Metagenomic analysis of the denture-associated oral microbiome in patients with denture stomatitis

Abstract

Denture stomatitis (DS) is an inflammatory condition that affect denture wearers and is characterized by erythema of the mucosa opposing the denture. DS is often associated with oral microbiome dysbiosis. We used shotgun metagenomics to investigate the association between the denture-associated oral microbiome (DAOM) and DS in older adults living in long-term care facilities. We included participants with DS (n = 28) and age-and sex-matched removable denture wearers without signs of DS (n = 28). Clinical oral examinations were performed, and demographic and medical data were obtained from medical records. Median (interquartile range) age of participants was 88 (9) years; 75% were females. Beta diversity differed between the DS and non-DS groups (Bray-Curtis dissimilarity, p = 0.01; Jaccard index, p = 0.004). Two phyla, nine genera, and 15 species differed significantly between groups, with the genera Candida and Scardovia, and species Candida albicans, Aggregatibacter actinomycetemcomitans, and Scardovia inopinata being enriched in DS. Network analysis revealed strongly interconnected microbial communities and more prominent bacterial-fungal co-occurrence in DS than in non-DS. These findings indicate that DS is associated with significant alterations in the DAOM, which may contribute to inflammation. Microbiome-targeted strategies are needed for the management of DS.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-16915-4.

Introduction

The oral cavity is a unique environment for microbial colonization, as it includes a combination of soft tissue and hard surfaces and a direct connection to the external environment outside the body. The non-shedding surface of the denture can be an ideal platform for microbial colonization. Accordingly, 15–70% of denture wearers are affected by denture stomatitis (DS), a common mucosal infection^1,2. DS is characterized by inflammation and erythema of the oral mucosa underneath a denture².

The aetiology of DS is multifactorial and is associated with various local factors, such as local mucosal trauma caused by an ill-fitting denture, continuous use of dentures, dry mouth, age of dentures, poor denture hygiene, contamination with microbial biofilm, a carbohydrate-rich diet, acidic salivary pH, and smoking^1,3–7. Additionally, microbial infection due to inadequate denture or oral hygiene contributes to DS development. Systemic factors that weaken disease resistance by lowering immunity further predispose individuals to DS⁸. DS can be symptomatic (such as a burning or salty sensation in the mouth) or asymptomatic².

The primary microorganism associated with the development of DS is yeast belonging to the genus Candida². However, species of the genus Candida also have potential interactions with key oral taxa. These interactions may shape the community structure and shift from health to disease opportunistically⁹. In vitro, the inclusion of Porphyromonas gingivalis in mixed-species biofilms reduced the virulence of Candida albicans^10,11. Interestingly, Aggregatibacter actinomycetemcomitans and P. gingivalis, species that were thought to disappear after removal or loss of all natural teeth, were detected from supragingival biofilm samples from dentures of edentulous subjects¹².

Sequencing approaches have revealed that DS development is more complex than a candidal infection. Microbial biofilms obtained from complete denture wearers with and without DS and analysed using 16 S rDNA sequencing revealed 27 bacterial species in both biofilms; 29 species were exclusively present in patients with DS and 26 only in healthy subjects¹³. Candida spp. detected by PCR were present in both biofilms. A significant decrease in the number of unique bacterial species has been observed in tongue samples of DS compared with non-DS analysed by 16 S rRNA sequencing, suggesting a reduction in biodiversity resulting in dysbiosis¹¹.

Although micro-organisms live in complex communities on the surface of the removable denture, the diversity is reduced in denture plaque than in dental plaque¹⁴. The presence of natural teeth has a significant effect on the overall oral microbial composition, and even the presence of a single tooth is sufficient to profoundly affect the microbiome composition and create a more pathogenic biofilm than when natural teeth are absent¹⁵. Wearing a removable partial denture seems to influence the oral ecosystem, changing it from a healthy to a diseased state via the increased presence of pathogens in the oral microbial community¹⁶. Shi B et al. (2016)¹⁷ found that the phylotype composition of the bacterial communities colonizing the dentures and remaining teeth of the same individuals were largely reflective of each other. This suggests that the overall microbial load may have a greater impact on DS development than the actual microbial composition of the mucosa-facing denture plaque.

Older adults, who are the primary users of removable dentures, are often medically compromised and more susceptible to infections¹⁸. Previous studies investigating denture biofilms in patients with or without DS have mainly utilized 16 S rRNA gene sequencing typically with relatively small sample sizes. To our knowledge, shotgun metagenomic sequencing has not been previously applied to this research area. In this study, we analysed the association between the denture-associated oral microbiome (DAOM) and DS using shotgun metagenomics in a sample of denture-using, dentate and edentate older adults living in long-term care (LTC) facilities.

Materials and methods

Study design and subjects

The present study, the Finnish Oral Health Studies in Older Adults (FINORAL), is a sub-study of randomly selected participants from a previous nutrition study¹⁹ that included individuals living in LTC (nursing homes and assisted living facilities) in the Helsinki capital area, Finland. Participants in the nutrition study were recruited in March 2017 and included individuals aged ≥ 65 years (N = 550). The FINORAL study included data from 393 participants²⁰. For this study, we excluded participants who had used antibiotics within < 2 months prior to sample collection, those requiring prophylactic antibiotics, individuals with major deficiencies or who refused clinical examination, those who died between the end of the nutrition study and the start of FINORAL, those with missing data on antibiotic use, and those who did not have removable dentures, complete or partial. This resulted in a final study population of 56 participants.

A questionnaire completed by a registered nurse was administered to all participants. The questionnaire included demographic characteristics (age and sex) and mobility status. Data on antibiotic use and medical diagnoses were obtained from participant medical records.

Clinical oral examination

Two qualified dentists conducted the clinical oral examinations between September 2017 and January 2019. Examinations were performed either at the bedside or while the participant was seated on a chair, depending on the individual’s condition and mobility. The number of teeth was recorded during examination. Participants were classified as dentate (at least one visible tooth or root was present) or edentate (completely toothless). The use and condition of removable dentures were assessed. Denture types were categorized as partial (involving one or two jaws), complete (full dentures in one or two jaws), or both (a combination of partial and complete dentures). Denture conditions were evaluated and classified as usable, in need of minor repair, in need of major repair, or unusable. In addition, the oral mucosa was examined and categorized as healthy, with lesions unrelated to denture use, or with lesions related to denture use. A clinical estimation of oral wetness was also recorded to assess the dryness of the oral cavity²¹.

Participants (N = 56) were divided into two groups for this study. We included 28 patients with DS based on published diagnostic criteria of the Newton classification (Type 1, localized inflammation or pinpoint hyperaemia; Type 2, more diffuse erythema [redness] involving part or all of the mucosa which is covered by the denture; and Type 3, inflammatory nodular/papillary hyperplasia)²² and 28 age- and sex-matched individuals who were denture wearers without signs of DS (hereafter called ‘non-DS’).

Microbiome sample collection and nucleic acid extraction

After the intraoral examination, microbial sampling was performed from the fitting surface of the denture using a sterile curette. A dentist used the curette to scratch the pink acrylic of a removable denture. Scratching was performed for approximately 5 s to ensure adequate biofilm collection. Samples were pooled into 1.5 mL microcentrifuge tubes containing PCR-grade water. The collected samples were initially frozen at − 20 °C and later transferred to − 80 °C until DNA extraction.

A total of 500 µL of microbial samples in microcentrifuge tubes were mixed with 500 µL of lysis buffer in NucleoSpin Bead Tubes Type B (Macherey-Nagel). Metagenomic DNA was purified using the Chemagic DNA Blood 400-H96 Kit, optimized for the Chemagic™ 360 instrument (PerkinElmer), with the Chemagic Saliva600 pre-filled protocol following the manufacturer’s instructions. DNA was diluted in elution buffer (PerkinElmer), and DNA concentrations were determined fluorometrically via a Qubit dsDNA Broad Range Assay Kit (Invitrogen, Carlsbad, CA, USA) with a DeNovix DS-11 FX + Fluorometer (DeNovix Inc., Wilmington, DE, USA). Libraries for next-generation sequencing were prepared using the NEBNext^® Ultra™ II FS DNA Library Prep Kit.

Sequencing and metagenomic data processing

Shotgun sequencing was performed on an Illumina NovaSeq 6000 instrument (Illumina, USA) with 150-bp paired-end sequencing. Metagenomic raw sequence quality was assessed using FastQC (v0.12.1)²³ and summarized with MultiQC (v1.14)²⁴. We filtered metagenomic reads to remove Illumina adapter sequences and low-quality bases using Trimmomatic (v0.39)²⁵. Host-associated reads were removed using KneadData (v0.12.0) (https://github.com/biobakery/kneaddata), with the default human genome reference database based on hg37. Processed reads were then mapped to a reference database using Kraken2 (v2.1.3)²⁶ for taxonomic classification. Species-level abundance estimation was performed using Bracken (v3.0)²⁷. Methods for taxonomic annotation have been described previously^28,29.

Statistical analysis and data visualization

Statistical analyses were performed using R (v. 4.2.2) and IBM SPSS Statistics (v.29.0.2.0). Continuous variables were summarized using median and interquartile range (IQR). Categorical variables were presented as counts and percentages. Mann-Whitney U test was used for continuous variables and Pearson χ² test was used for categorical variables. Before microbiome analysis, we removed samples with fewer than 1000 reads and standardized the sequencing by rarefying all samples to the same number of reads. We also retained only the taxa that appeared in at least 10% of the samples (prevalence threshold) to exclude rare or spurious taxa.

Alpha diversity was assessed using Chao1, Shannon, and Inverse Simpson indices via the phyloseq (v1.46) package in R³⁰. Differences between the DS and non-DS groups were evaluated using non-parametric Wilcoxon tests. Boxplots were generated using the ggplot2 package (v3.5.2)³¹ in R to visualize the diversity index distributions. Beta diversity was assessed using Bray-Curtis and Jaccard dissimilarity metrics. Permutational multivariate analysis of variance (PERMANOVA) was performed using the adonis2 function from the vegan package (v2.6.4)³² in R with 999 permutations to test for significant differences between groups. Principal coordinate analysis (PCoA) was performed using the vegan package (v2.6.4) to visualize the spatial separation between the DS and non-DS groups.

To perform differential abundance analysis, we used ANCOM-BC2³³ to identify differences in taxa at the phylum, genus, and species levels between the DS and non-DS groups. We adjusted for potential confounding factors including sex, age, and dentate or edentate status. P-values were adjusted using false discovery rate (FDR) to account for multiple comparisons; a p-value < 0.05 was considered statistically significant. A multivariate logistic regression model was constructed to assess the association between DS status and the relative abundances of significant microbial taxa. Model performance was evaluated using the area under the receiver operating characteristic curve (AUC).

For network analysis, the phyloseq object was converted to a microeco-compatible format. Correlations between microbial taxa were calculated using Spearman’s correlation followed by construction of the co-occurrence network using the microeco³⁴ R package (v1.14.0). The network was optimized with a p-value threshold of 0.01 and a correlation coefficient threshold of 0.9. Modules within the network were identified using the igraph package (v2.0.2). Chord diagrams were generated using the top phyla, with links representing their connections, and were visualized using the circlize (v0.4.16)³⁵ R package.

Results

Characteristics of the study participants

A total of 56 participants (28 with DS and 28 non-DS) were included in this study; 75% of the participants were female. Clinical and demographic characteristics of the study participants are presented in Table 1. Median (IQR) age was similar in the DS and non-DS groups [DS 87.5 (9) vs. non-DS 88 (9) years; p = 0.812]. Median (IQR) length of stay at the current facility was 24 (31) months, with no significant difference between groups (p = 0.684). Distribution of smoking status, diabetes, number of medications, cognitive and dementia status, mobility, and dietary habits were similar between the DS and non-DS groups.

Table 1.

Demographic and health characteristics of study participants.

	All n = 56	DS n = 28	Non-DS n = 28	p-value
	Median (IQR)
Age, years	88 (9)	87.5 (9)	88 (9)	0.812
Residency in current facility, months	24 (31)	27 (33)	24 (25)	0.684
	n (%)
Sex, female	42 (75)	21 (75)	21 (75)
Smoking				0.757
Never	27 (61.4)	14 (63.6)	13 (59.1)
Previous	17 (38.6)	8 (36.4)	9 (40.9)
Diabetes	11 (19.6)	8 (28.6)	3 (10.7)	0.093
Medications daily in use (months)				0.313
0	-	-	-
≤ 5	11 (19.6)	7 (25)	4 (14.3)
≥ 6	45 (80.4)	21 (75)	24 (85.7)
Dementia				0.330
No or mild	14 (26.9)	9 (32.1)	5 (20.9)
Moderate or severe	38 (73.1)	19 (67.9)	19 (79.1)
Cognition				0.659
Normal/near normal	30 (54.5)	16 (59.3)	14 (50)
Weakened	19 (34.5)	9 (33.3)	10 (35.7)
Very weak	6 (11.0)	2 (7.4)	4 (14.3)
Speech				0.345
Clear/understandable	44 (80)	23 (85.2)	21 (75)
Unclear or not speaking at all	11 (20)	4 (14.8)	7 (25)
Mobility				0.349
Independently or with aids	30 (54.5)	17 (60.7)	13 (48.1)
Supported or does not move	25 (45.5)	11 (39.3)	14 (51.9)
Diet				0.669
Ordinary	44 (83)	23 (85.2)	21 (80.8)
Soft	9 (17)	4 (14.8)	5 (19.2)
Eating				0.979
Independently or guided	53 (96.4)	27 (96.4)	26 (96.3)
Fully assisted	2 (3.6)	1 (3.6)	1 (3.7)
Dryness of mouth				0.433
No signs of dryness/normal	12 (22.6)	5 (19.2)	7 (25.9)
Somewhat dry	25 (47.2)	11 (42.3)	14 (51.9)
Dry	16 (30.2)	10 (38.5)	6 (22.2)

DS was more common among dentate (53.6%) than edentate (46.4%) study participants (p = 0.032). No significant differences were found in denture cleaning, daily oral hygiene, night-time wearing of dentures, denture type, or denture condition between groups. Dentate participants with DS had more residual roots (p = 0.048), higher gingival index (GI) value (p = 0.048), and gingivitis and mild periodontitis more frequently (p = 0.009) than dentate non-DS participants. No participants had moderate or severe periodontitis (Table 2).

Table 2.

Dental and periodontal characteristics of study participants.

	All n = 56	DS n = 28	Non-DS n = 28	p-value
	n (%)
Dentate or edentate				0.032
Edentate/toothless	26 (46.4)	9 (32.1)	17 (60.7)
Dentate (at least one tooth or root visible)	30 (53.6)	19 (67.9)	11 (39.3)
Denture cleaning				0.187
By participant	16 (32.7)	10 (41.7)	6 (24)
Someone else	33 (67.3)	14 (58.3)	19 (76)
Daily oral hygiene				0.151
By participant	16 (30.8)	11 (39.3)	5 (20.8)
Assisted or by someone else	36 (69.2)	17 (60.7)	19 (79.2)
Denture type				0.196
Partial (1–2 jaws)	8 (14.5)	4 (14.8)	4 (14.3)
Complete (1–2 jaws)	41 (74.5)	18 (66.7)	23 (82.1)
Both	6 (10.9)	5 (18.5)	1 (3.6)
Night-time wearing of dentures				0.301
Yes	20 (39.2)	12 (46.2)	8 (32)
No	31 (60.8)	14 (53.8)	17 (68)
Denture condition				0.610
No need of repair	32 (60.4)	16 (64)	16 (57.1)
In need of repair	21 (39.6)	9 (36)	12 (42.9)
Clinical diagnosis of periodontitis				0.009
Healthy periodontium	8 (27.6)	2 (11.8)	6 (50)
Gingivitis	13 (44.8)	7 (41.1)	6 (50)
Mild periodontitis	8 (27.6)	8 (47.1)	–
Moderate to severe periodontitis	–	–	–
	Median (IQR)
N of teeth	6.5 (5)	3.5 (8)	0 (6)	0.061
Residual roots	0 (0)	0 (2)	0 (0)	0.048
Probing pocket depth teeth
4–5 mm	0 (2)	0 (2)	0 (0)	0.209
≥6 mm	0 (0)	0 (0)	0 (0)	0.821
Teeth with increased mobility	0 (1)	0 (2)	0 (0)	0.293
Bleeding on probing, %	86 (100)	100 (100)	50 (100)	0.136
Plaque index	1.3 (2.91)	2 (1.96)	1 (2.47)	0.325
Gingival index	1 (2)	1.72 (1.44)	0.65 (1.25)	0.048
Root caries	0 (1)	0 (1)	0 (1)	0.478
Coronal caries	0 (0)	0 (0)	0 (0)	0.821

Taxonomic composition of DAOM

The DAOM primarily consisted of Bacillota, followed by Actinomycetota and Bacteroidota. Other phyla with moderate representation included Pseudomonadota, Fusobacteriota, and Ascomycota (Fig. S1a). Phyla accounting for < 1% of the total community included Candidatus Saccharibacteria, Spirochaetota, and Synergistota.

The dominant genera were Streptococcus, followed by Actinomyces, Rothia, Veillonella, Prevotella, Limosilactobacillus, Lactobacillus, Corynebacterium, Staphylococcus, and Schaalia (Fig. S1b).

Microbial diversity

All three alpha diversity indices were slightly higher in the DS group than in the non-DS group. The mean (SD) values were 659 (330) for Chao1, 2.72 (0.75) for Shannon index, and 8.79 (5.22) for Inverse Simpson index in the DS group. No significant differences were observed between groups (Fig. 1a; Table S1).

Fig. 1.

Alpha- and beta-diversity of the DAOM. (a) Alpha diversity indices (Chao1, Shannon, Inverse Simpson) comparing individuals with and without clinically diagnosed DS. (b) Beta diversity illustrated by Principal Coordinate Analysis (PCoA) based on Bray-Curtis and Jaccard distances, showing microbial community differences between DS and non-DS groups.

Beta diversity analyses showed significant differences in microbial community composition between DS and non-DS groups (Bray-Curtis dissimilarity: R²=0.0397, F = 2.23, adjusted p = 0.01; Jaccard index: R²=0.0324, F = 1.807, adjusted p = 0.004; Table S2). PCoA showed distinct clustering by stomatitis status (Fig. 1b).

DS signatures in DAOM composition

Although the relative abundance of the top abundant phyla and genera were similar between the DS and non-DS groups (Fig. S2), notable differences were observed, including variations in microbial taxa at the phylum, family, and genus levels within bacteria (Fig. 2a) and differences across other kingdoms, such as fungi, archaea, viruses, and protists (Fig. 2b).

Fig. 2.

Taxonomic differences in the DAOM between DS and non-DS groups. (a) Stacked bar plots showing mean relative abundance of microbial taxa at the phylum, family, and genus levels within the bacterial kingdom. (b) Stacked bar plots for other microbial kingdoms (including fungi, archaea, viruses, and protists) comparing individuals with clinically diagnosed DS and non-DS.

Two phyla and nine genera showed differences in abundance between groups, including a higher abundance of Candida (LFC = 2.861, p = 0.006), Scardovia (LFC = 2.313, p = 0.004), and Schleiferilactobacillus (LFC = 2.103, p = 0.029) and a lower abundance of Sphingomonas (LFC = − 1.312, p = 0.022) and Moraxella (LFC = − 1.367, p = 0.020) in the DS group than in the non-DS group (Fig. S3; Table S3).

Species-level analysis revealed that 15 species differed significantly between the DS and non-DS groups (Fig. S3). A. actinomycetemcomitans (LFC = 3.331, p = 0.017), C. albicans (LFC = 2.518, p = 0.015), Scardovia inopinata (LFC = 2.313, p = 0.005), and Aggregatibacter sp. oral taxon 513 (LFC = 1.964, p = 0.017) were more abundant in the DS group than in the non-DS group. In contrast, the abundance of Lactobacillus amylovorus (LFC = − 1.421, p = 0.028), Corynebacterium argentoratense (LFC = − 1.434, p = 0.031), and Lactobacillus sp. JM1 (LFC = − 1.444, p = 0.021) were more abundant in the non-DS than in the DS group (Table S3).

A multivariate logistic regression model was performed using six species significantly more abundant in the DS group to predict DS disease status (Table S4). The overall model demonstrated good discriminatory ability (AUC 0.853, 95% CI: 0.753–0.954) (Fig. 3). Among the taxa, C. albicans showed a significant positive association with DS (p = 0.016). S. inopinata (p = 0.046) and Aggregatibacter sp. oral taxon 513 (p = 0.112) also showed borderline association.

Fig. 3.

The area under the ROC curve (AUC) for DAOM-based DS classification. The blue line shows the model’s sensitivity versus false positive rate across thresholds, with a shaded grey area representing the 95% confidence interval (CI) of the ROC curve estimated by bootstrapping. The dashed diagonal line indicates random classification.

Microbial network modularity and interconnectivity in DS and non-DS groups

Network analyses were employed to explore alterations in microbial community interactions. The comparison of microbial network properties between DS and non-DS groups is shown in Table S5. In DS, a greater degree of modularity (0.692 vs. 0.567) and a higher clustering coefficient (0.674 vs. 0.594) were observed, indicating more distinct and tightly organized communities. DS also exhibited more interconnected microbial networks, with higher edge counts (1436 vs. 1277) and average degree (13.17 vs. 12.28) compared with non-DS individuals. Interconnected microbial biomarkers with distinct profiles were identified between DS and non-DS groups (Fig. 4). The non-DS group had high interaction especially between Bacteroidota and Bacillota, Fusobacteriota and Bacillota, and Campylobacterota and Bacteroidota, which were reduced in the DS group. The DS group had higher interconnectivity within phylum Pseudomonadota and within Actinomycetota than the non-DS group. Additionally, Ascomycota interacting with the phylum Bacillota was among the most abundant taxa with co-occurrence in the DS but not in the non-DS group.

Fig. 4.

Cord diagrams illustrating microbial taxa with strong positive correlations, summarizing co-occurrence interactions among the most abundant taxa. These diagrams reflect high interconnectivity across dominant phyla in (a) individuals with DS and (b) those without DS. Colours represent different phyla, and the widths of the links indicate the relative strength of interactions.

At the phylum level, Actinomycetota (407 vs. 227), Pseudomonadota (145 vs. 43), and Ascomycota (7 vs. 1) were more involved in DS-associated interactions, whereas Bacillota (559 vs. 701), Bacteroidota (214 vs. 457), and Fusobacteriota (80 vs. 262) were more involved in non-DS interactions (Table S5).

Discussion

In this study, we employed shotgun metagenomics to characterize the multi-kingdom DAOM and DS in older adults residing in LTC facilities. DS was more common among dentate participants than among edentate participants. Beta diversity differed significantly between individuals with and without DS. Two phyla, nine genera, and 15 species showed significant differences between the DS and non-DS groups, with Candida spp. being more abundant in the DS group. To distinguish DS and non-DS groups, combining the significant abundances of bacterial and fungal species provided a clinically relevant AUC of 0.853 (0.753–0.954) which was higher than the AUCs obtained by any single species. Bacterial-bacterial interactions were reduced, while bacterial-fungal interactions were more prominent in individuals with DS, suggesting a possible interruption of microbiome structure and a pathogenic role of multi-kingdom microbial interactions in DS.

The microbial communities colonizing dentures can affect the oral health of denture wearers. Our study did not show any significant differences in microbial community evenness and richness (alpha diversity) at the species level between the DS and non-DS groups, consistent with previous findings¹⁷. However, we observed significant differences in microbial community composition, as measured by beta diversity, between the DS and non-DS groups. Similarly, a recent pilot study¹¹ using 16 S rRNA sequencing reported reduced microbial diversity on the tongue in individuals with DS, along with distinct differences in the microbiota of denture-fitting surfaces and palatal mucosa compared with healthy individuals.

The DAOM primarily consisted of Bacillota, Actinomycetota, Bacteroidota, Pseudomonadota, and Fusobacteriota, which is consistent with the bacterial phylotypes found in healthy denture wearers¹³. Although bacterial components were in substantially greater abundance, minor amounts of fungi, archaea, viruses, and protists were also observed in DAOM. Similarly, Yacob and colleagues (2024)³⁶ reported that the genera with the highest relative abundance in the DAOM were Streptococcus, Actinomyces, Rothia, Veillonella, and Prevotella. DAOM may also harbour non-oral pathogenic microbiota that could increase the risk of systemic infections³⁷. We observed that eucaryotes Candida, Nakaseomyces, Babesia, Saccharomyces, and Malassezia were the predominant non-bacterial genera present in the DAOM. Biofilm formation on restorative materials positively correlates with increased surface roughness and higher surface free energy^17,38.

Dentures are used by both completely edentulous and partially dentate individuals. Although the phenomenon of biofilm formation is similar for denture and dental plaque, the environment between the denture and the opposing oral mucosal surface are microbially distinct^15,39. Candida species, particularly C. albicans exhibit varying morphologies and adherence patterns on different denture base materials⁴⁰. Candida-associated DS is a common fungal infection in patients with removable dentures. Factors that promote Candida virulence include denture wear, poor oral hygiene, reduced salivary flow, high carbohydrate intake, and plaque accumulation. Cardash (1989)⁴¹ showed that C. albicans was found in 100% of denture wearers with DS and in 82% in those without DS, and that the concentrations were higher on the denture base than the palate. Candida spp. can form biofilms on denture surfaces, primarily regulated by quorum sensing mechanisms. Transcription factors such as Efg1 and Bcr1 play key roles in biofilm formation and virulence during the development of DS⁴². In the present study, fungal species were more abundant in DS patients, which together with earlier studies suggests that fungal overgrowth plays a major role in DS pathogenesis^43–45.

Bacterial species that were particularly abundant in DS patients included A. actinomycetemcomitans, C. breve, S. inopinata, Aggregatibacter sp. oral taxon 513, A. pacaensis, C. curtum, and Capnocytophaga sp. oral taxon 902. Among these, A. actinomycetemcomitans is associated with periodontitis. In a population-based study, A. actinomycetemcomitans was detected in 20% of participants but only in 2% of subjects without teeth⁴⁶. Yasui et al. (2012)⁴⁷. also observed a lower abundance of A. actinomycetemcomitans in edentulous individuals than in dentate individuals. Könönen et al. (1991)⁴⁸ reported that A. actinomycetemcomitans was not detected in edentulous individuals with complete dentures. However, Sachdeo et al. (2008)¹² detected A. actinomycetemcomitans from biofilm samples from denture teeth of edentulous subjects. A. actinomycetemcomitans produces several virulence factors including leukotoxin A and lipopolysaccharide, which contribute to its pathogenicity, surface adhesion, and interactions with other microbial species⁴⁹. However, this species is serologically and genetically highly heterogeneous, including both benign and virulent phenotypes⁵⁰. Cryobacterium is a genus of gram-positive, rod-shaped, aerobic bacteria and C. breve is a psychrophilic species that has not previously been reported as part of the human oral microbiota. Our findings likely reflect C. breve is a transient colonizer in dentures, possibly introduced through food, water, or other environmental sources, rather than a stable oral resident. Species abundant in non-DS individuals included L. amylovorus, C. argentoratense, and L. falkenbergense. Of these, L. amylovorus and L. falkenbergense are lactic acid bacteria and are not stable oral residents, likely reflecting transient introduction through the diet, with uncertain relevance to denture colonization.

The DAOM has previously been shown to consist of the main genera Streptococcus, Actinomyces, and Rothia¹⁴. Similarly, in the present study, the five most abundant genera of the DAOM were Streptococcus, Actinomyces, Rothia, Veillonella, and Prevotella. Therefore, the DAOM differs from the subgingival microbiome, which consists of the main genera Fusobacterium, Prevotella, Streptococcus, Veillonella, and Capnocytophaga⁵¹; the subgingival microbiome thus includes a higher proportion of anaerobic species. Network analyses in patients with DS revealed a loss of interconnection between bacterial communities, such as between Bacillota and Bacteroidota or Fusobacteriota, suggesting a disruption in the interbacterial interactions, the major determinant of the microbial community composition. Instead, we observed stronger interactions spanning multiple kingdoms, especially between bacterial and fungal microbiomes. Bacterial-fungal interactions likely play a crucial role in biofilm formation and pathogenicity⁵². Synergistic interactions between bacteria, especially Streptococcus and Actinomyces, and the fungus Candida can enhance biofilm development on denture surfaces⁵². These complex interactions within the multi-kingdom microbiome, especially fungi and bacteria in denture biofilms, may contribute to increased pathogenicity and drug resistance⁵³. Overall, these findings support the hypothesis that denture wearing, particularly in DS, is accompanied by a dysbiotic condition where multikingdom microbial components interact¹⁵.

The strengths of our study include detailed clinical oral examinations, inclusion of a homogeneous age group of both sexes, extraction of information from medical records, and use of shotgun metagenomics for high-resolution multikingdom microbial profiling. The study population, namely very old adults living in LTC facilities, is a group that is typically prescribed multiple medications, functionally impaired, and particularly vulnerable to oral health complications and frailty. Nonetheless, this study had several limitations. The small sample size and cross-sectional design limit causal inference. Sampling conditions were adapted to housing unit operations, and participant instructions (e.g., wearing dentures for a set duration, avoiding eating, drinking, or cleaning beforehand) could not always be strictly enforced. Another limitation was the lack of detailed information on denture characteristics, including age, material, and surface properties, which could act as potential confounders. However, data on the need for denture repair were available, providing some insight into the condition of the dentures.

The novelty of our work lies in the simultaneous analysis of multiple microbial kingdoms, which revealed broader community alterations associated with DS beyond Candida spp., highlighting the importance of considering bacterial–fungal interactions in understanding the pathogenesis of DS. Despite these limitations, our findings revealed relevant clinical insights. Maintaining proper oral hygiene is crucial for preventing and treating DS in denture wearers⁵⁴. All dentate study subjects exhibited equally elevated plaque index values and generalized bleeding on probing, indicating a generally poor level of oral hygiene. Despite these factors and numerous oral health concerns, not all dentate participants had DS. The clinically observable differences in oral health included higher GI values, a greater number of root remnants, and a higher proportion of participants with clinically assessed mild periodontitis among those with DS compared with those without.

Conclusion

Our study identified a novel association between DS and significant alterations in DAOM in older adults residing in LTC facilities. Consistent with previous studies, we observed an increased abundance of Candida spp. in the DS group. Our multi-kingdom analysis revealed additional microbial species beyond Candida spp. providing new insights into the DS-associated shifts. These findings suggest that microbial alterations may contribute to DS pathogenesis through mechanisms such as mucosal inflammation, immune modulation, and potential systemic dissemination. Further research is needed to explore how these microbial changes influence DS progression over time and to evaluate targeted preventive interventions.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Conceptualization: PJP, RKS, KH, PM; Funding acquisition: PJP, PM; Project administration: PJP, RKS, KH, PM; Performed the laboratory work: MM, PJP, PM; Data analysis and interpretation: MM, PM; Writing – original draft preparation: MM, PM; Writing – review and editing: PJP, RKS, KH; All authors read and approved the final manuscript.

Funding

This work was supported by the Finnish Dental Society Apollonia, the Päivikki and Sakari Sohlberg Foundation, the Selma and Maja-Lisa Selander Foundation, the Sigrid Juselius foundation, and the Research Council of Finland (1340750 for PJP). Open access funded by Helsinki University Library. The funders had no role in the design or execution of the study; the collection, management, analysis, or interpretation of data; the preparation, review, or approval of the manuscript; or the decision to submit the manuscript for publication.

Data availability

The datasets generated and/or analyzed during the current study are not publicly available due to licensing restrictions from the City of Helsinki but are available from the corresponding author on reasonable request and with permission from the City of Helsinki.

Declarations

Ethics, consent, and permissions

The study was approved by the City of Helsinki and the Ethics Committee of the Hospital District of Helsinki and Uusimaa (HUS/2042/2016 and HUS/968/2017) and was conducted in accordance with Good Clinical Practice and the Declaration of Helsinki. Written informed consent was obtained from all participants or, when necessary, from a legal proxy if the participant was unable to comprehend the details of the study. Participation in the study was voluntary.

Competing interests

The authors declare no competing interests.