probiome Sign in
Strain

H. coagulans MTCC 5856

also LactoSpore · species Heyndrickxia coagulans
4 human trials · 1 review 4 single-strain · 0 in blends

Why does the strain code matter? Probiotics 101 →

In 219 products Multi-Strain Probiotic · InnovixLabsProbioMax IG 26 · XYMOGENBacillus Coagulans · Thorne All 219 →
probiome.ai Writtenfrom 4 study abstractsOct 16, 2026

What trials say about MTCC 5856

Most testing of MTCC 5856 looked at gut symptoms. All trials gave the strain on its own, not in a blend, and compared it with placebo . Both trials on bloating and gas, 106 people in all, reported improvement . For abdominal pain, 1 of 2 trials found an effect . One trial in 36 adults with diarrhea-predominant IBS (irritable bowel syndrome, a long-term gut condition) reported better symptoms . Neither of 2 trials on bowel habits, 100 people in all, found an effect .

One trial in 30 healthy adults found no clear difference in the makeup of gut bacteria . One trial in 56 healthy women tested a cream on the skin and reported less visible wrinkles and fine lines, but instrument readings of skin elasticity and moisture showed no clear difference .

On safety, 4 trials with 192 people in all measured side effects and body checks, and 0 of 4 found a difference from placebo .

probiome.ai

How this was made

What it is
An AI summary, written by probiome.ai. Not medical advice.
Written from
A table our code counted, and notes a model read from 4 study abstracts. Nothing else: no web search, no memory of other pages.
Numbers
Put in by code from our tables. The model can't add any; a sentence whose numbers don't match is dropped.
Checks
Every sentence cites a paper we hold. No advice, no cause words. All 8 sentences passed.
How
Drafted by probiome.ai. People at Probiome may refine it and regenerate it; every version is kept.
Rules it follows
evidence-grading v3: § 5 the words, § 7 summaries, § 14 bylines. Read the rules →
This version
v1 · Oct 16, 2026 · claude-sonnet-5-5 · prompt summary-v3

This describes what trials measured. It isn't advice about taking MTCC 5856.

The evidence behind it

Condition or outcomeTrialsPeopleFound an effectSingle strain · in a blendEvidence base
Side effects and safety 41920 of 44 single · 0 blend Some
Abdominal pain 21061 of 22 single · 0 blend Some
Bloating and gas 21062 of 22 single · 0 blend Some
Bowel habits 21000 of 22 single · 0 blend Some
Gut microbiome 1300 of 11 single · 0 blend Early
IBS 1361 of 11 single · 0 blend Early
Skin 1561 of 11 single · 0 blend Early
  • Side effects and safety 4 trials · 192 people · 0 found an effect 4 single · 0 blend Some
  • Abdominal pain 2 trials · 106 people · 1 found an effect 2 single · 0 blend Some
  • Bloating and gas 2 trials · 106 people · 2 found an effect 2 single · 0 blend Some
  • Bowel habits 2 trials · 100 people · 0 found an effect 2 single · 0 blend Some
  • Gut microbiome 1 trial · 30 people · 0 found an effect 1 single · 0 blend Early
  • IBS 1 trial · 36 people · 1 found an effect 1 single · 0 blend Early
  • Skin 1 trial · 56 people · 1 found an effect 1 single · 0 blend Early
  • Not tested in peopleNo trials in people.
  • EarlyOne single-strain trial, or only trials of blends.
  • SomeTwo or more single-strain trials.
  • Well studiedA meta-analysis or systematic review that reports this strain on its own, plus three or more single-strain trials with 300+ people in total.

Evidence base is how much testing there's been, not whether it works. Computed from the counts above by code, never by the AI. “Found an effect” counts trials that reported a significant result for that outcome. “People” adds up the trials' participants; a + means some trials didn't say how many.

Papers 5

Show All conditions Stomach painIBSDiarrheaDyspepsiaFlatulence AllSingle strain onlyBlends only

No papers match these filters.

Human trials, meta-analyses and systematic reviews naming the strain · titles and links only, never abstracts

How this was made

Papers come from PubMed: human trials, meta-analyses and systematic reviews that name the strain. A paper counts as single-strain when it names no other strain; the product tested may include other ingredients.

The summary was written by probiome.ai, Probiome's AI, in three steps. First, the model read each paper's title and abstract once and filled in a fixed form: who took part, how many, which strains, and what the authors reported for each outcome. Then code, not the model, counted those forms into the evidence table and graded each condition by the rule shown. Last, the model wrote sentences from the table and the forms only, citing a paper for every sentence.

Before anything is shown, code checks every sentence: each must cite papers we hold, every number must match those papers or the table, and it may not give advice, call the strain proven, or hide that a trial used a blend. Sentences that fail are dropped. It describes what trials measured; it isn't medical advice. The summary's own How this was made lists its sources, rules and versions.

Report a problem
Earlier versions · 1
  1. Version 1 (current) Oct 16, 2026 · 4 abstracts First version
Abstracts from PubMed, courtesy of the U.S. National Library of Medicine · summaries restate findings in our own words