Salivary Dysbiosis Aligns with an Olfactory-Cognitive Phenotype in Aging

BackgroundScalable, non-invasive markers for cognitive-decline risk are limited. Olfactory dysfunction is predictive, and oral dysbiosis is mechanistically linked to neurocognitive pathways. Hence, we tested whether pairing smell and global cognition with salivary microbiome profiling yields a targeted, clinically useful signal. MethodsWe enrolled 113 Memory Center attendees and community controls. Same-day MMSE, UPSIT, and saliva were obtained for 16S rRNA gene sequencing and cytokine measurement. Unsupervised k-means clustering on standardized MMSE-UPSIT defined two groups of participants: CNN (cognitively normal, normosmia) and CIH (cognitively impaired, hyposmia). Ordination and elastic-net models adjusted for age, sex, BMI, and sequencing depth. Functions were inferred with PICRUSt2 and were integrated with taxa via DIABLO. ResultsOverall, the 16S-based microbial community structure was similar between groups, indicating minor compositional shifts. CIH showed enrichment of periodontal anaerobes (Porphyromonas, Treponema and Prevotella), whereas CNN retained nitrate-reducing commensals (e.g. Neisseria subflava, Aggregatibacter aphrophilus). Functional shifts showed mixed consistency with literature, aligning for outer membrane usher proteins and alkyldihydroxy phosphate synthase, but diverging for thiaminase, alpha-glucuronidase, and chemotaxis protein CheX. Most salivary cytokines levels did not differ between groups. ConclusionsThis integrated smell, cognition, and saliva workflow delineates an olfactory- cognitive phenotype linked to a targeted, potentially modifiable salivary dysbiosis, periodontal anaerobes vs nitrate-reducers, rather than diffuse salivary inflammatory elevation. This approach may support non-invasive triage and monitoring along the oral- brain axis, pending independent, longitudinal validation.