Oral multi-niche microbiota and age stratification: a novel strategy for precision screening of periodontitis.
Chen Yinan Y, Mei Enhua E, Zhang Wei W, Wang Yan Y et al.
Periodontitis is a prevalent chronic inflammatory disease driven by oral microbiota dysbiosis, with profound oral and systemic impacts that impair healthy aging. Conventional periodontal assessment requires calibrated clinical examination, is time-consuming, and may cause discomfort or bleeding on probing, which limits its scalability for community-based screening. Microbiota-based strategies therefore show promise for accessible screening; however, the roles of oral niche heterogeneity and age-related microbial shifts remain unclear. A total of 120 participants were classified into low-risk/non-periodontal (NP, n = 67) and high-risk/periodontal (P, n = 53) groups according to a CPI-based index-tooth periodontal risk screening protocol combined with panoramic radiographic assessment. Easily accessible oral samples (saliva, supragingival dental plaque, and tongue coating) were collected for 16S rRNA gene sequencing. Bioinformatic analyses and machine-learning models were used to evaluate diagnostic efficacy, with age stratification (< 50 years vs. ≥ 50 years). Age differed significantly between the NP and P groups in the overall cohort (p = 0.004, t test). After age stratification, no significant within-stratum age differences were observed between the two periodontal groups (< 50 years: p = 0.122; ≥ 50 years: p = 0.948). Non-age-stratified exploratory models showed moderate but variable discriminatory performance, with LOOCV AUCs ranging from 0.616 to 0.740. After age stratification, the niche-specific models showed moderate and relatively consistent performance, with LOOCV AUCs ranging from 0.759 to 0.786 in the < 50-year stratum and from 0.749 to 0.760 in the ≥ 50-year stratum. Although several age-stratified models showed numerically higher AUCs than the corresponding non-age-stratified models, this apparent improvement should be interpreted cautiously because of the reduced subgroup sample sizes. Pairwise DeLong tests showed no statistically significant AUC differences among saliva, supragingival dental plaque, and tongue coating in either the non-age-stratified or age-stratified analyses. Accounting for age-related microbial variation may contribute to more consistent performance of microbiota-based periodontal screening models. However, the three accessible oral niches showed comparable moderate performance, and no single niche demonstrated statistically significant superiority. These findings are preliminary and require validation in larger, multicenter cohorts with independent external validation.