Clinical Manifestations of Subjective Sleep Disorders in Chinese Patients with Parkinson’s Disease and Their Relationship with Dopaminergic Drugs

Parkinson’s disease (PD) is a neurodegenerative disease commonly found in middle-aged and elderly individuals. Between 42% and 98% of PD patients have sleep disorders, including difficulty staying asleep, restless legs, vivid dreams, and excessive daytime sleepiness (EDS), which may be related to various nocturnal disturbances (such as difficulty turning over in bed, nocturia, and pain), anxiety and depression, and the effects of anti-PD drugs [1]. The etiology of sleep disorders in PD patients may be multifactorial and mutually influential. Many PD patients experience improved motor function after sleep, indicating that sleep affects the movement control of PD [2, 3]. Therefore, it is important to assess the sleep status of PD patients.

Anti-PD drugs include dopaminergic and non-dopaminergic drugs. The effect of dopaminergic drugs, including dopamine receptor agonists (DA) and levodopa preparations, on sleep in PD patients has not yet been well examined and conflicting effects have been reported. In clinical practice, dopaminergic drugs appear to exert different effects on sleep, with small doses promoting slow-wave sleep (SWS), rapid eye movement (REM) sleep, and possibly inducing drowsiness; and large doses reducing SWS, REM sleep, and increasing wakefulness [4]. Among them, non-ergot DA can alleviate some sleep disorders in PD patients [5‒7], while possibly inducing other sleep disorders [8]. Studies have found that the adverse effects of levodopa on sleep may be stronger than DA [9]. In addition, the dose, type, and time of taking dopaminergic drugs in relation to bedtime may affect the quality of sleep in PD patients. Elucidating the effects of dopaminergic drugs on sleep is important to determine strategies for optimizing sleep in PD patients.

This study retrospectively analyzed the clinical data of 248 PD patients treated in the PD clinic or hospitalized in the Department of Neurology of Beijing Hospital from 2012 to 2015. Parkinson’s disease sleep scale (PDSS) Chinese version was used to evaluate subjective sleep disorders in PD patients throughout the different stages of evolution, and the incidence and severity of each item of subjective sleep disorders were investigated. Combined with various disease characteristics of PD, the relationship between sleep disorders and the type, dosage, and time of taking dopaminergic drugs was analyzed to further clarify the influencing factors of subjective sleep disorder in PD patients.

We consecutively enrolled 248 PD patients who were treated in the PD clinic or hospitalized in the Department of Neurology of Beijing Hospital from 2012 to 2015. Among them, 16 patients were excluded due to incomplete questionnaire data. Finally, 232 patients were included in the study. The cohort comprised of 114 males and 118 females, aged 30–88 years, with an average age of 66.2 ± 9.7 years. The disease course ranged from 12 to 264 months, with a median of 73.5 (36.0, 115.5) months (M [Q1, Q3]). All the enrolled patients met the criteria of the UK Parkinson’s Disease Brain Bank published in 1992 [10]. The exclusion criteria were secondary PD syndrome, PD plus syndrome, history of deep brain stimulation, other neurological disorders, and mental disorders. Patients were also excluded if their score on the Mini-Mental State Examination (MMSE) was less than 24 points. This study was approved by the Ethics Committee of Beijing Hospital, and all enrolled subjects provided their informed consent.

Demographic and clinical information, including age, onset age, gender, comorbidities, and disease course, was collected. The Unified Parkinson Disease Rating Scale (UPDRS) and modified Hoehn-Yahr (H&Y) stage were used to assess disease severity. Cognitive function was assessed by the MMSE. The Hamilton Rating Scale for Depression (HAMD) and the Hamilton Rating Scale for Anxiety (HAMA) were used to assess the emotional state. The activities of daily living (ADL) scale was used to assess daily living ability. The patients were divided into 3 groups according to the H&Y stage: the PD-1 group included 62 patients with H&Y stages 1 and 1.5 (26.7%), the PD-2 group included 112 patients with H&Y stages 2 and 2.5 (48.3%), and the PD-3 group included 58 patients with H&Y stages 3 and 4 (25.0%). The clinical information of all PD patients and PD groups 1–3 is shown in Table 1.

The PDSS-Chinese version [11], which consists of 15 items, was used for evaluation. Scores for a given individual item range from 0 to 10: 0 represents the worst, 10 represents the best. The maximum total score is 150 points, which indicates that the patient is completely free of symptoms associated with sleep disorders. Patients were asked to rate their overall sleep over the prior month. The individual score of 15 items and total scores of each patient were calculated.

The types and dosages of anti-PD drugs were collected. The total intake of dopaminergic drugs over a 24-h period was converted to levodopa equivalent doses (LED-24 h) which uses commonly accepted conversion factors [12]. Patients were divided into 3 groups according to the type of dopaminergic drugs: the PD-A group = levodopa monotherapy, PD-B group = DA monotherapy, and PD-C group = combined therapy (levodopa and DA). The patients were asked if they take dopaminergic drugs within 4 h of bedtime, and the dose was converted into the LED-bb (before bedtime). In addition, the use of other related drugs (such as hypnotic and antidepressive drugs) was recorded and considered for analysis. All of the above questionnaires and scales were scored by a neurologist and performed during the “on” status.

SPSS 19.0 software was used for analysis. The Kolmogorov-Smirnov (K-S) normality test was performed on all quantitative variables. Data conforming to a normal distribution were expressed in mean ± standard deviation (x ± s), t test was used for comparison of two groups, and one-way analysis of variance (ANOVA) was used for comparison of three groups classified by H&Y or type of dopaminergic drug. When a significant difference was detected by ANOVA, a post hoc test (Bonferroni test/Tamhane test) was conducted. The Levene test was used to analyze the homogeneity of variances. Data that did not conform to the normal distribution were expressed as median (quartile) M (Q1, Q3), and Kruskal-Wallis tests were used for comparisons between the three groups. When a significant difference was detected, a Kruskal-Wallis 1-way ANOVA test was conducted. Frequency and percentage were calculated for qualitative variables, and χ² test was used for comparison among the three groups. When a significant difference was detected, a Bonferroni test was conducted.

Stepwise regression analyses were conducted to investigate the influencing factors of sleep disorders. The total PDSS score was used as the dependent variable, while age, disease duration, H&Y stage, use or not use of any non-dopaminergic anti-PD drugs, UPDRS total score, UPDRS III (motor function), scores of MMSE, ADL, HAMD, and HAMA, LED-24 h, and LED-bb were independent variables. The introduction variable p value was set to 0.05, and the exclusion variable p value was set to 0.10. Significance of differences was defined as two-tailed p < 0.05.

The average total score of PDSS in the 232 patients was 119.3 ± 19.7. Among the 15 items of sleep disorders assessed by PDSS, the top 3 with the highest incidence were item 1 (overall sleep, 89.2%), item 3 (sleep fragmentation, 80.6%), and item 8 (nocturia, 76.3%), and the top 3 with the lowest incidence were item 12 (painful posturing at awakening, 48.7%), item 9 (urinary incontinence, 35.8%), and item 7 (hallucinations, 30.6%) (Fig. 1). In terms of the severity of each item, the top 3 with the lowest average scores were item 8 (nocturia, 6.8 ± 3.0), item 3 (sleep fragmentation, 7.0 ± 2.6), and item 1 (overall sleep, 7.1 ± 1.9), while the top 3 with the highest average scores were item 7 (hallucinations, 9.2 ± 1.8), item 9 (urinary incontinence, 8.9 ± 2.1), and item 12 (painful posturing at awakening, 8.6 ± 2.1) (Fig. 2).

The PD patients’ clinical features and total PDSS scores classified by H&Y stage are shown in Table 1. ANOVA revealed that patients with higher H&Y stage were associated with lower PDSS, higher total UPDRS, UPDRS II, and UPDRS III scores, and higher LED-24 h. The post hoc test revealed that, except for the total PDSS score between PD-1 and PD-2 groups, PD-1 and PD-3 groups, and the LED-24 h between PD-2 and PD-3 groups, there were statistically significant differences between the groups (p < 0.05). Kruskal-Wallis test showed that patients with higher H&Y stage were associated with a longer disease course, lower MMSE score, higher UPDRS I, UPDRS IV, HAMD, HAMA, ADL scores, and dosage of dopaminergic drugs within 4 h before bed (LED-bb). The Kruskal-Wallis 1-way ANOVA test revealed that, except for the MMSE score between the PD-1 and PD-2 groups, PD-2 and PD-3 groups, the UPDRS I score between the PD-1 and PD-2 groups, the HAMD score between the PD-1 and PD-2 groups, the HAMA score between the PD-1 and PD-2 groups, PD-2 and PD-3 groups, LED-bb between the PD-1 and PD-2 groups, there were statistically significant differences between the groups (p < 0.05). The χ² test revealed that patients with higher H&Y stage were associated with increased proportion of dopaminergic medication intake within 4 h of sleep. Furthermore, the Bonferroni test demonstrated statistically significant differences between the PD-1 and PD-3 groups, PD-2 and PD-3 groups. There were no statistically significant differences in age or sex between the three groups.

The individual items of PDSS scores classified by the H&Y stage are shown in Table 2. ANOVA revealed that the scores of items 4 (nocturnal restlessness) and 15 (daytime sleepiness) showed statistically significant differences among the three groups (F = 3.70, p = 0.03; F = 3.81, p = 0.02). The post hoc test showed that the score of item 4 (nocturnal restlessness) was significantly different between the PD-1 and PD-3 groups (p = 0.02), as was the score of item 15 (daytime sleepiness) between the PD-2 and PD-3 groups (p = 0.03).

Of the 232 PD patients, 1 took only benzhexol, 2 took only selegiline, and the remaining 229 took anti-PD drugs containing dopaminergic drugs. Among them, 79 patients (34.1%) took levodopa in monotherapy, 21 patients (9.1%) took DA in monotherapy, and 129 patients (55.6%) took levodopa and DA combined therapy, of which, 40 patients took additional non-dopaminergic anti-PD drugs, including 27 (11.6%) patients who took amantadine, 35 (15.1%) patients took selegiline, and 12 (5.2%) patients took benzhexol. Overall, 150 (64.7%) patients took DA (86 pramipexole, 64 Piribedil). The mean dose of LED-24 h in all patients was 624.9 ± 346.2 mg, among which 158 patients (68.1%) took dopaminergic drugs within 4 h before bedtime, and LED-bb was 125.0 (0.0, 200.0) mg, see Table 1. In addition to anti-PD drugs, other drugs included 23 patients (9.9%) taking antidepressive drugs regularly and 32 patients (13.8%) taking hypnotic drugs occasionally. None of these treatments has had an effect on the scores obtained through the PDSS (t test: antidepressive drugs 118.3 ± 16.5, no antidepressive drugs 119.7 ± 22.1, p > 0.05; hypnotic drugs 120.9 ± 23.1, no hypnotic drug 118.1 ± 17.8, p > 0.05).

Patients with DA monotherapy were slightly younger than those with levodopa monotherapy (62.6 ± 10.7 vs. 67.2 ± 9.7; p = 0.16). The disease course of patients treated with DA monotherapy (56.9 ± 28.9 months), levodopa monotherapy (80.2 ± 58.0 months), and combined therapy (86.7 ± 51.6 months) was gradually prolonged (F = 2.87, p = 0.06), and there was no statistically significant difference in the disease course between the three groups (p > 0.05). The total and individual item scores of PDSS were compared among the three groups, and the differences were not statistically significant (Table 3).

Stepwise regression analysis with PDSS as the dependent variable identified that LED-bb (p < 0.00), LED-24 h (p < 0.00), and HAMD (p = 0.01) were determinants of PDSS (Table 4).

Some studies have reported that the incidence of sleep disorders in PD patients is 42∼98%. The most frequent complaints and symptoms are nocturia, difficulty to turn over in bed, painful muscle cramps, nightmares, limb or facial dystonia, leg jerks, visual hallucinations, vivid dreams, limb pain, and discomfort [13]. In addition, depression and dopaminergic therapy can also cause sleep disruption [14].

In this study, the sleep symptoms reported by PD patients, ranked from highest to lowest incidence, were (Fig. 1): sleep fragmentation (item 3), nocturia (item 8), daytime sleepiness (item 15), nocturnal restlessness (item 4), nightmare (item 6), difficulty falling asleep (item 2), painful muscle cramps (item 11), fidget in bed (item 5), sleep refreshment (item 14), numbness or tingling at night (item 10), tremor on waking (item 13), painful posturing at awakening (item 12), urinary incontinence (item 9), and hallucination (item 7). This result is largely consistent with the previous literature. Among the 15 items of PDSS, the top 3 with the lowest average scores and the most severe symptoms were item 8 (nocturia), item 3 (sleep fragmentation), and item 1 (overall sleep) (Fig. 2). Nocturia is a very common non-motor symptom, which is related to increased urine volume and decreased bladder capacity in PD patients at night [15]. The frequency of nocturia increases with the severity of PD, leading to difficulty maintaining sleep at night, thus directly affecting the overall sleep quality of patients at night. The score of sleep fragmentation (item 3) is lower than that of difficulty falling asleep (item 2), which is similar to the report by Chaudhuri et al. [15], suggesting that insomnia of PD patients is mainly caused by multiple nocturnal symptoms and sleep fragmentation at night, rather than difficulty falling asleep.

Nocturnal restlessness (item 4) reflects restlessness of the legs or arms at night in PD patients. In our study, the difference in the score of the three groups classified by H-Y stage was statistically significant (p < 0.05), and the patients with more severe disease were more susceptible to nocturnal restlessness, which in turn affected the quality of sleep. Increased daytime sleep (item 15) is also a common symptom of sleep disturbances in PD patients. In our cohort, 75% of patients had increased daytime sleep, which was similar to the rate of 87.5% of PD patients with varying degrees of daytime hypersomnia reported by Ondo et al. [16]. The difference between the three groups was statistically significant, suggesting that patients with more severe disease were more likely to have increased daytime sleep [17]. It has previously been reported that increased daytime sleep may also be associated with cognitive decline, the use of large dosage of dopaminergic drugs, anxiety and depression, visual hallucination, and restless leg syndrome [16].

As shown in Table 1 of this study, a higher H&Y stage in PD patients was correlated with a longer disease course, worse cognitive ability, more severe motor symptoms, more severe depression and anxiety, poorer activity of daily living, and larger doses of daily dopaminergic drugs and dopaminergic drugs taken at bedtime. The H&Y stage is not related to age or sex. There was statistical significance in PDSS scores between the three groups, suggesting that subjective sleep quality might get worse with the increase of H&Y stage.

The effects of dopaminergic drugs, including DA and levodopa preparations, on sleep in PD patients have been studied in recent years, but the results have been inconsistent. DA has been reported to cause increased daytime sleep in patients, especially non-ergot derivatives, which are associated with sleep attacks in PD patients [18, 19]. It has also been reported that patients treated with different DA drugs, whether ergot or not, and even levodopa can cause sleepiness [9], and there is no difference in the incidence of EDS among the different types of drugs [9, 20‒22]. Sleep attack and EDS are phenomenological effects that can be attributed to any dopaminergic treatment and are not unique to DA [23]. Long-term use of higher doses of levodopa has been reported to be associated with lower subjective sleep quality [24]. However, the objective measurement of parameters by polysomnography failed to find this result; Verbaan et al. [25] observed a negative association between levodopa dose and subjective sleep score.

In our study, no significant differences in overall or individual PDSS scores were found between the three groups of DA monotherapy, levodopa monotherapy, or combination therapy. Hallucination is one of the common neuropsychiatric non-motor symptoms in PD. It has been reported that hallucination occurs as both part of the pathological course of the disease itself and as a result of dopaminergic drug therapy. Approximately 40% of PD patients experience drug-induced hallucination [26], for which patients using DA are at increased risk [27, 28]. In this study, the incidence of hallucination was only 30.6%, and there was no significant difference between the three groups. This may be because patients with severe hallucination discontinue dopaminergic therapy, or that the severity of the disease in our enrolled patients has not yet reached the level of hallucination. The result of this study suggests that there is no significant difference between the use of two dopaminergic drugs alone or even in combination for subjective sleep disorders in patients, and thus, there is no difference in subjective sleep quality between patients taking DA or levodopa monotherapy.

Previous studies have reported the effect of dopaminergic drug dosage on sleep quality. Data from animal studies have suggested that low-dose dopaminergic drugs promote sleep by acting on presynaptic D2 receptors, while high-dose dopaminergic drugs inhibit sleep by acting on postsynaptic D1 and possibly D2 receptors [29]. Data from studies based on PD patients have shown that high-dose dopaminergic drugs adversely affect both objective sleep as assessed by ActiGraph [30, 31] and subjective sleep as assessed by PDSS [30]. An open-label study found that taking levodopa-controlled-release tablets before bed did not improve the sleep quality of PD patients [32]. Stepwise multiple regression analysis in this study showed that LED-bb and LED-24 h were the first and second main influencing factors of PDSS total score, that is, the higher the dosage of dopaminergic drugs taken within 4 h before sleep and the higher the total dosage of dopaminergic drugs taken daily, the more severe the subjective sleep disorder of PD patients. It has been reported that a high dosage of dopaminergic drugs before sleep is associated with an increased percentage of stage II sleep time and decreased percentage of REM sleep time [33], causing sleep disorders in PD patients, which was in line with our findings. However, some studies have shown conflicting results. For example, several studies have found that dopaminergic drugs are beneficial for sleep in patients with advanced PD [34‒36], which may be due to the fact that the benefits of dopaminergic drugs in improving nocturnal motor symptom exceed the negative effects of dopaminergic drugs on wake-promoting, thus improving sleep quality. However, our patients were mainly in H-Y stage 2–3, meaning we did not see the benefit of high-dose dopamine on sleep improvement.

In this study, the third determinant of subjective sleep disorder in PD patients was depression. Depression is associated with dopamine loss in the mesolimbic dopaminergic pathway of the bilateral ventral tegmental area (VTA), loss of serotonin (5-hydroxytryptamine, 5-HT) neurons in the dorsal raphe nucleus, decreased norepinephrine, and choline neurodegeneration [37]. PD patients with depression show greater dopamine loss in the mesolimbic dopaminergic pathway from VTA, as well as serotonin depletion in the dorsal raphe nucleus, noradrenalin depletion in the locus ceruleus, and acetylcholine depletion in the pedunculopontine nucleus and nucleus basalis of Meynert [38] compared to patients without depression [39, 40]. Decreased dopamine neurons in the midbrain can cause sleep-wake cycle disorders in PD patients [41]. From the perspective of sleep structure, a decrease of 5-HT can lead to a decrease of SWS, and choline neurodegeneration and a decrease of norepinephrine can also induce RBD [14]. Previous studies have reported that about 43% of PD patients suffer from depression, which is also related to sleep disorders [42]. Our results were consistent with previous reports.

In recent years, continuous drug delivery has emerged as a viable and practical treatment option for dopamine replacement therapy in advanced PD. A meta-analysis study of over 1,200 patients with advanced PD reported that levodopa-carbidopa intestinal gel significantly improved non-motor symptom burden, especially sleep outcomes such as EDS, insomnia, and nocturia [43]. Administration of nocturnal apomorphine infusion significantly reduced nocturnal complications. The mean PDSS score changed significantly compared with the control group [44]. A meta-analysis of 4,682 PD patients from 16 randomized controlled trials found that rotigotine transdermal patch significantly improved PDSS-2 scores, suggesting that rotigotine transdermal patch led to a remarkable improvement in sleep quality [45]. It can be seen that non-oral dopaminergic continuous drug delivery strategies may be an effective method to improve PD sleep disorders [46].

There are some limitations to this study. Polysomnography is the gold standard for assessing sleep disorders, but this requires a specialized laboratory and expensive equipment, and it cannot be practically used to measure the variable aspects of sleep disorders in PD patients. PDSS is a specialized sleep scale for PD, covering all aspects of PD sleep problems, especially the details of sleep disorders and nocturnal symptoms. With good reliability, validity, and feasibility, PDSS-Chinese version has been widely used in China [11]. Because of the small number of patients on dopaminergic monotherapy, our examination of the differential effects of different dopaminergic drugs was unlikely to be implemented. Previous studies have suggested that some non-dopaminergic anti-PD drugs also have potential effects on PD sleep [47]. However, our study did not reach such results. The effects of non-dopaminergic drugs on sleep should be categorized and compared.

Subjective sleep disorders in PD patients may be multifactorial. Our results indicate that a high dosage of dopaminergic drugs taken prior to sleep and daily total high dosage of dopaminergic drugs as well as depression have negative effects on the subjective sleep of PD patients. In particular, the timing and dosage of dopaminergic drugs before bedtime should be considered when treating sleep disorders of PD patients. It is necessary to further study whether the time and dosage of the last daily dopaminergic drug before sleep and different types of drugs have different effects on sleep in PD patients as well as the effects of other neurotransmitter systems drugs besides dopaminergic drugs on subjective sleep at night, combined with objective polysomnography monitoring data in PD patients. In clinics, attention should be paid to the sleep management of PD patients. While treating motor symptoms of PD patients, reasonable adjustment should be made to the timing and dosage of anti-PD medication as well as improve patients’ depression, which can effectively improve their subjective sleep disorders.

The authors thank the reviewers for their helpful comments. The authors also thank all of the participants for their hard work.

The study was conducted in accordance with the Declaration of Helsinki and approved by the Ethics Committee of Beijing Hospital, and the reference number is 2018BJYYEC-183-03. Written informed consent was obtained from all subjects involved in the study.

The authors declare that they have no financial or other conflicts of interest in relation to this research and its publication.

This work was supported by the project “National Major Multidisciplinary Cooperative Diagnosis and Treatment Capacity Building Project” from the National Health Commission of the People’s Republic of China and the National Key R&D Program of China under Grant (2017YFC1310200).

Dongdong Wu designed the study, analyzed the data, and drafted the manuscript for intellectual content; Jing He, Ying Jin, Huijing Liu, Xinxin Ma, and Yunfei Long were responsible for collection; Kai Li and Wei Du analyzed the data; Wen Su and Shuhua Li participated in the interpretation of the data; and Haibo chen supervised the finalization of the manuscript.

All data generated or analyzed during this study are included in this article. Further inquiries can be directed to the corresponding author.