In Parkinson's disease, OFF predictability and clinical features explain more OFF impact variance than OFF time.

Background: Historically, OFF burden in Parkinsons disease has been primarily attributed to motor features. Recent studies highlight that non motor symptoms, and the predictability of OFF episodes also drive functional impairment, yet they are rarely measured in clinical practice. Objective: To identify which clinical features are most closely associated with OFF time and OFF impact, and to quantify the added explanatory value of temporal predictability, non-motor, and behavioural domains beyond a core motor model. Methods: We analysed 1,252 OFF only visits from 430 PPMI participants. Outcomes were MDS UPDRS IV 4.3 (OFF time) and 4.4 (OFF impact). Linear mixed effects models with a participant random intercept were fitted. The core motor model included OFF state motor severity, freezing, tremor, levodopa responsiveness, and dyskinesia, plus covariates. Predictability (IV; 4.5), non motor (mood, fatigue/sleep, autonomic/GI), and behavioural (impulse control behaviours) domains were then added to assess added influence beyond motor. Analyses were stratified by time since diagnosis (Pooled; [≤] 4y; [≥] 6y). Results: Clinical features explained more variance in OFF impact than OFF time (25.9% vs 8.1%). OFF time was primarily linked to OFF state motor severity/freezing, with levodopa responsiveness important early. For OFF impact, predictability produced the largest increment in marginal R squared beyond the core motor model (pooled and Late). Within the core motor model, tremor was the largest contributor to OFF impact. Conclusions: Predictability is a prominent correlate of OFF impact. Asking about predictability may help tailor therapy, from timing optimisation to on demand rescue for unpredictable episodes.