Clinical Features and Correlates of Poor Nighttime Sleepiness in Patients with Parkinson's Disease

Objective The present study investigated the clinical features and correlates of poor nighttime sleepiness (PNS) in patients with Parkinson's disease (PD). Methods One hundred ten patients with PD (divided into PD-PNS group and PD-nPNS group) and forty-seven controls (nPD-PNS group) were enrolled in this study. Demographic information was collected. Patients were assessed according to the unified Parkinson's disease rating scale (UPDRS) and Hoehn–Yahr (H&Y) stage scale. Patients were also evaluated according to the Pittsburgh sleep quality index (PSQI), Epworth sleepiness scale (ESS), rapid eye movement sleep behavior disorder screening questionnaire (RBD-SQ), restless leg syndrome (RLS) diagnosis, Hamilton's depression scale (HAMD), and Hamilton's anxiety scale (HAMA). Results The prevalence of PNS was 55.45% (61/110) in patients with PD. The PD-PNS group tended to have a longer duration of disease, higher UPDRS-I and UPDRS-III scores, a higher percentage of RLS patients, and higher HAMA and HAMD scores than those of the PD-nPNS group. The PD-PNS group tended to have a higher percentage of RBD and RLS patients and higher HAMA and HAMD scores than those of the nPD-PNS group. Analysis of the PSQI components and PSQI impact factors showed that the PD-PNS group had worse subjective sleep quality (χ2 = −2.267, P = 0.023), shorter sleep latency (χ2 = −2.262, P = 0.024), fewer sleep medications (χ2 = −4.170, P ≤ 0.001), worse daytime functioning (χ2 = −2.347, P = 0.019), and an even higher prevalence of increased nocturia (χ2 = 4.447, P = 0.035), nightmares (χ2 = 7.887, P = 0.005), and pain (χ2 = 9.604, P = 0.002) than those of the nPD-PNS group. Analysis also indicated that the PSQI global score positively correlated with BMI (r = 0.216, P < 0.05), H&Y stage (r = 0.223, P < 0.05), UPDRS-I (r = 0.501, P < 0.01), UPDRS-III (r = 0.425, P < 0.01), ESS (r = −0.296, P < 0.01), RBD (r = 0.227, P < 0.05), RLS (r = 0.254, P < 0.01), HAMA (r = 0.329, P < 0.01), and HAMD (r = 0.466, P < 0.01). In the final model, H&Y stage, RLS, UPDRS-III, and HAMD remained associated with the PQSI score (P ≤ 0.001, P ≤ 0.001, P = 0.049, P ≤ 0.001, respectively). Conclusions Our data showed that PNS was common in patients with PD. H&Y stage, UPDRS-III, HAMD, and RLS were positively associated with PNS. Attention to the management of motor symptoms, RLS, and depression may be beneficial to nighttime sleep quality in patients with PD.


Introduction
Parkinson's disease (PD) is the second most common neurodegenerative disease and is characterized by motor and nonmotor dysfunctions [1]. Disordered sleep is one of the most frequent nonmotor symptoms in PD patients and has a significant negative impact on quality of life [2]. Disordered sleep affects 40-98% of PD patients worldwide and 47.66-89.10% of PD patients in China [3,4]. e possible pathogenesis of PD with sleep disturbance includes thalamocortical pathway degeneration and changes in neurotransmitter systems [5]. e etiology of sleep disturbance is multifactorial and involves the degeneration of areas regulating sleep, sleep structure affected by drugs, drug-induced sleep disturbance, and sleep fragmentation due to multiple factors [6]. Although many studies performed over the past decade have investigated the clinical characteristics of sleep disturbance among patients with PD, some issues remain unclear and warrant further delineation.
Sleep disorders in PD patients include insomnia, vivid dreams, restless legs syndrome (RLS), rapid eye movement sleep behavior disorder (RBD), periodic limb movements (PLM), circadian rhythm disruption, and excessive daytime sleepiness (EDS), which lead to nighttime and daytime sleep problems [7]. Nighttime and daytime sleep problems significantly impair quality of life, especially nighttime sleep problems, which increase the risk of cardiovascular and cerebrovascular events and lead to an economic burden for patients and their caregivers [8]. us, it is important to identify those patients with nighttime sleep problems early and provide potential treatment options as soon as possible. Identifying patients with unique clinical features can facilitate analyses of subtype specific biomarkers and epidemiological and clinical treatments. So far, few studies have focused on overall nighttime sleep quality in Chinese patients with PD, and clinical features and correlates of poor nighttime sleepiness in PD patients have not been investigated.
Currently, objective assessment of sleep physiology relies primarily on polysomnography (PSG), which records the electroencephalographic (EEG) activity, extraocular eye movements, mentalis muscle tone, air flow, respiratory effort, and cardiac rhythm to determine sleep staging and aids in diagnosing different types of sleep disorders. However, the assessment of sleep quality is usually based on patient self-reporting, interviews, and psychological variables. e latter can better reflect the subjective feelings of patients and directly represent clinical symptoms, which help researchers to easily obtain clinical characteristics and quickly screen target patients. e Pittsburgh sleep quality index (PSQI) is generally used to evaluate overall nighttime sleep quality and widely applied in clinical work and clinical research. Patients with PSQI scores higher than 7 are generally considered poor nighttime sleep (PNS) patients [9].
In general population, self-reported sleep disturbances are increasingly common with advancing age. In Outcomes of Sleep Disorders in Older Men (MrOS Sleep) study, up to 40.7% of older community-dwelling men were reported to have poor nighttime sleep quality (PSQI > 5) [10]. Additionally, age-related changes in sleep/wake patterns including lower sleep efficiency, longer sleep latency, greater nighttime wakefulness, and higher number of long wake episodes have been reported in population-based cohorts of older men and women [11]. Prior population-based studies in older people have also reported a high prevalence of sleepdisordered breathing and periodic leg movements in sleep [12]. To our knowledge, there are few studies investigating the nighttime sleep patterns of PD patients or discussing the differences of clinical features in sleep disorders that may exist between PD patients and general populations. e present study therefore aimed to characterize the clinical features of PNS in patients with PD and control subjects, to investigate the correlates of PNS among PD patients.

Patients.
A total of 110 PD patients were recruited from the clinic or impatient Department of Neurology, Xuzhou Central Hospital/Clinical Hospital of Xuzhou Medical University, from March 2017 to July 2018. e clinical diagnosis of idiopathic PD was determined based on the MDS clinical diagnostic criteria for Parkinson's disease [13]. Forty-nine simple poor nighttime sleepers without PD were recruited from the clinic and selected as controls. People with other diseases, such as respiratory diseases, urinary system diseases, cardiovascular and cerebrovascular diseases, and primary mental disorders, were excluded. People who were unable to finish the questionnaire were also excluded.
e Ethics Committee of Xuzhou Clinical Hospital of Xuzhou Medical University approved this study.

Assessments.
Each participant completed the Pittsburgh sleep quality index (PSQI), Epworth sleepiness scale (ESS), rapid eye movement sleep behavior disorder screening questionnaire (RBD-SQ), restless leg syndrome (RLS) diagnosis, Hamilton's depression scale (HAMD), and Hamilton's anxiety scale (HAMA). e PSQI is a self-report questionnaire that assesses nighttime sleep over a 1-month time period [9]. Nineteen individual items generate seven "component scores": subjective sleep quality, sleep latency, sleep duration, habitual sleep efficiency, sleep disturbances, sleep medication, and daytime dysfunction. e scores for these components range from 0 (no difficulty) to 3 (severe difficulty) and are summed to produce a global measure of sleep disturbance, with a global score ranging from 0 to 21. Higher scores represent poorer subjective sleep quality. Notably, the fifth item is composed of 9 nighttime symptoms that may affect sleep quality, including difficulty falling asleep, fragmented sleep or awakening earlier, increased nocturia, disturbance in respiration, cough or snoring, feeling cool, feeling hot, nightmares, and pain. Subjects in the present study with PSQI > 7 were considered poor nighttime sleepers [14]. According to their PSQI scores, PD subjects were divided into PD-PNS group (PSQI > 7) and PD-nPNS group (PSQI < 7). Subjects with PSQI > 7 but without Parkinson's disease were selected for the nPD-PNS group.
A movement disorder specialist clinically evaluated PD subjects in an "on" state. Clinical data were collected, including demographic information (age, sex, education), age of movement symptom onset, disease duration, concurrent diseases, medication information for PD (the levodopaequivalent dose (LED) was calculated based on previously reported conversion factors) [15], and detailed medical history. e unified Parkinson's disease rating scale (UPDRS) and Hoehn-Yahr stage (H&Y stage) scale were applied to all PD subjects in an "on" state [16]. Motor phenotypes were identified based on the ratio of the mean tremor score (sum of items 20 and 21 in the UPDRS-III divided by four) to the mean bradykinesia/rigid score (sum of items 22-27 and 31 in the UPDRS-III divided by 15).
All subjects underwent evaluations using the ESS, rapid eye movement (REM) RBD-SQ, RLS diagnosis, HAMD, and HAMA. e ESS is a widely used questionnaire for assessing the general level of daytime sleepiness. e ESS is composed of eight items that address typical day-to-day situations. Each item ranges from 0 to 3 points (0 � would never doze, 3 � high chance of dozing) to yield a total ESS score of 0-24 (lowest to highest sleep propensity). Subjects with ESS > 10 were considered excessive daytime sleepers (EDS), and normal sleep propensity was 0-10 [18]. e RBD-SQ is a valuable tool for screening rapid eye movement (REM) sleep behavior disorder (RBD) [19]. An RBD score of 5 or greater was defined as probable RBD. A diagnosis of RLS was made according to the RLS diagnostic criteria proposed by the International Restless Legs Syndrome Study Group (IRLSSG) in 2014, which is based on four essential features of the questionnaire after the exclusion of RLS mimics, such as positional discomfort, muscle cramp, venous stasis, vascular claudication, and peripheral neuropathy [20]. HAMD and HAMA were used to assess depression and anxiety, respectively [21]. Neurologists confirmed the final results.

Statistical Analysis.
e measurement data are expressed as means ± SD (standard deviation), and the enumeration data are shown as numbers (rate). Two independent sample t-tests were used to analyze the measurement data of two groups with a normal distribution. Non-normal distribution data were analyzed using nonparametric tests (Mann-Whitney test). e enumeration data were analyzed using the χ 2 test. We used Spearman's correlation to assess the correlations between the different factors and PSQI global score and PSQI components. e PSQI global score was the dependent variable, and clinical factors were independent variables. Line regression analysis was used to show the relationship of the parameters. P < 0.05 was considered statistically significant.

Demographic and Medication Data of PD Patients with and without PNS.
A total of 110 PD patients were included in this study. e demographic data and medication data of PD patients with PNS (PD-PNS) and without PNS (PD-nPNS) are shown in Table 1.

Demographic and Medication Data of PNS with or without PD.
e demographic and medication data of PNS patients with PD (PD-PNS) and without PD (nPD-PNS) are shown in Table 2.

Regression Analysis of Factors Associated with the PSQI Global Score of PD-PNS Patients.
Using the PSQI global score as the dependent variable and sex, age, disease duration, BMI, LED, H&Y stage, UPDRS I, UPDRS III, ESS, RBD, RLS, HAMA, and HAMD as independent variables, we performed stepwise linear regression analysis to identify the risk factors of the PSQI global score. e results showed that the included variables H&Y, UPDRS-III, RLS, and HAMD positively correlated with the PSQI global score (P ≤ 0.001, P ≤ 0.001, P � 0.049, and P ≤ 0.001, respectively), indicating that they are positive risk factors for the dependent variable. e R 2 was 0.577, and the adjusted R 2 was 0.541 in the model of regression (Table 4).

Discussion
e present cross-sectional study investigated the prevalence, clinical characteristics, and associated factors of PNS in PD patients and control subjects in a Chinese population. A wide range of evaluation methods were used to assess various factors that potentially influence PNS. Our findings did not show significant differences in demographic factors but showed significant differences in clinical symptoms between PD patients with and without PNS, indicating that PNS was associated with many variables. We also confirmed significantly higher prevalence of RBD, RLS, and depression in PD-PNS subjects than in age-matched and sex-matched nPD-PNS subjects.
e findings also showed significant differences in clinical symptoms between PNS patients with and without PD. e prevalence of PNS was 55.45% in patients with PD in the present study, which is higher than that reported in other studies [22]. is difference may be attributed to differences in the methodology and questionnaire interpretation. PD-PNS subjects showed specific clinical parameters. PD-PNS subjects tended to have higher UPDRS-I, UPDRS-III,  Parkinson's Disease 5 HAMA, and HAMD scores compared with those of PD-nPNS subjects. UPDRS-I and UPDRS-III represent diseaserelated psychological conditions and the disability of PD subjects, respectively. ese findings suggest that mental and motor impairment, together with anxiety and depression, are associated with PD-PNS. RLS is a common disorder characterized by a convincing urge to move the lower limbs accompanied by unpleasant sensations and symptoms that are aggravated during rest and alleviated by activity [23]. e association between RLS and Parkinson's disease (PD) is not clear, but dopaminergic hypofunction in the central nervous system is present in both diseases [24,25]. e prevalence of RLS in PD patients ranges from 8.41% to 34.85% in China and is about 15% in the world. e present study found that 23.64% of total PD patients had RLS, which is consistent with a previous study [23,26]. e PD-PNS group included a significantly higher percentage of RLS patients (33.79%) than PD-nPNS group (12.24%), which indicates that RLS affects the quality of nighttime sleep in PD patients. e prevalence of RLS in the general adult population in Asia is reported to be 1-4% [27,28], compared to approximately 7-10% in Europe and the United States [17]. Although RLS is a possible preclinical marker of PD or an early clinical feature of PD, a recent study has shown that the prevalence of RLS in PD patients is not significantly different from the general population, which contradicts the epidemiological association [25]. However, in our study, the proportion of RLS patients in PD-PNS group (33.79%) was significantly higher than nPD-PNS group (6.4%), indicating that these two PNS groupsmay not share the same mechanism, supporting the difference in RLS between PD patients and general populations.
RBD is characterized by the loss of normal muscle atonia during REM sleep. Patients often experience violent dreamenacting behaviors, which leads to disturbed sleep and potential injuries to themselves and their bed partners. Polysomnography (PSG) remains the diagnostic gold standard, but the diagnosis of probable RBD may be made based on clinical judgment or validated questionnaires [7]. RBD is an early sign of neurodegenerative disease, with a conversion rate from RBD to PD of 18.6-65.0% [29]. e association between PD and RBD can be explained as neurodegeneration in certain brainstem structures at Braak stage 1-2 [23]. It was reported that the prevalence of RBD in PD patients ranges from 22.2% to 60.0% in China [24]. e prevalence of RBD in the entire sample of PD patients in the present study was 27.27%, as judged by the RBD-SQ. ere  was no significant difference in the prevalence of RBD between PD-PNS and PD-nPNS groups, but significant differences between the PD-PNS and nPD-PNS groups were observed. RBD correlated with the PSQI global score and sleep disturbance, but it was not mainly associated with the PSQI global score in the final model. erefore, RBD is more likely to occur in PD patients and affect those patients' nighttime sleep. Excessive daytime sleepiness (EDS) was defined as an inability to maintain wakefulness and alertness during the major waking episodes of the day that resulted in periods of an irrepressible need for sleep or unintended lapses into drowsiness or sleep [7]. e present study used ESS to evaluate and calculate the prevalence of EDS. ere were no significant differences between PD-PNS and PD-nPNS groups or PD-PNS and nPD-PNS groups. Other studies reported that the prevalence of EDS was higher in PD patients than in the general population [30,31]. e present study finally indicated that ESS was associated with PQSI scores and some of PQSI components, including subjective sleep, sleep duration, sleep efficiency, and sleep disturbance, but was not the main factor associated with PNS in PD patients.
Multiple regression analysis showed that H&Y stage, UPDRS-III, RLS, and HAMD were the main factors associated with PNS in PD patients. ese findings suggested that the disease severity, motor impairments, RLS, and depression were the main factors affecting nighttime sleep quality of patients with PD. PD-PNS group had significantly higher UPDRS-Iand UPDRS-III scores, especially UPDRS-III, which represents the severity of motor symptoms. UPDRS-III was one of the main factors associated with PQSI scores in the final analysis, suggesting that motor symptoms have a significant positive effect on PNS; the worse the motor symptoms, the worse the nighttime sleep quality. Motor symptoms have been considered to be related to sleep disorders and cause nocturnal problems in other studies [32]. In a study by Suzuki et al., patients with PD-related sleep disorders had higher MDS-UPDRS part II, III, and IV scores. Notably, they found a significant link between the clinical motor subtypes (tremor dominant or postural instability and gait difficulty) and sleep-related symptoms [32].
RLS, correlating with the PSQI global score and daytime dysfunction in the present study, was also one of the main factors associated with the PSQI global score. PD patients with RLS exhibited higher PSQI score with poorer daytime dysfunction than PD patients without RLS and tended to manifest PNS. However, we did not find impacts of RBD and EDS on PNS in this study or find the difference in prevalence of RBD and EDS between the PD-PNS and PD-nPNS group.
is result is consistent with other studies [33,34]. Some studies pointed out that treatment for RLS could improve sleep quality and quality of life [35]. However, in other studies, the results have been contradictory or complementary [36]. In another study, researchers believed that RLS had less impact on sleep problems than EDS and RBD, and they explained that the results were influenced by the range of severity of RLS [32].
Among the psychiatric symptoms of PD, depression is the most common, followed by anxiety, or both coexisting. HAMD and HAMA were applied to assess depression and anxiety, respectively, in this study. We found that patients in PD-PNS group had significantly higher HAMD and HAMA scores than PD-nPNS group. In PD-PNS group, only the HAMD scores were significantly higher than those of nPD-PNS group. Both HAMD and HAMA were correlated with PSQI global score and subjective sleep, sleep latency, sleep latency, sleep efficiency, sleep disturbance, and daytime dysfunction. In the present study, final analysis revealed that depression, rather than anxiety, was one of the three main factors significantly associated with PNS. Previous studies have also found that depression or/and anxiety are important risk factors for poor sleep quality in patients with PD [37,38]. Contrary to our results, however, Menza and Rosen reported that depression did not significantly affect any of the sleep quality variables [39].
In addition, we found that PNS subjects had lower subjective sleep quality, shorter sleep latency, poorer daytime function, and fewer sleep medications. Night sleep quality in patients with PD may be affected by three main symptoms, namely, increased nocturia, nightmares, and pain. ese symptoms are also common nonmotor symptoms in patients with PD [40], suggesting that some nonmotor symptoms may affect the quality of nighttime sleep. Similar views have been noted in other studies that have shown significant correlations between nonmotor symptoms and PD-related sleep problems in PD patients but have not emphasized nighttime sleep problems [41].
In summary, PNS is common in patients with PD, with three main risk factors of severe motor symptoms, RLS, and depression. e differences in sleep pattern between PD-PNS and nPD-PNS patients spotlighted the fact that PD-PNS subjects had higher prevalence of RBD and RLS, higher scores of HAMD, lower subjective sleep quality, shorter sleep latency, poorer daytime function, and fewer sleeping medication. To our knowledge, the present study is a rare attempt to characterize PNS in patients with PD and to differentiate PNS between PD patients and the general population.
Considering the impact of PNS on patients' quality of life, physicians caring for PD patients should pay attention to nighttime sleep quality and actively prescribe appropriate medications for treatment, especially in patients with RLS, severe motor symptoms, and higher HAMD scores, who are at high risk for PNS.

Limitation.
Our study was designed as a questionnairebased interview investigation. Although it had its own advantages as we mentioned in the manuscript, it is to some degree lacking in strong objective data such as polysomnography or mean sleep latency tests, which limited the objectivity and accuracy of the findings. In addition, patients needed enough mental and physical condition to complete the questionnaires.