• bipolar depression
• antidepressant

Bipolar disorder (BD) is a chronic, recurrent, and debilitating illness characterized by manic or hypo-manic episodes interspersed with depressive episodes (Dell’Aglio et al., 2013).
Major depressive episodes (MDE) are the most frequent presentation of BD (Judd et al., 2002) and they have significant influence on the course of the disease and on global functioning of the patient (Altshuler LL et al., 2002; Rosa A.R. et al., 2010).
The target of treatment of depression in BD is not only to achieve full remission of acute episodes and to prevent suicides, but also to obtain long-term stabilization by preventing relapse.
The general agreement is that non-antidepressant treatments should be considered as mono-therapy before using antidepressant (AD) to treat BD (Pacchiarotti et al., 2013).
The aim of the study was to evaluate the use of AD in a naturalistic sample of 241 outpatients with MDE in MDD, BD-I or BD-II and to assess their short term effectiveness.
The sample was followed for a period of 12 weeks and naturalistically treated with mood stabilizers (lithium salt or anti-epileptics drug) antidepressant and antipsychotics.
The total sample was divided in two groups: patients treated with AD (pAD) and patients treated without AD (pNAD). Furthermore patients with BD were divided in two groups: patients treated with AD (BDAD) and treated without AD (BDNAD).
In this study we found that factors associated with AD treatment were older age, female sex, a negative family history of BD or of psychosis, a positive family history of depression. This is in line with another study (Williams et al., 2018).
However the differences between the two groups were not significant with the exception for marital status, in fact the number of married patients was higher in BDAD.
Analysing DSM-5 specifiers, anxiety features and seasonal pattern were significantly higher in BDAD while psychotic features was significantly higher in BDNAD.
Patients with anxiety features have greater depressive symptom severity, chronicity, more suicidal ideation or behavior and greater global impairment (Fava et al., 2006; Goldberg and Fawcett, 2012; Gasperz 2017).
The two groups differ significantly only for previous switches, that were higher in BDNAD. They didn’t significantly differed in term of suicide attempts and hospitalizations.
Regarding concomitant treatments, as expected, significant differences between the two groups were found in the more frequent use of mood stabilizers (lithium, valproic acid or carbamazepine) in BDNAD. This fact is supported by most guidelines that suggest AD as an adjuvant therapy with mood stabilizers in BD to decrease switch rates (Sachs et al., 2007, Pacchiarotti 2013) and reduce AD-emergent suicidality (Yerevanian and Choi 2013).
In our study comparing the two groups we didn’t found any significant differences in term of suicide attempts after 12 weeks of treatment.
Regarding mood switch during the first 12 weeks, we found no differences in the two groups. It could be explained by the association with mood stabilizers. In fact in a large study by Viktorin and colleagues, (2014) treatment with AD, when added to mood stabilizers, was not associated with higher switch rates than treatment without AD. Moreover it could be explained also by the fact that SSRI were the AD that were used the most. Indeed several placebo-controlled randomized trials did not find elevated switch rates with SSRI or bupropion, either as adjuncts to mood stabilizers or in mono-therapy (Tondo et al., 2010).
Regarding to symptomatology progression from baseline to twelve weeks of treatment there was a reduction of HDRS scores in the two groups from baseline to week 8, then there was an increase of scores in BDNAD and a stable score in BDAD until week 12.
Relative to median scores of YMRS there was a reduction in the two groups from baseline to week 8, then there was an increase of scores higher in BDAD.
No differences were found between the two groups in term of remission, response and CGI score after twelve weeks.
Several limitations should be considered in this study.
First, the sample was very small. Moreover the follow-up was in a short period of time (12 weeks).