What 105,000 People Over 30 Years Reveal About Aging Well

For thirty years, researchers followed a hundred thousand people to find out who reaches seventy with their mind, body and independence intact. Only about nine percent did. What separated them is more interesting, and more actionable, than the number suggests.

Lit Review Friday · What 105,000 People Over 30 Years Reveal About Aging Well · Published 2026 · 17 minute read

Update, August 2026: Three Papers That Published Since

Two teams, ten days apart, same answer

On 29 June 2026, Kwon and colleagues pooled 28 separate studies and reanalyzed them together. That is 955 people across 24 countries. They wanted to know when the gut community starts to change.

Here is what they found. Microbial diversity drops hardest at the very first step away from robust health (p = 0.0006). Not at the frail end. At the start, before a doctor would call anything wrong. Sixteen markers move with it. Core symbionts like Coprococcus eutactus thin out. Opportunists like Enterococcus gallinarum move in.

Nine days later, a group in Gothenburg published in Nature Communications. They had not read the first paper. They did not cite it. Vilar Geraldi and colleagues followed 2,081 Swedish women, all aged 75 to 80, for a median of about seven years, then checked their answer against 1,448 older adults in China aged 62 to 96. Four hundred and four species tracked with frailty, and most of them tracked the same way on both continents.

And because they followed people that long, they could say what came next. Higher diversity and higher gene richness both tracked with a lower risk of dying, and with a lower risk of a fall bad enough to put someone in hospital. Hip fracture showed nothing. 

Two teams, roughly 4,500 people, twenty-six countries, and neither group had any idea the other one was working on this. Ten days apart, they landed in the same place. I trust that a lot more than I would trust either one alone.

So why does this argue with our companion lit review?

If you read the companion post on what a microbiome test can and cannot see, you have met the other half of this already.

That post digs into a finding from Roager and colleagues in 2016. Some of the highest diversity scores ever recorded come from people whose gut moves very slowly. Nothing is thriving. Food sits. The fiber fermenters chew through what they have and come up empty. So the community switches to protein. More species show up in the count. The chemistry gets worse.

So one post says the diversity number is fragile and can go up for bad reasons. Two new papers say it goes down early and reliably. Both are right.

Here is the problem with the score: it counts who is in the room and tells you nothing about what anybody is doing in there. In the slow-transit study the number went up for a bad reason, and in the frailty studies it went down for an early one, and in both cases you had to know what moved it before it meant anything at all. That is not a tidy answer and I am not going to tidy it.

Can a gut test tell you how fast you are aging?

Not yet. But a third paper got closer than I expected.

Kunihiro and colleagues at the University of Hawaii at Manoa took 123 samples and measured two things in each one: which microbes were present, and how the person's DNA was chemically marked. Those marks are what an epigenetic clock reads. Their cohort included Native Hawaiian and Pacific Islander participants, which most microbiome studies cannot say.

The cool thing here is what failed. They tried three established clocks (Horvath, Levine, GrimAge2), all of which estimate how old you already are, and the gut data predicted none of them. Nothing. The only one that lit up was DunedinPACE, which measures how fast you are aging right now. Species-level rho was 0.408, p = 0.012. Which makes a kind of sense when you sit with it: your gut is not a record of your birthdays, it is what is going on in you this week.

They also checked whether the model was secretly just reading age. It was not. Adding age did not help it. And the single strongest microbe pointing toward slower aging was Bifidobacterium adolescentis.

What this changes about the post below

Nothing. The Tessier finding still stands. A hundred and five thousand people, up to thirty years, and how you eat still tracks with reaching seventy with your mind and body working.

What the new papers add is a second thing to watch. Diet is what goes in. Your microbes decide what comes out. They turn your food into molecules your own cells read as instructions. Two people can eat the same lunch and produce different chemistry. And the 2026 work says the gap between those two people opens earlier than anyone was looking for it.

Update references

1. Kwon C-Y, Choi YH, Kim H, Han K, Jang D, Hwangbo H. Gut microbial signature for frailty discrimination: a metagenomic meta-analysis of 28 independent cohorts. Experimental Gerontology. 2026;113223. https://doi.org/10.1016/j.exger.2026.113223
2. Vilar Geraldi M, Dwibedi C, Jaiswal R, et al. Gut microbiota associates with frailty in older women. Nature Communications. 2026;17(1). https://doi.org/10.1038/s41467-026-75176-5 · FREE FULL TEXT
3. Kunihiro BP, Yamamoto BY, Juarez R, Maunakea AK. Gut microbiome signatures associate with DNA methylation-based biological aging. Scientific Reports. 2026;16(1). https://doi.org/10.1038/s41598-026-52974-x · FREE FULL TEXT
4. Roager HM, Hansen LBS, Bahl MI, et al. Colonic transit time is related to bacterial metabolism and mucosal turnover in the gut. Nature Microbiology. 2016;1:16093. https://doi.org/10.1038/nmicrobiol.2016.93

The post supporting the 2025 Episode 5 podcast continues below, unchanged.

The Finding: About Nine Percent Made It

In March 2025, a group at the Harvard T.H. Chan School of Public Health published something unusual in Nature Medicine. Most nutrition research asks whether a diet prevents a particular disease. This one asked a different question: what does it take to arrive at old age still yourself?

They had the data to ask it. The Nurses' Health Study and the Health Professionals Follow-Up Study have been collecting detailed dietary questionnaires since 1986. Anne-Julie Tessier and colleagues pulled 105,015 participants out of those cohorts, 70,091 women and 34,924 men, average age 53 when the clock started, and followed them for up to thirty years.

By the time the researchers assessed outcomes in 2014 to 2016, 9,771 people, or 9.3 percent, met their definition of healthy aging. Ten point eight percent of the women. Six point two percent of the men.

That number deserves an immediate caveat, because it travels badly. This is not nine percent of people. It is nine percent of a cohort of American nurses and health professionals: educated, health-literate, largely of European ancestry, and already more likely than the general population to exercise, avoid smoking and see a doctor. The authors say so themselves and flag generalizability as a limitation. If anything, the number tells you how demanding their definition was.

What They Actually Measured

Healthy aging in this study was a five-part composite, and every part had to hold. A participant had to survive to age 70. They had to be free of eleven major chronic diseases. And they had to retain cognitive function, physical function and mental health.

The eleven diseases were cancer other than non-melanoma skin cancer, diabetes, myocardial infarction, coronary artery disease, congestive heart failure, stroke, kidney failure, chronic obstructive pulmonary disease, Parkinson's disease, multiple sclerosis and amyotrophic lateral sclerosis.

Notice what is on that list. Two neurodegenerative conditions and one demyelinating one, sitting alongside the cardiovascular and metabolic diseases you would expect. The researchers were not treating the aging brain as a separate department. Hold that thought; we come back to it.

Against that outcome they tested eight dietary patterns, not one. The Alternative Healthy Eating Index, the Mediterranean pattern, healthful and unhealthful plant-based indices, and several others. Diet was averaged across 1986 to 2010, from food-frequency questionnaires repeated every few years, with a six-year lag before outcome assessment so that early disease could not drive the eating pattern being measured.

Does Eating Well Matter More or Less as You Get Older?

Every one of the eight patterns showed the same direction. More adherence, better odds.

That last line is the one worth sitting with, because it points the opposite way from how this research usually gets summarized.

The familiar framing is that dietary effects are cumulative over decades, so you had better start young, and by implication there is a point after which the window has closed. The data here does not behave that way. The association did not fade at the older threshold. It got stronger. Diet adherence discriminated more sharply between people at 75 than it did at 70.

This is an observational finding, and observational findings carry the usual caution: people who eat well at 60 differ from people who do not in a hundred ways the model cannot fully adjust for. But the shape of the result is still real, and it is not the shape of a closing window.

The Ultra-Processed Side of the Ledger

Tessier's group also looked at the other direction. Comparing the highest fifth of ultra-processed food intake to the lowest, the odds of healthy aging were 32 percent lower (95% CI 27 to 37 percent).

This sits inside a much larger literature. In February 2024, Melissa Lane and colleagues published an umbrella review in the BMJ: forty-five pooled analyses drawn from existing meta-analyses, covering 9,888,373 participants. Direct associations turned up between ultra-processed food exposure and 32 of 45 health parameters, spanning mortality, cancer, and mental, respiratory, cardiovascular, gastrointestinal and metabolic outcomes. The evidence they graded as convincing covered cardiovascular-disease mortality, type 2 diabetes, and anxiety and common mental-disorder outcomes.

It matters what we mean by ultra-processed. The Nova classification is not about whether a food came out of a factory. Canned tomatoes are processed. Ultra-processed refers to industrial formulations built from ingredients you would not find in a home kitchen: maltodextrin, hydrogenated oils, protein isolates, modified starches, cosmetic additives. By that definition they supply about 57.9 percent of calories in the United States and 56.8 percent in the United Kingdom. Not a category on the edge of the diet. The majority of it.

How Does Processed Food Actually Reach the Gut Microbiome?

Epidemiology tells you that something is happening. It does not tell you how. For that you need studies that intervene, and the strongest of them are small.

Emulsifiers and the mucus layer

The colon lining is protected by a mucus layer that keeps bacteria at a distance from the epithelial cells beneath. Benoit Chassaing and colleagues showed in mice (Nature, 2015) that dietary emulsifiers thin that layer and let bacteria approach the gut wall, a phenomenon they called microbiota encroachment, with low-grade inflammation following.

Seven years later the same group ran it in people (Gastroenterology, 2022). Sixteen healthy adults, a randomized controlled feeding trial, fifteen grams a day of carboxymethylcellulose for eleven days. The treated group showed reduced microbial diversity, encroachment into the inner mucus layer in two of seven participants examined, and something more interesting for our purposes: reduced fecal short-chain fatty acids and reduced free amino acids.

Read that again. The additive did not only change which bacteria were present. It changed what they produced. Sixteen people over eleven days is a small study and should be held as such, but it is a direct human demonstration that a common food additive can lower the chemical output of the gut community.

Four days is enough to see it

Chunlong Zhu and colleagues (Nutrition Research, 2020) ran a crossover trial in ten people, four days on a fast-food pattern and four days on a Mediterranean pattern. The fast-food arm raised bile-tolerant organisms including Collinsella, Parabacteroides and Bilophila wadsworthia. The Mediterranean arm raised fiber fermenters in the Lachnospiraceae and Butyricicoccus groups, and raised two specific molecules: indole-3-lactic acid and indole-3-propionic acid.

Those two names are worth flagging. Indole-3-propionic acid is the molecule Episode 33 was entirely about: your own cells cannot make it, only certain gut bacteria can perform the final step, and it shows up in tissues throughout the body. Four days of eating differently moved it. That is how quickly the chemistry follows the food.

Speed, and what does not change

Lawrence David and colleagues (Nature, 2014) remains the reference point for how fast diet reshapes the gut, and it carries a caveat that usually gets dropped. Their entirely animal-based and entirely plant-based diets altered community structure and microbial gene expression within days, but they did not produce a significant change in alpha diversity, the count of different organisms present. What moved was composition and function, not richness.

One mechanism that is usually told wrong

The common version of the artificial-colors story is that gut bacteria break down synthetic dyes into inflammatory compounds. The best available work does not support that as stated. Yun Han Kwon and colleagues (Nature Communications, 2022) found that Allura Red AC disrupts the gut barrier through colonic serotonin and myosin light-chain kinase signaling, and critically, the effect appeared in germ-free mice as well. The pathway is at least partly independent of the microbiota.

There is a second detail in that paper that rarely gets quoted. Chronic exposure exacerbated colitis in the mouse models. Intermittent exposure did not. Dose and pattern mattered, not mere presence.

What Diversity Does and Does Not Mean

Two claims circulate in every conversation about the aging gut. Both are more interesting once you look at where they came from.

The thirty plants number

Eating thirty different plant foods a week is now repeated as though it were a clinical threshold. Its source is the American Gut Project, published by Daniel McDonald and colleagues in mSystems in 2018, and what that paper actually did is more useful than the slogan.

The researchers compared people reporting more than thirty plant types per week against people reporting ten or fewer. Those two bins were chosen for the comparison; they were not derived from the data as a tipping point. Within that comparison they found differential abundance of specific short-chain-fatty-acid-fermenting organisms, fewer antibiotic-resistance genes, and higher fecal conjugated linoleic acid.

They also found something the slogan leaves out entirely. Self-reported diet labels, vegan against omnivore, did not explain microbiome configuration. The plant-count question was informative where the identity label was not. That is the real finding: variety of inputs tracks with the community better than dietary identity does.

So thirty is a good heuristic. Keep using it. It is not a threshold, and this is self-reported cross-sectional citizen-science data, not a trial.

Do the oldest people have the most diverse microbiomes?

The tidy version says centenarians carry richer microbiomes than young adults, and that richness is what carried them there. The literature says something considerably stranger.

Elena Biagi and colleagues (2010) found centenarian microbiotas that looked, by conventional measures, compromised: reduced Faecalibacterium prausnitzii, one of the most reliably beneficial organisms in the human gut, alongside raised inflammatory markers, the pattern often called inflammaging. Their 2016 follow-up in Current Biology, extending to semi-supercentenarians past 105, described a core community that contracts with age while subdominant groups expand, with enrichment of Akkermansia, Bifidobacterium and Christensenellaceae.

Tomasz Wilmanski and colleagues (Nature Metabolism, 2021) reframed the whole question in more than nine thousand people. What tracked with healthy aging was not diversity but compositional uniqueness: healthy older adults drifted away from the population average, each toward a configuration of their own. And the inverse held with a sting. Retaining Bacteroides dominance into later life, which is to say staying typical, predicted decreased four-year survival.

Then Yuko Sato and colleagues (Nature, 2021) located something concrete. Centenarians were enriched for microbes producing a distinctive set of secondary bile acids, notably isoalloliphocholic acid, with antimicrobial activity against pathogens including Clostridioides difficile and Enterococcus faecium. Not more organisms. A particular chemistry.

What It Means

Put the pieces beside each other and a single thread runs through them.

A thirty-year cohort finds that dietary pattern tracks with arriving at old age intact, and tracks harder the further out you look. A human feeding trial finds that a single common additive lowers the chemical output of the gut community. A four-day crossover finds specific bacterial metabolites rising and falling with what people ate that week. And the microbiomes of the longest-lived people are distinguished not by how many species they hold but by what those species make.

The consistent variable is not the food itself and it is not the species list. It is the chemistry that emerges when a diverse community is fed diverse material. Fiber is the substrate for most of it, which is worth stating plainly in a culture currently preoccupied with protein: protein matters, and it is not what most of your gut community eats.

Where the Postbiotic Question Comes In

Follow the logic to its uncomfortable end. If what matters is the chemistry a community produces rather than the roster of the community itself, then a real question follows: what do you do when the producers are depleted?

Diet is the primary lever here and the most durable one, and it depends on being able to change what you eat, which is not available to everyone in every season of life. Probiotics take a different route: they deliver a handful of strains into an existing community that may or may not have room for them, and a handful of strains is not the same thing as the chemistry a whole community makes.

That gap is where our own work sits, and I should say plainly that I have an interest in it. If the microbes that produce the critical signals are depleted, then delivering the signals themselves is at least a coherent thing to try. A full-spectrum postbiotic carries the emergent chemistry that only shows up when a whole community is working, which is what a single strain cannot carry. ThaenaBiotic delivers more than 10,000 molecular signals on that basis. It is a dietary supplement, designed to support everyday gut resilience.

That is a thesis, and I want to be clear about what it is not. None of the work above was designed to test it. Those studies are about food, and about the chemistry a gut community makes. None of them studied our product, and none of them tell you anything about what it does. The human trial data that would test our thesis is part of what we are working toward. ThaenaBiotic is one tool in your toolbox, and the toolbox in this particular story is mostly plants.

The Honest Limitations

The anchor study is observational. It establishes association, not causation, and no amount of statistical adjustment converts one into the other. Diet was captured by food-frequency questionnaire, which is the best available tool at this scale and still relies on people accurately reporting what they ate.

The cohort is narrow: American nurses and health professionals, mostly of European ancestry, unusually health-literate. Whether the same patterns hold in populations with different food environments and different constraints is an open question, and the authors raise it themselves.

The mechanistic human work is small. Sixteen people for eleven days on the emulsifier trial, with encroachment observed in two of seven examined. Ten people for four days on the crossover. Small studies that agree with each other and with the larger epidemiology are worth something. They are not the same as a large trial.

And the centenarian literature is cross-sectional. People who reach 105 differ from the rest of us in ways that have nothing to do with bile acids.

Frequently Asked Questions

Is it too late to change my diet in my sixties or seventies?

Nothing in this research suggests a closing window. In Tessier and colleagues (Nature Medicine, 2025), the association between dietary pattern and healthy aging was stronger at a 75-year threshold than at 70. The study cannot prove that starting later causes benefit, but the common claim that the effects only accrue over decades and are therefore out of reach later is not what the data shows.

What counts as an ultra-processed food?

Under the Nova classification it means industrial formulations built largely from ingredients not used in home cooking, such as modified starches, hydrogenated oils, protein isolates and cosmetic additives. A canned tomato is processed. A shelf-stable snack built from a dozen refined components is ultra-processed. Classification is applied inconsistently across studies, which is a genuine limitation of the literature (Brichacek et al., Nutrients, 2024).

Do I need to eat thirty different plants a week?

Thirty is a useful target to aim at, not a validated threshold. In the American Gut Project (McDonald et al., mSystems, 2018), thirty was the upper comparison bin the researchers chose, not a number the data identified. What the study did show is that counting plant types was more informative about the gut community than dietary identity labels were.

Does gut microbial diversity actually matter, or is that a myth?

Diversity is a useful signal and a poor destination. In more than nine thousand people, healthy aging tracked with a gut community becoming more distinctive rather than more diverse, and staying close to the population average predicted worse survival (Wilmanski et al., Nature Metabolism, 2021). What the community produces appears to matter more than how many members it has.

Is a fecal transplant a way to restore gut chemistry?

Fecal microbiota transplantation is a clinical procedure performed by clinicians, and in the United States it is used for a narrow set of defined indications under FDA oversight. It transfers a whole living community rather than the molecules that community makes, and it carries the risk profile of a transplant. Nothing in the research covered here bears on using it for healthy aging.

Do older adults have worse microbiomes?

Different rather than worse. The gut community shifts with age, with a shrinking shared core and expanding individual variation. In people reaching very old age in good health, researchers have found distinctive metabolic capacities rather than simply more organisms, including a particular set of secondary bile acids with antimicrobial activity (Sato et al., Nature, 2021).

Is fiber more important than protein for gut health?

They answer different questions. Protein is what your own tissues are built from. Fiber is the substrate most of your gut community actually consumes, and it is the input that determines much of the chemistry they produce. In a four-day crossover, a plant-rich pattern raised specific bacterial metabolites that a fast-food pattern did not (Zhu et al., Nutrition Research, 2020).

The Bottom Line

One last thing worth naming. Parkinson's disease sits on the list of eleven conditions these researchers used to define healthy aging, alongside stroke and heart failure and diabetes. That is not a coincidence of study design. The next episode in this arc follows the gut into exactly that territory, where the story gets considerably less tidy.

References

1. Tessier AJ, Wang F, Ardisson Korat A, et al. Optimal dietary patterns for healthy aging. Nature Medicine. 2025;31(5):1644-1652. https://doi.org/10.1038/s41591-025-03570-5 · FREE FULL TEXT
2. Lane MM, Gamage E, Du S, et al. Ultra-processed food exposure and adverse health outcomes: umbrella review of epidemiological meta-analyses. BMJ. 2024;384:e077310. https://doi.org/10.1136/bmj-2023-077310 · FREE FULL TEXT
3. Chassaing B, Compher C, Bonhomme B, et al. Randomized controlled-feeding study of dietary emulsifier carboxymethylcellulose reveals detrimental impacts on the gut microbiota and metabolome. Gastroenterology. 2022;162(3):743-756. https://doi.org/10.1053/j.gastro.2021.11.006
4. Chassaing B, Koren O, Goodrich JK, et al. Dietary emulsifiers impact the mouse gut microbiota promoting colitis and metabolic syndrome. Nature. 2015;519(7541):92-96. https://doi.org/10.1038/nature14232
5. Zhu C, Sawrey-Kubicek L, Beals E, et al. Human gut microbiome composition and tryptophan metabolites were changed differently by fast food and Mediterranean diet in 4 days: a pilot study. Nutrition Research. 2020;77:62-72. https://doi.org/10.1016/j.nutres.2020.03.005
6. David LA, Maurice CF, Carmody RN, et al. Diet rapidly and reproducibly alters the human gut microbiome. Nature. 2014;505(7484):559-563. https://doi.org/10.1038/nature12820 · FREE FULL TEXT
7. Kwon YH, Banskota S, Wang H, et al. Chronic exposure to synthetic food colorant Allura Red AC promotes susceptibility to experimental colitis via intestinal serotonin in mice. Nature Communications. 2022;13:7617. https://doi.org/10.1038/s41467-022-35309-y · FREE FULL TEXT
8. McDonald D, Hyde E, Debelius JW, et al. American Gut: an open platform for citizen science microbiome research. mSystems. 2018;3(3):e00031-18. https://doi.org/10.1128/mSystems.00031-18 · FREE FULL TEXT
9. Biagi E, Nylund L, Candela M, et al. Through ageing, and beyond: gut microbiota and inflammatory status in seniors and centenarians. PLoS One. 2010;5(5):e10667. https://doi.org/10.1371/journal.pone.0010667 · FREE FULL TEXT
10. Biagi E, Franceschi C, Rampelli S, et al. Gut microbiota and extreme longevity. Current Biology. 2016;26(11):1480-1485. https://doi.org/10.1016/j.cub.2016.04.016
11. Wilmanski T, Diener C, Rappaport N, et al. Gut microbiome pattern reflects healthy ageing and predicts survival in humans. Nature Metabolism. 2021;3(2):274-286. https://doi.org/10.1038/s42255-021-00348-0
12. Sato Y, Atarashi K, Plichta DR, et al. Novel bile acid biosynthetic pathways are enriched in the microbiome of centenarians. Nature. 2021;599(7885):458-464. https://doi.org/10.1038/s41586-021-03832-5
13. Bradley E, Haran J. The human gut microbiome and aging. Gut Microbes. 2024;16(1). https://doi.org/10.1080/19490976.2024.2359677 · FREE FULL TEXT
14. Whelan K, Bancil AS, Lindsay JO, Chassaing B. Ultra-processed foods and food additives in gut health and disease. Nature Reviews Gastroenterology & Hepatology. 2024;21(6):406-427. https://doi.org/10.1038/s41575-024-00893-5
15. Srour B, Kordahi MC, Bonazzi E, et al. Ultra-processed foods and human health: from epidemiological evidence to mechanistic insights. Lancet Gastroenterology & Hepatology. 2022;7(12):1128-1140. https://doi.org/10.1016/S2468-1253(22)00169-8
16. Brichacek AL, Florkowski M, Abiona E, Frank KM. Ultra-processed foods: a narrative review of the impact on the human gut microbiome and variations in classification methods. Nutrients. 2024;16(11):1738. https://doi.org/10.3390/nu16111738 · FREE FULL TEXT
17. Martinez Steele E, Baraldi LG, Louzada ML, et al. Ultra-processed foods and added sugars in the US diet: evidence from a nationally representative cross-sectional study. BMJ Open. 2016;6(3):e009892. https://doi.org/10.1136/bmjopen-2015-009892 · FREE FULL TEXT
18. Rauber F, da Costa Louzada ML, Steele EM, et al. Ultra-processed food consumption and chronic non-communicable diseases-related dietary nutrient profile in the UK (2008-2014). Nutrients. 2018;10(5):587. https://doi.org/10.3390/nu10050587 · FREE FULL TEXT

This post accompanies the Lit Review Friday episode of Learn Something with Thaena.