Abstract
BACKGROUND: The ideal timing for initiating levodopa in newly diagnosed people with Parkinson's disease (PD) is uncertain due to limited data on the long-term effects of levodopa. OBJECTIVE: The aim was to investigate whether early levodopa initiation postpones mortality (primary outcome), the requirement of device-aided therapies, and the incidence of PD-related complications, such as fall-induced injuries. METHODS: Using nationwide claims data from Dutch hospitals (2012-2020), we grouped newly diagnosed PD individuals as "early initiators" (initiating levodopa within 2 years of diagnosis) or "nonearly initiators." We used the national death registry to assess mortality and health-care claims to assess PD-related complications and device-aided therapies. We used marginal structural models to compare mortality and device-aided therapy rates between groups, and a Poisson regression model to compare PD-related complication rates. RESULTS: Among 29,943 newly diagnosed PD individuals (mean age at diagnosis: 71.6, 38.5% female), there were 24,847 early and 5096 nonearly levodopa initiators. Over a median 4.25 years, 8109 (27.1%) died. The causal risk ratio for mortality was 1.04 (95% confidence interval [CI] 0.92-1.19) for early versus nonearly initiators. The risk ratio of receiving any device-aided therapy was 3.19 (95% CI 2.56-5.80). No association was observed with incidence of PD-related complications (incidence rate ratio: 1.00, 95% CI 0.96-1.05). CONCLUSIONS: Early levodopa initiation in PD does neither postpone nor accelerate mortality or PD-related complications, nor does it precipitate earlier occurrence of PD-related complications or mortality. However, we cannot exclude that the results were influenced by residual confounding due to unmeasured risk factors of mortality.