One Month of Vegan Eating Nudges the Body’s Aging Clocks

Twenty-four adults in Freiburg, Germany gave up meat entirely, and kept it up for four weeks. A coin flip, or its statistical equivalent, put them in the vegan arm of a randomized trial; another 24 were assigned to eat more than 150 grams of meat a day instead. Both groups ate to the same calorie target throughout, so nobody in this study was simply losing weight. What changed, according to a new genome-wide analysis of the trial, was something you cannot see on a bathroom scale: chemical tags scattered across more than 800,000 spots on their DNA.

Diet has long been linked to heart disease, diabetes and some cancers, but exactly how food talks to our genes has stayed murkier. Researchers from the University of California San Diego and the University of Freiburg set out to eavesdrop on that conversation using DNA methylation, the molecular switches that kick genes up or down without altering the genetic code underneath.

Blood was drawn from all 48 participants before the diet switch and again a month later, and the team combed through 812,934 CpG sites, no mean feat, looking for differences between the two diets. Before the intervention, 158 of those spots already differed between the groups by chance. Afterward, that number climbed to 182, a net rise of about 15 percent. None of it survived the standard statistical correction for testing 800,000-plus sites at once, a point the authors are upfront about in the paper, and one the press release rather steps around.

What the Blood Cells Gave Away

Still, a genome flooded with statistical noise can hide a real signal if you know roughly where to look, and the team had a hunch. Rather than staring at the whole array at once, they used the methylation data to estimate what kinds of white blood cells each sample contained, a trick called cell-type deconvolution.

The estimate predicted fewer neutrophils, the front-line inflammatory cells, and more regulatory CD4 T cells (a subset of white blood cells that help calm immune responses) in the vegan group after the month, a pattern consistent with the diet’s reputation for calming inflammation. Crucially, those predictions lined up closely with actual blood cell counts collected during the original trial, and that agreement is what let the researchers trust the rest of their genome-wide detective work. Lukas Karbacher, the study’s first author and a neuroscience graduate student at UC San Diego, said the match told the team “we were detecting real and meaningful biological changes.”

Jerome Mertens, a co-senior author and associate professor of neurosciences at UC San Diego, frames the whole result as evidence against genetic fatalism. “Our genes are not our destiny,” he said. It’s a tidy line, and also, in this particular case, a fairly literal description of what changed: not the DNA sequence itself, but the tags stuck to it.

Three Clocks, One Disagreement

The most striking part of the analysis concerns biological age itself, estimated not by counting birthdays but by reading methylation patterns through mathematical models called epigenetic clocks, ten of which were tested here. Two of these, PhenoAge and GrimAge, are tuned to predict health outcomes such as disease risk rather than the calendar, and both pointed the same way: the vegan group’s estimated biological age decelerated over the month (PhenoAge’s downward trend reached statistical significance, at negative 1.7 years against a rising 0.56 years in the meat-rich group), while GrimAge nudged in the same direction without reaching significance on its own. Directly comparing the two diet groups after the intervention, though, neither clock’s difference cleared the bar for statistical significance at this sample size, a nuance the press release’s own softer phrasing, that the diet suggested rather than proved a slowdown, mostly gets right. Then, the plot twist: the third clock. Blood&Skin, built specifically to track chronological age and the most accurate of the ten clocks tested, told the opposite story: an acceleration in the vegan group and a deceleration in the meat-rich one, and unlike the other two comparisons, this one was statistically significant both as an interaction effect and head to head.

The authors reckon this is nowhere near proof that a vegan diet extends life, prevents cancer or reverses aging. Every methylation change here rides on a randomized exposure, sure, but each individual reading is still a correlation, not a lab-confirmed mechanism. And the modest size of the shifts, spread across those 800,000-plus sites, echoes a separate cohort study that also found only subtle methylation differences between long-term vegans and meat eaters, meaning a bigger, longer trial is needed before anyone treats this as health advice. Hardly conclusive, by the team’s own admission. That caution lines up with an earlier twin study, which found comparable deceleration on the same clocks after just eight weeks on a vegan diet, though that trial’s vegan twins also lost more weight than their omnivore siblings, a confound this trial specifically avoided by keeping calories equal across both groups.

It also was not a health-food-store vegan diet. To hit the same calorie targets as the meat eaters, some vegan participants leaned on oils, nuts and the odd granola bar, landing somewhere between a whole-food vegan diet and a merely animal-free one, a distinction the authors flag as a real limitation. Not exactly a broccoli manifesto.

The team’s next questions are the obvious ones: do any of these epigenetic shifts stick around after the diet ends, and do they eventually show up as fewer actual disease diagnoses rather than friendlier lab numbers? Storz, from the University of Freiburg Medical Center, framed it simply: “Our study adds to growing evidence that nutrition can shape biology at the molecular level. The next step is to understand whether these epigenetic changes persist over time and whether they translate into meaningful improvements in long-term health. That will require larger clinical studies, but these findings provide an important foundation.”

He was careful to separate the result from calorie-cutting altogether, since both groups were fed to the same energy target throughout, and the effects tracked what people ate rather than how much.

A longer follow-up study from the same Freiburg group is already in the works, this time tracking what happens once the vegan phase ends. Whether the tags stuck to those 800,000-plus sites fade back to baseline, or stay put long after the granola bars are gone, is for now still an open question.

Frequently Asked Questions