If you have read any popular article about traditional Chinese medicine in the last decade, you have met the gut microbiome. Herbs "rebalance the microbiota." Formulas "reshape the gut ecosystem." The claim has become a reflex — and a review published in Frontiers in Microbiology on October 13, 2026 asks a question almost nobody bothers to ask: how would we know if any of that is actually true?

The authors — Yu Y, Wang J, Yuan Q, Yu L and Zhang C — did not set out to praise or debunk TCM. They set out to build a scoring system. Their target is a fashionable hypothesis: that TCM works by changing gut bacteria, which in turn changes the bile-acid pool, which then acts on host receptors like FXR and TGR5 to improve conditions such as fatty liver disease, obesity, type 2 diabetes, and inflammatory bowel disease. The biology behind each link is real and well-documented. The problem, they argue, is that papers routinely present the links as proven when only one or two of them were ever measured.

The Claim Most Papers Make — and the Part They Skip

The standard study looks like this: researchers give mice a TCM formula, observe that certain bacterial taxa become more abundant and that a disease marker improves, and conclude the herb worked "via modulation of the gut microbiota." The review's bluntest point is that this leap is not justified. As they put it, changes in microbial composition are often interpreted as mechanisms without evidence of altered microbial function, bile-acid profiles, receptor engagement, or clinical benefit. A bacterium growing more numerous is an observation. It is not a mechanism — any more than a crowded restaurant proves the food is good.

Consider what a genuine chain of evidence would require. You would need a chemically defined intervention — a named formula at a specified dose, from an authenticated, fingerprinted batch, because "Qushi Huayu" on one label is not chemically the same product as "Qushi Huayu" on another. You would need to show the microbiota actually changed in function, not just in headcount: measuring enzyme activities like bile salt hydrolase, or the bacterial bai genes that transform bile acids, rather than merely listing genus names. You would need to show the bile-acid profile changed in a relevant compartment — stool alone may say little about what reaches the liver or bloodstream. You would need evidence that a host receptor actually responded. And you would need a disease outcome that disappears when the microbes are removed — for instance, when antibiotics wipe out the gut flora, or when the benefit transfers to another animal via fecal microbiota transplantation.

Most published studies, the authors find, cover one or two of these links and gesture at the rest.

The Evidence Hierarchy: Grades A Through D

To make this concrete, the review proposes a simple evidence grading for intervention-to-pathway claims:

  • Level A: the gold standard — clinical efficacy demonstrated in a well-run human trial.
  • Level B: coordinated preclinical evidence with pathway-specific causal validation, such as blocking or adding back a specific bile-acid transformation to test whether the mechanism truly mediates the effect.
  • Level C: preclinical evidence without that pathway-specific validation — the formula helps animals, but nobody proved how.
  • Level D: microbiota-associated findings without a pathway-interruption test — the familiar "bacteria changed and the disease improved" pattern. Association, in other words.

Then they ran the published literature through this grader. The result is sobering. Across the eligible studies, only one human trial — a 246-participant randomized trial of the formula Qushi Huayu in people with nonalcoholic fatty liver disease — reported both clinical and microbiota-related outcomes. It earned Level A for clinical efficacy, but only Level D for the claim that the benefit actually traveled through the microbiota–bile-acid pathway, because its bile-acid measurements were incomplete. For obesity, type 2 diabetes, and IBD, the evidence was predominantly preclinical or observational.

The animal studies fared poorly on quality too. Assessed with the standard SYRCLE risk-of-bias tool, most were rated at "unclear risk" — randomization, allocation concealment, blinding, attrition, and even formula characterization were reported too thinly to judge. Uncertainty about the method, not just the result, is the norm in this literature.

Why This Review Is Useful — Even for Skeptics

It is tempting to read this as a debunking. It is not. The authors are careful to separate three things that other papers blend together: what TCM does directly to host tissues, what it does through the microbes, and what happens alongside treatment but is caused by something else entirely — diet changes, weight loss, concurrent medications, or simply getting better care.

That last point matters. In obesity and diabetes research, diet, exercise, metformin, and weight loss all reshape the microbiota and the bile-acid pool on their own. A study that attributes every microbial change to the herbal formula, without tracking what participants ate and which drugs they took, is assigning credit that may belong elsewhere. In inflammatory bowel disease the problem is even sharper: inflammation itself, surgery, and standard biologics all alter bile-acid absorption, so a post-treatment stool sample may reflect the disease getting better rather than the mechanism working.

Their proposed "minimum translational package" reads like a checklist for serious research: authenticated and chemically fingerprinted material; diet and co-medication records; function-resolved microbiome analysis plus targeted bile-acid measurements; a nominated host pathway; prespecified disease outcomes; and prospective safety monitoring, including herb–drug interaction checks. And the bar for claims is stated plainly: TCM should be considered a potential precision adjunct to standard care only when reproducible microbial-metabolic phenotypes are linked to clinically meaningful benefit. Not instead of standard care. Only alongside it, and only when the mechanism is pinned down.

What Readers Can Take From This

This review is not a treatment guide, and it offers no clinical recommendations — nor should it be read as one. Its value for a general reader is as a lens. The next time you encounter the sentence "this herb works by regulating the gut microbiota," you now have the questions to ask in return: which intervention, exactly, and was it chemically defined? Did the study measure microbial function or just count bacteria? Did it test whether removing the microbes erased the benefit? Was there a human outcome that actually mattered?

One trial does not prove a mechanism, and one changed bacterium does not explain a cured disease. The gut microbiome story in TCM is not false — the biology of the microbiota–bile-acid axis is some of the best-established in medicine. But an interesting biology and a proven mechanism are different things, and this review is, above all, an argument for keeping that distinction visible. The field grows up not when claims get bolder, but when they get more specific.

References

  1. Yu Y, Wang J, Yuan Q, Yu L and Zhang C. Evaluating traditional Chinese medicine modulation of the gut microbiota–bile acid axis in metabolic and inflammatory diseases: mechanistic evidence and translational priorities. Front Microbiol. 2026;17:1977691. doi:10.3389/fmicb.2026.1977691.
  2. Liu Q, et al. Qushi Huayu randomized clinical trial in adults with NAFLD. 2024 (cited in the review).
  3. Collins SL, Stine JG, Bisanz JE, Okafor CD, Patterson AD. Bile acids and the gut microbiota: metabolic interactions and impacts on disease. Nat Rev Microbiol. 2023;21:236–247. doi:10.1038/s41579-022-00805-x.
  4. Guzior DV, Quinn RA. Review: microbial transformations of human bile acids. Microbiome. 2021;9:140. doi:10.1186/s40168-021-01101-1.