HLA DQB1*05:01 alleles might explain at least part of this genetic susceptibility to NREM parasomnias (16), suggesting a possible involvement of immune-related mechanisms in motor control during sleep.

Developmental factors. The common pattern of onset is in childhood, and termination by late adolescence implicates a developmental factor. An immature form of some cortical GABAergic and cholinergic inhibitory circuits might be ineffective in stopping movements during sleep (16). It has also been suggested that disorders of arousal, including sleep terrors, may be associated with a dysfunction in the serotoninergic system, which plays a crucial role in generating slow wave sleep, arousal, and control of motor activity.

Psychological factors. Sleep terrors are common in patients with posttraumatic stress disorder and are probably increased in children raised in violent, abusive families. Some studies reported a significant association between anxiety level and sleep terror episodes. Children and adolescents with disorders of arousal exhibit specific psychobiological personality traits compared to age- and gender-matched control subjects (47), but psychopathology is extremely rare in children despite the impressive intensity of the attacks.

Finally, Boyden and colleagues proposed that sleep terrors might be an extreme response to the evolutionary–environmental mismatch that has resulted from changes in sleeping behavior, from co-sleeping, with decreased risk of predation of children having parents nearby, to sleeping separately (04; 16).

Organic factors. Sleep terrors occur more often in children with sleep-disordered breathing than in normal children and are increased in patients with obstructive sleep apnea, especially during nasal continuous positive airway pressure therapy, and in those who consume alcohol at bedtime or enjoy intense evening physical activities that promote increased slow wave sleep and NREM sleep instability. The prevalence of parasomnias, including sleep terrors, is greater among children with neurodevelopmental disorders, such as Down syndrome and cerebral palsy. These patients are more vulnerable to upper airway obstruction that can cause repeated arousals from sleep, triggering arousal disorders. Nevsimalova and colleagues suggested that childhood parasomnias, including sleep terror, can be regarded as a disorder of sleep maturation because they are frequently associated with perinatal risk factors and developmental comorbidities (40).

Sleep terrors starting in adulthood can be symptomatic of neurologic diseases. There are isolated documented cases of sleep terrors caused by a brainstem lesion or a thalamic tumor.

The pathogenesis of sleep terrors is unknown. Sleep terror is classified as a disorder of arousal based on the concept that disordered arousal mechanisms lead to behavioral and emotional activation but not to a full awakening (10; 39; 08; 32). Sleep disorders that are known to trigger arousals, like sleep-disordered breathing, may cause sleep terror in children. Sleep terrors usually occur during deep non-REM sleep (stage 3). High-voltage slow-wave activity may be seen immediately prior to the episode, and the EEG during an episode may show diffuse, hypersynchronous rhythmic delta, diffuse delta with intermixed faster frequencies in the theta and alpha range, or prominent alpha and beta activity.

Some authors failed to find a “delta wave build-up” prior to an arousal disorder, suggesting that this EEG pattern does not appear to be specific for an arousal disorder episode (45). Intracerebral EEG studies suggest that arousal disorders could be dissociated arousal states due to the coexistence of different, local, cerebral states of being (45; 02). Scalp EEG analysis reveals a localized decrease in slow wave activity over centro-parietal regions relative to the rest of the brain in patients with arousal disorders compared to good-sleeping healthy controls; also, these differences in local sleep were present in the absence of any detectable clinical or electrophysiological signs of arousal (11). These topographical changes in local EEG power persist during REM sleep and wakefulness, suggesting a trait-like functional change that crosses the boundaries of NREM sleep. Spectral analysis over 325 episodes of disorders of arousal showed an absolute significant increase in all frequency bands prior to episodes of disorders of arousal, excluding sigma, which displayed the opposite effect. In normalized maps, the increase was relatively higher over the central/anterior areas for both slow and fast frequency bands. Taken together, these results show that deep sleep and wake-like EEG rhythms coexist over overlapping areas before episodes of disorders of arousal, suggesting an alteration of local sleep mechanisms. Episodes of different complexity are preceded by a similar EEG activation, implying that they possibly share a similar pathophysiology (32).

Studies on adults suggest that an abnormal deep sleep associated with a high slow-wave sleep fragmentation might be responsible for the occurrence of sleep terror episodes (30). A video-polysomnographic (V-PSG) assessment to quantify slow-wave sleep interruptions (slow-wave sleep fragmentation index, slow/mixed and fast arousal ratios, and indexes per hour) and the associated behaviors in 60 adult patients with disorder of arousal showed that slow-wave sleep fragmentation index and the mixed, slow, and slow/mixed arousal indexes and ratios were higher in patients with disorders of arousal than controls (30). Usually, the increased slow-wave sleep fragmentation observed in patients with sleep terrors is not associated with the level of daytime sleepiness (29). In fact, daytime sleepiness in adult patients with sleep terrors and sleepwalking episodes seems to be associated with a specific polygraphic phenotype (rapid sleep onset, long sleep time, lower numbers of awakenings on N3), which is suggestive of a higher sleep propensity that may contribute to incomplete awakening from deep sleep (07).

Evaluating autonomic reactions in a small group of adult patients with sleepwalking and sleep terrors, Ledard and colleagues found an autonomic arousal occurring 4 seconds before motor arousal from N3 sleep (with a higher adrenergic reaction than in controls), suggesting that an alarming event during sleep (possibly a worrying sleep mentation or a local subcortical arousal) causes the motor arousal (22).

Epidemiology

Approximately 1% to 6% of prepubertal children have recurrent sleep terrors, with a peak incidence between 5 and 7 years of age (23). Episodes tend to decrease in frequency or cease during early adolescence such that 50% of children no longer have attacks by the age of 8, suggesting a disorder of maturation of the nervous system. The high amount of slow wave sleep in preschool and school aged children could be a predisposing factor for the occurrence of disorders of arousal. The decrease of delta sleep due to synaptic pruning during adolescence may also account for the disappearance of sleep terrors at this age (16). However, in some patients, episodes begin in adolescence or early adulthood (02; 39). The prevalence of disorders of arousal was 7.1% among boys and 7.7% among girls in a nationwide survey conducted among Japanese adolescents (20). The prevalence in adults is about 1%. Sex and racial or cultural differences do not appear to affect prevalence, although some sources indicate that boys seem to be more frequently affected than girls (16).

Many individuals share the condition with one or more family members, and the increased prevalence in first-degree relatives suggests an autosomal dominant pattern of inheritance. Again, these findings predate the discovery of autosomal-dominant nocturnal frontal lobe epilepsy and may refer to misdiagnosed seizures.

Prevention

Sleep deprivation, a big delay in sleep/wake schedule such as that of confinement due to COVID-19, emotional stress, alcohol use at bedtime, and febrile illness can influence the frequency and severity of episodes in susceptible individuals (33; 06). Avoidance of these precipitants may help to prevent sleep terrors. In some patients, the premenstrual period may be associated with more frequent episodes (35; 39).

Poorer sleep quality in children is often associated with maternal depressive symptoms, and parasomnias are more prevalent among children of mothers with chronic symptoms of depression (17).

Differential diagnosis

Confusing conditions

The differential diagnosis includes sleep-related epilepsy, REM sleep behavior disorder, nightmares, confusional arousals, nocturnal panic attacks, nocturnal delirium, and other sleep disorders that produce anxiety, including obstructive sleep apnea and nocturnal cardiac ischemia. Sleep terrors pose particular problems in their differential diagnosis with the sleep-related epileptic seizures, particularly sleep-related hypermotor epilepsy in which attacks commonly occur without scalp EEG epileptic abnormalities (46; 25; 39). Generally, the distinction between sleep terrors and epileptic seizures is based on clinical criteria. Features favoring the diagnosis of sleep-related hypermotor epilepsy rather than sleep terrors are a high rate of same-night recurrence, the presence of dystonic-dyskinetic motor pattern during the attacks, their stereotypical motor behavior, their response to antiepileptic medication, and onset or persistence into adulthood (43; 46; 39). Sleep terrors typically occur within the first few hours of sleep, whereas seizures may occur throughout the night (see Table 1). Nocturnal complex partial seizures may also be associated with fearful appearance, screaming, running, tachycardia, and vague frightening perceptions. However, it is not always possible to distinguish sleep terrors from seizures on the basis of history alone, and video-polysomnographic recording is mandatory in cases in which episodes are frequent and persisting in young adulthood with a violent motor behavior (39). In association with the episodes’ semiological features, sleep stage and the relative time of occurrence of minor and major motor manifestations during sleep represent useful criteria to discriminate sleep-related hypermotor epilepsy and disorders of arousal (42; 25). Analyzing the "event distribution index" during video-polysomnography recordings of 89 patients with a definite diagnosis of disorders of arousal (59) or sleep-related hypermotor epilepsy (30), the occurrence of at least one major event outside N3 was highly suggestive for sleep-related hypermotor epilepsy. The occurrence of at least one minor event during N3 was highly suggestive for disorders of arousal (42).

Sleep terrors usually occur within the first few hours of sleep and arise out of NREM sleep stages 3 (N3), whereas nightmares occur out of REM sleep during the middle or latter half of the night. Nightmares are not usually accompanied by major motor activity or severe anxiety, vocalization, and autonomic discharge, and they are less likely to begin with an intense scream than sleep terrors. The sleeper is more easily aroused and, when awakened from a nightmare, exhibits good intellectual function. Afterward, dream content can be recalled in vivid detail (see Table 1). Some patients, however, have clinical features that overlap between nightmares and sleep terrors.

Table 1. Differential Diagnosis of Sleep Terror

Confusional arousals are awakenings from slow-wave sleep without terror or ambulation. Patients may fumble with bedclothes and mumble incoherently but do not exhibit intense autonomic arousal or flight reactions. The features of sleep terrors, sleepwalking, and confusional arousals often overlap. For example, sleepwalking episodes and confusional arousals may be associated with whimpering or crying, fearful behavior, and mild autonomic arousal.

In nocturnal dissociative disorder, the patient is awake when the episode begins and behavior is purposeful, more complex, and longer lasting. The EEG is consistent with wakefulness.

REM sleep behavior disorder may be associated with violent behavior, running, or screaming; however, autonomic activation is usually absent or mild with, for example, little or no tachycardia. Episodes tend to occur later in the night and are often associated with dream recall. Observers often report that patients seem to be "acting out" their dreams. In some patients, however, behavior may be similar to that observed with sleep terrors, and some patients have an "overlap syndrome” (parasomnia overlap disorder) with elements of sleep terrors, somnambulism, and REM sleep behavior disorder (49).

It is, however, important to remember that disorders of arousal may begin in adulthood and also persist or arise in older adults. Motor patterns of disorders of arousal in older adults are similar to those in younger patients. A combined clinical examination and video polysomnography recording are crucial in establishing a definitive diagnosis of nocturnal motor behavior in all older adults and especially in those affected by neurodegenerative diseases (26).

Nocturnal panic attacks may at times clinically resemble sleep terrors, and episodes may arise out of NREM sleep, usually stage 2 or 3. Psychopathology is common, and there are always similar episodes in daytime wakefulness, which is not true in patients with sleep terrors.

In a single case, screaming/yelling episodes were due to a sporadic insulinoma, suggesting the inclusion of hypoglycemia in the differential diagnosis of new-onset or worsening seizures or night terrors (03).

Associated or underlying disorders

Comorbidities include obstructive sleep apnea, periodic limb movements during sleep, insomnia, restless legs syndrome, REM sleep behavior disorder (considered parasomnia overlap), and epilepsy (24).

Diagnostic workup

An accurate clinical interview is theoretically sufficient to confirm a diagnosis of disorders of arousal according to standard international criteria (34).

For otherwise normal children with typical behaviors occurring during the first third of the night, the diagnosis can usually be made based on clinical criteria. The International Classification of Sleep Disorders, third edition, criteria are adequate for a reliable diagnosis of disorders of arousal in adulthood, too (27). When the diagnosis is uncertain, video-polygraphic monitoring is indicated, particularly if the events are occurring several times per week. If nocturnal seizures are a diagnostic consideration, multiple EEG channels should be included in the recording montage, and the paper speed could be sufficient to identify epileptiform and ictal EEG activity (15 to 30 mm per second). A synchronized video recording of the patient is useful to observe the clinical manifestations of the motor attacks and their stereotypy if more than one episode is recorded (43; 39). Standard polysomnographic recording techniques used for identifying sleep apnea are generally insufficient when complex partial seizures are a consideration.

Polysomnographic monitoring usually demonstrates that sleep terrors consist of sudden and incomplete arousal from deep sleep. The onset of sleep terror episodes is usually within the first few hours of sleep, during stage 3 sleep. Prior to a sleep terror episode, the EEG may show high-voltage, generalized symmetrical, hypersynchronous slow-wave activity. During the episode, it often shows a regular, rhythmic delta activity pattern, associated with a marked increase in muscle tone and change in respiratory and heart rate. In these patients beyond full sleep terrors, partial arousals from slow-wave sleep without full terror are also common, and tachycardia usually occurs during both clinical episodes of sleep terror and partial arousals. Psychogenic dissociative episodes are associated with a waking EEG pattern.

In addition, the analysis of homemade video recordings of nocturnal episodes may be an important supportive diagnostic tool for disorders of arousal (34). Homemade video offers multiple advantages: wide availability, low cost, the possibility of recording patients in their usual sleep environment, and repeated recordings (31). Lopez and colleagues documented how home nocturnal infrared video recording has good feasibility and acceptability and may improve the evaluation of the phenotype and severity of nonrapid eye movement parasomnias and of treatment response in an ecological setting (28).

Psychiatric or psychological evaluation may be indicated in selected patients.

Management

As frightening as they are, parents and children usually only need to be reassured that the episodes are generally self-limiting and that the attacks rarely produce harm or injury. Attempts should be made to alleviate whatever stress may be going on in the child’s environment and to ensure that the child is getting adequate rest. When the behavior has the potential of injury or causes major disruption of family life, the pharmacological treatment most commonly includes melatonin and benzodiazepines; imipramine, mirtazapine, or ramelteon may be beneficial (18; 44; 13; 36).

Melatonin was reported to be efficacious and may serve as an initial pharmacological treatment strategy for patients with various phenotypes. Melatonin could improve the underlying circadian misalignment, which is usually a potential component to the development of NREM parasomnias (24). In other cases, melatonin may have been effective via partial treatment of sleep deprivation or insomnia, which could function as a precipitating factor for NREM episodes. Benzodiazepines are also helpful and can be used for long intervals with few complications in most patients; however, when the drug is withdrawn, the relapse rate is high. Among benzodiazepines, the most frequently used is clonazepam (0.5 to 2 mg at bedtime), which was reported to be effective in 40% of those treated (24). Diazepam (5 to 10 mg before retiring) could be useful, but to avoid daytime sedation due to the long half-life of diazepam and its metabolites, better results have been reported with shorter-acting benzodiazepines such as midazolam and oxazepam at the usual evening doses of 10 to 20 mg. The proposed mechanism of benzodiazepines in the treatment of disorders of arousal is reducing slow wave sleep, which NREM parasomnias generally arise from, or by increasing the arousal threshold, decreasing arousal, and stabilizing sleep. Appropriate caution in the use of benzodiazepines is recommended for patients with concurrent disorders, including advanced age, fall risks, abuse potential, and sleep-disordered breathing, as these outcomes could worsen respiratory depression and may result in the worsening of apnea and NREM parasomnia events (24).

Imipramine 50 to 100 mg at bedtime is sometimes effective, and hypnosis or other behavioral treatment may be helpful for some patients. A 6-month course of medication, followed by gradual withdrawal, is a typical program. In the wake of the hypothesis that manifestations of arousal disorders such as sleep terrors may be due to a conflict between the mechanisms generating slow-wave sleep and arousal, dependent on a dysfunction in the serotoninergic system, L-5-hydroxytryptophan, a precursor of serotonin, (2 mg/Kg at bedtime) has been proposed as highly effective in reducing the number of sleep terror episodes.

Treating comorbid conditions (eg, obstructive sleep apnea) is a recommended treatment strategy that is often associated with symptom improvement.

For adults, stress reduction through psychotherapy may be helpful when the disorder is linked to significant psychopathology, although studies evaluating the efficacy of different psychological approaches in adults (hypnosis, relaxation therapy, or cognitive behavioral therapy) provided contrasting results (36).

Special considerations

Pregnancy

The incidence of NREM parasomnias in pregnancy has not been systematically investigated (19). Few papers describe a decrease or an exacerbation of sleepwalking or sleep terrors during pregnancy. Studies comparing pregnant with nonpregnant women found no significant changes in the prevalence of NREM parasomnias in pregnancy. A survey of 325 women, using an online questionnaire and focusing on the prevalence of parasomnias 3 months before pregnancy, during pregnancy, and 3 months after delivery, documented a significant increase in the severity and frequency of sleepwalking, night terrors, vivid dreams, and nightmares during pregnancy (19). Although it could be difficult to distinguish NREM or REM parasomnia episodes without a video-polysomnographic recording of an episode, the increased prevalence of parasomnias during pregnancy needs to be targeted, especially by nonpharmacological approaches. The increased psychosocial stress associated with pregnancy as well as the impaired sleep quality and the increased incidence of sleep disorders, in particular sleep-disordered breathing, may be the cause of the parasomnia episodes among pregnant women.