Association Between Metabolic Syndrome and Periodontitis: A Systematic Review and Meta-analysis

Background:

Several epidemiological studies have reported an association between metabolic syndrome (MetS) and periodontal diseases (PDs). The aim of this systematic review was to investigate the existence and magnitude of this association.

Materials and Methods:

A systematic search of the literature was conducted looking for case-control, cross-sectional, cohort studies and population surveys including patients with measures of MetS and PD. Ovid MEDLINE, EMBASE, LILACS, and Cochrane library databases were used for the search by 2 independent reviewers. A meta-analysis was conducted to investigate the association for coexistence of MetS and PD.

Results:

A total of 20 studies were included in the review, from an initial search of 3486 titles. Only 1 study reported longitudinal data on the onset of MetS components in association with periodontal measures. However, several studies investigated coexistence. A random effects meta-analysis showed that the presence of MetS is associated with the presence of periodontitis in a total of 36 337 subjects (odds ratio = 1.71; 95% confidence interval = 1.42 to 2.03). When only studies with “secure” diagnoses were included (n = 16 405), the magnitude of association increased (odds ratio = 2.09; 95% confidence interval = 1.28 to 3.44). Moderate heterogeneity was detected (I² = 53.6%; P = .004).

Conclusions:

This review presents clear evidence for an association between MetS and periodontitis. The direction of the association and factors influencing it should be investigated by longitudinal and treatment studies. Periodontal diagnostic procedures should be routinely carried out in MetS patients.

Obesity, insulin resistance, hypertension, and atherogenic dyslipidemia seem to cluster in the same group of individuals and are now described as part of a condition termed metabolic syndrome (MetS), associated with a strong risk of developing diabetes and cardiovascular events (1). The MetS has an estimated prevalence of 17–32% in the general population, representing a considerable public health concern (2).

Epidemiological evidence accumulated over the last 2 decades suggests that periodontal diseases (PDs) are associated with dyslipidemia, glucose intolerance, hypertension, and a low-grade systemic inflammatory state (3–7) and with systemic diseases and conditions such as cardiovascular disease, diabetes, and obesity (8–10). Associations between periodontitis and MetS have been suggested prevalently by cross-sectional and longitudinal investigations (11–13). However, contradictory reports have also been published, where no association between PD and MetS was observed (14, 15). The supposed link between PD and systemic diseases (MetS in particular) may be related to shared risk factors including predisposing genetic variants (common susceptibility); health behavior factors such as tobacco smoking, diet, and physical exercise; and socioeconomic factors. However, pathogenic factors such as systemic release of periodontal bacteria (16), release of inflammatory cytokines by periodontal tissues or adipose tissues (17), oxidative stress (18), proatherogenic lipoproteins (19), abdominal obesity (20), and cross-reactivity and molecular mimicry (21) may also be responsible for these associations.

Overall, the magnitude of the association between PD and MetS has not been clearly estimated. Furthermore, to what extent periodontitis might contribute to the onset and progression of MetS or vice versa has not been systematically assessed. The aims of this systematic review were to establish the existence and magnitude of a possible association between the presence of periodontitis and the presence of MetS.

Materials and Methods

A PRISMA diagram is attached to follow the reporting of this systematic review (see Supplemental Data, published on The Endocrine Society's Journals Online web site at http://jcem.endojournals.org).

Research questions

A detailed protocol was written prior to starting this systematic review. In order not to introduce bias on the direction of causality between PD and MetS, we set the following research questions:

• •. 

  Broad question: is there any association between PDs and MetS?

• •. 

  Specific questions: 1) are patients with MetS more likely to have PD (compared to patients with no MetS); and 2) are patients with PD more likely to have MetS (compared to patients with no PD)?

The review followed the outlines of PECO:

• •. 

  Population: the population to be included in the review consisted of subjects diagnosed with PD or MetS and healthy controls.

• •. 

  Exposure: presence of MetS (or presence of PD) (depending on specific question 1 or 2 above).

• •. 

  Comparisons: subjects with MetS and no PD; subjects with PD and no MetS (depending on specific question 1 or 2 above).

• •. 

  Outcomes: any measure of PD diagnosis and/or extent and severity and MetS diagnosis and/or extent and severity.

Study selection

Published articles were considered eligible for this review according to the following characteristics:

• •. 

  Case-control, cross-sectional, cohort studies, and population surveys in humans.

• •. 

  Including subjects with measures of MetS and PDs and/or controls.

Search strategy

The search was conducted through the electronic databases MedLine, EMBASE, LILACS, Cochrane library, and unpublished databases, complemented by a search through the reference lists of included studies. No language restriction was included in the search. Among published literature, peer-reviewed studies, reports, book chapters, conference abstracts, and theses were screened. Narrative reviews on the topic were searched in order to identify suitable papers. Soon-to-be-published studies were sought by contacting editors from the journals with the highest impact factor related to our search (dental, metabolic, cardiovascular). In particular, we asked the editors if they were aware of papers in press with the keywords “periodontal disease” and “metabolic syndrome.” However, this strategy identified no such study. The search was updated on May 30, 2012.

The search strategy included the following search words:

• •. 

  MeSH terms in all trees/subheadings: “periodontal diseases” and “insulin resistance.”

• •. 

  Keywords for periodontal disease: “tooth loss,” “alveolar bone loss,” “periodont*,” and “gingiv*.”

• •. 

  Keywords for metabolic syndrome: “metabolic syndrome,” “syndrome X,” “obesity,” “hypertension,” “diabetes mellitus,” “insulin resistance,” “hypertriglyceridemia,” “hyperlipidemia,” “hypercholesterolemia,” “dyslipidemia,” “hyperglycemia,” and “hyperinsulinism.”

Study selection was conducted by 2 independent reviewers (L.N. and N.T.) in the following stages: 1) initial screening of potentially suitable titles and abstracts against the inclusion criteria to identify potentially relevant papers; and 2) screening of the full papers identified as possibly relevant in the initial screening. The help of translators was sought for studies not in English. After the screening of titles and abstracts, the studies included by both reviewers were compared, and a complete database was formed joining all studies selected by at least 1 reviewer. In case of a disagreement between reviewers, the decision about study eligibility was made by a consensus between the 2 reviewers. The inter-reviewer agreement was calculated by means of κ statistics after steps 1 and 2 described above.

Assessment of risk of bias

We aimed to assess precision, directness, applicability, and risk of bias of the included studies. The risk of bias was defined as a systematic error or deviation from the truth in results or inferences (22), whereas internal validity was defined as the extent to which the results of a study are correct for the circumstances being studied (23). The following issues were investigated: appropriateness of study design to the research objective, conduct of the study/internal validity, assessment of exposure, choice of outcome measure, reliability of diagnostic tests and criteria, statistical issues, random error, quality and completeness of reporting, external validity or applicability, and conflict of interest in the conduct of the study (24, 25).

Definitions of PDs and MetS

Owing to the lack of a general consensus for a diagnosis of PD and MetS and to the use of different definitions in the searched articles, we set the following diagnostic thresholds as part of an assessment of methodological quality and external validity:

1. Diagnosis of periodontitis

a. Secure periodontitis

• •. 

  At least 2 sites on different teeth with clinical attachment level (CAL) ≥6 mm and at least 1 site with probing pocket depth (PPD) ≥4 mm (26), or

• •. 

  At least 2 sites in nonadjacent teeth with proximal attachment loss of ≥3 mm (27), or

• •. 

  Community periodontal index (CPI) score of 4 in at least 1 quadrant (28)

• •. 

  In cases where no CAL or PPD is reported, radiographic alveolar bone loss ≥30% of root length or ≥5 mm in at least 2 teeth.

b. Insecure periodontitis

• i. 

  At least 2 sites on different teeth with periodontal CAL ≥4 mm or 1 site with PPD ≥4 mm (26); or

• ii. 

  CPI score 3 in at least 1 quadrant (28).

• iii. 

  “Alveolar bone loss” (not clearly defined or less than definition above).

• iv. 

  Unclear diagnostic criteria for periodontitis.

2. Diagnosis of gingivitis

• a. 

  Secure: at least 30% of sites with bleeding on probing or mean bleeding index = 1 (29) or at least 15 bleeding sites (30) but “periodontitis” excluded from participant selection as described in “Secure periodontitis” above.

• b. 

  Insecure: unspecified gingival inflammation (periodontitis excluded but not employing secure criteria above).

3. Diagnosis of MetS

a. Secure

• •. 

  Presence of any 3 of the following 5 risk factors: 1) elevated waist circumference (≥88 cm in women and ≥102 cm in men); 2) elevated fasting triglycerides (≥150 mg/dL or on drug treatment for elevated triglycerides); 3) reduced high-density lipoprotein(HDL)-cholesterol (<50 mg/dL in women and <40 mg/dL in men or on drug treatment for reduced HDL-cholesterol); 4) elevated blood pressure (BP) (systolic BP ≥130 mm Hg or diastolic BP ≥85 mm Hg, or on antihypertensive drug treatment in a patient with a history of hypertension); and 5) elevated fasting glucose (≥100 mg/dL or on drug treatment for elevated glucose) (31).

• •. 

  Presence of the following risk factors: 1) central obesity (body mass index >30 kg/m²) or waist circumference ≥94 cm for men and ≥80 cm for women; 2) raised triglycerides, >150 mg/dL (1.7 mmol/L), or specific treatment for this lipid abnormality; 3) reduced HDL cholesterol, <40 mg/dL (1.03 mmol/L) in males, <50 mg/dL (1.29 mmol/L) in females, or specific treatment for this lipid abnormality; 4) raised BP, systolic BP >130 mm Hg or diastolic BP >85 mm Hg, or treatment of previously diagnosed hypertension; and 5) raised fasting plasma glucose, >100 mg/dL (5.6 mmol/L), or previously diagnosed type 2 diabetes (32).

b. Insecure: insecure/unclear diagnosis of MetS (criteria not specified)

Data extraction

Two reviewers (L.N., N.T.) independently extracted the necessary data using a standardized form, and discrepancies were resolved by consensus. The following data were sought from each study: first author, journal, year of publication, source of funding, geographical setting, study design, ethnicity of participants, average age, number of smokers, number of patients with MetS, patients' demographics, definitions of MetS and PD, number of subjects affected by MetS and/or PD, number of healthy subjects (or number of MetS subjects with no PD or PD subjects with no MetS, depending on the studies), study outcomes, and odds ratios (ORs) for associations between PD and MetS. Data relative to risk of bias were extracted using the Newcastle Ottawa tool for case-control and cohort studies (25). In terms of study exclusion and inclusion, the κ scores for the 2 reviewers' agreement were 85% at the first step for screening titles and abstracts and 100% at the second step for screening full-text articles.

Statistical analysis

The outcome used in this meta-analysis was OR for MetS among people with or without a diagnosis of PD (and vice versa). We chose adjusted ORs with the adjustment of confounding variables, eg, ages, gender, socioeconomic status, etc., in multivariable analyses with MetS as the dependent variable and PD diagnosis as the risk factor. When the adjusted ORs were not reported, we used or derived unadjusted, crude OR using the number of patients with the diagnosis of MetS in patients with or without PDs. Preliminary results indicated moderate heterogeneity (P value for the χ² test = .004; I² = 53.6), so random effects meta-analyses were undertaken throughout this review. One study (20) reported ORs for men and women separately and therefore was treated as 2 studies. The inverses of variances for the log ORs were used as weights in the random effects analysis, and a stratified analysis was then undertaken according to whether the definitions of PD and MetS diagnosis were secure or not. Small study bias was examined using funnel plot and Egger's test. One study was omitted each time in a series of meta-analyses to identify influential studies. Meta-regression was undertaken to explore the impact of risk of bias of included studies, such as types of study design and quality of studies. All the statistical analyses were undertaken by use of the statistical software package STATA (version 12.1; StataCorp, College Station, Texas).

Results

From the initial 3486 titles screened, 40 were considered potentially suitable by at least 1 reviewer after initial screening (Figure 1). After full-text screening, 20 papers were excluded for the following reasons: 1) not reporting outcomes of interest (n = 12); 2) duplicate reports (n = 3); 3) reviews (n = 3); and 4) focused on medical subgroups (n = 2). Therefore, a total of 20 papers were included. Table 1 reports the main characteristics of the studies that were included. The 20 papers (11–15, 20, 33–46) included had been conducted in 12 different countries across Europe, Asia, North America, and South America. Fourteen were written in English, with the remaining 6 written in Spanish, Portuguese, French, Polish, Russian, and Chinese. The subject sample ranged from 45 patients in the study by Pozharitskaia et al (33) to 62 254 in the French survey by Guize et al (34) (median number of subjects, 803.5; interquartile range, 241 to 2157). One of the studies, Morita et al (12), was longitudinal, whereas all other studies were either case-control or cross-sectional surveys. Periodontitis was investigated in 19 of the studies, whereas one (34) only assessed gingivitis. The definition of periodontitis ranged from radiographic criteria (41) to clinical criteria based on CPI, diagnostic criteria reported by Page and Eke (26), Russell index (47), or arbitrary criteria based on PPD or CAL thresholds for number of affected sites or on full mouth PPD or CAL averages. No study had differentially diagnosed aggressive or chronic periodontitis. A range of 6.3 to 78.8% of subjects was classified by the authors as having periodontitis. The diagnosis of MetS was generally based on Adult Treatment Panel III (ATP III), National Cholesterol Education Program (NCEP) criteria, or those reported by the International Diabetes Federation (IDF). The percentage of subjects classified as having MetS across the included studies varied from 5.0 to 78.9%.

Figure 1.

Flow of reviewed studies from initial title and abstract screening (n = 3486) to final inclusion (n = 20).

Table 1.

Main Characteristics of Suitable Studies Included in the Review

First Author, Year (Ref.)	Country	Design	Sample Size (n)	PD Diagnosis	Subjects Diagnosed With PD (%)	MetS Diagnosis	Subjects Diagnosed With MetS (%)
Pozharitskaia, 2004 (33)	Russia	Case-control	45	Russell index	13.0	Unclear	66.7
Guize, 2006 (34)	France	Cross-sectional	62 254	Not reported	Not reported	ATP III	10.4
Borges, 2007 (15)	Brazil	Cross-sectional	1315	CPI codes 3–4	14.6	NCEP 2001	54.3
Shimazaki, 2007 (35)	Japan	Cross-sectional	584	Average CAL ≥3 mm	6.3	NCEP 2001	16.8
D'Aiuto, 2008 (11)	United States	Cross-sectional	13 677	Page & Eke 2007	14.0 (moderate to severe)	AHA 2005	Not reported
Khader, 2008 (36)	Jordan	Case-control	156	No categorical definition applied	Not reported	ATP III	50.0
Li, 2009 (37)	China	Case-control	192	>33% sites ≥3 mm CAL	72.4	IDF 2005	72.9
Kushiyama, 2009 (38)	Japan	Cross-sectional	1070	CPI code 4	29.5	NCEP 2001	5.0
Kowalski, 2009 (39)	Poland	Case-control	380	Russell Index	20.5 (severe)	IDF 2005	78.9
Morita, 2009 (40)	Japan	Cross-sectional	2478	CPI code 3–4	25.9	JSIM	8.2
Morita, 2010 (12)	Japan	Longitudinal	1023	CPI code 3–4	20.0	JSIM	0% at baseline
Nesbitt, 2010 (41)	United States	Cross-sectional	200	Radiographic distance CEJ-crestal bone ≥3 mm	21.5 (moderate to severe)	Modified ATP III	17.5
Benguigui, 2010 (14)	France	Cross-sectional	255	Page & Eke 2007	78.8 (moderate to severe)	ATP III	28.6
Han, 2010 (42)	Korea	Cross-sectional	1046	CPI code 3–4	34.0	IDF 2009	22.4
Timonen, 2010 (43)	Finland	Cross-sectional	2050	No categorical definition applied	Not reported	EGIR 2002	16.4
Andriankaya, 2010 (20)	United States	Cross-sectional	7431	Average PD, ≥2.5 mm	5.8	ATP III	19.7
Bensley, 2011 (44)	United States	Cross-sectional	481	Self-reported	27.0	AHA 2009	63.9
Kwon, 2011 (13)	Korea	Cross-sectional	6520	CPI code 3–4	32.9	ATP III	28.4
Bracho, 2011 (45)	Venezuela	Cross-sectional	297	CPI code 3–4	22.2	IDF 2009	43.8
Li, 2011 (46)	China	Cross-sectional	59	PD ≥5 mm, CAL ≥3 mm	57.6 (moderate to severe)	IDF 2005	44.1

Abbreviations: AHA, American Heart Association; EGIR, European Group for the Study of Insulin Resistance; JSIM, JAMA, Japanese Society of Internal Medicine; CEJ, cemento-enamel junction. Please note that for some studies several different diagnostic criteria for periodontitis were applied. Only one for each study is reported in this table. Some studies only reported continuous periodontal measures but no categorical subject-based data.

Study quality as assessed by the Newcastle Ottawa scale varied considerably across the studies, ranging from a score of 0/9 to 8/9, the main issues being lack of adequate diagnosis, definition of controls, and lack of adjustment for potential confounders (Table 2).

Table 2.

Study Quality as Assessed by the Newcastle Ottawa Scale as Judged by the 2 Reviewers Who Performed Data Extraction

First author, Year (Ref.)	Selection (Maximum, 4 Asterisks)	Comparability (Maximum, 2 Asterisks)	Exposure (Maximum, 3 Asterisks)
Pozharitskaia, 2004 (33)	*	*
Guize, 2006 (34)	**		*
Borges, 2007 (15)	*	*	**
Shimazaki, 2007 (35)	**	**	**
D'Aiuto, 2008 (11)	****	**	**
Khader, 2008 (36)		*	**
Li, 2009 (37)	**	*	**
Kushiyama, 2009 (38)	*	**	**
Kowalski, 2009 (39)
Morita, 2009 (40)	**	*	**
Morita, 2010 (12)	**	**	**
Nesbitt, 2010 (41)	**	*	*
Benguigui, 2010 (14)	****	**	**
Han, 2010 (42)	***	**	**
Timonen, 2010 (43)	***	*	*
Andriankaya, 2010 (20)	***	**	**
Bensley, 2011 (44)	**	*	**
Kwon, 2011 (13)	**	**	**
Bracho, 2011 (45)		*	*
Li, 2011 (46)	*	**	**

The meta-analysis based on cross-sectional and case-control studies revealed a positive association between coexistence of periodontitis and MetS (OR = 1.71; 95% confidence interval [CI] = 1.42 to 2.03) (Figure 2). Moderate heterogeneity was detected (I² = 53.6%; P = .004). Subanalyses by study design revealed higher heterogeneity for cross-sectional than for case-control studies (I² = 45.7 vs 0%). However, these latter studies had much smaller sample sizes and very wide CIs and accounted for only 1.5% of the total population in the meta-analysis. The pooled OR from the 3 small case-control studies was 14.57 (95% CI = 3.68 to 57.75), which was much higher than the pooled OR of 1.65 (95% CI = 1.42 to 1.92) from the 15 cross-sectional studies. Subanalyses for studies divided by “insecure” diagnoses and “secure” diagnoses confirmed these associations, showing a higher magnitude association when only secure diagnoses studies were considered (OR = 2.10; 95% CI = 1.28 to 3.44 for secure, vs OR = 1.67; 95% CI = 1.40 to 2.00, for insecure) (Figure 3).

Figure 2.

Forest plot for association between periodontitis and MetS (random effects model). The sizes of the gray boxes in the figure are in proportion to the weight assigned to each study, whereas the point within each box is the OR reported by each study. The horizontal bar is the reported 95% CI for each OR.

Figure 3.

Forest plot for association between periodontitis and MetS under the random effects meta-analysis for studies with secure (under the yes heading) and insecure diagnoses (under the no heading).

Egger's test shows that there might be small study bias (P = .081). Funnel plot (Supplemental Material, published on The Endocrine Society's Journals Online web site at http://jcem.endojournals.org) suggested that small studies tend to report greater ORs for MetS. Meta-regression showed that the risk of bias scores was not a significant factor for explaining the heterogeneity (−0.054; P = .363).

Discussion

The MetS is a common medical condition associated with high risk of developing diabetes and cardiovascular disease (1) and therefore with an unquestionable public health impact. This systematic review showed that subjects affected by MetS are nearly twice as likely to have periodontitis than the rest of the population. A very crude measure derived from this review is that nearly 40% (n = 1847 of 4798 total number of subjects with MetS included in the study) of the total MetS population included in the reviewed studies was diagnosed with periodontitis. This brings unequivocal evidence that PDs are a common feature of MetS.

Heterogeneity in periodontal and MetS diagnostic criteria emerged clearly from this review. In particular, a whole variety of diagnoses were used to define “periodontitis.” Some authors used well-defined criteria, including mainly the CPI as suggested by the World Health Organization or those suggested by Page and Eke (26) or Russell (47). However, certain studies used self-defined criteria based on thresholds of clinical data (eg, average PPD, percentage of sites with a certain CAL threshold, radiographic bone loss) or even self-reported information. This further testifies for the difficulties in having a universally accepted clinical case definition for periodontitis (48, 49). A rather more secure definition of MetS was adopted by most studies, using specific guidelines such as ATP III or IDF, with a few exceptions (as reported in Table 1). We included all studies with described definitions of “periodontitis” and “metabolic syndrome” but then performed a 2-threshold analysis, restricting the second analysis only to studies with predefined secure diagnostic criteria. In assessing this, we have to acknowledge that even our secure criteria were arbitrarily set (using high-threshold existing definitions). Interestingly, both analyses revealed a positive association between coexistence of the 2 conditions, with a higher magnitude for studies with a secure diagnosis. This supports the need for a reliable and secure diagnosis of these pathological entities in clinical research. No meta-analysis could be performed for the longitudinal associations between periodontitis and MetS, for a paucity of studies (12). A very limited number of studies reported gingivitis outcomes outside the context of periodontitis (eg, reporting data on gingival inflammation, such as gingival index, rather than simply CPI codes), showing an increased prevalence of gingival inflammation in subjects affected by MetS compared to healthy subjects (34, 39). Because of the paucity of studies investigating this association and the lack of a secure diagnosis of gingivitis, no meta-analysis was performed for this outcome. Among the other limitations of this review, we need to acknowledge the lack of adjusted ORs presented by some studies, which leaves potential for residual confounding in the meta-analysis. Furthermore, there still remains unexplained heterogeneity, which casts doubts on the reliability of the overall summary estimate.

The magnitude of association between PD and MetS emerging from this review (OR = 1.7 to 2.1) is in keeping with the described associations between PD and the separate components of MetS, namely obesity, hypercholesterolemia, insulin resistance, and hypertension. For example, a recent systematic review estimated an OR of 1.3 for the presence of periodontitis in overweight subjects and an OR of 1.8 in obese subjects (10). Most studies included in this review reported an apparent gradient effect for association with periodontitis in subjects with increasing number of positive MetS components (20, 38, 40–44). However, central obesity and glucose levels (OR = 1.5 and 1.7, respectively) both showed a higher association with PD than MetS (OR = 1.4) in subjects from the National Health and Nutrition Examination Survey (11). In a French population, HDL levels showed a higher association with PD than did MetS (14). Therefore, the question remains open on whether, from a biological standpoint, the presence of MetS rather than the sum of its individual components may have an additive effect on the presence of periodontitis. The biological rationale for the described association involves microbiological, inflammatory components as well as oxidative stress (18), proatherogenic lipoproteins (19), abdominal obesity (20), and cross-reactivity and molecular mimicry (21).

Indeed, questions remain on the directions of the PD-MetS associations and therefore on the possibility to prevent PD with an effective management of MetS and vice versa. A longitudinal study on 2973 nondiabetic individuals revealed an association between periodontitis and glycated hemoglobin increase over time (50). Initial evidence on MetS components comes from the longitudinal study by Morita et al (12) on 1023 Japanese subjects followed for 4 years, which showed that the presence of periodontal pockets was associated with a positive conversion of 1 or more metabolic components, suggesting that prevention and treatment of periodontitis may reduce the risk of developing MetS. A pilot treatment study in patients with PD and MetS showed a reduction in systemic levels of C-reactive protein, triglycerides, and leukocyte counts and an increase in HDL after successful periodontal treatment (51). A recent randomized controlled trial of 9-month duration conducted on 165 patients affected by PD and MetS revealed that reduction of periodontal inflammation by periodontal treatment (with or without the use of systemic antibiotics) was associated with reduction of systemic C-reactive protein levels (52). Periodontitis may have direct quality-of-life relevance as well as a potential copathogenic effect in patients with MetS. Although large-scale treatment studies need to be conducted to assess the effect of periodontal treatment in MetS, this review suggests that periodontitis is one of its common components, and as such oral health assessments and periodontal diagnosis (PPDs, periodontal attachment level, and bleeding on probing charts) should form part of the routine diagnostic procedures in MetS patients.

Acknowledgments

We acknowledge Dr Manar Aljateeli, Dr Linda Sun Liu, and Mrs. Agnieszka Lamb for their help with the translation of articles.

This study was supported by the Periodontal Research Fund of the University College London (UCL) Eastman Dental Institute, and it was undertaken at UCL, which received a proportion of funding from the Department of Health's National Institute of Health Research Biomedical Research Centres funding scheme. F.D. holds a Clinical Senior Lectureship Award supported by the UK Clinical Research Collaboration.

Disclosure Summary: The authors have nothing to declare.

Abbreviations

 
• BP

  blood pressure

 
• CAL

  clinical attachment level

 
• CI

  confidence interval

 
• CPI

  community periodontal index

 
• HDL

  high-density lipoprotein

 
• MetS

  metabolic syndrome

 
• OR

  odds ratio

 
• PD

  periodontal disease

 
• PPD

  probing pocket depth.