Best Probiotic for Men Evidence-graded, strain by strain

Probiotics for bloating and gas: the two strains that failed, and what the labels say

Two of the strains sold for this symptom have been measured on it in a large trial, and both of them came back negative. B. longum subsp. longum 35624 — the strain inside the product in this audit called Bloating Relief — was tested in a general population with the severity of abdominal discomfort and bloating as its primary endpoint, and the trial reported "no significant differences between the probiotic and the placebo groups in either abdominal discomfort or bloating (P>0.3)" (PMID 27845337). B. animalis subsp. lactis BB-12 was tested in 1,248 randomised adults, missed both of its pre-specified primary endpoints, and recorded that abdominal pain and bloating fell in every arm with "no difference between treatments was found" (PMID 26382580).

Neither of those is a small trial or an obscure one, and neither is disputed by the brands that print the strains. They are simply not the results the packaging implies. Both are graded here accordingly — BB-12 is Grade F for regularity and bloating, and 35624 holds a Grade C for global IBS symptoms that rests on a pivotal trial which enrolled women only.

And the ceiling over the whole symptom is worse than either strain. Pooling the field, the gastroenterology review that graded it found, verbatim, that "in the case of bloating, no category of probiotic demonstrated benefits in pooled analysis" (PMID 33315591). Not this strain rather than that one — no category.

, Independent analyst Last verified Disclosures

This is the buying page for the symptom. It reads the labels sold for bloating against what their own strains did when somebody measured bloating.

That does not settle whether a probiotic will do anything for you. It settles something narrower and more useful before you spend money: the strain on the front of the box is not carrying a bloating result, and in the two cases where somebody actually measured bloating, it is carrying a failed one. What is left to buy on is what the label discloses — the strain, and how much of it — because that is the part of the purchase that can be checked before you take it.

The two strains that were measured, and the labels that carry them

B. animalis subsp. lactis BB-12
3 of the 166 labels in this audit print it; 0 print a count of organisms for it 3 bury the organism inside a blend total, where the amount cannot be read at all.
B. longum subsp. longum 35624
5 of the 166 labels in this audit print it; 0 print a count of organisms for it 5 print milligrams and no count — a weight of powder, which is not the unit a trial doses in.

The trials behind both of those results — who was enrolled, who paid for them, and what the guideline bodies made of the field afterwards — are on the IBS and digestion evidence page. This page does not reprint them. What it does instead is read the bottles.

The 4 labels in this audit sold for gas or bloating

These are the labels in this audit whose own product name sells this symptom. Each one already has a full teardown on this site; what follows under each is that teardown's own opening finding, rendered from the record rather than rewritten here, so the sentence a reader gets is the sentence that was audited. The disclosure score beside each is a measure of what the label tells you, and it is not a measure of whether anything in the bottle works.

Each sentence above is the opening finding of that product's own teardown, rendered from the record rather than rewritten here, so it cannot drift from the page it came from. Follow the product name for the panel, the identifiers and the rest of what was checked.

The product named for bloating relief, read against its own panel

One of those four deserves reading twice, because it is the clean case. The product named for bloating relief carries 35624 — the strain whose general-population trial missed abdominal discomfort and bloating as its primary endpoint — and its panel then does something most panels in this audit do not: it prints the potency at both ends of the shelf life, and so discloses its own hundredfold decay.

Set that against the dose-finding in the strain's own literature. The pivotal trial tested three doses and only the 1×10⁸ arm beat placebo — the 1×10⁶ and 1×10¹⁰ arms did not (PMID 16863564). So the figure a buyer would want to compare is a count of organisms, and this label prints two of them, a hundredfold apart, one on each side of the figure that worked.

For scale on that disclosure: 11 of the 166 labels in this audit print potency at both endpoints — at manufacture and through the best-by date — and Align Probiotic Bloating Relief + Food Digestion, 28 Capsules is one of them. That is a real disclosure and it is why the decay is visible here at all. The full teardown, including the enzyme blend dosed in the wrong unit and the two different tenures the brand claims for its own recommendation ranking on the same page load, is on its own audit page.

The one place here where the claim and the trial endpoint match

There is one place in this audit where the claim on the box and the endpoint of a trial are the same thing. B. coagulans MTCC 5856, sold as LactoSpore, has a placebo-controlled human trial in adults with functional gas and bloating — not IBS in general, not regularity, that endpoint (PMID 36862903) — and a label carrying that strain advertises relief of occasional gas and bloating. Its teardown opens on exactly that, and the sentence is reproduced below from the record rather than rewritten here.

Read what that is, and stop there. It is a match between an advertised outcome and a trial's endpoint. It is not a report of what the trial found: this audit verified the strain, the endpoint and the identifier, and it did not read the results section, so no result is published here and none should be inferred from the match. A trial existing on the right question is a better starting position than the two strains above, which were asked the right question and answered it badly. It is not the same thing as an answer.

A handful of labels in this audit put that strain on a panel by itself, which matters more than it sounds: where a panel carries just that organism, the CFU figure on the front is the figure for the strain, and a blend total cannot bury it. The count is below.

4 of the 166 labels in this audit put that strain on a panel by itself. One of them is the label whose advertised outcome and whose strain’s trial endpoint are the same thing.

8 of the 166 labels print this designator, and 3 print a count of organisms for it. The trial dosed 2×10⁹ CFU a day. Read each printed figure below against that, and treat a milligram-only figure as what it is — an honest disclosure of a weight, silent on the count.

Two more strains a bloating shelf will sell you

Neither of these carries a bloating result either, and the counts below are larger than either of the two above — which is the shape of this whole category. HN019 is the strain marketed hardest for regularity, and its two larger clinical-endpoint trials both failed; GG has a mixed bloating and flatulence record that includes one clean null. Both are argued, with their trials and their funding, on the evidence page.

B. animalis subsp. lactis HN019
16 of the 166 labels print it; 7 print a count of organisms for it 8 bury the organism inside a blend total.
L. rhamnosus GG (ATCC 53103)
36 of the 166 labels print it; 9 print a count of organisms for it 26 bury the organism inside a blend total.

What to do with all of that

So the honest shopping rule for this symptom is short. Do not pay a premium for BB-12 or 35624 on a bloating claim — those are the two strains with a measured bloating result, and the results went the wrong way. If you want to try something anyway, buy a label that prints a strain designator and a count of organisms for it, so that you at least know what you took when you decide whether it worked.

And before any of that: bloating and gas that are new, persistent, painful, or come with weight loss or blood are a reason to be examined, not a reason to shop. Buying a supplement instead of getting examined costs a great deal more than the supplement.

Scope for every count on this page: the organism slots transcribed from 156 of the 166 labels in this audit, and the product names of all 166. A panel that is not fully transcribed can hide a line, so an absence here means “absent from the panels this audit has transcribed” and not “not sold anywhere”. 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.

Read next