Association of Periodontal Diseases with Intracranial Aneurysm Formation: Novel Predictive Indicators

Background and purpose: We investigated whether periodontal diseases, specically, periodontitis and gingivitis, could be risk factors of the incidence of intracranial aneurysms (IAs). Methods: We performed a case–control study to compare the differences in the periodontal disease parameters of 281 cases that were divided into the IA and non-IA groups. All cases underwent complete radiographic examination for IAs and examination for periodontal health. Results: Compared with those in the non-IA group, the cases in the IA group were older (53.95 ± 8.56 vs 47.79 ± 12.33, p < 0.001) and had a higher incidence of hypertension (76 vs 34, p = 0.006). Univariate analysis revealed that age (>50 years) and hypertension were risk factors of aneurysm formation (odds ratio [OR] 1.047, 95% condence interval [95%CI] 1.022–1.073, p < 0.001 and OR 2.047, 95%CI 1.232– 3.401, p = 0.006). In addition, univariate and multivariate logistic regression analyses showed that the parameters of periodontal diseases, including gingival index, plaque index, clinical attachment loss, and alveolar bone loss, were signicantly associated with the occurrence of IA (all p < 0.05). For further statistical investigation, the parameters of periodontal diseases were divided into four layers on the basis of their averages. Poor periodontal health condition was correlated with a high risk of IAs. Conclusion: Periodontal diseases, especially severe gingivitis or periodontitis combined with hypertension, were signicantly associated with the incidence of IAs. individuals with IAs and hospitalized volunteers (non-IA) who wanted to participate in this study, and (3) individuals who underwent the examination of periodontal health condition. Exclusion criteria included (1) individuals with acute ruptured aneurysms; (2) individuals receiving antihypertensive therapy with calcium channel blockers, such as Nifedipin; (3) individuals with severe cardiovascular diseases or cerebral ischemic stroke; (4) individuals with malignant diseases, chronic inammatory diseases, or antibiotic use within 2 weeks; and (5) individuals missing clinical follow-up. Baseline demographic information (including age, sex, clinical presentation, smoking and drinking history, diabetes, hypertension, and hypercholesterolemia) and IA characteristics (location, size, shape, and quantity) were recorded.


Introduction
Intracranial aneurysms (IAs) are pathological dilatations of cerebral arteries; they are most often saccular in shape and frequently found in proximal cerebral artery bifurcations [1] . Subarachnoid hemorrhages caused by IA rupture are rare events that affect 10-11/100,000 population per year in Western populations [2] . IA rupture and subsequent hemorrhage may account for a mortality rate of 35%, and most survivors of IAs are left with considerable neurological impairment [3][4] . Unruptured IAs are commonly treated through endovascular intervention or neurosurgical procedures to decrease the possibility of subarachnoid hemorrhage. However, most unruptured IAs are asymptomatic, and patients ignore potential risks. Thus, IAs must be screened before rupture, and the pathological mechanism of aneurysm formation should be emphasized.
Periodontitis and gingivitis are chronic in ammatory responses of the oral supporting tissues (gingiva, periodontal ligaments, and alveolar bone) surrounding the teeth [5][6][7] . Periodontitis results from the complex interplay between chronic bacterial infection and in ammatory host response and leads to the irreversible destruction of tooth-supporting tissues with tooth loss as a common end point [8] . Gingivitis develops gradually and includes in ammatory changes in the gingiva that are most commonly induced by the accumulation of dental plaque. In recent years, researchers have investigated the association between systematic vascular diseases and periodontitis or gingivitis. Lockhart et al. reported that oral bacterial deoxyribonucleic acid (DNA) has been found in arteries and suggested that oral infections may contribute to vascular wall in ammation [9] . Bacteria from periodontal pockets and their secreta, such as endotoxins, have been identi ed in cardiovascular atherosclerotic lesions [10][11][12] . Iwai investigated the potential role of periodontal diseases in the pathobiology of abdominal aortic aneurysms [13] . He detected periodontal bacterial DNA in abdominal aortic aneurysmal walls and then concluded that poor periodontal conditions have important effects on the progression of abdominal aortic aneurysms.
Nine studies have found signi cant associations between periodontal diseases and cerebravascular strokes [14][15][16][17][18][19][20][21][22] . Further clinical studies speculated that periodontal in ammation caused by Streptococcus mutans participates in the occurrence of intracranial hemorrhage strokes [23][24][25] . However, evidence for the role of periodontal diseases in IAs has received insu cient attention because of the availability of insu cient sample sizes. Therefore, we performed this case-control study to investigate whether periodontal diseases are associated with IA formation and to discover novel epidemiological evidence.

Study population
From September 2015 to September 2018, we collected cases that were admitted to our institution for IA detection via digital subtraction angiography or computer tomography angiography examination. During the same period, we also recruited cases that were diagnosed without IAs or several volunteers. A total of 281 cases provided written informed consent and were divided into two groups (166 cases in the IA group and 115 cases in the non-IA group). The privacy of the patients was strictly protected, and the protocol of this study was approved by our ethics committee. We subjected all recruited cases to periodontal health examination through a standardized approach in a dental unit by using a standard dental light, compressed air, a mouth mirror, and digital panoramic radiography free of charge. Periodontal disease was diagnosed on the basis of the clinical and radiographic criteria described by the 1999 Consensus Classi cation of Periodontal Diseases [26] .
Inclusion criteria were as follows: (1) individuals who were 18-80 years of age, (2) individuals with IAs and hospitalized volunteers (non-IA) who wanted to participate in this study, and (3) individuals who underwent the examination of periodontal health condition. Exclusion criteria included (1) individuals with acute ruptured aneurysms; (2) individuals receiving antihypertensive therapy with calcium channel blockers, such as Nifedipin; (3) individuals with severe cardiovascular diseases or cerebral ischemic stroke; (4) individuals with malignant diseases, chronic in ammatory diseases, or antibiotic use within 2 weeks; and (5) individuals missing clinical follow-up. Baseline demographic information (including age, sex, clinical presentation, smoking and drinking history, diabetes, hypertension, and hypercholesterolemia) and IA characteristics (location, size, shape, and quantity) were recorded.

Assessment of periodontal diseases
Periodontal diseases mainly include periodontitis and gingivitis. Gingivitis severity was assessed by using the gingival index (GI) system, which is based on the various tendencies of gingival bleeding after gingival irritation. A high index number (> 1.1) was de ned as severe gingivitis. Periodontitis was evaluated by using the plaque index (PI), clinical attachment loss (CAL), and alveolar bone loss (ABL). Plaque and gingivitis were scored at four sites per tooth (buccal, mesiolingual, lingual, and distolingual) and averaged for each subject. We presented these parameters in accordance with the following de nitions: First, the CAL was examined by inserting the tip of a CPI probe to measure the distance between the pocket and cementoenamel junction. Attachment levels were analyzed as continuous variables, and mean CAL > 4 mm was considered as severe periodontitis. Second, dental plaque was scored in accordance with the PI system, which is based on the same principle as the GI system. We divided PI into 4 sites (≤ 0.5, 0.51-1.0, 1.01-1.5, and > 1.5). A high score represented severe periodontitis.
Third, ABL levels were measured as the distance from the cementoenamel junction to the most apical extension of the bony defect. We strati ed the ABL into < 3, 3.00-4.00, 4.00-5.00, and > 5.00 mm. Among these periodontal parameters, the CAL was de ned as the primary criterion of periodontitis.

Statistical analysis
Data were presented as mean ± standard deviation or expressed in terms of frequencies and percentages. Independent sample t-tests and chi-square tests were performed to test for differences in continuous or categorical variables between the IA group and the non-IA group, respectively. Subgroup univariate and multivariate logistic regression analyses were performed to identify the independent contribution of gingivitis and periodontitis parameters (including GI, CAL, PI, and ABL) to the incidence of IAs. Odds ratios (ORs) and 95% con dence intervals (CIs) were given for all periodontal parameters. A two-sided p value of < 0.05 was considered to be signi cant. Statistical analysis was performed by using SPSS 22.0 software (SPSS Inc., Chicago, IL, USA).

Results
A total of 281 cases (including 174 [61.9%] females and 107 [38.1%] males) were recruited in this study. The ages of these cases ranged from 18 years to 79 years (51.34 ± 10.56 years). Table 1 shows the baseline characteristics and subgroup analysis results of the IA and non-IA group. No statistically signi cant differences were detected in the distributions of sex, diabetes mellitus, hyperlipidemia, cerebrovascular history (including cerebral ischemic and hemorrhage events), other personal history (such as alcohol, smoking, and BMI) in the two subgroups (all p > 0.05). Chi-square tests indicated that cases in the IA group were older than those in the non-IA group (53.95 ± 8.56 vs 47.79 ± 12.33, p < 0.001) and had a higher incidence of hypertension (76 vs 34, p = 0.006).  Table 2 provides the parameters of gingivitis and periodontitis. Signi cant statistical differences were found in GI, PI, CAL, and ABL (all p < 0.001) between the IA and non-IA group. IA cases had higher GI, PI, CAL, and ABL than non-IA cases. Univariate regression analysis revealed that these parameters, combined with age (over 50 years) and hypertension, were signi cantly associated with the occurrence of IAs (all p < 0.05) ( Table 3).  The parameters of periodontal diseases (GI, CAL, PI, and ABL) were divided into four layers on the basis of their averages and referenced literature [27] .  Table 4. All of the above parameters of severe periodontal diseases had p < 0.05. b "≤0.5" used as reference group in the binary logistic regression.
c "<3mm" used as reference group in the binary logistic regression. d "<3mm" used as reference group in the binary logistic regression.

Discussion
Periodontitis and gingivitis are associated with the risk of several diseases, such as rheumatoid arthritis and atherosclerosis [28][29] . Recent studies have discussed the role of periodontal diseases in causing cerebral ischemia. Chiu demonstrated that periodontitis is associated with stroke caused by large-artery atherosclerosis; this result supports the hypothesis of a possible link between periodontitis and atherosclerosis [30] . After adjusting for confounding vascular factors on the basis of etiologic subgroup analysis, Grau et al. concluded that severe periodontitis is an independent risk factor for IA with atherothrombotic origins (OR 2.35 [1.00-11.0] and OR 13.2 [2.68-64.7]) and that gingivitis is independently associated with cerebral ischemia given its value as an indicator of the actual status of periodontal in ammation [31] . These ndings suggest that chronic periodontal in ammatory responses contribute to a prothrombotic state via recurrent bacteremia and platelet and endothelial activation.
Plaque destabilization is a potential trigger of cardioembolism and cryptogenic stroke [32][33] . In addition, the accumulation of oxidized lipids in vascular walls and subsequent atherosclerotic remodeling may be altered indirectly by systemic immunization induced by local periodontitis.
Periodontal diseases, especially periodontitis, are associated with the remodeling of the aneurysmal wall in abdominal aortic aneurysms. In experimental models, periodontal bacteria promote the degeneration of the abdominal aortic aneurysmal wall by increasing the recruitment of neutrophils to the intraluminal thrombus that covers the inner portion of the abdominal aortic aneurysm [34] . The pathologies of IA and abdominal aortic aneurysm share several features [35] . The current understanding is that IAs form as the end result of ow-driven in ammatory cell-mediated cerebral artery wall remodeling at sites where high ow exerts high wall shear stress [1] . However, IA is not detected in all cases under high ow and shear stress in the bifurcations of cerebral arteries. Previous studies on the association between periodontal diseases and stroke focused on the pathological mechanism of IA. Pyysalo is the rst to detect the presence of oral bacteria DNA in ruptured and unruptured intracranial aneurysmal walls [36] . Hallikainen et al. speculated that periodontitis predisposes the artery wall toward aneurysm development [37] . However, few works have focused on the relationship between intracranial hemorrhagic diseases and periodontal diseases. In this case-control study, we focused on the potential association between periodontal diseases (including periodontitis and gingivitis) and IAs. Although GI, PI, CAL, and ABL have long been known as actual indicators of periodontal in ammation caused by multitudinous oral bacteria, we were the rst to demonstrate that these parameters were independent risk factors of IAs (p < 0.05, Table 3). Further subgroup analysis demonstrated that poor periodontitis parameters might be associated with the high incidence of IA. A past observation demonstrated that systemic elastase activity may play an important role given that increased serum elastase concentrations are associated with IAs, although the source of serum elastase is unknown [38] . Potential sources for circulating elastase are macrophages or neutrophils [39] . Tooth brushing or chewing disseminates periodontal Porphyromonas gingivalis (PG), especially in patients with periodontitis or gingivitis, to extraoral sites via circulation and then induces systemic in ammatory responses through transient bacteremia [40] . PG can modify dendritic cell function and induce proin ammatory cytokine production in macrophages. Thus, periodontal pathogens that in ltrate cerebral arteries likely lead to excessive collagen degradation and neutrophil accumulation in the thrombus, promoting changes in the course of cerebral artery remodeling [41] . In another mechanism, neutrophils induced by periodontal pathogens are a major source of proteolytic activity because they release the matrix metalloproteinases MMP-8 and MMP-9, myeloperoxidase, and elastase and then accelerate the course of proteolytic or cytotoxic injury [42] . The ultimate pathology is damage to the elastic bers of smooth muscles and the degeneration or necrosis of the intima medium. In combination with chronic hypertension, external vascular walls swell to form IAs. Previous literature has indicated that age (especially over 50 years) and hypertension are predictive risk factors of IA formation [43][44] . In the present study, we included patients who were older than 50 years or suffered hypertension and found that these cases had an increased incidence of IAs. Systemic hypertension not only affects tissue remodeling or vascular wall in ammation by exerting abnormal hemodynamic stresses but also activates the local renin-angiotensin system [45] . Moreover, hypertension may mediate vascular in ammation through the activation of NF-kappa B, which can further promote in ammation [46] . The continuous stimulatory response caused by hypertension and in ammation can result in the degeneration of vascular walls and then induce IA formation.
This study has several limitations. First, although the literature has indicated that periodontal treatment can induce bacteria to transfer to systematic circulation, we could not persuade all patients to undergo periodontal or caries treatment because of their poor obedience. Thus, we will further study whether periodontal treatment has low relevance to IA. Second, the speci c values of periodontal disease parameters that have the optimal predictive values for the incidence of IAs should be quanti ed through receiver operating characteristic curve analysis. Third, we did not perform further experimental research on in ammatory mediators that might cause IA directly. Fourth, other studies have demonstrated that age and hypertension are associated with the high risk of periodontal diseases. In this study, we only attempted to investigate the mechanism of aneurysm formation and neglected the relationship between these two factors and periodontal diseases.

Conclusion
Periodontal diseases were signi cantly associated with the high incidence of IA. Given our results, patients with severe gingivitis or periodontitis with hypertension should be encouraged to undergo cerebrovascular examination. Mechanistic experiments on in ammatory response will positively affect the risk prediction of IA.

List Of Abbreviations
ABL: alveolar bone loss BMI: body mass index