Impact on cognitive function

Feart et al.^{(Reference Feart, Samieri and Rondeau22)} were amongst the first to explore associations between adherence to the Mediterranean diet and cognition. Those authors found that higher adherence to a Mediterranean diet was associated with slower decline in performance on the mini mental state exam (MMSE) – an 11-item test designed to measure cognitive impairment – over a 5-year follow up in 1410 adults from Bordeaux, France. Since then, multiple epidemiological investigations have explored associations between adherence to a Mediterranean diet and cognition. Most studies have reported that higher adherence to a Mediterranean diet is associated with greater global or domain-specific cognitive function, or slower rates of cognitive decline, with effect sizes equivalent to several years of cognitive aging^{(Reference Jennings, Cunnane and Minihane13,Reference Townsend, Logan and O’Neill23–Reference Siervo, Shannon and Llewellyn25)} .

Several randomised controlled trials (RCTs) have also explored whether intervention with a Mediterranean diet impacts cognitive function/decline, with findings broadly aligned with the observational evidence. An analysis of data from the PREDIMED Navarra site suggested that intervention with a Mediterranean diet augmented with high levels of nuts or olive oil for ∼6·5 years, v. a low-fat control group, led to better performance on the MMSE and a Spanish version of the Clock Drawing test^{(Reference Martínez-Lapiscina, Clavero and Toledo26)}. Meanwhile, data from the PREDIMED Barcelona site – which included a more comprehensive suite of cognitive measures pre- and post-intervention – showed better cognitive trajectories over ∼4·1 years with the same Mediterranean diet interventions v. control^{(Reference Valls-Pedret, Sala-Vila and Serra-Mir27)}. Specifically, memory performance was significantly better in the Mediterranean diet plus nuts arm, whilst frontal and global cognitive performance were significantly better in the Mediterranean diet plus olive oil arm, compared with control^{(Reference Valls-Pedret, Sala-Vila and Serra-Mir27)}. Several other studies, including modified Mediterranean diets, have also shown cognitive benefits. This includes the ‘MedDairy’^{(Reference Wade, Davis and Dyer28)} (a Mediterranean diet intervention with 3–4 servings/day of diary products) and ‘MedPork’^{(Reference Wade, Davis and Dyer29)} (a Mediterranean diet intervention with 2–3 servings/week of pork) studies from Australia, which demonstrated greater processing speed and improved mood with the Mediterranean diet interventions v. control. Similarly, recent data from the DIRECT PLUS study, conducted in Israel, reported benefits of both a Mediterranean diet and a green Mediterranean diet (augmented with green tea and a green plant-based drink) on neuroimaging parameters v. a control diet, with the greatest benefits observed in the green Mediterranean diet group^{(Reference Kaplan, Zelicha and Yaskolka Meir30)}.

The recent MedEx-UK study^{(Reference Jennings, Shannon and Gillings31)}, conducted across 3 geographically distinct sites in the UK, showed that intervention with a Mediterranean diet alone, or in combination with increased physical activity, for 6 months resulted in better global cognitive performance and better memory v. a usual care control group. The intervention also improved vascular stiffness and pulse pressure variability in the combined Mediterranean diet and physical activity group, suggesting that the cognitive benefits could be, at least in part, underpinned by improvements in cerebrovascular health.

Despite these promising findings, it is important to note that not all RCTs have reported beneficial effects of a Mediterranean diet on cognition/brain health. For example, neither the MedLey study^{(Reference Knight, Bryan and Wilson32)} (6 month intervention in 137 older Australian adults) nor the much larger NU-AGE study (1 year intervention in 1279 older adults across Europe) reported cognitive benefits of a Mediterranean diet v. control. These results suggest that a Mediterranean diet can, but does not always, lead to tangible cognitive benefits. Potential reasons which could account for between-study differences in findings are explored later.

Impact on dementia incidence

Scarmeas and colleagues were the first to explore associations between Mediterranean diet adherence and Alzheimer’s disease incidence in 2006^{(Reference Scarmeas, Stern and Tang33)}. They followed 2258 participants from the Washington Heights-Inwood Columbia Aging Project (WHICAP) cohort in New York for an average of 4 years (range 0·2–13·9 years). During this time, 262 individuals developed Alzheimer’s disease. Compared with participants with the lowest level of Mediterranean diet adherence (tertile 1), individuals with moderate (tertile 2) and higher (tertile 3) adherence to this dietary pattern had 15 % and 40 % lower risk of Alzheimer’s disease, respectively. A later study by the same researchers using data from two cohorts recruited through the WHICAP project (1880 adults followed for an average of 5·4 years), demonstrated that both higher Mediterranean diet adherence (40 % lower risk in highest v. lowest tertile) and greater physical activity (33 % lower risk in highest v. lowest tertile) were associated with lower Alzheimer’s disease risk^{(Reference Scarmeas, Luchsinger and Schupf34)}. Participants who had both high Mediterranean diet adherence and high physical activity showed 35 % lower Alzheimer’s disease risk compared with individuals with low levels of both.

Multiple other studies have subsequently emerged exploring associations between Mediterranean diet adherence and either all-cause dementia or Alzheimer’s disease^{(Reference Fu, Tan and Lee24,Reference Nucci, Sommariva and Degoni35)} . Most have been characterised by relatively small sample sizes (e.g. 1000–6000 volunteers) with limited dementia cases (e.g. 20–400 all-cause dementia cases), although several studies leveraging large-scale population cohorts have begun to emerge. A large-scale study by Glans et al.^{(Reference Glans, Sonestedt and Nägga36)} reported no significant associations between Mediterranean diet adherence and either all-cause dementia or specific dementia sub-types in 28 025 Swedish participants (including 1943 dementia cases) over a 19·8-year median follow up. In contrast, in an analysis of data from 60 298 participants from the UK Biobank (including 882 dementia cases), our group observed 23 % lower risk of incident all-cause dementia with the highest (v. lowest) tertile of Mediterranean diet adherence^{(Reference Shannon, Ranson and Gregory37)}. Importantly, in that study, there was no significant interaction between Mediterranean diet adherence and polygenic risk for dementia, suggesting that following a Mediterranean diet could lower dementia risk irrespective of genetic vulnerability for this condition.

The overall associations between Mediterranean diet adherence and dementia incidence have been explored in various recent meta-analyses. For example, in a meta-analysis of data from 13 cohort, 6 cross-sectional and 1 nested case-control studies by Nucci et al.^{(Reference Nucci, Sommariva and Degoni35)}, higher adherence to a Mediterranean diet was associated with 11 % lower risk of all-cause dementia and 27 % lower risk of Alzheimer’s disease. Meanwhile, in a meta-analysis of 26 prospective cohort studies by Fu et al.^{(Reference Fu, Tan and Lee24)} higher (v. lower) adherence to a Mediterranean diet was associated with 15 % lower risk of MCI and 19 % lower risk of Alzheimer’s disease.

There are currently no RCTs on the effects of a Mediterranean diet on dementia incidence, which is likely due to the logistical (e.g. required study length, number of participants) and financial challenges of conducting such a study.

Geographical region

Data from a systematic review by Aridi et al.^{(Reference Aridi, Walker and Wright38)} suggests that the Mediterranean diet may be more effective at improving cognitive/brain health in Mediterranean compared with non-Mediterranean regions. Specifically, those authors found beneficial associations between Mediterranean diet adherence and cognitive health in ∼80 % of cohort studies conducted in the Mediterranean region, but only ∼50 % of studies conducted in non-Mediterranean regions. These results are also supported by recent analyses of a pan-European cohort by Gregory et al.^{(Reference Gregory, Ritchie and Ritchie39)}, in which dietary and cognitive methods were standardised across study centres. Those authors found that the beneficial associations between Mediterranean diet adherence and cognition were stronger and more consistently observed in Mediterranean v. non-Mediterranean countries^{(Reference Gregory, Ritchie and Ritchie39)}. Potential explanations include differences between participants from Mediterranean and non-Mediterranean countries in: (1) the composition or duration of adherence to this dietary pattern (e.g. lifelong v. recent adoption), (2) the absolute level of adherence achieved (with Mediterranean diet scores often several points higher in participants from Mediterranean countries), and (3) participation in other brain healthy lifestyle practices (e.g. being physically active and having large group meals which encourage socialisation) which may be more common in Mediterranean countries. Despite this, many studies have reported cognitive benefits associated with/consequent to consumption of a Mediterranean diet in non-Mediterranean regions (including the UK) with effect sizes that are likely to be clinically and practically meaningful (e.g.^{(Reference Shannon, Ranson and Gregory37,Reference Shannon, Stephan and Granic40)} ).

Mediterranean diet scoring methodology

A range of different tools have been used to define level of adherence to the Mediterranean diet (for review, see Siervo et al.^{(Reference Siervo, Shannon and Llewellyn25)}), and participants can achieve different Mediterranean diet adherence scores depending upon which tool is used. In our previous work using data from the EPIC-Norfolk cohort, we demonstrated that the strength of association between Mediterranean diet adherence and cognition varied depending upon the scoring tool used^{(Reference Shannon, Stephan and Granic40)}. Similarly, data from a systematic review by Townsend suggested that beneficial associations between Mediterranean diet adherence and cognitive outcomes were observed more commonly when the Panagiotakos et al.^{(Reference Panagiotakos, Pitsavos and Stefanadis41)} score rather than the Trichopoulou et al.^{(Reference Trichopoulou, Costacou and Bamia42)} score was used for defining level of Mediterranean diet adherence. One possible reason for this may be that the Panagiotakos et al.^{(Reference Panagiotakos, Pitsavos and Stefanadis41)} methodology assigns points ranging from 0–5 for intake of a range of different foods. As such, it can distinguish between small but potentially meaningful differences in intake^{(Reference Panagiotakos, Pitsavos and Stefanadis41)}. Conversely, the Trichopoulou et al. score assigns points of 0 or 1 based on intakes above/ below cohort-specific medians, which may be less sensitive at separating those with better/worse diets^{(Reference Trichopoulou, Costacou and Bamia42)}. The choice of tool used to measure Mediterranean diet adherence may be particularly relevant when conducting research in non-Mediterranean countries. Here, the dietary targets for achieving points, which are often based on norms in Mediterranean countries, may be unachievable for many individuals in non-Mediterranean countries. On the contrary, tools which award points proportional to the level of intake of a food (such as the Panagiotakos et al.^{(Reference Panagiotakos, Pitsavos and Stefanadis41)} score noted above) may help detect small but potentially meaningful differences in level of adherence to the Mediterranean diet.

Participant cohort

There is evidence to suggest that ‘at risk’ population groups, especially those with elevated CVD risk, may be more responsive to a Mediterranean diet. For example, in our analyses of the EPIC-Norfolk cohort, higher Mediterranean diet adherence was associated with lower risk of poor cognitive performance only in participants with an elevated cardiovascular risk (QRISK2) score^{(Reference Shannon, Stephan and Granic40)}. Similarly, RCTs reporting beneficial effects of a Mediterranean diet on cognition (e.g. PREDIMED, MedPork, MedDairy, DIRECT PLUS, MedEx-UK) have typically recruited participants with cardiometabolic co-morbidities (e.g. elevated blood pressure (BP), higher QRISK2 scores, abdominal obesity, dyslipidaemia). In contrast, studies such as MedLey^{(Reference Knight, Bryan and Wilson32)} and NU-AGE^{(Reference Marseglia, Xu and Fratiglioni43)}, which found no effects of a Mediterranean diet on cognition, recruited relatively healthy participants.

It is less clear whether genetic risk (e.g. APOE4 genotype or polygenic risk) modifies associations between Mediterranean diet adherence and cognition/dementia risk^{(Reference Shannon, Ranson and Gregory37,Reference Shannon, Stephan and Granic40,Reference Fote, Geller and Reyes-Ortiz44)} , and more research is needed on this topic.

Summary

Overall, the findings from observational studies and RCTs suggest that consumption of a Mediterranean diet could have beneficial effects on later life cognition and dementia risk, which might be particularly pronounced in certain population sub-groups (e.g. those with elevated cardiovascular risk profiles and individuals residing within the Mediterranean basin). However, these effects are not observed in all studies. Therefore, it will be important to establish the specific individuals for/circumstances under which this dietary pattern is beneficial. Such information would be important when designing for targeted or tailored/personalised dietary interventions to improve population health.

Cardiometabolic health

The most recent Lancet Commission Report on Dementia Prevention identified several vascular and metabolic risk factors for dementia. This includes hypertension, obesity, diabetes and high cholesterol^{(Reference Livingston, Huntley and Liu4)}, all of which could be targeted through consumption of healthy dietary patterns.

Hypertension affects over 1 billion individuals worldwide and is often undetected due to limited symptoms in most people^{(Reference Cherfan, Blacher and Asmar60)}. It is one of the leading risk factor for dementia with a recent meta-analysis estimating the weighted population-attributable fraction for hypertension and dementia to be 7·4 %^{(Reference Stephan, Cochrane and Kafadar52)}. This could be related to a greater white matter hyperintensities, smaller brain volumes and altered functional connectivity with hypertension^{(Reference Lane, Barnes and Nicholas61,Reference Carnevale, Maffei and Landolfi62)} . Meanwhile, BP reduction has been shown to significantly lower risk of MCI and combined MCI or probable dementia^{(Reference Williamson and Pajewski63)}. Healthy dietary patterns have consistently been associated with modest BP reduction (e.g. 1–5 mmHg), with greater effects emerging with longer duration intervention and in those with elevated baseline BP^{(Reference Cowell, Mistry and Deighton64,Reference Tomé-Carneiro and Visioli65)} . In addition, healthy dietary patterns have been linked with other beneficial cardiovascular changes, including improvements in endothelial function^{(Reference Shannon, Mendes and Köchl66)}. This could contribute further towards lower risk of dementia, especially vascular dementia and Alzheimer’s disease, both of which have strong vascular components^{(Reference Fang, Hsieh and Hu67)}. The BP-lowering effects of healthy dietary patterns have been linked to the typically high levels of fruit and vegetables, pulses, wholegrains and fish – sources of nutrients with anti-hypertensive properties such as fibre, inorganic nitrate, potassium, magnesium and n-3 fatty acids – alongside the typically low levels of sodium^{(Reference Cowell, Mistry and Deighton64,Reference Shannon, Stephan and Minihane68,Reference Shannon, Gregory and Siervo69)} .

Type 2 diabetes is an established risk factor for cognitive decline and dementia, with risk increasing in accordance with the duration and severity of diabetes^{(Reference Livingston, Huntley and Liu4)}. A meta-analysis of 14 studies including > 2 million participants suggested that type 2 diabetes increases all-cause dementia risk by 60 %^{(Reference Chatterjee, Peters and Woodward70)}. Imaging studies suggest that diabetes alters regional blood flow (including both hypo- and hyper-perfusion of different brain regions), increases the risk of vascular brain pathology, disturbs brain glucose metabolism, alters grey matter volume and impairs functional networks in the brain^{(Reference Meng, Liu and Li71,Reference van Harten, de Leeuw and Weinstein72)} . Meanwhile, consumption of a healthy dietary pattern has been shown to reduce diabetes risk in a dose-dependent manner in observational studies^{(Reference Zeraattalab-Motlagh, Jayedi and Shab-Bidar73)}. Similarly, data from the PREDIMED Rues cohort showed 52 % lower type 2 diabetes incidence over a ∼4 year period with a Mediterranean diet intervention^{(Reference Salas-Salvadó, Bulló and Babio74)}. Interestingly, a recent study by Lotan et al.^{(Reference Lotan, Ravona-Springer and Shakked75)} found that greater adherence to a Mediterranean diet was associated with better cognitive trajectories in individuals with type 2 diabetes. Collectively, these data suggest that consumption of a healthy dietary pattern could improve cognition and/or lower dementia risk both by preventing type 2 diabetes and by attenuating the effects of diabetes on the brain in individuals living with this condition.

Both the Mediterranean and MIND diets, even when individuals are allowed to eat ad libitum/ without deliberate energy intake restriction, have been associated with lower risk of obesity and more favourable blood lipid profiles^{(Reference Fateh, Muhammad and Kamari76–Reference Lotfi, Saneei and Hajhashemy79)} – both established risk factors for dementia^{(Reference Stephan, Cochrane and Kafadar52)}. This could be related to satiety-inducing effects (e.g. due to high intake of wholegrains, nuts and vegetables), impact on the gut microbiome and lower intake of energy-dense ‘convenience’ foods^{(Reference Lotfi, Saneei and Hajhashemy79)}. Similar features of these diets, alongside the high intake of unsaturated lipids through e.g. olive oil and oily fish, could also contribute towards the beneficial impacts on circulating cholesterol concentrations (e.g. lower LDL)^{(Reference Dinu, Pagliai and Casini80,Reference Martínez-González, Ruiz-Canela and Hruby81)} . A recent meta-analysis including data from over 1 million participants suggested that each 1 mmol increase in circulating LDL was associated with an 8 % increased risk of incident all-cause dementia^{(Reference Wee, Sukudom and Bhat82)}. Meanwhile, in another meta-analysis, statin use was associated with a 20 % lower risk of all-cause dementia and 32 % lower risk of Alzheimer’s disease^{(Reference Olmastroni, Molari and De Beni83)}. The precise mechanisms through which high LDL is linked to dementia risk remains uncertain, but has been proposed to include increased stroke risk (a known dementia risk factor), and increased deposition of amyloid-β and tau in the brain^{(Reference Wee, Sukudom and Bhat82)}.

Inflammation

Inflammation – both systemically and specific to the brain (i.e. neuroinflammation) – has been linked to cognitive decline and dementia risk^{(Reference Tangestani Fard and Stough84)}. In addition, diets rich in pro-inflammatory nutrients and foods have been shown to elevate markers of inflammation and increase risk of dementia^{(Reference Griffiths, Matu and Tang12,Reference Shi, Lin and Li85)} . For example, in an analysis of data from 166 377 participants from the UK Biobank, Shi et al.^{(Reference Shi, Lin and Li85)} found that for each unit increase in dietary inflammatory index score (which quantifies the inflammatory potential of the diet), the risk of all-cause dementia increased by 4·6 %.In addition, healthy dietary patterns, including the Mediterranean and Nordic diets, have been associated with lower concentrations of circulating inflammatory markers (e.g. c-reactive protein, IL-6 (IL-6) and TNF-alpha (TNF-a))^{(Reference Shannon, Ashor and Scialo86,Reference Lankinen, Uusitupa and Schwab87)} , which have been associated with higher risk of dementia^{(Reference Metti and Cauley88)}. Moreover, it has been suggested that the Mediterranean diet could impact the function of microglia – resident immune cells in the brain – promoting a shift from the pro-inflammatory and neurotoxic M1-phenotype to the anti-inflammatory and neuroprotective M2-phenotype^{(Reference Hornedo-Ortega, Cerezo and De Pablos89)}.

Gut microbiome

Several lines of evidence suggest that the gut microbiome could play a role in the development of Alzheimer’s disease, and could represent a potential target for prevention/treatment strategies (for review, see^{(Reference Seo and Holtzman90)}). For example, there are notable differences in the gut microbiome of individuals with Alzheimer’s disease, and pre-clinical Alzheimer’s disease, compared with healthy controls^{(Reference Seo and Holtzman90,Reference Ferreiro, Choi and Ryou91)} . Since it is possible that these differences could be due to reverse causality, more convincing evidence on the causal role of the gut microbiome in Alzheimer’s development comes from faecal microbiome transplantation studies. Transplanting the faecal microbiota from wild-type mice to transgenic mice expressing genes associated with Alzheimer’s disease development attenuates the formation of amyloid-β plaques and neurofibrillary tangles in the brain and minimises cognitive impairment^{(Reference Kim, Kim and Choi92)}. Recent Mendelian randomisation studies also suggest that the gut microbiome and its metabolites may be causally connected with cognitive performance^{(Reference Cao, Xing and Liang93)}. Healthy dietary patterns, particularly the Mediterranean diet, have been shown to impact the composition and metabolism of the gut microbiome, although effects are inconsistent between studies, which may be due to the specific constellation of foods included in this dietary pattern when administered in each study^{(Reference Kimble, Gouinguenet and Ashor94)}. There is also limited data on whether the benefits of healthy dietary patterns on cognition may be mediated, at least in part, by changes in the gut microbiome. The NU-AGE study, which involved a 1-year intervention with a Mediterranean-like diet, demonstrated diet-related alterations in the microbiome, with taxa enriched by the diet associated with improvements in cognition^{(Reference Ghosh, Rampelli and Jeffery95)}. In contrast, data from the MedDairy study showed modest changes in the abundance of select bacterial taxa with an 8-week Mediterranean diet plus dairy intervention. However, there were no significant associations between the altered taxa and cognitive outcomes^{(Reference Choo, Murphy and Wade96)}. Further research on the effects of healthy dietary patterns on the gut microbiome and cognition/dementia risk is needed.

Pace of biological aging

Healthy dietary patterns might also impact cognitive decline/dementia risk by influencing the pace of biological aging. The Mediterranean diet has been proposed to impact at least 9 of the hallmarks of aging^{(Reference Shannon, Ashor and Scialo86)} – that is, key determinants of the aging trajectory including genomic instability, telomere attrition, epigenetic alterations, loss of proteostasis, deregulated nutrient sensing, mitochondrial dysfunction, cellular senescence, stem cell exhaustion, and altered intercellular communication. This could contribute towards lower risk of age-related diseases. Moreover, a recent study found that a slower overall pace of aging (quantified using the DunedinPACE epigenetic clock) accounted for 27 % of the effect of a MIND diet and 22 % of the effect of the Mediterranean diet on risk of dementia^{(Reference Thomas, Ryan and Caspi97)}.

Brain-specific mechanisms of action

Consumption of healthy dietary patterns may have direct effects on the brain, which are detectable using neuroimaging or post-mortem techniques and could contribute to better cognition and/or lower dementia risk. For example, higher adherence to the Mediterranean and MIND diets, and several of their component foods, has been associated with lower Alzheimer’s disease pathology, including lower amyloid-β load in the brain^{(Reference Agarwal, Leurgans and Agrawal98–Reference Hill, Goodwill and Gorelik101)}. Whether this is due to reduced amyloid-β production and deposition, greater clearance, or a combination of both remains unclear. Both the Mediterranean and MIND diets have been associated, albeit inconsistently, with reduced brain atrophy and preserved cerebral microstructure^{(Reference Barnes, Dhana and Liu46,Reference Gregory, Pullen and Ritchie102–Reference Melo van Lent, O’Donnell and Beiser106)} . Using data from post-mortem prefrontal cortex tissue, Li et al.^{(Reference Li, Capuano and Agarwal107)} identified a transcriptomic profile, comprising 50 genes, which was correlated with level of adherence to the MIND diet. In further analyses, they found that this brain transcriptomic profile was associated with slower rates of cognitive decline and a lower risk of dementia. Higher Mediterranean diet adherence has been linked with improved cerebral bioenergetics, including greater brain glucose metabolism^{(Reference Matthews, Davies and Murray108)}, alongside greater structural connectivity^{(Reference Ruiz-Rizzo, Finke and Damoiseaux109,Reference Pelletier, Barul and Féart110)} . The Mediterranean diet also elevates circulating levels of brain-derived neurotrophic factor (BDNF), which plays a key role in synaptic plasticity, neurogenesis, gene transcription regulation and intracellular signalling cascade activation^{(Reference Xue, Waseem and Zhu111,Reference Sánchez-Villegas, Galbete and Martinez-González112)} . This could contribute towards the beneficial effects on brain health markers^{(Reference Xue, Waseem and Zhu111)}.

Additional large-scale, longer-duration interventions

At present, there are few intervention studies exploring the impact of healthy dietary patterns on cognitive aging, and none on dementia risk. Most of the current studies have used the Mediterranean diet as the choice healthy eating model and have been relatively short in duration (6–12 months), with some exceptions (e.g.^{(Reference Martínez-Lapiscina, Clavero and Toledo26,Reference Barnes, Dhana and Liu46)} ). They have also typically involved poor representation from minority ethnic communities and individuals from a lower socioeconomic background. As these individuals are disproportionately impacted by dementia^{(Reference Livingston, Huntley and Liu4)}, their inclusion in future trials, and the adaptation of trials to ensure the needs of these communities is met, should be a priority. Longer duration (e.g. several years), large-scale (e.g. adequately powered to detect cognitive and/or neuroimaging changes) studies comparing the feasibility, acceptability and efficacy of different dietary patterns on brain health are therefore warranted.

Utilisation of Mendelian randomisation

Whilst financial and logistical reasons mean that the emergence of additional large-scale dementia prevention trials is likely to be slow, Mendelian randomisation provides the opportunity to establish causal links between diet and dementia incidence and is likely to become increasingly popular. This approach uses genetic variants (e.g. single nucleotide polymorphisms) as instrumental variables to make cause-effect inferences about the link between an exposure (e.g. a dietary component/pattern) and outcome (e.g. dementia incidence). Mendelian randomisation studies are already beginning to shed new light on diet-dementia links. For example, a recent finding from Mendelian randomisation studies is that higher intake of alcohol is linearly associated with dementia risk amongst current drinkers^{(Reference Zheng, Liao and Luo113)}. This challenges the traditional perspective from observational research, which has posited a ‘J’ shaped relationship between alcohol intake and dementia risk. The findings suggest that there is no safe level of alcohol intake for dementia risk. The growing use of polygenic scores, which reflect overall diet quality as genetic proxies for dietary patterns, will allow causal evidence linking dietary patterns and dementia to be established.

Identification of the optimal window for intervention

At present, there is little agreement on the optimal window for intervening to lower dementia risk. Most interventions have focused on intervening in non-symptomatic individuals in mid-to-later life, although some studies (e.g. LipiDiDiet Trial^{(Reference Hendrix, Soininen and Solomon114)}) have shown benefits of dietary change in older individuals with prodromal Alzheimer’s disease. It is possible that the optimal time window for intervention might also differ between different individuals (e.g. depending upon sex, ethnicity, wider lifestyle factors or genetic predisposition). Recently, there has been a call to shift focus towards earlier adulthood, a time in which lifelong habits may be established and where many dementia risk factors (e.g. diabetes, overweight/obesity, high alcohol intake, smoking, high BP) are already present^{(Reference Farina, Bridgeman and Gregory115)}. Identifying the optimal (and most feasible) time(s) to intervene is an important priority for future studies.

Utilising state-of-the-art biomarkers

Blood-based biomarkers such as plasma amyloid-β 42/40 ratio, phosphorylated tau species (e.g. p-tau181 or p-tau217) and neurofilament light chain (NfL) are emerging as promising markers of Alzheimer’s disease risk^{(Reference Schöll, Verberk and Del Campo116)}. These biomarkers are viewed as accessible, cost-effective, and highly promising tools which could be used to (a) identify ‘at risk’ groups for trial enrolment, and (b) serve as a surrogate outcome in prevention trials. Their use is likely to become increasingly common in future years. However, various logistical challenges (e.g. establishing suitable cutoffs for different populations, confounding of levels by other health conditions, standardisation of analytic approaches)^{(Reference Schöll, Verberk and Del Campo116)} will need to be overcome before the full potential of these biomarkers can be realised.

Expanding evidence to non-Alzheimer’s dementia

Much of the (observational) research currently available on dietary patterns and dementia risk focuses on either all-cause dementia or Alzheimer’s disease. There are comparatively fewer studies exploring the impact of dietary patterns on vascular dementia – the second most common cause of dementia – and other dementia sub-types^{(Reference Griffiths, Matu and Tang12)}. Indeed, a recent systematic review by Griffiths et al.^{(Reference Griffiths, Matu and Tang12)} identified 16 studies focusing on foods or dietary patterns and vascular dementia risk, with only 1–2 studies (in some occasions with contradictory findings) for most dietary exposures. More research on vascular dementia and other dementias is therefore crucial to inform personalised and population-based dementia prevention initiatives.

Diet-drug interactions

As new pharmacological treatments for Alzheimer’s disease (e.g. anti-amyloid immunotherapies) begin to emerge, it is important to understand whether the efficacy and (often severe) side effect profile of these drugs can be mitigated by lifestyle factors, including diet. There is growing evidence that dietary factors, including healthy dietary patterns, have the potential to modulate the effectiveness of anti-cancer immunotherapies^{(Reference Golonko, Pienkowski and Swislocka117)}, which provides tentative proof of concept for exploring the interaction between diet and anti-amyloid immunotherapies (and other drugs) in the future.