Fang et al1 reported a meta-analysis of prospective studies to determine whether there is evidence of a quantitative dose-response association between physical activity and Parkinson disease (PD) risk. The analysis included 8 prospective studies with a total of more than 500 000 adults followed up for an average of 12 years, with more than 2100 PD cases. This meta-analysis provided compelling evidence that physical activity, particularly moderate to vigorous physical activity, was associated with a significantly reduced risk of developing PD.

The article by Fang et al1 is an excellent example showing meta-analysis in its best role because among the 8 studies, only 1 study found a statistically significant reduced PD risk when comparing the highest with the lowest category of physical activity; however, when analyzed together, a significantly reduced risk of PD was observed with the highest level of total physical activity compared with the lowest level (relative risk [RR], 0.79; 95% CI, 0.68-0.91). The interpretability of this RR is less than straightforward because each study assessed different domains of physical activity and used different cutoffs for classifying high vs low physical activity levels; however, a strength of the study by Fang et al was that they also conducted a quantitative meta-analysis using a more powerful common metric of physical activity that renders the results more understandable. They quantified physical activity for 6 of the 8 studies that had detailed information on frequency of engagement in physical activity categories according to their energy expenditure demands (metabolic equivalent of tasks [METs]). In analyses that included both men and women, each 10 MET-hours/week increase in total physical activity decreased the risk of PD by 9% (RR, 0.91; 95% CI, 0.86-0.96). For men, each 10 MET-hours/week increase in moderate to vigorous physical activity decreased the risk of PD by 17% (RR, 0.83; 95% CI, 0.76-0.90).

Since the article by Fang et al1 was submitted, Müller et al2 reported the results of a cohort study of 7347 male veterans (the Veterans Exercise Testing Study) in which physical fitness was measured objectively by maximal exercise testing. Exercise capacity was expressed as METs estimated from peak treadmill speed and grade. A strong inverse association was observed—men who had a high level (>12 METs) vs low level (<8 METs) of fitness on exercise testing had a 76% reduction in the incidence of PD (hazard ratio, 0.24; 95% CI, 0.08-0.73; P = .01). This study adds evidence that objectively measured physical fitness in men, not just self-reported moderate and vigorous physical activity, is inversely associated with PD in a dose-response fashion.

In the meta-analysis by Fang et al, no dose-response association with moderate to vigorous levels of physical activity in METs was observed for women; however, it would be premature to conclude that moderate to vigorous physical activity is not a protective factor for women because only 4 of the 8 studies included women and all but 1 of those studies had fewer than 150 women with PD. As a result, the meta-analysis was underpowered to observe a dose-response association for women. Importantly, in the largest and best-powered study to date among women, Xu et al3 observed a significant dose-response between the number of hours of moderate to vigorous physical activity per week and lower PD risk in 3 of the 4 age categories examined.

Physical activity is the latest in the list of lifestyle risk factors associated with a reduced risk of developing PD, which include cigarette smoking and coffee consumption, both of which have consistently been shown to be inversely associated with the risk of developing PD.4 Because the pathology of PD is characterized by progressive dopaminergic deficiency and selective loss of neurons in the substantia nigra, factors which reduce the risk of developing PD may also slow disease progression after diagnosis, as has been found for caffeine consumption and lower serum urate levels.4 If the causal inference about physical activity is correct, increasing physical activity might also slow the progression of PD.

How Robust Is This Finding to Other Noncausal Explanations?Reverse causality needs to be considered as a potential mechanism for the observed protective association between physical activity and PD, as the presence of preclinical PD at baseline measurements could possibly lead to lower levels of physical activity. Reverse causality is made less likely by the fact that physical activity data were collected many years prior to the diagnosis of PD (mean, 12 years; range, 6-22 years). Moreover, 4 studies excluded PD cases that occurred in the years immediately following baseline measurements and still showed an inverse association between physical activity and PD. Another type of plausible reverse causality is the possible presence of early nonmotor symptoms (hyposmia, constipation, and sleep disorders) that may precede PD onset by up to 2 decades and could affect the propensity to engage in moderate or vigorous physical activity. Arguing against this explanation is that some studies found strong inverse associations between PD and physical activity during much earlier periods of life (high school or college; ages 35-39 years).3,5

Is Physical Activity a Biologically Plausible Protective Factor for PD?

To our knowledge, the answer to this question is a qualified yes because there is accumulating evidence from both animal and human studies of the biological mechanisms by which physical activity could confer protection against PD.6,7 Evidence that physical activity may confer biological protection initially came from studies in rodent neurotoxicant-induced models that exercise reduced selective neurodegeneration of dopaminergic neurons and attenuated movement abnormalities.8,9 Physical exercise has also been shown in rodent models to have other beneficial effects including increased secretion of neurotrophic factors in the striatum and enhanced vesicular dopamine release.10 In humans, exercise also promotes the expression of the neural growth factors which could contribute to the neuroplasticity and survival of dopaminergic neurons.

Future Directions

Parkinson disease is notable in terms of the number of risk factors that are associated with a reduced risk of developing the disease, which include smoking, caffeine intake, and urate levels, all of which have been consistently shown to be inversely associated with the risk of developing PD.4 There is now evidence that a higher level of moderate to vigorous physical activities in middle or later life were associated with lower risk of PD. The finding that moderate to vigorous physical activity levels were associated with a lower incidence of PD among men meets 5 of the criteria for establishing causality10: strength of association, consistency of findings, temporality, biological gradient (demonstration of dose-response), and biological plausibility.

It is premature to exclude the possibility of an inverse association and possible protective effect of moderate or vigorous physical activity in women; further opportunities to carry out prospective investigations that include women should be sought. In addition to the MET hours per week quantitative measure used by Fang et al,1 it would be helpful to characterize the protective association of physical activity using a more relatable metric, such as number of minutes per week spent in moderate or vigorous physical activities. Based on the results summarized in the article by Fang et al,1 it seems clear that protection against PD can be added to the list of likely benefits of physical activity.

Article Information

Published: September 21, 2018. doi:10.1001/jamanetworkopen.2018.2633

Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2018 Nelson LM. JAMA Network Open.

Corresponding Author: Lorene M. Nelson, PhD, MS, Division of Epidemiology, Department of Health Research and Policy, Stanford University School of Medicine, HRP Redwood Bldg, Room T223, Stanford, CA 94305-5405 (

Conflict of Interest Disclosures: None reported.References1.Fang  X, Han  D, Cheng  Q,  et al.  Association of levels of physical activity with risk of Parkinson disease: a systematic review and meta-analysis.  JAMA Open Netw. 2018;1(5): e182421. doi:10.1001/jamanetworkopen.2018.2421ArticleGoogle Scholar2.Müller  J, Myers  J.  Association between physical fitness, cardiovascular risk factors, and Parkinson’s disease.  Eur J Prev Cardiol. 2018. doi:10.1177/2047487318771168PubMedGoogle Scholar3.Xu  Q, Park  Y, Huang  X,  et al.  Physical activities and future risk of Parkinson disease.  Neurology. 2010;75(4):341-348. doi:10.1212/WNL.0b013e3181ea1597PubMedGoogle ScholarCrossref4.Ascherio  A, Schwarzschild  MA.  The epidemiology of Parkinson’s disease: risk factors and prevention.  Lancet Neurol. 2016;15(12):1257-1272. doi:10.1016/S1474-4422(16)30230-7PubMedGoogle ScholarCrossref5.Chen  H, Zhang  SM, Schwarzschild  MA, Hernán  MA, Ascherio  A.  Physical activity and the risk of Parkinson disease.  Neurology. 2005;64(4):664-669. doi:10.1212/01.WNL.0000151960.28687.93PubMedGoogle ScholarCrossref6.Ahlskog  JE.  Does vigorous exercise have a neuroprotective effect in Parkinson disease?  Neurology. 2011;77(3):288-294. doi:10.1212/WNL.0b013e318225ab66PubMedGoogle ScholarCrossref7.Paillard  T, Rolland  Y, de Souto Barreto  P.  Protective effects of physical exercise in Alzheimer’s disease and Parkinson’s disease: a narrative review.  J Clin Neurol. 2015;11(3):212-219. doi:10.3988/jcn.2015.11.3.212PubMedGoogle ScholarCrossref8.Tillerson  JL, Caudle  WM, Reverón  ME, Miller  GW.  Exercise induces behavioral recovery and attenuates neurochemical deficits in rodent models of Parkinson’s disease.  Neuroscience. 2003;119(3):899-911. doi:10.1016/S0306-4522(03)00096-4PubMedGoogle ScholarCrossref9.Petzinger  GM, Walsh  JP, Akopian  G,  et al.  Effects of treadmill exercise on dopaminergic transmission in the 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine-lesioned mouse model of basal ganglia injury.  J Neurosci. 2007;27(20):5291-5300. doi:10.1523/JNEUROSCI.1069-07.2007PubMedGoogle ScholarCrossref10.Hill  AB.  The environment and disease: association or causation?  Proc R Soc Med. 1965;58:295-300.PubMedGoogle Scholar