Depression is the most common psychiatric disturbance in Parkinson's disease. We conducted a Cochrane systematic review to assess the efficacy and safety of antidepressant therapies in idiopathic Parkinson's disease. Relevant trials were identified from electronic databases, reference lists and queries to antidepressant manufacturers. Three randomised controlled trials examined oral antidepressants in 106 patients with Parkinson's disease. No eligible trials of electroconvulsive or behavioural therapy were found. In the first arm of the crossover trial by Andersen et al. (n=22), nortriptyline treated patients showed a larger improvement than placebo in a unique depression rating scale after 16 weeks although significance levels were not provided. A parallel group trial by Wermuth et al. (n=37) did not show any significant difference between citalopram and placebo in Hamilton score after 52 weeks. Rabey et al. (n=47) performed an open-label trial comparing fluvoxamine with amitriptyline. Similar numbers in each group had a 50% reduction in Hamilton score after 16 months. Major side effects including visual hallucinations and confusion were reported with fluvoxamine and amitriptyline. Insufficient data on the effectiveness and safety of antidepressant therapies in Parkinson's disease are available on which to make recommendations for their use. Large scale randomised controlled trials are urgently required.