Several studies have indicated a positive association between periodontitis and RA, sometimes referred to as a bidirectional interaction between the diseases (Cantley et al. 2011). On the other hand, most of the available population studies are small case–control studies providing limited evidence of an association between the two diseases. Some studies have demonstrated that patients with RA are more likely to acquire advanced periodontitis than individuals without RA. This has been shown for young adults (20–35 years) (Havemose-Poulsen et al. 2006) and for midlife to aged people (Käßer et al. 1997; Mercado et al. 2000; Mercado et al. 2001). Based on such findings, it has been proposed to develop systematic programs for prevention of periodontal complications in RA patients (Havemose-Poulsen et al. 2006). However, the outcomes of the many studies are seriously hampered by varying diagnostic criteria for both diseases. In one of the studies on periodontitis, cases were identified on the basis of mean clinical attachment loss being ≥4 mm, and the odds ratio for simultaneous occurrence of RA was as much as 6.09 (95 % CI 1.72–21.55), indicating a strong association although the confidence interval was wide. More extensive cross-sectional studies have provided weaker evidence of an association, the odds ratios being 1.82–1.94, and one of these had ranges of 95 % confidence indicating insignificance (Pablo et al. 2008; Demmer et al. 2011). A 20-year prospective follow-up study including 9,564 American adults defined periodontitis cases by tooth loss of four or more teeth with attachment loss or worse conditions. Baseline and incident cases of RA were defined on the basis of self-reported physician diagnosis or physical examination data corresponding to criteria 1–4 of the American Rheumatism Association 1987 criteria (Arnett et al. 1988). Incident RA was also defined on the basis of death certificate data or health-care facility discharge diagnosis of rheumatism. The adjusted odds ratios for incident RA were between 1.12 and 1.67, dependent on number of missing teeth among participants. Most odds ratios were statistically insignificant, and there was a lack of dose responsiveness (Demmer et al. 2011). A similar result was obtained in a comprehensive 12-year prospective follow-up study in American women (Arkema et al. 2010). Overall, the present data indicating an epidemiological association between periodontitis and RA are inconsistent. The varying case definitions used for both periodontitis and RA may, in part, explain the differences in results obtained. Also, patients with RA are usually receiving intensive anti-inflammatory treatment, which ameliorate periodontal disease progression. The inconsistent results may also be due to an inhomogeneous nature of the patients with the two disease categories, which may both contain patients with more than one disease.

A recent systematic review showed that seven out of ten case–control studies had found significantly more clinical tooth attachment loss in RA patients compared to controls (Kaur et al. 2013). The same review reported that five of seven studies found significantly increased tooth loss in RA patients compared to controls. When combining the results of the included studies in a meta-analysis, the weighed mean differences were significant in both clinical attachment level and tooth loss between RA patients and non-RA controls. This finding is further supported by a Dutch cross-sectional study, in which a significantly higher prevalence of severe periodontitis in RA patients (27 %) than in controls (12 %) was seen (De Smit et al. 2012). Furthermore, a case–control study compared the presence of severe periodontitis in 287 patients with RA with that in a noninflammatory arthritis control group of 330 patients with osteoarthritis, believed to be demographically similar to the RA group (Mikuls et al. 2014). Anti-cyclic citrullinated peptide antibody-positive patients were significantly more likely to have periodontitis (37 %) than the osteoarthritis controls (26.4 %). A multivariate analysis accounting for confounding factors showed that the anti-cyclic citrullinated peptide antibody-positive patients remained more likely to have periodontitis than controls, the significant odds ratio being 1.59. In this study, tooth loss in the RA patients was also more common than in the control group (Mikuls et al. 2014). Taken together, these studies strongly indicate that the periodontal status is worse in RA patients than in controls (Kaur et al. 2013; Payne et al. 2015). However, there is currently little evidence that periodontitis represents a risk factor for RA (Linden et al. 2013).

RA has several clinical and pathological characteristics in common with periodontitis. The diseases, although chronic in nature, show periodical flare-ups with increased tissue-destructive activity in some of the involved sites interposed by periods of relative quietness, and both diseases are quality of life hampering because they are associated with loss of function. Both diseases are characterized by their inflammatory nature with local degradation of collagen-rich soft and hard tissues mediated by cytokines and collagenolytic enzymes. Based on the above findings, it is likely that there is some degree of coexistence, although it is uncertain whether an association is causal or noncausal, for instance, due to shared environmental or other predisposing factors, i.e., smoking, and socioeconomic and genetic risk factors such as MHC class II HLA-DRB1 (Firatli et al. 1996; Katz et al. 1987; Marotte et al. 2006; Bonfil et al. 1999).

Both periodontitis and RA have cytokine profiles thought to be involved in the tissue-destructive inflammatory processes, including high production of TNF-α (Cantley et al. 2011). An important example is the common pathway of upregulated expression of receptor activator of nuclear factor kB ligand (RANKL) by fibroblasts and lymphocytes, essential for osteoclast formation (Crotti et al. 2003, Bartold et al. 2010a). Obviously, an exaggerated systemic inflammation induced by periodontal infection might worsen the immune-inflammatory reactions in the joints of RA patients and vice versa (Golub et al. 2006, reviewed by Payne et al. 2015).

In an attempt to identify similarities in the pathology of periodontitis and RA, hematological characteristics of patients with RA have been compared with those of aggressive periodontitis patients. Elevated levels of traditional markers of inflammation could be seen in patients with generalized aggressive periodontitis similar to patients with RA (Havemose-Poulsen et al. 2006). Other case–control studies have compared erythrocyte sedimentation rate, C-reactive protein, ACPAs (measured as autoantibodies to cyclic citrullinated peptides), rheumatoid factor, TNF-α, and interleukin (IL)-1β in RA patients with and without periodontitis, as systematically reviewed by Kaur et al. (2013). The outcome of the studies indicates that there is no good evidence for a correlation between increased levels of the majority of these factors and presence of periodontitis and RA. An exception is IL-1 level, which appears to be increased in patients with both diseases (Kaur et al. 2013). Moreover, dysregulation of immunoinflammatory responses has been found for both diseases in a number of studies (Mercado et al. 2001; Bartold et al. 2005; Havemose-Poulsen et al. 2005) including similar patterns of elevated IL-10 plasma levels in RA patients and in patients with aggressive periodontitis (Havemose-Poulsen et al. 2005). Thus, gene expression of pro- and anti-inflammatory cytokines in peripheral blood mononuclear cells might be a common denominator for the two diseases, but only few similarities between the two diseases with respect to these parameters have been found (Sørensen et al. 2009). Rheumatoid factors and ACPAs have been revealed in sera from periodontitis patients (Gargiulo et al. 1982; Thé and Ebersole 1991; Havemose-Poulsen et al. 2006), and levels of IgM- and IgA-rheumatoid factors in patients with RA were found to correlate with percentage of sites with clinical attachment loss≥2 mm, which is why these variables have been proposed as possible predictors of periodontal tissue destruction (Havemose-Poulsen et al. 2005) similarly to their use as predictors of joint erosions (Guillemin et al. 2003; Bukhari et al. 2002).

One of the most interesting aspects is the possible involvement of Porphyromonas gingivalis in the pathogenesis of RA (Rosentein et al. 2004). In some studies, the frequency of antibodies to P. gingivalis has been shown to be significantly higher in patients with RA than in controls (Mikuls et al. 2009; Okada et al. 2011), although this was not the case in another study based on a higher number of patients (Moen et al. 2003). Over recent years, there has been much speculation that P. gingivalis may play a role in generation of the citrullinated proteins, which are thought to constitute the pathogenic autoantigens in ACPA-positive patients. P. gingivalis has been shown to produce P. gingivalis peptidylarginine deiminase (PPAD), which, like human PADs, catalyzes citrullination of proteins. There is no amino acid sequence similarity between PPAD and human PADs, however, and while the bacterial enzyme targets carboxyterminal arginine residues (McGraw et al. 1999) (after cleavage of protein substrates by bacterial gingipains), human PADs efficiently deaminate internal arginine residues (Sugawara et al. 1982). Experimentally, PPAD is capable of citrullinating human fibrinogen and α-enolase, and it has been suggested that immune complexes formed between these citrullinated proteins and ACPAs play a pathogenic role in RA (Wegner et al. 2010