Background: The functional outcome of Bipolar Disorder (BD) is highly variable. This variability has been attributed to multiple demographic, clinical and cognitive factors. The critical next step is to identify combinations of predictors that can be used to specify prognostic subtypes, thus providing a basis for a staging classification in BD.
Methods: Latent Class Analysis was applied to multiple predictors of functional outcome in a sample of 106 remitted adults with BD.
Results: We identified two subtypes of patients presenting "good" (n=50; 47.6%) and "poor" (n=56; 52.4%) outcome. Episode density, level of residual depressive symptoms, estimated verbal intelligence and inhibitory control emerged as the most significant predictors of subtype membership at the p<0.05 level. Their odds ratio (OR) and confidence interval (CI) with reference to the "good" outcome group were: episode density (OR=4.622, CI 1.592-13.418), level of residual depressive symptoms (OR=1.543, CI 1.210-1.969), estimated verbal intelligence (OR=0.969; CI 0.945-0.995), and inhibitory control (OR=0.771, CI 0.656-0.907). Age, age of onset and duration of illness were comparable between prognostic groups.
Limitations: The longitudinal stability or evolution of the subtypes was not tested.
Conclusions: Our findings provide the first empirically derived staging classification of BD based on two underlying dimensions, one for illness severity and another for cognitive function. This approach can be further developed by expanding the dimensions included and testing the reproducibility and prospective prognostic value of the emerging classes. Developing a disease staging system for BD will allow individualised treatment planning for patients and selection of more homogeneous patient groups for research purposes.
Copyright © 2012 Elsevier B.V. All rights reserved.