Association between motor subtypes and longitudinal trajectories of non-motor symptoms in Parkinson’s disease

Parkinson’s disease (PD) exhibits considerable heterogeneity in both motor and non-motor symptoms (NMS), yet the longitudinal trajectories of NMS across motor subtypes remain poorly characterized. A total of 1384 PD patients were classified as tremor-dominant (TD, n = 935) or non-tremor-dominant (non-TD, n = 449). Cross-sectional associations between motor subtypes and NMS (ESS, REM, GDS, STAI, SCOPA-AUT, MoCA) were examined using multivariable linear regression adjusting for age, age at onset, race, body mass index (BMI), education, striatal binding ratio, Hoehn-Yahr stage, and UPDRS-III score. Longitudinal trajectories over up to 14 years were identified using latent class trajectory modeling, and associations with motor subtypes were assessed via logistic regression. Non-TD subtype was independently associated with higher baseline scores on ESS (β = 1.23), REM (β = 0.74), GDS (β = 1.10), STAI-total (β = 4.55), STAI-state (β = 1.86), STAI-trait (β = 2.69), and SCOPA-AUT (β = 2.18; all p < 0.01), but not with MoCA (p = 0.929). Latent class analyses identified distinct NMS trajectories. Non-TD subtype was associated with significantly higher odds of unfavorable trajectories for ESS (OR = 1.64), REM (OR = 1.39), STAI-total (OR = 1.58), STAI-trait (OR = 1.72), STAI-state (OR = 1.48), GDS (OR = 2.43 for highest vs. lowest class), and SCOPA-AUT (OR = 3.29; all p < 0.01). The non-tremor-dominant motor subtype is independently associated with more severe baseline NMS and adverse longitudinal trajectories of sleep, mood, and autonomic dysfunction. These findings identify non-TD patients as a high-risk phenotype warranting targeted NMS monitoring.