Objective: To determine how sensor-based measurement of motor performances relate to the heterogeneity of clinical progression in early Parkinson’s disease (PD).
Background: Observed disease progression in the timeframe of PD clinical trials is heterogeneous and generally subtle, with most participants lacking measurable change in disability based on the MDS-UPDRS clinical rating scale. Sensor-based digital measurement tools may have the potential to capture changes not reflected in MDS-UPDRS, but the responsiveness of digital measures of motor impairment remains to be further characterized in relation to patterns of clinical PD progression.
Method: Longitudinal data from PD participants (N=357) from the Phase II SPARK trial (NCT03318523) were analyzed to determine the relationship between digital measures of motor performances and identified clusters of clinical PD progression. The trial population was stratified based on their longitudinal clinical motor score (MDS-UPDRS part II+III) using growth mixture modeling, an unsupervised clustering method. Inertial measurement unit sensor-based kinematic measures of walking, turning, and wrist pronation-supination were compared using Cohen’s d as a measure of difference between identified clusters of PD progression. Data from clinic visits after initiation of symptomatic medications were censored.
Results: Model selection resulted in a two-group linear model: a more severe and rapidly progressing group (N=34; baseline MDS-UPDRS II+III = 41; progression rate = 31 pts/year) and a less severe, slowly progressing group (N=323; baseline = 26; progression rate = 6 pts/ year). At baseline, the faster progression group was significantly worse on all digital measures of upper extremity bradykinesia, turn dynamics, and steady-state gait. Longitudinally, the groups differed exclusively in digital measures of gait and turn dynamics, with gait speed and stride length having the largest effect sizes.
Conclusion: Unsupervised longitudinal clustering of PD trial participants based on their clinical motor trajectories identified two distinct subgroups. The faster clinical progression group had a digitally-defined phenotype of rapid gait deterioration. Further longitudinal evaluation of digital measures within the slower-progressing group is warranted.
To cite this abstract in AMA style:D. Rodriguez Duque, J. Edgerton, L. Zhu, C. Kanzler, C. de Moor, S. Belachew, T. Dam, T. Liu, C. Shen, M. Yang, F. Nahab. Sensor-based digital measurement of gait and wrist pronation-supination in clinically-defined classes of Parkinson’s disease progression identified by unsupervised clustering [abstract]. Mov Disord. 2023; 38 (suppl 1). https://www.mdsabstracts.org/abstract/sensor-based-digital-measurement-of-gait-and-wrist-pronation-supination-in-clinically-defined-classes-of-parkinsons-disease-progression-identified-by-unsupervised-clustering/. Accessed September 21, 2023.
« Back to 2023 International Congress
MDS Abstracts - https://www.mdsabstracts.org/abstract/sensor-based-digital-measurement-of-gait-and-wrist-pronation-supination-in-clinically-defined-classes-of-parkinsons-disease-progression-identified-by-unsupervised-clustering/