Evidence has shown that some patients with bipolar disorder have a relatively accurate sense of their cognitive abilities, whereas others may overreported or underreported cognitive difficulties, which causes a discrepancy in this measures.

To investigate concordance and discrepancy between subjective and objective cognitive measures, as well as to identify factors that could influence this discrepancy.

Patients who met DSM IV-TR criteria for bipolar disorder in partial or full remission (HDRS-17 score ≤ 12; YMRS score ≤ 7) were recruited from outpatient clinic at Barcelona and Porto Alegre. Objective cognitive assessment was performed by the Letter-Number Sequencing (LNS-WAIS III). Cognitive Complaints in Bipolar Disorder Rating Scale (COBRA) was used as a subjective cognitive measure.

Were included 179 patients. We found a concordance between COBRA and LNS in 62 cases, and discrepancy in 117 cases (Fig. 1). The incongruent group (COBRA–and LNS + ) have less years of study (8.10 ± 4.01) than the incongruent group (COBRA+ and LNS–) (13.44 ± 4.05, P = 0.001), and than congruent group (COBRA–and NLS–) (13.75 ± 4.04, P = 0.003). Finally, the congruent group (COBRA+ and LNS + ) was the group with higher functioning impairment.

A few number of false-negative cases were detected, suggesting that COBRA can be used as a screening instrument. A special attention should be provided for subjects with a few years of study, because possibly these subjects presents more difficulty in express its cognitive difficulties.

The authors have not supplied their declaration of competing interest.