Which mixed depression model? A comparison between DSM-5-defined mixed features and Koukopoulos' criteria

Background The criteria of the Diagnostic and Statistical Manual of Mental Disorders 5th edition "with mixed features specifier" (DSM-5 MFS) are considered controversial since they include only typical manic symptoms. By contrast, Koukopoulos developed an alternative model of mixed depression (MxD) focusing primarily on the excitatory component. Objective To compare DSM-5 MFS and Koukopoulos' MxD (KMxD) in terms of prevalence, associated clinical variables, and discriminative capacity for bipolar depression in patients with major depressive episode (MDE). Methods A total of 300 patients with MDE-155 with major depressive disorder and 145 with bipolar disorder (BD)-were recruited. The discriminative capacity of DSM-5 MFS and KMxD criteria for BD was estimated using the area under the curves of receiver operating characteristic (ROC_AUC). The clinical variables associated with these two diagnostic constructs were assessed by performing a logistic regression. Results A total of 44 and 165 patients met the DSM-5 MFS and KMxD criteria, respectively. The ROC_AUCs and their confidence intervals for BD according to DSM-5 MFS and KMxD were 77.0% (72.0%-82.1%) and 71.9% (66.2%-77.7%), respectively. The optimal thresholds (combining sensitivity and specificity measures) for BD diagnosis were >= 1 (77%/68%) for DSM-5 MFS and >= 3 symptoms (78%/66%) for KMxD. However, considering the DSM-5 MFS cut-off (>= 3 symptoms), the specificity (97%) increased at the expense of sensitivity (26%). Conclusions KMxD and DSM-5-MFS showed an overlapping discriminative capacity for bipolar depression. The current diagnostic threshold of DSM-5 MFS did not prove to be very inclusive, if compared with the greater diagnostic sensitivity of KMxD, which also yielded better association with clinical variables related to mixedness.