Abstract
Although studies have shown an association between oral health and dementia, causality and underlying mechanisms remain under debate. Reverse causality and unmeasured confounding factors, such as childhood cognitive function, raise questions about causality. Regarding theoretical mechanisms, essential oral functions such as eating and speaking are rarely discussed. The aim of this review was 2-fold: 1) to explain how recent epidemiologic studies tried to address these criticisms and 2) to suggest future research directions. To address reverse causality, studies used repeated surveys over time. Some studies considered the bidirectional relationship between oral health and dementia using appropriate methods. However, even in these studies, cognitive function prior to baseline was not incorporated. To infer the influence of unmeasured confounders, quantitative bias analysis using the E value is recommended, wherein the E value indicates the minimum strength of an unmeasured confounder needed to invalidate an observed association. Furthermore, methods that can ignore the effects of unmeasured confounding are good options, including fixed effect analysis and the instrumental variable method. However, few studies have applied these methods, yielding mixed results. Regarding mechanisms, although eating and speaking are essential oral functions, they have often been overlooked as potential mechanisms. These functions have a social aspect that facilitates interpersonal interactions and can reduce social isolation. The expert commission reported that social isolation in later life is 1 of the 14 modifiable risk factors for dementia. When we consider multilayered direct and indirect mechanisms of dementia throughout the life course in addition to the previously proposed mechanisms, such as periodontal inflammation, we find that poor oral health possibly increases dementia through social isolation via eating and speaking problems. In conclusion, based on causal inference studies and theoretical frameworks, oral health may be a modifiable risk factor for dementia. Methodologically and theoretically robust studies considering these points are needed to determine causality between oral health and dementia.