Symptom-based depression subtypes and sex: a mixture modeling approach integrating biopsychosocial aspects

Abstract

Major depressive disorder (MDD) is a highly heterogeneous psychopathological construct that risks subsuming different types of psychiatric conditions. A possibility in order to deal with this lack of specificity is to definemore homogeneous depression subtypes,such as symptom-­‐based subtypes (Baumeister & Parker, 2012). Mixture modelling (e.g.,latent class analysis (LCA) and latent transition analysis (LTA)) displays a promising person-­‐centered statistical approach enabling theextraction of homogeneous subgroups of depressed subjects. Prior mixture modelling studies havederivedtypical (also referred to asmelancholic), atypical, and moderatedepression subtypes, which were not only discriminable by differing symptom profiles, but alsoby specific biopsychosocial characteristics.However, despite thewell-­‐established impact of sex on depression there is still a lack of knowledge aboutempirically derived sex-­‐related and sex-­‐specific depression subtypesto date. In particular, population-­‐based studies examining this issueare required.The present doctoral thesis aimed to contribute to this research gap by three studiesprovidingthe basic research for futureadvances fosteringthe effectiveness and specificity of depression treatment.In the first study (manuscript I), the stability and transition patterns of the three empirically derived symptom-­‐based subtypes severe typical, severe atypical, and moderate depression were examined over the time span of 20 years using data from the Zurich Study.LTAsunder consideration of the predictor sexwere fitted to population-­‐based data of 322 subjects manifesting depressive symptoms. Subjects withtemporally stable depressive phenomenologywere furthermorecharacterized by psychosocial characteristics. Results indicatedthat the severe atypical subtype and the moderate subtypes displayed a high temporal stability, which even increased over time,whereas stability of the severetypical subtype was relatively low.The latter was associated with male sex, while females were more prone to belong to the severe atypical subtype. This study was the first to detectrelevant sex-­‐related differences in long-­‐term stability and transition patternsof depression subtypes. Males exhibited a more pronounced stability of depression subtypes than females. In contrast,females displayed more transitions between the depression subtypes over time; transitions from the severe typical to the severe atypical subtype and vice versa were particularly prominent.With regard to the psychosocial characteristics, both stable severe subtypes showed significant associations with psychosis syndromes, while eating disorders wereonly significantly associated withthe stablesevere atypical subtype. The second study (manuscript II) aimed at replicating the empirically found depressionsubtypes from the first study by applyingLCAsto another cross-­‐sectional community sampleincluding subjects with depressive symptoms (373 males; 443females). TheZInEP (German: ZurcherImpulsprogramm zur nachhaltigen Entwicklung der Psychiatrie)surveywas parallelised regarding sex-­‐age composition to the longitudinal cohort used in study one. Additionalsex-­‐specificdepressionsubtypes were expected due to the higher statistical power of the large database. This studysuccessfully replicated the severe male-­‐related typical subtype, the severe female-­‐related atypical subtype, and the non-­‐sex-­‐related moderate subtype. Furthermore, two male-­‐specific depressive subtypeswere derived: a large severe irritable/angry-­‐rejection sensitive (IARS)subtype, partly resembling Rutz’s male depressive syndrome (Rutz, 1999; Rutz, von Knorring, Pihlgren, Rihmer, & Walinder, 1995)and a small psychomotor retarded subtype.With regard to associated psychosocial characteristics, there were similarities and discrepancies between the sexes. Males with membershipinthe severe typical subtype displayedthe lowest masculine gender role orientation,while typical depressed females revealedmore anxiety disorders. The severe atypical subtype wasmarkedlyassociated with eating disorders in both sexes, however only in females with substance use (alcohol and drugs).Incontrast, substanceusewas linkedtothe severe IARS subtype in males.Compared tothestable, long-­‐term depression subtypes of study one, the comorbidity patterns of the depressive subtypes in the cross-­‐sectional study were much more pronounced. This may be the result of ‘switchers’fluctuating between typical and atypical episodes(and perhaps further depressive subtypes)over time, which could only be captured in the longitudinal study design. The author proposesthat the issue of ‘switchers’ has been underestimated in depression research to date. Finally, the third study (manuscript III) accounted for a biological characterisation of the data-­‐driven depression subtypesfrom study two by comparing the available gonadal hormones of 60 depressive males with 15 healthy subjects. Both the severe typical subtype and severe IARSsubtypeshowed significantly higher serum testosteronelevels compared to the controls. The lower testosteroneconcentrationof the severe atypical subtype was obviously confounded by the high body mass index(BMI) associated with this subtype.Overall,the empirically derived depression subtypes under consideration of the factor sex are replicable in both cross-­‐sectional and longitudinal samples andarequite well characterizable by biopsychosocial correlates,supporting the assumptionthat MDD consists of several homogeneous,etiopathogenetic conditions. Combiningthenew data-­‐driven findings with the existing paradigms and theories has the potential forinnovative inspiring perspectives. Theextended knowledge onsymptom-­‐based depression subtypes may provide the basic research for future preventionsand more specific depression treatments.