| Sign In to gain access to subscriptions and/or personal tools. |
Periodontal Disease and Biomarkers Related to Cardiovascular Disease
1 Department of Oral Health Policy and Epidemiology, Harvard School of Dental Medicine; Correspondence: * corresponding author, kjoshipura{at}hsdm.harvard.edu
Periodontal disease is a chronic infection of the gums characterized by a loss of attachment between the tooth and bone, and by bone loss. We evaluated cross-sectionally the association between periodontal disease and C-reactive protein (CRP), fibrinogen, factor VII, tissue plasminogen activator (t-PA), LDL-C, von Willebrand factor, and soluble tumor necrosis factor receptors 1 and 2. The final sample consisted of 468 men (ages 47–80 yrs), participating in the Health Professional Follow-up Study, who provided blood and were free of CVD, diabetes, and cancer. In multivariate regression models controlling for age, cigarette smoking, alcohol intake, physical activity, and aspirin intake, self-reported periodontal disease was associated with significantly higher levels of CRP (30% higher among periodontal cases compared with non-cases), t-PA (11% higher), and LDL-C (11% higher). Based on our data, periodontal disease showed significant associations with biomarkers of endothelial dysfunction and dyslipidemia, which may potentially mediate the association between periodontal and cardiovascular disease.
Key Words: periodontal disease • tooth loss • cardiovascular disease • biomarkers • inflammation
The association between periodontal disease and cardiovascular disease (CVD) has received considerable attention, although it is not clear whether there is a causal component (Joshipura et al., 2000). It is also unclear whether the association between tooth loss and CVD is caused by antecedent periodontal disease, decay, the process of tooth removal, dietary changes following tooth loss, or other causes (Joshipura et al., 1998). There is growing evidence that periodontal disease may be associated with biomarkers of inflammation, hemostasis, and dyslipidemia (Kweider et al., 1993; Slade et al., 2000; Wu et al., 2000). Moderate elevation of serum C-reactive protein (CRP) is associated with increased cardiovascular disease risk among otherwise healthy individuals (Ridker et al., 1998, 2000). Elevated plasma fibrinogen levels are also known to be independently associated with CVD (Thompson et al., 1995). It is known that tissue plasminogen activator (t-PA) plays a key role in the lysis of blood clots; increased plasma t-PA, possibly a marker of endothelial dysfunction, is associated with increased risk of CHD (Thompson et al., 1995). The association between periodontal disease and tissue plasminogen activator (t-PA) has not been evaluated. von Willebrand factor (vWF) is released by platelets and mediates early hemostasis (Mattila et al., 1989). Tumor necrosis factor (TNF) plays a key role in the initiation of the inflammatory response (Bazzoni and Beutler, 1996). TNF receptors (sTNF-R1 and sTNF-R2) are markers of TNF activity. TNF-  has been associated with CVD risk factors, and with carotid intima-media thickness (Skoog et al., 2002). High-serum/low-density lipoprotein cholesterol (LDL-C) is an established CVD risk factor (Wilson et al., 1998; Ridker et al., 2000). Hence, if periodontal disease is found to be associated with these biomarkers, they may be potential mediators for the association between periodontal disease and CVD. Our study aims to evaluate associations among periodontal disease, tooth loss, and specific biomarkers in blood within the Health Professional Follow-up Study (HPFS).
The HPFS is a prospective investigation of 51,529 male health professionals who were 40–75 years of age when the study began in 1986. The cohort includes dentists, veterinarians, pharmacists, optometrists, osteopaths, and podiatrists. Participants completed detailed dietary and medical history questionnaires, and 18,225 of them subsequently provided a blood sample. The study was approved by the Institutional Review Board, and the return of the questionnaire(s) constituted informed consent. This is a secondary analysis of data that were collected primarily for the evaluation of the association between alcohol and biomarkers. Since the sub-study involving the biomarker analyses was originally initiated to evaluate the association between alcohol and biomarkers, we excluded 8922 men who had incomplete information on diet, cigarette smoking, alcohol consumption, and physical activity from 1986 to 1994. We also excluded 208 men with cardiovascular disease, diabetes, gastric or duodenal ulcer, liver disease, and cancer (except non-melanoma skin cancer). Of the remaining men, 468 were then randomly sampled within 7 clusters defined by their self-reported alcohol consumption pattern (abstain, light, moderate, etc.). The distributions of dietary and lifestyle characteristics were similar between those of the population who did and did not return a blood sample, and between the 468 men in this sample and the HPFS population (Fung et al., 2000). At baseline (1986), participants were asked, "Have you had periodontal disease with bone loss?" Participants were sent follow-up questionnaires to ascertain professionally diagnosed periodontal disease during each two-year follow-up period. Self-reported periodontal disease measures were found to be valid against radiographs in this cohort (Joshipura et al., 1996, 2002). We classified a person as having periodontal disease if he reported being professionally diagnosed with periodontal disease through 1994 (since biomarkers were assessed from blood collected between 1993 and 1995); otherwise he was considered negative for periodontal disease.
We evaluated self-reported numbers of natural teeth at baseline, dichotomized into 0–25 vs. Participants were sent a kit with instructions and supplies to have blood drawn. Blood samples were returned to our laboratory via overnight courier; over 95% arrived within 24 hrs of blood being drawn. The blood samples were centrifuged, aliquoted, and stored in liquid nitrogen (–150°C). Fewer than 15% of the samples were slightly hemolyzed, and very few were moderately hemolyzed (< 3%), lipemic (< 1%), or not cooled upon arrival (< 0.5%). CRP was measured by an ultra-sensitive immunotechnique on the Behring BNII analyzer. LDL-C was precipitated by the addition of phosphotungstic acid and magnesium ions. Fibrinogen was measured by the Clauss method. Factor VII, t-PA, sTNF-R, and vWF antigen concentrations were determined by ELISA. The intra-assay CV was 2.6% for fibrinogen, 5.5% for t-PA antigen, 3.0% for factor VII, 8.8% for vWF, 5.6% for CRP, 3.1% for LDL-C, 8.7% for TNFR-1, and 9.3% for TNFR-2.
Statistical Analysis We controlled for factors associated with the biomarkers that are likely to confound the relationship between periodontal disease and/or tooth loss and the biomarkers. All models controlled for age, alcohol, cigarette smoking, BMI, physical activity, and regular aspirin intake, with the 1994 questionnaire as the index year. Each participant was asked to report his height and weight. Alcohol intake was calculated by summing the frequencies and amounts of beer, red wine, white wine, and spirits as reported by the participants on the food-frequency questionnaire (SFFQ). Self-reported anthropometric measures and alcohol consumption have been shown to be valid in this cohort (Giovannucci et al., 1991; Rimm et al., 1991). Metabolic equivalents for the task (METs) are defined for each type of physical activity as a multiple of the energy requirement of sitting quietly for 1 hr and cumulated over various activities. We also included aspirin intake, since aspirin may lower the levels of biomarkers such as CRP (Ikonomidis et al., 1999).
From the 468 men who provided blood samples, 91 (19%) reported a positive history of periodontal disease up to 1994. As shown in Table 1  , men with periodontal disease were slightly older and had higher plasma levels of CRP, LDL-C, vWF antigen, and t-PA antigen.
Based on the AHA guidelines (Pearson et al., 2003), a cut-off of < 0.1 mg/dL for CRP is considered low-risk; 50% of our population falls into this category; 15% are in the high-risk category, defined as > 0.3 mg/dL, and the remaining 35% are in the average risk category. Based on the third report of the National Cholesterol Education Program (NCEP, 2001), 8.8% of our population has optimal LDL-C levels (< 100 mg/dL), 40% have high (  160 mg/dL), and 13.5% have very high LDL-C levels ( 190 mg/dL).
In multivariate models (Table 2
In separate models, we evaluated the association of tooth loss with the biomarkers (Table 3 ). Plasma t-PA was 7% higher (p = 0.07) and LDL-C was 9% higher (p = 0.01) among men with 24 or fewer natural teeth. CRP and other biomarkers did not show any association with tooth loss.
The mean plasma levels in the blood cohort were compared by their a.m./p.m. blood drawn status. Men whose blood was drawn at "a.m." compared with "p.m." had significantly higher levels of t-PA antigen (mg/dL), sTNF-R1, and sTNF-R2. Most of the other plasma levels taken at "a.m." were slightly higher compared with "p.m." but not significant. Controlling for time of blood drawn did not change the associations in the multivariate analysis. Because fasting status before the blood draw may affect the biomarker levels, we conducted a sub-analysis by excluding the men who provided a non-fasting blood sample (blood drawn < 6 hrs since last meal, n = 269). Limiting to the sub-group with the fasting sample increased the association between periodontal disease and CRP to 45%, and decreased the association with t-PA to 6% and LDL-C to 9%. The associations remained significant for CRP and LDL-C, and showed borderline significance for t-PA (p = 0.11).
In this cross-sectional study of US health professional men, we found that periodontal disease was associated with elevated serum levels of CRP, t-PA, and LDL-C. The analysis was independent of age, cigarette smoking, alcohol, BMI, regular aspirin use, and physical activity. The associations were similar among men who provided a fasting blood sample. Our results for CRP are consistent with the existing literature. A study (Wu et al., 2000) based on data from the Third National Health and Nutrition Examination Survey (1988–1994) found significant associations between periodontal status and increased CRP (Slade et al., 2000; Wu et al., 2000). The prevalence of elevated CRP was doubled among the individuals with periodontal disease compared with healthy individuals. A recent study found that elevated CRP values in periodontitis patients were significantly higher than in periodontally healthy people (38% vs. 17% had elevated CRP) (Noack et al., 2001). Our smaller associations of 30% difference in CRP levels between people with and those without periodontal disease may be explained by the use of self-reported periodontal measures in our study, or more rigorous control of confounding due to our homogenous population of health professionals. We are not aware of any study that directly evaluates the association between periodontal disease and t-PA. The association between periodontal disease and LDL-C levels has been seen in a few studies (Pohl et al., 1995; Netea et al., 1997; Cutler et al., 1999a,b; Losche et al., 2000; Katz et al., 2002); however, none of these studies controls for dietary factors. In a study of 10,590 Israeli military men and women, the association persisted after investigators controlled for BMI, age, diastolic blood pressure, and smoking (Katz et al., 2002). Wu et al.(2000) found that periodontal disease was related to a significantly higher level of total cholesterol, but that there was no association with HDL cholesterol, also implying that periodontal disease was related to an increase in LDL-C. Hence, our results on LDL-C are consistent with reports in the literature and persisted after we adjusted for dietary factors. Many studies showed associations between periodontal disease and cardiovascular diseases. Increased levels of serum LDL-C, CRP, and t-PA have been related to increased cardiovascular risk (Thompson et al., 1995; Ridker et al., 1998, 2000; Wilson et al., 1998). Our results are of clinical significance, since the American Heart Association now advocates the use of CRP in addition to LDL-C as a predictor of heart disease risk (Pearson et al., 2003).
Our study did not show any association between periodontal disease and fibrinogen, factor VII, or TNF receptors. We have over 80% power to detect 10% or greater differences for all biomarkers except CRP and vWF, where we have power to detect 20% and 15% differences, respectively. Other than two reports (Kweider et al., 1993; Wu et al., 2000), which showed an association with fibrinogen, there is not much literature directly relating periodontal disease with these biomarkers in blood. Hence, our results suggest that biomarkers of inflammation, dyslipidemia, and endothelial dysfunction, rather than hemostatic factors, may be potential mediators of the periodontal cardiovascular disease relationship. Further studies are needed to confirm the associations and to ascertain whether serum TNF The HPFS cohort does not represent a random sample of US men. Therefore, the biomarker levels and other lifestyle characteristics may not reflect those of the general US male population. However, there is no reason to believe that the effect of periodontal disease on biomarkers is different in our population. The homogeneity of the cohort with respect to education and socio-economic status reduces the likelihood that unaccounted factors related to SES have biased our results. Because this study was a secondary analysis of data collected for a study relating alcohol to biomarkers, the population did not include CVD cases. This exclusion may attenuate the results, since people with high levels of CRP may be more likely to have both periodontal disease and cardiovascular disease. The fact that this sample was selected on the basis of alcohol consumption is unlikely to affect the association, since we stratified and controlled for alcohol consumption in the analyses. Our study is cross-sectional; hence we cannot establish whether periodontal disease resulted in the change in biomarkers or vice versa. However, our periodontal measure reflects cumulative periodontal disease; hence it seems likely that periodontal disease may have led to a change in biomarkers. In conclusion, our results suggest that periodontal disease is related to biomarkers of endothelial dysfunction and dyslipidemia such as CRP, t-PA, and LDL-C, which are known risk factors for cardiovascular disease.
This research was supported by Grants HL35464, CA 55075, and DE12102 from the National Institutes of Health, Bethesda, MD. We are indebted to the participants of the Health Professionals’ Follow-up Study for their continued cooperation and participation; to Al Wing and Mira Kaufman for computer assistance; to Jill Arnold, Betsy Frost-Hawes, Kerry Demers, and Mitzi Wolff for their assistance in the compilation of data; and to Laura Sampson, RD, for maintaining our food composition tables. Received for publication February 10, 2003. Revision received August 1, 2003. Accepted for publication November 14, 2003.
Journal of Dental Research, Vol. 83, No. 2,
151-155 (2004) This article has been cited by other articles:
|
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
|
|
 |
|
Sign In to gain access to subscriptions and/or personal tools.
TOP
ABSTRACT
INTRODUCTION
