Treatment of severe periodontitis may improve clinical disease activity in otherwise treatment-refractory rheumatoid arthritis patients

Rheumatology key message

• RA patients with treatable periodontitis are highly prevalent among those with multiple resistance to DMARDs.

Sir, Immune responses against a large variety of citrullinated peptides are a hallmark of RA, with the periodontal tissue proposed as a key source of antigens. Dysbiosis in the oral microbiome during periodontitis is also associated with the onset of RA and auto-antibody generation; however, links between different microbial communities and distinct autoantibody fine specificities require further clarification [1, 2].

Treatment of periodontitis may induce remission in early RA [3], but did not affect disease activity in established RA in the ESPERA trial [4]. In this registered study (NCT02096120; ethical approval: KEK-BE 130/13) we investigated the hypothesis that persistent infectious triggers of RA disease activity accumulate in the periodontium of people who cannot be successfully treated with multiple anti-rheumatic therapies. We recruited patients with inadequate response to at least two conventional synthetic DMARD and at least two immune target-specific biological DMARDs. All patients had erosive RA, moderate to high disease activity and high levels of both antibodies CCP and RF. Eight out of 10 screened patients [mean DAS based on 28 joints (DAS28) 5.1 (interquartile range 4.8, 5.8)] exhibited loss of periodontal clinical attachment level ⩾5 mm at least in two separate sites, fulfilling treatment criteria. After obtaining their written informed consent, treatment comprised one-stage full mouth disinfection, oral hygiene instruction, plus empiric antibiotic treatment of either 1 week of amoxicillin (500 mg, three times a day for 7 days prescribed in seven cases) or alternatively azithromycin (500 mg once daily for 3 days given in one patient with a history of intolerance to beta-lactam antibiotics), plus metronidazole therapy (500 mg three times daily for 7 days in all eight treated patients) [5]. Azithromycin may, in addition to its antibiotic action, also convey anti-inflammatory properties. At baseline, and 3 and 6 months after treatment, periodontal inflammation was monitored by assessing clinical attachment level, probing depth and bleeding on probing rates, along with RA disease activity measured by the DAS28 [6], microbial colonization in the gingival crevicular fluid and serum antibody repertoire against specific citrullinated peptides.

Successful periodontal treatment was demonstrated by improved mean clinical attachment level (P = 0.018), and reduced mean probing depth (P = 0.051) and mean bleeding on probing (P = 0.018) at 3 months in all treated patients. Five out of eight patients exhibited an improvement in DAS28_ESR of at least 0.6 points at 3 months. We considered these individuals to be clinical responders [7]. Periodontitis was more severe at baseline, but improved more profoundly upon intervention, in DAS28 responders (Table 1). Eleven periodontopathogenic bacterial species were monitored in the subgingival biofilm, using a semiquantitative nucleic acid-based detection method (Micro-IDent^®plus11, Hain Lifescience, Nehren, Germany). Aggregatibacter actinomycetemcomitans was not detected in any sample at levels above 10⁶ bacteria. In two patients Porphyromas gingivalis was present at baseline at levels higher than 10⁶, and in both subjects levels were reduced 3 months after treatment. However, P. gingivalis colonization at month 6 returned to baseline levels. Of these two patients, one had a clinical response at month 3. This patient was treated with azithromycin. The other of these two patients was a non-DAS28 responder. Prevotella intermedia was present in high amounts at baseline in one patient. Following treatment, P. intermedia DNA levels were reduced only in this patient with maintained DAS28 response over 6 months, characterized by a course in DAS28_ESR of 5.0 at baseline to 2.7 at month 3 and 2.3 at month 6, with undetectable bacterial load observed at month 3 and maintained at month 6. Capnocytophaga, a group of periodontopathogenic species linked with RF positivity in the Third National Health and Nutrition Examination Survey (NHANES-III) cohort without RA [8], was observed significantly more frequently at baseline in patients who went on to show a DAS28 response >0.6 points, compared with non-responders. Colonization by Capnocytophaga species was reduced 3 months after intervention in three out of four DAS28 responders; the fourth patient declined dental follow-up examinations. Serum concentrations of anti-citrullinated tenascin-C (cTNC5) antibodies were significantly higher at baseline in patients without a DAS28 response than in DAS28 responders (Table 1). However, none of the tested autoantibody titres significantly changed upon periodontal therapy.

Table 1.

Baseline characteristics (non-italic) and change in secondary outcome parameters from baseline to month 3

Parameter	Unit	Non-responder (n = 3)	Responder (n = 5)
Age	years	62 (13)	61 (10)
Female	n	3	3
Body mass index	kg/m2	27 (7)	29 (5)
Disease duration	years	18 (10)	10 (6)
Previous csDMARDs	n, median	2	2
Previous bDMARDs	n, median	6	5.5
DAS28ESR*	Mean (s.d.)	5.7 (0.4)	4.8 (0.3)
Total number of teeth	Median (IQR)	20 (8, 21)	14 (8, 16)
Number of pockets >4 mm	Median (IQR)	15 (7, 17)	14 (8, 15)
Probing depth	mm, median (IQR)	3.1 (2.6, 3.4)	3.9 (3.8, 3.9)
Change in probing depth	mm, median (IQR)	1.0 (0.9–1.1)  *	1.4 (1.2–1.5), n = 4
Clinical Attachment level (CAL)**	mm, median (IQR)	−3.5 (−5, −2.2)	−6.9 (−7, −6.6)
Change in CAL	mm, median (IQR)	0.2 (0–1.1)  *	1.2 (0.9–1.4), n = 4
BOP	Sites in %, median (IQR)	27 (18, 83)	40 (35, 46)
Change in BOP	Sites in %, median (IQR)	−19 (−48, −7)	−32 (−40, −20), n = 4
O’Leary plaque index	Median (IQR)	34 (16, 52)	80 (0, 91)
A. actinomycetemcomitans	N	0	0
P. gingivalis	N	1	1
P. intermedia	N	0	1
Capnocytophaga sp.***	N	1	4
RF	IU/ml	420 (63, 777)	157 (137, 561)
CCP antibodies	IU/ml	169 (155, 183)	190 (188, 191)
rVim	OD, median (IQR)	0 (0, 0.008)	0 (0, 0)
cVim	OD, median (IQR)	0.026 (0.01, 1.344)	0.078 (0.043, 0.098)
rFib	OD, median (IQR)	0 (0, 0.038)	0 (0, 0)
cFib	OD, median (IQR)	0.286 (0.001, 0.694)	0.158 (0.109, 0.674)
REP-1	OD, median (IQR)	0.008 (0.003, 0.012)	0.004 (0.001, 0.006)
CEP-1	OD, median (IQR)	0.026 (0.01, 0.349)	0.013 (0.013, 0.452)
rTNC5	OD, median (IQR)	0.061 (0.037, 0.063)	0.038 (0.029, 0.041)
cTNC5**	OD, median (IQR)	0.085 (0.074, 0.215)	0.051 (0.037, 0.054)

*P = 0.03,

**P = 0.05,

***P = 0.014.

Statistics for binary variables were Fisher’s exact text and for quantitative data Wilcoxon’s rank sum test. bDMARD: biological DMARD; BOP: bleeding on probing; CAL: clinical attachment level; csDMARD: conventional synthetic DMARD; cVim: antibodies against citrullinated vimentin peptides; IU: international unit; IQR: interquartile range; OD: optical density; REP-1 and CEP-1: antibodies against enolase and rTNC5 and cTNC5 against tenascin-C; rFib and cFib: antibodies against fibrinogen; rVim: antibodies against arginine baring.

In summary, we identified a substantial proportion of cases with a high degree of periodontitis in this cohort of people with RA who do not respond well to several rounds of treatment with different DMARDs. Besides the known association between periodontitis and RA, this study might also imply negative effects of an intensive programme of immune-modulating RA therapy on oral health. DAS28 response in five out of these eight patients following treatment for their periodontitis appears promising, although clinical responses did not track back to specific periodontopathogens or changes in antibody titres against specific citrullinated epitopes. These data reveal that consideration of the clinical severity of periodontitis, and associated microbial communities, may be important co-factors for the design of future interventional trials.