Prevalence of Periodontal Disease and Periodontopathic Bacteria in Anti–Cyclic Citrullinated Protein Antibody–Positive At-Risk Adults Without Arthritis

Key Points Question What is the prevalence of periodontal disease and citrullinating periodontopathic bacteria in anti–cyclic citrullinated protein–positive at-risk individuals (CCP+ at-risk) compared with a healthy control group and patients with early rheumatoid arthritis (RA)? Findings This cross-sectional study identified an increased prevalence of periodontal disease sites, clinical periodontitis, and periodontal inflamed surface area in CCP+ at-risk individuals and those with early RA compared with a control group. Results showed that CCP+ at-risk individuals had increased abundance of Porphyromonas gingivalis at healthy periodontal sites compared with the control group and patients with early RA. Meaning In individuals at risk of RA, periodontitis and P gingivalis were increased before joint disease and may be a target for prevention.


Introduction
Autoantibodies associated with rheumatoid arthritis (RA) can be detected in the serum years before patients develop joint inflammation, [1][2][3] suggesting the joints may be a target rather than the primary cause of this disease. Such observations suggest a preclinical phase of RA and, importantly, raise the possibility of disease prevention. The enrichment of serum IgA anticitrullinated protein antibodies (ACPA) in individuals at risk of RA suggests mucosal sites (eg, oral mucosa) may be important in the earliest phase of RA. 4,5 There is good evidence that periodontitis and RA are clinically associated. [6][7][8] Furthermore, periodontitis is associated with a specific bacterial signature characterized by the increased abundance of the pathogenic organism Porphyromonas gingivalis alongside a community of other, predominantly anaerobic, organisms. 9 Porphyromonas gingivalis is capable of citrullinating local antigens by virtue of its peptidylarginine deiminase enzyme. 10 In a putative etiological model, virulent strains of P gingivalis at inflamed periodontal sites generate novel citrullinated antigens that trigger a mucosal immune response in certain individuals, possibly those with genetic predispositions. 11 Recent data suggest the periodontopathic bacterium Aggregatibacter actinomycetemcomitans may also directly induce neutrophil citrullination at the periodontium 12 and therefore potentially initiate ACPA.
Despite these observations, to our knowledge, periodontitis and citrullinating bacteria have not been described in individuals at risk of RA. We sought to comprehensively measure periodontitis and the abundance of key citrullinating bacteria in individuals who were ACPA positive (ie, individuals positive for anti-cyclic citrullinated protein [CCP] without synovitis and at risk of RA), patients with anti-CCP-positive early RA (ERA), and healthy control individuals. We hypothesized that (1) periodontitis would be similarly increased in CCP+ at-risk individuals and those with ERA compared with healthy individuals and (2) there would be an increased abundance of citrullinating periodontopathic bacteria in CCP+ at-risk individuals and patients with ERA compared with healthy individuals.

Methods Design
A cross-sectional study of periodontal and clinical parameters was performed in CCP+ at-risk individuals, patients with ERA, and healthy individuals between April 27, 2015, and May 8, 2017. In this exploratory study, we aimed for 30 participants per group, in line with recommendations for pilot studies. Groups were approximately frequency matched during recruitment for age, sex, and smoking status. After 20 CCP+ at-risk individuals, 20 healthy individuals, and 10 patients with ERA were recruited, demographic and smoking data were reviewed. Approximate frequency matching was performed to recruit remaining healthy individuals and patients with ERA within the age range of 31 to 70 years and to recruit balanced numbers within the tertiles of age calculated in the first 20 CCP+ at-risk individual (first and third tertiles 52 and 60 years, respectively). We approximately matched the proportion of participants who had ever smoked, which was 60% in the first 20 CCP+ at-risk individuals.
A shotgun metagenomic analysis was performed on subgingival plaque samples collected during the periodontal assessments.
Ethical approval for this study was provided by the National Research Ethics Service Committee Yorkshire and the Humber, Leeds West. Written informed consent was received from all participants.
This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline. 13

Clinical Participants
Anti-CCP-positive at-risk individuals with musculoskeletal symptoms but no clinical synovitis, patients with ERA, and asymptomatic healthy individuals were recruited at Chapel Allerton Hospital, Leeds, United Kingdom.
We recruited CCP+ at-risk participants from the Leeds at-risk cohort. 14,15 Patients older than 18 years presenting to general practitioners or other health professionals with new-onset musculoskeletal symptoms but no clinical synovitis were invited to participate. Primary care recruitment was adopted nationally by the UK Primary Care Clinical Research Network. Anti-CCP testing was performed centrally using the Bioplex 2200 kit (BioRad). Those with a positive test result were invited to a dedicated research clinic in Leeds where recruitment for this study was undertaken.
Patients from the Leeds early arthritis clinic who were anti-CCP-positive but did not have clinical synovitis were also recruited. Patients with ERA were all anti-CCP-positive and within the first 3 months of disease-modifying antirheumatic drug (DMARD) therapy. Control participants had no joint disease or history of inflammatory arthritis (and no affected first-degree relatives). Control individuals included coworkers at the University of Leeds (eg, academics, administrative workers, laboratory staff, cleaning staff) and people from the local community (eg, lay members of the Leeds Biomedical Research Centre Patient and Public Involvement group, their contacts, and local community groups). Control participants were typical of the general population with a range of socioeconomic groups represented.
Demographic details and a serum IgG anti-CCP2 test (Immunocap; Phadia) were taken at the time of the periodontal assessment.

Periodontal Assessment
Periodontal assessments were performed at Chapel Allerton Hospital, Leeds. Periodontal status was assessed by 3 experienced dentists (V.C., A.S., and A.T.). Dentists were blinded to the RA status of the participants. A full-mouth examination of 6 sites on each natural tooth (recorded as a 6-point pocket chart) was performed on each participant. The 6 sites measured were the 4 corners of the tooth and the midpoint between the buccal and lingual aspects of the tooth. The following parameters were recorded at each available dental site: probing pocket depth (PPD) (millimeters), clinical attachment level (CAL) (millimeters), and presence of bleeding on probing (BOP) (indicated as present or absent). The PCP10 periodontal probe (Hu-Friedy) was used for assessment of BOP and PPD. Measurements of PPD were taken along the vertical axis of the tooth at each site using approximately 0.25 N of force.
Periodontal disease sites were determined according to the recent update to the 1999 American Academy of Periodontology Classification of Periodontal Diseases and Conditions. 16 To ensure high sensitivity, thresholds for CAL and PPD were deliberately set so that sites with mild, moderate, and severe periodontitis would all be captured. Periodontal sites with 2 mm or greater CAL and 4 mm or greater PPD were defined as periodontitis sites (PDD) and considered to represent sites of current or past (including treated) periodontitis. Periodontal sites with 2 mm or greater CAL, 4 mm or greater PPD, and BOP were defined as active PDD and considered to represent current active periodontitis.
In addition to these parameters, dentists also classified all participants according to overall clinical periodontal status. The periodontal medical record for each participant was reviewed by the 3 dentists who were blinded to all patient details. In each case, clinical periodontal status was agreed by consensus and was classified as follows: (1) healthy (no periodontitis or gingivitis), (2) gingivitis only (no periodontitis), or (3) periodontitis with or without gingivitis. Classification was based on clinical judgment and the update to the 1999 American Academy of Periodontology Classification of Periodontal Diseases and Conditions, 16 taking into account the distribution, extent, and severity of periodontitis and also the need for treatment.

Periodontal Inflamed Surface Area
To quantify the total burden of periodontal inflammation, the total periodontal inflamed surface area (PISA) was calculated for each participant from PPD and CAL measurements at each dental site using the method described by Nesse et al. 17 This index has been proposed as a way of more accurately quantifying inflamed periodontal tissues. 17

Subgingival Plaque Collection
Healthy and diseased periodontal sites suitable for subgingival plaque collection were identified by dentists during the periodontal examination. Supragingival plaque was removed with cotton-wool pledgets prior to sample collection (eAppendix 2 in the Supplement).

Shotgun Metagenomic Data Processing
Sequence data were uploaded to the MG-RAST metagenomics analysis pipeline (version 4.03) for quality processing and basic taxonomic analysis. 20 Low-quality regions (bases with quality scores <15) and reads shorter than 15 bp were discarded. Artificial replicate sequences and host-specific species sequences (eg, plant, human, or mouse) were also removed. Taxonomic abundance profiles at species level were generated by annotation against the RefSeq database housed within MG-RAST

Periodontal Assessment
The percentage of periodontal sites with CAL 2 mm or greater, PPD 4 mm or greater, BOP, PDD, and active PDD were all greater in anti-CCP+ at-risk individuals compared with healthy individuals  Table 2). In contrast, there were no differences in any of these parameters between anti-CCP+ at-risk individuals and patients with ERA. The number of missing teeth was higher in patients with ERA compared with healthy individuals, likely reflecting the higher mean age in this group (54 and 49 years, respectively).

Ultrasonographic Assessment
All CCP+ at-risk individuals underwent US assessment. Synovitis was defined as the presence of gray scale synovial hypertrophy greater than or equal to 1 and PD signal greater than or equal to 1 (gray scale Ն1 and PD Ն1) at the same joint. Of 48 CCP+ at-risk individuals, 46 (96%) had no US synovitis,  suggesting the absence of both clinical and subclinical joint inflammation in these subjects. Of the 2 patients who had US synovitis, 1 had synovitis in both first metatarsophalangeal joints and the other in both wrist joints.  (Figure). Interestingly, at diseased periodontal sites, both CCP+ at-risk and healthy participants had a greater abundance of P gingivalis compared with patients with ERA (effect size = 5.12; 95% CI, 3.34-6.90 and 4.60; 95% CI, 2.54-6.66; P < .001, respectively). There were no differences in the abundance of A actinomycetemcomitans or F alocis according to RA status at either healthy or diseased periodontal sites.

Discussion
In this exploratory cross-sectional study, we have identified an increased prevalence of periodontal disease and periodontal inflammation in CCP+ at-risk individuals without joint inflammation.
Furthermore, we identified an increased abundance of the periodontopathic bacterium P gingivalis at the healthy periodontal sites of CCP+ at-risk participants compared with control individuals and patients with ERA. The microbiome in periodontal disease is distinct from that seen in periodontal health, with an increased abundance of pathogenic bacterial communities. 32,33 Porphyromonas gingivalis is a key component of the so-called red complex of gram-negative bacteria associated with periodontal disease. 9 It has been hypothesized that P gingivalis, through posttranslational citrullination of periodontal mucosal proteins, may provide an antigenic source for ACPA in RA. 11 A recent study 34 showed oral priming with P gingivalis can trigger an erosive ACPA-positive arthritis in an in vivo animal model. In the current study, we found an increased abundance of P gingivalis in subgingival plaque from the healthy periodontal sites of CCP+ at-risk participants compared with healthy individuals and patients with ERA. This association was not seen at diseased periodontal sites where P gingivalis was identified at similar levels in CCP+ at-risk and healthy participants. The reason for the lower abundance of P gingivalis at diseased sites in patients with ERA compared with the other groups is not fully clear, but this may be due to an early effect of therapy; 62% of these patients were receiving DMARDs and most had also received corticosteroids. It is possible that these treatments may have had an early influence on the periodontal microbiome. As periodontitis may be caused by a dysregulated inflammatory response initiated by the biofilm, it is possible that immunomodulatory therapies could have a direct effect on periodontal inflammation, with consequent effects on the microbiome. Also, DMARDs are believed to have antimicrobial properties. 35-37 It is possible that P gingivalis may be affected in this way, and this would be an interesting area to explore in future work. matching during the recruitment period, the final group numbers were unequal; despite the planned number of participants being exceeded, fewer patients were identified and recruited in the healthy control group and ERA group compared with the CCP+ at-risk group. Some individuals from all groups declined study participation. Unfortunately, details of eligible participants who declined in each group are not available. It is possible that this self-selection may have introduced bias. However, a participant survey conducted prior to periodontal assessment suggests no difference in access to dental care or self-reported oral symptoms between CCP+ at-risk and healthy control groups (eAppendix 3 and the eTable in the Supplement). Although participant groups were matched for smoking status, we did not match for diabetes, which is also associated with periodontal disease.
However, the prevalence of diabetes in our participants was low and unlikely to have influenced the data (eAppendix 4 in the Supplement). We acknowledge that other rare systemic conditions are associated with periodontal disease (eg, neutrophil disorders, epidermolysis bullosa, Ehlers-Danlos syndrome, hematological malignant neoplasms). 40 We were not aware of our participants being affected by these conditions, but they were not specifically excluded. These limitations mean that the findings of this study must be considered as exploratory. These data should be validated in other cohorts, and longitudinal follow-up will be important to assess whether periodontal disease predicts the onset of clinical arthritis in CCP+ at-risk individuals.

Conclusions
This study is the first, to our knowledge, to demonstrate an increased prevalence of periodontal disease together with an increased abundance of P gingivalis in anti-CCP-positive individuals at risk of RA. These data suggest periodontal inflammation and the enrichment of P gingivalis may precede joint inflammation in RA and support an association between these risk factors and disease initiation.
This study adds to an emerging evidence base linking periodontal and systemic disease and, therefore, further highlights the potential importance of improving dental health and reducing the burden of periodontal disease on the risk of chronic systemic diseases such as RA. Importantly, these findings suggest periodontal inflammation may be a legitimate target to explore for preventive intervention in RA.