Deep brain stimulation for Parkinson's disease: meta-analysis of results of randomized trials at varying lengths of follow-up.

OBJECTIVE Deep brain stimulation (DBS) is effective in the management of patients with advanced Parkinson's disease (PD). While both the globus pallidus pars interna (GPi) and the subthalamic nucleus (STN) are accepted targets, their relative efficacy in randomized controlled trials (RCTs) has not been established beyond 12 months. The objective of this study was to conduct a meta-analysis of RCTs to compare outcomes among adults with PD undergoing DBS of GPi or STN at various time points, including 36 months of follow-up. METHODS The MEDLINE, Embase, CENTRAL, Web of Science, and CINAHL databases were searched. Registries for clinical trials, selected conference proceedings, and the table of contents for selected journals were also searched. Screens were conducted independently and in duplicate. Among the 623 studies initially identified (615 through database search, 7 through manual review of bibliographies, and 1 through a repeat screen of literature prior to submission), 19 underwent full-text review; 13 of these were included in the quantitative meta-analysis. Data were extracted independently and in duplicate. The Cochrane Collaboration tool was used to assess the risk of bias. The GRADE evidence profile tool was used to assess the quality of the evidence. Motor scores, medication dosage reduction, activities of daily living, depression, dyskinesias, and adverse events were compared. The influence of disease duration (a priori) and the proportion of male patients within a study (post hoc) were explored as potential subgroups. RESULTS Thirteen studies (6 original cohorts) were identified. No difference in motor scores or activities of daily living was identified at 36 months. Medications were significantly reduced with STN stimulation (5 studies, weighted mean difference [WMD] -365.46, 95% CI -599.48 to -131.44, p = 0.002). Beck Depression Inventory scores were significantly better with GPi stimulation (3 studies; WMD 2.53, 95% CI 0.99-4.06 p = 0.001). The motor benefits of GPi and STN DBS for PD are similar. CONCLUSIONS The motor benefits achieved with GPi and STN DBS for PD are similar. DBS of STN allows for a greater reduction of medication, but not as significant an advantage as DBS of GPi with respect to mood. This difference is sustained at 36 months. Further long-term studies are necessary.