Background
Currently, early and accurate diagnosis of PD remains a challenge. Recent studies have increasingly focused on the non-motor symptoms (NMS) of Parkinson’s disease (PD) [
1]. Sleep disorders are considered to be one of the most prevalent NMS of PD and cover a wide range of conditions, such as rapid eye movement (REM) sleep behavior disorder (RBD), insomnia, restless legs syndrome (RLS), obstructive sleep apnea (OSA), and excessive daytime sleepiness (EDS) [
2]. RBD is a reliable clinical predictor of neurodegenerative disease. Previous studies have shown that RBD is one of the earliest sign of alpha-synucleinopathies, characterized by an intermittent loss of normal muscle tone that happens mainly in the REM sleep phase [
3]. In addition, EDS is considered to be one of the earliest NMS of PD and a significant cause of disability [
4,
5]. The main characteristic of EDS is the incapacity to remain awake and attentive throughout the waking hours of the day, which causes unintentional lapses into sleepiness or sleep and adversely affects safety and quality of life [
6]. Sleep disorders have become a focal point of research in the field of neurodegenerative diseases. However, the precise mechanisms by which they reflect or influence the neurodegenerative process in PD remain incompletely understood. Although sleep disorders are not exclusive to PD and lack specificity across neurodegenerative conditions, they represent one of the most common NMS in PD. They sometimes precede motor symptoms in the early stages of PD and, as the disease progresses, can have a profoundly negative impact on patients’ physical function, quality of life, and overall health [
7], potentially exceeding the disabling capacity of motor symptoms. Furthermore, it is important to note that sleep disorders not only serve as an early symptom of PD but also significantly contribute to the disease’s progression. By assessing patients’ risk before the typical motor symptoms of PD manifest, we can facilitate early intervention, thereby opening up new avenues for enhancing the overall health and quality of life of PD patients. Currently, the potential of sleep disorders as predictors of neuronal damage remains largely unexplored. Establishing this connection provides a novel perspective on potential alterations within the nervous system associated with early NMS.
As one of the most promising biomarkers of axonal injury, elevated blood or cerebrospinal fluid (CSF) concentrations of NfL in PD individuals were found to be associated with greater PD severity, reduced life expectancy, and an increased susceptibility to motor and cognitive impairment [
8,
9]. In recent years, an increasing body of research has shown a robust correlation between serum NfL and CSF NfL and that the diagnostic accuracy of serum NfL is the same as that of CSF NfL in the discrimination between PD and atypical parkinsonian disorders (APD) [
10]. Whether sleep disorders could increase the level of serum NfL has not been fully investigated.
Our research aims to determine whether sleep disorders can predict elevated serum NfL levels, aiming to establish a correlation between PD clinical symptoms and the pathological alterations stemming from neuronal damage. Early detection and diagnosis of PD are critical for effective disease management. By identifying sleep disorders early and monitoring NfL levels, we can evaluate patients’ susceptibility before noticeable PD symptoms become prominent, enabling early intervention. In addition, the use of NfL as a biomarker for monitoring sleep disorders, tracking disease progression and evaluating treatment efficacy can help to make timely adjustments to treatment regimens in order to slow down disease progression. This approach not only optimizes the treatment and management of PD patients but also provide new therapeutic targets for specific sleep disorders, opening up new possibilities for improving the overall health and quality of life of individuals with PD. We have conducted baseline and longitudinal analyses of individuals belonging to three distinct categorical groups (early PD, prodromal PD and HCs) to examine the baseline and longitudinal connections between sleep disorders and serum NfL levels. Given the gender differences in sleep disorders, we further conducted a subgroup analysis stratified by gender.
Discussion
The present study revealed that different RBD and ESS items in three diagnostic groups were strongly correlated with serum NfL levels. Moreover, significant disparities were seen in the baseline levels of serum NfL across the three diagnostic groups, with the PD group having the highest serum NfL levels. The findings of our research indicated that serum NfL levels in PD individuals were influenced by RBD and EDS behaviors, where there were gender differences. In prodromal PD participants, significant associations were seen between some specific RBD behaviors and ESS items and higher serum NfL levels. In HCs, we only discovered a significant correlation between possibilities of daytime sleepiness when sitting more quietly after lunch and more elevated serum NFL levels, which was still significant in the male subgroup. These findings indicated that sleep disorders might be a marker of the severity of neurological damage and PD progression. Good sleep may have neuroprotective effects, thus attention should paid to sleep management for PD prevention.
Clinical studies on the pathogenesis of RBD have suggested that the potential cause of RBD may include the degeneration or dysfunction of the brainstem circuits responsible for regulating REM sleep [
20,
21]. Aside from that, previous research has demonstrated that RBD falls under the category of “diffuse malignant phenotype”, which designates a more severe subtype of synaptic neuropathy with more pronounced dopaminergic impairments [
22]. Therefore, we can infer that PD patients with the RBD subtype are more severely affected, which is also supported by pathological autopsy studies [
23,
24]. Nevertheless, serum NfL has not been adequately studied as a promising biomarker for monitoring the course of the disease and its correlation with RBD. Serum NfL is considered a potentially useful biomarker for the RBD subgroup in a study on the associations of serum NfL and glial fibrillary acidic protein with the RBD subtype of PD, which is consistent with our results [
25]. Another study exploring the relationship between plasma NfL levels and cognitive function in PD individuals also observed significantly higher plasma NfL levels in individuals with RBD compared to those without RBD [
26]. In contrast, a newly published study on the correlation of non-motor markers and neuronal damage in patients with early Parkinson’s disease found no link between plasma NfL levels and the presence of an RBD [
27]. This discrepancy might result from the bias in the assessment of sleep traits and variations in sample sizes.
NfL can serve as a marker of axonal injury because it is stable in axons under physiologic circumstances [
28]. In contrast, following axonal damage, neurofilaments diffuse into the CSF and then they are discharged into blood circulation through the arachnoid villi [
29,
30]. Significantly, a longitudinal study involving 40 individuals with chronic insomnia disorder (CID) revealed a correlation between heightened levels of serum NfL (indicating functional and structural damage to neurons, axons, and glial cells) and both subjective and objective sleep parameters among CID individuals [
31]. A possible mechanism linking sleep disorders with neuronal damage has been proposed by previous studies. Since a good night’s sleep can enhance the clearance of potentially neurotoxic waste products from the glymphatic system that accumulate during wakefulness, sleep disorders can increase the production of reactive oxygen species (a kind of toxic metabolites) which damage neurons [
32]. Therefore, as a type of sleep disorders, RBD may affect cerebrospinal fluid flow and promote glymphatic system dysfunction, which might inhibit the clearance of metabolic waste and eventually cause neuronal damage [
33].
There are several potential mechanisms underlying the associations between EDS-related brain structural alterations and neurodegeneration. Firstly, PD progression may be accompanied by degeneration of the neurons that affect wakefulness and sleep, which raises the possibility of sleep disorders (including EDS) in PD patients. Previous studies have found that CSF production correlates with changes in circadian rhythms, and human CSF production peaks during sleep after midnight, as measured by magnetic resonance imaging [
34,
35]. EDS is one of the common symptoms of circadian rhythm disturbance and it may predate Lewy pathology (LP), a marker of PD pathogenesis [
36]. The relationship between EDS and widespread topographic LP expansion might further support the early finding of an association between EDS and PD [
37]. Since the clock gene is central to circadian rhythms, its altered expression in PD disrupts the 24-hour cycle, leading to impaired nighttime sleep and further impaired function of the lymphatic system [
38].
Although the mechanisms underlying the association of EDS with elevated serum NfL levels have not been elucidated, currently there are two reasons to account for the association. Firstly, evidence showed that EDS might be an indication of inadequate clearance of neurotoxic metabolic by-products during sleep or neurodegeneration in areas associated with the maintenance of the wake state [
33]. In addition, a previous study has shown that as people with EDS get older or enter late middle age, their cortical thickness decreases [
39]. A community-based longitudinal study (the MAP cohort) found that lower inferior lateral orbitofrontal cortex and inferior frontal orbital gyrus grey matter volumes were associated with greater sleep fragmentation (including RBD) in older community adults [
40]. And a short-term follow-up study also found that increased plasma NfL was associated with cortical thinning [
41]. Therefore, we speculate that EDS is associated with increased serum NfL levels, possibly via EDS-related brain structural alterations. This is consistent with previous findings on the correlation between EDS and CSF biomarkers of inflammation and axonal integrity in cognitively unimpaired older adults [
39]. The correlation between EDS and NfL in HC group can thus be explained. Secondly, EDS is thought to be a consequence of OSA [
22,
42]. OSA induces cerebral hypoperfusion, leading to increased oxidative stress and subsequently neuronal and axonal damage [
43]. Upon axonal damage, the release of NfL occurred first into CSF and then into the blood, leading to increased levels of NfL in the serum [
44].
The association between sleep disorders and serum NfL levels was investigated using a cross-sectional and prospective follow-up study design in an effort to improve the study’s scientific validity. Moreover, unlike CSF, which must be collected via a problematic process that frequently involves a puncture, serum biomarkers are simple to acquire.
This study represents the first comprehensive exploration of the association between sleep disorders and serum NfL levels in individuals experiencing early and prodromal PD, thus addressing a notable research gap in this domain. Moreover, the cross-sectional and prospective follow-up design employed in this investigation distinctly bolsters the scientific validity and effectiveness of our findings. Our study offers a more nuanced depiction of the relationship between sleep disorders and serum NfL levels. Compared with biomarkers that require complex procedures such as cerebrospinal fluid puncture for collection, serum samples are simple to obtain, non-invasive, and easily applicable in a wide range of clinical and research settings. Despite these advantages, the study faced some limitations that warrant consideration in future studies. Firstly, assessments such as RBDSQ and ESS rely on self-reported sleep questionnaires, which are subject to various subjective factors. Therefore, future studies should incorporate objective indicators such as nocturnal polysomnography and pathological evidence. Secondly, potential missing data in our follow-up assessments may affect the reliability of our results. Thus, a wider cohort study is needed in the future with measures to minimize missing data and ensure the stability and credibility of the results. Lastly, it is important to note that our blood samples were stored at -80℃ for a period, and previous related studies have suggested that this may lead to alterations in concentration [
45]. Future studies should control for this variable. Finally, the sample size of participants with both EDS and pRBD was small in this study. Therefore, it is necessary to validate our findings in a larger sample size in order to enhance the generalisability and applicability of the findings.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.