Murdoch University Research Repository

Welcome to the Murdoch University Research Repository

The Murdoch University Research Repository is an open access digital collection of research
created by Murdoch University staff, researchers and postgraduate students.

Learn more

Instrumented gait analysis identifies potential predictors for Parkinson’s disease converters [abstract]

Del Din, S., Elshehabi, M., Galna, B.ORCID: 0000-0002-5890-1894, Hansen, C., Hobert, M., Suenkel, U., Berg, D., Rochester, L. and Maetzler, W. (2018) Instrumented gait analysis identifies potential predictors for Parkinson’s disease converters [abstract]. Movement Disorders, 33 (suppl 2). Abstract Number: 1098.


Objective: This longitudinal prospective observational study investigated if gait can predict Parkinson’s disease (PD) conversion from a cohort of community-dwelling older adults.

Background: PD is a progressive disorder including a prodromal period when definitive motor/non-motor symptoms to permit a diagnosis have not yet appeared. Quantification of gait with wearable technology (WT) may serve as an accurate tool to identify surrogate markers of incipient disease manifestation. Recently arm swing and selective gait characteristics measured with WT have been shown to be potential prodromal markers for people at risk for PD [1]; however these data were obtained from a cross-sectional assessment; the potential of gait to predict PD conversion has not been investigated yet in a longitudinal cohort.

Methods: 16 participants (69±5 years (yrs)) who were diagnosed with PD on average 4.5 yrs after baseline assessment (converters (PDC)) and 48 age-matched old healthy adults (HA) recruited in the TREND study were included. Assessments were performed longitudinally 4 times at 2-year intervals. Participants were asked to walk at their preferred speed, performing 2 straight-line trials over 20m with a WT device placed on the lower back; 14 validated clinically relevant gait characteristics were quantified [2]. ANCOVA was used to examine gait between-group differences; the value of baseline gait in predicting PDC was explored using AUC and stepwise, forward, logistic regression analyses. Random effects linear mixed-models (RELM) were used to predict latency gait deterioration and diagnosis of PD.

Results: PDC walked with significantly lower pace, higher variability and asymmetry than HA (p≤0.027). Pace, variability and asymmetry characteristics were able to significantly predict PDC (AUC≥0.695). Step time variability was the best predictor for the stepwise, forward, logistic regression (sensitivity 25%, specificity 98%, accuracy of 80%). RELMs indicate gait impairment (step velocity and step length) is evident 4-6 yrs prior to diagnosis.

Conclusions: Our preliminary results suggest that pace, variability and asymmetry of gait represent sensitive predictors of prodromal PD and that gait impairment starts 4-5 years prior to diagnosis. Therefore, gait assessment may play an important role in concert with other biomarkers to identify people at high risk of PD and aid early diagnosis.

Item Type: Journal Article
Publisher: John Wiley & Sons, Inc.
Publisher's Website:
Other Information: Part of: 2018 International Congress, 5-9 October 2018, Hong Kong.
Item Control Page Item Control Page