Reviews/Probiotics and Gut Microbiome/Probiotics prevent antibiotic-associated diarrhea
Clinical · PGM-GI-02 · v2
The corpus tests whether taking probiotics alongside antibiotics stops the diarrhea antibiotics can cause. The best-run and largest trials that used a strict definition of that diarrhea found no benefit: a multispecies product in children [1], a yeast product in hospitalized adults [4], and a yogurt-based product in children [18] all matched placebo. Benefit did appear, but only for particular strains or for looser measures, such as fewer overall loose stools [1], shorter diarrhea [6], or diarrhea during a stomach-infection treatment [10], [11]. So the effect is real in some settings but not uniform, and one trial [1] showed the answer even flips depending on how diarrhea is defined. Because a blanket 'probiotics prevent it' claim is contradicted by the strongest trials yet supported by narrower ones, the panel treated the sweeping claim as overstated rather than clearly true or false.
On the strict antibiotic-associated diarrhea (AAD) endpoint the primary trials are predominantly null: [1] reported no effect (RR 0.81; 95% CI 0.49-1.33), [4] was null (HR 1.02; 95% CI .55-1.90; P=.94) and stopped early for futility, and [18] found nearly identical 2% rates (3 active vs 2 control). Benefit is confined to specific strains or softer endpoints: [1] reduced overall diarrhea (RR 0.65; 95% CI 0.44-0.94), [6] cut diarrhea-like duration in a post hoc analysis (2.70 vs 3.71 d; P=0.037), [10] reduced diarrhea during H. pylori eradication (11.2 vs 21.2%; p=0.012), and [11] improved AAD with A. clausii. A large ICU trial [8] showed no diarrhea difference among its secondary outcomes. The signal is therefore genuine but heterogeneous and context-dependent rather than universal. The contested point: The division is between endpoint and strain: the strict AAD definition yields null results in the largest trials ([1], [4], [18]) while looser endpoints and named strains ([1] overall diarrhea, [6], [10], [11]) show benefit. [1] itself demonstrates that changing the AAD definition flips the result, which is the crux of whether the universal claim can stand.
All 20 papers behind this claim with their DOIs and what each does to it, the panel’s arbitration, the evidence gap map and the open research questions. Free, and sent to your inbox.
We’ll also send you other claims from Probiotics and Gut Microbiome, one at a time, and tell you when a release or a pivotal trial changes a grade. Unsubscribe by replying.
One reviewer's only job is to attack the finding. The challenges below were weighed before the grade was set.
The claim is stated as a universal preventive effect, but the largest, most rigorous blinded RCTs on the strict AAD endpoint are null: [1] (RR 0.81, CI crosses 1), [4] (HR 1.02, stopped for futility), and [18] (2% vs 2%).
Positive signals are strain- and endpoint-specific ([10] S. boulardii in H. pylori therapy; [11] A. clausii) and cannot support a genus-wide claim; [1] itself demonstrates the result flips with the AAD definition used.
Where it remains uncertain. The effect depends on how AAD is defined and which strain is used — [1] shows the result flips with the endpoint definition, so a genus-wide preventive claim is not supported by the primary trials.
Aleatory
When antibiotics barely disrupt the microbiota, AAD is rare (2% in [18]), leaving little room to detect prevention and adding chance variability.
Epistemic
Benefit is documented only for particular strains ([10] S. boulardii, [11] A. clausii), so whether other strains prevent AAD is unestablished by the corpus.
Inferential
[16] measures GI complaints after macrolides and [6]'s benefit is a post hoc duration outcome, not the pre-specified AAD endpoint, so reading them as AAD prevention is an inferential leap.
The load-bearing papers from a corpus of 20, each classified by what it does to the claim. Every DOI is taken from the corpus the claim was graded on.
Largest paediatric RCT: null on strict AAD, positive on overall diarrhea — shows the effect is definition-dependent
“the probiotic (n = 158) had no effect on risk of AAD (relative risk [RR], 0.81; 95% CI, 0.49-1.33). However, children in the probiotic group had a lower risk of diarrhea regardless of the etiology (RR, 0.65; 95% CI, 0.44-0.94).”
Panel auditInternal validity: high · scope of inference: Children receiving antibiotics, AAD defined stringently as C. difficile or unexplained diarrhea.Key limitations: null result depends on stringent AAD definition · broader diarrhea outcome did show benefit (RR 0.65)
Large adult RCT, null and halted for futility — direct contradiction of the claim for S. boulardii in hospitalised adults
“The hazard ratio of AAD in the S boulardii group compared with the placebo group was 1.02 (95% confidence interval, .55-1.90; P = .94).”
Panel auditInternal validity: medium · scope of inference: Hospitalized adults on systemic antibiotics without particular risk factors, S. boulardii only.Key limitations: stopped early for futility · single strain, single population · underpowered
Primary endpoints null; benefit only on post hoc duration measure
“Primary study outcomes of consistency and frequency of bowel movements were not significantly different between the probiotic and placebo groups.”
Panel auditInternal validity: medium · scope of inference: Healthy adults on a single antibiotic regimen; prevention of AAD vs reduction of episode duration.Key limitations: primary outcome null · positive duration finding is post hoc · single antibiotic regimen
Large trial: diarrhea (secondary) not reduced by LGG in ICU patients
“None of the 20 prespecified secondary outcomes, including other ICU-acquired infections, diarrhea, antimicrobial use, mortality, or length of stay showed a significant difference.”
Panel auditInternal validity: high · scope of inference: Critically ill mechanically ventilated adults; diarrhea as a secondary ICU outcome, not AAD-specific.Key limitations: diarrhea is a secondary, not AAD-specific, outcome · primary outcome was VAP · probiotic-associated infections observed
4 of 20 papers shown. The full review lists all of them, with the panel’s arbitration →
Claims are graded at each quarterly release, and re-graded between releases only when a pivotal trial lands. Every change is a new version of the claim.
Well-conducted large RCTs contradict the strict endpoint while strain-specific trials support it, so the broad claim is neither established nor refuted on this corpus.
Release 2026-q4 re-graded every claim from its corpus under protocol v0.19. An earlier release's grade is shown for the record; it came from an earlier protocol and is not a movement.
Assumptions that carry weight but are not themselves established.
The panel records the gap that would most move the claim, and the study that would close it.
Strain-specific strict-endpoint RCT. Positive signals are tied to individual strains ([10], [11]) while pooled products are null on strict AAD ([1], [18]); a single-strain trial on the strict endpoint would separate strain effect from noise. The panel rates this the highest-importance open gap for the claim, and names the study design that would close it: Large blinded RCT of one defined strain (e.g. S. boulardii or A. clausii) powered on the strict AAD definition in a microbiome-disrupting antibiotic regimen.
Open question: Does a single defined probiotic strain prevent AAD on the strict endpoint in a population whose antibiotics disrupt the microbiome? Why it is still open: The strain-positive trials ([10], [11]) use softer or indirect endpoints while the strict-endpoint trials ([1], [18]) test pooled products, leaving the single-strain strict-endpoint case untested.
Priors is a sceptical first reader, not peer review, and not medical advice.
Release 2026-q4 · graded 18 Sep 2026 · protocol v0.19