Background/Purpose:   Periodontitis (PD) is an inflammatory disease of tissues supporting the teeth, caused by microorganisms that adhere to and grow along the tooth’s surface.  Both PD and pathogens linked to PD have been proposed as risk factors in rheumatoid arthritis (RA).  Previous studies have primarily targeted select organisms for investigation, failing to account for the complex nature of the local microbiome involved in PD.  Thus, we characterized the subgingival microbiome in patients with established RA, comparing resident microbial communities with those of controls.

Methods:   RA patients satisfying the 1987 ACR classification criteria and osteoarthritis (OA) controls underwent full-mouth examination and were classified as having PD based on Machtei criteria.  Subgingival plaque was collected from up to 4 sites and pooled.  Bacterial DNA was isolated and sequenced using 454 pyrosequencing; trimmed reads were searched against the Human Oral Microbiome Database.  Hierarchical cluster analysis with averaging-linkage agglomeration on microbiome profiles were used to classify the samples into 3 clusters. Data were standardized by samples prior to the clustering.  With only a 1 patient with membership, Cluster 3 was excluded from further analysis.  Logistic regression was used to determine factors associated with membership in Cluster 1 vs. Cluster 2.  Factors chosen a priori for analysis included: RA vs. OA case status, PD, smoking status, age, gender, marital status, education, race, body mass index (BMI), and diabetes.

Results:   RA (n=260) and OA (n=296) patients were similar in terms of sociodemographic factors except marital status; OA patients were more likely to have diabetes (25% vs. 18%, p=0.04) and had a higher BMI (31.7 vs. 29.8 kg/m², p=0.0008); RA patients were more likely to have ever smoked (62% vs. 46%, p=0.0002), have PD (38% vs. 27%, p=0.007 ), and be married (69% vs. 61%, p=0.06). Patients segregated into 3 groups based on subgingival microbial composition with Clusters 1 and 2 accounting for all but one individual.  Factors significantly associated with Cluster membership included PD, age, smoking, and race (Table).  There was no association of RA-OA case status with Cluster membership, suggesting similar overall subgingival microbiome composition by arthritis diagnosis after accounting for other factors.  Results were unchanged with RA cases limited to those with a positive anti-CCP antibody.

Conclusion:   Patients with established RA demonstrated a subgingival microbial composition that was similar to patients with OA after accounting for the presence of PD and other PD risk factors.  Further study will be needed to examine whether an abundance of individual bacterial species or microbial diversity is impacted in established RA or whether the overall microbial composition in RA evolves over time with the use of RA therapies and changes in underlying disease activity.