Oral microbiome dysbiosis and oral-gut microbial network disruption in hand osteoarthritis: data from the Xiangya Osteoarthritis Study

Abstract

Objectives

The oral microbiome plays a critical role in modulating systemic inflammation, partly through its interactions with the gut microbiome. Although gut microbiome dysbiosis has been implicated in symptomatic hand osteoarthritis (SHOA), the role of oral microbiome dysbiosis in SHOA and its relationship with gut microbiome dysbiosis remain unclear. Elucidating these associations could provide novel insights into SHOA pathogenesis.

Methods

Participants were recruited from the Xiangya Osteoarthritis (XO) Study, an ongoing community-based observational study. Saliva samples were analysed using 16S ribosomal RNA gene sequencing. Oral microbial richness, composition and relative abundance of specific taxa were compared between SHOA participants and controls without SHOA. Correlations within the oral-gut microbiome network were also assessed and compared between groups.

Results

Compared with controls (n=712), participants with SHOA (n=52) exhibited significantly lower oral microbial richness (p=0.007) and altered composition (p=0.007). The relative abundance of the genus Trichococcus was significantly higher in SHOA participants (β=0.437 (95% CI 0.174 to 0.699), p=0.001, Q=0.073) and positively associated with SHOA severity. Furthermore, the number of significant correlations within the oral-gut microbiome network was markedly reduced in SHOA participants compared with controls. Notably, Trichococcus abundance in the oral microbiome correlated positively with the gut microbial KEGG pathway of tyrosine metabolism (r=0.137, p=0.001, Q=0.047), both linked to SHOA.

Conclusion

Oral microbiome dysbiosis and disruption of the oral-gut microbiome network are associated with prevalent SHOA. These findings suggest a potential role of the oral-gut microbiome axis in SHOA pathogenesis. Larger studies are needed to confirm these associations.

Trial registration number

NCT04033757.

WHAT IS ALREADY KNOWN ON THIS TOPIC

• Hand osteoarthritis (OA) affects hundreds of millions of people worldwide, leading to a substantial disease burden, yet its pathogenesis remains poorly understood.

• The oral microbiome plays a role in systemic inflammation and interacts closely with the gut microbiome, which is significantly associated with SHOA; therefore, the oral microbiome potentially could contribute to the pathogenesis of SHOA.

WHAT THIS STUDY ADDS

• Using data from a community-based observational study, we found that oral microbiome dysbiosis and a disrupted oral-gut microbiome network are associated with prevalent SHOA.

• The genus Trichococcus in the oral microbiome was significantly enriched in SHOA participants and positively correlated with disease severity.

• Two oral microbial KEGG pathways related to alkaloid biosynthesis were significantly reduced in individuals with SHOA.

• Reduced oral-gut microbiome network correlations were observed in SHOA.

• Additionally, Trichococcus abundance in the oral microbiome was positively correlated with the gut microbial KEGG pathway of tyrosine metabolism.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• The findings provide new insights into the role of oral microbiome alterations in SHOA and their correlation with gut microbiome dysbiosis, suggesting the potential involvement of the oral-gut microbiome axis in SHOA pathogenesis.

Introduction

Hand osteoarthritis (OA) is a common musculoskeletal disorder affecting over 142 million individuals globally.^{1 2} It is characterised by hand pain, stiffness, reduced grip strength, functional limitations and impaired quality of life, leading to a substantial clinical and socioeconomic burden.^2–5 Although emerging evidence links gut microbiome dysbiosis and related metabolite alterations to hand OA,^{6 7} the underlying risk factors and pathogenic mechanisms remain poorly understood. Therefore, there is a need to further elucidate these factors which could inform the potential treatment targets.

The oral microbiome, a highly diverse and dynamic ecosystem, plays a critical role in regulating systemic inflammation and maintaining immune homeostasis through interactions with host cells and distal organs.^8–10 Unlike the gut microbiome, the oral microbiome is directly and continuously exposed to external factors such as diet, smoking and alcohol consumption,¹¹ potentially making it more sensitive to environmental perturbations. Notably, as two of the largest microbial communities in the human body, the oral and gut microbiomes share a complex and interconnected relationship despite distinct microbial compositions across anatomical sites.^{10 12} Population-based studies have linked concurrent dysbiosis in these microbiomes to various systemic diseases.^{13 14} While gut microbiome dysbiosis has been implicated in hand OA, the role of oral microbiome dysbiosis and its relationship with gut microbiota in this context remains unexplored. Elucidating these relationships and underlying mechanisms may offer valuable insights for population screening and targeted interventions, particularly as saliva represents a more accessible and acceptable sample type than faeces.

To address this knowledge gap, we used data from the Xiangya OA (XO) Study, a population-based observational study, to examine the association between the oral microbiome and hand OA and to explore the correlation between the oral and gut microbiomes in individuals with hand OA.

Methods

Study participants

The XO Study is a large community-based longitudinal study designed to investigate the natural history and risk factors of OA in a rural area of China (ClinicalTrials.gov identifier: NCT04033757). As described in previous publications,6 7 15 16 the XO Study recruited a randomly selected sample of residents aged 50 years or older from Longshan County, Hunan Province. The study consists of three subcohorts: subcohort I (n=1469, initiated in 2015), subcohort II (n=1271, initiated in 2018) and subcohort III (n=1340, initiated in 2019). For the present analysis, only participants from subcohort II were included, as saliva samples were collected at baseline in this group.

Participants were excluded if they met any of the following criteria: (1) absence of hand posteroanterior radiographic images or hand symptom assessment; (2) lack of saliva samples for 16S ribosomal RNA (rRNA) sequencing; (3) use of antibiotics within 1 month before saliva sample collection; (4) self-reported history of cancer or (5) missing dietary intake data. To investigate oral-gut microbiome correlations, we further excluded participants without stool samples available for shotgun metagenomic sequencing. The study workflow is illustrated in figure 1.

Figure 1.

Summary of the present study. KEGG, Kyoto Encyclopaedia of Genes and Genomes; rRNA, ribosomal RNA; SHOA, symptomatic hand osteoarthritis; XO, Xiangya osteoarthritis.

Assessment of symptomatic hand osteoarthritis

Posteroanterior radiographs of both hands were obtained for all participants. A modified Kellgren/Lawrence (K/L) scale was used to assess the second to fifth distal interphalangeal joints, second to fifth proximal interphalangeal joints, first to fifth metacarpophalangeal joints, thumb interphalangeal joints and thumb base (carpometacarpal) joints in each hand.¹⁷ Radiographic hand OA was defined as a K/L grade of ≥2 in any of these joints. The hand radiographs were primarily evaluated by an experienced musculoskeletal radiologist with specialised expertise in K/L grading, who was blinded to all clinical and demographic information. To evaluate intra-rater reliability, each batch of 50 new radiographs included five previously scored images randomly selected for repeat evaluation by the same reader. Additionally, a senior musculoskeletal radiologist independently evaluated a subset of 30 radiographs to assess inter-rater reliability. The intra-rater and inter-rater reliability for identifying radiographic hand OA were high, with kappa values of 0.91 (95% CI 0.83 to 0.99) and 0.71 (95% CI 0.45 to 0.96), respectively.⁷ Hand symptoms were considered present if participants answered ‘yes’ to the question: “On most days, do you have pain, aching, or stiffness in your left/right hand?” Symptomatic hand OA (SHOA) was defined as the presence of both self-reported hand symptoms and radiographic OA (ie, K/L grade ≥2) in at least one joint of the same hand.¹⁸ Participants without SHOA in either hand were classified as the control group.

Saliva sample collection and DNA extraction

Saliva samples were collected in the morning following an overnight fast of at least 8 hours. Participants were instructed to avoid activities that could stimulate oral secretions, such as tooth brushing or mouthwash use, before sample collection. Samples were immediately frozen on-site, transported on dry ice within 20 min and stored at −80°C until analysis.

Genomic DNA was extracted from 5 mL of each saliva sample using the TIANamp Soil DNA Kit (DP336) according to the manufacturer’s instructions. DNA concentration was quantified using a Qubit 2.0 fluorometer (Thermo Fisher Scientific, Waltham, Massachusetts, USA).

16S rRNA gene sequencing

The 16S rRNA gene was amplified using primers 341F (5’-CCTACGGGNGGCWGCAG-3’)/805R (5’-GACTACHVGGGTATCTAATCC-3’), targeting the V3–V4 hypervariable region. Sequencing was performed on an Illumina NovaSeq PE250 platform according to the manufacturer’s protocol. Paired-end reads were demultiplexed using unique barcodes and barcodes and primer sequences were removed. Reads were then merged using Fast Length Adjustment of SHort reads. Quality filtering of raw reads was performed with fqtrim (V.0.94) to generate high-quality clean tags, and chimeric sequences were filtered using Vsearch software (V.2.3.4). Amplicon sequence variants (ASVs) were inferred and dereplicated using Divisive Amplicon Denoising Algorithm 2, resulting in a feature table and representative feature sequences. Sequence alignment was performed using Basic Local Alignment Search Tool and taxonomic annotation of ASVs was conducted using the SILVA database (Release 138, https://www.arb-silva.de/documentation/release-138/). Alpha-diversity and β-diversity analyses were conducted using Quantitative Insights Into Microbial Ecology. Alpha-diversity, reflecting microbial richness and evenness, was assessed using the Shannon index. Beta-diversity, representing differences in community composition, was evaluated using Bray-Curtis distance. All diversity analyses were performed after rarefying samples to a uniform sequencing depth of 17 630 sequences per sample, corresponding to the minimum sequencing depth observed across all samples, to ensure comparability.

Stool sample collection and shotgun metagenomic sequencing

Stool and saliva samples were collected during the same visit. Stool samples were obtained on-site, immediately frozen, transported on dry ice within 20 min and stored at −80°C until further processing. DNA was extracted from 200 mg of stool using the Magen HiPure Soil DNA Kit (Magen, Guangzhou, China), following the manufacturer’s instructions. Library preparation and shotgun metagenomic sequencing were performed on the MGISEQ-2000 platform (BGI-Shenzhen, China), to generate 150 bp paired-end reads. Quality control procedures included removal of low-quality reads, adapter trimming and host DNA filtering. Taxonomic annotation was performed using Kraken2 based on the Unified Human Gastrointestinal Genome, and functional profiling was conducted using Kyoto Encyclopaedia of Genes and Genomes (KEGG) pathway mapping. Detailed methods have been described previously.⁶

Assessment of covariates

Demographic and lifestyle information was collected through face-to-face interviews conducted by trained health professionals using standard questionnaires.¹⁵ Data included age, sex, body mass index (BMI), smoking status, alcohol consumption and dietary intake frequency of meat/eggs, dairy products and vegetables. Height and weight were measured, and BMI was calculated as weight (kg) divided by height squared (m²).

Statistical analysis

The diversity of oral microbiota at the genus level was compared between participants with SHOA (ie, the SHOA group) and those without SHOA (ie, the control group) using α-diversity (Shannon index) and β-diversity (Bray-Curtis distance). Differences in α-diversity were assessed using the Wilcoxon rank-sum test, while differences in β-diversity were evaluated using permutation multivariate analysis of variance in combination with principal coordinates analysis.

To enhance robustness, a 10% prevalence filter was applied to remove rare taxa and microbial function.¹⁹ For oral microbial genera present in ≥10% of samples, associations between relative abundance and SHOA were assessed using Tweedie compound Poisson generalised linear mixed-effects models (GLMMs), implemented in the Microbiome Multivariable Associations with Linear Models (MaAsLin2 V.1.16.0) R package.²⁰ Models were adjusted for age, sex, BMI, smoking status, alcohol consumption and frequency of dietary intake of meat/eggs, dairy products and vegetables, based on previous studies.6 7 21 22 Relative abundance data were normalised using total-sum scaling. GLMMs were applied to account for the non-normal distribution of abundance data and multiple sources of variability.²³ To evaluate the robustness of the main findings, we performed a sensitivity analysis using a 1:4 age-matched and sex-matched case-control design without replacement. The same approach was used to evaluate the association between oral microbiota and SHOA severity. SHOA participants were stratified according to: (1) unilateral versus bilateral SHOA; (2) maximum hand K/L grade (2 vs >2); (3) maximum visual analogue scale (VAS) pain score (<4 vs ≥4) and (4) number of hand joints with K/L grade ≥2 (one vs more than one joint). Additionally, to assess whether radiographic hand OA shares similar oral microbiome profiles with SHOA, we further investigated the associations of oral microbiome α-diversity, β-diversity and taxonomic composition between participants with radiographic hand OA and those without radiographic hand OA and without pain, using the same analytical strategies applied in the SHOA analyses. Furthermore, MaAsLin2 was applied to assess differences in microbial functions, defined as KEGG level 3 pathways present in ≥10% of samples, between the SHOA and control groups, adjusting for the same set of confounders.

To explore ecological links between the oral microbiome and gut microbiome, Spearman’s correlation analyses were performed separately in the SHOA and control groups, adjusting for the same confounders. Gut microbiome data were derived from metagenomic sequencing of stool samples previously collected in the XO Study.^{6 16} Only oral microbial genera with a relative abundance >0.01% and present in >70% of samples were included. Microbiome correlation networks were visualised using Cytoscape (V.3.9.1) and network connectedness was calculated as the ratio of significant correlations (edges) to the number of taxa (nodes).²⁴ Differences in the number of edges between groups were assessed using the Wilcoxon rank-sum test. Spearman’s correlation analyses were applied to assess the correlations between SHOA-related oral microbial genus and SHOA-related gut microbiome taxa and microbial functions (ie, KEGG pathways), adjusting for the same covariates.^{6 16}

All statistical analyses were conducted using SAS V.9.4 (SAS Institute, Cary, North Carolina, USA) and R V.4.3.3 (R Foundation, Vienna, Austria). A two-sided p value <0.05 was considered statistically significant. For high-throughput sequencing data, multiple testing was adjusted using the Benjamini-Hochberg false discovery rate (FDR) method,²⁵ with FDR-adjusted p values (Q values) <0.1 deemed statistically significant.¹⁴

Results

Basic characteristics

Of the 1271 participants initially enrolled in subcohort II of the XO Study, we excluded 31 individuals due to missing hand radiographs or symptom assessments, 230 who did not provide saliva samples, 242 reporting antibiotic use within 1 month before saliva sample collection, three with a history of cancer and one with missing dietary intake data. This left a total of 764 participants for inclusion in the final analysis. Baseline characteristics are summarised in table 1. Compared with controls (n=712), participants with SHOA (n=52) were older (median age: 71.0 years vs 62.0 years; p<0.001), more often women (78.8% vs 56.3%; p=0.002), non-smokers (80.8% vs 61.2%; p=0.002) and non-drinkers (65.4% vs 51.1%; p=0.043).

Table 1.

Basic characteristics of included participants

Variable	SHOA (n=52)	Control (n=712)
Age, median (IQR), years	71.0 (64.5, 77.0)	62.0 (55.0, 68.0)
Sex, n (%)
Male	11 (21.2)	311 (43.7)
Female	41 (78.8)	401 (56.3)
BMI, median (IQR), kg/m2	23.5 (21.0, 27.6)	23.6 (21.4, 26.0)
Smoking status, n (%)
Never	42 (80.8)	433 (61.2)
Past	3 (5.8)	17 (2.4)
Current	7 (13.5)	258 (36.4)
Alcohol drinking status, n (%)
Never	34 (65.4)	362 (51.1)
Past	6 (11.5)	59 (8.3)
Current	12 (23.1)	288 (40.6)
Dietary intake of meat/eggs, median (IQR), times/week	2.0 (2.0, 4.5)	4.5 (2.0, 4.5)
Dietary intake of vegetables, median (IQR), times/week	7.0 (4.5, 14.0)	4.5 (4.5, 14.0)
Dietary intake of dairy products, median (IQR), times/week	0.0 (0.0, 0.0)	0.0 (0.0, 0.0)

Missing data were observed for BMI (n=15), smoking status (n=4) and drinking status (n=3).

BMI, body mass index; SHOA, symptomatic hand osteoarthritis.

Oral microbiome and SHOA

Significant differences in oral microbial diversity were observed between participants with and without SHOA. Specifically, α-diversity, as measured by the Shannon index, was significantly lower in the SHOA group (p=0.007), with a mean±SD of 6.54±0.63 and a median (IQR) of 6.62 (6.19–6.93), compared with 6.78±0.56 and 6.84 (6.45–7.15) in the control group (figure 2A). Similarly, β-diversity, measured by Bray-Curtis distance, differed significantly between groups (p=0.007) (figure 2B). A Venn diagram showed that 911 ASVs and eight genera were unique to the SHOA group (figure 2C).

Figure 2.

Comparison of oral microbiome diversity and composition between participants with SHOA and controls. (A) Violin plots showing α-diversity measured by the Shannon index between participants with SHOA and controls. (B) PCoA plots comparing β-diversity measured by Bray-Curtis distance between participants with SHOA and controls. (C) Venn diagram illustrating shared and exclusive ASVs and microbial genera between participants with SHOA and controls. (D) Relative abundance of the top 20 microbial genera in the oral microbiome composition of participants with SHOA and controls. (E) Volcano plot depicting differences in the microbial genera between participants with SHOA and controls. The red region indicates statistically significant differences (Q<0.1). (F) Relative abundance of differential microbial genus between participants with SHOA and controls. ASV, amplicon sequence variant; PCoA, principal coordinates analysis; SHOA, symptomatic hand osteoarthritis.

The dominant genera in both groups included Neisseria, Streptococcus, Haemophilus, Porphyromonas and Fusobacterium (figure 2D). Compared with controls, participants with SHOA showed a higher relative abundance of Trichococcus (β=0.437 (95% CI 0.174 to 0.699), p=0.001, Q=0.073) (figure 2E,F). Detailed genus-level associations with SHOA are presented in online supplemental table S1. In the 1:4 age-matched and sex-matched sensitivity analysis, the results were broadly consistent, with Trichococcus remaining positively associated with SHOA (β=0.456 (95% CI 0.124 to 0.787), p=0.008). Moreover, higher Trichococcus abundance was significantly associated with greater SHOA severity across multiple clinical measures, specifically laterality (β=0.093 (95% CI 0.027 to 0.158), p=0.006), maximum K/L grade (β=0.090 (95% CI 0.025 to 0.156), p=0.007), VAS pain score (β=0.110 (95% CI 0.043 to 0.176), p=0.001) and number of affected joints (β=0.095 (95% CI 0.029 to 0.161), p=0.005) (table 2). When comparing participants with radiographic hand OA with those without radiographic hand OA and without pain, α-diversity did not differ significantly between groups (p=0.379), whereas β-diversity showed a significant difference (p=0.001). However, we did not observe a significant association between the relative abundance of Trichococcus and radiographic hand OA (online supplemental table S2).

Table 2.

Association between the oral microbial genus Trichococcus and the severity of SHOA

Severity	Number	Coefficient (95% CI)	P value
Laterality
Non-SHOA	712	0.093 (0.027 to 0.158)	0.006
Unilateral	31
Bilateral	21
Maximum K/L grade of both hands
<2	712	0.090 (0.025 to 0.156)	0.007
2	32
>2	20
Maximum VAS score of both hands
0	712	0.110 (0.043 to 0.176)	0.001
<4	13
≥4	39
Number of hand joints with K/L grade ≥2
None	712	0.095 (0.029 to 0.161)	0.005
One joint	19
Two or more joints	33

K/L grade, Kellgren/Lawrence grade; SHOA, symptomatic hand osteoarthritis; VAS, visual analogue scale.

At the functional level, compared with controls, participants with SHOA exhibited a significantly lower relative abundance of KEGG level 3 pathways related to tropane, piperidine and pyridine alkaloid biosynthesis (β=−0.033 (95% CI −0.055 to −0.010), p=0.006, Q=0.084) and indole alkaloid biosynthesis (β=−0.849 (95% CI −1.456 to −0.241), p=0.006, Q=0.090) (figure 3, online supplemental table S3).

Figure 3.

Associations of oral microbial KEGG level 3 pathways with SHOA. The bar plot illustrates a significant (*Q<0.1) decrease in the relative abundance of tropane, piperidine and pyridine alkaloid biosynthesis, as well as indole alkaloid biosynthesis in the SHOA group compared with the control group. KEGG, Kyoto Encyclopaedia of Genes and Genomes; SHOA, symptomatic hand osteoarthritis.

Oral-gut microbiome correlation network and SHOA

The oral-gut microbiome correlation network at the genus level differed significantly between SHOA participants and controls (p=0.037) (figure 4). The SHOA group exhibited a smaller network (nodes=225; edges=586; relative connectedness ratio=2.6) compared with the control group (nodes=238; edges=788; relative connectedness ratio=3.3). Additionally, the average number of neighbours per node was lower in the SHOA group (5.2 vs 6.6). Detailed correlations within each group are provided in online supplemental tables S4 and S5. Furthermore, as shown in figure 5A–C, we examined the association between the SHOA-related oral microbial genus (ie, Trichococcus) and the SHOA-related gut microbiome taxa and KEGG pathways. Although no significant associations were found between SHOA-related oral microbial genus and gut microbiome taxa, a significant positive association was observed between Trichococcus and the gut microbial functional pathway for tyrosine metabolism (r=0.137, p=0.001, Q=0.047) (figure 5D). Additional associations between Trichococcus and SHOA-related gut microbial functions (ie, KEGG pathways) are provided in online supplemental table S6.

Figure 4.

Oral-gut microbiome correlation networks at the genus level in SHOA and control groups. Only genera with a relative abundance >0.01 and present in >70% of samples were considered. Solid lines indicate significant correlations (p<0.05) between oral and gut microbial genera, with line width indicating correlation magnitude. Red and blue edges represent positive and inverse correlations, respectively. Red triangles (circles) denote oral (gut) microbial genera with significant differences in relative abundance between the SHOA group and the control group. Yellow triangles (circles) represent oral (gut) microbial genera with no significant differences in relative abundance between the two groups. SHOA, symptomatic hand osteoarthritis.

Figure 5.

Association between SHOA-related oral microbial genus and SHOA-related gut microbiome taxa and microbial functions. (A) Heatmap-circle showing correlations between SHOA-related oral microbial genus (ie, Trichococcus) and SHOA-related gut microbial genera. (B) Heatmap-circle showing correlations between SHOA-related oral microbial genus (ie, Trichococcus) and SHOA-related gut microbial species. (C) Heatmap-circle showing correlations between SHOA-related oral microbial genus (ie, Trichococcus) and SHOA-related gut microbial KEGG pathways. A significant (*Q<0.1) positive association was observed between Trichococcus and tyrosine metabolism. (D) Scatter plot illustrating the correlation between the SHOA-related oral microbial genus (ie, Trichococcus) and the SHOA-related KEGG pathway (ie, tyrosine metabolism), which was identified as a significant association. KEGG, Kyoto Encyclopaedia of Genes and Genomes; SHOA, symptomatic hand osteoarthritis.

Discussion

Using data from a population-based observational study, we found significant alterations in the oral microbiome of individuals with SHOA. This SHOA-related oral microbial dysbiosis was characterised by distinct structural and compositional differences at the genus level, with a notably increased relative abundance of Trichococcus, which also correlated positively with SHOA severity. Additionally, two oral microbial KEGG pathways, tropane, piperidine and pyridine alkaloid biosynthesis and indole alkaloid biosynthesis, were significantly decreased in participants with SHOA. The oral-gut microbiome correlations in SHOA individuals markedly reduced, suggesting a disrupted balance between these two microbial communities. Notably, Trichococcus was positively associated with the tyrosine metabolism pathway, a gut microbiome metabolic pathway previously implicated in SHOA,^{6 7} further supporting a potential link between oral and gut microbiome dysbiosis in SHOA.

Comparison with previous studies

To date, no studies have investigated the oral microbiome in individuals with SHOA. One previous study analysed saliva samples from 110 patients with rheumatoid arthritis, 67 patients with knee OA and 155 healthy individuals using 16S rRNA gene sequencing.²⁶ That study reported significant differences in oral microbiome profiles among these groups. However, owing to the distinct biomechanical and pathophysiological characteristics of knee and hand OA, their underlying mechanisms may differ, leaving the association between oral microbiome dysbiosis and hand OA remains poorly understood. Our study identified oral microbiome alterations in SHOA participants, suggesting that common systemic mechanisms, such as chronic inflammation, immune dysregulation or microbial metabolites, may mediate oral microbiome effects across both weight-bearing and non-weight-bearing joints. Notably, the specific microbial taxa associated with knee OA in previous studies differ from those associated with hand OA here, potentially reflecting differences in disease aetiology or population heterogeneity. Further research is warranted to determine whether distinct joint locations exhibit overlapping or divergent oral microbiome signatures and to elucidate the mechanisms linking oral microbiome dysbiosis to OA pathogenesis at different joint sites. Moreover, the oral-gut microbiome axis plays a critical role in human health, with alterations in the gut microbiome often accompanied by changes in the oral microbiome.¹⁰ Our previous research demonstrated an association between gut microbiome dysbiosis and prevalent SHOA.^{6 7} The present study adds new evidence linking oral microbiome dysbiosis to prevalent SHOA and suggests that disruption of the oral-gut microbiome network may contribute to SHOA pathogenesis.

Possible explanations

The oral cavity serves as a gateway for systemic inflammation, as oral bacteria can translocate to distant sites and promote disease development.¹⁰ Oral microbiome dysbiosis is known to trigger local and systemic inflammatory responses and has been implicated in the pathogenesis of various systemic conditions.⁹ Given that hand OA has been linked to systemic inflammation,²⁷ it is possible that oral microbiome dysbiosis contributes to SHOA through similar inflammatory pathways. In our study, we observed a higher relative abundance of the genus Trichococcus in the saliva of individuals with SHOA than controls. Notably, a previous animal study reported that enrichment of Trichococcus occurred in mice fed a high-fat diet,²⁸ while its abundance decreased following treatment in Zucker diabetic fatty rats.²⁹ These findings suggest a potential association between Trichococcus and unhealthy dietary patterns. However, the specific mechanisms by which Trichococcus affects host health and disease processes remain to be elucidated.

Moreover, our study revealed that individuals with SHOA exhibited a reduced oral-gut microbiome correlation network, compared with controls. Although the oral and gut microbiomes represent distinct microbial ecosystems,¹² they are closely interconnected and influence each other through multiple pathways.¹⁰ The observed reduction in network complexity may reflect disrupted microbial communication between these communities. Given the critical role of the oral-gut microbiome axis and its bidirectional cross-talk in maintaining host physiological balance,¹⁰ such a disruption could impair microbial homeostasis and reduce the host’s resilience to systemic inflammation. Further research is needed to elucidate specific microbial interactions and their physiological consequences. Additionally, by analysing the correlation between the differential oral genus Trichococcus and previously identified gut microbiome metabolic pathways in SHOA participants and controls,⁶ we observed a positive correlation between Trichococcus and the tyrosine metabolism pathway. Tyrosine plays key roles in human physiology and has been implicated in OA. For example, elevated tyrosine levels have been reported in individuals with knee OA,³⁰ and Mendelian randomisation analyses have shown that genetically predicted tyrosine levels were significantly associated with increased OA risk, especially in the knee and hip.³¹ Tyrosine is also a precursor to catecholamines,³² which can activate T and B lymphocytes via protein kinase A signalling and contribute to persistent synovial inflammation.³³ Our findings suggest that the observed association between oral microbial Trichococcus and the gut microbial tyrosine metabolism pathway may play a role in SHOA pathogenesis. However, mechanistic studies are needed to clarify causal relationships and uncover the underlying biological pathways.

Strengths and limitations

Several strengths of our study are worth noting. First, we identified robust associations between the oral microbiome, particularly Trichococcus, and SHOA with stronger correlations observed in participants with more severe SHOA across multiple classification methods. This consistency enhances the credibility and reliability of our findings. Second, our study provides novel evidence linking the oral microbiome to SHOA and highlights the disruption of the oral-gut microbiome network, offering a promising direction for future research into OA pathogenesis. Third, we identified several altered KEGG metabolic pathways in both the oral and gut microbiomes that were associated with SHOA, which may have translational implications for biomarker discovery and potential therapeutic targets.

Nonetheless, several limitations should be acknowledged. First, we profiled the oral microbiota using 16S rRNA gene sequencing, which provides taxonomic and compositional insights but lacks the resolution to identify specific bacterial species or strains. Future studies employing metagenomic and targeted metabolomic approaches are needed to characterise functional microbial profiles and clarify their roles in SHOA pathogenesis. Second, owing to the cross-sectional design of our study, we could not establish a temporal or causal relationship between oral microbiome alterations and SHOA. Longitudinal cohort studies are required to determine whether these microbial changes precede SHOA development or are consequences of the disease. Third, in the absence of an appropriate animal model for hand OA, we were unable to investigate the mechanistic links between the oral and gut microbiomes and their potential impact on SHOA development. Studies focusing on other joint tissues may help validate and further elucidate these associations. Fourth, although we adjusted for several known confounders, residual confounding from unmeasured variables cannot be excluded, an inherent limitation of observational research. Fifth, SHOA was diagnosed at the hand level rather than at the individual joint level, which may have introduced some degree of misclassification. Sixth, the relatively small sample size of SHOA cases may limit statistical power and could prevent detection of additional microbial taxa with more modest associations with SHOA pathogenesis. Accordingly, the present findings should be regarded as exploratory and warrant validation in future studies with larger sample sizes.

Clinical and research implications

This study provides the first evidence linking oral microbiome alterations to SHOA and highlights their correlation with gut microbiome dysbiosis, suggesting a potential role for the oral-gut microbiome axis in SHOA pathogenesis. Our findings underscore the importance of maintaining the balance between oral and gut microbiomes for overall health. Future research integrating oral microbiome profiling with analyses of microbial metabolic products along the oral-gut microbiome axis may offer deeper insights into the mechanisms underlying OA. If confirmed by longitudinal studies, mechanistic investigations and interventional trials, these findings could inform the development of novel microbiome-targeted strategies for the prevention and management of SHOA.

Conclusions

Oral microbiome dysbiosis and disruption of the oral-gut microbiome network are associated with prevalent SHOA. These findings suggest a potential role of the oral-gut microbiome axis in SHOA pathogenesis. Larger studies are needed to confirm these associations.

Data availability statement

Data are available in a public, open access repository. Data are available on reasonable request. The 16S rRNA gene sequencing data for the oral microbiome and the metagenomic sequencing data for the gut microbiome presented in this study have been deposited in the Genome Sequence Archive (GSA) at the National Genomics Data Center, China National Center for Bioinformation, under accession numbers CRA029367 (oral microbiome) and CRA012188 (gut microbiome), and are publicly accessible at https://ngdc.cncb.ac.cn/gsa. Additional data are available from the corresponding authors on reasonable request.

Ethics statements

Patient consent for publication

Not applicable.

Ethics approval

This study was approved by the Research Ethics Committee of Xiangya Hospital, Central South University (approval number 201510506). All participants provided written informed consent before participating in the study.