Poor Oral Health May Flag Higher Dementia Risk

Poor oral health may be a dementia risk signal—but it is not a proven cause

Oral health belongs on the long list of things that track with healthy aging. A large UK Biobank analysis found that people with prevalent poor oral health problems had a 17% higher adjusted risk of later all-cause dementia. The same study identified several blood proteins that could help explain the link, led by GDF15. That is a useful clue, not proof that gum disease causes dementia or that a dental cleaning prevents it.

The distinction matters. Dementia develops over decades, and poor oral health shares many drivers with cognitive decline: smoking, vascular disease, diet, education, income, frailty, disability, and access to medical care. Early cognitive decline can also make brushing, flossing, appointments, and symptom reporting harder. Observational studies can adjust for some of this; they cannot erase it.

Still, the oral-systemic-brain connection has become harder to dismiss as a random correlation. Cohort studies, neuroimaging analyses, and evidence syntheses repeatedly find that worse oral health travels with worse cognitive outcomes. The prudent conclusion is practical rather than miraculous: take oral health seriously because it matters in its own right, and because it may be one more marker of long-term brain risk.

What the UK Biobank proteomics study found

The most detailed study comes from Beydoun et al.'s 2024 UK Biobank cohort analysis. Researchers followed 37,269 adults aged 50 to 74 years who had baseline plasma proteomic measurements. They excluded people with dementia at baseline and followed the remaining participants for up to 15 years.

During follow-up, 988 participants developed incident dementia, a cumulative incidence of 2.65% in the analytic sample. Participants classified as having prevalent poor oral health problems had a 17% higher adjusted risk of all-cause dementia after the investigators accounted for demographic, socioeconomic, behavioral, and cardiometabolic factors.

The available report does not provide the exact adjusted hazard ratio and 95% confidence interval in the accessible text, nor does it provide oral-health-group-specific absolute dementia risks. Those omissions limit how precisely readers can judge the magnitude and certainty of the association. A 17% relative increase can matter at the population level, but it does not tell an individual how much their personal dementia risk will change.

Why the protein findings are interesting

The investigators measured 1,463 plasma proteins and used mediation analyses to ask whether circulating proteins might statistically sit between poor oral health and later dementia. GDF15 produced the largest single mediation signal, accounting for 28% of the total association as a pure indirect effect. A principal component built from GDF15, IL19, MMP12, and ACVRL1 explained 11% of the total association through a positive pure indirect effect.

Those proteins point toward plausible inflammatory, metabolic, and vascular biology. The broader pathway analysis retained 173 proteins and found enrichment in immune-system, signal-transduction, metabolic, disease, and gene-expression pathways. Within the GDF15-related cluster, transmembrane receptor protein tyrosine kinase signaling was the leading Gene Ontology term.

But a statistical mediator is not a biological witness who saw the crime. GDF15 may lie on a causal pathway, reflect systemic illness that affects both oral health and dementia risk, or partly capture residual confounding. Screening hundreds of proteins also creates opportunities for false-positive and model-selection findings, even when researchers apply correction procedures. These results deserve replication with repeated measurements and better clinical oral-health assessments.

Other human studies support a brain-relevant association

Evidence from several populations points in the same direction, though each study measures a different piece of the problem. In Tsimane forager-farmers, Trumble et al. reported in 2024 that poor oral health was associated with higher inflammation, greater aortic valve calcification, and lower total brain volume. This observational study cannot determine which came first, but it makes a purely dental explanation less convincing. Oral disease may travel with systemic inflammatory and vascular injury signals in settings far removed from the UK Biobank.

A separate 2024 UK Biobank neuroimaging analysis found that poor oral health was associated with markers of white matter injury in middle-aged adults. White matter injury is not dementia, and imaging findings are surrogate outcomes rather than patient outcomes. It is, however, a plausible bridge between vascular risk and later cognitive decline.

Prospective studies of dental care and diagnosed periodontal disease add context. A longitudinal study of older Japanese adults found that dental-care utilization tracked with incident dementia risk. Dental attendance is an imperfect exposure: it can reflect money, mobility, education, health literacy, and healthcare engagement as much as it reflects tooth or gum disease.

That issue becomes even clearer in a prospective All of Us analysis, where unmet dental-care needs due to cost were associated with incident dementia and cardiovascular disease. Cost barriers can worsen oral disease, but they also measure socioeconomic disadvantage and gaps in care. The finding supports the idea that oral and brain health share social and vascular pathways; it does not isolate a specific bacterial or inflammatory cause.

A more direct disease measure came from a 15-year prospective cohort of older men in Northern Ireland. Farsi et al. found that periodontitis was associated with incident cognitive decline and dementia. That is relevant because periodontitis is a clinical condition rather than a proxy for access or behavior. Yet the population was older men, and smoking, diet, frailty, cardiovascular disease, and other shared risks can still confound even a long-running cohort.

The broader literature finds a pattern, not a treatment

Two 2024 evidence syntheses support taking the association seriously. An umbrella review by Lin et al. found recurring links between oral-health factors and cognitive dysfunction across reviews and meta-analyses, but also judged the evidence heterogeneous and lower certainty. A meta-analysis in older adults likewise reported a correlation between worse periodontal health and cognitive impairment.

Heterogeneity is not a technical footnote. Studies define oral health differently: self-reported problems, tooth loss, periodontal disease, dental visits, and inability to afford care are related but not interchangeable. They also measure different outcomes, including cognitive impairment, cognitive decline, imaging markers, and diagnosed dementia. Combining these results into a single claim that gum disease causes Alzheimer's disease would be scientifically sloppy.

The primary UK Biobank study has a similar limitation. Its exposure was a broad prevalent poor-oral-health-problems construct, not a standardized periodontal examination. The accessible publication text does not provide the item-by-item definition or per-item effect estimates. That leaves uncertainty about which oral conditions drove the association and whether the result would look the same with formal periodontal measurements.

What should you do with this information?

Do not buy a "dementia-proof mouth" routine. No study here shows that brushing, flossing, periodontal treatment, implants, or more dental visits prevent dementia. No randomized human trial in this evidence base demonstrates that improving oral health changes incident dementia, cognitive decline, or a validated dementia-prevention outcome.

Do maintain oral health as ordinary preventive care. Seek evaluation for bleeding gums, persistent bad breath, loose teeth, pain, swelling, or difficulty chewing. Keep regular dental care when feasible, clean between teeth in the way your dental professional recommends, and address smoking and cardiometabolic risk factors. These steps already have dental and general-health reasons behind them.

Poor oral health is a credible warning signal for worse cognitive aging, with plausible inflammatory, vascular, and proteomic links. It is not yet a proven dementia cause or a proven treatment target for prevention.

The next useful studies will use stronger clinical oral-health phenotyping, repeat biomarker measurements over time, replicate the protein findings in independent cohorts, and test whether periodontal or dental treatment changes dementia-relevant biomarkers, white matter injury, cognitive decline, or dementia itself. Until then, GDF15 and its fellow proteins remain research leads—not a reason to confuse a correlation with a cure.