Sleep Disorders in Atypical Parkinsonism


  • Relevant disclosures and conflicts of interest are listed at the end of this article.


Sleep disorders are commonly observed in atypical parkinsonism, with particular disorders occurring more frequently in specific parkinsonian disorders. Multiple system atrophy (MSA) is a synucleinopathy often associated with nocturnal stridor, which is a serious, but treatable, condition highly specific to MSA. In addition, this disorder is strongly associated with rapid eye movement sleep behavior disorder (RBD), which is also observed in dementia with Lewy bodies. RBD is far less prevalent in progressive supranuclear palsy (PSP), which is a tauopathy. Insomnia and impaired sleep architecture are the most common sleep abnormalities observed in PSP. Corticobasilar degeneration is also a tauopathy, but has far fewer sleep complaints associated with it than PSP. In this article, we review the spectrum of sleep dysfunction across the atypical parkinsonian disorders, emphasize the importance of evaluating for sleep disorders in patients with parkinsonian symptoms, and point to sleep characteristics that can provide diagnostic clues to the underlying parkinsonian disorder.

Disrupted sleep and alertness are increasingly being recognized as common symptoms in neurodegenerative disorders that negatively affect the quality of life and safety of patients. Furthermore, disturbed sleep-wake cycles may be an early marker of ongoing neurodegeneration, providing a window of opportunity to apply disease-modifying interventions early in the course of neurodegeneration. Parkinson's disease (PD), the second-most common neurodegenerative disorder, may be looked upon as an excellent example where efforts to understand impaired sleep-wake homeostasis led to significant advances in understanding the disease biology and progression. In contrast to PD, sleep disorders in atypical parkinsonian syndromes, such as Multiple system atrophy (MSA), dementia with Lewy Bodies (DLB), progressive supranuclear palsy (PSP), and corticobasal degeneration (CBD), have not been systematically studied. Sleep dysfunction can frequently be observed in these patient populations. In one survey of 57 patients with MSA, compared to 62 patients with PD, sleep disturbances were found in 70% of MSA patients, compared to only 62% of PD patients.[1] Whereas there is substantial overlap in the types of sleep problems encountered across atypical parkinsonism, there are some sleep disorders that are either unique to, or far more prevalent in certain parkinsonian disorders. Efforts to understand these differences may provide clues to the unique pathophysiology of each of these disorders. This review highlights sleep disorders associated with atypical parkinsonism.


MSA is a neurodegenerative disorder characterized by a combination of autonomic dysfunction, parkinsonism, and/or ataxia. Based on the combination of symptoms, MSA is divided into MSA-P (parkinsonism subtype) and MSA-C (cerebellar subtype). MSA, a synucleinopathy, is characterized microscopically by glial cytoplasmic inclusions.[2] Macroscopic changes vary depending on the subtype, with cerebellar atrophy in MSA-C and pronounced substantia nigra (SN) pallor in MSA-P.[3] The underlying neuropathology may contribute to the sleep disorders that are unique to this disorder. Sleep dysfunction is a well-recognized comorbidity associated with MSA.

Sleep Disordered Breathing

Sleep disordered breathing of various types can occur in MSA. Obstructive sleep apnea (OSA) can be observed, but there are also case reports of central apneas developing after treatment of the OSA[4] or of central sleep apnea occurring as the first symptom of MSA.[5] Nocturnal stridor is one of the most serious sleep disorders associated with MSA. Though laryngeal dysfunction was reported on in the early descriptions of MSA,[6, 7] its clinical significance was not recognized until many years later. Multiple case reports have described vocal cord palsy presenting as nocturnal stridor in patients with MSA. In the largest study of 42 patients with MSA, 17 had nocturnal stridor, which was associated with a significantly shorter survival period.[8] Another study of 19 MSA patients showed that 42% had nocturnal stridor.[9] Laryngoscopy has demonstrated variable degrees of pathology in patients with MSA, ranging from unilateral palsy of the cricothyroid muscle with normal abductor function[10] to normal vocal cord function during wakefulness.[11] If present, the unilateral vocal cord paralysis does not improve with vocal cord pinning.[12] These findings suggest that laryngeal dysfunction either worsens or is only present during sleep. Recent studies that compared laryngoscopic evaluation during wake and sleep demonstrated paradoxical movement of the vocal cords during sleep, which was not observed during waking.[13, 14] Nocturnal stridor is quite serious and potentially fatal, with several case reports of untreated MSA patients dying during sleep.[15] In a longitudinal study of 45 patients with MSA, among 10 patients who died, 6 had sudden death during sleep.[16]

The pathophysiology of nocturnal stridor in MSA is not fully elucidated. Autopsy studies have shown neurogenic atrophy of the posterior cricoarytenoid muscle.[17] Electromyography (EMG) of the laryngeal muscles during sleep demonstrates increased activation of the thyroarytenoid muscle in association with nocturnal stridor, which improves with continuous positive airway pressure (CPAP).[18] This respiratory dysfunction has been attributed to loss of neurons in the nucleus ambiguus (NA) within the brainstem, however in a study that compared MSA and PD patients without stridor, no significant difference in the number of neurons in the NA was observed.[19] Two autopsy studies demonstrated depletion of cholinergic neurons in the arcuate nucleus in patients with MSA, which may in part be responsible for the observed respiratory disturbances.[20, 21]

Tracheostomy has previously been used as a treatment for sleep disordered breathing in MSA and has been demonstrated in at least one case to improve the previously observed paradoxical vocal cord activation.[22] However, the use of tracheostomy is somewhat controversial, because there are rare reports of central sleep apnea worsening subsequent to tracheostomy,[23] but the general consensus is that tracheostomy should be considered for any patient with daytime stridor or evidence of immobile vocal cords.[24] CPAP has also been demonstrated to be effective at eliminating the nocturnal stridor,[25] with continued efficacy demonstrated on long-term follow-up.[26] Several recent studies demonstrated benefits of bilevel positive airway pressure, optimally titrated using laryngoscopy,[27] and adaptive servoventilation may also be an option for patients with predominantly central apneas. However, despite these treatment options, there are several case reports of patients with nocturnal stridor that still died suddenly during sleep, despite treatment with either tracheostomy or noninvasive ventilation.[16]

Rapid Eye Movement Sleep Behavior Disorder

Rapid eye movement (REM) sleep behavior disorder (RBD) is a parasomnia characterized by loss of the normal muscle atonia during REM sleep, resulting in potentially injurious dream enactment behavior. RBD was initially described in 1986[28] and has been associated with the later development of neurodegenerative disorders,[29] with MSA being one of the most common. For example, in one case series an increase in sleep talking was one of the earliest manifestations of MSA.[30] Case reports of RBD are present across the clinical continuum of MSA, including MSA-P[31] and MSA-C.[32] Early series reported a clinical prevalence of RBD in up to 69% of MSA patients. REM sleep without atonia is even more prevalent and was found in 90% of polysomnograms (PSGs) performed on this population.[33] Though RBD appears to be more frequent in MSA patients, compared to PD patients, the overall severity of the disease may not differ significantly between these two populations.[34] Neuroanatomically, the severity of RBD in MSA patients correlates with the degree of loss of striatal monoaminergic neurons, but not with the degree of loss of thalamic cholinergic neurons.[35] It was initially thought that RBD could help distinguish patients with MSA from those affected with pure autonomic failure (PAF).[36] However, a later study demonstrated that nocturnal vocalizations were just as frequent in PAF as in MSA.[37] Of interest are reports documenting improvement in waking motor impairments during episodes of RBD in patients with MSA.[38, 39] These observations raise questions regarding the differences in regulation of muscle control during REM sleep and wakefulness.

From a treatment standpoint, there are two primary treatments that have been studied for RBD occurring outside of the context of DLB, which will be detailed further in the next section. Clonazepam was the first medication described for use in RBD and remains the mainstay of treatment for many physicians.[28] However, there is also promising evidence for the use of melatonin in the treatment of RBD, with a double-blind, placebo-controlled trial using 3 mg of melatonin demonstrating significant improvement in symptoms, as well as a decrease in the tonic EMG activity recorded on PSG, which persisted even after stopping the melatonin.[40] More recently, a chart review of 45 patients with RBD demonstrated that both clonazepam and melatonin improved the frequency and severity of dream enactment behavior to a similar degree, though fewer injuries and adverse effects were noted in the patients treated with melatonin (average dose: 6 mg).[41] However, because most evidence to date is observational, the International RBD Study Group recently issued a consensus statement with a proposed design for a large, controlled treatment study for comparing the use of clonazepam and melatonin for treatment of RBD.[42]

Excessive Daytime Sleepiness

Excessive daytime sleepiness (EDS) has been reported in 28% of Caucasian patients with MSA[43] and 24% of Japanese patients with MSA.[44] The etiology of EDS in MSA is likely multifactorial and includes medication effects, the intrinsic pathology of the disease, and coexistent primary sleep disorders. Sleep attacks during a levodopa challenge have been reported in the MSA population.[45, 46] PSG recordings of patients with MSA reveal an overall normal sleep architecture structure, but an unexplained decreased nocturnal sleep time, which may contribute to daytime sleepiness.[47] One of the potential neuroanatomic substrates for EDS in MSA is the hypocretin neuronal network within the lateral hypothalamus, which plays an important role in promoting alertness. In an autopsy study of 7 MSA patients, the number of hypocretin neurons was significantly lower, compared to controls.[21] Other studies, however, showed that MSA patients have normal hypocretin levels in the cerebrospinal fluid (CSF),[48, 49] suggesting that it may require substantial neuronal degeneration before a measurable difference in CSF levels is observed. Alternatively, the excessive daytime sleepiness observed in MSA could be related to loss of cholinergic alerting signals from the brainstem, since it has been demonstrated that patients with MSA have a significant loss of cholinergic neurons within the laterodorsal tegmental (LDTg) and pedunculopontine tegmental nuclei (PPTg) in the pons.[50] Because these nuclei are active during wake and REM sleep, this cell loss has been proposed to be important both for the excessive daytime sleepiness and the RBD commonly observed in patients with MSA.

Other Sleep Disorders

There is limited literature on periodic limb movements in MSA. In a study comparing patients with PD, MSA, and healthy controls, MSA patients had fewer periodic limb movements of sleep (PLMS) and waking than PD patients, but more than controls.[51] MSA patients may have a decrease in the cortical and autonomic arousal responses to periodic leg movements, suggesting that MSA patients may experience less-disrupted sleep from limb movements, when compared to controls.[52]

The function of the circadian system in MSA has not been systematically investigated. Impaired circadian rhythm of blood pressure and a high prevalence of nocturia in the MSA population provide indirect evidence that circadian disruption may persist in MSA. One of the hallmark symptoms of MSA is a loss of the normal nocturnal dip in blood pressure, with early case reports of MSA demonstrating an overnight rise in both systolic and diastolic blood pressure,[53] which appears to be much more pronounced in patients with MSA, compared to PD.[54] Nocturia is also a common complaint in MSA. Normal circadian rhythms of vasopressin that control urine volume production appear to be altered in MSA. Arginine vasopressin levels in MSA are high during the daytime and low at night, which represents a reversal of normal diurnal vasopressin rhythms.[55-57] Two autopsy studies demonstrated a decrease in vasopressin neurons in the suprachiasmatic nucleus, the primary circadian pacemaker, in patients with MSA.[58, 59] Based on these findings, a small trial of nighttime intranasal administration of desmopressin resulted in improvement in nocturia in patients with MSA.[60] Several studies reported impaired circadian rhythms of growth hormone secretion,[61] blunted 24-hour core body temperature amplitude,[62] blunted circadian variation in gastric motility,[63] and decreased morning cortical secretion[64] in patients with MSA. These studies provide further evidence for circadian disruption in MSA.


DLB is a synucleinopathy clinically characterized by fluctuating cognitive performance and alertness, visual hallucinations, and parkinsonism. Sleep disorders in DLB are common, with the reported prevalence of sleep dysfunction in the DLB population ranging from 44% to 55%.[65, 66] Although there is significant overlap between sleep disorders reported in DLB and PD, the presence of sleep dysfunction can facilitate differentiation between DLB and other dementias. For example, sleep disturbances appear to be significantly more common in DLB, when compared to Alzheimer's disease (AD; odds ratio = 2.93).[67]


The majority of the literature related to sleep dysfunction in DLB has focused on RBD as a symptom to distinguish DLB from the other dementias. Initially reported in 1995, the first case report of RBD in DLB revealed marked neuronal loss in the locus ceruleus and SN.[68] Several case series followed and reported that all[69] or nearly all[70] demented patients with RBD met the clinical diagnostic criteria for DLB, which was later confirmed by autopsy.[69] Among 360 patients with dementia, those with DLB were more likely to have RBD, when compared to patients with AD and mild cognitive impairment (56% vs. 2%).[71] A recent study of PSGs in 72 patients with DLB found that 96% had reports of dream enactment behavior and 83% had REM sleep without atonia demonstrated on PSG. In addition, a large case series was conducted evaluating the clinical and neuropathologic diagnoses of 172 patients with RBD, which found that the most common diagnosis (45% of cases) was DLB.[72] Overall, the strong association between RBD and DLB prompted the inclusion of RBD as a core clinical feature for the diagnosis of DLB in 2011.[73] Similar to patients with MSA, the presence of RBD in DLB is neuropathologically associated with marked loss of cholinergic neurons from the PPTg in the brainstem.[74, 75] Though the traditional treatment choices of clonazepam or melatonin can be used, additional treatment options for RBD in DLB have been reported in several case series. Rivastigmine, clonazepam, donepezil, and l-dopa have all been reported to be beneficial for RBD associated with DLB.[76-78] There is evidence that treatment with rivastigmine normalized sleep in a small number of these patients.[79] A study of 3 DLB patients found that treatment with either clonazepam or donepezil resulted in significant improvement in the number of RBD episodes.[80] Another case report of an individual who was unable to tolerate clonazepam as a result of hallucinations demonstrated that l-dopa may be effective for treating RBD.[76] In 20 patients with either PD or DLB and RBD, memantine significantly decreased abnormal movements during sleep, when compared with placebo.[77]


EDS is more common in DLB, compared to AD (75% DLB vs. 48% AD),[78] so it may be helpful in distinguishing between DLB and AD. Unlike some of the other atypical parkinsonian disorders, DLB is not associated with decreased CSF hypocretin levels.[81, 82] This is in contrast to the decreased hypocretin immunoreactivity observed in neuropathological studies of autopsied brains from DLB and AD patients.[83] This is similar to what is observed in MSA, and this discrepancy may reflect the need for significant neuronal loss before noticeable changes in CSF hypocretin levels can be detected.[84] Because some features of DLB are similar to narcolepsy, including hypersomnia and visual hallucinations, and there is loss of hypocretin neurons in both, an evaluation was performed looking at the prevalence in DLB patients of the human leukocyte antigen haplotype, DR2-DQ6, which is more common in narcolepsy. In this study, no significant increase in this specific haplotype in DLB patients was observed, suggesting that these two disorders do not share the same mechanism of hypocretin neuronal loss.[85]

Other Sleep Disorders

Other sleep disorders can also be observed in DLB, including PLMS. In one study comparing 9 DLB patients, 12 AD patients, and 10 nondemented patients, PLMS were significantly higher in DLB patients.[86]


PSP, initially described by Steele et al. in 1964, is characterized by supranuclear ophthalmoplegia, postural instability, parkinsonism, and pseudobulbar palsy.[87] PSP is a tauopathy resulting in marked atrophy of the midbrain.[88] Tauopathies result from abnormal accumulation of tau, a microtubule-associated protein, resulting in neurofibrillary tangles.

Sleep Architecture Changes

Abnormalities in sleep architecture and insomnia are more frequently described in PSP than in the other neurodegenerative disorders. In one of the first case series to evaluate sleep architecture in PSP using PSG, all 4 patients had significantly decreased total sleep time and REM sleep, along with a significant increase in the number and duration of nocturnal awakenings. Non-REM sleep was also abnormal, with decreased spindles and an increase in alpha-frequency activity throughout sleep.[89] Additional case reports have also reported decreased total sleep time[90] and fragmented sleep with decreased slow wave and REM sleep.[91] A larger case series of 10 patients found a correlation between the severity of the disease and the decrease in total sleep time.[92] Quantitative EEG on 6 subjects with PSP confirmed findings of previous studies, demonstrating a shorter total sleep time, decreased sleep efficiency, sleep spindles, and REM sleep.[93] These findings are likely explained by the preferential degeneration of the pontine tegemental nuclei in the brainstem, particularly the PPTg, thought to be important for the generation of REM sleep, and the connections between the PPTg and the reticular nuclei of the thalamus are important for generating spindles.[94] Interestingly, despite objective data showing a significant decrease in sleep efficiency, when compared to patients with PD, subjective reports of sleep quality using the Parkinson's Disease Sleep Scale did not differ significantly between patients with PD and PSP.[95] One study that evaluated the levels of hypocretin in the CSF of patients with PSP found levels in PSP patients to be significantly lower, compared to PD patients, and levels were inversely correlated with disease duration.[96]

Other Sleep Disorders

Sleep disordered breathing does not appear to be significantly associated with PSP. Among 11 patients with PSP, there was no evidence of clinically significant sleep disordered breathing, though many patients had disturbances in voluntary respiration.[97] In another comparative study of PSP and PD, there was no significant difference in the proportion of sleep disordered breathing in the PSP patients, compared to the PD patients.[95]

Although more commonly reported in synucleinopathies, RBD has been observed in PSP as well. In a relatively small study of 15 PSP patients, 15 PD patients, and 15 age-matched controls, REM sleep without atonia and RBD were equally present among PSP and PD patients (33% vs. 28%).[98] However, in another study that examined 20 PSP patients and 93 PD patients, RBD and REM sleep without atonia were present in 60% of PD patients and only 20% of PSP patients.[99]

Circadian rhythms are not well characterized in PSP. Decreased amplitude of the core body temperature has been reported in patients with PSP, when compared to patients with PD. The significance of this comparison is unclear, because patients with PD have also been found to have circadian rhythm abnormalities in core body temperature.[100] Similar to patients with MSA, patients with PSP have been found to have a loss of the nocturnal fall in blood pressure, though PSP patients are not as severe as MSA patients.[54]


CBD is a tauopathy that is much less common than the other atypical parkinsonian disorders. Clinical features include a combination of both cortical and parkinsonian features.[101] Gross neuropathology is notable for asymmetric focal cortical atrophy, whereas the hallmark microscopic pathologic findings are ballooned neurons in the cortex, along with astrocytic plaques.[102]

The literature on sleep disorders in CBD is primarily limited to case reports, likely a result of the rarity of CBD. Two case reports of RBD in CBD have been reported on.[103, 104] Neither CBD case was confirmed by autopsy, which raises the question of whether these cases were truly CBD or a different parkinsonian disorder. In the largest case series of CBD patients, only 1 of 11 had RBD. This patient did have neuropathology-confirmed CBD.[105] In a descriptive study of 5 patients with CBD, all 5 had insomnia, 4 had periodic limb movements during sleep, 2 had sleep disordered breathing, and none had symptoms suggestive of RBD.[106] Interestingly, in several of these patients, the periodic limb movements were more severe on the side dominantly affected by parkinsonism.[107]

CSF hypocretin levels have also been evaluated in CBD and found to be significantly lower than in patients with PD. The study was quite small (n = 7), and the majority of the effect may have been the result of primarily 1 outlier.[96]


In this review, we highlighted the sleep disorders observed in atypical parkinsonism. Though there is significant overlap in the spectrum of sleep disorders observed within atypical parkinsonism, each disorder has a unique profile of sleep dysfunction associated with it (Fig. 1). These profiles can provide insight into the specific neuropathology and pathophysiology of each disorder. In addition, in cases where the clinical diagnosis may be uncertain, the presence or absence of these various sleep disorders may provide additional clues to the underlying diagnosis. Furthermore, some sleep disorders, such as RBD, may provide an early marker of developing neurodegeneration, similarly to idiopathic PD.

Figure 1.

Overlap between sleep disorders in the atypical parkinsonisms. MSA and LBD are synucleinopathies (dark gray), whereas PSP and CDB are tauopathies (light gray). EDS can be observed in all disorders, whereas RBD is more common in the synucleinopathies. LBD, Lewy body dementia.

MSA and DLB are synucleinopathies. There is growing evidence that RBD is much more prevalent among synucleinopathies, compared with other atypical parkinsonian syndromes.[72] Current theories suggest that RBD may result from damage to the sublaterodorsal nucleus in the brainstem,[108] which is one of the areas affected early in synuclein-associated neurodegeneration. Cholinergic neurons appear to be affected in MSA. The loss of these neurons from the arcuate nucleus may be contributing to the nocturnal stridor, whereas loss of neurons from the PPTg and LDTg, areas that are both active during REM sleep, may contribute to changes in overall sleep architecture.[94] Furthermore, hypothalamic neuronal loss, generally attributed to the autonomic dysfunction in MSA, could also contribute to circadian disruption.[109]

PSP and CBD are tauopathies. Sleep dysfunction appears to be less prevalent in CBD, compared with PSP. The most striking sleep abnormalities in PSP are alterations of sleep architecture, with decreased spindles and REM sleep. Degeneration of the PPTg and LDTg, which normally contain REM-on neurons,[94] as well as degeneration of the thalamus,[110] which is normally responsible for the generation of sleep spindles, may explain these alterations in the sleep of PSP patients. There is evidence for differences in the intracellular processing of tau aggregates in CBD, compared to PSP, resulting in alterations in the underlying neuropathology,[111] which may explain the predilection for the cortex in CBD, compared to the brainstem in PSP. Interestingly, there is evidence that PLMSs are worse on the side of the cortical and basal ganglia atrophy, which may provide some insight into the underlying pathophysiology of the PLMS.[106]

There are limitations to the current studies. Numerous comorbidities are associated with the atypical parkinsonian disorders, including anxiety, depression, psychosis, and medication side effects, which can all contribute to poor overnight sleep and/or daytime sleepiness. It is therefore challenging to characterize sleep phenotypes without significant confounders. The majority of studies to date involved small patient cohorts and were based primarily on clinical diagnosis, lacking neuropathological confirmation. Clinicopathologic studies have shown that the accuracy of the clinical diagnosis ranges from 48% for CBD[112] to 78% for PSP,[113] which indicates that studies not including postmortem confirmation of diagnosis may have inaccurately classified individuals. DLB is even more difficult to classify, and there is significant overlap with individuals diagnosed with PD with dementia.

Important future directions for the study of sleep in atypical parkinsonism will be to prospectively assess sleep-wake cycles in larger cohorts of patients using self-reported and objective methods of sleep and alertness. These studies should also incorporate treatment algorithms, because there are only a limited number of studies to date that have evaluated treatment options for sleep disorders associated with atypical parkinsonism. Early presentation of sleep disorders, such as RBD, in the process of neurodegeneration provides an excellent opportunity to intervene with neuroprotective agents, when available.


Overall, the atypical parkinsonism disorders can sometimes be difficult to distinguish clinically. Including assessment for sleep disorders in the clinical evaluation may provide further insight into the underlying movement disorder and, in the case of nocturnal stridor in MSA, may also have important treatment outcomes.

Author Roles

(1) Research Project: A. Conception, B. Organization, C. Execution; (2) Statistical Analysis: A. Design, B. Execution, C. Review and Critique; (3) Manuscript: A. Writing of the First Draft, B. Review and Critique.

S.M.A.: 3A  A.V.: 3B


Funding Sources and Conflicts of Interest: This study was supported by the National Institute of Neurological Disorders and Stroke (T32 NS047987 [to S.M.A.] and K23 NS072283 [to A.V.]). The authors report no conflicts of interest.

Financial Disclosures for previous 12 months: The authors report no financial disclosures.