Objective: To compare the distribution of Parkinson’s disease (PD) risk markers and comprehensive risk scores in an online-cohort recruited with open invitations through a media campaign versus a ‘true’ population-based sample.

Background: Validations of risk scores for prodromal PD yielded variable performances depending on the cohort in which they were applied. Different cohort types and recruitment designs are likely to influence distribution of risk markers within risk scores and performance of the scores.

Method: Analysis was performed in the Tyrol cohort of the Healthy Brain Ageing (HeBA) study, a European multi-centre study on the prevalence and predictivity of risk factors for PD; and the Bruneck Study cohort, a prospective study on neurological and cardiovascular diseases representative of the general elderly population. Data in the HeBA cohort has been collected starting from April 2022 via online-questionnaires and subsequent remote smell testing (n=1275). For the Bruneck Study sample data from past in-person follow-up examinations were used (n=540), recruited initially through invitation from the official population register as an age- and sex-stratified random sample. For the present analysis we focused on simple and established risk factors (family history, olfaction, probable REM-sleep behaviour disorder [RBD], depression, constipation and orthostatic hypotension) and applied the updated MDS research criteria for prodromal PD.

Results: Participants of the HeBA cohort were younger and had a female preponderance compared with the Bruneck cohort (median age 61 [25^th–75^th percentile: 56–67] vs. 67 [61–76] years, p<0.001; % females 59% vs. 54%, p=0.029). Positive family history (16.1% vs. 7.6%, p<0.001), symptoms of RBD (21.1% vs. 7.7%, p<0.001), constipation (12.4% vs. 9.3%, p=0.055), and orthostatic hypotension (21.0% vs. 9.6%, p<0.001) were more frequent in the HeBA cohort, while symptoms of depression (42.4% vs. 29.8%, p<0.001) were more frequent in the Bruneck cohort. Equal distribution was found for hyposmia (40.0 vs. 39.6%, p=0.906). Probabilities for prodromal PD were higher in the Bruneck cohort (2.4% [0.5–10.7%] vs. 1.0% [0.3–3.1%], p<0.001).

Conclusion: Distributions of risk factors and risk scores differ in the compared cohorts, probably owing to differences in cohort characteristics including age and recruitment strategies.

« Back to 2024 International Congress

MDS Abstracts - https://www.mdsabstracts.org/abstract/cohort-effects-on-risk-scores-and-risk-markers-for-parkinsons-disease-comparison-of-an-online-recruited-versus-population-based-cohort/