Objective: To investigate potentially modifiable predictors of loss to follow up (LTFU) three years after inclusion in a nationwide mono-centric Parkinson’s cohort study.

Background: LTFU is a big challenge in longitudinal studies of progressive neurodegenerative diseases such as Parkinson’s disease (PD) and leads to a loss of information, possibly biased results and decreased statistical power1. The assessment of the reasons of LTFU in people with an early disease stage, allows identification of predictors for this subgroup less impaired by the disease.

Method: People with typical PD or Parkinson’s disease Dementia from the Luxembourg Parkinson’s Study, a nation-wide, monocentric, observational, longitudinal-prospective study3 were eligible. The published protocol2 describes participation recruitment and engagement strategies. Data analysis was carried out by the “missing_compare”-function of the “finalfit”-package in R. We applied “Kruskal-Wallis test” for numerical variables and “chi-squared test” for categorical variables to investigate the predictors of loss to follow-up three years after study inclusion. We counteracted the problem of multiple comparisons by Bonferroni correction.

Results: Out of 736 eligible people with PD and PDD, 207 joined the cohort less than three years ago and 174 had a Hoehn and Yahr (H&Y) disease stage score higher than 2. Finally, longitudinal data of 242 patients with a complete follow-up and 113 patients who were lost to follow-up were analyzed. Out of those 113 patients, 25 died (22.1%), and one (0.9%) was excluded. 
The results show that only Starkstein Apathy Score at baseline was Bonferroni-significant (p-value <= 0.05/21) associated with LTFU.

Conclusion: Although the role of confounders and covariates like emotional wellbeing and social support (PDQ-39 subscales), depression (Beck Depression Inventory), cognition (MoCA) and years of education need to be clarified by further research applying multiple regression, patient-reported apathy (Starkstein Apathy Scale) at baseline is associated with LTFU three years after study inclusion. Consequently, people presenting apathy at study inclusion require an individualized approach to prevent LTFU in longitudinal cohorts. Our results will be taken into account for the design of future participation recruitment and engagement strategies of Luxembourg Parkinson’s study.

References: 1. Haberstumpf S, Leinweber J, Lauer M, Polak T, Deckert J, Herrmann MJ. Factors associated with dropout in the longitudinal Vogel study of cognitive decline. Eur J Neurosci. 2021 Sep 7. doi: 10.1111/ejn.15446. Epub ahead of print. PMID: 34490950.

2. Hipp, G., Vaillant, M., Diederich, N. J., Roomp, K., Satagopam, V. P., Banda, P., . . . Kruger, R. (2018). The Luxembourg Parkinson’s Study: A Comprehensive Approach for Stratification and Early Diagnosis. Front Aging Neurosci, 10, 326. doi:10.3389/fnagi.2018.00326

« Back to 2022 International Congress

MDS Abstracts - https://www.mdsabstracts.org/abstract/predictors-of-loss-to-follow-up-three-years-after-inclusion-in-a-prospective-longitudinal-cohort-study-in-people-with-early-parkinsons-disease/