Association between periodontal disease and subclinical atherosclerosis : the ELSA-Brasil study

The aim of this study was to investigate the relationship between periodontal disease and increased thickness of the carotid artery intimamedia complex. A cross-sectional study was conducted with 220 adults (age ≥ 35 years) among participants in the Brazilian Longitudinal Study of Adult Health (ELSA-Brasil). Carotid artery ultrasound and periodontal clinical examinations were conducted and included visible plaque index, gingival bleeding on probing index, probing pocket depth (PPD), and clinical attachment level (CAL). Individuals with increased carotid artery thickness showed fewer teeth and higher frequency of CAL ≥ 3mm, CAL ≥ 4mm, CAL ≥ 5mm, and CAL ≥ 6mm and PPD ≥ 4mm (p < 0.05). Despite the use of 18 definitions for periodontal disease, only one confirmed the hypothesis of an association between periodontal disease and subclinical atherosclerosis. Individuals with 10% or more sites with CAL ≥ 4mm were more likely to present carotid thickening. Atherosclerosis; Carotid Artery Diseases; Periodontics Introduction Atherosclerosis is one of the leading causes of adult mortality and morbidity in Latin America 1. Due to its different etiologies, many underlying causes remain unknown 2. Studies have suggested that periodontal disease is associated with the early development of atherosclerotic lesions in the carotid artery 3,4,5,6. Periodontal disease is a chronic multifactorial immune disease that occurs in response to periodontopathogenic antigens. The condition begins with inflammation of the tissues surrounding and supporting the teeth and can progress until involving the entire periodontium, including the alveolar bone, cementum, and periodontal ligament, leading to irreversible loss of the periodontium and eventually tooth loss in the more advanced stages 7. The association between periodontal disease and atherosclerotic cardiovascular disease has received considerable attention 8,9, but these studies’ findings are not consistent. Some studies report that periodontal disease, through an inflammatory process, can cause infection of the vascular endothelium and contribute to the occurrence of atherosclerosis, increasing the risk of myocardial ischemia and infarction, always preceded by thromboembolic events 10. In addition to this hypothesis, it has been suggested that periodontal microorganisms can induce or accelerate atherosclerosis through different ARTIGO ARTICLE R ET R A TE D A R TI C LE Batista RM et al. 966 Cad. Saúde Pública, Rio de Janeiro, 28(5):965-976, mai, 2012 mechanisms, for example: favoring the local increase in lymphocytes, macrophages, and production of tissue growth factors; local release of endotoxin (lipopolysaccharides), and molecular mimicry between microbial and human heat shock protein 60, inducing an autoimmune reaction. In addition, the systemic increase in cytokines with activation of inflammatory markers and stimulation of pro-coagulants can lead to thrombosis and acute ischemia, besides inducing changes in lipoproteins, resulting in pre-atherosclerotic conditions 11. Mild to moderate periodontal disease affects 30-50% of the adult population, while the severe generalized form affects 5-15% of adults in the United States 12. The multifactorial etiology of periodontal disease includes both specific subgingival bacteria and individual factors such as age, race, and gender, and systemic factors like smoking, diabetes, osteoporosis, and stress. In addition, poor diet, low socioeconomic status, and limited access to health services have been associated with its occurrence 12. Among the risk factors presented for periodontal disease, age, smoking, and diabetes mellitus are generally considered potential confounding variables in studies on associations in periodontal disease 10. Inflammation is a common characteristic, both in atherosclerosis and periodontal disease, and thus a common mediator of these two conditions 13,14. There is no consensus concerning the possible influence of periodontal disease on the thickness of the carotid artery intima-media complex (IMC). The majority of studies that have evaluated this association present methodological limitations, like small sample size, lack of control of confounding variables, presence of bias, and discrepancies in the case definition for periodontal disease 15,16, thus limiting the internal validity and potentially jeopardizing the conclusions. The majority of the literature reviewed in the studies 8,17 supports a modest association between periodontal disease and atherosclerosis. However, lack of standardization of measures and definitions of periodontal disease, as well as the potential confounding factors common to both conditions, hinder interpretation of the results 18. A better understanding of the relationship between periodontal disease and the risks of subclinical atherosclerosis motivated the current study, the objective of which was to investigate the relationship between clinical parameters of periodontal disease and thickening of the carotid IMC, used as a proxy for diagnosing subclinical vascular disease in adults. Methods


Introduction
Atherosclerosis is one of the leading causes of adult mortality and morbidity in Latin America 1 .Due to its different etiologies, many underlying causes remain unknown 2 .
Studies have suggested that periodontal disease is associated with the early development of atherosclerotic lesions in the carotid artery 3,4,5,6 .Periodontal disease is a chronic multifactorial immune disease that occurs in response to periodontopathogenic antigens.The condition begins with inflammation of the tissues surrounding and supporting the teeth and can progress until involving the entire periodontium, including the alveolar bone, cementum, and periodontal ligament, leading to irreversible loss of the periodontium and eventually tooth loss in the more advanced stages 7 .
The association between periodontal disease and atherosclerotic cardiovascular disease has received considerable attention 8,9 , but these studies' findings are not consistent.Some studies report that periodontal disease, through an inflammatory process, can cause infection of the vascular endothelium and contribute to the occurrence of atherosclerosis, increasing the risk of myocardial ischemia and infarction, always preceded by thromboembolic events 10 .In addition to this hypothesis, it has been suggested that periodontal microorganisms can induce or accelerate atherosclerosis through different ARTIGO ARTICLE

R E T R A C T E D A R T I C L E
mechanisms, for example: favoring the local increase in lymphocytes, macrophages, and production of tissue growth factors; local release of endotoxin (lipopolysaccharides), and molecular mimicry between microbial and human heat shock protein 60, inducing an autoimmune reaction.In addition, the systemic increase in cytokines with activation of inflammatory markers and stimulation of pro-coagulants can lead to thrombosis and acute ischemia, besides inducing changes in lipoproteins, resulting in pre-atherosclerotic conditions 11 .
Mild to moderate periodontal disease affects 30-50% of the adult population, while the severe generalized form affects 5-15% of adults in the United States 12 .The multifactorial etiology of periodontal disease includes both specific subgingival bacteria and individual factors such as age, race, and gender, and systemic factors like smoking, diabetes, osteoporosis, and stress.In addition, poor diet, low socioeconomic status, and limited access to health services have been associated with its occurrence 12 .
Among the risk factors presented for periodontal disease, age, smoking, and diabetes mellitus are generally considered potential confounding variables in studies on associations in periodontal disease 10 .
Inflammation is a common characteristic, both in atherosclerosis and periodontal disease, and thus a common mediator of these two conditions 13,14 .There is no consensus concerning the possible influence of periodontal disease on the thickness of the carotid artery intima-media complex (IMC).The majority of studies that have evaluated this association present methodological limitations, like small sample size, lack of control of confounding variables, presence of bias, and discrepancies in the case definition for periodontal disease 15,16 , thus limiting the internal validity and potentially jeopardizing the conclusions.
The majority of the literature reviewed in the studies 8,17 supports a modest association between periodontal disease and atherosclerosis.However, lack of standardization of measures and definitions of periodontal disease, as well as the potential confounding factors common to both conditions, hinder interpretation of the results 18 .
A better understanding of the relationship between periodontal disease and the risks of subclinical atherosclerosis motivated the current study, the objective of which was to investigate the relationship between clinical parameters of periodontal disease and thickening of the carotid IMC, used as a proxy for diagnosing subclinical vascular disease in adults.

Participants
A cross-sectional study was conducted among participants in the Brazilian Longitudinal Study of Adult Health (the ELSA-Brasil).The ELSA-Brasil study is a multicenter cohort study consisting of 15 thousand employees from public institutions of higher learning and research in the Northeast, South, and Southeast regions of Brazil.The purpose of the study is to investigate the incidence and risk factors for chronic diseases, in particular cardiovascular diseases and diabetes.The supplemental oral health study for ELSA-Brasil was based on data collected in the ELSA project in the State of Espírito Santo in 2009-2010, after anthropometric examination and carotid ultrasound.This study's target population consisted of male and female active and retired employees from 35 to 74 years of age, from the ELSA project/ Espírito Santo.The sample universe consisted of the 497 participants in the ELSA project that were treated and had undergone ultrasound of the common carotid artery by a single examiner.Providing for possible losses, all were invited to participate in the study.The only exclusion criterion was the use of total upper and lower dental prostheses.
The final sample consisted of 220 participants (78 with increased IMC thickness and 142 with normal IMC).This sample size allowed detecting an expected 25% (estimated) difference between the proportion of periodontal disease in the sample with normal carotid IMC and in the sample with altered carotid IMC, with a minimum power of 80% and 5% significance.
An intra-examiner clinical calibration study for the periodontal examination was conducted previously.A single examiner responsible for examination in the principal study performed the intra-examiner calibration, under the supervision of the external observer, in the same location and under the same conditions of lighting and instrumentation, with employees from the same age bracket as the study participants.Seventeen patients were examined, with 48-hour intervals between examinations.The clinical parameters considered for calibration were: probing pocket depth (PPD) and clinical attachment level (CAL) of teeth 16, 11, 27, 31, 37, and 46, with a total of 96 measurements performed (48 measurements at two moments).Reliability between measurements was tested using the intra-class correlation coefficient (ICC) and comparison of the means with the matched t test.Of 48 comparisons, 41 showed statistically significant correlations, of which 5 with ICC > 0.80 and 3 with ICC

R E T R A C T E D
A R T I C L E from 0.50 to 0.80.The analysis did not include three comparisons, because there was no variation in the measurements.In the comparison of means, 47 of the 48 comparisons were statistically significant.

Periodontal clinical examination
Participants were submitted to periodontal examination by a single calibrated examiner who was blind to the carotid ultrasound data.
Periodontal clinical measurements used in the study were: plaque index (PI) 19 , gingival bleeding index (GBI) 19 , PPD, and CAL.The PI and GBI parameters were measured dichotomously as the presence or absence of bacterial plaque and bleeding after periodontal probing.
For PPD, CAL, GBI, and PI, six measurements were taken per tooth, corresponding to the mesiobuccal, midbuccal, distobuccal, mesiolingual, midlingual, and distolingual surfaces.All the clinical measurements were taken in all the teeth, except for the third molars.
PPD measurements were recorded in millimeters from the free gingival margin to the bottom of the gingival sulcus or periodontal pocket.In the CAL measurements, the cementumdentin junction and gingival level were used as reference points.PPD and CAL measurements were recorded with a North Carolina model periodontal probe (Hu-Friedy, Chicago, USA), with 1-millimeter markings over a total of 15 mm and a tip with 0.35mm diameter.
When the gingival margin or cementum-dentin junction was located between two marks on the periodontal probe, the value from the deepest marking was recorded.In addition to the periodontal measurements, the number of natural teeth was also recorded.All measurements were taken in a single sitting.

Ultrasound examination
The arterial wall consists of three layers: the intima, a thin endothelial layer that borders on the blood-filled lumen, the adventitia, the external layer bordering on the surrounding tissue, and the media, a layer between these internal and external layers.Due to the easily discernible borders between the lumen and the intima and between the media and adventitia, the intima and media are generally evaluated together, which provides the measurement of the IMC thickness.
Progression of atherosclerotic disease is frequently accompanied by a visible thickening of the IMC and the presence of fibrotic or calcified plaques 20 .
Measuring the carotid artery IMC thickness is a way of evaluating one of the most important indicators of cardiovascular disease in patients.Its use is based on the possibility of predicting the outcomes of future cardiovascular events with a non-invasive technique, namely ultrasound (US), providing a diagnostic gain based on the importance of IMC thickening as a risk factor for cardiovascular disease 20 .
IMC thickness was measured bilaterally next to the carotid sinus at the base of the neck, using Toshiba Ultrasound Aplio XG (Model SSA, 790 A; Toshiba, Barueri, Brazil).The test was performed by a researcher who was blind to the periodontal parameters and other patient characteristics.

Comparison groups
There is still much disagreement on which IMC thickness offers the best result for methodological standardization.Standardization of a protocol to measure IMC changes would facilitate the comparison of results from studies using this technique 20 .
According to the consensus statement of the American Society of Echocardiography on the use of carotid ultrasound to identify subclinical vascular disease, the sample's 75 th percentile corresponds to the values adopted as possible future risk of cardiovascular disease 20 .
This study used the value corresponding to the sample's 75 th percentile as altered carotid IMC thickness, which varied from 0.95mm in the right common carotid artery to 0.98mm in the left common carotid artery.Thus, participants with right common carotid artery ≥ 0.95mm and left common carotid artery ≥ 0.98mm were defined as having altered carotid IMC.

Covariables
The sociodemographic characteristics were collected through a structured interview conducted by trained interviewers and included: age, gender, marital status, self-reported race/skin color, schooling, family income, smoking, and self-reported hypertension, diabetes, and insulin use.The anthropometric characteristic evaluated by the study was body mass index (BMI = kg/m 2 ), or weight (kg) divided by height squared (m 2 ).The equipment used to obtain these measurements included a stadiometer accurate to 0.1cm, electronic scales calibrated in grams (Toledo 200kg; Toledo do Brasil Indústria de Balanças Ltda., São Bernardo do Campo, Brazil), anthropometric tape measure, 150cm -Mabbis, Gulick model (CARDIOMED, Curitiba, Brazil), 20kg test weight -total 80kg, mobile footrest with height on 8cm

R E T R A C T E D A R T I C L E
plane, and 1.20x0.50cmmirror.BMI was classified according to the World Health Organization (WHO) classification 21 .

Periodontal clinical parameters and definition of periodontal disease
All the clinical parameters, PI, GBI, PPD, and CAL, were computed for each study subject and later for each group (normal and altered carotid IMC thickness).Clinical parameters analyzed by the study were number and frequency of sites with PPD and CAL ≥ 3mm, ≥ 4mm, ≥ 5mm, and ≥ 6mm.

Statistical analysis
Comparisons between IMC groups were performed with the chi-square test, Fisher's exact test, Mann-Whitney test, and t test.Correlation between the periodontal clinical parameters and carotid artery IMC thickness was evaluated with the Spearman correlation coefficient.Simple logistic regression was performed to test the association between different definitions of periodontal disease and altered carotid IMC thickness through odds ratios (OR) with 95% confidence intervals (95%CI).Multivariate analyses were then used to adjust covariables.All analyses used SPSS, version 18.0 (SPSS Inc., Chicago, USA).Statistical significance was set at 0.05 (p ≤ 0.05).

Research ethics issues
Participants were required to sign a prior informed consent form.The project was approved by the Institutional Review Board of the Federal University in Espírito Santo (Universidade Federal do Espírito Santo; protocol n o .145/08).

Results
The sample consisted of 220 individuals among the 497 participants in the ELSA-ES project.It was not possible to contact 160 participants (32.1%).
Of the 337 individuals that were invited, 61 (18.1%) refused to participate.Of the 276 that agreed, 45 (16%) failed to appear for the examination and 11 (4.34%) were excluded because they were using total upper and lower dental prostheses (Figure 1).
There were proportionally more females among the group that agreed to participate (58.2%).65.8% of the study sample were married.White skin color/race was the most common (45.4%), followed by brown race (pardo) (33.5%).Table 1 shows the participants' sociodemographic characteristics.
Table 2 shows the sociodemographic characteristics and BMI according to IMC thickness (altered versus normal).Mean age and the proportion of male individuals and those with diabetes, hypertension, and overweight/obesity were statistically higher in the group with altered IMC.Smoking did not differ statistically between groups (p = 0.5604).
Table 3 compares number of teeth and periodontal parameters between groups.Mean number of teeth present, number of sites with CAL ≥ 3mm, CAL ≥ 4mm, CAL ≥ 5mm, and CAL ≥ 6mm, and frequency of sites with CAL ≥ 4mm, CAL ≥ 5mm, and CAL ≥ 6mm and PPD ≥ 4mm were statistically higher in the group with altered IMC thickness as compared to the group with normal IMC.Associations were observed between periodontal clinical parameters and carotid IMC thickness, based on analyses of correlations.Right carotid artery IMC thickening was correlated with PPD ≥ 5mm and CAL ≥ 5mm and CAL ≥ 6mm (p < 0.05), while left carotid artery IMC thickening was associated with PPD ≥ 4mm, PPD ≥ 5mm, and CAL ≥ 5mm and CAL ≥ 6mm (p < 0.05) (Table 4).
Table 5 shows the crude and adjusted associations between different definitions of periodontal disease and altered carotid IMC thickness.
Among the 18 definitions of periodontal disease, seven showed a positive unadjusted association between periodontal disease and altered carotid IMC thickness.In the multivariate analysis, which included the confounding variables, only definition 5 of periodontal disease remained associated with the outcome.The odds of detecting altered carotid IMC thickness were 2.56 times higher in those with at least 10% of sites with CAL ≥ 4mm (95%CI: 1.12-5.88).
Table 6 shows the results of multivariate analysis between the variables age, gender, hypertension, diabetes, BMI, periodontal disease, and altered carotid IMC thickness.In the final logistic regression model, in addition to periodontal disease (definition 5), only age remained associated with altered carotid IMC thickness (OR = 1.04; 95%CI: 1.01-1.08).more frequent in individuals with altered mean carotid IMC thickness.Although 18 definitions of periodontal disease were used, only one confirmed the hypothesis of an association between periodontal disease and subclinical atherosclerosis.Individuals with 10% or more sites with CAL ≥ 4mm showed higher odds of carotid IMC thickening.

R E T R A C T E D A R T I C L E
As in the study by Demmer et al. 4 , measures of severity and extent of CAL and higher tooth loss were associated with chronic markers such as altered IMC thickness, while the same was not observed with PPD.
Findings from studies on the association between periodontal disease and atherosclerosis are highly dependent on how periodontal disease is defined.There is no consensus on the standard definition for diagnosing periodontal disease in these studies.In addition, other limitations were observed, such as indirect, inadequate, or incomplete measurements of periodontal disease, as well as self-reported periodontal disease 33 .This scenario produces inconsistent findings, leading researchers to question both the magnitude and significance of the observed associations 8,17 .
Although there are studies on the association between periodontal disease and cardiovascular diseases, little research has specifically addressed the possible systemic effects of the systemic dissemination of oral microorganisms in humans 4,13 .
In a longitudinal study that defined bacterial exposure using immunoglobulin G (IgG) serum antibodies, the odds of carotid atherosclerosis were higher in subjects with elevated levels of oral microorganisms.This association was independent of smoking, suggesting that the latter was not a confounding or effect-modifying variable 14 .Likewise, the present study found no association between smoking and subclinical atherosclerosis.

R E T R A C T E D A R T I C L E
Recent studies found modest associations between advanced periodontal disease and tooth loss and carotid IMC thickening in developed countries 4,6 .Other studies were conducted in individuals with associated conditions such as chronic kidney disease 3 or kidney transplant 34 .
The observed association between severe periodontal disease and increased carotid IMC thickness had been reported in a previous study 15 .Severe periodontal disease was associated with increased mean IMC thickness (OR = 1.312; 95%CI: 1.03-1.66),but this association did not remain after the adjusted analysis.
Severe periodontal disease was also associated with subclinical atherosclerosis (IMC thickness ≥ 0.82mm) in systemically healthy young individuals and with age ≤ 40 years 5 .Even using different age brackets, both the previous study and our study found an association, so age is an important confounding variable 16 .Another dif-ference that appeared when comparing our findings with those of previous studies relates to the cutoff between altered and normal IMC thickness 5 .It is difficult to compare studies that use the altered versus normal classification of IMC thickness, because there is no consensus on the cutoff to define the altered group, since this is done with the 75 th percentile of IMC thickness for the sample.Thus, the classification of altered carotid IMC thickness depends on the distribution of measurements of IMC thickness in each study.
A limitation of the current study was its crosssectional design, which does not allow inferences on causality.Importantly, only one of the 18 definitions of periodontal disease was associated with carotid IMC thickening after multivariate analysis, and in this measurement only CAL was considered.Although there is no clinical parameter that indicates activity of periodontal disease,

R E T R A C T E D
A R T I C L E

R E T R A C T E D A R T I C L E
definitions that include PPD and bleeding are considered more robust, since CAL is known to be related to history and sequelae of periodontal disease.
Another limitation of this study is that only clinical measurements of periodontal disease were evaluated.Microbiological data and inflammatory markers for periodontal disease have been proven as more specific indicators 13,14 as compared to clinical parameters.The study also failed to measure such risk parameters for cardiovascular disease as C-reactive protein, cholesterol (LDL and HDL), and triglycerides.
A recent study reported the effect of periodontal treatment on changes in the carotid.
The observations indicated that changes in IMC thickness after periodontal treatment are possible in systemically healthy individuals 9 .Future intervention studies are necessary to better characterize the role of periodontal disease and its products on clinical and subclinical cardiovascular events.
Based on the current study's findings, we conclude that different measurements for the severity of periodontal disease and number of teeth were associated with carotid thickness.In addition, individuals with 10% or more sites with CAL ≥ 4mm showed higher odds of altered carotid thickness after controlling for confounding factors.
Aterosclerose; Doença das Artérias Carótidas; Periodontia Contributors R. M. Batista participated in the study conceptualization and design, data collection, analysis, and interpretation, and writing of the article and was responsible for the study as a whole.E. P. Rosetti collaborated in the study conceptualization and design, data analysis and interpretation, writing of the article, critical revision of the text, and final approval of the article.E. Zandonade contributed to the study conceptualization and design, data analysis and interpretation, statistical analysis, writing of the article, critical revision of the text, and final approval.L. H. Roelke participated in the data collection and critical revision of the text.M. V. Vettore contributed to the data analysis and interpretation and critical revision of the text.A. E. Oliveira collaborated in the study conceptualization and design, writing of the article, and critical revision of the text.

Table 1
Participants' characteristics: variables and covariables of the 220 adults examined in the study.
* Mean and standard deviation.

Table 2
Risk fact ors of the 220 adults according to carotid intima-media complex (IMC) group.

Table 3
Periodontal clinical parameters according to normal versus altered carotid intima-media complex (IMC).

Table 5
Crude and adjusted odds ratios (OR) and respective 95% confi dence intervals (95%CI) for relationship between altered carotid intima-media complex (IMC) and periodontal disease, age, gender, hypertension, diabetes, and body mass index (BMI).

Table 6 Results
of multiple logistic regression analysis, crude and adjusted odds ratios (OR) with respective 95% confi dence intervals (95%CI) for relationship between altered carotid artery intima-media complex (IMC) (dependent variable) and periodontal disease, age, gender, hypertension, diabetes, and body mass index (BMI) (independent variables) for the best observed model (defi nition 5).* p < 0.05 in the adjusted analysis.Note: defi nition 5: ≥ 10% of sites with clinical attachment level (CAL) ≥ 4mm.