Clinical manifestations

Presentation and course

Table 1. REM Sleep Behavior Disorder Criteria (ICSD-3)

RBD presents with a complaint of dream enactment behavior (see Table 1 for ICSD-3 criteria) or sleep-related injury. The spectrum of dream enactment behavior varies from small hand movements to leaping out of bed. The frequency of episodes ranges from once every few months to multiple nightly episodes. Often, there is a prolonged prodromal period lasting years with progressively more prominent dream enactment behavior. Some patients adopt extraordinary measures to prevent sleep-related injury: they may place obstacles to hinder exiting the bed or sleep in a room devoid of furniture. Prolonged diagnostic delay is common, and some families deal with these behaviors for decades prior to seeking medical attention (66).

RBD movements appear purposeful, such as throwing a ball or flailing to protect oneself (02). These movements, with their complex and repetitive nature, are often distinguished from other nocturnal behaviors on polysomnography based on appearance alone. Unlike NREM parasomnias, the duration of behaviors in RBD is brief, and on awakening from an episode, there is usually a rapid return to alertness and orientation. These characteristics are not unexpected as unlike NREM sleep, a low arousal threshold characterizes REM.

The reported motor activity usually correlates with remembered dream mentation, leading to the patient’s (or the patient’s bedpartner) complaint of "acting out dreams." Interestingly, unlike normal dream mentation, which is often forgotten by dreamers within a few minutes of awakening, RBD patients often recall dreams with striking clarity for weeks or years (66). Importantly, RBD patients do not have more violent dreams than normal individuals; instead, they merely act them out. Further, during the daytime, they are not more aggressive, nor do they suffer from personality disturbances (93).

In RBD, sleep-related vocalizations are typically loud and aggressive (expletives are not uncommon). This is most often discordant from waking personality. Further, RBD vocalizations need to be distinguished from sleep talking, which is common (during both NREM and REM), more typical of daytime conversation, and does not, in itself, represent pathology.

Due to its association with REM sleep, it is not unexpected that dream enactment behavior appears more often during the later periods of the sleep period. Even so, events in the latter half of the night REM sleep are more likely to demonstrate violent-aggressive features. RBD, like REM sleep, rarely occurs during daytime naps, with a significant exception for cases associated with narcolepsy (66).

In the absence of injury, most RBD patients attribute only minimal daytime consequences, if any, to their nocturnal behaviors. One study noted that 70% of patients reported good sleep quality. However, this is in contrast to the patients’ bed partners; 44% of patients were unaware of dream enactment until their bed partners told them (42).

Ancillary clinical features of RBD. Patients with RBD demonstrate various features of Parkinson disease and other Lewy body diseases (see Etiology section). Investigations have illuminated these motor, cognitive, and autonomic impairments.

Motor testing reveals that patients with RBD demonstrate abnormalities often imperceptible on a general clinical exam. These are notable approximately 5 to 7 years prior to the diagnosis of Parkinson disease, with subtle abnormalities on the alternate tap test being the first motor abnormality to appear as early as 13 years prior to Parkinson diagnosis (39). It should be noted that typical cardinal features of Parkinson disease, such as rigidity and tremor, typically appear 2 to 3 years prior to Parkinson disease diagnosis in patients with iRBD (39).

Patients with RBD have cognitive impairments, in particular visuospatial deficits and executive dysfunction, which correlate with mild predominantly occipital hypometabolism on FDG-PET scans (22). In particular, progressive abnormalities have been found in visual perception, visuospatial learning, facial expression recognition, and color identification (66; 125). Other investigations have revealed impairments in attention and executive function, with executive dysfunction and visuospatial impairment predicting more rapid phenoconversion (15; 125a; 156). Further, the presence of pareidolic responses in patients with isolated RBD correlates with underlying cognitive impairment and may predict future development of clinically overt Lewy body disease (139). Consistent with these findings, in vivo PET imaging of acetylcholinesterase activity has shown reduced superior temporal cortex, occipital cortex, cingulate cortex, and dorsolateral prefrontal cortex cholinergic activity (53).

The cognitive deficits among RBD patients with dementia are more similar to dementia with Lewy bodies (DLB) than to Alzheimer disease (15; 121). Further, among patients with mild cognitive impairment who were later proven to have dementia with Lewy bodies, nearly all had symptoms of RBD (15). Also, the presence of RBD symptoms in patients with dementia predicts dementia with Lewy bodies on postmortem examination (108). The presence of RBD in a patient with cognitive impairment consistent with Alzheimer disease dementia is highly suggestive of underlying comorbid synuclein pathology (17).

Prior to the onset of Parkinson disease, RBD subjects demonstrate mild postural and gait abnormalities. While attempting to stand motionless, patients with RBD have postural instability when distracted (25). Then, during gait initiation, RBD patients show a pattern of abnormal force generation consistent with FOG (161). Subsequently, while ambulating, there is a measurable decline in velocity and cadence, with an increase in stride and swing variability of limb movements in RBD (104).

In general, Parkinson disease patients with RBD appear to have a more malignant and aggressive disease characterized by a higher Hoehn and Yahr stage, more cognitive impairment, worse quality of life, increased postural instability, and gait dysfunction compared with Parkinson disease patients without RBD (24; 66; 125c; 136). Neuroimaging studies of Parkinson disease patients with RBD show more widespread cortical and subcortical atrophy than Parkinson disease patients without RBD (133). Further, Parkinson disease patients with RBD appear to be more likely to have freezing of gait (FOG) and a higher frequency of falls compared with Parkinson patients without RBD (58). Studies have suggested a Parkinson disease classification scheme of mild-motor predominant, diffuse malignant, and intermediate phenotypes (40). Patients with RBD fall into the diffuse malignant or intermediate phenotypes, which are associated with more rapid disease progression and earlier mortality compared with mild-motor predominant Parkinson disease (32).

Comorbid hyposmia is frequently noted in cases of spontaneous RBD and is likely the first symptom (apart from RBD) suggestive of neurodegeneration, presenting greater than 20 years prior to phenoconversion (40). These impairments in olfaction are likely representative of evolving Lewy body disease and not multiple system atrophy, in which olfaction remains normal (132).

Autonomic dysfunction as measured by heart rate variability, cardiac scintigraphy, and orthostatic responses is frequently present in RBD patients, with some studies showing that abnormalities in cardiac scintigraphy precede changes in DaTscan (50; 148; 11; 72).

Prior to the onset of motor features, constipation is the most prominent autonomic symptom among patients with RBD and suggests early alpha-synuclein degeneration of enteric neurons, presenting approximately 10 years prior to phenoconversion (41; 39). Additionally, alpha synuclein depositions in the enteric nervous system have been found in patients with Parkinson disease and Parkinson disease with RBD with slower colonic transit time, suggesting the presence of RBD indicates the manifestation of a widespread synucleinopathy (82). This has been confirmed by the presence of phosphorylated alpha-synuclein in skin biopsies of patients with iRBD and an association with more severe autonomic dysfunction on standardized autonomic testing (105).

One notable exception to the link between RBD and Parkinson disease are patients with familial cases of Parkinson disease. The most studied familial cases involve patients with the LRRK-2 mutation, who are less likely to have RBD (122). On the other hand, mutations in glucocerebrosidase (GBA) are associated with both Parkinson disease and RBD, with the presence of GBA mutations in patients with iRBD, suggesting a more rapid rate of phenoconversion (62).

Prognosis and complications

Sleep-related injuries may result from behaviors such as punching, kicking, or leaping out of bed, with greater than 50% of patients or bedpartners reporting injury (99). Examples of sleep-related injury include subdural hematoma, shoulder dislocation, cervical fracture, as well as lacerations severing arteries, tendons, and nerves. As bed partners are frequently the target of violent dream enactment, RBD may have forensic implications, with patients wrongly in the criminal justice system (66).

Initially, most patients with RBD do well with environmental changes and pharmacotherapy. The most commonly prescribed medications include bedtime oral clonazepam or melatonin (15; 81; 66). However, the long-term efficacy and safety of these agents in the setting of progressive dementing illnesses are uncertain. One follow-up study of clonazepam demonstrated that half of all patients would ultimately either decrease the nighttime dose due to side effects or discontinue the medication (03). RBD has also been reported to be an independent predictor of mortality in Parkinson disease and typically is associated with a poor prognosis in Parkinson disease (164). Conversely, patients with RBD have increased mortality when compared with the general population only if they develop a clinical neurodegenerative disease (170).

Ultimately a patient’s prognosis depends on the underlying pathological process. Both toxic RBD and dream enactment behavior due to sleep apnea are typically resolved by eliminating the offending agent or treating the sleep-disordered breathing, respectively (66). Among neurodegenerative cases, the interval between the onset of RBD and disabling neurologic symptoms varies from months to decades.

Predicting the rate of phenoconversion. Numerous ancillary clinical features can help predict how rapid a patient with RBD will develop a parkinsonian disorder. Patients who are older and have family history of dementia are more likely to convert within 4 years (125a). In addition, subtle motor or cognitive dysfunction, orthostatic hypotension, constipation, hyposmia, impaired color vision, and nonuse of antidepressants predicts more rapid conversion (91; 125b; 09; 38). Stratification with these markers increases risk of conversion by 200% (125b). RBD patients with errors in the trail making test (part B) predicted the development of dementia, whereas verbal fluency (semantic) and verbal episodic learning tests appear best to monitor changes in cognition of patients with RBD over time (92).

Decreased dopamine uptake (as measured by DAT-SPECT scan) in the putamen is associated with increased short-term risk of phenoconversion to Parkinson disease, suggesting serial DATSCANs may be useful in monitoring disease progression in patients with isolated RBD (83). Specifically, quantitative analysis of DAT tracer uptake in the putamen appears to be a robust predictor of short-term phenoconversion, with additional variables such as age greater than 70 at the time of DATSCAN and the presence of constipation increasing the risk even further (06). [18 F]fluorodeoxyglucose-PET (FDG-PET) scans of the brain with a Parkinson disease-related brain pattern at isolated RBD diagnosis also appear to predict more rapid phenoconversion, with serial FDG-PET scans showing a higher frequency or intensity of the Parkinson disease-related brain pattern over time (79).

CSF alpha-synuclein positivity as measured by real-time quaking-induced conversion (RT-QuIC) has been shown to be associated with a higher risk of phenoconversion compared with patients with isolated RBD who were RT-QuIC negative at diagnosis (70). It should be noted that although the risk and rate of phenoconversion in the typical RBD patient (ie, older than 65 years old at diagnosis/symptom onset and male gender), the significance of RBD in patients younger than 50, associated with antidepressant use or autoimmune disease, is much less well known. Prospective data in patients with longstanding RBD (ie, disease free 10 years after RBD diagnosis) suggest that these patients still have similar prodromal biomarkers of neurodegeneration or evidence of subclinical Parkinson disease similar to patients who developed neurodegenerative disease within 4 years of RBD diagnosis (71; 165).

Overall, the ideal method of predicting phenoconversion remains controversial. From a practical standpoint, the presence of older age at diagnosis in combination with more abnormal biomarkers (abnormal DATSCAN, FDG-PET scan, anosmia, constipation, motor impairment, autonomic dysfunction, mild cognitive impairment, increased REM sleep without atonia) likely confers a higher risk of more imminent phenoconversion.

Clinical vignette

A 71-year-old retired police officer returned to the sleep center 10 years after being diagnosed with obstructive sleep apnea and treated with continuous positive airway pressure (CPAP). The patient indicated that despite excellent compliance over the last decade, he suspected that he no longer had obstructive sleep apnea. In particular, his CPAP machine broke 6 months prior, and he had been asymptomatic (no daytime sleepiness) despite a lack of treatment. Further, his wife stated that he no longer snored, nor did he appear to have labored breathing. His weight had not significantly changed.

Interestingly, the patient’s spouse also indicated that he had developed a 2-year history of dream enacting behaviors. These behaviors had become progressively more disruptive, initially characterized by loud vocalization and random twitches during sleep, then progressing to yelling, swearing, and flailing. The vocalizations suggested that the patient was dreaming of his police work. One event wherein he jumped out of bed resulted in fractures of two lumbar vertebrae. If awakened following these episodes, he was immediately awake, alert, and oriented and recalled his dream mentation.

Polysomnography revealed prominent tonic and phasic muscle activity during REM sleep associated with vocalization and kicking motions. Despite spending the entire night supine, the patient did not have sleep disordered breathing.

The patient’s dream enactment behavior was initially treated with bedtime oral clonazepam. However, the patient was unable to tolerate residual morning sleep inertia and switched to high-dose melatonin. Melatonin at 12 mg eliminated the majority of dream enactment behaviors; however, he still left the bed during a dream approximately once a week. Both the patient and spouse were concerned about possible sleep-related injury.

The patient was given a customized bed alarm, with voice recording unit. Subsequently, when the patient sat upright during a dream the pressure sensing pad triggered his wife’s admonition from the bedside speaker, “Pat, you’re having a dream go back to sleep.” At 6 months, he had not had any further sleep-related injuries.

Biological basis

Etiology and pathogenesis

RBD represents the final common pathway of several diverse pathologies, all of which result in a failure to inhibit central motor pattern generators and spinal motoneurons. These pathologies manifest in a loss of behavioral control during REM sleep and the enactment of dream mentation (15). However, it has been postulated that dream enactment behaviors in RBD may occur secondary to a “bottom up” process rather than a “top down” (14). In this proposed model, loss of REM muscle atonia due to brainstem dysfunction results in dream enactment behavior that is interpreted into dreams (rather than dreaming first and acting out what one dreams). This is based on studies of brainstem stimulation that can produce defensive and aggressive behaviors in rats and monkeys that are similar to those seen in RBD. Therefore, sensory feedback from moving limbs may influence dream mentation.

The suppression of motor activity during normal REM sleep is the cumulative result of multiple, currently poorly understood pathways that ultimately terminate on spinal motor neurons, most notably via the magnocellular reticular formation in the medulla. Multiple areas of the brainstem may influence muscle tone during REM sleep. These include pontine REM-on (precoeruleus and sublateral dorsal) and REM-off (ventral lateral portion of the periaqueductal grey matter and lateral pontine tegmentum) nuclei (15). Various other brainstem structures mediate REM paralysis at least partially through GABA and glycine inhibition of motor neurons (67). Genetic inactivation of the rat sublateral dorsal nucleus (analogous to the human locus subcoeruleus) causes REM sleep without atonia but does not impact cortical features of REM sleep (160). Dysfunction in these REM modulating structures as well as their related neurotransmitters and pathways can result in REM sleep without atonia (15; 75; 67).

With the exception of cases caused by a focal lesion, such as a stroke, it is difficult to localize RBD pathology. However, investigations are beginning to focus on the coeruleus complex. Evaluating for neuromelanin with 3T MRI, changes were seen in the coeruleus area among patients with RBD and Parkinson disease as well as RBD alone (35). A neuroimaging study utilizing 7T MRI to assess brainstem connectivity with tractography in patients with iRBD has shown decreased excitatory connectivity between the locus coeruleus/subcoeruleus complex and ventromedial medullary nuclei known to be integral for REM sleep atonia generation, mirroring what has been shown in animal studies of RBD (51).

Synuclein neurodegenerative etiologies. The majority of spontaneously developing RBD cases are related to Lewy body CNS pathologies. These conditions include Parkinson disease, multiple system atrophy, dementia with Lewy bodies, and pure autonomic failure. Patients with these disorders have increased alpha-synuclein oligomers in the CSF, a marker of neurodegeneration, and demonstrate pathological alpha-synuclein deposition in the CNS on postmortem examination. These findings have also been described in RBD (66; 68). Further, some patients with isolated RBD demonstrate synuclein deposition in the olfactory mucosa, skin, and enteric plexus. The brainstem nuclei that control REM sleep (see Pathogenesis and pathophysiology section) are often involved early in the natural history of synucleinopathies (15).

The interval between the onset of RBD and diagnosis of Parkinson disease or related disorders can vary from months to decades. The largest multicenter study of greater than 1000 patients with iRBD reported a 6.3% per year risk of phenoconversion, with 73.5% of patients phenoconverting within 12 years from RBD diagnosis (130). Among the rare cases of apparently persistent isolated RBD, biomarkers still demonstrate impending neurodegeneration (165). Further, the finding of diffuse Lewy body disease on postmortem evaluation in otherwise asymptomatic patients with RBD suggests that in the setting of very slow disease progression, dream enactment may be the sole clinical manifestation of alpha-synuclein pathology.

By the time motor abnormalities develop in patients with Parkinson disease, the vast majority (up to 90%) of dopaminergic neurons in the substantial nigra are irreversibly dysfunctional (80). As RBD is a prodromal syndrome of Parkinson disease, neuroimaging and transcranial magnetic stimulation studies have demonstrated coincident and progressive dopaminergic, cholinergic, and other related parkinsonian abnormalities (109; 61; 133; 31).

Forebrain pathologies are also implicated in the failure to suppress motor activity in RBD, presumably by unleashing central pattern generators, functional groups of neurons that give rise to stereotyped patterns of behavior such as startling, punching, or jumping. This is the purported pathway by which clonazepam suppresses dream enactment behavior as clonazepam does not promote tonic REM sleep atonia on polysomnography. Dream enactment behaviors bypass the basal ganglia in contrast to wakeful movements (94; 57). This helps explain why RBD actions are often not parkinsonian, in contrast to the bradykinetic movements these Parkinson disease patients may have during wakeful hours.

Non-synuclein neurodegenerative etiologies. RBD has been associated with other neurodegenerative pathologies. RBD has been reported in cases of Alzheimer disease, tauopathy-related parkinsonian syndromes (progressive supranuclear palsy, corticobasal degeneration, and Guadaloupean parkinsonism), and TDP-43-opathies (frontotemporal dementia, amyotrophic lateral sclerosis), although the presence of RBD in these cases is most likely related to co-mingling alpha-synucleinopathy pathology. However, studies of RSWA comparing patients with synucleinopathy and non-synucleinopathy diagnoses have shown similar levels of RSWA in non-synucleinopathies compared with non-RBD control patients, which brings into question whether or not RBD reported in non-synucleinopathies could represent other diagnoses (periodic limb movement disorder, non-REM parasomnia, obstructive sleep apnea) (97; 102).

Also, RBD has been noted in some trinucleotide and tetranucleotide repeat disorders, including spinal cerebellar ataxia types 2 (SCA2), 3 (SCA3), 31 (SCA31), Huntington disease, and myotonic dystrophy type 2. In addition, RBD has been reported to occur in cases of Wilson disease and pantothenate kinase-associated neurodegeneration. With the notable exception of SCA3, none of these disorders have prevalence rates similar to synuclein disorders. Moreover, non-synuclein conditions are not typically preceded by RBD, but instead, develop RBD coincidentally or following other neurologic deficits (15; 26; 34; 88; 113; 17; 157; 66; 119).

State boundary dysfunction. REM sleep behavior disorder and narcolepsy are both due to dysfunction in state boundary control, meaning the brain fails to separate features of REM sleep and wakefulness. In narcolepsy, a persistence of REM sleep atonia into wakefulness manifests as sleep paralysis and cataplexy. In RBD, wakeful muscle tone intrudes on REM sleep and dream enactment behavior emerges. Thus, it is not surprising that approximately 50% of patients with narcolepsy also have RBD symptoms (30). This association is strongest among narcolepsy patients with cataplexy, otherwise known as narcolepsy type I (29; 110).

The mechanism of RBD in narcolepsy is likely different from other forms of RBD and related to the failure of orexin to stabilize REM sleep. RBD in patients with narcolepsy demonstrates unique features compared to synuclein etiologies. In particular, dream enactment tends to present earlier, be composed of more simple movements, and is less violent. Further, in addition to REM sleep without atonia RBD patients with narcolepsy have frequent shifts between REM and NREM sleep consistent sleep state boundary dysfunction (29). The risk of future neurodegenerative disease in narcolepsy patients with RBD is not well understood, but likely less than those with iRBD, as phosphorylated alpha-synuclein has not been reported in skin biopsies of narcoleptic patients compared with iRBD (04).

Medication-associated RBD. Psychoactive medications, in particular antidepressants, have long been noted to acutely precipitate or exacerbate dream enactment behavior. Implicated medication classes include tricyclic and tetracyclic antidepressants, monoamine oxidase (MAO) inhibitors, selective-serotonin reuptake inhibitors, serotonin-norepinephrine reuptake inhibitors, and an acetylcholinesterase inhibitor (63; 75; 49). In fact, the induction of RBD-like behavior by phenelzine (an MAO inhibitor) and clomipramine (a tricyclic antidepressant) was repeatedly described in the two decades preceding the seminal 1986 RBD publication (01; 13; 12; 141; 63). Despite the initial reports involving tricyclic antidepressants, given their ubiquity in current clinical practice, selective-serotonin reuptake inhibitors are most commonly implicated in medication-associated RBD. On PSG, drug-induced RBD has more REM sleep-related motor activity in the lower extremities than nondrug-induced RBD (101).

Medication-induced RBD may, in fact, be the most prevalent form of RBD, especially among the young (155; 75). Approximately 10% of depressed patients have an increase in REM sleep motor tone after treatment with sertraline compared to 3% of patients at baseline prior to treatment (167).

Evidence suggests that antidepressants do not represent a de novo induction of RBD, as these patients would have otherwise developed RBD, and ultimately a parkinsonian syndrome later. One study demonstrated that patients with antidepressant-associated RBD have other prodromal markers of alpha-synuclein neurodegeneration, such as hyposmia, constipation, as well as visual and subtle motor impairments (128). In another study, patients with medication-associated RBD had reduced striatal dopamine dysfunction as measured by (18)F-DOPA PET imaging (163). These findings imply that antidepressants are not causing RBD in isolation but instead are unmasking individuals at risk of a neurodegenerative syndrome.

Lesional RBD. Occasionally lesions affecting the pons, medulla, or limbic system, such as malignancies, aneurysms, or white matter disease, will cause RBD, and these cases do not appear to progress to neurodegeneration (103). The first series of patients with RBD investigated with MRI demonstrated dorsal pontomesencephalic area infarctions in 3 of 6 patients (28). Subsequently, reports have emerged of dream enactment behavior following a focal CNS insult from various vascular, demyelinating, neoplastic, and traumatic etiologies (153; 66; 103). Cranial imaging has typically demonstrated pontine tegmentum pathology, although limbic lesions have also been reported. These lesional studies expand nicely on Jouvet’s 1965 feline studies that induced dream enactment after pontine lesions near the locus coeruleus (74).

Inflammatory RBD. RBD can be mediated by inflammatory processes, which implies that immunomodulating therapies may be a plausible therapeutic option. Patients with antibodies against a cell adhesion molecule (IgLON5) commonly have RBD (and RSWA along with severely disturbed sleep architecture and central sleep apnea) that may partially respond to immunosuppression, although most patients relentlessly progress and expire despite immunotherapy. Postmortem examinations reveal hyperphosphorylated tau in the brainstem, suggesting an overlap between autoimmunity and neurodegeneration in these patients (135). Additionally, two cases of RBD, occurring in the setting of paraneoplastic cerebellar degeneration, were responsive to intravenous immunoglobulin (159). Antibodies to leucine rich glioma inactivated-1(LGi1), Ma-2, and dipeptidyl-peptidase-like protein 6 (DPPX) have also been reported to be associated with the development of RBD.

Isolated RBD. Previously, the chronic form of RBD was felt to be idiopathic in the majority of cases. However, the recognition of numerous associated etiologies (neurodegenerative, toxic, lesional) underlying RBD cases have questioned that assumption (66). Further, as noted above, at least the majority of surviving patients with RBD will progress to a parkinsonian or other neurodegenerative condition, leading some to argue that idiopathic RBD does not exist and the term isolated has been recommended to replace idiopathic (66; 60).

Similar to Parkinson disease, RBD is associated with various environmental exposures and behavioral risk factors. In particular, patients with RBD are more likely to smoke, have a history of traumatic brain injury, have fewer years of education, have worked as farmers, and have been exposed to pesticides (131). Heavy alcohol use has also been associated with increased risk of RBD (89).

Epidemiology

Previous reports of RBD prevalence vary depending on whether a measurement of REM sleep EMG tone by polysomnography was included in the diagnosis. Surveys have revealed that some dream enactment behavior by clinical history alone is nearly universal. In a study of 1140 college-aged students, 98% acknowledged a history of at least one dream enactment behavior symptom (112). Complex dream enactment behavior in the setting of anxious dreams is particularly common in recently postpartum women (111). The first PSG population-based study in RBD found that 1.06% of individuals met PSG criteria for RBD. Importantly, this study showed an equal gender prevalence compared with previous studies that were predominantly male. This is clinically relevant as this suggests that RBD is likely equally common in women, but women are less recognized as they don’t have violent dream enactment behaviors, although they are at equal risk of future neurodegenerative disease. This suggests that great efforts should be made to identify women with RBD in hopes of implementation or study of neuroprotective medications (56).

RBD is more common among the elderly and patients with comorbid neurodegenerative disorders. A general elderly population study utilizing PSG identified RBD in 2% and REM sleep without atonia in the absence of the RBD clinical syndrome in 5% (77). Among Parkinson disease patients, approximately one-third to one-half have RBD (121; 123). The frequency of RBD is even higher among other alpha-synuclein pathologies, with 50% to 80% in dementia with Lewy bodies and 80% to 95% in multiple system atrophy (116).

Among younger adults (younger than 40 years old), RBD is more likely due to antidepressant medications or narcolepsy (15; 87). Fifty percent of patients with narcolepsy have abnormally high REM sleep motor activity (30). Among patients with young onset Parkinson disease, RBD is uncommon (90).

Although RBD is common, with 35 million expected patients worldwide, it is challenging to diagnose, and the vast majority of cases go unrecognized (02). Not unexpectedly, RBD is more commonly diagnosed among patients with other sleep complaints. In a study of 703 patients arriving at a sleep disorders center, 34 (5%) were eventually diagnosed with RBD (47). Importantly, of these 34 RBD patients, only six were referred to the sleep disorder center for suspected RBD; the rest were discovered only after a comprehensive sleep history was taken and polysomnogram performed.

Prevention

At this time, there are no interventions to prevent the development of RBD associated with neurodegeneration. Further, once RBD has developed, it is uncertain what measures may be taken to prevent or delay the onset of a neurodegenerative disorder. However, RBD is currently used as a biophysiological marker to identify patients at high risk of alpha-synuclein disease, and, hopefully in the future, will help identify neuroprotective therapies that impede parkinsonian disorders. A consortium of multinational sleep investigators, the International Rapid Eye Movement Sleep Behavior Disorder Study Group (IRBDSG), assembles annually in order to promote the development of collaborative clinical trials (145).

Growing evidence suggests that increased levels of exercise in early and mid-life decreases the risk of Parkinson disease and Alzheimer disease. Further, moderate to high intensity aerobic exercise slows the progression of Parkinson disease. Altogether, this suggests that exercise could be a potential therapeutic option to slow RBD progression, although this has not been formally studied (95).

A single proof of concept randomized control trial study evaluating selegiline as a neuroprotective agent in 30 patients with isolated RBD failed to show any difference in rate of phenoconversion between the control and selegiline arms (07). However, there were differences in DAT uptake in the caudate nucleus, but not the putamen between treatment arms at follow-up, suggesting a possibly protective effect and more importantly highlighting the need for biomarker derived measures of disease progression rather than clinical syndrome.

Another intriguing compound, already in use among RBD patients, is melatonin. This agent in high doses (6 to 15 mg) at bedtime abates dream enactment behavior (see Management section). Melatonin has been demonstrated to have antioxidant properties and to protect mitochondrial function, suggesting potential as a neuroprotective agent. However, no large studies of patients with RBD taking melatonin have shown a lower risk of phenoconversion associated with melatonin use (149). Obviously, prospective randomized controlled trials are needed prior to concluding that any agent may prevent or slow the onset of neurodegeneration.

Avoiding antidepressant medications when clinically appropriate and taking measures to reduce risk of stroke can prevent toxic and lesional RBD, respectively. Given the increased risk of RBD in heavy drinkers, decreasing alcohol intake may reduce risk of RBD.

Differential diagnosis

Confusing conditions

The most common disorders that need to be distinguished from RBD include the NREM parasomnias: confusional arousals, sleepwalking, and night terrors. Because of the disproportionate amount of REM sleep at the end of the sleep period, RBD episodes typically occur during the second half of the night in contrast to NREM parasomnias that more often occur in the first half of the night. Patients with RBD are more likely to act out and report dream mentation when awakened from an episode. Dream mentation is more aggressive in RBD, and violent punching and kicking are common. When sleepwalking does result in sleep related injury it is usually sustained during a sudden rush to escape the bedroom, often into or through a window (158).

Behaviors that may mimic RBD can occur during the REM sleep fragmentation associated with obstructive sleep apnea or sleep-related gastroesophageal reflux (15; 66). Further, patients with severe periodic limb movement sleep disorder may also report unpleasant dreams and simulate RBD-like behavior, requiring video polysomnography to distinguish between the two disorders (52).

Less likely etiologies include sleep-related dissociative disorder or nocturnal epilepsy. The behavior in sleep-related dissociative disorder is often prolonged, and polysomnography demonstrates wakefulness throughout the episode. Nocturnal epilepsy is characterized by stereotyped, recurrent, abnormal behaviors, and the EEG may demonstrate epileptic activity (64).

Parasomnia overlap disorder, a combination of both an NREM parasomnia and RBD, is frequently noted and should be considered as well (140). In a series, approximately 21% of all RBD cases and 28% of all sleepwalking/sleep terror cases were later determined to have parasomnia overlap disorder (140). In another report of 93 patients with RBD, 10 patients also had a history of sleepwalking or nocturnal wandering behavior (115).

Trauma associated sleep disorder (TSD) represents an overlap between REM sleep behavior disorder and posttraumatic stress disorder, sharing features with each but has been proposed as a separate entity. Patients with trauma associated sleep disorder have loss of normal skeletal muscle paralysis similar to RBD patients; however, unlike patients with RBD, patients with trauma associated sleep disorder have a history of trauma preceding the development dream enactment behaviors, and dream content typically mirrors that of the individual’s traumatic experience (37).

Diagnostic workup

Diagnosing RBD requires a thorough clinical evaluation, with a detailed review of the sleep-wake complaints followed by a neuropsychiatric history and examination. A report from a bed partner is particularly helpful as many patients are unable to properly recall the sleep-related events by the time they are discussed with a clinician.

Brief dream enactment behavior occurring in the latter half of the sleep period followed by alertness and orientation on awakening are features that help to distinguish RBD from other parasomnias. This presentation contrasts with sleepwalking in which there is often a lifelong history of prolonged, complex, nonviolent activities, typically emanating from the first half of the sleep period (02).

It is also useful to inquire about ancillary alpha-synuclein symptoms, such as difficulty with smell and bowel motility. When chronic unexplained hyposmia and constipation coexist with RBD, they are highly suggestive of an impending synuclein disorder.

Among patients with Parkinson disease, falling out of bed during sleep is highly suggestive of RBD and a risk factor for sleep-related injury (162).

Diagnostic tools. Screening surveys are available to help identify RBD among certain patient populations. The RBD Questionnaire-Hong Kong (RBDQ-HK) is a validated 13-item measure with scores ranging from 0-100. A higher score indicates greater severity of parasomnia symptoms (85). With a threshold score of 18, the RBDQ-HK has a positive predictive value of 86% and a negative predictive value of 83% (85). The REM Sleep Behavior Disorder Screening Questionnaire (RBDSQ) is a 10-item screen, and when administered in combination with a sleep history, a score greater than 4 has a sensitivity of 90% and a specificity of 87% (151). However, it should be pointed out that the RBDSQ performs much better when screening for RBD in the general population when compared with specific patient populations such as Parkinson disease (86).

Intriguingly, two separate groups of investigators have demonstrated an ability to detect RBD with only one question. The Mayo Sleep Questionnaire (MSQ) is a comprehensive sleep health survey, filled out by bed partners, that includes the following question: “Have you ever seen the patient appear to ‘act out his/her dreams’ while sleeping? (punched or flailed arms in the air, shouted or screamed)” (16). A separate group asked a separate question, the RBD Single-Question Screen (RBD1Q): “Have you ever been told, or suspected yourself, that you seem to ‘act out your dreams’ while asleep (for example, punching, flailing your arms in the air, making running movements, etc.)?” (126). The original studies demonstrated high diagnostic sensitivities (MSQ=100%, RBD1Q=94%) and specificities (MSQ=95%, RBD1Q=87%) for these questionnaires. However, both the RBD1Q and the RBD question in the MSQ were noted to have significantly lower specificities when compared to polysomnography (PSG) (MSQ=64%, RBD1Q=68%) (18). Conversely, it has been noted that PSG is not a faultless investigation as it can miss up to 20% of RBD cases (106). Thus, a high number of false negative RBD cases on PSG could explain the seemingly low specificity of the helpful, simple, clinical tools (124).

Visual actigraphy analysis has been shown to accurately diagnose RBD in patients with PSG-proven RBD, especially in combination with clinical history consistent with dream enactment behaviors, providing a potential low-cost alternative for RBD diagnosis when compared with PSG (150). However, further validation of this method is needed.

Polysomnography (PSG). According to current diagnostic criteria, PSG is necessary for definitive diagnosis (69; 02) and is helpful for ruling out other conditions (66). Even when abnormal behavior does not occur during a single night study, RSWA is often present, which in combination with clinical history of dream enactment behavior, can establish a diagnosis (168). Surface EMG of the submentalis, flexor digitorum superficialis, and anterior tibialis muscles are the most commonly analyzed muscles for RSWA. Submentalis and flexor digitorum superficialis are the most sensitive and specific muscles for identification of RBD, and submentalis RSWA appears to be highly specific for underlying synucleinopathy neurodegeneration, even in the absence of clinical dream enactment (48; 97; 102).

By American Academy of Sleep Medicine criteria, a 30-second PSG epoch is scored as RSWA when there is either a sustained elevation of EMG activity (greater than 50% of the 30-second epoch duration compared to minimal amplitude in NREM sleep) or excessive bursts of transient muscle activity (at least half of all 3-second mini epochs on a 30-second page).

RSWA is further quantified by the REM sleep atonia index (RAI) ranging from 0.0 to 1.0 (1.0 completely normal atonia). The RAI demonstrates night to night consistency with minimal variability (19). Defining what quantifies as excessive REM sleep motor activity is challenging, as some degree of REM sleep muscle tone is normal. In order to objectively quantify excessive motor activity, computer-assisted scoring methods have been developed, and an abnormal REM sleep atonia index is often defined as (< 0.88) (43; 49; 99; 46).

Often, early in the course of disease, patients with intermittent REM sleep-related motor behaviors consistent with dream enactment do not have enough RSWA to meet RBD criteria. In two years, however, follow-up PSG testing demonstrates that 69% of patients with these REM sleep behavioral events will convert to RBD (107).

Careful review of the PSG video can discern RBD from other motor parasomnias, and RBD is discernible from other parasomnias as the motor activity is more typically repetitive and pseudohallucinatory, frequently with hand babbling (limb wrist, flexed fingers-like a baby). This subtle dream enactment often involves only the upper extremities and, thus, EMG monitoring of forearm musculature should be included (66). Other REM sleep phenomena are often present, including snoring, and in males, penile tumescence.

PSG is also helpful in ruling out conditions such as sleep-disordered breathing, periodic limb movement disorder, and nocturnal epilepsy as a cause of the nocturnal behaviors. The sleep fragmentation of obstructive sleep apnea during REM sleep can lead to dream enactment behavior but resolves with treatment. Periodic limb movement disorder, primarily a NREM sleep phenomena, can be confused with RBD by history, but is easily discernible on polysomnography. Nocturnal epilepsy should be suspected, and a 32-lead EEG montage added, especially if there is a history of stereotyped, abnormal, and repetitive behaviors (64).

Other clinical investigations. Additional testing such as neuroimaging, EEG, and neuropsychological batteries are warranted only if there is further evidence suggesting a neuropsychiatric disorder or focal neurologic deficits on examination. The practical clinical utility of tests that identify CNS dopamine dysfunction, such as a DaTSCAN, or other markers of synuclein pathology, such as MIBG cardiac scintography, FDG-PET scan, or regional cerebral blood flow has yet to be determined clinically, although they are active areas of investigation in research (59).

Management

Management should initially focus on patient and bed partner safety, by modifying the sleeping environment. Subsequently, the clinician should eliminate aggravating agents, as well as identify and treat comorbid sleep disorders. Unfortunately, frequency of dream enactment behaviors is not predictive of injury, so bedroom safety must be addressed with all patients with RBD at a minimum (100). Most cases of toxic RBD are self-limited following discontinuation of offending medication, and dream enactment behavior typically resolves if underlying obstructive sleep apnea is treated. When violent nocturnal behaviors persist despite these interventions or in situations with a high probability of injury, pharmacotherapy is appropriate. However, there is an absence of large randomized controlled trials, and systematic reviews as well as professional societies have found insufficient evidence at present to make definitive conclusions regarding RBD therapy, although small randomized, placebo-controlled studies of clonazepam and melatonin have failed to show significant improvement in RBD symptoms with clonazepam or melatonin compared to placebo; however, it is possible that in the case of melatonin that patients were simply insufficiently dosed (146; 134; 76; 147). Instead, a consensus has arisen based on original reports, case series, and small clinical trials. Ultimately, clonazepam and melatonin, the mainstays of treatment for RBD, reduce the frequency and severity of dream enactment behaviors as well as injury potential, but complete elimination of dream enactment behaviors is rare.

Clonazepam. Clonazepam has been the most widely prescribed agent for RBD, and approximately 90% of patients initially respond well to low doses (0.5 to 1.0 mg) (115; 144). Clonazepam decreases the risk for dream enactment injury when administered at bedtime (42). The agent also appears to be effective in cases of Parkinson disease and narcolepsy (144; 142).

Clonazepam’s therapeutic mechanism in RBD is not fully understood, although it is thought that clonazepam may decrease the frequency of unpleasant nightmares, thus, decreasing dream enactment behaviors. In general, it is thought that clonazepam does not impact REM sleep without atonia, but some studies have suggested it may have an impact (84; 21). Long-term studies of clonazepam range from sustained benefit without dose escalation to others with a high incidence of dose escalation and treatment failure (142; 144; 55; 03; 66). In a series, 58% of patients on clonazepam reported clinically significant adverse effects, with 50% either stopping the agent or reducing the dose (03). Further, a prospective study of clonazepam in treatment-naive patients did not demonstrate an improvement in symptoms (44). Also, clonazepam is particularly problematic in the setting of advanced neurodegenerative disease where it’s prolonged duration of action may result in morning sedation as well as cognitive and gait impairment (66). A double-blind placebo-controlled trial of 0.5 mg of clonazepam in Parkinson disease patients with probable RBD found no difference between clonazepam and placebo-treated patients (147).

Melatonin. Several studies have suggested that melatonin may be an effective and safe first-line treatment for RBD. Melatonin is an endogenous hormone normally secreted by the pineal gland in response to evening darkness and helps entrain circadian rhythms. Exogenous immediate-release melatonin in high doses at bedtime (6 to 15 mg) augments REM sleep atonia and improves RBD symptoms (66). Although the exact mechanism of melatonin in RBD is unknown, it may be realigning the circadian rhythmicity of REM sleep. One interesting study demonstrated that patients with untreated RBD lose circadian-dependent features of REM sleep and often have a phase delayed circadian rhythm (08), whereas another study reported smaller pineal gland volume in Alzheimer disease patients with probable RBD who did not meet criteria for dementia with Lewy bodies (117).

Investigators have reported that melatonin is effective either in combination with clonazepam or as sole therapy (152; 81). Importantly, a direct comparison study noted that melatonin was equal to clonazepam in treatment efficacy and superior in side effect profile. Patients on melatonin reported fewer adverse effects, in particular falls and injuries, compared to clonazepam (96). In the setting of Parkinson disease, melatonin is a particularly intriguing option as it is only mildly sedating. A meta-analysis concluded that melatonin was efficacious in RBD among patients with neurodegeneration (169). On the other hand, two placebo-controlled trials of 2 mg, 4 mg, and 6 mg of prolonged release melatonin found no significant improvement in RBD symptoms in patients with either iRBD or Parkinson disease-RBD (76; 54). The reason prolonged-release melatonin has been less effective for RBD symptoms than immediate-release melatonin is unclear. It is likely that trial design has a strong influence (prolonged-release melatonin has been studied in double-blind placebo-controlled studies whereas immediate-release melatonin has been reported in observational or open-label studies, which are prone to bias), but it is also possible that doses trialed for prolonged-release melatonin have been insufficient as the median effective dose for melatonin has been reported to be 6 mg, which is the maximum dose studied for prolonged-release melatonin. Future controlled dose-finding trials comparing immediate- and prolonged-release melatonin are needed to answer this important question.

Melatonin suppresses both phasic and tonic REM sleep motor activity, and its effect persists for weeks after the agent is discontinued (15; 81). Also, ramelteon, a melatonin agonist, was reported to effectively treat RBD in two cases, but was ineffective in reducing objective dream enactment behaviors or altering RSWA in a small open-label clinical trial (114; 36).

Rivastigmine. Cholinergic agents may be useful in RBD among patients who have failed conventional therapy. Two placebo-controlled crossover trials noted that the cholinesterase inhibitor rivastigmine reduced the number of dream enactment behavior episodes among patients who already had mild cognitive impairment, thus an indication for rivastigmine (33; 20).

Pramipexole. Pramipexole may be effective in mild cases of RBD, in particular, those associated with frequent periodic limb movements. One investigation noted that compared to clonazepam-responsive patients, pramipexole-responsive patients had more mild disease at baseline as measured by REM sleep atonia (138). Similar to treating obstructive sleep apnea in RBD, pramipexole may decrease nocturnal behaviors by reversing a sleep fragmenting condition, periodic limb movements. Another pramipexole study in patients with RBD reported a decrease in distressing nocturnal behaviors along with a decrease in periodic limb movements, but no effect on REM sleep atonia (137). Alternatively, it may be that only RBD patients with abnormal dopamine uptake on DATSCAN may find RBD treatment response with dopamine agonists such as pramipexole, although this requires further study.

Other therapies. Other agents with some reported success include prazosin, imipramine, carbamazepine, rotigotine, levodopa, donepezil, sodium oxybate, triazolam, zopiclone, quetiapine, clozapine, and cannabidiol (23; 66). However, the data supporting efficacy of these agents are weak.

Surgical treatments for Parkinson disease including deep brain stimulation do not appear to strikingly improve RBD. There is often an improvement in subjective sleep quality and sleep architecture on polysomnography; however, there has been no noted improvement in dream enactment behavior, and in some cases, it can actually worsen (27; 78). These findings are not unexpected as the target of deep brain stimulation, the subthalamic nucleus, does not have a known effect on REM sleep.

Interestingly, RBD symptoms tend to decrease with progression of the underlying neurodegenerative condition, whereas RSWA increases, and RBD patients with a diagnosed neurodegenerative disease are less likely to be injured by their dream enactment behaviors than those with iRBD (99; 45; 166). Although the mechanism of this moderation of dream enactment behavior is uncertain, it may be due to diffuse impairment of motor circuits with advanced disease, leading to the replacement of strong phasic bursts of RSWA seen in iRBD with low amplitude continuous tonic muscle activity that is more characteristically seen in RBD patients with comorbid neurodegenerative disease (98; 45).

Refractory RBD and bed alarm therapy. Medication refractory RBD is a daunting and potentially life-threatening condition with limited management options. Exiting the bed while acting out a dream is a particularly high-risk behavior and may result in severe traumatic injury (143). Intriguingly, the low arousal threshold and rapid transition to alert wakefulness from REM sleep offers a therapeutic window to halt behavior prior to sleep-related injury (73). Despite apparent unconsciousness during REM sleep, the brain is readily responsive to complex auditory sound processing (73; 10). This contrasts with the high arousal threshold of NREM sleep often demonstrated by the inability to redirect or wake up sleepwalkers (a NREM parasomnia) (120). Further, other REM pathologies, such as nightmare disorder, can be treated with verbal redirection (64).

A study of patients with medication refractory RBD and sleep-related injury demonstrated the utility of a customized bed alarm that delivered a calming message at the onset of dream enactment behavior (65). Ideal voices, typically those of family members, were identified and commands to halt dream enactment behavior were then recorded (eg, “Dave, you are having a dream, lay back down”). Subsequently, when the patient arose during sleep, the voice emanated from a bedside speaker on a repeating loop until the patient returned to lying down on the pressure pad.

In conclusion, there is a clinical consensus that clonazepam and melatonin are reasonable first-line therapies; however, more methodologically rigorous studies are needed to convincingly prove efficacy. Importantly, agencies that regulate drug approval do not recognize clonazepam, or any other agent, so its use is off-label. The American Academy of Sleep Medicine has assembled a task force to review and provide management recommendations for RBD. The task force is expecting to publish their findings in 2021.

Patient counseling. The diagnosis of isolated RBD provides an ethical dilemma for many clinicians in terms of who to inform of the neurodegenerative risk associated with RBD. Although circumstances surrounding each individual patient should be taken into account when deciding to disclose the risk for future neurodegenerative disease, general consensus is to inform patients of this risk; however, the degree of granularity in which this risk is discussed is variable (154). In particular, if patients develop RBD after age 60, are not taking antidepressants, and already have nonmotor features of synucleinopathy, they should certainly be informed of the risk of neurodegenerative disease and referred to a center specializing in RBD. For individuals diagnosed with RBD at a young age (ie, less than 50 years of age), are associated with antidepressant medications, and have the presence of structural lesions or autoimmunity, the risk of neurodegeneration is less clear, and clinicians should use their own judgement whether to inform patients of neurodegenerative risk, although it would be reasonable to observe these patients closely for evolving signs of neurodegeneration (05).