Current Update on Clinically Relevant Sleep Issues in Parkinson’s Disease: A Narrative Review

Screening tool for PD-related sleep problems

Regarding the scales used for the assessment of sleep problems, the PDSS, the Pittsburgh Sleep Quality Index (PSQI), and Scales for Outcomes in Parkinson’s Disease–Sleep (SCOPA-sleep) are recommended for rating overall sleep problems, and the Epworth Sleepiness Scale (ESS) is recommended for assessing daytime sleepiness [22]. The ESS and PSQI are widely used but were not developed for PD patients. The PDSS-2 is a revised version of the PDSS that evaluates 15 PD-related sleep and nocturnal problems using a frequency measure ranging from 0 (never) to 4 (very frequent) [23]. Total PDSS-2 scores range from 0 to 60, with higher scores indicating greater severity [24]. Cutoff values for the PDSS-2 total score, namely, ≥15 for the detection of poor sleep [25] and≥18 for clinically relevant sleep problems [26], have been reported. When sleep-maintenance insomnia or early awakening is identified by PDSS-2 subitems, screening for motor symptoms (wearing off) and other nonmotor symptoms and adjustment of the dopaminergic medication dosage should be performed.

Treatment of insomnia

The first step in the treatment of insomnia is screening for and managing nocturnal symptoms. In patients with sleep-onset insomnia, RLS should be ruled out. Sleep-maintenance insomnia can result from various problems, such as motor symptoms, pain, nocturia, sleep apnea, anxiety, and depression. Initially, sleep hygiene instructions should be given, such as refraining from consuming caffeine and alcohol; refraining from smoking; avoiding naps; adjusting the bedroom environment; and getting regular exercise. In a meta-analysis on the effects of exercise on sleep quality in PD patients, regular exercise at a moderate to maximum level of intensity improved subjective sleep quality, but exercise at a mild to moderate level of intensity showed nonsignificant effects [27]. In a randomized controlled trial with PD patients assigned to an exercise group, which performed supervised exercise 3 times a week for 16 weeks, and a no exercise group, which engaged in in-person discussions and monthly phone calls about sleep hygiene, significant improvements in sleep efficiency, total sleep time, wake after sleep onset, and slow-wave sleep (SWS) were observed in the exercise group compared with the no exercise group [28].

Long-acting dopamine agonists, such as rotigotine, are useful for treating nocturnal PD-related symptoms [29]. In a 6-week, randomized, placebo-controlled trial in 30 PD patients, eszopiclone did not increase the total sleep time but improved sleep quality and decreased the number of awakenings [30]. Prolonged-release melatonin [31] and melatonin [32] have also been shown to be effective as treatments for insomnia in patients with PD in randomized, double-blind, placebo-controlled trials. Dual orexin receptor antagonists such as suvorexant and lemborexant, which have different actions relative to GABA_A agonists, had fewer next-day residual effects [33] and were safe for non-PD individuals with comorbid sleep apnea [34]. Because of their safety and tolerability profile and lack of dependence, dual orexin receptor antagonists may be useful and recommended for treating insomnia in patients with various psychiatric and neurodegenerative diseases [35]. The treatment of irregular sleep-wake rhythm disorder associated with mild-to-moderate Alzheimer’s disease with lemborexant is under investigation [36].

Pathology and imaging findings relevant to EDS

In PD, dopamine neurons of the substantia nigra pars compacta (SNpc) and dopamine content in the striatum are reduced by 60%and 80%, respectively, by the time motor symptoms appear [37]. The SNpc dopaminergic neurons project directly to the thalamocortical circuitry that promotes arousal and feed back to the ascending reticular activating system along with the ventral tegmental area (VTA) and the ventral periaqueductal gray matter. In PD, it is possible that the degeneration of these dopaminergic neurons can cause EDS [38]. In a study of 21 patients with early PD who underwent FP-CIT SPECT, there was no correlation between dopamine transporter (DAT) binding and motor function or disease severity, but the ESS scores were negatively correlated with DAT binding in the striatum, caudate nucleus, and putamen [39]. In an MPTP-induced PD model, EDS and sleep-onset REM periods were reversible with levodopa [40]. These observations may support dopaminergic contributions to EDS.

In PD, in addition to the dopaminergic system, the brainstem and hypothalamic nuclei, which are involved in sleep-wake regulation, are also affected. The degree of neuronal depletion involved in sleep-wake regulation is 37–55%in the serotoninergic raphe nuclei, 32–46%in the basal forebrain cholinergic system of the nucleus basalis of Meynert, 46%in the pedunculopontine tegmental nucleus/laterodorsal tegmental nucleus, and 63–78%in the noradrenergic locus coeruleus [38]. These findings may support a link between PD-related pathology and EDS. However, the relevance of the involvement of the brainstem and hypothalamic nuclei in the clinical manifestations of EDS is still uncertain. EDS was found relatively more frequently in patients with PD and PD with dementia; however, CSF hypocretin-1 (orexin-A) levels did not differ among controls, patients with PD and patients with PD with dementia, and there was no correlation between CSF orexin levels and ESS scores or sleep latency on the MSLT [41]. Regarding the interpretation of these findings, it should also be noted that in animal studies, a 50%decrease in CSF orexin levels requires the degeneration of 70%or more of the hypothalamic orexin neurons [42].

The serotonin transporter availability of dorsal raphe nuclei measured with ¹²³I-FP-CIT SPECT was reduced in PD patients with resting tremor compared with those without resting tremor, and serotonin transporter availability was correlated with tremor amplitude but not ESS or RBD screening questionnaire scores [43]. On the other hand, a study using [¹¹C]DASB PET showed that serotonin transporter availability in the caudate nucleus, putamen, ventral striatum, thalamus, hypothalamus, and raphe nuclei was significantly reduced in PD patients with sleep disturbances (PDSS scores < 82) compared with PD patients without sleep disturbances [44]. Further brain imaging studies on EDS and the clinical correlates are warranted.

Natural course of and risk factors for EDS in PD

In a multicenter study, EDS was found in 165 (37.8%) of 436 PD patients, and logistic regression analysis revealed that male sex, Movement Disorder Society (MDS)-sponsored revision of the Unified Parkinson’s Disease Rating Scale (MDS-UPDRS) III scores (motor symptoms), hallucinations, and MDS-UPDRS II scores (aspects of motor symptoms experienced in daily life) were associated with EDS [13]. In the analysis by clinical subtype, ESS scores were significantly higher in the patients with the postural instability and gait disturbance (PIGD) type than in those with the tremor-dominant type. The prevalence of EDS increases after the onset of PD motor symptoms. In an 8-year follow-up study of 232 PD patients, the proportion of patients with EDS increased from 5.6%at the beginning of the study to 22.5%at 4 years and 40.8%at 8 years [45]. Logistic regression analysis showed that factors contributing to EDS were older age, male sex, and dopamine agonist use, but in a subanalysis of the patients who were never treated with dopamine agonists, the predictor of EDS was Hoehn-Yahr (HY) stage alone. This result indicates that EDS is related to disease severity, independent of the effects of dopamine agonists.

In a 5-year follow-up study of 413 PD patients, EDS (Scales for Outcomes in Parkinson’s Disease-Sleep-Daytime Sleepiness (SCOPA-SLEEP-DS) score≥5) was found in 43%of the patients, and 46%of the patients without EDS at the start of the study subsequently developed EDS [46]. The factors associated with the development of EDS over time included male sex, poor nighttime sleep quality, cognitive impairment, autonomic dysfunction, hallucinations, mild dyskinesia, dopamine agonist use, and antihypertensive use. The predictors of the future development of EDS in patients without EDS at baseline were a higher baseline SCOPA-SLEEP-DS score, a PIGD phenotype and urinary tract symptoms. Furthermore, in a follow-up study of untreated PD patients, the proportion of patients with EDS increased from 11%at the start of the study to 23.4%at 5 years. Risk factors for the development of EDS at 5 years included an ESS score > 5 at baseline, male sex, dopamine agonist use, depressive symptoms and high UPDRS-Activities of Daily Living (ADL) scores [47]. Similarly, a prospective study using data from the Parkinson’s Progression Markers Initiative, which included 423 untreated PD patients within 2 years of diagnosis, showed an increase in ESS scores from baseline to 3 years [48]. EDS was related to a non-tremor-dominant phenotype, autonomic dysfunction, depression, anxiety, probable RBD, and dopaminergic therapy in a dose-dependent manner.

In a review article analyzing the results of 18 longitudinal studies on nonmotor symptoms, the risk factors for EDS were male sex, dopamine agonist use, insomnia, impairments in activities of daily living, cognitive impairment, and depression [14]. Therefore, disease progression, the use of dopaminergic treatment, susceptibility to sleepiness, and cognitive and psychiatric symptoms may all interact synergistically and contribute to the emergence of EDS.

Sleep episodes (or sleep attacks)

Sudden-onset sleep episodes (or sleep attacks) are potentially life-threatening problems that occur abruptly during daily activities. In the first report by Frucht et al. [49] in 1999, car accidents caused by sleep episodes in eight patients with PD who were taking nonergot dopamine agonists, such as pramipexole and ropinirole, were described. Subsequent studies reported that regardless of the type of specific dopamine agonist, the use of multiple dopamine agonists and higher doses of dopaminergic drugs were associated with an increased risk of sleep episodes [50]. In a multicenter study, the prevalence of sleep episodes was higher in PD patients than in controls (6.4%vs. 0.7%), and sleep episodes were more likely to occur in patients with EDS than in those without EDS (22.5%vs. 2.0%) [51]. The factors associated with sleep episodes were UPDRS part 1 scores (mental status), high ESS scores, and the use of high doses of dopaminergic drugs; an ESS score≥10 was predictive of sleep episodes, with a sensitivity of 75%and specificity of 82.4%. Similarly, Tan et al. [52] reported that sudden-onset sleep episodes were present in 13.4%of 201 patients with PD, and an ESS score≥10 was useful for predicting sleep episodes (sensitivity, 71.4%; specificity, 88.4%). In a study performed by the Canadian Movement Disorders Group including 638 PD patients [53], 3.8%of 420 drivers with PD experienced sudden-onset sleep episodes while driving, and 0.7%had no warning signs. An ESS score > 7 had a sensitivity of 75%and specificity of 50%for predicting sudden sleep episodes related to motor vehicle accidents. In contrast, a systematic review of sleep attacks and dopaminergic medications showed that 13%of PD patients on dopaminergic medications experienced sleep attacks. However, there were no significant differences in the ESS scores between patients with and without sleep attacks [54]. Thus, it should be noted that patients with EDS tend to have sudden-onset sleep episodes, but subjective sleepiness cannot always predict sudden-onset sleep episodes.

Disease-dopaminergic drug interactions and EDS

In the advanced stages of PD, the dosage of dopaminergic agents increases with the severity of the disease, and therefore, EDS and sleep episodes should always be monitored. EDS or sleep episodes can occur not only when taking higher dopamine agonist dosages but also when switching between the same type of dopamine agonists at equivalent dosages [55]. In a study including 54 levodopa-treated PD patients complaining of EDS, the mean sleep latency (MSL) was 6.3±0.6 min, 50%had an MSL less than 5 min, and 39%had a narcolepsy phenotype (2 or more sleep-onset REM periods during the 5 sessions of the MSLT) [9]. There was no correlation between the MSL and the daily dose of dopamine agonists or the total levodopa equivalent daily dose (LEDD); however, the levodopa daily dose was positively correlated with the MSL. In clinical studies, all classes of dopaminergic medications have been associated with EDS [17, 56].

In animal studies, D1 receptor agonists induce an increase in wakefulness and reductions in SWS and REM sleep, and D3 receptor agonists induce increases in SWS and REM sleep [57]. D2 receptor agonists induce biphasic effects: lower doses reduce wakefulness and increase SWS and REM sleep predominantly through the activation of D2 autoreceptors, whereas higher doses induce the opposite effect predominantly through the facilitation of D2 postsynaptic receptors. However, some PD patients in clinical settings have experienced increases in daytime sleepiness with the administration of higher dosages of dopaminergic treatment, and the effect was dose-dependent [58, 59]. Interestingly, a study showed that in patients with PD, higher doses of dopamine agonists were associated with reduced daytime alertness, whereas higher doses of levodopa were associated with enhanced alertness [60].

A study of 46 levodopa-treated patients with advanced PD found abnormal objective sleepiness (MSL < 5 min) in 41.3%of the patients on MSLT. Although 70%of those with ESS scores≥10 (i.e., with subjective daytime sleepiness) had an MSL < 10 min, subjective sleepiness did not correlate with the MSL and tended to be underestimated. The intensity of sleepiness (Karolinska Sleepiness Scale) just before the MSLT did not predict sleep latency [10]. Liguori et al. [61] compared daytime sleepiness between PD patients treated with the nocturnal administration of rotigotine and those receiving a placebo for 6 to 10 weeks. Rotigotine improved sleep efficiency, shortened sleep latency and reduced wake time after sleep onset. However, daytime sleepiness did not change, regardless of whether it was measured subjectively (ESS) or objectively (MSLT). The results implied that daytime sleepiness may be caused by an impairment of wakefulness related to PD pathology that is unrelated to nocturnal symptoms or sleep status.

Taken together, it can be speculated that EDS increases along with progressive neurodegeneration in the brainstem and hypothalamus pathology that regulates sleep/wake states and is complexly related to increased dosages of dopaminergic medication over the disease course. Furthermore, it seems that some subsets of PD patients have a narcoleptic phenotype.

Treatment for EDS and sleep episodes

Patients with EDS or sleep episodes (sleep attacks) should be instructed not to drive or engage in potentially hazardous tasks until their symptoms improve. Motor and nonmotor symptoms during the nighttime and comorbid sleep disorders such as RBD, RLS and sleep apnea should be identified, as sleep disturbance can cause EDS. In patients taking dopamine agonists, a dose reduction should be considered. Especially in patients whose EDS is temporally associated with the addition of or change in dopamine agonists, adjustments should be made, such as reducing the dose or discontinuing the possible causative agent. In a 6-week placebo-controlled, randomized trial of PD patients complaining of daytime somnolence (ESS≥10), drowsiness on the clinical global impression scale was significantly improved in the caffeine-treated group compared to the placebo group, although the ESS scores did not significantly change [62]. Some reports have shown the efficacy of modafinil for EDS [63]. In a 3-month open-label study, the adenosine A2A receptor antagonist istradefylline improved motor symptoms and daytime sleepiness without worsening nighttime sleep disturbances [64]. In a 12-week open-label study, the addition of the MAO-B inhibitor selegiline, for which a major metabolite is amphetamine, allowed a reduction in the dose of dopamine agonist, resulting in improvement of EDS [65]. In a double-blind, placebo-controlled study of 30 PD patients with sleep disorders treated with 1 mg of rasagiline, an MAO-B inhibitor, or a placebo for 8 weeks, the rasagiline group showed a decrease in wake after sleep onset and stage N1 duration on polysomnography (PSG) and improvements in daytime sleepiness [66]. Additionally, in a small-sample study, another MAO-B inhibitor, safinamide, improved subjective daytime sleepiness [67]. Thus, although sufficient evidence is lacking, the addition of nondopaminergic drugs and decreased doses of dopamine agonists could be a potential treatment for EDS in patients with PD.

Lesions responsible for RBD

There are several neuronal populations in the brainstem that regulate REM sleep. During REM sleep, ventrolateral periaqueductal gray neurons, the locus coeruleus, the raphe nucleus and the tuberomammillary nucleus are inactive, whereas the sublaterodorsal nucleus (locus subcoeruleus) and cholinergic pedunculopontine systems are active [71]. The sublaterodorsal nuclei have a role in initiating REM sleep and transitioning from NREM sleep, and for normal REM atonia. Glutamatergic projections from the sublaterodorsal nuclei activate gamma-aminobutyric acid/glycinergic ventral gigantocellular nuclei in the medulla, resulting in suppression of spinal motor neurons [72, 73]. In RBD, degeneration of the sublaterodorsal nuclei results in an inability to activate ventral gigantocellular nuclei, allowing the activation of spinal motor neurons, which results in abnormal motor behavior during REM sleep. Studies using neuromelanin-sensitive MRI revealed a decrease in intensity in a region of the coeruleus/subcoeruleus complex in iRBD patients [74] and in PD patients with RBD [75]. In a study using ¹⁸F-fluoroethoxybenzovesamicol PET imaging, cholinergic alterations, i.e., compensatory cholinergic upregulation, were found in multiple brain areas, such as brainstem structures involved in REM sleep regulation and muscle atonia in iRBD patients [76].

RBD and neurodegeneration

In a study reported by McCarter et al. [77], RWA analysis of the submentalis muscles (SM) was positive in 58.2%of 53 patients with synucleinopathies (33 PD and 20 multiple system atrophy (MSA) patients) and in 3.2%of 24 patients with tauopathies (17 with progressive supranuclear palsy and 7 with corticobasal degeneration). This study reported a sensitivity of 70%and a specificity of 100%for phasic and tonic muscle activities in SM (“SM any”) and a sensitivity of 77%and a specificity of 100%for SM phasic activity for distinguishing synucleinopathies from tauopathies, but RWA analysis of the anterior tibialis muscle (AT) was not useful for differential diagnosis.

Different criteria have been proposed and discussed regarding the definition of RWA on PSG and the appropriate cutoff for RWA for the diagnosis of RBD. In a study performed by a Canadian group, a rate of tonic muscle activity≥30%or excessive phasic muscle activity≥15%during the entire REM sleep period based on submental electromyography was useful for distinguishing iRBD patients from healthy individuals [78]. An Austrian group reported that a criterion of≥27%combined tonic and excessive phasic muscle activity in the submental muscles and excessive phasic muscle activity in the forearm (flexor digitorum superficialis) by 30-second epoch judgments was useful for the diagnosis of RBD [79]. In a study including 60 patients with PSG-confirmed iRBD followed for at least 3 years, 43%developed parkinsonism or mild cognitive impairment. Kaplan-Meier analysis showed that the phenoconversion risk was 20%and 35%at 3 and 5 years, respectively. A cutoff of RWA in SM+AT any > 46.4%was associated with 25%and 45%phenoconversion risks at 3 and 5 years, and when applying this RWA cutoff in patients > 65 years of age at diagnosis, the phenoconversion risks increased to 30%and 55%at 3 and 5 years, respectively [80].

RBD has received considerable attention as a precursor to synucleinopathies such as PD, dementia with Lewy bodies (DLB) and MSA [81]. Schenck et al. [82] reported that 38%of 29 male patients with iRBD aged over 50 years developed Parkinsonian disorders at a mean follow-up of 3.7±1.4 years, and after 16 years of follow-up, 81%of the patients initially diagnosed with iRBD eventually developed neurodegenerative diseases such as PD and dementia [83]. Other studies from Canada and Spain have confirmed the risk of developing neurodegenerative disease in iRBD patients using the Kaplan-Meier method [84, 85]. A recent systematic review and meta-analysis showed that the risk of developing neurodegenerative diseases was 33.5%at 5 years, 82.4%at 10.5 years and 96.6%at 14 years of follow-up, with an average conversion rate of 32.0%after a mean duration of 4.8±2.4 years of follow-up [86]. Among the patients who converted to neurodegenerative diseases, the most common disease was PD (44%), followed by DLB (25%), unspecified dementia (7%), MSA (5%), Alzheimer’s disease (3%), mild cognitive impairment (3%) and others (0.5%). A clinicopathological study showed that among 170 patients with neurodegenerative disorders associated with RBD, 160 (94%) had synucleinopathies, and 83%had Lewy body disease (PD, PD with dementia, or DLB) [87]. Although RBD frequently co-occurs in patients with MSA (73–88%) when assessed in a cross-sectional manner [88], prospective studies have consistently shown that patients with RBD more frequently convert to PD or DLB than MSA among synucleinopathies [86, 90]. Several clinical markers indicative of Lewy body diseases, such as cognitive impairment, olfactory impairment, visual impairment and reduced meta-iodobenzylguanidine (MIBG) myocardial scintigraphy accumulation, have been observed in iRBD patients [68, 91]. The frequency of olfactory impairment in iRBD patients who eventually developed MSA (50%) [92] seems to be lower than that in those who converted to PD or DLB (80%) [90].

In a study comparing annual changes in the precursor markers of neurodegeneration, including motor function, color vision, olfaction, cognition, autonomic dysfunction and glucocerebrosidase (GBA) mutation status, between patients with iRBD who had not progressed to neurodegenerative disease for more than 10 years (longstanding iRBD) and those with RBD who had progressed to neurodegenerative disease within 4 years (early converters), the rate of increase in the probability of prodromal PD (%), as calculated by the MDS research criteria for prodromal PD [93], was significantly greater in the early converter group than in the longstanding iRBD group. However, there were no differences between the two groups in terms of the baseline precursor markers, with the exception of olfactory impairments [94]. This suggests that neurodegeneration may insidiously progress at different rates in patients with iRBD.

Several studies have shown a higher rate of alpha-synuclein pathology outside the central nervous systems in patients with iRBD than in healthy controls. Iranzo et al. [95] reported that 50%of patients with iRBD, 54%of those with PD, 50%of those with DLB, and 3%of the controls had phosphorylated α-synuclein deposits in the labial salivary glands. Other studies of the submandibular gland [96] and skin biopsies [97] have shown that alpha-synuclein pathology was significantly more common in patients with iRBD than in healthy subjects. Of note, in a study of colon biopsies, immunostaining of tissue with serine 129-phosphorylated α-synuclein antibody was more specific than that with 15G7 α-synuclein antibody for differentiating iRBD patients from healthy subjects [98]. These biopsy studies provide pathological evidence that iRBD represents synucleinopathy and that a less invasive and more specific method of detecting synuclein pathology may be useful in future neuroprotective trials in patients with iRBD.

RBD with PD

When assessed in a cross-sectional manner, RBD is found in up to 50%of PD patients [68, 91]. Compared with those without RBD, PD patients with RBD have more rapid progression [99], a higher risk of dementia [100] and more severe synuclein pathology, including in the brainstem region [101]. Studies using a data-driven clustering approach identified 3 PD clinical subtypes: mild motor predominant, 49–53%; intermediate, 35–39%; and diffuse malignant, 9–16%. The mild motor predominant type had slow disease progression with a good response to dopaminergic treatment and was related to onset at a young age and mild motor symptoms, while the diffuse malignant type showed rapid disease progression and poor response to dopaminergic treatment and was associated with RBD, mild cognitive impairment, orthostatic impairment and severe motor symptoms [102, 103]. During 5 years of follow-up, the presence of RBD at baseline in early PD patients was associated with an increased risk of functional dependency as measured by the Schwab and England score [104]. In a meta-analysis addressing clinical differences between PD patients with and without RBD, PD patients with PSG-confirmed RBD were older, more likely to be male, and had a longer disease duration, higher HY scale scores (disease severity), and higher UPDRS III scores (motor signs) [105].

In a two-year prospective study of 123 early PD patients and 106 age-matched controls that monitored 30 candidate markers of disease progression, 10 progression markers in PD were identified, which included RBD symptoms and daytime sleepiness [106]. Figorilli et al. [107] prospectively studied 22 PD patients with PSG-confirmed RBD, and while the changes in RBD symptoms varied widely (27%, improved; 27%, worsened; 45%, unchanged), the rate of RWA increased in all patients. Changes in RWA were correlated with the degree of dyskinesia, motor fluctuations, and worsening of executive and visuospatial-cognitive functions. While RBD symptoms may fluctuate depending on the patient, RWA could be a marker of disease progression.

A close relationship between alpha-synuclein pathology and RBD has been recently shown using a novel prodromal PD mouse model. Taguchi et al. [108] successfully produced RWA, an essential RBD feature, and hyposmia using alpha-synuclein (SNCA) bacterial artificial chromosome transgenic mice, which showed no parkinsonism with only a 17.1%decrease in dopaminergic neurons at 18 months of age compared with wild-type mice.

Prodromal state of RBD

Nonviolent behaviors are observed in 18%of RBD patients [109], including jerky, simple, and minor movements of the limbs that are common in RBD patients [110]. It is possible that nonviolent mild motor behaviors may occur in the early phase of RBD or precede full-blown RBD. As a prodromal state of RBD, REM sleep behavioral events (RBE), involving seemingly purposeful motor behaviors and vocalizations with increased tonic and phasic electromyographic (EMG) activity during REM sleep that do not fulfill the RBD criteria, have been proposed [68, 112]. RBE was found in half of de novo PD patients [111], and in early PD patients, 38%of patients with RBE developed RBD within two years [112].

Whether isolated RWA (iRWA), a condition in which no motor behaviors or vocalization are shown other than increased EMG tone during REM sleep, could be a very early state of RBD and precede RBE has attracted attention. In a study with 14 subjects with iRWA, 1 (7.3%) developed RBD after 8.6±0.9 years of follow-up. The rate of RWA significantly increased from 32.5±9.4%to 52.2±16.6%over time. Neurodegenerative markers, including mild cognitive impairment and impaired olfaction, were found in 72%of the patients [113]. In a prospective study including 31 iRBD and 67 iRWA subjects, 18 subjects with iRWA (26.8%) developed iRBD within 2.6±2.2 years, and 6 patients with iRWA (8.9%) developed neurodegenerative disorders (4 developed PD and 2 developed Alzheimer’s disease) within 2.4±1.5 years. In contrast, 8 (25.4%) patients with iRBD developed PD within 2.6±2.2 years. Phenoconversion was positively correlated with age and periodic limb movements in sleep in both the iRBD and iRWA groups [114]. However, a study comparing 49 iRWA subjects and 41 healthy controls found that compared with the healthy controls, iRWA subjects had higher REM density, which is reported to be lower in PD patients, and normal heart rate variability, which is known to be decreased in PD; these findings may weaken the hypothesis that iRWA is a prodromal state of PD [115].

Genetic similarities between RBD and PD

GBA mutations, which cause Gaucher disease, were more common in PD patients than in controls (7%vs. 1%), and GBA mutations were associated with a 5.43-fold higher risk of PD [116]. In addition, GBA mutations were more commonly found in idiopathic RBD patients than in healthy subjects (2.6%vs. 0.4%) [117], suggesting an association between RBD and GBA mutations. In other studies, PD patients with GBA mutations had more severe motor symptoms and higher rates of RBD, hallucinations and other nonmotor symptoms, such as depression and olfactory impairments, than PD patients without GBA mutations [118]. Thus, GBA mutations may affect the clinical features of PD.

A study showed a significant association between Lewy-type alpha-synucleinopathy (LTS) pathology and the prevalence of probable premortem RBD in various neurodegenerative diseases, including progressive supranuclear palsy (with LTS, 25.0%; without LTS, 6.1%) [119]. The association of RBD with MAPT is unclear [120]. Interestingly, carriers of different haplotypes of LRRK2 mutations may have different risks of RBD: the p.N551K-p.R1398H-p.K1423K haplotype was associated with a lower risk of RBD, and the common variant p. S1647T was associated with an increased risk of RBD [121]. Full sequencing and genotyping of SNCA in patients with iRBD, PD, and DLB and controls showed that a 5’-region SNCA variant (rs10005233) was associated with an increased risk of iRBD and probable RBD with PD [122]. Zhao et al. [123] recently found SNCA hypomethylation in leukocytes in patients with iRBD and PD.

Treatment of RBD

First, patients are asked to adjust their bedroom environment, including removing objects from the bedroom, such as stands, pens and furniture, that could cause trauma if abnormal behavior or movements occur, and the patient is instructed to sleep on a mattress on the floor to prevent him/her from falling out of bed. The patient’s bed partner should sleep in a separate room until the patient’s abnormal behavior is improved.

RBD can be induced by antidepressants such as tricyclic antidepressants and serotonin reuptake inhibitors; their use should be reviewed and, if possible, reduced or discontinued. A small dose of clonazepam (0.5 mg to 1 mg) or melatonin (3–12 mg) administered before bedtime is effective as a treatment for RBD [124]. A multicenter open-label study showed that the use of the melatonin receptor agonist ramelteon at 8 mg/day for 12 weeks improved RBD symptoms in patients with PD [125]. Although clonazepam is highly effective at reducing RBD symptoms, it should be noted that it may cause falls, next-day residual sleepiness and worsening of apnea in patients with comorbid sleep apnea syndrome. In addition, dopamine agonists (pramipexole and rotigotine [126]) and yi-gan san have been reported to be efficacious [124]. In a double-blind, crossover trial including 12 PD patients with PSG-confirmed RBD in whom RBD symptoms were refractory to melatonin and clonazepam, rivastigmine treatment (4.6 mg/24 hours) for 3 weeks significantly reduced RBD episodes compared with the placebo [127]. Another placebo-controlled, crossover pilot trial of rivastigmine in 25 consecutive patients with mild cognitive impairment who showed RBD symptoms refractory to first-line treatments such as melatonin and clonazepam was reported. The results showed that rivastigmine treatment resulted in a significant reduction in RBD episodes compared with the placebo [128]. Thus, cholinesterase inhibitors may be an alternative treatment for patients with refractory RBD.

Pathophysiology of RLS

The pathogenesis of RLS is mainly brain iron dysfunction, in addition to dysfunction in various nervous system and neurotransmitter systems such as dopamine, glutamate, orexin and adenosine [137, 138], and the involvement of peripheral hypoxia [139] has been suggested. Serum levels of hepcidin, a hormone involved in iron regulation, were reported to be higher in an RLS group than in a healthy control group [140]. The effects of dopamine agonists suggest dopamine impairment in the central nervous system. However, brain imaging studies of the nigrostriatal system have not provided consistent results [136].

In an autopsy study of patients with idiopathic RLS, tyrosine hydroxylase (TH) staining did not reveal impairments in dopaminergic areas, including the hypothalamus, but iron and H-ferritin staining in the substantia nigra were lower than in controls [141]. However, other autopsy studies showed that the dopaminergic markers TH and phosphorylated TH were increased in the substantia nigra and putamen in RLS patients [142]. Furthermore, D2 receptor expression was negatively correlated with premortem RLS symptoms, and severe RLS symptoms were related to lower D2 receptor expression. These results are consistent with the iron deficiency model of RLS, and a decrease in D2/D3 receptors in the mesolimbic system and an increase in CSF levels of homovanillic acid, a dopamine metabolite, and 3-O-methyl dopa, a metabolite of levodopa [138], suggest activation of the dopaminergic system in RLS [137]. The possible explanation for the nighttime occurrence of RLS in a hyperdopaminergic state may be that it is a reflection of the circadian profile of dopamine, including decreased dopaminergic activity and the downregulation of dopamine receptors at night, which results in the occurrence of RLS symptoms predominantly at night [138].

RLS with PD

In a recent meta-analysis, the prevalence of RLS was 2.86 times higher in the PD group than in the healthy control group, with a prevalence of 15%in treated PD patients and 11%in drug-naïve PD patients [143]. Patients with comorbid RLS and PD have been reported to be less likely to have a family history of RLS than patients with idiopathic RLS, and they have an older age at onset, have mild RLS symptoms, and more often develop RLS after the onset of PD [144, 145]. In a cohort of 113 patients with PD, 24%had RLS, 61%of whom had RLS symptoms associated with wearing off phenomena, suggesting the importance of monitoring off-related symptoms [146]. An increase in the prevalence of RLS over time has been reported in a 4-year prospective study including untreated patients with PD at baseline [147].

Leg motor restlessness (LMR), which is characterized by the main symptom of RLS, “an urge to move”, but does not meet the diagnostic criteria for RLS, was reported to be three times more likely to occur in untreated patients with PD than in healthy subjects [148]. RLS variants and LMR are potential targets for treatment because they, similar to RLS, can cause insomnia. We investigated RLS, LMR and their variants in patients with PD-related disorders [149]. The results showed that in 63 PD patients, the proportions of patients with comorbid RLS, LMR, RLS variants and LMR variants were 12.7%, 11.1%, 0%, and 1.6%, respectively, with these RLS-related symptoms being associated with autonomic symptoms, sleep disturbances and depressive symptoms. Having any RLS-related symptoms was more prevalent in patients with PD (50.8%) or MSA (29.4%) than in patients with progressive supranuclear palsy (9.1%), with a similar trend observed in a subanalysis of the untreated group, suggesting that screening for RLS-related symptoms could be useful for differentiating PD or MSA from progressive supranuclear palsy.

The pathology of concomitant PD and RLS is distinct from that of idiopathic RLS. Patients with RLS with PD show degeneration of the nigrostriatal dopaminergic neuron and increased iron in the substantia nigra, in contrast to those with idiopathic RLS [130, 150]. In addition, it is important to note that dopaminergic treatment in PD may lead to an underestimation of RLS (dopamine agonists reduce RLS) or an overestimation of RLS (long-term treatment may cause augmentation and unmask subclinical RLS). In a study of 80 PD patients, 30 PD patients with RLS were older and had a longer duration of PD, higher nonmotor symptom scores (NMSQ), and a higher rate of impulse control disorders (50%vs. 26%) than those without RLS, and RLS was an independent risk factor for impulse control disorders (odds ratio 5.9) after adjusting for confounding factors [151]. Impairments in posterior hypothalamic A11 dopaminergic neurons in RLS have been suggested from animal studies [152]. The nucleus accumbens, which participates in the regulation of reward and motivation, receives dense projections from the paraventricular nucleus and dopaminergic fibers in the paraventricular nucleus originating from a group of hypothalamic dopamine cells, including A11 [153]. Therefore, impaired dopaminergic neurons in the hypothalamic A11 cell population in PD patients with RLS may show impaired projections from the paraventricular nucleus to the nucleus accumbens, which may contribute to the development of impulse control disorders. In a study of 143 PD patients, the association between sleep disturbance and impulsivity was investigated, and PD patients with impulsivity had more complaints of sleep disturbance and daytime sleepiness and a higher prevalence of RLS than PD patients without impulsivity [154]. In a hierarchical linear model, decreased sleep efficiency was most strongly associated with impulsivity. Therefore, the possibility of concomitant impulsivity in PD patients with RLS may deserve attention in clinical practice when adding dopamine agonists.

Treatment of RLS

Blood tests should be performed to screen for iron deficiency anemia, renal dysfunction and low serum ferritin. In patients with low serum ferritin (<50–75μg/L), iron supplementation is recommended. However, in patients with PD, whether iron should be added is debatable, as PD is related to oxidative damage via iron accumulation in the substantia nigra [155]. Alcohol or caffeine consumption, stress and intense exercise before bedtime can trigger RLS symptoms. Antidepressants, dopamine antagonists or antihistamines may induce RLS, and their use should be reduced or discontinued. As a nonpharmacological treatment, the efficacy of exercise with regard to the reduction in RLS symptoms has been reported. A randomized controlled trial showed that an exercise group engaging in aerobic and lower-body resistance training 3 days per week had significant improvements in RLS symptoms compared with a placebo group [156]. Additionally, exercise is considered likely to be efficacious for the treatment of RLS symptoms in uremic patients [157].

Long-acting dopamine agonists, such as rotigotine, should be added before bedtime for PD patients with RLS, LMR, or RLS symptoms related to the wear-off phenomenon. An alpha-2 δ ligand (gabapentin enacarbil at 300–600 mg) or clonazepam (0.5 mg) should also be considered in cases in which sufficient dopamine agonists are already being administered. In severe RLS, oxycodone–naloxone has been reported to be efficacious when used at low doses twice a day [158]. STN DBS may improve the symptoms of RLS, although the effect of concomitant changes in the dose of dopaminergic treatment should be considered [159].

Treatment of obstructive sleep apnea

Oral appliances are used in mild cases, and continuous positive airway pressure (CPAP) is used in moderate or severe cases. Weight loss is recommended for obese patients. The significance of CPAP therapy for OSA has been highlighted by several studies: one study described associations among the narcoleptic phenotype, SOREMP on MSLT, and the presence of sleep-disordered breathing in 46 patients with advanced PD [10], and another prospective study that included 38 PD patients with a mean disease duration of 5.3 years showed that CPAP treatment improved OSA and daytime sleepiness measured by MSLT (baseline, MSL 8.7 min; 3 weeks, 11.3 min) [170]. Interestingly, positive effects of long-acting levodopa at bedtime on OSA have been reported [171]. Additionally, CPAP treatment of OSA in patients with PD with a mean usage of 3 hours 36 min was associated with improved sleep quality, anxiety, and cognitive function over a 12-month period [172]. In a study from a tertiary sleep center, 89.2%of PSG-confirmed RBD patients had concomitant OSA, and CPAP therapy improved RBD symptoms in 45.8%of those using CPAP [173]. Thus, in patients with RBD with OSA, CPAP should be tried first. However, long-term adherence to CPAP is a significant problem, considering the progressive motor and nonmotor symptoms in PD, particularly in the later stages. Cognitive impairment, motor disability, nocturia and RBD were reported to be factors influencing CPAP adherence [174].

Treatment of circadian abnormalities

Exposure to one hour of bright light (5,000 lux) once a day during the evening hours for 3 months improved sleep problems assessed by PDSS-2 scores, with restoration of circadian phases based on the assessment of the expression levels of three clock genes (Per3, Nr1d1, and Nr1d2) in 16 patients with PD [183]. In a randomized trial including 31 treated PD patients with EDS (ESS≥12) who received 1 hour of bright light therapy (LT) (10,000 lux) or dim-red LT (control condition, <300 lux) in the morning and in the afternoon (twice daily) for 2 weeks, bright LT improved EDS, sleep quality, and sleep fragmentation and increased daily physical activity in patients with PD [184].