Identifying antecedent and illness course variables differentiating bipolar I, bipolar II and unipolar disorders

Background: Clinical differentiation of bipolar conditions (and especially bipolar II disorder) from unipolar conditions is not always straightforward. We sought to identify illness antecedents and correlates that may assist their differentiation and complement clinical symptoms. Methods: We undertook detailed comparative analyses of comprehensive data obtained from patients diagnosed with a bipolar or unipolar mood disorder. Results: The sample comprised 138 bipolar (45 bipolar I and 93 bipolar II) and 214 unipolar participants. Univariate analyses identified numerous differentiating variables, while multivariate analyses generated a refined variable list to determine discriminatory capacity. Controlling for all other factors, those with a bipolar (I or II) condition were more likely than the unipolar sub-set to report a family history of bipolar disorder, experiencing bullying at school, to make a suicide/self-harm attempt, and be less likely to be clinically judged as having 'problematic' personality traits. Factors differentiating bipolar II from unipolar sub-sets included the aforementioned variables, as well as higher rates of lifetime heavy drinking and female gender, and briefer depressive episodes in the bipolar II group. Bipolar I and II sub-sets differentiated solely by higher rates of hospitalization in the former group. Limitations: Some study variables (e.g., hospitalization) may merely reflect DSM-IV diagnostic criteria or consequences rather than illness antecedents or correlates. Other self-reported variables (e.g., bullying) are subject to memory biases, and may reflect higher-order variables (e.g., early problematic personality traits). Conclusions: Study findings provide assistance to determining non-symptom features that may improve discrimination of the bipolar disorders from themselves and from unipolar conditions. (c) 2012 Elsevier B.V. All rights reserved.