In an attempt to find diagnostic predictors of treatment response, 45 inpatients with major depression in a randomized trial received amitriptyline or imipramine for 23 days after a washout period of 16 days. A polydiagnostic approach was applied to the classification of endogenous depression. Only a minority of diagnostic schedules for endogenous depression proved predictive of a more favorable course during antidepressant treatment: the criteria of D. F. Klein, the Newcastle Scale II, and the Vienna Research Criteria. The symptoms of nonreactivity of mood and diurnal variation were the only endogenous symptoms with predictive ability. In addition, psychotic features and secondary depression predicted an unfavorable course, whereas characteristics of previous course (bipolarity, recurrence) were not predictive.
