To the Editor: Dr. Kocsis urges that DSM-5 should retain the status quo of melancholia as a specifier for major depressive disorder, whereas we recommended, based on much evidence (1, 2), that melancholia be positioned as a distinct mood disorder.

The current characterization of major depressive disorder is the key problem. In DSM criteria for major depressive disorder, many features of melancholia have been adopted, while allowing many nonmelancholic features. Of nine qualifying items, only five are required for major depressive disorder. This disjunctive format guarantees heterogeneity within the diagnosis. Moreover, because the melancholia specifier criteria overlap so much with the criteria for unspecified major depressive disorder, it becomes logically problematic to establish melancholia as a distinctive disorder in evaluating biomarkers or treatment specificity. When alternate criteria for melancholia versus nonmelancholic major depressive disorder are employed, differential treatment outcomes emerge. For example, the Newcastle Scale predicts ECT response (3); the meta-analysis by Perry (4) reported tricyclic antidepressants as distinctly superior to selective serotonin reuptake inhibitors; and higher ratings on the Clinical Outcome in Routine Evaluation show superior responses to antidepressant drugs and to ECT (1, pp. 160—171).

The Mallinckrodt et al. study cited by Dr. Kocsis misapplied even the limited DSM criteria for the melancholia specifier. Mallinckrodt et al. used an idiosyncratic Mini-International Neuropsychiatric Interview that does not measure the criteria central to the melancholia definition. While the usefulness of neuroendocrine markers is rejected by Dr. Kocsis, the evidence is strong that the Dexamethasone Suppression Test does differentiate ICD-8-defined melancholic and nonmelancholic depression (2). And, in challenging the argument that melancholia responds poorly to placebo, he ignores evidence presented by Brown (5).

In our view, melancholia is a distinctive syndrome clinically defined by specific and pervasive disturbances in affect, psychomotor activity, vegetative functions, and cognition— and, in a subset of patients, by psychosis. The risk of suicide is high. Tricyclic antidepressants and ECT are differentially superior treatments. To advance diagnosis-specific treatment and research, melancholia should be acknowledged as an entity, not marginalized as a secondary specifier. Given the stakes, the importance of getting the diagnosis right is considerable.