Parasomnias: the spectrum of things that go bump in the night
From ghoulies and ghosties, And three leggity beasties, And things that go bump in the night, Good Lord, deliver us. Old Scottish Prayer
The traditional view that parasomnias reflect the unwelcome intrusion of some nocturnal spirit or supernatural force into sleeping people has largely been superseded. However, sleep-related behavioural phenomena are still a common cause of considerable anguish and perplexity to the sufferers, their bed partners, and clinicians alike. A common misperception among the last is that detailed and expensive investigations are invariably needed for precise diagnosis. However, I hope this article will demonstrate that a working knowledge of parasomnias and the spectrum of their presentation will allow a confident diagnosis and treatment, if needed, in most cases from the history alone. In complex situations, especially if there is no witness or bed partner, the input from a sleep clinic experienced in managing parasomnias is undoubtedly helpful. But in practice, it is relatively rare for overnight polysomnography complete with full EEG recording, video monitoring, and multiple limb EMG channels to provide additional worthwhile diagnostic information.
Parasomnias have a considerable range of type and severity. Most classifications are based on the stage of sleep from which the events are most likely to occur (table 1). Overall, the non-rapid eye movement (REM) parasomnias are more common although REM parasomnias are more likely to be seen in general neurological practice.
With notable exceptions (such as REM sleep behaviour disorder), it is rarely appropriate to treat parasomnias with medication, certainly not for prolonged periods. In fact the evidence base for effective drug therapy is virtually non-existent. Some claim major improvements with alternative approaches such as hypnotism or behaviour therapy although it is usually very difficult to find practitioners with expertise in these areas.
WHAT HAPPENS IN NORMAL SLEEP?
Common experience suggests that someone in normal sleep occasionally moves or twitches and it is therefore worth considering what aspects of sleep-related movements fall within the range of normal. It is, for example, almost a universal experience to feel occasional unsettling sudden starts at the point of sleep onset, as if “falling through space”. These hypnic jerks are entirely benign and are more likely in the young, especially if sleep deprived. Mild restlessness or subtle shifts of position tend to occur at sleep stage transitions, on average every 15 minutes or so, without producing arousals. Furthermore, during REM sleep, minor facial or limb jerks may be seen and autonomic changes, including penile erection, are quite normal. In the very deepest stages of non-REM sleep, some people have short lived confusion if suddenly awoken—so-called “sleep drunkenness”. This is more common in the younger age group and within the normal spectrum.
If closely monitored, very few adults have “perfect” sleep with no EEG evidence for arousals during the night. Indeed, up to 10 brief awakenings through the night may occur without recall or recognition, and these can be considered to be normal.
WAKE-SLEEP TRANSITION DISORDERS
Over the last 20 years, it has been increasingly realised that sleep is not simply an absence of wakefulness. Recent evidence has revealed that specific areas of the brain, such as the ventrolateral preoptic area in the hypothalamus, become very active at the point of sleep onset and can be appropriately considered as “sleep centres”. Given the likely presence of brainstem “sleep switches” that orchestrate smooth yet quick transitions between the mutually exclusive states of wakefulness, non-REM sleep and REM sleep,1,2 it is perhaps surprising that the mechanism does not go wrong more frequently.
Hypnic jerks and sensory starts
In the transition from wake to sleep, hypnic jerks are rarely disturbing or frequent enough to cause insomnia and therefore to merit therapy. If really needed, a short acting hypnotic such as zolpidem is usually helpful. Some people describe more elaborate hypnic phenomena with sensory symptoms as the main manifestation, so-called “sensory starts”. When there is a dramatic explosive quality to the auditory or visual experiences, the term “exploding head syndrome” has been used. Importantly, headache per se is not a feature but the episodes can certainly be alarming and interfere with sleep onset. If there are repeated awakenings throughout the night, the symptom can recur at each wake-sleep transition, leading to significant sleep fragmentation.
A poorly defined parasomnia at sleep onset is head banging, formerly called “jactatio capitis nocturna”. This most commonly affects young children, particularly those with autism or learning disability. Some authorities view the movements as semi-voluntary or tic-like, helping to somehow relax or comfort the child before sleep. Apart from measures to reduce the likelihood of injury, no specific treatment is usually indicated and the phenomenon often resolves spontaneously with time. However, in some adults the abnormal movements may persist and evolve into so-called “rhythmic movement disorder” that can recur through the night, even in REM sleep where complete paralysis is the norm.3 Various movements have been described from minor foot flexions to dramatic head shakes that can disturb bed partners. The subjects themselves are rarely concerned. Empirical and intermittent use of hypnotic agents such as clonazepam in severe cases may help reduce the disturbances but there is very little evidence to guide drug management.
This is a rare phenomenon that occurs in profound drowsiness, usually just at the point of sleep; there are recurrent axial jerks or unpleasant sensorimotor symptoms that descend through the body and prevent sleep.4 The movements are slower and more focal than the more common experience of hypnic jerks which are brief and generalised. If propriospinal myoclonus is a possibility, it merits spinal cord imaging as there can be a structural focus for the myoclonic movements.5
PERIODIC LIMB MOVEMENTS OF SLEEP: A PARASOMNIA ARISING FROM LIGHT SLEEP
Abnormal movements that arise most commonly from light non-REM sleep are periodic limb movements of sleep. These usually involve one or both lower limbs and manifest as a slow withdrawal reflex that starts with foot dorsiflexion and sometimes spreads to involve the whole limb over 0.5 to 5 seconds. By definition, the stereotyped movements occur at regular intervals, typically every 20–40 seconds, in bursts through the night. It can be very difficult to ascertain their clinical significance because they accompany many primary sleep disorders and are increasingly common with age, affecting up to 50% of the elderly.6 A rate of more than five movements per hour can be considered abnormal but this remains extremely controversial. Occasionally, there are clear EEG arousals following periodic movements suggesting that they are disrupting sleep. More often there is no clear correlation and some view periodic limb movements of sleep as epiphenomena, sometimes following arousals due to other causes such as hypopnea. Even in apparently homogenous populations with the “pure” syndrome, it is difficult to correlate the apparent severity of the condition with any parameter of sleepiness or sleep efficiency.7 As in restless legs syndrome, with which periodic leg movements in sleep are strongly associated, the best available evidence for treatment suggests the use of dopaminergic agonists before bed, although benzodiazepines also offer symptomatic control. Before drug treatments are considered it is appropriate to check ferritin levels since iron deficiency is a recognised reversible trigger. Furthermore, tricyclic antidepressants, often prescribed to improve sleep, can actually worsen periodic limb movements and should be discontinued if possible.
PARASOMNIAS ARISING FROM DEEP NON-REM SLEEP: DISORDERS OF AROUSAL
Disorders of arousal refer to parasomnias that stem suddenly and without warning from deep non-REM or slow wave sleep (fig 1) They are extremely common in children, affecting at least 6% on a regular basis.8 Although data are scarce, the problem may persist into adulthood in 1–2% of the population. The behavioural manifestations are extremely variable although they can be clinically divided, somewhat artificially, into night terrors, confusional arousals, and sleep walking.
Night terrors may affect 1% of young children and are usually characterised by a sudden piercing scream about one hour into sleep, followed by intense autonomic and motor manifestations of non-specific fear. The child has total amnesia for the event in contrast to the usual alarm and distress reported in the parents.
Confusional arousals consist of brief episodes in which the subject simply appears to wake from deep sleep staring ahead, sometimes following apparent visual stimuli around the room. There are no complex movements and little emotional disturbance.
Sleepwalking in children is usually limited to wandering from the bedroom to the living room although attempts to leave the building are not rare. The child is able to navigate appropriately and can converse to a limited extent, appearing vague and often fatuous. Sleepwalking may persist into adulthood in up to 10% of cases.
The pathophysiology of this type of parasomnia is not understood but can perhaps be regarded as an incomplete arousal from deep non-REM sleep such that the subject is “entrapped” in an abnormal state of semi-awareness. Subsequently, there is typically no clear memory at all for the event. One intriguing SPECT study managed to capture a teenage subject sleepwalking and fuelled the notion that sleepwalking is a dissociated state with shared features of waking and sleep.9 In particular, compared to deep sleep, increased blood flow was seen in posterior cingulate cortex and the anterior cerebellum while most of the frontoparietal cortex was underactive compared to quiet wakefulness.
Some authorities propose that sleepwalkers have unusually deep sleep. In support of this, sleep deprivation or an unusual sleep-wake cycle may provoke episodes in those predisposed, as may alcohol and hypnotic drug therapy. Conversely, factors interrupting deep sleep such as apnoea, nocturnal pain, periodic leg movements, environmental stimuli, or stress also increase the likelihood of sleepwalking episodes. In addition, genetic factors are clearly important with most patients reporting events in first degree relatives.
A confident diagnosis of an arousal disorder is usually made from the history alone. Indeed, it is rare for overnight recordings to capture typical events although polysomnography may be indicated in adults, if only to exclude potential provoking stimuli such as apnoeas. Nevertheless, the sleep architecture in sleepwalkers on nights where sleep is deemed normal may reveal subtle abnormalities.10 In particular, the percentage of slow wave sleep is reportedly higher as is the number of EEG arousals from it. However, these findings, along with more detailed computerised analysis of the sleep pattern, are not specific or sensitive enough to have major diagnostic implications. Similarly, forced arousals from deep sleep used by some sleep laboratories to provoke episodes rarely provide useful information.
Particularly in adults, it has become increasingly recognised that extremely complex behaviours can be exhibited during sleep without conscious or voluntary control.
Sleep-related eating disorder, also known as nocturnal binge eating disorder, usually occurs on the background of more conventional sleepwalking.11 Sufferers typically rise from sleep and eat food, often after elaborate preparation but without subsequent recall. It is said that idiosyncratic foodstuffs such as cat food can be consumed, often in a messy manner. Depression and significant weight gain are common sequelae and treatment can be difficult. Benzodiapepines may worsen the situation and psychotherapy is generally unhelpful. Of interest, success with topiramate has been reported in a few cases.
Violence or sexual assault may also occur in the context of a parasomnia arising from deep non-REM sleep and this controversial area is increasingly discussed in the medicolegal arena. Several high profile cases of manslaughter have been successfully defended on the grounds of a “non-insane automatism”. A large survey on sleep behaviour from the UK in 1997 questioned 5000 adults and produced the surprising result that 2% admitted to injurious behaviour to themselves or their bed partner.12 Young adult males were particularly likely to report such events. The triggers for violent acts during sleep are varied. Sometimes the behaviour can be triggered by the bed partner trying to wake or restrain the sleepwalker. Often, however, the act is senseless or motiveless and the subject has no clear recall, usually reacting with guilt and perplexity on hearing an account of the events. Injuries incurred by sleepwalkers tend to occur when abrupt arousals from deep sleep cause the subject to leap from the bed in a seemingly desperate attempt to escape a perceived dangerous situation.
Other parasomnias usually arising from deep non-REM sleep
Nocturnal visual hallucinations are a rare parasomnia, usually affecting the elderly either in the context of another parasomnia such as REM sleep behaviour disorder, or in the severely visually impaired. In the latter situation, the phenomena appear very similar to Charles-Bonnet hallucinations. The images are perceived on waking from deep sleep and can be frightening, lasting several minutes, with full recall. Typically, they disappear if the lights are switched on. There is usually no dream narrative and the images appear isolated and immobile. Beta-blockers can sometimes precipitate this parasomnia.
Nocturnal panic attacks can occur in combination with daytime panic disorder or as an isolated phenomenon. Typically, a subject will wake at the point of entering deep non-REM sleep with non-specific extreme fear, associated with palpitations, paraesthesiae, and dyspnoea. It is important to exclude sleep apnoea, nocturnal stridor and REM sleep phenomena before making the diagnosis. Management is similar to daytime panic disorder.
Distinguishing non-REM parasomnias from nocturnal partial seizures
Sometimes, there can be doubt, both in children and adults, whether abnormal motor behaviour during sleep reflects an epileptic disorder or a parasomnia. There are several pointers from the history that help although sometimes a confident diagnosis can only be made after video analysis of the attack semiology (table 2).
PARASOMNIAS ARISING FROM REM OR DREAM SLEEP
At least 20% of our sleep is spent in the curious and highly distinct state of REM sleep. Dreams associated with a clear narrative and vivid imagery occur exclusively during this stage, although they are not generally recalled unless the individual is aroused to wakefulness during the dream.
Considerable advances in the neurobiology of REM sleep have been made over the last four decades although the three basic parameters by which it is defined were first established over 50 years ago. In brief, these are:
a small amplitude desynchronised surface EEG
phasic episodes of saccadic eye movements seen in the oculographic EEG leads
and, other than occasional “twitches” or myoclonic jerks, profound muscle inactivity, as evidenced by a virtually silent EMG trace obtained from chin or shin electrodes.
Largely through work on cats, this third aspect of REM sleep has been shown clearly to reflect active descending inhibition of voluntary muscle activity. In particular, during REM sleep, a group of brainstem neurons mostly in the upper medulla become active, sending descending impulses that directly influence the skeletal alpha motor neurons in the anterior spinal cord. Via this predominantly glycinergic inhibitory mechanism, the motor neurons are switched “off”, rendering the subject totally floppy and atonic. If it were feasible to elicit peripheral reflexes, the sleeping subject would be found temporarily areflexic. Clearly, this mechanism spares respiratory muscles and, by definition, the extraocular musculature. It can be surmised that this complex system to induce paralysis has evolved as a protective mechanism to prevent subjects moving their limbs injuriously during their dreams, a likely advantage especially if the dream was violent or emotionally charged.
Disturbing dreams that wake the individual from REM sleep are such a common experience as to be considered normal, especially in children. However, the experiences in some can be so frequent and intrusive to merit treatment. Post-traumatic stress disorder is particularly associated with recurrent nightmares which may respond to behavioural therapies such as asking the patient to rewrite their dreams with a happy ending, rehearsing these before sleep. Drugs such as beta-blockers may be implicated in triggering disturbing dreams.
REM sleep paralysis
A significant proportion of the normal population report very occasional episodes of profound voluntary muscle paralysis on waking from sleep in the morning or during the night. These episodes are generally very disturbing, especially since the subject often reports an inability to breathe deeply or voluntarily. The paralysis lasts for a few seconds only and can be terminated by the bed partner touching or shaking the subject. Rarely, the problem recurs on a regular basis and leads to neurological assessment. This phenomenon of sleep paralysis reflects intrusion of REM sleep atonia into wakefulness and can sometimes be associated with frank hallucinatory visual or tactile experiences. There is a familial form of REM sleep paralysis. It is also sometimes a prominent component of the narcolepsy syndrome in which case it is commoner for the events to occur at initial sleep onset. The REM sleep suppressing properties of most antidepressants will usually abolish the episodes if specific treatment is thought necessary. Traditionally, as with cataplexy, tricyclics such as clomipramine are generally used first line.
REM sleep behaviour disorder
Early pioneering lesion studies from Michel Jouvet’s laboratory in the mid 1960s, exploring the neuroanatomy of sleep, produced some interesting results in the cat.13 Using electrodes to destroy small and discrete areas in the cat brainstem, it was found that bilateral lesions in the region of the locus coeruleus reliably produced a state in which the unfortunate cat appeared to be having dream activity within REM sleep yet was physically highly active, exhibiting so-called “oneiric behaviour”. Slightly larger lesions in the same area, however, abolished REM altogether. This early work confirmed the notion of brainstem inhibitory “centres”, active during REM sleep, capable of producing profound voluntary muscle atonia. Unknowingly at the time, Jouvet had produced an animal model in search of a human disease. Although sporadic reports were subsequently published, demonstrating occasional similar examples of dream enactments in human subjects, it wasn’t until 1986 when this new parasomnia was formally described by Carlos Schenck in five elderly patients.14 Within four years of this report, REM sleep behaviour disorder had been incorporated into the international classification of sleep disorders.
Several large series from various centres have now been published, all of which agree on the clinical features, such as the striking predominance of elderly males (table 3). It should be emphasised that the disorder has a wide clinical spectrum, ranging from occasional vocalisations and minor twitches to violent thrashing upper limb movements capable of causing serious injuries to sufferer and bed partner alike. In fact, it is usually the latter who is responsible for bringing the problem to medical attention. There is little true interaction with the external environment in subjects experiencing REM sleep behaviour disorder. The various behaviours appear brief, random, and unfocussed. Unless the movements are particularly vigorous, the subject does not usually wake and can remain blissfully unaware of any nocturnal disturbance (fig 2). However, if roused from sleep to wakefulness, often by a disturbed bed partner, dream recall is the norm. In severe cases, the behaviours recur with each REM period, often becoming more elaborate, vocal, or violent as the night proceeds. Perhaps surprisingly, apart from the potential effects of any injuries, there are usually few effects on daytime functioning and it is commoner for the concerned spouse to have daytime somnolence as a result of sleep disruption than the “sufferer”.
The diagnosis is usually strongly suspected from just the history although it can be difficult if the subject sleeps alone, for example. One important differential diagnosis is recurrent arousals at the termination of severe episodes of obstructive sleep apnoea which can sometimes be disturbing enough to mimic REM sleep behaviour disorder. Furthermore, this syndrome is not infrequently misdiagnosed as a nocturnal seizure disorder, resulting in inappropriate treatments and driving restrictions. For these reasons, if facilities exist and, perhaps more importantly, there are interested clinicians, it can be useful to monitor selected patients overnight. Ideally, multiple electrode placements over several muscle groups should be used to demonstrate the primary abnormality of abnormal tone or phasic movements during REM sleep. Monitoring can also be useful to confirm the absence of epileptiform EEG activity during REM. Very rarely, parasomnias such as so-called rhythmic movement disorder, can be limited to the REM stage of sleep and cause some diagnostic difficulty. However, at a practical level, the diagnosis is usually secure when a report is obtained of intermittent and variably violent, brief movements usually of the arms, together with vocalisations and observable dream enactment.
The link with parkinsonism
Probably most REM sleep behaviour disorder patients do not seek medical attention because the symptoms are mild and go unnoticed. But, of those cases reported in the literature, around half have been in association with a neurological disorder—the remainder appearing as apparently isolated examples.15 Although REM sleep behaviour disorder often accompanies primary sleep disorders such as narcolepsy, the most frequent association is with parkinsonian syndromes. In fact, REM sleep behaviour disorder is relatively common in all the synucleinopathies, including idiopathic Parkinson’s disease, dementia with Lewy bodies16 and multiple system atrophy.17
It is clearly of interest if a substantial proportion of apparently isolated REM sleep behaviour disorder patients followed up over a period of years will eventually develop clinical signs of parkinsonism with or without dementia. The original study18 suggested a figure of 38% although it seems that this figure may rise to at least 65% with even longer follow up.19 Another indication that many idiopathic REM sleep behaviour disorder cases are harbouring subclinical parkinsonism comes from SPECT imaging that suggests an early dopamine deficiency state.20 Furthermore, a very recent study has demonstrated olfactory dysfunction in REM sleep behaviour disorder without parkinsonism similar to that seen in Parkinson’s disease itself.21 Confusingly, however, this abnormality was also seen in REM sleep behaviour disorder cases who presented in the context of narcolepsy and presumably are not at extra risk of subsequently developing a movement disorder.
If REM sleep behaviour disorder and Parkinson’s disease coexist, whether or not the former precedes the latter, there is some evidence that the parkinsonian syndrome may run a more complex course than average. In particular, a recent study suggested that REM sleep behaviour disorder predicted a more likely occurrence of subsequent neuropsychiatric disturbance, particularly hallucinations.22 This is particularly relevant to the observation that parkinsonian visual hallucinosis probably reflects intrusion of REM sleep dream imagery into wakefulness, further supporting the notion that REM processes are frequently disturbed by Parkinson’s disease pathology, presumably in the brainstem. The further association between frank dementia and hallucinations in Parkinson’s disease and sleep disruption raises the possibility that the presence of REM sleep behaviour disorder may also be more likely to herald cognitive impairment. Whether this pessimistic prediction is true remains to be established by prospective studies. Even in apparently idiopathic cases of REM sleep behaviour disorder, however, there is some evidence of subtle cognitive impairment, albeit confined to the visuospatial domain.23
Initially, the robust association of REM sleep behaviour disorder with Parkinson’s disease and related conditions was thought to separate the synucleinopathies from other neurodegenerative diseases such as tauopathies. This dichotomy was heralded as a potentially useful clinical tool if there were diagnostic doubts. However, if the possibility of REM sleep behaviour disorder is systematically assessed, then it is also relatively common in conditions such as progressive supranuclear palsy which do not have primary synuclein pathology.24 Furthermore, REM sleep behaviour disorder has also been observed in genetic forms of Parkinson’s disease without synuclein pathology and also in several types of spinocerebellar ataxia.
Not all patients with REM sleep behaviour disorder require drug treatment. Improving the safety of the sleep environment may be the only intervention necessary. Partners moving to another bed or room may be another appropriate practical strategy. It should also be borne in mind that most classes of antidepressant are capable of triggering or worsening REM sleep behaviour disorder with resolution after drug discontinuation. If active treatment is thought necessary, although there is no controlled evidence, clonazepam is generally regarded as the first-line drug, taken before bedtime in a dose range 0.25–2 mg. Whether clonazepam has particular properties that distinguish it from other benzodiazepines is unclear. It is effective in around 90% of cases, usually with minimal adverse effects, although the dose may need to be increased as a result of tolerance. Typically, the clinical problem is lessened although minor twitches and vocalisations still occur with continuing evidence for REM without atonia on polysomnographic recording.
When prescribing clonazepam, particular care needs to be taken in patients with cognitive impairment or if there is additional sleep disordered breathing. The latter caveat may limit its usefulness in multiple system atrophy, for example, where nocturnal stridor is not uncommon. If the drug is prescribed in primary care, it needs to be emphasised that it is not simply a “sleeping tablet” but is being recommended long term to treat a sleep disorder.
There are sporadic reports of REM sleep behaviour disorder in complex parkinsonian syndromes apparently responding to atypical neuroleptic agents such as quetiapine, and cholinesterase inhibitors including donepezil. Furthermore, L-dopa itself has been reported to ameliorate REM sleep behaviour disorder in Parkinson’s disease which is interesting from a pathophysiological perspective. Finally, in idiopathic cases of REM sleep behaviour disorder, several small open trials have reported success with melatonin (3–12 mg, oral) before bed.
Sleep-related painful erections
These are characterised by penile pain during erections associated with REM sleep episodes. There is usually no local pathology although Peyronie’s disease is an association. It is very rare and often responds to REM suppressant therapies such as tricyclic antidepressants.
Nocturnal clenching or grinding of teeth is surprisingly common, and often overlooked. As well as contributing to poor quality sleep, it can fuel a variety of headache syndromes. It occurs at any age and during any sleep stage. General anxiety or arousals due to breathing difficulties commonly provoke or worsen symptoms. If there is no clear history of bruxism, it can be picked up as prominent 1 Hz rhythmic activity reflecting movement artefact on an overnight EEG recording. Dental occlusive appliances are usually effective in treating the problem.
Catathrenia (nocturnal expiratory groaning)
This rare but striking parasomnia affects the younger age group and is characterised by high pitched monotonous groaning that starts within an hour or two of sleep onset.25 The groans follow deep inspirations and can recur in clusters, each lasting up to 30 seconds or so. They may be loosely associated with REM sleep although there are no other identifiable sleep abnormalities. Understandably, the bed partner is often severely disturbed by this phenomenon. Unfortunately, a variety of empirical drug treatments and nasal continuous positive airways pressure have generally been unsuccessful.
A confident diagnosis can often be made from a detailed history from the patient and, if possible, the parents or bed partner.
Detailed overnight investigations of parasomnias are not usually needed.
REM sleep behaviour disorder is common and should be sought in all neurodegenerative diseases.
Obstructive sleep apnoea can often provoke parasomnias in adults and should not be overlooked as an independent treatable factor.
If treatment is thought necessary, only short-term courses are usually appropriate with the exception of REM sleep behaviour disorder.
Clonazepam is effective in most parasomnias and can be regarded as the “steroid” of sleep medicine in a dose range 0.25–2 mg.
If the parasomnia is at the wake-sleep transition, a short acting hypnotic such as zopiclone or zolpidem is preferred.
REM sleep parasomnias such as sleep paralysis usually respond to antidepressants such as tricyclics.
Tricyclics can worsen periodic leg movements of sleep.
Periodic leg movements usually respond to dopaminergic therapy in much the same way as the restless legs syndrome.
This article was reviewed by Professor Adam Zeman, Exeter, UK.