Background Development of objective, reliable and easy-to-use methods to detect the onset of motor changes in Parkinson's disease (PD) is required to identify the temporal window in which neuromodulatory therapies could be implemented. Turning impairments are present at early stages of PD. However, it is unclear, to date, whether circular walking is also altered in prodromal PD. Objective Explore the predictive potential of circular walking in prodromal PD. Methods We included 102 subjects from a nine-year prospective cohort study (with 712 participants) in the current nested case-control analysis: 16 diagnosed with PD during follow-up (incident PD) and 96 healthy controls, matched in gender, age, and education with a 1:6 ratio. Forty-one gait features were extracted from baseline assessments with accelerometers under single and dual-tasking conditions. A Cox proportional hazards regression analysis was used to test the temporal association of non-correlated gait features to the probability of being diagnosed with PD. Results We identified associations between time from baseline measurement to PD diagnosis for eleven gait features, mostly based on harmonic ratios, step and stride variability, and index of harmonicity, partially in combination with gait speed. Most significant associations indicated that low gait symmetry and low rhythmicity were associated with larger hazard of being diagnosed with PD. Area under the curve ranged 0.63-0.69. Conclusions Despite low sensitivity and specificity, the findings potentially reflect prodromal motor impairments of PD manifested during circular walking, assessed quantitatively with a low-cost and wearable instrument. This will contribute to the characterization of pre-diagnostic PD motor symptoms.
Read full abstract