In a small observational secondary analysis of PKG wearable data (n=25–30 complete cases), increased percent time in dyskinesia tended to associate with less improvement in total MDS‑UPDRS over routine treat‑to‑target care, but the adjusted effect was borderline (β=0.25 per 1% PTD, p=0.064) and…

Indicates wearable-derived dyskinesia burden may serve as a pragmatic digital biomarker to monitor and potentially guide PD treatment, offering translational clinical utility if confirmed in larger prospective studies.

Wearable monitoring offers a way to quantify motor-state burden in Parkinson's disease during routine care, but the clinical meaning of longitudinal change in dyskinesia burden remains uncertain. This secondary analysis used a publicly available real-world treat-to-target dataset from people with Parkinson's disease managed with serial Personal KinetiGraph (PKG) monitoring. Participants with paired baseline and final total Movement Disorder Society-sponsored Unified Parkinson's Disease Rating Scale (MDS-UPDRS) scores were included in the paired clinical analyses. The primary adjusted regression used a complete-case sample restricted to participants with paired percent time in dyskinesia (PTD) values and complete covariate data. Change-score linear regression was used to examine whether longitudinal change in PTD was associated with change in total MDS-UPDRS score, adjusting for baseline total MDS-UPDRS, sex, disease duration, and baseline PTD. Among 30 participants with paired MDS-UPDRS data, total scores improved from 65.3 (SD 19.2) at baseline to 53.9 (SD 23.9) at final assessment (mean change - 11.4, 95% CI - 16.7 to - 6.0; p < 0.001). In the primary adjusted regression (n = 25), each 1-percentage-point increase in PTD corresponded to an estimated 0.25-point less favourable change in total MDS-UPDRS score (β = 0.25, SE 0.13, 95% CI - 0.02 to 0.52; p = 0.064). The estimate was in the expected direction, but the confidence interval crossed zero and the result was not conventionally statistically significant. Sensitivity analyses, including ANCOVA-type modelling and bootstrap resampling, were supportive only and did not materially change the overall interpretation. In this exploratory secondary analysis, increasing wearable-derived dyskinesia burden tended to accompany less favourable change in disability during routine treat-to-target care for Parkinson's disease. These findings should be interpreted cautiously because of the small complete-case sample and the observational design. Larger prospective studies are needed to clarify whether wearable-derived dyskinesia burden has clinically useful prognostic or treatment-guiding value and to determine whether clinically meaningful thresholds can be defined.