The gut after antibiotics: what can support the rebuild and what slows it down

The infection is gone, the belly feels different

The last tablet sits in an empty blister. The fever has been gone for days, the cough as well. The matter is, in principle, settled.

Only your belly has not noticed yet.

It makes sounds it did not make before. The bowel movement is softer than usual, or it does not come at all. After eating there is a tight feeling. And somewhere at the back of your mind sits the sentence you picked up between two appointments: after antibiotics you have to rebuild the gut.

Many people know this pattern. In my consultation it comes up almost weekly, usually with the same question in tow. Which product should I take.

First a sentence that stands above everything else. Antibiotics save lives. This text does not judge any medical prescription and at no point advises stopping, shortening or refusing a prescribed antibiotic. A course that has been started is finished as it was prescribed. Everything written here comes afterwards.

And one more sentence about the term itself. Gut restoration is not a defined medical procedure. There is no guideline for it, no fixed sequence and no endpoint that is examined anywhere. The term describes an intention, not a method. I use it here because you searched for it, and because you should know which data sit behind the word and which do not.

And now to the question you are actually here for. What is going on down there.

What antibiotics actually change in the microbiome

Imagine a meadow with a few hundred plant species growing on it. Some in large patches, many only in single spots. An antibiotic is not a lawnmower that cuts everything to the same height. It is more like a very selective herbicide that hits certain species hard and others barely at all.

What happens afterwards can be described on four levels, and these four levels recover at different speeds. That is the reason for the contradictory time frames online.

Level one: diversity

Diversity means how many species are there and how evenly they are distributed. It drops quickly after an antibiotic and usually comes back. But not immediately, and not in everyone to the same degree.

Level two: composition

Diversity can come back without the same species coming back. Two meadows with the same number of plant species can look completely different.

The same group looked more closely three years later, over ten months with two courses of ciprofloxacin in between. Per person 52 to 56 stool samples, more than 1.7 million sequences in total. The effect set in within 3 to 4 days of starting. About a week after the end of each course the return began, but it often stayed incomplete. In the end the composition had stabilised in all three, yet it was changed compared with the starting state. The authors describe this as a possible switch into an alternative stable state.

The second observation matters: the reaction differed between the three people, even between the two courses in the same person. There is no standard course. That is uncomfortable and honest.

Level three: individual species that do not come back

The longest documented time span comes from a Swedish study, and it is the reason why some texts speak of years.

The methodologically cleanest study in this block is a different one, and it supports the same point with considerably more people.

Level four: the resistance genes

This is the half of the story that practically never appears in guides. An antibiotic does not only change who lives in the gut. It also changes which tools these residents carry with them.

Resistance genes are blueprints with which bacteria can render an antibiotic harmless or pump it out of the cell. They can be passed on between bacteria, including between different species. Taking an antibiotic can give exactly these carriers an advantage. In Palleja, species with beta lactam resistance genes were positively selected, in Jernberg the genes stayed elevated for two years, in Zaura they were enriched.

This is no reason for a guilty conscience and no argument against a necessary course. It is the reason why careful use of antibiotics is a medical goal.

And now you know why the question how long does this take has no single answer. It has five, and each of them is right for its own measure.

The finding that puts the common recommendation in doubt

If you have read this far, one conclusion suggests itself. If antibiotics thin out diversity, then you simply top it up. You give bacteria back. That sounds so logical that hardly anyone questions it.

In 2018 an Israeli group tested exactly that. And the result is uncomfortable.

Before this turns into a campaign against probiotics, four limitations belong with it.

The sister study from the same programme also explains why people react so differently to the same product.

This leaves two camps facing each other. One says probiotics belong after every antibiotic. The other says they do nothing. Both cite real studies, and both are talking about different things.

The other side: where probiotics have good data

Now column A. And it is strong.

Antibiotic associated diarrhoea is not a marginal phenomenon. The German guideline gives a frequency of about 10 percent under parenteral antibiotic therapy, and more than 75 percent of these cases are not infectious. For exactly this problem there is a large landscape of studies.

The picture becomes considerably sharper when you look at a single named organism instead of the category.

And now comes the number that explains why the German guideline still gives no recommendation.

How far this carries is shown by the largest single trial in this field. It found nothing, and that is an important result.

What the guidelines make of this

And here it becomes interesting. The guidelines say neither the one nor the other. They say a third thing.

The decisive part is not in this sentence but in the reasoning behind it. The guideline sets out that the Cochrane benefit only appears from a frequency of at least 5 percent in the treated population. And that such a frequency is reached in Germany almost only in one high risk group, namely people with acute leukaemia or after allogeneic stem cell transplantation. On top of that, it remains unclear which probiotic would be meant, because the meta-analyses pool different products.

That is not a rejection. That is arithmetic.

Two other professional societies reach slightly different conclusions from the same evidence. The European society for paediatric gastroenterology formulates recommendations only where at least two randomised trials exist for a strain named with its strain designation. If risk factors are present in children, for example antibiotic class, duration of treatment, age, hospital care or previous episodes, it names Lactobacillus rhamnosus GG or Saccharomyces boulardii, each with moderate evidence quality and a strong recommendation. Without risk factors, it does not. American gastroenterology likewise decides by indication and strain, and names the low certainty of evidence explicitly.

What form and strain even mean is set out in the article on probiotics in spore form versus capsule form. Here it stays at two sentences, because this article answers a different question.

And now you know why you find two camps online, both pointing at real studies. They asked different questions.

What slows the rebuild, often without you noticing

The more interesting part of the question is rarely what you take in addition. It is what works against recovery at the same time. While you are busy with a product, several brakes are often running in the background. Some of them are even on a prescription.

The second course of antibiotics

The study by Dethlefsen and Relman over ten months included two courses of ciprofloxacin. After the second one, recovery was not the same as after the first. A meadow that is treated twice in the same year looks different afterwards than one that is treated once.

For the very long time frame there is a large observational study that I place carefully.

The strongest single brake sits on a prescription

This is the point where many guides stop and where it actually starts to become interesting.

Why stomach acid matters for the gut at all can be said in one sentence. It is the first barrier between what you eat and what arrives in the small intestine. If there is less acid up top, more can arrive further down that does not belong there.

Alcohol, a one sided diet, lack of sleep, stress

On alcohol there is a study with mucosal biopsies. Ece Mutlu and colleagues examined colonic biopsies from 48 people with alcohol dependence and 18 healthy controls. A subgroup showed an altered colonic microbiota, with fewer Bacteroidetes and more Proteobacteria, in part linked with high endotoxin levels in the blood. The connectedness of the community was reduced, even after longer abstinence.

Two limitations definitely belong with this: it was about alcohol dependence, not about a glass of wine in the evening. And only a subgroup was affected. A number for moderate consumption cannot be derived from it. As a direction the finding still holds.

A very one sided diet is the second brake, and the most common one. Anyone living on toast, rusks and bananas for two weeks after an infection feeds a very small selection of residents. For a few days that is fine. As a permanent state it takes the basis away from exactly those species that are trying to come back.

Lack of sleep and stress belong on this list as well, without me pulling in a single study for it here. For the link between sleep and the microbiome I have gathered the evidence in the article on sleep and the gut microbiome. For the connection through the nervous system the same applies in the article on the gut brain axis and vagus stimulation. Both are mechanistically well founded and thinner in human evidence than the guides suggest. That is exactly how it belongs named.

And now you know why the question what do I take is often the smaller half. The bigger one is what is working against it at the same time.

What can support the rebuild

Now the part you were probably waiting for. It begins with a limitation that applies to the whole section.

None of the studies in this section was carried out in people after a course of antibiotics. All of them are analogies. That is no reason to leave them out, but a reason to label them. Where the label [Analogy, not tested in people after antibiotics] appears, that is exactly what I mean.

Substrate variety instead of a single number

Bacteria do not eat fibre in general. They eat particular sugar chains, and different species eat different ones. A meadow with thirty plant species needs more than one fertiliser. Variety on the plate is therefore the more plausible target than a number of grams. How robust the famous number 30 is can be found in the article on the fibre myths. Here the finding of a single randomised trial is enough, because it is unexpectedly fine grained.

In practice this does not mean: eat two kilos of sauerkraut now. It means: in the trial cited it was the fermented foods group in which the described changes appeared, so yoghurt, kefir, sauerkraut, kimchi and miso. For the time after a course of antibiotics this has not been tested, and no effect on your recovery can be promised from it. It is the route with the least uncertainty and not the route with proof of effect. Variety on the plate beats a single number.

Two limits belong with this. In pregnancy, unpasteurised fermented and raw milk products do not belong on the plate because of listeria, and pasteurised versions are the way here. And anyone who reacts sensitively to histamine often tolerates matured fermented products poorly.

The convenient way out that is not one

At this point many texts add the sentence that you should simply take prebiotics instead of probiotics. So not bacteria, but their food. It sounds elegant. It is less well studied than it sounds.

What prebiotics and resistant starch actually feed in the normal case, and where their strengths lie, is a topic of its own. In the situation directly after a course of antibiotics I name the gap instead of filling it. That is the more honest answer.

Movement

Jacob Allen and colleagues had 32 previously inactive adults, 18 of them lean and 14 with obesity, train under supervision three times a week for six weeks. Six weeks without training followed, and diet was controlled before each sample.

In the lean participants the short chain fatty acids in stool rose, in those with obesity they did not. The practically most important finding comes at the end: the changes largely receded once the training stopped. Two honest sentences follow from this. Movement can shift the microbiome measurably, but apparently only while it is happening. And the effect is not the same in everyone. [Analogy, not tested in people after antibiotics]

Sleep and time

Sleep belongs on this list because it is the condition for almost everything else. I give no numbers, because for the question after antibiotics I found no verifiable human study that I could cite in good conscience.

And then time. The least popular point, and probably the most important. The time axis here lies at weeks to months, not at days. An ecosystem that is sorting itself out again is hard to hurry along. Anyone who tries the next product after ten days out of dissatisfaction has never let the first step run its course.

If symptoms persist for weeks, something other than plain recovery can be behind them. Bacterial overgrowth in the small intestine is one of the possibilities that can be examined medically, and there is the article on SIBO for that. Overgrowth with yeasts is also often discussed, and what needs saying about it is in the article on Candida in the gut. And coeliac disease can hide behind persistent diarrhoea too, particularly when the symptoms are first noticed after an infection.

Two further possibilities I name here expressly, because they are often noticed late. Microscopic colitis causes watery diarrhoea without blood, often over weeks, and it cannot be seen with the naked eye during a colonoscopy. It shows up only in the tissue samples taken during it. If a colonoscopy is coming up for you and the diarrhoea persists, that is a good point to raise beforehand, and there is the article on microscopic colitis for it. The second is a disturbed bile acid cycle, which can likewise cause watery diarrhoea and responds to something entirely different, and there is more on that in the article on bile and bile acids. Both belong in a medical assessment and not with preparations from the internet.

And now you know why the most honest advice in this section sounds boring. It is boring because it is documented.

Clostridioides difficile: the case that belongs in medical hands

Up to here it was about nuances. Now it is about a clear line.

There is one situation after antibiotics in which self experiments are the wrong route and delay can do harm. It has a name.

How strongly an antibiotic is linked with this infection differs considerably between the classes.

One note that goes beyond the microbiome and belongs here. All the substances named are prescription only, and the choice is made by the prescribing doctor according to pathogen and resistance pattern, not according to the microbiome. For fluoroquinolones, which include ciprofloxacin, the European authorities restricted use in 2019. The reason is rare but partly long lasting and sometimes not reversible problems affecting tendons, muscles, joints and the nervous system, along with a signal for bulges of the main artery. They should therefore no longer be used for mild infections where an alternative is available.

If you notice tendon pain or abnormal sensations during or after such a course, that belongs in a prompt medical assessment. With sudden severe pain in chest, abdomen or back, call the emergency number 112 and do not wait. That is not a reason to stop a started course on your own, but a reason to discuss it medically.

The standard has changed recently. In its current version the German guideline names fidaxomicin as the preferred first line treatment, a prescription only antibiotic directed specifically at Clostridioides difficile which, on the available data, affects the rest of the gut flora less strongly than the earlier standard agents. Like any antibiotic it can have side effects, among them nausea, vomiting and abdominal pain, and it is not suitable where a hypersensitivity is known. Which agent fits in an individual case is decided by the treating doctor according to severity, history and accompanying conditions. Doses I deliberately do not give here.

With repeated relapses a faecal microbiota transfer can be considered, in Germany so far as an individual treatment attempt. Where the diagnosis is suspected or confirmed, stopping proton pump inhibitors should also be considered. Each of these points is a medical decision, not a self treatment measure.

How much a complete community can achieve is shown by the study that gave the procedure its breakthrough.

And now you know where the line between self care and diagnostics lies. It is not blurred. It is sharp.

The bigger question that does not belong in your belly

There is a sentence that is constantly misused in this field, and I would like to write it down properly.

Antibiotics are one of the most effective inventions in medicine. Pneumonia, blood poisoning, meningitis, a severe wound infection: in all of these situations an antibiotic decides whether someone survives. That is the starting point, and nothing that follows changes it.

At the same time there are numbers showing that they are used more often than guidelines would consider appropriate. These numbers describe a system, not a single person and not a single prescription.

So that the article does not argue with American numbers alone, here is the European addition, in which German practices also took part.

In practice this conversation takes pressure off both sides. Saying that you would also go home content without a prescription, provided that is medically defensible, changes the situation. And anyone with a clear indication gets their antibiotic and takes it in full.

What you can realistically expect

That leaves the question you probably came with. When is this over. The honest answer has four parts.

First, what comes back. The basic pattern, diversity as a number, function. In young healthy adults usually within about six weeks.

Second, what often does not come back. Individual species. In the study of twelve men there were nine of them missing in most participants after half a year. Whether that matters for you personally has not been studied so far.

Third, why you do not feel it. In Dethlefsen a third of the taxa were shifted, and all participants reported normal bowel function. Symptoms and microbiome can be decoupled, in both directions. If your belly causes symptoms after a course of antibiotics, the cause is not automatically the species list. It can also lie in motility, in the gut lining, in what you ate while ill, or in something else entirely.

Fourth, why there is no test for it. There is no blood value and no validated stool test that proves a microbiome has been rebuilt. Even the accepted definition for it is missing. And the sister study to the Suez work shows that stool does not even reveal whether a probiotic has colonised the mucosa. Which questions a stool test can answer is set out in the article on stool testing and dysbiosis diagnostics.

If after a course of antibiotics you want to sort out not just a single topic but the whole picture, you will find the frame in the article on the gut reset and holistic gut treatment. It orders the building blocks that are only touched on here.

Frequently asked questions

Where this topic connects to the rest of the body

The gut after a course of antibiotics does not stand on its own. It hangs on what lands on the plate, on sleep, on the nervous system and on the question of how well nutrients are taken up while the lining is busy.