The interactions between non-motor symptoms of Parkinson’s disease

KEYWORDS:

• Parkinson’s disease
• non-motor symptoms
• cognition
• mood
• sleep
• autonomic
• sensory

1. Overview

Non-motor symptoms (NMS) are an important, yet underappreciated, feature of Parkinson’s disease (PD). Many of these varied symptoms are central determinants of quality of life. This burden is in large part due to the ineffective management strategies that are rooted in our limited understanding of the pathophysiological mechanisms which underlie the many NMS that manifest in PD. Here, we highlight several common NMS, their respective assessment methods and evidence that they are often inter-related. We then suggest several strategies that may help to interrogate the complexities of NMS interactions. We end by concluding that future research should focus on the relationships between NMS – how they interact, coexist, and/or amplify one another – and whether NMS may also have phenotypic profiles that are useful to clinicians and researchers for future clinical trial selection, prodromal biomarkers profiles, or determining different pathological progression.

2. NMS complexity: the elephant in the room

NMS in PD include a wide variety of symptoms, many of which were initially described by James Parkinson himself [1]. NMS affect nearly the entire body and almost all patients with PD suffers from at least one (and often many more) NMS. These symptoms begin at varying times in the disease process (e.g. from prodromal to late-stage disease) and cross many domains: including affect (e.g. depression, anxiety, apathy), cognition (e.g. attention, memory, hallucinations, dementia, psychosis), sleep (e.g. rapid eye movement [REM] sleep behavior disorder (RBD), excessive daytime sleepiness (EDS), insomnia, restless leg syndrome), the senses (loss of smell, visual disturbances, proprioceptive deficits, pain), and a variety of other autonomic bodily functions (e.g. constipation, urinary urgency, excessive sweating, sexual dysfunction, orthostatic hypotension). Given their diverse nature, it is perhaps not surprising that the specific origins and neuropathobiology underlying the majority of NMS remain enigmatic [2–4], for full review of the pathologic substrates of NM manifestations refer to the study of Jellinger [5]. Furthermore, there are many barriers in this field, including a lack of animal models together with the lack of appropriate biomarkers and clinical scales specifically validated in PD, which currently limit our ability to fully disentangle the underlying mechanisms of different NMS. There is growing evidence that NMS may arise from pathological damage to multiple neurotransmitter systems or to key hubs within the structural backbone of the brain, particularly within the brainstem [6,7], however, the mechanisms by which Lewy pathology spreads through the brain over time have recently been called into question [8]. In summary, NMS represent a complicated and challenging problem for scientists and clinicians alike. More importantly, given that NMS occur frequently in PD, and often present in the earliest stages of the disease, identifying, tracking, and treating NMS represents an important piece toward solving the puzzle of heterogeneity across the spectrum of PD.

2.1. Multiple ways to assess NMS: adding fuel to the fire

Perhaps the most popular approach to this problem has been to isolate and investigate individual NMS in isolation. These studies often focus on ways to characterize the frequency and severity, usually in the form of a questionnaire. While this approach certainly contains merit (e.g. to foster investigations into pathological, behavioral, or neural correlates of each symptom), this strategy tends to ultimately manifest with a series of individually unique, but globally overlapping, measures. As of this writing, we are aware of over 10 different questionnaire-based assessments to quantify anxiety and depression (NPI, HADS, DASS, BAI/BDI, PAS, GDS, STAI, Ham-A/Ham-D) and seven different scales to assess apathy or fatigue (AES-C/AES-S, AS, LARS, MFI, PFS, FSS, PACIT-F). That is, there have been many tools developed to clarify symptom prevalence, but no gold standard measurements have been identified. This problem is especially pertinent given that each idiosyncratic questionnaire tends to uniquely measure variable amounts of utility, validity, specificity, and effectiveness. To address these issues going forward, an IP-MDS task force has been initiated to bring together a panel of experts to review such scales and assess their strengths and limitations and ultimately put forward recommendations for their use in future Parkinson’s research.

Classical holistic tools that measure NMS in a more comprehensive way rather than in isolation include: rater-based questionnaires such as NMSS and PDCS, clinician based assessments such as the MDS-UPDRS part I, and self-reported questionnaires such as the NMSQuest [9]. The development of these tools has improved our ability to quantify the burden of NMS, as well as revolutionized the management of NMS for patient and evaluate the effectiveness of NMS treatment with clinical trials [10]. While many of these more global assessments of NMS provide an overall picture of the NMS severity and burden, the detailed assessment of individual NMS symptoms is compromised. Moreover, both of these prevailing approaches are essentially ignorant to the interactions between NMS, which often co-occur in the same individuals, and thus may end up effectively describing the tip of an immense iceberg.

Indeed, recent research has discovered many cross-symptom interactions when examining a particular NMS. Epidemiological evidence and associations within various research studies has shown many symptoms that coexist (e.g. anxiety and depression), symptoms that relate to one another (e.g. apathy and fatigue), and symptoms that may even contribute to the pathophysiology of other NMS (e.g. visual disturbance and hallucinations). For example, a recent study of de novo PD patients found that individuals with distressing fatigue also had greater sleepiness (ESS), depression (BDI), anxiety (PAS-total), and apathy (AES-total). In addition, a logistic regression model could accurately distinguish 90% of patients with distressing fatigue using scores on the ESS, PAS-episodic, and AES-cognitive [11]. In a similar fashion, Lenka and colleagues [12] considered whether other NMS might provide insight as to why only a subset of patients with PD develop psychosis. This symptom varies in onset (early or late in the disease) and is related to REM sleep behavior disorder, excessive daytime sleepiness, visual disturbances, and the great deal of overlap in neuropychological deficits and abnormalities in cholinergic transmission. Indeed, a significantly higher proportion of patients with early onset psychosis were found to have RBD, EDS, as well as poor scores on FAB and higher intake of anti-cholinergic agents compared to those with late onset psychosis who were taking greater dopamine doses. These works underscore the importance of considering the interactions between NMS as well as dopaminergic and non-dopaminergics medication usage.

2.2. Interactions with Parkinson’s therapeutics: the jury is out

NMS have been largely presumed to originate from non-dopaminergic pathways, as a range of studies have failed to find associations between NMS and dopamine transporter imaging or symptom severity (UPDRS-III) [13]; and NMS subdomains and progression are often relatively unaffected by dopaminergic therapy [13,14]. Although a review by Chaudhuri and Schapira [15] suggest that many NMS (e.g. depression, anxiety, apathy, anhedonia, restless legs and periodic limb movements during sleep, nocturia, constipation, fatigue, central pain related to PD), indeed are responsive to dopaminergic treatment. For full review of the complex relationship between NMS and dopamine transmission in PD, please consult [15].

Conflicting results suggest that further research is needed to tease apart the specific domains of NMS that could indeed be linked to degree of dopamine deficiency. Recent work by Standaert and colleagues [16] showed that after 12 weeks, patients treated with levodopa–carbidopa intestinal gel showed significant reductions in NMSS total scores (which remained a significant reduction for 60 weeks). Additionally, 5/9 domain scores were significantly decreased at week 60, compared to baseline: sleep/fatigue, attention/memory, gastrointestinal tract, sexual function, and miscellaneous. Similarly, a large multi-center open-label, nonrandomized, comparative study contrasting subcutaneous apomorphine infusion and intrajejunal levodopa infusion found a significant improvement of total NMSS compared to baseline for both treatments [17]. One notable caveat is that of both of these studies were uncontrolled, and thus may simply reflect a placebo effect. In fact, a double-blind randomized, placebo-controlled trial evaluating the effect of rotigotine on NMS using the NMSS reported that rotigotine did not statistically effect patients’ NMSS total score differently than the placebo, since the placebo effect was quite large [18].

Consideration of placebo effects is also important when evaluating the effects of deep brain stimulation on NMS. Two uncontrolled studies evaluated the effects of acute subthalamic nucleus deep brain stimulation (STN-DBS) as well as the longitudinal changes after 24 months of STN-DBS. The first study reported significant acute effects specifically in anxiety and fatigue [19], whereas the second study reported significant improvements in the total NMSS score, as well as beneficial effects of STN-DBS on many individual NMS, such as daytime sleepiness, fatigue, urinary symptoms, cardiovascular symptoms, among others after 24 months [20]. In addition to lacking a control group, both studies were also unblinded, and thus the results must be interpreted with caution.

Only recently, fluctuations of NMS have become a focus of interest, and recognized as a separate entity from motor fluctuations, despite their frequent temporal co-occurrence [21]. Pain, anxiety, drenching sweats, slowness of thinking, fatigue, and akathisia were all complaints in more than half of all patients studied. Together, these studies suggest that traditional endophenotypes are likely dominated by NMS, yet these remain under explored. Further work is needed to examine ‘non-motor’ subtypes of PD, to help predict future disease progression as well as perhaps aid in selecting effective therapeutic agents and deep brain surgeries.

2.3. Exposing the true dimensions of NMS phenotypes: an art or a science?

A sensible approach to this dizzying array of complexity is to use dimensionality reduction techniques (such as cluster analysis) to examine whether subtypes may be present in the broad spectrum of NMS (for a full review of non-motor features of PD subtypes, please refer to [22]). For example, a recent latent class analysis identified four different subtypes of mood disorders in PD: one was deemed psychologically healthy (approximately 50%), whereas the other three classes were associated with psychological distress – one with moderate anxiety alone (approximately 20%), and two with moderate levels of depression plus moderate or severe anxiety [23]. Additionally, when both non-motor and motor symptoms were considered using a cluster analysis, a non-motor dominant cluster was found, highlighting the importance of NMS in understanding the heterogeneity of PD [24,25]. We predict that shifting our focus toward analyses that use dimension reduction techniques to discover the underlying dimensions of NMS will usher in an improved understanding of the clinical heterogeneity, onset, origins, and progression of NMS, while also helping to gain insights into the integrity of non-dopaminergic systems and course of degeneration.

Fereshtehnejad and colleagues [26] took this exact approach to cluster and classify early PD patients. In doing so, they found that non-motor features (cognitive impairment, RBD, and dysautonomia) were critical for the definition of three distinct subtypes of PD (mild motor-predominant, diffuse malignant, intermediate), which also corresponded to substantially differing levels of brain network atrophy, and disease progression over time. In addition to using clustering and classifications, one can apply multivariate methods for dimension reduction in order to relate sets of variables to one another, and then examine linear combinations of variables. For example, recent work examined the relationship between atrophy in brain networks and their associated clinical motor and non-motor phenotypes [27]. Thus, rather than focusing on a particular symptom of PD and studying brain alterations, these authors were able to investigate whole brain alterations while considering multiple clinical aspects simultaneously.

3. Conclusion

While the goal of each of these approaches is to build a foundation of knowledge of the system (from the cellular level to the large-scale brain network level) that may contribute to individual NMS as well as better explain heterogeneity in PD, in the future we must avoid overlooking the substantial interactions that exist between NMS. We maintain that multivariate analyses are more likely to bring us closer to the biological underpinnings of NMS, and also to help with inclusion criteria that may be useful in designing more effective clinical trials [13]. In other words, considering NMS as an emergent property of a dynamic system with interactions and interdependence rather than a simple linear sum of the component parts may be the crucial next step toward understanding how to identify and treat these troubling symptoms. Pairing these multivariate approaches with neuroimaging, genetics, and other biomarkers will also be useful, helping us to learn more about what patterns of NMS can tell us about PD and thus, to optimally manage individuals’ unique symptom profiles. Moreover, given that many NMS pre-date the diagnosis of PD, phenotypes of NMS are crucial to predict disease course and optimize early intervention strategies, and thus emphasize the importance of fully understanding the underlying patterns of NMS and what they may reflect in the brain.

Declaration of interest

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties. Peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Additional information

Funding

This paper was not funded.