| Sign In to gain access to subscriptions and/or personal tools. |
Relationship of Periodontal Disease and Edentulism to Stroke/TIA
1 School of Dentistry, CB#7450, University of North Carolina, Chapel Hill, NC 27599-7450, USA; and Correspondence: *corresponding author, john_elter{at}unc.edu
Periodontitis has been shown to increase the systemic inflammatory response, which has been implicated in atherosclerosis and cerebrovascular events. We hypothesized an association between periodontitis or edentulism and Stroke/TIA in the ARIC Study. Data on 9415 dentate and 1491 edentulous adults included demographics, cardiovascular outcomes, lifestyle, laboratory measures, and, for 6436 of the dentate, a dental examination. The dependent variable was Stroke/TIA, and the exposure was extent (%) of attachment level 3+ millimeters (AL). Quartiles of AL and edentulism were compared for Stroke/TIA using odds ratios (OR) and 95% confidence intervals (CI), and confounders were controlled by logistic regression. Stroke/TIA was prevalent in 13.5% of periodontal examinees, 15.6% of dentate non-examinees, and 22.5% of edentulous persons. The highest quartile of AL (OR 1.3, CI 1.02-1.7) and edentulism (OR 1.4, CI 1.5-2.0) were associated with Stroke/TIA.
Key Words: periodontal diseases • edentulism • stroke • transient ischemic attack • epidemiology
Periodontitis has been related to stroke (Syrjanen et al., 1989; Beck et al., 1996; Grau et al., 1997; Ziegler et al., 1998; Wu et al., 2000), but findings have been inconsistent (Howell et al., 2001). Stroke is defined as rapidly developing signs of focal (or global) disturbance of cerebral function lasting more than 24 hrs, when the blood supply to the brain is interrupted (Thorvaldsen et al., 1995). A transient ischemic attack (TIA) presents with similar symptoms, which often resolve within one hour, although they may persist for up to 24 hrs. Several risk factors have been identified for stroke: coronary heart disease (CHD) (National Institute of Neurological Diseases and Stroke, 2002a), male sex (Williams et al., 1999; Kolominsky-Rabas et al., 2001; Wolfe et al., 2002), African-American race (Wolfe et al., 2002), age (Williams et al., 1999), smoking (Hankey, 2000), alcohol (for hemorrhagic stroke) (McDonnell et al., 2000), hypertension (Beckett, 2001), hypercholesterolemia (Ansell, 2000), diabetes (Sacco, 2002), and carotid artery disease (National Institute of Neurological Diseases and Stroke, 2002a). The purpose of this cross-sectional study was to determine whether persons with prevalent Stroke/TIA in the ARIC Study were more likely to have periodontitis or be edentulous compared with persons without Stroke/TIA.
The ARIC Study investigated the etiology and natural history of atherosclerosis and clinical cardiovascular disease in four US communities (ARIC Investigators, 1989). The Dental ARIC Study was performed on a subset of dentate participants (1996-1998), and all information for this cross-sectional study was collected at Visit 4. The study sample consisted of 6436 persons with a periodontal evaluation and determination of Stroke/TIA. An additional 1491 edentulous and all dentate (N = 9415) were compared for Stroke/TIA. Human subjects participated in ARIC after providing informed consent to a protocol that was reviewed and approved by the UNC School of Dentistry Committee on Research Involving Human Subjects. Periodontal attachment level was measured at 6 sites per tooth for all teeth (Beck et al., 2001). Extent of attachment level 3+ millimeters was the main exposure variable, and was the number of sites with AL 3+ millimeters divided by the number of measured sites, times 100. Extent of AL was categorized into quartiles: 0% to < 6.5%, 6.5% to < 15.4%, 15.4% to < 31.4%, and > 31.4%. For a separate comparison among the entire ARIC Visit 4 cohort, the exposure variable was self-reported edentulous vs. dentate. The outcome was prevalent stroke (ischemic or hemorrhagic) or TIA at Visit 4, which consisted of self-reported history of Stroke/TIA diagnosed by a physician, neurological deficits of rapid onset assessed by a structured interview, and Stroke/TIA captured by ARIC hospital surveillance up to Visit 4 (The ARIC Investigators, 1987b). Covariates included sex; age; race and ARIC field center; education (< 11 yrs, 12-16 yrs, or 17+ yrs); hypertension (defined as systolic blood pressure > 140 mm/hg, or diastolic blood pressure > 90, or self-reported antihypertensive medications in the preceding 2 wks (The ARIC Investigators, 1987a)); smoking (current, former, never) stratified by intensity (light, < 20 pack-years; heavy, 20+ pack-years); diabetes [fasting blood glucose > 126 mg/dL (> 200 mg/dL if not fasting) or diabetes medication]; prevalent coronary heart disease (CHD); plasma LDL and HDL and triglycerides in mg/dL; and body mass index (BMI). Responses to the questions "Do you currently have a dentist?", method of payment for health care (insurance without Medicare, insurance and Medicare, Medicaid, no insurance, or other insurance), and income (< $12,000, $12,000-$50,000, or $50,000+) were also included. We performed analyses to determine associations of AL with Stroke/TIA among the dentate examined for periodontal disease, and separately for edentulism among all edentulous and dentate. Logistic regression was used to calculate odds ratios (OR) and 95% confidence intervals (CI). Confounders were selected for their relationship to AL and Stroke/TIA, and interactions of covariates with AL were also evaluated. Covariates were retained if significant at P < 0.05 or if they confounded the AL-Stroke/TIA relationship (defined as change of 10% in the coefficient for AL). The association of AL with Stroke/TIA was evaluated for trend by logistic regression.
Stroke was prevalent in 7.4%, TIA in 6.4%, and Stroke/TIA in 13.5% of the 6436 periodontal examinees (Appendix Table 1, www.dentalresearch.org). All three outcomes were more prevalent among the 2979 dentate who were not examined for periodontal disease (Stroke 8.4%, TIA 7.4%, and Stroke/TIA 15.6%), and most prevalent among the 1491 edentulous (Stroke 11.7%, TIA 11.0%, Stroke/TIA 22.5%). Stroke/TIA was associated with higher extent of AL (P = 0.0003, Appendix Table 1, www.dentalresearch.org). Prevalence of Stroke/TIA increased with each quartile of AL (P = 0.0016, Appendix Table 1, www.dentalresearch.org). Also related to stroke/TIA were edentulism, female gender, education, smoking, diabetes, hypertension, Medicaid, income, age, and CHD. The factor most strongly related to Stroke/TIA was CHD. Health insurance other than Medicare was related to lower prevalence of Stroke/TIA. In a crude model for AL and Stroke/TIA, the highest quartile of AL had about one and a half times the odds of Stroke/TIA, compared with the lowest quartile (Table). The addition of age, sex, race/center, education, income, method of payment, and having no dentist to the model had no effect. Income, method of payment, and not having a dentist were not significant and dropped out of the model. The addition of smoking, hypertension, diabetes, lipids, CHD, and BMI had no further confounding effect, and lipids and BMI were dropped from the final model (Appendix Table 2, www.dentalresearch.org). In separate models, edentulous persons had 1.7 times the odds of stroke/TIA compared with all dentate (Table, last column); however, the addition of covariates weakened the association. Details of the edentulous model are presented in Appendix Table 7 (www.dentalresearch.org). Interactions of AL with covariates were evaluated, but none was significant. "Missing teeth" was not significant in a model with or without AL (Appendix Tables 5 and 6, www.dentalresearch.org), nor was the interaction of AL and missing teeth.
The concept underlying this study is that stroke is accompanied by significant morbidity and mortality, and that periodontitis may be an under-recognized risk factor creating a burden of bacterial pathogens, antigens, endotoxins, and inflammatory cytokines that contribute to the process of atherogenesis and thrombo-embolic events. Certain persons exhibit a higher expression of local gingival crevicular fluid and systemic mediators of inflammation. Persons who exhibit this exaggerated response to periodontitis may be at increased risk for stroke, due to the effects of circulating inflammatory products on the formation of atherosclerotic plaques (Offenbacher, 1996). This study shows a weak independent association of AL with Stroke/TIA. This association persisted despite adjustment for confounding by traditional CVD risk factors. Although every attempt was made to control for confounders, a weak association such as that observed in this study could be due to residual confounding.
An adjusted logistic model showed a significant trend of increasing odds of Stroke/TIA for each quartile or quintile of AL (P = 0.04, Appendix Tables 2 and 3, www.dentalresearch.org). A slightly stronger association was obtained for edentulous vs. dentate (Table
Our results are consistent with those of the previous reports examining this relationship (Syrjanen et al., 1989; Grau et al., 1997; Ziegler et al., 1998; Wu et al., 2000). One study (Howell et al., 2001) found a non-significant trend between self-reported periodontal disease and incident, non-fatal stroke among US male physicians, possibly because use of self-reported periodontal disease as an exposure variable may have led to misclassification. We chose not to set a priori definitions for periodontal disease, because there has been little agreement about which definition to use. AL is indicative of past periodontal disease experience more than it indicates the current burden of oral infection. If the systemic effects of chronic oral infection are quantified by extent of AL, that measure has been affected by loss of teeth, especially those lost to periodontal disease. While it is impossible to determine why teeth were lost, it may be that teeth with extensive AL had a greater probability of having been lost due to periodontal disease. In fact, each higher quartile of AL had 1.7 times fewer teeth remaining (P < 0.0001). This may be why edentulous persons had greater odds of Stroke/TIA compared with the dentate, even though their current burden of infection may have been minimal. However, the reasons for edentulism are impossible to determine, and may be related to caries or poor nutrition, rather than to periodontal disease (Joshipura et al., 1998). This may result in the relationship of edentulism to Stroke/TIA being confounded. We know that the edentulous are very different from the ARIC dental sample, in that their health is poorer in almost all respects. For these reasons, the observed relationship between edentulism and Stroke/TIA may actually be due to the presence of common risk factors rather than to any direct etiological mechanism. The dental ARIC sub-sample is known to be much healthier than the greater ARIC sample. This disparity results in part from the exclusion criteria for the periodontal examination, which included any need for antibiotic pre-medication. The ARIC Study discriminated between ischemic and hemorrhagic strokes in its hospital surveillance program, but not for the collection of self-reported stroke information. Therefore, it was not possible to determine the proportion of strokes of the hemorrhagic form, an event that has not been found to be associated with periodontal disease (Wu et al., 2000). We found that hemorrhagic strokes comprised 17.7% of strokes captured by the ARIC surveillance, which is consistent with reports from other studies (National Institute of Neurological Diseases and Stroke, 2002b). It is possible that the lack of specificity in the prevalent outcome may have attenuated the reported association between AL and Stroke/TIA. Other than AL, only hypertension, diabetes, education, CHD, and female sex were associated with Stroke/TIA in the final model (Appendix Table 2, www.dentalresearch.org). Except for female sex, these factors are consistent with well-established risk factors for stroke. The seemingly anomalous finding for sex may be attributed to the use of a combined outcome: While females had a higher rate for TIA, males had a higher prevalence of stroke. Stroke alone was not independently associated with AL. Selective survival and morbidity from stroke may have resulted in the inclusion of persons with less extensive periodontal disease, thereby biasing the association with stroke. Even though TIA and stroke share a common atherosclerosis etiology, TIA does not result in persistent morbidity. Thus, any relationship between periodontitis and TIA would be less susceptible to bias. No inference about causality should be made from the results of this cross-sectional study. ARIC is currently following participants for incident CVD events, and that report will appear shortly.
The authors thank the staff and ARIC participants for their important contributions. This research was supported by NIDCR grant R01DE 11551 and by contracts from NHLBI. Received for publication March 18, 2003. Revision received September 11, 2003. Accepted for publication September 17, 2003.
Journal of Dental Research, Vol. 82, No. 12,
998-1001 (2003) This article has been cited by other articles:
|
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
|
|
 |
|
Sign In to gain access to subscriptions and/or personal tools.
TOP
ABSTRACT
INTRODUCTION
