The best probiotic to take with antibiotics, and what its label leaves out

Yes. This is an outcome where a probiotic has a real result behind it, and it is worth buying one for. The strain class is Saccharomyces boulardii — a yeast, not a bacterium — and the outcome it earns its grade for is diarrhoea while you are taking antibiotics. Everything that qualifies that sentence is in the branch below and in the sections after it, and none of it is a reason to skip the answer: if you are starting a course of antibiotics and you want to do one thing, this is the one thing.

What this page adds to that is the part a shelf will not tell you: which bottles print the strain, which print an amount for it, and whether those two facts ever appear on the same label. They are different questions from "does it work", they are the questions a label can actually settle, and the counts below are recomputed from the audited panels every time this page is built.

Roman Sydorenko, Independent analyst · Last verified 30 July 2026 · Disclosures

This is the buying page for the antibiotics occasion. It reads the audited labels against the strain class that earns a grade here, and sends you to the trial record rather than reprinting it.

What the grade covers, and the outcome it does not #

The grade on this site describes the strength of evidence for a named strain against a named outcome, and it does not describe a product. Two rows of the outcome table matter to a reader standing in a pharmacy, and they point in opposite directions: the diarrhoea row is the reason to buy, and the C. difficile row is the reason not to expect what the box implies. Both are reproduced below rather than summarised, because a buying page that quotes the row it likes and paraphrases the row it does not is the thing this site was built against.

Both rows are assigned and argued on the antibiotic diarrhoea evidence page — the pooled result, its certainty rating, the two independent adult trials that came out null, and the class-level C. difficile figure that is real but points at no product. Every number this page uses comes from there or from the label records below. If you want the literature rather than the shopping, that is the page to read, and this one will keep pointing you at it.

What this page is not for: the course you have finished #

This page is about taking a probiotic while you are on antibiotics. It is not about repairing anything afterwards, and that is a scope limit rather than an oversight: the promise that a probiotic restores a gut the antibiotics emptied is not a finding this fact base holds, and where the outcomes behind that promise have been measured at class level the results are set out on the evidence page. If you have already finished your course and nothing is wrong, there is no purchase on this page for you.

The recommendation, as published, and the bottle it names #

What follows is the buying branch as it was published on the home page, unedited, and then the labels it points at. The branch names a product and states four facts about that product's panel; every one of them is checked against the teardown record on each build, and a changed record fails the build instead of quietly shipping a recommendation the record no longer supports.

If you are about to start a course of antibiotics — this is the one good reason. The strain class with the best-powered adult result is Saccharomyces boulardii: 17.4% → 8.2% antibiotic-associated diarrhoea, RR 0.49 (95% CI 0.38–0.63), across 15 RCTs and 3,114 adults (PMID 26216624 (opens in a new tab)). Look for a single-organism S. boulardii product, ideally strain CNCM I-745, and start it early — the largest pragmatic trial in the field recorded a mean delay of 29.4 hours from first antibiotic to first probiotic dose (PMID 38073551 (opens in a new tab)). Caveat that must travel with this: S. boulardii's adult result is for AAD, not for C. difficile — for CDAD in adults it is not significant (RR 0.80, 0.47–1.34, 9 RCTs, n=1,441; PMID 26216624 (opens in a new tab)). None of the 27 products in the frozen founding cohort is an S. boulardii single-organism product, and the ones that contain it bury it in a multi-organism blend with no per-strain quantity — one of them counting a yeast inside a stated total of "13 Strains of Probiotic Bacteria." Products torn down since that cohort was frozen include single-organism S. boulardii SKUs, and the branch below names one.

The product that branch names is Florastor Dual Action Probiotic Supplement, and its full teardown — including what its panel does not print — is on its own page. Disclosure score 74.4, rank 46. This site earns nothing from that link or from any purchase; the address of the product is published as text on the teardown, never as a link out.

Who should not take this at all #

The case reports behind that, the regulator's own wording, and the population-scale figures that argue the risk is smaller than the case reports make it sound are set out together on the safety page. They are not reprinted here, because a shopping page is the worst possible place to meet a safety signal for the first time in summary form.

Which audited labels print it, and which print the strain #

Here is the split that decides what you can and cannot check. A panel can print the species — two Latin words naming a yeast — or it can print the species and the strain designator, the code a trial is published under. The counts below separate the two, because in this literature the grade is attached to a designated strain and a species line does not carry it. A bottle printing the species alone has told you what kind of organism is inside. It has not told you that the organism in it is the one the trials were run on, and no amount of reading the box harder will settle that.

Labels printing this organism at all

20 of 166

Labels printing the designated strain the grade attaches to

3 of 20 The code a published trial can be looked up by. The rest name the species; one of them prints a house code of its own against the organism, which looks like the same kind of fact and is not one, and it is quoted as printed in the list below.

Labels printing an amount for this organism

9 of 20 5 print a count of organisms and 4 print milligrams only. The remaining 11 bury the organism inside a blend total, where the amount cannot be read at all.

Labels printing a designator and a count of organisms together

0 of 166 Not one, on the panels this audit has transcribed. A panel here tells you which strain, or how many organisms, and the two facts do not meet.

Labels carrying this organism and nothing else

6 of 20 What the branch above means by a single-organism product, resolved against the panels rather than listed by hand.

Every label in this audit that prints the organism is below, best-disclosed first, with what its own panel says beside it. A panel printing a designator is a fact about the panel: it is not evidence that the product was tested, and not evidence that it delivers the amount that was studied.

• Florastor Dual Action Probiotic Supplement Panel prints “Saccharomyces boulardii CNCM I-745®”, 500 mg (per 2-capsule serving; 250 mg per capsule) for this organism. Disclosure score 74.4, rank 46.
• Florastor Florastor Probiotic, 30 Capsules Panel prints “Saccharomyces boulardii CNCM I-745”, 250 mg per capsule (500 mg per 2-capsule serving) for this organism. this record carries a verification flag. Disclosure score 74.4, rank 46.
• Ortho Molecular Products Ortho Biotic Capsules (SKU 527030 / 527060) Panel prints “Saccharomyces boulardii” and no strain designator against it, 200 mg (3 Billion CFU) for this organism. this record carries a verification flag. Disclosure score 72.8, rank 49.
• Florastor Advanced Immunity Probiotic & Immune Supplement, 30 vegetarian capsules (SKU 58928) Panel prints “Saccharomyces boulardii CNCM I-745”, 500 mg per 2-capsule serving (250 mg per capsule) for this organism. Disclosure score 66.7, rank 58.
• Xymogen ProbioMax Sb 35B, 30 acid-resistant vegetarian capsules Panel prints “Saccharomyces boulardii” and no strain designator against it, 250 mg (5 Billion CFU†) for this organism. Disclosure score 50.5, rank 92.
• Metagenics UltraFlora Spectrum Probiotic Panel prints “Saccharomyces boulardii” and no strain designator against it, no amount for this organism — it is inside a blend total, so the dose cannot be read off the label. this record carries a verification flag. Disclosure score 49.1, rank 95.
• Jarrow Formulas Saccharomyces Boulardii + MOS, 5 Billion CFU, 90 Delayed-Release Veggie Caps Panel prints “Saccharomyces boulardii (an S. cerevisiae strain)” and no strain designator against it, 250 mg (5 Billion CFU) for this organism. Disclosure score 48.8, rank 96.
• Renew Life Ultimate Flora Saccharomyces Boulardii Go Pack, 20 vegan capsules (RN15417) Panel prints “Saccharomyces boulardii” and no strain designator against it, 5 Billion CFU, 250 mg for this organism. Disclosure score 48.8, rank 96.
• Designs for Health FloraMyces, 60 vegetarian capsules (SKU FLM060) Panel prints “Saccharomyces boulardii” and no strain designator against it, 500 mg per 2-capsule serving for this organism. this record carries a verification flag. Disclosure score 34.7, rank 112.
• Biotics 8 Men's High Performance Probiotic Formula Panel prints “Saccharomyces Boulardii” and no strain designator against it, 2 Billion CFU for this organism. Disclosure score 31.8, rank 116.
• Ancient Nutrition SBO Probiotics Men's Once Daily Panel prints “Saccharomyces boulardii” and no strain designator against it, no amount for this organism — it is inside a blend total, so the dose cannot be read off the label. Disclosure score 17.1, rank 135.
• Ancient Nutrition SBO Probiotics Ultimate Panel prints “Saccharomyces boulardii” and no strain designator against it, no amount for this organism — it is inside a blend total, so the dose cannot be read off the label. this record carries a verification flag. Disclosure score 17.1, rank 135.
• Garden of Life Primal Defense Ultra, 15 Billion CFU Plus HSOs, 90 Capsules Panel prints “Saccharomyces boulardii” and no strain designator against it, no amount for this organism — it is inside a blend total, so the dose cannot be read off the label. this record carries a verification flag. Disclosure score 17.1, rank 135.
• Vitamin Bounty Pro-50 Probiotic (Raw Probiotics), 60 veggie capsules Panel prints “Saccharomyces boulardii” and no strain designator against it, no amount for this organism — it is inside a blend total, so the dose cannot be read off the label. Disclosure score 17.1, rank 135.
• Vitamin Bounty Pro-Daily Probiotic Panel prints “Saccharomyces boulardii” and no strain designator against it, no amount for this organism — it is inside a blend total, so the dose cannot be read off the label. Disclosure score 17.1, rank 135.
• Vitamin Bounty Pro-100 Probiotic Panel prints “Saccharomyces boulardii” and no strain designator against it, no amount for this organism — it is inside a blend total, so the dose cannot be read off the label. Disclosure score 14.7, rank 142.
• DRFOSTER Probiotics for Men Digestive Health, 120 Billion CFUs with Saw Palmetto & Cranberry Panel prints “Saccharomyces boulardii” and no strain designator against it, no amount for this organism — it is inside a blend total, so the dose cannot be read off the label. Disclosure score 8.5, rank 147.
• Nutricost Probiotic Complex 50 Billion CFU, 60 Capsules Panel prints “Saccharomyces boulardii” and no strain designator against it, no amount for this organism — it is inside a blend total, so the dose cannot be read off the label. this record carries a verification flag. Disclosure score 8.5, rank 147.
• Onnit Total GUT HEALTH Panel prints “Saccharomyces boulardii” and no strain designator against it, no amount for this organism — it is inside a blend total, so the dose cannot be read off the label. Unranked — this label answers too little of the rubric to carry a disclosure score.
• YourBiology Gut+ (1 Month Supply) Panel prints “Saccharomyces boulardii” and the designator “SB01” against it, which is not the strain this section is about, no amount for this organism — it is inside a blend total, so the dose cannot be read off the label. this record carries a verification flag. Unranked — this label answers too little of the rubric to carry a disclosure score.

How much — and why this page will not tell you #

The next question a buyer asks is how much, and this page cannot answer it from the record. This fact base holds no published per-day amount for the pooled adult result — so there is no tested figure here to hold a panel against, and inventing one from a bottle's directions would be exactly the substitution this site exists to catch. What the panels themselves disclose is below, and it is worth reading for a different reason.

Amounts on these labels come in two incompatible units. Some panels print milligrams, a weight of the dried preparation. Others print a count of organisms — CFU, live cells, billions. No arithmetic a reader can do converts one into the other, so two bottles of the same species can be genuinely incomparable on the shelf. Both counts are below, split by unit, and the split is the finding.

When to start, which is the part most people get wrong #

The branch above ends in an instruction rather than a result, and it is the part of this page most likely to change what a reader actually does: start it early. The figure it cites is not an effect size and must not be read as one — nobody randomised early against late, so what it records is how far real practice drifted from the intention, in the largest pragmatic trial in the field. Treat it as a reason to open the box the day the antibiotics start, not as a demonstration that timing is what makes it work.

What the shelf sells for this instead, and how many bottles carry it #

These are the strains a shelf offers instead when the word "antibiotics" is on the box, and none of them is behind the graded row above. They are here because a buyer will meet them and because the count of labels carrying each one shows where the marketing went. What happened to each of them in adults — including the one that fails on this exact outcome while remaining significant in children — is argued strain by strain on the evidence page, with the certainty ratings attached.

L. rhamnosus GG (ATCC 53103)

36 labels print it; 9 print a count of organisms for it 1 prints milligrams and no count. 26 bury the organism inside a blend total.

L. casei DN-114 001

0 labels print it; 0 print a count of organisms for it Not printed on any panel this audit has transcribed.

The Bio-K+ line — L. acidophilus CL1285, L. casei LBC80R, L. rhamnosus CLR2

3 labels print it; 0 print a count of organisms for it 3 bury the organism inside a blend total.

What to do with all of that #

So, standing in a pharmacy with a prescription in your hand: buy a single-organism S. boulardii product whose panel prints a strain designator, start it the day the antibiotics start, and stop expecting it to do anything beyond the one outcome it is graded for. Those are three things a label and a calendar can settle before you swallow anything, which is more than most of this category offers.

Two things it is not for. It is not protection against C. difficile in adults — the branch above prints that result in the same paragraph as the good one, and it is not significant. And it is not a repair job for after the course, which is the promise this category sells hardest and the one the measured outcomes do not support. If diarrhoea while taking antibiotics is not your problem, this is not your purchase.

Scope for every count on this page: the organism slots transcribed from 156 of the 166 labels in this audit. A panel that is not fully transcribed can hide a line, so “not printed on any panel this audit has transcribed” means exactly that and not “not sold anywhere”. None of these counts is a statement about efficacy, and a product appearing on this page is not a recommendation of that product.

Read next #

• The antibiotic diarrhoea trial record The pooled result with its confidence interval and certainty rating, the two independent adult trials that failed, and what the class-level C. difficile figure does and does not support — including the question of what happens after the course.
• Are probiotics safe, and for whom The contraindications above in full, with the case reports, the regulator's wording and the population-level counts that put them in proportion.
• The best probiotic for IBS, and whether the dose is on the label The other buying page built on a graded row, where the answer comes out the other way: two strains earn a grade and neither is cleanly buyable.
• Every audited label, filterable All 166 teardowns, one row per product and one column per question, so you can sort on disclosure instead of on marketing.

The buying branch above was published on the home page under #buy-antibiotics and is relocated here unchanged, together with the six-assertion gate that checks the product it names against that product's own record. The counts around it are computed from the product collection on every build; nothing on this page is typed across from another one.