Edentulous disparities among geriatric population according to the sexual difference in South Korea: a nationwide population-based study

The proportion aged 60 years or older in the world's population is expected to double by 2050. In general, they have many complex diseases and poor oral health status. Oral health is one of the important health indicators of elderly people and it is affected by diverse factors, such as socioeconomic status. In this study, sexual difference was considered as an associated factor that is closely related to edentulism. The sexual difference might be more influential within the geriatric population because of lower economic and educational backgrounds at this stage. Edentulism was significantly higher among elderly females than males when combined with the education level. The lower the level of education, the higher the prevalence of edentulism as much as 24 ~ 28 times, especially in females (P = 0.002). These findings suggest a more complex relationship between oral health, socioeconomic status, and sexual difference.


Prevalence of edentulism in elderly individuals.
The proportion of the geriatric population without teeth was 9.5%. This increased with age, with 18.1% of participants aged 80 years being edentulous; the female showed a higher rate of edentulism than the male which was not statistically significant ( Table 2). While concerning educational level, the prevalence of edentulism was 11.1% in people who finished elementary school, which was twice as high as that among high school graduates and those with higher education. The gap in the prevalence of edentulism by education level was much higher and statistically different in females compared to males, who did not show statistical significance. The higher the income, the lower the prevalence of edentulism, especially in females; the gap between the first and fifth quartiles was almost double compared to males. The participants who lived alone had a slightly higher prevalence of edentulism than participants who lived with someone else. There was no statistical difference among participants with hypertension in general. However, females with this metabolic disease had a higher prevalence of edentulism. The participants who did not have a history of drinking monthly had a higher prevalence of edentulism, but there was no difference according to the sexual difference. The percent of people with no teeth was 11.4% in smokers, which was higher than that in non-smokers, as much as twice in both males and females. In the case of physical activity, the male who were not engaged in physical activity had a higher rate of edentulism. Participants, both male and female, who brushed lesser than twice a day had a five times higher likelihood of edentulism than those who brushed more. The rate of edentulism was almost five times higher in participants who did not receive oral examinations once a year; the gap was bigger in males, as ten times more likely to be edentulous.
Logistic regressions by risk factors associated with edentulism in elderly individuals. The logistic regression analysis for the factors affecting the prevalence of edentulism among elderly individuals is presented in Table 3. Age, educational level, income, living alone, and lifetime smoking were significantly associated with edentulism among elderly individuals. Participants who brushed fewer than twice a day were five times more likely to be edentulous than participants who did, and participants who did not receive yearly oral examinations were six times more likely to be edentulous than participants who did. After adjustment for all risk factors, the geriatric population who did not receive yearly oral examinations showed a lower odds ratio: almost four times that before. The coefficient of determination of NagelKerke R2 in this model was 0.185. There were different patterns according to the sexual difference in the prevalence of edentulism and educational level in Tables 1 and 2. In the third model considering the interaction effect of educational level to edentulism depending on the sexual difference, the odds ratio of edentulism decreased for females with the opposite results of those from the second model. Compared to the group of elderly female participants who graduated high school, people who finished middle school showed 29 times higher risk of being edentulous, and people who graduated elementary school were at a 24-fold higher risk of being edentulous. The lower the level of education, the higher the prevalence of edentulism, especially in females. The coefficient of determination of NagelKerke R2 in the final model was 0.198 and the p-value was 0.006 for the test of model effects of interaction between the sexual difference and educational level.

Discussion
The proportion of the world's population aged 60 years or older is expected to double by 2050. As a result, population aging is recognized as an important problem worldwide 31 . Oral health is an essential health indicator among elderly people 9,10 . A gap is reported in the oral health of geriatrics for various factors including socioeconomic Table 1. Characteristics of the study population of elderly between 2016 to 2018. *Medium-intensity physical activity for more than 2 h and 30 min, high-intensity physical activity for more than 1 h and 15 min, or mixed intensity per week (1 min of high-intensity physical activity is equal for 2 min of medium-intensity). www.nature.com/scientificreports/ and behavioral factors. In this study, sexual difference was considered the main factor that affected the prevalence of edentulism among those aged 65 years or above. There was no difference in the prevalence of edentulism according to the sexual difference until the interaction between the sexual difference and education level was considered, which was significant. Edentulism among elderly females was significantly higher than that among males when categorized by education level. The prevalence of edentulism among middle school or elementary school graduates and participants with lower education levels was 24-28 times that of high school graduates or participants with higher education in an elderly female. Table 2. Prevalence of edentulism according to characteristics of the study population of elderly between 2016 to 2018. *Medium-intensity physical activity for more than 2 h and 30 min, high-intensity physical activity for more than 1 h and 15 min, or mixed intensity per week (1 min of high-intensity physical activity is equal for 2 min of medium-intensity). www.nature.com/scientificreports/ The prevalence of edentulism among elderly individuals in South Korea was 9.5%, which was lower than the prevalence of 12.9% observed in the United States 32 , but higher than the prevalence of 8.1% observed in Japan 33 . The proportion of elderly individuals is expected to reach 46.5% in 2067 34 , of which the proportion of female elderly individuals is expected to be about 55%, higher than that of male elderly individuals in Korea. Sexual difference may be an important factor associated with edentulism among the geriatric population, especially for economic and educational backgrounds. Some studies have reported that the lower the education or income, the higher the prevalence of edentulism [35][36][37][38] . This population is also reported to have lower subjective oral health, which might affect care for oral health 39 . Most of them showed a lower tendency to visit dental clinics 40 , which might have worsened oral health conditions. There was a significant difference in edentulism by socioeconomic level in an elderly female. This is the reason that the interaction between sex differences and socioeconomic factors is assumed to exist. There might be a more complex relationship between oral health, socioeconomic status, and sexual difference. Among the geriatric Korean population, a female had a significantly less chance than a male of being educated. They also lived longer than males and many of them lived alone 41 . Elderly females were found to be more sensitive to socioeconomic factors as well. The gap in economic poverty appeared in poorer elderly www.nature.com/scientificreports/ females, which worsened oral health 42 . Paola et al. reported that elderly females did more preventive oral care than males, but complex mechanisms working in oral health make necessary careful management 43 . The oral health gap should be considered not just by the sexual difference itself but also by the contextual mechanism around it. Regular tooth brushing twice a day and yearly oral examination was found to have effects on edentulism. Healthy behaviors were reported as closely related to oral health 44 . The prevalence of edentulism was higher in participants with a history of smoking, brushing fewer than twice a day, and not taking yearly oral examinations 45,46 . These oral health behaviors showed different patterns depending on the sexual difference, and females are known to visit the dentist more frequently, perform better oral care 47 , and have better knowledge, attitude, and behavior about oral health than males 48,49 . The elderly female showed about a 10% higher rate of brushing their teeth twice or more a day than males in this study. Likewise, the sexual difference in elderly individuals had significant associations with different oral health behavior factors, such as physical activity for males and hypertension for females. The type of residence also showed a significant association. Elderly individuals living alone were at an almost two times higher risk of edentulism. Kim et al. reported that elderly people who lived alone had a higher possibility of needing dentures and experiencing difficulty in mastication than geriatric people living with their families 50 . Elderly males living without families showed a higher level of edentulism. Elderly individuals living alone might have fewer opportunities to acquire and exchange information about oral health and more difficulties in practicing healthy oral behavior 51 .
There are several limitations of this study. First, social relationships and support might influence physical, mental, and oral health and behavior, especially in geriatric people 52,53 . The data used in this study did not include social capital variables such as the number of friends and meetings. It is necessary to understand the social relationship between elderly individuals with oral health in depth in the future. Second, it is difficult to identify the causal relationship because the data was a cross-sectional study. The present education or income levels were surveyed but the past socioeconomic status or oral health behavior of the participant could not be included in this survey. The causal relationship from previous conditions would be important in the study of the geriatric population. Third, this study focused on edentulism disparities among the geriatric population according to sexual differences and educational levels. There might have been other confounding factors with sex, such as hypertension and aerobic physical activity. When they were added to interaction terms analysis and few minor numeric changes of coefficients in the model existed and the p values lay on the borderline of statistical significance. More to the point, it was not clear whether sex was a confounding factor with these variables in the contextual framework. Therefore, only educational level was included as an interaction term in the final model. Future studies must design a longitudinal data set such as a cohort framework to identify the influential relationship after identifying the related factors including diverse confounders. Even with these limitations, this study established the fundamentally important health characteristics of an aging society according to the sexual difference with oral health status using nationally representative data.

Conclusion
This study revealed the association between sex differences and socioeconomic factors related to the oral health of elderly individuals. In edentulism in elderly individuals, there was no difference according to the sexual difference, initially. After considering the interaction between sex difference and education, an elderly female had a lower possibility of being edentulous than a male. The prevalence of edentulism in the elderly female who graduated middle school or elementary school was 24 ~ 28 times that of high school graduates. The sexual difference in the geriatric population had significant associations with different oral health behavior factors, such as physical activity for males and hypertension for females. In conclusion, in the plan for oral health improvement in elderly individuals, the sexual difference must be identified, such as socioeconomic factors for females and health behavior factors for males.

Methods
Study subjects. This study analyzed oral health conditions by the sexual difference in the geriatric population aged 65 years or older using the Korean National Health and Nutrition Survey (KNHANES VII, 2016-2018) data. The sampling frame was layered based on the size of the area (cities, provinces, and districts) and housing types (general housing, apartments). The ratio of residential area and educational background of household owners was used as the intrinsic stratification criteria. Finally, 576 districts were surveyed over 3 years, with 10,611 households participating in the study. A cohort of 24,269 participants was recruited with a response rate of 76.6%. Among them, 3426 people were elderly, which was 21.0% of all subjects.

Study variable.
The general characteristics of the study sample were sex, age, education or income level, household type, region, comorbidities such as hypertension or diabetes, health-related behavior such as drinking, smoking, aerobic physical activity, brushing teeth, or visiting a dental clinic for oral examination per year. The independent variables are classified as shown in Table 1. The type of household was categorized by the number of family members-one as living alone and two or more as a family living together. A person involved in as much as medium-intensity physical activity for 2 h and 30 min, high-intensity physical activity for 1 h and 15 min, or medium and high-intensity physical activity per week (1-min-high intensity for 2 min) was defined as properly engaging in physical activity. Edentulism was defined as a person who had none of the natural teeth including third molars, calculated based on the result of the oral examination. A complex sample logistic regression could not check the collinearity of the variables in the options of the statistical package. Therefore, we applied the collinearity option in the general logistic regression with study variables. Variance inflation factors (VIF) were used to assess multicollinearity among the socioeconomic variables. VIF > 10 indicated the presence of multicollinearity 54 . However, no indicators of multicollinearity were identified as all the VIFs were under 5. www.nature.com/scientificreports/ Statistical analysis. A complex sample analysis was used because the KNHANES was a two-stage stratified cluster sampling. It was conducted by generating an integrated weight based on the prepared analysis plan file. The plan file adapted analysis weight with "oral examination weight", design of the strata with "variance estimation" and "group aged 65 years or above, and cluster as primary sampling unit with the district. The KNHANES recommended analyzing the data by reflecting this sample design (stratum, cluster, weight). Especially if only a part of the data in a complex sample is analyzed, the standard error of the estimate might be biased due to missing data information. The group variable as the elderly "aged 65 years or above" was created and adapted for a subgroup analysis 55,56 . Complex sample frequency and chi-square tests were conducted to find out the difference in the prevalence of edentulism according to socioeconomic factors, chronic diseases, and health-related behaviors. A complex sample logistical regressions were performed with an unadjusted model with univariate variables, a fully adjusted model including all variables (Model 1), and an advanced model with interaction by the sexual difference and education (Model 2). An interaction analysis was conducted to confirm the relationship between education level and edentulism depending on the sexual difference. After including the interaction terms of sex and education variables in the complex logistic regression model, the significance of each interaction term confirmed whether edentulism appeared differently by education level according to the sexual difference. All analyses used SPSS (Statistical Packages for Social Science 26.0. SPSS Inc., USA) and statistical significance was set to α = 0.05.
Ethical approval and informed consent. This study used the dataset obtained from the KNHANES VII, 2016 to 2018. All KNHANES were conducted with participants' informed consent after approval by the Research Ethics Review Committee of the Korea Disease Control and Prevention Agency (KDCA) (IRB No. 2018-03-P-A for the KNHANES VII). This analytical study was approved again by the institutional review board (IRB) of Kyung Hee University (IRB No. KHSIRB-21-337(EA)) as exemption of the review because this retrospective analysis included the dataset of national surveillance and did not contain personally identifiable information. All methods were carried out following the KNHANES analytic guidelines and regulations.

Data availability
The data that supports the findings of this study are available from the Korean Disease Control and Prevention Agency (KDCA), but restrictions apply to the availability of data, which was used with permission for the current study and therefore not publicly available. Data is however available upon reasonable request and with permission of KDCA.