59 studies, 7,463 patients: what stroke really does to your gut bacteria

If you have had a stroke, or you are watching your blood pressure and cholesterol numbers and wondering what else you can move, you have probably run into the claim that the gut is involved. It is a genuinely plausible idea, and it now has the largest evidence synthesis it has ever had. A new meta-analysis pooling 59 studies and 7,463 ischemic stroke patients finds that stroke patients' gut bacteria really do look different from healthy controls' in specific, repeatable ways. It also finds that the single number most people reach for, overall microbial diversity, does not budge at all. And it cannot tell you whether any of it came first.

Key takeaways

• Across 59 pooled studies, overall gut microbial diversity was no different between ischemic stroke patients and controls. The popular "stroke means lower gut diversity" shorthand does not survive pooling.
• What did shift was composition: more Enterobacteriaceae, Pseudomonas and Streptococcus; less Lachnospiraceae, Prevotellaceae, Akkermansia, Bacteroides, Coprococcus and Roseburia, several of which are the gut's main butyrate producers.
• Reduced richness showed up only in the subgroups who also developed cognitive impairment or mood problems after their stroke.
• The headline caveat: every pooled study is a snapshot taken after the stroke. Nothing here shows dysbiosis causes stroke, and no gut treatment has been shown to change stroke outcomes.

Ischemic stroke shows taxonomic dysbiosis without a diversity drop

Ischemic stroke patients carry a measurably different mix of gut bacteria than controls, but not a less diverse one. That is the core result of a systematic review and meta-analysis published in Frontiers in Aging Neuroscience (DOI 10.3389/fnagi.2026.1914614) by Jiagui Huang, Yan He and colleagues at The Second People's Hospital of Yibin City in Yibin, China.

The team tested alpha diversity, which is a measure of how many different bacterial types live in one person's gut and how evenly they are spread, using five standard indices: Shannon, Chao1, ACE, Simpson and observed species richness. None of them separated stroke patients from controls, with all P values above 0.05. That is a negative result worth sitting with, because "low gut diversity" is the single most over-used shorthand in consumer microbiome writing.

Composition was a different story. At the phylum level, the broadest tier of bacterial classification, there was again nothing. One rung down, at the family level, stroke patients carried more Enterobacteriaceae (a family that includes E. coli and other opportunists that thrive in an inflamed gut) and less Lachnospiraceae and Prevotellaceae. At the genus level, Pseudomonas and Streptococcus were elevated, while Akkermansia, Bacteroides, Coprococcus and Roseburia were depleted.

Why stroke survivors and people managing cardiovascular risk should care

The depleted genera are not a random assortment. Roseburia, Coprococcus and much of Lachnospiraceae are the gut's principal butyrate producers, meaning they ferment dietary fibre into a short-chain fatty acid that feeds the cells lining the colon and helps hold the gut barrier shut. Akkermansia does related barrier-maintenance work. Losing them while gaining Enterobacteriaceae is the classic signature of a leakier, more inflammatory gut, which is exactly the state that could, in principle, feed systemic inflammation relevant to vascular disease.

The subgroup results are the part most likely to matter to a survivor or a carer. Overall richness did not change, but observed species richness was significantly altered in patients with post-stroke cognitive impairment (95% CI [-0.80, -0.13], P = 0.006), and the Chao1 richness index was altered in those with mood disorders (95% CI [0.01, 0.37], P = 0.04). Post-stroke depression and post-stroke cognitive decline are two of the most common and least well-managed complications of surviving a stroke, and this is a signal that they track with gut ecology. It is not a signal about what to do next.

Pooling 59 studies from 17,281 screened records

The methodological move that makes this worth reading is scale plus screening discipline. The authors followed PRISMA reporting standards, searched PubMed, Embase and the Cochrane Library, and deliberately excluded animal and in vitro work so that only human observational data entered the pool. Two reviewers independently screened and extracted data, and study quality was appraised with the Newcastle-Ottawa Scale, the standard tool for grading observational studies.

From 17,281 initially identified records, 59 studies survived to the final analysis, covering 7,463 patients. Analyses used random-effects models in RevMan 5.3, with heterogeneity quantified by the I-squared statistic and publication bias formally assessed. Random-effects modelling matters here: it assumes the true effect genuinely varies between studies rather than pretending a Chinese hospital cohort and a European one are measuring the same thing, which is the honest assumption for microbiome data.

What a stroke survivor can and cannot act on

Nothing in this paper justifies buying a probiotic to prevent or recover from a stroke. That needs saying plainly, because the taxa named here map neatly onto products already marketed for exactly that, and the mapping is marketing, not evidence. No randomised trial has shown that a probiotic, a prebiotic fibre or a faecal transplant changes stroke risk, stroke severity or post-stroke recovery.

What is reasonable is more modest. Eating enough fermentable fibre to support butyrate production is a well-supported general gut-health move that costs nothing and is safe for most people, and it happens to be the intervention most likely to support the genera this analysis found depleted. If you are recovering from a stroke and struggling with mood or memory, this paper is a legitimate reason to raise those symptoms with your neurologist as their own problem worth treating, rather than something to endure. And if you are on repeated antibiotic courses during a hospital stay, that is worth a conversation with the treating team about necessity, though it is a conversation about antibiotic stewardship, not about your microbiome specifically.

The reverse-causation problem this design cannot escape

Every one of the 59 pooled studies is observational and cross-sectional, which means each is a snapshot of people who have already had a stroke. That leaves the causal arrow pointing in either direction, and the case for it pointing backwards is strong. Acute stroke triggers a sympathetic stress response that slows gut motility. Hospitalisation brings antibiotics, proton-pump inhibitors, tube feeding, sedation and near-total immobility, all within the days when stool samples typically get collected. Every one of those independently shifts the gut toward more Enterobacteriaceae and fewer fibre-fermenters. The pattern this meta-analysis found is a pattern you would expect to see in a stroke ward even if the gut had nothing to do with why anyone was admitted.

This is not a hypothetical objection. An earlier case-control study of stroke and TIA patients in the Journal of the American Heart Association found the same broad shift and explicitly conceded it could not tell whether the changes preceded the stroke or followed it. Against that, mouse work in Nature Medicine has shown that altering intestinal flora genuinely changes ischemic brain injury, transmissibly by faecal transplant, via intestinal T cells. So a causal pathway exists in animals. Whether it is the one operating in these human cohorts is unknown.

Two further limitations are worth flagging. The available version reports the direction of every taxonomic shift but not pooled effect sizes for them, so the magnitude is unquantified. And the paper is currently posted as a provisional accepted version, without the formatted results, discussion, heterogeneity values, funding statement or competing-interests declaration.

What would actually settle this: longitudinal cohorts and trials

The authors' own conclusion is that future multi-omics and large cohort studies are needed to establish causality and identify therapeutic targets. Practically, that means two things. First, prospective cohorts that sample the gut in healthy people and follow them until some have strokes, which is the only observational design that can rule out reverse causation. Second, an intervention trial that changes the gut and measures a stroke outcome, which nobody has run.

A middle path is already being explored: work such as the Stroke Dysbiosis Index has tried to compress these compositional differences into a single score linked to brain injury and prognosis. If that holds up, gut composition could become a prognostic marker, telling clinicians which survivors are at higher risk of a bad course, long before anyone knows whether changing it helps.

Common questions

Does a stroke damage your gut bacteria? This analysis cannot answer that directly, but it is consistent with it. Stroke patients show more Enterobacteriaceae and fewer butyrate-producing bacteria than controls, and because samples were taken after the stroke, the stroke and its treatment are plausible causes of that shift rather than consequences.

Can improving my gut health prevent a stroke? There is no evidence for that yet. No randomised trial has tested whether a gut-directed intervention lowers stroke risk. Established stroke prevention still runs through blood pressure, atrial fibrillation, cholesterol, diabetes control and smoking.

Should I take a probiotic after a stroke? Not on the basis of this paper. It identified which bacteria differ, not whether replacing them helps, and probiotic supplements are not the same organisms in the same doses as the depleted genera. Ask your own clinical team, particularly if you are immunocompromised or tube-fed, where probiotics carry real risks.

Why does it matter that overall diversity did not change? Because it means "low diversity" is the wrong lens here. The gut of a stroke patient is not emptier than a healthy person's, it is differently populated, with beneficial fibre-fermenters swapped for opportunists. That is a more specific and more actionable target than a single diversity score.

References

1. J Huang, H Yang, X Li, Z Cao, X Zeng, Y He (2026). The Gut Microbiota Dysbiosis in Ischemic Stroke: a Systematic Review and META-ANALYSIS of Microbial Diversity and Taxonomic Alterations. Frontiers in Aging Neuroscience. https://doi.org/10.3389/fnagi.2026.1914614
2. Commensal microbiota affects ischemic stroke outcome by regulating intestinal gamma-delta T cells (Nature Medicine, 2016). Nature. https://www.nature.com/articles/nm.4068
3. Geng-Hong Xia, Chao You, Xu-Xuan Gao, et al. (2019). Stroke Dysbiosis Index (SDI) in Gut Microbiome Are Associated With Brain Injury and Prognosis of Stroke. Frontiers in Neurology. https://doi.org/10.3389/fneur.2019.00397