Oral Rinse Sourced Microbiota in Oral Health and Disease in a Representative U.S. Adult Population

This study reports genus-level oral rinse microbiota profiles in a population of 3,770 U.S. adults from NHANES 2009–2012. Oral conditions explained more microbial variance than host factors, with diversity and composition differing across health, caries, periodontitis, co-occurring disease, and edentulism. In periodontitis, diversity increased alongside a shift toward anaerobic, inflammation-adapted communities. Additionally, dysbiosis-associated genera increased with disease severity gradient while health-associated ones decreased. A machine learning model based on microbial features achieved moderate accuracy in identifying severe periodontitis (AUROC in leave-one-dataset-out validation = 0.81, 95% CI 0.75-0.87; AUROC in external validation = 0.83, 95% CI 0.77-0.88). Key contributing taxa included Desulfobulbus, Defluviitaleaceae UCG-011, Treponema, Filifactor, Pseudoramibacter, Mycoplasma, Porphyromonas, Bergeyella, and Bifidobacterium were defined in the model. These findings support oral rinses as a non-invasive tool for monitoring oral microbial ecology and assessing the presence and severity of oral disease at the population level.