Inside the Study: Who Was Measured and How

The discovery group came from the Rhineland Study, a large population study in Bonn run by DZNE. The analysis included 6,470 adults with an average age of 56.3, ranging from 30 to 95, and 57% were women. A replication group of 1,034 adults came from EPIC-Potsdam, run by the German Institute of Human Nutrition, with an average age of 49.9. Everyone completed a food frequency questionnaire and gave a blood sample.

The researchers read DNA methylation, small chemical tags that switch genes on and off, at roughly 850,000 sites in the genome, about twice as many as most earlier studies, according to DZNE. They then fed those patterns into three epigenetic clocks: GrimAge and PhenoAge, which estimate how much older or younger the body looks than its calendar age, and DunedinPACE, which estimates how fast someone is aging right now.

What an Epigenetic Clock Can and Cannot Tell You

Epigenetic clocks are statistical tools built by matching methylation patterns to age, health outcomes or the pace of physical decline in large groups of people. They are useful for research because they respond to lifestyle factors such as smoking and can be compared across thousands of people. They are not a diagnosis. A small shift in a clock, like the ones reported here, describes a difference between groups on average, not a precise number of years gained or lost by any individual. That is why the authors describe their results in standardized units rather than in years.

The 10 Eating Patterns, Decoded

The “10 diets” are really 10 scoring systems that researchers use to grade how closely a person eats like a given pattern. Most people score well on some and poorly on others. Lead author Juliana Tavares, a doctoral researcher at DZNE, noted that “hardly anyone counts as a healthy eater by every standard at once.”

Which Patterns Stood Out, and by How Much

The researchers reported results per one standard deviation increase in each diet score, meaning roughly the difference between an average eater and someone noticeably healthier than average by that measure. Effects are also in standardized units, so they describe relative shifts rather than years of life. DASH showed the largest link: each step up was associated with a DunedinPACE value 0.10 standard deviations slower.

When the two study groups were pooled, DASH, Nordic, the Alternate Healthy Eating Index and the Dietary Inflammatory Index remained significant for both GrimAge and PhenoAge after correction for multiple testing. The EAT-Lancet diet was the outlier: it was linked to DunedinPACE but not to the other two clocks, and none of its DNA-site signals replicated.

Where the Diets Overlap

More than 70% of the biological pathways affected were shared across all 10 patterns, DZNE reported, involving cell structure, cell signaling and metabolism. Some diet-specific signals appeared too: heart-related pathways for DASH, and memory and cognition pathways for MIND, which was designed with brain health in mind. Breteler summarized the takeaway: “there is probably no single ‘correct’ diet.”

Replication at the level of individual DNA sites was partial: 39 of 240 sites reproduced in the smaller group. But the headline diet-and-clock associations were similar in direction and size across both cohorts, with correlations of 0.79 to 0.91. The authors note that EPIC-Potsdam diet data were collected in the 1990s, which may explain some differences.

The Limits the Authors Spell Out

The researchers list several caveats. The participants were predominantly German and of European ancestry, so the results have not been tested in other populations or food cultures, including the more varied diets common in the United States. Food frequency questionnaires rely on memory and can miss less common foods. Residual confounding, meaning other lifestyle differences between healthy and less healthy eaters, cannot be ruled out. And the analysis captured a single snapshot in time, so it cannot show that changing diet changes the clocks. None of that erases the consistency of the signal, but it does set its size in context.

The Smoker Surprise

Associations between healthy eating and slower aging were stronger among the 767 current smokers than among former or never smokers, for six of the diet scores: Mediterranean, DASH, AHEI, Nordic, the plant-based index and the inflammatory index. Smokers did not generally eat more healthfully. “There is just more health benefit to gain for smokers by eating healthy because of the enormous detrimental effects of smoking,” Tavares said. The researchers have not identified the mechanism.

Turning the Findings Into Dinner

Because more than 70% of the biological effects were shared, the practical lesson is to eat the things these patterns have in common rather than chase one branded diet. Those shared foods are vegetables, fruits, whole grains, legumes, nuts, fish and plant oils, with less processed meat, refined grains and sugary drinks.

How the Patterns Translate to a Week

The MIND diet’s original scoring offers concrete weekly targets that overlap with the others: leafy greens at least six times a week, berries at least twice, nuts at least five times, beans at least three times, fish at least once and poultry at least twice. Nordic-style eating adds rye or oats, berries and canola oil. DASH emphasizes low-fat dairy and keeping sodium in check.

1. Fill half your plate with plantsVegetables, fruit, beans and whole grains drive most of the 10 scores.
2. Swap one refined grain a dayWhite bread or rice for whole grain improves hPDI, AHEI and DASH scores.
3. Eat fish at least once a weekIt counts toward MIND, Nordic and Mediterranean patterns.
4. Choose plant oilsOlive or canola oil instead of butter fits Mediterranean, MIND and Nordic.
5. Cut sugary drinks firstThey raise unhealthful plant-based and inflammatory scores.
6. Watch sodiumThe standard DASH plan caps it at 2,300 mg a day.

How This Research Unfolded

• 1990sEPIC-Potsdam collects diet data later used for replication.
• May 2025An earlier version of the analysis is posted as a medRxiv preprint.
• Aug. 29, 2026Peer-reviewed paper published in Nature Communications.
• Sept. 17 to 18, 2026DZNE findings reported by trade and U.S. news outlets.
• Oct. 20, 2026ScienceDaily runs the DZNE summary, widening coverage.
• NextResearchers call for studies in more diverse populations and long-term trials.

• Leafy greens 6 times a weekMIND diet scoring target.
• Berries 2 times a weekMIND and Nordic favorite.
• Nuts 5 times a weekMIND and Mediterranean staple.
• Beans 3 times a weekCounts in nearly every score.
• Fish 1 time a week or moreMIND, Nordic, Mediterranean.
• Sodium under 2,300 mgStandard DASH limit.
• Sugary drinks near 0Lowers uPDI and inflammatory scores.
• No smokingThe biggest aging lever of all.