Objective: DSM-5 introduced the "with mixed features" specifier for major depressive episodes. The authors assessed the prevalence and phenomenology of mixed depression among bipolar disorder patients and qualitatively compared a range of diagnostic thresholds for mixed depression. Method: In a naturalistic study, 907 adult outpatients with bipolar disorder participating in the Stanley Foundation Bipolar Network were followed longitudinally across 14,310 visits from 1995 to 2002. The Inventory of Depressive Symptomatology-Clinician-Rated Version (IDS-C) and the Young Mania Rating Scale (YMRS) were administered at each visit. Results: Mixed depression, defined as an IDS-C score ≥15 and a YMRS score >2 and <12 at the same visit, was observed in 2,139 visits (14.9% of total visits, and 43.5% of visits with depression) by 584 patients (64.4% of all patients). Women were significantly more likely than men to experience sub threshold hypomania during visits with depression (40.7% compared with 34.4%). Patients with one or more mixed depression visits had more symptomatic visits and fewer euthymic visits compared with those with no mixed depression visits. DSM-5-based definitions of mixed depression (ranging from narrower definitions requiring ≥3 non overlapping YMR Sitems concurrent with an IDS-C score≥15, to broader definitions requiring ≥2 non overlapping YMRS items) yielded lower mixed depression prevalence rates (6.3% and 10.8% of visits, respectively) but were found to have similar relationships to gender and longitudinal symptom severity. Conclusions: Among outpatients with bipolar disorder, concurrent hypomanic symptoms observed during visits with depression were common, particularly in women. The DSM-5 diagnostic criteria for depression with mixed features may yield inadequate sensitivity to detect patients with mixed depression.