Delineating bipolar II mixed states in the Ravenna–San Diego collaborative study: the relative prevalence and diagnostic significance of hypomanic features during major depressive episodes

Background: Depressive mixed state (DMX), defined by hypomanic features during a major depressive episode (MDE) is under-researched. Accordingly, study aims were to find DMX prevalence in unipolar major depressive disorder (MDD) and bipolar II depressive phase, to delineate the most common hypomanic signs and symptoms during DMX, and to assess their sensitivity and specificity for the diagnosis of DMX and bipolar II. Methods: 161 unipolar and bipolar II MDE psychotropic drug- and substance-free consecutive outpatients were interviewed during an MDE with the Structured Clinical Interview for DSM-IV. DMX was defined at two threshold levels as an MDE with two or more (DMX2), and with three or more (DMX3) simultaneous intra-episode hypomanic signs and symptoms. Results: DMX2 was present in 73.1% of bipolar II, and in 42.1% of unipolar MDD (P<0.000); DMX3 was present in 46.3% of bipolar II, and in 7.8% of unipolar MDD (P<0.000). The most common hypomanic manifestations during MDE were irritability, distractibility, and racing thoughts. Irritability had the best combination of sensitivity and specificity for the diagnosis of DMX2 and DMX3. Various combinations of irritability, distractibility, and racing thoughts correctly classified the highest number of DMX2 and DMX3, and had the strongest predictive power. DMX2 had high sensitivity and low specificity for bipolar II, whereas DMX3 had low sensitivity (46.3%) and high specificity (92.1%). Limitations: Single interviewer, cross-sectional assessment, and interviewing clinician not blind to patients’ unipolar vs. bipolar status. Conclusions: When conservatively defined (≥3 intra-episode hypomanic signs and symptoms during MDE), DMX is prevalent in the natural history of bipolar II but uncommon in unipolar MDD. These findings have treatment implications, because of growing concerns that antidepressants may worsen DMX, which in turn may respond better to mood stabilizers. These data also have methodological implications for diagnostic practice: rather than solely depending on the vagaries of the patient’s memory for past hypomanic episodes, the search for hypomanic features — ostensibly elation would not be one of those — during an index depressive episode could enhance the detection of bipolar II in otherwise pseudo-unipolar patients. Strict adherence to current clinical diagnostic interview instruments (e.g. the SCID) would make such detection difficult, if not impossible.