The study’s methodological backbone is the Mini-Mental State Examination, or MMSE, a widely used screening instrument that probes orientation, memory, attention, language and visuospatial skills with a maximum score of 30 points. Scores falling below established cut-offs flag possible cognitive impairment, ranging from mild deficits to dementia-compatible profiles. By combining MMSE results with participants’ self-reported counts of remaining natural teeth, the team was able to classify individuals along two independent dimensions: oral status and cognitive status. Self-reported tooth counts are a limitation worth noting, since recall can be imperfect, but they are a standard approach in large population surveys where full dental examinations are impractical, and prior validation work in comparable cohorts has generally found reasonable agreement between what people report and what clinicians observe.

What elevates this analysis above a simple two-by-two comparison is its use of restricted cubic spline modeling, a statistical technique that allows researchers to plot the shape of a relationship between an exposure and an outcome without forcing it into a preconceived straight line or threshold model. Splines divide the exposure variable into segments joined smoothly at knots, letting the data reveal whether the risk rises steeply at first, plateaus at some critical tooth count, or accelerates only after severe tooth loss. In this case, the spline analysis returned something notable for its simplicity: a linear dose–response association. There was no apparent safe threshold below which tooth loss began to matter, and no plateau where losing further teeth stopped adding risk. The association climbed steadily across the spectrum of remaining dentition, which the authors interpret as evidence that oral health status and cognitive status are connected in a graded, continuous fashion rather than through a single catastrophic tipping point.

The multivariate logistic regression models underpinning these findings adjusted for a battery of potential confounders, the standard arsenal in observational aging research, including demographic characteristics, socioeconomic indicators, health behaviors and chronic disease burden. This matters because tooth loss is not randomly distributed across populations. People with less education, lower income, rural residence, smoking histories and higher loads of chronic illness tend to lose teeth earlier and also face elevated dementia risk for reasons that have nothing to do with their mouths. Statistical adjustment cannot eliminate all such confounding, and the authors are appropriately cautious, framing their results as an association that provides a basis for longitudinal work rather than proof of causation. Still, the persistence of a graded relationship after adjustment strengthens the case that something about oral health itself, or a mechanism tightly linked to it, is doing real work here.

Perhaps the most practically interesting result concerns dentures. The researchers used modified Poisson regression with robust variance, a technique that yields prevalence ratios more interpretable than odds ratios when outcomes are common, to examine the joint association of tooth loss and denture use with cognitive impairment. The comparison group was participants with 20 or more remaining natural teeth who used dentures, a combination reflecting relatively preserved oral function. At the opposite extreme stood participants with no remaining natural teeth who did not use dentures. That group showed the highest prevalence of cognitive impairment in the entire cohort, with an adjusted prevalence ratio of 2.801, meaning their impairment prevalence was nearly three times that of the reference group, with a 95 percent confidence interval of 1.974 to 3.974. The interval lies entirely above 1, indicating a statistically robust elevation.

Yet the denture story comes with an important statistical subtlety. When the team formally tested for additive interaction, the epidemiological concept describing whether two exposures together produce more risk than the sum of their individual effects, they found no statistically significant interaction. In plain terms, tooth loss and non-use of dentures appear to each carry their own association with cognitive impairment, and combining them does not produce a risk that exceeds what their separate contributions would predict. This distinction matters for public health messaging. It suggests that denture use is not a magic shield whose benefit multiplies in the edentulous, but rather that maintaining functional dentition, whether natural or prosthetic, tracks with better cognitive outcomes across the board. Encouraging denture adoption among those who have lost teeth remains a reasonable, low-risk recommendation, but the study cannot yet claim that providing dentures would directly prevent cognitive decline.

Why would missing teeth and a struggling brain travel together? Several biologically plausible mechanisms have been proposed in the broader literature, and this study, while not designed to test them, sits squarely within that framework. One leading hypothesis centers on mastication. Chewing exerts mechanical load on the jaw and periodontal tissues, and animal and human imaging studies have linked reduced masticatory function to decreased regional cerebral blood flow and altered hippocampal activity. Losing teeth reduces chewing efficiency, and dentures restore it only partially, which could translate into chronic under-stimulation of neural circuits involved in memory. A second pathway involves inflammation and infection. Periodontal disease, the leading cause of tooth loss, sustains systemic inflammatory signaling, and inflammatory markers have been implicated in neurodegeneration. Pathogenic oral bacteria have also been detected in brain tissue in some studies of Alzheimer’s disease, though their causal role remains contested. A third route is nutritional: severe tooth loss changes diet composition, often reducing intake of fibrous fruits, vegetables and protein-rich foods, potentially creating deficiencies that compound over decades.

Reverse causation looms over all such studies and deserves honest acknowledgment. Cognitive decline itself may drive tooth loss. People developing dementia may neglect oral hygiene, miss dental appointments, lose the executive function needed to manage dentures, or change their diets in ways that accelerate dental deterioration. A cross-sectional design, which captures a single moment in time, cannot disentangle which came first. The authors explicitly position their work as a foundation for longitudinal studies that follow participants over years, measuring tooth loss and cognitive trajectories in sequence. If prospective cohorts confirm that tooth loss precedes cognitive decline after accounting for baseline cognition, the case for oral health as a modifiable dementia risk factor would strengthen considerably, joining hearing loss, hypertension, smoking and social isolation on the list of targets identified in major dementia prevention commissions.

The population context adds weight to the findings. China is aging at extraordinary speed, with hundreds of millions of people now over 60, and edentulism and severe tooth loss remain common among older Chinese adults, particularly in rural areas where dental care access has historically been limited. The Chinese Longitudinal Healthy Longevity Survey, ethically approved by Peking University’s ethics committee and conducted with written informed consent from all participants, offers an unusually rich window into this population because it deliberately includes the very old, a group often underrepresented in health surveys. If even a fraction of the observed association reflects a causal pathway, interventions promoting tooth retention, periodontal treatment and denture rehabilitation could represent scalable, inexpensive components of dementia prevention strategies, complementing pharmacological approaches that have so far delivered only modest benefits.

For readers, the takeaway is neither panic nor complacency. The study does not claim that brushing your teeth guarantees a protected brain, nor that dentures reverse cognitive impairment. What it does establish, with the technical rigor of spline modeling, adjusted regression and sensitivity to interaction effects, is that the association between remaining teeth and cognitive health is real, graded and linear in this large national sample, and that the worst oral health profile, complete toothlessness without prosthetic replacement, coincides with nearly triple the prevalence of cognitive impairment seen in the best-preserved mouths. The mouth, long treated by medicine as separate from the brain, is increasingly revealing itself as part of the same aging story. As the authors conclude, oral health deserves recognition as a genuine component of healthy aging, and the coming decade of longitudinal research will determine whether the dentist’s chair belongs alongside the cardiologist’s and the neurologist’s in the fight against cognitive decline.

Subject of Research: The association between tooth loss and cognitive impairment in older adults

Article Title: Dose‒response relationship between tooth loss and cognitive impairment in older adults

Article References: Zhang, Y., Shan, X., Qian, Y., Xing, Y., Bao, L., Yang, K., Jin, H., & Yang, F. (2026). Dose‒response relationship between tooth loss and cognitive impairment in older adults. BMC Public Health. https://doi.org/10.1186/s12889-026-29602-4

Image Credits: AI Generated

DOI: 10.1186/s12889-026-29602-4

Keywords: tooth loss, cognitive impairment, oral health, MMSE, dose-response, older adults, dementia, dentures, Chinese Longitudinal Healthy Longevity Survey, periodontal disease, healthy aging, epidemiology