Dietary Interventions and Cognitive Function across the Dementia Continuum: A Systematic Review, Meta-Analysis, Meta-Regression and Call to Action for Research Reform
Bibliographic record
Abstract
Abstract Background Dementia cases are projected to rise to 153 million worldwide by 2050. Diet is a modifiable risk factor, and interventions may delay cognitive decline. No review has synthesised the full randomised controlled trial (RCT) evidence across all dietary interventions and dementia stages. Objective To critically appraise and synthesise evidence of dietary interventions on cognitive function across the dementia disease continuum (healthy, at-risk/preclinical, mild cognitive impairment, clinical). Methods Following PRISMA guidelines, seven databases and three trial registries were searched for RCTs from inception to 28 January 2025. Eligible RCTs evaluated multidomain (e.g., diet plus exercise), whole dietary pattern (e.g., Mediterranean diet), or single food (e.g., blueberries) interventions ≥ 2-weeks in adults, 18+ years. Data extraction was triple coded; risk of bias was assessed using the Cochrane Risk of Bias (RoB2) tool. Narrative synthesis was complemented by pooled random-effects meta-analysis. Prespecified meta-regression explored reasons for heterogeneity. Sensitivity analyses were performed. Results Eighty-three RCTs (110 diet intervention comparisons; n = 24 063 participants) met inclusion criteria. Overall, 190 cognitive measures were extracted and mapped onto the six DSM-V neurocognitive domains. A total of 885 cognitive, neuroimaging, and blood biomarker outcomes were assessed. Pooled meta-analysis ( k = 35) showed a small, significant improvement in global cognition (Hedges’ g = 0.25, 95%CI: 0.15 to 0.36). Multidomain ( g = 0.25, 95%CI: 0.10 to 0.41) and total diet ( g = 0.27, 95%CI: 0.10 to 0.44) interventions showed benefit; single food interventions were non-significant. Trial durations ≤ 12 weeks (β = 0.59) and presence of cardiometabolic comorbidities at baseline (β = 0.45) predicted greater cognitive improvements in multidomain interventions; for total diet interventions, lower risk of bias (β = 0.51) moderated effects. Neuroimaging revealed benefits in MCI across all intervention types. Certainty of evidence was rated very low. Discussion Dietary interventions modestly improve cognitive function across dementia stages, with multidomain and total diet interventions showing the most benefit. RCTs during the preclinical phase (e.g., MCI) of dementia in adults with co-occurring cardiometabolic risk are recommended. Given substantial heterogeneity, risk of bias, and measurement inconsistency, international consensus is urgently needed to standardise intervention protocols (e.g., duration, type, dosage) and cognitive outcome measures to strengthen evidence for dementia prevention and care. Funding NIHR Applied Research Collaboration Oxford and Thames Valley, NIHR Oxford Health Biomedical Research Centre (BRC) Registration NIHR PROSPERO database (registration: CRD42023488336)
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.087 | 0.135 |
| Meta-epidemiology (narrow) | 0.004 | 0.002 |
| Meta-epidemiology (broad) | 0.026 | 0.045 |
| Bibliometrics | 0.010 | 0.008 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.006 | 0.005 |
| Open science | 0.004 | 0.003 |
| Research integrity | 0.003 | 0.003 |
| Insufficient payload (model declined to judge) | 0.005 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".