Dysbiosis

Your gut houses an entire metabolic organ — trillions of microbes that ferment your fibre, train your immune system and broadcast chemical signals to your brain, muscles and metabolism. With age that ecosystem drifts, and where the gut wall starts to leak (clearly in old mice, less clearly in people), the leak feeds the body-wide inflammation of aging — which makes this one of the more modifiable hallmarks, though the human trial evidence is thinner than the enthusiasm around it.

Dysbiosis is the twelfth of the twelve hallmarks of aging, added in the 2023 expansion of the framework. The gut microbiome co-evolves with you across life: rapid diversification in infancy, a stable symbiotic ecosystem through adulthood, and then a progressive remodelling in late life sometimes called biome-aging — a gradual loss of beneficial mutualists, a drop in the microbes that make protective short-chain fatty acids, and an expansion of opportunistic, pro-inflammatory species known as pathobionts.[1] In mice, that imbalance looks more like a cause than a passenger: transplanting gut microbes from old mice into young ones sped up inflammation in the brain and retina and made the gut wall leakier, and microbes from young donors reversed those changes.[2] The proposed route from gut to decline is by feeding inflammaging.[3] In people the evidence is so far only associational.

How confident should you be

Mechanism — Moderate. Human intervention — Weak to Moderate, and weaker than the field's enthusiasm. Consumer products — Caution, in the sense that they are sold well ahead of their evidence.

The mechanism on this page is coherent and well supported in animals: butyrate feeds the gut lining, the lining fails without it, and in old mice what leaks through drives inflammation; in people that chain shows up mainly in the frail and ill. Where the evidence thins is the step everyone wants — that deliberately changing your microbiome changes how you age. The best human trial on this page ran for a year in 612 people and its authors describe their own associations as "relatively weak." The centenarian data is cross-sectional, so it tells you what long-lived guts look like, not how they got that way, and the best long-term human data, a 15-year Finnish cohort, links a pathobiont-rich gut to earlier death without showing that one causes the other. And the field has a large commercial fringe — gut-testing kits, personalised supplement subscriptions — whose claims are not supported by any of this.

One possible reversal worth carrying into the sections below: more short-chain fatty acid may not be uniformly better. In mouse models of Alzheimer's and Parkinson's disease, adding these molecules made the disease worse, although in people with Parkinson's the gut appears to make too little of them, not too much.

How the aging gut drifts

Moderate. Repeatedly seen, but descriptive, and much of it tracks frailty, care setting and medicines rather than age itself.

Several age-related changes — slower gut motility, a thinning mucosal layer, a less vigilant immune system, dietary shifts and rising numbers of prescription medicines — together nudge the gut ecosystem from cooperative balance toward pathobiont dominance.[4] The changes reported most often in older guts are a relative loss of Faecalibacterium and the Lachnospiraceae family, both major producers of short-chain fatty acids (SCFAs), and, especially in the very old, more Proteobacteria, a group linked to gut inflammation. But a systematic review of 27 human studies found the specific changes varied from study to study, that the mucus-dwelling Akkermansia tended to rise rather than fall, and that every study was a snapshot rather than people followed over time.[5]

Much of that pattern follows health rather than age. Among 178 older Irish adults, gut microbes grouped by where people lived. Those in long-stay care had markedly less diverse communities than people living at home, diet split people along the same lines, and losing the home-dwellers' type of microbiota went with greater frailty.[6] Among 2,081 Swedish women all aged 75–80, frailer women had less diverse, less capable gut communities, even after accounting for statins, acid-suppressing drugs and metformin, and most of the frailty-linked species showed the same pattern in 1,448 older adults in China.[7] Because these women were all about the same age, the differences point to health rather than birthdays. Still, a single snapshot cannot show whether the microbes change first or the frailty does.

The strongest human evidence is prospective rather than cross-sectional, though still observational. In 7,211 Finnish adults followed for 15 years, a gut community rich in Enterobacteriaceae (a family that includes many potential pathobionts) went with a higher risk of death: about 14% higher for each step up a composite microbiome score, and about a third higher in people with the most Enterobacteriaceae than in those with the least. Diet, recent antibiotic use and the usual risk factors did not explain the link away, though the authors say trials are still needed to show cause.[8] In 256 community-dwelling American men aged 85 and over, a microbiome that kept drifting away from the common Bacteroides-dominated pattern predicted better survival over four years, a link that did not show up when younger men were included.[9]

The central mechanism: butyrate, the barrier, and the leak

Moderate. Well characterised in cell and animal work; the human evidence is correlational.

The reason this matters runs through one molecule and one wall. When commensal bacteria ferment dietary fibre they produce short-chain fatty acids — chiefly acetate, propionate and butyrate. Butyrate is the main fuel of the cells lining the colon. In germ-free mice, which have no gut bacteria to make it, colon cells fall short on energy, and adding butyrate rescues them.[10] In cultured human intestinal cells, butyrate also helps assemble the tight-junction proteins that seal the gut wall.[11] As the butyrate-producing species decline with age, the lining is starved of its fuel, the seal weakens, and the barrier becomes permeable — the "leaky gut" state. That sequence is best documented in animals. In people, a direct test of gut permeability in 48 healthy 65–75-year-olds and 52 healthy 18–40-year-olds found the barrier working about as well in the older group, including in the colon, where butyrate does its work. Only among people with irritable bowel syndrome (IBS), 21 older and 27 younger, did the older patients leak more, and then in the small intestine rather than the colon, mainly in the diarrhoea-predominant type. The authors found no sign that ageing on its own weakens the barrier, though they noted that medicines and other illnesses may still affect it.[12]

Once the barrier leaks, bacterial products can get into the blood and switch on inflammation. The clearest demonstration is in mice: as they age, their gut wall becomes leakier and a bacterial cell-wall fragment shows up in their blood alongside a rise in inflammatory signals, while germ-free mice show neither, and housing germ-free mice with old conventional mice is enough to raise those signals.[13] Leaked bacterial DNA matters as well as cell-wall fragments: in ageing mice, gut bacterial DNA travels in the blood and builds up in heart-muscle cells, where it inflames the heart through cGAS-STING, the cell's sensor for stray DNA; bacterial DNA is also enriched in the heart-muscle cells of older people, though that human finding is descriptive only.[14] The same sensor is tripped by DNA escaping from damaged mitochondria in the brains of old mice, which ties mitochondrial dysfunction into the same inflammatory loop,[15] and its response to DNA from damaged nuclei is covered under genomic instability. In people, the leak-and-inflame link shows up mainly in those already frail or ill. A review of 11 studies found that frail older adults had higher blood levels of zonulin, a proposed marker of a leaky gut, and of inflammatory signals than healthy peers.[16] Even in illness the markers are shaky: a 2024 meta-analysis of 10 studies in older people with Parkinson's, Alzheimer's and related conditions found no difference in stool zonulin in Parkinson's; another stool leak marker, alpha-1-antitrypsin, was higher in Parkinson's, but the authors judged the evidence for it too limited to rely on yet.[17] So the gut's contribution to inflammaging is well shown in mice and plausible, but unproven, in healthy older people. The same leak-and-inflame logic operates from the mouth — periodontal disease lets oral pathobionts seed systemic inflammation — which is part of why oral health tracks with longevity.

The gut talks to the rest of the body

Moderate as mechanism; the gut–brain reversal below is animal and observational.

The microbiome is wired into the body through several signalling "axes," and each frays with dysbiosis.

Gut–brain. In a healthy state, commensal bacteria make tryptophan-derived molecules and short-chain fatty acids that reach the brain and restrain activation of its resident immune cells, keeping neuroinflammation in check. But the relationship depends on context, and this is the one place on the page where the intuitive story may reverse, so far only in mice. In germ-free mice engineered to overproduce tau and to carry the human ApoE4 risk variant, adding short-chain fatty acids to drinking water increased brain inflammation and the tangle-forming form of tau in older animals.[18] In a separate mouse model that builds amyloid plaques, supplementation raised plaque load in germ-free mice to the level of normally colonised mice, and raised it further still in normally colonised ones.[19] In germ-free mice that overproduce alpha-synuclein, the protein that clumps in Parkinson's disease, feeding these bacterial products promoted brain inflammation and worsened movement problems.[20] People with Parkinson's disease look different. Pooled across 9 studies of 485 patients and 338 healthy controls, patients' stool held less acetate, propionate and butyrate, while blood levels of propionate, butyrate and isobutyrate ran modestly higher.[21] In one study of 95 patients and 33 controls, lower stool levels went with more severe disease, and so did higher blood levels, which the authors read as a leakier gut wall letting the acids through rather than the gut making too much.[22] The one trial so far gave 72 people with Parkinson's a prebiotic, short-chain fatty acids or both for 6 months; the prebiotic alone or combined with the acids eased constipation, but the trial had no placebo group and did not test effects on the disease itself.[23] So the warning that more short-chain fatty acid can do harm comes from germ-free and genetically engineered mice; in people with Parkinson's, the evidence so far points to too little made in the gut plus a leaky barrier, not too much.

On the helpful side, human trials of feeding gut bacteria hint at a small payoff for thinking. In 36 pairs of twins aged 60 and over, 12 weeks of an inulin and fructo-oligosaccharide prebiotic did not improve the trial's main muscle outcome, but it did improve a combined memory-and-thinking score compared with placebo, driven mainly by fewer errors on a visual learning test; the authors could not say whether the gain matters in daily life.[24] A review of 14 trials in adults aged 60 and over, mostly of probiotics, pooled the 6 that used the 30-point Mini-Mental State Examination and found a gain of just over 1 point, with low-to-moderate certainty, results that varied widely between trials, and benefits seen more consistently in people without dementia.[25] A pooling of 20 probiotic trials in older adults leaned the same way but found no statistically clear effect.[26]

Gut–muscle. In 88 Taiwanese adults over 65, those with low muscle mass had less butyrate in their stool, and the amount tracked with their muscle mass.[27] In 823 Japanese adults aged 60 and over who started with normal grip strength and were followed for about eight years, those who ate more short-chain fatty acids were less likely to develop weak grip: about 23% lower odds for every extra 268 milligrams a day.[28] The step from there to muscle growth signalling and mitochondrial building is so far mouse work: a short-chain fatty acid mix given to fast-ageing sarcopenic mice increased muscle mass and grip strength by switching on the mTOR growth pathway.[29] Whether feeding gut bacteria builds muscle in people is not yet shown. In the twin trial described above, where everyone also did resistance exercise and took a protein supplement, the prebiotic changed the gut bacteria but did not improve chair-rise time, grip strength or a standard physical performance score.[30] Pooling four trials of probiotics in older adults found walking speed about 0.1 metres per second faster, while an apparent grip-strength gain did not hold up in a sensitivity analysis; evidence for prebiotics and synbiotics is still thin.[31] The training side of that story lives under resistance training.

Gut–circadian. The microbiome and the body clock form a two-way loop: feeding–fasting cycles and host clock genes shape which microbes thrive hour by hour, while microbial metabolites — short-chain fatty acids, secondary bile acids, indoles — feed back to entrain the clock genes in the liver and gut lining.[32] Shift work, light at night and erratic eating are thought to disrupt this loop.[33] In mice, jet-lag-style schedules wipe out the daily rhythm of gut microbes, and constant light shifts the microbial mix in ways that track with weight gain and insulin resistance; late-night eating also alters microbial rhythms.[34] Mice whose gut-lining clock is switched off develop worse chemically induced colitis.[35] Whether the same chain leads to a leakier gut and more inflammation in people is not yet shown, but it is one more reason consistent, circadian-aligned sleep and meal timing are sensible.

What the longest-lived guts look like

Weak as guidance; Moderate as description. Almost all of it is cross-sectional — it describes long-lived guts, not how they got that way.

Centenarians are a natural experiment in successful aging, and their microbiomes are distinctive, but not in the simple "keep your youthful species" way, and the two main functional studies do not fully agree. A very small Italian study sequenced the gut microbes of three centenarians, five people aged 59 to 75 and one 38-year-old. The centenarians' microbes carried fewer genes for fermenting carbohydrates and making short-chain fatty acids, more genes for breaking down protein, and more pathobionts such as Escherichia, while the microbes of the 59-to-75-year-olds still looked much like those of the young adult.[36] A larger cross-sectional study of 19 Sardinian centenarians, compared with 23 healthy older adults and 17 young adults, agreed that centenarians' microbes are low in genes for degrading carbohydrates, including fibre, and found them often short of the classic health-associated species Faecalibacterium prausnitzii. Unlike the Italian study, though, it found a high capacity for central metabolism, especially glycolysis and fermentation to short-chain fatty acids. The same Sardinian cohort was enriched for taxa including Bifidobacterium adolescentis, Methanobrevibacter smithii and the genus Desulfovibrio.[37]

The most striking centenarian finding is a specialised bile-acid chemistry. Compared with older and young adults, Japanese centenarians carried gut microbes better able to make unusual secondary bile acids. From one centenarian's stool, researchers isolated strains in the family Odoribacteraceae that make isoallolithocholic acid, which in lab tests strongly inhibited Gram-positive, drug-resistant bacteria including Clostridioides difficile, the organism behind severe antibiotic-associated colitis.[38] The authors suggest this may help long-lived guts keep disease-causing opportunists in check, but that protective effect has not been shown in centenarians themselves. The bile-acid side of this is covered further under bile and metabolism.

Larger studies complicate the picture. Among 297 centenarians in southern China, guts looked in some ways younger than those of other older adults: more often dominated by Bacteroides, more even in their mix of species and with fewer potential pathobionts.[39] Across eight long-lived populations, a handful of species, including Methanobrevibacter smithii and Desulfovibrio fairfieldensis, were enriched again and again, and Akkermansia muciniphila in the three largest groups.[40] Yet in the American study of men aged 85 and over described above, it was losing Bacteroides dominance that went with longer survival.[41] The safest reading is that suppressing pathobionts and keeping up fermentation matter more than any one recipe of species, which seems to differ between populations.

What actually helps

Diet — Moderate, and weaker than it is usually reported. Exercise — Weak to Moderate. Probiotics — Weak and strain-specific.

The gut is genuinely responsive to ordinary behaviour, and the best evidence is for whole dietary patterns rather than pills.

Fibre and a plant-rich, Mediterranean pattern are the foundation. Diverse plant fibres are the fuel that SCFA-producing commensals need, and polyphenols — in berries, nuts and olive oil — selectively favour beneficial taxa.[42] The landmark human evidence is the NU-AGE trial, designed explicitly to test whether a whole-diet approach could damp down inflammaging.[43] Its microbiome analysis covered 612 non-frail and pre-frail older adults across five European countries, randomised single-blind to a year of Mediterranean-pattern eating or control.[44]

It is worth stating what that trial did and did not show, because it is routinely reported more strongly than it reads. Greater adherence shifted the gut community toward diet-responsive bacterial groups, and those groups were consistently associated with lower high-sensitivity C-reactive protein and interleukin-17 and with less frailty. But overall microbiome diversity did not significantly increase in any country, in either arm — what higher adherence did was slow the age-related loss of diversity. And the authors state plainly that they could not observe a direct association between dietary adherence and frailty; the frailty link is an inferred, indirect path running through the microbiome. Their own summary of the magnitude is that "the absolute values of the associations were relatively weak," reflecting the small effect of one year of diet on an established microbial community.[45] Real, directionally encouraging, modest — and the same Mediterranean dietary pattern that anchors the rest of the site, with fibre doing much of the work.

Fermented foods are a reasonable addition, but the evidence is thinner than it sounds. In one small trial in healthy adults (18 per arm), working up to about six servings a day of yoghurt, kefir, kombucha and fermented vegetables over ten weeks raised gut-microbiome diversity and lowered a set of inflammatory proteins. But the trial's main pre-specified immune measure did not change, it had no true control group, and the finding has not been directly repeated, in older adults or anyone else.[46] As covered under fermented foods, the food matrix generally matters more than live colony counts.

Exercise shifts the community too. Regular moderate physical activity tends to enrich short-chain-fatty-acid producers; a rise in Akkermansia muciniphila, a mucin-dwelling species linked to a healthy gut barrier, has been seen too, but only in a few small human studies.[47] A 2025 review proposes that exercise may also partly offset the gut disruption of poor sleep, but the idea rests mostly on animal work and observational human studies, and its authors note the field still "relies heavily on correlational findings".[48] Two honest caveats: in human studies the effect falls on composition more than on overall diversity, and it is modest and depends on intensity.[49] Metabolic benefit can also come without much microbiome change: in a 24-week randomised trial in 61 sedentary adults aged 18–25, moderate or vigorous training shifted the gut community only slightly, and none of those shifts tracked the heart and metabolic gains,[50] while a 3-week trial in 30 sedentary adults with overweight or obesity, combining vigorous walking with a large calorie cut, improved insulin sensitivity and cholesterol with no detectable microbiome change at all.[51] The exercise effect is also not cleanly separable from the diet that usually accompanies it.

Faecal microbiota transplantation (moving gut microbes from a healthy donor) is an established treatment for recurrent C. difficile infection, but it has not been shown to help healthy aging in people. One of the first placebo-controlled trials, in 80 community-dwelling adults aged 65 to 84, gives capsules from young, physically active donors after a short antibiotic course and measures safety, everyday functional ability and muscle strength; it has finished enrolling and is not due to finish until September 2027.[52]

Probiotics are narrower than the marketing suggests. Specific strains have shown specific, modest benefits for specific endpoints. Pooled across 155 randomised trials in about 9,900 adults, probiotics lowered LDL ("bad") cholesterol by only about 0.15 mmol/L, and the authors call the size and consistency of the effect uncertain.[53] In trials of adults with an average age of 50 or more, probiotics added about 1.9 kg of grip strength and 0.08 metres per second of walking speed, though most studies carried a high risk of bias,[54] and in another pooled analysis of older adults only the walking-speed gain survived a robustness check.[55] For inflammation and thinking the picture is mixed: a pooled analysis of 20 trials in older adults found no significant effect on either,[56] and a single-strain trial in 66 healthy adults over 70, partly run by the strain's maker, did not significantly lower blood C-reactive protein, though it lowered a stool marker of gut inflammation.[57] Probiotics are also not risk-free: in critically ill patients, the capsule's own strain has been traced into the bloodstream.[58] The durable lever is the diet that feeds the microbes you already have, not a capsule of new ones.

Medicines matter too. In about 1,900 Dutch adults, including people with inflammatory bowel disease or irritable bowel syndrome, acid-suppressing proton-pump inhibitors, metformin, antibiotics and laxatives had the strongest links to an altered gut community of any commonly used drugs.[59] In about 1,800 adults, the 211 taking proton-pump inhibitors had less diverse gut bacteria, with more mouth-derived and potentially harmful species, and the authors judged the effect larger than that of antibiotics across the population.[60] Among 76 hospitalised patients averaging 83 years old, those taking more regular drugs had a less diverse gut.[61] These are snapshot studies that show association, not proof that the drugs caused the changes. Still, avoiding antibiotics you do not need, and asking your doctor whether a long-running acid-suppressing prescription is still necessary, are cheap levers that no supplement replaces.

What this does and doesn't tell you

What it tells you: dysbiosis is a plausible and modifiable contributor to aging. In mice, losing butyrate producers weakens the gut barrier and the leak feeds body-wide inflammation; in people that chain shows up mainly in the frail and ill, and the links to the brain, muscle and metabolism are still mostly associations. The centenarian data, small and partly conflicting, suggests the goal is functional resilience, keeping up fermentation and keeping pathobionts in check, rather than preserving a fixed set of "young" species. And the levers with the best evidence are fibre and a Mediterranean pattern, with exercise, regular sleep, fermented foods and avoiding unneeded antibiotics and acid suppressants as sensible extras, all squarely within reach.

What it doesn't tell you: that a probiotic supplement or a microbiome test will extend your life. Consumer gut tests run well ahead of the evidence. When identical stool reference samples were sent to seven direct-to-consumer services, the differences between companies were as large as the differences between different people's guts, and one company's repeat results on the same sample barely agreed with each other.[62] An international expert consensus concluded there is not enough evidence to recommend routine microbiome testing in clinical practice, and advised against testing companies giving treatment advice or reporting single "dysbiosis" scores.[63] Most probiotic products are likewise ahead of their data: trial benefits are strain- and endpoint-specific, and "precision geronutrition" remains an aspiration rather than a validated clinical service. It also does not tell you that more is better — the gut–brain data shows short-chain fatty acid supplements worsening disease only in mouse models, while people with Parkinson's appear to make too little of them in the gut — nor that the diet evidence is strong: the flagship year-long trial found effects its own authors call relatively weak, and no significant gain in diversity. The honest message is the familiar one: feed your microbiome a diverse, plant-rich diet, move, and keep regular hours. That does more for your gut than anything sold in a bottle.

Further reading

  • Kadyan S, et al. Microbiome-based therapeutics towards healthier aging and longevity. Genome Med 2025.[64]
  • Tseng CH, et al. From dysbiosis to longevity: a narrative review into the gut microbiome's impact on aging. J Biomed Sci 2025.[65]
  • Abahussin HM, et al. Exploring the intricate link between gut microbiota dysbiosis and the aging process: implications for age-related diseases. Gut Pathog 2025.[66]
  • Badal VD, et al. The Gut Microbiome, Aging, and Longevity: A Systematic Review. Nutrients 2020.[67]
  • Sato Y, et al. Novel bile acid biosynthetic pathways are enriched in the microbiome of centenarians. Nature 2021.[68]
  • Rampelli S, et al. Functional metagenomic profiling of intestinal microbiome in extreme ageing. Aging (Albany NY) 2013.[69]
  • Wu L, et al. A Cross-Sectional Study of Compositional and Functional Profiles of Gut Microbiota in Sardinian Centenarians. mSystems 2019.[70]
  • Zheng B, et al. The molecular interplay between the gut microbiome and circadian rhythms: an integrated review. Front Microbiol 2025.[71]
  • Zhang M, et al. Interactions between Gut Microbiota, Host Circadian Rhythms, and Metabolic Diseases. Adv Nutr 2025.[72]
  • Ghosh TS, Rampelli S, Jeffery IB, et al. Mediterranean diet intervention alters the gut microbiome in older people reducing frailty and improving health status: the NU-AGE 1-year dietary intervention across five European countries. Gut 2020 — n=612; read the limitations, the associations are weak.[73]
  • Santoro A, et al. Combating inflammaging through a Mediterranean whole diet approach: the NU-AGE project's conceptual framework and design. Mech Ageing Dev 2014.[74]
  • Zhao B, et al. Exercise as a modulator of gut microbiota for improvement of sleep quality: a narrative review. Front Neurosci 2025.[75]
  • Dziewiecka H, Buttar HS, et al. Physical activity induced alterations of gut microbiota in humans: a systematic review. BMC Sports Sci Med Rehabil 2022.[76]

— § —