Bipolar disorder (BD) commonly co-occurs with anxiety disorders (AD) and this comorbidity has been associated to worse clinical outcomes in BD patients. However frequency and clinical outcomes data might be biased once some of them had been originated from acutely ill bipolar patients as well as those in remission.

To assess AD in bipolar sample in order to know if this frequency vary according illness phase and identify possible clinical outcomes related to this comorbidity.

A population of 355 outpatients with bipolar I disorder were evaluated using structured instruments (SCID). The sample was split into four groups considering the illness phase and the presence of AD: euthymic without AD (euthymic/non- AD); euthymic with AD (euthimic/AD); non-euthymic without AD (non-euthymic/non-AD) and non-euthymic with AD (noneuthymic/ AD) groups. Clinical and socio-demographics differences between the groups were compared through bivariate analysis.

There was a progressive and significantly increasing in some worst clinical outcomes frequency, such as rapid cycling, suicide attempts, substance and alcohol use disorders, history of antidepressant use and decreased scores in all domains of QoL, from control group (euthymic /non-AD) through out intermediate group (non-euthymic/non-AD) until the group most affected (non-euthymic/AD). Both euthymic/non-AD and euthimic/AD groups did not differ.

We identified a group of severe bipolar patients that have worse clinical outcomes, and among them, the anxiety comorbidity. But, contrary to what has been stated, AD appears to be related with poor bipolar prognosis only when other severity clinical factors are presents.