SEARCH

SEARCH BY CITATION

Keywords:

  • body mass index;
  • periodontal infection;
  • older people

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. Author contributions
  10. References

Objective

The aim of this study was to investigate the association between BMI and periodontal infection in a sample of non-smoking individuals aged 75 years or older.

Subjects and Methods

The study sample included 157 non-smoking dentate persons (110 women, 47 men, mean age 80.6 years) belonging to the Geriatric Multidisciplinary Strategy for the Good Care of Older People study in Kuopio, Finland. The data were gathered by interview together with geriatric and oral clinical examination. The outcome variable was the number of teeth with periodontal pockets measuring 4 mm or more in depth. Poisson regression models were used to estimate relative risk (RR) and 95% confidence intervals (CI).

Results

After adjustment for confounding factors, the relative risk for the number of teeth with deepened periodontal pockets (≥4 mm) was 0.7 (CI: 0.6–0.9) among those with a BMI 25–29.99 and 1.1 (CI: 0.8–1.4) among those with a BMI ≥30, compared with those having a BMI <25.

Conclusion

Within the limitations of this study, including small sample size, possibility of confounding and other biases, the results do not provide evidence that elevated body weight would be a risk for periodontal infection among older people.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. Author contributions
  10. References

Many studies (Saito et al, 2001; Wood et al, 2003; Genco et al, 2005; Sarlati et al, 2008; Ylöstalo et al, 2008; Khader et al, 2009) have shown that body weight is associated with periodontal infection. The biological mechanisms by which body weight affects the development and/or progression of periodontal infection may be the harmful effects of proinflammatory cytokines or adipokines secreted by adipose tissue (Karthikeyan and Pradeep, 2007; Saito and Shimazaki, 2007).

To date, only a few studies (Al-Zahrani et al, 2003; Torrungruang et al, 2005; Linden et al, 2007) have analysed the role of high body weight as a risk for periodontal disease among older people, and the results of these studies are in contradiction. The reason for these contradictory results is not known, but they may arise from differences between study populations, such as difference in age, number of teeth and risk factors for periodontitis. Also, increased individual susceptibility to infections or cumulative effects of extraction-oriented dental treatment, which cannot properly be taken into account in the analyses, may explain the variation in the findings of earlier studies.

Based on the biological harmful effect of adipose tissue, we hypothesized that subjects with high body weight are more likely to have periodontal infection or severe periodontal infection than subjects with normal body weight. The aim of this paper was to study the association between BMI and periodontal infection in a homogenous group of non-smoking subjects aged 75 years or older.

Material and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. Author contributions
  10. References

Study population

The study population consisted of 157 non-smoking dentate people (110 women, 47 men, mean age 80.6 ± 3.7 years) belonging to the Geriatric Multidisciplinary Strategy for the Good Care of Older People (GeMS) study. The GeMS study was based on a sample of 1000 subjects aged 75 or older on the 1 November 2003, living in the city of Kuopio in eastern Finland, which was randomly drawn from the Population Register Centre and randomized into an intervention group (n = 500) and a control group (n = 500). A geriatric examination was carried out for 404 persons in the intervention group (participation rate 80.8%). A more detailed description of the GeMS study is presented by Lampela et al (2010).

A baseline oral clinical examination was carried out for 354 of the 404 participants belonging to the intervention group (23 died before the oral clinical examination; 27 refused to come to the oral clinical examination) in 2004–2005 (participation rate 70.8%). Among these, there were 180 dentate persons of whom 165 home-dwelling participants underwent a periodontal examination. Finally, smokers and one person without BMI variable were excluded from the final study population, resulting in 157 study subjects (Figure 1).

image

Figure 1. Flow chart of participation in the periodontal examination

Download figure to PowerPoint

Written informed consent was received from the study subjects or their relatives. The ethics committee of the University of Kuopio and Kuopio University Hospital approved the study protocol.

Comprehensive geriatric assessment

In 2004, two physicians specialising in geriatrics and two trained nurses conducted a comprehensive geriatric assessment (CGA) of the participants at the health centre of Kuopio or at the homes of the participants in cases where the person was unable to visit the health centre. The participants were interviewed for their health behaviour, health and social life. A close relative or caregiver gave the information in cases where the person was not able to answer the questions. Drugs used and overall physical and mental status were examined by a physician specializing in geriatrics.

Oral clinical examination

Oral clinical examinations were carried out in 2004–2005 by two dentists at the dental clinic of the municipal health centre of Kuopio or at the homes of the participants if they preferred a dentist's home visit (127 study subjects were examined at the dental clinic, 25 at home and 9 persons in some other place, such as in sheltered home).

The oral clinical examinations were based on written instructions and performed in a standardized way. At the beginning of the study, the dentists examined 7 study subjects together to improve agreement between the examiners on their assessment of the condition of teeth and periodontium as well as the assessment of the participants' treatment need. Because of the length of the examination and the age of the subjects, repeated examinations were not performed.

The oral clinical examination was carried out in a dental unit having a dental chair, a unit lamp, a syringe and saliva suction, using a WHO colour-coded periodontal probe, a gauze pad and a mouth mirror. The extraoral clinical examination included contemplation of the face and lips, measuring the maximum opening of the mouth and palpation of the main occlusal muscles and condyles. In the intraoral examination, the dentist recorded the cleanliness, condition and functioning of the dentures, as well as the periodontal and dental status. In addition to the clinical examination, the dentist conducted a structured interview about oral health habits.

Outcome variable

The outcome variable was the number of teeth having periodontal pockets of a pocket depth of 4 mm or more, which was used to measure the presence and extent of periodontal infection. The periodontal pockets were probed on the mesiobuccal and distolingual/distopalatal surfaces of all teeth, and the deepest pocket depth of each tooth was recorded.

Explanatory variable

Trained study nurses measured body weight and height from all the study subjects during the comprehensive geriatric assessment. BMI was calculated as the body weight in kilograms/height in metres squared (kg/m2). BMI was classified into three groups according to the World Health Organization (World Health Organization, 2000) classification: less than 25, 25.0–29.99 and 30 or more.

Potential confounding factors

Toothbrushing frequency was classified as toothbrushing twice a day or more often vs once a day or more rarely. Dental visiting habit was categorized as regular dental visits vs symptom-based visits or no visits.

A crude estimate of the amount of dental plaque was made based on visual observation of the presence of dental plaque on the palatal and buccal surfaces of all teeth after light drying with air syringes. The variable was classified as <20% of teeth with dental plaque, 20–50% of teeth with dental plaque and >50% of teeth with dental plaque.

Diabetes was defined on the basis of a recorded diagnosis, reimbursable medication or information obtained from the geriatric examination. A recorded medical diagnosis was used to define rheumatoid diseases.

The Lawton Instrumental Activities of Daily Living Scale (IADL) was used to evaluate functional status and independent living skills (Lawton and Brody, 1969). The IADL questionnaire included 8 domains, and the sum score of the IADL ranged from 0 (dependence in all domains) to 8 (totally independent). The IADL score was classified as 0–6 vs 7–8. Cohabitation was classified as living with somebody vs living alone.

Duration of education was classified into two categories: 7 years or more vs less than 7 years. Smoking habits were asked as to whether the study subjects had never smoked, smoked previously but had quit smoking, smoked occasionally or smoked daily. Because there were only very few smokers among the study subjects, they were excluded from the study (Figure 1).

Statistical methods

Poisson regression models were used to estimate relative risk (RR) and 95% confidence intervals (CI). The models were adjusted for factors, which according to current knowledge, are predictive of the outcome. These included gender, age, education, cohabitation, presence of dental plaque, dental visiting habits, toothbrushing frequency, physical activity (IADL score), diabetes, rheumatoid diseases and number of teeth (offset variable). SPSS 16.0 for Windows (SPSS Inc, 2008) was used in the statistical analyses.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. Author contributions
  10. References

Of the participants of this study, 39% had a BMI less than 25, 41% had a BMI of 25–29.99, and 20% of the participants had a BMI of 30 or more. The mean number of teeth in the respective BMI categories was 14.7, 14.4 and 14.7. The mean number of teeth with periodontal pockets measuring 4 mm or more was 2.8, 2.4 and 3.1, respectively.

The median number of teeth with periodontal pockets measuring 4 mm or more was 2.0 among participants with a BMI less than 25, 1.0 among participants with a BMI of 25–29.99 and 1.0 among participants with a BMI of 30 or more (Table 1).

Table 1. Basic characteristic of the study population by different categories of BMI
 BMI
Total n = 157<25 n = 61 (38.9%)25 to <30 n = 64 (40.8%)≥30 n = 32 (20.4%)
  1. a

    Education was missing from 2 subjects, cohabitation from 2 subjects, dental visits from 2 subjects, toothbrushing from 2 subjects, IADL score from 1 subject and rheumatoid diseases from 13 subjects.

  2. b

    Data in 2006.

Sociodemographic variables
 Age (mean ± s.d.)80.6 ± 3.780.3 ± 3.381.1 ± 4.280.1 ± 3.3
 ≥85 years (n,%)22/157 (14.0)7/61 (11.5)11/64 (17.2)4/32 (12.5)
 Gender, proportion of males (n,%)47/157 (29.9)20/61 (32.8)24/64 (37.5)3/32 (15.2)
 Education ≥7 yearsa (n,%)87/155 (56.1)37/60 (61.7)33/63 (52.4)17/32 (53.1)
Cohabitationa
 Living alone (n,%)81/155 (52.3)32/60 (53.3)28/64 (43.8)21/31 (67.7)
 Living with somebody (n,%)74/155 (47.7)28/60 (46.7)36/64 (56.3)10/31 (32.3)
Dental variables
 Number of teeth (mean ± s.d.)14.6 ± 8.214.7 ± 8.814.4 ± 7.914.7 ± 7.7
 Number of teeth (min, max)1, 291, 281, 272, 29
Number of teeth
 1–9 teeth (n,%)56/157 (35.7)24/61 (39.3)22/64 (34.4.)10/32 (31.2)
 10–19 teeth (n,%)43/157 (27.4)13/61 (21.3)20/64 (31.2)10/32 (31.2)
 ≥20 teeth (n,%)58/157 (36.9)24/61 (39.3)22/64 (34.4.)12/32 (37.5)
Number of teeth with periodontal pockets ≥4 mm (mean ± s.d.)2.7 ± 3.72.8 ± 3.22.4 ± 4.03.1 ± 4.1
Number of teeth with periodontal pockets ≥4 mm (min, max)0, 260, 120, 260, 14
Number of teeth with periodontal pockets ≥4 mm (median)1.02.01.01.0
Proportion of participants with ≥1 tooth with periodontal pockets ≥4 mm (n,%)92/157 (58.6)39/61 (63.9)34/64 (53.1)19/32 (59.4)
Number of carious teeth (mean ± s.d.)1.2 ± 2.10.9 ± 1.81.7 ± 2.60.9 ± 1.3
Teeth with dental plaque
 ≤20% of teeth with dental plaque (n,%)54/157 (34.4)22/61 (36.1)21/64 (32.8)11/32 (34.4)
 21–50% of teeth with dental plaque (n,%)38/157 (24.2)18/61 (29.5)12/64 (18.8)8/32 (25.0)
 >50% of teeth with dental plaque (n,%)65/157 (41.4)21/61 (34.4)31/64 (48.4)13/32 (40.6)
 Regular dental visitsa (n,%)88/155 (56.8)35/61 (57.4)36/63 (57.1)17/31 (54.8)
 Toothbrushing twice a daya (n,%)130/155 (83.9)57/61 (93.4)49/63 (77.8)24/31 (77.4)
General health-related variables
 Diabetes (n,%)18/157 (11.5)5/61 (8.2)7/64 (10.9)6/32 (18.8)
 IADL score 0–6a (n,%)36/156 (23.1)7/61 (11.5)22/63 (34.9)7/32 (21.9)
 Rheumatoid diseasesb (n,%)17/144 (11.8)7/54 (13.0)9/59 (15.3)1/31 (3.2)

Unadjusted risk estimates for explanatory variables are presented in Table 2. There was no consistent, exposure-dependent association between body mass index and the number of teeth with deepened periodontal pockets after adjustment for age, gender, education, cohabitation, presence of dental plaque, dental visiting habits, toothbrushing frequency, physical activity (IADL score), diabetes, rheumatoid diseases and number of teeth. It was found that subjects with a BMI of 25–29.99 were about 30 per cent less likely to have teeth with deepened periodontal pockets (RR 0.7, CI: 0.6–0.9) compared with the reference category (BMI less than 25), whereas those with a BMI of 30 or more were about 10 per cent more likely to have periodontal infection compared with the reference category (RR 1.1, CI: 0.8–1.4).

Table 2. Association between BMI and the number of teeth with a pocket depth of 4 mm or more. Unadjusted relative risks (RR) with 95% confidence intervals (CI) of Poisson regression modelsa
VariableNumber of teeth with periodontal pockets ≥4 mm
Unadjusted RR (95% CI) n = 157
  1. a

    Number of teeth as an offset variable.

  2. Education was missing from 2 subjects, cohabitation from 2 subjects, dental visits from 2 subjects, toothbrushing from 2 subjects, IADL score from 1 subject and rheumatoid diseases from 13 subjects.

BMI
 <251.0
 25–<300.9 (0.7–1.1)
 ≥301.1 (0.9–1.4)
Gender
 Female1.0
 Male0.9 (0.7–1.1)
Age, continuous1.0 (1.0–1.0)
Education
 ≥7 years1.0
 7 years0.9 (0.7–1.1)
Cohabitation
 Living alone1.0
 Living with somebody1.1 (0.9–1.4)
Teeth with dental plaque
 ≤20% of teeth with dental plaque1.0
 21–50% of teeth with dental plaque1.6 (1.3–2.1)
 >50% of teeth with dental plaque1.7 (1.3–2.1)
Dental visits
 Regular1.0
 Symptom-based or never1.7 (1.4–2.0)
Toothbrushing
 Twice a day or more often1.0
 Once a day or more seldom0.7 (0.5–1.0)
IADL score
 7–81.0
 0–61.2 (0.9–1.5)
Diabetes
 No1.0
 Yes1.0 (0.8–1.4)
Rheumatoid diseases
 No1.0
 Yes0.5 (0.4–0.8)

The role of underweight, BMI <18.5, was also studied. There were only five underweight participants, and they did not have more periodontal disease than other participants; unadjusted RR for deepened periodontal pockets was 0.6 (CI: 0.4–1.2) (other participants used as a reference category).

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. Author contributions
  10. References

The main finding was that 75-year-old or older persons with elevated body weight were not more likely to have signs of periodontal infection. This finding is in accordance with previous studies showing no association between BMI and the combined measure of periodontal pocket depth and attachment loss among 60–90-year-old persons (Al-Zahrani et al, 2003) and attachment level among 50–73-year-old persons (Torrungruang et al, 2005), but contrary to the findings of Linden et al (2007), who reported an association between BMI and periodontitis in 60–70-year-old men.

In this study, an attempt was made to study the role of underweight, BMI <18.5 in relation to periodontal infection. There were only five persons who were underweight, and for this reason, we cannot make any conclusions about the role of underweight as a risk for periodontal infection.

Several earlier studies (reviewed by Chaffee and Weston, 2010; Suvan et al, 2011), including also longitudinal studies (Saxlin et al, 2010; Morita et al, 2011), have reported an association between body weight and periodontal infection among young and middle-aged persons, whereas the findings of earlier studies made among older populations are less consistent. The reason why an association has not been found among older people is not known, but it is easy to speculate that it is caused by age-related physiological changes such as changes in body composition, especially a loss of subcutaneous fat and a gain of visceral fat, which have been shown to act in a less functional manner among older persons than among younger persons (Zafon, 2009). This explanation is also in accordance with the findings of Saito et al (2001), who have earlier reported that visceral fat is associated with periodontal infection in younger people, but not in middle-aged or older age groups.

Role of confounding

Another possibility is that there is no true age modification, but rather the contradiction between the results from different age groups is due to biases or imperfect controlling for confounders related, for example, to oral health behaviour or non-oral diseases. In fact, the possibility of confounding as an explanation is supported by the observation that participants with a BMI of 30 or more had, for example, more diabetes and unhealthier oral health habits than subjects with a BMI less than 30. Although we do not consider that diabetes essentially confounded the association between BMI and the number of teeth of deepened periodontal pockets–because diabetes seemed not be associated with the outcome variable (Table 2)–there may be other non-oral diseases or health-related behaviour that may cause confounding.

The role of confounding was reduced by using a study population, which was quite homogenous in terms of age, geographical distribution and ethnic origin. In addition, the effect of confounding was controlled by making several restrictions. In this study, we were able to exclude smokers and institutionalized participants, but not diabetics due to the high prevalence of diabetes, 18 diabetics.

We controlled for the effect of potential confounders, those factors that are known to be predictive for periodontal infection and which were not controlled by restriction, by means of multivariate models. Despite restriction and the use of multivariate models, it is possible that confounding existed. It may be related to the number of remaining teeth or previous smoking. The former is due to the mechanistic nature of statistical methods, which are not capable of fully taking into account all aspects related to the number of remaining teeth. The latter, because even after quitting, smoking may exert an effect on the periodontium. However, there were only a few subjects who had quit smoking quite recently, one in 2002, one in 2000, one in 1989 and others before that, meaning that in most cases, the time elapsed since quitting was long enough to eliminate the toxic effect of smoking on the periodontium. Also the fact that the outcome is the number of teeth with deepened pockets, which reflects the current condition of the periodontium, reduces the effect of past smoking.

Role of selection bias

The magnitude of the effect of selection bias is difficult to estimate. In Finland, where edentulousness is common, older people with their own natural teeth represent a selected older population. This selection can be related to both behavioural and constitutional factors. The former would mean that the study population overrepresents subjects with very good health behaviour, related possibly to underlying attitudinal factors. The latter would mean that selection can be related to susceptibility to infectious diseases, for example. This can lead to a situation where the most vulnerable subjects would not belong to the study population because they have lost their teeth, are hospitalized or have died earlier.

There may also exist selection at tooth level because it is possible that vulnerable subjects may have lost their teeth selectively, of which effect it is impossible to assess. This potential selection may be related to the anatomy of teeth and their positions in the alveolar ridge. On the other hand, as seen in Table 1, the means and distributions of teeth were fairly similar in the BMI categories indicating that no essential differences lie between subjects with different BMI at least in relation to the number of teeth.

Selection bias may also be related to selective participation. In this study, 27 subjects refused to come to the oral clinical examination, the reasons being, for example, ‘being tired’ or ‘having no teeth’. The effect of non-participation cannot be assessed, but some attempts were made to find out whether selective participation had any effect. In these analyses, it was found that gender did not have an effect on participation, but it seemed that those who participated were younger and more educated, on average. Bearing in mind the fact that high age and low education are both related to periodontal diseases as well as non-oral diseases, it can reasonably be expected that non-participation by such individuals does not decrease, but rather increases the credibility of the results by reducing the magnitude of confounding.

Other alternative explanations

In this study, the lowest risk estimates for periodontal infection were found in the BMI category of 25–29. This could be due to the fact that the optimal BMI in relation to overall health for old people of this age is indeed higher than in the general population. This finding would be in line with an earlier observation where, among old people, the optimal BMI for survival was found to be 24–29 (Dey et al, 2001). Another explanation, of course, is that the lowest risk estimates in the BMI category 25–29 are due to chance. And, it must be said that despite the fairly high participation rate in our study, the overall number of participants was fairly low, which increased the role of chance in the results.

To exclude the possibility that the reference category did not represent subjects that can be considered normal, the role of underweight subjects in this study population was studied. There were only 5 persons whose BMI was under 18.5, and when they were contrasted to other subjects, it was found that they did not have more periodontal infection than the others. This together with the low number of underweight persons suggests that the fairly low risk estimates in other groups are not due to an abnormal reference category.

Strengths and limitations

It is also worthwhile to emphasize that measuring of periodontal infection differs between studies, although we feel the discrepancy in the results is not primarily due to different measurements of periodontitis/periodontal infection. In this study, we used a continuous outcome, that is, the number of teeth with deepened periodontal pockets, which focuses on the presence and extent of current infection of the periodontium. Besides taking both extent and presence into account, using a continuous variable increased the accuracy of the measurement of periodontal infection compared with a situation where the outcome is based on a clinical diagnosis of periodontitis or is otherwise a categorical variable. One limitation related to the measurement of periodontal infection is that only the deepest periodontal pocket of each tooth was registered, which might have underestimated the overall number of periodontal pockets for a participant. However, the effects are most likely small, because most of study subjects had a fairly low number of teeth with periodontal pockets and there were a fairly large number of subjects without any teeth with deepened periodontal pockets.

To restrict the time taken to complete the clinical examination, clinical attachment level (CAL) was not measured and for that reason it was not available as an outcome measure. The reason pocket depth was preferred instead of CAL was that the data were based on the GeMS study, which was an intervention study aimed at improving oral health, and the presence of deepened periodontal pockets is a standardized and generally accepted measure for assessing periodontal treatment need.

Lastly, one limitation of the study is a lack of assessment of repeatability (intraexaminer kappa) and concordance between examiners (interexaminer kappa), which could not be assessed because no repeated measurements of dental status were made.

Conclusion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. Author contributions
  10. References

It can be concluded that within the limitations of this study, including small sample size, possibility of confounding and other biases, the results do not provide evidence that elevated body weight would be a risk for periodontal infection among older people. More studies, preferably with large study populations, are needed on the role of elevated body weight as a risk for periodontal infection among older people.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. Author contributions
  10. References

This GEMS study was financially supported by the Social Insurance Institute and the City of Kuopio.

Author contributions

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. Author contributions
  10. References

RO has made partly the drafting. He has participated in the interpretation of the data, revised the article and participated in final approval of the article. A-MS has made partly the drafting. She has participated in the interpretation of the data. She has revised the article and participated in final approval of the article. KK has planned the study, made oral clinical examinations. She has revised the article and participated to final approval. MK has participated in interpretation, revision and final approval of the article. PR has planned the study, made oral clinical examinations. She has revised the article and participated to final approval. SH has contacted A-MS and given the data her to analyse it. She has revised the article and participated to final approval. RS has planned the study and participated to implementation of the study. He has made revision of the article and participated to final approval. PY has participated in interpretation and revision of the article and participated to final approval.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. Author contributions
  10. References