Objectives: The aim of the present study was to identify the onset predictors of a change in chewing ability over a 3-year period in the frail elderly.
Methods: The subjects were frail elderly people living in southern Japan. Data were collected at baseline (n = 92) and 3 years later (n = 60). The dependent variable was a change in self-rated chewing ability. The independent variables were age, gender, number of present teeth, dentition, maximum bite force (evaluated using a pressure-sensitive foil), basic activities of daily living, self-rated general health status, higher level of competence (evaluated using Tokyo Metropolitan Institute of Gerontology index), cognitive status (evaluated by revised Hasegawa Dementia Rating scale), and quality of life (evaluated using Philadelphia Geriatric Center morale scale) at baseline. To identify the most reliable predictors, bivariate analysis and multiple logistic regression analysis were performed, with the change of chewing ability as the dependent variable.
Results: Bivariate analysis showed the change in chewing ability was significantly associated with general health status (p < 0.01), number of present teeth (p < 0.05) and maximum bite force (p < 0.01). Backward logistic regression analysis revealed that the most reliable predictor of a change in chewing ability at 3 years was general health status (odds ratio = 6.35, 95% CI = 1.69–23.88).
Conclusion: The present findings suggest that general health status at baseline produces a change in chewing ability.