Predictors of the Onset of Manic Symptoms and a (Hypo)Manic Episode in Patients with Major Depressive Disorder

OBJECTIVE One third of patients with a major depressive episode also experience manic symptoms or, even, a (hypo)manic episode. Retrospective studies on the temporal sequencing of symptomatology suggest that the majority of these patients report depressive symptoms before the onset of manic symptoms. However, prospective studies are scarce and this study will, therefore, prospectively examine the onset of either manic symptoms or a (hypo)manic episode in patients with a major depressive disorder. In addition, we will consider the impact of a large set of potential risk factors on both outcomes. METHODOLOGY Four-year follow-up data were used to determine the onset of manic symptoms as well as a CIDI-based (hypo)manic episode in a large sample (n = 889, age: 18-65 years) of outpatients with a major depressive disorder and without manic symptoms at baseline. Baseline vulnerability (i.e., sociodemographics, family history of depression, childhood trauma, life-events) and clinical (i.e., isolated manic symptoms, depression characteristics, and psychiatric comorbidity) factors were considered as potential risk factors. RESULTS In our sample of depressed patients, 15.9% developed manic symptoms and an additional 4.7% developed a (hypo)manic episode during four years. Baseline isolated manic symptoms and comorbid alcohol dependence predicted both the onset of manic symptoms and a (hypo)manic episode. Low education only predicted the onset of manic symptoms, whereas male gender, childhood trauma and severity of depressive symptoms showed strong associations with, especially, the onset of (hypo)manic episodes. CONCLUSIONS A substantial proportion (20.6%) of patients with a major depressive disorder later developed manic symptoms or a (hypo)manic episode. Interestingly, some identified risk factors differed for the two outcomes, which may indicate that pathways leading to the onset of manic symptoms or a (hypo)manic episode might be different. Our findings indirectly support a clinical staging model.