Best Probiotic for Men Evidence-graded, strain by strain

Do probiotics make you poop?

Not in the way the question usually means it. A probiotic is not a laxative, and on regularity and constipation this site grades the whole category D — "class-level pooled evidence only. Real, but it transfers to no product on a shelf, because the pool contains dozens of different organisms." The strain marketed hardest for "regularity" failed both of its clinical-endpoint trials, and the largest digestive probiotic trial ever run missed both of its pre-specified primary endpoints, one of which was defecation frequency. The bowel-habit change that does have strong evidence behind it runs the other way: less diarrhoea while you are taking a course of antibiotics.

, Independent analyst Last verified Disclosures

The strain sold to you for this, and what happened when it was tested

On regularity, the strain marketed hardest for it has the weakest case. B. animalis subsp. lactis HN019 has zero IBS trials ("HN019" AND "irritable bowel" = 0 records, re-verified 29 July 2026), and both larger constipation trials failed: "There were no statistically significant differences in the primary or secondary outcomes between interventions" (n=228, author at DuPont, the manufacturer; PMID 29227175) and "HN019 was not found to be superior in managing FC compared to placebo… This study did not support the expected effect of HN019 on improving CSBM frequency" (authors at International Flavors & Fragrances; PMID 40320938). The one positive result is a 14-day whole-gut-transit-time surrogate, funded by the Fonterra Research Centre with capsules supplied by Danisco USA, in a paper stating "The authors report no conflicts of interest" (PMID 21663486).

The numbers in that one positive trial, and the reason they are weaker than they look. 100 people were randomised for 14 days and whole-gut transit time was measured with radiopaque markers. The high-dose arm (17.2 billion CFU) went from 49±30 to 21±32 hours, p<0.001; the low-dose arm (1.8 billion) from 60±33 to 41±39 hours, p=0.01; placebo from 43±31 to 44±33 hours. The placebo arm started faster than either active arm, which is exactly the setup in which regression to the mean produces a result, and the symptom findings were within-group comparisons rather than tests against placebo (PMID 21663486; PMC3171707).

And it is a trial that collected the male answer and did not publish it. Waller 2011 randomised by sex, and reported that transit time at baseline differed by sex — women 55±34 hours versus men 41±24 hours, p=0.02 — but published no sex-stratified outcome. The data to answer "does this do anything for a man" were collected and never reported. That is the second time in this literature the same thing happened; the first is in the section below.

The largest trial in the field measured exactly this, and missed

Exactly one trial in the entire digestive-probiotic literature ran a pre-defined sex comparison, and it was null. BB-12, n=1,248 randomised / 1,000 per-protocol — the largest probiotic digestive RCT ever run. Randomisation was stratified by sex and hormonal status, so the sex subgroup was pre-defined, not post hoc. Men were 20.3–20.4% (~254). The result, verbatim: "There was no significant difference between the subgroups of sex/hormonal status (P=0·44)." Funding, verbatim: "Design, conduct, analysis and reporting of the study, as well as writing of this manuscript, were funded in full by Chr. Hansen A/S", with four employee authors (PMID 26382580).

That same trial missed both of its pre-specified primary endpoints. Defecation-frequency responder OR 1.31 (0.98–1.75), P=0.071; GI well-being not significant. The widely quoted OR 1.55 (1.22–1.96), P=0.0003 rests on a responder definition tightened after the fact, which the paper itself labels post hoc. And verbatim: "Abdominal pain and bloating decreased during the study in all the groups… but no difference between treatments was found."

What does change how often you go

One bowel-habit effect in this whole corpus reaches Grade B, and it is about stopping diarrhoea rather than starting a bowel movement.

S. boulardii for AAD in adults is significant and well-powered: 17.4% → 8.2%, RR 0.49 (95% CI 0.38–0.63), 15 RCTs, n=3,114 (PMID 26216624). That is the best-powered adult probiotic result in this entire corpus.

Why the effect stops when you stop

Because in most people it never took up residence in the first place. ISAPP, the field's own scientific association, verbatim: "Probiotics typically do not take up residence in your gut and may not evoke any detectable change in the microbes that are normally present." That is expert consensus on a consumer-facing page, not a systematic review, and it is graded that way here.

The measured version of the same point. In the 2018 Cell pair, mucosal colonisation was "person-, region- and strain-specific… indistinguishable by probiotics presence in stool", and the authors concluded that "empiric probiotics supplementation may be limited in universally and persistently impacting the gut mucosa." Two things follow that a monthly subscription does not advertise: what shows up in your stool does not tell you what reached your gut wall, and whatever effect you get is unlikely to outlast the capsules.

What to do with all that

  1. If you are going to try it anyway, at least buy a dose that matches a trial. Three products in this audit disclose HN019 at a figure close to the arm that produced the positive transit result — 2 billion CFU against the trial's 1.8 billion low-dose arm; one of them is in the frozen founding cohort and two were torn down after it. That is a dose match to a 14-day transit surrogate, and the same strain's two clinical-endpoint trials both failed. It is not a match to a clinical benefit and must not be read as one.
  2. Read the potency basis before the CFU number. A count measured at time of manufacture makes no commitment about what is alive when you swallow it. Which products state which is a column in the label audit.
  3. Give it a fixed trial and then stop. BSG's guidance on the neighbouring indication is to take a probiotic for up to 12 weeks and stop if there is no improvement, and NICE names no strain at all. Neither body is telling you to stay on one indefinitely.
  4. If a change in bowel habit is what sent you here, that is a symptom and it needs a diagnosis. A capsule cannot tell you whether the thing you have is the thing you think you have.

Which probiotic for constipation — and what this audit actually contains

This section does not end in a recommendation, because the grade above does not support one. What it can do is show you the shelf against the record. B. animalis subsp. lactis HN019 is the designator the regularity marketing rests on, so "which probiotic for constipation" becomes a question a label can settle: which panels in this audit print that code, what amount each one prints against it, how much of the rubric each label answers, and what taking it costs a day.

Read the amount, not the bottle, and read the unit before the number. A panel that prints this organism inside a blend total has disclosed the organism and withheld the dose. A panel that prints milligrams has weighed a powder, and no arithmetic converts a weight into the count of organisms a trial doses in. An amount stated as a count is the form that can be held against the 1.8-billion low-dose arm at all — and that arm produced a 14-day whole-gut-transit-time surrogate, not a clinical endpoint, in a trial whose placebo group started faster than either active group. The same strain's two larger clinical-endpoint trials both failed.

16 of the 166 labels in this audit print this designator, and 8 of those print an amount for this organism. 7 of the 8 give a count of organisms — the unit the trial dosed in — and 1 gives milligrams only. The remaining 8 bury the organism inside a blend total, where the amount cannot be read at all. They are ordered below by how much of the rubric each label answers, best disclosure first, because a disclosure ranking is the only ordering this page is entitled to offer.

A panel printing this designator is a fact about the panel. It is not evidence that the product was tested, not evidence that it delivers the amount any trial used, and not a reason to buy it. The roster is rebuilt from the transcribed organism slots on each build, so a label that stops printing the code leaves this list without anybody editing the page, and a label that starts printing it joins.

Scope for the counts above: 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 a product absent from this list is a product whose transcribed slots do not print the code — never a statement that no such product is sold.

Every identifier above is a link to the record it names. If a number here is wrong, the corrections page says how corrections are handled: dated, permanent, and quoting the original error.