Abstract
BACKGROUND AND OBJECTIVES
Recent reports have suggested that levodopa exposure may contribute to freezing of gait (FOG), but this relationship remains controversial because of confounding by disease severity and duration. Our goal was to assess the relationship between levodopa use and FOG incidence and severity.
METHODS
We studied 2 prospective, multicentric cohorts: NS-Park and Parkinson's Progression Markers Initiative (PPMI). Patients with ≥2 visits were selected. Three nested subcohorts were defined based on disease duration, baseline FOG, and levodopa exposure. FOG was assessed using the Movement Disorder Society-Unified Parkinson's Disease Rating Scale (MDS-UPDRS) Part II 2.13. Analyses included mixed-effects ordered logistic regression to evaluate FOG severity and Kaplan-Meier and Cox proportional hazards models to assess incident FOG. Models adjusted for disease duration, Hoehn and Yahr (H&Y) stage, and MDS-UPDRS Part III scores.
RESULTS
A total of 25,602 patients with Parkinson disease from the NS-Park (mean age 68.3 ± 29.1, 42% female) and 1,441 from the PPMI (mean age 62.9 ± 9.8, 37% female) cohorts were included. In unadjusted analyses with the NS-Park cohort, levodopa use was associated with higher FOG severity and incidence, but this association was no longer statistically significant in the PPMI cohort. In mixed-effects ordinal logistic regression models adjusting for disease duration/severity, levodopa status showed a nonsignificant effect size on FOG (NS-Park odds ratio [OR] 1.84, 95% CI 0.94-3.61, p = 0.074; PPMI OR 0.76, 95% CI 0.38-1.51, p = 0.430), while disease duration (NS-Park OR 4.4, 95% CI 3.01-6.29, p = 6.0e-15; PPMI OR 6.97, 95% CI 5.05-9.63, p = 4.7e-32), Hoehn & Yahr stage (NS-Park OR 6.8, 95% CI 3.52-13.08, p = 1.059e-08; PPMI OR 1.5, 95% CI 0.96-2.22, p = 0.073), and MDS-UPDRS-III score (NS-Park OR 1.85, 95% CI 1.38-2.49, p = 4.34e-05; PPMI OR 2.00, 95% CI 1.67-2.39, p = 2.4e-14) were independently associated with greater FOG severity. No statistically significant association was observed between future levodopa exposure and de novo FOG incidence. In turn, FOG incidence was driven by the H&Y (NS-Park: hazard ratio [HR] 1.94, 95% CI 1.68-2.23; PPMI: HR 1.78, 95% CI 1.47-2.17) and MDS-UPDRS III (NS-Park: HR 1.40, 95% CI 1.07-1.83; PPMI: HR 1.26, 95% CI 1.17-1.36).
DISCUSSION
No significant or consistent association between levodopa exposure and FOG severity or incidence was observed after adjustment for disease duration and severity, with the effect estimates significantly attenuated despite residual uncertainty.