The association between Parkinson's disease (PD) and the G196A polymorphism in the brain-derived neurotrophic factor (BDNF) gene has been investigated in several case-control studies, producing contradictory results: one study indicated that homozygocity of AA is associated with PD, another study produced the opposite result, whereas other studies found no association. To investigate these contradictory findings, a meta-analysis of all available association studies between the G196A polymorphism and the risk of developing PD was conducted. Four out of six identified studies included populations of East Asian descent, and two included populations of European descent (Whites). Overall, the meta-analysis of the allele contrast (A vs G) suggested large heterogeneity between studies (P=0.07, I2 =51%) and no association between G196A and the risk of developing PD: random effects odds ratio (OR)=1.00 [95% CI (0.85, 1.18)]. The sensitivity analysis (exclusion of two studies: one East Asian and one White) with the controls not in Hardy-Weinberg equilibrium showed large heterogeneity (P=0.10, I2 =52%) and no significant association: random effects OR=0.94 [95% CI (0.77, 1.15)]. The subgroup analyses for East Asians and Whites produced no significant association. In addition, the contrast of homozygotes, and the dominant and recessive models for allele A did not support a major role for this polymorphism in the pathogenesis of PD. There were no sources of bias in the selected studies, and the differential magnitude of effect in large vs small studies was not significant. The meta-analysis results suggest that the involvement of the BDNF gene in susceptibility to PD merits further exploration with larger and more rigorous population association studies.