Antidepressants may induce manic or hypomanic episodes. The identification of predictors of antidepressant- induced mania (AIM) is essential to improve the management of bipolar disorder (BD). However, the rare studies on AIM are generally characterized by small sample sizes, varying definitions of AIM and heterogeneous groups of patients, thus leading to conflicting results.

To compare a population of AIM(+) to AIM(-) patients in order to identify specific clinical factors associated with AIM.

All 252 participants met the DSM-IV criteria for BD. Only patients who reported AIM in the 90 days after the beginning of an antidepressant (with or without a mood stabilizer) were diagnosed as AIM (+) and those without any lifetime history of AIM despite lifetime antidepressant prescription were considered AIM (-). Sociodemographic and clinical factors were collected using the DIGS, ALS, AIS BIS and WURS.

AIM(+) (N=74) and AIM(-) (N=178) patients did not differ significantly in terms of age, gender distribution, bipolar disorder duration and age of onset, ALS, BIS and WURS score. However, the rates of rapid cyclers, lifetime history of suicidal acts, alcohol use disorder and AIS score were significantly higher in the AIM(+)group. The type of polarity of onset was significantly different in both groups.

A history of rapid cycling, of suicidal acts and of alcohol use disorder could be considered as risk factors of AIM in BD. Patients with these factors could therefore be identified as a vulnerable subgroup prone to manic switch with antidepressant.