Periodontal disease and cancer: Epidemiologic studies and possible mechanisms

Abstract Epidemiologic and cancer control studies on the association of periodontal disease and cancer risk mostly suggest a positive association with overall cancer risk and certain specific types of cancer. These findings are generally consistent among cross‐sectional and longitudinal studies. In this paper, we review epidemiologic studies and current knowledge on periodontal disease and cancer, with a focus on those studies conducted in the years following the Joint European Federation of Periodontology/American Academy of Periodontology Workshop on “Periodontitis and Systemic Diseases” in November 2012. This review also explores the role of chronic inflammation as a biologically plausible mechanistic link between periodontal disease and risk of cancer. Furthermore, it highlights studies that have examined the potential importance of certain periodontal pathogens in this association.


| INTRODUC TI ON
Periodontal disease is a prototype of a locally destructive, chronic, low-grade inflammatory process, and has been linked to increased cancer risk in a variety of epidemiologic studies. 1 In addition, differences in study populations, sampling strategies, study designs, and methodologies make comparisons across studies challenging. It is worth noting, however, that newer epidemiologic studies are increasingly exploring alternative exposure measures, such as the oral microbiome, and specific groups of established periodontal pathogens, in an attempt to further understand the potential role of periodontal disease in cancer development.
This review provides a comprehensive assessment of existing epidemiologic evidence on the association between periodontal disease and cancer risk, with emphasis on research findings published periodontal disease and cancer risk are discussed, as well as important knowledge gaps which might inform future directions for research and clinical applications.

| PERI ODONTAL D IS E A S E AND TOTAL C AN CER RIS K
Few published epidemiologic studies on periodontal disease and incident total cancer exist and these mostly point to a positive association. See Tables 1 and 2 for details. One early study from 2008 included 48 375 US male health professionals who were followed for an average of 17.7 years. The authors reported a statistically significant increased risk of total cancer among those with a history of periodontal disease (hazard ratio 1.14, 95% confidence interval 1.07-1.22) and individual cancer sites of the lung, kidney, pancreas, and hematological system. 1 However, the positive associations remained only for overall cancer risk and hematological cancers when restricted to never-smokers. In a population-based cohort of 15 333 Swedish twins, investigators found an increased overall cancer risk (hazard ratio 1.15, 95% confidence interval 1.01-1.32) among participants with self-reported tooth mobility involving at least half of their dentition after adjusting for established risk factors including smoking. 2 Statistically significant associations were also observed TA B L E 1 Periodontal disease and total (incident) cancer estimates By contrast, findings from the Glasgow alumni cohort study found no association between the number of missing teeth and overall cancer mortality (hazard ratio 1.00, 95% confidence interval 0.98-1.02) or lung cancer mortality. 17 Two other studies reported no associations between the number of missing teeth (used as a proxy measure for periodontal disease) and cancer mortality among Swedish women and an older Japanese population, respectively. 18, 19 Abnet et al 20

| Newer epidemiologic evidence
Research evidence from recent years repeatedly indicates that periodontal disease is linked to an increased risk of head and neck cancer. In 2013, three different meta-analyses were conducted into this association. One of the meta-analyses examined periodontal disease and risk of head and neck cancer. 42 The other two evaluated tooth loss with head and neck cancer risk. 43 Similarly, findings from other, more recent meta-analyses 45,46 and one systematic review 47 all point to an association between periodontal disease and oral cancer risk. In a meta-regression analysis involving seven case-control studies, the authors observed a linear-dose response between the number of missing teeth and oral cancer risk. For each additional missing tooth, the odds ratio significantly increased by 0.03 (95% confidence interval 0.01-0.05), although moderate heterogeneity (I 2 = 67.5%, P = 0.003) was present. 9 There was, however, no linear-dose effect observed with respect to head and neck cancer risk.
A retrospective study comprising 178 oral cancer cases and 123 controls examined the relationship between chronic periodontitis (evidenced by bone loss) and oral squamous cell carcinoma.
Higher mean bone loss was related to an increased oral cancer risk (odds ratio 2.4, 95% confidence interval 1.5-3.8). 14 Similar positive associations were reported in other case-control studies between history of periodontal disease and head and neck cancer risk, 48 and generalized gingival recession in relation to oral cancer risk. 49 Those study findings are, however, in contradiction to a case-control study among the Han Chinese population that found no significant association between tooth loss and risk of oral cancer. 50 Other published epidemiologic studies include two large prospective studies. A Taiwanese

| PERI ODONTAL D IS E A S E AND RIS K OF C AN CER S OF THE D I G E S TIVE TR AC T
A number of studies have examined the association between periodontal disease and cancers of the digestive tract. Results from early studies indicate periodontal disease is linked to an increased risk of upper gastrointestinal tract cancer incidence and mortality. 51 A more recent study involving elderly Japanese found a positive association between tooth loss and orodigestive cancer mortality. 56 The latest study on periodontal disease and orodigestive cancer risk relied on clinical measures based on the Center for Disease Control/ American Academy of Periodontology definition and found no association between mild, moderate, or severe periodontitis relative to risk of orodigestive cancer. 6 Individuals with no evidence of periodontitis were the reference group. No differences in risk were observed for race or in never-smokers.
The majority of evidence for periodontal disease and orodigestive/upper gastrointestinal cancer suggests a positive association, but most of the data are based on mortality rather than incidence.

| PERI ODONTAL D IS E A S E AND RIS K OF C AN CER S OF THE D I G E S TIVE TR AC T -E SOPHAG E AL C AN CER
Studies assessing the risk of esophageal cancers have been largely population-based involving participants primarily from China, 51,57,58 Iran, 59

| Newer epidemiologic evidence
More recent studies reporting on the association between periodon- confidence interval 1.00-1.02). In the third meta-analysis, which was drawn from three cohort studies, five case-control studies, and one cross-sectional study, tooth loss was also positively linked to an increased risk of esophageal cancer in their pooled analysis (relative risk 1.30, 95% confidence interval 1.06-1.60, I 2 = 13.5%). A significant dose-response relationship was present (relative risk 1.01, 95% confidence interval 1.00-1.03; P for nonlinearity = .45). 64 In these latter two meta-analyses, inconsistencies were noted across stud-

| PERI ODONTAL D IS E A S E AND RIS K OF C AN CER S OF THE D I G E S TIVE TR AC T -G A S TRI C C AN CER
There has been at least one study linking periodontal disease with an increased risk of premalignant gastric lesions, 66

| PERI ODONTAL D IS E A S E AND RIS K OF C AN CER S OF THE D I G E S TIVE TR AC T -PAN CRE ATI C C AN CER
Some early epidemiologic studies have demonstrated a positive relationship between periodontal disease and pancreatic cancer. These include a study of 29  The precise nature of the association between periodontal disease and pancreatic cancer risk has yet to be fully determined. The role of smoking relative to this association has not been clearly delineated. However, the fact that periodontal disease has been positively linked to pancreatic cancer risk among never-smokers in a few large prospective studies may be an indication that periodontal disease has a role in the development of pancreatic cancer independent of smoking status, and should be explored further.

| Newer epidemiologic evidence
A meta-analysis conducted in 2017 investigated the association between periodontal disease and colorectal cancer based on data obtained from four cohort studies. 9 One study focused on risk of death from colorectal cancer rather than incident colorectal can-  confidence interval 1.14-3.30) and never-smokers (odds ratio 0.58, 95% confidence interval 0.12-2.78). This is an indication that smoking may be a strong contributory factor in the association between periodontal disease and lung cancer. For Tu et al, 17 however, lung cancer mortality was not associated at all with periodontal disease.

| Newer epidemiologic evidence
Few published articles have emerged on the association between periodontal disease and lung cancer risk in recent years.
Summary estimates from a meta-analysis based on the assessment of five cohort studies associated increased lung cancer risk with the presence of periodontal disease (hazard ratio 1.24, 95% confidence interval 1.13-1.36; P for heterogeneity = .22). 77 Results from another recent meta-analysis, also based on five cohort studies, showed a positive association between periodontal disease and lung cancer risk, pooled relative risk 1.33, 95% confidence interval 1.19-1.49; with no statistical heterogeneity detected (I 2 = 0, P = 0.58). 9 These results need to be interpreted cautiously since each meta-analysis only included a few observational studies that adopted different methods of periodontal disease assessment. In one of the studies included in the metaanalysis by Michaud et al, 9 although the authors controlled for smoking status and duration, the association between periodontal disease (determined via self-report) and lung cancer risk

| PERI ODONTAL D IS E A S E AND B RE A S T C AN CER RIS K
There is a paucity of scientific literature on the relationship between breast cancer and periodontal disease, and results have varied.
One of the older studies examined the association between tooth loss (obtained via self-report) and the risk of cancer at 14 common sites, including breast cancer, in a Japanese population.

| PERI ODONTAL D IS E A S E AND C AN CER RIS K : OTHER LE SS COMMON MALIG NAN CIE S
For a number of cancer sites, there is very little published epidemiologic data primarily because the number of documented cancer cases at these sites in individuals with periodontal disease are often too limited in number to allow for any robust statistical analyses or meaningful conclusions to be drawn. Nevertheless, some such uncommon cancer sites in relation to periodontal disease are briefly described below.

| Gall bladder cancer
The only published study in relation to gall bladder cancer risk reported a significantly greater risk of diagnosis of gall bladder cancer (n = 60) among those who self-reported a history of periodontal disease compared with those with no history of periodontal disease

| Prostate cancer
Lee et al 86

| Genitourinary cancers
Periodontal disease has been linked to an elevated risk of genitourinary cancers combined (hazard ratio 1.30, 95% confidence interval 1.21-1.39). 11 No associations were observed with cancers of the

| PL AUS IB LE MECHANIS TI C LINK S B E T WEEN PERI ODONTAL D IS E A S E AND C AN CER RIS K
The mechanism through which cancer may develop among individuals who have periodontal disease is not entirely clear. A number of plausible mechanisms have been proposed and inflammation appears to play a significant role in many of these postulated mechanisms. This is not surprising as periodontal disease is a prototype of an infectious process that induces chronic low-grade inflammation if left untreated. Infection has been known to promote inflammation, and persistent low-grade inflammation has been linked to cancer. 89

| Oral microbiome and inflammation in relation to cancer risk
In periodontitis, subgingival biofilms serve as reservoirs of anaerobic, gram-negative bacteria. 98 Periodontal pathogens such as Porphyromonas gingivalis and Aggregatibacter actinomycetemcomitans release enzymes that aid in digesting extracellular matrix components including collagen in order to produce substrates for their own nutrition and to enhance tissue invasion. 99 These bacterial enzymes, as well as other bacterial components such as endotoxins, and metabolic by-products that are naturally toxic to tissues, may cause direct DNA damage to neighboring epithelial cells. They can induce mutations in proto-oncogenes and tumor suppressor genes, or interfere with the molecular pathways involved in cell proliferation and/or survival. 100 Notable differences have been observed in the composition of the bacterial microflora of tumor tissue within the oral cavity compared with nontumor sites, and this shift in bacterial colonization may be associated with an increased risk of oral squamous cell carcinomas. 101 Oral bacteria may promote oral carcinogenesis by constitutively activating toll-like receptors (such as toll-like receptor-5). 102 Toll-like receptors usually present on the surfaces of cells of the innate immune system have also been associated with epithelial and cancer cells, 103 and are implicated in inflammation, cellular proliferation, invasion, and evasion of antitumoral immune responses. 104,105 Periodontal pathogens such as Fusobacterium nucleatum have been isolated from inflammatory bowel disease conditions, including Crohn's disease 106 and ulcerative colitis. 107 They have also been detected in premalignant lesions such as colorectal adenomas 108 and colorectal cancers. 109

| Periodontal disease and cancer risk -the oral-systemic link
Dense collections of gram-negative anaerobic bacteria that have been detected in the periodontal pockets of people with periodontitis 117 may become detached and be micro-aspirated or ingested. Alternatively, the ulcerated periodontal pocket walls may unwittingly serve as a potential avenue for the escape of toxic metabolites, 34 or the oral bacteria themselves and their associated components, into the systemic circulation, to lodge at distant body sites. 118 This proposed mechanism is supported by the fact that transient bacteremia has been reported following routine periodontal examination, 119 and during daily activities such as toothbrushing, and may occur more frequently among individuals with gingival inflammation. 120 Masticatory forces also appear to enhance the release of significant amounts of oral bacterial proinflammatory components such as lipopolysaccharides into the systemic circulation in individuals with severe periodontitis. 121 Alhough the release of these oral bacteria into the blood stream is temporary, periodontal pathogens such as P. gingivalis can evade their uptake and destruction by phagocytes via their protease production. 122  It is especially important to note that oral organisms have also been found in precancerous lesions of the stomach 137 and of the colon (ie, colorectal adenomas), 108 and certain cancers such as colorectal carcinomas, 109,110 oral carcinomas, 138 esophageal, and gastric cancers. 139,140 Oral bacteria in the bloodstream may cause harm via their lipopolysaccharide component, by inducing either a systemic 141 or local inflammatory response at the site where they lodge.
Lipopolysaccharide is a prototype of a pathogen-associated molecular pattern 142 and a potent inducer of inflammation, even in tiny amounts. 143,144 Pathogen-associated molecular patterns serve as ligands for toll-like receptors present on cell surfaces of the innate immune system known to instigate inflammation. 142 Okuda et al 145  cases of chronic respiratory disease, n = 386). They posited that inflammatory mediators from the diseased periodontium could be aspirated into the lungs to promote bacterial pneumonia. [147][148][149] Local inflammatory mediators produced in response to periodontal disease such as interleukin-6, tumor necrosis factor-alpha, and prostaglandin E 2 , escape through the damaged periodontal tissue pockets into the systemic circulation to produce their systemic effects at remote body sites. 150 Loos et al 150  have been a few large studies. 151,153,158 Furthermore, increased serum C-reactive protein levels have been linked to increased risk of precursor lesions, 159 as well as various cancers. 160,161 If the periodontal disease is untreated, persistence of the initiating factors is likely to occur and inflammation may fail to resolve. Over time, the chronic stimulus from the diseased periodontium could lead to the generation of persistent, low-grade inflammation that may contribute to the carcinogenic process. 80,162 This is a likely mechanism through which periodontal disease may be linked to an increased risk of cancer.

| Other plausible mechanistic links between periodontal disease and cancer
Refractory periodontitis has been linked to phenotypic changes in the mononuclear cell-cytokine system resulting in a much stronger inflammatory response than usual upon exposure to certain bacterial stimuli such as lipopolysaccharide. 163 Genetic polymorphisms involving inflammatory cytokines may contribute to individual susceptibility to disease severity 90 and possibly cancer. Certain individuals with chronic periodontitis also possess an inherent defect in their immune system, particularly with regard to bacterial clearance and tumor immune surveillance. 164 This may increase their susceptibility to cancer. Matrix metalloproteinases play a crucial role in extracellular matrix and basement membrane degradation. 165 This property aids in periodontal tissue destruction, 166,167 as well as in cancer progression and metastasis, by causing tissue dissolution enabling tumor invasion. 168,169 Susceptibility to chronic periodontitis has been demonstrated among Chinese with certain matrix metalloproteinase genetic polymorphisms. 170 This enhanced feature of matrix metalloproteinases may have implications for increased cancer risk, too.
In hyperglycemic states, as observed with diabetic patients, nonenzymatic glycation and oxidation of proteins and lipids occurs, leading to the formation of advanced glycation end products (advanced glycation end products). These substances can accumulate and cause pathogenic changes through interaction with their associated receptor, receptor for advanced glycation end products. Advanced glycation end product receptor is expressed on various cell surfaces and has been implicated in several disease conditions including inflammation, periodontal disease, diabetes, and cancer. 171,172 Advanced glycation end product-advanced glycation end product receptor interactions can upregulate certain inflammatory cytokines like tumor necrosis factor-alpha and other inflammatory mediators such as lipopolysaccharides, leading to overresponsive immune responses in the diabetic patient, thereby promoting bone destruction, such as observed in severe periodontal disease. 172 Advanced glycation end product receptor ligands are also secreted by cancer cells and help to promote carcinogenesis by stimulating cancer cells directly, causing them to act autonomously. They also modulate various cell types within the tumor microenvironment, such as fibroblasts, leukocytes, and vascular cells, resulting in increased fibrosis, inflammation, and angiogenesis. 171 Therefore, the interrelationships between periodontal disease, diabetes, and cancer risk should be given further consideration.
Advanced periodontal disease is often accompanied by tooth loss. When severe, it may reduce masticatory efficiency and lead to the avoidance of chewing tough foods like fruits and fibrous vegetables. 173,174 Unfortunately, these foods are often more nutritious and rich in cancer-fighting agents such as antioxidants. Studies have shown that fruits and nonstarchy vegetables are associated with a decreased risk of certain cancers (oral cavity, pharynx, esophagus, stomach, and lung). 175 Soft diets are less nutritious but more energy-dense. High calorific diets are described as proinflammatory because they are associated with higher levels of C-reactive protein and interleukin-6. 176 Diets with higher dietary inflammatory index scores may promote cancer risk. [177][178][179] Finally, epigenetic changes resulting in the hypermethylation of E-cadherin and cyclooxygenase-2 have been associated with chronic periodontitis, and may be linked with increased cancer risk. 180

| CON CLUS IONS
Our understanding of the relationship between periodontal disease and the risk of developing certain cancers is still evolving. Available data from epidemiologic evidence on periodontal disease and cancer risk mostly points to a positive association. It appears that risk may be higher for certain anatomic sites, particularly those in close proximity to the oral cavity (esophagus, upper gastrointenstinal tract).
However, proving the unequivocal presence of an association between periodontal disease and cancer risk remains a difficult challenge. The majority of epidemiologic data have been derived from observational studies with attendant difficulties relating to the use of diverse proxy measures for periodontal disease ascertainment.
Other challenges include issues with residual confounding by smoking in study populations involving smokers, and proper accountability of other unanticipated extraneous factors that might contribute to the association. There is therefore a need for further exploration using randomized controlled clinical trials and well-designed, large prospective studies to help clarify the nature of association and, in particular, add evidence to support causality between periodontal disease and risk of cancer. These studies should strive to incorporate standardized clinical measurements in ascertaining periodontal disease status. More investigation is also needed to assess how improved periodontal disease prevention and management strategies may impact cancer risk.
Ever since the Joint European Federation of Periodontology/ American Academy of Periodontology Workshop on "Periodontitis and Systemic Diseases" in November 2012, several other studies have added to the evidence that could support biological plausibility. The role of periodontal disease and oral microorganisms needs to be studied further, including how they are involved in the association between periodontal disease and cancer risk. Such studies may lead to novel approaches to prevent some cancers. Since there is also some evidence that indicates inflammation may mediate the association between periodontal disease and/or its pathogens and cancer risk, more research may help gain a better understanding in this regard.