Survival of Melancholia: a Retrospective Study of Patients with Depressive Disorders

The DSM-IV and the DSM-5 eliminated the importance of the syndromal identity of melancholic depression in favour of a dimensional model within the domain of major depressive disorders. Melancholic depression was excluded from DSM as a distinct disorder owing to the impact of ageing, genetics, and course of illness. We challenge these assertions using retrospective data collected from patients with depression.

Electronic medical records of 1073 patients with depressive-spectrum disorders in 12 centres across Germany spanning from January 2010 to June 2013 were retrospectively reviewed. The diagnosis of melancholia was made using the Hamilton Depression Rating Scale 21 items (HAMD-21). Patients were followed up every 2 weeks and yearly until discharge from inpatient units. The final dataset consisted of 1014 patients; each had received a minimum of two complete observations.

At baseline, patients with melancholic depression had higher HAMD-21 score than did patients with non-melancholic depression (32.6 vs 23.13, p < 0.001). At the final visit, patients with melancholic depression responded to treatment more often than did patients with non-melancholic depression (81.3% vs 69.04%, p = 0.0156), whereas the two groups were comparable in terms of remission status (50.55 vs 48.68%, p = 0.1943). The relapse rate was higher in patients with melancholic depression than in patients with non-melancholic depression after 1 year (60% vs 45.01%, p = 0.0599), 2 years (77.78% vs 60.36%, p = 0.0233), and 4 years (80% vs 64.45%, p = 0.0452).

Melancholic depression has an identifiable constellation of symptoms and it is not just a severe form of major depression. Melancholic depression is not the result of age-related or pathoplastic changes. We advocate including melancholia as its own illness entity in the next edition of the DSM.