The plot is pretty straightforward. A lean, disciplined little bacterium lives in your mucus layer. It is scarce in people with obesity, scarce in people with type 2 diabetes, scarce in people with inflammatory bowel disease. Feed it, supplement it, restore it, and metabolic order returns. Roll credits.

In today’s newsletter, I want to complicate that story, because the complication is where the clinical usefulness lies.

To do this, two things have to be squared. First, a trial published last year found that this bacterium produced real, measurable metabolic improvements in people with type 2 diabetes, but only in those who didn’t already have much of it. Give the identical capsule to someone whose gut was already well stocked, and nothing happened. Nothing at all.

The second is that this same organism, so consistently helpful in metabolic disease, shows up enriched in the stool of people with Parkinson’s disease. Not depleted but enriched, across four continents, alongside a collapse of the bacteria that make butyrate.

Both observations appear contradictory, but neither cancels the other. Together they tell you that we have been asking the wrong question about this microbe entirely.

What This Organism Actually Does For A Living

Akkermansia muciniphila makes up roughly one to five percent of the gut bacteria in a healthy adult. It does not float around in the middle of your colon with the fibre eaters. It lives pressed up against your gut wall, inside the slippery mucus layer that coats it, and it makes its living by eating that mucus.

Read that again, because it is the most important fact in this article. The organism eats your gut barrier.

Under normal conditions this is not a problem, instead, it is considered housekeeping. Your gut wall secretes fresh mucus continuously, old mucus has to be cleared, and A. muciniphila does that clearing while releasing two useful by-products, acetate and propionate. Those are then picked up by other bacteria, particularly Faecalibacterium prausnitzii and Roseburia intestinalis, which convert them into butyrate, the fuel your colon cells run on. This is why A. muciniphila gets called a keystone species rather than simply a good one. It is because much of its value is second-hand: it feeds the neighbours.

But a bacterium who takes your gut walls apart is useful only for as long as the demolition rate matches the rebuild rate. Change what is on the menu and the arrangement changes character. Hold that thought.

Four Trials, One Pattern

2019, Belgium, 32 people. The study that launched the field. Adults who were overweight and insulin resistant, meaning their bodies still made plenty of insulin but had stopped listening to it, took live A. muciniphila, heat-killed A. muciniphila, or a dummy capsule daily for three months. Nobody, patient or researcher, knew who got what.

The heat-killed group improved their insulin sensitivity, meaning how well their cells responded when insulin told them to pull sugar out of the bloodstream, by about 29 percent. Circulating insulin fell by about 34 percent, which in practice means the pancreas no longer had to shout as loudly to get the same job done. Total cholesterol dropped about 9 percent. Weight, fat mass and hip measurements moved in the right direction but not far enough to rule out chance in a group that small. The live capsule did not clearly separate from placebo.

For six years that was the entire human story, and a great deal was built on top of it.

2025, China, 58 people. A team in Shanghai gave live A. muciniphila or placebo for twelve weeks to adults with type 2 diabetes who were taking no diabetes medication at all, which matters because metformin raises this bacterium on its own and muddies every study that includes it.

The headline result was a disappointment. Both groups lost weight, both improved their HbA1c, no meaningful difference between them. On a conventional reading, it failed.

Then the researchers split the participants by how much A. muciniphila each person had before starting, and the picture reversed. Among those who began low, the supplement took hold and multiplied. Weight fell, fat mass and visceral fat fell, HbA1c came down by more than half a percentage point, LDL cholesterol dropped, diastolic blood pressure fell by around five points. None of that happened on placebo. In a sealed metabolic chamber, this group shifted measurably toward burning fat rather than carbohydrate. Among those who began high, the supplement barely colonized, and not one of those improvements appeared.

They then proved the mechanism rather than merely arguing it. Stool from a low donor and a high donor went into germ-free mice, and both groups got the same supplement. Only the low-donor mice took up the strain, and only they improved, with a thicker mucus layer, tighter seals between gut cells, and lower inflammatory markers.

Two more trials were published this year, and together they turn the finding above into something you can act on.

In the rest of today’s letter for paid members, we’ll explore:

• Why the largest trial of Akkermansia to date missed its primary endpoint

• Live or heat-killed A. muciniphila probiotic: one of them won, and it was not the obvious one

• Why this “beneficial” microbe is elevated in Parkinson’s disease

• How to read an Akkermansia result on a stool test, who should supplement and who shouldn’t.

If you genuinely enjoy learning more about practical ways gut health and biology impact overall well-being, you should consider upgrading to a paid subscription. I go beyond sharing tips. I do deep dives, share what works and what doesn’t, and highlight what is solidly backed by evidence, what isn’t, and the in-betweens. A paid subscription costs the price of one cup of Starbucks coffee per month. I will do everything possible to give you the best value.

You don’t need a clinical background to belong here. Clinicians, health enthusiasts, and anyone simply trying to make better decisions about their own health will find something useful. I write so the science is rigorous but never out of reach.

I really do hope you join our community.

Read more

Before going deeper, it is worth grounding this discussion in a rather surprising origin story: longevity science began, in part, with the gut microbiome. Élie Metchnikoff, the Russian zoologist who coined the term gerontology and later won the Nobel Prize for his discovery of phagocytosis, helped shift the cultural understanding of aging, by observing how gut microbes influence it. Rather than viewing aging as an inevitable or divinely ordained decline, he argued that it could be studied, influenced, and potentially treated as a medical condition.

The Concept of Longevity Science Emerged from Gut Biology

In the early 1900s, Élie Metchnikoff noticed something curious about the peasant populations of rural Bulgaria. Despite hard labor, modest means, and little access to medicine, an unusual number of them lived well into their 80s, 90s, and beyond. What set them apart, he observed, was that they consumed fermented sour milk daily, and in generous quantities. Metchnikoff eventually isolated the organism responsible for the fermentation, a lactic acid bacterium he called the Bulgarian bacillus, known today as Lactobacillus delbrueckii subsp. Bulgaricus.

His interpretation was remarkably ahead of its time. Metchnikoff proposed that aging was accelerated by toxic compounds produced by putrefactive bacteria in the colon, a process he called intestinal autointoxication, and that regularly seeding the gut with lactic acid bacteria could crowd out these harmful microbes and slow the decline. He was so convinced that he drank sour milk daily for the rest of his life. Admittedly, some of his mechanistic details did not survive scrutiny, but the central intuition did: the microbial residents of our gut are active players in how we age, and they can be deliberately shaped to our benefit. More than a century later, with sequencing tools Metchnikoff could never have imagined, we are essentially confirming, refining, and extending the hypothesis he built from a bowl of Bulgarian yogurt.

Five Ways the Gut Microbiome Shapes the Way You Age

When researchers study centenarians, people who have lived past 100, they keep finding something similar. These individuals carry gut microbiomes that look remarkably youthful. They are usually diverse, stable, unique, rich in beneficial species, and humming with metabolic activity. Why is this likely the case, you may ask? Well, it is because the gut microbiome sits at the intersection of nearly every system that determines how we age — immunity, metabolism, the gut barrier, and even the brain.

In this article, I delve into five key ways your gut microbiome influences how long, and how well, you live. This does not in any way suggest that the microbiome alone is what influences aging, rather, my goal is to illuminate the biology and the mechanism with which our gut health modulates the process of aging. It also puts specificity to how each lever of our gut biology can be optimized to enhance healthy aging.

Read more

I was 9 years old when Ginika passed but I was well aware of the cause of her demise. I understood what typhoid fever meant, as it was quite endemic in Nigeria: almost every person would contract the disease at least once in their lifetime and I was no exception. Its symptoms were telling and terrible: prolonged high fever, severe headache, extreme fatigue, altered bowel movements and gastrointestinal distress. It was so familiar that lay people would attempt to treat it without testing, and quite often, they guessed right. Typhoid fever is a life-threatening illness caused by the bacterium Salmonella enterica serovar Typhi (S. Typhi). It typically spreads through food or water contaminated with the feces or urine of an infected person or a chronic carrier.

One major reason why a bacterium like S. Typhi would develop resistance is the indiscriminate use of antibiotics. Continued drug use forces the bugs to develop intelligent strategies to circumvent the antibiotic. Typically, they do this by modifying parts of their genetic make-up, conferring a resistance phenotype to them.

While researchers were focused on the drug-driven mutations within bacterial cell genomes, antibiotics were busy wreaking greater havoc on the larger community of microbes in the gut: the microbiota. Antibiotics are well-known to have off-target effects, meaning that if they were intended to kill bacteria A, they end up wiping out bacteria B, C, D and E as well, all of which may be beneficial or health-promoting species. When this collateral damage happens at scale, the gut microbiome is reshaped into a state that can no longer support optimal health. And the result is the emergence and exacerbation of various chronic diseases. This underscores why chronic diseases like obesity, asthma, type 2 diabetes, inflammatory bowel disease and even colon cancer have all been associated with previous antibiotic exposure.

So how do we increase the chances of selective targeting of pathogens without collateral damage? That is the primary purpose of today’s newsletter. It is to explore how to target problematic bacterium A, while sparing the beneficial bacteria B, C, D and E. But to explore this, we must first address two key questions:

1) How do bacterial cells behave when they must compete for nutrients, colonization niches and limited resources? This matters because bacterial strains living with other bugs within the same environment, quite often, use antibacterial compounds that they synthesize, to selectively target their competitors. Usually, the producing bacterium and its kin are immune to the compounds that they make, implying that these antibacterial molecules were evolutionarily designed to protect the bacterium and its lineage from extinction due to competition. Which raises the follow-up: can this lethal weaponry be mapped to a gene or set of genes in the organism?

2) If bacteria indeed synthesize these antimicrobial molecules against competitors, how do they ensure that a narrow spectrum of killing is maintained without any off-target effects? This is the key feature that an antimicrobial must possess to distinguish its activity from that of a broad-spectrum antibiotic.

In the following paragraphs, I address two killer molecular nanomachines that bacterial cells have employed for billions of years to take out their rivals, even when they are hiding in a teeming crowd of other microbes. Given the precision with which these two lethal weapons selectively target bacteria down to strain level, they present themselves as the future anti-bacterial tools that can be correctly applied to complex biological ecosystems like the gut microbiome, without having to nuke other beneficial microbial residents in the gut.

Bacteria Have Been Killing Each Other Precisely for Billions of Years.

Long before the first antibiotic reached a patient, medicine had stumbled on something in nature that killed bacteria with startling precision. It just could not see what it was.

In the summer of 1915, severe dysentery tore through French troops garrisoned at Maisons-Laffitte, outside Paris. A scientist named Félix d’Hérelle, working at the Pasteur Institute, noticed something odd in the feces of the soldiers who were surviving the disease. He observed that when he filtered their feces and dripped the clear liquid onto a culture of Shigella dysenteriae grown on a lab plate, the bacteria growing on the plate died. Whatever was doing the killing was invisible under his microscope, slipped through filters fine enough to trap bacteria, and multiplied only when the bacteria causing the dysentery were there to feed on. In 1917, he published the observation and gave the invisible agent a name: bacteriophage (phage, for short), literally, eater of bacteria.

Phages are viruses, but they infect only bacteria. They cannot enter a human cell. So where do they come from? Well, they are stretches of DNA written into the chromosomes of bacteria. So when you read through the chromosome of a bacterium, you are likely to find phage genes in there. Not existing independently, but written in the genome of the host bacteria. Bacteria have evolutionarily harbored phages in their chromosomes as part of their permanent armory. They are quietly passed from one generation to the next, until stress, chemical signaling or competition wakes them. When they wake, the host cell carrying them sacrifices itself and releases a burst of phage particles that destroy neighboring competitors, while its own kin, carrying the same instructions, are immune. So, bacteria have been settling territorial disputes this way for billions of years and our gut has been the battlefield.

A phage has a head made of a protein shell. Inside that shell is its own genetic material, its DNA. The head, through a tube, is joined to a tail that works like a syringe. For a phage to target a bacterial cell, its tail must first recognize a specific receptor on the surface of its target. It then docks onto it, injects its DNA, hijacks the bacterial cell’s machinery to build hundreds of copies of itself, and bursts the cell open on the way out. As you can probably tell now, phages are precise. They do not need to insert their full selves into the bacteria. They only need to inject their DNA into the bacteria, then while inside their host, they express their own genes to make a head, tail and other essential proteins to create a full phage, all with the help of the host bacterium’s replicative machinery.

Really cool. But I want you to pay attention to the stage where a phage initially recognizes the surface of the bacterium to infect, because that is the main point you should not miss. A phage that reads the surface of Salmonella enterica serovar Typhi will drift past Escherichia coli and past every beneficial resident of the gut without touching them. This is killing at the strain level, not the phylum level, and it is precisely the property that broad-spectrum antibiotics lack. That is the level of precision you want to see when you are interested in targeting one single strain amidst a plethora of bugs.

In 1919, d’Hérelle treated his first patient, a boy with dysentery, in Paris, and by the Second World War the Red Army was using phage preparations against dysentery and infected wounds, with mass production improvised in a Stalingrad basement in 1942.

Then penicillin arrived, and the West largely forgot about phages.

To be fair, the forgetting was not entirely unjustified. A phage is alive in the ways that matter clinically. It replicates unpredictably, so dosing can be tricky. Also, there was worry that bacteria could evolve altered surfaces and escape it after repeated usage. Worse, phages that integrate into the bacterial chromosome can ferry genes between bacteria, including toxin and resistance genes, which is exactly how Escherichia coli O157:H7 acquired the Shiga toxin that makes it lethal.

This presents a new challenge. Yes, phages can reliably bypass the off-target limitation known with antibiotics. They can selectively target bacteria using their syringe-like tails, but some of them can transfer lethal bacterial genes from one host to the other. The question then becomes: can we just keep the syringe-like tail to selectively kill pathogenic bacteria and discard the phage genome that could introduce unwanted features into the target organism?

Keeping the Syringe and Discarding the Genome

It turns out that bacteria have more than phages in their arsenal. They also have entities resembling phages, more like phage cousins. However, these cousins have unique features. Their tails look the same as phages, but they are headless and have no genetic material. And because they look like phage tails, they are called tailocins. Whether tailocins existed before phages, we do not know, but like phages, they are also written permanently within the chromosomes of several bacteria. Tailocins have a narrow killing spectrum against strains of the same species. And the conditions that trigger a bacterial cell to induce its phage genes are the same ones required for tailocin induction: stress, chemical signaling and competition.

My doctoral research at the University of Toronto was on tailocins. During this time, I induced different tailocin types from multiple bacterial strains (see Saha et al., Journal of Bacteriology, 2023; reference 8 below). If I am being conservative, I’d say I induced tailocins over a thousand times, so I do have first-hand experience working with them. Like phages, they use special binding proteins on their tails to attach to a compatible receptor, such as the LPS, on the target bacterial surface. Once they dock, they insert their core into the bacterial cell, disrupting its membrane, thereby causing ions to leak out of the cell. This compromises important cellular processes such as the ability of the bacteria to make essential proteins, so they eventually die. This attribute is what separates tailocins from phages, whose primary purpose of infection is to insert and replicate their DNA within the bacterial cell.

One key feature that makes tailocins attractive for targeted therapy is their malleability to genetic engineering. One of the successes I had in the lab was to engineer them to kill both clinical and environmental strains of Pseudomonas aeruginosa growing on a lab plate, while sparing other non-target bacterial strains. What this implies is that in clinical settings, tailocins can be genetically designed to target a pathogenic strain, e.g. E. coli O157:H7, while bypassing its beneficial counterpart E. coli Nissle 1917. Same tailocin, different activity toward strains of the same species. That discrimination is exactly what antibiotics have never managed.

The use of tailocins has been extended beyond testing them on organisms growing on lab plates. They have been used to treat Clostridioides difficile, Pseudomonas aeruginosa and E. coli infections in animal models, all with surgical precision and ultra-specificity. These pre-clinical successes give immense hope about what the future of targeted therapy can be. Mice were used in the Clostridioides difficile and Pseudomonas aeruginosa studies; the E. coli O157:H7 work used infant rabbits.

While we wait on this future, key questions remain to be investigated: i) Can we reproduce in humans the tailocin killing efficacy seen in mice? ii) Do tailocins trigger severe immune responses or cytokine release? iii) How will tailocins behave in the bloodstream? iv) Do they clear too fast via the immune system? v) Can they penetrate dense biofilms? vi) How fast do target bacteria evolve resistance to them? All very exciting questions, which make the future of targeted therapy more exciting to anticipate.

What Ginika Did Not Have

Every patient who must deal with antibiotic resistance is fighting on two fronts. One is the pathogen’s resistance to the drug itself. And the second is the annihilation of the gut microbiome diversity due to antibiotic use. That is the double cost of antibiotic resistance, and it is why I find tailocins so compelling and promising. They are ancient, they are already written into the chromosomes of bacteria living in your colon right now, and they have been settling territorial disputes with strain-level accuracy since long before the first human swallowed the first sulfa drug. What they offer is precisely the thing Ginika never had: a way to remove one organism from a crowded ecosystem and leave the crowd intact.

We are not there yet. The questions listed above are real and unanswered, and I would rather say so plainly than oversell a preclinical result. But the direction of travel matters. If the next generation of antibacterials can be aimed rather than sprayed, the gut microbiome stops being collateral damage in the fight against infection and starts being an ally worth protecting. To me, this is the difference between treating an infection and treating a person.

Ginika would have turned 46 this year. I would like the medicine that failed her to be unrecognizable by the time it reaches my own children.

References

1. GBD 2021 Antimicrobial Resistance Collaborators. Global burden of bacterial antimicrobial resistance 1990-2021: a systematic analysis with forecasts to 2050. Lancet. 2024;404(10459):1199-1226. doi:10.1016/S0140-6736(24)01867-1

2. d’Herelle F. Sur un microbe invisible antagoniste des bacilles dysenteriques. C R Acad Sci Paris. 1917;165:373-375. English translation: On an invisible microbe antagonistic toward dysenteric bacilli. Res Microbiol. 2007;158(7):553-554. doi:10.1016/j.resmic.2007.07.005

3. Summers WC. Felix d’Herelle and the Origins of Molecular Biology. New Haven, CT: Yale University Press; 1999. (Source for the Maisons-Laffitte outbreak and the 1919 first patient.)

4. Chanishvili N. Phage therapy: history from Twort and d’Herelle through Soviet experience to current approaches. Adv Virus Res. 2012;83:3-40. doi:10.1016/B978-0-12-394438-2.00001-3 (Source for Red Army use and wartime production.)

5. Gebhart D, Lok S, Clare S, et al. A modified R-type bacteriocin specifically targeting Clostridium difficile prevents colonization of mice without affecting gut microbiota diversity. mBio. 2015;6(2):e02368-14. doi:10.1128/mBio.02368-14

6. Scholl D, Martin DW Jr. Antibacterial efficacy of R-type pyocins towards Pseudomonas aeruginosa in a murine peritonitis model. Antimicrob Agents Chemother. 2008;52(5):1647-1652. doi:10.1128/AAC.01479-07

7. Ritchie JM, Greenwich JL, Davis BM, et al. An Escherichia coli O157-specific engineered pyocin prevents and ameliorates infection by E. coli O157:H7 in an animal model of diarrheal disease. Antimicrob Agents Chemother. 2011;55(12):5469-5474. doi:10.1128/AAC.05031-11 (infant rabbit model)

8. Saha S, Ojobor CD, Li ASC, Mackinnon E, North OI, Bondy-Denomy J, Lam JS, Ensminger AW, Maxwell KL, Davidson AR. F-type pyocins are diverse noncontractile phage tail-like weapons for killing Pseudomonas aeruginosa. J Bacteriol. 2023;205(6):e00029-23. doi:10.1128/jb.00029-23

I do not see patients (I am PhD, not MD). I spend some of my days training clinicians, teaching them to read the microbiome and turn it into better care. But my cousin’s call is exactly the kind of story I carry into every training room, because it shows how much of a woman’s hormonal life is decided in a place no one is taught to look. So let me make each of those places obvious, one at a time.

The gut decides how much estrogen a woman actually keeps

We talk about a woman’s hormones as though they are made in the ovaries and the story ends there. It does not. After estrogen has done its work, the liver tags it for disposal. Then a community of gut bacteria carrying an enzyme called beta-glucuronidase snips that tag off and sets the estrogen free again, back into circulation. This bacterial workforce is the estrobolome, and it is the most direct line that exists between the microbiome and a woman’s hormones. When it is overactive, too much estrogen is reclaimed, tilting a younger woman toward the estrogen-dominant picture behind some endometriosis and heavy, painful cycles. When it fades, as it often does across the transition, a woman loses even the estrogen she has left, and she suffers more than her bloodwork predicts. This is why two women with identical ovaries can have entirely different menopauses.

Whether soy helps a woman depends on her bacteria, not her diet

In a training session last year, a physician described a patient who had done everything the magazines advised. Soy milk, edamame, tofu three nights a week, and no relief at all. The patient assumed her body was broken. But it was not. The benefit of soy for hot flushes does not come from soy. It comes from a metabolite called equol, and equol is made only by women who carry the right bacteria species like Slackia isoflavoniconvertens and Adlercreutzia equolifaciens. Only a third to a half of women host them. The rest can eat soy by the bowlful and make almost none. Once a clinician can tell a producer from a non-producer, she stops asking a woman’s diet to do something her microbiome cannot, and that single result reshapes the advice she gives.

Read more

In my talk, I addressed key points:

1. The first 1,000 days and why early colonization of microbes dictate lifetime trajectory of babies.

2. Characterization of a healthy pediatric microbiome with respect to microbial succession and microbiome maturation.

3. Recognizing pediatric dysbiosis signatures through the lens of two disease models: Allergic diseases and Necrotizing Enterocolitis.

4. The key modulators of the early-life microbiome changes.

5. Targeted interventions: Researched prebiotics, probiotics and FMTs, for correcting pediatric dysbiosis and improving health outcomes.

6. I responded to questions asked by participants in the last 10 mins of the presentation.

This is a one-hour long talk, hosted on CAHN-Pro’s Youtube page. You can find it here.

Note: The video length is 1 hr; not 1 hr. 30 mins. The organizers inadvertently left the recoding on for another 30 mins after my talk.

Enjoy and let me know the part of the presentation you find most interesting.

Happy listening!

Most clinicians know the IBS loop without having to think about it. A patient presents with a symptom flare. You reach for the empiric tools that work for the average patient: low-FODMAP for most, a fiber supplement, an off-the-shelf probiotic, an antispasmodic. The patient improves, partially or temporarily. Then they cycle back. You adjust, they improve again, they recur. IBS affects roughly 4 to 11 percent of the population and dominates GI referral volume, and for a large share of those patients the defining feature of their care is not the diagnosis. Response is real but shallow, and recurrence is the rule rather than the exception.

That loop is not a failure of clinical effort but the failure of resolution. We are treating a label, and the label was never designed to tell us what to do next.

The label tells you what the patient feels, not why

Rome IV and the IBS-D, IBS-C, IBS-M split are genuinely useful for what they were built to do, which is to organize symptoms and exclude alarm features. What they do not do is predict who responds to which therapy. A symptom subtype describes the output of the system. It says nothing about the engine driving it.

Picture two patients who walk into your clinic with an identical IBS-D presentation. Same Rome IV criteria, same Bristol stool scores, same story of urgency after meals. On paper they are interchangeable. Underneath, one is being driven by a hydrogen sulfide problem and the other by depleted butyrate producers and a leaky barrier. These are not variations on a theme. They are opposite biological drivers that call for opposite interventions. Push fermentable fiber at the first patient and you feed the very process making them worse. Withhold it from the second and you starve the repair you are trying to support. Identical symptoms, entirely opposite biology, and the symptom label cannot tell them apart.

This is the central problem with symptom-based subtyping. It is not that it is wrong. It is that it is silent on the one question that determines whether your next move helps.

IBS is mechanistically diverse, and the microbiome is the layer that explains the “why”

It is more accurate to think of “IBS” as a single label sitting on top of several distinct biological drivers that happen to converge on a similar set of complaints. In the talk I gave to this group, I laid these out as five layers stacked beneath the standard label: methanogen burden, bile acid handling, fermentation capacity, barrier integrity, and histamine signaling. Each one is measurable. Each one produces a different functional fingerprint, whether that is elevated methane, a dysregulated fecal bile acid profile, impaired short-chain fatty acid production, compromised permeability, or activated histamine pathways.

The microbiome is the critical layer because it is where the “why” lives. When you can see which of these engines is running hot in each patient, the symptom label stops being a dead end and starts being a starting point.

What a microbiome test is, and what it is not

Before going mechanism by mechanism, I want to be exact about the claim I am making, because the field is full of overreach and I have no interest in adding to it.

A microbiome test is not a standalone diagnostic. It is not a replacement for clinical judgment, and it is not a replacement for Rome IV. There is no single IBS signature waiting to be discovered, and any test that claims to have found one should be treated with suspicion. I am not positioning these tests as the thing that tells you a patient has IBS.

What they are is adjunctive and hypothesis-generating. They are a stratification tool. They give you high-resolution profiling of the functional layer beneath the symptom, and from that they offer a guide for management refinement. The value is not in a diagnosis. The value is in seeing under the label well enough to form a meaningful hypothesis, and a meaningful hypothesis is what raises your probability of choosing the intervention that actually holds. That is the entire argument. Everything below is an illustration of it.

Mechanism 1: Methanogens and constipation

The first engine is methane. Methanobrevibacter smithii is an archaeon, not a bacterium, and it is the dominant methanogen in the human gut. When it overgrows, breath methane rises, intestinal transit slows and the clinical consequence is constipation. Patients with methane-predominant overgrowth are several times more likely to be constipated than those with a hydrogen-predominant pattern, and the degree of methane production tracks with the severity of constipation. This is the biology behind the classic IBS-C presentation in a meaningful subset of patients.

What does the test change? It flags candidates for a methane-directed workup rather than another reflexive laxative trial, and it sets honest expectations that standard fiber and osmotic agents may underperform in exactly this group. For example, in the randomized work from Ghoshal and colleagues, reducing methane production with rifaximin shortened transit and improved constipation, which is precisely the result you would predict if methane were the driver and not a bystander. You are no longer guessing whether this is a fiber problem or a methanogen problem. You can form the hypothesis and test it.

Evidence: Kim G, et al. Dig Dis Sci. 2012;57:3213-3218. Ghoshal UC, et al. Gut Liver. 2016. Ghoshal UC, et al. (rifaximin and methane). Rezaie A, et al. Am J Gastroenterol. 2017;112:775-784 (IMO definition).

Mechanism 2: Hydrogen sulfide and sulfate-reducing organisms

The second engine sits at the opposite end of the symptom spectrum. Bilophila wadsworthia and Desulfovibrio are sulfate-reducing organisms. They take sulfur compounds, many of them delivered by a diet high in animal protein and fat and reduce them to hydrogen sulfide gas. At low levels hydrogen sulfide is part of normal colonic physiology. At high levels it becomes a problem. It is cytotoxic to the colonic epithelium, it interferes with the colonocyte’s ability to oxidize butyrate for fuel, and it is associated with the kind of urgency and bloating that brings an IBS-D patient back to your office.

This is the patient who looks, on symptom criteria alone, exactly like the depleted-butyrate patient I will describe next, and who would be harmed by the same intervention that helps that patient. Here the test points you toward moderating the dietary sulfur load rather than reflexively adding fermentable fiber. Adding more fermentable substrate to a sulfur-driven, hydrogen-sulfide-rich gut is not a neutral move. It can feed the wrong process. Seeing the sulfate-reducer signature is what stops you from making that mistake, and there is no symptom on the Rome IV checklist that would have warned you.

Mechanism 3: Depleted butyrate producers

The third engine is loss rather than overgrowth. Butyrate is the primary fuel for colonocytes and a central regulator of barrier integrity and mucosal tone. The organisms that produce it, principally Faecalibacterium prausnitzii, Roseburia, and Eubacterium rectale, are among the most consistently informative species in the entire profile. When they are depleted, the barrier loosens. Low-grade inflammation and visceral hypersensitivity follow, and the patient presents with the heightened pain sensitivity typical of IBS-D and IBS-M, often against a background of lower overall diversity.

This is not speculative. Pozuelo and colleagues profiled 113 patients with IBS against 66 healthy controls in 2015 and found that lower microbial diversity in IBS tracked with a lower relative abundance of butyrate-producing bacteria, an effect that was most pronounced in the IBS-D and IBS-M subtypes, with several taxa correlating directly with flatulence and abdominal pain.

The management implication is specific and easy to get wrong. The depleted-butyrate patient argues for a titrated butyrogenic strategy built on resistant starch and specific, well-chosen fibers, introduced carefully. It also explains, mechanistically, why aggressive fermentable fiber backfires in some patients. A gut without the species to ferment that fiber cleanly does not respond to it the way the textbook promises. You are not feeding butyrate producers that are not there. You are feeding gas, distension, and pain. Knowing which patient is sitting in front of you is the difference between a strategy and a gamble.

Evidence: Pozuelo M, et al. Sci Rep. 2015 (113 IBS patients).

Mechanism 4: Fermentation capacity and diet candidacy

The fourth engine reframes the most common intervention in all of IBS care. Low-FODMAP is effective, but it is effective for a particular kind of patient, and we have historically applied it to everyone because we had no way to tell the two kinds apart.

Vervier and colleagues changed that in their 2021 work in Gut. Using metagenomics on IBS cases and household controls, they identified two distinct microbiome profiles at baseline. Roughly half the patients carried what they called the IBS-P, or pathogenic-like, signature: enriched in Firmicutes and in genes for amino acid and carbohydrate metabolism, depleted in Bacteroidetes, and associated with higher symptom severity. The other half carried the IBS-H, or health-like, signature, with a microbiome that closely resembled healthy controls. When both groups went on a low-FODMAP diet, the IBS-H and control microbiomes were essentially unaffected, while the IBS-P signature shifted measurably toward a health-associated profile, with a rise in Bacteroidetes, normalization of metabolic genes and significant alleviation of symptoms. The 2024 volatome work in eBioMedicine extended this, showing that the IBS-P group carried a fermentative, short-chain-fatty-acid-rich metabolic signature that dropped on FODMAP restriction, and that a small panel of metabolites could predict subtype with useful accuracy.

Read those two studies together and the implication is hard to miss. Low-FODMAP is not a permanent diet and it is not for everyone. It is best understood as a time-limited diagnostic phase. In the IBS-P patient it does real work, both on symptoms and on the underlying microbial signature. In the IBS-H patient it offers a weak response while carrying a real risk of depleting an already healthy microbiome, putting a patient through a burdensome, restrictive elimination for little gain. The test helps you decide who actually benefits before you commit them to it. That is management refinement in its purest form: same prescription, but offered to the patient whose biology predicts it will land.

Evidence: Vervier K, et al. Gut. 2021 (IBS-P vs IBS-H). Volatome metabotype study, eBioMedicine. 2024.

Mechanism 5: Histamine-producing organisms and pain flares

The fifth engine is the one that most often hides in plain sight: the patient whose pain flares track with meals. De Palma and colleagues, working across McMaster and Queen’s and publishing in Science Translational Medicine in 2022, identified a specific culprit. Klebsiella aerogenes, the McMaster-Queen strain, carries a histidine decarboxylase variant that lets it convert dietary histidine, an amino acid abundant in animal and plant protein, into histamine. They found this organism in up to 25 percent of gut microbiota samples from patients with IBS.

The downstream cascade is what makes this clinically important. Bacterial histamine activates the histamine-4 receptor, which recruits mast cells into the intestinal wall. Those mast cells release more histamine and additional pain-signaling mediators, producing visceral hypersensitivity. The team demonstrated the mechanism cleanly in germ-free mice colonized with stool from patients with high urinary histamine, and they showed something that closes the loop for the clinician: in their patient cohort, stool histamine was high when patients reported severe pain and low when they were pain-free. They also showed that reducing fermentable carbohydrate intake lowered urinary histamine and visceral hypersensitivity, which connects this mechanism back to the fermentation-capacity story and to the enrichment of histidine-related pathways Vervier saw in the IBS-P group. The biology rhymes across mechanisms more often than the symptom labels would ever suggest.

For management, identifying this signature flags candidates for a histamine-aware approach, whether that is a lower-histidine dietary strategy, a mast cell stabilizer, an antimicrobial targeting K. aerogenes or an antihistamine. It also carries a specific warning that I think is underappreciated in practice: it argues against recommending histamine-producing probiotic strains to exactly this patient. Some commercially common probiotic species are histamine producers. In a histamine-driven patient, the well-intentioned probiotic can worsen the situation.

Evidence: De Palma G, et al. Sci Transl Med. 2022.

What this actually changes in clinic

Step back from the five engines and the pattern is the same in every case. The symptom label puts a heterogeneous group of patients into one box. The microbiome profile tells you, for the individual in front of you, which engine is most likely running the show. From that you form a hypothesis: this looks like a methane problem, this looks like a sulfide problem, this patient is a strong low-FODMAP candidate and that one is not, this patient should not be anywhere near a histamine-producing probiotic. None of those is a diagnosis. Every one of them is a better-informed bet than the empiric default, and better bets, compounded across a practice, are what break the trial-and-error loop.

I want to end where I started, because it matters. I am not asking you to replace your clinical judgment, your Rome IV workup, or your physical exam with a test result. I am asking you to add a layer of resolution that the symptom label was never able to give you. The microbiome test does not tell you that a patient has IBS. It helps you see what is underneath the IBS, and it hands you the hypotheses that make your next decision more likely to be the right one. In a condition defined by recurrence and low durability of response, that shift, from treating the label to reading the biology beneath it, is where better management truly comes from.

This is a closer look at how early colonization sets the trajectory of human health, what a healthy pediatric microbiome actually looks like as it matures, and where clinicians and parents have meaningful room to intervene.

A Non-Negotiable Window of Opportunity

The first 1,000 days, from conception to a child’s third birthday, represent a developmental window that does not reopen. During this period the infant immune system is uniquely permissive, learning in real time how to distinguish dangerous pathogens from harmless environmental exposures. The microbes colonizing the gut are the primary teachers in that classroom.

Just how essential this microbial education is becomes obvious when it is taken away. One of the ways we know this is because germ-free mice, raised entirely without a microbiome, display severe growth stunting, impaired blood-brain barrier formation, and profound immune deficits. The gut microbiota influences the somatotropic axis directly, regulating insulin-like growth factor 1 (IGF-1) and growth hormone production. Without microbial input, the body cannot fully build itself.

When the microbiome is disrupted within these first 1,000 days, immune calibration falters. The downstream consequences show up later as allergies, asthma, obesity, and autoimmune disease, sometimes irreversibly so.

A Choreographed Succession, Not a Static Picture

A healthy pediatric microbiome is not one fixed community of bacteria. It is a dynamic, tightly choreographed succession of ecosystems that matures alongside the child, and each phase has a specific developmental job.

The neonatal phase, roughly the first month of life, begins in a gut that is still oxygen-rich. The earliest colonizers are typically facultative anaerobes such as Enterobacteriaceae, and their first task is to consume that available oxygen, transforming the gut into the strictly anaerobic environment the next wave of microbes requires to thrive.

From around one to six months, in a healthy breastfed infant, strict anaerobes take over. Bifidobacterium and Bacteroides come to dominate the landscape. Historically, and in modern non-industrialized populations, this phase is overwhelmingly dominated by Bifidobacterium infantis, a keystone species uniquely equipped with gene clusters to digest the complex human milk oligosaccharides that an infant cannot break down on their own. The microbiome at this stage is intentionally low in diversity and developmentally immature, and that immaturity is a feature rather than a flaw. It provides critical colonization resistance against pathogens.

What has changed, profoundly, is the cast of organisms occupying that phase in the industrialized world. We are witnessing a generational loss of B. infantis itself. In a comparative study that metagenomically sequenced infant stool across industrialized, transitional, and non-industrialized populations, it was found that industrialized infants are heavily depleted of B. infantis and its HMO-utilization gene cassettes, which have been largely replaced by B. breve, a species with far less capacity to digest breast milk sugars. The downstream signature of this loss is measurable in something as fundamental as infant stool pH, which has risen from roughly 5.0 to 6.5 over the last century. That shift tracks with modern gut dysfunction, diminished antibody responses to vaccines, and an increased risk of atopic and immune-mediated disease. A keystone species that fed on the very sugars evolution placed in breast milk has been quietly disappearing from the infants who need it most.

Between six and thirty-six months, the introduction of solid foods and the eventual end of breastfeeding trigger a structural reorganization of the ecosystem. The microbiome diversifies rapidly. Bifidobacterium dominance recedes, making room for complex carbohydrate-degrading bacteria from the Clostridiales and Firmicutes groups, which produce the short-chain fatty acids, particularly butyrate, that fuel colonocytes and shape immune tone. By around age three, the microbiome has largely stabilized into an adult-like configuration.

When the Sequence Goes Wrong: The Asthma Signal

Dysbiosis in childhood often shows up not as a dramatic overgrowth, but as a quiet absence. Specific immune-modulating bacteria fail to arrive on schedule, and the immune system never receives the signal it was waiting for.

The clearest illustration of this comes from the Canadian Healthy Infant Longitudinal Development (CHILD) cohort. These researchers tracked 319 infants and compared the gut microbiomes at three months and one year of age in those who went on to develop atopy and wheeze by age one against healthy controls. They wanted to know whether dysbiosis precedes asthma, or simply accompanies it.

The findings were striking. Infants at high risk for asthma showed a transient but profound dysbiosis specifically within the first 100 days of life, marked by depletion of four bacterial genera: Faecalibacterium, Lachnospira, Veillonella, and Rothia, collectively abbreviated FLVR. Alongside this taxonomic loss came significantly reduced fecal acetate, a key anti-inflammatory short-chain fatty acid. By one year of age, the differences had largely faded, which is itself the point. The first 100 days operate as a fleeting critical window for immune education, and a deficit during that window leaves a mark even after the ecosystem rebalances. To establish causation, the team inoculated germ-free mice with the four FLVR organisms and found it ameliorated airway inflammation in their offspring.

What Pushes the Sequence Off Course

The modern pediatric microbiome is under near-constant pressure from lifestyle and environmental inputs that can derail normal assembly.

Birth mode is one of the most consequential variables. Cesarean delivery bypasses the birth canal, depriving the infant of maternal vaginal and fecal microbes. C-section infants show delayed colonization by Bacteroides and Bifidobacterium and an overrepresentation of hospital-associated organisms and opportunistic pathogens.

Diet is another master regulator. Exclusively breastfed infants maintain a lower, more stable microbiota age, which is protective against diarrheal dysbiosis and inflammation. Formula introduction prematurely accelerates microbiome maturation, expanding Firmicutes and Proteobacteria before the immune system is ready to interpret those signals correctly.

Early antibiotic exposure is perhaps the most clinically relevant disruptor. A single course of common agents such as amoxicillin or macrolides can produce severe, long-term depletion of Bifidobacteria, with opportunistic Enterobacteria and Clostridia blooming in their place. The downstream associations include elevated risk of obesity, asthma, and inflammatory bowel disease later in life.

That said, not every environmental input pushes in the wrong direction. For instance, exposure to household pets and older siblings is associated with faster microbiome maturation; which is likely a consequence of broader environmental microbial exposure, and is generally protective against allergic disease.

Restoring What Was Lost: Targeted Interventions

Because the early-life microbiome is malleable, it is also responsive. The same plasticity that makes it vulnerable to disruption makes it open to repair, and the most useful interventions follow the natural lines of microbial transfer that biology already uses.

The first of those lines runs through the mother. Maternal probiotic supplementation has emerged as a powerful tool precisely because the mother’s microbiome is intimately connected to the infant’s through the gut-breast axis. Recent systematic reviews and meta-analyses of randomized controlled trials examined what happens when pregnant and lactating women are supplemented with strains of Lactobacillus and Bifidobacterium. The maternal milk microbiome is remodeled in clinically meaningful ways: beneficial bacterial abundance rises, and pathogenic species such as Staphylococcus are suppressed, which carries particular value for mothers managing lactational mastitis. Through breastfeeding, those beneficial organisms are vertically transferred to the infant, and the downstream effects show up where it matters. Infants of supplemented mothers demonstrate more optimized, controlled weight gain, significantly reduced risk of infantile colic, and lower incidence of atopic conditions including eczema. The intervention works not because we are introducing the infant to something foreign, but because we are reinforcing the biological route that was always meant to carry these microbes.

The second line runs directly to the infant. Given the generational loss of B. infantis in industrialized populations, restoring this keystone species directly is becoming clinically important. A randomized, placebo-controlled trial in 40 exclusively breastfed infants aged 2 to 4 months evaluated whether direct supplementation with B. infantis EVC001 could establish stable colonization, and what the minimum effective dose might be. Across low, medium, and high doses given for 28 days, every tested dose successfully restored fecal B. infantis and produced a nearly two-fold increase in the broader Bifidobacteriaceae family. Crucially, the colonization persisted robustly at least one month after supplementation ended, which suggests the infant gut remains receptive to targeted microbial rescue well beyond the immediate newborn window. Replenishing B. infantis reduces enteric inflammation, limits the presence of antibiotic-resistant bacteria, and trains the developing immune system toward a healthy TH1 response.

For infants exposed to antibiotics, supplementation with specific Bifidobacterium strains may help restore an age-appropriate community structure and mitigate inflammatory outgrowths. Human breast milk remains the ultimate prebiotic, with its oligosaccharides selectively feeding the organisms an infant gut is designed to host.

For more severe dysbiosis, direct microbial transfer is moving from concept to clinic. To address the microbial deficits of C-section delivery, one study carefully screened seven mothers and gave their C-section infants a highly diluted sample of maternal feces mixed with breast milk shortly after birth. The intervention was safe, and it successfully shifted the infants’ microbiome trajectories to closely resemble those of vaginally born infants, restoring the Bacteroides populations that are typically missing after C-section.

In older children, the picture is more complex but equally promising. Autism Spectrum Disorder, which is highly comorbid with severe gastrointestinal dysbiosis, has become a focus for more intensive interventions. In a study of 18 autistic children, researchers tested a 10-week Microbiota Transfer Therapy protocol consisting of two weeks of antibiotic treatment to clear pathogenic organisms, a bowel cleanse, and high-dose standardized human gut microbiota from healthy donors. At the two-year follow-up, participants maintained an 80 percent reduction in gastrointestinal symptoms, and their core autism behavioral symptoms showed slow, steady, and significant improvement, alongside a durable increase in gut microbial diversity.

These are not magic-bullet stories. They are proof-of-principle that the developing microbiome will accept correction when the intervention is biologically coherent and properly timed.

Where Clinical Practice Is Headed

Direct-to-consumer microbiome testing for infants is not yet a standard clinical recommendation, and we should be honest about that. But the trajectory is clear. The first 1,000 days are the critical intervention window, and the science is moving quickly toward early screening that can flag elevated Proteobacteria, which has predictive value for pediatric Crohn’s disease and necrotizing enterocolitis, or the absence of FLVR organisms in the asthma context, before clinical disease ever manifests.

When that kind of screening becomes routine, the conversation will shift from treating pediatric inflammatory and allergic disease to preventing it.

The Bigger Picture

What this body of research ultimately reframes is the question itself. We have spent decades asking why some children develop asthma, obesity, IBD, or neurodevelopmental conditions and others do not, and we have searched for answers almost entirely in the genome and the household. Both matter. But neither, on its own, explains what we see.

The microbiome offers a third explanation, and it is one that is modifiable, partly inheritable, and entirely consequential. A child’s earliest microbial encounters help decide how their immune system will read the world, how their metabolism will partition energy, and how their brain will respond to stress. That is an enormous amount of biology riding on a window that closes around the third birthday.

For clinicians, this is a call to think upstream, to ask about birth mode, feeding history, and antibiotic exposure with the same rigor we ask about family history. For parents, it is a reminder that the small, ordinary decisions of early life, breastfeeding when possible, holding the line on unnecessary antibiotics, supporting the maternal microbiome through pregnancy and lactation, and letting children encounter the microbial world rather than sanitizing it away, are not small at all.

We are not just treating pediatric symptoms anymore. We are cultivating the ecological foundation of a human life.

In the 1980s, epidemiological surveys out of Japan revealed something striking. Only 13.5% to 15.2% of menopausal Japanese women reported experiencing hot flashes. Meanwhile, women in North America and Europe were (and still are) moving through the menopausal transition with vasomotor symptoms (hot flashes, night sweats) so intense and so prolonged that many spend a decade of their lives managing them.

The easy explanation was diet. Japanese women ate soy. Western women did not. The conclusion felt tidy, and the supplement industry ran with it.

But something kept not adding up. When Western women started taking isolated soy extract pills, the results were messy. Some women got striking relief. Others got none at all. The trials were inconsistent, and the data from one study would contradict the next. For a molecule that was supposed to be a clean hormonal lever, soy was behaving strangely.

Why? Because the answer to the paradox was not in the soybean. The answer was sitting quietly in the gut.

Soy Is Not the Active Ingredient. Equol Is.

Soy is rich in isoflavones, particularly one called daidzein. On its own, daidzein has a relatively weak estrogenic effect in the human body. It is not the active molecule most women need.

The molecule that actually does the work is equol. Equol is a nonsteroidal compound that is produced when specific gut bacteria metabolize daidzein. And equol is not just a slightly better version of daidzein. It is a different class of molecule entirely.

Of all soy isoflavone metabolites, equol has the strongest binding affinity to estrogen receptor-beta. Because it preferentially binds to this specific receptor and not estrogen receptor-alpha, it behaves like a natural selective estrogen receptor modulator (SERM). That means it delivers the cooling, symptom-relieving benefits of estrogen without the proliferative effects on breast and uterine tissue.

This is what every woman reaching for soy is actually hoping to get. Equol. Not daidzein.

However, here is the problem - humans cannot make equol on their own. We are entirely dependent on gut bacteria to do it for us. And unfortunately, not every woman has those bacteria.

The Two Hidden Populations Inside Every Soy Trial

Have you ever wondered why there seem to be inconsistencies in soy trials?

Well, it is because every clinical trial that has ever tested soy for menopausal symptoms has, without knowing it, been studying two completely different populations at the same time. There are equol producers and there are non-producers, and they respond to soy as if they were in two different studies.

Only 20% to 30% of Western adults harbor the gut bacteria required to convert daidzein into equol. In Asian populations, that number climbs to 50% to 60%. When a clinical trial recruits a Western cohort and does not stratify participants upstream, the equol producers get real symptomatic relief, the non-producers get essentially nothing, and the two results cancel each other out in the final statistical analysis.

The trial then gets published as “inconclusive.” But the real story is that the microbiome of roughly three out of four women was never equipped to generate the active compound in the first place.

This is what I mean when I say the gut microbiome is what determines the fate of responders versus non-responders.

Meet the Microscopic Machinery

The bacteria doing this conversion are a small and specialized group. Most of them belong to the family Eggerthellaceae (historically grouped under the Coriobacteriaceae), and the species level matters, because this is where you start to see how narrow the machinery actually is.

The equol-producing species isolated from the human gut so far include:

• Adlercreutzia equolifaciens

• Slackia equolifaciens

• Slackia isoflavoniconvertens

• Asaccharobacter celatus

• Enterorhabdus mucosicola

• Eggerthella sp. YY7918

A few other strains have been identified but not yet fully named, including Paraeggerthella sp. SNR40-432. Most of these organisms are strict anaerobes, which means they are fragile, oxygen-sensitive, and easily suppressed by antibiotics, alcohol, and inflammatory conditions.

Making equol from daidzein is not a one-step reaction. It is a multi-step enzymatic assembly line. Three reductases (DZNR, DHDR, and THDR) and one racemase work in sequence to progressively modify daidzein until it becomes (S)-equol, the biologically active enantiomer that the human body absorbs and uses.

Some bacterial strains carry all the enzymes. Others only carry a few. When a strain is missing an enzyme, it cannot complete the conversion on its own, which means multiple strains often have to collaborate in the same gut to get the job done. This is microbial cross-feeding in action, and it is one of the reasons why diversity in the Eggerthellaceae family matters more than the presence of any single species.

When that ecological collaboration breaks down, equol production stops.

The WAVS Trial and Why Whole Soybeans Outperform Pills

If equol production depends on having the right bacteria, then the question becomes obvious: can we feed those bacteria to make soy work better?

The Women’s Study for the Alleviation of Vasomotor Symptoms (WAVS) answered this question with one of the most striking datasets in menopause nutrition.

Postmenopausal women who were experiencing multiple daily hot flashes were placed on a low-fat, vegan diet that included just half a cup (86 grams) of cooked whole soybeans per day. According to the WAVS trial, total hot flashes decreased by 79% and moderate-to-severe hot flashes dropped by 84%. Within 12 weeks, 59% of women in the dietary intervention group became completely free of moderate-to-severe hot flashes.

This is not a result you see with isolated isoflavone supplements. Something fundamentally different happens when the whole food is consumed alongside a fiber-dense, plant-forward eating pattern.

What I believe is happening is that the diet itself is doing two jobs at once. It is delivering the daidzein substrate, and it is simultaneously reshaping the microbial environment to favor the proliferation of equol-producing microbes. Fiber feeds the Lachnospiraceae and Ruminococcaceae families, which in turn create the anaerobic, SCFA-rich terrain in which Adlercreutzia and Slackia species can thrive. Isolated isoflavone pills deliver the substrate but do nothing for the ecosystem.

You cannot supplement your way around a bacterial deficit. You have to feed it back.

Menopause Itself Sabotages the Very Bacteria You Need

There is even a further twist to this story!

The menopausal transition itself is rough on the exact microbial communities that produce equol. Aging and the hormonal shifts of perimenopause destabilize the gut ecosystem at the worst possible time. In studies of menopausal women taking soy supplements, those who were non-producers actually saw a significant decrease in their overall cultivable microbial populations over a 6-month period. Equol producers, on the other hand, saw their microbial populations thrive.

To put this differently: if a woman enters perimenopause without the baseline equol-producing bacteria, the transition itself makes it harder, not easier, to build them up. The microbiome of a non-producer on soy drifts further from the equol-producing phenotype over time.

This is why I believe that the best window to assess and support a woman’s microbiome is before the symptoms become entrenched. Pre-habilitating the gut microbiome with the baseline species at perimenopause is crucial for postmenopausal health.

“Am I a Producer?” The Clinical Question Worth Asking

At this point, the practical question becomes unavoidable. How does a woman know whether she is equipped to convert daidzein into equol?

There are two clean ways to phenotype this.

The classical method is a soy challenge. A patient consumes a standardized dose of soy isoflavones (typically daidzein) for about three days, after which urine is collected and analyzed for equol and daidzein. Rather than relying on an absolute concentration, the diagnostic marker is the ratio of equol to daidzein excreted. If equol accounts for roughly 1.8 percent or more of the daidzein excreted, the patient is identified as an equol producer.

The newer and more elegant approach uses an in vitro fecal incubation to directly assess whether a woman’s microbial community can convert daidzein into equol under controlled conditions. This bypasses the need for a dietary intervention and can be run alongside a standard microbiome assessment.

In practice, I think every woman considering soy as a menopause strategy should be asking this question, and every practitioner recommending soy to a perimenopausal patient should be thinking about it.

What to Do If You Are a Non-Producer

For the roughly 70% of Western women who do not naturally produce equol, there are two scientifically supported paths forward.

The first is to bypass the gut entirely with a natural S-equol supplement. Clinical trials using SE5-OH, a supplement made by fermenting soy germ with equol-producing bacteria, have shown that a daily 10 mg dose significantly reduces hot flash frequency and severity, and even relieves neck and shoulder stiffness in non-producers. This is the most direct workaround.

The second path is slower but more foundational. Remodel the microbiome. Women who consume higher amounts of polyunsaturated fatty acids and follow plant-heavy diets are significantly more likely to harbor equol-producing bacteria. Transitioning to a high-fiber, low-fat, plant-forward eating pattern appears to recruit and feed the species that make the conversion possible. It is not a guaranteed fix, and it takes time, but it addresses the terrain rather than patching over it.

In my own thinking, these two strategies are not either-or. A non-producer can take S-equol for symptom relief now, while simultaneously remodeling the microbiome over the course of six to twelve months to build the native capacity.

The Bigger Frame: Soy Is an Ecological Intervention

This is where I want to land.

We have to stop thinking about soy as a nutrient and certainly stop thinking about it as a drug. Soy is an ecological intervention for the gut.

When a woman eats whole soybeans, she is not just ingesting isoflavones. She is delivering prebiotic substrates and alternative energy sources that actively shape which microbial communities thrive. The Lachnospiraceae and Ruminococcaceae families expand. The terrain shifts. And in some women, that shift recruits and feeds the Adlercreutzia, Slackia, and Eggerthella species that generate equol.

The unique composition of a woman’s microbiome is what decides whether soy is medicine or sham. Two women can eat the exact same meal and experience two fundamentally different hormonal outcomes, and neither of them has done anything wrong.

This is the frame I want every practitioner and every perimenopausal woman to hold. The microbiome is not a downstream detail in menopause care. It is the deciding factor in whether a dietary strategy works at all.

And once you see menopause through that lens, the mixed results in the soy literature stop being a contradiction. They start being exactly what you would expect.

References

1. Lampe, J. W. (2009). Is equol the key to the efficacy of soy foods? American Journal of Clinical Nutrition, 89(5), 1664S–1667S.

2. Setchell, K. D. R., Brown, N. M., & Lydeking-Olsen, E. (2002). The clinical importance of the metabolite equol: a clue to the effectiveness of soy and its isoflavones. Journal of Nutrition, 132(12), 3577–3584.

3. Barnard, N. D., Kahleova, H., Holtz, D. N., et al. (2021). The Women’s Study for the Alleviation of Vasomotor Symptoms (WAVS): a randomized, controlled trial of a plant-based diet and whole soybeans for postmenopausal women. Menopause, 28(10), 1150–1156.

4. Matthies, A., Blaut, M., & Braune, A. (2009). Isolation of a human intestinal bacterium capable of daidzein and genistein conversion. Applied and Environmental Microbiology, 75(6), 1740–1744.

5. Yokoyama, S., & Suzuki, T. (2008). Isolation and characterization of a novel equol-producing bacterium from human feces. Bioscience, Biotechnology, and Biochemistry, 72(10), 2660–2666.

6. Shimada, Y., Takahashi, M., Miyazawa, N., Abiru, Y., Uchiyama, S., & Hishigaki, H. (2012). Identification of a novel dihydrodaidzein racemase essential for biosynthesis of equol from daidzein in Lactococcus sp. strain 20-92. Applied and Environmental Microbiology, 78(14), 4902–4907.

7. Aso, T., Uchiyama, S., Matsumura, Y., et al. (2012). A natural S-equol supplement alleviates hot flushes and other menopausal symptoms in equol nonproducing postmenopausal Japanese women. Journal of Women’s Health, 21(1), 92–100.

8. Mayo, B., Vázquez, L., & Flórez, A. B. (2019). Equol: a bacterial metabolite from the daidzein isoflavone and its presumed beneficial health effects. Nutrients, 11(9), 2231.

9. Frankenfeld, C. L. (2017). Cardiometabolic risk and gut microbial phytoestrogen metabolite phenotypes. Molecular Nutrition & Food Research, 61(1), 1500900.

Now, I want to be clear: the microbiome is not the whole picture. Weight health is multifactorial. Behavioral psychology, hormonal status, sleep architecture, physical activity, medication effects, socioeconomic access to food, all of these matter, and there are well-established clinical approaches addressing each. What I am offering here is the microbiome perspective, a lens that is often missing from the conversation despite strong and growing evidence that it shapes metabolic outcomes in ways that standard interventions alone do not explain. This post is not a replacement for comprehensive weight management. It is a complement to it.

That said, in the context of practical application, many clinicians and patients still struggle with identifying and tracking the microbiome markers most relevant to weight gain and obesity. That is why I am writing this — to elaborate on the key microbiome markers of weight health and to discuss the clinical reasoning and nuances behind these signatures. The fundamental clinical value of microbiome testing for weight loss patients lies in its ability to replace generalized dietary advice with precision interventions. Two patients presenting with identical BMI, dietary habits, and exercise patterns may have fundamentally different microbiome landscapes driving their metabolic dysfunction. One may have a Proteobacteria-dominant, LPS-driven inflammatory profile requiring barrier repair and antimicrobial botanicals. The other may have SCFA depletion requiring aggressive prebiotic therapy and resistant starch loading. Without testing, both patients receive the same intervention, and one fails.

It is also important to note that microbiome changes can either modulate or result from weight gain. A single gut test cannot tell you whether these markers are causes or effects, and realistically, that distinction does not matter as much as we think. What matters is that these changes create a continuing feedback loop that dysregulates metabolic function, exacerbating the condition. The goal is to break that loop by correcting the dysbiosis signatures that lead to or reinforce weight gain and weight loss resistance.

In this article, I share the various microbiome archetypes of weight gain, obesity, and weight loss resistance. I derived these archetypes from analyzing hundreds of microbiome profiles associated with weight phenotypes. As you will soon learn, there is no universal microbiome profile for weight gain or obesity, so a one-size-fits-all approach will likely fail. For each archetype, I also explain the mechanistic and clinical reasoning for why that profile underlies weight loss resistance. I then suggest interventional approaches to support each archetype. Consider this a framework, not a fixed protocol. Typically, other factors, including health history, comorbidities, and intolerances, determine the final protocol the clinician lands on.

The 7R Framework

The framework I use is built around seven clinical tools, each beginning with R.

Reset primes motility, enzyme output, and bile flow so the gut can respond to what comes next. Reduce clears pathobionts and biofilm-forming organisms that are actively driving dysfunction. Reseed introduces targeted probiotic strains selected by phenotype, not by label. Refeed supplies the prebiotic substrates those organisms need to establish and produce meaningful metabolic output. Repair restores barrier integrity and supplies the cofactors the ecosystem cannot function without. Rebalance addresses the higher-order signaling layers such as hormones, bile acids, neurotransmitters, and oxidative stress. Resilience is the long game. It includes dietary diversity, exercise, sleep, stress management, and serial retesting to confirm that what you built holds.

In every case, Reset is assessed universally before any archetype-specific work begins. And Resilience, while always the last step executed, is planned from the first consultation. Serial retesting confirms whether the protocol is working or whether you need to cycle back.

These are not phases to march through in order. They are tools. The archetype determines which tools you pull, in what sequence, and which you skip entirely.

The Various Microbiome Archetypes of Weight Gain & Interventional Approaches

Archetype 1: The Inflammatory-Dominant Profile

Elevated inflammatory markers, Proteobacteria expansion, depleted barrier-protective species, and metabolic endotoxemia.

Clinical Reasoning: Why This Profile Drives Weight Loss Resistance

Expansion of LPS-producing Proteobacteria combined with a compromised gut barrier allows bacterial endotoxin (lipopolysaccharide) to translocate into systemic circulation. This triggers chronic low-grade inflammation, which directly promotes insulin resistance, leptin resistance, and adipogenesis. The patient is in a state of metabolic endotoxemia: their immune system is constantly activated by microbial products that should never have crossed the epithelial barrier.

In this state, caloric restriction and exercise produce diminished returns because inflammatory signaling overrides normal metabolic regulation. Until the source of endotoxemia is resolved (pathobiont reduction plus barrier repair), the patient’s metabolic physiology actively opposes weight loss.

Characteristic Biomarker Pattern

• Stool markers: Calprotectin >80 ug/g; sIgA > 275 mg/dL; total SCFAs <5.0 mg/mL

• Microbiome composition markers: Proteobacteria = high; Faecalibacterium, Akkermansia, Roseburia, and other keystone species = low

• Microbiome functional capacity: LPS = high

Interventional Approach

The Inflammatory-Dominant Profile follows the sequence: Reduce → Repair → Reseed → Refeed → Resilience. When Proteobacteria are expanded and LPS potential is high, the priority is clearing the endotoxin-producing organisms before anything else. You cannot repair a barrier while the organisms destroying it are still thriving. Once pathobionts are reduced, barrier repair with tributyrin and zinc carnosine stabilizes the epithelium so that probiotics introduced in the reseeding phase actually have a surface to colonize. Prebiotics come last and are staged cautiously, because feeding an inflamed, leaky gut fuels the wrong organisms. PHGG is the safest starting point here. High-dose inulin waits until inflammation has resolved.

Archetype 2: The SCFA-Depleted / Low Fermentation Profile

Depleted SCFA production, low diversity, insufficient fermentation capacity, impaired GLP-1 activation. No significant pathobiont overgrowth.

Clinical Reasoning: Why This Profile Drives Weight Loss Resistance

Short-chain fatty acids, particularly propionate and butyrate, are the primary microbial signals that stimulate GLP-1 secretion from intestinal L-cells. GLP-1 suppresses appetite, slows gastric emptying, and improves insulin sensitivity. When the microbiome lacks the fermenters to produce adequate SCFAs, endogenous GLP-1 output is impaired. The patient experiences poor satiety, dysregulated blood glucose, and increased caloric intake despite conscious effort.

This is the same pathway targeted by GLP-1 receptor agonist drugs (semaglutide, tirzepatide), meaning a depleted SCFA profile represents a functional deficiency in the body’s own appetite regulation system. Restoring fermentation capacity through targeted reseeding and prebiotic loading can rebuild this endogenous pathway.

Characteristic Biomarker Pattern

• Stool markers: Total SCFAs <5.0 mg/mL; stool pH alkaline (>7.0)

• Microbiome composition markers: Diversity score ≤3; low levels of Faecalibacterium, Roseburia, and other keystone species; low Prevotella/Bacteroides ratio

• Microbiome functional capacity: Endogenous GLP-1 inducers (propionate, butyrate) = low

Interventional Approach

The SCFA-Depleted / Low Fermentation Profile follows a different logic entirely: Reseed → Refeed → Repair → Rebalance → Resilience. The Reduce step is skipped altogether. There are no pathobionts to clear, and deploying antimicrobials into an already impoverished ecosystem would strip out the little diversity that remains. Instead, you establish fermenters first with Bifido-dominant and Lactobacillus strains, then immediately feed them with resistant starch and PHGG to kickstart the cross-feeding cascade that produces butyrate. Barrier repair runs concurrently as endogenous SCFA production builds, bridging the colonocyte fuel gap with exogenous tributyrin until the gut can supply its own. This archetype has the most predictable recovery trajectory, typically responding within 8 to 12 weeks.

Archetype 3: The Methane-Dominant / Slow Transit Profile

Elevated Methanobrevibacter smithii, constipation, bloating, slow transit, and paradoxical weight gain from increased caloric absorption. Note: Reset (motility priming) is especially critical for this archetype.

Clinical Reasoning: Why This Profile Drives Weight Loss Resistance

Methanobrevibacter smithii produces methane gas in the colon, which directly slows intestinal transit by inhibiting serotonin-mediated peristalsis. Delayed transit time increases the duration of contact between luminal contents and the absorptive epithelium, leading to greater caloric extraction from the same amount of food. A patient with methane-dominant constipation may absorb significantly more calories from an identical meal compared to a patient with normal transit.

This creates a paradox: the patient gains weight or fails to lose weight despite genuine caloric restriction. Additionally, slow transit promotes putrefactive fermentation (elevated ammonia, hydrogen sulfide), which further damages the colonic environment. These patients often present with frustration and distrust of dietary approaches because their lived experience is that diets do not work. The clinical explanation of the methane-transit-calorie absorption mechanism restores trust and provides the rationale for a motility-first protocol.

Characteristic Biomarker Pattern

• Stool biomarker: Stool consistency, hard/lumpy; pH may be alkaline

• Microbiome composition marker: M. smithii elevated

• Microbiome functional capacity: Serotonin and/or acetylcholine = low; methane = high

Interventional Approach

The Methane-Dominant / Slow Transit Profile requires aggressive motility priming (Reset) before anything archetype-specific begins. After that, the sequence is: Reduce → Reseed → Refeed → Rebalance → Resilience. The Reduce step here is not antibacterial but anti-archaeal, because Methanobrevibacter smithii is an archaeon that resists standard antimicrobial botanicals. Allicin is the primary agent. Reseeding then targets strains with specific evidence in constipation-predominant profiles, not just general colonization. Only soluble fiber during the Refeed phase, since insoluble fiber worsens bloating in slow-transit patients. The Rebalance step focuses on sustained motility support through exercise, vagal tone work, and serotonin pathway support, because without ongoing motility maintenance, methanogens recolonize.

Archetype 4: The Hormonal-Metabolic Profile

Beta-glucuronidase dysregulation, low GLP-1 activation, impaired neurotransmitter signaling, and visceral adiposity. Common in perimenopausal patients and those on hormonal therapies. Reduce is deployed only if the report shows concurrent pathobiont elevation.

Clinical Reasoning: Why This Profile Drives Weight Loss Resistance

This archetype operates through two interconnected axes. First, elevated beta-glucuronidase activity in the gut deconjugates estrogen metabolites that were intended for fecal excretion, allowing them to be reabsorbed into circulation through the estrobolome. This estrogen recirculation promotes estrogen dominance, which drives preferential visceral fat deposition, particularly in perimenopausal women and patients on hormonal therapies.

Second, depleted microbial production of serotonin, dopamine, and GABA impairs the gut-brain signaling that regulates appetite, mood, and stress response. Low serotonin reduces satiety. Low dopamine drives reward-seeking eating behavior. Low GABA elevates anxiety, which activates the HPA axis and increases cortisol. Cortisol directly promotes visceral fat accumulation and insulin resistance.

The patient experiences weight gain that is hormonally mediated, not calorically mediated, which is why standard dietary approaches fail. The protocol must address the hormonal and neurotransmitter signaling layer first, before rebuilding the microbial ecosystem.

Characteristic Biomarker Pattern

• Stool biomarker: Beta-glucuronidase > 8000 U/h*g

• Microbiome composition markers: Bacteroides, E. coli = high (both are high beta-glucuronidase producers in the gut)

• Microbiome functional capacity: GLP-1, serotonin, dopamine = low; vitamin B6/B9 = low

Interventional Approach

The Hormonal-Metabolic Profile flips the conventional order: Rebalance → Reseed → Refeed → Repair → Resilience. Rebalance is moved to the front because the primary driver is not microbial composition only but hormonal and neurotransmitter signaling dysfunction as well. Elevated beta-glucuronidase is recycling estrogen through the estrobolome, driving visceral fat deposition. Depleted serotonin, dopamine, and GABA production impair appetite regulation and amplify cortisol-driven weight gain. Calcium-D-glucarate addresses beta-glucuronidase directly, while adaptogenic and antihistamine support targets the HPA axis. Only after the signaling layer is addressed do you reseed with strains selected for neurotransmitter modulation, and strain selection must account for histamine status. Retest intervals are longer here, a minimum of 16 weeks, because estrobolome shifts are slow.

Archetype 5: The GLP-1 Pharmacotherapy Adjunct Profile

Patients on or transitioning to/from GLP-1 receptor agonist therapy (semaglutide, tirzepatide). Microbiome optimization maximizes pharmacological response, mitigates side effects, and builds an exit strategy. Reduce is deployed only if concurrent pathobiont overgrowth exists.

Clinical Reasoning: Why Microbiome Optimization Matters for This Profile

GLP-1 receptor agonists produce weight loss by suppressing appetite and slowing gastric emptying, but they do not address the underlying microbiome dysfunction that contributed to metabolic disease. Patients on GLP-1 RAs frequently reduce their dietary intake to very low levels, which inadvertently results in suboptimal daily intake of fiber and nutrients. This may lead to low microbial diversity, weakened mucosal barrier, and depleted SCFA-producing species that drive endogenous GLP-1 secretion.

The result is a patient who loses weight pharmacologically while their microbiome deteriorates. When the drug is discontinued, the patient lacks the endogenous GLP-1 capacity to maintain satiety and metabolic regulation, leading to rapid weight regain. Additionally, GLP-1 RA-mediated weight loss is often disproportionately lean mass rather than fat mass, compounding the metabolic vulnerability.

The microbiome protocol for this archetype serves two purposes: mitigate the collateral damage of pharmacotherapy (barrier weakening, diversity loss) and build endogenous GLP-1 production capacity through SCFA optimization, so the patient has a viable exit strategy from the drug.

Characteristic Biomarker Pattern

• Microbiome functional marker: GLP-1 activators, SCFAs = low

• Microbiome composition marker: Akkermansia muciniphila = high; Diversity score may decline due to reduced dietary intake; variable pathobiont picture (overlay with any other archetype is possible)

Interventional Approach

The GLP-1 Pharmacotherapy Adjunct Profile sequences as: Repair → Reseed → Refeed → Rebalance → Resilience. Patients on semaglutide or tirzepatide are eating significantly less, which inadvertently starves the microbiome of dietary diversity and weakens the mucosal barrier. Repair is prioritized early to counteract this collateral damage, with emphasis on adequate protein to prevent the disproportionate lean mass loss that GLP-1 receptor agonists are known to cause. Reseeding and Refeeding then build endogenous GLP-1 production capacity through SCFA optimization. This is the exit strategy. The entire point of the microbiome protocol in this archetype is to ensure the patient has a functioning endogenous appetite regulation system when the drug is eventually discontinued. Without it, weight regain is predictable.

Conclusion: The Pattern Is the Protocol

If there is one thing I want you to take away from this post, it is this: there is no single gut microbiome profile for weight gain. There are patterns, and each pattern has its own logic, its own drivers, and its own sequence of intervention.

The patient whose metabolism is being hijacked by endotoxemia does not need the same protocol as the patient whose fermentation capacity has collapsed. The patient with methane-dominant constipation absorbing extra calories from every meal is not the same as the patient whose estrobolome is recycling estrogen into visceral fat. And the patient on semaglutide who is losing weight while their microbiome quietly deteriorates needs a plan for what happens after the prescription ends.

This is why I built the archetype framework, and why I treat the 7R system as a toolkit, not a checklist. The archetype tells you which tools to reach for. The clinical picture tells you when to reach for them.

If you are a clinician, I hope this gives you a working lens to look at your patients’ microbiome reports differently, not just reading the numbers, but reading the metabolic story those numbers are telling. If you are someone navigating your own weight health journey, I hope this helps you understand why the approaches you have tried may not have worked, and why the answer may not be to try harder, but to look deeper.

I would love to hear from you. Which of these archetypes resonates most with your clinical experience or your own health story? If you are a practitioner, have you seen patients who clearly fit one of these profiles, or who overlap between archetypes? Drop your thoughts in the comments. The more we share, the sharper this framework gets.

Disclaimer: This article is for educational purposes only and does not constitute medical advice. It is not intended to replace individualized clinical assessment or treatment. Always consult a qualified healthcare professional before making changes to your health protocol.

I get it. The one-size-fits-all approach does not work for human biology. Everything – the patient’s history, clinical presentations, lifestyle, and even the methodology behind the microbiome report itself has to be well contextualized to make any meaningful change.

But here is what I have learned from those 150-plus consultations: regardless of the protocol a practitioner eventually lands on, there are non-negotiable priorities that must be addressed before any intervention begins. Skip them, and the protocol will likely fail; not because the protocol was wrong, but because the terrain was not ready for it.

The microbiome does not function in isolation. It modulates and is modulated by digestion, immune tone, motility, and neurochemistry. If those systems are compromised, even the best-designed protocol is building on an unstable foundation.

A quick disclaimer: This is not medical advice. The purpose is to educate both the practitioner and patient on key considerations to increase the chances of success. By training, I am a scientist (PhD in Molecular Genetics with specialization in Molecular Microbiology), and I run a microbiome company, so I have had the privilege to review more microbiome reports than most people. I am also in constant interaction with clinicians of diverse backgrounds (MDs, naturopaths, dietitians, health coaches, longevity specialists, name it), exchanging ideas and notes. Hence, my thoughts are founded on systems biology thinking, well-tested in clinical settings.

Here are the five priorities I walk practitioners through before they write a single recommendation.

Priority 1: Safety – Is There an Active Infection?

This is non-negotiable. If you see a known pathogen on the report such as Clostridioides difficile, Salmonella, pathogenic E. coli, Campylobacter, or elevated parasitic signatures, you stop everything else. Active infections override optimization protocols, full stop.

The patient is in crisis mode, and the immune system is actively fighting an invader. Loading prebiotics or diversity-building fibers on top of an active infection is not just unhelpful, it can make things worse. You could be feeding the very organism you need to eliminate, or provoking an immune response that amplifies the damage.

What to look for on the report: Presence of known pathogens via PCR or sequencing-based detection. If your test includes virulence factor profiling or toxin gene detection, even better. Usually, active infections are associated with clinical presentations, so the presence of a disease agent matched with apparent symptoms is important.

The clinical move: Treat the infection first. Confirm clearance. Then proceed to the rest of the framework.

Priority 2: Digestion – Can This Patient Actually Break Down What You Are Prescribing?

This is the one I find myself repeating most often. A practitioner designs an elegant protocol, e.g. targeted fibers, polyphenol-rich foods, resistant starch, and the patient comes back eight weeks later with bloating, discomfort, and no meaningful change. Why? Because they could not digest any of it.

Pancreatic elastase is the biomarker here. It is a measure of exocrine pancreatic function – essentially, how well the pancreas is producing digestive enzymes. If pancreatic elastase is below 200 µg/g, you are looking at pancreatic exocrine insufficiency. The patient is not breaking down proteins, fats, or complex carbohydrates properly. And if they cannot break them down, the microbiome cannot access the substrates it needs from them either.

What to look for on the report: Faecal pancreatic elastase levels. Below 200 µg/g is insufficient. Below 100 µg/g is severe.

The clinical move: Supplement with pancreatic enzyme replacement therapy (PERT) before introducing any prebiotic, fibre, or dietary intervention. In situations like this, this is not optional. It is a prerequisite. You are not just feeding the patient. You are feeding the microbiome through the patient’s digestive capacity.

Read more

If you work in microbiome science long enough, you start to appreciate that gut bacteria are not passive tenants. They are biochemical factories — producing metabolites, signaling molecules, and even toxins that can influence organs far removed from the intestinal lumen. Most of the time, these microbial products stay local. But sometimes, they find a highway out.

All three times I tested my friend’s child’s gut microbiome, she had remarkably elevated levels of three key genera: Alistipes, Sarcina, and Clostridium - with species within the Clostridium cluster being the most prevalent. Her levels were also substantially higher than those seen in her healthy sibling who did not have ASD. When I compared her microbiome profile to a healthy control group (not age-matched or sex-matched), her Clostridium levels sat at the 99th percentile. Upon further research, I found that her profile aligned well with findings from numerous published studies, suggesting that the pattern of dysbiosis I was seeing in my N-of-1 experiment may not be artifactual. So my goal became to pursue a theoretical hypothesis for why the Clostridium dominance seen in the gut microbiome of autistic children could modulate their symptoms and behavior.

Let me be clear upfront: this is a hypothesis. It has not been proven. Autism is highly heritable, polygenic, and multifactorial. No single microbial toxin causes autism. But as someone who spends every day looking at microbiome data and thinking about how microbial ecosystems interact with human physiology, I believe this hypothesis deserves a serious scientific conversation. The molecular biology underpinning each step is real. The question is whether the complete chain operates in living humans at the relevant scale.

At the end of this article, I also discuss arguments that strengthen and weaken the hypothesis. I am hoping it challenges you, the reader, to support, refute, or synthesize altogether a different premise through which this complex linkage should be viewed.

Where Did This Idea Come From?

The hypothesis was most notably articulated by Ellen Bolte in a 1998 paper published in Medical Hypotheses. Bolte was not an academic microbiologist - she was a mother whose son had developed autism following a course of antibiotics, and she connected the dots between clostridial overgrowth, neurotoxin production, and behavioral regression. The idea was later explored and expanded upon by Sydney Finegold and others, who documented elevated Clostridium species in the stool of children with ASD compared to neurotypical controls.

What made this hypothesis stand out is its specificity. It does not end with “gut bacteria are involved.” It proposes a complete mechanistic chain - from dysbiosis to neurotoxin to brain dysfunction. That specificity is both the hypothesis’s greatest strength and its greatest vulnerability, because every link can be independently tested and potentially disproved.

To unpack this, I will break down this mechanism into six key events linking how elevated levels of specific Clostridium species in the gut can potentially mediate autistic behavior.

The Mechanistic Chain: Six Links from Gut to Brain

Link 1: Dysbiotic Clostridial Overgrowth

The story begins with disruption of the developing gut microbiome. In susceptible infants and toddlers, particularly following broad-spectrum antibiotic use, the loss of colonization resistance from protective commensals like Bacteroides and Bifidobacterium opens ecological niches for spore-forming Clostridia.

Multiple studies have documented elevated counts of Clostridium clusters I and XI in the stool of autistic children versus matched controls, including work by Finegold et al. (2002), Parracho et al. (2005), and Song et al. (2004). Spore formers are notoriously resilient — they survive antibiotic assault and recolonize aggressively once the competitive landscape is cleared.

Link 2: Local Tetanus Neurotoxin Production

Clostridium tetani is the classic producer of tetanus neurotoxin (TeNT), but it is not alone. Several related clostridia carry tetanus toxin gene homologs or produce structurally similar zinc-metalloprotease neurotoxins. The hypothesis proposes that these organisms, once established in the gut, produce TeNT or TeNT-like toxins at sub-clinical levels — not enough to cause the full-blown spastic paralysis of classical tetanus, but enough to interact with the dense neuronal network embedded in the intestinal wall.

If you think about this critically, you would not rule this out as being far-fetched. The enteric nervous system - sometimes called the “second brain”, contains roughly 500 million neurons, and it sits just millimeters away from the luminal contents. This means the local toxin–nerve interaction does not necessarily require the toxin to be absorbed into the bloodstream. It just needs to reach the nerve endings lining the gut wall.

Link 3: Retrograde Axonal Transport to the CNS

This is where the molecular biology gets elegant. TeNT has a well-characterized mode of entry: it binds to ganglioside receptors on presynaptic nerve terminals, undergoes receptor-mediated endocytosis, and is then retrogradely transported along the axon toward the cell body. This is the same mechanism that allows classical tetanus to occur — where infection in a wound allows toxin to climb peripheral nerves to the spinal cord.

In the gut context, the vagus nerve is the highway. It provides a direct anatomical connection from the enteric nervous system to the brainstem. From there, connections radiate to the limbic system, basal ganglia, and cortex. Low doses of tetanus toxin, hitching a ride up the vagus, could reach precisely the brain regions that show altered function in autism: the amygdala, temporal cortex, cerebellum, and prefrontal circuits.

Link 4: Disruption of Inhibitory Neurotransmission

Once inside the central nervous system (CNS), TeNT does one thing with extraordinary precision: it cleaves VAMP2 (also known as synaptobrevin), a protein essential for neuronal communication.

Think of your nervous system like a massive postal network. Brain neurons need to send chemical “letters” called neurotransmitters to each other across small gaps between them. To deliver these chemical letters, neurons use a special release system. Imagine it like a zipper that opens to let packages out. VAMP2 is a crucial part of that “zipper system” - it is essentially the pull-tab. Without it working properly, the zipper cannot open and the neurotransmitter packages cannot be delivered.

TeNT acts like a very precise pair of molecular scissors that cuts only that specific pull-tab (VAMP2). It is remarkably accurate: like scissors that can cut only red ribbon in a room full of different colored ribbons. And critically, TeNT preferentially targets GABA-producing neurons (called GABAergic neurons).

Why does this matter? Because the brain has two fundamental types of signaling neurons. Some act like “gas pedals” (excitatory neurons) that make things happen, and others act like “brakes” (inhibitory neurons) that calm things down. The “brake” neurons use chemicals called GABA and glycine to tell other brain cells to slow down or stop firing. TeNT specifically targets these brake neurons. When it cuts their delivery system, they cannot release their “slow down” messages properly.

Now apply this to a developing brain. If sub-clinical levels of TeNT selectively impair GABA release from inhibitory interneurons, you get a shift in the excitatory/inhibitory (E/I) balance toward excitation. And E/I imbalance is one of the most consistently observed neurophysiological signatures in autism. EEG studies, MEG recordings, and postmortem analyses have all documented reduced GABAergic tone, decreased interneuron density, and altered GABA receptor expression in individuals with ASD.

What this hypothesis makes us ponder is this: what if a microbial neurotoxin is one of the upstream drivers of that imbalance?

Link 5: Developmental Timing Is Everything

The developing brain in the first two to three years of life is undergoing massive synaptogenesis, pruning, and circuit consolidation. Think of it as a massive construction site. Workers are frantically building new roads (connections between brain cells), tearing down temporary scaffolding, and deciding which roads become permanent highways.

Remember those “brake” chemicals (GABA) we talked about? In babies, they actually work backwards. Instead of saying “slow down,” they are saying “speed up and grow!” It is like having construction workers who are supposed to direct traffic, but in babies, they are telling everyone to build faster. Around age 2–3, something remarkable happens: the baby’s brain essentially flips a switch, and suddenly those same GABA workers change their job — they go from saying “build faster” to “slow down and be calm.” This transition is controlled by two molecular “switches” called NKCC1 and KCC2.

Now imagine that TeNT damages this system during that critical construction period. If the GABA workers cannot do their job properly during the “build faster” phase, the brain’s construction goes haywire. Neural connections get built in the wrong places or not at all. And when it is time for the big switchover to “slow down” mode, the system is already damaged. The result could be a severely altered brain circuit that cannot calm itself properly.

The hypothesis gains biological plausibility precisely because the proposed insult (impaired GABA release) intersects with a known vulnerability window.

Link 6: The Behavioral Endpoint

The brain regions most accessible via retrograde vagal transport — brainstem nuclei, amygdala, cerebellum, temporal cortex map directly onto the core behavioral domains of autism: social cognition (amygdala, temporal cortex), language development (temporal–frontal circuits), repetitive behaviors (basal ganglia–cerebellar loops), and sensory hypersensitivity (brainstem, thalamus).

The anatomical targeting of the proposed toxin delivery route aligns with the neuroanatomy of autism. This is a striking alignment, even if it is coincidental.

Strengths of the Hypothesis

1. Mechanistic specificity. Unlike vague “gut–brain connection” claims, this hypothesis lays out a testable molecular chain: organism → toxin → receptor → transport mechanism → target neurons → biochemical consequence → phenotype. Every step can be experimentally tested.

2. Solid molecular biology at each link. TeNT’s binding to gangliosides, retrograde transport, VAMP2 cleavage, and preferential accumulation in inhibitory neurons are all well-established in neurotoxicology. None of these individual steps require speculative leaps.

3. Consistent with observed dysbiosis. Elevated Clostridium species in ASD stool samples have been reported across multiple independent studies. The ecological starting point is empirically supported.

4. E/I imbalance convergence. The proposed mechanism converges on excitatory/inhibitory imbalance, which is independently one of the strongest neurophysiological findings in autism research. The hypothesis offers a potential upstream driver for an observation that currently lacks a unifying cause.

5. Anatomical plausibility. The vagus nerve provides a direct, well-documented conduit from gut to brainstem. Retrograde toxin transport via this route is not speculative. It is the mechanism of classical tetanus pathology, just applied to a different anatomical starting point.

6. Vancomycin trial suggestive evidence. Sandler et al. (2000) showed transient behavioral improvement in some ASD children treated with oral vancomycin, with relapse after cessation. This is consistent with an ongoing microbial contribution rather than fixed damage.

Weaknesses and Open Questions

1. No direct detection of TeNT in the gut of autistic children. This is the single biggest gap. Despite decades of interest, no study has demonstrated TeNT protein or biologically active TeNT-like toxin in the intestinal lumen or mucosa of children with ASD. Without this, the entire chain rests on inference.

2. Dosing and survival questions. The gut lumen is a harsh environment. It is full of proteases, extreme pH gradients, and a mucus layer. How much biologically active toxin could survive long enough to reach enteric nerve terminals? The concentrations required for VAMP2 cleavage may be orders of magnitude higher than what commensal-level clostridia can produce locally.

3. No animal model confirmation. No study has shown that chronic low-dose enteric TeNT exposure during development produces an autism-like behavioral phenotype in rodents or any other animal model. This would be the most definitive preclinical test, and it has not been done. What has been done is injecting TeNT directly into the brain (intrahippocampal or intracerebroventricular) in rats, but these are epilepsy models, not autism models.

4. Correlation vs. causation in dysbiosis data. Yes, children with ASD often have elevated Clostridium counts. But ASD children also frequently have restricted diets, altered GI motility, and different medication histories. The elevated Clostridium could be a consequence of behavioral differences, not a cause. This confounding is difficult to resolve without prospective longitudinal cohorts.

5. Genetic heritability dominates. Twin studies consistently estimate ASD heritability at 60–90%. Hundreds of risk genes have been identified (many in synaptic, chromatin remodeling, and transcriptional regulation pathways). Any environmental contributor, including a microbial one, would need to operate within and alongside this strong genetic architecture, likely as a modifier rather than a primary driver.

6. Vancomycin evidence is limited and unreplicated. The Sandler trial was small, unblinded, and has not been robustly replicated. Behavioral improvements from vancomycin could reflect changes in many clostridial metabolites (propionic acid, p-cresol, phenols), not specifically TeNT reduction.

What Would It Take to Test This Properly?

This is where I think the conversation should shift. Instead of debating whether the hypothesis is “right,” we should ask: what experiments would move the needle?

First, we need metagenomics-level surveillance for tetanus toxin gene clusters (tetR, tetX, tent) in pediatric gut samples, particularly in ASD and neurotypical cohorts. Shotgun metagenomics is now affordable enough to do this at scale.

Second, we need sensitive assays for TeNT protein or TeNT-like activity in intestinal fluid or mucosal biopsies - not just stool, but closer to the nerve interface.

Third, we need the animal experiment: colonize germ-free or antibiotic-treated neonatal mice with TeNT-producing clostridia and look for behavioral, electrophysiological, and histological evidence of altered inhibitory neurotransmission during development.

Until those experiments are done, the hypothesis remains exactly that: a hypothesis. An elegant, mechanistically detailed, and biologically plausible one but still unproven.

The Bottom Line

The tetanus toxin hypothesis of autism is a theoretical framework that proposes a specific mechanistic route from gut dysbiosis to neurodevelopmental dysfunction: clostridial overgrowth → local TeNT production → retrograde vagal transport → VAMP2 cleavage in inhibitory interneurons → E/I imbalance → disrupted circuit formation during a critical developmental window.

Each individual link has strong molecular biology behind it. The full chain has never been demonstrated end to end.

The hypothesis’s greatest value may not be its ultimate correctness, but the way it forces us to think mechanistically about gut–brain interactions in neurodevelopment — to move beyond vague correlations and ask exactly how, exactly where, and at exactly what dose a microbial product could alter brain function.

The premise was simple, almost shockingly so. Babies born vaginally emerge coated in their mother’s microbes - a process researchers have affectionately dubbed the “bacterial baptism”. It’s messy, ancient, and according to accumulating evidence, it might be critical. C-section babies, however, miss this entirely. They enter the world pristine, untouched by the microbial ecosystem that has been waiting for them.

So the researchers asked a provocative question: What if we could trick biology? What if we manually swabbed a C-section baby with their mother’s vaginal fluids immediately after birth? Could we fool their immune system into thinking they’d traveled the traditional route?

This concept, known as “vaginal seeding,” ignited a global conversation. It promised a tantalizingly low-tech solution to a high-tech problem. Parents started doing it at home. Scientists started debating it in journals. Medical organizations started warning against it.

Nearly a decade later, the science has evolved from a fascinating theory into something far more complex and far more interesting. Here’s what we actually know now.

The Biological Rationale: Why Would Anyone Want To Do That?

Your body hosts trillions of microbes, and your first major dose happens at birth. It’s not incidental exposure. Rather, it is more like an inoculation - a ‘welcome to the world’ ritual that the vaginal microbes perform for little humans passing through.

When a baby descends through the birth canal, they don’t just pass through; they’re coated in maternal bacteria - Lactobacillus, Prevotella, Sneathia and the likes. These pioneer microbes are thought to train the infant’s developing immune system in its most plastic, receptive state, teaching it the crucial distinction between friend and foe.

C-section babies get a different welcome party. Instead of vaginal microbes, they’re often colonized by whatever’s in the operating room - skin bacteria like Staphylococcus and Corynebacterium, environmental strains and the works.

Why does this matter? Because epidemiological studies have long shown that C-section babies face slightly elevated risks of immune and metabolic disorders such as asthma, allergies and obesity down the line. Not massively higher. Not guaranteed. But consistently, measurably higher.

The hypothesis driving vaginal seeding is elegantly simple: these “missing microbes” are likely the culprit. Restore them, and maybe you close the gap.

The Evidence: From Proof of Concept to Observational Studies and RCTs

The Pilot (2016): Proof of Concept

The study that started it all, by Dominguez-Bello and colleagues, was tiny - just four seeded infants. But it was influential for a reason.

The researchers took sterile gauze, incubated it in the mother’s vagina for an hour before the C-section, then swabbed the baby’s mouth, face, and body immediately after birth. They tracked the babies’ microbiomes for a month.

The result? The seeded babies’ skin and oral bacteria looked more like vaginally born babies than their unseeded C-section peers. Not identical but closer to it. This was enough to raise eyebrows and questions in equal measure.

The Longitudinal Follow-Up (2021): Does It Last?

As interest grew, so did the studies. In 2021, Song and colleagues followed 177 babies over their first year, tracking how their microbiomes evolved.

They found something intriguing: C-section babies who received seeding saw their microbiomes “naturalize” - particularly on the skin and in the gut. Their bacterial communities began tracking more closely with the vaginal-birth trajectory, as if the swab had nudged them onto a different developmental path.

But questions remained. Were these changes meaningful? Did they actually affect health outcomes? And most importantly, were they safe?

The Randomized Controlled Trials (2023): The Gold Standard Weighs In

Then came the studies scientists had been waiting for: randomized controlled trials (RCTs), the gold standard of medical research.

The “Modest” Result: Mueller and colleagues ran a blinded RCT and found that seeding increased the engraftment of maternal bacteria and reduced potential pathogens like Enterobacter in infant stool. Good news, right?

Not so fast. The seeded infants also showed reduced gut diversity compared to controls. It is worthy of note that this reduction isn’t necessarily bad. Breastfed vaginally-born infants typically have lower diversity early on too, dominated by specific beneficial bacteria. So consider this more like a feature and not a bug.

The “Neuro” Surprise: Then Zhou and colleagues dropped a bombshell with their triple-blind RCT.

C-section infants who received vaginal seeding scored significantly higher on neurodevelopmental tests involving communication and motor skills at six months compared to those who received a saline placebo. This finding sent shockwaves through the field, further splitting scientists on either side of the debate - understandably so, because what could be a bolder claim than vaginal seeding influencing the brain development of infants?

The Counter-evidence: Are The Claims Actually Replicable?

Just when you thought the science was trending positive, Liu and colleagues joined with a 2023 randomized clinical trial that throws cold water on the entire enterprise.

Their study followed 120 infants for two full years - the most rigorous long-term investigation of vaginal seeding to date. And their findings were very profound.

The “Microbiome Makeover” Never Materialized

The fundamental promise of vaginal seeding is straightforward: coat the baby in mom’s bacteria, colonize their gut, change their microbial destiny. Liu’s team tested this directly, analyzing infant gut bacteria at birth, 6, 12, 18, and 24 months.

The result? Vaginal seeding made no real difference.

The overall structure and diversity of gut bacteria were essentially identical between swabbed babies and those who received standard care. Yes, there was a tiny uptick in specific bacteria such as Lactobacillus and Bacteroides at birth and 6 months. But these differences were subtle, almost whisper-quiet in the data. And by the time the children were toddlers, they’d vanished entirely.

The “seed,” quite literally, didn’t take root.

It Didn’t Prevent the “C-Section Penalty”

Here’s where it gets even more interesting. Proponents argue that seeding should prevent the well-documented health risks associated with C-section birth: asthma, allergies, obesity. This was the real test.

Liu’s team measured everything over two years:

Allergies: At 18 months, they tested for immune reactions to 20 common allergens. There was no statistical difference between C-section babies that were seeded and those who were not.

Weight and Metabolism: They tracked BMI and age-adjusted BMI scores at multiple time points. At the 6-month mark, the seeded group did have fewer overweight infants - a finding that briefly sparked hope. But then 12 months rolled around. Then 18. Then 24. The effect had completely disappeared. The authors suggest, with characteristic scientific understatement, that this was likely a statistical fluke rather than a real protective benefit.

Why the Intervention Likely Failed: A Reality Check

Liu’s team offers a sobering explanation that cuts to the heart of the matter: a 20-second swab is not birth.

During vaginal delivery, the baby experiences hours of microbial exposure. They’re not just briefly touched by bacteria; they’re bathed in it repeatedly as they move through the birth canal. They ingest amniotic fluid mixed with maternal microbes. Their skin is compressed against vaginal walls for extended periods.

A swab, even a carefully applied one, can’t replicate this. It’s like comparing a cup of water thrown at you versus swimming in an ocean. The scale, duration, and intensity of exposure are fundamentally different.

Moreover, the researchers note that babies aren’t just passive recipients of bacteria. The process of labor itself - the stress hormones, the physical compression, the timing of first breath, all of that, may prime the infant’s immune system to accept and integrate those microbes in ways that a post-surgical swab simply cannot trigger.

In other words, we might have been trying to hack biology with a shortcut that biology doesn’t recognize.

New Perspective: Are We Seeding the Wrong Thing?

One argument I want to make here is that perhaps, we might be swabbing from the wrong source entirely. The vagina might not be the main actor here. A work that I recently reviewed was from one of my hires who freshly completed her PhD at my alma mater, University of Toronto. In her brilliant work, she showed that the faecal microbiomes of vaginally delivered newborns, from day zero until six months post-delivery, looked more like their mother’s faecal microbiome than vaginal microbiome.

Think about the mechanics of vaginal birth. Yes, the baby passes through the vagina. But there’s also inevitable exposure to maternal fecal microbes such as Bacteroides and Bifidobacterium. Thanks to sheer anatomical proximity. Yes, these gut-associated bacteria are notoriously depleted in C-section babies. But here’s the kicker - studies show that vaginal seeding often fails to restore these specific populations because they aren’t abundant in the vagina to begin with.

The work of Korpela and colleagues in 2020 strongly supports this. They used maternal fecal microbiota transplantation (FMT) in Cesarean-born infants to rapidly restore normal gut microbiome development. Yes, you read that correctly. A pilot study tested feeding C-section babies a tiny, diluted amount of maternal stool mixed in breast milk.

The results were striking. FMT restored the gut microbiome much more rapidly and completely than vaginal seeding ever did. It’s gross and controversial. But it might be more biologically accurate than the original intervention.

The Safety Debate: “DIY” vs. Medical Supervision

This is where things get serious and personal for many parents.

The Case for Caution

Major medical organizations, including the American College of Obstetricians and Gynecologists (ACOG), have issued clear warnings: do not perform vaginal seeding outside of research protocols.

The reason? Infection risk.

A mother can be an asymptomatic carrier of pathogens that are harmless to her but devastating to a newborn: Group B Streptococcus (GBS), Herpes Simplex Virus (HSV), Chlamydia, Gonorrhea. Transferring these to a newborn’s eyes, mouth, and open airways can be catastrophic.

In fact, there is a documented case of a healthy infant developing severe herpes infection following a “DIY” vaginal seeding procedure. The parents had no idea the mother had a history of cold sores - HSV-1, typically benign in adults.

The Argument for Safety (In Clinical Settings)

In the clinical trials (not the DIYs) mentioned above, safety was paramount. Mothers underwent rigorous screening for STIs and other pathogens that went beyond standard prenatal care.

Under these strict protocols, no serious adverse events related to seeding were reported across hundreds of participants. So it appears that the procedure can be safe but only with comprehensive screening, medical supervision, and careful execution. Home experiments are a different story entirely.

What We’re Still Waiting For

We’re in what scientists call the “wait and see” phase, which admittedly, isn’t very satisfying.

We know vaginal seeding has been reported to change the microbiome. There are hypotheses that it might help neurodevelopment. But we still don’t know if it actually prevents the big three C-section-associated risks - obesity, asthma, and autoimmune diseases.

Ongoing trials in the United States are tracking these exact outcomes. Until those results come in, likely in the next few years, vaginal seeding remains firmly in the “experimental procedure” category, not a standard medical recommendation.

The Bottom Line (For Now)

If you’re a parent planning a C-section and wondering whether to try vaginal seeding, here’s the honest answer: we don’t have enough evidence yet to recommend it routinely.

What we do have solid evidence for:

• Breastfeeding helps correct microbial imbalances in C-section babies

• Avoiding unnecessary antibiotics (for both mother and baby) preserves healthy microbiome development

• Skin-to-skin contact immediately after birth supports beneficial bacterial colonization

The story of vaginal seeding is a perfect illustration of how science actually works - messy, contradictory, evolving. It started with a compelling biological hypothesis. Early studies showed promise. Then a rigorous trial suggested it might not work at all. Now we’re in the uncomfortable middle ground, waiting for more data.

But here’s what makes this saga so fascinating: even if vaginal seeding itself doesn’t pan out, the questions it raised have revolutionized how we think about birth. The idea that those first microbial exposures might shape a child’s health for years and maybe even their brain development has opened entirely new fields of research.

The bacterial baptism, it turns out, might matter more than we ever imagined. We’re just still figuring out whether we can replicate it with a swab.

Or whether we should even try at all.

Based on clinical trials and meta-analyses, approximately 9% to 12% of statin users may experience an increased risk of developing type II diabetes and/or insulin resistance.

The onset and exacerbation of insulin resistance and hyperglycemia in statin users is aided and driven by changes in the gut microbiome.

Clinical studies have shown that statins directly inhibit the growth of the gut bacteria — Clostridium species, in a dose-dependent manner. Why does this matter?

Because, these Clostridium species encode enzymes that transform a primary bile acid -Chenodeoxycholic acid (CDCA) to a secondary bile acid, ursodeoxycholic acid (UDCA).

Therefore, the inhibition of Clostridium species by statins leads to the depletion of UDCA. This secondary bile acid — UDCA, is an important player here.

UDCA is a signaling molecule for GLP-1 production. It activates TGR5 receptors in the colon to produce GLP-1.

Reduced UDCA levels lead to downregulated GLP-1 levels. Increased insulin resistance and elevated HbA1c levels.

If you have read this post to this point, then you have the high-level summary of this newsletter. However, if you are someone who likes ‘nerding’ through raw scientific data, then you may find the next few paragraphs enchanting. Otherwise, please feel free to skip to the last paragraph of this article.

Evidence From Human Study

The figure below illustrates the results of a well-conducted clinical study observing human patients taking atorvastatin (a common statin) over 16 weeks.

In simple terms, the figure tells a story of “good news” regarding cholesterol and “bad news” regarding blood sugar regulation. Here is a breakdown of the findings:

1. The Study Setup (Panel A): The researchers followed 30 patients taking atorvastatin and compared them to 10 healthy control subjects who were not taking the drug. They monitored them for 16 weeks.

2. The Drug Worked for Cholesterol (Panels B & C): The statin did exactly what it was supposed to do for heart health. Within just one week of starting the medication, the patients’ total cholesterol and LDL (often called “bad cholesterol”) dropped significantly.

3. The Drug Negatively Affected Blood Sugar (Panels D, E, F, & G): While cholesterol went down, markers related to diabetes risk went up:

   • HbA1c (Long-term blood sugar): This marker increased significantly starting at week 4 and remained high through week 16.

   • Insulin Resistance: The patients’ insulin levels and C-peptide levels increased. Consequently, their HOMA-IR score (a calculation used to measure insulin resistance) also went up. This indicates their bodies were struggling to process sugar efficiently.

4. A Key Hormone Dropped (Panel H): The researchers found a potential cause for the blood sugar issues. A hormone called GLP-1 (which helps the body release insulin and control sugar) decreased significantly 4 weeks after the patients started taking the statin.

Essentially, this figure shows that while atorvastatin successfully lowered cholesterol, it simultaneously caused the patients to develop signs of insulin resistance and glucose intolerance, likely because the drug suppressed the beneficial hormone GLP-1.

Statin Usage Disrupts The Microbiome In A Way That Permits Metabolic Side Effects

These metabolic defects did not occur without a cause. They were aided by changes in the gut microbiome composition and function. In subsequent figures (2 & 3), the researchers analyzed the faecal content of the statin users and their healthy controls. It was seen that the onset of the metabolic defects reported at week 4 (as shown in figure 1 above) coincided perfectly with the depletion of Clostridium species and UDCA (the secondary bile acids that Clostridium produces) among statin users, further emphasizing that the gut microbiome mediates the metabolic side effects seen in the statin user group.

UDCA Supplementation Ameliorates Metabolic Side Effects In Statin Users.

The researchers utilized specific interventions to counteract the insulin resistance and hyperglycemia caused by statin therapy. The interventions focused on restoring the “Clostridium–bile acid axis,” which statins disrupt. If this ‘restoration’ reverses the defective metabolic markers (HbA1c, GLP-1, serum insulin and C-peptide), it would mean that the gut microbiome function was indeed playing a role in the side effects experienced by some statin users.

Here is a summary of what was done to ameliorate these side effects in both humans and mice:

In Humans

• UDCA Supplementation: The researchers administered ursodeoxycholic acid (UDCA) to patients who had been on statins for over 6 months and exhibited elevated HbA1c levels.

  • Dosage: Patients received 500 mg of UDCA per day alongside their continued statin therapy.

  • Results: After a 2-month follow-up, the patients showed substantially decreased levels of HbA1c, serum insulin, and C-peptide, alongside increased levels of the beneficial hormone GLP-1.

  • Safety & Efficacy: Crucially, this combination therapy improved glucose homeostasis without impairing the lipid-lowering effect of the statin (cholesterol levels remained controlled). Liver function markers (ALT and AST) also showed improvement.

In Mice

• UDCA Supplementation: Similar to the human trial, mice fed a high-fat diet with atorvastatin were given oral UDCA (50 mg/kg) for 4 weeks. This treatment rapidly restored GLP-1 secretion and significantly improved glucose intolerance and insulin resistance.

• Bacterial Transplantation (Clostridium sp.): The researchers transplanted specific Clostridium species (which statins usually kill) into statin-treated mice.

  • Mechanism: This colonization restored the production of enzymes necessary to create bile acids.

  • Results: The transplant reversed statin-induced hyperglycemia, restored GLP-1 levels, and increased energy expenditure and oxygen consumption.

The Bottom Line: Your Gut Microbes Are Speaking. Are You Listening?

Here’s the uncomfortable truth that millions of statin users need to hear — the very drug protecting your heart may be quietly dismantling the microbial machinery that keeps your blood sugar in check. And the most unsettling part? Most people and most doctors aren’t looking for it.

This isn’t anti-statin fearmongering. Statins save lives. The data on cardiovascular protection is robust and undeniable. But science doesn’t ask us to choose between protecting the heart and protecting metabolism. It asks us to be smarter — to see the whole system, not just the cholesterol panel.

The research is strikingly clear: statins deplete Clostridium species, UDCA drops, GLP-1 falls, and then insulin resistance rises. It’s not a mystery. It’s a chain reaction and every link in that chain is now a potential point of intervention.

If you’re on a statin, don’t stop taking it. But start asking better questions. Get your HbA1c monitored regularly. Talk to your physician about UDCA supplementation. Pay attention to your gut. Because the difference between a statin that only protects your heart and one that silently pushes you toward type II diabetes may come down to something as ancient and overlooked as the bacteria living in your colon.

A new study published in Cell Metabolism brings this notion to the forefront. The paper, titled “The gut microbiota shapes the human and murine volatilome“ provides definitive evidence that the bacteria residing in your intestines directly influence the chemical composition of the air you exhale.

For years, science has chased the idea of “breathprinting” – using exhaled chemicals to diagnose diseases like cancer or diabetes. However, a major piece of the puzzle was missing: where exactly are these compounds coming from? This new research confirms that your breath is not just a byproduct of your own lungs and metabolism. It’s also a broadcast signal from your gut microbiome.

The Goal: A Window into the Gut

The primary objective of this study was to determine if the chemicals in our breath could reveal what’s happening inside our intestines.

Current methods for analyzing the microbiome like sequencing stool samples are slow and require unpleasant collection methods. The researchers hypothesized that since gut microbes produce gases that absorb into the bloodstream, travel to the lungs, and get exhaled, then the composition of exhaled air could be a proxy tool for monitoring gut health.

Imagine skipping the awkward stool sample and just breathing into a tube instead – that’s the promise here.

The Findings: Humans, Mice, and Microbes

The study was divided into two distinct phases: a correlational study in humans and a causative study in mice.

1. The Human Connection

The team recruited 27 healthy children (ages 6–12) and collected paired samples: their breath and their stool. By sequencing the DNA in the stool and running the breath through specialized equipment that can identify individual chemicals in breath samples, they looked for patterns.

They found a strong, statistically significant correlation between the specific bacteria in a child’s gut and the chemicals in their breath. This wasn’t just a vague association; specific bacteria were linked to specific gases:

• Isoprene: A common compound in human breath, found to correlate with the abundance of Roseburia bacteria and Asaccharobacter celatus.

• Acetic Acid: Linked to Bacteroides ovatus, a known fermenter in the gut.

• Terpenes: Here’s where it gets interesting. Terpenes are compounds usually associated with plants. Think of that fresh pine or citrus smell. They found these plant-derived scents in human breath, shaped by our gut bacteria breaking down what we eat. This suggests that our gut microbes may play a major role in metabolizing dietary compounds into breath signatures.

2. Establishing Causality: The “Clean” Mouse

Correlations in humans are fascinating, but they don’t prove that bacteria cause the breath changes. To prove this, the researchers turned to germ-free mice – that is, animals raised in completely sterile conditions with no bacteria whatsoever.

They developed a novel engineering solution: a custom breathing apparatus that collects air from deep in the lungs via a tube inserted into the trachea. This was critical because it bypassed the mouth, ensuring that the chemicals collected came from the internal system, not from oral bacteria or food residue stuck in the teeth.

When they performed fecal transplants – which is moving the microbiome of a “conventional” mouse into a germ-free mouse, the recipient’s breath profile shifted to match the donor’s. This proved that the microbiome is sufficient to reshape the chemical components of exhaled air.

The Most Compelling Evidence: From Test Tube to Breath

Here’s where the researchers did something clever: they wanted to prove beyond doubt that specific bacteria create specific gases that end up in breath. So they ran a three-step test:

Step 1: Grow individual gut bacteria (like E. coli and Collinsella aerofaciens) in sealed jars and measure what gases they produce.

Step 2: Put those same bacteria into sterile mice.

Step 3: Check if the mice now breathe out those exact same gases.

And it worked for several key compounds:

• Ethyl Acetate: E. coli produced this in the jar, and mice colonized with E. coli exhaled high levels of it.

• Toluene: Often thought of as an industrial pollutant or paint thinner, this compound was produced by C. aerofaciens sitting alone in a culture dish and then found in the breath of mice carrying that bacterium.

• Benzothiazole: Enriched in both E. coli cultures and in the breath of E. coli-colonized mice.

This is the smoking gun: a direct line from a specific microbe to a specific chemical signature in breath.

The pathway is clear: Microbe (Gut) → Chemical (Blood) → Exhalation (Lungs).

Translational Potential: The Future of Diagnostics

The implications of this work are significant for clinical medicine. To demonstrate proof of principle, the authors looked at a small group of children with asthma. Previous research has shown that asthmatic children often have an abundance of a bacterium called Eubacterium siraeum in their gut.

Using only the breath data (specifically a panel of 4 to 10 chemicals), the researchers were able to predict the abundance of E. siraeum in the children’s guts with reasonable accuracy.

This suggests a future where:

• Neonatal Care: Picture a scenario where rather than collecting stool samples from sick premature babies (which can be difficult and stressful), nurses could simply analyze the air in their incubators to detect dangerous gut infections days before the baby shows symptoms. Conditions like necrotizing enterocolitis could be caught earlier, when treatment is most effective.

• Gut Health Monitoring: Patients with inflammatory bowel disease or those recovering from antibiotics could blow into a device to check if their “good bacteria” have repopulated, using markers like ethyl acetate. No shipping stool samples to a lab. No waiting weeks for results.

Limitations and Caveats

While this paper is a technological tour de force, there are limitations that warrant caution before we expect a “Gut Breathalyzer” at the pharmacy.

• The “Noise” of Real Life: The human study was small – 27 healthy children and 14 with asthma. While the children fasted for two hours before testing, diet and oral hygiene weren’t strictly standardized. Food is a massive source of breath chemicals (especially those terpenes), and oral bacteria produce gases too. Distinguishing a “gut signal” from a “lunch signal” or a “mouth signal” in the general population will be difficult.

• Indirect Mechanisms: Not every change in breath is a gas produced directly by bacteria. Some breath chemicals appear because bacteria change how your own body processes things – like how your liver handles certain compounds. The study notes that microbes can alter the host’s metabolism, which secondarily changes the breath. Telling these two pathways apart isn’t easy yet.

• Mouse Stress: The collection method for mice involved anesthesia and a breathing tube. Physiological stress can alter metabolism, potentially skewing breath profiles in ways that might not reflect normal conditions.

• Complexity of Dysbiosis: The study successfully predicted E. siraeum abundance, but predicting complex community shifts involving hundreds of species interacting simultaneously is a much harder computational challenge. Real gut problems rarely involve just one bacterium going rogue.

The Verdict

The authors have successfully laid the biological foundation for breath-based microbiome diagnostics. They’ve moved the field from vague correlations (”sick people smell different”) to specific biological attributions (”this bacterium produces this gas”).

Here’s the thing: this work implies that we are constantly exhaling a stream of data about our internal inhabitants. With the right sensors, our breath could become the most convenient window we have into the complex world of the gut microbiome. For now, these are possibilities. Ultimately, the future will tell!

Reference:

1. Hernandez-Leyva, A. J., et al. (2026). The gut microbiota shapes the human and murine breath volatilome. Cell Metabolism, 38, 1–15.

This specialized microbial network is known as the estrobolome.

And once you understand it, women’s health looks very different.

From fertility challenges and endometriosis to cancer risk and the menopausal transition, the estrobolome offers a new biological lens through which we can understand hormone balance; not as a purely endocrine process, but as a host–microbe partnership.

Your hormones don’t act alone. Your gut has a vote.

The Recycling Center: What Is the Estrobolome?

To understand the estrobolome, we first need to understand how the body normally processes estrogen.

Estrogens circulate through the bloodstream, acting on tissues like the breast, brain, uterus, bones, and cardiovascular system. Eventually, the liver filters these hormones and converts them into a conjugated form — essentially packaging them so they become inactive and water-soluble. These packaged estrogens are secreted into bile and delivered into the intestine for elimination.

In a simple system, that would be the end of the story: hormones in, hormones used, hormones out.

But the gut microbiome rewrites the ending.

Certain gut bacteria produce an enzyme called β-glucuronidase. This enzyme “unpacks” (deconjugates) estrogen, converting it back into its active form. Once reactivated, estrogen can be reabsorbed through the intestinal wall, enter the portal circulation, and recirculate throughout the body — a process known as enterohepatic circulation.

In other words, your gut microbes decide whether estrogen is:

• eliminated as waste

• or recycled back into your bloodstream

Your microbiome functions less like plumbing… and more like a hormone recycling plant.

That microbial machinery is the estrobolome.

Healthy vs. Unhealthy Estrobolome

A healthy estrobolome acts as a finely tuned regulator. It maintains hormonal homeostasis allowing enough estrogen recycling to support physiological needs like bone density, fertility, and cognitive health, without allowing levels to become excessive or pathological. This state is typically associated with high microbial diversity, metabolic resilience, and strong gut barrier integrity.

A dysregulated estrobolome, however, reflects microbial imbalance (dysbiosis) and reduced diversity. And this imbalance can push estrogen biology in two dangerous directions:

1) Hyperestrogenic State

Excessive β-glucuronidase activity leads to increased estrogen reactivation and reabsorption, raising circulating estrogen levels. This is associated with estrogen-driven proliferative conditions such as:

• endometriosis

• uterine fibroids

• estrogen-receptor-positive breast cancer

• endometrial hyperplasia

2) Hypoestrogenic State

Insufficient estrogen recycling leads to excessive elimination, contributing to low-estrogen states associated with:

• metabolic dysfunction

• obesity

• insulin resistance

• bone loss

• cognitive decline

• cardiovascular risk

Same system. Opposite pathologies.

Key Microbial Players

While the estrobolome is a functional system rather than a single species, certain microbial patterns repeatedly emerge:

Healthy State

A balanced microbiome is typically dominated by Bacteroidetes and Firmicutes, with high overall diversity. Beneficial genera such as Lactobacillus and Bifidobacterium support gut barrier integrity, immune regulation, and metabolic balance.

Notably, Bifidobacterium animalis has been associated with healthier estrogen profiles in menopausal women and correlates positively with estradiol levels.

Dysbiotic States

PCOS
Enrichment of Bacteroides vulgatus, Escherichia, and Shigella, with depletion of protective taxa like Akkermansia and Ruminococcaceae.

Breast Cancer
Increased Clostridia and Ruminococcaceae, taxa associated with higher β-glucuronidase activity and elevated estrogen recycling.

Endometrial & Cervical Cancer
Overrepresentation of Prevotella and Porphyromonas, with depletion of Lactobacillus species that normally dominate healthy vaginal and gut ecosystems.

These patterns suggest that cancer risk is not only genetic or hormonal. It is also microbial.

The Estrobolome in Disease States

Because estrogen receptors are distributed throughout the body, estrobolome dysfunction has systemic consequences.

Polycystic Ovary Syndrome (PCOS)

PCOS is a complex endocrine-metabolic disorder marked by insulin resistance, hyperandrogenism, and chronic inflammation. Women with PCOS consistently show reduced gut microbial diversity.

Dysbiosis increases gut permeability, allowing lipopolysaccharides (LPS) to enter circulation, driving inflammation and insulin resistance. This metabolic inflammation stimulates ovarian androgen production, worsening hormonal imbalance.

This is not just a reproductive disorder. It is a gut-immune-endocrine disorder.

Endometriosis

Endometriosis is an estrogen-dependent condition. A gut microbiome with high β-glucuronidase activity increases estrogen recycling, maintaining elevated systemic estrogen levels that fuel lesion growth and inflammation.

Animal and primate models show reduced Lactobacillus and increased Gram-negative bacteria, reinforcing the gut-immune-estrogen axis in disease progression.

Cancer

Estrogen-receptor-positive breast cancer is particularly sensitive to estrogen recycling. A microbiome that aggressively reactivates estrogen increases systemic exposure and oncogenic signaling. Dysbiosis has also been observed in ovarian, cervical, and endometrial cancers, suggesting the estrobolome may function as both a risk modifier and a biomarker system.

The Estrobolome and the Menopausal Transition

Menopause is not just a hormonal transition; it is a microbial transition.

As ovarian estrogen production declines, the body becomes more dependent on peripheral estrogen production and microbial recycling. Yet paradoxically, the postmenopausal gut microbiome becomes less diverse and shifts toward a microbial profile more similar to that of males.

Beneficial taxa such as Bifidobacterium animalis decline, correlating with reductions in circulating sex hormones. This microbial shift reduces estrogen reabsorption, exacerbating:

• vasomotor symptoms

• bone loss

• cardiovascular risk

• cognitive decline

• systemic inflammation

Menopause, therefore, is not only endocrine, it is endo-microbial.

Maintaining a Healthy Estrobolome

The most empowering truth about the estrobolome is this: it is modifiable.

1) Prioritize Fiber and Prebiotics

Diet is the dominant driver of microbial composition. Fiber supports microbial diversity and short-chain fatty acid production (especially butyrate), strengthening the gut barrier and regulating estrogen metabolism.

2) Targeted Probiotics

Specific Lactobacillus and Bifidobacterium strains improve metabolic and hormonal markers in PCOS and improve vascular function in postmenopausal women.

3) Phytoestrogens and Soy

Soy isoflavones can be converted by certain gut bacteria into equol, a potent estrogenic and antioxidant compound. Only some women have equol-producing microbiomes but dietary exposure can help select for them.

4) Minimize Antibiotics and Ultra-Processed Foods

Both reduce microbial diversity and impair estrogen metabolism, disrupting estrobolome function.

The Estrobolome Reframes Women’s Health

The estrobolome shows us that hormone balance is not just about glands and receptors, but also about microbes, enzymes, metabolism, immunity, and ecology. Your gut is not just digesting food; it is actively shaping your endocrine system.

Whether navigating fertility, managing estrogen-driven disease, or transitioning through menopause, the gut microbiome offers a powerful, non-invasive lever for hormonal health.

You are not just your hormones.
You are your hormones plus your microbes.
And your microbes are doing far more work than anyone ever gave them credit for.

When fat enters the small intestine, the gallbladder contracts and releases bile. Most bile acids are released in a conjugated form, meaning they are chemically linked to a small helper molecule – usually taurine or glycine. This modification makes bile acids more water-soluble and ready to do their work.

But before bile acids can be recycled and reused by the body, they must be “unlocked.” This unlocking step is called deconjugation and this is where the gut microbiome becomes essential.

Where Gut Microbes Step In

A group of beneficial gut microbes, primarily residing in the small intestine, produce a specialized enzyme called bile salt hydrolase (BSH). BSH removes the helper molecule from conjugated bile acids, freeing them so they can be reabsorbed and reused by the body.

This process is mutually beneficial. Humans gain more efficient fat digestion and nutrient absorption, while microbes gain access to bile acids that help shape their ecological niche in the gut.

Roughly 90% of bile acids are reabsorbed in the small intestine and returned to the liver through a tightly regulated recycling system known as enterohepatic circulation – one of the most elegant efficiency systems in human physiology.

The Trip to the Large Intestine

The remaining ~10% of bile acids escape reabsorption and travel into the colon (large intestine). This is where things get especially interesting.

Certain gut microbes in the colon convert primary bile acids into secondary bile acids – a biochemical upgrade with profound metabolic consequences. These secondary bile acids are not just digestive byproducts; they act as signaling molecules.

Some secondary bile acids activate a receptor called TGR5, found on intestinal cells known as L-cells. When TGR5 is activated, it triggers the release of GLP-1, a hormone that:

• Reduces appetite

• Improves blood sugar control

• Enhances insulin sensitivity

• Supports overall metabolic health

In simple terms, your gut microbes transform leftover bile acids into metabolic signals that help regulate weight and glucose balance.

Why This Matters

When the gut microbiome is balanced, the right microbes are present to:

• Properly deconjugate bile acids

• Convert primary bile acids into beneficial secondary bile acids

• Stimulate GLP-1 production naturally

When the microbiome is disrupted – by poor diet, chronic stress, infections, or medications, these conversions decline. The metabolic benefits are blunted, and the signaling system becomes less effective.

Better microbiome → better bile acid transformation → better metabolic health.

Secondary Bile Acids and Weight Health: Supporting Evidence

Bariatric Surgery, Secondary Bile Acids, and the Microbiome

Sleeve gastrectomy is a weight-loss surgery in which approximately 70–80% of the stomach is removed, creating a narrow, tube-shaped stomach. The essence of the procedure is durable weight loss driven by both stomach restriction and hormonal changes, including reduced hunger due to lower ghrelin production.

In a 2025 study, Zhao and colleagues followed 108 obese adults for 12 months after sleeve gastrectomy. Compared to controls, post-surgery participants showed:

• Reduced levels of primary bile acids

• Increased levels of conjugated secondary bile acids, including glycodeoxycholic acid, taurodeoxycholic acid, and taurolithocholic acid
  

Higher levels of these secondary bile acids were strongly associated with lower BMI, body weight, and waist circumference.

Crucially, these bile acid changes tracked closely with shifts in the gut microbiome. There was a selective increase in members of the Clostridia class – microbes known to carry enzymes involved in bile acid deconjugation and secondary bile acid synthesis. Higher abundance of these microbes correlated with improved weight and metabolic markers.

To establish causality, researchers transplanted fecal microbiota from post-surgery patients into mice. The recipient mice accumulated less body fat, confirming that changes in the microbiome and the secondary bile acids it produces actively contribute to weight regulation.

Targeted Colonic Delivery of Bile Acids

In a 2020 study, Calderon and colleagues investigated whether delivering a conjugated secondary bile acid – taurodeoxycholate (TDCA), directly to the colon could influence metabolism.

They found that colonic delivery of TDCA:

• Increased post-meal GLP-1 levels

• Promoted weight loss in individuals with obesity and type 2 diabetes

The proposed mechanism is elegant: delivering bile acids directly to the colon places them close to L-cells, where they can activate TGR5 receptors on the basolateral membrane and trigger GLP-1 release.

While not yet standard clinical practice, this work highlights the therapeutic potential of bile acid signaling when properly targeted.

UDCA and Metabolic Health

Another secondary bile acid, ursodeoxycholic acid (UDCA), has also been studied in metabolic disease.

In one trial, obese women with non-alcoholic fatty liver disease (NAFLD) followed a 1,200-kcal/day diet and were randomized to receive either UDCA or placebo. After six weeks:

• Both groups lost similar amounts of weight

• Only the UDCA group showed significant improvements in cholesterol and glucose markers

This suggests that UDCA can enhance metabolic health during weight loss, even if it does not independently amplify weight reduction.

Optimizing the Gut Microbiome to Support Weight Health

The goal is not simply “more bile acids,” but better bile acid signaling—specifically, biasing signaling toward TGR5 activation (which enhances GLP-1) while avoiding excessive activation of FXR, a receptor that can suppress GLP-1 signaling.

Plausible strategies include:

A. Diverse fermentable fibers
A mix of soluble and insoluble fibers supports microbial groups involved in bile acid transformation and short-chain fatty acid production. Together, these influence FXR and TGR5 signaling and overall energy balance. Aim for fibers from legumes, oats, barley, flax, vegetables, and low-sugar fruits. Ensure they are introduced gradually to minimize GI discomfort.

B. Moderate, unsaturated fat intake
Bile acids are released in response to dietary fat. However, chronic very high-fat diets – especially those rich in saturated fats, can disrupt bile acid composition and promote obesity in animal models. Emphasize unsaturated fats such as olive oil, nuts, seeds, and fish while avoiding sustained excess saturated fat intake.

C. Probiotics, prebiotics, and targeted modulation
Although human data are still emerging, several microbial strains – particularly Lactobacillus and Bifidobacterium species, possess bile salt hydrolase activity. These microbes can influence bile acid pools and signaling pathways, offering a promising route for microbiome-guided metabolic support.

Gut Microbes are Cornerstones of Weight Health & Metabolic Intelligence

What emerges from this growing body of evidence is a powerful idea: the gut microbiome does not just respond to weight loss, rather, it also helps to drive it. By transforming bile acids into metabolic signals like GLP-1, gut microbes act as unseen partners in weight regulation.

As research continues to unravel these pathways, one thing is clear: supporting a healthy, diverse microbiome goes beyond gut health. It is about metabolic intelligence. And bile acids may be one of the most underappreciated messengers in that conversation.

So, although practitioners know that microbiome testing can meaningfully shape the way they deliver care, the barrier to adoption remains high. Practitioners’ needs have moved far beyond simply knowing what to tell patients to eat or avoid. What they need now is clarity on how much, how long, which prebiotic or antimicrobial, what should be avoided, what is contraindicated, and essentially, a stepwise action plan that is truly practicable.

When you run a gut test, the result should guide how interventions are approached. Otherwise, it is useless. The purpose of testing is to shine light on the problem you are trying to solve so that you’re not shooting darts in the dark. You should not support Patient A, who has a high pathogen/pathobiont signature, the same way you would support Patient B, who merely needs optimization and has no significant overgrowth. Microbiome-based interventions are not “one-size-fits-all,” and the frustration clinicians face stems from using the same protocol across entirely different patient archetypes. More often than not, the outcomes don’t move the needle, and even when they do, they sometimes make things worse.

To truly personalize interventions, protocols must be built on a gut healing framework. Think of this framework as a toolbox, not a rigid sequence. Some gut profiles will require more tools than others, and even when they require the same tools, the order may differ slightly.

The 7R Framework of Gut Healing

No two guts are alike. Protocols must be tailored to each person’s microbiome phenotype. That’s why I created the 7R Framework, an adaptation of the 4R framework commonly used in functional medicine for gut restoration. This is a structured, high-resolution approach to gut healing that integrates stool microbiome data, clinical symptoms, medical history, and targeted dietary and lifestyle changes.

Each “R” is not a rigid step but a phase that adapts to the individual’s phenotype – dysbiosis type, motility status, inflammatory load, histamine sensitivity, bile acid patterns, metabolic dysfunction, and more. Again, this framework is a toolbox, not a linear checklist. Some phenotypes will require four Rs; others may need five or six. It all depends on who is sitting in front of you.

Because there is a lot to unpack, I will discuss this framework at a high level and emphasize why each component is clinically important. I’ll drill down into each R in future posts.

R1 – Reset Motility & Secretions (Prime)

“Before you fix a house, you clear the hallways and open the windows.”

Gut healing begins with flow. When motility is sluggish or gastric secretions are compromised, everything else stalls – fermentation increases, pathogens overgrow, and nutrients are poorly absorbed.

This phase supports the Migrating Motor Complex (MMC), the electrical “sweeper wave” that clears debris between meals and ensures adequate digestive secretions.

Why this matters clinically

Optimizing motility reduces stasis, gas, reflux, and SIBO tendencies. It increases resilience against opportunistic blooms commonly seen on microbiome tests – particularly Enterobacteriaceae, Proteobacteria, and methane-producers.

R2 – Reduce / Remove (Targeted)

“Kill what harms, but don’t scorch the earth.”

This phase strategically targets pathobionts. Not aggressively or indiscriminately. Over-treatment destabilizes an already fragile ecosystem, so precision matters.

It is strongly discouraged to run a “kill phase” without motility support; otherwise, dead microbial debris accumulates, and patients feel worse.

R3 – Re-seed (Strain-Specific Probiotics)

“Rebuild the ecosystem with the right citizens—not just more citizens.”

Probiotic selection should be phenotype-driven, not guided by brand popularity. When properly personalized, probiotics can be transformative; when not, they can be counterproductive.

For example:

• Histamine-sensitive individuals should avoid histamine-producing strains.

• Certain strains are well-studied for constipation; others are not.

Importantly, probiotics rarely colonize permanently – they influence digestion, immune signaling, gut-brain communication, and inflammatory tone.

R4 – Re-feed (Prebiotics, Fiber & Polyphenols)

“If you plant a garden but never feed it, nothing grows.”

Once the ecosystem is quieter and more stable, you feed beneficial microbes so they flourish and outcompete opportunists. This phase rebuilds richness, diversity, and SCFA production. Low-gas fibers such as resistant starches (RS2 or RS3) and PHGG are typically well-tolerated. Inulin, GOS, and FOS can be layered later.

Do not introduce prebiotics early in severe dysbiosis – they can feed the wrong organisms.

R5 – Repair (Barrier & Mucosal Healing)

“The gut lining is the firewall; once breached, everything downstream suffers.”

When the epithelial barrier is compromised, bacterial byproducts, food antigens, and inflammatory triggers cross into circulation. Repairing this layer is non-negotiable.

Core components of gut repair include:

• Short-chain fatty acid support (tight junction regulation)

• Mucosal repair

• Anti-inflammatory nutrients

• Micronutrients that modulate immune function and tissue regeneration

R6 – Rebalance Bile/Acids & Redox

“The gut is a biochemical factory – when bile acids, sulfur, and redox pathways misfire, metabolism follows.”

This phase addresses deeper biochemical patterns: bile acid metabolism, sulfur/sulfide pathways, and oxidative stress.

Stool chemistry and microbiome data often reflect these imbalances. For example, low levels of Akkermansia or certain species of Bifidiobacterium can signal poor bile acid recycling. Similarly, oxidative stress may arise from excessive LPS levels in the gut, and steatorrhea, which is excessive fat in poop, may indicate low bile acid activity.

This phase directly influences glucose regulation, lipid metabolism, and inflammation.

R7 – Resilience & Relapse-Proofing

“Gut health is not an event. It’s a rhythm.”

Once the gut stabilizes, the focus shifts to maintenance – dietary diversity, sleep, exercise, and healthy lifestyle patterns. The goal is to build an ecosystem that is self-correcting – one that rebounds smoothly from stress, travel, dietary shifts, or infections.

The Gut Does Not Heal By Chance. It Heals By Strategy

In the end, microbiome testing becomes truly powerful only when it translates into action. Not generic handouts, not “avoid this food” checklists, but a structured, phenotype-driven roadmap rooted in biology. The 7R Framework gives clinicians that roadmap – clear enough to act on, flexible enough to personalize, and grounded enough to deliver results that are predictable and repeatable.

When practitioners stop guessing and start strategizing, the gut becomes not a black box, but a system we can finally work with – intelligently, precisely, and with confidence.

The catastrophe hinged on the microbial transformation of Sorivudine in the gut into a compound called BVU (E-5-(2-bromovinyl)uracil). Once absorbed and transported to the liver, BVU irreversibly inhibits a crucial enzyme – dihydropyrimidine dehydrogenase (DPD), the enzyme responsible for breaking down 5-FU. With DPD shut down, 5-FU accumulated to lethal concentrations, destroying bone marrow, gut lining, and immune defenses until patients succumbed.

Today, we now understand that the gut microbiome can shape inter-individual drug responses in remarkable ways. Since 1993, hundreds of studies have demonstrated the microbiome’s central role not only in drug-drug interactions, but also in therapeutic outcomes. Many findings have moved from “interesting correlation” to clear causation.

Why would patient A respond beautifully to a treatment while patient B with the same diagnosis shows no response at all?

Increasingly, the answer lies in their gut microbiomes.

The Gut Microbiome Can Determine Cancer Treatment Efficacy

In 2018, a landmark paper published in Science showed that gut microbiome composition was predictive of response to anti-PD-1 treatment in melanoma.

Melanoma is an aggressive type of skin cancer originating from pigment-producing cells. The human immune system relies on T cells - elite soldiers trained to detect and kill abnormal cells. But cancer has learned to exploit immune “checkpoints,” the brakes that prevent T cells from attacking healthy tissue. Melanoma cells frequently hijack these brakes, convincing T cells that the cancer is “normal.”

Anti-PD-1 therapy works by cutting the brake line. Once the brakes are removed, T cells wake up, recognize the cancer, and attack it aggressively.

In the study, patients with metastatic melanoma received anti-PD-1 immunotherapy. When researchers analyzed their stool samples, they found clear differences between responders and non-responders. Responders had higher abundance of certain bacteria, most notably Faecalibacterium.

By day 600 of the study, more than 80% of patients with high levels of Faecalibacterium remained progression-free. Meanwhile, over half of those with low Faecalibacterium saw their cancer progress again in just ~200 days.

In contrast, patients whose guts were dominated by bacteria from the order Bacteroidales generally had poorer responses, shorter progression-free survival, and weaker anti-tumor immunity.

To test whether this relationship was truly causal, researchers transplanted stool from responders and non-responders into germ-free mice. The result was clear -
Mice receiving microbiota from responders had smaller tumors and stronger responses to immunotherapy.

Antibiotics Can Worsen Immunotherapy Outcomes

These observations were reinforced by Pinato et al. (2019), who studied patients with non-small-cell lung cancer, melanoma, and other malignancies receiving immune checkpoint inhibitors.

Patients were divided into three groups:

• those who received antibiotics before immunotherapy

• those who received antibiotics during treatment

• those with no antibiotic exposure

The findings were striking:

• Antibiotics before immunotherapy: median survival ~2 months

• No antibiotics: median survival ~26 months

• Antibiotics during treatment: ~22 months

The results indicate that disrupting the microbiome before immunotherapy impairs anti-cancer immune readiness, likely undermining treatment efficacy. On the other hand, antibiotics after treatment initiation still have an effect, but the immune system may have already been “reprogrammed” by the immunotherapy.

A related study by Taur et al. (2014) showed that patients with high gut microbiome diversity had significantly better overall survival and fewer transplant-related deaths following allogeneic stem cell transplantation.

Can We Clinically Shift the Microbiome to Improve Immunotherapy?

Yes, and dietary fiber appears to be a powerful tool.

Spencer et al. (2021) showed that late-stage melanoma patients on anti-PD-1 therapy who consumed ≥20 g/day of dietary fiber had significantly longer progression-free survival and a trend toward better response rates. In fact, every extra 5 g/day of fiber was linked to a ~30% lower risk of progression or death. In mouse experiments, the same research team confirmed that high-fiber diets improved tumor control in a microbiome-dependent manner.

Why does fiber matter?

Because fiber-fermenting microbes such as Faecalibacterium and members of Ruminococcaceae produce short-chain fatty acids (notably butyrate) that support anti-inflammatory, anti-cancer immune responses and enhance T-cell activity. Simply put –
Feed the right microbes, and they help your body fight cancer more effectively.

The Microbiome Full Circle Moment

The story of Sorivudine in 1993 was a tragic reminder that the gut microbiome can turn a safe drug into a lethal one. Today, we’ve come full circle: the same microbial ecosystem that once amplified toxicity is now helping determine who responds to life-saving immunotherapies. The science is clear – the microbiome is not a side character in medicine. It is a central decision-maker. As we move into a future of personalized oncology, understanding and modulating the gut microbiome may become just as essential as the drugs themselves.

Our understanding of the complexity of the gut microbiome makes it apparent that the idea of a single “healthy microbiome” will remain elusive because there isn’t one in the first place. We now know that every individual has a unique microbiome and, at best, we will share 40% of the same strains. Trying to define a single “healthy microbiome” is like asking, “What does a healthy forest look like?” Some thrive as dense, humid rainforests. Others flourish as dry savannas, mossy boreal woods, or open pine valleys. They don’t look the same. They don’t have the same trees or animals. But each can be perfectly healthy for the ecosystem it belongs to.

That said, even though a rainforest and a savanna look nothing alike, certain problems can cut across both ecosystems that throw them out of balance. These are universal danger signals - patterns that tell you an ecosystem, of any kind, is under stress. Loss of key stabilizers, takeover by opportunists, disrupted nutrient cycles, weakened resilience, and rising ecological stress - these are all signs of universal danger. The gut microbiome behaves the same way - there are signatures that define imbalanced gut microbiomes. These are the signals we should be looking for when examining microbiome test results. Having seen and analyzed tens of thousands of microbiome profiles with their associated health data, the patterns can be quite obvious.

How do we spot these universal dangers in the gut microbiome?

• Loss of Key Functional Players: A rainforest and a savanna each depend on certain species to maintain stability. If those players disappear, the system destabilizes regardless of diversity level. Similarly, all healthy microbiomes have keystone species - specific microbes that play monumental roles in maintaining gut health, ecological stability, and the functional capacity of the intestinal ecosystem, even if they are not always the most abundant. These species are essential because they contribute to well-defined functions, including breaking down dietary fiber into vital short-chain fatty acids (SCFAs), regulating the immune system, synthesizing essential vitamins, and protecting against disease-causing bacteria.

  Some consistently identified keystone species in the human gut include Akkermansia muciniphila, Faecalibacterium prausnitzii, Bacteroides species (B. uniformis, B. thetaiotaomicron, B. fragilis), Bifidobacterium species (B. longum, B. adolescentis, B. breve), Roseburia intestinalis, and a few others. Although there are likely more keystone species than we currently know, the absence of these species is often associated with dysbiosis-related symptoms.

• Over-expansion of Opportunistic Pathogens: In rainforests, we would consider these opportunists to be invasive vines and pests, whereas in savannas, they’d be thorny shrubs. In the gut microbiome, the overgrowth of certain microbes such as Candida and species of Proteobacteria are indicators of opportunist expansion. The implications of these overgrowths can be grave - they can drive the build-up of gases, slowed motility, breakdown of the intestinal barrier, excessive LPS-driven inflammation, and a dysregulated immune system. More often than not, overgrowth of these opportunists also corresponds to the loss of beneficial keystone species. This is because the conditions favoring the growth of keystone species and opportunistic pathogens often differ.

  For instance, a healthy gut lumen is naturally airless (hypoxic) - with virtually no oxygen, creating the perfect sanctuary for your most important keystone species like the butyrate-producing Faecalibacterium prausnitzii. But when that oxygen gradient breaks - whether from inflammation, barrier disruption, or metabolic stress, luminal oxygen levels can increase. And the moment air finds its way into that ecosystem, the rules change. Your beneficial microbes - the peacekeepers of the gut, most of which thrive in the absence of oxygen, start to wither. In their place, the opportunists move in, especially the oxygen-tolerant members of the Proteobacteria phylum. It’s one of the clearest microbial fingerprints of a gut that’s shifting from harmony to imbalance.

• Severely Low Diversity: Diversity means the total number of microbial species in the gut and how spread out they are. High diversity can be a hallmark of a healthy microbiome. However, the fact that person A has more species than person B does not necessarily indicate that person A is healthier. In fact, there are studies that do not see differences in the microbiome diversity of healthy vs. disease groups. So why does diversity matter? The reason why higher diversity is desirable is because it allows for the functional redundancy of the microbiome. This is a situation where an ecosystem with different species has overlapping functional capabilities such that when one is decreased or wiped out, its functional “look-alike” can continue to support microbiome health.

  Assuming for some reason you had to take antibiotics for a UTI infection, some keystone species keeping an anti-inflammatory state in your gut may be wiped out in the process. Individuals with higher diversity are more likely to have other species unscathed by the drug, that can continue to support their health until the microbiome returns to its previous state (or one similar to it). This is essential to the stability and resilience of the gut microbiome. And the lower the diversity, the less likely you are to achieve functional redundancy.

• Altered Metabolic Outputs: Microbiome metabolic output refers to the collection of biochemical substances (metabolites) and signaling molecules that gut microbes produce as they break down food, interact with each other, and communicate with the host. These outputs include short-chain fatty acids, gases, vitamins, neurotransmitter precursors, bile-acid derivatives, and many other metabolites that directly influence digestion, immunity, metabolism, inflammation, and brain function.

  An imbalance in the gut microbiome can alter metabolic outputs in ways that impact human health negatively. For instance, low levels of butyrate production lead to fatigue, impaired intestinal permeability, neuroinflammation, and so on. Excessive build-up of hydrogen sulfide contributes to bloating and brain fog. Sometimes, these metabolic fingerprints correlate more with the symptoms than the microbes themselves. Other times, different alterations in metabolic outputs can present with the same symptoms - for instance, decreased gut motility could arise due to excessive intestinal methane build-up or as a result of severely low levels of butyrate. With a good microbiome test, you can map more specifically, the imbalances driving the symptoms.

• Increased Intestinal Permeability (Leaky Gut): Leaky gut is a well-known hallmark of an imbalanced gut microbiome. It is often characterized by disruption in the tight junctions of the intestinal walls, thinning of the mucus layer, and the transport of lipopolysaccharides (LPS) and toxins from the gut into the bloodstream - all of which set the stage for chronic inflammation.

  Although there are several ways to investigate intestinal permeability, the gold standard is the Lactulose-Mannitol (L/M) test. The way it works is that you ingest two sugars — mannitol and lactulose. Mannitol is a small molecule that is easily absorbed, whereas lactulose is a large sugar molecule that should not be absorbed unless tight junctions are loose. After about six hours, urine is collected and probed for the ratio of lactulose to mannitol, which reflects how much the gut barrier “leaks.”

  While the L/M test is localized to the small intestine, its results often correlate strongly with colonic dysbiosis because the gut is a continuous tube; hence, permeability is rarely restricted to a single segment. Moreover, tight junction proteins are regulated systemically, so a breach in the small intestine often reflects a global mucosal dysfunction.

  Conversely, the loss of butyrate-producing species such as Faecalibacterium prausnitzii, Roseburia intestinalis, in the colon (large intestine) can drive leakiness in the gut. Butyrate is the primary energy source for intestinal cells (colonocytes), so when butyrate levels are suboptimal, colonocytes starve, leading to the loss of maintenance of tight-junction proteins (leakiness). Butyrate also signals intestinal goblet cells to produce the mucin that forms the protective mucus layer. Without butyrate, the mucus layer will eventually dissolve and expose the gut lining.

  Leaky Gut Flags in Microbiome Tests That You Mustn’t Ignore

  Proteobacteria (e.g., E. coli, Klebsiella, Citrobacter, Enterobacter, Morganella, Salmonella, Pseudomonas) become dominant when the gut barrier is compromised. They don’t just correlate with leaky gut; they help create it, so their expansion is likely an indication that it’s already happening. You should treat Proteobacteria overgrowth as a leaky-gut “flag” because no other microbial group proliferates so dramatically when oxygen levels rise in the lumen, mucosa is inflamed, butyrate is low, mucin is eroded, tight junctions are compromised, or immune tolerance is disrupted. Their continued shedding of LPS triggers mucosal injury and increases leakiness.

  So, is it really true that we don’t know what a healthy microbiome is made of?

  The answer is both yes and no. We may never define a single, universal blueprint just as we can’t define one ideal forest. But we absolutely can identify when an ecosystem is thriving and when it’s drifting toward collapse. The real question was never “What does a healthy microbiome look like?” but “How do we recognize when it’s no longer healthy?” And on that, the science is remarkably clear.

In just five years, America has fallen head-over-heels for GLP-1 drugs. Once reserved for diabetes, they’ve become the new gold standard for weight loss and the numbers are staggering. Today, 1 in 8 U.S. adults and nearly 1 in 5 women aged 50 to 64 use GLP-1 receptor agonists like Ozempic or Wegovy. National spending on these drugs has exploded more than 500%, soaring from $13.7 billion in 2018 to $71.7 billion in 2023. Between 2019 and 2024 alone, the share of adults with obesity who received a GLP-1 prescription (without surgery) quadrupled from 2.5% to 11.2%.

But here’s the twist: while drug-induced GLP-1 signaling is reshaping waistlines across the nation, your gut microbes have been doing it naturally all along.

Now, let me be clear, I reckon there are many effective, non-microbiome-based strategies for weight management: calorie control, exercise, hormonal balancing, sleep, and stress modulation, to name a few. However, this article focuses on an often-overlooked player - the gut microbiome, and how it can profoundly support healthy weight through its influence on GLP-1 production.

What Are GLP-1/GLP-1RA Drugs?

GLP-1 receptor agonists (GLP-1RAs) mimic a hormone your gut already makes - glucagon-like peptide-1 (GLP-1). This natural hormone:

• Helps the pancreas release insulin when blood sugar is high

• Slows down stomach emptying (so you feel full sooner and longer)

• Suppresses appetite by acting on brain pathways

• Lowers glucagon (which otherwise raises blood sugar)

Given how GLP-1 RA drugs work as listed above, it is obvious why they are popular treatments for diabetes and obesity.

While these drugs have their place in people’s weight health journeys, your gut microbiome, when healthy and nourished, can achieve similar effects from the inside out.

How Your Microbiome Influences GLP-1 and Weight Health

Your gut influences weight through multiple biological routes, including:

1. Short-chain fatty acids (SCFAs)

2. The Prevotella-to-Bacteroides (P/B) ratio

3. Secondary bile acids

4. Lipopolysaccharides (LPS)

In this article, I’ll focus on the first two. (The others will be explored in a future post.)

1. Short-Chain Fatty Acids (SCFAs): Your Gut’s Hidden Hormone Hackers

When you eat food rich in soluble fiber, the beneficial bacteria in your large intestine (colon) ferment them into compounds called short-chain fatty acids - mainly acetate, propionate, and butyrate. All three stimulate GLP-1 production, but propionate stands out as the most potent.

Propionate triggers GLP-1 release by binding to receptors (FFAR2 and FFAR3) on specialized hormone-secreting gut cells called L cells. Think of L cells as “smart sensors” that line your lower intestine. When they “taste” propionate, they respond by releasing GLP-1 into the bloodstream.

Knowing this, you might assume taking propionate supplements would boost GLP-1 but it’s not that simple. Propionate taken orally is quickly absorbed in the small intestine, never reaching the colon where it’s most needed. That’s why feeding the right gut bacteria with the right fibers remains the most effective approach.

The Research Evidence

In a 24-week randomized controlled trial, overweight adults who took 10 grams per day of inulin-propionate ester (IPE) - a compound that delivers propionate specifically to the colon showed:

• Prevention of weight gain

• Reduced visceral fat

• Greater satiety

• Improved insulin sensitivity
  compared to those given inulin alone.

The 10 g of IPE delivered around 2.4 g of propionate directly to the colon, achieving the same propionate levels you’d get from about 60 g/day of regular fiber.

Interestingly, a follow-up study found this effect was age-dependent, as this outcome was more pronounced in middle-aged adults than in younger individuals.

But a word of caution - more propionate is not always better. Excess levels have been linked to neuroinflammation, mitochondrial stress, and in some contexts, neurodevelopmental disorders. Like most things in biology, balance is key.

2. The Prevotella/Bacteroides Ratio: Why Some People Lose Weight Faster Than Others

Ever wonder why two people can follow the same “healthy” diet - yet only one loses weight? The answer may lie in their Prevotella-to-Bacteroides ratio (P/B ratio). People with a Prevotella-dominant microbiome (high P/B ratio) tend to lose more weight on high-fiber, plant-based diets than those dominated by Bacteroides (low P/B ratio).

To clarify, Prevotella is a bacteria species that thrives on complex carbohydrates and fibers, producing SCFAs like propionate and acetate that boost GLP-1 and satiety. Bacteroides, on the other hand, are more common in western-style diets high in fat and animal protein, specialize in breaking down proteins and fats, yielding fewer beneficial SCFAs.

Key Human Studies

• Christensen et al., 2019 - In a 6-week dietary trial, individuals with a high P/B ratio on a whole grain diet (high fiber content) lost 1.8 kg more body weight compared to those on a refined wheat diet. Conversely, participants with a low P/B ratio maintained a stable weight.

• Hjorth et al., 2017 - In a 6-month study comparing a high-fiber Nordic Diet to a standard Danish diet, those with a high P/B ratio lost 3.15 kg more body fat.

• During a one-year follow-up, only the high-P/B participants maintained their weight loss.

These findings highlight a powerful truth: your microbiome can amplify or blunt your response to dietary change.

Why a High P/B Ratio Promotes Weight Loss

1. Superior Fiber Fermentation
   Prevotella species efficiently break down resistant starches, pectins, and β-glucans, generating more propionate - the same GLP-1-boosting metabolite discussed earlier.

2. Reduced Energy Harvest
   Prevotella-driven fermentation tends to produce metabolites that are linked to lower energy extraction efficiency, meaning fewer calories are absorbed from the same amount of food. Bacteroides dominant individuals may derive more usable calories from protein/fat diets, reducing weight loss potential under high-fiber dietary interventions.
   

How to Cultivate the “Lean Prevotella-Propionate Axis”

• Increase soluble fiber intake: Prioritize inulin, resistant starches, pectins, arabinoxylans, and β-glucans.

• Add prebiotics that enrich propionate producers: Include foods like rye, barley, legumes, and Jerusalem artichoke (rich in FOS and inulin).

• Balance protein intake: Too much animal protein promotes proteolytic fermentation; combine moderate protein with complex carbs and fiber.

• Avoid broad-spectrum antibiotics when unnecessary: These can deplete anaerobic propionate producers.

• Limit ultra-processed foods and excessive saturated fats: These favor Bacteroides dominance and reduce SCFA diversity.

• Encourage cross-feeding: Combine fermented foods (rich in Lactobacillus) with fiber sources to support bacteria like Veillonella that convert lactate into propionate.
  

Rethinking the Weight-Loss Narrative

Weight loss has long been framed as a battle of willpower, calories, and hormones. But the truth is, it’s also a story of ecosystems - the microscopic kind that lives inside us.

Your gut microbiome is an active participant in how your body senses food, regulates hunger, stores fat, and generates energy. While GLP-1 drugs may have given medicine a revolutionary tool, they are, in essence, imitating what a healthy gut has done for millennia.

So perhaps the next frontier of sustainable weight management isn’t just about stronger drugs or stricter diets. It is about restoring the microbial orchestra that keeps our biology in rhythm. Because when your gut bacteria thrive, your metabolism listens.

If you’re intrigued by the intersection of science and storytelling, subscribe to my weekly newsletter. I translate the complexities of your gut’s inner world into clear, accessible language, revealing its profound impact on your energy, mood, longevity and overall health.

Studies cited

Chambers ES, Viardot A, Psichas A, Morrison DJ, Murphy KG, Zac-Varghese SE, MacDougall K, Preston T, Tedford C, Finlayson GS, Blundell JE, Bell JD, Thomas EL, Mt-Isa S, Ashby D, Gibson GR, Kolida S, Dhillo WS, Bloom SR, Morley W, Clegg S, Frost G. Effects of targeted delivery of propionate to the human colon on appetite regulation, body weight maintenance and adiposity in overweight adults. Gut. 2015 Nov;64(11):1744-54. doi: 10.1136/gutjnl-2014-307913. Epub 2014 Dec 10. PMID: 25500202; PMCID: PMC4680171.

Christensen L, Vuholm S, Roager HM, Nielsen DS, Krych L, Kristensen M, Astrup A, Hjorth MF. Prevotella Abundance Predicts Weight Loss Success in Healthy, Overweight Adults Consuming a Whole-Grain Diet Ad Libitum: A Post Hoc Analysis of a 6-Wk Randomized Controlled Trial. J Nutr. 2019 Dec 1;149(12):2174-2181. doi: 10.1093/jn/nxz198. PMID: 31504699.

Hjorth MF, Blædel T, Bendtsen LQ, Lorenzen JK, Holm JB, Kiilerich P, Roager HM, Kristiansen K, Larsen LH, Astrup A. Prevotella-to-Bacteroides ratio predicts body weight and fat loss success on 24-week diets varying in macronutrient composition and dietary fiber: results from a post-hoc analysis. Int J Obes (Lond). 2019 Jan;43(1):149-157. doi: 10.1038/s41366-018-0093-2. Epub 2018 May 17. PMID: 29777234; PMCID: PMC6331389.

Pugh JE, Petropoulou K, Vasconcelos JC, Anjum A, Thom G, McCombie L, Tashkova M, Alshehhi S, Babalis D, Holroyd L, Sadiq BA, Prechtl C, Preston T, Chambers E, Lean MJ, Dhillo W, Prevost AT, Morrison D, Frost G. Increase in colonic PRopionate as a method of prEVENTing weight gain over 12 months in adults aged 20-40 years (iPREVENT): a multi-centre, double-blind, randomised, parallel-group trial. EClinicalMedicine. 2024 Sep 25;76:102844. doi: 10.1016/j.eclinm.2024.102844. PMID: 39391015; PMCID: PMC11466568.