Introduction

The concept of an association between rheumatoid arthritis (RA) and periodontitis (PD) has been discussed for over two centuries. In the 1949 Proceedings of the Royal Society of Medicine, it was suggested that dental infections may play a role in various rheumatic syndromes [1]. Following this statement, there has been a steady growth of evidence linking rheumatic disorders with periodontal inflammation and infection. Among the rheumatic diseases, RA has gained most attention due to the similarities in the pathological features between RA and PD such as being chronic inflammatory in nature and it involves dysregulation of the immune system as well as destruction of soft and hard tissues [2].

Thus far, proposed explanations for the relationship include the direct action of PD-associated bacteria, autoantigen production during the course of PD and the shared inflammatory pathway exacerbating both diseases [3,4,5,6]. However, in 2013, the European Federation of Periodontology and American Association of Periodontology workshop on PD and systemic diseases conveyed their concern on the limited sound epidemiological evidence to support the role of PD as a risk factor for RA and vice versa [7]. In response to this statement, extensive work has been carried out to further explore the plausible biological link between RA and PD. A recent review by Bartold and Lopez-Olivia in Periodontology 2000 thoroughly reviewed the available evidence published until 2017 and concluded that there was remarkable progress in evaluating the link between RA and PD and most of the studies showed a trend of supporting the association [8].

From January 2018 to August 2019, an additional 55 studies concerning RA and PD have been published (Table 1). Thirteen of these publications are review articles including two systematic reviews analysing the cohort studies and microbiological evidence. Similar to publications from the year 2012 till 2017, recent evidence of RA and PD over the past 2 years has resulted from 10 publications on population-based cohort studies and also cross-sectional studies examining the specific microbiological or molecular aspects of the association. However, there has been a shift in the trend of experimental studies conducted, moving from animal studies to human studies especially in the form of clinical trials to assess the effect of periodontal treatment on RA or the affect anti-rheumatic drugs in RA patients with PD. Therefore, following the comprehensive review by Bartold and Lopez-Olivia (2020), this article aims to summarize the most recent scientific evidence and further appraise the association between RA and PD. The studies covered in this review are summarized in Table 2 which excludes narrative reviews.

Table 1 Publications concerning link between RA and PD in 2018–2019
Full size table
Table 2 Study design and findings from studies concerning link between RA and PD in 2018–2019
Full size table

Epidemiologic Evidence on Association Between RA and PD

From the 11 published cohort studies, 4 publications were cross-sectional studies, 3 were case-control studies and 1 systematic review. Fifty percent of these studies were carried out in the Asia Pacific region. Among the cross-sectional studies, 4 out of 6 studies reported an association between RA and PD [9,10,11,12]. The two studies that did not find any association between the two diseases was a cross-sectional study carried out in Malaysia which only found 2 out of 44 RA patients recruited who presented with pocket depths deeper than 3 mm [13] and a study carried out in South Korea of 20,297 subjects which reported that RA was only associated with tooth loss in younger adults [14]. All case-control studies demonstrated a positive association between RA and PD [15,16,17].

Interestingly, in three studies where the sample size of RA patients was less than 100 and recruited from a single rheumatology centre, the prevalence of those suffering PD was either very low [13] or very high [9, 11, 18]. However, in a larger population using data from the Korea National Health and Nutrition Examination Survey, the prevalence was only approximately 28% [12] to 32% [14] compared with 60 to 70% in the former studies [9, 11].

Jung and colleagues found 1.5% of their subjects had RA, which was in agreement with previous prevalence studies [19, 20]. Using a univariate analysis, the risk for periodontal disease was shown to be 1.64 times higher in subjects with RA than those without RA, and 1.97 times higher when adjusted for sex and age, which was statistically significant [12]. This was also in agreement with another prevalence study carried out in the USA [21] and Taiwan [22]. A recent systematic review suggested an association between RA and PD by the common pro-inflammatory profiles. A further meta-analysis also showed a higher RA prevalence for subjects with PD (OR 1.97; CI 1.68–2.31; p < 0.00001) although considerable heterogeneity among studies was significant [23].

Genetic Basis of Relationship Between RA and PD

Previous studies carried out prior to 2018 have mostly reported on the genetic association between RA, PD and the highly polymorphic human leucocyte antigen (HLA) class II molecules, in particular the HLA-DRB1 phenotype [8]. In addition to this, the possible role of other RA-related genetic polymorphisms has also been investigated to analyse the association with the destruction of the periodontium as well. Previous studies have concluded that interleukin-1 (IL-1) and interleukin-10 (IL-10) gene polymorphisms may have some correlations between PD and RA [8].

In a recent case-control study by Kobayashi and colleagues, their multiple logistic regression analysis showed that the single nucleotide polymorphisms (SNP) rs2237892 of KCNQ1 gene also resulted in a significant association with the coexistence of RA and chronic PD, which suggests that individuals carrying the T allele may likely have both diseases [24]. Another case-control study by Schulz and colleagues further emphasized the association of genetic variation in TNFα and IFNγ, as well as cytokine receptor IL4R-α with RA and PD. However, their multivariate analysis showed that only the A allele of IFN-gamma appeared to be a significant marker of RA and PD comorbidities [25]. Genetic studies in relation to the association between RA and PD still warrant further investigations as most studies to date have been carried out in a relatively small sample sizes.

Plausible Mechanisms Linking RA and PD

Role of PD-Associated Bacteria

Since 2018, a further 4 animal studies have been reported investigating PD-associated bacteria in the link between PD and RA. The human studies looking at this association have however progressed from hypothesis-driven studies such as PCR studies looking mainly at the presence of Porphyromonas gingivalis to hypothesis-free studies performing sequencing.

Animal Studies

Of the four latest animal studies investigating the association between periodontal disease and RA, all have confirmed the contributing role of P. gingivalis in this association [26,27,28,29]. Inoculation of mice with P. gingivalis has been shown to increased serum levels of anti-citrullinated protein autoantibodies (ACPA) and an increasing trend is also seen in the saliva [28]. In the aetio-pathogenesis of periodontal disease, P. gingivalis fimbriae are the major extracellular components that adhere to the host cell surfaces. Using a mouse model of arthritis, Jeong et al. (2018) demonstrated that inhibiting P. gingivalis adhesion using a FimA antibody (Ab) prevented RA progression. They also showed that orally inoculated P. gingivalis may utilize dendritic cells, macrophages and neutrophils to migrate to the joints of collagen-induced arthritis mice which will then result in synovial inflammation [26]. Similarly, eradication of bacteria by using chlorhexidine prior to collagen-induced arthritis was shown to reduce the incidence of experimental arthritis in mice [29].

Human Studies

The development of a subgingival bacterial biofilm has been recognized as one of the essential factors in the initiation and progression of periodontitis. Numerous bacteria have been implicated as contributory agents in the aetio-pathogenesis of RA [3, 30]. This assumption arises following the detection of bacterial DNA and high levels of oral anaerobic bacterial antibodies of P. gingivalis within the synovial fluid of inflamed joints in RA subjects [3, 30]. There is an assumption that genetic material is carried from the periodontium to joints in free form of DNA via bloodstream [31,32,33].

P. gingivalis has been one of the most widely studied bacteria in relation to the association between PD with RA. It is the only known microorganism with the ability to produce a peptidylarginine deiminase enzyme (PPAD) that catalyses citrullination of arginine in both host and bacterial proteins [34]. This citrullinated peptide antigen has been reported to be present in the inflamed periodontium of PD subjects and thus further activates the adaptive immune response that is selective to RA [35]. Therefore, it has been proposed to play an active role in RA development [36, 37]. Since PPAD is not calcium dependent, it can also be auto-citrullinated but this finding has not been proven to occur in-vivo [8].

Aggregatibacter actinomycetemcomitans (Aa) has also been widely studied as it has been shown to be able to dysregulate peptidylarginine deiminases (PAD) in neutrophils, leading to the extracellular release of autoantigens following neutrophil apoptosis via Aa-induced hypercitrullination. This process is mediated by its major virulence factor, leukotoxin A (LtxA), which forms pores on the cell membrane of neutrophils, allowing PAD activation and citrullination of a broad range of peptides [38]. This process provides an additional biologically plausible oral microbiological link (other than P. gingivalis) between a bacteria and the promotion of autoimmunity directed against citrullinated proteins.

Prior to 2018, other oral bacteria that have been studied widely and found to be present in the synovium of RA subjects, or these subjects having reported to have increased antibody levels against these bacteria, are Prevotella intermedia, Tannarella forsythia, Prevotella melaninogenica, Filifactor alocis, Prevotella spp. and Leptotrichia spp. [8]. However, next generation sequencing studies which have been done to look for associations between the two diseases have however found no evidence for this association [8].

Since 2018, nine studies have investigated the role of oral bacteria in the association between PD and RA. One systematic review confirmed a significant influence of P. gingivalis on the pathogenesis of RA as higher frequency of P. ginigivalis DNA was found in the serum and synovial fluid of RA patients and thus supports its role in the pathogenesis of RA [39]. The presence of P. gingivalis fimA genotype II has also been reported to be more frequently detected in RA subjects with PD [40]. In a pilot study, Bender et al. (2019) also reported a potential link between glutaminyl cyclase synthesized by P. gingivalis in patients with RA based on its similarity to human glutaminyl cyclases (QC and isoQC) which play an important role in maintaining inflammatory conditions [41]. However, of the 9 studies reported between 2018 and 2019, only 3 studies found a higher prevalence for P. gingivalis in RA with PD subjects [40,41,42]. Thus the actual role of P. gingivalis in the aetio-pathogenesis of RA still needs further validation.

RA subjects have been reported to have a higher bacterial load, a more diverse microbiota and an increase in bacterial species associated with periodontal disease [43]. Among the species reported are Capnocytophaga sp., Tannerella forsythia, Desulfobulbus sp., Prevotella sp., NA sp., Bulleidia sp., Filifactor alocis and Dialister pneumosintes [44, 45]. However, the subgingival microbiome in RA subjects with active disease on anti-inflammatory therapy was also reported to be healthier and comprised of Corynebacterium matruchotii, Actinomyces, Veillonella and Streptococcus as compared to RA subjects in remission especially those who were smokers [46]. They have suggested that the potential role of microbial community types in patient stratification and personalized therapy should be further assessed in longitudinal studies. In periodontally healthy individuals, it has been reported that RA is able to enrich the oral microbiome for inflammophilic and citrulline-producing organisms such as Cryptobacterium curtum, another organism capable of producing large amounts of citrulline [47].

Despite all the studies that have reported the presence of bacteria and the role of PPAD or dysbiosis in the association between PD and RA, these associations have yet to be biologically and clinically proven. The looming question still remains as to whether the association is truly related to PPAD causing a change in the immune tolerance or whether it is due to PPAD causing an increase in virulence of P. gingivalis and subsequently an increase in the severity of the periodontal inflammation which ultimately leads to increase in RA severity [8]. Similarly, the presence of other bacteria may indeed cause the dysbiosis that ultimately leads to an increase in the periodontal inflammation present and the subsequent increase in RA severity (Fig. 1).

Fig. 1
figure 1

Possible mechanisms on how PD aggravates RA (taken from Lee et al. (2019) [97]). The role of P. gingivalis in contributing to the formation of autoantigens either directly through protein citrullination or indirectly through inflammation-mediated carbamylation (1). P. gingivalis as a keystone pathogen in causing alterations to the gut microflora and potentially leading to endotoxemia and the persistence of low-grade systemic inflammation which may exacerbate the inflammatory response within the joint (2). Extracellular release of autoantigens following neutrophil apoptosis as a result of A. actinomycetemcomitans-induced hypercitrullination (3)

Full size image

PD and Autoantigen Production

Citrullination

Citrullination is a conversion process of positively charged arginine residues to neutral citrulline residues by a family of enzymes called PAD. This process takes place in the presence of calcium as part of the normal physiological process and functioning of the immune system [48]. Apart from the involvement in physiological processes, citrullination also takes part in pathological inflammatory conditions as part of the innate response to bacterial infection and cell death mechanism. For example, PAD-4-induced citrullination is known to play an important role in altering chemokines’ function and has been actively participating in neutrophils extracellular traps (NETs) formation as part of the antibacterial mechanism. It has often been linked to chronic inflammatory disorders such as multiple sclerosis, Alzheimer’s disease, psoriasis, and RA [48, 49].

The enzymatic deimination of arginine residues to citrulline by PAD (post-translational modification) alters the tertiary structure, antigenicity, and function of proteins [50, 51]. Subsequently, this may expose the previously hidden immune epitopes and induce an autoimmune response as citrulline is not a standard amino acid of proteins [52]. In a susceptible patient, these citrullinated peptides will act as an antigenic determinant that could break the immunological tolerance and evoke an autoimmune response by binding onto the antigen presenting cells. As a result, pathogenic T and B cells will be activated, leading to the RA-specific ACPA formation [50]. These ACPA will then form immune complexes with citrullinated peptides, resulting in the production of inflammatory mediators and ultimately causing joint destruction in RA [36, 53]. ACPA has emerged as a specific serological marker for RA due to its high specificity and predictive value for RA [54].

Apart from inflamed synovial joints in RA, citrullinated proteins have also been detected in inflamed periodontal tissues of patients with PD. The presence of citrullinated proteins in periodontal tissues was first revealed by Nesse et al. [55]. Citrullinated proteins, as well as PAD-2 and PAD-4 (both these PADs are associated with citrullination in RA) were also found in inflamed periodontal tissues in PD patients without any signs of RA [56]. Based on these findings, it was postulated that PD could provide a conducive environment for citrullination and initiation of anti-CitP targeting citrullinated peptides in joints. Prior to 2018, studies looking at citrullination and ACPA in RA and PD have focused on either periodontal inflammation as the driving force for citrullination or periodontal pathogens (as explained in the previous section) as the primary driver of citrullination [8].

Since 2018, there have been 5 studies looking at citrullination and ACPA in the link between PD and RA. Microbes such as P. gingivalis, Aa, and citrulline-producing microorganisms have been reported to be drivers of citrullination in 3 studies [32, 42, 47]. Two studies reported that inflammation was the primary driver for ACPA formation [44, 57]. In their study on first-degree relatives of RA subjects, Louton et al. (2019) have demonstrated that severe PD was associated with ACPA+ve subjects as compared to mild PD which was associated more with ACPA−ve subjects [57]. Further studies need to be done in the search for the primary driving force in associating citrullination in the link between PD and RA.

Carbamylation

Carbamylation is a non-enzymatic myeloperoxidase (MPO)-dependent post-translational modification of free amino acid lysine into homocitrulline [58]. Bright et al. proposed inflamed periodontal tissues as a source of carbamylated proteins due to high production of MPO by neutrophils during the release of NETs [59]. The formation of carbamylated proteins in addition to citrullinated proteins in PD may break the immune tolerance leading to production of autoantibodies [59,60,61].

Previously, an animal study demonstrated the presence of anti-carbamylated proteins antibodies precedes signs of joint damage [62]. Since 2018, studies on RA patients with PD-verified production of carbamylated proteins in PD and the circulating level of carbamylated proteins and NETs are associated with severity of PD in RA patients that can be significantly improved with non-surgical periodontal treatment (NSPT) [63, 64]. Therefore, inflammatory-driven carbamylation of proteins may explain the plausible link between RA and PD.

Molecular Mechanisms Uniting PD and RA

The underlying molecular mechanism in the relationship between PD and RA resides in the concept that both conditions involve overproduction of pro-inflammatory mediators and dysregulation of cytokine networks [65, 66]. The two-hit model, proposed to explain the association between RA and PD, describes local inflammatory lesions in periodontal tissues triggered by subgingival microflora (1st hit) and the release of pro-inflammatory cytokines from joints (2nd hit), escalating the systemic inflammatory cascades that further amplify local inflammatory or osteoclastogenic mediators such as prostaglandins, receptor activator of nuclear factor κ-B ligand (RANKL), and matrix metalloproteinases (MMPs) in synovium or periodontal tissues [5]. The connection between RA and PD may also lie in mutual underlying inflammatory pathways that are either driven by, or results in, production of common mediators [6].

Before 2018, two systematic reviews had been published which analysed the cytokine profile in serum and gingival crevicular fluid of RA patients with PD [67, 68]. A number of inflammatory mediators or enzymes such as TNFα, interleukins (IL), RANKL, and MMPs in RA patients with PD have been studied with regard to possible molecular mechanisms linking between two diseases [68]. Subsequent to these two systematic reviews, further works including genetic studies have been conducted to explore the role of TNFα, IL-1 and IL-6 but the results are still unclear to understand the common pathological pathway between these two diseases [8].

Since 2018, 5 studies have scrutinized the molecular aspects of the association between PD and RA and specific attention has been given to the role of proteolytic enzymes and osteoclastogenic mediators such as MMPs and RANKL. TNFα and IL-1β are still the major cytokines under investigation since both have been shown to be elevated in RA and PD patients and both cytokines have pro-inflammatory effect that promotes connective tissue turnover and bone resorption [69, 70]. A recent study assessing serum TNFα level in RA patients with PD who were not on TNFα inhibitors substantiated that the presence of PD in these subjects was associated with overproduction of TNFα [71]. However, the correlation between pro-inflammatory cytokines and parameters of PD and RA is still unclear. On the other hand, the serum level of anti-inflammatory cytokine, IL-10 in RA patients with PD was described as very low. However after periodontal therapy, the level of IL-10 in systemically healthy PD patients increased as compared to RA-PD subjects [72••].

MMP-8 is a collagenase that is involved in connective tissue degradation and is frequently found in the synovial fluid of RA patients and GCF in PD patients [69, 70]. In line with previous evidence, latest investigations revealed raised levels of MMP-8 in the serum and GCF of RA patients with PD compared to systemically and periodontally healthy subjects [73••, 74]. The serum level of MMP-8 is associated with periodontal inflammation and clinical attachment loss [74]. However, in contrast to early RA patients, established RA patients with high disease activity have persistent elevated levels of MMP-8 regardless of additional biologic therapy to conventional disease modifying anti-rheumatoid drugs (DMARDs) [73••].

Recent studies have also shown that individuals with either RA or PD or both have increased serum levels of RANKL [75, 76]. This suggests that dysregulation of inflammatory cascades in patients with RA and PD leads to sustained overproduction of osteoclastogenic mediators and enzymes such as RANKL and MMPs.

An environmental study on the effect of diet on the inflammatory response in RA patients experiencing PD reported an association between increase in serum omega-3 levels due to diet or supplements and favourable outcome of RA [77]. However, it is premature to conclude that the anti-inflammatory effect of omega-3 improves disease activity of RA and PD, but it may suggest that diet could influence the inflammatory pathway in these patients. In addition, DMARDs, such as combination of methotrexate and TNF-inhibitors, have shown to affect the biochemical markers as well as periodontal inflammation [78••]. Therefore, further work to evaluate and confirm the biomarkers’ profile of RA patients with PD is needed to control confounders including the effect of diet and DMARDs.

Periodontal Therapy in RA Patients with PD

Over the past decade, numerous studies have been conducted to assess the effect of periodontal treatment on RA. Most of the studies tested NSPT in RA patients with PD and observed a clinical change in RA disease activity. NSPT has been established to be successful in reducing periodontal pocket depths and improving periodontal inflammatory status in PD patients [79]. Similar findings can be found when NSPT was performed on PD patients who concurrently suffer from RA demonstrating that RA status does not hamper the periodontal outcomes of NSPT [67, 80, 72••].

Prior to 2018, systematic reviews and interventional studies suggested that NSPT may have beneficial effects on periodontal and RA disease activities, but more thorough and larger controlled trials are needed to convincingly substantiate the effect of NSPT on improvement of RA disease activity [8]. Over the last 2 years, no systematic review has been published concerning NSPT on RA patients. There have been five interventional studies published evaluating the effect of periodontal therapy on rheumatologic disease activity parameters, but all the studies involved very low number of subjects [80,81,82, 72••] except one multicentre randomized trial nested in a cohort study in France which involved 472 subjects [83]. However, this French study only looked at oral hygiene instructions in their test group while the control group did not receive any therapy. The findings in this study indicated no significant improvement in RA disease activity. It was also noted that the oral hygiene of all subjects was good at baseline [83] in contrary to an observational study by Hashimoto et al. (2019) that suggested poor oral hygiene could explain the severity of PD in RA patients [84].

With reference to the role of NSPT in improving RA disease activity, currently available studies have produced conflicting results. This may be due to a number of factors including the issue of small sample size, discrepancies in evaluation periods, and case definitions of PD which were adopted by the studies. Only one study incorporated bleeding on probing (BOP) at sites with attachment loss as one of the criteria to define a PD case [81]. BOP is an established quantitative tool to assess gingival inflammation which can discriminate periodontal health and disease [85]. There is a marked discrepancy in mean BOP scores at baseline among the intervention studies which ranged from 12 to 97% [82, 86]. The magnitude of changes in BOP following NSPT is associated with significant improvement in DAS28 scores [87]. Hence, the difference in extent of periodontal inflammation at baseline among the studies might explain the heterogeneous RA-related outcomes of NSPT. The inflammatory burden in the periodontal tissues may reflect or affect the disease activity of RA considering inflammation plays a key role in orchestrating the patho-mechanism link between RA and PD [8]. Therefore, to observe greater magnitude of periodontal therapy impact on RA disease activity, future studies need to assess the effect of periodontal therapy in active RA patients with extensive periodontal inflammation, and aim for subjects attaining periodontal health and not delivery of periodontal treatment per se.

Effect of DMARDs in RA Patients With PD.

In view of the association between RA and PD, the effect of RA treatment on periodontal status can be reasonably anticipated. Host modulation drugs such as DMARDs have been used extensively in the management of RA and as a result these drugs have stimulated research interest in their use as adjuncts in periodontal therapy [88, 89].

RA patients tend to have higher levels of TNFα in oral fluids [68]. Consequently, previous studies have assessed the effect of TNFα inhibitor in RA patients with PD. The interventional studies have demonstrated that the potential benefit of TNFα inhibitor drugs on periodontal inflammatory status and clinical attachment levels was not independent of periodontal therapy [90, 91]. Studies on other conventional or biologic DMARDs such as IL-6 inhibitor and anti-B lymphocytes demonstrated similar effects on periodontal clinical parameters [92, 93]. DMARDs may be advantageous as an adjunct to NSPT in improving periodontal parameters in RA patients with PD [8]. Since 2018, two studies have reported findings supporting the beneficial effect of DMARDs on periodontal status. A clinical trial on RA patients revealed that a cocktail of DMARDs was an assistance in reduction of periodontal pocket depth in conjunction to NSPT as compared to DMARDs used as monotherapy [94].

Similarly, a longitudinal study examining the impact of infliximab therapy on PD-related biochemical markers in RA patients found significant improvements in serum levels of MMP-3 after therapy [95]. This improvement was seen even in subjects with the presence of concurrent PD which contradicts a previous study by Savioli et al. [96]. Nevertheless, with scant evidence and inconsistent findings, the impact of periodontal inflammation on RA treatment outcome is still equivocal. A larger controlled trial is required to verify this potential issue.

Concluding Comments

In this review we have highlighted the most recent progress in the last 20 months (January 2018 to August 2019) in the field studying the relationship between PD and RA. This field continues to be a very vibrant of investigation at the population, biochemical, cellular, genetic, microbiological and clinical levels. There is now unequivocal evidence to indicate that this relationship is very strong. However, studies to date have not been able to identify any unifying link between these two chronic inflammatory conditions. The evidence assessed in this review indicates that while there is strong evidence to support a role for underlying dysregulation of inflammatory processes and autoimmune responses in the relationship between PD and RA, the precise processes that link them remain elusive. It is clear that while the association can be bi-directional there is no evidence to suggest that either condition is causative of the other. It seems most likely that there will be subsets of individuals who suffer specifically from both conditions and it is these individuals for whom our attention should now focus.