Why probiotic evidence is strain-specific, and does it hold?

Probiotic research results are usually described as strain-specific: a benefit shown in trials of one named organism is treated as evidence for that organism, not for probiotics in general or for a close relative. A large 2018 meta-analysis supports that rule across many conditions in adults and children. A 2023 review focused on children reaches a more cautious conclusion: the rule is widely accepted, but there is not yet enough direct evidence in pediatric trials to support or refute it.

An earlier article on what research shows about probiotics in children used strain specificity to explain why pediatric trials disagree. This one examines the idea itself and how well it has been tested.

Why do researchers say probiotic evidence is strain-specific?

A probiotic is a live microorganism intended to have a health benefit when consumed or applied to the body. The National Institutes of Health's National Center for Complementary and Integrative Health (NCCIH) notes that the most common probiotics are bacteria from the groups Lactobacillus and Bifidobacterium, and that some are yeasts, such as Saccharomyces boulardii (NIH NCCIH, last updated August 2019).

The same page states the core idea plainly: different types of probiotics may have different effects. If a specific kind of Lactobacillus helps prevent an illness, that does not necessarily mean another kind of Lactobacillus, or any Bifidobacterium, would do the same.

That is the strain-specificity rule: a trial result stays attached to the exact organism tested. It is a rule about how far evidence can travel.

What do the genus, species, and strain labels tell you?

Microorganisms are named in a hierarchy, and each level narrows the group:

• Genus is a broad group of related organisms, such as Lactobacillus.
• Species is a narrower group within a genus. Lactobacillus rhamnosus is one species.
• Strain is a specific lineage within a species, identified by a designation of letters or numbers. In L. rhamnosus GG, "GG" is the strain designation.

Two strains of one species can still differ in ways that may matter for a health effect, so a label naming only a genus, or only a species, leaves open which trial evidence applies.

Names are also less stable than they look. The authors of the 2018 meta-analysis described below list the changing taxonomy of bacterial and fungal species, and the lack of a global consensus on strain designations, among the limits of their own work (McFarland et al., 2018). The 2023 pediatric review, for example, writes Lacticaseibacillus paracasei, a genus name that does not appear in the 2018 paper.

What did the largest strain-level meta-analysis find?

A systematic review gathers every study meeting prespecified criteria; a meta-analysis combines their results statistically. In 2018, McFarland, Evans, and Goldstein published one designed specifically to test strain specificity in Frontiers in Medicine (PubMed 29868585).

Their stated motivation was that reviews often pool different probiotics together, which can mislead. They searched literature from 1970 to 2017 and included only trials of probiotics with an identifiable strain (a single strain or a defined mixture), and only where at least two randomized controlled trials existed for the same disease use.

They included 228 trials. Significant efficacy evidence was found for 7 (70%) of the probiotic strains assessed across four preventive uses, and for 11 (65%) across five treatment uses. Put the other way, about a third of the strains assessed in each group did not show significant efficacy.

The adult antibiotic-associated diarrhea results show what strain specificity looks like in data. Antibiotic-associated diarrhea is diarrhea that develops during or after a course of antibiotics. Within the Lactobacillus genus, a three-strain mixture and L. reuteri ATCC 55730 each showed a significant reduction in pooled trials, while L. rhamnosus GG, pooled across six trials, did not reach statistical significance for this use in adults (risk ratio 0.55, 95% confidence interval 0.25 to 1.18). The authors also report that L. rhamnosus GG and S. boulardii CNCM I-745 showed significant variation in efficacy from one disease to another, which is the second half of their conclusion: efficacy is both strain-specific and disease-specific.

A risk ratio compares how often an outcome occurs in treated and comparison groups; a 95% confidence interval that crosses 1, as this one does, means the data are compatible with no effect.

The authors were candid about limits. Few trials used the same strain or mixture for each specific disease. Strain designations were often missing from original publications, so the team traced product histories to assign them. Outcome measures varied widely, which limited pooling. Their conflict of interest statement also discloses advisory, lecturing, or board roles with probiotic companies, which readers can weigh alongside the methods.

How well has strain specificity been tested in children?

In 2023, Jankiewicz and colleagues published a rapid review, a streamlined systematic review completed on a shorter timeline, in the Journal of Pediatric Gastroenterology and Nutrition (PubMed 36563089). It opens by stating that the dogma of probiotic strain specificity is widely accepted, but only genus- and species-specific effects are supported by evidence from clinical trials.

The authors searched three databases through August 2022 for randomized trials in children aged 0 to 18 years, across conditions including antibiotic-associated diarrhea, acute diarrhea, necrotizing enterocolitis, respiratory tract infections, Helicobacter pylori infection, and atopic dermatitis. A trial was eligible only if at least one other trial tested a different strain of the same species against the same comparator, and trials without proper strain-level data were excluded.

Twenty-three trials qualified. Because no trial compared two strains directly, the authors made 11 indirect comparisons. With one exception, two Lacticaseibacillus paracasei strains in reducing atopic dermatitis symptoms, they found no significant differences between the clinical effects of different strains of the same species.

Their conclusion is carefully balanced: there is insufficient clinical evidence to support or refute the claim that probiotic effects in children are strain-specific. The authors reported no conflicts of interest.

Two method points keep this result in proportion:

• Indirect comparisons are weaker than head-to-head trials. Comparing strains across separate trials also compares their children, doses, settings, and outcome definitions. The authors themselves say a head-to-head comparison is the optimal design, and that such comparisons are lacking.
• "No significant difference" is not proof that strains are the same. Few, small trials often lack the power to detect real differences, so a null comparison means the data cannot tell the strains apart, not that they are equivalent.

The question resembles the one in pediatric extrapolation: when is it reasonable to let evidence from one source stand in for another, and what evidence would justify it?

Do the two reviews contradict each other?

McFarland and colleagues asked whether pooling different probiotics together hides real differences, and found that within broad groups, some strains showed efficacy for a given use and others did not. Jankiewicz and colleagues asked a narrower question in children: when two strains of the same species are tested for the same condition, do their results differ? In the available pediatric trials, mostly they could not tell.

Both reviews point to the same gap. Neither had many direct, strain-against-strain trials to work with, and both say so. NCCIH's broader summary fits the same picture: for irritable bowel syndrome, it reports that a 2018 review of 53 studies could not identify which species, strains, or combinations were most likely to be helpful (NIH NCCIH).

Strain specificity is best read as a sound default for reporting and interpreting evidence, not a settled finding for every condition and age group. The broader pediatric health and GI research on this site meets the same lesson elsewhere: how a question is defined shapes the answer, as with how researchers define functional GI disorders in children.

What does this mean when reading a probiotic study?

These questions help a reader judge how far a probiotic result can travel:

1. Is the strain named, or only the genus or species? Without a strain designation, it is hard to know which body of evidence the result belongs to.
2. Is the condition specific? Prevention during antibiotics, treatment of an infection, and a chronic condition are different questions.
3. Were strains compared directly? A head-to-head trial carries more weight than a comparison across separate trials.
4. How many trials, and how large? A conclusion resting on two small trials is fragile, and pooled reviews should still report results strain by strain.

Frequently asked questions

What is the difference between a species and a strain? A species is a group of closely related organisms, such as Lactobacillus rhamnosus. A strain is one specific lineage within that species, identified by a designation such as GG.

Does strain-specific mean other strains do not work? Not necessarily. It means evidence for one strain does not automatically transfer to another. Whether a different strain works is a separate question that needs its own trials.

A sound rule that still needs testing

Strain specificity keeps readers from treating "probiotics" as one thing, and the 2018 evidence supports it. The 2023 pediatric review is a reminder that a widely accepted rule is not the same as a tested one. Readers interested in how pediatric and gastrointestinal research questions are framed can explore the peer-reviewed publications listed on this site.