Diet, inflammation, and dementia risk: what does the evidence actually show?

The idea that what we eat might influence whether we develop dementia is not new, but the evidence underpinning it has become substantially more detailed over the past decade. We now have plausible biological mechanisms, consistent epidemiological signals, and a small but growing body of dietary intervention data. We do not yet have proof that dietary change prevents Alzheimer's disease. That distinction matters, and this article holds it throughout.

What the evidence supports is a coherent picture in which chronic low-grade systemic inflammation, sustained over years or decades, contributes to the neuroinflammatory processes that characterise Alzheimer's disease (AD). Diet is one of the most modifiable determinants of that inflammatory state. Understanding the relationship between dietary patterns, inflammation, and cognitive outcomes is therefore clinically relevant, even in the absence of definitive trial evidence.

The inflammatory basis of Alzheimer's disease

Alzheimer's disease is not simply a disease of amyloid-beta plaques and neurofibrillary tangles. Neuroinflammation is now recognised as a central and early feature of AD pathogenesis, not a secondary response to established pathology. Activated microglia, elevated pro-inflammatory cytokines including TNF-alpha, IL-1beta, and IL-6, and impaired resolution of inflammation have all been identified in AD brain tissue and in the cerebrospinal fluid of people in preclinical stages of the disease.

This matters for a dietary discussion because systemic inflammatory mediators can cross the blood-brain barrier and activate central neuroinflammatory pathways. A sustained pro-inflammatory state in the periphery, driven by diet, metabolic dysfunction, or chronic infection, is therefore a plausible upstream contributor to neurodegeneration. The evidence for this pathway is strongest in epidemiological and mechanistic studies; it has not yet been tested in randomised trials with cognitive outcomes as the primary endpoint.

Dietary patterns and cognitive decline: the epidemiological signal

The most consistent epidemiological evidence comes from studies of dietary patterns rather than individual nutrients. Two patterns have received the most rigorous investigation: the Mediterranean diet and the MIND diet (Mediterranean-DASH Intervention for Neurodegenerative Delay).

The Mediterranean diet is characterised by high intake of vegetables, legumes, fruit, whole grains, fish, and olive oil, with moderate wine consumption and low intake of red meat and processed foods. A systematic review and meta-analysis by Singh et al. (2014; Epidemiology and Psychiatric Sciences) found that higher adherence to a Mediterranean dietary pattern was associated with a reduced risk of cognitive impairment and AD, with pooled risk reductions in the range of 33 to 40% across included studies. These are observational estimates and carry the confounding limitations common to dietary epidemiology: people who adhere closely to a Mediterranean diet differ from those who do not in income, education, physical activity, and other health behaviours. The signal is consistent but cannot be attributed causally to diet alone.

The MIND diet was developed specifically to incorporate foods with the strongest evidence for neuroprotection. A prospective cohort study by Morris et al. (2015; Alzheimer's and Dementia, 11(9), 1007-1014) found that high MIND diet adherence was associated with a cognitive age approximately 7.5 years younger than those with low adherence, and with a 53% lower rate of AD over a mean follow-up of 4.7 years. Again, this is observational. A subsequent randomised controlled trial of the MIND diet (the MIND trial; Bhupathiraju et al., 2023) found that intensive MIND diet counselling did not significantly slow cognitive decline compared to a healthy diet control in participants without dementia at baseline over three years. The observational benefit may partly reflect residual confounding; the RCT result should temper the confidence with which dietary pattern change is recommended as a dementia prevention strategy.

Ultra-processed food, systemic inflammation, and cognitive risk

Ultra-processed food consumption has emerged as a consistent predictor of elevated inflammatory biomarkers and adverse cognitive outcomes in prospective cohort data. A cohort study by Gonçalves et al. (2023; JAMA Neurology, 80(2), 142-152) found that each 10% increase in ultra-processed food consumption was associated with a 16% higher rate of cognitive decline and a 25% higher rate of global cognitive decline, over an eight-year follow-up in a Brazilian cohort of over 10,000 adults. Mechanistically, ultra-processed foods promote systemic inflammation through multiple pathways: high refined carbohydrate content drives postprandial glucose and insulin spikes, advanced glycation end-products promote oxidative stress, and additives including emulsifiers may disrupt gut barrier integrity and promote dysbiosis.

The gut-brain axis is relevant here. A disrupted gut microbiome, driven by low-fibre, high-sugar dietary patterns, is associated with increased intestinal permeability and systemic translocation of bacterial lipopolysaccharides. These act as potent Toll-like receptor agonists, sustaining a chronic low-grade inflammatory state that may contribute to central neuroinflammation over time. This pathway is mechanistically plausible and supported by animal data; its specific contribution to human AD pathogenesis has not been directly quantified.

The oral-gut-brain axis: a specific mechanistic pathway

One of the more precisely characterised inflammatory pathways linking the environment to the brain runs through the oral cavity. Periodontitis, established gum disease affecting the bone and tissue supporting the teeth, creates a chronically inflamed mucosal surface that serves as a reservoir for gram-negative anaerobic bacteria. These pathogens, particularly Porphyromonas gingivalis (Pg), can enter the systemic circulation and, in experimental models, access the central nervous system.

A landmark study by Dominy et al. (2019; Science Advances, 5(1), eaau3333) detected Pg DNA and its toxic proteases, known as gingipains, in post-mortem brain tissue from AD patients in 96% of 53 samples examined. Higher gingipain levels correlated with tau and ubiquitin pathology. This is hypothesis-supporting evidence, not proof of causation in living patients, and the translational gap between post-mortem detection and active disease contribution has not been closed. Animal studies have, however, confirmed that oral Pg inoculation induces amyloid-beta accumulation and tau hyperphosphorylation in murine models, lending mechanistic credibility to the human tissue observations.

The epidemiological signal is consistent with this pathway. A systematic review and meta-analysis by Leira et al. (2017; Neuroepidemiology, 48(1-2), 21-31) found that individuals with periodontitis were associated with a pooled odds ratio of 1.67 (95% CI: 1.21 to 2.32) for developing dementia. Absolute risk differences across the contributing cohorts are small and vary widely; relative measures of this kind should not be interpreted as large population-level effects. To illustrate the scale: in a population with an annual dementia incidence of roughly 1 to 2% in those aged over 65, a relative risk increase of 67% would translate to an absolute increase in annual risk of approximately 0.7 to 1.3 percentage points. This contextualisation matters for patient communication. A population-based analysis using NHANES 2011 to 2014 data (Beydoun et al., 2020; EBioMedicine, 59, 102936) confirmed the association between periodontal status and cognitive performance at a national sample level.

The relevance to diet is direct. The same dietary patterns that drive systemic inflammation also promote oral dysbiosis. High refined carbohydrate intake, low fibre consumption, and micronutrient deficiency (particularly vitamins C, D, and folate) are established contributors to periodontal disease severity. Diet therefore acts both proximally, through its effects on systemic and gut inflammation, and distally, through its influence on the oral microbiome and periodontal health, as a potential upstream modifier of neuroinflammatory risk.

Periodontal disease is modifiable. A prospective 12-year cohort study found that participants who received periodontal treatment were associated with a 38% lower observed incidence of dementia (hazard ratio 0.62; 95% CI: 0.41 to 0.93). This is an association within treatment patterns in an observational cohort, not a proven treatment effect, and healthy-user bias cannot be excluded. The first randomised controlled trial of periodontal treatment as a dementia intervention, the PETAL trial (NCT04120831), is ongoing.

Specific nutrients with mechanistic and clinical relevance

Beyond dietary patterns, several individual nutrients have both plausible mechanisms and at least moderate human evidence relevant to cognitive outcomes.

Omega-3 long-chain polyunsaturated fatty acids (EPA and DHA) have anti-inflammatory properties and are structural components of neuronal membranes. Epidemiological evidence consistently associates higher oily fish intake with reduced cognitive decline. A prospective analysis in the Nurses' Health Study found that women with the highest fish intake had cognitive scores equivalent to those approximately 1.7 years younger than women with the lowest intake. Randomised trial evidence is more mixed: a Cochrane review of omega-3 supplementation in cognitively healthy older adults found no significant effect on cognitive function over the follow-up periods studied, while some trials in populations with mild cognitive impairment have shown modest benefits on specific memory outcomes. The evidence is sufficient to support dietary adequacy but not to recommend therapeutic supplementation as a dementia prevention strategy in the general population.

Folate and B vitamins have a well-established role in homocysteine metabolism. Elevated homocysteine is an independent risk factor for cognitive decline and is associated with accelerated brain atrophy on neuroimaging. A randomised trial by Smith et al. (2010; PLOS ONE, 5(9), e12244) found that B vitamin supplementation in older adults with mild cognitive impairment significantly slowed brain atrophy rates compared to placebo. The clinical significance of this biomarker effect in terms of dementia incidence has not been established in longer-term trials with clinical endpoints.

Polyphenols, including curcumin, resveratrol, and quercetin, have anti-inflammatory and antioxidant properties with mechanistic relevance to neuroinflammation. Human evidence for cognitive benefit from these compounds remains limited. Most trials are small, short in duration, and use biomarker rather than clinical endpoints. They are discussed in more detail in the relevant library entries linked below.

Vitamin D deficiency is associated with elevated inflammatory markers and with higher rates of cognitive decline in prospective cohort studies. A meta-analysis by Balion et al. (2012; Neurology) found that lower vitamin D levels were associated with poorer performance on cognitive tests, though effect sizes were modest and heterogeneity was substantial across studies. Randomised trial evidence for a cognitive benefit of supplementation in deficient populations is limited, and the large VITAL trial found no significant cognitive benefit over five years in the general population. The relationship may be bidirectional: reduced outdoor activity and dietary variety in those with early cognitive decline may themselves reduce vitamin D status.

Multiple converging mechanistic pathways have been identified, which strengthens the biological plausibility of the diet-inflammation-dementia relationship. It is worth being explicit, however, that convergence of proposed mechanisms does not increase causal probability without corresponding human interventional evidence. Each pathway described above is hypothesis-supporting; their accumulation in a diagram or review does not constitute cumulative proof of causation.

What the evidence does not support

Several popular claims in this area exceed what the evidence can currently sustain.

The claim that following a Mediterranean or MIND diet prevents Alzheimer's disease is not supported by the available RCT evidence. The observational signal is consistent but confounded. The one large RCT (the MIND trial) did not show a significant cognitive benefit over a healthy diet control.

The claim that omega-3 supplementation improves cognitive function in healthy adults is not supported by systematic review evidence. The therapeutic case in mild cognitive impairment is more plausible but not established.

The claim that treating periodontal disease prevents dementia is not established. The observational associations are consistent and the mechanistic pathways are plausible, but no completed randomised trial has demonstrated that periodontal treatment alters the trajectory of cognitive decline.

Risk factors: modifiable and otherwise

The Lancet Commission on dementia prevention, intervention, and care (2020) identified twelve potentially modifiable risk factors accounting for approximately 40% of dementia cases worldwide. Dietary factors contribute to several of these: obesity, diabetes, physical inactivity, and excessive alcohol consumption all have strong dietary components. Hypertension, another modifiable risk factor, is itself partially diet-responsive.

Periodontitis was not included in the Lancet Commission 2020 framework, though subsequent modelling has suggested it may account for a meaningful proportion of attributable risk if the observational associations are causal. The 2024 updated Commission report is expected to address this more explicitly.

Non-modifiable risk factors, including age, genetics (particularly the APOE epsilon-4 allele), and family history, dominate individual risk and cannot be addressed through diet. Dietary modification is best understood as one component of a broader modifiable risk reduction strategy rather than as a primary prevention approach with proven efficacy.

What this means in practice

The practical implication is proportionate. A dietary pattern characterised by high vegetable, legume, fish, and olive oil intake, low ultra-processed food consumption, and adequate micronutrient status is associated with lower systemic inflammation and with more favourable cognitive trajectories in prospective data. That is a reasonable basis for dietary guidance, independent of whether dementia prevention is proven.

Periodontal health should be considered part of the same systemic picture. Diet influences oral microbiome composition, periodontal disease severity, and systemic inflammatory burden simultaneously. Patients asking about diet and brain health should be aware that oral hygiene and periodontal status are part of that conversation, not a separate clinical domain.

Where evidence is limited or outcomes are uncertain, conclusions should be treated as provisional and subject to revision as the evidence base develops.

Key references

Beydoun, M.A., Beydoun, H.A., Weiss, J., et al. (2020) Clinical and periodontal disease measures and cognitive performance: NHANES 2011-2014. EBioMedicine, 59, 102936. https://doi.org/10.1016/j.ebiom.2020.102936

Dominy, S.S. et al. (2019) Porphyromonas gingivalis in Alzheimer's disease brains: evidence for disease causation and treatment with small-molecule inhibitors. Science Advances, 5(1), eaau3333. https://doi.org/10.1126/sciadv.aau3333

Gonçalves, N.G. et al. (2023) Association between consumption of ultraprocessed foods and cognitive decline. JAMA Neurology, 80(2), 142-152. https://doi.org/10.1001/jamaneurol.2022.4397

Leira, Y., Domínguez, C., Seoane, J., et al. (2017) Is periodontal disease associated with Alzheimer's disease? A systematic review with meta-analysis. Neuroepidemiology, 48(1-2), 21-31. https://doi.org/10.1159/000458411

Livingston, G. et al. (2020) Dementia prevention, intervention, and care: 2020 report of the Lancet Commission. The Lancet, 396(10248), 413-446. https://doi.org/10.1016/S0140-6736(20)30367-6

Morris, M.C. et al. (2015) MIND diet associated with reduced incidence of Alzheimer's disease. Alzheimer's and Dementia, 11(9), 1007-1014. https://doi.org/10.1016/j.jalz.2014.11.009

Singh, B. et al. (2014) Association of Mediterranean diet with mild cognitive impairment and Alzheimer's disease: a systematic review and meta-analysis. Journal of Alzheimer's Disease, 39(2), 271-282. https://doi.org/10.3233/JAD-130830

Smith, A.D. et al. (2010) Homocysteine-lowering by B vitamins slows the rate of accelerated brain atrophy in mild cognitive impairment: a randomized controlled trial. PLOS ONE, 5(9), e12244. https://doi.org/10.1371/journal.pone.0012244

For omega-3 evidence, see the Omega-3 entry in the Evidentia library.

For folate and B vitamin evidence, see the Folate and Vitamin B12 entries in the Evidentia library.

For curcumin evidence, see the Curcumin entry in the Evidentia library.

For resveratrol evidence, see the Resveratrol entry in the Evidentia library.

For vitamin D evidence, see the Vitamin D entry in the Evidentia library.