Major depressive disorder (MDD) is a risk factor for neurodegeneration, yet its heterogeneity makes identifying at-risk subtype challenging. Notably, MDD frequently co-occurs with REM sleep behavior disorder (RBD), a specific prodrome of α-synucleinopathy. It remains unclear whether comorbid MDD + RBD reflects a benign antidepressant effect, or higher neurodegenerative risk. Given growing recognition of gut-brain axis in neuropsychiatry, we aimed to delineate microbial signatures of MDD + RBD. We employed a four-group case-control design (N = 420) comprising 124 healthy controls (HC); 80 MDD without RBD features (MDD-only); 82 MDD + RBD; and 134 iRBD without psychiatric disease. All participants underwent clinical evaluation and provided fecal samples for metagenomic sequencing. Random Forest model was used to distinguish MDD + RBD, and further assessed in a validation dataset of 65 participants with MDD + RBD (n = 31) and MDD-only (n = 34). MDD + RBD exhibited prodromal neurodegenerative features, including elevated total likelihood ratio of prodromal Parkinson's Disease, olfactory deficits, and subtle motor signs. The microbial composition in MDD + RBD differed from HC and MDD-only, while resembling iRBD. Taxonomically, MDD + RBD exhibited an iRBD-like dysbiosis (e.g., enriched Akkermansia muciniphila, Ruthenibacterium lactatiformans; depleted Faecalibacterium prausnitzii), alongside depression-associated shifts (e.g., Streptococcus parasanguinis and Actinomyces oris). Functionally, MDD + RBD showed attenuated capacity of B‑vitamin biosynthesis and polysaccharides degradation, mirroring iRBD. The Random Forest machine-learning model distinguished MDD + RBD in older adults from MDD-only with an AUC of 0.73 in cross-validation and 0.79 in the validation dataset. MDD + RBD may represent a biologically distinct depression subtype associated with potential neurodegenerative risk. Gut microbiome provides a candidate approach for potential risk stratification in psychiatric populations.
Gut-Brain Axis