Does tooth loss increase the risk of dementia? Can improvements in chewing ability prevent cognitive impairment or ameliorate cognitive decline? The association between the brain and the stomatognathic system, which plays a key role in chewing and swallowing [
1], has recently been hotly debated in the media [
2‐
5]. Behind these arguments is the emerging concept of the ‘brain-stomatognathic axis’, generally defined as a complex communication network between the brain, including both cortical and subcortical structures, and the stomatognathic system. The top-down control from the brain to the stomatognathic system, such as the coordination between jaw motion and tongue movement, has been established [
1,
6]. However, what has remained unclear is whether input from peripheral structures, such as the sensory signals from the jaw and teeth, can likewise affect the brain. Aging is associated with a decline in both stomatognathic (e.g., tooth loss) [
7] and brain functions (e.g., cognitive impairment or dementia) [
8]. Therefore, the mechanisms underlying the brain-stomatognathic axis have emerged as critical issues in neuroscience as well as in orofacial and geriatric medicine.
Accumulating evidence suggests that cognitive decline may be associated with masticatory dysfunction [
9‐
15]. The term ‘cognitive decline’ generally refers to the decreased cognitive abilities, including short-term and long-term memory, reasoning, and language abilities, which can be associated with normal aging or dementia [
16]. ‘Masticatory dysfunction’, as an umbrella term, refers to a debilitating condition in which normal masticatory function is compromised due to structural factors (e.g., tooth loss) or functional factors (e.g., weaker biting force or poorer masticatory performance) [
17]. The association between cognitive decline and masticatory dysfunction was highlighted in the famous Nun study, which revealed that the number of missing teeth was associated with an increased risk of dementia [
18]. The conclusion was supported by recent meta-analytical findings [
9] and evidence from clinical [
19] and animal research [
10]. These data all suggested a close association between cognitive decline and masticatory dysfunction. For the general public, it is tempting to think that improving masticatory function may be a new path for preventing or ameliorating cognitive decline in elderly individuals [
2‐
5].
In this review, we argue that the current evidence has not provided adequate information regarding the mechanisms underlying the relationship between cognitive decline and masticatory dysfunction. These findings may suggest a stomatognathic-to-brain effect, i.e., poorer masticatory conditions predispose individuals to cognitive decline. However, the cause-and-effect relationship is unclear, and a gap remains between theoretical and clinical investigations. To assess this issue, we first systematically reviewed the recent evidence in the past five years (2012.10.15 ~ 2017.10.15) regarding the association between cognitive decline and masticatory dysfunction, focusing on the findings from clinical/epidemiological and animal research (for a comprehensive review of this topic based on earlier publications, please see [
15,
20]). The limitations of these research models will be discussed. Second, we reviewed the evidence from brain neuroimaging studies. We propose that neuroimaging would be an ideal tool for bridging the gap between clinical/epidemiological evidence and animal research. Third, we summarized and put forward three hypotheses regarding the mechanisms underlying the brain-stomatognathic axis. These hypotheses can be tested using neuroimaging methods. Finally, we summarized the future directions for research into the brain-stomatognathic axis.
Evidence from clinical and epidemiological research
First, we reviewed the findings from systematic reviews or meta-analyses published in the past five years regarding the association between cognitive decline and masticatory dysfunction (for detailed procedures of the literature search and screening, please see Additional file
1). As shown in Table
1, two studies supported for the notion that decreased masticatory function is associated with diminished cognitive functions [
9,
11], whereas two studies concluded that the association remained ‘inconclusive’ or ‘unclear’ [
12,
13]. Another study showed no statistically significant difference in the number of teeth between elderly individuals with and without dementia [
14]. It should be noted that all five of the reviews stated some potential confounding effects or between-study heterogeneity in study design [
9,
11‐
14] As a cause-and-effect relationship cannot be inferred from the findings of the cross-sectional studies [
11,
13], we restrict our discussion to studies with a longitudinal, prospective design. In total, 13 prospective studies ([
21‐
33]) were reported on by the five reviews [
9,
11‐
14]. Nine studies reported that a worse dental status was associated with cognitive decline ([
22,
25,
27‐
29,
32,
33], one study showed a borderline significance of the association [
31]), three studies reported no significant association [
24,
26,
30], and one reported a negative association, i.e., more missing teeth was associated with a lower risk of dementia [
21]. Though the majority of the studies revealed an association between cognitive decline and masticatory dysfunction, a great heterogeneity existed in the outcome assessment regarding masticatory function. For example, the number of missing teeth was adopted as a primary index related to masticatory function [
17]. In one study, the condition of tooth loss was categorized based on the number of teeth (fewer than 20 or not) [
30], while another study assessed the number of missing teeth per decade of follow up [
25].
Table 1
Conclusion from the systematic review and meta-analysis on clinical and epidemiological research published in the past five years
Reference | Criteria of study selection | Number of studies included | Major findings or conclusion (direct quotation) |
Tonsekar et al. 2017 [ 13] | Publications on the relation between periodontitis, tooth loss and dementia | Total: 8a PT:4 RT:3 | ‘The literature on chronic periodontitis and multiple tooth loss as risk factors to dementia remains inconclusive.’ |
| Publications that assessed associations between mastication and cognitive function, cognitive decline and dementia among population over 40 years old | Total: 33 CS:22 PT:11 | ‘Most studies point to a positive association between mastication and cognitive function, including dementia among elderly people.’ |
| Publications that examined the effect of oral health on change in cognitive health or dementia incidence, or the publications that examined the reverse effect. | Total: 11b (all longitudinal studies) | ‘Similarly, cognitive decline was not consistently associated with greater loss of teeth or number of decayed teeth.’ |
| Publications about oral health and orofacial pain Comparison was made between the older people with and without dementia. | Total: 19c CS: 9 CC: 3 RCT: 1 Longitudinal: 6 | ‘……they had an equivalent number of teeth present, similar rate of edentulousness, and equivalent decayed missing filled teeth index.’ |
Cerutti- Kopplin et al. 2016 [ 9] | Publications on the association between oral health and cognitive function, via prospective cohort study designs | Total: 10 PT: 10 | ‘Within the limits of the quality of published evidence, this meta-analysis lends further support to the hypothesis that tooth loss is associated with an increased risk of cognitive impairment and dementia.’ |
Second, it is noteworthy that while most of the studies showed an association between tooth loss and cognitive decline, the elderly participants may have undergone long-term adaptation to their edentulousness/tooth-loss condition during chewing. Therefore, though being a feasible and reliable index, the number of teeth lost may not fully capture the change in masticatory function, and functional assessments (e.g., masticatory performance) may provide a better index for masticatory dysfunction [
11]. To understand the association between masticatory performance and cognitive conditions, we performed a review by systematically searching for and screening the original studies that directly investigated the association between cognitive decline and masticatory ability using a masticatory performance test (Table
2). We found five studies published in the past five years that objectively quantified masticatory performance using functional assessments, including the color-changeable chewing gum test [
34,
35], the two-color chewing gum test [
36,
37], and the Optocal chewing test and the sieve fractionation test [
38]. Masticatory performance decreased in patients with dementia compared to the controls [
36,
38] and was associated with performance on cognitive tests [
34,
35,
37] (Table
2). These findings suggested that functional assessments may be useful for assessing practical chewing performance.
Table 2
Findings from the clinical/epidemiological research that objectively quantified masticatory performance using functional assessments published in the past five years
Reference | Study group | Outcome assessment | Major findings (direct quotation) |
| 295 participants (age ≥ 70 years), a rural city of Korea | Color- changeable chewing gum, MMSE-DS, ADL, MNA | ‘Our findings suggest that poor chewing ability is associated with cognitive impairment or dementia in the elderly living in rural area.’ |
| 16 AD patients (mean age = 76.7yers) and 16 controls (mean age = 75.2 years) | Optocal chewable test, sieve fractionation test, MMSE | ‘Compared to controls, mild AD patients had decreased MP (P < 0.01) and MMSE (P = 0.01). MP showed a moderate negative correlation with MMSE (r = −0.69).’ |
Weijenberg et al. 2015 [ 37] | 114 patients with dementia (age 66–97 years) | Two-color gum mixing ability test, a multi-domain neuropsychological test battery | ‘Significant relationships were observed between masticatory performance and general cognition and between masticatory performance and verbal fluency.’ |
| 29 patients with dementia (age ≥ 75 years), 22 controlsa | Two-color mixing test, dental and nutritional assessment | ‘The chewing efficiency by visual inspection proved worse in participants with dementia than in the controls (p < 0.011) and explained 9.3% of the variance in the diagnosis of dementia.’ |
| 269 community-dwelling elderly aged ≥75 living in Tosa, Japan | Color-changeable chewing gum, MMSE, HDSR and FAB, ADL, QOL, FDSK-11 | ‘Lower cognitive functions were significantly related to low chewing ability; MMSE (P = 0.022), HDSR (P = 0.017) and FAB (P = 0.002).’ |
In general, the findings from recent reviews (Table
1) and studies of masticatory performance (Table
2) suggested that masticatory dysfunction was a potential risk factor for cognitive impairment in elderly individuals. However, these conclusions must be carefully interpreted due to the following limitations:
(A) As previously stated, there is a great heterogeneity among the methods for assessing masticatory function. In fact, only a small number of epidemiological surveys have directly assessed masticatory function, such as masticatory cutting ability, mixing ability or bite force, all of which represent different aspects of masticatory performance [
17,
39]. Moreover, the self-reported chewing experience was not consistent with the results of masticatory performance assessments [
40]. Therefore, it is questionable whether the number of teeth lost, a simplified epidemiological index of the dental condition, can fully capture the difficulty experienced during eating.
(B) Elderly individuals may develop some adaptive strategies to cope with tooth loss during eating. For example, they may carefully choose the type of food they eat and a method for processing it in order to more easily chew and swallow. Therefore, in regard to the association between cognitive decline and masticatory dysfunction, an individual’s nutritional status, eating habits and general physical condition are critical factors to be adjusted for, especially in a longitudinal study [
9].
(C) It is noteworthy that not all of the clinical/epidemiological studies included baseline or follow-up assessments of cognitive functions. Therefore, long-term changes in cognitive functions, which can be partly explained by normal aging, may not be adequately evaluated [
16]. In addition, exercise is regarded as a critical factor to both physical fitness and cognitive ability in elderly individuals [
41]. Long-term follow up on these factors would help to control for the effects of the baseline changes in general physical conditions.
Evidence from animal research
Compared to clinical and epidemiological studies, animal research provides benefits in examining the relevant mechanistic cause-effect relationships, e.g., the cellular or neurochemical changes associated with the studied behavioral deficits [
42‐
60]. These findings from animal research published in the past five years are summarized in Table
3. In general, these studies have supported the hypothesis that cognitive decline is associated with masticatory dysfunction. These studies have demonstrated that cognitive decline is related to cellular and neurochemical change in the hippocampus, including decreased cellular proliferation [
47,
48,
50], decreased levels of brain-derived neurotrophic factor [
44,
47,
59], as well as increased nitrous oxide levels [
55] and extracellular dopamine levels [
52]. These findings suggested that the hippocampus-dependent deficits in learning and memory may contribute to the association between cognitive decline and masticatory dysfunction [
10]. Importantly, using animal models, researchers were able to investigate the interactional effect between masticatory dysfunction and other factors, such as the type of diet [
46,
59], environmental stimuli [
47], and stress [
52]. These factors can partially ameliorate the cognitive deficits induced by masticatory dysfunction [
46,
47]. These findings imply that masticatory dysfunction per se may not be the only determinant to cognitive decline. Rather, the interaction of this dysfunction with other factors can partially account for the observed cognitive deficits.
Table 3
Findings from the animal research published in the past five years
Reference | Strain/Experimental manipulation | Behavioral findings | Cellular /neurochemical findings |
Fukushima-Nakayama et al. 2017 [ 44] | C57BL/6 J mice/Normal (N) or solid (S) diet | Passive avoidance test (−) and Object location test (−) in S group | Hippocampal neurons, neurogenesis, neuronal activity (−), BDNF expression (−) in S group |
| SAMP8 mice/Tooth loss soon after tooth eruption | Morris water maze (−) | Cell proliferation/cell survival in DG (−) Synaptophysin expression in HIP (−) Newborn cell differentiation in DG (X) |
Avivi-Arber et al. 2016 [ 43] | 7 strains of female mice/Molar extraction | | Regional and voxel-wise volumes of cortical brain regions (−) and subcortical, sensorimotor, temporal limbic regions (+) |
| C57BL/6 J mice/Molar extraction (E) and powder (P) or solid (S) diet | Step-through passive avoidance test (−) for E/S and E/P groups 16 weeks later | BDNF-related mRNA in HIP (−) for E/S and E/P groups 16 weeks later |
| APP transgenic mice/Molar extraction | Step-through passive avoidance test (X) | Amount of Aβ and number of pyramidal cells in HIP (X) |
| SAMP8 mice/Molar extraction (E) or intact (I) and standard (S) or enriched (R) environment | Morris water maze (−) in E group | Proliferation and survival of newborn cells in DG (−) and BDNF levels in HIP (−) in E group |
The effect was attenuated in E/R group |
| Tooth extraction (E) and zinc-deficient (ZD) or zinc-sufficient (ZS) | Spatial memory (−) in E/ZD group; recovered in E/ZS group | Astrocytic density in CA1 (+) in ZD group |
| KM mice/Molar extraction | Morris water maze (−) | Levels of NO and inducible nitric oxide synthase in HIP (+) |
| Stress condition (S), stress with voluntary chewing condition (SC), and control condition (C) | Open-Field Test and Elevated Plus Maze Test (+) in SC vs. S | Extracellular dopamine concentration in HIP (+) in SC vs. S |
| CD1 mice/Molar extraction | | Density and absorbance of doublecortin- and neuronal nuclear antigen-positive cells (−) |
| C57BL/6 J mice/Chow diet (C) or liquid diet (L) | Passive avoidance test (−) in L vs. C | BDNF level (+), TrkB (−), and number of pyramidal neurons (−) in HIP in L vs. C |
| Hard (H) or soft (S) diet | Avoidance of butyric acid (−) in S vs. H group | Expression of Fos-ir cells at the Pr5 (+) and the density of BrdU-ir cells in SVZ and OB (+) in H vs. S group |
In S group, avoidance of butyric acid and responses to odors and neurogenesis in SVZ were reversed after hard-dieting for 3 months |
Nose-Ishibashi et al. 2014 [ 50] | C57BL6/J mice/Post-weaning and Hard (H) or soft (S) diet | Home cage activity (−), open fielf test (+), prepulse inhibition (−), learning and memory tests (X) in S vs. H group | Cell proliferation, BDNF and Akt1 gene expression (−) in HIP in S vs. H group |
Kawahata et al. 2014 [ 45] | SAMP8 mice/molar extraction (E) or intact (I) | open-field test (+), object-recognition test (−) and weight (−) in E vs. I group | |
| Sprague-Dawley rats/Solid diet (S) or liquid diet (L) | | Neuronal differentiation and survival (X), HPA-axis function (X), cell proliferation in HIP and hypothalamus (−) in L vs. S group |
Niijima-Yaoita et al. 2013 [ 49] | Powdered (P) or standard (S) diet | Social interaction time (+) | Dopamine turnover (+) and D4 receptor expression (−) in frontal cortex in P vs. S group |
| Transgenic mice/molar extraction (E) or intact (I) | Passive avoidance test (−) in E vs. I group | Neuronal cell number in CA1/CA3 (−) and Aβ, Aβ40, and Aβ4 level (X) in E vs. I group |
| Hard (H) or normal (N) diet | Morris water maze (+) in H vs. N group | Expression of glutamate receptor 1 mRNA (+) in DG in H vs. N group |
Although animal research has provided a great deal of evidence regarding the mechanisms underlying the association between cognitive decline and masticatory dysfunction, several aspects need further clarification:
(A) One of the major challenges faced in interpreting these results is external validity, i.e., the extent to which we can generalize the findings from animal models to human subjects. Most of the animal studies adopted behavioral tasks, such as the Morris water maze and the passive avoidance task, which evaluated spatial and associative learning. One should bear in mind that a poor performance in these tasks does not necessarily reflect cognitive decline in elderly human subjects. The latter is a more complex condition, consisting of changes in short- and long-term memory, language, and reasoning.
(B) Most animal studies adopted tooth extraction as the experimental model to induce masticatory dysfunction, assuming that fewer molars leads to poorer chewing ability. In contrast, human subjects may develop adaptive strategies to cope with tooth loss. For instance, an earlier investigation on 315 edentulous elderly individuals showed that approximately 40% of them did not report chewing difficulty [
61]. A recent investigation of elderly individuals revealed that chewing difficulty associated with a decreased number of functional units depended on the choice of food [
62]. Notably, because experimental animals have a shorter lifespan, it may be difficult to evaluate the long-term adaptive effect of masticatory function using animal models.
Limitations: Beyond the brain-stomatognathic axis
The abovementioned hypotheses focused on the brain mechanisms that underlie the brain-stomatognathic axis. We limited our search to recent studies that focused on the relevant brain and stomatognathic mechanisms. However, it should be noted that other factors may play key roles in the association between masticatory dysfunction and cognitive decline. (A) First, both tooth loss or poor masticatory performance are related to oral hygiene skills and behavior, which may be compromised in the elderly individuals with cognitive impairment [
11]. (B) Second, nutritional biomarkers (e.g., cholesterol) were reported to be independent risk markers of cognitive decline [
93], and in elderly individuals, masticatory ability may explain part of the variance in nutrient intake [
94]. (C) Third, among the major causes of tooth loss are periodontal diseases, which are associated with inflammation related to periodontal pathogens. The inflammatory damage of small blood vessels may play a key role in the pathogenesis of Alzheimer’s disease dementia [
95]. Therefore, microbiological and immunological aspects, particularly periodontal conditions, should be considered [
11,
13]. (D) Evidence from animal research has revealed that chewing may mediate the hippocampus-dependent cognitive deficit by suppressing HPA axis hyperactivity in the hippocampus [
20,
80]. These considerations indicate that an integrative, multi-disciplinary investigation –including behavioral, nutritional, immunological and hormonal research – is necessary for a full understanding of the brain-stomatognathic axis.
Considerations for future research
Based on the current evidence from the clinical, epidemiological, animal, and neuroimaging studies, we argue that the mechanisms underlying the brain-stomatognathic axis have not been elucidated, and the cause-and-effect relationship between cognitive decline and masticatory dysfunction requires more investigation. We suggest that the following aspects of this field must be highlighted in future research:
Refinement of behavioral assessments
In terms of ‘masticatory dysfunction’ most recent studies focused only on the anatomical deficit of the stomatognathic system, using the number of missing teeth as an index. Again, the anatomical deficit does not necessarily reflect the subjective chewing experience or objective masticatory performance. A more comprehensive assessment of the specific elements of stomatognathic function (e.g., masticatory performance, biting force, oral stereognosis, and masticatory muscles) and their interaction should be considered in future research. In terms of ‘cognitive decline’, our review revealed a substantial gap between the cognitive abilities assessed in human subjects and those assessed in animal research. A better reconciliation of the assessments between different species would increase the validity of the conclusions drawn from animal research.
Inclusion of baseline changes in mental and physical conditions
As revealed by animal research, the effect of masticatory dysfunction on cognitive decline may interact with the nutritional [
46,
59] and mental condition of the organism [
47,
52]. Notably, these factors are associated with the physical fitness of the elderly, and the interactional effect of these factors on cognitive decline should be considered.
Experimental designs that include prospective, longitudinal observations
One of the key requirements for understanding the mechanisms the brain-stomatognathic axis is the ability to differentiate between normal and pathological aging. The effect of long-term alterations, such as neuroplasticity effects in the brain, cannot be inferred from cross-sectional findings. A systematic collection of the results from masticatory and cognitive assessments would help clarify the effect of normal aging.
A precise estimation of the effect size
From a clinical perspective, the critical question may not always be to determine ‘If X and Y is associated’. Rather, to make a proper diagnosis and prognosis, clinicians need to know ‘To what extent does the change in X contribute the change in Y’, i.e., an estimation of the effect size of the association. As shown in a recent meta-analysis, suboptimal dentition (i.e., having <20 teeth) was associated with a 20% higher risk of cognitive decline [
9,
52]. However, it is unclear if this is a pure effect from tooth loss or a combined effect due to other confounding factors. To estimate the actual effect that masticatory dysfunction has on cognitive decline is a clinically significant question.