Introduction: Several epidemiological studies have suggested that a link between periodontal disease and rheumatoid arthritis. In cohorts, the risk of development of rheumatoid arthritis is increased in subjects with periodontal disease. The main bacterium implicated in chronic periodontal disease is Porphyromas gingivalis, a commensal microorganism of the mouth. P. gingivalis is the only microorganism with a deiminase capable of transforming arginine into citrulline, and is suspected of to play a major role in the production of anti-citrulline antibodies, the principal diagnostic marker of rheumatoid arthritis.
Hypotheses: The regular professional and individual plaque control in patients with rheumatoid arthritis (standard of care) - i.e. decreasing oral bacterial load, including P. gingivalis - reduces the activity of rheumatoid arthritis Principal objective: To assess the impact of professional periodontal scaling every 6 months combined with the individual plaque control, i.e. toothbrushing plus the use of mouthrinses on rheumatoid arthritis activity.
Methodology: The trial design is innovative and based on random selection of a sample of patients in an observational cohort of patients as described in the recent publication "Cohort Multiple Randomised Controlled Trials Design" (Ref.: BMJ 2010; 340: c1066). A randomly selected sample of eligible patients from the cohort will be asked to take part in the study. Changes in the randomised sample will be compared to those of other potentially eligible patients from the cohort. The cohort used will be the ESPOIR cohort, a cohort including 813 patients who initially presented with early undifferentiated polyarthritis that developed into rheumatoid arthritis with anti-citrulline antibodies in approximately 40 to 50% of cases.
Procedures studied: Periodontal scaling every 6 months and twice daily use of a mouthwash containing essential oils (Listerine®) plus a triclosan/copolymer toothpaste (Colgate Total®).
Sample size calculation: The principal assessment criterion is DAS variations at 1 year. The expected statistical power is 80% (the type I risk being set at 5%) and the standard deviation common to the two samples for DAS variation is evaluated at 1.5. With these hypotheses, the inclusion of 100 patients in the study group and 100 patients in the control group will produce an effect size of 0.40 (i.e. a difference in DAS between the two groups equal to 0.6).
Statistical analysis: A linear, mixed-effects model will be used to compare DAS variations between the two groups. In addition, the Mixed Model for Repeated Measurements (MMRM) based on a linear mixed model will include two other random effects, i.e.: the centre effect and the centre\*treatment interaction effect, which will be used to measure heterogeneity between the centres and heterogeneity of the effect of the procedure. Finally, propensity scores with adjustment by weighting will be used to take into account any potential imbalances between the two study arms.
