Abstracts from this study were presented at the Cognitive Aging Conference 2008 (http://www.cos.gatech.edu/cac/program.pdf; p. 7) and the International Association for Dental Research, 2008 (http://iadr.confex.com/iadr/2008Toronto/techprogram/abstract_104698.htm).
Tooth Loss and Periodontal Disease Predict Poor Cognitive Function in Older Men
Article first published online: 1 APR 2010
© 2010, Copyright the Authors. Journal compilation © 2010, The American Geriatrics Society
Journal of the American Geriatrics Society
Volume 58, Issue 4, pages 713–718, April 2010
How to Cite
Kaye, E. K., Valencia, A., Baba, N., Spiro, A., Dietrich, T. and Garcia, R. I. (2010), Tooth Loss and Periodontal Disease Predict Poor Cognitive Function in Older Men. Journal of the American Geriatrics Society, 58: 713–718. doi: 10.1111/j.1532-5415.2010.02788.x
- Issue published online: 1 APR 2010
- Article first published online: 1 APR 2010
- tooth loss;
- periodontal diseases;
- dental caries;
- Mini-Mental State Examination;
- cognitive impairment
OBJECTIVES: To determine whether rates of tooth loss, periodontal disease progression, and caries incidence predict cognitive decline in men.
DESIGN: Prospective study.
SETTING: Community-dwelling men enrolled in the Veterans Affairs Dental Longitudinal Study.
PARTICIPANTS: Five hundred ninety-seven dentate men aged 28 to 70 at study baseline who have been followed up to 32 years.
MEASUREMENTS: Oral examinations were conducted approximately every 3 years. Periodontal disease measures included probing pocket depth and radiographic alveolar bone height. Participants underwent cognitive testing beginning in 1993. Low cognitive status was defined as less than 25 points or less than 90% of the age- and education-specific median on the Mini-Mental State Examination (MMSE) and less than 10 points on a spatial copying task.
RESULTS: Each tooth lost per decade since the baseline dental examination increased the risks of low MMSE score (hazard ratio (HR)=1.09, 95% confidence interval (CI)=1.01–1.18) and low spatial copying score (HR=1.12, CI=1.05–1.18). Risks were greater per additional tooth with progression of alveolar bone loss (spatial copying: HR=1.03, CI=1.01–1.06), probing pocket depth (MMSE: HR=1.04, CI=1.01–1.09; spatial copying: HR=1.04, CI=1.01–1.06), and caries (spatial copying: HR=1.05, CI=1.01–1.08). Risks were consistently higher in men who were older than 45.5 at baseline than in younger men.
CONCLUSION: Risk of cognitive decline in older men increases as more teeth are lost. Periodontal disease and caries, major reasons for tooth loss, are also related to cognitive decline.
Emerging evidence suggests that a history of tooth loss may be associated with cognitive impairment in community-dwelling older adults.1–7 Cross-sectional analyses of nationally representative data from several countries using a variety of cognitive function tests show greater odds of impaired test performance or prevalence of dementia associated with a low number of teeth remaining.1–5 In a case–control study of elderly identical twins discordant for dementia, loss of at least half of the teeth before age 35 was significantly related to risk of Alzheimer's disease and other dementias,5 although an examination of data from the National Health and Nutrition Examination Surveys (NHANES III, 1988–1994) found that poverty, cardiovascular risk factors, and physical activity attenuated the positive associations between number of missing teeth and poor performance on three tests of cognitive function.3 Findings from prospective studies are inconclusive. Elderly women with nine or fewer teeth had twice the risk of developing dementia as those with 10 or more teeth.6 Edentulous persons aged 70 and older scored lower on a battery of cognitive tests than those with teeth, but the difference did not persist after adjustment for age and intelligence quotient.4
A decline in cognitive function would be expected to disrupt normal activities of daily living such as oral hygiene and dental care, and this would explain the cross-sectional associations. The prospective findings from a previous study5 suggest that the reverse direction of events is also plausible—that tooth loss precedes and plays a causal role in cognitive decline. Several physiological mechanisms have been proposed to explain a causal pathway. Low intake of several B vitamins has been linked to cognitive decline,8 and loss of teeth negatively affects ability to consume recommended levels of many foods and nutrients.2,9 Pro-inflammatory factors derived from the body's response to chronic periodontal infection may travel through the systemic circulation to the brain where they exacerbate inflammatory processes and vascular pathologies.10,11 Reports that cognitive function is related to measures of periodontal disease as well as tooth loss strengthen this latter hypothesis. In the NHANES 1988 to 1994 and 2001 to 2002 surveys, poor performance on the Digit Symbol Substitution Test or Serial Digit Learning Test was associated with clinical periodontal disease measures after adjusting for multiple confounding factors.3,12 In further analyses, greater odds of difficulty recalling a three-word sequence or performing serial subtractions in older adults with the highest Porphyromonas gingivalis exposure levels was reported,13 although only one study has prospectively examined periodontal disease and cognitive status. Women with moderate or severe alveolar bone loss tended to have greater risk of developing dementia than women with less-severe disease, but the finding was not statistically significant.6
The purpose of this study was to determine whether rate of tooth loss predicted poor cognitive function in a cohort of community-dwelling men. In addition, whether periodontitis and caries, the primary causes of tooth loss, predict cognitive function was examined.
Participants were members of the U.S. Department of Veterans Affairs (VA) Dental Longitudinal Study (DLS), a prospective observational study of oral health in men that began in 1968 as an ancillary to the Normative Aging Study.14 The original DLS cohort comprised 1,231 medically healthy men. This closed-panel longitudinal study cohort is a convenience sample of adult men, originally aged 24 to 84, recruited from the Boston, Massachusetts, metropolitan area. The men are not patients of VA hospitals, but receive dental and medical care in the private sector and are fairly representative of the adult male population in the greater Boston area. The cohort has been well characterized with regard to health-related predictors of cognitive function.8,15,16 Data from the dental examinations and cognitive testing were available for 597 dentate men.
Comprehensive dental and physical examinations were conducted approximately every 3 years using commonly available, standardized measurement techniques. At each dental examination, a calibrated periodontist assessed number of teeth remaining and scored levels of maximum probing pocket depth on each tooth following criteria used consistently over the course of the study.17,18
Caries and restorations on each of five tooth surfaces were recorded. Alveolar bone loss (ABL), as a percentage of root length in 20% increments, was measured on the mesial and distal aspects of each tooth by superimposing a ruler onto periapical radiographs.19 Participants completed questionnaires on oral hygiene practices, professional dental care, and smoking history. Dental examinations conducted between 1968 and 2002 are included in this report.
Beginning in 1993, the Mini-Mental State Examination (MMSE) and a spatial copying task (SCT) were administered at each examination. The MMSE is widely used as a global test of cognitive function to screen for dementia and tests orientation, immediate memory, attention and calculation, short-term recall, and language.20 Correlations between MMSEs repeated at short intervals in cognitively healthy populations are generally 0.80 to 0.95.21 Possible scores range from 0 to 30; mild impairment is indicated by scores between 21 and 24. MMSE scores were also converted to a percentage of the median for the participant's age and educational level,22 resulting in values that ranged from 54% to 116%. For the SCT, a battery of figures derived from the Consortium to Establish a Registry for Alzheimer's Disease neuropsychological battery23 and the Developmental Test of Visual-Motor Integration,24 the participant is shown and asked to copy nine geometric designs. The designs have an increasing level of difficulty, and test scores reflect the number of figures correctly drawn, weighted according to difficulty.8 Two raters scored each participant's SCT figures, and agreement was achieved between them. Possible SCT scores range from 0 to 26. Cognitive test results used in the analyses were obtained at between 1993 and 2001.
Data on alcohol and dietary intake were obtained from food frequency questionnaires administered at each examination since 1984. Diagnoses of comorbidities (cardiovascular disease (CVD), coronary heart disease (CHD), cancer, diabetes mellitus) were obtained during examination by the study physician or from death certificates where applicable. Years of education, smoking, and frequency and amounts of all prescription and nonprescription medications were obtained using the questionnaire at each examination. Weight (pounds) and height (inches) were measured on a beam balance and stadiometer, respectively, and converted to metric units. Body mass index was computed as weight divided by height squared. The institutional review board of the VA Boston Healthcare System approved the protocols. All participants gave informed consent on approved forms before each examination.
Tooth loss rate after baseline was computed as the number of teeth lost per decade of follow-up. Alveolar bone loss progression and probing pocket depth progression were defined, respectively, as numbers of teeth that showed decrease in bone height of at least 40% from baseline or were lost, and increase in pocket depth of 2 mm or more from baseline or were lost. Caries incidence was estimated according to number of sound teeth at baseline that developed caries, were restored, or lost. Tooth loss was included in the periodontal disease and caries progression estimates to avoid underestimating disease due to missing values for absent teeth.
A low outcome on the MMSE was defined as follows. For men in this cohort with scores of 21 to 24, the age- and education-specific median ranged from 72% to 89%. Therefore, a low MMSE score was defined as less than 25 points or less than 90% of the age- and education-specific median. According to this criterion, 95 men had a low MMSE score. A low SCT outcome was defined as a score more than 1 standard deviation below the mean, yielding a cutoff of 10 points and 185 men with a score less than 10. Several other cutoff points were also examined for each cognitive function test. Two hundred fifty-five and 25 men scored below 95% and 80%, respectively, of the age- and education-specific median MMSE score, and 211 and 124 scored below 11 and 9 points, respectively, on the SCT.
Potential confounding was examined according to age; years of education; smoking; body mass index; regular medication use (aspirin, nonsteroidal anti-inflammatory drugs); development of CHD, stroke, hypertension, CVD, cancer, or diabetes mellitus during follow-up; and intake of alcohol, coffee and tea, folate, vitamin B6, and vitamin B12 using bivariate correlations and independent sample t-tests. Based on the results, variables were included as covariates in multivariable models if they were significant in bivariate analyses and remained in the multivariable models at P<.15.
Risks of developing low MMSE or SCT scores in relation to tooth loss, progression of periodontal disease, and progression of caries were estimated using adjusted hazard ratios (HRs) and 95% confidence intervals (CIs) obtained from Cox proportional hazards regression models stratified according to median age (≤45.5 vs >45.5). Time to event, which varied from 20 to 32 years, was defined as the interval between the DLS baseline and year of the lowest cognitive test score. Analyses were performed using SAS, version 9.1 (SAS Institute, Inc., Cary, NC).
Between 1993 and 2001, 387 men had at least one repeat cognitive test. The attributable risk percentage of cognitive test score deterioration that could be attributed to loss of any teeth was estimated.25
Of the 1,231 original enrollees, 659 dentate men participated in dental examinations between 1993 and 2001 at the time when cognitive tests were also administered. Of these, 62 did not undergo any cognitive testing, leaving 597 for the analyses. The 62 men without cognitive test results were older at the dental study baseline but otherwise were comparable with respect to prevalence of smoking and CHD, mean years of education, alcohol intake, tooth loss, and progression of periodontal disease and caries.
Men who developed low cognitive test status were older; had less education; and were more likely to have CHD, lose more teeth, and develop more caries than men whose test scores remained normal (Table 1).
|Low (n=95)||Normal (n=500)||Low (n=185)||Normal (n=412)|
|At dental baseline (1968–1973)|
|Age, mean ± SD||48.4 ± 7.7||45.3 ± 6.9||47.8 ± 7.2||44.8 ± 6.9|
|Education, years, mean ± SD||13.3 ± 2.6||14.6 ± 2.6||14.1 ± 2.6||14.6 ± 2.6|
|Ever smoked, %||41||45||40||46|
|Number of teeth, mean ± SD||23.8 ± 5.6||24.2 ± 5.3||24.1 ± 5.3||24.2 ± 5.4|
|Number of teeth with ≥1 decayed or filled surfaces, mean ± SD||15.1 ± 5.3||15.3 ± 5.1||14.9 ± 4.8||15.5 ± 5.2|
|Number of teeth with probing pocket depth ≥4 mm, mean ± SD||4.0 ± 3.9||3.7 ± 4.3||4.3 ± 4.6||3.5 ± 4.0|
|Number of teeth with alveolar bone loss ≥40%, mean ± SD||0.7 ± 1.4||0.6 ± 1.7||0.9 ± 2.3||0.5 ± 1.3|
|At time of first cognitive test (1993–2001)|
|Age, mean ± SD||71.7 ± 8.3||68.9 ± 7.3||71.4 ± 7.7||68.4 ± 7.2|
|Years elapsed since baseline, mean ± SD||23.8 ± 1.8||24.1 ± 1.8||24.0 ± 1.8||24.1 ± 1.8|
|MMSE score, mean ± SD||23.8 ± 1.8||28.1 ± 1.2||26.8 ± 2.4||27.9 ± 1.7|
|MMSE score as % of age and education-specific median, mean ± SD||86.4 ± 8.8||97.8 ± 4.2||93.8 ± 8.2||97.0 ± 5.6|
|Weighted SCT score, mean ± SD||10.0 ± 5.5||15.0 ± 6.0||6.8 ± 2.2||17.5 ± 4.2|
|Alcohol intake, g/day, mean ± SD||7.5 ± 10.9||13.2 ± 15.6||11.1 ± 14.3||12.8 ± 15.4|
|History of coronary heart disease, %||24||15||22||15|
|Any tooth loss, %||85||74||81||73|
|Number of teeth lost, mean ± SD||5.7 ± 6.0||3.8 ± 4.8||5.0 ± 5.6||3.7 ± 4.8|
|Number of teeth with new caries or restorations, mean ± SD||13.2 ± 5.1||11.7 ± 5.6||12.7 ± 5.2||11.6 ± 5.6|
|Number of teeth with probing pocket depth progression, mean ± SD||15.8 ± 6.6||15.1 ± 8.0||15.4 ± 7.8||15.2 ± 7.8|
|Number of teeth with alveolar bone loss progression, mean ± SD||11.8 ± 7.7||10.9 ± 9.0||11.8 ± 8.9||10.7 ± 8.8|
Men who were above the median age at baseline (45.5) were nearly 10 years older at the time of cognitive testing, 1.9 times as likely to have a low SCT score (P<.001) and 1.7 times as likely to have a low MMSE score (P=.01) as men below the median age. On average, this older group had more teeth with 40% or more alveolar bone loss existing at baseline (1 vs 0.5 in younger men) and experienced more tooth loss (5 teeth lost per decade vs 3 per decade in younger men) during follow-up, but rates of periodontal disease progression were similar.
Higher rates of tooth loss and periodontal disease progression independently increased the risks of low cognitive test scores during follow-up (Table 2). For each tooth lost per decade, the overall risk of a low MMSE score (<25 points or <90% of age- and education-specific median) or low SCT score (<10 points) increased 9% to 12%. For each tooth that had progression of alveolar bone loss or probing pocket depth, the overall risks of low scores increased 2% to 5%. Development of new caries or restorations was associated with greater risk of a low SCT score. When stratified according to baseline age, the risks were consistently higher in the older than the younger men. The results did not change when smoking, diabetes mellitus, additional comorbidities, medication use, or dietary intakes were included in the models. The HR estimates were similar when the definitions of low scores were changed to less than 95% or less than 80% on the MMSE, and less than 11 or less than 9 on the SCT.
|Risk per Additional Tooth Lost||Adjusted Hazard Ratio (95% Confidence Interval)|
|Low MMSE Score*||Low SCT Score†|
|All (n=595)||≤45.5 (n=296)||>45.5 (n=299)||All (n=597)||≤45.5 (n=297)||>45.5 (n=300)|
|Per decade||1.09 (1.01–1.18)||1.08 (0.92–1.26)||1.10 (1.01–1.21)||1.12 (1.05–1.18)||1.10 (1.01–1.20)||1.13 (1.05–1.21)|
|With alveolar bone loss progression per decade||1.03 (1.00–1.07)||1.00 (0.94–1.07)||1.05 (1.01–1.09)||1.03 (1.01–1.06)||1.02 (0.97–1.06)||1.04 (1.02–1.07)|
|With pocket depth progression per decade||1.04 (1.01–1.09)||1.02 (0.95–1.10)||1.05 (1.01–1.10)||1.04 (1.01–1.06)||1.02 (0.96–1.07)||1.04 (1.01–1.07)|
|With new caries or restorations per decade||1.02 (0.97–1.08)||1.01 (0.94–1.09)||1.06 (0.96–1.16)||1.05 (1.01–1.08)||1.02 (0.98–1.06)||1.13 (1.07–1.21)|
Of the men who scored 26 points or higher on the initial MMSE, 23 (5%) later fell below this cutoff point, and of those who scored 10 points or higher on the initial SCT, 59 (13%) later fell below this score. The attributable risk percentages for low scores on the MMSE and SCT due to tooth loss were 46% and 9%, respectively. That is, of the men in the cohort who lost any teeth, 9% to 46% of those who scored low on a cognitive function test would not have done so had they kept all the teeth they had at baseline.
These findings indicate that, in older men, rates of tooth loss and periodontal disease progression during adulthood independently predict performance on the MMSE and SCT cognitive tests. Development of caries appeared to increase the risk of poor performance as well. The greater risks were most consistent in men who were older than 45.5 at the dental baseline. This finding may reflect the higher prevalence of periodontal disease that existed in older men at baseline or some other unique characteristic of the older cohort.
These results indicate that, for each tooth lost per decade, the risk of a low cognitive test score increased 9% to 12%. Based on these estimates, if 12 teeth were lost per decade, the risk of impaired cognition would approach 100%. This prediction was borne out in this cohort. For example, nearly all (86%) of the participants whose tooth loss rate was 12 per decade or more were classified as having a low SCT score. Such a rapid rate of loss in older adults is most probably a sign of severe periodontal disease.
A unique strength of this study is the prospective design, with comprehensive data on tooth loss, periodontal disease, and caries recorded by trained dental examiners during a 32-year follow-up interval. The measures of periodontal disease included radiographic alveolar bone loss, which is an indicator of cumulative periodontitis history. In addition, there was extensive information on important covariates obtained during serial physical examinations. Cognitive test scores at the dental baseline in 1968 to 1973 were not available to exclude preexisting dementia, although it is unlikely that these men were cognitively impaired at baseline because they continued to participate in examinations, mail surveys, and many other assessments for at least 20 years before cognition testing began. Other limitations of this study include the absence of data on women and ethnic minorities and that the men in the analyses were younger than those who did not complete the cognitive tests. As a result, few of the men were moderately or severely impaired.
These findings are consistent with previous literature that showed greater risk of impaired cognitive performance associated with various indicators of poor oral health.1–7,12,13 Those findings were also extended by demonstrating that, in this cohort, the detrimental changes in oral health preceded rather than resulted from cognitive decline.
Dementia and cognitive impairment are growing worldwide health problems. By 2040, it is expected that dementia will affect more than 81 million people worldwide.26 In the United States, Alzheimer's disease, the most common form of dementia, rose from being the 14th most common cause of death in 1995 to 7th just 10 years later.27 Because there are no effective treatments for reversing dementia, it is important to identify modifiable risk factors in an attempt to intervene and reduce the incidence of disease. Attributable risk estimates from this cohort suggest that at least 14 cases of cognitive decline and perhaps as many as 37, detected using the SCT and MMSE tests, respectively, would not have occurred if tooth loss had been prevented.
An inflammatory model has been invoked to explain degeneration of brain neuronal cells and risk of dementia,10 but a number of other mechanisms might also explain these findings. The similarities between risk factors for CVD and Alzheimer's disease raise the possibility of a causal biological pathway linking these two common disorders, whereas some investigators propose that the two diseases are independent outcomes of the atherogenic process.28 Poor nutritional status, especially in relation to B vitamins,8,29 may also play a role in the progression of both diseases. Although the analyses in the current study evaluated these nutritional factors and controlled for the development of CHD and its common predictors, there may be other underlying but unmeasured risk factors that contribute to the associations found. Another possibility is that a pro-inflammatory phenotype is the common factor linking periodontal disease, CVD, and cognitive decline.30 Presence of the apo-E å4 genetic marker, a risk factor for Alzheimer's disease, attenuated the association between number of teeth present and prevalence of dementia.6
This study of community-dwelling men showed that rates of tooth loss and periodontal disease progression predicted subsequent decline in cognitive function. These findings lend support to the hypothesis that oral health is an important modifiable determinant of cognitive function and suggests that peripheral inflammation contributes to the development of dementia and cognitive impairment.
The views expressed in this paper are those of the authors and do not necessarily represent the views of the VA.
Conflict of Interest: The editor in chief has reviewed the conflict of interest checklist provided by the authors and has determined that the authors have no financial or any other kind of personal conflicts with this paper.
Dr. Spiro is the recipient of a VA Clinical Science Research and Development Research Career Scientist award. Dr. Garcia is a recipient of a VA Career Development Award in Health Services Research from the VA Health Services Research and Developemnt Service, and National Institute of Dental and Craniofacial Research Midcareer Investigator Award in Patient-oriented Research (K24 DE00419).
Financial Disclosure: The Dental Longitudinal Study and Normative Aging Study are components of the Massachusetts Veterans Epidemiology Research and Information Center, which is supported by the VA Cooperative Studies Program. Cognitive data collection was supported by the Clinical Science Research and Development Service of the VA, the National Institutes of Health (Grants R01-AA08941, R01-AG13006, R01-AG14345, R01-AG18436, 5-P42-ES05947, R01-ES05257), and the U.S. Department of Agriculture, Agricultural Research Service (Contract 53-K06-510).
Author Contributions: Drs. Spiro and Garcia proposed the study. Drs. Garcia, Krall Kaye, and Spiro acquired the data. Dr. Krall Kaye and Ms. Valencia conducted statistical analyses. Dr. Krall Kaye drafted the report, and all authors contributed to its revision. Dr. Krall Kaye had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Sponsor's Role: The study sponsors had no role in design and conduct of the study; collection, management, analysis, and interpretation of the data; and preparation, review, or approval of the manuscript.
- 14The Veterans Administration Longitudinal Study of Oral Health. Aging Hum Devel 1972;3:125–137., , et al.
- 19Alveolar bone loss as related to oral hygiene and age. J Periodontol 1959;30:7–16.
- 24The Beery-Buktenica Developmental Test of Visual-Motor Integration (VMI), 4th Ed. Carlsbad, CA: Modern Curriculum Press, 1996.
- 25Epidemiology in Medicine. Boston: Little, Brown & Co, 1987.,
- 26Global prevalence of dementia: A Delphi consensus study. Lancet 2005;366:12–17., , et al.