Some modifiable risk factors for dementia are closely related to oral health. Although eating and speaking abilities are fundamental oral functions, limited studies have focused on the effect of malnutrition and lack of social interaction between oral health and dementia. We investigated the mediating effects of nutritional and social factors on the association between the number of teeth and the incidence of dementia. This 6-y cohort study used data from the Japan Gerontological Evaluation Study targeting older adults aged 65 y and above. The number of teeth (exposure) and covariates in 2010 (baseline survey), mediators (weight loss, vegetable and fruit intake, homeboundness, social network) in 2013, and the onset of dementia (outcome) between 2013 and 2016 were obtained. The Karlson-Holm-Breen mediation method was applied. A total of 35,744 participants were included (54.0% women). The mean age at baseline was 73.1 ± 5.5 y for men and 73.2 ± 5.5 y for women. A total of 1,776 participants (5.0%) had dementia during the follow-up period. There was a significant total effect of the number of teeth on the onset of dementia (hazard ratio, 1.14; 95% CI, 1.01-1.28). Controlling for nutritional and social mediators, the effect of the number of teeth was reduced to 1.10 (95% CI, 0.98-1.25), leaving an indirect effect of 1.03 (95% CI, 1.02-1.04). In the sex-stratified analysis, the proportion mediated by weight loss was 6.35% for men and 4.07% for women. The proportions mediated by vegetable and fruit intake and homeboundness were 4.44% and 4.83% for men and 8.45% and 0.93% for women, respectively. Furthermore, the proportion mediated by social networks was 13.79% for men and 4.00% for women. Tooth loss was associated with the onset of dementia. Nutritional and social factors partially mediated this association.
Keywords: cohort studies; dental public health; epidemiology; gerontology; nutrition/nutritional sciences; psychosocial factors.