Parkinson's disease (PD) is a common chronic neurodegenerative disease that places an enormous material and mental burden on patients (1, 2). Although PD is characterized by motor symptoms, a broad spectrum of non-motor symptoms (NMS), including sleep disturbance, neuropsychiatric symptoms, and autonomic dysfunction, can be identified in people with PD (3). Sleep disturbances in PD include insomnia, rapid eye movement (REM) sleep behavior disorder, sleep-disordered breathing, restless legs syndrome, and circadian dysregulation, and affect a high percentage of people with PD (4, 5). The decline in sleep quality is a very common complaint among PD patients. One possible reason for the high prevalence of sleep complaints is that neurodegenerative diseases disrupt the circadian rhythm system at a faster rate than normal aging (6). Furthermore, there are strong correlations between sleep disturbances and other non-motor as well as motor symptoms such as nocturia, dystonia, and pain (7). Worse sleep quality has recently been reported to aggravate motor and non-motor symptoms in PD (8, 9). Many sleep problems may occur in the early stages of the disease and worsen as the disease progresses (10, 11), necessitating proper vigilance.

In the past, the evaluation of sleep quality relied mainly on a series of sleep scales. Many studies have used the Pittsburgh Sleep Quality Index (PSQI) and the Parkinson's Disease Sleep Scale (PDSS) for sleep assessment in PD patients (12). Although these instruments are rapidly administered and efficient, a detailed understanding of the sleep structure of patients with sleep complaints is still lacking. Currently, video-polysomnography (v-PSG), which is considered to be the most valuable method of diagnosing sleep disorders, is widely applied (13). PSG can not only quantitatively analyze the sleep structure but can also sensitively record the occurrence of specific sleep events. Previous studies using PSG revealed that, compared with healthy controls, patients with PD have shorter total sleep time, lower sleep efficiency, and increased sleep latency and wakefulness after sleep onset (14–16). However, anti-PD medication has an inevitable effect on sleep structures. The use of MAO-B inhibitors is related to sleep stages 2 and 3, while sleep stage 1 can be predicted by dopamine agonists (17).

In this respect, a cohort of drug-naive patients is essential to bringing new insights into the field of the non-motor spectrum of PD. This study aimed to investigate changes in the sleep structure in drug-naive PD patients with poor subjective sleep quality compared to those with good sleep quality using PSG technology and to explore the association between clinical features and sleep structure.

A total of 44 patients were enrolled in this study, including 24 (54.5%) in the good sleeper group and 20 (45.5%) in the poor sleeper group. We observed that poor sleepers had severe non-motor symptoms and poor quality of life. They had longer wake-up times and lower sleep efficiency during night sleep according to PSG results. Furthermore, we found significant associations between nocturnal arousal events and motor symptom scores in all drug-naive PD patients. Our findings provide new insights into sleep problems in Parkinson's disease.

Our study evaluated the prevalence of three major sleep disorders in PD, including RBD (36.4%), PLMD (27.3%), and OSA (27.3%). The results are consistent with the frequencies reported in previous studies (29, 30). We did not find significant differences in PSQI or PDSS scores between the groups with and without these three sleep disorders. The results are consistent with those reported in some studies (31, 32), but contrary to other studies (33, 34). Differences between the results obtained in this study and other studies may be explained by the possible coexistence of other sleep problems in different PD cohorts and by the fact that only one or two questions on the PSQI and PDSS scales are strongly associated with specific types of sleep disorders. A previous study that tested the criterion validity of the PSQI in diagnosing the sleep disorders mentioned above found that it was not highly accurate for predicting the existence of these disorders (35). Thus, the PSQI is more suitable for screening for the presence of sleep alterations than for diagnosing specific types of sleep disorders. In our study, we divided the patients into two groups based on their PSQI scores. We compared the distribution of sleep disorders in good and poor sleepers, and no significant difference was found.

Changes in PSG in our study focused on drug-naive PD patients. After considering the distribution of different sleep disorders, we wanted to investigate the features of disturbed objective sleep parameters in patients suffering from reduced subjective sleep quality. Previous studies have reported that, compared with healthy controls, there were significant differences in sleep architecture patterns in PD patients (16); these differences include a decrease in TST, SE, N2 sleep percentage, and REM sleep percentage and an increase in WASO, N1 sleep percentage, REM sleep latency, and PLMI. Our study compared the sleep structure between good and poor sleepers in drug-naive patients. We observed that as patients' sleep quality began to decline, the main changes were longer wake-up times after falling asleep and lower sleep efficiency. The differences in the findings may be explained by the progression of the disease and the use of anti-PD medication. In terms of neuropathologic evidence, sleep–wake disturbances in PD may reflect disruption of the neural circuitry that controls circadian rhythms (6). Altered peripheral clock gene expression and decreased function of the hypothalamic suprachiasmatic nucleus (SCN) are believed to be responsible for reduced melatonin output and sleep–wake disruption in PD patients (36, 37). Our results also revealed significant associations between the percentage of REM sleep and the H-Y stage in poor sleepers. Sleep-controlling neurons have recently been reported to be sensitive and vulnerable to α-synuclein (38), and the spread of pathology to the coeruleus/subcoeruleus complex may further disrupt REM sleep (39–41). With regard to anti-PD medication, one study has reported that higher amounts of dopaminergic medications, especially L-dopa, were related to sleep fragmentation and insomnia (42). These changes have been described as a progressive destructuring of the sleep structure (43). The mechanisms underlying sleep disturbances in PD remain speculative. Our observations may be an early objective manifestation of sleep disturbance in original ecological patients.

We consider sleep disturbances an integral part of PD rather than comorbidity. Our research explored the association between sleep parameters and typical clinical characteristics of PD, including the H-Y stage, the UPDRS-III score, the NMSQ score, and the PDQ-39 score. We also conducted an analysis of correlations between disease and PSG parameters, and no significant correlations were found. However, we observed that MAI was positively related to the UPDRS-III score in good sleepers and that WASO increased with the UPDRS-III score in poor sleepers. Both MAI and WASO were nocturnal arousal events, and dopamine may play a crucial role in the close association between these events and motor dysfunction. As we all know, PD is characterized by a progressive loss of dopaminergic neurons in the substantia nigra, and its metabolism and activity are also strongly influenced by the circadian clock (44). On the contrary, sleep dysfunction can disrupt the circadian rhythms that are generated by SCN, and the dysfunction of sleep and peripheral clocks, such as those in microglial and neuronal cells, leads to subsequent changes within and outside of the brain, including neuroinflammation, increased oxidative stress, and reduced metabolic clearance. These outcomes have been proposed to exacerbate Parkinson's disease progression (45–48). Therefore, the increase in nocturnal arousal events may be of great value in predicting the progression of motor dysfunction in PD. Our recent study found that N3 sleep (slow-wave sleep, SWS) is one of the important risk factors in PD patients with hallucinations (49). The deconstruction of the sleep structure is a process affected by many complex factors. The decrease in light sleep (including N1 sleep and N2 sleep) in the early stages of the disease causes more daytime fatigue. This may aggravate non-motor symptoms or quality of life burden in drug-naive PD patients.

Periodic limb movements during sleep (PLMS) have been reported to be associated with greater symptom severity, more subjective sleep disturbance, and decreased quality of life in PD patients (34, 50). We did not observe significant differences in PLMI variation in drug-naive PD patients with different subjective sleep quality. However, our results revealed an association between PLMI and non-motor symptoms in drug-naive patients with poor sleep quality. Evidence has confirmed that there are common pathophysiological pathways between PLMS and PD (51). Nigrostriatal degeneration may promote PLMS, whereas dopaminergic enhancement may control PLMS. In our drug-naive PD cohort, the regression model suggested that increasing PLMI was an independent variable predicting a higher NMSQ score in poor sleepers.

Several methodological considerations should be mentioned. First, although there have been some studies investigating the characteristics of objective sleep parameters in drug-naive PD patients, we conducted a more comprehensive clinical evaluation of these patients and further explored the relationship between sleep parameters and clinical features. Second, our study was based on PSG data. The majority of published studies related to sleep problems used subjective tools to measure sleep quality, while only a few such studies used objective methods. However, patients in our study did not undergo a second PSG examination to eliminate the first-night effect, which should be considered as a methodological and ethnic bias. In addition, we did not include healthy controls with and without sleep disorders such as OSA, RBD, and PLMS. Finally, our cohort consisted of a relatively small number of patients and was recruited in a single center. Further studies should include larger cohorts and conduct longitudinal follow-ups to confirm the correlations between sleep parameters and clinical characteristics.

Our results showed that more than 50% of drug-naive PD patients have poor subjective sleep quality, mainly manifested by increased nocturnal awakening time and decreased sleep efficiency. Poor sleepers have severe non-motor symptoms and poor quality of life. Nocturnal arousal events, including MAI and WASO, may predict the progression of motor dysfunction. We provide new evidence for associations between sleep parameters and clinical characteristics in PD patients before they begin medication interference.

The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author.

The study was approved by the Ethics Committee of the Affiliated Brain Hospital of Nanjing Medical University. All procedures were in accordance with the Declaration of Helsinki.

YJ and LZ conceived and designed the study. LZ, XJ, and JY obtained the funding. YJ, SZ, RG, YW, YC, DL, JZ, XJ, BS, and YP collected the data. YJ, SZ, RG, and YP conducted the data analysis. YJ drafted the manuscript. All authors contributed to the article and approved the submitted version.