The Mediating Effects of Nutritional Status on the Relationship between Number of Residual Teeth and Cognitive Function among Older Adults: A Cross-Sectional Multicenter Study

The underlying mechanisms of the relationship between the number of teeth and cognition is still unclear. We aimed to construct a mediation model between the number of residual teeth and cognitive function, using nutritional status as a mediating factor. This study was completed using the West China Health and Aging Trend cohort. A total of 6634 multi-ethnic older adults, aged 50 years or older, were included. This study measured cognitive function using the Short-Portable Mental Status Questionnaire, and nutritional status was assessed using the Mini Nutritional Assessment-Short Form. The mediation analysis examined the potential mediating role of nutritional status. The pathway analysis was supplemented and validated using the structural equation modelling framework. Multiple linear regression demonstrated that a higher number of residual teeth was correlated with enhanced cognitive function (β = −0.15; 95% CI: −0.19 to −0.111). The mediation model, from the number of residual teeth to cognitive impairment, was partially mediated by nutritional status (β = −0.0608; 95% CI: −0.0762 to −0.0461). The proportion of the mediating effect, expressed as a percentage, was 40.66%. Furthermore, the estimated coefficients for the number of residual teeth and nutritional status varied across ethnic groups. This study indicated that enhancing the nutrition of older adults could reduce the adverse effects of the number of residual teeth on cognitive function among older adults.


Introduction
In aging populations, an increasing number of older adults are living with cognitive impairments. The pooled prevalence of cognitive impairment among Chinese older adults was 15.4% in 2021 [1]. Among older adults living in rural areas, the prevalence of cognitive impairment was 26.48% [2], with a higher prevalence reported in western China [3]. Moreover, cognitive impairment imposes an enormous disease burden [4]. Identifying the risk factors of cognitive impairment and early interventions can improve the quality of life of older adults and their caregivers.
An increasing number of studies have suggested that tooth loss in later life could increase the risk of cognitive impairment in older adults [5][6][7]. Studies have indicated that tooth loss could be an early indicator of accelerated aging [8,9]. A meta-analysis study [7] reported that older adults with more teeth had a decreased risk of dementia by approximately 50%. Another meta-analysis study [6] which examined longitudinal studies in Asian countries, also indicated that tooth loss is a risk factor for dementia.
However, the underlying mechanisms of the relationship between the number of residual teeth and cognitive function in older adults is still unclear. Studies have suggested that nutritional status plays an explanatory role in this relationship. More specifically, impaired dentition status is associated with poor dietary intake and malnutrition [10]. Malnutrition risks were associated with worse cognitive function [11]. Some studies suggested that numerous nutrients affect cognitive function [12,13]. Currently, only a limited-scale study [14] has examined nutrition as a mediating factor when investigating the relationship between the number of teeth, denture wearing, and cognition. Further verifications of this relationship are needed with studies that use larger samples. Consequently, we hypothesize that the association between the number of residual teeth and cognitive function is possibly mediated by nutritional pathways among 6634 multi-ethnic older adults from western China.

Study Protocol and Population
The data in this cross-sectional study were obtained from a West China Health and Aging Trend (WCHAT) multicenter cohort study which focused on aging and health status in older adults. The cohort collected baseline data from July to December 2018, and the study was conducted in four provinces in western China, including Yunnan, Guizhou, Sichuan, and Xinjiang. Considering the topographical characteristics and scattered living patterns in different ethnic regions in western China, the county's medical institution was selected as the primary venue for conducting the study. Ethics approval for this research was granted by the West China Hospital of Sichuan University (reference: 2017-445) on 31 January 2018. The register ID is 1800018895 in the Chinese Clinical Trial Registry. Informed consent was obtained from all respondents. Details of the protocol have been reported elsewhere [15]. The researchers, who majored in geriatrics from Sichuan University, completed comprehensive questionnaires to record older adults' demographic characteristics and health-related data. All researchers were trained for 2 days in how to collect questionnaire data via face to face and one-on-one personal interviews. The interrater reliability achieved 90% through practical training on questionnaires. The questionnaire underwent cultural adaptations and preliminary testing to ensure its relevance to the specific cultural, lifestyle, and linguistic characteristics of western China. To be eligible to participate, the following inclusion criteria were established: (1) willingness to participate and to have provided signed informed consent, and (2) to be 50 years or older. The exclusion criteria were as follows: (1) absence of cognitive function assessment, and (2) lack of covariate data. From an initial sample of 7536 participants, 902 were excluded after applying the inclusion and exclusion criteria, and a total of 6634 participants were included in the study (Figure 1).

Data Collection
The number of residual teeth was calculated via a specialized oral examination. Cognitive function was assessed using the Chinese version of the 10-item Short Portable Mental Status Questionnaire (SPMSQ), which evaluates memory, orientation to time and place, current event information, and calculation. Higher error rates indicate more severe cognitive impairment. Pfeiffer's criteria [16] classify cognitive function as follows: normal cognition (0-2), mild cognitive impairment (3)(4), and moderate and severe cognitive impairment (5 or higher). This scoring adjusts for the biasing effects of educational background. The nutrition status was estimated by the Mini Nutritional Assessment-Short Form (MNA-SF). The MNA-SF is a reliable tool for distinguishing older adults who are malnourished or at risk of malnutrition. MNA-SF scores are categorized as follows: normal nutrition (12)(13)(14), at risk of malnutrition (8)(9)(10)(11), and malnutrition (0-7) [17]. The flowchart of study participants; 6634 participants were analyzed in this study. Among those who were excluded from the study, 97 participants were less than 50 years old, 619 participants presented without cognition data, and 186 participants presented without covariates data.

Data Collection
The number of residual teeth was calculated via a specialized oral examination. Cognitive function was assessed using the Chinese version of the 10-item Short Portable Mental Status Questionnaire (SPMSQ), which evaluates memory, orientation to time and place, current event information, and calculation. Higher error rates indicate more severe cognitive impairment. Pfeiffer's criteria [16] classify cognitive function as follows: normal cognition (0-2), mild cognitive impairment (3)(4), and moderate and severe cognitive impairment (5 or higher). This scoring adjusts for the biasing effects of educational background. The nutrition status was estimated by the Mini Nutritional Assessment-Short Form (MNA-SF). The MNA-SF is a reliable tool for distinguishing older adults who are malnourished or at risk of malnutrition. MNA-SF scores are categorized as follows: normal nutrition (12)(13)(14), at risk of malnutrition (8)(9)(10)(11), and malnutrition (0-7) [17].
Additional sociodemographic covariates were age, sex, educational level, occupation, spouse status, longevity of family members, and ethnic groups (Han, Tibetan, Qiang, Yi, Uyghur, and others). "Others" refers to a few ethnic groups combined with the Zhuang, Manchu, Hui, Mongolian, and Tujia groups; they are classified as such due to the number of participants being less than 200. Basic lifestyle information was also collected: type of drinking water and history of smoking, alcohol use, and drinking tea.

Statistical Analysis
Descriptive data were summarized using proportions for categorical data and means with standard deviation (SD) for continuous data. Differences between the categories of cognitive impairment and the variables studied were analyzed using one-way ANOVA for continuous variables and the chi-squared test for categorical variables. Multiple comparisons were conducted using independent sample t-tests and chi-square tests. The Bonferroni correction was used to correct p-values for multiple comparisons.
The association between the number of residual teeth, MNA-SF scores, and cognitive scores was analyzed using linear regression, which used three separate models. We entered the number of residual teeth as a predictor variable, and cognitive scores as outcome variables, using two separate models: Model 1 was adjusted in accordance with age, sex, ethnic group, educational level, occupation, spouse status, longevity of family members, denture usage, and lifestyles (type of drinking water, history of smoking, alcohol use, and drinking tea); Model 2 was adjusted using Model 1 and the MNA-SF score. We entered Figure 1. The flowchart of study participants; 6634 participants were analyzed in this study. Among those who were excluded from the study, 97 participants were less than 50 years old, 619 participants presented without cognition data, and 186 participants presented without covariates data.
Additional sociodemographic covariates were age, sex, educational level, occupation, spouse status, longevity of family members, and ethnic groups (Han, Tibetan, Qiang, Yi, Uyghur, and others). "Others" refers to a few ethnic groups combined with the Zhuang, Manchu, Hui, Mongolian, and Tujia groups; they are classified as such due to the number of participants being less than 200. Basic lifestyle information was also collected: type of drinking water and history of smoking, alcohol use, and drinking tea.

Statistical Analysis
Descriptive data were summarized using proportions for categorical data and means with standard deviation (SD) for continuous data. Differences between the categories of cognitive impairment and the variables studied were analyzed using one-way ANOVA for continuous variables and the chi-squared test for categorical variables. Multiple comparisons were conducted using independent sample t-tests and chi-square tests. The Bonferroni correction was used to correct p-values for multiple comparisons.
The association between the number of residual teeth, MNA-SF scores, and cognitive scores was analyzed using linear regression, which used three separate models. We entered the number of residual teeth as a predictor variable, and cognitive scores as outcome variables, using two separate models: Model 1 was adjusted in accordance with age, sex, ethnic group, educational level, occupation, spouse status, longevity of family members, denture usage, and lifestyles (type of drinking water, history of smoking, alcohol use, and drinking tea); Model 2 was adjusted using Model 1 and the MNA-SF score. We entered the number of residual teeth as a predictor variable and MNA-SF scores as outcome variables in accordance with Model 3, which was adjusted using the same confounding factors as Model 1. R 4.1.3 was used to perform all statistical analyses in this study. We used the mediation R package to evaluate the influence of the MNA-SF score on the relationship between the number of residual teeth and cognitive scores. The path analysis was conducted (lavaan package) using the structural equation model (SEM) framework. The results were statistically significant at p < 0.05, and all p values were two-sided.

Results
The descriptive analyses of demographic characteristics and lifestyles were stratified in accordance with level of cognitive impairment, and they are summarized in Table 1. The participants had an average age of 62.39 (8.26) years, they were predominantly female (62.5%), and were ethnically Han (36.1%). The average number of residual teeth was 21.20 (9.52). The percentage of participants wearing a denture was 38%. The mean MNA-SF score was 12.58 (1.58 Table S1 showed the results of the multiple comparisons among the three groups. The multiple comparisons revealed significant differences between participants with mild cognitive impairments and those with normal cognitive function. b The significant differences between participants with moderate to severe cognitive impairments and those with normal cognitive function. c The significant differences between participants with moderate to severe cognitive impairments and those with mild cognitive impairments. Multiple linear regression demonstrated a significant negative association between the number of residual teeth and cognitive impairment (Model 1 β = −0.150; 95% CI: −0.190 to −0.111) ( Table 2), and this association was independent of covariates. However, the coefficient of the association between the number of residual teeth and cognitive impairment decreased when the nutrition score was included as a parameter (Model 2 β = −0.089; 95% CI: −0.126 to −0.052). Moreover, the number of residual teeth was significantly associated with nutritional status (Model 3 β = 0.197; 95% CI: 0.153 to 0.241). This suggested that nutritional status may partly influence the association between the number of residual teeth and cognitive impairment.  Mediation model analysis was performed to determine if nutritional status was the mediator of the effects regarding the number of residual teeth and cognitive impairment. As shown in Figure 2, mediation model analysis showed that a decline in the number of residual teeth was correlated with worse nutritional status (β = 0.197; 95% CI: 0.153 to 0.241); the nutritional status score showed a significant negative association with cognitive impairment (β = −0.309; 95% CI: −0.330 to −0.289). The mediating effects of the number of residual teeth on cognitive impairment, mediated by nutritional status, were statistically significant (β = −0.061; 95% CI: −0.076 to −0.046), and the direct effects were also statistically significant (β = −0.089; 95% CI: −0.126 to −0.050). The total effect of the number of residual teeth on cognitive impairment was statistically significant (β = −0.150; 95% CI: −0.186 to −0.110). This suggested that nutritional status partly mediates the association between cognitive impairment and the number of residual teeth. Expressed as a percentage, the effect of the number of residual teeth on cognitive impairment, mediated by nutritional status, was 40.66% (95% CI: 30.00% to 55.44%) (Supplementary Material Table S2). The mediation analysis' statistical details were adjusted in accordance with the covariates in Table S2. of residual teeth. The remaining covariates (including age, sex, ethnic group, occupation, educational level, spouse status, longevity of family members, denture status, type of drinking water, and history of smoking, alcohol use, and drinking tea) were adjusted in all models. We conducted an additional mediation analysis by stratifying the sample into four subgroups based on age (Figure 3 and Supplementary Material Table S3). Among older adults in the four groups, negative correlations were identified with the indirect effects that were mediated through nutritional status, as well as the direct and total effects that the number of residual teeth have on cognitive impairment. Interestingly, expressed as a percentage, the mediating effect of nutritional status became more potent with age among We conducted an additional mediation analysis by stratifying the sample into four subgroups based on age (Figure 3 and Supplementary Material Table S3). Among older adults in the four groups, negative correlations were identified with the indirect effects that were mediated through nutritional status, as well as the direct and total effects that the number of residual teeth have on cognitive impairment. Interestingly, expressed as a percentage, the mediating effect of nutritional status became more potent with age among individuals aged 50 to 79 years (50-59 group: 25.48% (95% CI: 11.37% to 43.35%); 60-69 group: 44.87% (95% CI: 32.56% to 64.93%); 70-79 group: 57.77% (95% CI: 38.89% to 93.50%)). However, in the subgroup of elderly individuals aged 80 years and older, the mediating effect became weaker (46.99% (95% CI: 25.89% to 84.36%)). The SEM framework was used for path analysis (Figure 4). The correlations between the number of residual teeth and cognitive impairment, as well as nutritional status and cognitive impairment, were both negative. (SEM standardized co-efficient: −0.096 and −0.386). The correlation between the number of residual teeth and nutritional status was positive (SEM standardized co-efficient: 0.134). Furthermore, age, sex, and ethnic groups showed different estimate coefficients compared with the number of residual teeth and nutritional status. In particular, Uyghur groups showed the most obvious coefficients, with regard to the number of residual teeth, of all the ethnic groups (SEM standardized co-efficient: −0.233). Moreover, the strongest coefficients between ethnicity and nutritional status occurred in the Yi group (SEM standardized co-efficient: −0.145). The p value of the entire SEM pathway was statistically significant except for the p value of the Qiang groups ( Figure 4 and Supplementary Material Table S4). These results further confirmed the association between the number of residual teeth, nutritional status, and cognitive function.  . A pathway analysis using the SEM framework. This framework shows that the estimated coefficient of the number of residual teeth to cognitive impairment was −0.096, thus, it exhibited a negative influence. The SEM coefficient of the number of residual teeth to nutrition status was 0.134, thus, it exhibited a positive influence. The SEM coefficient of nutrition status to cognitive impairment was −0.386, thus, it exhibited a negative influence. Age, sex, and ethnic groups have different standardized coefficients to the number of the residual teeth and nutritional status. The p values of all the pathways shown in the framework were significant except for the Qiang groups (** p < 0.005; *** p < 0.001).

Discussion
Based on this cross-sectional study of 6634 older adults from multi-ethnic groups in western China, we identified the partial mediating role of nutritional status in the relationship between the number of residual teeth and cognitive impairment.
The mediation model in this study established a pathway from the number of residual teeth to cognitive impairment, partially mediated by nutritional status. This result aligns well with previous studies. A prior study [14] found that the mediating effect of nutritional status, expressed as a percentage, on the number of teeth, denture wearing, and severe cognitive impairment, was 23.1%. In this study, the mediating effect in older adults in western China, expressed as a percentage, was 40.66%; this was greater than previous results. The differences in cognitive assessment methods may have impacted the results of this study. In our study, the SPMSQ utilized was more sensitive than the Clinical Dementia Rating (CDR), particularly when screening mild cognitive impairment among older adults [18,19]. Therefore, the number of older adults that were identified as having mild cognitive impairments in this study is larger compared with previous studies. We speculated that nutrition may have a stronger influence on individuals with milder cognitive impairments, and thus, their cognitive function.
This study demonstrated a positive correlation between a greater number of remaining teeth, better nutritional status, and improved cognitive function in older adults. We believe that this mediation model consolidates and validates the perspectives of previous studies. First, tooth loss causes profound damage to the acquisition and utilization of nutrients. Older adults without teeth prefer soft food and they consume greater quantities of sugar [10,20] instead of foods such as fresh fruits and vegetables [21]. Yoshihara et al. [22] also suggested that the number of residual teeth was related to the intake of several nutrients; in particular, total protein, sodium, vitamin, and niacin. Second, suboptimal nutritional status impaired cognitive function. Some studies suggested that adherence to a healthy diet pattern, such as a Mediterranean diet, is associated with a lower risk of cognitive impairment [23,24]. This is attributed to the presence of several neuro-nutritional components in the diet [25]. For example, the consumption of omega-3 fatty acids from fish and plant sources can enhance cognitive health and mental health [26,27].
We revealed that the association between the number of residual teeth and cognitive impairment is mediated through nutritional status; this association was strengthened with increasing age. This finding might be attributed to the rate of tooth loss, which increased with age. Previous studies indicated that age is a significant predictor of tooth loss [28]. Older adults with 10 or fewer teeth find it difficult to meet dietary recommendations, and the number of such individuals has increased [29]. Therefore, we inferred that the substantial loss of teeth in relatively senior older adults may exacerbate nutritional issues and make cognitive impairments more severe. However, the mediating effect was reduced among older adults aged 80 years and older. This could be due to the limited representation of individuals in this age group, resulting in increased variability and larger errors in the analysis.
This study also found direct effects between the number of residual teeth and cognitive function. We considered that a higher number of residual teeth is directly related to enhanced cognitive performance to some extent. This finding is consistent with prior studies. Patients suffering from Alzheimer's disease are characterized by a lower number of residual teeth [30]. Yusuke [31] also found that participants with some natural teeth had a 0.40 point higher cognitive function score than those with complete tooth loss. We assumed that this direct correlation may be due to genetic regulation in humans. Genetic factors affect the number of tooth phenotypes [32] and cognitive function [33]. For example, the matrix Gla protein gene (MGP), as a candidate gene, is implicated in the pathogenesis of tooth loss and cognitive impairment through the encoding of the expressed vitamin K binding protein [34][35][36].
Another interesting finding of our study is that ethnic groups had different effects on the number of residual teeth among older adults in western China, particularly in Uyghurs. Currently, there is insufficient research available to provide an explanation for this result. In previous studies, researchers reported that variations caused by genetic factors could influence the number of residual teeth and tooth anatomy [34]. Tan et al. [37] found that the EDARV370A genetic variant in Uyghurs played a key role in dental morphology. Therefore, we speculated that this genetic variation is one of the risk factors for the reduced number of residual teeth in Uygur elderly individuals compared with other ethnic groups. Further research is required to investigate additional factors (such as economic status, occupation, and lifestyle habits) that may influence the number of residual teeth. Furthermore, this study found that nutritional status produced different effects on various ethnic groups, with the Yi group exhibiting the poorest nutritional status. This result can be attributed to the relatively primitive lifestyle and unique dietary culture and customs maintained by the aboriginal Yi ethnic group. Yao et al. [38] reported that the prevalence of underweight people was 9.5% in the Yi Autonomous Prefecture. Dietary nutrient intakes of adult Yi people were varied, and there were differences between rural and urban areas [39]. Based on the abovementioned characteristics of ethnic groups, individualized assessments of health conditions are needed for minority ethnic older adults. Caregivers should implement targeted interventions that are tailored to meet their specific needs.
This study has several limitations that need to be acknowledged. Firstly, the absence of comprehensive oral health data, such as odontalgia, swallowing, and chewing abilities, may introduce potential bias. Additionally, the data on cognitive function and nutritional status were collected by self-report questionnaires, which could be subject to recall bias and may not fully capture the true status of the participants. Therefore, it is necessary to be careful when interpreting the results. Future studies should aim to comprehensively evaluate nutritional status, oral health, and cognitive function using objective measures, and this mediation model should be validated in a longitudinal study.

Conclusions
The mediation model in this study established a pathway from the number of residual teeth to cognitive impairment, partially mediated by nutritional status. This finding indicated that enhancing the nutrition of older adults could reduce the adverse effects of the number of residual teeth on cognitive function among older adults in Western China. The findings could provide guidance for caregivers when providing effective strategies to preserve cognitive function in older adults with few residual teeth. Timely nutrition interventions can help prevent degradations in cognitive function.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/nu15143089/s1, Table S1: The results of multiple comparisons and the Bonferroni correction; Table S2: MNA-SF score as a mediator of the effect of residual teeth on cognitive function; Table S3: Mediating analysis in older adults stratified by age; Table S4  Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.

Data Availability Statement:
The data that support the findings of this study are available from the corresponding author, upon reasonable request.