Abstract
Objectives
While risk factors for tooth loss in adults have been identified, limited studies describing factors associated with incident tooth loss in postmenopausal women exist. This study assessed both clinical and non-clinical risk factors for incident tooth loss.
Methods
Postmenopausal women (N= 1,341) were recruited between 1997–2000 from 1847 eligible Observational Study participants of the Buffalo, NY center of the Women’s Health Initiative who had complete dental examinations to assess alveolar bone height, soft tissue attachment and general oral health, and completed questionnaires concerning demographics, general health, lifestyle and oral health (72.6% participation rate). Five years later (2002–2005), 1021 women (76.1%) repeated these examinations and questionnaires. Incident tooth loss was determined by oral examination
Results
After an average 5.1 years of follow-up (SD, 0.38), a total of 323 teeth were lost in 293 women, resulting in 28.7% of women with incident loss of at least one tooth. In multivariable models, diabetes history, gum disease history, smoking, previous tooth loss, BMI and plaque index, baseline clinical measures including alveolar crestal height (ACH) (OR=1.22 per mm loss, 95% CI 1.11, 1.35), clinical attachment loss (CAL) (OR=1.13 per mm loss, 95% CI 1.05, 1.23) and pocket depth (PD) (OR=1.26 per mm loss, 95% CI 1.13, 1.41) were significant risk factors of incident tooth loss. In a community model that included no clinical measures, diabetes history (OR=2.45, 95% CI 1.26, 4.77), prior gum disease (OR=1.97, 95% CI 1.43, 2.70), ever smoking (OR=1.42, 95% CI 1,06, 1.89), number of teeth lost at baseline (OR=1.05 per tooth, 95% CI 1.02, 1.08) and BMI (OR=1.15 per 5 km/m2 increase, 95% CI 1.01, 1.33) were associated with an increased risk of incident tooth loss.
Conclusions
Clinical and questionnaire based models were found to provide similar risk estimates for incident tooth loss in postmenopausal women. These models identified high risk postmenopausal women where preventive strategies may be targeted.
Introduction
Tooth loss is associated with problems for people at any age, including significant morbidity among older adults. Tooth loss affects quality of life and chronic conditions such as heart disease, diabetes and even mortality (1–5). As a result, partial and complete edentulism in older men and women may pose a significant public health problem in the United States as the population ages. Greater emphasis on improving oral health has been suggested (6–8). Understanding factors that are associated with incident tooth loss will aid in developing these prevention strategies. To this end, associations of numerous exposures and tooth loss have been explored and include various demographic, lifestyle and clinical characteristics (9–13). Some studies have explored incident tooth loss among adults (14–17), where risk increases with age and postmenopausal status, the latter potentially due to factors such as bone loss and estrogen deficiency (18,19). Despite having better oral health than men, women more often experience tooth loss (20–22). Longitudinal studies of incident tooth loss among postmenopausal women are limited (23,24). The purpose of this study was to evaluate a series of clinical and non-clinical models of risk factors of incident tooth loss in postmenopausal women, and identify those which could be informative to researchers studying this population or to clinicians managing their patients. We hypothesized that certain clinical and personal characteristics could characterize those women at greater risk of tooth loss.
Methods
Sample
This study was conducted in participants of the Women’s Health Initiative Observational Study (WHIOS) enrolled at the University at Buffalo clinical center (N=2249) (Figure 1). These postmenopausal women were contacted between 1997–2000 and invited to participate in an ancillary study of osteoporosis and oral bone loss (Osteo Perio Ancillary Study) (25). Those interested and eligible completed a series of questionnaires, had an oral examination and underwent systemic bone density screening. A total of 1362 women participated in this ancillary study. Of those, 1341 who had complete baseline questionnaires (5 missing questionnaires) and oral radiographs (16 incomplete radiographs) were invited back approximately 5 years later for a follow-up study where questionnaire information and standardized examinations were repeated. Among the 1341 participants from the baseline study invited to the follow-up examination, 101 were ineligible, 51 were determined to be deceased, 151 were not interested in participating, 3 withdrew from the original WHI observational study, and 9 were unable to be contacted. An additional 5 consented but had an incomplete dental exam at the follow-up visit. These analyses are based on the remaining 1,021 individuals who participated in the follow-up study. Participants were ineligible for the baseline study if they had less than 6 teeth remaining, history of bone disease, bilateral hip replacement, cancer diagnosis in the 10 years prior to interview, and certain other serious illnesses (25). For the follow-up study, women were ineligible if during the follow-up they developed cancer or an immunosuppressive disorder (eg. transplant recipient), were on long term antibiotic treatment or received recent dental x-rays within the prior year. All participants provided signed informed consent for the WHI-OS, as well as the baseline and follow-up ancillary studies. All studies were approved by the University at Buffalo’s Health Sciences Institutional Review Board.
Figure 1.
Recruitment flowchart for prospective study of postmenopausal women and incident tooth loss
Interview
At both the baseline and follow-up visits, participants provided self-reported responses to lifestyle and demographic questionnaires, had physical measures taken, completed a bone density assessment and had an oral examination (25,26). All questionnaires and examinations were conducted using standardized protocols. Demographic and other variables to be examined included age, self described race/ethnicity (American Indian/Alaskan Native, Asian/Pacific Islander, Black or African American, Hispanic/Latino, White), education level (high school or less, college, graduate school), marital status (never married, divorced/separated, widowed, currently married, marriage-like relation), income (less than $10,000, $10–19,999, $20–34,999, $35–49,999, $50–74,999, $75–99,999, $100,000 or more), cigarette smoking (ever/never), frequency of brushing (≤1 times per day, ≥2 times per day), flossing (daily, >1 time per week, 1 time per week., not every week) and dental visits (only with a problem or never, ≥1 time per year), body mass index (BMI in kilograms/meter2), type 2 diabetes history (yes or no), years since menopause without use of hormone therapy (HT), current HT use (yes or no), current HT and/or bone drug use, worst T-score category from DXA of all sites measured (normal, T=>−1, osteopenia, −2.5<=T<−1, osteoporosis, T=<−2.5), tooth loss at baseline (any, none), number of teeth remaining at baseline visit, history of surgery for gum disease (ever, never) and history of tooth extraction at baseline due to gum disease (ever, never).
Clinical Exam
Tooth presence was recorded as part of oral examination by a dental examiner both at baseline and follow-up. For the purpose of these analyses, tooth loss was defined as determination of no tooth present at follow-up that was present at baseline examination. The primary outcome for these analyses was the dichotomous variable indicating any tooth loss (Yes/No) during the study period. The oral examination included oral radiographs for determination of alveolar crestal height (ACH). ACH was measured in all teeth present in up to 48 sites from 24 teeth from 11 oral radiographs using a method developed by Hausmann et al (27–29). Radiographs were digitized and ACH was measured as the distance in millimeters from the cemento-enamel junction to the most coronal part of the alveolar crest in a plane parallel to the long axis of the tooth for 2 sites per tooth. Larger ACH values represent worse bone loss. Three ACH measures were created from these data: whole mouth mean (mm) of all sites measured, worst site measured (mm) and number of sites measured with loss of 4mm or greater. Probing pocket depth (PPD) was measured in millimeters using a constant force probe (The Florida Probe System®, Gainsville, FL). Inter- and intra-examiner reliability of the PPD measurements have been previously reported to be within 0.5mm (30) and 0.57 mm intra-examiner reliability of clinical attachment loss (CAL) (31). Internal quality control of ACH measures revealed a coefficient of variation of 5%. Clinical attachment loss (CAL) was measured by subtracting the distance between the cemento-enamel junction and the gingival margin from PPD in millimeters (26). For both PPD and CAL six sites per tooth were measured. Whole mouth (mean) and worst site measured were created using all sites for CAL and PPD. Calculus presence per tooth was assessed on examination, recorded as absent, supragingival only or supragingival and/or subgingival and used here as whole mouth mean in all teeth assessed. Plaque and gingival bleeding were recorded as absent or present on 3 sites per tooth, expressed as mean (%) of the whole mouth. Presence of gingival bleeding was assessed using a Michigan O periodontal probe inserted at each site and recorded as present or absent. Reasons for tooth loss were self-report during the oral examination on query of the dental examiner for each tooth noted missing on examination. Recorded reasons included accident, caries, congenitally missing, ortho/malposed, periodontal disease, unerupted, root canal and unable to determine.
Statistical Analyses
The analytic strategy included a review of the published literature and identification of factors that were known or suspected as risk factors for tooth loss. Those factors that were available in our dataset were considered in the analysis. Following this selection, we determined univariate statistical associations of these study variables for incident tooth loss (Yes/No) using Student’s t-test or Pearson’s Chi square. Risk factors found to be associated with tooth loss in univariate analyses (p≤0.20) were attempted in a series of multivariate logistic regression models. We chose to use logistic regression because the outcome of interest is dichotomous (any tooth loss, yes or no) and logistic regression is an appropriate model for binary outcomes. Four models were created. Three of the models were clinically based with one model built for each clinical measure (ACH, CAL, PPD) separately as the main independent variable, and then determining which other variables identified in univariate analyses added significantly to these models. A 4th model was built attempting only those variables identified in the univariate analyses that could be assessed without a clinical oral examination, described as the “community” model. Forward stepwise logistic regression was used to create these models, attempting entry at a significance level of p=0.20 and removal at p=0.05. Each step of a regression model was assessed against the others by comparing the changes in the individual significance of the independent variables. This approach incorporated a priori knowledge of known, important risk factors as well as an objective selection of factors which were statistically significant. After all stepwise procedures were completed, models were re-run using the “enter” method to maximize the number of participants included in the final models, as a result of missing data for individuals who did not get included in the forward models. Using the same process, the community-based model was developed. The forward selection process allowed for a step-by-step assessment of the importance of each factor in contributing to the overall model. In each of the clinical models, ACH, CAL and PPD, only one clinical measure was included to avoid collinearity with the other related oral measures. Within our data set, correlations among the independent predictors other than ACH, CAL and PPD were all quite small to moderate, indicating a lack of multicolliearity. Linearity between continuous independent variables and tooth loss (dichotomous outcome) was assessed by regressing each individual predictor against the log odds of tooth loss for the entered independent variable. All statistical analyses were conducted using SPSS 15.0.1 for Windows.
Results
Study participant baseline characteristics are shown in Table 1. At baseline the mean number of teeth present per person was 23.5 (±5.2, range 6–28). While the baseline study required that only women with at least 6 teeth present were eligible, only 1% had this few teeth at enrollment and 91% began the study with more than half (16 to 28) of their teeth intact.
Table 1.
Participant characteristics of postmenopausal women in osteoporosis prospective study at baseline enrollment overall, and by tooth loss status at 5-year follow-up
| Baseline Measures |
Total Sample (N=1021) n (%)* Mean (SD) |
Tooth Loss – Yes (N=293) n (%) Mean (SD) |
Tooth Loss – No (N=728) n (%) Mean (SD) |
P-Value (t- test or Chi Square) |
|---|---|---|---|---|
| Demographics | ||||
| Age at baseline visit (yrs) | 65.88 (6.70) | 66.74 (6.95) | 65.53 (6.57) | 0.01 |
| Race (White) | 1000 (97.9) | 286 (97.6) | 714 (98.1) | 0.59 |
| Education | ||||
| High school or less | 212 (21.2) | 65 (22.6) | 147 (20.6) | 0.73 |
| College or higher | 790 (78.8) | 223 (77.4) | 567 (79.4) | |
| Marital status | ||||
| Currently married or cohabitating |
728 (71.3) | 216 (73.7) | 512 (70.3) | 0.21 |
| Other | 293 (28.7) | 77 (26.3) | 216 (29.7) | |
| Total family income last year | ||||
| Less than $19,999 | 127 (13.1) | 42 (15.1) | 85 (12.3) | 0.22 |
| $20–44,999 | 451 (46.5) | 142 (51.0) | 309 (44.7) | |
| $50 or more | 391 (40.4) | 94 (33.9) | 297 (43.0) | |
| Lifestyle | ||||
| Ever smoke cigarettes (yes) | 469 (45.9) | 158 (53.9) | 311 (42.7) | 0.001 |
| Brush Teeth | ||||
| 1× or less per day | 234 (22.9) | 78 (26.6) | 156 (21.4) | 0.07 |
| 2× or more per day | 788 (77.1) | 215 (73.4) | 572 (78.6) | |
| Floss Teeth | ||||
| Not every week | 181 (17.8) | 59 (20.3) | 122 (16.8) | 0.18 |
| 1× per week | 93 (9.2) | 20 (6.9) | 73 (10.1) | |
| More than 1× per week | 293 (28.9) | 77 (26.6) | 216 (29.8) | |
| Every day | 448 (44.1) | 134 (46.2) | 314 (43.3) | |
| Visit Dentist | ||||
| Once per year or more | 939 (92.0) | 261 (89.1) | 678 (93.1) | 0.03 |
| Only with problem or never | 82 (8.0) | 32 (10.9) | 50 (6.9) | |
| Health | ||||
| BMI (kg/m2) | 26.57 (5.12) | 27.44 (5.60) | 26.23 (4.88) | 0.001 |
| Diabetes history (yes) | 43 (4.2) | 22 (7.5) | 21 (2.9) | <0.01 |
| Years past menopause without hormone therapy |
12.35 (9.75) | 13.12 (9.92) | 12.02 (9.66) | 0.11 |
| Current hormone therapy (yes) | 491 (48.1) | 121 (41.3) | 370 (50.8) | 0.01 |
| Current HT or bone drug use (yes) |
552 (54.1) | 142 (48.5) | 410 (56.3) | 0.23 |
| Worst T-score | ||||
| Normal, T=>−1.0 | 173 (16.9) | 42 (14.3) | 131 (18.0) | 0.30 |
| Osteopenia, −2.5<T<−1.0 | 469 (45.9) | 143 (48.8) | 326 (44.8) | |
| Osteoporosis, T=<−2.5 | 379 (37.1) | 108 (36.9) | 271 (37.2) | |
| Oral Health | ||||
| Tooth loss (yes) | 837 (82.0) | 256 (87.4) | 581 (79.8) | <0.01 |
| Mean number of teeth | 23.5 (5.21) | 22.3 (5.42) | 23.9 (5.05) | <0.001 |
| Whole mouth mean alveolar crestal height (ACH) (mm) |
2.44 (0.76) | 2.67 (0.83) | 2.35 (0.71) | <0.01 |
| Worst ACH site (mm) | 4.60 (1.54) | 5.10 (1.77) | 4.40 (1.39) | <0.01 |
| Mean clinical attachment loss (CAL) (mm) |
2.27 (0.64) | 2.40 (0.7) | 2.21 (0.6) | <0.01 |
| Worst CAL (mm) | 5.91 (1.89) | 6.40 (2.06) | 5.71 (1.78) | <0.01 |
| Mean Pocket Depth (PPD) (mm) |
1.93 (0.35) | 1.99 (0.4) | 1.91 (0.3) | <0.01 |
| Worst PPD (mm) | 4.94 (1.37) | 5.30 (1.58) | 4.79 (1.24) | <0.01 |
| Oral surgery or extraction due to gum disease |
||||
| Ever | 236 (23.8) | 97 (34.5) | 139 (19.6) | <0.01 |
| Never | 754 (76.2) | 184 (65.5) | 570 (80.4) | |
| Mean Plaque Index (%) | 51.27 (24.7) | 55.75 (25.3) | 49.48 (24.3) | <0.01 |
| Mean Gingival Bleeding (%) | 33.51 (22.4) | 36.07 (25.8) | 32.48 (20.9) | 0.02 |
| Mean Calculus Index | 0.63 (0.57) | 0.69 (0.6) | 0.60 (0.6) | 0.03 |
| Number of sites with 4 mm or more |
2.67 (4.14) | 3.64 (4.81) | 2.28 (3.78) | <0.01 |
| Follow-up Measures | ||||
| Mean number of teeth at follow- up |
22.92 (5.55) | 20.41 (5.93) | 23.9 (5.05) | <0.01 |
| Follow-up time (years) | 5.07 (0.38) | 5.09 (0.41) | 5.06 (0.37) | 0.22 |
Percentages out of valid responses; Total Missing: Education=19, Income=52, Flossing=6, Oral Surgery=31; there were no significant differences of percentage Total Missing between the two groups
The most significant differences between those who lost any teeth and those who did not were seen in the proportion of ever-smokers (54% and 43%, respectively), higher mean BMI (27.4 compared to 26.2) and number of teeth at baseline (22.3 compared to 23.9); all p≤0.001. Other factors significantly associated with incident tooth loss included older age, lower frequency of dental visits, diabetes history, not currently using hormone therapy, worse ACH, CAL, PD, plaque or gingivitis, and any history of oral surgery or extraction due to gum disease. Follow-up time was similar in those who lost any teeth (5.09 years) and those who did not (5.06 years) (p=0.22) (Table 1).
Oral clinical measures are also included in Table 1. The mean baseline ‘worst ACH’ site among all participants was 4.60 mm (±1.5), with a significant difference (p<0.01) between those experiencing subsequent tooth loss (5.10 ±1.8) and those who did not (4.40 ±1.4). The mean number of sites with 4 mm or more of bone loss was 2.7 with a range of zero to 31. CAL at worst site at baseline was significantly higher among those with incident tooth loss compared to no loss (6.4 ±2.1 and 5.7 ±1.8, respectively; p<0.01); as was PPD at worst site (with loss, 5.3 ±1.6 vs. no loss, 4.8 ±1.2; p<0.01). Among all women, the mean proportion of teeth with any plaque was 51.3% (range 0–100%) and mean gingivitis (gingival bleeding) was 33.5% (range 0–100%).
The 5-year cumulative incidence of any tooth loss was 293 in 1,021 participants (28.7 per hundred women) (Table 2). Of the 293 participants with tooth loss at follow-up, 175 (59.7%) had lost only one tooth, 67 (22.9%) had lost 2 teeth and 51 (17.4%) had lost 3 or more teeth. Among participants losing teeth, an average loss of 1.85 teeth per person (SD = 1.63, range 1–14) occurred. One participant who lost all their teeth during the study period (7 teeth lost out of 7 at baseline). The main reason reported for tooth loss during follow-up caries (64.4%), followed by periodontal disease (13.6%). Other reasons given for incident tooth loss are reported in table 2.
Table 2.
Distribution of tooth loss, reported reasons for loss, and time to follow-up from baseline among postmenopausal women during 5-year follow-up
| Number of teeth lost | Total participants n (%) |
Reasons for tooth loss | Total number of teeth lost = 323 n (%)† |
|---|---|---|---|
| 1 | 175 (59.7) | Caries | 209 (64.7) |
| 2 | 67 (22.9) | Periodontal disease | 44 (13.6) |
| 3 | 25 (8.5) | Retained root | 23 (7.1) |
| 4 | 10 (3.4) | Root canal | 19 (5.9) |
| 5 | 4 (1.4) | Fracture/accident | 11 (3.4) |
| 6 | 4 (1.4) | Implant | 9 (2.8) |
| 7 | 3 (1.0) | Ortho/malposed | 6 (1.9) |
| 8 | 1 (0.3) | Unable to determine | 2 (<1.0) |
| 10 | 3 (1.0) | ||
| 14 | 1 (0.3) | ||
| Total with any tooth loss | 323 in 293 women |
Participants who lost more than one tooth may report more than one reason for the losses – percentages out of total teeth lost (n=323) by entire group.
Results from the logistic regression analyses are shown in Table 3. Each of the 3 clinical measures (ACH, CAL, PPD) were significant risk factors for incident tooth loss. Other variables that were also statistically significant risk factors of tooth loss in the multivariable adjusted analyses included diabetes history, prior gum disease/surgery, ever smoking, number of teeth lost at baseline, BMI and mean plaque index. The OR estimates for the non-clinical factors (in addition to each of the 3 clinical measures) across the three models were all very similar. Each of the models showed the greatest contribution to risk of tooth loss coming from diabetes history with about a 2.5 fold increase in risk. Surgery or extraction due to gum disease was associated with a 55% to 64% increased risk in the clinical models.
Table 3.
Multivariate logistic regression models for assessing risk of tooth loss among postmenopausal women
| Included in analyses n=981 | Odds Ratio Point Estimate |
95% CI Lower Limit |
95% CI Upper Limit |
P-value |
|---|---|---|---|---|
| ACH Model | ||||
| Worst ACH (per mm) | 1.22 | 1.11 | 1.35 | <0.001 |
| Diabetes history (yes) | 2.52 | 1.28 | 4.95 | 0.007 |
| Gum disease/surgery (yes) | 1.56 | 1.10 | 2.19 | 0.012 |
| Ever smoke (yes) | 1.40 | 1.05 | 1.88 | 0.024 |
| Teeth lost at baseline (per tooth) | 1.03 | 1.00 | 1.06 | 0.030 |
| BMI (per 5 kg/m2 increase) | 1.16 | 1.01 | 1.34 | 0.036 |
|
Mean plaque index (per 10% increase) |
1.08 | 1.02 | 1.15 | 0.009 |
| CAL Model | ||||
| Worst CAL (per mm) | 1.13 | 1.05 | 1.23 | 0.002 |
| Diabetes history (yes) | 2.40 | 1.23 | 4.70 | 0.011 |
| Gum disease/surgery (yes) | 1.64 | 1.17 | 2.31 | 0.004 |
| Ever smoke (yes) | 1.43 | 1.06 | 1.91 | 0.018 |
| Teeth lost at baseline (per tooth) | 1.04 | 1.01 | 1.07 | 0.010 |
| BMI (per 5 kg/m2 increase) | 1.16 | 1.01 | 1.34 | 0.037 |
|
Mean plaque index (per 10% increase) |
1.07 | 1.01 | 1.15 | 0.018 |
| PPD Model | ||||
| Worst PPD (per mm) | 1.26 | 1.13 | 1.41 | <0.001 |
| Diabetes history (yes) | 2.48 | 1.27 | 4.86 | 0.008 |
| Gum disease/surgery (yes) | 1.62 | 1.16 | 2.27 | 0.005 |
| Ever smoke (y/n) | 1.39 | 1.03 | 1.86 | 0.030 |
| Teeth lost at baseline (per tooth) | 1.05 | 1.02 | 1.08 | <0.001 |
| BMI (per 5 kg/m2 increase) | 1.14 | 1.00 | 1.31 | 0.070 |
|
Mean plaque index (per 10% increase) |
1.07 | 1.00 | 1.14 | 0.036 |
| Community Model(n=990) | ||||
| Diabetes history (yes) | 2.45 | 1.26 | 4.77 | 0.008 |
| Gum disease/surgery (yes) | 1.97 | 1.43 | 2.70 | <0.001 |
| Ever smoke (yes) | 1.42 | 1.06 | 1.89 | 0.018 |
| Teeth lost at baseline (per tooth) | 1.05 | 1.02 | 1.08 | <0.001 |
| BMI (per 5 kg/m2 increase) | 1.15 | 1.01 | 1.33 | 0.038 |
In the ACH multivariable adjusted model, each additional 1mm of ACH at baseline was associated with a 22% increase in risk for tooth loss. Similar findings were found in the PPD and CAL models. In the CAL model, each 1 mm worse attachment loss was associated with a 13% increased risk in tooth loss, while each 1 mm worse pocket depth was associated with a 26% increased risk of tooth loss in the next 5 years. The community model that used information from self-reported questionnaires included diabetes history, prior gum disease/surgery, ever smoking, number of teeth lost at baseline and BMI as significant risk factors for incident tooth loss (Table 3).
Table 4 shows the multivariable adjusted results from subgroup analyses stratified on tooth loss history. At baseline, a total of 184 (18%) women had no missing teeth. Among these women, risk of incident tooth loss was significantly associated in a multivariable model including ever smoking, worst ACH and prior hormone therapy use. Among those beginning the study period with at least 1 prior tooth loss (n=837, 82%), risk of incident tooth loss during the follow-up period included diabetes history, prior gum disease/surgery, worst pocket depth, worst ACH, BMI, and number of teeth lost at baseline.
Table 4.
Multivariable logistic regression estimates for assessing risk of tooth loss in postmenopausal women according to tooth loss at baseline
| Odds Ratio Point Estimate |
95% CI Lower Limit |
95% CI Upper Limit |
P-value | |
|---|---|---|---|---|
|
No Missing Teeth at Baseline n = 184 |
||||
| Ever smoke (yes) | 2.40 | 1.08 | 5.33 | 0.032 |
| Worst ACH (per mm) | 1.64 | 1.20 | 2.24 | 0.002 |
| Hormone therapy ever (yes) | 0.34 | 0.15 | 0.79 | 0.012 |
|
Any Missing Teeth at Baseline n = 837 |
||||
| Diabetes history (yes) | 2.45 | 1.22 | 4.94 | 0.012 |
| Gum disease/surgery (yes) | 1.52 | 1.05 | 2.20 | 0.025 |
| Worst PPD (per mm) | 1.22 | 1.08 | 1.38 | 0.001 |
| Worst ACH (per mm) | 1.13 | 1.01 | 1.26 | 0.031 |
| BMI (per 5 kg/m2) | 1.28 | 1.10 | 1.47 | 0.002 |
| Teeth lost at baseline (per tooth) | 1.04 | 1.01 | 1.07 | 0.008 |
Discussion
In the current study we examined risk factors of incident tooth loss in a prospective cohort of postmenopausal women. The observed cumulative rate of person-level tooth loss of nearly 30% within a 5 year period is compatible with other published findings, underscoring the fact that tooth loss occurs frequently and affects a large number of people, including postmenopausal women (23,32–36).
Our study may be the first to evaluate three separate clinical measures in a series of multivariate models in one prospective cohort. In addition, this study was able to evaluate a non-clinical model that used information solely from questionnaire data in the same cohort. All 4 models yielded similar results in their ability to simultaneously assess the contributions from multiple factors towards risk of incident tooth loss in these postmenopausal women.
Although prior studies evaluating ACH, CAL and PPD in relation to tooth loss have been conducted and found positive associations (12,23,33), few if any have been able to evaluate the individual effect magnitude of these clinical measures in one cohort, and be able to assess the magnitude of other important factors that contribute to these models at the same time. In addition, some previous studies have been relatively small (33). In our study, a fourth non-clinical community model was evaluated in the same cohort and found similar strengths as the clinically based models. This non-clinical model is relevant and may prove useful in settings where no clinical measures are available or attainable. Even within clinical settings, not all measures (ACH, CAL, PPD) may be available in all patients. Therefore, the consistency of all 4 models in assessing risk of tooth loss makes these findings relevant to most dental practitioners.
In the three clinical models, a set of additional factors that were associated with risk of incident tooth loss included history of diabetes, prior gum disease/surgery, cigarette smoking, number of missing teeth at baseline examination, BMI, and plaque. In the community model, significant risk factors for incident tooth loss included diabetes history, prior gum disease/surgery, smoking, number of teeth missing at baseline and BMI. Many of these factors have been previously reported as individual risk factors for tooth loss. However, our study is one of the first to assess 4 multivariate models prospectively in a cohort of postmenopausal women. Factors that may be important in this group are likely different from those in younger populations and men.
Consistent with our findings, many of the additional risk factors we describe have been identified in other prospective studies (9,11–13,15,16,23,36–40). Factors such as diabetes (41–43) and higher BMI (44) were important risk factors for tooth loss. Other studies have found oral and other factors such as hormone use, smoking, diabetes, partial endentulism, gum disease/surgery, bone or attachment loss, probing depth and plaque, are related to subsequent tooth loss (4,5,45–53).
A retrospective study within the Framingham Heart Study supported our findings that estrogen use is protective against tooth loss (54), and four cross-sectional studies identified gender (female), years since menopause, smoking, socio-economic status and attachment loss to correlate with tooth loss (55–58). However, not all studies have found associations with these factors. A retrospective study of postmenopausal Japanese women did not find that estrogen use, years since menopause or BMI were associated with tooth loss (59). However, over half of this group had prior hysterectomy or oophorectomy; smokers, diabetics and past estrogen users were excluded, limiting comparability. Since our study was restricted to postmenopausal females, not all study findings could be compared. Other previous studies have included men, different age groups, other ethnic and racial groups and populations with notably different prevalence rates of other diseases. Our cohort was a relatively healthy and ambulatory group of women who had relatively good oral hygiene habits.
Interestingly, risk factors of incident tooth loss varied by prevalence of tooth loss at baseline enrollment. In analyses stratified on prevalence of tooth loss at baseline (Table 4), smoking status showed the most risk of tooth loss among women who had no tooth loss at baseline, along with ACH, while hormone therapy use was a significant protective factor in those women. Among women with prevalent tooth loss at baseline, history of diabetes emerged as the strongest risk factor for future tooth loss. However, oral health indicators (ACH, PPD, prior oral surgery), BMI, and number of previous teeth lost were also significantly associated with incident tooth loss. As an illustration in those without prevalent tooth loss at baseline, a non-smoking, postmenopausal woman with average ‘worst ACH’ (5 mm), having previously taken estrogen, would have a 14% chance of losing at least one tooth in the next 5 years, while a smoker and non-estrogen user would have a 54% chance of tooth loss in this time period. In those with prevalent tooth loss at baseline, a non-diabetic, postmenopausal woman with a BMI (30 kg/m2), average levels of ‘worst ACH’ (5 mm), number of teeth currently lost (5), PPD (5 mm) and no prior gum surgery or tooth extraction due to gum disease would have a 30% probability of losing a tooth within 5 years. However, when a positive history of diabetes is included, the same individual would have an 51% chance of tooth loss. These data suggest there may be different etiologies at play for those experiencing tooth loss earlier in life as opposed to in later postmenopausal years. In addition, differences in risk were identified according to number of teeth lost at baseline. The women with 20 or fewer teeth present at baseline (n=191) lost an average of 1.04 teeth during follow-up, compared 0.41 teeth in those with 21 or more teeth at baseline (n=830). This finding supports the fact that prevalent tooth loss predicts future loss, even though those individuals have fewer teeth remaining to lose.
While these models will require validation to assess their predictive value in other cohorts, the following additional illustrations may be informative to future development of such predictive models for postmenopausal women. Using the ACH multivariate risk model as an example, the probability of tooth loss for a non-diabetic, non-smoking postmenopausal woman with a BMI (30 kg/m2), average levels of ‘worst ACH’ (5 mm), number of teeth currently lost (5), plaque index (50%), and no prior gum surgery or tooth extraction due to gum disease has a 25% probability of losing a tooth within 5 years. The results using the clinically derived CAL and PPD measures and same ‘average’ conditions are virtually the same, with probabilities of tooth loss between 21–24%. Using the community model, the same individual as previously described would have a 23% chance of tooth loss. By using the average of the highest (worst) quintile for each variable ACH (7 mm), CAL (7 mm), PPD (7 mm), number of teeth lost at baseline (14), and BMI (35 kg/m2), for a postmenopausal woman with diabetes, smoker, history of gum surgery or tooth extraction due to gum disease, and plaque index (90%), we calculate an 80–87% probability for the loss of at least one tooth in the next 5 years. The community model, which omits the clinical measures, gives a similar estimate for likelihood of tooth loss, 79%.
In considering these findings, some potential limitations should be acknowledged. Our participants were fairly homogeneous and our results may not be generalizable to other racial or SES groups, or to the U.S. population at large. We were unable to carefully examine differences by factors such as race, education, income, type of insurance or marital status. In many longitudinal studies, bias due to nonresponse may be of concern. However, within our study population we maintained follow-up with over 75% of participants (Figure 1), reducing the potential influence of non-response bias. Women, and those who are married, or with higher education or income, are more likely than men to seek dental care (60), especially if dental coverage is carried (61), which could lead to more extractions. However, we did not capture the type of clinical center where these women received treatment or the type of dental practitioners sought out, thus were unable to assess this potential predictor. Our participants did have a relatively high proportion indicating preventive care within the last 12 months. Studies that enroll participants from dental clinics alone could lead to biased interpretations related to reasons for tooth loss. However, we did not recruit from dental clinics or based on dental care, we enrolled women from the original Buffalo cohort of the WHI that was recruited from the western NY community using flyers and letters sent to households and through regional advertisement. In the original WHI enrollment there was no focus on dental outcomes. Also, the majority of tooth extractions in adults are typically due to periodontal disease or caries, and more often extractions are recommended by the dentist than requested by the patient (62–68).
Strengths of the study are its large sample size of community-dwelling, postmenopausal women. We maintained a high response rate throughout follow-up. The follow-up duration allowed for sufficient events to conduct some sub-group analyses. The examinations were done using standardized protocols and trained professionals, reducing potential for misclassification and some sources of bias. By assessing clinical and socio-demographic baseline exposures, including tooth loss, we were able to establish temporal associations.
We developed a series of clinical and questionnaire-based multivariate risk models for incident tooth loss in older women. This information may help to identify individuals at increased risk, where preventive strategies may be targeted.
Acknowledgements
Grant Support: This study was supported by funding from National Heart, Lung, and Blood Institute (NHLBI) contract N01WH32122, NIH/NIDCR grants 1R01-DE13505 and DE04898, and USARMC grant DAMD 17-96-1-6319. Mr. Bole is supported on NCI training grant R25CA113951.
References
- 1.Elter JR, Champagne CME, Offenbacher S, Beck JD. Relationship of periodontal disease and tooth loss to prevalence of coronary heart disease. Journal of Periodontology. 2004;75(6):782–790. doi: 10.1902/jop.2004.75.6.782. [DOI] [PubMed] [Google Scholar]
- 2.Hamalainen P, Meurman JH, Keskinen M, Heikkinen E. Relationship between dental health and 10-year mortality in a cohort of community-dwelling elderly people. European Journal of Oral Sciences. 2003;111(4):291–296. doi: 10.1034/j.1600-0722.2003.00055.x. [DOI] [PubMed] [Google Scholar]
- 3.Marcenes W, Steele JG, Sheiham A, Walls AW. The relationship between dental status, food selection, nutrient intake, nutritional status, and body mass index in older people. Cadernos de saude publica / Ministerio da Saude, Fundacao Oswaldo Cruz, Escola Nacional de Saude Publica. 2003;19(3):809–816. doi: 10.1590/s0102-311x2003000300013. [DOI] [PubMed] [Google Scholar]
- 4.Kapp JM, Boren SA, Yun S, LeMaster J. Diabetes and tooth loss in a national sample of dentate adults reporting annual dental visits. Preventing chronic disease. 2007;4(3) [PMC free article] [PubMed] [Google Scholar]
- 5.Sheiham A, Steele JG, Marcenes W, Finch S, Walls AWG. The relationship between oral health status and Body Mass Index among older people: A national survey of older people in Great Britain. British Dental Journal. 2002;192(12):703–706. doi: 10.1038/sj.bdj.4801461. [DOI] [PubMed] [Google Scholar]
- 6.Petersen PE, Yamamoto T. Improving the oral health of older people: The approach of the WHO Global Oral Health Programme. Community Dentistry and Oral Epidemiology. 2005;33(2):81–92. doi: 10.1111/j.1600-0528.2004.00219.x. [DOI] [PubMed] [Google Scholar]
- 7.Lamster IB. Oral Health Care Services for Older Adults: A Looming Crisis. American Journal of Public Health. 2004;94(5):699–702. doi: 10.2105/ajph.94.5.699. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Dye BA, Tan S, Smith V, Lewis BG, Barker LK, Thornton-Evans G, et al. Trends in oral health status: United States 1988–1994 and 1999–2004 Vital and health statistics. Series 11, Data from the national health survey. 2007;(248):1–92. [PubMed] [Google Scholar]
- 9.Copeland LB, Krall EA, Brown LJ, Garcia RI, Streckfus CF. Predictors of tooth loss in two US adult populations. Journal of Public Health Dentistry. 2004;64(1):31–37. doi: 10.1111/j.1752-7325.2004.tb02723.x. [DOI] [PubMed] [Google Scholar]
- 10.Muzzi L, Nieri M, Cattabriga M, Rotundo R, Cairo F, Pini Prato GP. The potential prognostic value of some periodontal factors for tooth loss: A retrospective multilevel analysis on periodontal patients treated and maintained over 10 years. Journal of Periodontology. 2006;77(12):2084–2089. doi: 10.1902/jop.2006.050227. [DOI] [PubMed] [Google Scholar]
- 11.Gilbert GH, Shelton BJ, Chavers LS, Bradford EH., Jr Predicting tooth loss during a population-based study: Role of attachment level in the presence of other dental conditions. Journal of Periodontology. 2002;73(12):1427–1436. doi: 10.1902/jop.2002.73.12.1427. [DOI] [PubMed] [Google Scholar]
- 12.Machtei EE, Hausmann E, Dunford R, Grossi S, Ho A, Davis G, et al. Longitudinal study of predictive factors for periodontal disease and tooth loss. Journal of Clinical Periodontology. 1999;26(6):374–380. doi: 10.1034/j.1600-051x.1999.260607.x. [DOI] [PubMed] [Google Scholar]
- 13.Gilbert GH, Miller MK, Duncan RP, Ringelberg ML, Dolan TA, Foerster U. Tooth-specific and person-level predictors of 24-month tooth loss among older adults. Community Dentistry and Oral Epidemiology. 1999;27(5):372–385. doi: 10.1111/j.1600-0528.1999.tb02034.x. [DOI] [PubMed] [Google Scholar]
- 14.Slade GD, Gansky SA, Spencer AJ. Two-year incidence of tooth loss among South Australians aged 60+ years. Community Dentistry and Oral Epidemiology. 1997;25(6):429–437. doi: 10.1111/j.1600-0528.1997.tb01734.x. [DOI] [PubMed] [Google Scholar]
- 15.Beck JD, Sharp T, Koch GG, Offenbacher S. A 5-year study of attachment loss and tooth loss in community-dwelling older adults. Journal of Periodontal Research. 1997;32(6):516–523. doi: 10.1111/j.1600-0765.1997.tb00567.x. [DOI] [PubMed] [Google Scholar]
- 16.Baelum V, Luan W-, Chen X, Fejerskov O. Predictors of tooth loss over 10 years in adult and elderly Chinese. Community Dentistry and Oral Epidemiology. 1997;25(3):204–210. doi: 10.1111/j.1600-0528.1997.tb00927.x. [DOI] [PubMed] [Google Scholar]
- 17.Nordstrom G, Bergman B, Borg K, Nilsson H, Tillberg A, Wenslov J- A 9-year longitudinal study of reported oral problems and dental and periodontal status in 70- and 79-year-old city cohorts in northern Sweden. Acta Odontologica Scandinavica. 1998;56(2):76–84. doi: 10.1080/00016359850136021. [DOI] [PubMed] [Google Scholar]
- 18.Meisel P, Reifenberger J, Haase R, Nauck M, Bandt C, Kocher T. Women are periodontally healthier than men, but why don't they have more teeth than men? Menopause. 2008;15(2):270–275. doi: 10.1097/gme.0b013e31811ece0a. [DOI] [PubMed] [Google Scholar]
- 19.Kaye EK. Bone health and oral health. J.Am.Dent.Assoc. 2007;138(5):616–619. doi: 10.14219/jada.archive.2007.0232. [DOI] [PubMed] [Google Scholar]
- 20.Soder PO, Jin LJ, Soder B, Wikner S. Periodontal status in an urban adult population in Sweden. Community Dent.Oral Epidemiol. 1994;22(2):106–111. doi: 10.1111/j.1600-0528.1994.tb01582.x. [DOI] [PubMed] [Google Scholar]
- 21.Desvarieux M, Schwahn C, Volzke H, Demmer RT, Ludemann J, Kessler C, et al. Gender differences in the relationship between periodontal disease, tooth loss, and atherosclerosis. Stroke. 2004;35(9):2029–2035. doi: 10.1161/01.STR.0000136767.71518.36. [DOI] [PubMed] [Google Scholar]
- 22.Haikola B, Oikarinen K, Söderholm A-, Remes-Lyly T, Sipilä K. Prevalence of edentulousness and related factors among elderly Finns. J.Oral Rehabil. 2008;35(11):827–835. doi: 10.1111/j.1365-2842.2008.01873.x. [DOI] [PubMed] [Google Scholar]
- 23.Warren JJ, Watkins CA, Cowen HJ, Hand JS, Levy SM, Kuthy RA. Tooth loss in the very old: 13–15-year incidence among elderly Iowans. Community Dentistry and Oral Epidemiology. 2002;30(1):29–37. doi: 10.1034/j.1600-0528.2002.300105.x. [DOI] [PubMed] [Google Scholar]
- 24.Hamalainen P, Meurman JH, Keskinen M, Heikkinen E. Changes in dental status over 10 years in 80-year-old people: A prospective cohort study. Community Dentistry and Oral Epidemiology. 2004;32(5):374–384. doi: 10.1111/j.1600-0528.2004.00178.x. [DOI] [PubMed] [Google Scholar]
- 25.Wactawski-Wende J, Hausmann E, Hovey K, Trevisan M, Grossi S, Genco RJ. The association between osteoporosis and alveolar crestal height in postmenopausal women. Journal of Periodontology. 2005;76(11 SUPPL):2116–2124. doi: 10.1902/jop.2005.76.11-S.2116. [DOI] [PubMed] [Google Scholar]
- 26.Brennan RM, Genco RJ, Hovey KM, Trevisan M, Wactawski-Wende J. Clinical attachment loss, systemic bone density, and subgingival calculus in postmenopausal women. Journal of Periodontology. 2007;78(11):2104–2111. doi: 10.1902/jop.2007.070155. [DOI] [PubMed] [Google Scholar]
- 27.Hausmann E, Allen K, Christersson L, Genco RJ. Effect of x-ray beam vertical angulation on radiographic alveolar crest level measurement. Journal of Periodontal Research. 1989;24(1):8–19. doi: 10.1111/j.1600-0765.1989.tb00852.x. [DOI] [PubMed] [Google Scholar]
- 28.Hausmann E, Allen K, Clerehugh V. What alveolar crest level on a bite-wing radiograph represents bone loss? Journal of Periodontology. 1991;62(9):570–572. doi: 10.1902/jop.1991.62.9.570. [DOI] [PubMed] [Google Scholar]
- 29.Hausmann E, Allen K, Dunford R, Christersson L. A reliable computerized method to determine the level of the radiographic alveolar crest. Journal of Periodontal Research. 1989;24(6):368–369. doi: 10.1111/j.1600-0765.1989.tb00884.x. [DOI] [PubMed] [Google Scholar]
- 30.Araujo MWB, Hovey KM, Benedek JR, Grossi SG, Dorn J, Wactawski-Wende J, et al. Reproducibility of probing depth measurements using a constant-force electronic probe: Analysis of inter- and intraexaminer variability. J.Periodontol. 2003;74(12):1736–1740. doi: 10.1902/jop.2003.74.12.1736. [DOI] [PubMed] [Google Scholar]
- 31.Grossi Sg, Dunford RG, Ho A, Koch G, Machtei EE, Genco RJ. Sources of error for periodontal probing measurements. Journal of Periodontal Research. 1996;31:330–336. doi: 10.1111/j.1600-0765.1996.tb00500.x. [DOI] [PubMed] [Google Scholar]
- 32.Fure S, Zickert I. Incidence of tooth loss and dental caries in 60-, 70- and 80-year-old Swedish individuals. Community Dentistry and Oral Epidemiology. 1997;25(2):137–142. doi: 10.1111/j.1600-0528.1997.tb00911.x. [DOI] [PubMed] [Google Scholar]
- 33.Tezal M, Wactawski-Wende J, Grossi SG, Dmochowski J, Genco RJ. Periodontal disease and the incidence of tooth loss in postmenopausal women. Journal of Periodontology. 2005;76(7):1123–1128. doi: 10.1902/jop.2005.76.7.1123. [DOI] [PubMed] [Google Scholar]
- 34.Hand JS, Hunt RJ, Kohout FJ. Five-year incidence of tooth loss in Iowans aged 65 and older. Community dentistry and oral epidemiology. 1991;19(1):48–51. doi: 10.1111/j.1600-0528.1991.tb00105.x. [DOI] [PubMed] [Google Scholar]
- 35.Drake CW, Hunt RJ, Koch GG. Three-year tooth loss among black and white older adults in North Carolina. Journal of Dental Research. 1995;74(2):675–680. doi: 10.1177/00220345950740020801. [DOI] [PubMed] [Google Scholar]
- 36.Locker D, Ford J, Leake JL. Incidence of and risk factors for tooth loss in a population of older Canadians. Journal of Dental Research. 1996;75(2):783–789. doi: 10.1177/00220345960750020801. [DOI] [PubMed] [Google Scholar]
- 37.Grodstein F, Colditz GA, Stampfer MJ. Compendium of continuing education in dentistry. 22. Vol. 1998. Jamesburg NJ: 1995. Tooth loss and hormone use in postmenopausal women; pp. S9–S16. [PubMed] [Google Scholar]
- 38.Shimazaki Y, Soh I, Koga T, Miyazaki H, Takehara T. Risk factors for tooth loss in the institutionalised elderly; a six-year cohort study. Community Dental Health. 2003;20(2):123–127. [PubMed] [Google Scholar]
- 39.Neely AL, Holford TR, Loe H, Anerud A, Boysen H. The natural history of periodontal disease in humans: Risk factors for tooth loss in caries-free subjects receiving no oral health care. Journal of Clinical Periodontology. 2005;32(9):984–993. doi: 10.1111/j.1600-051X.2005.00797.x. [DOI] [PubMed] [Google Scholar]
- 40.Mundt T, Schwahn C, Mack F, Polzer I, Samietz S, Kocher T, et al. Risk indicators for missing teeth in working-age pomeranians - An evaluation of high-risk populations. Journal of Public Health Dentistry. 2007;67(4):243–249. doi: 10.1111/j.1752-7325.2007.00041.x. [DOI] [PubMed] [Google Scholar]
- 41.Klein BEK, Klein R, Knudtson MD. Life-style correlates of tooth loss in an adult midwestern population. Journal of Public Health Dentistry. 2004;64(3):145–150. doi: 10.1111/j.1752-7325.2004.tb02744.x. [DOI] [PubMed] [Google Scholar]
- 42.Faggion CM, Jr, Petersilka G, Lange DE, Gerss J, Flemmig TF. Prognostic model for tooth survival in patients treated for periodontitis. Journal of Clinical Periodontology. 2007;34(3):226–231. doi: 10.1111/j.1600-051X.2006.01045.x. [DOI] [PubMed] [Google Scholar]
- 43.Taylor GW, Manz MC, Borgnakke WS. Compendium of continuing education in dentistry. Jamesburg N.J: 1995. Diabetes periodontal diseases dental caries, and tooth loss: a review of the literature. 2004 25 3:179–184, 186–188, 190; quiz 192. [PubMed] [Google Scholar]
- 44.Grodstein F, Colditz GA, Stampfer MJ. Post-menopausal hormone use, and tooth loss: A prospective study. J AmDent Assoc. 1996;127(3):370–378. doi: 10.14219/jada.archive.1996.0208. [DOI] [PubMed] [Google Scholar]
- 45.Eklund SA, Burt BA. Risk factors for total tooth loss in the United States; longitudinal analysis of national data. Journal of public health dentistry. 1994;54(1):5–14. doi: 10.1111/j.1752-7325.1994.tb01173.x. [DOI] [PubMed] [Google Scholar]
- 46.Al-Shammari KF, Al-Khabbaz AK, Al-Ansari JM, Neiva R, Wang H- Risk indicators for tooth loss due to periodontal disease. Journal of Periodontology. 2005;76(11):1910–1918. doi: 10.1902/jop.2005.76.11.1910. [DOI] [PubMed] [Google Scholar]
- 47.Taguchi A, Tsuda M, Ohtsuka M, Nakamoto T, Inagaki K, Noguchi T, et al. Menopause. 3 Pt 1. Vol. 14. New York, N.Y.: 2007. Interaction of obesity and skeletal bone mineral density in tooth retention in Japanese postmenopausal women; pp. 500–504. [DOI] [PubMed] [Google Scholar]
- 48.Hanioka T, Ojima M, Tanaka K, Aoyama H. Relationship between smoking status and tooth loss: Findings from national databases in Japan. Journal of Epidemiology. 2007;17(4):125–132. doi: 10.2188/jea.17.125. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 49.Tanaka K, Miyake Y, Sasaki S, Ohya Y, Miyamoto S, Matsunaga I, et al. Active and passive smoking and tooth loss in Japanese women: Baseline data from the Osaka maternal and child health study. Annals of Epidemiology. 2005;15(5):358–364. doi: 10.1016/j.annepidem.2004.12.005. [DOI] [PubMed] [Google Scholar]
- 50.Tonetti MS. Cigarette smoking and periodontal diseases: etiology and management of disease. Annals of periodontology / the American Academy of Periodontology. 1998;3(1):88–101. doi: 10.1902/annals.1998.3.1.88. [DOI] [PubMed] [Google Scholar]
- 51.Jansson L, Lavstedt S, Frithiof L, Theobald H. Relationship between oral health and mortality in Cardiovascular Diseases. Journal of Clinical Periodontology. 2001;28(8):762–768. doi: 10.1034/j.1600-051x.2001.280807.x. [DOI] [PubMed] [Google Scholar]
- 52.Jansson L, Lavstedt S, Zimmerman M. Prediction of marginal bone loss and tooth loss - A prospective study over 20 years. Journal of Clinical Periodontology. 2002;29(8):672–678. doi: 10.1034/j.1600-051x.2002.290802.x. [DOI] [PubMed] [Google Scholar]
- 53.Axelsson P, Nystrom B, Lindhe J. The long-term effect of a plaque control program on tooth mortality, caries and periodontal disease in adults: Results after 30 years of maintenance. Journal of Clinical Periodontology. 2004;31(9):749–757. doi: 10.1111/j.1600-051X.2004.00563.x. [DOI] [PubMed] [Google Scholar]
- 54.Krall EA, Dawson-Hughes B, Hannan MT, Wilson PWF, Kiel DP. Postmenopausal estrogen replacement and tooth retention. American Journal of Medicine. 1997;102(6):536–542. doi: 10.1016/s0002-9343(97)00045-4. [DOI] [PubMed] [Google Scholar]
- 55.Susin C, Haas AN, Opermann RV, Albandar JM. Tooth loss in a young population from south Brazil. Journal of Public Health Dentistry. 2006;66(2):110–114. doi: 10.1111/j.1752-7325.2006.tb02565.x. [DOI] [PubMed] [Google Scholar]
- 56.Susin C, Oppermann RV, Haugejorden O, Albandar JM. Tooth loss and associated risk indicators in an adult urban population from south Brazil. Acta Odontologica Scandinavica. 2005;63(2):85–93. doi: 10.1080/00016350510019694. [DOI] [PubMed] [Google Scholar]
- 57.Musacchio E, Perissinotto E, Binotto P, Sartori L, Silva-Netto F, Zambon S, et al. Tooth loss in the elderly and its association with nutritional status, socio-economic and lifestyle factors. Acta Odontologica Scandinavica. 2007;65(2):78–86. doi: 10.1080/00016350601058069. [DOI] [PubMed] [Google Scholar]
- 58.Morita I, Nakagaki H, Toyama A, Hayashi M, Shimozato M, Watanabe T, et al. Behavioral factors to include in guidelines for lifelong oral healthiness: An observational study in Japanese adults. BMC Oral Health. 2006:6. doi: 10.1186/1472-6831-6-15. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 59.Taguchi A, Sanada M, Suei Y, Ohtsuka M, Nakamoto T, Lee K, et al. Effect of estrogen use on tooth retention, oral bone height, and oral bone porosity in Japanese postmenopausal women. Menopause. 2004;11(5):556–562. doi: 10.1097/01.gme.0000113845.74462.bf. [DOI] [PubMed] [Google Scholar]
- 60.Christensen LB, Petersen PE, Steding-Jessen M. Consumption of dental services among adults in Denmark 1994–2003. Eur.J.Oral Sci. 2007;115(3):174–179. doi: 10.1111/j.1600-0722.2007.00453.x. [DOI] [PubMed] [Google Scholar]
- 61.Xu X, Patel DA, Vahratian A, Ransom SB. Insurance coverage and health care use among near-elderly women. Womens Health Issues. 2006;16(3):139–148. doi: 10.1016/j.whi.2006.02.005. [DOI] [PubMed] [Google Scholar]
- 62.Chestnutt IG, Binnie VI, Taylor MM. Reasons for tooth extraction in Scotland. J.Dent. 2000;28(4):295–297. doi: 10.1016/s0300-5712(99)00069-x. [DOI] [PubMed] [Google Scholar]
- 63.McCaul LK. The reasons for extraction of permanent teeth in Scotland: A 15-year follow-up study. Br.Dent.J. 2001;190(12):658–662. doi: 10.1038/sj.bdj.4801068. [DOI] [PubMed] [Google Scholar]
- 64.Murray H, Clarke M, Locker D, Kay EJ. Reasons for tooth extractions in dental practices in Ontario, Canada according to tooth type. Int.Dent.J. 1997;47(1):3–8. doi: 10.1111/j.1875-595x.1997.tb00670.x. [DOI] [PubMed] [Google Scholar]
- 65.Morita M, Kimura T, Kanegae M, Ishikawa A, Watanabe T. Reasons for extraction of permanent teeth in Japan. Community Dent.Oral Epidemiol. 1994;22(5 Pt 1):303–306. doi: 10.1111/j.1600-0528.1994.tb02056.x. [DOI] [PubMed] [Google Scholar]
- 66.Murray H, Locker D, Kay EJ. Patterns of and reasons for tooth extractions in general dental practice in Ontario, Canada. Community Dent.Oral Epidemiol. 1996;24(3):196–200. doi: 10.1111/j.1600-0528.1996.tb00841.x. [DOI] [PubMed] [Google Scholar]
- 67.Ong G, Yeo J-, Bhole S. A survey of reasons for extraction of permanent teeth in Singapore. Community Dent.Oral Epidemiol. 1996;24(2):124–127. doi: 10.1111/j.1600-0528.1996.tb00828.x. [DOI] [PubMed] [Google Scholar]
- 68.Klock KS. Patients' perceptions of the decision-making process leading to extraction of permanent teeth in Norway. Community Dent.Oral Epidemiol. 1995;23(3):165–169. doi: 10.1111/j.1600-0528.1995.tb00222.x. [DOI] [PubMed] [Google Scholar]