Turning Slowly Predicts Future Diagnosis of Parkinson's Disease: A Decade-Long Longitudinal Analysis

Abstract

Objective: Wearable technology allows accurate measurement of turning while walking, with cross-sectional studies indicating that difficulty turning presents even in preclinical phases of Parkinson's disease. The aim of our study was to quantify rate of change of turning performance in a cohort of older adults, and test whether turning decline can predict future diagnosis of Parkinson's disease. Methods: A total of 1,051 participants from the Tübingen Evaluation of Risk Factors for Early Detection of Neurodegeneration (TREND) study were included for a 5-visit analysis over 10 years, with development of clinically evident Parkinson's disease tracked. Participants walked a 20-m hallway for 1 minute at their preferred pace, with a wearable device on the lower back. Longitudinal trajectories of turning performance were modelled using random effects linear mixed models to establish the interval between initial turning changes and Parkinson's disease diagnosis. Cox regression assessed whether initial turning measures could predict time to Parkinson's disease onset, controlling for age and sex. Results: Of all participants, 23 were diagnosed with Parkinson's disease an average of 5.3 years post-baseline. Slower peak angular velocity at baseline was associated with a higher hazard of Parkinson's disease diagnosis, with deviations from controls emerging approximately 8.8 years before diagnosis. Additional analysis with a machine learning model using baseline characteristics of age, sex and peak angular velocity, identified 60% of prediagnostic Parkinson's disease (sensitivity: 0.600) and 80.5% non-prediagnostic Parkinson's disease (specificity: 0.805), with an area under the curve of 80.5%. Interpretation: Peak angular velocity during turning shows promise identifying and tracking motor progression in the pre-diagnostic phase of Parkinson's disease. ANN NEUROL 2025.