Abstract
We used National Performance Evaluation Survey data to estimate the prevalence and associated factors of edentulism among noninstitutionalized adults aged 35 years and older in Mexico. Statistically, the variables positively associated with edentulism were older age (odds ratio [OR]=1.08) and female gender (OR=1.79). Nonsmoking status (OR=0.70) and having a higher wealth index score (OR=0.72) were negatively associated. This information constitutes the first large-scale evaluation in Mexico for one of the World Health Organization’s priority oral health problems.
The World Health Organization (WHO) considers edentulism (complete absence of natural teeth) a poor public health outcome that substantially affects oral and general health status, as well as quality of life. It is an important but often-overlooked public health issue, especially for the elderly.1,2 Oral health status plays an important role in the nutrition of adults and older people.3 Tooth loss has been associated with changes in food taste and food preferences and nutritional deficiency.3,4 Little oral health epidemiological information is available for the Mexican populations: specifically, no information is available on prevalence of edentulism in adults and the elderly. The objective of this study was to determine the prevalence and associated factors of edentulism among a large-scale sample of noninstitutionalized people aged 35 years and older in Mexico.
The National Performance Evaluation Survey 2002–2003 (a countrywide evaluation of health services in Mexico) was part of the technical collaboration between the Ministry of Health of Mexico and WHO, which used the survey instrument and sampling strategies developed by WHO for the World Health Survey. The National Institute of Public Health and the Ministry of Health of Mexico General Directorate of Performance Evaluation undertook the survey. The National Performance Evaluation Survey provides policymakers with reliable and internationally comparable baseline information on a variety of health indicators, including general population health measures and the effectiveness of health systems.
The National Performance Evaluation Survey was conducted between November 2002 and April 2003 and collected information from 38 746 households in 32 states in Mexico, with a mean of 1250 households per state, although only 20 Mexican states collected dental data. The sampling frame was probabilistic, stratified, used a multistage cluster design and was calculated to provide representative information at both the state level and by urban and rural areas.5
The survey comprised a face-to-face interview divided into household and individual questionnaires. (Both questionnaires and a discussion of general methodological issues are available on WHO’s Web page: http://www3.who.int/whs/P/instrumentandrel8293.html.) The household survey includes general topics, such as physical characteristics of the household and ownership of consumer goods, which were combined in the polychoric principal components analyses to construct a wealth index.6
The dependent variable in our analyses was self-reported edentulism (individuals who stated that they had lost all of their natural teeth). Weighted analyses were conducted in Stata software 8.2 (StataCorp LP, College Station, Tex) with the module svy Bivariate and multivariate logistic regression analyses were undertaken. To avoid multicollinearity in the variables included in the final model, the variance inflation factor test was performed. Finally, the Box-Tidwell test was performed to evaluate whether continuous variables in the final model were related to the logit of the dependent variable lineally. The multivariate logistic regression model incorporated all variables available (socio-demographic and socioeconomic) that were thought to be related to edentulism and associated at the bivariate analysis level with P < .25 to control for possible confusion.
National Performance Evaluation Survey data on dental conditions were available only for subjects aged 18 years and older, representing a total of 24 159 households. In the current round of analyses, only survey respondents who were aged older than 35 years (n = 13 966) were included, representing 29 853 607 adult country inhabitants. All results presented are weighted. The mean age was 50.95 ± 12.98 SD (ranging from 35 to 98 years old), with 57.9% women (n = 17 296 022) and 42.1% men (n = 12 557 585). Overall prevalence of edentulism in this adult population was 10.2%; when the WHO’s “age index” for prevalence of edentulism was used, the proportion for the elderly ( > 65 years old) was 30.6% and for the younger adults (35–44 years old) was 2.4%. Figure 1 shows the geographic distribution of edentulism in Mexico, with prevalence ranging from 5.0% in Tlaxcala (central Mexico) to 16.7% in Michoacán (western Mexico). Table 1 presents descriptive results for the study population.
In the bivariate analyses (Table 1), we observed that age, gender, smoking, diabetes, health insurance status, marital status, occupation, and wealth index were positively associated with edentulism, whereas preventive attitude and higher level of schooling were negatively associated with edentulism. The multivariate logistic regression model (Table 2) adjusted for schooling and Indian ethnic status showed that for each year of age increase, the odds of being edentulous increased 8% (95% confidence interval [CI]=1.07, 1.09), with women being 1.79 times more likely (95% CI=1.50, 2.13) to be edentulous, but nonsmokers (odds ratio [OR]=0.70; 95% CI=0.54, 0.90) and better-off people (higher wealth index) (OR=0.72; 95% CI=0.53, 0.98) being less likely to lose all of their natural teeth. No plausible interaction terms were significant. There was no effect when differentiating urban and rural areas.
The importance of preserving teeth in good condition not only may be restricted to quality of life in the elderly1,2 but also may be an indicator of general health because edentulism is associated with systemic, chronic diseases.7–9 Our study provided large-scale data for the first time in Mexico on one of the priority oral health indicators proposed by WHO. Most of the studies performed in other countries have included only people aged 60 years and older.
The prevalence observed in our study for the elderly (30.6%) can be contrasted with reports from other countries; both similarities and disparities occur. Our findings fall within the range derived from a telephone survey in the United States10: in 1997, the prevalence ranged from 13.9% (in Hawaii) to 47.9% (in West Virginia), and in 2002, it ranged from 13% (in Hawaii and California) to 42% (in Kentucky).11 The Mexican prevalence is also similar to that observed in Australians (also obtained through self-report in a telephone survey)1 but much higher than the 7.0% reported in Japanese adults.12 As in the other studies, our results were based on self-reports that, although not verified by dental personnel, have been shown to have high correlation with clinical examinations in these age groups.13 It can be safely assumed that the lay public knows when no teeth are present.
As previously reported, age is strongly associated with edentulism1,14; this finding is not surprising, because the cumulative effects of dental caries and periodontal diseases,15,16 as well as treatment decisions associated with these 2 main reasons for tooth loss, increase with age. Differential treatment choices across the socioeconomic continuum, such as endodontic treatment versus tooth extraction, also may play a role in the overall tooth loss phenomenon. As reported by other authors, women are more likely to be edentulous than are men.1 Because cigarette smoking is a risk factor for adult periodontitis and tooth loss,10 the higher prevalence of edentulism among smokers may be directly related to the adverse effects of smoking on periodontal health. This association also has been observed in other studies.10,17 The periodontal status of former smokers ranks between that of never smokers and current smokers, suggesting that smoking causes changes in the periodontium but that the deterioration may not continue after tobacco use cessation.18
As in other countries, we found that edentulism is closely associated with socioeconomic variables. Epidemiological studies show that persons of low social class or low income and individuals with few years of schooling are more likely to be edentulous than are persons of higher social class, income, and educational attainment.14
Our study had certain limitations that emphasize a cautious interpretation of results. A cross-sectional study measures associations at 1 point in time, introducing the problem of temporal ambiguity and the inability to establish causal relationships. Furthermore, this study cannot be considered truly nationally representative because 12 of 32 states (mostly northern states) were excluded because of survey logistics. Contrast of results with other countries is limited by necessity because studies in comparable settings in emerging economies are not available.
This baseline information constitutes the first national evaluation in Mexico for one of the priority WHO oral health problems. The observed prevalence of edentulism was moderate to low in the adult and elderly groups (compared with other reports in non-Mexican populations). This important oral health condition was strongly linked to birth cohorts and possibly associated with increasing age. This study of noninstitutionalized adults found a significant relation between sociodemographic variables (e.g., age and gender) and edentulism, with socioeconomic status (as measured by the wealth index) playing a role in the prevalence of edentulism. Nonsmoking status had an effect on edentulism, separate from socioeconomic and sociodemographic factors. Future studies should examine risk factors for tooth loss to effect changes in health systems and increase tooth longevity.
Note. OR = odds ratio; CI = confidence interval. * P < .05; **P < .01; ***P < .001. Note. OR = odds ratio; CI = confidence interval. Box–Tidwell for age: P = .075. Hosmer–Lemeshow goodness-of-fit: χ28 = 7.06; P = .5304. * P < .05; **P < .01; ***P < .001. FIGURE 1— Geographic distribution of edentulism in 20 of 32 states of Mexico, as a weighted percentage of the population.
n % Weighted OR (95% CI) Age 13 966 100 1.08 (1.07, 1.09)*** Gender Men 5975 42.1 1.00 Women 7991 57.9 1.48 (1.27, 1.73)*** Smoking Never 10 872 77.3 1.00 Sometimes 1964 15.4 1.09 (0.82, 1.46) Daily 1–5 times 614 4.2 1.02 (0.71, 1.46) Daily > 5 times 516 3.1 1.53 (1.04, 2.23)* Years smoking daily < 10 251 24.1 1.00 11–20 334 29.3 1.96 (0.79, 4.87) 21–30 247 24.2 2.52 (1.05, 6.06)* > 30 298 22.4 4.34 (2.05, 9.19)*** Diabetes No 12 782 91.6 1.00 Yes 1184 8.4 1.82 (1.39, 2.38)*** Preventive attitude No 8944 61.8 1.00 Yes 5022 38.2 0.59 (0.48, 0.72)*** Health insurance Without insurance 8426 61.2 1.00 With insurance 5510 38.8 1.31 (1.08, 1.59)** Schooling Elementary incomplete 3151 21.1 1.00 Elementary complete 6834 48.9 0.47 (0.38, 0.59)*** Middle complete 1911 14.7 0.31 (0.21, 0.45)*** High school 1176 9.0 0.24 (0.14, 0.41)*** College 894 6.2 0.32 (0.22, 0.49)*** Marital status Married 8867 67.2 1.00 Single 997 6.0 1.15 (0.85, 1.56) Divorced 1011 6.2 1.49 (1.06, 2.08)* Widowed 2023 11.2 3.88 (3.05, 4.95)*** Cohabitant 1068 9.3 0.78 (0.48, 1.25) Occupation Government employee 1243 7.7 1.00 Nongovernment employee 1349 10.3 1.13 (0.73, 1.74) Self-employed 4079 29.3 1.85 (1.22, 2.79)** Does not work/voluntary worker 7295 52.7 3.87 (2.64, 5.68)*** Residence Rural 4128 26.6 1.00 Urban 9831 73.3 1.06 (0.81, 1.40) Indian ethnic status No 13 054 92.5 1.00 Yes 801 7.5 0.75 (0.51, 1.12) Wealth index 1 (poorest) 3734 28.4 1.22 (0.93, 1.59) 2 3509 24.1 1.41 (1.12, 1.78)** 3 3440 24.0 0.92 (0.64, 1.31) 4 (highest) 3036 23.5 1.00 Adjusted OR (95% CI) Years of age 1.08 (1.07, 1.09)*** Gender Men 1.00 Women 1.79 (1.50, 2.13)*** Smoking Sometimes, daily 1.00 No, never 0.70 (0.54, 0.90)** Wealth index First, second, or third quartile 1.00 Fourth quartile 0.72 (0.53, 0.98)* 
