A case-control study of topical and supplemental fluoride use and osteosarcoma risk.

BACKGROUND: A relationship between fluoride and osteosarcoma has been hypothesized but not validated. To the authors' knowledge, there are no published studies examining topical fluoride or dietary fluoride supplements and osteosarcoma risk. The purpose of this study was to examine the association between ever or never use of topical and dietary fluoride supplements and osteosarcoma. METHODS: The authors performed a secondary data analysis on data from 2 separate but linked studies. Patients for Phase 1 and Phase 2 were selected from US hospitals using a hospital-based matched case-control study design. Case patients were those who had received diagnoses of osteosarcoma, and control patients were those who had received diagnoses of other bone tumors or nonneoplastic conditions. In Phase 1, case patients (N = 209) and control patients (N = 440) were those seeking treatment at orthopedic departments from 1989 through 1993. In Phase 2, incident case patients (N = 108) and control patients (N = 296) were identified and treated by physicians from 1994 through 2000. This analysis included all patients who met eligibility criteria and on whom the authors had complete data on exposure, outcome, and covariates. The authors used conditional logistic regression to estimate odds ratios and 95% confidence intervals (CIs) for the association of topical fluoride use and supplemental fluoride use with osteosarcoma. RESULTS: The adjusted odds ratios were 0.94 (95% CI, 0.60 to 1.46) and 0.78 (95% CI, 0.46 to 1.33) for topical fluoride and supplemental fluoride, respectively. CONCLUSIONS: Neither topical nor dietary fluoride supplements are associated with an increased risk of developing osteosarcoma. PRACTICAL IMPLICATIONS: Supplemental and topical fluorides used in the dental office and in over-the-counter products are not related to an increased risk of developing osteosarcoma.

Comprehensive, School-Based Preventive Dentistry: Program Details and Students' Unmet Dental Needs.

BACKGROUND: In this paper, we describe the design, program details, and baseline demographics and oral health of participants in ForsythKids, a regional, comprehensive, school-based mobile caries prevention program. METHODS: We solicited all Massachusetts elementary schools with greater than 50% of students receiving free or reduced-price meals. Six schools initially elected to participate, ultimately followed by over 50 schools. Interventions were based on systematic reviews and randomized controlled caries prevention trials. Participating students received semiannual dental examinations, followed by comprehensive preventive care. Summary statistics regarding oral health indicators were derived from individual tooth- and surface-level data. RESULTS: Over a 6-year period, data were collected on 6927 children. The number of students per school ranged from 58 to 681. The overall participation rate was 15%, ranging from 10% to 29%. Overall, 57% of the children were younger than 8 years at baseline. Approximately, 54% of children experienced dental decay on any tooth at baseline; 32% had untreated decay on any tooth, 29% had untreated decay on primary teeth, and 10% untreated decay on permanent teeth. CONCLUSIONS: Untreated dental decay was double the national average, even in schools within several blocks of community dental clinics. These data demonstrate the need for caries prevention beyond the traditional dental practice.

Do genetic factors explain the association between poor oral health and cardiovascular disease? A prospective study among Swedish twins.

Epidemiologic studies suggest positive associations between poor oral health and cardiovascular disease. The authors undertook a prospective study among 15,273 Swedish twins (1963-2000) to examine whether genetic factors underlying the 2 diseases could explain previous associations. They estimated hazard ratios and 95% confidence intervals controlling for individual factors and stratifying on twin pairs to control for familial effects. Quantitative genetic analyses estimated genetic correlations between oral diseases and cardiovascular disease outcomes. Tooth loss (hazard ratio (HR) = 1.2, 95% confidence interval (CI): 1.1, 1.4) and periodontal disease (HR = 1.3, 95% CI: 1.0, 1.4) were associated with small excess risks of cardiovascular disease; periodontal disease was also associated with coronary heart disease (HR = 1.4, 95% CI: 1.1, 1.6). Adjustment for genetic factors in co-twin analyses did not appreciably change estimates. In contrast, tooth loss was more strongly associated with coronary heart disease in twin models (HR = 2.1, 95% CI: 1.2, 3.8) compared with adjusting for individual factors alone (HR = 1.3, 95% CI: 1.1, 1.4). There was evidence of shared genetic factors between cardiovascular disease and tooth loss (r(G) = 0.18) and periodontal disease (r(G) = 0.29). Oral disease was associated with excess cardiovascular disease risk, independent of genetic factors. There appear to be common pathogenetic mechanisms between poor oral health and cardiovascular disease.

The 3' UTR IGF2R-A2/B2 variant is associated with increased tumor growth and advanced stages in non-small cell lung cancer.

Normal function of insulin-like growth factor II receptor (IGF2R) gene has been associated with negative control of tumor growth in vivo and in vitro. Rare alleles at a 3' UTR short tandem repeat polymorphism of IGF2R are known to decrease transcript stability. One such allele (A2/B2) increases significantly the risk of oral squamous cell carcinoma and non-small cell lung carcinoma (NSCLC) in Caucasians. To determine potential association(s) between A2/B2 presence and development and/or progression of disease, we examined in 103 NSCLC patients, free of IGF2R allelic imbalance aberrations, the 3' UTR allelic status in relation to tumor kinetic parameters (proliferation index-PI and apoptotic index-AI) and clinicopathological data. PCR and automated sequence analyses were employed to genotype the IGF2R 3' UTR polymorphism. Given that, oncogenic mitogens, which escape degradation by IGF2R, can also activate p53 through a DNA damage response, the patterns between p53 status and IGF2R genetic constitution were also evaluated in relation to the above parameters. The A2/B2 variant was significantly more common (p=0.005, chi2-test) in lung cancer patients (25% vs 15%). Its presence was accompanied by high cellular proliferation (p=0.028, t-test) along with increased tumor cell growth (GI=PI/AI) (p=0.022, t-test) and it was significantly found in advanced stages. Also, patients carrying the A2/B2 in their genetic constitution that exhibit aberrant p53 expression have faster growing tumors and progress more rapidly to advanced stages. In conclusion, the IGF2R-A2/B2 variant probably provides a selective advantage for NSCLC progression through increased tumor growth.

Periodontitis and the risk for non-fatal stroke in Korean adults.

BACKGROUND: The association between periodontal inflammation and non-fatal stroke is still controversial and limited to evidence in Western countries. The aim of this study was to investigate whether periodontitis is independently associated with non-fatal stroke in Korean adults. METHODS: A case-control study was conducted on 265 non-fatal chronic stroke cases at the National Rehabilitation Center, Seoul, Korea, and 214 non-stroke population controls. Medical specialists diagnosed stroke by using brain imaging from magnetic resonance imaging and/or computerized tomography. A dentist recorded the clinical attachment level (CAL), the distance between the cemento-enamel junction and the probed base of the periodontal pocket, using a University of North Carolina-15 manual probe. An interview assessed 17 sociodemographic, behavioral, systemic/oral health-related possible confounders. Multiple logistic regression analysis was used to evaluate the association between periodontitis and stroke while controlling for age, gender, income, education, smoking, drinking, history of systemic disease, body mass index, familial cardiovascular risk factors, and oral health behaviors. Subgroup analyses were also performed. RESULTS: Stroke was strongly associated with periodontitis (presence of CAL > or =6 mm): the odds ratio was 4.0 (95% confidence interval: 2.3 to 7.0) after controlling for all possible confounders. The association with periodontitis (tertiary percentage of CAL > or =5 mm) had a dose-response effect. The association between periodontitis and stroke was higher among adults younger than age 60 (6.0 versus 2.6) and normotensives (4.8 versus 3.2). CONCLUSION: Our data suggested that periodontitis is independently associated with non-fatal stroke, and its impact seems to be greater among younger or normotensive Korean adults.

A reproducibility study for a fluoride assay in bone.

Osteosarcoma is a rare malignancy of largely unknown etiology. Although there is no consistent evidence for an association between fluoridation and cancer, some concerns remain about osteosarcoma. As part of the design of a collaborative study, bone samples were collected to allow for an evaluation of the association between osteosarcoma risk and individual fluoride exposure measured by levels of fluoride in bone. In this report, we provide the results of pilot experiments to consider issues that arose during the study design and to assess the reliability of the bone assays. Correlations of fluoride levels between normal bone near the affected area and iliac crest bone were strong and positive. The day-to-day laboratory analysis of fluoride in human and deer jaw bone yielded acceptable average coefficients of variation below 10% and an overall estimate of 5%. The intraclass correlation (ICC) is of particular importance to epidemiologists because it indicates the effect of measurement error on study results. Here, the estimated ICC is 0.86, and the estimated downward bias is only 14%. Hence, the ICC is strong enough so that the estimates of the relative risk will suffer little attenuation from lab measurements.

Periodontitis and plasma C-reactive protein during pregnancy.

BACKGROUND: Periodontitis has been associated with increased risk of adverse pregnancy outcomes and elevated C-reactive protein (CRP) concentrations in non-pregnant adults. We examined the relationship between periodontitis and CRP among women who provided dental radiographs and had blood collected during early pregnancy, excluding smokers and diabetic patients. METHODS: From Project Viva, an ongoing cohort study, we measured plasma CRP in 35 subjects with periodontitis (i.e., at least one site with > or =3 mm of alveolar bone loss) and a random sample of 66 periodontally healthy subjects matched on age and race/ethnicity. We performed linear regression analysis with log-transformed CRP levels as the outcome. RESULTS: The mean (+/- SE) CRP level was 65% higher (95% confidence interval: -2%, 180%; P = 0.06) in women with periodontitis (2.46 +/- 0.52 mg/l) than in controls (1.49 +/- 0.22 mg/l), adjusting for factors related to CRP levels, including age, race/ethnicity, pre-pregnancy body mass index, alcohol intake, education, income, and gestational age at blood collection. CONCLUSIONS: These findings suggest that periodontitis may increase CRP levels in pregnancy. CRP could potentially mediate the association of periodontitis with adverse pregnancy outcomes.

Pulpal inflammation and incidence of coronary heart disease.

Pulpal inflammation is primarily caused by coronal caries, and leads to root canal therapy (RCT). Chronic inflammation has been associated with various cardiovascular diseases. This study evaluates the association between pulpal inflammation (using RCT as a surrogate) and incident coronary heart disease (CHD). We report results among males from the Health Professionals Follow-Up Study (HPFS), excluding participants with prior cardiovascular disease or diabetes. We obtained RCT data from the HPFS cohort (n = 34,683). Compared to men without RCT, those with >/=1 RCT had a multivariate RR of 1.21 (95% CI 1.05-1.40) for CHD. The association was limited to dentists (RR = 1.38; 95% CI 1.14-1.67). There was no association among nondentists (RR = 1.03). Dental caries was not associated with CHD. The results suggest a possible modest association between pulpal inflammation and CHD.

Risk assessment and management of periodontal disease.

BACKGROUND: As our understanding of periodontal diseases has increased, it has become clear that certain risk factors are associated with the diseases' incidence, severity and progression. This article focuses on the role of risk assessment and disease management in improving patient outcomes, both in the general population and in specific population groups with an increased risk of developing periodontal disease or with associated comorbidities. TYPES OF STUDIES REVIEWED: The author reviewed literature related to the efficacy of risk assessment and periodontal disease management in improving clinical outcomes. In addition, he examined studies demonstrating a link between periodontal disease and specific patient populations and other comorbidities. CONCLUSIONS: Risk assessment can help predict a patient's risk of developing periodontal disease and improve clinical decision making. In turn, patient adherence to a self-care oral health regimen is a key component to successful periodontal disease management. CLINICAL IMPLICATIONS: The clinical practice of risk assessment may reduce the need for complex periodontal therapy, improve patient outcomes and ultimately reduce oral health care costs. Patients are encouraged to become actively involved in periodontal disease management by following a daily three-step regimen of brushing, flossing and rinsing with an antimicrobial mouthrinse.

Insulin-like growth factor II receptor gene-167 genotype increases the risk of oral squamous cell carcinoma in humans.

Our purpose was to evaluate inherited short tandem repeat polymorphisms of the insulin-like growth factor II receptor gene (IGF2R) in oral cancer risk. The 197 individuals that consented to participate in a hospital-based, case-control study were interviewed with a structured questionnaire and provided blood and saliva. DNA was extracted for genotyping using a PCR-based method. Odds ratios were calculated using multivariate logistic regression. Subjects carrying the heterozygous 167-bp IGF2R genotype had a 2.7-fold higher risk of oral cancer compared with subjects with other genotypes (odds ratio = 2.7, 95% confidence interval: 1.16-6.48), controlling for major confounders. Our results suggest that genetic variation of IGF2R may influence significantly the risk of oral cancer.

Risk assessment criteria applied to a screening exam: implications for improving the efficiency of a sealant program.

OBJECTIVES: The present study aimed to determine the ability of first-grade screening exam findings to predict carious lesions in permanent first molars by fourth grade. The ability to identify high-risk children would be useful in the rational allocation of limited public health resources, such as the application of sealants. METHODS: Screening exams were performed on 204 children in 1st grade and 4th grade. Analyses were conducted at both the child- and molar- levels. Cross-tabulations were used to identify the threshold dmfs + DMFS in first grade that had the highest sensitivity and highest negative predictive value for discriminating between children with and without permanent caries (1) in the entire population and (2) in subpopulations defined by race/ethnicity, gender, and language spoken at home. This threshold then was entered into logistic regression models. RESULTS: On the child and molar level, the study determined that dmfs + DMFS > 0 had the highest sensitivity (child: 69.4%, molar: 74.7%) and negative predictive value (child: 87.8%, molar: 94.4%). Using this test criterion, univariate logistic regression of the child-level data revealed an odds ratio of 2.72 (p = 0.012) for presence of permanent tooth caries experience. Controlling for gender, grade, race/ethnicity, and language spoken at home, the odds ratio increased slightly to 2.76 (p = 0.012). CONCLUSIONS: There is a relationship between carious lesion experience in 1st grade and carious lesion incidence in the first permanent molars by 4th grade. Using the dmfs + DMFS > 0 criterion may help public health providers determine which children should receive sealants when resources do not allow the delivery of sealants to all children.

Clinical prevention and population health: curriculum framework for health professions.

The Clinical Prevention and Population Health Curriculum Framework is the initial product of the Healthy People Curriculum Task Force convened by the Association of Teachers of Preventive Medicine and the Association of Academic Health Centers. The Task Force includes representatives of allopathic and osteopathic medicine, nursing and nurse practitioners, dentistry, pharmacy, and physician assistants. The Task Force aims to accomplish the Healthy People 2010 goal of increasing the prevention content of clinical health professional education. The Curriculum Framework provides a structure for organizing curriculum, monitoring curriculum, and communicating within and among professions. The Framework contains four components: evidence base for practice, clinical preventive services-health promotion, health systems and health policy, and community aspects of practice. The full Framework includes 19 domains. The title "Clinical Prevention and Population Health" has been carefully chosen to include both individual- and population-oriented prevention efforts. It is recommended that all participating clinical health professions use this title when referring to this area of curriculum. The Task Force recommends that each profession systematically determine whether appropriate items in the Curriculum Framework are included in its standardized examinations for licensure and certification and for program accreditation.

Oral hygiene practices and periodontitis in health care professionals.

BACKGROUND: One in 3 adults in the United States has periodontitis, yet there are few widely applied methods for its prevention. Few studies have assessed oral hygiene practices related to periodontitis. METHODS: There were 533 participants in this study (14.8% male dentists; 13.7% male non-dentist health care professionals; and 71.5% female nurses), all of whom are enrolled in ongoing cohort studies in the United States which have inquired about their oral hygiene practices and whether they have professionally diagnosed periodontitis. We studied the relationship between oral hygiene practices and periodontitis, controlling for potential confounders. RESULTS: The participants were 69 years old on average, 84.2% white, 12.4% current smokers, and 8.3% diabetic. Seventy percent of the dentists and nurses brushed at least twice a day compared to 56% of the other health professionals. Two-thirds of the nurses, 56.3% dentists, and 36.4% other health professionals flossed at least once daily. Persons brushing twice daily were as likely to have periodontitis as those brushing once or less daily (odds ratio [OR] = 1.16; 95% confidence interval [CI]: 0.64 to 2.10); persons flossing less than once a day were as likely to have periodontitis as those who flossed daily (OR = 1.16, 95% CI: 0.63 to 2.13) after controlling for profession, age, gender, smoking, diabetes, coronary heart disease, history of periodontal surgery, and number of teeth present. CONCLUSIONS: People who reported better oral hygiene practices did not report less periodontitis, after controlling for potential confounders. Oral hygiene practices were not associated with periodontitis in this population.

Can nonstandardized bitewing radiographs be used to assess the presence of alveolar bone loss in epidemiologic studies?

OBJECTIVE: To compare periodontitis-associated alveolar bone loss assessment by standardized and nonstandardized radiographs in clinical and epidemiologic studies. METHODS: Participants included 37 patients aged 21-66 years with prior nonstandardized bitewing radiographs scheduled to receive bitewing radiographs as part of their next routine dental care visit. Standardized bitewing radiographs were taken with a Rinn film holder to position the film in the mouth and align the X-rays so that they were at 90 degrees to the film. Before taking the radiograph the bite was registered in centric relation using a polyether impression material. One registered dental hygienist took and processed all the standardized radiographs. One dentist read all radiographs using a viewing box, magnifying lens, and periodontal probe with William's markings. Radiographic bone loss was measured to the closest millimeter at mesial and distal sites of the posterior teeth excluding third molars. The examining dentist was blinded to the participant's name, age, gender, or if the radiograph was standardized or nonstandardized. RESULTS: Mean bone loss (+/- SD) was similar in the standardized and nonstandardized groups (1.60 +/- 0.72 mm versus 1.64 +/- 0.85 mm), and the correlation was high (r = 0.95). Periodontitis was defined as present if the participant had at least one site with 3, 4 and 5 mm bone loss. The Kappa statistics for concordance using these three cutoffs were good and ranged from 0.60 to 0.65. The sensitivity ranged from 72.7 to 80.8% and specificity from 88.5 to 90.9%. CONCLUSIONS: Periodontitis assessed as mean alveolar bone loss or the prevalence of disease based on alveolar bone loss can be accurately and reliably evaluated from nonstandardized radiographs.

Associations between occupational health behaviors and occupational dental erosion.

OBJECTIVES: The aim of this study was to evaluate the associations between occupational health behaviors and occupational dental erosion. METHODS: Using data for 943 workers among 34 factories, selected by three-stage stratified cluster sampling from 888 factories using acids, two sets of modified case-control studies were performed. The cases were 242 workers with any dental erosion (G1-5) and 78 with severe dental erosion (G3-5); the controls were 701 workers with no erosion (GO) and 864 workers with no or mild erosion grades, GO-2, respectively. The main explanatory variables were behaviors such as wearing a respiratory mask and gargling at work. The results were adjusted for employment, age, sex, knowledge, and opinion about occupational health, attrition, and abrasion. Bivariate and multivariate logistic regression analyses were conducted. RESULTS: The odds of overall occupational dental erosion (G1-5) was 0.63 (95% CI = 0.42, 0.94) for respiratory mask wearers compared to nonwearers; the odds of severe occupational dental erosion (G3-5) was not significantly less in respiratory mask wearers (OR = 0.94; 95% CI = 0.53, 1.67). Gargling did not show a significant association with occupational dental erosion in this study. CONCLUSIONS: Among occupational health behaviors, wearing personal protective respiratory masks in work was significantly associated with less overall occupational dental erosion.

Validation of self-reported periodontal measures among health professionals.

OBJECTIVE: The objective of this study was to determine the validity of self-reported periodontal measures among nondentist health professionals. Valid self-reported measures could provide a time- and cost-efficient alternative for large epidemiologic studies. METHODS: A subsample of 212 male nondentists sampled on the basis of their reported periodontal severity from the Health Professional Follow-up Study (HPFS) provided dental radiographs and completed questionnaires assessing self-reported oral health. Alveolar bone loss was evaluated from the radiographs at 32 posterior sites and used as the standard measure of cumulative periodontal disease. RESULTS: The self-reported ordinal periodontal measure had a linear relationship with mean radiographic bone loss (r = .61). The positive and negative predictive values of the dichotomized self-reported periodontal measures were 83 percent and 69 percent. Self-reported history of periodontal surgery was also a good surrogate for bone loss (predictive value positive 78 percent and negative 71 percent). CONCLUSIONS: Self-reports can provide discrimination and ranking information of cumulative periodontal disease among health professionals and can be used to provide valid results in etiologic studies in health professionals' populations.

Validation of self-reported oral health measures.

OBJECTIVES: To assess the validity of self-reported oral disease and health care measures in two populations. METHODS: Telephone interviews were conducted among a subsample of participants in the VA Dental Longitudinal Study (VADLS) asking them about periodontal disease status and treatment. Radiographic alveolar bone loss evaluated at all the interproximal sites was used as the standard. A separate study was carried out among first-time patients at the Harvard School of Dental Medicine (HSDM) student clinic. Self-reported measures were obtained by a self-administered questionnaire and compared with clinical and radiograph examinations. The measures used were based on published work that demonstrated good validity of self-reported periodontal measures among health professionals. RESULTS: Among 145 VADLS participants, self-reports of periodontal disease showed a good specificity (59.8%-90.7%), but low sensitivity (17.7%-64.7%). Among 58 HSDM patients, the self-reported numbers of remaining teeth, fillings, root canal therapy, and prosthesis were strongly correlated with clinical records (r = 0.74-1.0); self-report was less accurate for measures of periodontal disease (r = 0.56) and decayed teeth (r = 0.47). CONCLUSIONS: Self-reports provide reasonably valid estimates for numbers of remaining teeth, fillings, root canal therapy, and fixed and removable prostheses. However, they appear to be less useful for the assessment of dental caries and periodontal disease in the two populations we have studied. There remains a need and potential to further develop self-report oral health measures that are valid for use in large population studies. Such self-report measures would yield great cost and time savings.

Problems in exposure assessment of fluoride in drinking water.

OBJECTIVES: We developed a source documentation approach that identified fluoride content of drinking water at the state or local level to estimate fluoride concentrations from public water systems. We then compared estimates from this approach with estimates obtained from a single source, the 1992 Centers for Disease Control and Prevention (CDC) Fluoridation Census. METHODS: We used residential histories from a case-control study. For each residence we attempted to determine fluoride concentrations using the 1992 CDC Fluoridation Census. For the source documentation method we utilized multiple sources from state and local contacts to verify and collect additional data. We compared the fluoride estimates obtained by the two methods. RESULTS: When fluoride values were found using both methods, there was good correlation (Kendall's tau = 0.85; 95% confidence interval = 0.79, 0.90) and concordance was 96 percent. We obtained over 99 percent of the fluoride values needed using source documentation as compared to 49 percent of the values needed when we used a single publication. When fluoride values were missing using the 1992 CDC Fluoridation Census, 21 percent had source documentation estimates of at least 0.7 ppm. CONCLUSIONS: Researchers need to consider limitations of using a secondary data source to estimate fluoride in drinking water, particularly in studies where exposure to fluoride is the primary exposure of interest.

Use of preexisting radiographs for assessing periodontal disease in epidemiologic studies.

OBJECTIVES: This study evaluates the feasibility of obtaining preexisting dental radiographs by mail, the validity of assessing alveolar bone loss from posterior radiographs compared to full mouth, and the validity of alveolar bone loss assessed from radiographs taken at different times. METHODS: This investigation uses data obtained for a study evaluating associations between oral conditions, blood biomarkers, and coronary heart disease within two large cohorts: the Nurses' Health Study and the Health Professionals Follow-up Study. If consenting participants had dental radiographs, we requested these radiographs from their dentists. Some dentists returned multiple sets of radiographs, which enabled us to conduct this study. A calibrated periodontist read all radiographs with good intraexaminer reliability (r=0.91). We compared posterior radiographs to full mouth (n=121 sets), as well as radiographs taken at different times (mean difference of 5 years) (n=102 pairs). RESULTS: Of the 812 participants, 81 percent consented and 66 percent provided radiographs. Posterior radiographs underestimated periodontitis (> or =1 site with > or =5 mm alveolar bone loss) prevalence by 6 percent (53.7% vs 57.0%) compared to full mouth, with sensitivity of 0.94 (95% confidence interval [CI]=0.86, 0.98) and specificity of 1. Spearman correlation coefficients comparing mean alveolar bone loss were 0.70 for anterior versus posterior teeth (mean difference=0.48), 0.92 for posterior teeth versus full mouth (mean difference=0.25), and 0.78 for pairs of radiographs taken at different times (mean difference=0.01). The kappa statistic was 0.70 comparing radiographs taken within 5 years and 0.29 when the period extended beyond 5 years. CONCLUSION: Preexisting radiographs are feasible for use in epidemiologic studies and provide valid assessments of periodontal disease.

The relationship between amalgam restorations and mercury levels in male dentists and nondental health professionals.

OBJECTIVES: The objectives of this study were: (1) to compare the mercury levels in general dentists with the mercury levels in other health professionals using toenail clippings as a biomarker, (2) to identify risk factors associated with high mercury levels, and (3) to compare practice characteristics of dentists with high and low mercury levels. METHODS: A sample of 579 men was randomly selected from the 33,737 men participating in the Health Professionals Follow-up Study who had provided toenail samples in 1987. A questionnaire was sent to these male subjects in 1991 to obtain information on fish consumption, toothbrushing frequency, number of teeth, number of amalgam restorations, general practice or specialty status, number of amalgam restorations placed and removed per week, mercury storage and handling procedures, and mercury spillage incidents. A measure of long-term mercury exposure was obtained from toenail samples using neutron activation analysis for the 410 respondents (71% response rate). The 90th percentile mercury level in toenails (0.88 ppm) was selected as the threshold for elevated toenail mercury level. RESULTS: No relationship was found between the number of dental amalgams and toenail mercury levels among general dentists, dental specialists, and nondental health professionals. General dentists were found to have more than twice the level of mercury in toenails than nondental health professionals (mean level = 0.94 vs 0.45) and 60 percent higher than dental specialists (mean = 0.59). The combined use of disposable capsules and water storage of scrap amalgam appeared to reduce the risk of elevated mercury levels. Regardless of professional status, consumption of tuna and saltwater fish were the primary exposure factors that were positively associated with toenail mercury levels. CONCLUSIONS: As shown by the associations with dental profession and fish consumption, the mercury content of toenails is a stable biomarker of cumulative long-term mercury exposure. The lack of association between nail mercury levels and number of amalgam restorations suggests that avoidance of mercury amalgam restorative materials cannot be justified by the presence of mercury released from dental amalgams.