The effectiveness of oral levodopa in complex Parkinson's disease (PD) is limited by its short half-life, and the resulting pulsatile dopaminergic stimulation leads to complex motor fluctuations and dyskinesia. Several treatments provide more continuous/less pulsatile dopaminergic stimulation by modifying the pharmacokinetics of levodopa or dopamine; however, patients with advanced disease can be refractory to these treatments. In such cases infusion therapies (apomorphine and intraduodenal levodopa) and neurosurgery (deep brain stimulation [DBS]) may be used. The purpose of this systematic review is to assess, as far as possible, the relative effectiveness of these therapies. There were no randomised controlled trials comparing the three treatment modalities or any directly comparable studies, therefore a descriptive analysis of the data was performed. Studies identified for levodopa infusion and DBS supported a significant benefit compared with best medical management in terms of improvements in the proportion of the waking day in a functional "on" state, activities of daily living and motor score. This finding was supported in observational studies for all three therapies. Adverse events were not adequately reported in the majority of included studies and it was therefore not possible to obtain a reliable tolerability profile of the different treatment options. The absence of direct comparative data means that, for the immediate future at least, treatment choices for advanced PD will be determined by clinical judgement and patient preference. There is an urgent need for well-designed clinical trials to generate reliable data to inform the clinical management of this difficult-to-treat subgroup of PD patients.