Lofepramine, an imipramine analogue, was compared with imipramine in a multicentre, double‐blind clinical trial. The 62 patients (31 in each of two treatment groups) had a depressive syndrome that normally would have been treated with a tricyclic antidepressant. These patients had not received any adequate treatment for their present depressive episode. After a wash‐out period and once a week during treatment (up to 5 weeks), routine laboratory tests and electrocardiograms were done. The dosage was 50 mg t.i.d. for imipramine and 70 mg t.i.d. for lofepramine.Depression ratings with the Cronholm‐Ottosson depression rating scale were performed before treatment and once weekly for 3 weeks and then in the 5th week. The last four ratings were combined with rating of side‐effects.In the 5th week of treatment 15 out of 31 in the lofepramine group and 18 out of 31 in the imipramine group had recovered. This difference was not significant, nor did the median values of individual symptoms differ between the groups. The side‐effects were moderate and the two groups only differed significantly in the items “dry mouth” and accommodation disturbances in favour of lofepramine. The drug compliance was checked by plasma levels of desmethylimipramine in the imipramine group, parent compound, and “apparent” desmethylimipramine in the lofepramine group.The relationship between plasma drug levels, the effect on noradrenaline uptakein vitroand amelioration is discussed inSiwers et al. (1977) in this issue.The clinical outcome in the two groups did not differ significantly; interpretation of this result is discussed in relation to the reliability of rating and sele