Periodontitis and Oral Bacteria Related to Rheumatoid Arthritis: A Case-Control Study

Abstract

Introduction and aims

This study compared periodontal status and oral bacteria between rheumatoid arthritis (RA) patients and healthy controls (HCs), and examined the influence of oral bacteria on the association between periodontitis and RA.

Methods

In total, 85 patients with RA and 119 HCs were enrolled. The oral microflora DNA test was used to quantify the oral bacterial species detected in gingival crevicular fluid. Probing depth and the clinical attachment level of the periodontal ligament were taken as parameters of periodontal status. Height, body weight, medical history, family history of RA, lifestyle habits, and stress were evaluated using a self-administered questionnaire. Univariate and multivariate logistic regression analyses were conducted to assess the association between RA and periodontal status/oral bacteria.

Results

RA patients exhibited significantly greater probing depth than HCs. The HCs demonstrated higher abundances of Fusobacterium nucleatum subsp. polymorphum, Fusobacterium periodonticum, Campylobacter showae, Campylobacter gracilis, Eikenella corrodens, Streptococcus mitis, Streptococcus mitis bv 2, and Actinomyces naeslundii II. In forward stepwise multivariate analysis, the odds ratios (ORs) for RA were significantly higher for patients with a family history of RA, smokers, those with deep periodontal pockets, and those with a larger population of F. nucleatum subsp. animalis and Veillonella parvula. Patients with more Campylobacter gracilis had a significantly lower OR for RA.

Conclusion

A comparison of the oral bacteria of RA patients and HCs suggests that F. nucleatum subsp. animalis and V. parvula are involved in RA patients. However, there are still many unknowns about the relationship between oral bacteria and RA, and further research is needed.

Introduction

Periodontitis is a chronic inflammatory disease induced by periodontal bacteria in dental plaque characterised by the destruction of periodontal tissues.^1,2 It is associated with many systemic diseases,3, 4, 5 including rheumatoid arthritis (RA). RA is an immune disorder associated with swelling and joint pain.⁶ Its prevalence in Japan is estimated to be 0.5-1.0%, with a predilection for women in their 30s to 50s.⁷ RA causes chronic synovitis, in which cytokines and enzymes are produced, resulting in the proliferation of synovial cells in the joints and destruction of cartilage and bone.⁶

There are many reports of the association between RA and periodontitis.8, 9, 10 Among RA patients, those with poor periodontal status tend to have more severe disease.¹¹ In a follow-up study of arthralgia patients without RA, those with periodontal pockets ≥ 4 mm were at higher risk of developing RA.¹² Therefore, periodontitis is considered to be conducive to the development and aggravation of RA. Indeed, RA patients have a higher incidence of periodontitis than healthy individuals.¹³

Periodontitis involves many oral bacteria.14, 15, 16 Porphyromonas gingivalis (P. gingivalis) is thought to be involved in citrullination of periodontal tissue proteins and evokes an autoimmune response in RA.¹⁷ In addition, Aggregatibacter actinomycetemcomitans produces leukotoxins and activates the citrullinated conversion enzyme in neutrophils, enhancing the production of anti-citrullinated peptide antibody (ACPA).¹⁸ Therefore, the oral bacteria involved in periodontitis may also be involved in the development and aggravation of RA. However, some studies have reported contrary results, for example that P. gingivalis does not lead to the citrullination of periodontal tissue proteins.^19,20 Overall, the influence of oral bacteria on the association between periodontitis and RA remains unresolved.

This study compared periodontal status and oral bacteria between RA patients and healthy controls (HCs) and examined the influence of oral bacteria on the association between periodontitis and RA.

Methods

Participants

We conducted a case–control study including 85 patients with RA (18 males and 67 females aged 27 to 85 years; mean age: 61.9 ± 12.4) and 119 HCs (14 males and 105 females aged 58-74 years; mean age: 66.6 ± 10.8). RA affects women more frequently than men.²¹ In a study based on the National Database of Health Insurance Claims and Specific Health Checkup of Japan, the prevalence of RA was estimated, revealing that 76.27% of patients were female, a proportion comparable to that observed in the present study.²¹

RA patients who were undergoing treatment and routine follow-up at an orthopaedic clinic in Aichi Prefecture, Japan, were enrolled in the RA group. Classification of the RA patients was performed according to the 1987 American College of Rheumatology (ACR) RA criteria²² or the 2010 ACR/European League Against Rheumatism (EULAR) RA criteria.²³ Patients aged ≥18 years were included. Pregnant and edentulous patients were excluded.

HCs without RA were recruited from a university dental hospital in Aichi Prefecture, Japan (HC group). Individuals aged ≥18 years who had never been diagnosed with RA in the past or present were included. Pregnant individuals and those who were edentulous were excluded, consistent with the exclusion criteria applied to the RA group.

This study will be conducted from October 2019 to December 2021 and was approved by the Ethics Committee of Aichi Gakuin University, School of Dentistry (approval number 572). All study participants provided written informed consent.

Oral health examination

The oral health examination in the RA group was performed by 1 dentist in an orthopaedic clinic, using a portable dental chair and adequate artificial lighting. The dentist was blinded to the patient’s health data, including the RA data. The examination in the HC group was performed by 3 dentists in the treatment room of the university dental hospital. Tooth condition and periodontal status were evaluated. Each tooth’s condition was recorded as sound, decayed, filled, or missing (DMF). The total number of sound, decayed, and filled teeth (excluding third molars) was taken as the total number of teeth. For investigation of periodontal status, the periodontal probing depth (PD) and clinical attachment level (CAL) of the periodontal ligament were assessed at 6 points per tooth using a periodontal probe (PCPN15; Hu-Friedy, Chicago, IL, USA). Three dentists with calibrated inter-examiner kappa index values of ≥ 0.83 examined the participants for DMF teeth. Intra-class correlation coefficients (ICCs) for repeated measures ranged between 0.70 and 0.79 for PD and between 0.69 and 0.79 for CAL.

Microbial analysis

• 1)Collection of samples for microbiome analyses.

  For the extraction of microbiologic samples, the tooth with the deepest PD in each participant was selected. After removing the plaque above the gumline, a subgingival biofilm sample was taken. Then, these samples were combined using 2 sterile pieces of absorbent paper inserted for 10 s after ensuring the area was dry.

• 2)Quantitative detection of 28 species of oral bacteria.

  We used the oral microflora DNA test (Periodontal Pathogenic Bacteria Test kit; GC, Tokyo, Japan) for quantitative detection of 28 species of oral bacteria and total bacterial counts in gingival crevicular fluid using a DNA chip.^24,25 DNA was extracted from cultured bacteria or GCF samples using QuickGene-800 (Fujifilm, Japan) and fluorescently labelled. Competitive polymerase chain reaction (PCR) and hybridization were carried out in the following steps. For PCR, the V3 forward primer (5’-Cy5-TACGGGAGGCAGCAG-3’) and V4 reverse primer (5’-TACCIGGGTATCTAATCC- 3’) were used. PCR was conducted using 0.5 amol of control DNA, 20 pmol of each primer, 10 µL 2 × PCR solution Premix Ex Taq, Hot-start version (Takara, Shiga, Japan), and template (as described below) in a total volume of 20 µL. The reaction was started through initial denaturation for 1 min at 95°C, followed by 40 cycles at 98°C for 10 s, 55°C for 30 s, and 72°C for 20 s. The amplicon length is approximately 440 bp. The PCR product was directly suspended in 180 µL hybridisation solution (48 µL 1 M tris-HCl pH7.5, 48 µL 1 M NaCl, 20 µL 0.5% Tween-20, and 64 µL Milli-Q water), hybridised with the probes on the DNA Chip at 50°C for 16 h, and washed with the Genopal instrument system (Mitsubishi Chemical, Tokyo, Japan). Hybridization signal intensity (SI) was determined using multi-beam excitation technology and the Genopal reader (Mitsubishi Chemical). SI for subsequent analyses was obtained by deducting the median SI of the background spots from the median SI of the 5 spots on each probe. For each array, the median SI of the background spots + 3σ was treated as the detection limit value. As PCR templates, MSA-1003 containing mixed genomic material of 20 strains (American Type Culture Collection, Manassas, VA, USA), plasmid DNA, or subgingival plaque samples were used. As the first step in quantitative detection, we measured the total amount of 16S rRNA using a standard calibration curve plotted with reference to a previous method.²⁵ Next, we determined the number of each bacterial species using species-specific probes. Then, these results were corrected using the hybridization affinity ratio. Data from the Ribosomal RNA Database (version 5.5; Schmidt Laboratory at the University of Michigan, Ann Arbor, MI, USA) were used to determine the number of copies of 16S rRNA. In the absence of appropriate information, the median value for the genus was used. To calculate the total number of bacteria in samples, the 16S rRNA copy number relative to genomic DNA was assumed to be 4.5, calculated based on a weighted average reported in a study in which the predominant and prevalent bacterial species in the saliva of orally healthy subjects were determined via pyrosequencing.²⁶ The bacterial counts were calculated by multiplying the molecular weight of the genome by Avogadro’s constant (i.e. the molecular weight of 16S rRNA divided by the number of 16S rRNA copies).

Questionnaire

Height, body weight, medical history, family history of RA, lifestyle habits, and stress were evaluated using a self-administered questionnaire. Body mass index (BMI) is calculated by dividing body weight by height squared (kg/m²). Information about the family history of RA was obtained via the question “Does anyone in your family have RA?,” with answer options of “yes” and “no.” Regarding lifestyle habits, we asked about the frequency of alcohol consumption (every day, 3-4 times/week, 1-2 times/week, rarely, or never) and smoking habits (never a smoker, former smoker, or current smoker). Information about stress was obtained via the question “Do you usually feel stressed?” Stress was categorised as “never feel,” “rarely feel,” “sometimes feel,” and “always feel.”

Statistical analysis

We used the mean values of the PD and CAL measurements in all analyses. The counts of oral bacteria were logarithmically transformed to base 10. The frequency of alcohol consumption was divided into 2 categories (no, <1 time/week; yes, ≥1 time/week). The stress level was divided into 2 categories (no: never or rarely; yes: sometimes or always).

The Mann-Whitney U test and the chi-square test were used to compare the groups. We performed univariate and forward stepwise multivariate logistic regression analyses to determine the effect of oral bacteria on RA, calculating the odds ratios (ORs) and 95% confidence intervals (CIs). Age, BMI, sex, and factors with a P value < 0.2 between the groups in univariate analyses were included as independent variables. The adequacy of the sample size for multivariate logistic regression analysis was confirmed by ensuring that the number of minority categories in the dependent variable was at least 10-fold higher than the number of independent variables included in the analysis.²⁷ We performed Spearman rank correlation to confirm relationships among oral bacteria. Because Streptococcus mitis (S. mitis) and Streptococcus mitis bv 2 (S. mitis bv 2), as well as Campylobacter showae (C. showae) and Campylobacter gracilis (C. gracilis), had high correlation coefficients, and because S. mitis bv 2 and C. showae showed weak effects on RA, they were not included as independent variables. In addition, Spearman rank correlation was used to examine the relationship between mean PD, mean CAL, and oral bacteria. Multivariable linear regression analysis was performed to determine the effect of oral bacteria on the mean PD or the mean CAL. Age, BMI, sex, and factors with a P value < .2 in univariate analyses were included as independent variables. In the multivariable analyses, we used 3 models depending on the oral bacteria included in the analysis. Model 1 included only P. gingivalis, Model 2 included the red complex, and Model 3 included oral bacteria with a P value < .2 in univariate analyses as independent variables. The minimum required sample size for multivariable regression analysis was calculated using G*Power 3.1.^28,29 A sample size of 163 participants was calculated based on an effect size (f²) of 0.15,³⁰ 22 independent variables, statistical power of 0.80, and an alpha error of 0.05. For analyses involving 13 independent variables, 131 participants were needed. The statistical analyses were performed using SPSS ver. 26.0 (IBM, Tokyo, Japan). P values < .05 indicate statistical significance.

Results

Table 1 shows the characteristics of each group. Participants in the RA group tended to be younger (P = .005), more likely to have a family history of RA (P = .011), more likely to be smokers (P < .001), and were more stressed with a greater mean PD (P = .033).

Table 1.

Characteristics of the study population.

	RA group (n = 85)	HC group (n = 119)	P value
	Median (25 percentile, 75 percentile)
Age	63.0 (53.5, 71.0)	69.0 (58.0, 74.0)	.005
BMI (kg/m2)	22.2 (19.9, 24.6)	21.2 (19.1, 23.7)	.119
	N (%)
Sex
Male	18 (21.2)	14 (11.8)	.080
Female	67 (78.8)	105 (88.2)
Diabetes
No	78 (91.8)	116 (97.5)	.097
Yes	7 (8.2)	3 (1.5)
Family history of RA
No	70 (82.4)	112 (94.1)	.011
Yes	15 (17.6)	7 (5.9)
Daily alcohol drinking
No	61 (71.8)	78 (65.5)	.365
Yes	24 (28.2)	41 (34.5)
Smoking habit
Never smoker	48 (56.5)	102 (85.7)	<.001
Former smoker	24 (28.2)	16 (13.4)
Current smoker	13 (15.3)	1 (0.8)
Having stress
No	31 (36.5)	62 (52.1)	.033
Yes	54 (63.5)	57 (47.9)
DMF teeth	18.0 (12.0, 23.0)	17.0 (12.0, 21.0)	.506
Mean PD (mm)	2.38 (2.19, 2.54)	2.18 (1.90, 2.38)	<.001
Mean CAL (mm)	2.42 (2.27, 2.63)	2.51 (2.28, 2.77)	.214

RA, rheumatoid arthritis; HC, healthy control; BMI, body mass index; DMF, decayed, missing, or filled; PD, pocket depth; CAL, clinical attachment level.

Table 2 shows the numbers of oral bacteria in each group. The HCs had more Fusobacterium nucleatum subsp. polymorphum (P = .027), Fusobacterium periodonticum (F. periodonticum) (P = .001), C. showae (P < .001), C. gracilis (P < .001), and Eikenella corrodens (E. corrodens) (P < .001), S. mitis (P < .001), and S. mitis bv 2 (P < .001), and Actinomyces naeslundii II (A. naeslundii II) (P = .010).

Table 2.

The number of each oral bacteria of RA patients and controls.

Number of oral bacteria (Logarithmic scale)	RA group	HC group	P value
	Median (25 percentile, 75 percentile)
Porphyromonas gingivalis	4.01 (0, 5.56)	3.52 (0, 4.94)	.314
Tannerella forsythia	4.72 (3.52, 5.49)	4.54 (3.27, 5.68)	.792
Treponema denticola	4.13 (3.31, 5.44)	3.85 (3.08, 5.38)	.632
Campylobacter rectus	4.96 (4.15, 6.05)	4.81 (0, 6.10)	.443
Fusobacterium nucleatum subsp. polymorphum	4.42 (0, 5.18)	4.91 (4.11, 5.48)	.027
Fusobacterium nucleatum subsp. animalis	5.89 (5.20, 6.34)	5.74 (4.81, 6.42)	.441
Fusobacterium nucleatum subsp. nucleatum	5.60 (4.86, 5.92)	5.36 (4.59, 5.92)	.464
Fusobacterium periodonticum	5.03 (4.62, 5.52)	5.46 (4.95, 5.82)	.001
Fusobacterium nucleatum subsp. vincentii	4.74 (4.22, 5.21)	4.93 (4.46, 5.32)	.076
Prevotella nigrescens	4.64 (0, 5.60)	4.32 (0. 5.51)	.489
Prevotella intermedia	3.53 (0, 5.15)	0 (0, 4.67)	.019
Streptococcus constellatus	0 (0, 4.50)	0 (0, 4.11)	.462
Campylobacter showae	0 (0, 4.34)	4.40 (0, 4.83)	<.001
Campylobacter gracilis	0 (0, 4.49)	4.34 (0, 4.81)	<.001
Aggregatibacter actinomycetemcomitans	0 (0, 2.78)	0 (0, 2.80)	.966
Campylobacter concisus	3.06 (0, 4.09)	2.40 (0, 3.96)	.590
Capnocytophaga gingivalis	3.56 (0, 4.09)	3.47 (0, 4.33)	.641
Capnocytophaga ochracea	3.89 (0, 4.70)	4.17 (0, 4.59)	.356
Capnocytophaga sputigena	4.38 (0, 4.93)	4.54 (0, 5.01)	.196
Eikenella corrodens	0 (0, 3.55)	3.60 (0, 4.06)	<.001
Streptococcus intermedius	3.56 (1.22, 4.61)	3.93 (2.91, 4.72)	.330
Streptococcus gordonii	4.21 (3.77, 4.65)	4.15 (3.51, 4.64)	.624
Streptococcus mitis	3.45 (1.55, 4.12)	4.18 (3.63, 4.64)	<.001
Streptococcus mitis bv 2	3.60 (3.18, 4.20)	4.10 (3.58, 4.54)	<.001
Actinomyces odontolyticus	2.68 (2.48, 2.99)	2.76 (2.50, 2.99)	.443
Veillonella parvula	3.97 (0, 4.72)	0 (0. 4.55)	.105
Actinomyces naeslundii II	3.71 (0, 4.19)	3.99 (2.98, 4.60)	.010
Selenomonas noxia	4.12 (0, 4.83)	3.67 (0, 4.78)	.203

RA, rheumatoid arthritis; HC, healthy control.

Number of each oral bacteria was logarithmically converted to 10 as a base.

The results of univariate and forward stepwise multivariate logistic regression analyses are shown in Table 3. In the forward stepwise multivariate analysis, the ORs of RA were significantly higher among individuals with a family history of RA (OR = 4.18; 95% CI = 1.23-14.2; P = .022), former smokers (OR = 3.49; 95% CI = 1.37-8.93; P = .009), current smokers (OR = 20.1; 95% CI = 1.94-207.3; P = .012), those with a high mean PD (OR = 12.4; 95% CI = 3.05-50.0; P = < .001), and those with a greater abundance of Fusobacterium nucleatum subsp. animalis (F. nucleatum subsp. animalis) (OR = 1.76; 95% CI = 1.17-2.64; P = .006), and those with a greater abundance of Veillonella parvula (V. parvula) (OR = 1.26; 95% CI = 1.05-1.50; P = .012). Conversely, a greater abundance of C. gracilis was associated with significantly lower ORs for RA (OR = 0.50; 95% CI = 0.40-0.63; P = < .001).

Table 3.

Factors associated with RA in the logistic regression analysis.

	Dependent variable (HC group = 0, RA group = 1)
Independent variable	Crude OR (95% CI)	P value	Adjusted OR (95% CI)	P value
Age	0.97 (0.94, 0.99)	.005
BMI (kg/m2)	1.05 (0.97, 1.14)	.210
Sex
Male	1
Female	0.50 (0.23, 1.06)	.072
Diabetes
No	1
Yes	3.47 (0.87, 13.9)	.078
Family history of RA
No	1		1
Yes	3.43 (1.33, 8.83)	.011	4.18 (1.23, 14.2)	.022
Smoking habit
Never smoker	1		1
Former smoker	3.12 (1.56, 6.55)	.002	3.49 (1.37, 8.93)	.009
Current smoker	27.7 (3.51, 217.3)	.002	20.1 (1.94, 207.3)	.012
Having stress
No	1
Yes	1.90 (1.07, 3.35)	.028
Mean PD (mm)	15.0 (5.02, 44.6)	<.001	12.4 (3.05, 50.0)	<.001
Number of oral bacteria (Logarithmic scale)
Campylobacter rectus	1.08 (0.96, 1.22)	.197
Fusobacterium nucleatum subsp. polymorphum	0.87 (0.77, 0.99)	.031
Fusobacterium nucleatum subsp. animalis	1.20 (0.95, 1.52)	.120	1.76 (1.17, 2.64)	.006
Prevotella intermedia	1.14 (1.02, 1.27)	.018
Campylobacter gracilis	0.71 (0.62, 0.81)	<.001	0.50 (0.40, 0.63)	<.001
Capnocytophaga ochracea	0.91 (0.80, 1.03)	.144
Eikenella corrodens	0.74 (0.64, 0.86)	<.001
Streptococcus mitis	0.69 (0.57, 0.83)	<.001
Veillonella parvula	1.11 (0.99, 1.26)	.081	1.26 (1.05, 1.50)	.012
Actinomyces naeslundii II	0.85 (0.73, 0.98)	.024
Selenomonas noxia	1.10 (0.97, 1.25)	.158

HC, healthy control; OR, odds ratios; CI, confidence interval; RA, rheumatoid arthritis; PD, pocket depth.

Tables 4 and 5 show the correlations between mean PD and mean CAL and the numbers of each bacteria. P. gingivalis, Tannerella forsythia (T. forsythia), Treponema denticola (T. denticola), and F. nucleatum subsp. animalis, Fusobacterium nucleatum subsp. nucleatum (F. nucleatum subsp. nucleatum), Prevotella nigrescens (P. nigrescens), Prevotella intermedia (P. intermedia), Streptococcus constellatus (S. constellatus), V. parvula, and red complex were significantly positively correlated with mean PD in all individuals (P < .001, < .001, = .001, = .001, = .005, < .001, = .001, = .017, = .038, and < .001, respectively) . E. corrodens, S. mitis, and A. naeslundii II were significantly negatively correlated with mean PD in all individuals (P = .043, = .019, and = .011, respectively). P. gingivalis, T. forsythia, T. denticola, Campylobacter rectus, and F. nucleatum subsp. animalis, F. nucleatum subsp. nucleatum, P. nigrescens, S. constellatus, Streptococcus gordonii, and red complex were significantly positively correlated with mean PD in the RA group (P < .001, < .001, = .017, = .044, = .009, = .021, = .006, = .037, = .021, and < .001, respectively). T. forsythia, T. denticola, P. nigrescens, P. intermedia, and red complex were significantly positively correlated with mean PD in the HC group (P = .025, = .027, = .037, = .034, and = .011, respectively). Actinomyces odontolyticus and A. naeslundii II were significantly negatively correlated with mean PD in the HC group (P = .021 and = .005, respectively). P. gingivalis, T. forsythia, T. denticola, and F. nucleatum subsp. animalis, F. nucleatum subsp. nucleatum, F. periodonticum, and Fusobacterium nucleatum subsp. vincentii (F. nucleatum subsp. vincentii), P. intermedia, S. constellatus, and red complex were significantly positively correlated with mean CAL in all individuals (P < .001, < .001, = .029, = .006, = .010, = .027, = .008, = .024, = .007, and < .001, respectively). P. gingivalis, T. forsythia, and red complex were significantly positively correlated with mean CAL in the RA group (P < .001, = .001, and < .001, respectively). P. gingivalis, T. forsythia, F. nucleatum subsp. animalis, F. nucleatum subsp. nucleatum, F. nucleatum subsp. vincentii, P. intermedia, S. constellatus, and red complex were significantly positively correlated with mean CAL in the HC group (P = .008, = .015, = .011, = .023, = .027, = .012, = .035, and = .009, respectively).

Table 4.

Correlation between mean PD and the number of each oral bacteria.

	Dependent variable: Mean PD (mm)
	All subjects		RA group		HC group
	(n = 204)	P value	(n = 85)	P value	(n = 119)	P value
Number of oral bacteria (Logarithmic scale)
Porphyromonas gingivalis	r = 0.28	<.001	r = 0.50	<.001	r = 0.12	.179
Tannerella forsythia	r = 0.28	<.001	r = 0.50	<.001	r = 0.21	.025
Treponema denticola	r = 0.23	.001	r = 0.26	.017	r = 0.20	.027
Campylobacter rectus	r = 0.13	.075	r = 0.22	.044	r = 0.05	.596
Fusobacterium nucleatum subsp. polymorphum	r = 0.05	.461	r = 0.20	.071	r = 0.05	.620
Fusobacterium nucleatum subsp. animalis	r = 0.24	.001	r = 0.28	.009	r = 0.18	.051
Fusobacterium nucleatum subsp. nucleatum	r = 0.20	.005	r = 0.25	.021	r = 0.14	.130
Fusobacterium periodonticum	r = 0.05	.503	r = 0.21	.059	r = 0.10	.292
Fusobacterium nucleatum subsp. vincentii	r = 0.11	.108	r = 0.20	.063	r = 0.14	.141
Prevotella nigrescens	r = 0.24	<.001	r = 0.29	.006	r = 0.19	.037
Prevotella intermedia	r = 0.23	.001	r = 0.17	.127	r = 0.20	.034
Streptococcus constellatus	r = 0.17	.017	r = 0.23	.037	r = 0.13	.147
Campylobacter showae	r = −0.004	.958	r = 0.19	.090	r = 0.07	.475
Campylobacter gracilis	r = −0.03	.722	r = 0.14	.204	r = 0.04	.655
Aggregatibacter actinomycetemcomitans	r = 0.03	.638	r = 0.09	.399	r = 0.02	.839
Campylobacter concisus	r = −0.06	.415	r = −0.09	.403	r = −0.07	.482
Capnocytophaga gingivalis	r = −0.02	.821	r = 0.08	.465	r = −0.05	.600
Capnocytophaga ochracea	r = −0.12	.094	r = −0.13	.254	r = −0.11	.220
Capnocytophaga sputigena	r = −0.08	.232	r = 0.001	.996	r = −0.10	.272
Eikenella corrodens	r = −0.14	.043	r = −0.05	.647	r = −0.02	.843
Streptococcus intermedius	r = 0.02	.799	r = 0.08	.449	r = 0.03	.777
Streptococcus gordonii	r = 0.03	.720	r = 0.25	.021	r = −0.11	.227
Streptococcus mitis	r = −0.16	.019	r = 0.11	.302	r = −0.15	.100
Streptococcus mitis bv 2	r = −0.12	.091	r = 0.19	.075	r = −0.14	.132
Actinomyces odontolyticus	r = −0.10	.144	r = 0.09	.391	r = −0.21	.021
Veillonella parvula	r = 0.15	.038	r = 0.12	.278	r = 0.10	.276
Actinomyces naeslundii II	r = −0.18	.011	r = 0.04	.714	r = −0.26	.005
Selenomonas noxia	r = 0.10	.165	r = 0.09	.392	r = 0.02	.806
Red complex*	r = 0.31	<.001	r = 0.49	<.001	r = 0.23	.011

PD, pocket depth; RA, rheumatoid arthritis; HC, healthy control.

^⁎Red complex includes Porphyromonas gingivalis, Tannerella forsythia, and Treponema denticola.

Table 5.

Correlation between mean CAL and the number of each oral bacteria.

	Dependent variable: Mean CAL (mm)
	All subjects		RA group		HC group
	(n = 204)	P value	(n = 85)	P value	(n = 119)	P value
Number of oral bacteria (Logarithmic scale)
Porphyromonas gingivalis	r = 0.31	<.001	r = 0.43	<.001	r = 0.24	.008
Tannerella forsythia	r = 0.27	<.001	r = 0.36	.001	r = 0.22	.015
Treponema denticola	r = 0.15	.029	r = 0.15	.168	r = 0.18	.051
Campylobacter rectus	r = 0.08	.248	r = 0.12	.281	r = 0.07	.452
Fusobacterium nucleatum subsp. polymorphum	r = 0.10	.141	r = 0.17	.126	r = 0.03	.740
Fusobacterium nucleatum subsp. animalis	r = 0.19	.006	r = 0.16	.148	r = 0.23	.011
Fusobacterium nucleatum subsp. nucleatum	r = 0.18	.010	r = 0.16	.134	r = 0.21	.023
Fusobacterium periodonticum	r = 0.16	.027	r = 0.09	.412	r = 0.18	.053
Fusobacterium nucleatum subsp. vincentii	r = 0.18	.008	r = 0.14	.212	r = 0.20	.027
Prevotella nigrescens	r = 0.12	.102	r = 0.17	.118	r = 0.09	.329
Prevotella intermedia	r = 0.16	.024	r = 0.14	.203	r = 0.23	.012
Streptococcus constellatus	r = 0.19	.007	r = 0.19	.090	r = 0.19	.035
Campylobacter showae	r = 0.14	.046	r = 0.14	.206	r = 0.09	.328
Campylobacter gracilis	r = 0.10	.152	r = 0.07	.526	r = 0.08	.397
Aggregatibacter actinomycetemcomitans	r = −0.02	.743	r = 0.02	.849	r = −0.04	.688
Campylobacter concisus	r = −0.01	.859	r = −0.08	.449	r = 0.06	.528
Capnocytophaga gingivalis	r = −0.08	.284	r = −0.003	.977	r = −0.10	.258
Capnocytophaga ochracea	r = −0.14	.046	r = −0.20	.072	r = −0.12	.190
Capnocytophaga sputigena	r = −0.11	.136	r = −0.10	.376	r = −0.11	.241
Eikenella corrodens	r = −0.11	.109	r = −0.12	.275	r = −0.16	.093
Streptococcus intermedius	r = 0.04	.551	r = 0.07	.556	r = 0.03	.747
Streptococcus gordonii	r = 0.05	.514	r = 0.16	.136	r = −0.01	.882
Streptococcus mitis	r = 0.02	.837	r = 0.06	.611	r = −0.04	.643
Streptococcus mitis bv 2	r = 0.05	.522	r = 0.14	.188	r = −0.04	.686
Actinomyces odontolyticus	r = −0.01	.900	r = 0.11	.306	r = −0.11	.254
Veillonella parvula	r = 0.03	.645	r = 0.05	.622	r = 0.03	.714
Actinomyces naeslundii II	r = 0.04	.554	r = 0.02	.873	r = 0.03	.725
Selenomonas noxia	r = 0.03	.706	r = 0.04	.697	r = 0.03	.735
Red complex*	r = 0.28	<.001	r = 0.39	<.001	r = 0.23	.009

CAL, clinical attachment level; RA, rheumatoid arthritis; HC, healthy control.

^⁎Red complex includes Porphyromonas gingivalis, Tannerella forsythia, and Treponema denticola.

Table 6 shows the results of the multivariable linear regression analysis of each independent variable, with the mean PD as the dependent variable. In Model 1, the mean PD increased by 0.02 mm for every 10-fold increase in P. gingivalis (β = 0.02, 95% CI = 0.004-0.04, P = .014). In the RA group, the mean PD was 0.19 mm greater than in the HC group (β = 0.19, 95% CI = 0.11-0.28, P < .001). In Model 2, the mean PD increased by 0.05 mm for every 10-fold increase in the red complex (β = 0.05, 95% CI = 0.03-0.08, P < .001). In the RA group, the mean PD was 0.20 mm greater, as compared to the HC group (β = 0.20, 95% CI = 0.11-0.28, P = < .001). In Model 3, the mean PD increased by 0.02 mm for every 10-fold increase in P. nigrescens (β = 0.02, 95% CI = 0.00-0.04, P = .046). Conversely, the mean PD decreased by 0.03 mm for every 10-fold increase in Capnocytophaga ochracea (β = -0.03, 95% CI = -0.05 to -0.003, P = .028). In the RA group, the mean PD was 0.19 mm greater than in the HC group (β = 0.19, 95% CI = 0.1-0.28, P = < .001).

Table 6.

Regression coefficients of independent variables against mean PD in multivariate linear regression models.

	Dependent variable: Mean PD (mm)
	Model 1		Model 2		Model 3
Independent variable	β (95% CI)	P value	β (95% CI)	P value	β (95% CI)	P value
Age (years)	0.002 (−0.002, 0.006)	.318	0.001 (−0.003, 0.004)	.773	0.001 (−0.003, 0.005)	.679
BMI (kg/m2)	0.01 (−0.002, 0.02)	.094	0.01 (−0.003, 0.02)	.140	0.01 (−0.001, 0.02)	.076
Sex: female (reference: male)	−0.11 (−0.24, 0.01)	.072	−0.11 (−0.23, 0.10)	.072	−0.12 (−0.25, 0.003)	.055
Diabetes: yes (reference: no)	0.08 (−0.11, 0.27)	.396	0.09 (−0.10, 0.28)	.350	0.07 (−0.13, 0.27)	.482
Smoking status (reference: never smoker)
Former smoker	0.005 (−0.11, 0.12)	.926	−0.01 (−0.12, 0.1)	.849	−0.01 (−0.13, 0.10)	.842
Current smoker	0.15 (−0.03, 0.33)	.093	0.16 (−0.01, 0.34)	.069	0.15 (−0.03, 0.33)	.096
Number of oral bacteria (Logarithmic scale)
Porphyromonas gingivalis	0.02 (0.004, 0.04)	.014			0.01 (−0.01, 0.03)	.322
Tannerella forsythia					0.04 (−0.01, 0.08)	.084
Treponema denticola					−0.001 (−0.04, 0.04)	.943
Campylobacter rectus					−0.005 (−0.03, 0.02)	.615
Fusobacterium nucleatum subsp. animalis					−0.02 (−0.09, 0.06)	.599
Fusobacterium nucleatum subsp. nucleatum					−0.03 (−0.11, 0.05)	.399
Fusobacterium nucleatum subsp. vincentii					0.04 (−0.02, 0.09)	.240
Prevotella nigrescens					0.02 (0.00, 0.04)	.046
Prevotella intermedia					0.003 (−0.02, 0.02)	.755
Streptococcus constellatus					0.001 (−0.02, 0.02)	.917
Capnocytophaga ochracea					−0.03 (−0.05, −0.003)	.028
Capnocytophaga sputigena					−0.002 (−0.03, 0.02)	.888
Eikenella corrodens					−0.001 (−0.03, 0.02)	.958
Streptococcus mitis					0.008 (−0.02, 0.04)	.572
Actinomyces naeslundii II					−0.02 (−0.04, 0.01)	.151
Red complex*			0.05 (0.03, 0.08)	<.001
RA group (reference: HC group)	0.19 (0.11, 0.28)	<.001	0.20 (0.11, 0.28)	<.001	0.19 (0.1, 0.28)	<.001

PD, pocket depth; CI, confidence interval; BMI, body mass index; RA, rheumatoid arthritis; HC, healthy control.

^⁎Red complex includes Porphyromonas gingivalis, Tannerella forsythia, and Treponema denticola.

Table 7 shows the results of multivariable linear regression of each independent variable, with mean CAL as the dependent variable. In Model 1, the mean CAL increased by 0.01 mm for each 1-year increase in age (β = 0.01, 95% CI = 0.005-0.02, P < .001) and was 0.64 mm greater in current smokers than in never smokers (β = 0.64, 95% CI = 0.35-0.94, P < .001). In Model 2, the mean CAL increased by 0.01 mm for each 1-year increase in age (β = 0.01, 95% CI = 0.006-0.02, P < .001) and was 0.67 mm greater in current smokers than in never smokers (β = 0.67, 95% CI = 0.37-0.97, P < .001). In Model 3, the mean CAL increased by 0.01 mm for each 1-year increase in age (β = 0.01, 95% CI = 0.004-0.02, P = .001). The mean CAL of females was 0.22 mm lower than that of males (β = -0.22, 95% CI = -0.43 to -0.02, P = .033). Additionally, the mean CAL of current smokers was 0.60 mm greater than that of never smokers (β = 0.60, 95% CI = 0.30-0.90, P < .001).

Table 7.

Regression coefficients of independent variables against mean CAL in multivariate linear regression models.

	Dependent variable: Mean CAL (mm)
	Model 1		Model 2		Model 3
Independent variable	β (95% CI)	P value	β (95% CI)	P value	β (95% CI)	P value
Age (years)	0.01 (0.005, 0.02)	<.001	0.01 (0.006, 0.02)	<.001	0.01 (0.004, 0.02)	.001
BMI (kg/m2)	−0.002 (−0.02, 0.02)	.858	0.002 (−0.02, 0.02)	.841	0.001 (−0.02, 0.02)	.906
Sex: female (reference: male)	−0.21 (−0.41, −0.002)	.048	−0.20 (−0.41, 0.01)	.055	−0.22 (−0.43, −0.02)	.033
Smoking status (reference: never smoker)
Former smoker	0.06 (−0.13, 0.24)	.555	0.07 (−0.12, 0.26)	.445	0.03 (−0.16, 0.22)	.747
Current smoker	0.64 (0.35, 0.94)	<.001	0.67 (0.37, 0.97)	<.001	0.60 (0.30, 0.90)	<.001
Having stress: yes (reference: no)	−0.05 (−0.19, 0.08)	.441	−0.04 (−0.19, 0.10)	.535	−0.05 (−0.20, 0.09)	.446
Number of oral bacteria (Logarithmic scale)
Porphyromonas gingivalis	0.02 (−0.004, 0.05)	.092			0.02 (−0.01, 0.05)	.142
Tannerella forsythia					−0.002 (−0.06, 0.06)	.957
Fusobacterium nucleatum subsp. nucleatum					0.00 (−0.09, 0.09)	.994
Streptococcus constellatus					0.01 (−0.02, 0.05)	.473
Capnocytophaga gingivalis					−0.03 (−0.07, 0.00)	.052
Capnocytophaga ochracea					−0.02 (−0.05, 0.01)	.277
Red complex*			-0.006 (0.06, 0.04)	.798
RA group (reference: HC group)	-0.04 (-0.189, 0.10)	.540	-0.04 (0.18, 0.10)	.588	-0.05 (-0.19, 0.09)	.474

CAL, clinical attachment level; CI, confidence interval; BMI, body mass index; RA, rheumatoid arthritis; HC, healthy control.

^⁎Red complex includes Porphyromonas gingivalis, Tannerella forsythia, and Treponema denticola.

Discussion

Our results showed that F. nucleatum subsp. animalis and V. parvula are associated with RA, independently of other RA-related factors, indicating that these species may be related to RA development and aggravation. We did not find a direct association between P. gingivalis and RA but our results imply that it may aggravate periodontal status, which may eventually lead to RA.

F. nucleatum has been implicated in RA in a previous study.³¹ A recent study showed that F. nucleatum was enriched in the gut of RA patients, as compared to HCs, with its abundance positively correlated with RA disease activity, and that F. nucleatum aggravated arthritis.³² In another study, a triple oral inoculation of pathobionts (P. gingivalis, F. nucleatum, and A. actinomycetemcomintans) combined with collagen induced arthritis in mice, and oral inoculation with either F. nucleatum or A. actinomycetemcomintans alone accelerated arthritis onset and progression.³³ Comparison of salivary oral flora among those with early RA, suspicion of RA, and HCs showed that the first 2 groups had more V. parvula than the latter group.³⁴ In a study of patients with Sjögren’s syndrome complicated with collagen diseases, such as RA or systemic lupus erythematosus, the patients had higher rates of V. parvula than HCs.³⁵ In another study, mice with antigen-induced arthritis had elevated levels of V. parvula.³⁶ In our study, RA patients had a greater abundance of these species than did HCs. Because oral bacteria, including periodontal bacteria, may affect not only oral disease but also the development or aggravation of RA, appropriate oral self-care and professional care can help prevent the onset and aggravation of RA. Encouraging RA patients or suspected RA patients to visit dental clinics may prevent the onset and aggravation of RA. However, further research is needed.

In this study, C. gracilis, which is present in the gingival sulcus and has been associated with periodontal disease,³⁷ was found more often in HCs than in RA patients. In a study examining the association between the genus Campylobacter and periodontitis, C. gracilis was detected in supragingival pockets; however, the proportion of individuals harbouring colonies of C. gracilis was higher in healthy individuals and those with gingivitis than in those with early or chronic periodontitis.³⁷ A previous study found worse periodontal status of RA patients compared to HCs.¹³ Similarly, in the present study, RA patients had a higher mean PD than HCs. Therefore, RA patients with poor periodontal status may have a lower abundance of C. gracilis compared to HCs.

P. gingivalis is the only oral bacteria known to produce peptidylarginine deiminase (PAD); the bacterial PAD (PPAD) produced by P. gingivalis citrullinates periodontal tissue proteins and triggers an autoimmune response in RA. Therefore, it has been suggested that patients with periodontitis may develop RA via P. gingivalis.¹⁷ However, in our study, there was no significant difference in P. gingivalis counts between RA patients and HCs. Notably, when the association between mean PD and oral bacteria was examined separately in RA patients and in HCs, stronger correlations between P. gingivalis and red complex abundances, and the mean PD and the mean CAL, were seen in RA patients than in HCs.¹⁶ In addition, multivariable linear regression analyses with the mean PD as the dependent variable showed that participants with more P. gingivalis and red complex had deeper mean PD values. These results imply that P. gingivalis may be involved in the formation of periodontal pockets and the aggravation of periodontitis. Previous studies have shown that the development and aggravation of RA occur more often in individuals with periodontitis than in those without periodontitis.^12,38 Considering that a correlation between mean PD and RA was also observed in the present study, it is possible that P. gingivalis influences the onset of RA indirectly. However, the results of multivariable linear regression analysis, which included all oral bacteria associated with mean PD as independent variables, showed no significant association between P. gingivalis and mean PD, indicating that P. gingivalis alone was not related to periodontitis aggravation or RA onset. Although no direct association between P. gingivalis and RA was observed in this study, other studies have also questioned the role of PPAD in the association between P. gingivalis and RA.^19,20 Further studies are needed to clarify the association. However, reducing inflammation in the periodontal tissues of patients with, or suspected of having, RA may prevent the onset and aggravation of RA.

In this study, when comparing the RA and HC groups, PD was significantly higher in RA compared to HCs, while no significant difference was observed in CAL between the 2 groups. A previous study has shown that RA patients have worse periodontal conditions than HCs.¹¹ A significant difference in CAL between the RA and HC groups might not have been observed because many HCs may have had a history of periodontitis but were currently in good periodontal condition. Stepwise logistic regression analysis demonstrated that individuals with higher PD had greater ORs for RA, independently of other factors. Therefore, it is essential to provide oral hygiene instructions to RA patients to suppress gingival inflammation.

RA is an autoimmune disease, but genetic factors are involved, and a large cohort study found that close relatives of RA patients had approximately 3 times the risk of developing RA compared to those without such relatives.³⁹ In our study, RA patients were more likely to have a family history of RA than were HCs. Periodontitis and peri-implantitis also involve genetic and epigenetic factors.⁴⁰ Therefore, individuals who are genetically predisposed to periodontitis may be more susceptible to developing RA. In addition, they were more likely to be smokers than HCs. Smoking is an independent risk factor for developing RA.⁴¹ Tobacco is thought to mediate airway protein citrullination by the PAD enzyme in the lungs, and smokers with both ACPA and human leukocyte antigen (HLA)-DRB have a more than 20-fold increased risk of developing RA.⁴² In addition, smokers have a poorer periodontal status,⁴³ and people with periodontitis are at higher risk of developing RA.^44,45 Therefore, while smoking is of course involved in the development of RA, future studies should examine the effects of smoking on RA development, including the association between smoking and periodontitis.

This study had some limitations. In the future, longitudinal studies are needed to clarify the direction of the association between RA and oral bacteria and the effect of periodontal bacteria on the development of RA. In our study, the RA group comprised patients who were able to visit an orthopaedic clinic, thus excluding severe RA patients requiring hospitalisation. On the other hand, the HC group included patients visiting a specific department at our dental hospital. To clarify the relationship between RA and periodontal pathogens, large-scale studies including large numbers of RA patients with varying disease severity are needed. In this study, the RA and HC groups were surveyed at different locations; the oral status of the RA group was evaluated by a single examiner, whereas the HC group was evaluated by 3 examiners, including the one who evaluated the RA group. Therefore, there may have been bias due to differences in evaluators between the RA and HC groups. The questionnaire included general items commonly used in other studies, but we did not pretest the questionnaire or confirm its validity. In addition, we collected oral bacteria from the deepest parts of the periodontal pockets of participants and searched for specific bacterial species; we did not conduct an exhaustive survey of bacterial flora. The effects of many other bacterial species on RA should be investigated in the future.

Conclusion

In this study, RA patients had more F. nucleatum subsp. animalis and V. parvula than HCs. C. gracilis was found more often in HCs than in RA patients. Our findings suggest that F. nucleatum subsp. animalis and V. parvula may be related to the development and aggravation of RA. Although we did not find a direct association between P. gingivalis and RA, P. gingivalis may aggravate periodontitis and tissue inflammation, which may affect RA. However, there are still many unknowns about the relationship between oral bacteria and RA, and further research is needed.

Funding

This study was supported by the Japan Society for the Promotion of Science (JSPS) KAKENHI (grant number: 22K17295) and by the department budget of Aichi Gakuin University.

Data availability statement

The datasets generated during and / or analyzed during the current study are available from the corresponding author on reasonable request.

Author contributions

HH designed the study, performed the periodontal examinations, performed data analysis, and wrote the manuscript. TN performed the periodontal examinations and critically revised the manuscript. YT performed the periodontal examinations and critically revised the manuscript. SH contributed to the acquisition of data and critically revised the manuscript. YS designed the study and critically revised the manuscript. All authors have seen and approved the final version of the manuscript.

Conflict of interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.