Hierarchical Inflammatory Phenotypes of Depression: A Novel Approach across Five Independent Samples and 27,730 Adults

Abstract

Background: Although characterizing associations between inflammation and depression may prove critical for informing theory, research, and treatment decisions, extant research has been limited by not testing whether inflammation is associated with depression broadly or with a subset of symptoms in the same model. This lack of direct comparison has hampered attempts to understand inflammatory phenotypes of depression and, critically, fails to consider that inflammation might be uniquely associated with both depression broadly and individual symptoms. Methods: We used moderated nonlinear factor analysis in five population-based cohorts (N=27,730, 51% female, mean age=46 years). Results: C-reactive protein is simultaneously associated with latent depression, appetite, and fatigue. Specifically, C-reactive protein was associated with latent depression in all five samples (rs: .044-.089; ps: <.001-.002) and was associated with both appetite (significant rs: .031-.049, significant ps: .001-.007) and fatigue (significant rs: .030-.054, significant ps: <.001-.029) in four samples. These results were largely robust while controlling for gender, age, and disease burden. Conclusions: Methodologically, these models indicate that the PHQ-9 is scalar noninvariant as a function of CRP. Therefore, mean-comparisons of depression total scores and CRP might be misleading without accounting for symptom-specific associations. Conceptually, these findings indicate that studies investigating inflammatory phenotypes of depression should examine how inflammation is simultaneously related to both depression broadly in addition to specific symptoms, and whether these relations function via different mechanistic pathways. Doing so has the potential to yield new theoretical insights and may lead to novel therapeutic strategies for reducing inflammation-related symptoms of depression.