The antidepressive effect of citalopram in depressed patients with idiopathic Parkinson's disease (PD) was evaluated in a placebo-controlled trial. Before treatment the patients had a mean Hamilton Depression Scale (HAM-D) score around 17. In the acute 6-week phase a statistically significant decrease in HAM-D was obtained in both groups of patients but without difference between the groups. End-point HAM-D scores were around 10. In the continuation phase many patients dropped out because of the lack of response. The dose of citalopram was rather low, and the drug was well tolerated. Thus, no exacerbation of PD was seen when citalopram was compared with placebo. During the trial the applicability of the HAM-D in patients with PD was questioned because around 56% of the patients had recurrent brief depression of 2-3 days' duration, which might explain the negative results of the present study.