Probiotics after antibiotics: what the evidence actually shows
Antibiotics are among the most important tools in medicine. They are also among the most disruptive interventions to the gut microbiome, reducing microbial diversity, altering species composition, and creating conditions that allow opportunistic organisms to establish themselves. Antibiotic-associated diarrhoea affects somewhere between five and thirty percent of people taking antibiotics, depending on the drug class, dose, duration, and individual baseline microbiota. It is an uncomfortable and sometimes clinically significant problem, and the appeal of a simple intervention to prevent it is obvious.
Probiotics are the most commonly reached-for solution. The advice to take probiotics alongside or after a course of antibiotics is widespread, dispensed by pharmacists, printed on leaflets, and repeated across consumer health content. The evidence underlying this advice is real but considerably more specific than the general recommendation implies. Not all probiotics work. Timing matters more than most guidance acknowledges. And one strand of evidence raises a genuinely uncomfortable question about whether post-antibiotic probiotic use might, in some circumstances, delay rather than accelerate gut microbiome recovery.
What antibiotics actually do to the gut
Antibiotics act on bacteria. They cannot distinguish between pathogenic bacteria causing an infection and the commensal bacteria that constitute a healthy gut microbiome. Broad-spectrum antibiotics in particular, amoxicillin-clavulanate, fluoroquinolones, clindamycin, and broad-spectrum cephalosporins among the most impactful, sweep through the gut and reduce microbial diversity substantially. This disruption typically begins within the first day or two of a course and can persist for weeks to months after the antibiotics are stopped. Some studies have documented detectable microbiota changes for up to two years following a course of antibiotics, though the clinical significance of long-term low-level changes is debated.
The disruption creates two related problems. The first is diarrhoea from loss of the microbial communities that normally regulate gut motility, fermentation, and fluid absorption. The second is the window of susceptibility it creates for opportunistic pathogens, most significantly Clostridioides difficile, which can colonise the disrupted gut and cause a spectrum of illness ranging from mild diarrhoea to severe colitis. These are separate endpoints, and the evidence for probiotics differs meaningfully between them.
The evidence for preventing antibiotic-associated diarrhoea
This is the context with the strongest probiotic evidence, and it is still not straightforward.
The evidence here needs to be read carefully, because the headline numbers are more uncertain than they appear.
A 2025 umbrella review of meta-analyses identified LGG and Saccharomyces boulardii CNCM I-745 as the most extensively studied strains with the most consistent signals. A 2024 meta-analysis of fifteen randomised controlled trials in 7,427 adult participants found a relative risk of 0.6 (95% CI: 0.43 to 0.82) for antibiotic-associated diarrhoea with probiotic use compared to placebo. A strain-specific analysis in outpatient populations found that diarrhoea occurred in 8.0 percent of participants in the probiotic group compared to 17.7 percent in controls, an absolute reduction of around 9.7 percentage points in that population. Strain-specific subgroup analyses suggested larger directional effects for LGG and S. boulardii individually, but these subgroups were substantially smaller and the estimates are likely imprecise; they are better read as supportive signals than as reliable point estimates.
Heterogeneity across the broader literature is substantial, driven by differences in antibiotic class, study population, probiotic dose, strain, duration, and how diarrhoea was defined. Critically, the absolute risk reduction depends heavily on baseline risk: a relative reduction of this magnitude translates to a much smaller absolute benefit in populations where antibiotic-associated diarrhoea occurs in only five to ten percent of people than in higher-risk populations where baseline rates reach twenty to thirty percent. The studies with low risk of bias, when examined separately, do not consistently show the same effect sizes as pooled analyses. And guideline bodies have reached genuinely different conclusions from the same evidence base.
The American Gastroenterological Association's 2020 clinical practice guideline did not make a recommendation for or against probiotics for the prevention of antibiotic-associated diarrhoea as a general endpoint. Notably, it assessed C. difficile-associated diarrhoea as a separate outcome and made conditional recommendations for specific strains in that context. The American College of Gastroenterology has taken a more cautious position, recommending against probiotic use for C. difficile prevention. The Cochrane Collaboration and World Gastroenterology Organisation have reached more supportive conclusions, with the Cochrane review of 2021 finding moderate certainty evidence that probiotics are effective for preventing C. difficile-associated diarrhoea. This is not a settled area, and the divergence between authoritative bodies reflects genuine uncertainty in the evidence rather than straightforward disagreement.
Which strains have evidence
The evidence for antibiotic-associated diarrhoea prevention is not a class effect. It belongs to specific strains that have been studied in this context, and the effect size reported for those strains cannot be assumed to apply to other organisms marketed under similar names.
Lactobacillus rhamnosus GG has the most extensive evidence base of any single probiotic strain studied in this indication. It is a bacterial strain and is therefore susceptible to antibiotics, a point with practical consequences for how it should be timed, discussed below. S. boulardii CNCM I-745 is the most studied Saccharomyces strain and has a specific and important advantage: it is a yeast, not a bacterium, and is therefore not killed by antibiotics. This means it can be taken at the same time as an antibiotic without the survival concerns that apply to bacterial strains. S. boulardii also has a reasonable evidence base for C. difficile-associated diarrhoea prevention specifically.
No other strains have a comparable depth of evidence specifically for this indication. Some strains have emerging or limited trial support, but the evidence base for LGG and S. boulardii in this context is substantially more extensive than for any other single strain or combination. Consumer products often combine multiple strains from Lactobacillus and Bifidobacterium genera with no established evidence for antibiotic-associated diarrhoea. The presence of a large CFU count and multiple strains does not substitute for evidence in the relevant context.
Timing: when to start and what it means
The timing question is one of the most practically important and least clearly communicated aspects of probiotic use during antibiotic courses.
The evidence supports starting probiotics at the same time as the antibiotic course, not after it finishes. A meta-analysis examining timing found that probiotics started within two days of antibiotic initiation appeared substantially more effective than those started later. Meta-regression analyses have suggested risk reductions in the region of seventy percent when started early compared to around thirty percent when started after two days, though these are estimates from indirect comparisons rather than head-to-head trials and should be treated as indicative of the direction and magnitude rather than precise figures. The rationale is consistent with the underlying biology: preventing the window of pathogen susceptibility from establishing is more tractable than attempting to reverse disruption that has already occurred.
For bacterial strains such as LGG, a separation of at least two hours between the antibiotic dose and the probiotic dose is conventionally recommended, on the basis that the antibiotic concentration in the gut will be lower after this interval and more bacteria will survive. The evidence base for exactly this two-hour window is pharmacological reasoning rather than direct trial data. Trials have generally not compared one-hour versus two-hour separation. But the principle is reasonable and the recommendation is widely supported in clinical guidance. S. boulardii requires no such separation as antibiotics do not affect it.
How long to continue after the antibiotic course ends is less clearly established. A practical approach supported by general clinical guidance is to continue for one to two weeks after completing the antibiotic course. Longer continuation has not been shown to provide additional benefit for diarrhoea prevention in healthy adults.
The microbiome recovery question
Here the picture becomes more complicated, and it is important to engage with it honestly.
A study from the Weizmann Institute of Science examined what happened to the gut microbiome in two groups following a course of antibiotics: one group given an eleven-strain probiotic supplement for four weeks and one group given no probiotics. The probiotic group had their gut successfully colonised by the probiotic strains, but the presence of those strains appeared to delay the return of the person's original microbiome for the duration of the study, up to six months. The control group, by contrast, had their original microbiome substantially restored within three weeks.
This finding has been used in some quarters to argue against post-antibiotic probiotic use entirely. That conclusion goes further than the study supports. The study used one specific eleven-strain product, examined microbiome composition rather than clinical endpoints, and the clinical significance of delayed microbiome recovery measured by sequencing is not established. Microbiome sequencing tells you which organisms are present and in what proportion; it does not directly measure what those organisms are doing functionally, and a microbiome that looks different by sequencing may or may not be functionally different in ways that affect health. It is not known whether the delayed return of the original microbiome in the probiotic group translated into any difference in health outcomes. Furthermore, the strains and dose used in that study are not the same as those studied in the AAD prevention trials, so the findings cannot be straightforwardly applied to products like LGG or S. boulardii used specifically for diarrhoea prevention.
What the study does establish is that the common consumer framing, that probiotics help restore your normal gut microbiome after antibiotics, is not clearly supported, and may in some formulations be counterproductive. The evidence-supported use case is preventing diarrhoea during and immediately after the antibiotic course, not accelerating broad microbiome restoration. These are different goals, and the distinction matters.
What the guideline divergence tells us
The fact that major gastroenterological bodies have reached different recommendations from the same evidence is informative rather than merely confusing. It reflects several real tensions in the literature: heterogeneity across trials, uncertainty about which endpoint to prioritise, the influence of industry funding on some primary studies, and legitimate differences in how organisations weigh low-certainty evidence against cost and patient preference.
The practical consequence is that probiotic use for antibiotic-associated diarrhoea prevention sits in a space where reasonable clinicians disagree. The evidence for benefit is real but not definitive. For most healthy adults taking a standard antibiotic course, the risk is low and the cost of a short probiotic course is modest, so the balance tilts towards use of a well-characterised strain if the person is motivated. For people at higher baseline risk, those on broad-spectrum antibiotics, those with prior C. difficile infection, those on prolonged courses, the case is somewhat stronger. For immunocompromised individuals, the general caution about probiotic safety in at-risk populations applies and clinical guidance is appropriate before supplementing.
Practical implications
A reasonable evidence-based approach is as follows. Start with a strain that has been studied for this indication specifically, LGG or S. boulardii CNCM I-745 are the most defensible choices based on the current evidence. Start at the same time as the antibiotic course, not after it. If using a bacterial strain like LGG, separate the probiotic dose from the antibiotic dose by at least two hours. Continue for one to two weeks after completing the course. Be sceptical of products that combine many strains without identifying them to the strain level, or that make broad microbiome restoration claims not supported by the trial evidence.
There is also a reasonable non-supplemental approach. Dietary fibre, fermented foods, and adequate hydration all support gut microbiota recovery after antibiotics. These are not equivalent to the specific trial-tested strains for diarrhoea prevention, but for people who do not develop diarrhoea and whose microbiome disruption is mild, dietary support alongside the antibiotic course may be sufficient.
The evidence supports benefit from specific strains in specific contexts, primarily prevention of antibiotic-associated diarrhoea when the right strain is started early, not the broader claim that any probiotic helps restore gut health after antibiotics. The generic advice to take any probiotic after antibiotics does not reflect what the evidence actually shows.
For detailed evidence profiles on the strains discussed in this article, see Lactobacillus rhamnosus GG and Saccharomyces boulardii. For a broader overview of how probiotic evidence works, see the Probiotics hub and How to choose a probiotic.