Objective: To examine the association between time-dependent physical activity (PA) and incidence of Parkinson’s disease (PD) in women from the E3N cohort over 29 years of follow-up using lagged analyses to address the potential for reverse causation.
Background: Previous studies found an inverse association between PA and PD, with a stronger association in men than women based on the few studies that examined sex-differences. In addition, most studies used a single PA measure and lagged analyses with short lags (≤5 years) were performed to account for reverse causation. Hence, additional prospective studies with long follow-up are needed in women.
Method: We followed 96,665 women (1990-2018) through biennial self-administered questionnaires and drug-claim databases. PD was ascertained through a multistep validation process based on drug claim databases and validation of medical records by an expert panel. PA level was assessed in 6 questionnaires, but questions were not the same throughout all questionnaires. We therefore created a time-dependent latent PA variable using latent-process mixed models. We used time-dependent Cox proportional hazards models with age as the time scale and adjusted for confounders (rural residence, parity, smoking, menopausal status, body mass index) to estimate hazard ratios (HR) and 95% confidence intervals (CI). In our main analysis, we lagged PA (in quartiles) by 5 years to account for reverse causation; in sensitivity analysis we used lags of 10-, 15-, and 20-years. We used linear mixed models adjusted for confounders to examine trajectories of continuous PA prior to PD diagnosis over 29 years of follow-up within a nested case-control study.
Results: Over the follow-up (mean of 22 ± 4.1 years), 1,163 women developed PD. In our main analysis (5-years lag), PD incidence was lower (HR=0.75, 95%CI=0.63-0.88, P-value=0.0007) among those who practiced high levels of PA compared with those in lowest category. Similar results were found using longer lags. Analyses of PA trajectories showed that the average PA level was lower in cases compared with controls throughout the follow-up.
Conclusion: Based on a large cohort of women with a long follow-up, a high PA level was associated with a lower hazard of PD, even when PA was assessed long before PD diagnosis. Further studies are needed to understand the mechanisms underlying this association and which types of activities are beneficial.
References: 1. Yang F, Trolle Lagerros Y, Bellocco R, Adami HO, Fang F, Pedersen NL, et al. Physical activity and risk of Parkinson’s disease in the Swedish National March Cohort. Brain. 2015;138(Pt 2):269-75.
2. Canonico M, Artaud F, Degaey I, Moisan F, Kabore R, Portugal B, et al. Incidence of Parkinson’s disease in French women from the E3N cohort study over 27 years of follow-up. Eur J Epidemiol. 2022.
To cite this abstract in AMA style:B. Portugal, F. Artaud, I. Degaey, E. Roze, G. Severi, M. Canonico, C. Proust-Lima, A. Elbaz. Physical activity and incidence of Parkinson’s disease in French women from the E3N cohort study [abstract]. Mov Disord. 2022; 37 (suppl 2). https://www.mdsabstracts.org/abstract/physical-activity-and-incidence-of-parkinsons-disease-in-french-women-from-the-e3n-cohort-study/. Accessed September 21, 2023.
« Back to 2022 International Congress
MDS Abstracts - https://www.mdsabstracts.org/abstract/physical-activity-and-incidence-of-parkinsons-disease-in-french-women-from-the-e3n-cohort-study/