Rheumatoid Arthritis and Periodontal Disease: A Rheumatologist's Perspective

Abstract

Despite advances in our understanding of the inflammatory events that underlie rheumatoid arthritis (RA), which have led to targeted therapies that more effectively control the condition, the etiology of RA is not fully understood. With the discovery that serum antibodies to citrullinated peptides (ACPA) are highly specific for RA and that Porphyromonas gingivalis, the major pathogen responsible for periodontitis (PD), contains the enzyme responsible for the citrullination of peptides, a plausible explanation for observations of increased incidence and severity of PD in RA patients and an appreciation of pathogenic similarities between the two conditions has emerged. Studies of the effect of RA treatment on the severity of PD have been limited and conflicting, especially with respect to anti-TNF agents, but indicate the potential for IL-6 as a therapeutic target for both conditions. PD treatment appears to improve clinical and laboratory evidence of RA disease activity, and the response of RA to anti-TNF therapy is abrogated by the presence of PD. Thus, evaluation and treatment of PD can be recommended for all RA patients.

Introduction

Rheumatoid arthritis (RA) and periodontal disease (PD) are common chronic inflammatory diseases, with pathophysiological similarities that have been appreciated for over 30 years [1]. In the past decade, major advances in our understanding of common risk factors and underlying biochemical and immunological processes have supported a direct, possibly causal association between the two conditions. To understand the potential connection between the two conditions, some basic knowledge of RA is necessary.

RA is a systemic autoimmune disease characterized by chronic joint inflammation, progressive destruction of articular cartilage, and bone erosion, ultimately leading to loss of joint integrity. Although RA manifests primarily as articular disease, periarticular manifestations such as bursitis, tenosynovitis, and carpal tunnel syndrome may also be seen. Beyond joint involvement, RA can cause constitutional symptoms such as malaise, fatigue, anemia, and thrombocytosis, all of which represent systemic manifestations of the condition. In addition, it has effects on multiple organ systems (e.g., keratoconjunctivitis sicca, xerostomia [Sjögren’s syndrome], lymphadenopathy, pleuropericarditis, splenomegaly, peripheral neuropathy, and vasculitis) [2].

RA is the most common inflammatory arthritis, affecting approximately 1 % of the adult population worldwide. Its peak onset is usually between the ages of 30 and 50 years, but it can occur at virtually any age, including in childhood. The prevalence in the United States is estimated at 2.1 million (25 new cases per 100,000 persons for men and 54 per 100,000 persons for women annually), with women two to three times more commonly affected than men [2].

RA is responsible for a high level of functional impairment and work disability, and it contributes to a reduction in quality of life due to factors such as pain, fatigue, depression, anxiety, separation and divorce, job change, and dependence on others [3]. The life expectancy of patients with RA may be reduced by as much as 5 to 15 years [4, 5]. Cardiovascular disease is generally considered to be the leading cause of mortality in patients with RA, accounting for about half of all deaths [5].

RA behaves differently in different individuals, and should really be considered a syndrome with common manifestations. The final common pathway includes persistent synovial inflammation and associated cartilage and bone damage, but different disease subsets can be separated by clinical phenotype (erosion or extra-articular manifestations), particularly by the presence or absence of “specific” autoantibodies (rheumatoid factor [RF] and antibodies to citrullinated peptides [ACPA]) [2].

Who Gets Rheumatoid Arthritis, and Why?

Despite impressive strides in our understanding of the inflammatory events that underlie RA, and the therapeutic interventions that much more effectively control the condition, the etiology of RA is not fully understood. As has been contemplated for most autoimmune diseases, both environmental influences and genetic factors contribute to the development of the condition. More than 100 genetic loci have been linked with RA susceptibility [6]. The most important genetic risk, seen in approximately 80 % of white patients with RA, is the conserved genetic sequence (“shared epitope”) of the amino acid sequence at positions 70–74 of the DRβ1 chain, located in the HLA-D region of chromosome 6 [7], although this is also found in 20 % of the general population [8]. Genome-wide association studies, however, can only explain about one-quarter of the disease variance [9]. Studies in monozygotic twins revealed a concordance rate of only up to 15–30 %, suggesting that environmental or perhaps epigenetic factors, or both, are required for the development of clinically evident RA [9].

Some ethnic groups, particularly some Native American populations (Pima, Tlingit, Yakima, and Chippewa) [10–12], have a strikingly high prevalence of the disease, leading to the speculation that RA is a disease of the New World that was “transmitted” to European and other populations after the period of European colonization [13, 14]. This conclusion is supported by paleopathology demonstrating evidence of RA in North American native groups dating back thousands of years, but only in Europe since the 1780s. Weissmann has suggested that this corresponds with the availability of sugar and molasses from the West Indies and the resultant higher incidence of dental caries and PD among the European bourgeoisie [14]. Interestingly, the North American native populations with a high prevalence of RA have a similarly high prevalence of the associated “shared epitope” [15], and have been reported to have a particularly high incidence of PD [16, 17].

Rheumatoid Arthritis and Periodontal Disease: Rheumatoid Factor and ACPA, via P. Gingivalis

The concept that intraoral infection is implicated in RA pathogenesis is not original. In the early days of the 20th century, the “oral sepsis” hypothesis was proposed, implicating apical infections of the teeth as the cause of RA [18]. This led to the use of dental extraction as a therapeutic option, which continued for at least four decades until it was eventually deemed ineffective.

In the 1980s, the concept that the inflammatory mechanisms underlying RA and PD shared common features was first introduced by Snyderman and McCarty [1], noting that both conditions involved self-sustaining inflammation in a fluid-filled compartment, in which the inflammatory cells and other mediators eventually lead to erosion of adjacent bone. Studies have identified inflammatory mediators, including tumor necrosis factor (TNF)-α, IL-1, IL-6, prostaglandin E2, reactive oxygen species, and matrix metalloproteinases, that can lead to the bone destruction shared by the two diseases [14, 19]. The similarities between the two conditions were further emphasized by Greenwald and Kirkwood, who suggested that proof-of-concept studies could be performed in patients with PD in order to validate therapeutic approaches prior to embarking on clinical trials in patients with RA [20].

In has long been known that patients with active RA demonstrate significantly increased frequency and severity of periodontal diseases as well as tooth and alveolar bone loss, with several recent studies substantiating the presence of PD in patients with RA early in the course of disease [21–23, 24••]. The early development of RA argues for either a causal relationship between the two conditions or a shared etiopathogenesis, or both. A recent study linked the presence of alveolar bone loss with diagnosis of RA compared with osteoarthritis controls, and the magnitude of alveolar bone loss with RA disease activity, including autoantibody levels [25••].

RF is the hallmark autoantibody linked with RA. These antibodies, typically IgM or IgA targeting the Fc portion of IgG, have long been identified as predictors of more severe disease and extra-articular manifestations, but their pathogenetic significance is still not completely understood [9, 11]. In addition to their presence in patients with RA, RFs have been detected in the gingiva, subgingival plaque, saliva, and serum of patients with PD [26, 27], and are thought, through complement activation, to help modulate the immune response in periodontal lesions [28].

The antibody system that may directly link the shared pathogenesis of RA and PD is the ACPA. ACPAs are immunoglobulins that are highly specific for the diagnosis of RA [8, 29]. Although the direct role of ACPA in RA is not entirely understood, it has become clear that post-translational changes to proteins occur in sites of tissue injury or inflammation [30, 31]. These post-translational modifications, which include citrullination, alter the biochemical and antigenic nature of the peptide structure, exposing neo-epitopes, and result in the generation of autoantibodies to these altered peptides [32]. There are at least five isotypes of the mammalian enzyme (peptidylarginine deiminase [PAD]) capable of inducing citrullination of mammalian proteins (PAD-1, 2, 3, 4, and 6) [33]. In humans, PAD-2 and PAD-4 are associated with the production of citrullinated proteins in RA [34]. In genetically predisposed individuals—possibly due to binding of citrullinated peptides to the shared epitope—generation of ACPAs is directed toward particular citrullinated proteins (e.g., α-enolase, vimentin, fibrinogen, and type II collagen). Studies have identified ACPA levels as specific serological markers for RA that correlate with disease severity as well as the development of erosive and extra-articular disease [35–37]. In some RA patients, ACPA can be detected in the circulation many years prior to the development of clinically manifest disease (i.e., preclinical RA) [38, 39]. During this phase, the synovial architecture shows little or no indication of inflammation or activation of immune cells [40–42].

In 2004, after the clinical utility of ACPA in the diagnosis and prognosis of RA and the elucidation of the enzymatic apparatus of P. gingivalis had been established [43], we proposed that it was actually the humoral immune response to P. gingivalis that provided a stimulus for the development of RA [44]. We posited that periodontitis, and exposure to P. gingivalis and its own PAD (PPAD), could break immune tolerance and trigger a latent antibody response against citrullinated proteins prior to the onset of RA [44]. PPAD is unique to P. gingivalis and shares sequence homology with other bacterial arginine deiminases, and is not evolutionarily related to the mammalian family of PAD enzymes [43].

P. gingivalis has been identified in subgingival biofilms of patients with RA. With the use of high-throughput bacterial DNA sequencing, Scher et al. showed that about 55 % of new-onset untreated RA patients carried P. gingivalis, twice as many as seen among healthy controls [24••]. Other groups have also reported a trend toward increased incidence of P. gingivalis in patients with RA [45••], particularly early in the disease course [46••, 47••].

In addition to PPAD, recent studies have confirmed the presence of PAD-2 and PAD-4 in inflamed periodontal tissues [48, 49]. Wegner et al. demonstrated that through PPAD, P. gingivalis could induce citrullination in both bacterial and host peptides, such as fibrinogen and α-enolase. The ability to generate citrullinated peptides by both host and bacterial PAD in periodontal tissues provides strong evidence of a role for periodontal citrullination in the pathogenesis of RA [50] P. gingivalis has also been shown to have an enolase, which when auto-citrullinated results in antibodies that are able to cross-react with antibodies to α-enolase found in RA patients [51]. Antibodies to P. gingivalis have been shown to correlate with the presence of ACPAs in patients with RA and their family members [52]. Even in the absence of RA, ACPAs have been reported in the serum of individuals suffering from PD [35, 53]. Thus, there is considerable potential for significant citrullination to occur within inflamed periodontal tissues. The generation of ACPA directed against periodontal antigens may thus prime susceptible individuals to a heightened ACPA reaction, often directed toward additional targets, as a consequence of epitope spreading prior to RA disease onset [37, 54, 55, 56••].

The ACPA link between PD and RA appears to be a two-way street, in that the predilection for ACPA formation in patients with RA seems to be linked to the degree of alveolar bone loss. The University of Nebraska group recently published studies assessing the incidence of PD among patients with RA and correlation with RA disease activity and serologies. The degree of alveolar bone loss was associated with ACPA levels, clinical and radiographic measures of RA disease activity, and level of disability. Alveolar bone loss was associated with the presence of ACPAs targeting citrullinated vimentin and histone. The levels of specific ACPAs were associated with the presence of subgingival P. gingivalis and with higher levels of anti-P. gingivalis antibodies [25••, 46••].

The centrality of citrullination in both PD and RA has been demonstrated by recent animal studies. Preexistent periodontitis was shown to exacerbate experimental arthritis in a murine model [57••]. Infection with viable P. gingivalis exacerbated collagen-induced arthritis, and was linked with the presence of PPAD [58••]. In another murine model of both PD and RA, PPAD-deficient strains of P. gingivalis were associated with significantly reduced periodontal inflammation and lowered ACPA levels, in addition to a reduction in joint swelling and radiographic damage [59••].

Effects of Periodontitis Treatment on Rheumatoid Arthritis Outcomes

Periodontal care involving scaling and root planing to eliminate dental plaque and calculus formation has been shown not only to improve local gingival and periodontal inflammation, but to reduce systemic mediators of inflammation that might underlie the association of PD with atherosclerosis and other systemic conditions [60]. Furthermore, if PD is a contributor to the underlying systemic inflammatory process in RA, the expectation would be that periodontal care should help ameliorate RA disease activity.

Several studies have now demonstrated that nonsurgical periodontal treatment is able to reduce serum TNF-α, IL-6, and C-reactive protein levels, and to moderate RA disease activity, in patients with moderate to severe PD [60–62]. Ortiz et al. evaluated disease outcomes in 40 RA patients, half of whom were treated with conventional disease-modifying antirheumatic drugs (DMARDs), and half who received DMARDs and anti-TNF treatment. Patients in each group were randomized to either scaling and root planing or no PD treatment. Those who received PD treatment showed a significant decrease in clinical measures of RA disease activity and laboratory abnormalities compared with those without PD therapy. Interestingly, those who received anti-TNF therapy showed improvement in clinical PD activity [61]. Similar findings indicating improved RA outcomes in patients receiving nonsurgical PD therapy have been demonstrated in studies from other geographic regions [62–64, 65••, 66, 67]. Another study, in RA patients who were about to commence biologic treatment, found that good RA outcomes were only achieved in patients without established PD, suggesting that the presence of PD could interfere with the success of anti-TNF-therapy [65••].

Bartold’s group performed a meta-analysis of five studies [6, 61, 62, 64, 67] and similarly concluded that nonsurgical periodontal treatment could lead to improvement in markers of disease activity in RA, noting that the studies were of short duration (<6 months), included few subjects, and did not sufficiently assess potentially confounding comorbidities [68••].

Effects of Rheumatoid Arthritis Treatment on Periodontitis Outcomes

Pharmacological treatment of RA involves the use of DMARDs. Methotrexate is the drug most often utilized, and is often characterized as the foundation upon which other therapeutic decisions are made. The history of DMARD treatment of RA, however, has witnessed an evolution, from the empirical use of small molecules, often thought to treat possible infectious agents (gold salts, antimalarials [hydroxychloroquine], and sulfasalazine), to broad-spectrum immunomodulatory medications (methotrexate, cyclosporine, and leflunomide), to biologic therapies (monoclonal antibodies, soluble receptors, fusion proteins) that target specific proinflammatory cytokines (TNF-α, IL-1, IL-6) or cell surface molecules, which either define cell populations that can be depleted (i.e., CD20 on B lymphocytes) or are required for T cell co-stimulation and activation. The use of several available biologic agents, either alone or in conjunction with methotrexate, has dramatically reduced the rate of joint destruction that is linked with disability as well as mortality in RA. Since RA and PD share several common pathogenic mechanisms, and since treatment of PD may exert a beneficial effect upon RA (see above), one might expect that treatments that are effective in RA would also improve PD.

There is substantial evidence for the utility of the prostaglandin inhibitors, non-steroidal anti-inflammatory drugs (NSAIDs) in the treatment of both conditions. However, NSAIDs have not been determined to have disease-modifying effects in the treatment of RA, and will not be further reviewed here [69].

Of medications that do have clinical evidence for disease modification in RA, the semi-synthetic tetracyclines (minocycline and doxycycline) presumably work through both anti-inflammatory pathways and inhibition of collagenase and other matrix metalloproteinases, as has been demonstrated for subantimicrobial concentrations of tetracyclines in treatment of PD [70]. Despite substantial laboratory evidence of their effectiveness as inhibitors and successful, albeit small, clinical trials [71], these medications have not been perceived to have a major clinical benefit and are not commonly used in the treatment of RA [72].

With the exception of rare case reports, information on the effects of other small-molecule DMARDs on PD is limited to a few animal studies evaluating the effects of gold and methotrexate. In a ligature-induced experimental model of PD in squirrel monkeys, maintaining therapeutic levels of gold salts (gold sodium thiomalate, 5 mg/kg every 4 days for 12 days) resulted in smaller areas of infiltrated supracrestal connective tissue, less loss of connective tissue attachment, and less coronal alveolar bone loss compared with placebo [73]. This disease-modifying effect appeared to be related to a selective decrease in infiltrating polymorphonuclear (PMN) leukocytes, presumably due to gold’s inhibitory effect on PMN chemotaxis [74]. In a similar model of periodontitis in Wistar rats, intraperitoneal administration of methotrexate at doses of 0.1 mg/kg, 0.5 mg/kg (doses in a range approximating those used to treat RA), and 1 mg/kg (a dose sufficient to cause neutropenia) for 28 days was no more effective than placebo in preventing alveolar bone loss [75].

The TNF-α inhibitors were the first biologics introduced for the treatment of RA, and remain the first-line and most commonly prescribed biologics for this disorder. There have been no placebo-controlled trials of TNF inhibitors for PD, and disappointingly, the few clinical trials of this class of biologics reported to date have yielded conflicting and—at best—only suggestive evidence of a possible beneficial effect on PD. In an observational study of RA patients treated with infliximab (a chimeric anti-TNF monoclonal antibody), anti-TNF-α-naïve RA patients, and healthy controls (10 in each group), infliximab-treated patients had significantly less attachment loss (p = 0.027) than TNF-α-naïve and healthy controls and, as might have been predicted, lower levels of TNF-α in gingival crevicular fluid. [76]. In a cross-sectional study in which 20 RA patients who had received a mean 40 months of therapy with infliximab were compared with 20 infliximab-naïve subjects, treatment with infliximab was associated with a significant increase in gingival inflammation as assessed by modified gingival score and papillary bleeding index, but had no effect on the severity of PD as judged by attachment loss and probing depth. From the group of 20 infliximab-naïve patients, nine patients with PD were then treated with infliximab for nine months and reevaluated. While an increase in gingival inflammation was seen once again, confirming observations from the cross-sectional part of the study, the clinical attachment loss was significantly (p < 0.05) reduced after treatment with infliximab, while probing depth remained constant. [77]. Quite opposite effects were recently reported for adalimumab (a fully humanized anti-TNF monoclonal antibody). Among 20 Japanese RA patients treated for three months, a significant improvement in measures of gingivitis (gingival index, bleeding on probing) was seen. While there was a significant reduction in probing depth, however, no effect upon clinical attachment level was observed [78].

An impressive effect of IL-6 inhibition on PD was evaluated in a unique eight-week clinical trial in which 28 RA patients who had completed 20 months of therapy with tocilizumab, and 27 RA biologic DMARD-naïve “control” patients, were evaluated at baseline for clinical (gingival index, bleeding on probing, probing depth, clinical attachment level) and biomarker PD parameters (serum IL-6, MMP-3 levels). RA disease activity was also evaluated by a composite clinical and laboratory assessment. These clinical and biomarker studies were repeated eight weeks later while patients continued their treatment regimens. Thus this study, unlike the anti-TNF trials, approximated a placebo-controlled trial. Patients treated with tocilizumab showed a significantly greater decrease in gingival index (p < 0.01), bleeding on probing (p < 0.003), probing depth (p < 0.0001), and clinical attachment level (p < 0.005) compared to the control group. This robust effect on all aspects of PD clinical activity occurred while only a trend toward improvement in RA disease activity (p < 0.06) was observed [79••].

Modulation of T cell activation through inhibition of the B7-CD28 costimulatory pathway has provided another therapeutic target in RA with the use of abatacept, a CTLA4-IgG Fc fragment fusion protein that inhibits interaction between B7 (CD80/86) expressed on antigen-presenting cells with CD28 expressed on CD4 helper T cells. A requirement for B7 costimulation was demonstrated in a cell transfer model of antigen and LPS-induced PD in rodents. The inhibition of alveolar bone loss in this model with either local or systemic administration of CTLA4-Ig suggested that abatacept treatment of RA could also ameliorate PD [80]. To date, no human clinical trials have been conducted. The effect of B cell depletion with rituximab has not yet been studied in PD.

In conclusion, there are no data in humans to assess the effects of non-biologic DMARDs on PD. Although there is more information from a limited number of studies of biologic DMARDS, the number of patients studied has been small, and there have been no randomized double-blind placebo-controlled trials. The use of anti-TNF antibodies has not consistently prevented alveolar bone loss, and has either aggravated or improved gingival inflammation, depending upon which drug was used. Whether these divergent results reflect differences in structure of the anti-TNFs studied (infliximab is chimeric and adalimumab is fully humanized) is not clear. To date, the most impressive beneficial effect was seen with IL-6 inhibition reported with tocilizumab, an IL-6 receptor antibody. Further studies with tocilizumab, as well as observations with the use of anti-IL-6 antibodies, may further elucidate the importance of this particular cytokine in the pathogenesis of PD.

Periodontal Disease is not Sine qua non for Development of Rheumatoid Arthritis

One can make a compelling argument for PD as a proximate contributing factor in the breakdown of tolerance to citrullinated proteins and the ultimate development of RA. But PD is neither necessary nor sufficient for the development of RA. We have noted that the preponderance of data connecting PD and RA is reserved for those patients that are seropositive (ACPA-positive) and likely possess the shared epitope, the major genetic predisposition for the development of RA. These connections may not apply to patients who are seronegative. Furthermore, many of the studies that have supported this clinical association are fraught with potential inaccuracies—small sample size, lack of uniformity in the definitions of PD and RA (patients asked to self-identify their underlying condition, or no measurement of the associated serologic abnormalities [RF or ACPA] to substantiate the diagnosis), and clinical settings that likely represent a disparate population (e.g., Veterans Administration Medical Centers likely to have an abundance of male patients) [81].

Perhaps other oral organisms serve as antigenic triggers and contribute to the development of RA, either independently or in conjunction with P. gingivalis. Scher et al. identified additional oral pathogens that correlate with either the development of RA-associated autoantibodies (Anaeroglobus geminatus) or early RA (i.e., Prevotella, Leptotrichia, and Tannerella species) [24••]. Using a murine model of PD and RA, de Aquino et al. were able to generate subgingival and synovial inflammation, requiring the presence of both P. gingivalis and Prevotella nigrescens [82••].

Other mucosal sites exposed to high loads of bacterial antigens, such as lung and gut, may represent alternative sites of generation of immune stimulation and autoimmune reactants [83••]. The lung is particularly suspect, as recent epidemiological data, particularly in the context of genetic predisposition (shared epitope) and ACPA seropositivity, support a connection with tobacco smoking [84]. It is well known that smoking is a risk factor for periodontal inflammation and destruction [85], but smoking also turns out to be the best-established environmental risk factor for development of RA. Smoking raises the risk of seropositivity not only in RA patients, but also among smokers with aggressive PD without clinical evidence of RA [86, 87••]. Population-based case-control studies [88] have shown that smokers have twice the risk of developing seropositive RA compared with never-smokers, and the risk is dose-dependent on lifetime exposure to smoking [89, 90]. Bronchiectasis has also been identified as a risk factor for RA, with a tenfold higher prevalence compared to osteoarthritic patients [91] and an association with elevated levels of ACPA [92••]. In studies utilizing high-resolution computed tomography of the lungs, ACPA positivity in the absence of clinical arthritis has been associated with greater airway abnormalities than seen with seronegative individuals [93]. In a manner analogous to Weissmann’s hypothesis regarding the role of sugar and the risk of RA [14], Aceves-Avila suggested that the availability of tobacco may have altered the predisposition to RA among European populations [94].

Alterations in intestinal microbiome homeostasis have also been linked with the development of RA [95]. Some studies have demonstrated an association with a reduction in fecal bacterial counts or a greater diversity of Lactobacillus varieties in patients with early RA [96]. Scher’s group documented an abundance of Prevotella copri among RA patients, identifying the organism in 75 % of patients with new-onset RA, compared to 13 % of the non-affected population [24••].

Rheumatoid Arthritis: The Elusive Second Hit

The relationship between PD and RA has been hypothesized to represent a "two-hit" model [19, 97, 98]. Whether one accepts the breakdown of tolerance due to citrullination in the periodontium [44], or identifies the lungs as the source of ACPA or the gut as a site of immune dysregulation representing the “first hit,” there remains tremendous uncertainty as to how disease manifestations materialize in the joints. The second joint “insult,” whether infectious or biomechanical or microvascular, has yet to be identified. It has been proposed that oral pathogens can directly inoculate the joints, with reports of detection of DNA from P. gingivalis isolated from rheumatoid synovium [99]. The relevance of this remains unclear, without identifying histological or microbiological evidence of an active anaerobic septic arthritis. Local production of ACPA directed against type II collagen, fibrinogen, α-enolase, or vimentin may occur as citrullinated neo-epitopes of these proteins have been detected in the joints of animal models of RA and in RA patients [100, 101••, 102]. ACPAs may also be first generated or propagated by neutrophil cell death or NETosis, which results in post-translational citrullination of histones and the extracellular release of chromatin, and occurs at the site of tissue injury and inflammation [103••, 104]. Yet the presence of ACPA cannot simply be a nonspecific result of synovial inflammation, as citrullinated peptides and ACPAs seem to be absent from synovial tissue in inflammatory and non-inflammatory joint diseases other than RA [32].

Similarly, other autoantibody systems have been implicated in the development of RA. Antibodies that recognize carbamylated proteins (anti-CarPA) that are generated against peptides that have been homocitrullinated (post-translational modification of lysine residues) have been identified in RA patients [105, 106]. In an animal model, the presence of anti-CarPA increased susceptibility to synovitis in the presence of citrullinated peptides [107]. This mechanism is particularly attractive in that carbamylated proteins have been identified in inflamed gingival tissue, and may further link the presence of PD with the development of RA [108].

Other infectious agents and their products (heat shock proteins) have been found to be associated with the development of RA, with molecular mimicry often the postulated connection [2, 9]. In particular, infection with an arthritogenic strain of virus (e.g., parvovirus B19) [109] or accumulation of Epstein--Barr virus-activated immune cells in synovium may serve as the stimulus for synovial inflammation [110].

Rheumatoid Arthritis and Periodontal Disease: Practical Recommendations

With the availability of biologic DMARDs and the potential to achieve more profound disease control, a more aggressive approach to managing RA has been adopted (i.e., “treat to target”) [111]. If we are to accomplish that goal, we cannot lose sight of the non-pharmacologic measures that might favorably impact disease control [112].

• Fatty acid supplementation: The anti-inflammatory potential of diets rich in select n-3 and n-6 polyunsaturated fatty acids, which can alter eicosanoid production, has been demonstrated in vitro and in many disease states. Supplementation with borage seed oil as a source of gamma-linolenic acid, and fish oil for eicosapentaenoic acid and docosahexaenoic acid, has been shown to improve outcomes in RA patients by decreasing the need for TNF antagonist use, improving the response to combination DMARD therapy, and achieving a higher rate of clinical remission [113, 114]. Benefits have also been shown in the treatment of PD [115, 116].

• Smoking cessation: Although the connection between smoking and RA is well-substantiated [117], the beneficial effects of smoking cessation are less clear [118]. However, a recent study in women demonstrated that the risk of RA decreases over time after smoking cessation [119], and smoking cessation seems to improve outcomes in PD as well [120].

• Weight loss: Likewise, studies have supported a connection between overweight/obesity status and RA disease activity, disability, and the likelihood of achieving good disease control [121, 122]. Although weight loss has yet to be shown to improve RA disease activity, in light of the excess cardiovascular morbidity seen in RA patients, weight loss regimens should be instituted in order to reduce modifiable risk factors. Similarly, PD has been associated with cardiovascular morbidities, and obesity has also been linked with development of PD [123].

• Evaluation and treatment of periodontal disease: As summarized above, Bartold’s group concluded that nonsurgical periodontal treatment can improve RA disease activity. Furthermore, the response of RA to anti-TNF therapy can be abrogated by the presence of PD [65••, 68••]. Early referral for dental evaluation and treatment may improve the course of RA [47••].

Conclusions

Although there is compelling data supporting a link between the development of PD and RA, cementing the association will require further extensive investigative work. In particular, this will include longitudinal observational studies and microbiome analysis to elucidate the temporal relationship between PD and RA; manipulations of the activity of PPAD in P. gingivalis and the development of ACPA; interventional (pharmacologic or immunization) trials in PD-RA experimental animal models; and ultimately, the treatment of subclinical PD in individuals at highest risk for development of RA (e.g., monozygotic twins, ACPA-seropositive first-degree family members).

As rheumatologists, we now discuss the possibility of remission with RA patients. A successful understanding of the link between PD and RA offers the prospect of soon discussing the prevention of RA.