If dementia is all you are worried about, a study of British civil servants suggests you may as well double down on pork pies and fish and chips. In the March 12 JAMA, researchers led by Tasnime Akbaraly at Université Montpellier, France, reported that midlife eating habits among people in the Whitehall cohort had no bearing on their risk of dementia up to 25 years later. A healthier diet did lessen mortality, suggesting at least some benefit from veggies, fruits, and healthy fats.

  • Midlife diet had no correlation to dementia risk over 25 years.
  • Diet got worse in decade prior to dementia onset.
  • A healthier diet did protect against dementia in people who had cardiovascular disease at baseline.

Most commentators hastened to note that despite its size, length, and careful analysis, this observational study cannot provide the final word on the relationship between diet and dementia. “At first glance, the findings seem surprising and liberating for those concerned enough to restrict their diet to healthier choices,” commented Gregory Cole of the University of California, Los Angeles. “But before giving up supposedly healthy diets, readers should consider some of the study’s important limitations.” To Cole, these include a predominantly male cohort, no data for ApoE4 carriers, and missing dietary details that might have fished out a cognitive benefit.

Despite numerous past studies, the relationship between diet and cognitive decline is anything but clear. In intervention studies, most diets, with the exception of the Mediterranean, have failed to slow cognitive decline (Andrieu et al., 2017Marseglia et al., 2018; Nov 2017 news; Valls-Pedret et al., 2015). Some observational studies have reported cognitive benefits of some diets; however, this type of study is fraught with confounders and most follow people for less than 10 years (van de Rest et al., 2015; Cao et al., 2016; Aug 2009 news). 

The researchers sought to address these problems by tapping data from one of the longest-running observational studies of both dietary and cognitive data. The Whitehall II study enrolled 8,225 British civil servants between 1991 and 1993, when their average age was 50. Two-thirds were men. They underwent clinical evaluations approximately every five years, for a total of 24.8 years on average. At three visits—in 1991–1993, 1997–1999, and 2002–2004—they filled out food-frequency questionnaires, from which the researchers generated an Alternative Healthy Eating Index (AHEI) score. The index ranges from zero to 110 points, and was developed as part of the Nurse’s Health Study, which reported that higher scores correlated with a lower incidence of chronic diseases such as cardiovascular disease and cancer (Chiuve et al., 2012). On the AHEI, more points are given for higher intake of foods considered healthy, including whole fruits and vegetables, whole grains, nuts and legumes, and long chain omega-3 fatty acids. Foods deemed unhealthy, such as red and processed meat, trans fat, and sodium, score lower. The researchers modified their questionnaire to reflect foods commonly eaten in England. At baseline, participants had an average AHEI score of 52.

The primary outcome was incidence of dementia ascertained by electronic health records. Cognitive decline, assessed by change in global cognitive scores over an average of 18 years, was a secondary outcome. The researchers controlled for numerous sociodemographic factors, including marital status, occupation, education, and ethnicity, as well as behavioral and health factors, such as smoking, alcohol consumption, physical activity, diabetes, cardiovascular disease, and ApoE4 genotype.

Among the 6,961 participants with sufficient dietary and cognitive data, 344 cases of dementia were reported by 2017, with 75 percent of those recorded after 2010. The researchers found no correlation between AHEI tertile at any visit and dementia risk over the course of the study. They also found no association between tertiles and cognitive decline. However, people in the highest tertile had a lower risk of mortality by 2017, suggesting some correlation between this dietary index and health in this cohort. Because the researchers did not report whether higher AHEI scores correlated with decreased risk of diet-related diseases, such as diabetes, stroke, or cardiovascular disease, it is unclear to what extent the score reflects diet quality.

To address the possibility that the AHEI score did not accurately capture the dietary patterns of the participants, the researchers also used their food-questionnaire answers to generate dietary patterns a posteriori. In a principle component analysis, which identifies dominant patterns within complex datasets, two patterns emerged: a healthy diet, which featured fruits, vegetables, and fish; and a Western-type diet, which was heavy on fried food, processed and red meat, pies, chocolate, high-fat dairy, and refined grains. However, similar to the AHEI, the degree to which participants adhered to either pattern did not correlate with dementia risk. There was one exception. In a post hoc analysis, the researchers found a lower risk of dementia among 41 people with a history of cardiovascular disease who reportedly adhered to a healthy food diet pattern at the time of their 2002–2004 visit. People without a history of cardiovascular disease, and/or who reported a healthy diet only in earlier years, were not protected.

The researchers also explored the possibility that people who developed dementia might veer toward a less-healthy diet. To address this, the researchers compared the trajectories of AHEI scores in people who later developed dementia with those who did not. Their scores were neck and neck for the first nine years of the study, but by 10 years, AHEI scores started to dip slightly in people who later developed dementia. This suggests that a change in diet quality is a feature of preclinical dementia, the researchers wrote. These findings are concerning, noted Nikolaos Scarmeas of Columbia University in New York, as they support the idea of reverse causality, which has long been suspected of confounding observational studies.

While most commentators considered the study important, they noted caveats that preclude firm conclusions. Karen Murphy of the University of South Australia in Adelaide believes the AHEI may not have captured some critical components of a healthy diet. For example, were fruit-eaters noshing on dark berries, which are rich in antioxidants and flavonoids, or bananas, which aren’t? Were chocolate-eaters consuming antioxidant-rich dark chocolate, or gorging on the sugary milk variety? Cole agreed that the devil could be in the details. “Fish consumption may be protective, but it is not clear what this means in the U.K., with its fried fish and chips tradition,” he wrote. In the Women’s Health Initiative cohort study, fish reduced heart failure risk, while fried fish increased it (Belin et al., 2011). 

Cole added that people who developed dementia had more cardiovascular risk factors. “This suggests dementia risk in the study is substantially driven by cardiovascular risk factors, but those with the healthiest diets were not protected from cardiovascular disease or dementia. In contrast, some major studies of Mediterranean diet interventions have shown very clear protection against CVD risk,” he wrote. This suggests that even those in the highest AHEI tertile may not have had healthy diets.

Cole also noted that while the researchers adjusted for ApoE4 status, they did not provide the data broken down by genotype, which might have revealed correlations. In other studies, ApoE4 carriers have derived greater benefit from dietary and other lifestyle interventions than noncarriers. “This study population is 69.1 percent male while dementia risk is substantially higher in females, particularly females with ApoE4,” he added.

Claire McEvoy of Queen’s University in Belfast bemoaned a lack of coherence between dietary indices used across studies. “Defining a dietary index specifically aimed at improving cognitive performance would be desirable for future studies so that the findings can be more easily compared,” she noted. Two previous observational studies also found no association between AHEI scores and dementia (Haring et al., 2016; Richard et al., 2018). 

Just this month, the CARDIA study reported that, among a racially diverse population of participants, greater adherence to the Mediterranean diet or A Priori Diet Quality Score (APDQS) in adulthood associated with greater cognitive function in midlife (McEvoy et al., 2019). The numerous differences between the McEvoy and Whitehall studies—including the diversity and age of the populations, how diet was assessed, and how cognition was measured—preclude a direct comparison between their results, McEvoy wrote. Another important difference between the studies is that CARDIA was racially diverse. “Different ethnicities may suggest different genetic susceptibilities and different cognition-related pathologies and hence different responses to nutritional factors,” he wrote.

Michelle Luciano, University of Edinburgh, agreed that the Whitehall cohort, consisting of predominantly white London office workers, does not reflect broader populations. “Further study needs to include a more varied demographic that might have a wider range of dietary habits,” she wrote. “Other large prospective population studies like this are needed to resolve the debate on whether a healthy dietary style can protect against dementia.”—Jessica Shugart


  1. These results from the recent report from the Whitehall study suggest a null association between midlife dietary patterns and later-life cognitive performance/dementia incidence. Findings are quite disappointing, raising questions and concerns as to whether a true underlying biological association between nutrition and cognition/dementia really exists. The trajectories of dietary habits with aging, leading to an increasing divergence/separation of adherence with advancing age among subjects who become demented versus those who remain cognitively unimpaired add additional concerns since they support the idea of reverse causality (i.e., dietary habits may be altered as subjects approach clinical dementia onset).

    On the other hand, healthy dietary patterns are not all the same. More extensively explored dietary patterns in this field, such as the Mediterranean diet, are briefly investigated in supplementary analyses and only in relation to incident dementia (i.e., not in relation to the more sensitive measures of cognitive decline). Additionally, although there are clear advantages to the dietary pattern approach, it would be interesting to know whether particular elements of these patterns and specific foods or nutrients had notable associations with cognition and dementia.

    Similar to the current study, some previous reports exploring the role of diet and cognition/dementia at middle age also have been negative. But others have been supportive. Actually, in recently published findings from the CARDIA study, adherence to specific healthy dietary patterns (a Mediterranean-style diet and an A Priori Diet Quality Score) in early adulthood was associated with better cognitive health in midlife (McEvoy et al., 2019; Bowman and Scarmeas, 2019). Beyond the difference in age epochs examined, another major difference with the present study and CARDIA is that the latter was racially diverse. Different ethnicities may suggest different genetic susceptibilities and different cognition-related pathologies and hence different responses to nutritional factors. Overall, the literature on nutrition, particularly at middle or earlier age, and cognitive performance is quite conflicting. Underlying clinical, epidemiological, and biological links remain to be explored.


    . Dietary patterns during adulthood and cognitive performance in midlife: The CARDIA study. Neurology. 2019 Mar 6; PubMed.

    . Dietary patterns in early life pay dividends for midlife cognitive performance. Neurology. 2019 Apr 2;92(14):645-646. Epub 2019 Mar 6 PubMed.

Make a Comment

To make a comment you must login or register.


News Citations

  1. LipiDiDiet Data Published
  2. Pass the Bordeaux—Mediterranean Diet, Exercise Reduce AD Risk

Paper Citations

  1. . Effect of long-term omega 3 polyunsaturated fatty acid supplementation with or without multidomain intervention on cognitive function in elderly adults with memory complaints (MAPT): a randomised, placebo-controlled trial. Lancet Neurol. 2017 May;16(5):377-389. Epub 2017 Mar 27 PubMed.
  2. . Effect of the NU-AGE Diet on Cognitive Functioning in Older Adults: A Randomized Controlled Trial. Front Physiol. 2018;9:349. Epub 2018 Apr 4 PubMed.
  3. . Mediterranean Diet and Age-Related Cognitive Decline: A Randomized Clinical Trial. JAMA Intern Med. 2015 Jul;175(7):1094-103. PubMed.
  4. . Dietary patterns, cognitive decline, and dementia: a systematic review. Adv Nutr. 2015 Mar;6(2):154-68. Print 2015 Mar PubMed.
  5. . Dietary Patterns and Risk of Dementia: a Systematic Review and Meta-Analysis of Cohort Studies. Mol Neurobiol. 2016 Nov;53(9):6144-6154. Epub 2015 Nov 9 PubMed.
  6. . Alternative dietary indices both strongly predict risk of chronic disease. J Nutr. 2012 Jun;142(6):1009-18. Epub 2012 Apr 18 PubMed.
  7. . Fish intake and the risk of incident heart failure: the Women's Health Initiative. Circ Heart Fail. 2011 Jul;4(4):404-13. Epub 2011 May 24 PubMed.
  8. . No Association between Dietary Patterns and Risk for Cognitive Decline in Older Women with 9-Year Follow-Up: Data from the Women's Health Initiative Memory Study. J Acad Nutr Diet. 2016 Jun;116(6):921-930.e1. Epub 2016 Mar 31 PubMed.
  9. . Dietary Patterns and Cognitive Function among Older Community-Dwelling Adults. Nutrients. 2018 Aug 14;10(8) PubMed.
  10. . Dietary patterns during adulthood and cognitive performance in midlife: The CARDIA study. Neurology. 2019 Mar 6; PubMed.

Further Reading

Primary Papers

  1. . Association of Midlife Diet With Subsequent Risk for Dementia. JAMA. 2019 Mar 12;321(10):957-968. PubMed.