Press releases

Women and Parkinson’s Disease: Physical Activity Found to Be Beneficial More Than Twenty Years Before Diagnosis

17 May 2023 | By Inserm (Newsroom) | Public health

marche sportive retraitée© AdobeStock

Although physical activity is regularly discussed as being one of the avenues for preventing Parkinson’s disease (PD), studies so far had been unable to eliminate certain biases or had not focused on its specific role in women. Researchers from Inserm, Université Paris-Saclay and Université de Versailles Saint-Quentin-en-Yvelines (UVSQ) with the Gustave Roussy Institute studied the impact of physical activity on the development of PD in nearly 100,000 women from the French cohort E3N followed up over 29 years. They observed that the more the participants had been active in their lives, the lower was their risk of developing the disease – with such a benefit already present more than 20 years before diagnosis. The researchers have also shown that in women with PD there had been a significant decrease in physical activity in the 10 years prior to diagnosis, probably due to bothersome precursor symptoms. These findings, to be published in Neurology, call for consideration to be given to the rollout of preventive physical activity-based programs in people at risk of PD.

Parkinson’s disease (PD) is a progressive neurodegenerative disease characterized by the destruction of certain neurons in the brain and the accumulation of proteins toxic to them. It is the second most common neurodegenerative disease in France and there is still no cure; therefore it is essential to identify avenues for prevention.

Previous studies had shown a reduced risk of developing PD in people with who engage in high levels of physical activity. However, the limited number of such studies involving women did not make it possible to confirm this association in that population. In addition, these studies generally had a relatively short participant follow-up period with only one evaluation of physical activity, which did not make it possible to eliminate certain biases – particularly that of reverse causation.

Reverse causation is expressed as follows: precursor symptoms of PD (constipation, sleep disorders, smell disorders, subtle motor signs, etc.) may be present several years before the disease is diagnosed. Their effects could lead people to modify their behavior (such as how much physical activity they do) prior to diagnosis, which is liable to distort the statistical analyses that study the relationship between these behaviors and the risk of developing the disease.

A team led by Alexis Elbaz, Inserm researcher at the Epidemiology and Population Health Research Center (Inserm/Université Paris-Saclay/UVSQ), studied the impact of physical activity on the onset of PD in women from the E3N cohort over a follow-up period of 29 years. E3N is a French cohort of nearly 100,000 women who have been followed-up since 1990. With nearly 1,200 women with PD identified in 2018[1], it is currently the world’s largest female prospective cohort for this disease.

The changes in the physical activity[2] of each participant throughout the 29 years of follow-up were estimated based on the information collected via six individual questionnaires completed at different stages of the follow-up. The data obtained compared the practice of physical activity prior to diagnosis in the participants with PD with those of the participants of the same age without PD.

In order to reduce the risk of a reverse causation bias resulting from the potential influence of the precursor symptoms on physical activity in the years prior to diagnosis, the scientists examined the impact of physical activity evaluated more than 5, 10, 15 and 20 years prior to diagnosis on the risk of developing PD.

What they observed was that the more physically active the participants, the less likely they were to develop PD, even when the physical activity had been evaluated more than 20 years before diagnosis. The most active women therefore had a 25% less risk of developing the disease compared with the least active.

Overall, the participants with PD had been less active than the others throughout the follow-up period, including more than 20 years before diagnosis. This gap between women with and without PD increased further in the 10 years before diagnosis, suggesting that the precursor symptoms occurring in this interval could indeed be responsible for a decrease in physical activity in the women who will go on to develop the disease but who have not yet been diagnosed.


“These findings support the protective effect in women of physical activity against the risk of developing Parkinson’s disease, even over the very long term, specifies Berta Portugal, PhD student and first author of this research. They support the utility of implementing physical activity programs to prevent Parkinson’s disease in people at risk and call for conducting other studies in order to understand what activity type and level of intensity are the most beneficial,” she adds.


“This study demonstrates the importance of studies with a lengthy follow-up within large-scale cohorts to better account for the precursor symptoms and the statistical bias they cause in evaluating the benefit of physical activity on the development of Parkinson’s disease,” concludes Elbaz.


[1]The percentage of women with PD in the cohort is similar to that of the general population in Europe over the same period of time.

[2]The concept of “physical activity” goes beyond sport alone. In this study, it is quantified by the distance walked daily, the number of flights of stairs climbed, journeys taken by bicycle, time spent doing light and heavy household activities, time spent doing recreational activities, whether light (gardening, sports of moderate intensity) or vigorous (intensive sports).

Researcher Contact

Alexis Elbaz

Inserm researcher

Inserm Unit 1018

Center for Research in Epidemiology and Public Health (CESP)


Press Contact



Association of physical activity and Parkinson’s disease in women: Long-term

follow-up of the E3N cohort study


Berta Portugal, PhD candidate1; Fanny Artaud, PhD1; Isabelle Degaey, MD1; Emmanuel Roze, MD, PhD2; Agnes Fournier, PhD1; Gianluca Severi, PhD1,3; Marianne Canonico, PhD1; Cécile Proust-Lima, PhD4; Alexis Elbaz, MD, PhD1


  1. Université Paris-Saclay, UVSQ, Gustave Roussy, Inserm, U1018, Team « Exposome, Heredity, Cancer, and Health », CESP, 94807, Villejuif, France.
  2. AP-HP, Hôpital Pitié-Salpêtrière, Neurology Department, Paris, France; Sorbonne Université, France; Inserm U1127, CNRS 7225, Brain Institute, Paris, France.
  3. Department of Statistics, Computer Science, Applications G. Parenti (DISIA), University of Florence, 50134 Florence, Italy.
  4. Université Bordeaux, Inserm, Bordeaux Population Health Research Center, UMR1219, F-33000 Bordeaux, France