Association between overweight/obesity and periodontitis in adults. A systematic review


Jean Suvan, Unit of Periodontology, University College London (UCL), Eastman Dental Institute, 256 Gray's Inn Road, London, WC1X 8LD, UK. E-mail:


Overweight and obesity have been suggested to be associated with periodontitis as published in studies and narrative summaries. This project presents results of a systematic review investigating the association between overweight or obesity (as defined by the World Health Organization) and periodontitis.

Search strategy included electronic and hand searching to December 2009. Ovid MEDLINE, EMBASE, LILACS, and SIGLE were searched. RCTs, cohort, case–control and cross-sectional study designs that included measures of periodontitis and body composition were eligible. Duplicate, independent screening and data abstraction were performed. Meta-analyses were performed when appropriate.

A total of 526 titles and abstracts were screened, resulting in 61 full text articles and abstracts assessed for eligibility with 33 being included. Nineteen studies provided sufficient information for inclusion in meta-analyses. Meta-analyses indicated statistically significant associations between periodontitis and body mass index (BMI) category obese OR 1.81(1.42, 2.30), overweight OR 1.27(1.06, 1.51) and obese and overweight combined OR 2.13(1.40, 3.26).

In conclusion, these results support an association between BMI overweight and obesity and periodontitis although the magnitude is unclear. Additional prospective studies to further quantify, or understand the mechanisms, of this association are merited. There is insufficient evidence to provide guidelines to clinicians on the clinical management of periodontitis in overweight and obese individuals.


Chronic non-communicable diseases have overtaken communicable diseases as leading health problems in most parts of the world (1). Recently published evidence has suggested an association between obesity and periodontitis; two diseases classed by the World Health Organization (WHO) as chronic non-communicable diseases and best addressed via a common risk factor approach due to the multiple factors implicated in their onset and progression (2).

Periodontitis, a chronic inflammatory disease initiated by dental plaque biofilm and perpetuated by a deregulated immune response, is considered to be one of the two most important global oral health burdens (3). Its reported prevalence ranges from 20% to 50% in the general population (4,5).

More recently, an emerging epidemic public health problem has occurred as overweight and obesity prevalence has trebled since the 1980s, especially in developed countries (6). Overweight and obesity are defined as abnormal or excessive fat accumulation that represents a risk to general health (7). Furthermore, increasing knowledge of the active role fat cells play in the regulation of inflammation and immunity has labelled obesity as a chronic disease (8). Indeed, obesity's direct adverse effects on general health have been identified including altered blood pressure, insulin resistance, dyslipidemia and a state of low-grade inflammation (9,10). In addition, numerous other comorbidities are currently being investigated including diabetes, coronary artery disease, stroke, respiratory disorders, cancers, osteoarthritis, liver and gall bladder disease (10).

Based upon the hypothesis that systemic inflammation associated with obesity may affect susceptibility to chronic infectious diseases, the question of an association between obesity and periodontitis has arisen (9). An increasing body of literature investigating this association has been published over the last 10 years, including various study designs providing associated levels of evidence (11–19). Numerous narrative reviews exist (20–26); however, non-standardized methods and subjective approaches render these prone to bias (27). No previous existing systematic reviews of this topic were identified despite extensive searching of the literature described below.

The aim of this study was to provide a systematic review of the evidence to date investigating the association between overweight/obesity (as defined by body mass index – BMI) and periodontitis (6). In addition, a secondary objective was to summarize measures used to assess body composition and periodontitis in the included investigations to facilitate interpretation of current evidence and guide future study design. The systematic review question was: ‘What is the relationship between overweight or obesity (based on BMI) associated with excessive body fat and periodontitis in terms of clinical periodontal outcomes in adults?’

Materials and methods


The focus of the review was to provide a comprehensive summary of the existing evidence of an association between exposure and disease therefore inclusion criteria were set to be broad and inclusive. All study designs were included with the exception of case studies and case series.

Body composition was considered to be the exposure. For this review, body composition was defined to include any measure of body weight (defined as weight of a person's body) and size directed towards determining the ‘make up of lean and fat tissue in the body’(28). This included such measures as BMI, waist circumference, waist hip ratio, and body fat percent and others as reported. Exposure was not limited to obesity (defined as the chronic condition characterized by an excess amount of body fat). Periodontitis was defined to include all forms of periodontitis excluding gingivitis, periodontitis associated with specific syndromes or that manifest in children and adolescents. Therefore, eligibility criteria were set to include individuals from age 18 years onward. All oral health outcomes measures reported were eligible for inclusion.

Search and screening

Preliminary searches designed to locate possible review articles, narrative or systematic were conducted to facilitate development of the electronic search strategy. The strategy was developed using MeSH and free text terms then piloted to confirm high sensitivity over high precision in search results. The strategy was customized as appropriate before application to each database. Electronic searching included the search of electronic databases to December 2009 using terms and strategy set a priori according to each database (Ovid MEDLINE, EMBASE, LILACS and SIGLE). No language or year restrictions were applied. An example of search strategy applied is found in Table 1.

Table 1.  Search strategy example for Ovid MEDLINE
Search strategy
 2. periodont$.mp.
 3. gingiv$.mp.
 4. (gingiv$ adj health).mp.
 5. (ging$ adj disease).mp.
 6. (inflamm$ adj gingiv$).mp.
 9. 1 or 2 or 3 or 4 or 5 or 6 or 7
10. 8 and 9
12. Obes$.mp.
13. Body
14. Body
15. Body
17. adipos$.mp.
19. over
22. binge eating
23. weight
24. weight
25. fat overload
26. body mass
27. waist
28. waist
29. (weight adj over).mp.
30. (weight adj reduc$).mp.
31. (weight adj reduc$).mp.
32. (weight adj loss).mp.
33. (weight adj maint$).mp.
34. (weight adj decreas$).mp.
35. (weight adj watch$).mp.
36. (weight adj control).mp.
38. body weight [mp = ti, ot, ab, nm, hw]
40. metabol$.mp.
41. metabolic
42. 11 or 12 or 13 or 14 or 15 or 16 or 17 or 18 or 19 or 20 or 21 or 22 or 23 or 24 or 25 or 26 or 27 or 28 or 29 or 30 or 31 or 32 or 33 or 34 or 35 or 36 or 37 or 38 or 39 or 40 or 41
43. 8 and 42
44. 10 and 43

Hand searching was comprised of checking bibliographic references of included articles and related review articles. In addition, online hand searching of issues within the past 10 years of key periodontal and obesity journals was performed (JClinPerio, JPerio, JDentRes, Obes Rev, Obesity, Int J Obes, Diabetes Obes Metab).

Titles and abstracts (when available) of all reports identified through the search were scanned by two reviewers independently (JS and FD). Full reports were obtained and reviewed independently for studies appearing to meet the inclusion criteria or for which there was insufficient information in the title and abstract to allow a clear decision (JS and FD).

Bias-protection assessment

Bias-protection assessment of included studies was undertaken independently and in duplicate by two reviewers. Cohort and case–control studies were assessed using the validated Newcastle-Ottawa Quality Assessment Scale (29) as recommended by the Cochrane Collaboration Guidelines for the assessment of non-randomized studies (30). These tools award stars (*) in three categories for each study based on incorporation of design elements associated with minimising bias.

A validated assessment tool to assess cross-sectional studies was not located. The quality of cross-sectional studies were assessed using questions from the Newcastle-Ottawa Quality Assessment Scale for cohort study design that were deemed applicable to the cross-sectional study design (29).

Data abstraction

Data abstraction forms were developed to collect data outlined in the systematic review protocol. The forms were piloted on five studies by both investigators involved in data abstraction. Following comparison and discussion of the five studies, the forms were revised to facilitate accuracy and clarity in collection of data. Data abstraction was then performed on all included studies independently and in duplicate (JS and FD). Completed forms were compared to validate accuracy of the data abstraction. Disagreements were resolved through discussion with a third person (DM).

Data synthesis

Descriptive methods

Descriptive summary was performed by summarizing the studies in evidence tables to determine the quantity of data, checking further for variations in study characteristics (populations, exposures, outcomes, design, quality and results). Bias-protection assessment was also summarized in table format. Evidence tables provided the framework to assess data suitability for further quantitative analyses such as meta-analysis.

Quantitative methods

Stata 10 statistical software (StataCorp) was used for statistical analyses. Pooled estimates of odds ratios and corresponding 95% confidence intervals, were calculated for dichotomous data. Studies were grouped for meta-analysis according to the outcome measure reported. Where reported, adjusted odds ratios in the different studies were used to perform the meta-analyses. Three meta-analyses were performed comparing obese with non-obese, overweight with non-overweight and overweight/obese combined compared with non-overweight/obese in relation to the associated odds of periodontitis. There were insufficient study numbers to perform sub-group analyses according to periodontal assessment parameter within these groups. Pooled mean differences and 95% confidence intervals were planned for continuous data; however, these data were not sufficiently similar to warrant performing a meta-analysis.


Pooled estimates of the odds ratios were calculated using random effects models rather than fixed effects models due to the high levels of heterogeneity amongst the study data as confirmed in descriptive methods (see evidence Tables 2–4).

Table 2.  Characteristics of included cohort studies
Author (country)Population/SampleBody Composition (exposure)Periodontal outcome (dependent variables)Covariates included for adjustmentPublication conclusionsEffect OR (95% CI) or other as stated
  1. BMI, body mass index; CAL, clinical attachment level; CI, confidence interval; NS, not significant; OR, odds ratio; PPD, pocket probing depth; WHO, World Health Organization.

Linden et al. 2007 (62) (Northern Ireland, UK)Subset of Prospective Epidemiological Study of Myocardial Infarction (PRIME) with periodontal assessment at 10 year visit
n = 1362 men
59–72 years
WHO categories:
Underweight < 18.5 kg m−2
 Normal 18.5–24.9 (reference)
 Overweight 25–29.9
 Obese ≥ 30
PPD and CAL 4 sites/ tooth
Categorized as
 Low threshold periodontitis – minimum 2 teeth with ≥ 6 mm attachment loss and minimum 1 site with PPD ≥5 mm
High threshold periodontitis –≥15% of all sites showed attachment loss ≥ 6 mm
Age, smoking, diabetes, education, socioeconomic status, dental attendances, brushing frequencyPrevalence
 Obesity was associated with periodontitis in the homogeneous group of 60–70 years old European men and high BMI levels in early life did not predict periodontitis in later life in the this sample
 Low threshold periodontitis
  Overweight 1.23 (0.88, 1.71)
  Obese 1.77 (1.20, 2.63) P < 0.01
 High threshold periodontitis
  Overweight 1.07 (0.61, 1.88)
  Obese 1.55 (1.20, 2.93)
Change weight since 21
 Low threshold periodontitis
  Overweight 0.99 (0.72, 1.36)
  Obese 1.33 (0.95, 1.86) NS
 High threshold periodontitis
  Overweight 0.88 (0.49, 1.55)
  Obese 1.65 (0.95, 2.87) NS
Table 3.  Characteristics of included case–control studies
Author (country)Population/SampleBody Composition (exposure)Periodontal outcome (dependent variables)Covariates included for adjustmentPublication conclusionsEffect OR (95% CI) or other as stated
  1. BF, body fat; BMI, body mass index; BOP, bleeding on probing; CAL, clinical attachment level; CI, confidence interval; N/A, no data; OR, odds ratio; PD, pocket depth; PPD, pocket probing depth; SD, standard deviation; WC, waist circumference; WHO, World Health Organization; WHR, waist/hip ratio.

Saito et al. 2008(84) (Japan)Subset of Hisayama prospective cohort study of cardiovascular diseases
n = 76 healthy women with minimum 22 teeth
Categorized as periodontitis cases (34) or periodontally healthy (42)
50–59 years
BF (bio impedance method)
Blood sample to assess resistin and adiponectin levels
PD and CAL recorded two sites per tooth (mesial and mesio-buccal) in two contra-lateral quadrants
Periodontitis defined as minimum 1 tooth with ≥ 6 mm PD and/or minimum 3 teeth with ≥ 4 mm PD
Periodontal health defined as
All teeth with ≤ 3 mm PD and maximum two teeth with bleeding upon probing
Smoking status, agePrevalence
 Serum resistin level was associated with periodontitis in middle-aged Japanese women. Resistin may mediate the triangular relationship that exists among obesity, type 2 diabetes, and periodontal disease
Healthy gingival n = 42
Mean ± SD
 Resistin 8.0 ± 5.2
 Adiponectin 12.4 ± 5.1
 BMI 21.9 ± 3.2
 BF 26.4 ± 6.0
 WC 78.9 ± 10.7
 WHR 0.89 ± 0.06
Periodontitis N = 34
 Resistin 9.9 ± 5.2 P = 0.037
 Adiponectin 11.5 ± 5.0 P = 0.46
 BMI 23.3 ± 2.6 P = 0.013
 BF 28.7 ± 4.6 P = 0.079
 WC 81.0 ± 8.0 P = 0.39
 WHR 0.89 ± 0.07 P = 0.92
Sarlati et al. 2008 (58) (Iran)Sample of young Iranian population attending Department of Nutrition Research, Shahid Beheshti University
n = 80, 40 normal and 40 overweight and obese
18–34 years
WHO categories:
Underweight < 18.5 kg m−2
 Normal 18.5–24.9
 Overweight 25–29.9
 Obese ≥ 30
WHO categories
 Women cut off > 88 cm
 Men cut off > 102 cm
PD and CAL measured at four sites around all teeth (one lingual)
Plaque index (PI) as plaque present or absent (O'Leary)
Age, gender, education, time elapsed since last dental visit, smoking and diabetesSeverity
 Overall and abdominal obesity were associated with the extent of periodontal disease in a sample of Iranian young individuals
Spearman's rank correlation between variables (P < 0.05)
 BMI and PPD 0.33 P = 0.003
 BMI and CAL 0.39 P = 0.000
 WC and PPD 0.323 P = 0.003
 WC and CAL 0.44 P = 0.000
Mean (SD) of periodontal parameters in:
  CAL 1.61 (0.33)
  PPD 2.55 (0.32)
  PI (%) 66.5 (19.7)
  CAL 1.61 (0.33) P = 0.000
  PPD 2.55 (0.32) P = 0.000
  PI (%) 66.5 (19.7) P = 0.054
Gursoy et al. 2006 (55) (Turkey)n = 39 chronic periodontitis patients from university/hospital setting
n = 8 obese, type 2 diabetic
n = 12 non obese type 2 diabetic
n = 8 obese, non diabetic
n = 11 controls non obese, non diabetic
35–65 years
Periodontitis defined as minimum four sites with PPD ≥ 4 mm
 Obesity did not seem to be a significant factor in any parameters evaluated
Obese type 2 diabetic (mean, SD)
 PI 1.86 ± 0.67
 GI 1.99 ± 0.41
 PPD 2.85 ± 0.55 P < 0.05 (compared to control)
Non-obese type 2 diabetic (mean, SD)
 PI.80 ± 0.60
 GI 2.03 ± 0.56
 PPD 2.79 ± 0.73
Obese (mean, SD)
 PI 1.49 ± 0.42
 GI 1.64 ± 0.41
 PPD 2.33 ± 0.46
Control (mean, SD)
 PI 1.60 ± 0.62
 GI 1.77 ± 0.39
 PPD 2.20 ± 0.35
Alabdulkarim et al. 2005 (12) (USA)Subset of University patient records reviewed retrospectively
n = 400
Selected according to BMI
Obese = 200
Non-obese = 200
≥18 years
<25 Non-obese
≥30 Obese
Mean Alveolar Bone Score (ABS)
(Bjorn and Holmberg 1966)
full mouth radiograph evaluation
Periodontitis classed as ABS < 60
Age, gender, smoking, diabetes mellitus, employment, marital status, number of teeth presentPrevalence
 Obesity associated with increased prevalence of periodontitis (measured by radiographic bone loss) especially in younger population
Obese (non obese as reference)
 1.86 (0.99, 3.51) P = 0.054
Obese stratified by age, gender, smoking status (adjusted excluding stratifying variable)
  <40 years 2.67 (1.09, 6.58) P = 0.028
  ≤40 years 1.06 (0.57, 1.95) P = 0.860
  Male 1.95 (1.09, 3.49) P = 0.020
  Female 3.14 (1.64, 6.05) P = 0.000
 Smoking status
  Yes 2.22 (1.19, 4.12) P = 0.011
  No 3.36 (1.74, 6.50) P = 0.000
Socransky and Haffajee 2005(85) (USA)n = 415 systemically healthy subjects
n = 329 chronic perio
n = 86 periodontal health
≥18 years
Three categories:
 Normal < 25
 Overweight 25–30
 Obese ≥ 30
PPD and CAL, PI measured at six sites per tooth
Plaque sample from mesial of each tooth
Age, smoking statusPrevalence
 Obese subjects more likely to exhibit periodontitis, had greater mean PD and percentage sites with plaque and higher proportions of T.forsythia than overweight or normal weight subjects
Proportion of normal , overweight and obese difference in periodontally diseased and healthy groups
 Periodontal health group
  Normal 63%
  Overweight 24 %
  Obese 13%
  % normal significantly higher P < 0.001
 Periodontitis group
  Normal 30%
  Overweight 35 %
  Obese 35%
  % obese significantly higher P < 0.001
Buhlin et al. 2003 (63) (Sweden)n = 96
50 severe periodontitis cases (periodontal specialty practice)
46 periodontally healthy controls (hospital employees)
36–70 years
 Overweight defined as BMI > 26 in men, >25 in women
Full mouth PPD and CAL measurements (six sites/tooth)
  Horizontal loss of supporting tissue by 1/3 or more of the root length with bleeding on probing, furcation involvement or angular bony defects
  No PDs over 5 mm and no clinical attachment loss
Age, gender, smokingPrevalence
 Analysis revealed association between prevalence of periodontitis and BMI
BMI risk for periodontitis (OR)
 4.54 (1.588, 13.00)
Table 4.  Characteristics of included cross-sectional studies
Author (country)Population/SampleBody Composition (exposure)Periodontal outcome (dependent variables)Covariates included for adjustmentPublication conclusionsEffect OR (95% CI) or other as stated
  1. BF, body fat; BMI, body mass index; BOP, bleeding on probing; CAL, clinical attachment level; CI, confidence interval; CPITN, Community Periodontal Index of Treatment Needs; GB, gingival bleeding index; N/A, no data; NHANES III, Third National Health and Nutrition Examination Survey; OR, odds ratio; PD, pocket depth; PI, plaque index; PPD, pocket probing depth; RR; relative risk; SD, standard deviation; SE, standard error; VFA, visceral fat area; WC, waist circumference; WHO, World Health Organization; WHR, waist/hip ratio.

Haffajee and Socransky 2009 (67) (USA)Retrospective analysis of subjects included in previous trials at The Forsyth Institute
n = 695
Periodontally healthy/gingivitis = 121
Chronic periodontitis = 574
18–86 years
WHO categories
Underweight < 18.5 kg m−2
 Normal 18.5–24.9 (reference)
 Overweight 25–29.9
 Obese ≥ 30
Body fat %
Full mouth assessment of plaque, bleeding, PD, and CAL Periodontally healthy/gingivitis defined as minimum 24 teeth present with <2% of sites with PD ≥ 4 mm
Chronic periodontitis defined as minimum 20 teeth present with >5% of sites with PD ≥ 4 mm and/or >5% sites with CAL ≥ 4 mm
 Subjects, particularly younger females who were overweight or obese, were at greater risk for periodontitis than subjects with normal BMI after adjusting for age, gender and smoking
BMI and risk for periodontitis (ref. normal)
 Overweight 3.07 (1.91, 4.84)
 Obese 5.31 (1.41, 1.644)
BMI adjusted for gender and age (ref. overweight and normal)
 Obese 2.31 (1.19, 4.49) P = 0.0086
 Periodontally healthy/gingivitis
 Normal 65.3%
 Overweight 24.0%
 Obese 10.7%
Chronic periodontitis
 Normal 32.9%
 Overweight 37.8%
 Obese 29.3%
Mean BF %
 Healthy 25.0 ± 10.2
 Periodontitis 29.5 ± 10.2
Han et al. 2009 (64) (Korea)Sihwa-Banwol Environmental Health Cohort
n = 1046
15–84 years
(separate data reported on adults 45–54 years)
 Overweight 23–25
 Obese ≥ 25
Waist Circumference
 ≥85 high for women
 ≥90 high for men
 ≥0.85 high for women
 ≥0.90 high for men
Obesity ≥ 100 cm2
PPD scored one code per sextant based on deepest pocket using WHO probe
CPITN Code 0–2 no periodontitis (PPD ≤ 3.5 mm)
CPITN Code 3 or 4 periodontitis
(PPD ≥ 4.0 mm)
Age, gender, monthly family income, smoking, drinking, frequency of daily brushing, physical activityPrevalence
 Obesity was associated with periodontitis
BMI and risk for periodontitis (ref. normal)

 Males 45–54 years (n = 96)
  Overweight 1.11 (0.33, 3.74)
  Obese 2.50 (0.85, 7.38)
 Females (n-102)
  Overweight 1.14 (0.39, 3.28)
  Obese 1.47 (0.52, 4.16)
 Males 45–54 years (n = 96)
  Obese 1.46 (0.60, 3.59)
 Females (n-102)
  Obese 2.76 (1.05, 7.27)
Hilgert et al. 2009 (60) (Brazil)n = 783 Elderly Southern Brazilian people
60–91 years
WHO category obese
 Non-obese < 30
 Obese ≥ 30
Examined for number of teeth and use of prothesesDemographic and socio-economic parameters and smokingPoorer oral status represented by tooth loss was associated with obesity in Southern Brazil older peopleBMI and risk for upper denture only OR (ref. non-obese)
 Obese 2.34 (1.18, 4.27)
BMI and risk of ≤8 teeth present OR (ref. non-obese)
 Obese 2.96 (1.68, 5.19)
Khader et al. 2009 (69) (Jordan)Individuals accompanying patients during their visit to outpatient clinics to Jordan University Medical Centre
n = 340
18–70 years
WHO categories
Underweight < 18.5 kg m−2
 Normal 18.5 – 24.9 (reference)
 Overweight 25–29.9
 Obese ≥ 30
 >88 high for women
 >102 high for men
BF %
 <30% normal
 >30% high
WHR obesity
 >0.90 men
 >0.87 women
Full mouth assessment for PPD, CAL
PII and GI on Ramjford teeth
Periodontitis defined as presence of >4 teeth with minimum one site with PPD ≥ 4 mm and CAL ≥ 3 mm
Age, plaque index, number of missing teethPrevalence
 BMI defined obesity, high WC, and high fat per cent were significantly associated with increased odds of having periodontitis
BMI and risk for periodontitis (ref. normal)
 Overweight 1.40 (0.70, 3.0) P = 0.330
 Obese 2.9 (1.3, 6.1) P = 0.006
WC and risk for periodontitis (ref. normal)
 High 2.1 (1.2, 3.7) P = 0.009
WHR and risk for periodontitis (ref. normal)
 High 1.4 (0.8, 2.4) P = 0.221
Body Fat% and risk for periodontitis (ref. normal)
 High 1.8 (1.03,.3.3) P = 0.039
Presence of periodontitis in BMI categories
 Normal 14.0%
 Overweight 29.6%
 Obese 51.9% P < 0.005
Presence of periodontitis in WC categories
 Normal 19.5%
 High 46.5% P < 0.005
Presence of periodontitis in WHR categories
 Normal 24.7%
 High 38.2% P = 0.008
Kongstad et al. 2009 (70) (Denmark)Subset of Copenhagen Heart Study 2004–2007
n = 1504
20–95 years
BMI – self report
WHO categories
Underweight < 18.5 kg m−2
 Normal 18.5–24.9 (reference)
 Overweight 25–29.9
 Obese ≥ 30
Full mouth assessment for PPD, CAL, BOP, and PI
Full mouth mean CAL dichotomized as
≥3 mm
<3 mm
Age, gender
Age, gender, smoking, educational level, income, alcohol consumption, physical activity, diabetes, number of teeth, and plaque score
 BMI be inversely associated with clinical AL but positively related to BOP
BMI and risk for periodontitis (ref. normal)
 CAL ≥ 3 mm adjusted for age and gender
  Overweight 1.03 (0.76, 1.39)
  Obese 1.08 (0.71, 1.64)
 CAL ≥ 3 mm adjusted for all covariates
  Overweight 0.87 (0.61, 1.23)
  Obese 0.60 (0.36, 0.99)
Kushiyama et al. 2009 (65) (Japan)n = 1070 residents of Miyazaki City (subset of Health Study who had oral examination)
281 males, 789 females
40–70 years
 ≥25 Overweight and Obese
 <25 Non-overweight or obese
 PPD scored one code per sextant based on deepest pocket using WHO probe
 CPITN Code 0–2 no periodontitis (PPD ≤ 3.5 mm)
 CPITN Code 3 or 4 periodontitis (PPD ≥ 4.0 mm)
Age, gender and smokingPrevalence
 Study results support suspected relationship between metabolic syndrome and periodontal disease
BMI and risk for periodontitis CPITN = 4
(BMI ref. <25)
 Overweight and Obese 1.09 (0.77, 1.53) P = 0.64
Nucci Da Silva et al. 2009 (53) (Brazil)n = 214
29–65 years
WHO categories
Underweight < 18.5 kg m−2
 Normal 18.5–24.9 (reference)
 Overweight 25–29.9
 Obese ≥ 30
BF %
Full mouth periodontal examination
Periodontitis defined as presence
≥4 teeth with minimum one or more sties with probing PD ≥ 4 mm and CAL ≥ 3 mm
Not reportedPrevalence
 BMI defined obesity , high WC, and high fat percent were significantly associated with increased odds of having periodontitis
Prevalence of periodontitis by BMI category
 Normal 9.5%
 Overweight 39.6%
 Obese 51.9 %
BMI and risk for periodontitis OR (ref. normal)
 Obese 2.7 (1.8, 6.4)
Östberg et al. 2009 (61) (Sweden)n = 2761
Data from Skaraborg Project 2001–2005
30–74 years
WHO categories
Underweight < 18.5 kg m−2
 Normal 18.5–24.9 (reference)
 Overweight 25–29.9
 Obese ≥ 30
 >88 high for women
 >102 high for men
BF %
Tooth loss dichotomized as 0–19 teeth remaining or >20 teeth remaining (evaluated by patient questionnaire)Age, genderTooth loss is associated with both general and abdominal obesity in participants < 60 yearsBMI risk of <20 teeth remaining (OR)
 Obese 1.25 (0.94, 1.65)
WC risk of <20 teeth remaining (OR)
 Obese 1.43 (1.10, 1.85)
Saxlin et al. 2009 (86) (Finland)Finland Health 2000 Survey
n = 425
45–64 years
Never smoker
 Used BMI as continuous measure
Full mouth assessment for PPD, CAL
Analyzed by number of pockets > 4 mm and >6 mm
Age, gender, education, presence of plaque and number of teethPrevalence
 Serum IL-6 was associated with periodontal infection
BMI and no. teeth with PPD ≥ 4 mm
 RR = 1.03 (1.00, 1.06)
Ekuni et al. 2008 (14) (Japan)Subset of Health and Medical Center of Okayama University health survey who had dental examination
n = 618
18–24 years
Males = 296
Females = 322
WHO categories:
Underweight < 18.5 kg m−2
 Normal 18.5–24.9 (reference)
 Overweight 25–29.9
 Obese ≥ 30
BF %
PPD scored one code per sextant based on deepest pocket using WHO probe
CPITN Code 0–2 no periodontitis (PPD ≤ 3.5 mm)
CPITN Code 3 or 4 periodontitis
(PPD ≥ 4.0 mm)
Age, genderPrevalence
 BMI could be a potential risk factor for periodontitis among healthy young individuals
Risk for periodontitis per 1 kg m−2
 16% increased risk for periodontitis per 1 kg m−2
 Adjusted OR 1.16 (1.03, 1.31)
Ylöstalo et al. 2008 (59) (Finland)Dentate, non-diabetic subset of Health 2000 Health Examination Survey
n = 2842
30–49 years
Categorized in quintiles:
WHO probe measuring four sites per teeth recording only deepest site per tooth
Analyzed as number of teeth with probing pockets of 4 mm or deeper and 6 mm or deeper
Age, gender, smoking, education, dental attendance pattern, tooth brushing frequency, presence of plaque and no. of teethPrevalence, Extent and Severity
 Results showed an association between body weight and periodontal infection among the non-diabetic, non-smoking population aged 30–49
RR for BMI and no. teeth with PD > 4 mm in total population:
 <22.3 reference
 22.3–24.3 1.1 (0.9, 1.3)
 24.3–26.3 1.0 (0.8, 1.1)
 26.3–29.1 1.0 (0.9, 1.2)
 >29.1 1.2 (1.0, 1.4)
 Continuous variable 1.01 (1.00, 1.02)
RR for BMI and no. teeth with PD > 6 mm in total population:
 <22.3 reference
 22.3–24.3 0.9 (0.6, 1.4)
 24.3–26.3 0.6 (0.4, 1.0)
 26.3–29.1 0.8 (0.5, 1.3)
 >29.1 1.0 (0.7, 1.6)
 Continuous variable 1.01 (0.96 , 1.05)
Shimazaki et al. 2007(87) (Japan)Subset Hisayama community health study
n = 584 women with minimum 10 teeth
40–79 years
Categorized as
 ≤88 reference
 >88 high
PD and clinical attachment level measured mesio-buccal and mid-buccal site of all teeth in two quadrants (one max, one mand)
Categorized using mean as
PD < 2.0 or ≥2.0 mm and
CAL < 3 mm or ≥3.0 mm
Age and smoking statusPrevalence
 Results indicated an association between components of metabolic syndrome and periodontitis in a female population
Risk for mean PD ≥ 2.0 mm with WC > 88 (OR)
 1.8 (1.2,2.8)
Risk for mean CAL ≥ 3.0 mm
 0.9 (0.4, 1.9)
Mean (SD) WC in groups by:
 Mean PD < 2.0 WC 82.7 (10.1)
 Mean PD ≥ 2.0 WC 85.8 (8.9) P,0.01
 Mean CAL < 3.0 WC 83.2 (10.0)
 Mean CAL ≥ 3.0 WC 84.6 (9.8) NS
Borges-Yanez et al. 2006 (54) (Mexico)Older age subset of Cross-cultural research on nutrition study
n = 473 non-diabetic
>60 years
defined as BMI ≥ 30
PD and clinical attachment level measured mesio-buccal and mid-buccal site of all teeth in two quadrants (one max, one mand)
Moderate periodontitis – minimum two sites with attachment loss ≥ 4 mm
Severe periodontitis – minimum one site with attachment loss ≥ 6 mm
Calculus indexPrevalence
 Combined obesity and high calculus index associated with high probability of having periodontitis
BMI ≥ 30
Risk (OR) for moderate or severe periodontitis
 0.9325 (0.8524, 1.0170)
BMI and high calculus index
 1.1367 (1.0000, 1.3058)
No periodontitis
 Mean BMI 25.6 (SD 4.84) n = 152
Moderate or Severe Periodontitis
 Mean BMI 25.7 (SD 4.9) P = 0.084 n = 169
Al-Zahrani et al. 2005 (37) (USA)NHANES III subset with periodontal measurements and similar physical activity level for ≥10 years
n = 2521
>18 years
BMIPD and clinical attachment level measured mesio-buccal and mid-buccal site of all teeth in two quadrants (one max, one mand)
Periodontitis classed as minimum one site with attachment loss ≥ 3 mm and PPD ≥ 4 mm
Age, gender, smoking, race, poverty index, education, vitamin use, healthy eating index, time since last dental visit, gingival bleeding, dental calculusNone stated relating to body composition
Prevalence results reported
Bivariate association of BMI and periodontitis
No periodontitis
 Mean BMI 25.8 (se 0.2)
 Mean BMI 26.6 (se 0.3) P = 0.039
Multivariable association.
 OR 1.01 (0.98, 1.05) P value 0.524
Chapper et al. 2005 (57) (Brazil)Sample from Gestational Diabetes Outpatient Unit
n = 60 women with gestational diabetes
Non smokers
19–45 years
n = 18 normal BMI
n = 15 Overweight
n = 27 Obese
WHO categories
 18.5–24.9 normal
 25–29.9 overweight
 ≥30 obese
CAL recording deepest site of six/tooth
GB and BOP
Only female, non smokers with diabetes in sampleSeverity
 Patients with gestational diabetes and pregestational obesity had more gingival inflammation and attachment loss than those with normal pregestational BMI
GB (mean, SD)
 Normal 52.76 ± 27.99*
 Overweight 65.64 ± 23.31
 Obese 78.85 ± 27.44*P < 0.05
BOP (mean, SD)
 Normal 55.65 ± 27.65
 Overweight 71.47 ± 20.35
 Obese 75.31 ± 30.33 P = 0.062
CAL (mean, SD)
 Normal 2.21 ± 0.41*
 Overweight 2.40 ± 0.53
 Obese 2.61 ± 0.54*P < 0.05
No. of teeth (mean, SD)
 Normal 24.44 ± 6.43
 Overweight 25.27 ± 5.24
 Obese 26.25 ± 5.10 P = 0.556
Dalla Vecchia et al. 2005 (13) (Brazil)Subset of a larger representative sample of metropolitan area of Porto Alegre, RioGrande do Sul, Brazil
n = 706
30–65 years
Categorized as:
 <18.5 underweight
 18.5 to 24.9 normal
 25 to 29.9 overweight
 ≥30 obese
Full mouth periodontal assessment (PPD and CAL) at six sites per tooth
Presence of periodontitis classed as ≥30% teeth with attachment loss ≥ 5 mm
Male and Female separate analysis
Age, smoking, race, socioeconomic status, presence of calculus, frequency of dental visits
 Obesity was significantly associated with occurrence of periodontitis in adult, non-smoker women
BMI Risk for periodontitis
Males (n = 329)
 Normal reference
 Overweight 1.1 ( 0.4, 3.3)
 Obese 1.0 ( 0.5, 1.7)
Females (n = 377)
 Normal reference
 Overweight 1.3 (0.8, 2.2)
 Obese 2.1 ( 1.1, 3.9) P < 0.05
Genco et al. 2005 (15) (USA)Part I
 n = 12,367 non-diabetic
 20–90 years
Part I
 ≥27 Overweight
 <27 Non-overweight
Part I
 PD and attachment level measured mesio-buccal and mid-buccal site of all teeth in 2 quadrants(one max, one mand)
Mean attachment loss ≥ 1.5 mm defined as periodontal case
Age, gender, smoking, education, race/ethnicityPrevalence
 Obesity is a significant predictor of periodontal disease
Part I
Risk (OR) for periodontitis
 BMI ≥ 27 1.48 (1.13, 1.93)
Mean % individuals with periodontal disease in groups:
 ≥27 Overweight 24.3
 <27 Non-overweight 26.9
Nishida et al. 2005 (66) (Japan)n = 372 factory workers with complete data from a sample of 409
20–59 years
BMI subdivided in six groups
PPD assessed at 6 sites/teeth, only deepest site/tooth recorded
Calculated % of teeth with PPD > 3.5 mm
Periodontitis = % teeth with PPD > 3.5 mm ≥20%
Non-periodontitis = % teeth with PPD > 3.5 mm <20%
Age, gender, smoking, alcohol consumption, frequency of tooth brushingPrevalence
 Smoking displays the greatest impact on periodontitis among lifestyle-related factors. Both smoking and obesity are independent risk indicators for periodontitis; moreover, these parameters exhibit a dose-response relationship with respect to periodontitis risk
Risk (OR) for periodontitis BMI ≥ 25
 Reference BMI < 25
 3.17 (1.79, 5.61)
Risk for periodontitis by BMI category
 <20 reference
 20–21.9 1.88 (0.66, 5.37)
 22–23.9 1.74 (0.61, 4.97)
 24–25.9 2.68 (0.90, 7.98)
 26–27.9 3.89 (1.20, 12.56)
 ≥28 4.40 (1.18, 16.44)
 P value for trend P = 0.0057
Saito et al. 2005 (17) (Japan)Subset Hisayama community health study
n = 584 women with minimum 10 teeth
40–79 years
All measures divided into quartiles for analysis
 BF (measured by bio-impedance)
PD and clinical attachment level measured mesio-buccal and mid-buccal site of all teeth in two quadrants (one max, one mand)
Categorized in quintiles based on calculated mean PPD and CAL
Age, smoking, plaque index, socioeconomic status, glucose tolerance testPrevalence
 Obesity was associated with deep pockets in Japanese women, even after adjusting for oral glucose tolerance test
Highest quintile PPD (≥1.9 mm) and BMI quartile
 BMI 20.8–22.7 3.0 (1.4, 6.3) P < 0.01
 BMI 22.7–24.9 2.3 (1.1, 5.0) P < 0.05
 BMI 25.0–46.7 4.3 (2.1, 8.9) P < 0.001
Highest quintile PPD (≥1.9 mm) and Body Fat quartile
 BF 24.2–27.9 2.6 (1.2, 5.3) P < 0.05
 BF 28.0–32.5 2.8 (1.3, 5.7) P < 0.01
 BF 32.6–52.5 3.3 (1.6, 6.8) P < 0.01
Highest quintile PPD (≥1.9 mm) and WHR quartile
 WHR 0.89–0.94 1.4 (0.7, 2.8)
 WHR 0.94–0.97 2.8 (0.6, 2.4)
 WHR 0.97–1.12 2.1 (1.1, 4.1) P < 0.05
Highest quintile CAL (≥2.42 mm) and BMI quartile
 BMI 20.8–22.7 1.6 (0.8, 3.1)
 BMI 22.7–24.9 1.3 (0.7, 2.6)
 BMI 25.0–46.7 1.8 (0.9, 3.4)
Mean (SD) body composition in groups by:
Mean PD < 1.9
 BMI 22.9 (3.5)
 BF 28.0 (6.1)
 WHR 0.93 (0.06)
Mean PD ≥ 1.9
 BMI 24.1 (2.9) P = 0.0004
 BF 30.4 (5.7) P = 0.0002
 WHR 0.94 (0.05) P = 0.027
Torrungruang et al. 2005 (18) (Thailand)n = 2005 participants
50–73 years
Survey of senior employees and retired personnel group of Thai workers
Categorized as
 <18.5 underweight
 18.5 to 22.9 normal
 23 to 24.9 at risk
 25 to 29.9 obese I
 ≥30 obese II
 With cut-off point of ≥80 for women and ≥90 for men
Two randomly selected quadrants
Plaque index (PI) plaque present or absent, two buccal sites/tooth one quad and two lingual sites/tooth one quad
PD six sites/tooth
CAL six sites/tooth
Categorized as
Mild CAL < 2.5 mm (reference)
Moderate CAL 2.5 to 3.9 mm
Severe CAL ≥ 4.0 mm
Age, gender, oral hygiene status, education level, income, smoking status, alcohol consumption, diabetes mellitusSeverity
 Age, gender, education, oral hygiene status, smoking, and diabetes are significantly associated with periodontal disease severity in this study group
BMI Risk for moderate periodontitis
 normal 1.1 (0.8, 1.4)
 obese I 1.0 (0.8, 1.3)
 obese II 1.4 (0.9, 2.3)
High WC risk for moderate periodontitis
BMI Risk for severe periodontitis
 normal 1.0 (0.7, 1.4)
 obese I 1.0 (0.8, 1.5)
 obese II 1.5 (0.8, 2.7)
High WC risk for moderate periodontitis
 0.8 (0.6–1.1)
Al-Zahrani et al. 2003(11) (USA)NHANES III subset with complete periodontal measurements
n = 13,665
18–90 years
Categorized as:
 <18.5 underweight
 18.5 to 24.9 normal
 25 to 29.9 overweight
 ≥30 obese
 >88 high for women
 >102 high for men
PD and clinical attachment level measured mesio-buccal and mid-buccal site of all teeth in two quadrants (one max, one mand)
Periodontitis classed as minimum one site with attachment loss ≥ 3 mm and PPD ≥ 4 mm
Age, gender, smoking, race, poverty index, education, diabetes, time since last dental visitPrevalence
 BMI > 30 and WC > 88 cm associated with increased prevalence of periodontal disease in younger population
 Overweight 1.06 (0.912, 1.235) P = 0.434
 Obese 1.37 (1.41, 1.644) P = 0.001
 High 1.33 (1.113, 1.601) P = 0.002
 Stratified by Age
   Overweight 1.00 (0.705, 1.407)
   Obese 1.76 (1.187, 2.612) P < 0.01
   Overweight 0.98 (0.812, 1.191)
   Obese 1.12 (0.820, 1.522)
   Overweight 1.20 (0.882, 1.626)
   Obese 1.20 (0.892, 1.608)
Prevalence of Periodontal Disease % (SE)
 <18.5 underweight 9.15 (2.488) n = 301
 18.5 to 24.9 normal 11.55 (0.828) n = 5289
 25 to 29.9 overweight 14.59 (0.856) n = 4623
 ≥30 obese 17.65 (1.008) n = 3427
Wood et al. 2003 (19) (USA)Caucasian, >18 years, participants of NHANES III (sample of general population)
n = 8032 overall (varied by availability of outcome data)
Subcutaneous fat (S)
Fat-free mass (FFM)
PD and clinical attachment level recorded mesial and mesio-buccal in two contra-lateral quadrantsAge, gender, history of diabetes, current smoking, and socioeconomic statusPrevalence,
and Severity
 Significant correlations between body composition and periodontal disease (with WHR most significant, followed by BMI, FFM, and S) showed similarities to those observed in other obesity-related health problems
Correlation coefficient:
Between mean periodontal attachment loss and
 BMI 0.1512 (P < 0.01)
 WHR 0.2459 (P < 0.01)
 Log sum of S 0.0866 (P < 0.01)
 FFM 0.0553 (NS)
Between mean PD and
 BMI 0.1306 (P < 0.01)
 WHR 0.1596 (P < 0.01)
 Log sum of S 0.1100 (P < 0.01)
 FFM -0.0560 (NS)
Between mean gingival bleeding and
 BMI 0.0945 (P < 0.01)
 WHR 0.2081 (P < 0.01)
 Log sum of S 0.0413 (NS)
 FFM 0.1872 (P < 0.01)
Saito et al. 2001 (16) (Japan)Subset of Fukuoka Health Promotion Centre health promotion program
n = 643 healthy subjects (512 females, 131 males)
19–79 years
Categorized as
 Upper body obesity defined as WHR ≥0.8 in females or WHR ≥ 0.9 in males
 X-ray absorpiometry (% BF) categorized
 For men (for women):
  Low ≤ 19.9 (≤29.9)
  Mid 20–29.9 (30–39.9)
  High 30–39.9 (40–49.9)
  Very high ≥ 40 (≥50)
10 teeth probed (1st and 2nd molars, upper right incisor, and lower left incisor) with WHO probe 6 sites per tooth giving a score/sextant
Code 0 PPD < 3.5 mm
Code 1 PPD 3.5–5.5 mm
Code 2 PPD ≥ 6 mm
Perio case classed as at least one sextant with code 1 or 2
Age, gender, smoking, socioeconomic status, diabetes, frequency of oral hygienePrevalence
 In subjects with high waist-hip ratio, higher categories of BMI (or body fat) significantly increased the adjusted risk of periodontitis, compared with subjects with low waist-hip ratios and the lowest category of BMI (or body fat)
High WHR (Upper body obesity)
Risk for periodontitis (OR)
 2.0 (1.4, 2.9)
High WHR and BMI category
Risk for periodontitis
 BMI 22–24.9 2.0 (1.1, 3.4)
 BMI 25–29.9 3.3 (1.9, 5.6)
 BMI ≥ 30 4.3 (1.6, 11.7)
Saito et al. 1998 (68) (Japan)Subset of Fukuoka Health Promotion Centre health promotion program
n = 241 healthy subjects (172 females, 69 males)
20–59 years
Categorized as
 X-ray absorpiometry (% BF)
PPD scored one code per sextant based on deepest pocket using WHO probe
CPITN Code 0–2 no periodontitis (PPD ≤ 3.5 mm)
CPITN Code 3 or 4 periodontitis
(PPD ≥ 4.0 mm)
Age, gender, smoking, oral hygiene frequencyPrevalence
 Obesity and periodontitis are related
Relative Risk for periodontitis
 BMI 20–24.9 1.7 (0.7, 3.8)
 BMI 25–29.9 3.4 (1.2, 9.6)
 BMI ≥ 30 8.6 (1.4, 51.4)
 P = 0.02
RR for each 5% increase in body fat
 1.3 (1.0, 1.6) P = 0.02
Johansson et al. 1994 (72) (Sweden)Swedish subset of sample of the Multinational Monitoring of Trends and Determinants in Cardiovascular Disease study
n = 1583
25–64 years
Data collected at medical examination visit
Dental self-assessment of dentate status based on 3 questions:
1/ Do you only have your natural teeth?
2/ Do you have natural teeth and a removable partial denture?
3/ Do you wear full dentures?
Age, education, diet, HDL cholesterol, smoking
 Age adjusted odds ratios to be edentulous increased for, both sexes by increasing quintiles of BMI, waist hip ratio.
Risk for edentulism by increasing quintile BMI(OR):
  1st quintile – reference
  4th quintile – 1.19 (0.66, 2.14)
  5th quintile – 2.07 (1.17, 3.63)
  4th quintile – 1.85 (0.99, 3.46)
  5th quintile – 3.44 (1.87, 6.32)
Risk for edentulism by increasing quintile of WHR
  1st quintile – reference
  4th quintile – 1.66 (0.92, 3.00)
  5th quintile – 2.38 (1.34, 4.22)
  4th quintile – 4.40 (2.34, 8.30)
  5th quintile – 5.98 (3.19, 11.2)
  BMI Risk to be edentulous (with analysis adjustment for all factors listed)
 Men 1.13 (0.94, 1.35) Women 1.34 (1.09, 1.64)
Mean BMI in:
 Dentate men 25.8 (SE 0.1)
Edentulous men 6.7 (SE 0.3) P < 0.01
 Dentate women 25.0 (SE 0.2)
 Edentulous women 26.8 (SE 0.3) P < 0.001
Summers and Oberman 1968 (56) (USA)Subset of Tecumseh, Michigan health study
408 participants had dental exam
84 edentulous excluded
n = 324
>20 years
Males = 154
Females = 170
Relative weight
Periodontal Disease Index (PDI)
Based on examination of Ramjord teeth (UR molar, UL central, UL bicuspid, LL molar, LR central, LL bicuspid
Categorical score given assessing periodontal status based on gingival inflammation, crevice depth, plaque and calculus
Genders analyzed separately
 The relationship between periodontal disease and certain physiologic and social variables might be worthy of longitudinal study
Correlation (simple correlation coefficient):
between weight and PDI score in men
 −0.78 (NS)
between weight and PDI score in females
 0.13 (NS)
between relative weight and PDI score in men
 0.82 (NS)
between weight and PDI score in females
 −0.32 (NS)


The combined total of references obtained from the electronic search strategy modified for each database was 541 citations, with removal of 34 duplicates resulting in 507 citations. Electronic search for grey literature resulted in 0 hits, and hand searching provided 20 additional citations. This totalled 527 titles and abstracts to be screened.

The 527 titles and abstracts were screened in duplicate by JS and FD. Kappa score calculated for screening agreement was 0.95 (95% CI 0.90 to 0.99). Disagreement on two articles was resolved by discussion and warranted a change in the protocol to exclude studies that investigated the effect of oral health (exposure) on obesity (outcome). Figure 1 summarises the screening process showing the number of citations passing through each step.

Figure 1.

Study identification flow chart.

The search retrieved a large number of relevant articles together with a substantial number of irrelevant hits confirming the high sensitivity and relatively low precision of the search. The majority of these contained information pertaining to body composition and other chronic diseases. Articles of possible mechanisms, laboratory studies, the relationship between body composition and caries, and narrative reviews were also amongst the excluded publications. 8 foreign language publications were located (Portuguese) and excluded at the abstract screening stage with the assistance of translation.

All 61 potentially relevant full text articles and abstracts (where only abstracts were published) were screened independently in duplicate according to the eligibility criteria outlined in the protocol. Reviewers were in full agreement on inclusion of articles and abstracts.

Twenty-eight of the 61 full text articles and abstracts screened were excluded for the following reasons:

  • • Two were animal or in vitro studies (31,32)
  • • Four included adolescents with adults in analyses (33–36)
  • • Fourteen did not report outcomes relevant to the review (37–50)
  • • Two were not relevant or investigated the effect of being dentate on weight (51,52)
  • • Six were narrative reviews or editorials (20,21,23–26)

Where abstracts only were published and were eligible for inclusion based on information available (one abstract) (53), author contact was attempted to obtain additional data required to allow inclusion of the study. This resulted in the final inclusion of 32 full text articles and one abstract in the systematic review.

Descriptive results

Descriptive summaries of the 33 included studies are presented as evidence tables in Tables 2–4. Studies are listed chronologically from the most recent based on publication date (thereafter alphabetically within same year). The first author is listed together with the country location of the population studied. The studies are grouped into separate tables according to study design beginning with cohort studies as the highest level of evidence retrieved then moving down the hierarchy of evidence (see Table 2–4). The tables are organized according to categories of data and structured to facilitate the description of potential sources of heterogeneity.

Study characteristics

An overall summary of noteworthy study characteristics appears in Table 5. Further overview of results highlights that three studies reported no evidence of an association between body composition and periodontitis (54–56). The remaining 30 studies reported a positive association between body composition and periodontitis of which 25 were focused on prevalence, five studies reported on extent and severity of periodontitis (19,36,57–59) and two on tooth loss as an indicator of periodontitis (60,61). One study investigated a possible time dependent association between weight and periodontitis but did not show a positive correlation over time (62).

Table 5.  Summary of included study characteristics
Study CharacteristicNo. of studies (n = 33)Citations
  1. CPITN, Community Periodontal Index of Treatment Needs; NHANES III, Third National Health and Nutrition Examination Survey; PDI, Periodontal Disease Index.

Publication Year  
 Not published between 1999 and 20093(54,55,74)
 Published in 2008/200913(14,36,50,53,56–65)
Study Design  
Secondary analysis of community health database21(11,13–17,19,54–57,60–62,64,65,70,71,74–76)
 Analysis of NHANES III4(11,15,19,37)
 Analysis of Hisayama Community Health Study3(17,62,75)
USA or Japan study sample16(11,12,14–17,19,55,56,62,69,70,73–76)
Periodontal status assessment based on: Full mouth11(13,53,56,59,60,63,64,66–69)
 Two sites/tooth, two contra-lateral quadrants9(11,15,17–19,62,70,71,75)
 Deepest site per tooth recorded3(65,72,73)
Body composition assessment based on:All except 
 Body mass index2(55,75)
 Waist circumference8(18,53,57,59,61–63,75)
 Waist hip ratio7(18,53,57,59,61–63,75)
 Body fat %8(14,16,17,53,56,59,62,74)

Prevalence data reported odds ratios ranging from 0.93 to 5.31 for periodontitis associated with overweight or obese. One study on extent and severity of periodontitis reported no significant association between BMI and number of pockets with pocket probing depth (PPD) >6 mm (59). One study reported a Spearman's correlation coefficient of 0.33 (P = 0.003) between the severity of periodontitis and BMI (58).

Bias assessment protection

Based on the Newcastle-Ottawa Quality Assessment Scale for the respective study design (29), all studies met the criteria to be categorized as low risk of bias within the respective study design. Five out of six of the case–control studies recruited controls from hospital environments rather than the community and hence, would be considered to be of lesser quality than those with controls recruited from the community. All cross-sectional studies would be considered comparable in quality.

Quantitative results

Six studies included in a meta-analysis (15,17,63–66) compared the odds of having periodontitis if an individual was overweight or obese rather than of normal weight using WHO BMI categories as the exposure measure. Individual study odds ratios ranged from 1.09 to 4.54, with not all providing statistically significant evidence of an association between overweight/obesity and periodontitis (Fig. 2 Forest plot overweight or obese). Precision of the studies varied substantially as seen in the varying span of the confidence intervals. The overall summary estimate based on a random effects model, deemed more appropriate due to the level of heterogeneity in the studies, demonstrated the odds of having periodontitis if an individual is overweight or obese to be 2.13 (95% CI 1.40, 3.26) times the odds if they have a normal BMI. This result was statistically significant (P < 0.001). A chi-squared test and I2 test for heterogeneity confirmed statistically significant heterogeneity (P = 0.001 and I2 = 73.0%).

Figure 2.

Forest plot overweight and obese compared with normal (body mass index).

Twelve studies included in a meta-analysis compared the odds of having periodontitis if an individual was in an overweight BMI category (but not obese) rather than in a normal BMI category. Individual study odds ratios ranged from 0.93 to 3.40, with only three of the 12 studies (16,67,68) providing statistically significant evidence of an association between BMI category overweight and periodontitis (Fig. 3 Forest Plot Overweight). Precision of the studies varied substantially as demonstrated by the confidence intervals. The overall summary estimate based on a random-effects model demonstrated the odds of having periodontitis if an individual is overweight (but not obese) to be 1.27 (95% CI 1.06, 1.51) times that of a BMI normal category individual. This result was statistically significant (P = 0.008). Chi-squared test and I2 test for heterogeneity confirmed statistically significant heterogeneity (P < .001 and I2 = 78.8%).

Figure 3.

Forest plot overweight compared with normal (body mass index).

Twelve studies included in a meta-analysis compared the odds of having periodontitis if an individual was in an obese BMI category to that if they were in a normal BMI category. Individual study odds ratios ranged from 1.00 to 8.60, with 10 of the 12 (11–13,16,18,53,59,62,67–70) studies providing statistically significant evidence of an association between BMI category obesity and periodontitis (Fig. 4 Forest plot obese). The overall summary estimate based on random-effects model demonstrated the odds of having periodontitis if an individual was obese to be 1.81 (95% CI 1.42, 2.30) times that if they had a normal BMI. This result was statistically significant (P < 0.001). Chi-squared test and I2 test for heterogeneity confirmed statistically significant heterogeneity (P < 0.001 and I2 = 72.6%).

Figure 4.

Forest plot obese compared with normal (body mass index).


The overall results, based on the descriptive and quantitative summaries, indicate evidence of a positive association between overweight and obesity, and/or excessive body fat and prevalence of periodontitis, although largely based on data from study designs low in the hierarchy of evidence. Reports of the relationship of body composition and extent and severity of periodontitis are inconclusive with some demonstrating a positive association and others the contrary (19,36,55,57–59). In the studies that incorporated multiple measures of body composition, the measures and magnitude of association varied within the same study dependent on the measure of body composition. Results were also inconsistent between studies when comparing similar measures of body composition. This variability suggests a need to continue to include a battery of measures of body composition in future clinical studies estimating periodontitis outcomes to allow validation and increased understanding of the interpretation of both existing and newer methods of body composition assessment.

Many factors contributed to the heterogeneity of the data included in the evidence tables and meta-analysis. Elements that differed across studies included; variations in statistical models and tests used in analyses, covariates included in the analyses (e.g. some studies reported as few as one explanatory variable with some reports based on inclusion of eight to 11 factors), sampling frames, definition basis of the presence of periodontitis (number of sites above a threshold, CPITN, PPD, CAL, radiographs, tooth loss), scope of periodontal status assessment (e.g. full mouth vs. selected sites), and variations in cut-points for BMI categories. The WHO has defined BMI categories to facilitate analyses of the association BMI may have with health or disease (Table 6) (71). Some of the included studies grouped BMI according to different thresholds rather than based on WHO thresholds (15,17,63–66,72). The basis for BMI thresholds chosen in setting BMI categories in these studies was not clear. Similar to other outcomes, the cut-point chosen in setting thresholds will have a direct influence on the magnitude of effect size, therefore contributing to the heterogeneity in these studies. Despite the high heterogeneity, a consistent pattern remains supporting an increased risk for periodontitis for individuals who are overweight or obese. The heterogeneity may impact the magnitude of the increased risk, rather than precluding the risk.

Table 6.  The WHO international classification of adult underweight, overweight and obesity according to BMI (WHO 2008)
ClassificationBMI(kg m−2)
Principal cut-off pointsAdditional cut-off points
 Severe thinness<16.00<16.00
 Moderate thinness16.00–16.9916.00–16.99
 Mild thinness17.00–18.4917.00–18.49
Normal range18.50–24.9918.50–22.99
  Obese class I30.00–34–9930.00–32.49
  Obese class II35.00–39.9935.00–37.49
  Obese class III≥40.00≥40.00

The evidence resulting from this review is based on study designs valuable to generating hypotheses but weak for confirming an association between the exposure and disease or for studying the temporal relationship between the putative exposure and outcome. New studies of higher levels of evidence (e.g. prospective) designed according to pre-specified hypotheses are necessary to confirm and further understand the association between body composition and the onset or extent and severity of periodontitis. Such designs should incorporate comprehensive full mouth periodontal measurements and/or alveolar bone levels to allow more precise assessment of the periodontal status (73,74). This systematic review suggests there is currently sufficient observational evidence to argue further research is needed based on a clinical and patient-centred approach as compared to an epidemiological framework. The diversity of results according to study outcomes, covariates, and results (as outlined in the preceding paragraph) suggest that numerous factors may be implicated in the modification of the effect of overweight/obesity and excessive body fat on periodontitis.

The meta-analyses demonstrated a range of summary estimates of odds ratios. The highest summary estimate resulted from combining studies that reported overweight/obesity categories together yielding an overall odds ratio of 2.1 (CI 1.4, 3.3), P < 0.001. Confidence intervals indicated variability in the precision of the studies with two having less precision than the remaining four. With only six studies in this meta-analysis, these results may be less conclusive.

Results of the meta-analyses performed in the present study indicated a stronger association between obesity and periodontitis than that between overweight and periodontitis. The odds of having periodontitis if an individual is obese is 1.8 times greater than that if the individual has a normal BMI (P < 0.001), whereas the odds is only 1.3 times greater in the overweight compared to those with normal weight (P = 0.008). These data may infer an increasing risk for periodontitis with increased BMI (i.e. a ‘dose–response’), although it is important to recognize BMI values and thresholds have their origin in epidemiological reports which lack interpretation of individual clinical meaning.

Understanding the reasons for the variability in the odds ratio summary estimates between overweight and obese may lie in the consideration of possible mechanisms behind the hypothesized association. A report from Amar and co-workers confirmed a dysregulation of immune responses in a periodontitis animal model resulting in increased bone loss in obese rats (75). In humans, possible mechanisms have been suggested to be partially linked to the elevation of levels of acute phase proteins, pro-inflammatory cytokines and leukocyte counts in obesity as a result of adipose tissue endocrine activity (76). Obesity has been shown to modulate host immune response resulting in increased susceptibility to various infections and exaggerated host immune responses to them (9). Furthermore, it been shown that adipose tissue (adipocytes) secretes several pro-inflammatory factors, also implicated in periodontitis, including cytokines (e.g. IL-6), chemokines and can affect T-cell function (9).

Clinical evidence seems to suggest that obese individuals have an increased local inflammatory response (34) as well as possibly an altered periodontal microflora (77). As host response to local bacterial challenge is a key factor in determining periodontitis susceptibility (78), an increased inflammatory state as that found in obese individuals could predispose them to increased periodontal tissue destruction (22).

An alternative mechanism behind this association could be the altered insulin sensitivity state found in obese individuals (8). A body of evidence suggests that reduced insulin sensitivity coupled with increased production and accumulation of advanced glycation end-products at the gingival level in people with diabetes can result in greater periodontal tissue destruction (15,79). Chronic periodontitis has been suggested as a distant source of inflammatory by-products affecting insulin sensitivity (15). D'Aiuto et al. have more recently presented data suggestive of a moderate association between severe periodontitis and metabolic syndrome including insulin resistance and obesity (80). Tumour necrosis factor-α (TNF-α) is one of the cytokines secreted in particular by visceral adipose tissue and is implicated in mediation of endotoxin induced injury in periodontal tissue (81). Concentrations of TNF-α have been reported to be higher in periodontitis patients than periodontally healthy patients and have been shown to be reduced following periodontal therapy (77). TNF-α in gingival crevicular fluid (GCF) has also been shown to be associated with BMI ≥ 40 (34). Association of TNF-α in gingival crevicular fluid and BMI has also been demonstrated in obese subjects without periodontitis (34). It has been suggested that the concentration of TNF-α in the GCF is resulting from the secretions of the adipose tissue (34). Leptin, also secreted from adipose tissue, plays a protective role in immune system function and is present at higher levels in minimally inflamed gingival with decreased concentrations present in sites with increased probing pocket depth (82).

A number of plausible biological pathways for a possible association between overweight/obesity and periodontitis have been suggested. However, it should be acknowledged that each of these chronic diseases is also independently associated with other factors, some of which are common to both diseases such as smoking, diet and insulin sensitivity. Further understanding of the impact these factors may have on overweight/obesity, periodontitis, or the interaction of these factors together with these two chronic diseases, is still to be ascertained.

Limitations of this study have resulted from the challenge of locating evidence specific to obesity and periodontitis. The search strategy identified a broad spectrum of articles; however, still more data may lie embedded in publications focusing on related elements studies of comorbidities such as cardiovascular disease, or diabetes and periodontitis. Hand searching of bibliographies and recent related journals contributed additional articles; however, much data may remain hidden. The results and value of these collateral data existing as part of previously conducted investigations remains unknown. It is also possible that additional researchers have investigated this association but not reported it due to a lack of positive findings, therefore resulting in a possible publication bias. In addition, the meta-analyses performed required some level of assumption in grouping studies together based on outcome reported and similar BMI grouping. High levels of heterogeneity in the elements of study characteristics suggest caution in the interpretation of the pooled estimates as it is feasible that the heterogeneity renders the magnitude of the estimates unreliable.

In accordance with the secondary objective of the review to facilitate design and conduct of further research in this area, suggestions are presented in Table 7, following the structure suggested by Brown et al. 2006 for formulating research recommendations in systematic reviews (83).

Table 7.  Research Recommendations (based on format from Brown et al. 2006)
Core ElementsRecommendation for Further Research
  1. BMI, body mass index; WC, waist circumference; WHO, World Health Organization; WHR, waist/hip ratio.

(E) Evidence (current)Systematic review identified predominantly cross sectional and a few case control studies (a number of secondary analyzes of health databases)
(P) PopulationAnalysis or recruitment of homogeneous age groups including adolescence, otherwise healthy individuals then progress to other population groups (e.g. diabetes sample)
(I) Intervention/ExposureWHO categories of BMI as basic standard approach to BMI categories
(C) ComparisonInclusion of alternate measures of body composition such as WC (using WHO categories), WHR (using WHO categories), %body fat
(T) Time stampDecember 2009
(O) OutcomesTooth loss, full mouth periodontal assessment including bleeding. Include age, gender, smoking as covariates in analyses.
(d) Disease burdenPrevalence of overweight and obesity 1.7 billion individuals worldwide
Prevalence of periodontitis ranging from 20% to 50% worldwide.
(t) TimelinessCohort follow-up of young adults over a period of years, 6-month follow-up of treatment groups
(s) Study typeProspective cohort or treatment studies

In conclusion, results of this study indicate a positive association of BMI categories overweight and obese with presence of periodontitis, although the magnitude of the association is still unclear. This evidence suggests the need and merit of robust prospective studies in humans designed to further quantify and/or understand the mechanisms of this association. There is insufficient evidence at this time to provide specific guidance to clinicians on the clinical management of overweight and obese individuals with periodontitis.

Conflicts of interest

None for all authors.


The kind assistance of Donna Moskal-Fitzpatrick, Unit of Periodontology, UCL, Eastman Dental Institute is gratefully acknowledged. This work was undertaken at UCLH/UCL who received a proportion of funding from the Department of Health's NIHR Biomedical Research Centres funding scheme. This project was supported in part by an educational grant provided by Johnson & Johnson Consumer Services EAME Limited. Dr D'Aiuto holds a Clinical Senior Lectureship Award supported by the UK Clinical Research Collaboration.