An analysis of influencing factors of oral frailty in the elderly in the community

Using the convenience sampling method, the elderly in three communities of our city were selected as study participants. The inclusion criteria were as follows: (1) ≥ 60 years old; (2) long-term residence in the area (for more than 6 months within 1 year); and (3) people with normal language expression and communication skills. The exclusion criteria were as follows: (1) those who had been diagnosed with severe mental disorders, physical disorders and dementia by professional medical institutions; (2) those who could not participate in and complete the questionnaire survey and a physical examination; (3) and those with severe heart, brain, kidney and other organ dysfunction or in the acute phase of the disease. A total of 380 elderly people were finally included in the study, which was approved by the Ethics Committee of the hospital. All subjects gave informed consent and signed a written consent form.

A paper questionnaire was distributed on the spot by investigators who were trained in a uniform manner, and the same guidance language was used to help the elderly participants complete it. After completion, the questionnaire was collected in and checked on the spot. If there were missing items or incomplete answers, the questionnaire was returned to the respondent for amendment. The investigator assisted those with poor literacy or vision by reading out the questions and responses. A total of 400 questionnaires were distributed, and 380 valid questionnaires were recovered, with an effective questionnaire recovery rate of 95%.

Demographic data of the participants were collected, including age, gender, smoking history, drinking history, education level, marital status, living status and per capita annual income of the family. Disease-related data was also collected, including number of chronic diseases, payment method of medical expenses, number of teeth, number of dentures, dry mouth issues, subjective chewing difficulties, OF score, PF status, oral health, nutritional status, and sleep quality.

Oral frailty was evaluated using the OF checklist proposed by Tanaka et al. [17]. The checklist consists of 8 items: (1) whether it is harder to eat solid food than it was half a year ago; (2) whether they sometimes choke on tea or soup; (3) whether they have false teeth; (4) whether they have dry mouth symptoms; (5) whether the number social outings has decreased compared with half a year ago; (6) whether they can chew hard food, such as peanuts or pickled radish; (7) whether they have brushed their teeth at least twice a day; and (8) whether they see a dentist at least once a year. The score ranges from 0 to 11 points, with 0–2 points indicating a low OF risk, 3 points indicating a moderate risk and ≥ 4 points indicating a high risk. Receiver operating characteristic (ROC) curve analysis was used to evaluate the predictive value of each of the 8 independent variables for OF occurrence, with an OF score of 4 or higher indicating the existence of OF. The area under the curve (AUC), sensitivity, specificity, positive predictive value and negative predictive value of each variable were calculated with respect to predicting the occurrence of OF.

The study employed several forms of assessment. The Frailty Scale, developed by the International Association for Nutrition, Health and Aging, assesses five key areas: fatigue, resistance, ambulation, illness, and weight loss,each item on the scale is scored 1 point,a total score of 3 or more points indicates the presence of frailty [18]. The Oral Health Assessment Tool (OHAT) was revised by Chalmers et al. [19] and consists of a concise oral health checklist with 8 items: lips, tongue, gingival tissue, saliva, natural teeth, dentures, oral cleaning and toothache. The score for each item is 0 (normal), 1 (lesion) or 2 (abnormal), and the total score is 0–16 points; the higher the total score, the worse the oral health. The Mini-nutritional Assessment Short Form was developed by Rubenstein et al. [20] and contains 6 items: body mass index, recent weight change, acute diseases or major psychological changes, activity ability, neuropsychiatric diseases and food intake. The total score can be 0–14 points, with ≥ 11 points indicating normal nutrition and < 11 points indicating malnutrition. Finally, sleep quality was evaluated using the Pittsburgh Sleep Quality Index [21], which has 7 components: subjective sleep quality, sleep latency, sleep duration, habitual sleep efficiency, sleep disturbances, use of sleeping medication and daytime dysfunction. The total score was 0–21, and the higher the score, the worse the sleep quality: 0–5 for excellent sleep quality, 6–10 indicating good sleep quality, 11–15 representing moderate sleep quality, and 16–21 denoting poor sleep quality.

The sample size was calculated using the G*Power software (version 3.1.9.7, Heinrich Heine University Düsseldorf, Germany). Based on a significance level of 0.05, a power of 0.80 and an effect size of 0.15 for multiple linear regression analysis, the minimum sample size was estimated to be 368. Since 380 elderly people participated in the study, it exceeded the minimum requirement. Statistical analysis was performed using SPSS 26.0 statistical software. Normally distributed measurement data were expressed as (x ± s), and analysis of variance was used for comparison between groups. The count data were expressed as frequency (n) or rate (%). The χ2 test was used for those who met the conditions, and the Fisher exact probability method was used for those who did not meet the conditions. The chi-square test was used for the comparison of categorical variables between groups, and the Fisher exact probability method was used when the expected frequency of any cell was less than five.

Receiver operating characteristic curve analysis was used to evaluate the predictive value of each of the 8 independent variables on the occurrence of OF. The dependent variable was the occurrence of OF, defined as having an OF score of 4 or higher. The independent variables were the 10 factors that were included in the multiple linear regression model.

There were 278 community elderly in the frailty group, 137 males and 141 females, with an average age of 77.28 ± 4.67 years and an OF score of 7.28 ± 4.67; there were 58 community elderly in the pre-frailty group, 32 males and 26 females, with an average age of 70.362 ± 4.87 years and an OF score of 6.71 ± 2.03; and there were a total of 44 community elderly people in the non-PF group, 40 males and 4 females, with an average age of 66.50 ± 3.63 years and an OF score of 6.16 ± 1.54. There were statistically significant differences in age, height, sex ratio, residence ratio, living situation, occupation, number of chronic diseases, number of teeth, number of dentures, proportion of dry mouth, subjective chewing difficulties, OF score, oral health score, nutritional status and sleep quality among the three groups (P < 0.05). However, there was no significant difference in the other factors (P > 0.05) (see Table 1).

To further explore the influencing factors of OF, multiple linear regression analysis was used to find out meaningful suspicious risk factors. Stepwise regression analysis was performed with the OF score of the respondents as the dependent variable and the collected variables as the independent variables. The stepwise regression method was used to include and exclude the independent variables (α_entry = 0.05, α_removal = 0.1), and the influencing factors with interaction were eliminated. The results showed that a total of 10 variables could be included, namely age, gender, education level, frailty stage, frailty score, number of dentures, dry mouth, subjective chewing difficulties, oral health score and sleep quality, and the other variables were to be excluded. The model R² = 0.767 indicates that these 10 variables can explain 76.7% of the factors affecting the OF score, and F = 434.73, P < 0.05, indicates that the dependent variable and the 10 variables fit well; the Debin–Watson index is 1.985, which shows that there is no correlation between the independent variables of the model. The significance test results of the 10 independent variables in the model were all P < 0.05, which proved that the 10 independent variables were statistically significant in the model and should be retained. In addition, the variance inflation factor values of the 10 independent variables are all less than 10, so there is no collinear relationship between the respective variables. The multiple linear regression equation obtained by fitting is: Y = 14.614 + 0.336X₁ + 0.431X₂ + 0.276X₃ + 0.518X₄ + 0.183X₅ + 0.389X₆ + 0.517X₇ + 0.640X₈ + 0.245X₉ + 0.483X₁₀, where Y is the OF score, X₁ is age, X₂ is sex, X₃ is standard of culture, X₄ is PF score, X₅ is physical weakness period, X₆ is number of dentures, X₇ is xerostomia, X₈ is subjective chewing difficulty, X₉ is oral health score and X₁₀ is sleep quality (see Table 2).

The results showed that the 10 variables had certain predictive values for the occurrence of OF. Among them, the frailty score, subjective chewing difficulty and sleep quality had higher predictive values for OF. The AUC of frailty predicting OF was 0.751 (95% CI: 0.683–0.862); the AUC of subjective chewing difficulty in predicting OF was 0.765 (95% CI: 0.655–0.831); and, the AUC of sleep quality in predicting OF was 0.736 (95% CI: 0.652–0.781). The predictive value of other factors for the occurrence of OF is shown in Table 3, and the ROC curve is shown in Fig. 1.