A double-blind randomised controlled trial of the effect of low dose lofepramine (70 mg once daily) against placebo was carried out on depressed elderly inpatients on general medical wards for the elderly, comparing measures of depression and side-effects between the randomised groups. Patients were identified for the study using the Geriatric Depression Scale (GDS) and the Brief Assessment Schedule Depression Cards (BASDEC). Sixty-three subjects were randomised: 46 patients completed the entire trial of 28 days treatment. BASDEC and GDS were administered on day 8 post-admission, and depressed patients were randomised double-blind to either low dose lofepramine (70 mg daily) (n = 23) or placebo (n = 23). Assessment of changes in depressive states were made using the Montgomery Asberg Depression Rating Scale (MADRS) on days 8, 18 and 36 post-admission. Both groups improved by a similar amount during the trial. Lofepramine tended to be more effective than placebo in those patients who were more depressed (GDS > or = 18). On the other hand, subjects who were less depressed (i.e. GDS < 18) improved more on placebo than lofepramine. Low dose lofepramine may prove useful in moderately or severely depressed patients treated for only 4 weeks. However, low dose lofepramine is not indicated for mild (GDS 15-18) depression.