Dietary patterns and healthy aging, what the evidence supports

Nutrition research changed direction, in official U.S. guidance, in favor of talking about whole dietary patterns rather than individual nutrients, and this page follows that framing rather than isolating any single food or supplement. It also includes the field’s most instructive case study in scientific self-correction, a landmark trial that was formally retracted and republished, because getting that story right matters more than smoothing it over.

What “dietary pattern” means in current guidance

The Dietary Guidelines for Americans, 2025-2030, the current federal edition, released in January 2026 and replacing the 2020-2025 edition, frames its core message around eating patterns built from whole, nutrient-dense foods, protein, dairy, vegetables, fruits, healthy fats, and whole grains, while minimizing highly processed foods high in refined carbohydrates, added sugars, excess sodium, and unhealthy fats. That is a shift in framing worth noting on its own: the prior edition described several named model patterns, including a Healthy Mediterranean-Style pattern, side by side; the current edition consolidates that guidance around flexible, whole-food principles rather than promoting one named pattern above the others. A page describing “the Mediterranean diet” as the current official U.S. recommendation would be describing the previous edition’s framing, not the current one.

What randomized trials actually show

Two trials carry most of the randomized weight behind dietary-pattern claims in aging and chronic disease.

The DASH trial (Appel and colleagues, NEJM, 1997) randomized participants to a control diet, a diet high in fruits and vegetables, or the full DASH combination diet (high in fruits, vegetables, and low-fat dairy, reduced in saturated and total fat). The full combination diet reduced systolic blood pressure by 5.5 mm Hg and diastolic blood pressure by 3.0 mm Hg more than the control diet, both highly significant results, with a larger effect (11.4 and 5.5 mm Hg) among participants who had hypertension at enrollment. This is genuine randomized trial evidence for a specific, measurable outcome, blood pressure, over a period of weeks.

The PREDIMED trial (Estruch and colleagues, NEJM, 2013) randomized 7,447 people in Spain at high cardiovascular risk to a Mediterranean diet supplemented with extra-virgin olive oil, a Mediterranean diet supplemented with mixed nuts, or a control diet. Over a median 4.8 years of follow-up, the trial reported a major cardiovascular event in 96 participants in the olive oil group and 83 in the nuts group, versus 109 in the control group, corresponding to multivariable-adjusted hazard ratios of 0.70 (95% CI 0.54 to 0.92) for the olive oil group and 0.72 (95% CI 0.54 to 0.96) for the nuts group, meaning roughly a 28% to 30% lower rate of major cardiovascular events in each Mediterranean-diet group compared with control.

The PREDIMED retraction, explained honestly

In June 2018, five years after its original publication, NEJM formally retracted and simultaneously republished the PREDIMED paper. The issue was not fabricated data or a false result; it was a randomization irregularity affecting a subset of participants, roughly 1,588 of the 7,447 enrolled, at some of the trial’s study sites, where participants were reportedly enrolled by household or by clinic in some cases rather than individually randomized as the original protocol described. Under a randomized design, that irregularity threatened the statistical basis for treating the groups as genuinely comparable.

The investigators reanalyzed the data with statistical adjustment for the affected sites and ran additional sensitivity analyses, and republished the corrected analysis the same day the original was retracted. According to a summary from the Harvard T.H. Chan School of Public Health, which reviewed the correction directly, the corrected analysis reported effect sizes in a similar range to the original, and the trial’s qualitative conclusion, a lower rate of major cardiovascular events on a Mediterranean diet supplemented with olive oil or nuts compared with a lower-fat control diet, held up.

The accurate way to describe this history is that the 2013 paper was retracted for a real methodological irregularity, not that the trial’s findings were disproven, and not that nothing happened. Both overstatements are common online. PREDIMED remains one of the largest randomized dietary-pattern trials completed to date, and its history is also a reminder that even a landmark trial’s precise effect size can carry more uncertainty than a single number implies.

Where the rest of the evidence is observational

Beyond DASH and PREDIMED, most of the evidence connecting Mediterranean- or DASH-style eating to broader outcomes, overall mortality, cancer, or neurodegenerative disease, comes from observational cohort studies using adherence scores rather than from randomized trials. A widely cited meta-analysis by Sofi and colleagues (BMJ, 2008), pooling 12 cohort studies and over 1.5 million participants, found that a 2-point increase on a Mediterranean-diet adherence score was associated with a 9% lower relative risk of overall mortality (pooled RR 0.91, 95% CI 0.89 to 0.94), a 9% lower relative risk of cardiovascular mortality, and a 13% lower relative risk of neurodegenerative disease incidence, including Alzheimer’s and Parkinson’s disease. Those are real, replicated associations from a very large combined sample, but they are associations between self-reported eating habits and outcomes in free-living populations, not results from randomly assigning people to diets.

Where a popular claim was tested and did not hold up

Cognitive decline is the clearest place where public expectation has outrun the trial evidence. The MIND diet trial (Barnes and colleagues, NEJM, 2023), a 3-year randomized trial in 604 older adults with a family history of dementia, compared the MIND diet (a hybrid of the Mediterranean and DASH patterns) against a control diet that also included mild caloric restriction. Both groups improved on cognitive testing over the 3 years, and the difference between them was not statistically significant. The trial’s own authors concluded that “changes in cognition and brain MRI outcomes from baseline to year 3 did not differ significantly between those who followed the MIND diet and those who followed the control diet.” A specific dietary pattern preventing cognitive decline, stated as a settled fact, is not what this randomized trial found; it is a claim the observational literature suggests and this particular trial did not confirm.

Weighing this evidence against an individual situation

The strongest results here, DASH’s blood pressure effect and PREDIMED’s cardiovascular event reduction, were measured in specific trial populations, people with hypertension or existing cardiovascular risk factors, over specific time periods, and using specific foods provided or subsidized as part of the trial protocol. How a dietary pattern should be adapted for a particular medical condition, medication (including interactions with foods like grapefruit or vitamin K-rich vegetables and certain blood thinners), or nutrient deficiency is a question this population-level trial evidence was not designed to answer, and is a matter for a clinician or registered dietitian who knows the relevant history.