In a randomised double-blind comparative trial, the antidepressant efficacy of a daily dose of 800 mg of the St. John's wort extract LoHyp-57 (dry extract of St. John's wort, drug extrakt ratio 5-7:1, solvent, ethanol 60% [w/w]) was shown to be equivalent to that of 20 mg fluoxetine (CAS 54910-89-3) in elderly patients with mild or moderate depressive episodes according to ICD 10 (International Statistical Classification of Diseases and Related Health Problems). Treatment was given for six weeks. 149 out-patients (129 females and 20 males) were included in the intention-to-treat analysis. 72 of these patients were assigned to the ICD 10 diagnostic criterion F32.0 (mild depressive episode), while 77 patients were suffering from moderate depressive episodes, corresponding to F32.1. The principal target criterion was the patient's global score on the HAMILTON Depression Scale (items 1-17). During the six-week course of treatment with LoHyp-57, the HAMILTON global score fell from 16.60 points at entry to 7.91 points, and in the fluoxetine sample it fell from 17.18 to 8.11 points. In the group of patients with mild depressive episodes, the score showed a mean fall from 14.21 to 6.21 points on LoHyp-57, and from 15.21 to 7.46 points on fluoxetine. In patients with moderate depressive episodes, the score showed a mean fall from 18.73 to 9.43 points on LoHyp-57 and from 19.10 to 8.75 points on fluoxetine. The efficacy of both medications was found to be equivalent both in mild and moderate depressive episodes. Both treatment groups showed adverse drug reactions (ADRs). Twelve ADRs with a possible relationship to the study medication were reported during treatment with LoHyp-57. Six patients were prematurely withdrawn from treatment with the study medication for this reason. On fluoxetine 17 ADRs occurred with a possible relationship to the study medication. These led to abandonment of treatment and therefore premature withdrawal from the study in 8 cases.