Datasets from two recently conducted phase 3 clinical trials and three ongoing observational studies, all enrolling participants with early Parkinson’s disease, have been cleaned, re-coded, and consolidated for disease progression modeling. The clinical trial datasets include data from STEADY-PD III (n=336) and SURE-PD 3 (n=298). The observational study datasets come from the Parkinson’s Progression Marker Initiative (PPMI; n=413), Incidence of Cognitive Impairment in Cohorts with Longitudinal Evaluation (ICICLE; n=314), and Cambridgeshire Parkinson’s Incidence from GP to Neurologist (CamPaIGN; n=142). We plan to expand this dataset by adding additional clinical trial and observational study data that are immediately available to our team.
The original Ambulatory Capacity Measure is a construct to measure gait and ambulation ability in Parkinson’s disease, derived from the original UPDRS. The contemporary MDS-UPDRS differs from the UPDRS in scaling and some individual items and there is no congruent measure of ambulatory capacity using it. Therefore, we created an updated ACM (uACM) from select MDS-UPDRS items. Our analysis with uACM scores have shown they correlate with other scores that measure disease progression and disability, and the scores progress steadily over time despite increasing dopaminergic medications. This latter point is important, as the MDS-UPDRS is highly sensitive to dopaminergic medications, which introduces significant variability when trying to model MDS-UPDRS scores over time.Therefore, we believe uACM scores to be an attractive candidate as an outcome measure for disease progression modeling.
Developed disease progression trajectory models We have developed latent class mixed models to identify Parkinson’s disease progression subgroups demonstrating distinct patterns of motor and non-motor symptoms. Separate models were constructed for MDS-UPDRS part II scores, part III scores, uACM scores, and Montreal Cognitive Assessment (MoCA) scores over four years. For each outcome, we have identified the optimal number of trajectory classes. For example, for MoCA scores, three different trajectory classes emerge as the optimal number of classes (Figure): stable progression, moderate progression, and rapid progression.