C. difficile in the gut: when an antibiotic triggers a colitis

When the antibiotic works and the gut pays the price

The tooth root was inflamed, or the bladder, or the lung. You were given an antibiotic, and it did what it was meant to do. Five days later you are sitting on the toilet for the sixth time this morning. The stool is watery, the belly cramps, the smell is unfamiliar.

Many people know this pattern. Most of them are lucky.

An antibiotic can unsettle the gut along several routes. The most harmless one is purely osmotic: sugars that are normally metabolised by bacteria stay behind and pull water after them. A second route runs through the movement of the bowel. Both are unpleasant and usually stop again on their own. Usually is not always: there are further forms of antibiotic associated bowel inflammation, among them a bloody form described above all after penicillin derivatives. Blood in the stool, fever or severe abdominal pain are therefore a reason to have things checked medically and not to wait, even without a clostridial finding.

A part of these diarrhoeas is something else. Here a bacterium has spread that previously had no chance: Clostridioides difficile, formerly Clostridium difficile. A small part of that becomes serious.

How big is the problem? The best view of the order of magnitude comes from a population based surveillance programme running over seven years.

And now you know why this text does not start with the organism, but with the ground it grows on.

What C. difficile is, and what toxins A and B do

Picture a bacterium that knows two states. In the first it is active and divides. In the second it withdraws into a capsule, shuts its metabolism down almost entirely and waits. Months. Years.

This capsule is called a spore. It survives heat, dryness, stomach acid and alcohol. It is the reason for almost everything that is difficult about this infection. Clostridioides difficile is strictly anaerobic, and in its active form it dies from oxygen. That is exactly why it travels as a spore. And it is why the hand sanitiser on the wall is not enough, more on that later.

Two toxins, one mechanism

What makes you ill is not the bacterium itself. What makes you ill are two large proteins that it releases: toxin A, classically described as the enterotoxin, and toxin B, the cytotoxin. At heart both do the same thing.

Inside each of your gut cells sit small switch proteins, the Rho GTPases. They are the site managers of the cell skeleton and determine where supporting fibres are stretched and how firmly the cell stays connected to its neighbours. The toxins attach a sugar group to these switches. With that, the site managers are shut down.

The decisive sentence comes now. Only strains that can produce these toxins cause disease. Strains without the toxin genes can colonise your gut, show up in the laboratory and do you no harm. That is the reason for the whole diagnostic effort further down.

How this was discovered

Until the nineteen seventies, pseudomembranous colitis was a riddle. The plaques were visible at colonoscopy and the link with antibiotics was known. What exactly was happening there, nobody could say.

And now you know why there are three different tests further down and not one.

Why an antibiotic of all things opens the door

Ask yourself something: why does a substantial share of people carry this bacterium around without ever hearing about it? And why does that tip over in some people within a few days? The answer is called colonisation resistance. Clumsy word, simple idea.

Your gut flora is not a collection, it is an occupied surface. Every niche is taken, every nutrient has a customer, and the residents alter the chemistry so that newcomers have a hard time. A new bacterium does not land in an empty room, it lands in a full flat.

The bile acid arc

The best studied part of this resistance runs through bile acids. The short version in four sentences.

Your liver builds primary bile acids from cholesterol, which reach the gut as taurocholate and glycocholate. In the large bowel certain bacteria turn them into secondary bile acids, above all deoxycholate. The spore needs a start signal to germinate, and that signal is precisely the primary bile acids. Secondary deoxycholate by contrast slows the growth of the germinated form.

So you have a balance that has nothing to do with willpower and everything to do with metabolism. And you can see at once what happens when the bacteria that perform the second step are missing.

And in humans? Weingarden 2014 analysed stool samples before and after a faecal transplantation. Before, almost only primary bile acids, with secondary ones barely detectable. Afterwards the secondary ones dominated, as in the donors. Small numbers, no clinical endpoint, but the bridge from animal to human. PMID: 24284963 · DOI: 10.1152/ajpgi.00282.2013 [Cohort, small]

How bile acids work in general, and why losing them can cause diarrhoea even without an infection, is covered in Bile, bile acids and TUDCA. Rebuilding after a course of antibiotics is covered in The gut after antibiotics.

One note belongs here: it explicitly does not follow from this that you should tinker with your bile acids yourself while treatment is running. Bile acid binders, for instance the prescription only cholestyramine, do not belong together with vancomycin or fidaxomicin according to current expert advice, because they can weaken the treatment, PMID: 42383946. The rebuilding comes afterwards.

And now you know why the time after the antibiotic matters more than the time during it.

The spectrum: from harmless diarrhoea to pseudomembranous colitis

On the internet you will find two kinds of text about C. difficile. Some reassure you to death, others frighten you. Both act as if there were one clinical picture. There is a spectrum.

Stage 4 deserves a sentence I do not want to soften. A toxic megacolon is life threatening. The large bowel widens, its wall becomes thin, and it can tear. Sometimes surgery is needed. This happens rarely. But it happens. And it becomes more dangerous if someone with high fever and severe abdominal pain first spends three days waiting and trying home remedies alone. Not because those remedies are bad, but because they are not made for this situation.

One sentence on the often mentioned sweetish, foul smell: it appears in textbooks, but it does not work as a criterion. A reason to ask, not a finding.

If your diarrhoea lasts longer without a preceding antibiotic, it belongs in a broader search for causes: Chronic diarrhoea. If inflammatory bowel disease is a possibility, the framing sits in Crohn's disease and ulcerative colitis.

And now you know why the first question is not which remedy would be good now, but which stage you are standing on.

Who carries a higher risk

Not every antibiotic carries the same risk. One of the most useful pieces of information in this field, and it appears in almost no patient text.

What else counts

Age counts as a factor in its own right, as do a stay in hospital or a care home and immunosuppression, whether from illness or from medication.

Acid blockers: three sentences and one important note

Proton pump inhibitors are linked with a raised risk in meta-analyses. Deshpande 2012 found an odds ratio of 2.15 across 30 studies, Kwok 2012 across 42 studies found 1.74 for a first infection and 2.51 for a recurrence. Both with very high heterogeneity.

What acid blockers are good for, where they are over prescribed and what a supervised taper looks like is covered at length in its own article: Understanding heartburn and acid blockers. The counter question, whether there is enough stomach acid at all, always comes first in my practice, and you will find it in Low stomach acid and betaine HCl.

Where the organisms come from is more open than assumed

And now you know why the question about the antibiotic class brings more than the question about the door handle.

Diagnosis: the two step algorithm and the most expensive thinking error

Imagine a report arrives in the post. On it stands: Clostridioides difficile, positive. Your heart sinks. Yet that sentence can mean three completely different things.

This section is the heart of the text. It is somewhat technical and worth it anyway, because it can spare you from being treated on the strength of a lab slip that says nothing about you.

Rule zero: what actually gets tested

That is how the guideline of the American professional societies puts it, and the European documents say the same in substance. Anyone who sends in formed stool produces a result without meaning. And once in the world, such a result develops a gravity of its own.

The most expensive thinking error

Now comes the study that carries this section. Its result changed the way diagnosis is done.

Furuya-Kanamori 2015 reviewed the literature from 1980 to 2015 on carriage without illness. There is not even a uniform definition, prevalence varies widely with setting and method and is probably underestimated. Some colonised people even appear to be protected by it. In short: detection without symptoms is not a diagnosis, and screening without symptoms has no proven benefit. PMID: 26573915 · DOI: 10.1186/s12879-015-1258-4 [Systematic Review]

Why stool PCR panels in general find more than they can interpret is covered in Stool testing, PCR and dysbiosis diagnostics. The underlying idea is the same.

One note that matters to me: nothing in this section argues for skipping a recommended examination. If a colonoscopy, a sigmoidoscopy or a laboratory check is proposed, there are good reasons for it. Scepticism towards a single test result is something different from scepticism towards being assessed.

And now you know why the question about the toxin matters more than the question about the organism.

Guideline based treatment: fidaxomicin, vancomycin, and why metronidazole stepped back

There is something paradoxical about it: an illness triggered by antibiotics is treated with antibiotics. Many people stumble over that, and I understand why.

The difference lies in the target and the place. The triggering antibiotic was meant for an infection somewhere in the body and took the gut along as collateral damage. The agents used here stay in the gut and aim at this one organism.

The two first choices

Oral vancomycin is the older mainstay. Swallowed, it stays in the gut lumen and is barely taken up into the blood. That is exactly what is wanted here, because the damage sits in the large bowel. Fidaxomicin is the newer agent with a narrower spectrum, which spares the anaerobic bacteria of the flora more. That is where its advantage comes from.

Cornely 2012 repeated this comparison double blind across 86 centres in Europe, Canada and the USA: 91.7 versus 90.6 percent, non inferiority met. One subgroup was notable. People who needed other antibiotics at the same time responded more often on fidaxomicin, 90.2 versus 73.3 percent. Those are exactly the difficult cases in everyday care. PMID: 22321770 · DOI: 10.1016/S1473-3099(11)70374-7 [RCT, n=535]

Why metronidazole stepped back

For decades metronidazole was the first choice, and in older patient leaflets it still is. The change did not come overnight, it came through several studies pointing in the same direction.

Stevens 2017 compared 47,471 treated patients in the US veterans system using propensity matching. Vancomycin went together with lower 30 day mortality, while no meaningful difference was found in recurrence risk. Large observational data therefore support the direction of the randomised trials on survival. The recurrence problem remains untouched by it. PMID: 28166328 · DOI: 10.1001/jamainternmed.2016.9045 [Cohort, n=47471]

The guideline map, and where it diverges

An honest text also says where the professional societies do not write the same thing. And here they do not, not entirely.

The difference between the ACG on one side and the DGVS, IDSA and ESCMID on the other is not a trifle. So it also depends on which framework applies in the treating institution. And the cost difference belongs in the picture as a reality, because fidaxomicin is considerably dearer.

What else belongs to it

Where possible the triggering antibiotic is stopped or changed. That is a medical decision, because the infection it was prescribed for is still there. Severe and fulminant courses belong in hospital. Motility inhibitors such as loperamide do not belong in self treatment when a Clostridioides difficile infection is suspected, at no level of severity. They can still the bowel and hold the toxins where they do harm. Whether anything is given against the diarrhoea in an individual case is decided by the treating doctor, not by the home medicine cabinet. Further agents with thinner data appear in the Cochrane review, there with low to very low evidence quality.

And now you know why the choice of agent is above all a decision about the coming weeks.

When it comes back: the recurrence circle and how it is broken

The diarrhoea was gone. For ten days you felt better. And then, three weeks later, you are sitting there again. This is where the despair sits in this illness.

After the first episode the infection returns in roughly 15 to 25 percent, and after the second the risk rises clearly. The often quoted 40 to 60 percent come from recurrence cohorts and placebo arms, so from selected groups. The direction still holds: every further episode makes the next one more likely.

Why

Three things come together. The spores outlast every course of antibiotics, because in their resting state they cannot be attacked. The flora is not yet back at work afterwards, so the secondary bile acids are still missing. And the antibody response against toxin B turns out weak in some people, as the cohort of Loo 2011 suggests. The antibiotic tidies up and at the same time keeps the ground ready. A circle that closes.

The taper and pulse regimen

One answer to this is strikingly simple as an idea: do not stop the drug abruptly, let it fade out over weeks at ever wider intervals. The spores germinate in waves, each dose hits the freshly germinated ones, and during the gaps the flora gets time.

The antibody against toxin B

Faecal transplantation: the strongest evidence of all FMT applications

Faecal transplantation is discussed for many illnesses, mostly on thin data. In recurrent Clostridioides difficile infection that is different, and the difference deserves emphasis.

Standardised preparations take the chance element out of a single stool donation. Feuerstadt 2022 tested purified spores from donor stool and found recurrences in 12 versus 40 percent on placebo, PMID: 35045228 [RCT, n=182]. Khanna 2022 tested a broad microbial consortium as a single enema, with a modelled success rate of 70.6 versus 57.5 percent, PMID: 36287379 [RCT, n=267]. Dubberke 2023 added safety data over 24 months, with no signal for late effects, PMID: 36544075 [RCT, long term data].

The placebo rate of 57.5 percent in Khanna is the most important incidental finding of this section. A substantial share of people do not relapse even without an additional procedure.

How the procedure runs and where the research stands for other indications is covered in Faecal transplantation: what it can do and what it cannot.

And now you know why the question at relapse is not which antibiotic is stronger still.

Probiotics honestly, hygiene at home, and the way back

Three topics that come up almost every time. Kept apart, because the evidence for them differs.

Block A: probiotics without gloss

Here there is a real tension in the data. I leave it standing instead of picking a side.

For Saccharomyces boulardii things look better at first glance. Szajewska 2015 summarised 21 randomised trials with 4780 participants and found a fall in antibiotic associated diarrhoea from 18.7 to 8.5 percent, relative risk 0.47, PMID: 26216624 [Meta-analysis, k=21, n=4780]. The catch sits in the endpoint: what is meant is antibiotic associated diarrhoea as a whole, so predominantly the harmless part of the spectrum. For C. difficile infection alone the data are thinner.

In clinical practice I observe that the balance looks different for a 78 year old on their fourth course of antibiotics this year than for a healthy 30 year old on a short course. That is judgement from the consulting room, not a study result, and it applies only to the question of what runs alongside. What each dosage form can do is covered in Probiotics: spore form versus capsule.

Block B: hygiene at home

The gap that hardly any page fills.

This combination of measure and reassurance also appears in the current expert review of the American gastroenterology society, which lists household hygiene expressly as a point of its own, PMID: 42383946 [Guideline]. The two belong together, otherwise either carelessness or fear takes hold.

One sentence on notification: under the German Infection Protection Act a case is reported by name where the course is clinically severe, for instance on hospital admission because of a community acquired infection, on transfer to intensive care, on surgery of the large bowel, or on death within 30 days. The source is the guidance document of the Robert Koch Institute.

Block C: the way back

First the honest answer: there is no robust study on a particular diet after this infection. Everything that follows is clinical experience, not study evidence.

In the acute phase fluid and electrolytes are what count. Then a stepwise, well tolerated diet. And only later, when the gut is calmer, the patient rebuilding of variety and fibre. Too much fibre too quickly usually brings nothing but bloating in an irritated gut.

What data exist on this are covered in The gut after antibiotics. Whether the famous number 30 for plants a week holds up is examined in Fibre myths.

What happens mechanistically afterwards, from visceral hypersensitivity to altered motility, is covered in IBS: finding the causes.

The next course of antibiotics

Finally the point where you can shape the most. This is not about demonising antibiotics. They are one of the greatest achievements of medicine, and a bacterial infection that needs treating gets treated.

It is about three questions that are fair to ask. Is an antibiotic needed here? Does it have to be this broad? How long does it have to be? You put these questions to your treating doctor, and the decision is made together.

And one practical thing: tell every doctor who prescribes you an antibiotic in future that you have had a C. difficile episode. It rarely appears in the notes by itself, and it can influence the choice of agent.

And now you know why the most important part of this story begins after the last day of treatment.

Common questions about Clostridioides difficile

Where this infection connects to the rest of the gut

A C. difficile episode does not stand on its own. It hangs on the antibiotic history, on the acid situation in the stomach, on the diagnostic culture of the laboratory, and on how patiently the rebuilding runs afterwards.