| Sign In to gain access to subscriptions and/or personal tools. |
Oral Health Indicators Poorly Predict Coronary Heart Disease Deaths
1 Department of Public Health, University of Helsinki, PO Box 41, 00014 University of Helsinki, and Correspondence: * corresponding author, Risto.Tuominen{at}Helsinki.Fi
Several earlier studies have suggested that development of coronary heart disease (CHD) is causally related to oral infections. The aim of this study was to investigate the association between oral health indicators and CHD deaths. Out of a nationally representative sample, 6527 men and women aged 30–69 years participated in the health examination with a dental check. Detailed oral health data included caries, periodontal and dental plaque status, presence of remaining teeth, and various types of dentures. Over a mean 12-year follow-up, persons dying of CHD were older and more often smoked, had hypertension, hypercholesterolemia, diabetes, and only a basic education compared with other persons. In univariate analyses, several oral health indicators were associated with CHD deaths. Adjustment for the established CHD risk factors reduced all these associations to statistical non-significance. The associations between oral health indicators and CHD are mostly explained by confounding factors, particularly those relating to health behavior.
Key Words: coronary heart disease • oral health • epidemiology
Several studies over the last decade have suggested that development of coronary heart disease (CHD) is related to various types of oral infections (Mattila et al., 1989, 1995; DeStefano et al., 1993; Genco, 1996; Loesche et al., 1998). Periodontal infections, in particular, have been suspected as a risk factor for cardiovascular diseases (Mattila et al., 1993; Beck et al., 1996, 1998; Grau et al., 1997; Beck and Offenbacher, 1998). Several hypothetical pathogenetic mechanisms have been postulated to explain the association between infectious agents and complications of atherosclerosis (Mattila et al., 1998; Mehta et al., 1998). Infectious agents relevant in oral health, such as Streptococcus sanguis (Herzberg and Meyer, 1996) and Actinobacillus actinomycetemcomitans (Mattila et al., 1998), have also been shown to have possible direct effects contributing to pathogenesis of atherosclerosis and thrombosis. The associations between infectious agents found in the oral cavity and myocardial events are also assumed to be reflected by the positive association observed between white blood cell count and serum albumin and myocardial events (Kweider et al., 1993; Gillum, 1994). Many of the studies demonstrating associations between oral infections and atherosclerosis have been based on small, selected samples (Mattila et al., 1989, 1993, 1995; Grau et al., 1997; Loesche et al., 1998). Oral infection measures have been estimated by indices which do not necessarily measure level or severity of infection (Mattila et al., 1989, 1993, 1995; Mattila, 1993; Paunio et al., 1993; Grau et al., 1997). Furthermore, in studies with a high proportion of older subjects, bias may have occurred because CHD is a common cause of mortality and morbidity among the elderly. In addition, many earlier studies have included only male subjects or few women. Previous studies have failed to show the relationship between oral infections and CHD among women. In a large non-concurrent cohort study by DeStefano et al. (1993), periodontal conditions appeared to be associated with CHD mortality. However, the oral hygiene index seemed to have an even stronger association, indicating that oral health indices may be general indicators of personal health behavior and practice, instead of being causally related to CHD risk. Another large epidemiological study (Joshipura et al., 1996) among male health care workers found no association between self-reported periodontal disease and CHD. The aim of this study was to investigate the association between oral health indicators and CHD deaths among Finnish adults aged 30–69 years during a 12-year follow-up.
The baseline data for this study are based on the Mini-Finland Health Survey, conducted during 1978–80. The study group in this comprehensive health survey was a two-stage cluster sample drawn from the population register and stratified to represent Finns aged 30 years and over. The first stage consisted of the selection of 40 representative areas. In the second stage, a systematic sample of inhabitants was drawn from each area. The sample size was 8000, of whom 7217 (90.2%) participated in the health examination. The subjects were first interviewed for socio-economic factors, health behavior, and other health determinants at home. The participants then underwent a health examination, including physical and dental examinations. This study group was comprised of all 3091 men and 3436 women aged 30–69 years with complete examination data. Participation in the study was strictly voluntary, and all invited persons were informed by an invitation letter and during the examination regarding the use of the data. As a rule, in the 1970s, no written informed consent was obtained in epidemiological studies. However, the current interpretation of authorities is that in these early studies voluntary participation can be equated with the present practice of informed consent.
Dental Data Periodontal status was registered according to the modified Periodontal Treatment Need System (PTNS) (Johansen et al., 1973). All four surfaces of the teeth were measured, except the distal surface of the last existing molar tooth. For periodontal recordings, the mouth was divided into four quadrants which were used as the basic units for registration. Gingival inflammation was registered according to the method of Löe and Silness (1963). Each jaw quadrant was examined and classified into one of four categories: (1) no inflammation, (2) gingival inflammation, (3) periodontal pocket(s) from 4 to 6 mm deep, and (4) pocket depth(s) exceeding 6 mm. For periodontal status, subjects were categorized according to the worst periodontal condition in the mouth. Dental plaque status was measured separately for each jaw quadrant. Total and partial edentulousness and the presence of various types of dentures were recorded separately for both jaws. The presence of fibroma and denture stomatitis was recorded. In the health interview, participants were asked about the frequency of dental attendance. They were defined as having a regular pattern of dental attendance if they claimed to have a dental check-up at least once a year; otherwise, dental attendance was defined as "irregular".
Other Baseline Data
Follow-up
Statistical Methods Relative mortality risks by specific oral health indicators adjusted for potential confounders, and the 95% confidence intervals were computed according to Cox’s proportional hazards model. Variable selection for the final models was based on both full models, including all studied variables, and careful analyses of possible collinearity and confounding. Multicollinearity did not disturb the final modeling. The SAS statistical package was used (SAS Institute Inc., Cary, NC, USA).
Persons dying of CHD were older and more often smoked, had hypertension, hypercholesterolemia, diabetes, and only a basic education compared with other persons (Table 1  ). Distribution of coronary risk factors by oral health indicators is depicted in Table 2. Prevalence of hypertension did not differ by oral health indicators in men, except that those with deep periodontal pockets had a higher prevalence of hypertension than did other men. Prevalence of hypertension in women was higher in those with fewer remaining teeth, in those who reported irregular dental attendance, and in edentulous women compared with other women. Prevalence of hypercholesterolemia did not appreciably vary by oral health indicators. Smoking and low education level were more prevalent in those belonging to almost any poor oral health indicator group than in those belonging to other groups. Prevalence of diabetes did not appreciably vary with oral health indicators in men, but did somewhat in women.
Edentulousness, wearing removable dentures, denture stomatitis, fibroma, and dental plaque were not associated with CHD mortality, and thus, these indicators were excluded from the final analyses. Analyses were restricted to those participants having at least one remaining tooth.
CHD death cases had fewer remaining teeth, an increased number of carious teeth, fewer filled teeth, and more retained roots than did non-cases. The differences were particularly evident in men. However, the number of subjects with periodontal pockets did not differ between cases and non-cases (Table 3
When only age was adjusted for, the number of remaining teeth was inversely associated with CHD mortality in men (Table 4 ). Numbers of carious teeth and retained roots were significantly positively associated with CHD mortality in men. No association was observed between the depth of periodontal pockets as an indicator of severity of periodontal diseases and CHD death in men. In women, no significant association was observed with any oral health indicator (Table 4). After adjustment for age and other oral health indicators, only the number of filled teeth remained significantly inversely associated with CHD death in men. Adjustment for education level and other CHD risk factors reduced the inverse association to statistical non-significance (Table 4).
Similar analyses were also performed with a shorter follow-up, excluding from end points all those who died after 5 years of follow-up. The associations remained essentially similar, as described above.
Oral health indicators were not independent risk factors for CHD deaths. In men, high numbers of carious teeth and retained roots, indicating poor oral health care, were associated with CHD risk when only age was adjusted for, but not after adjustment for confounding factors. The number of filled teeth was also inversely associated with CHD death in men, but after confounders were adjusted for, this association turned out to be statistically non-significant. No association was observed between periodontal pocket depths and CHD death, although a priori assumptions have been made (DeStefano et al., 1993; Beck et al., 1996, 1998; Beck and Offenbacher, 1998; Thoden van Velzen et al., 1984; Loesche, 1994) that deep periodontal pockets are especially viable portals of entry for bacteria into the systemic blood circulation. Our study was based on a large nation-wide representative sample of an adult population. The follow-up regarding deaths was complete. Information on general and oral health indicators was available only at the beginning of follow-up. Because the strength of some indicators may be diluted during a long follow-up, the analyses were also repeated including only deaths during the first 5 years; the results remained virtually unchanged. The presence of clinical oral infections can be reliably assessed, e.g., by the use of measures of caries status, retained roots, and markers of periodontal infections, as was done in the present study. In several previous studies (Mattila et al., 1989, 1993, 1995; Gillum, 1994; Grau et al., 1997), the Total Dental Index (TDI) has been used as a measure of oral infection load. However, the relative weights of selected oral health indicators in TDI are not based on empirically validated data. Missing teeth are also frequently considered as an indicator of oral infection (Paunio et al., 1993). Several reasons for missing teeth are not, however, dependent on oral infections. Thus, many indicators used in previous studies are not valid indicators of oral infection, and their use in prospective studies is questionable (Joshipura et al., 1998). Several confounders have associations with both oral health indicators and risk of CHD (Hujoel et al., 2000, 2001; Kolltveit and Eriksen, 2001). Some of these—particularly age, smoking, and low socio-economic status—were also observed to be confounders in our study. The observation that several of the oral indicators were associated with CHD risk when studied alone but not after adjustment for other risk factors indicates that the causal link with oral indicators is weak or non-existent, as also stated in other studies (Hujoel et al., 2000; Howell et al., 2001; Muller, 2001). The inverse association of filled teeth with CHD risk, observed also in one previous study (Ahlqwist et al., 1993), can be explained as an indicator of health consciousness of the individual, and thus, as a marker for generally decreased risk of CHD. Most of the previous studies suggesting a relationship between oral health status and CHD risk were comprised of only men. In our study, the association was studied in both women and men. The associations were apparent only in men, whereas even detailed oral health data failed to uncover an association with CHD risk in women. Despite the relatively small number of end-point cases among women, the discrepancy between the results by gender is another indication against a causal relationship. Generally poorer health-consciousness of men and more prevalent risk behavior may partly explain the gender differences. Since access to care and attitudes toward health care may influence the choice of oral health services, as well as the risk of CHD, one needs to be cautious of confounding by behaviorally related factors (Joshipura et al., 1998). Of the several previously studied oral health indicators, periodontal infections have been suggested to have a unique position as a risk factor for CHD (Thoden van Velzen et al., 1984; DeStefano et al., 1993; Loesche, 1994; Beck et al., 1996, 1998; Beck and Offenbacher, 1998). Periodontal infections, measured as depth of periodontal pockets, were very common in our study population, with only a minority of the population having healthy periodontal tissues. The use of mouth quadrant as the basic unit of registration for periodontal pockets may have caused overestimation of the severity of periodontitis. Because the majority of subjects manifested periodontal pockets, it was difficult to demonstrate gradual worsening of periodontal status and subsequent gradual increase in CHD risk. At the time of the baseline study, the PTNS method was commonly used for epidemiologic studies. However, current opinion is that use of attachment loss would be a better proxy for the bacterial and inflammatory burden over the years. The oral health status of the study population was determined by a thorough structured clinical examination by dentists. However, we had no opportunity to study radiological evidence of alveolar bone loss as a possible indicator of severe periodontal infections, because no orthopantomograms were available. Despite this limitation, accurate diagnoses of periodontal pockets were made. Given the associations of oral health indicators with CHD death in age-adjusted but not in multivariate analyses, we conclude that the associations can be explained by confounding factors, particularly those relating to health behavior. Although our results do not allow us to exclude the possibility of an etiologic role of oral infections in CHD, they suggest that the role of such infections must be rather limited.
This study has been supported by the Academy of Finland and by the Social Insurance Institution of Finland. Received for publication September 16, 2002. Revision received May 29, 2003. Accepted for publication June 10, 2003.
Journal of Dental Research, Vol. 82, No. 9,
713-718 (2003) This article has been cited by other articles:
|
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
|
|
 |
|
Sign In to gain access to subscriptions and/or personal tools.
TOP
ABSTRACT
INTRODUCTION
