Potential future treatments for Parkinson's disease (PD) include those that not only provide symptomatic relief to patients but are also neuroprotective and/or neurorestorative. Glial cell line-derived neurotrophic factor (GDNF) may be a valuable candidate in this regard. Positive results using monkeys have encouraged the use of GDNF in human trials. These trials have unfortunately shown mixed results, illustrating the influence that various parameters of administration can have on clinical outcome. The aim of this review is to compare the findings of these clinical studies with available primate data. While bolus intraventricular injections of GDNF in primates have shown some behavioural efficacy, there was no clinical benefit in the first human trial using this method, which was most likely a result of inefficient GDNF distribution in the striatal parenchyma. In primates, however, continuous (rather than bolus) delivery of GDNF into the ventricles results in significant distribution in the striatum. While chronic delivery of GDNF into the ventricles has not been assessed in humans, intraputamenal protein delivery in two Phase I trials have demonstrated that GDNF considerably reduces PD symptoms, suggesting that the putamen is the optimal location for delivery. Primate studies have shown that vector mediated delivery of GDNF may provide a suitable means for long-term intraputamenal delivery. However, the possibility of high levels of GDNF resulting in widespread distribution of GDNF to non-targeted areas is a cause of concern, and vectors with transgene regulation are needed. The development of these vectors may be the way forward for GDNF treatment.