Progressive dopamine and hypocretin deficiencies in Parkinson’s disease: is there an impact on sleep and wakefulness?


Christian Baumann MD, Department of Neurology, University Hospital Zurich, Frauenklinikstrasse 26, 8091 Zurich, Switzerland. Tel.: 41-44-255-5503; fax: 41-44-255-4380; e-mail:


Sleep–wake disturbances are frequent in patients with Parkinson’s disease, but prospective controlled electrophysiological studies of sleep in those patients are surprisingly sparse, and the pathophysiology of sleep–wake disturbances in Parkinson’s disease remains largely elusive. In particular, the impact of impaired dopaminergic and hypocretin (orexin) signalling on sleep and wakefulness in Parkinson’s disease is still unknown. We performed a prospective, controlled electrophysiological study in patients with early and advanced Parkinson’s disease, e.g. in subjects with presumably different levels of dopamine and hypocretin cell loss. We compared sleep laboratory tests and cerebrospinal fluid levels with hypocretin-deficient patients with narcolepsy with cataplexy, and with matched controls. Nocturnal sleep efficiency was most decreased in advanced Parkinson patients, and still lower in early Parkinson patients than in narcolepsy subjects. Excessive daytime sleepiness was most severe in narcolepsy patients. In Parkinson patients, objective sleepiness correlated with decrease of cerebrospinal fluid hypocretin levels, and repeated hypocretin measurements in two Parkinson patients revealed a decrease of levels over years. This suggests that dopamine and hypocretin deficiency differentially affect sleep and wakefulness in Parkinson’s disease. Poorer sleep quality is linked to dopamine deficiency and other disease-related factors. Despite hypocretin cell loss in Parkinson’s disease being only partial, disturbed hypocretin signalling is likely to contribute to excessive daytime sleepiness in Parkinson patients.


Patients with Parkinson’s disease (PD) often suffer from a variety of sleep–wake disturbances (SWD), but prospective controlled electrophysiological studies on SWD in defined stages of PD are surprisingly sparse, and the pathophysiology of SWD in PD remains largely elusive. Young et al. (2002) examined 11 patients with early PD and seven patients with advanced PD, all with excessive daytime sleepiness (EDS). Compared to controls, they found fragmented nocturnal sleep in both groups, but sleep disruption was not associated with severity of PD. One paper on SWD in patients with untreated early PD found a higher prevalence of rapid eye movement (REM) sleep without atonia than in controls (Buskova et al., 2011). Another controlled study focused on obstructive sleep apnoea in 55 PD patients and observed that sleep-related breathing disorders are not more prevalent or more severe than in controls (Trotti and Bliwise, 2010). Conversely, an earlier case–control study in 49 patients with PD in the early or middle stage found that non-obese PD patients often had polysomnographic findings of sleep apnoea (Diederich et al., 2005). Further studies also gave important insights into sleep and wakefulness in PD but were questionnaire-based only, non-controlled or retrospective, or included mixed stages of PD (Arnulf et al., 2002; Cochen De Cock et al., 2010; Comella et al., 1993; De Cock et al., 2007; Gagnon et al., 2002; Happe et al., 2005; Monaca et al., 2006; Razmy et al., 2004; Rye et al., 2000; Sixel-Döring et al., 2011; Stevens et al., 2004; Valko et al., 2010).

Multiple factors, including motor symptoms, nocturia, psychiatric comorbidities and medication, may cause or aggravate SWD in PD, but their pathophysiology is still not well understood. In particular, it is not known how neurodegeneration of differential neurotransmitter systems impacts SWD in PD. Together with altered dopaminergic signalling, partial loss of wake-promoting hypocretin cells in the hypothalamus of PD patients might contribute to nocturnal SWD and particularly EDS (Baumann et al., 2005; Fronczek et al., 2007; Thannickal et al., 2007). Knowledge of the in-vivo influence of dopaminergic and hypocretin signalling on EDS and nocturnal sleep in PD, however, is sparse.

In order to understand more clearly the influences of impaired dopaminergic and hypocretin signalling on sleep and wakefulness in PD, we recorded prospectively measures of sleep and wakefulness in patients with defined stages of PD and compared the results not only to matched healthy controls, but also to hypocretin-deficient narcolepsy patients with cataplexy (NC). PD and NC share multiple disturbances of both sleep and wakefulness (Arnulf et al., 2000). In NC patients, a marked loss of arousal-promoting hypocretin cells cause EDS and fragmented nocturnal sleep (Baumann and Bassetti, 2005). Also, REM sleep behaviour disorder (RBD) can be observed in NC, but the dopaminergic system in the substantia nigra is spared.

Based on previous observations, we hypothesized that nocturnal sleep is fragmented in both PD and NC, but more severely in dopamine-deficient PD, particularly in advanced PD (aPD). With regard to EDS, we hypothesized that this symptom was present in both disorders, but most pronounced in hypocretin-deficient NC, and least in early PD (ePD).

Patients and Methods

In this prospective study, we examined 10 consecutive patients with advanced PD (aPD: 10 male, mean age 64 years, range 54–80) and 10 early and treatment-naive PD patients (ePD: five male, mean age 59 years, range 48–73). We compared all results with prospective data of 10 healthy age-matched controls (Co: five male, mean age 63 years, range 46–75). In addition, we compared sleep measures to retrospective data of 10 untreated patients with NC who have been published before (NC: three male, mean age 39 years, range 24–57) (Dauvilliers et al., 2003). Stimulant or sodium oxybate treatment was discontinued 7 days before all examinations. Cerebrospinal fluid (CSF) hypocretin levels were decreased pathologically in all NC patients.

All PD patients were recruited in the movement disorders outpatient unit of the Department of Neurology at the University Hospital Zurich. PD was diagnosed according to international criteria (Gelb et al., 1999). For ePD patients, a positive levodopa challenge test to prove responsiveness to dopaminergic agents was mandatory for inclusion.

Inclusion criteria for ePD patients included disease duration ≤6 years and UPDRS III values ≤28. For aPD patients, disease duration ≥9 years and UPDRS III ≥35 were required. To rule out the influence of dementia on sleep–wake parameters and to ensure optimal compliance with all study procedures, a score on the Mini Mental State Examination (MMSE) of ≥27 was mandatory for inclusion of subjects. ePD patients and narcolepsy patients were treatment-naive for all study examinations.

Exclusion criteria included a previously diagnosed nocturnal sleep–wake disorder (except for RBD in PD patients, and for narcolepsy in the NC group), and the diagnosis of any neurological disease other than PD or NC. Two days before and during all study procedures, the use of hypnotic medications was not allowed. All patients gave written informed consent to participate in the study, which was approved by the local ethical committee.

Clinical assessments

We obtained detailed sleep histories by means of standard sleep questionnaires, including the Epworth Sleepiness Scale, and specific questions regarding sleep paralysis, hallucinations and RBD. We also assessed detailed histories of current medication. Complete neurological assessment, including MMSE, was performed in all patients. Motor impairment was assessed off medication with the Unified Parkinson’s Disease Rating Scale, part III (UPDRS III), and PD severity was classified according to the Hoehn and Yahr scale (Fahn et al., 1987; Hoehn and Yahr, 2001). We calculated levodopa equivalent doses (LED) according to international standards (Tomlinson et al., 2010). Symptoms of depression were assessed with the Beck Depression Inventory.

Sleep laboratory examinations

To rule out sleep deprivation prior to sleep laboratory recordings, we recorded sleep and physical activity levels over 14 days with sleep logs, and verified them with wrist actigraphy which was performed at the same time. Actigraphy recordings were performed as reported previously (Baumann et al., 2007).

Conventional overnight video polysomnography (PSG) was performed from 23:00 h to 07:00 h, as described previously (Baumann et al., 2007). We scored sleep stages visually in 30-s epochs according to standard criteria (Iber et al., 2007). Polysomnographic scoring of aPD patients was performed along the suggestions of Bliwise et al. (2000). A polysomnographic diagnosis of RBD was made along previous recommendations (Montplaisir et al., 2010). In addition, we assessed the periodic limb movement arousal index (PLMAI), which was defined as the total number of periodic leg movements per hour of sleep in which EEG arousal occurred within 3 s of movement termination (Claman et al., 2006).

The Multiple Sleep Latency Test (MSLT) included four sleep opportunities every 2 h, and was started 2 h after completion of PSG. Mean sleep latencies below 8 min were considered to reflect objective daytime sleepiness. In addition, we assessed the presence of sleep onset REM periods.

Laboratory tests

In 16 of 20 PD patients, we assessed CSF hypocretin-1 levels as described previously (Baumann et al., 2005). In two patients, we measured CSF hypocretin-1 levels twice during the course of their disease, and these measurements were performed in the same assay. Based on our control group data, we consider hypocretin levels <320 pg mL−1 as abnormally low (Baumann et al., 2004). The hypocretin results of our retrospectively assessed NC patients have been reported previously (Baumann et al., 2006).


Differences within all groups were computed with one-way analysis of variance (anova), and in the case of non-parametric variables with the Kruskal–Wallis test. Multiple comparisons between groups were made with Fisher’s least significant difference (LSD). Non-parametric tests were used when appropriate, including the Mann–Whitney U-test. Otherwise, Student’s t-test was used for comparisons of means. Correlations of parametric data were performed with the Pearson test, and Spearman’s test was applied for non-parametric data. Stepwise regression was used for the backward elimination of non-significant candidate variables. All statistical analyses were made with spss version 19.


Study population, clinical and demographic variables

Thirty-one PD patients and 13 controls were approached for this project. Ten patients declined to participate. We excluded one PD patient because of a MMSE score of 24 points. Three healthy controls were excluded because polysomnography revealed sleep apnoea syndrome. Actigraphy revealed that all participants were sleep-satiated before all study procedures.

Table 1 summarizes demographic and clinical variables of all patients and controls. All aPD patients were treated with dopaminergic drugs (levodopa, = 10; cabergolide, = 1; ropinirole, = 4; pramipexole, = 4; rotigotine, = 1).

Table 1.   Detailed demography, clinical and sleep latoratory findings in 10 patients with advanced Parkisonson’s disease (aPD), 10 patients with early Parkinson’s disease (ePD), 10 patients with narcolepsy with cataplexy (NC) and 10 matched controls (Co)
 aPDePDNCCo P Post-hoc significance (LSD, Fisher post-hoc)
  1. Results are presented as means ± standard deviation. NA, not available; LSD, least significant difference; UPDRS III, unified Parkinson’s disease rating scale, part III; DA, dopamine agonist; LED, levodopa equivalent dose; BDI, Beck Depression Inventory (≥10 points indicating depression). Type of antidepressants—T, tetracyclic antidepressant/mirtazapine; S, serotonin reuptake inhibitor; PLMS, periodic limb movements during sleep; PLMAI, periodic limb movement arousal index; Δ WE-WD, difference in sleep times between weekends and weekdays; SOREMPs, sleep onset rapid eye movement sleep periods; NS, not significant; NREM, non-rapid eye movement. Assumption of disturbed neurotransmitter signalling based on published evidence. *P < 0.05; **< 0.01; ***< 0.001.

Dopamine cells↓↓↓NormalNormal  
Hypocretin cells↓↓↓↓↓Normal  
n 10101010  
Age (years)63.8 ± 7.658.7 ± 7.539.2 ± 11.063.4 ± 8.0<0.001A > C***, B > C***
Sex (male/female)10/05/53/75/5  
Disease duration (years)14.6 ± 5.43.3 ± 1.3  <0.001 
UPDRS III (on)28.7 ± 7.210.0 ± 2.4  <0.001 
UPDRS III (off)50.1 ± 10.220.4 ± 5.7  <0.001 
Hoehn and Yahr3.1 ± 0.71.0 ± 0.0  <0.001 
Levodopa (mg day−1)845 ± 4790 ± 0  <0.001 
DA dose (mg day−1)8.7 ± 9.40 ± 0  0.003 
LED (mg day−1)1189 ± 4200 ± 0  <0.001 
Dementia0/100/10  NS 
Antidepressants6/10 (4 T, 2 S)2/10 (1 T, 1 S)    
Total sleep time (min)344 ± 90358 ± 79360 ± 33396 ± 68NS 
NREM latency (min)26 ± 2533 ± 217 ± 523 ± 220.002A > C*, B > C***
REM latency (min)180 ± 110115 ± 5326 ± 31115 ± 590.001A > C**, B > C***
Arousal Index7.0 ± 5.17.0 ± 8.311.0 ± 6.413.5 ± 10.0NS 
Wake (%)32 ± 1424 ± 1412 ± 715 ± 100.004A > C**, B > C*, A > D*
NREM 1 (%)24 ± 199 ± 314 ± 49 ± 40.006A > B*, B < C**, A > D*
NREM 2 (%)32 ± 1545 ± 1341 ± 1048 ± 120.04A < B*, A < D**
NREM 3 (%)5 ± 47 ± 616 ± 812 ± 60.004A < C**, B < C**, A < D*
REM (%)7 ± 616 ± 817 ± 515 ± 50.05A < B*, A < C**, A < D**
Sleep efficiency (%)68 ± 1476 ± 1488 ± 784 ± 100.005A < C**, B < C*, A < D*
Apnoea–hypopnoea index16 ± 235 ± 53 ± 65 ± 4NS 
PLMS index11 ± 2312 ± 1817 ± 2726 ± 21NS 
PLMAI4 ± 66 ± 68 ± 98 ± 7NS 
Multiple Sleep Latency Test      
Mean sleep lat. (min)6.0 ± 3.411.5 ± 4.22.1 ± 1.512.1 ± 3.4<0.001A < B/D**, A > C**, B > C***
Mean sleep lat. <8 min701010010  
SOREMPs (mean)0.403.30.1<0.001A < C**, B < C***
Epworth Sleepiness Scale      
Total scores10 ± 57 ± 516 ± 57 ± 50.002A < C**, B < C**
Scores ≥10 (%)50109040  
Inactivity/24 h (%)46 ± 1145 ± 1235 ± 738 ± 70.05A > C*, B > C*
Δ WE-WD (h)0.4 ± 0.90.3 ± 1.11.1 ± 1.50.6 ± 0.8NS 
Cerebrospinal fluid hypocretin-1 levels      
n 7710   
Levels (pg mL−1)487 ± 57525 ± 6314 ± 21NA<0.001A > C***, B > C***

Nocturnal sleep

We hypothesized that nocturnal sleep is disturbed by frequent awakenings and RBD in both PD and NC, but more severely in dopamine-deficient PD, particularly in aPD.

During detailed interviews, one ePD and five aPD patients reported symptoms compatible with RBD, two aPD patients reported hypnagogic hallucinations, one of those with sleep paralysis, and eight aPD suffered from insomnia. None of the NC patients reported RBD-like symptoms, but six reported hypnagogic hallucinations and seven sleep paralysis; all patients reported fragmented nocturnal sleep.

An overview on polysomnography findings is given in Table 1. Compared to healthy controls, aPD patients had a lower sleep efficiency (= 0.02), less REM sleep (= 0.01) and less deep sleep (= 0.01). ePD patients and healthy controls had similar results in all polysomnographic parameters.

Comparing aPD and ePD, there were differences in respect to sleep architecture. ePD patients revealed normal sleep architecture with four to five sleep cycles, whereas this cyclic pattern was not preserved in seven of 10 aPD patients. aPD patients stayed longer in Stage 1 (= 0.02), but shorter in REM sleep (= 0.03).

Compared to NC patients, both aPD and ePD had longer sleep latencies to Stage 2 (= 0.015 for aPD, < 0.001 for ePD) and REM sleep (= 0.003 for aPD, < 0.001 for ePD), spent less time in deep sleep stage (= 0.003 for aPD, = 0.009 for ePD) and had lower sleep efficiencies (= 0.001 for aPD, = 0.023 for ePD). Furthermore, aPD patients showed less REM sleep (= 0.001), and ePD patients less NREM Stage 1 than NC patients (= 0.008).

In 12 of 20 PD patients (60%), polysomnography revealed RBD (nine in aPD, three in ePD). Another 15% had EMG correlates of impaired REM sleep atonia, but without typical behavioural manifestations of RBD, and 25% did not have any signs of RBD. In NC patients, polysomnography correlates of RBD were not observed.

Excessive daytime sleepiness

We hypothesized that EDS can be observed in both disorders, but is most pronounced in hypocretin-deficient NC and least in ePD.

Data on Epworth Sleepiness Scale scores and MSLT are listed in Table 1. Compared to ePD patients and healthy controls (mean sleep latencies 11.5 ± 4.2 and 12.1  ± 3.4 min), aPD patients were significantly sleepier on the MSLT (6.0 ± 3.4 min, = 0.009). However, NC patients were even sleepier than aPD patients (2.1 ± 1.5 min, = 0.007). Multiple sleep onset REM periods (SOREMs) in MSLT were more frequent in NC (eight of 10 patients, mean: 3.3) than in aPD (two of 10 patients, mean: 0.4), but not found in controls (mean: 0.1) and in ePD patients (mean: 0). We found no correlations between ESS and MSLT results.

In PD patients, stepwise linear regression analysis revealed that mean sleep latencies on MSLT were correlated negatively with disease duration (= −0.58, < 0.002) and motor impairment as assessed with UPDRS III (= −0.76, < 0.001). Age, gender, body mass indices (BMI), differences of sleep times between weekends and weekdays (as assessed with actigraphy), mean LED and mean total dopamine agonist doses were not associated with mean sleep latencies on MSLT. However, aPD patients on dopamine agonists were significantly sleepier on MSLT than aPD patients without dopamine agonist treatment (4.9 ± 2.6 versus 9.1 ± 3.1 min, = 0.045).

Hypocretin levels in the cerebrospinal fluid

CSF hypocretin-1 levels were normal (>320 pg mL−1) in all PD patients (Table 1, Fig. 1). Levels in aPD patients (487±57 pg mL−1) were not lower than in ePD patients (525 ± 63 pg mL−1, = 0.1). Levels in NC patients (14 ± 21) were lower than in PD patients (< 0.001). In PD patients, levels correlated with mean sleep latencies on MSLT (= 0.60, = 0.02) (Fig. 2). Conversely, no correlation was seen between CSF hypocretin and disease duration, age, LED or dopamine agonist dose. In two aPD cases, repeated CSF hypocretin measurements were available with time intervals of 4 and 5 years, respectively. In both patients, CSF hypocretin levels decreased by more than 100 mL pg−1 (Fig. 1).

Figure 1.

 Cerebrospinal fluid (CSF) hypocretin-1 levels (in pg mL−1) in seven patients with early Parkinson’s disease (ePD), seven patients with advanced Parkinson’s disease (aPD) and 10 patients with narcolepsy with cataplexy (NC). Each dot represents one patient, and horizontal lines display mean values. Note that repeated measurements were available in two aPD patients, with a time interval of 4–5 years between the first (aPD 1) and the second (aPD 2) measurement. Dotted lines connect values from the same patients.

Figure 2.

 Cerebrospinal fluid (CSF) hypocretin-1 levels (in pg mL−1) and mean sleep latencies on multiple sleep latency tests (MSLT) correlated in Parkinson’s disease (PD) patients (= 14, = 0.60, = 0.02). Each dot represents one patient, and a linear regression line is included.


With this prospective, controlled study, we aimed at examining nocturnal sleep and wakefulness in consecutive patients with presumably different levels of dopaminergic and hypocretinergic signalling deficits, i.e. patients suffering from early and late PD, NC and age-matched healthy controls. Cerebrospinal fluid hypocretin levels were lowest in NC patients, and there was a trend towards higher levels in ePD than in aPD patients. Furthermore, repeated CSF hypocretin measurements in two PD patients revealed a marked decrease of concentrations over time. Such an intra-individual decrease over time has not been reported previously in PD, but is in agreement with a previous autoptic study performing hypocretin measurements of ventricular CSF in aPD patients, which revealed lower levels than in controls, which decreased with the severity of the disease (Fronczek et al., 2007; Thannickal et al., 2007). These and our results suggest that there is progressive loss of hypocretin cells over the course of PD.

Confirming our hypothesis, we found that PD patients with predominant loss of dopamine-producing cells suffered from most disturbed nocturnal sleep: both deep sleep and sleep efficiency were lowest in aPD. However, compared to age-matched controls, partial loss of dopamine-producing cells in patients with ePD does not cause major changes of nocturnal sleep, but moderate disturbances such as a trend towards lower sleep efficiency and decreased amounts of deep sleep can be observed. With further progression and neuronal degeneration, sleep architecture changes or even disappears, and both sleep efficiency and deep sleep decline markedly.

The aetiology of nocturnal sleep disturbances in PD is complex, and therefore it must not be concluded that deficient dopaminergic signalling has a linear effect on nocturnal sleep. For instance, motor symptoms such as hypokinesia or tremor, as well as non-motor symptoms such as nocturia or pain, are likely to contribute to fragmented nocturnal sleep (Arnulf et al., 2000; Iranzo et al., 2002). There is, however, growing evidence that neurodegeneration in PD involves more brain structures than only the substantia nigra (Braak et al., 2003). The observation of progressive partial hypocretin and melanin concentrating hormone cell loss in the hypothalamus suggests that other sleep–wake modulating neuronal centres may also be involved. In this line, and because of our observation of an abolition of sleep architecture in only a subset of aPD patients, we assume that differential involvement of key structures for sleep–wake regulation causes different levels of nocturnal sleep disruption in PD patients, together with nocturnal motor symptoms, neuropsychiatric comorbidities, medication and co-existing sleep–wake disturbances including sleep apnoea and periodic limb movements during sleep.

In our PD patients, both sleep apnoea and periodic limb movements were neither more frequent nor more severe than in our healthy controls. The question of whether or not these SWD are more prevalent in PD than in controls is subject of an ongoing debate, and needs to be addressed by large controlled studies. RBD in our PD patients was more often present in aPD than in ePD patients. This observation is in agreement with a previous study in 457 PD patients, which found that this parasomnia increases with age and with disease duration (Sixel-Döring et al., 2011).

EDS is common in PD, particularly in the advanced stage of the disease (Arnulf et al., 2002; Gjerstad et al., 2006). Our MSLT results confirm earlier findings: in aPD patients, EDS was present in 70%, whereas mean sleep latencies in ePD patients were similar to healthy controls (EDS present in 10%). Nevertheless, EDS and enhanced REM sleep pressure were most pronounced in NC patients, i.e. in patients with subtotal loss of hypocretin neurones, but spared dopamine cells. In aPD patients, EDS was milder than in NC, irrespective of dopaminergic treatment.

The aetiology of EDS in PD patients is not understood fully, but is probably multi-factorial. First, dopaminergic treatment, particularly with dopamine agonists, is associated with EDS (Gjerstad et al., 2006). As assessed by stepwise linear regression analysis, we did not find a correlation between LED or total dose of dopamine agonists and EDS as assessed with MSLT or the Epworth Sleepiness Scale. Instead, our study revealed associations of EDS with disease duration and motor impairment as assessed with the UPDRS III scale. Along the same lines, Roth et al. (2003)found that EDS is unrelated to the use of a specific dopamine agonist or to polysomnographic findings of disturbed night sleep (Roth et al. 2003).

These findings suggest that neurodegeneration of sleep–wake modulating neuronal systems may also contribute to EDS in PD. Autopsy studies in PD patients have shown that approximately 40% of hypocretin neurones are lost in advanced stages (Fronczek et al., 2007; Thannickal et al., 2007). Our finding of a correlation between hypocretin levels and objective EDS as assessed with MSLT corroborates the hypothesis that degeneration of hypocretin neurones is contributing to EDS in PD. Similarly, in narcolepsy, lower CSF hypocretin levels are linked to more severe EDS (Baumann et al., 2006). Conversely, a study performing MSLT in eight PD patients without dementia and in seven with dementia revealed similar CSF hypocretin levels in both groups and no statistical relation between mean sleep latencies and CSF hypocretin levels (Compta et al., 2009). This discrepancy with our own results might be a result of the smaller number of examined patients and the inclusion of both demented and non-demented PD subjects. The two patients with lowest levels, however, had the shortest mean sleep latency or multiple sleep onset REM periods. The authors concluded that mechanisms other than disturbed hypocretin signalling may also be responsible for EDS. In fact, besides hypocretin neurones, degeneration of other neuronal systems, particularly monoaminergic and cholinergic nuclei, may also contribute to EDS in PD patients.

Conversely, and in contrast to NC patients, REM sleep pressure as assessed by the number of SOREMs on MSLT in PD patients was mainly normal, both in nocturnal PSG and in daytime MSLT. However, despite the fact that PD patients did not express significantly more SOREMs than controls, we observed multiple SOREMs in two aPD patients but in no controls. This is in line with previous findings, including those of Rye et al. (2000), who reported multiple SOREMs and short mean sleep latencies on MSLT in four (15%) of 27 PD patients. This proportion appears to be higher than findings in healthy controls, as shown in a large community cohort study (Mignot et al., 2006): in this study, multiple SOREMs were observed in 13% of males and 6% of females. Thus, it remains unclear whether partial hypocretin loss in aPD is not only associated with EDS, but also with enhanced REM sleep pressure. To confirm (or refute) the hypothesis that partial hypocretin loss contributes to enhanced REM sleep pressure, the present study was probably underpowered.

Last, but not least, disturbed nocturnal sleep might contribute to EDS in PD. However, in accordance with previous studies, we did not find that polysomnographic variables such as sleep efficiency, arousal indices, apnoea–hypopnoea indices and periodic limb movement indices were associated with EDS (Arnulf et al., 2002; Baumann et al., 2005).

This study has limitations. First, the groups of prospective patients, although small, were studied carefully, and age and gender were not matched between NC and PD patients. This might contribute to the differences seen between the examined patients’ groups. Furthermore, the higher LED in aPD patients compared to ePD patients might also have influenced sleep–wake parameters. As a further possible limitation, we excluded controls with sleep apnoea. Despite the fact that apnoea–hypopnoea parameters were similar in all groups, the slightly higher apnoea severity in aPD patients might have affected sleep fragmentation and structure. In future studies, a second control group with matched apnoea severity might also be included.

In conclusion, our study reveals that nocturnal sleep is more disturbed in dopamine-deficient PD, and excessive daytime sleepiness is more severe in hypocretin-deficient NC. However, in PD patients, partial and progressive hypocretin signalling deficiency was also linked to EDS, as quantified with MSLT. Because of degeneration of multiple brain areas in PD, including probable lesions to other sleep–wake modulating neuronal populations, we may not conclude directly that dopamine deficiency causes primarily nocturnal sleep disruption in contrast to hypocretin deficiency, which causes hypo-arousal and EDS.


This study was supported by a competitive grant from the Swiss Parkinson Foundation (

Disclosure Statement