The Impact of Physical Activity on Non-Motor Symptoms in Parkinson’s Disease: A Systematic Review

Parkinson’s disease (PD) is a neurological disorder that is associated with both motor and non-motor symptoms (NMS). The management of PD is primarily via pharmaceutical treatment; however, non-pharmaceutical interventions have become increasingly recognized in the management of motor and NMS. In this review, the efficacy of physical activity, including physiotherapy and occupational therapy, as an intervention in NMS will be assessed. The papers were extracted between the 20th and 22nd of June 2016 from PubMed, Web of Science, Medline, Ovid, SportsDiscuss, and Scopus using the MeSH search terms “Parkinson’s,” “Parkinson,” and “Parkinsonism” in conjunction with “exercise,” “physical activity,” “physiotherapy,” “occupational therapy,” “physical therapy,” “rehabilitation,” “dance,” and “martial arts.” Twenty studies matched inclusion criteria of having 10 or more participants with diagnosed idiopathic PD participating in the intervention as well as having to evaluate the effects of physical activity on NMS in PD as controlled, randomized intervention studies. The outcomes of interest were NMS, including depression, cognition, fatigue, apathy, anxiety, and sleep. Risk of bias in the studies was evaluated using the Cochrane Collaboration’s tool for assessing risk of bias. Comparability of the various intervention methods, however, was challenging due to demographic variability and methodological differences. Nevertheless, physical activity can positively impact the global NMS burden including depression, apathy, fatigue, day time sleepiness, sleep, and cognition, thus supporting its therapeutic potential in neurodegenerative conditions such as PD. It is recommended that further adequately powered studies are conducted to assess the therapeutic role of physical activity on both motor and non-motor aspects of PD. These studies should be optimally designed to assess non-motor elements of disease using instruments validated in PD.

with PD can be challenging and non-pharmacological options such as education, support services, and exercise are likely to be underutilized due to various reasons that include limited knowledge on its therapeutic potential. In recent times, there is a growing body of evidence which supports the beneficial effects of non-pharmacological therapy, in particular, the effect of exercise on both motor and NMS (4).
Physical activity has been found to influence the brain's neurochemistry and plasticity, through the upregulation of neurotrophins such as brain-derived nerve factor (BDRF) and nerve growth factor (NGF) in rat models (5,6). BDRF has been suggested to increase turnover rate of dopamine in vitro and provide a neuroprotective role in nigral dopamine neurons (7). Additionally, BDRF has been thought to regulate branching and remodeling of axons and dendrites, whose length is reduced in PD (8,9). NGF has also been linked to neuroprotection through stabilizing intracellular calcium which may influence L-type calcium channels know to mediate synaptic transmission of dopamine in rat midbrains (10,11). The increased presence of such chemicals in the brain could help impede the progression of PD and potentially provide a neuroprotective effect.
Additionally, the relationship between exercise and endorphins has long been studied, as the opioid peptide can produce feelings of euphoria and well-being. The hormone is produced by the pituitary gland and hypothalamus during exercise, as well as excitement and pain (12). Endorphins are hypothesized to improve mood, such as depression and anxiety, via two different mechanisms; through binding to opioid receptors in the frontal cortex and limbic region, areas involved with mood; and by interacting with other neurotransmitters, such as dopamine, which also partake in improvement of mood (13).
Physical activity is becoming more popular for the treatment of chronic disease such as PD. However, the focus of many studies, to date, has primarily been on motor symptoms, with fewer studies assessing its effect on NMS. This review is aimed at determining the best available current evidence on the effects of various forms of exercise on NMS in PD.

Literature Search
This review included articles which met all of the following criteria: 10 or more participants diagnosed with idiopathic PD completing the intervention, participants needed to be randomized, and studies must evaluate the effect of physical activity on NMS in PD. For the purpose of this review, physical activity is defined as "any bodily movement produced by skeletal muscles that result in energy expenditure" (14). Papers were excluded if published prior to 1996 and did not assess global or specific NMS as a variable of interest. Additionally, papers needed to score a minimum score of 2 out of a maximum 5 according to the JADAD Scale, a short and widely used method to assess the quality of a report (15). The literature search was conducted using six databases between the 20th and 22nd of June 2016 (PubMed, Medline, Ovid, SportsDiscuss, Scopus, and Web of Science). Search terms included the MeSH terms for "Parkinson's, " "Parkinson, " and "Parkinsonism" combined with "exercise, " "physical activity, " "dance, " "physiotherapy, " "occupational therapy, " "physical therapy, " "rehabilitation, " and "martial arts. " Titles and abstracts were read with papers not meeting selection criteria being discarded, and those remaining were read in full to check for suitability, in accordance with the Preferred Reporting Items for Systematic and Meta-Analyses (PRISMA) (see Figure S1A in Supplementary Material) (16). The protocol for this systematic review was registered with PROSPERO on the 8th of April 2016.

Data extraction
Data extraction was completed by one reviewer confirmed by a fellow author. Relevant articles meeting the inclusion criteria were reviewed with all relevant information, such as type of intervention, frequency, duration, and mode of delivery, along with outcomes.

Outcomes of interest
The outcomes of interest were both global and specific NMS. Those assessed globally included instruments such as the Unified Parkinson's Disease Rating Scale Part 1 (UPDRS-1) and the Non-Motor Symptom Scale (NMSS), as well as more focused assessments of cognition, depression, fatigue, apathy, anxiety, and sleep using validated instruments.

ReSULTS
The database search yielded 20 papers which met inclusion criteria (see Table 1).

Risk of Bias
Risk of bias was assessed using the Cochrane tool for assessing risk of bias, which assesses bias in a number of different aspects that include random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, blinding of outcome data, selective reporting, and other bias (37). Only four studies had a low risk of bias in all sections (19,32,34,35), while majority had at least one criterion with unknown bias (17, 18, 20, 23-29, 31, 33). Four studies did have high risk of bias in one section, three for selective reporting due to presenting only some of the data (21,30,36), and one for blinding of outcome assessment as assessors were not blinded to allocation (22). The criterion which was least reported, and thus had the highest number of unknown risk of bias was allocation concealment (17, 18, 20-31, 33, 36). For more information, see Figure S2A in Supplementary Material.

DiSCUSSiOn
In more recent times, non-pharmacological therapies in PD have become increasingly acknowledged as beneficial with various modalities offered to patient populations.

intervention Design
The study design varied greatly, in particular, when looking at frequency and length of intervention. Various studies showed a significant improvement in NMS, including two short duration (20,25) and one medium duration studies (36) involving 2-3 sessions/week lasting 45 min to 1 h and a study with a low frequency of 6-h long classes (32). Depression was one of the most widely studied outcomes, showing significant improvements in a number of different activities (23,25,32,36). Significant improvements were also seen in sleep (24,32), fatigue, apathy (25), and cognition (36).
Although it is widely regarded that exercise affects brain plasticity, different exercise types have been found to selectively affect various brain regions. Aerobic training demonstrated its importance in the aging brain, showing the most benefit in brain regions most affected by aging, including prefrontal, superior parietal, and temporal cortices in gray matter, and anterior and transverse tracts between the frontal and parietal lobes in white matter, which are areas involved in cognition and everyday functioning (38). Another study on resistance training had shown to significantly change brain regions involved with response inhibition, including the left anterior insula which extends into the lateral orbital frontal cortex as well as the anterior part of the left middle temporal gyrus (39). Participants who underwent training twice weekly significantly improved in the Flanker test when compared with participants who had 12 months of twice weekly balance and tone training. However, participants who only performed resistance training once per week did not show significant improvements from the control group. This suggests that while different exercise types may beneficially affect different brain regions, the improvement was not distinguishable in this review as different types of exercise showed positive effects in global and specific NMS such as sleep. However, fatigue and apathy only showed significant improvement in aerobic exercise, possibly due to the brain regions primarily involved.

Risk of Bias
Three of the studies showed a high risk of bias in selective reporting of outcome measures, highlighting the trends of importance to the authors, in particular, those with positive results (21,30,36). With regard to health-care interventions, it is important that all data are fully and clearly reported to help guide clinicians in decision-making (40). However, the majority of data were present or able to be retrieved from the authors in regard to the outcomes of interest of this study, alleviating the potential bias as a result of underreporting. One study did not blind the assessors for data collection, which may have implications on data collected (22). Unblinded assessors may exhibit prejudice to expectant results (41).
In the MDS taskforce review of ASs, AS was ranked over LARS as its sensitivity was yet to be determined (47). The fatigue instruments analyzed by the MDS taskforce were the FSS and PFS, where FSS is validated for both screening and rating severity, while PFS was validated as screening tool. The KFS has also been independently validated for the screening of fatigue in chronic disease though not specifically validated in PD (48). Sleep instruments were also reviewed (49) with the ESS been shown to be both valid and reliable (50) and PDSS-2 also validated (51).
Validated depression scales reviewed and validated by the MDS taskforce included the BDI, HADS, and Ham-D. The Ham-D was shown to be an optimal instrument, followed by the BDI and HADS (52). HADS has also been suggested by the MDS taskforce for rating anxiety (53). The LPDQ has been validated for use in a general population and not in PD (54).
The MDS diagnostic criteria for mild cognitive impairment recommend that detailed level 2 assessment requires a minimum of two tests per cognitive domain, but stated that excessive or imbalanced number of tests per cognitive domain could cause bias (55). The MoCA has adequate metrics as a brief assessment of global cognition in PD (56), whereas the MMSE is thought to be less sensitive to mild cognitive impairment in PD (57). The CAB has been shown to be sensitive in mild cognitive impairment and dementia but is yet to be validated in PD (58).
The environment can have numerous effects on NMS such as cognition. One study found that interaction with natural environments as opposed to urban environments can improve cognitive control (59). This could provide a basis for why Nordic walking, which was performed at a city park, had significant improvements in the NMS tested (25). Group settings and family support can also contribute to aspects such as exercise frequency and involvement. It was previously found that social support networks are low in older adults compared with younger individuals, possibly due to issues such as peers becoming less active and family falling into social stereotypes fearing injury for a frail elderly population (60). This could suggest that studies which comprise group settings, as well as including family members, could be beneficial to the participant.

Limitations
This review does present a number of limitations. First, the study sample size was modest, with 20 papers reviewed, in-part due to the strict criterion. Additionally, the majority of papers include aerobic or custom training programs, with only a small fraction looking at strength and balance training. Furthermore, variable assessment of NMS using different instruments made it challenging for comparing study outcomes. Due to these limitations, a meta-analytical approach to this topic will be suitable when more well-designed studies are made available.

COnCLUSiOn
Physical activity may be a suitable non-pharmacological therapy in PD. Global and specific NMS involving depression, apathy, fatigue, cognition, and sleep were significantly improved by some form of physical activity. However, the synergistic effects of both pharmacological and non-pharmacological treatments in PD are still unclear. For an unbiased appraisal of each activity type, further research is needed which prioritizes NMS to help determine the most beneficial effects of physical activity in this complex condition.

AUTHOR COnTRiBUTiOnS
MC conducted the literature review and drafted the manuscript. KD reviewed the manuscript. TK supervised and reviewed the manuscript.

SUPPLeMenTARY MATeRiAL
The Supplementary Material for this article can be found online at http://journal.frontiersin.org/article/10.3389/fmed.2016.00035