A total of 56 patients attending a general practitioner for treatment of depression, most of whom met the criteria for major depression, were included in this double-blind, parallel group, 6-week study, in which the selective MAO-A inhibitor moclobemide (MOC; maximum dose 600 mg) was compared with the tricyclic antidepressant doxepin (DOX; maximum dose 250 mg). Thirty patients on MOC and 23 on DOX were assessed after treatment for at least 1 week and are included in the response evaluation. Improvement was assessed primarily with the Montgomery-Asberg Depression Rating Scale (MADRS). There were only 4 drop-outs in the MOC group and three in the DOX group after 1 week. Overall improvement measures showed a nonsignificant difference in favor of DOX. Two factors were found to have prognostic significance: (1) previous or present panic attacks (10 patients in the MOC group and--by chance--only one in the DOX group) were associated with significantly lower improvement within the MOC group. Since we had no a priori hypothesis about this effect, it could be a chance finding. (2) Improvement was negatively correlated with age; this was statistically significant in the total group as well as in the MOC group, with a nonsignificant trend in the same direction in the DOX group. Side effects differed little between the two groups; only dryness of mouth appeared with markedly higher frequency in the DOX group.