Levodopa-induced dyskinesia in Parkinson's disease: still no proof? A meta-analysis.

We investigated whether there is a linear relationship between levodopa (LD) dose and treatment duration, and the development of levodopa-induced dyskinesia (LID) among patients with early untreated Parkinson's disease (PD). We performed a meta-analysis of randomized-controlled trials (RCTs) comparing LD monotherapy to any other antiparkinsonian treatment in early PD patients. Meta-regressions were conducted including as covariates the effects of LD dose, treatment duration, and age. We further proceeded in subgroup analyses based on the type of medications in the non-LD monotherapy (control) group and on whether patients in the control group received additional levodopa or not. Thirteen eligible RCTs were included, which revealed a significantly higher risk for dyskinesia in patients initially treated with LD monotherapy compared to any other treatment (OR = 2.82). None of the subsequent meta-regressions revealed any significant relationship with dose, treatment duration or age. Patients treated on LD monotherapy or MAOΙ plus LD were at a greater risk to develop LID than patients who received DA only or DA plus supplemental LD. The increased heterogeneity compromised the robustness of the results. The alleged correlation between LID and LD dose and treatment duration cannot be verified based on the data available so far. Well-designed, large-scale, long-term, RCTs on drug-naïve PD patients could allow the better comprehension of the pattern of the association between LID and LD treatment parameters.