Abstract
Background
To investigate the association between salivary oral microbiome diversity and periodontitis in a nationally representative U.S. population.
Methods
A total of 5323 participants with both periodontal examination and oral microbiome data from NHANES 2009–2012 were included. Periodontitis was defined using 2012 CDC/AAP criteria. Four alpha diversity indices (operational taxonomic unit [OTU] richness, Faith's phylogenetic diversity [FPD], the Shannon‐Weiner index [SWI], and the inverse Simpson index [ISI]) were calculated. Beta diversity metrics (Bray–Curtis, weighted and unweighted UniFrac) were assessed using PERMANOVA for group comparisons and hierarchical clustering to identify subgroups. Ordered logistic regression was employed to examine the associations between periodontitis severity and microbial alpha diversity, beta diversity clusters, and bacterial genus abundances.
Results
Higher alpha diversity was positively associated with greater periodontitis severity (OTU richness: OR = 1.01, 95%CI: 1.01–1.02; FPD: OR = 1.21, 95%CI: 1.19–1.24; SWI: OR = 2.03, 95%CI: 1.81–2.30; ISI: OR = 16.61, 95%CI: 1.75–157.59). Beta diversity analyses revealed significant community compositional differences across periodontitis severity levels (PERMANOVA, all p = 0.001). Hierarchical clustering identified distinct microbial subgroups with varying odds of belonging to more severe periodontitis levels. Eighty‐two bacterial genera were significantly associated with periodontitis, including taxa showing strong positive associations consistent with the red and orange complexes (e.g., Porphyromonas, Tannerella, Treponema) and taxa exhibiting negative associations involving health‐associated genera (e.g., Rothia, Veillonella).
Conclusions
Salivary microbiome diversity, particularly higher alpha diversity and specific beta diversity clusters, was associated with periodontitis severity, supporting its potential relevance as a non‐invasive indicator of periodontal health status.
Plain language summary
This national study examined the relationship between the types and quantities of microbes present in saliva and the severity of periodontitis in more than 5,300 U.S. adults. We found that the more diverse the microbial community in a person's saliva—meaning a greater variety and number of different types—the more severe their periodontitis tended to be. We also found that the composition of these microbes differed depending on the severity of the periodontitis. Specifically, certain clusters of microbes, including well‐known disease‐causing bacteria such as Porphyromonas and Treponema (often called the red and orange complexes), were strongly associated with more severe periodontitis. Conversely, other common, generally healthy bacteria were less abundant. These findings suggest that analyzing the diversity and composition of microbes in saliva could be a valuable, noninvasive way to check for and potentially monitor the severity of periodontitis.
Keywords: alpha diversity, beta diversity, NHANES, oral microbiome, periodontitis
1. INTRODUCTION
Periodontitis is a chronic inflammatory disease linked to microbial dysbiosis within dental plaque biofilms, leading to progressive destruction of the periodontal supporting tissues. 1 The composition of supra‐ and subgingival plaque microbiota differs between individuals with and without periodontitis, as periodontitis patients exhibit higher levels of Porphyromonas gingivalis, Tannerella forsythia, and Treponema denticola in both supra‐ and subgingival plaque. 2 Based on DNA‐DNA hybridization techniques, Socransky et al. identified specific periodontal pathogenic species contributing to the development of periodontitis, including the red and orange complexes as key contributors. 3
With the advent of high‐throughput bacterial sequencing technologies, 16S rRNA sequencing has emerged as a powerful approach to characterize the diversity and dynamics of the oral microbiota and subgingival microbiota with higher taxonomic resolution. 4 Approximately 774 species‐level taxa have been identified in the oral microbiome, which collectively maintain a beneficial equilibrium to contribute to immune homeostasis. 5 This has shifted the focus from the specific plaque hypothesis to the polymicrobial synergy and dysbiosis model. 6 According to this model, the dysbiosis of the oral microbial community generates microenvironments that favor inflammophilic periodontal pathogens and trigger host inflammatory responses. 7
Although subgingival biofilm dysbiosis has been proven to be significantly associated with periodontitis, it remains unclear whether salivary microbial imbalance is related to periodontal health status. 8 The National Health and Nutrition Examination Survey (NHANES) 2009–2012 oral microbiome dataset represents the first nationally representative oral microbiome sequencing resource with a large sample size, enabling population‐level inferences beyond those of earlier initiatives such as the Human Microbiome Project (HMP). 9 These data provide an unprecedented opportunity to examine salivary microbial diversity in relation to periodontitis.
The aim of this present study was to assess the association between salivary microbiome diversity and the severity of periodontitis in a nationally representative sample.
2. METHODS
2.1. Study design and participants
Data for this study were obtained from the NHANES, conducted by the U.S. Centers for Disease Control and Prevention (CDC). The NHANES protocol was approved by the National Center for Health Statistics (NCHS) Institutional Review Board, and all participants provided written informed consent. The NHANES data used in this study covered the period from 2009 to 2012. Participants were included if they had (1) complete periodontal examination data and (2) available oral microbiome data. Among the 20,293 individuals who participated in the NHANES 2009–2012, 13,203 individuals were excluded due to incomplete periodontal examination data, and 1,767 individuals were excluded due to unavailable oral microbiome data. The final analytic sample consisted of 5,323 eligible participants, as shown in Figure 1.
FIGURE 1.
Flowchart of the included subjects.
2.2. Oral microbiome diversity
Oral rinse samples collected in NHANES 2009–2012 were used to characterize the oral microbiome. Participants rinsed with mouthwash for 5 seconds and gargled for another 5 seconds, repeating the procedure three times before expectorating into a sterile collection cup. DNA was extracted using the Puregene DNA purification kit, followed by PCR amplification and sequencing of the V4 region of the 16S rRNA gene. Sequencing reads were processed with DADA2 and QIIME to generate amplicon sequence variants (ASVs) and to assign taxonomic identities.
Oral microbiome data were obtained from the NHANES Oral Microbiome Project (OMP). The ASV tables were processed and annotated using the Human Oral Microbiome Database (HOMD) as a reference. Rarefied genus‐level abundance tables were then used for all subsequent analyses.
Alpha diversity, representing the richness and abundance of microbial species within a sample, was quantified using four metrics: operational taxonomic unit (OTU) richness, Faith's phylogenetic diversity (FPD), the Shannon‐Weiner index (SWI), and the inverse Simpson index (ISI). OTU richness and FPD reflect species richness, while SWI and ISI account for both richness and evenness. All metrics were calculated using a rarefaction depth of 10,000 reads, to ensure sufficient species richness and a large enough sample size were captured.
Beta diversity, reflecting differences in microbial community composition between samples, was assessed using Bray–Curtis, weighted UniFrac, and unweighted UniFrac distances.
2.3. Diagnosis of periodontitis
Periodontitis was classified according to the 2012 CDC/AAP case definitions for population‐based surveillance. 10 Mild periodontitis was defined as the presence of at least two interproximal sites (not on the same tooth) with clinical attachment loss (CAL) of at least 3 mm and at least two interproximal sites with a probing depth (PD) of at least 4 mm, or one site with a PD of at least 5 mm; moderate periodontitis was defined as having at least two interproximal sites (not on the same tooth) with CAL of at least 4 mm or having at least two interproximal sites with PD of at least 5 mm; severe periodontitis was defined as having at least two interproximal sites with CAL of at least 6 mm and at least one interproximal site with PD of at least 5 mm. No periodontitis was defined as no evidence of mild, moderate, or severe periodontitis.
2.4. Covariates
Covariates were selected based on established risk factors for periodontitis reported in previous literature, including age, sex, race/ethnicity, body mass index (BMI), educational level, household income, diabetes mellitus (DM), hypertension, smoking, and alcohol consumption. 11 Sex was categorized as male or female; race/ethnicity was divided into non‐Hispanic Whites, non‐Hispanic Blacks, Hispanic Americans, other Hispanics, and other races; BMI was calculated as weight (kg) divided by height (m) squared; educational level was categorized into less than high school, high school graduation, or above high school; annual household income was divided into less than $25,000, $25,000–$34,999, $35,000–$74,999, and $75,000 or more; DM and hypertension were defined as self‐reported doctor's diagnosis; the frequency of alcohol intake (times per day) was derived from the NHANES dietary questionnaire and categorized into < 2, 2–5, and > 5 times/day according to the data distribution; and smoking status was categorized as never, former, or current.
2.5. Statistical analysis
Descriptive statistics were reported as mean and standard deviation (SD) for continuous variables and as frequency and percentage for categorical variables. Group differences were assessed using t‐tests for continuous variables and chi‐square tests for categorical variables.
To analyze associations of alpha and beta microbiome diversity with periodontitis severity, three ordered logistic regression models were constructed. Model 1 adjusted for no covariates; Model 2 adjusted for age, sex, race/ethnicity, education level, household income, and BMI; Model 3 adjusted for age, sex, race/ethnicity, education level, household income, BMI, DM, hypertension, alcohol consumption, and smoking status. Alpha diversity indices were analyzed as continuous variables, categorical quartiles, and linear trends across quartiles (P for trend).
For beta diversity, three distance matrices (Bray–Curtis, weighted UniFrac, and unweighted UniFrac) were used in separate hierarchical clustering analyses to classify individuals into subgroups based on microbial composition. The optimal number of clusters (four) was determined using the Elbow method. Additionally, the three beta diversity distance matrices were visualized as principal coordinates analysis (PCoA) ordination plots, and group differences across periodontitis severity levels were tested using PERMANOVA.
Differential genera were identified using ordered logistic regression models implemented with the R Package MASS, adjusting for potential confounders as in Model 3. False discovery rate (FDR) correction using the Benjamini–Hochberg method was employed to account for multiple comparisons.
All statistical tests were two‐tailed, and p < 0.05 was considered statistically significant.
3. RESULTS
3.1. Baseline characteristics
Table 1 summarizes the characteristics of the 5,323 participants. The mean age was 47.51 years, and 50.0% were female. Non‐Hispanic White participants comprised the largest race/ethnicity group (66.0%). The mean BMI was 29.13. In terms of education, 16.01% of participants had less than a high school education, while 63.56% had attained education beyond high school. Household income varied, with the largest proportion (37.17%) reporting an annual income over $75,000. Additionally, 12.33% had DM and 35.17% had hypertension. Smoking status was distributed as never (56.21%), former (24.35%), and current smokers (19.42%). All covariates showed statistically significant differences between participants with and without periodontitis (p < 0.05).
TABLE 1.
Baseline characteristics of the included participants.
| Variable | None | Mild | Moderate | Severe | p |
|---|---|---|---|---|---|
| n (%) | 2551 (56.84) | 369 (6.51) | 1725 (27.81) | 678 (8.85) | |
| Age (years) | 45.36 (0.38) | 46.02 (0.70) | 51.01 (0.41) | 51.38 (0.48) | < 0.01 |
| 30–39 | 1070 (38.31) | 157 (37.16) | 346 (19.99) | 76 (13.19) | <0.01 |
| 40–49 | 721 (31.02) | 116 (31.60) | 426 (25.77) | 189 (33.30) | |
| 50–59 | 447 (20.06) | 66 (22.21) | 506 (31.76) | 240 (35.42) | |
| ≥ 60 | 313 (10.61) | 30 (9.04) | 447 (22.49) | 173 (18.09) | |
| Sex | <0.01 | ||||
| Female | 1556 (58.41) | 165 (42.41) | 765 (42.94) | 180 (23.91) | |
| Male | 995 (41.59) | 204 (57.59) | 960 (57.06) | 498 (76.09) | |
| Race/Ethnicity | <0.01 | ||||
| Mexican American | 303 (5.84) | 87 (13.39) | 346 (12.33) | 163 (15.89) | |
| Non‐Hispanic Black | 463 (8.68) | 86 (15.02) | 466 (14.44) | 208 (20.28) | |
| Non‐Hispanic White | 1216 (73.87) | 131 (59.15) | 510 (57.47) | 156 (47.32) | |
| Other Hispanic | 267 (4.90) | 40 (7.59) | 200 (6.71) | 60 (5.66) | |
| Other Races | 302 (6.71) | 25 (4.85) | 203 (9.05) | 91 (10.85) | |
| BMI | 28.86 (0.16) | 30.37 (0.58) | 29.47 (0.26) | 28.86 (0.29) | 0.02 |
| Education level | <0.01 | ||||
| Below high school | 382 (9.20) | 90 (14.53) | 569 (25.03) | 278 (32.73) | |
| High school | 449 (16.71) | 92 (26.56) | 407 (23.69) | 180 (28.84) | |
| Above high school | 1718 (74.09) | 187 (58.91) | 745 (51.28) | 220 (38.43) | |
| Family income (dollars) | <0.01 | ||||
| Under 25000 | 559 (13.60) | 102 (21.58) | 598 (27.10) | 264 (33.97) | |
| 25000–35000 | 314 (9.54) | 70 (17.07) | 263 (13.68) | 115 (17.80) | |
| 35000–75000 | 683 (29.02) | 99 (29.35) | 462 (32.16) | 173 (32.94) | |
| 75000 and over | 924 (47.85) | 84 (32.00) | 316 (27.06) | 81 (15.30) | |
| Diabetes mellitus | <0.01 | ||||
| No | 2228 (91.71) | 314 (85.22) | 1333 (82.54) | 494 (78.67) | |
| Yes | 289 (8.29) | 52 (14.78) | 388 (17.46) | 183 (21.33) | |
| Hypertension | <0.01 | ||||
| No | 1744 (69.75) | 240 (64.39) | 937 (58.49) | 347 (53.45) | |
| Yes | 807 (30.25) | 129 (35.61) | 787 (41.51) | 331 (46.55) | |
| Frequency of alcohol intake (per day) | <0.01 | ||||
| < 2 | 667 (38.95) | 80 (29.96) | 329 (33.12) | 93 (16.77) | |
| 2–5 | 940 (54.72) | 140 (58.11) | 531 (51.42) | 239 (64.67) | |
| > 5 | 135 (6.33) | 33 (11.93) | 166 (15.47) | 95 (18.56) | |
| Smoking status | <0.01 | ||||
| Former | 532 (23.84) | 63 (18.58) | 418 (25.56) | 170 (28.17) | |
| Never | 1667 (64.41) | 229 (62.66) | 848 (45.74) | 255 (31.74) | |
| Current | 352 (11.75) | 77 (18.76) | 457 (28.71) | 253 (40.09) |
Abbreviation: BMI, body mass index.
3.2. Association analyses of alpha diversity with periodontitis severity
Alpha diversity dispersion comparisons among participants without periodontitis and those with three levels of periodontitis severity are shown in Figure S1 in the online Journal of Periodontology. All alpha diversity indices, whether analyzed as continuous variables, categorical quartiles, or linear trends, showed positive associations with periodontitis severity in the fully adjusted model (Model 3), as shown in Table 2. Continuous OTU richness exhibited a weak positive but significant correlation with periodontitis (OR: 1.01, 95%CI: 1.01–1.02), while continuous FPD and SWI showed stronger correlations (OR: 1.21, 95%CI: 1.19–1.24; OR: 2.03, 95%CI: 1.81–2.30). In addition, ISI exhibited a particularly strong positive association with periodontitis (OR: 16.61, 95% CI: 1.75–157.59). The wide 95%CI for ISI is due to its low variance in this study, which results in unstable coefficient estimates rather than true measurement imprecision. Participants in the highest quartile (Q4) of all indices had significantly higher periodontitis severity compared with those in the lowest quartile (Q1). Trend analyses further confirmed significant linear associations of four indices with periodontitis severity. Overall, increasing microbial richness and evenness were associated with a greater periodontitis severity.
TABLE 2.
Ordered logistic regressions of alpha diversity indices and periodontitis.
| Model 1 | Model 2 | Model 3 | ||||
|---|---|---|---|---|---|---|
| OR (95%CI) | p | OR (95%CI) | p | OR (95%CI) | p | |
| OTU | 1.01 (1.01, 1.02) | <0.01 | 1.02 (1.01, 1.02) | <0.01 | 1.01 (1.01, 1.02) | <0.01 |
| Q1 | Ref | Ref | Ref | |||
| Q2 | 1.23 (1.06, 1.43) | <0.01 | 1.28 (1.08, 1.51) | <0.01 | 1.38 (1.11, 1.71) | <0.01 |
| Q3 | 2.21 (1.91, 2.56) | <0.01 | 2.39 (2.02, 2.82) | <0.01 | 2.43 (1.96, 3.01) | <0.01 |
| Q4 | 4.85 (4.17, 5.64) | <0.01 | 5.17 (4.35, 6.16) | <0.01 | 4.97 (3.98, 6.22) | <0.01 |
| P trend | <0.01 | <0.01 | <0.01 | |||
| FPD | 1.21 (1.19, 1.23) | <0.01 | 1.22 (1.20, 1.24) | <0.01 | 1.21 (1.19, 1.24) | <0.01 |
| Q1 | Ref | Ref | Ref | |||
| Q2 | 1.30 (1.12, 1.52) | <0.01 | 1.28 (1.08, 1.52) | <0.01 | 1.28 (1.03, 1.60) | 0.02 |
| Q3 | 2.30 (1.98, 2.67) | <0.01 | 2.42 (2.04, 2.86) | <0.01 | 2.51 (2.02, 3.12) | <0.01 |
| Q4 | 5.32 (4.57, 6.19) | <0.01 | 5.48 (4.62, 6.53) | <0.01 | 5.11 (4.09, 6.39) | <0.01 |
| P trend | <0.01 | <0.01 | <0.01 | |||
| SWI | 2.06 (1.90, 2.24) | <0.01 | 1.93 (1.77, 2.12) | <0.01 | 2.03 (1.81, 2.30) | <0.01 |
| Q1 | Ref | Ref | Ref | |||
| Q2 | 1.29 (1.11, 1.50) | <0.01 | 1.24 (1.05, 1.46) | <0.01 | 1.24 (1.00, 1.53) | 0.04 |
| Q3 | 1.76 (1.52, 2.04) | <0.01 | 1.67 (1.42, 1.96) | <0.01 | 1.76 (1.43, 2.17) | <0.01 |
| Q4 | 3.63 (3.14, 4.21) | <0.01 | 3.23 (2.74, 3.81) | <0.01 | 3.22 (2.61, 3.98) | <0.01 |
| P trend | <0.01 | <0.01 | <0.01 | |||
| ISI | 102.32 (39.21, 276.63) | <0.01 | 9.62 (1.41, 65.35) | 0.02 | 16.61 (1.75, 157.59) | 0.02 |
| Q1 | Ref | Ref | Ref | |||
| Q2 | 1.17 (1.02, 1.36) | 0.03 | 1.04 (0.88, 1.22) | 0.63 | 1.03 (0.83, 1.26) | 0.81 |
| Q3 | 1.36 (1.18, 1.57) | <0.01 | 1.18 (1.00, 1.38) | 0.04 | 1.16 (0.94, 1.42) | 0.15 |
| Q4 | 2.23 (1.93, 2.58) | <0.01 | 1.92 (1.64, 2.24) | <0.01 | 1.84 (1.50, 2.25) | <0.01 |
| P trend | <0.01 | <0.01 | <0.01 | |||
Abbreviations: CI, confidence interval; FPD, Faith's phylogenetic diversity; ISI, the inverse Simpson index (ISI); OR, odds ratio; OTU, operational taxonomic unit richness; Q1, the first quartile; Q2, the second quartile; Q3, the third quartile; Q4, the fourth quartile; SWI, the Shannon‐Weiner index.
Model 1 was the crude model.
Model 2 was adjusted for age, sex, race/ethnicity, body mass index, education level, and family income.
Model 3 was adjusted for age, sex, race/ethnicity, body mass index, education level, family income, diabetes mellitus, hypertension, alcohol intake, and smoking status.
3.3. Association analyses of beta diversity with periodontitis severity
Beta diversity dispersion, assessed by distance‐to‐centroid comparisons among participants without periodontitis and those with three levels of periodontitis severity, is shown in Figure S2 in the online Journal of Periodontology. Hierarchical clustering analysis of the dissimilarity matrices identified four clusters, is shown in Figure S3 in the online Journal of Periodontology. As Table 3 shows, based on Bray–Curtis dissimilarity, Cluster C was significantly associated with higher odds of being classified into a more severe periodontitis category compared with Cluster A (OR = 1.60; 95%CI: 1.02–2.50). Weighted UniFrac‐based clustering identified a significant association between Cluster D and lower odds of belonging to a more severe category (OR = 0.79; 95%CI: 0.65–0.95). Unweighted UniFrac‐based clustering revealed that both Cluster B (OR: 1.94, 95%CI: 1.58–2.37) and Cluster C (OR: 2.78, 95%CI: 2.28–3.40) were significantly associated with higher odds of being assigned to more severe periodontitis levels.
TABLE 3.
Ordered logistic regressions of beta diversity clusters and periodontitis.
| Beta diversity | Model 1 | Model 2 | Model 3 | |||
|---|---|---|---|---|---|---|
| OR (95%CI) | p | OR (95%CI) | p | OR (95%CI) | p | |
| Bray–Curtis | ||||||
| Cluster A | Ref | Ref | Ref | |||
| Cluster B | 0.81 (0.70, 0.92) | <0.01 | 1.00 (0.86, 1.16) | <0.01 | 1.06 (0.84, 1.34) | 0.59 |
| Cluster C | 1.59 (1.34, 1.90) | <0.01 | 1.65 (1.36, 1.99) | <0.01 | 1.60 (1.02, 2.50) | 0.04 |
| Cluster D | 0.93 (0.82, 1.06) | 0.30 | 0.98 (0.85, 1.13) | <0.01 | 0.93 (0.70, 1.23) | 0.58 |
| Weighted UniFrac | ||||||
| Cluster A | Ref | Ref | Ref | |||
| Cluster B | 0.64 (0.53, 0.76) | <0.01 | 0.83 (0.66, 0.96) | 0.02 | 0.81 (0.63, 1.03) | 0.09 |
| Cluster C | 0.95 (0.83, 1.08) | 0.05 | 0.91 (0.79, 1.06) | 0.02 | 1.02 (0.85, 1.21) | 0.08 |
| Cluster D | 0.61 (0.53, 0.69) | <0.01 | 0.68 (0.58, 0.78) | <0.01 | 0.79 (0.65, 0.95) | 0.01 |
| Unweighted UniFrac | ||||||
| Cluster A | Ref | Ref | Ref | |||
| Cluster B | 1.98 (1.71, 2.29) | <0.01 | 1.95 (1.66, 2.28) | <0.01 | 1.94 (1.58, 2.37) | <0.01 |
| Cluster C | 3.05 (2.66, 3.15) | <0.01 | 2.71 (2.32, 3.17) | <0.01 | 2.78 (2.28, 3.40) | <0.01 |
| Cluster D | 0.75 (0.63, 0.91) | <0.01 | 0.85 (0.69, 1.04) | 0.11 | 0.95 (0.74, 1.23) | 0.74 |
Abbreviations: CI, confidence interval; OR, odds ratio; Ref, the reference group.
Model 1 was the crude model.
Model 2 was adjusted for age, sex, race/ethnicity, BMI, education level, and family income.
Model 3 was adjusted for age, sex, race/ethnicity, BMI, education level, family income, diabetes mellitus, hypertension, alcohol intake, and smoking status.
PCoA ordination plots based on Bray–Curtis, weighted UniFrac, and unweighted UniFrac distances revealed visible separations of microbial community structures across periodontitis severity groups. PERMANOVA tests confirmed that overall community composition differed significantly among the four groups for all three distance metrics (Figure 2, all p = 0.001).
FIGURE 2.
Principal coordinates analysis (PCoA) ordination plots dependent on beta diversity with PERMANOVA tests. (Top Panel) Bray–Curtis distance; (Middle Panel) Unweighted UniFrac distance; (Bottom Panel) Weighted UniFrac distance.
3.4. Association analyses of bacterial genera with periodontitis severity
The microbial profiles across the four periodontitis severity groups revealed significant shifts in the relative abundances of multiple bacterial genera (Figure 3, Table S1 in the online Journal of Periodontology). Association analyses identified 82 bacterial genera whose relative abundances were significantly correlated with periodontitis severity (p FDR < 0.05). A consistent enrichment of well‐recognized periodontopathogens was observed along this gradient. Genera belonging to the red and orange complexes, such as Treponema_2 (OR = 1.65, 95%CI: 1.58–1.75), Porphyromonas (OR = 1.39, 95%CI: 1.31–1.47), Tannerella (OR = 1.51, 95%CI: 1.41–1.62), Fusobacterium (OR = 1.43, 95%CI: 1.32–1.54), Prevotella (OR = 1.27, 95% CI 1.18–1.36), Parvimonas (OR = 1.43, 95% CI: 1.34–1.51), and Campylobacter (OR = 1.33, 95% CI: 1.23–1.44), and the Eubacterium nodatum group (OR = 1.37, 95% CI: 1.29–1.47) demonstrated strong positive associations with periodontitis severity. In addition, Filifactor (OR = 1.55, 95% CI 1.45–1.65), a Gram‐positive anaerobic rod from the Peptostreptococcaceae family, showed a similarly strong positive association.
FIGURE 3.
Differential genera associated with periodontal disease severity. The vertical axis represents phylum, and the horizontal axis represents genus. ***P fdr < 0.001; **P fdr < 0.01; *P fdr < 0.05.
In contrast, certain genera typically associated with oral health were notably reduced in more severe groups. For instance, Rothia (OR = 0.83, 95%CI: 0.77–0.89), Veillonella (OR = 0.85, 95%CI: 0.78–0.92), and the Lachnospiraceae_NK3A20_group (OR = 0.40, 95%CI: 0.23–0.67) exhibited strong negative associations.
These findings illustrate a structured ecological gradient characterized by the enrichment of red/orange complex pathogens and other emerging anaerobes, accompanied by the depletion of health‐associated commensals along the continuum of periodontitis severity.
3.5. Interaction and subgroup analyses
Interaction analyses identified a significant effect modification by both sex and BMI across multiple alpha diversity indices (all p for interaction < 0.05; Table S2 in the online Journal of Periodontology). Subgroup analyses further showed that participants with obesity (BMI > 28) exhibited stronger associations, particularly for the FPD and SWI indices, suggesting that metabolic status may amplify these associations (Table S3 in the online Journal of Periodontology). In addition, a clear sex‐specific pattern was observed, with markedly stronger associations among females than males, most notably for ISI, implying potential sex‐related differences in host susceptibility to microbial dysbiosis (Table S3).
4. DISCUSSION
To our knowledge, this is the first study to investigate the association between salivary microbiome diversity and periodontitis using a nationally representative oral microbiome dataset. Our findings demonstrate that periodontitis severity is strongly associated with oral microbiome diversity, particularly with increased alpha diversity and specific microbial clusters identified through beta diversity metrics. These results underscore the importance of maintaining microbial homeostasis for periodontal health.
Our study corroborates a previous study showing that patients with periodontitis exhibit higher subgingival alpha microbial diversity, both in richness and evenness, compared with those without periodontitis. 7 , 12 , 13 , 14 However, most earlier studies were limited by small sample sizes (< 100 individuals with periodontitis). 15 , 16 , 17 , 18 , 19 By utilizing NHANES data, we provide the first large‐scale population‐based evidence confirming a positive association between periodontitis and alpha diversity. Although saliva‐based microbiome profiling differs from subgingival sampling, salivary microbiota can reflect periodontitis‐associated microbial changes and periodontal treatment responses, 20 , 21 supporting its potential utility as a non‐invasive indicator of periodontal health.
In this study, OTU richness, FPD, SWI, and ISI were all positively associated with periodontitis severity, which is consistent with the microbial succession theory and ecological plaque hypothesis. 22 The observed increase in alpha diversity likely reflects ecological shifts driven by the expansion of diverse anaerobic taxa that are frequently associated with periodontal disease under inflammatory conditions, rather than the simple loss of health‐associated taxa. 7 This pattern suggests that the progression of periodontitis involves a community‐wide ecological shift toward a more complex but dysbiotic microbial network, rather than replacement by a few dominant pathogenic taxa. Notably, while our analysis identified several genera traditionally linked to periodontitis, it is important to recognize that the NHANES database provides taxonomic resolution at the genus level. Because individual species within the same genus can exhibit divergent roles—ranging from commensalism to high pathogenicity—the observed associations at the genus level should be interpreted as potential markers for dysbiosis that require further validation at the species level.
Beyond alpha diversity, beta diversity analyses further supported distinct microbial community structures across disease severity levels. PCoA ordination plots and PERMANOVA testing based on Bray–Curtis distance, unweighted Unifrac distance, and weighted Unifrac distance demonstrated significant separation in community composition among healthy, mild, moderate, and severe periodontitis groups. These results indicate that as periodontitis severity increases, the overall salivary microbial architecture becomes progressively distinct, suggesting ecological divergence of microbial communities. Hierarchical clustering of beta diversity dissimilarity matrices identified distinct microbial subgroups, among which certain clusters were significantly associated with greater odds of being classified into more severe periodontitis categories. This aligns with previous research showing that periodontal status is the strongest explanatory factor for variation in subgingival microbiome beta diversity, while age and tobacco use also contributed partially. 23 Such clustering patterns may represent microbial community compositions reflecting underlying host–microbiome ecological interactions. 24
At the taxonomic level, 82 bacterial genera were significantly associated with periodontitis severity.
Genera commonly assigned to the red and orange complexes, such as Porphyromonas, Tannerella, Treponema, Fusobacterium, and the Eubacterium nodatum group, showed strong positive associations with more severe periodontitis. These genera have been repeatedly implicated in periodontal dysbiosis at the community level, although pathogenicity is ultimately determined at the species or strain level. 3
In contrast, health‐associated or commensal taxa including Rothia, Veillonella, and members of the Lachnospiraceae_NK3A20_group exhibited negative associations, highlighting the loss of community stability and health‐related taxa as inflammation intensifies. 25 Together, these alpha and beta diversity findings illustrate a structured ecological gradient from balanced, diverse health‐associated communities to assemblages enriched in taxa associated with periodontal dysbiosis that underpin the pathogenesis of periodontitis.
Besides, interaction and subgroup analyses further identified BMI as a significant modifier of the relationship between alpha diversity and periodontitis, with stronger associations observed in individuals with higher BMI. This finding is biologically plausible and consistent with previous studies reporting increased salivary microbiome diversity and elevated systemic inflammation in obese individuals. 26 Obesity‐related metabolic and inflammatory alterations may amplify microbial dysbiosis and ecological instability, 27 thereby strengthening the link between microbial diversity and periodontal inflammation. These results highlight the importance of considering host metabolic status when interpreting the association between oral microbiome and disease and suggest that systemic metabolic conditions may shape the susceptibility and ecological response of the oral microbiome to periodontal dysbiosis. 28 In addition, our analyses also revealed a pronounced sex‐related modification of the association between salivary alpha diversity and periodontitis, with substantially stronger associations observed among females. Several factors may underlie these sex‐specific differences. Host characteristics known to influence the salivary microbiome differ substantially between men and women, including female‐specific physiological factors (e.g., age at menarche, menstrual cycle dynamics, pregnancy), as well as sex‐dependent variations in diet and oral hygiene behaviors (such as brushing frequency and flossing habits). 29 Furthermore, hormones, such as aldosterone, cortisone, and dehydroepiandrosterone, exhibited stronger causal enrichment for salivary and tongue dorsum microbiome composition in females than in males, 29 suggesting that sex‐specific hormonal–microbial interactions may contribute to the enhanced association between microbial diversity and periodontitis observed in women.
In all, our findings suggest that high alpha diversity can serve as an indicator of increased periodontitis severity. Similar to the subgingival microbiome dysbiosis index (SMDI), alpha diversity metrics may hold value for both the diagnosis and prognosis of periodontitis, warranting further validation. 30 From a translational perspective, salivary microbial diversity may hold promise as a non‐invasive ecological indicator for periodontal health assessment. Unlike molecular biomarkers such as salivary MMP‐8, which reflect host inflammatory responses, diversity‐based indices capture community‐level microbial imbalance and ecological instability. 31 Given that saliva sampling is simple, rapid, and patient‐friendly, diversity metrics derived from salivary sequencing could be incorporated into adjunctive diagnostic frameworks for large‐scale screening or early disease detection. In clinical settings, such indices could complement established molecular or microbial biomarkers by providing a holistic ecological context for interpreting oral health status. For instance, while molecular markers quantify host response intensity, microbial diversity reflects the underlying ecological dysbiosis that drives disease initiation. Future translational studies integrating salivary diversity metrics with molecular biomarkers and machine learning‐based prediction models may yield comprehensive, non‐invasive diagnostic tools capable of improving early identification, risk stratification, and monitoring of periodontitis.
There are several limitations to be declared. First, NHANES utilized salivary rinse samples rather than subgingival plaque, which may be more directly linked to disease progression. Nevertheless, salivary rinses also demonstrate potential value as a non‐invasive tool for monitoring periodontal health. Second, sequencing of the V1–V3 region could provide a more comprehensive assessment of oral microbial community ecology, but the reliance on V4 region sequencing in NHANES might limit taxonomic resolution at the species level. Third, the cross‐sectional design of this study precludes temporal analysis of the transition from the state of periodontal health to periodontitis, thereby limiting our ability to capture the dynamic ecological shifts in the oral microbiome over time. Finally, this study did not include biomarker exploration or validation analyses, such as predictive model construction or external cohort validation; therefore, the potential diagnostic utility of salivary microbial diversity remains to be confirmed in future research.
5. CONCLUSION
This nationally representative analysis demonstrates that higher salivary microbiome alpha diversity is positively associated with periodontitis severity and identifies microbial clusters, based on beta diversity metrics, that are linked to more severe disease. These findings underscore the potential utility of salivary microbiome diversity as a non‐invasive indicator for periodontal health assessment and as a risk stratification tool in periodontal research.
AUTHOR CONTRIBUTIONS
Rui Pu: Conceptualization; data curation; methodology, writing—original draft. Zhikang Wang: Conceptualization; data curation; formal analysis; writing—original draft. Yunxuan Chen: Formal analysis; investigation. Danhong Zhou: Data curation; investigation. Guoli Yang: Project administration; supervision; writing—review & editing. Zhiwei Jiang: Project administration; supervision; writing—review & editing. All authors approved the final version of the article and agreed to accept responsibility for the work.
CONFLICT OF INTEREST STATEMENT
The authors declare no conflicts of interest.
ETHICS STATEMENT
This study utilized publicly available, fully de‐identified data from the National Health and Nutrition Examination Survey (NHANES), which was approved by the National Center for Health Statistics Research Ethics Review Board, and written informed consent was obtained from all participants. According to institutional policy, secondary analyses of de‐identified public registries do not constitute human subjects research locally and are exempt from standard institutional IRB review.
Supporting information
Supporting Information
ACKNOWLEDGMENTS
We would like to thank all the participants, investigators, and members of the NHANES research group, and the National Center for Health Statistics, U.S. Centers for Disease Control and Prevention. This work was supported by the National Natural Science Foundation of China (Nos. 82370991 and 82571129).
Contributor Information
Guoli Yang, Email: guo_li1214@zju.edu.cn.
Zhiwei Jiang, Email: jzw0913@zju.edu.cn.
DATA AVAILABILITY STATEMENT
The data were derived from sources in the public domain: NHANES (https://www.cdc.gov/nchs/nhanes/index.html).
REFERENCES
- 1. Papapanou PN, Sanz M, Buduneli N, et al. Periodontitis: consensus report of workgroup 2 of the 2017 world workshop on the classification of periodontal and peri‐implant diseases and conditions. J Periodontol. 2018;89(Suppl 1):S173‐S182. [DOI] [PubMed] [Google Scholar]
- 2. Ximénez‐Fyvie LA, Haffajee AD, Socransky SS. Comparison of the microbiota of supra‐ and subgingival plaque in health and periodontitis. J Clin Periodontol. 2000;27:648‐657. [DOI] [PubMed] [Google Scholar]
- 3. Socransky SS, Haffajee AD, Cugini MA, Smith C, jr Kent RL. Microbial complexes in subgingival plaque. J Clin Periodontol. 1998;25:134‐144. [DOI] [PubMed] [Google Scholar]
- 4. Santi‐Rocca J, Martín‐García DF, Lorca‐Alonso I, et al. Microbial complexes in subgingival plaque: a bacterial meta‐taxonomic study. J Clin Periodontol. 2025;52:983‐998. [DOI] [PubMed] [Google Scholar]
- 5. Escapa IF, Chen T, Huang Y, Gajare P, Dewhirst FE, Lemon KP. New insights into human nostril microbiome from the expanded Human Oral Microbiome Database (eHOMD): a resource for the microbiome of the human aerodigestive tract. mSystems. 2018;3:e00187‐18. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Manoil D, Parga A, Bostanci N, Belibasakis GN, Microbial diagnostics in periodontal diseases. Periodontol 2000. 2024;95:176‐193. [DOI] [PubMed] [Google Scholar]
- 7. Abusleme L, Dupuy AK, Dutzan N, et al. The subgingival microbiome in health and periodontitis and its relationship with community biomass and inflammation. ISME J. 2013;7:1016‐1025. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Chew RJJ, Tan KS, Chen T, Al‐Hebshi NN, Goh CE. Quantifying periodontitis‐associated oral dysbiosis in tongue and saliva microbiomes‐an integrated data analysis. J Periodontol. 2025;96:55‐66. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Vogtmann E, Chaturvedi AK, Blaser MJ, et al. Representative oral microbiome data for the US population: the National Health and Nutrition Examination Survey. Lancet Microbe. 2023;4:e60‐e61. [DOI] [PubMed] [Google Scholar]
- 10. Eke PI, Page RC, Wei L, Thornton‐Evans G, Genco RJ. Update of the case definitions for population‐based surveillance of periodontitis. J Periodontol. 2012;83:1449‐1454. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Darby I. Risk factors for periodontitis & peri‐implantitis. Periodontol 2000. 2022;90:9‐12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Chen C, Hemme C, Beleno J, et al. Oral microbiota of periodontal health and disease and their changes after nonsurgical periodontal therapy. ISMJ J. 2018;12:1210‐1224. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Akatsu T, Souno H, Fujii A, Minegishi Y, Ota N, Yamashita Y. Characteristics of subgingival plaque microbiome in Japanese older adults with healthy gingiva. J Clin Periodontol. 2025;52(9):1314‐1326. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Duran‐Pinedo A, Solbiati J, Teles F, Teles R, Zang Y, Frias‐Lopez J. Long‐term dynamics of the human oral microbiome during clinical disease progression. BMC Biol. 2021;19:240. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Kistler JO, Booth V, Bradshaw DJ, Wade WG. Bacterial community development in experimental gingivitis. PLoS One. 2013;8:e71227. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Kirst ME, Li EC, Alfant B, et al. Dysbiosis and alterations in predicted functions of the subgingival microbiome in chronic periodontitis. Appl Environ Microbiol. 2015;81:783‐793. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Camelo‐Castillo A, Novoa L, Balsa‐Castro C, Blanco J, Mira A, Tomás I. Relationship between periodontitis‐associated subgingival microbiota and clinical inflammation by 16S pyrosequencing. J Clin Periodontol. 2015;42:1074‐1082. [DOI] [PubMed] [Google Scholar]
- 18. Griffen AL, Beall CJ, Campbell JH, et al. Distinct and complex bacterial profiles in human periodontitis and health revealed by 16S pyrosequencing. ISMJ J. 2012;6:1176‐1185. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Schincaglia GP, Hong BY, Rosania A, et al. Clinical, immune, and microbiome traits of gingivitis and peri‐implant mucositis. J Dent Res. 2017;96:47‐55. [DOI] [PubMed] [Google Scholar]
- 20. Jung JS, Kook JK, Park SN, et al. Salivary microbiota reflecting changes in subgingival microbiota. Microbiol Spectr. 2024;12:e0103024. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Kageyama S, Takeshita T, Asakawa M, et al. Relative abundance of total subgingival plaque‐specific bacteria in salivary microbiota reflects the overall periodontal condition in patients with periodontitis. PLoS One. 2017;12:e0174782. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Marsh PD. Microbial ecology of dental plaque and its significance in health and disease. Adv Dent Res. 1994;8:263‐271. [DOI] [PubMed] [Google Scholar]
- 23. Iniesta M, Chamorro C, Ambrosio N, Marín MJ, Sanz M, Herrera D. Subgingival microbiome in periodontal health, gingivitis and different stages of periodontitis. J Clin Periodontol. 2023;50:905‐920. [DOI] [PubMed] [Google Scholar]
- 24. Genco RJ, LaMonte MJ, McSkimming DI, et al. The subgingival microbiome relationship to periodontal disease in older women. J Dent Res. 2019;98:975‐984. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Abusleme L, Hoare A, Hong BY, Diaz PI. Microbial signatures of health, gingivitis, and periodontitis. Periodontol 2000. 2021;86:57‐78. [DOI] [PubMed] [Google Scholar]
- 26. Kim KS, Na HS, Oh TJ, et al. Oral microbiome changes in subjects with obesity following bariatric surgery compared to lean counterparts. Front Microbiol. 2025;16:1553404. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Verdam FJ, Fuentes S, de Jonge C, et al. Human intestinal microbiota composition is associated with local and systemic inflammation in obesity. Obesity (Silver Spring). 2013;21:E607‐615. [DOI] [PubMed] [Google Scholar]
- 28. Marruganti C, Suvan JE, D'Aiuto F. Periodontitis and metabolic diseases (diabetes and obesity): tackling multimorbidity. Periodontol 2000. Published online October 16 2023. [DOI] [PubMed] [Google Scholar]
- 29. Liu X, Tong X, Jie Z, et al. Sex differences in the oral microbiome, host traits, and their causal relationships. iScience. 2023;26:105839. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Chew RJJ, Goh CE, Lin XYS, et al. Prognostic impact of microbiome dysbiosis: a prospective study. J Clin Periodontol. 2025;52:215‐224. [DOI] [PubMed] [Google Scholar]
- 31. Rakic M, Calciolari E, Grant MM, Radovanovic S, Bostanci N, Preshaw PM. Host markers of periodontal diseases: meta‐analysis of diagnostic accuracy studies. J Clin Periodontol. 2025;52 (Suppl 29):155‐181. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supporting Information
Data Availability Statement
The data were derived from sources in the public domain: NHANES (https://www.cdc.gov/nchs/nhanes/index.html).