Objective: To determine whether risk estimates calculated in PREDICT-PD are associated with objective motor features on clinical examination.

Background: PREDICT-PD aims to identify a group of people at increased risk of Parkinson’s disease (PD), by combining early non-motor features and risk factors assessed online.(1) We sought to determine whether higher risk participants were more likely to have subtle motor features of PD compared with lower risk.

Methods: Existing participants in the PREDICT-PD study agreed to a home visit during which a neurological examination was undertaken (AJN). Part 3 of the MDS-UPDRS was recorded on video and scored independently by two movement disorders experts (AJL and AS) blind to risk status. Scores were compared between raters and where no significant disparity existed (<5 point difference), scores were averaged. Where there was disparity in scores, the two raters re-scored, again independently. Where disparity remained, raters re-scored jointly to obtain consensus. 209 participants were assessed at home between November 2012 and September 2015 (most were assessed in the first 10 months of this period). Comparisons were made between part 3 scores and log-transformed risk estimates calculated using survey data from year 1 of follow-up (Winter 2012/3). Higher risk subjects were defined as those above the 15th centile of risk estimates.

Results: 7 subjects independently diagnosed with PD during follow-up were excluded, as where 2 subjects taking medication associated with parkinsonism (neuroleptics and lithium). After round 1 of scoring, 24 out of 200 subjects (12%) were re-scored independently, and after round 2, 10 of these were scored jointly. Part 3 UPDRS scores ranged from 0 to 19. Of the 200 subjects, 188 had risk estimates available from year 1 for comparative analysis. Median UPDRS part 3 score was significantly greater (3; IQR 1-5.5) in higher risk than lower risk subjects (1 (IQR 0-2.75); p<0.0001). Linear regression analysis demonstrated a continuous relationship between log-transformed risk estimates and part 3 scores (β=1.83, p<0.0001). 32 of 76 higher risk subjects (42%) exceeded published cut-offs for sub-threshold parkinsonism (UPDRS>3 excluding action tremor).(2)

Conclusions: These results provide preliminary evidence that the PREDICT-PD algorithm identifies a population enriched for early motor features of PD. References: 1) Noyce et al. JNNP 2013. 2) Berg et al. Mov Disord 2015.

« Back to 2016 International Congress

MDS Abstracts - https://www.mdsabstracts.org/abstract/blinded-video-assessment-of-subtle-parkinsonian-signs-in-predict-pd-participants/