|
Citation
|
Brancati GE, Vieta E, Azorin JM, Angst J, Bowden CL, Mosolov S, Young AH, Perugi G. J. Psychiatr. Res. 2019; 115: 151-157. |
|
Abstract
|
BACKGROUND: DSM-5 and ICD-11 define mixed depression as the presence of non-overlapping symptoms of opposite polarity during a major depressive episode. However, such a definition has generated controversy.
METHODS: 2720 patients with major depression, enrolled in BRIDGE-II-MIX cross-sectional study, were subdivided in clusters using a k-medoids algorithm based on 32 clinical features. Clinical variables were compared among clusters. Stepwise logistic regression and random forest predictor importance estimates were used to identify which features best predicted cluster membership. Data-driven criteria were compared with DSM-5 mixed specifier and previously proposed research-based criteria (RBDC).
RESULTS: Two clusters were identified (MDE ± MX), mainly reflecting differences in current manic symptoms. As expected, MDE + MX showed higher rates of comorbidities and bipolar features, more previous depressive episodes and suicide attempts, shorter duration of current MDE and lower age at onset. Seven clinical features among the original 32 proved to be the best predictors of cluster membership. Derived criteria perfectly allocated subjects in clusters, requiring at least four features out of the following seven: irritability, emotional lability, psychomotor agitation, distractibility, mood reactivity, absence of reduced appetite, and absence of psychomotor retardation. RBDC showed a better performance than DSM-5 in identifying MDE + MX subjects.
CONCLUSION: Our results strongly suggest a predominant role for overlapping "manic" symptoms in defining mixed depressive states. Mixed depression is better identified by the presence of excitatory features shared with mania and atypical features rather than by non-overlapping manic symptoms.
Copyright © 2019. Published by Elsevier Ltd.
Language: en |