Parkinson’s disease (PD) is the most common movement disorder and second-most common neurodegenerative disorder in the world after Alzheimer’s disease, affecting >1% of the population aged ≥65 years and expected to double in prevalence by 2030 (Aarsland et al., 2021). PD is pathologically characterized by dopaminergic neuronal loss in the substantia nigra pars compacta and the presence of intracellular inclusions containing α-synuclein aggregates (Armstrong and Okun, 2020; Liu et al., 2022). Considering that the progression of PD can span decades, it has adverse and profound consequences for patients, caregivers, and the society (Bloem et al., 2021).

Apart from cardinal motor features, such as bradykinesia (slowness of movement), rigidity, and resting tremor, PD is also associated with a heterogeneous spectrum of non-motor symptoms that contribute greatly to the overall disease burden (Schapira et al., 2017; Aarsland et al., 2021). Notably, fatigue is a common but less visible symptom of PD and affects approximately half of all patients (Kluger, 2017), which is often considered by patients with PD to be one of the most disabling symptoms affecting their daily activities and quality of life (Stocchi et al., 2014). As demonstrated by the growing number of research articles, interest in the subject is increasing. Furthermore, expert opinions on raising awareness on fatigue have been published (Lazcano-Ocampo et al., 2020). Considering that fatigue has diverse clinical manifestation, large clinical heterogeneity with many confounding factors, and lacks of clear pathogenesis and treatment to date, the understanding of fatigue in PD is insufficient. Although previous studies on fatigue in PD have mostly focused on the clinical features and factors associated with PD, fatigue has also been shown to be associated with other common non-motor symptoms of PD, including cognitive impairment, depression, apathy, anxiety, autonomic dysfunction, daytime somnolence, and sleep disturbances, which adversely affects patients’ quality of life (Stocchi et al., 2014; Fu et al., 2016; Kluger et al., 2017; Siciliano et al., 2018; Ongre et al., 2021; Ou et al., 2021; Zhou et al., 2021; Béreau et al., 2022). However, the results of these studies are not always consistent. For example, Ongre et al. reported that a higher level of fatigue was associated with better cognitive functioning, whereas Kluger et al. and Siciliano et al. found that fatigue is associated with cognitive decline. Moreover, most published investigations were limited in size (sample size <500), and knowledge regarding the longitudinal progression of fatigue (Ongre et al., 2021; Ou et al., 2021) in PD is currently limited. Furthermore, the potential influence of the presence of concomitant fatigue in PD on the rate of progression of other phenotypic features has not been characterized.

In the light of these discrepancies and gaps in the literature, we conducted a longitudinal cohort study of patients with PD with a 2-year follow-up period. The aims of this study were to assess the prevalence and evolution, clinical characteristics, correlates and predictors of fatigue, as well as to investigate the influence of comorbid fatigue on the longitudinal changes in motor and non-motor symptoms in a large cohort of patients with PD.

This longitudinal cohort study systematically investigated the prevalence, evolution, and clinical correlates of fatigue among PD patients as well as assessed differences in disease progression across PD fatigue subgroups. Overall, approximately one-third of patients included in our study experienced fatigue. We found that PD patients with fatigue were more likely to be older, have a longer disease duration and higher baseline LEDD, and manifested more severe motor and non-motor phenotypes than PD patients without fatigue. We also found that increased disease severity, presence of the PIGD subtype, EDS, and wearing-off were associated with fatigue in patients with PD. To the best of our knowledge, this is the first study to assess longitudinal progression across different fatigue subgroups of PD patients. We observed significant subgroup-level differences in the progression of motor and non-motor symptoms. Our study provides evidence for the longitudinal progression of clinical symptoms in different fatigue subgroups of patients with PD and will facilitate the clinical management of fatigue.

We found that the overall frequency of fatigue in patients with PD was 36.8%, which is consistent with the study of Stocchi et al. (2014) and provides further evidence for fatigue as one of several common non-motor symptoms in patients with PD. The prevalence of fatigue in previous studies (Friedman et al., 2016; Siciliano et al., 2018) has been found to range from 33 to 70%, and the wide range of values may reflect differences in the study designs, especially the use of different fatigue rating scales and definitions of fatigue as well as the populations tested. We also found that the frequency of fatigue increased with the severity of PD, which is consistent with findings from previous studies (Stocchi et al., 2014; Fu et al., 2016; Ou et al., 2021). However, the results of the meta-analysis (Siciliano et al., 2018) showed that the prevalence of fatigue was not moderated by the disease severity.

Our study found that PD patients with fatigue were older, had longer disease duration, and had higher LEDD than those without fatigue. Additionally, PD patients with fatigue showed significantly greater disease severity scores and a higher prevalence of all motor and non-motor symptoms than those without fatigue. These findings indicate that fatigue exacerbates the symptoms and burden of PD patients and are in agreement with previous studies (Stocchi et al., 2014; Fu et al., 2016; Siciliano et al., 2018).

Additionally, we investigated the factors associated with fatigue in order to improve the management of fatigue in patients with PD. In the multivariate logistic regression for the entire cohort, we found that the total UPDRS score, NMSS score and the presence of the PIGD subtype, EDS, and wearing-off were significantly associated with fatigue in patients with PD. Notably, fatigue has been reported to be reduced by dopaminergic drugs, such as levodopa (Lou et al., 2003), suggesting that dopaminergic deficits are involved in the development of fatigue in PD, which is further supported by our finding that more severe motor disabilities were associated with fatigue. However, the non-persistent nature of fatigue observed in many studies (Ongre et al., 2017; Ou et al., 2021; including our study) also suggests that dopaminergic mechanisms only partially contribute to the development of fatigue in PD. In fact, neither LEDD, Hoehn and Yahr stage nor disease duration were associated with fatigue in our study, providing indirect support for the hypothesis that fatigue in PD may result from the disruption of nondopaminergic pathways (Kluger, 2017; Siciliano et al., 2018). This implies that fatigue is independent of dopaminergic treatment in PD and instead an independent non-motor symptom of PD. Although “significant” in UPDRS and NMSS scores associated with fatigue, the odds ratios were close to 1.0 and confidence intervals were narrow, it is likely to be a very small association that may have limited clinical significance. Because UPDRS and NMSS scores are continuous variables, differences in scores between different patients can be large or small. Patients with severe symptoms can roughly be considered as more prone to fatigue than patients with very mild symptoms in clinical practice.

Our study also revealed a relationship between EDS and fatigue, which is consistent with the results of previous studies (Stocchi et al., 2014; Siciliano et al., 2018). Nevertheless, there are a few inconsistent findings in our study regarding factors associated with fatigue that should be noted. For example, some studies have shown that poor quality of life, sleep disorders, and depression (Ongre et al., 2021; Ou et al., 2021; Béreau et al., 2022) are related to fatigue, whereas we did not find an association between fatigue and the patients’ PDQ-39 score, PDSS score, the presence of pRBD, or depression as a whole. In order to determine whether the results of related factors are different due to different disease stages, we performed subgroup analyses of correlated factors grouped by disease duration and Hoehn and Yahr stages. We found that the total UPDRS score, NMSS score, PIGD subtype, and the absence of FOG were significantly associated with fatigue in PD patients with disease duration less than or equal to 3 years. The PDSS score was found to be associated with fatigue in patients with a disease course longer than 10 years; however, the OR value was close to 1 with little effect. No relationship was found between fatigue and PDQ-39, pRBD, or depression in each subgroup analysis, which may be ascribed to different inclusion criteria and confounding factors considered in previous studies (Ongre et al., 2021; Ou et al., 2021; Béreau et al., 2022). Our findings also differed from those of previous studies in that we found that the presence of the PIGD subtype and wearing-off were associated with fatigue. The link between fatigue and wearing-off further implicates that dopaminergic factors—either via direct dopaminergic neurotransmission or via indirect mechanisms—are key for the development of fatigue (Franke and Storch, 2017). Unfortunately, to date, there is no effective treatment for fatigue in patients with PD (Elbers et al., 2016), and rasagiline was only proposed as “possibly useful” (Seppi et al., 2019), deep brain stimulation may help improve fatigue in patients with PD (Lazcano-Ocampo et al., 2021). Therefore, PD patients with fatigue may benefit from the management of motor symptoms, wearing-off, and EDS because they are all associated with fatigue in these patients.

Our study also demonstrated that nearly all PD motor and non-motor symptoms significantly deteriorated over time, and that the occurrence of fatigue in PD was associated with worse baseline findings and faster progression rates, which are consistent with findings from previous studies. For example, data from the Oxford Parkinson’s Disease Centre Discovery cohort demonstrated that almost all PD motor and non-motor symptoms measured in the study significantly increased over time (Liu et al., 2021). Other studies have also confirmed the longitudinal progression of both motor and non-motor symptoms, including autonomic dysfunction, EDS, and cognitive impairment (Amara et al., 2017; Simuni et al., 2018; Stanković et al., 2019; Myers et al., 2022). It should be noted that a subset of symptoms have shown stability, relief, and improvement in longitudinal studies (Cilia et al., 2020; Simuni et al., 2022). However, despite these inconsistencies, most studies have reported an overall deterioration of motor and non-motor symptoms, consistent with our data. In the subgroup analysis of our study, the persistent-fatigue group showed significantly greater progression in most symptoms, in comparison with the never-fatigue group. In comparison with the never-fatigue group, the nonpersistent-fatigue group showed more rapid progression in the freezing of gait, pRBD and total SCOPA-AUT score. However, other clinical features (including motor symptoms, wearing-off, dyskinesia, EDS, depression, quality of life and cognitive symptoms) were not significantly different in the evolution of these symptoms between these two groups. This reflected indirectly that the prognosis of patients with non-persistent fatigue was better than that of patients with persistent fatigue in the overall cohort, which further proves that fatigue can aggravate disease progression. Moreover, this suggests that the improvement of fatigue may alleviate motor and other non-motor symptoms to some extent. The results also provide indirect support for the hypothesis that progression during early PD in the persistent fatigue group may be faster than that in the non-persistent fatigue group.

The relationship between treatment and relief or aggravation of fatigue has not been thoroughly studied in the past. The factors influencing fatigue are quite numerous, and may include (in addition to the drugs regime) the amount of physical exercise and physical labor, emotional aspects, and diet, among others (Fu et al., 2016; Ongre et al., 2017; Lin et al., 2021; Ongre et al., 2021). A previous study analyzed factors associated with fatigue improvement, and found that patients with fatigue at baseline received higher doses of dopaminergic medication during follow-up than those without fatigue. Moreover, the study showed that the improvement of fatigue could not necessarily be attributed to changes in disease severity or other non-motor symptoms (such as depressive symptoms, sleep disorder, apathy, or cognitive impairment), and suggested an improved efficacy of dopamine agonists over levodopa in PD patients with fatigue (Ongre et al., 2017). However, the authors of this study did not take into account some additional confounding factors. Rigorous prospective cohort studies, with very strict design, precise group allocation, and with more confounding factors being considered, are needed in the future to further explore the relationship between the drug regime or other alleviating/aggravation factors, and the evolution of fatigue. Proposed pathophysiologic mechanisms of fatigue currently include increased circulating proinflammatory cytokines, dysfunction in nigrostriatal and extrastriatal dopaminergic pathways, involvement of nondopaminergic (particularly serotonergic) pathways, executive/prefrontal pathology, and involvement of the autonomic nervous system (Kluger, 2017; Lazcano-Ocampo et al., 2020). Further longitudinal studies are needed to corroborate our results and to investigate the specific neurobiological pathways related to fatigue in PD. To the best of our knowledge, no other studies have evaluated the longitudinal progression in PD patients with and without fatigue. The above findings suggest that screening for fatigue symptoms at baseline could improve the estimation of the subsequent disease trajectory, enabling better individualized treatment, information, and support to be offered to patients and their caregivers.

This study had some limitations. First, although the 2-year follow-up period was relatively short, we plan to address this by continuing to follow up our cohort of patients. Second, our study involved heterogeneity in clinical data collection and the data quality obtained from different investigators. Nevertheless, all researchers involved in the study received standardized and unified training to ensure consistency and quality of the data as much as possible. Third, PFS-16 was developed to assess a single construct reflecting the physical aspects of fatigue in patients with PD and to measure both the presence of fatigue and its severity. But this scale may not adequately reflect clinically significant non-physical aspects of fatigue and does not define fatigue. However, PFS-16 is a brief and easily completed scale developed specifically for PD, and meets the criteria for the designation of “recommended” as defined by the Movement Disorder Society for screening fatigue (Friedman et al., 2010). Fourth, the study was lacking in biomarkers for the characterization of fatigue. Future studies combining clinical information and additional variables (i.e., genetics, neuroimaging markers, and other biomarkers) are warranted.

Increased disease severity, presence of the PIGD subtype, EDS, and wearing-off were associated with fatigue in patients with PD. Significant subgroup-level differences were observed in the progression of motor and non-motor symptoms across different fatigue subgroups of PD patients.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The studies involving human participants were reviewed and approved by Ethics Committee of Xiangya Hospital of Central South University. The patients/participants provided their written informed consent to participate in this study.

XZ: research project conception, organization, execution, statistical analysis design and execution of the study, patients’ enrollment and follow up, and writing of the first manuscript draft. YX, TS, YZ, and HP: research project execution and patients’ enrollment and follow up. QX, YC, QS, XY, JG, and BT: research project organization, execution, and manuscript review and critique. XW: statistical analysis design, execution, review and critique of the study and manuscript review and critique. LL and ZL: research project conception, organization, execution, and manuscript review and critique. All authors contributed to the article and approved the submitted version.

This study was supported by the Hunan Innovative Province Construction Project (Grant Nos. 2019SK2335 and 2021SK1010), Natural Science Foundation of Hunan Province (Grant No. 2022JJ40843), National Key Research and Development Program of China (Grant Nos. 2021YFC2501204 and 2016YFC306000), National Natural Science Foundation of China (Grant No. 82001359), and Scientific Research Project of Hunan Provincial Health Commission (Grant No. 202203074637).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.