Probiotics
What probiotics are
Probiotics are live microorganisms that, when administered in adequate amounts, confer a health benefit on the host. This definition, established by the World Health Organization and the Food and Agriculture Organization, sets a high bar: to qualify as a probiotic, a microorganism must be live, delivered in sufficient quantity, and demonstrated to produce a specific health benefit. Products that simply contain bacteria do not automatically meet it.
The term covers a wide range of organisms. The most commonly studied genera in human trials are Lactobacillus and Bifidobacterium, both lactic acid bacteria with long histories of use in fermented foods. Saccharomyces boulardii, a yeast rather than a bacterium, is also well-studied and has distinct properties that make it particularly useful in specific clinical contexts. Within each genus, individual species are divided further into strains, and this level of distinction matters enormously: evidence generated for one strain cannot be assumed to apply to another, even within the same species. A product labelled simply "Lactobacillus acidophilus" tells you almost nothing clinically useful about what it will do.
The mechanisms through which probiotics influence human health are varied and not fully understood. Proposed pathways include competitive exclusion of pathogenic organisms, modulation of the mucosal immune system, production of short-chain fatty acids and other bioactive compounds, strengthening of the intestinal epithelial barrier, and indirect effects on systemic inflammatory signalling. The relative importance of each mechanism differs by strain and by clinical context, and for most consumer-facing probiotic products the mechanistic basis for any claimed benefit remains uncharacterised at the strain level.
Probiotics are found naturally in fermented foods including yogurt, kefir, sauerkraut, kimchi, and miso, though the strains present in these foods are generally not the same as those studied in clinical trials, and their quantity and viability are variable. Supplemental probiotics are available as capsules, sachets, liquids, and chewables. Formulation quality varies substantially between products, and the number of viable organisms at expiration rather than at manufacture is the relevant figure for dosing purposes.
Why strain specificity matters
The most important concept in probiotics science is that findings from trials of one strain cannot be extrapolated to other strains, even closely related ones. A systematic review and meta-analysis examining this principle directly found strong evidence that probiotic efficacy is both strain-specific and disease-specific (McFarland et al., Frontiers in Medicine, 2018). This has direct consequences for how evidence is interpreted and how products are evaluated.
In practice, strain specificity means that a well-conducted trial showing that Lactobacillus rhamnosus GG reduces antibiotic-associated diarrhoea does not support the claim that a different Lactobacillus rhamnosus product will do the same. It means that a product containing ten strains is not straightforwardly superior to a product containing one, because the additional strains may have no evidence for the intended outcome, and high strain counts can sometimes reduce colonisation of any individual strain. It also means that most consumer product labels, which identify strains vaguely or not at all, make it effectively impossible for a purchaser to assess whether the product reflects any of the relevant trial evidence.
The practical implication is that choosing a probiotic by genus or CFU count rather than by strain and indication is unlikely to produce reliable results. The evidence base for individual strains, reviewed in the linked entries below, is the appropriate starting point.
What the evidence shows
The evidence for probiotics is stronger in some contexts than in others, and the quality of that evidence varies considerably across conditions.
The areas with the most replicated clinical evidence are antibiotic-associated diarrhoea prevention, acute infectious diarrhoea treatment in children, and certain applications in irritable bowel syndrome. For antibiotic-associated diarrhoea, both Lactobacillus rhamnosus GG and Saccharomyces boulardii have multiple randomised controlled trials and meta-analyses reporting positive effects, with relative risk reductions in the region of 30 to 60 percent depending on strain, population, and antibiotic type. Effect sizes are directionally consistent but vary substantially across trials. Guideline bodies have reached different conclusions on this evidence: the AGA's 2020 clinical practice guideline recommended against routine probiotic use for antibiotic-associated diarrhoea in many contexts, citing high heterogeneity and low certainty, while other bodies have reached more permissive positions. This reflects a genuinely mixed evidence landscape rather than a settled one. For acute infectious diarrhoea in children, several strains, including LGG, S. boulardii, and Lactobacillus reuteri DSM 17938, have moderate-to-strong trial support for reducing duration, and ESPGHAN guidance has been more supportive of specific strains in this paediatric context. The IBS evidence is more problematic than its frequent citation implies: heterogeneity across trials is high, placebo response rates are substantial, and strain-level replication is inconsistent. Some strains show signal for specific symptom clusters; overall certainty is low.
The evidence for broader claims, general gut health maintenance, immune support in healthy adults, mental health benefits, weight management, is considerably weaker. A 2026 meta-analysis found that probiotic supplementation did not significantly alter gut microbiota alpha-diversity in healthy individuals (BMC Medicine, 2026). This is a meaningful finding, though it requires careful interpretation: alpha-diversity is a measure of within-sample species richness and evenness, and its absence of change does not rule out functional changes in microbiota activity or composition at other levels. What it does challenge is the specific consumer framing that probiotics build a more diverse and resilient microbiome, for which the evidence in healthy individuals is not established. Immune-related claims in healthy populations are generally supported only by biomarker data, changes in cytokine profiles or immune cell counts, for example, rather than clinical outcomes such as reduced infection rates. The gut-brain axis hypothesis is biologically plausible and has some clinical support in specific populations, but evidence in healthy individuals is preliminary.
Safety in healthy adults is generally good. Common adverse effects are minor and self-limiting, typically transient gastrointestinal symptoms such as bloating or gas in the first days of use. There are meaningful safety considerations in immunocompromised individuals, critically ill patients, and those with central venous catheters, where bacteraemia and fungaemia have been reported. Probiotics should be used with caution in these populations and only under clinical supervision.
Understanding the five strains covered in this library
Evidentia covers five of the most clinically studied probiotic strains. Each has its own evidence profile, primary use cases, and limitations. The summaries below are intended to help orient readers before consulting the full entries.
Lactobacillus rhamnosus GG (LGG) is one of the most extensively studied probiotic strains in human trials. Its strongest evidence sits in the prevention and treatment of antibiotic-associated diarrhoea and the management of acute infectious diarrhoea in children. Evidence in IBS is mixed. It is a bacterial strain and is therefore killed by antibiotics if taken simultaneously rather than spaced appropriately.
Saccharomyces boulardii is a yeast, not a bacterium, which gives it a distinct clinical property: antibiotics do not affect it, making it particularly suitable for use during antibiotic courses. Its strongest evidence is also in antibiotic-associated diarrhoea and acute diarrhoea. It is not appropriate for immunocompromised patients.
Lactobacillus reuteri has an unusually broad range of studied applications, including infant colic, oral health, H. pylori eradication as an adjunct, and some cardiometabolic outcomes. The specific strain designation matters substantially here, as DSM 17938 and ATCC PTA 6475 have meaningfully different evidence profiles. Evidence across most applications is moderate and requires cautious interpretation.
Lactobacillus acidophilus is one of the most commercially ubiquitous probiotic species, found in the majority of multi-strain consumer products. Its clinical evidence base is more limited than its market presence implies. Strain specificity is especially important here: most commercial products do not specify the strain used, and studies of uncharacterised L. acidophilus have produced inconsistent results.
Bifidobacterium lactis BB-12 has a reasonable evidence base, particularly in digestive regularity, immune modulation in healthy adults, and some infant health applications. It is one of the better-studied Bifidobacterium strains and appears in several multi-strain products where the strain is properly characterised.
How to read probiotic product labels
Several factors distinguish products that reflect the trial evidence from those that do not. The strain designation, full genus, species, and alphanumeric strain identifier, should be listed on the label. Without the strain identifier, it is not possible to determine whether the product contains the organism studied in any given trial. CFU count should be guaranteed at expiration, not manufacture; a product stating "50 billion CFU at time of manufacture" offers no assurance about viability at the point of consumption. Delivery mechanism matters: stomach acid degrades many bacterial strains before they reach the small intestine, and acid-resistant or delayed-release capsules substantially improve delivery for susceptible strains. Third-party testing by a named certifier, NSF International, USP, or equivalent, provides meaningful quality assurance; "third-party tested" without naming the certifier is an unverifiable claim.
Multi-strain products with high CFU counts are heavily marketed but not necessarily superior for specific indications. Higher strain counts do not straightforwardly produce broader effects, and for any condition with specific strain-level trial evidence, a well-characterised single-strain product reflecting that evidence is generally more appropriate than a broad-spectrum supplement.
Individual variation
Response to probiotic supplementation varies between individuals and is influenced by baseline gut microbiota composition, immune status, diet, age, and the specific health context. A concept receiving increasing attention in the literature is the distinction between responders and non-responders: individuals vary substantially in whether and how their gut microbiota changes in response to a given probiotic, and this variation is not reliably predictable from clinical or demographic characteristics. There is no reliable method for identifying in advance which individuals will respond to a given strain, and inter-individual variability in trial outcomes is consistently high, contributing to heterogeneous results across studies.
Women have specific contexts where probiotic evidence is more directly applicable, particularly in vaginal microbiome support and urinary tract health, where Lactobacillus-dominant communities play a well-established role. Pregnancy represents a context where certain strains have been studied for safety and some potential benefits, though evidence remains limited and clinical oversight is appropriate. Perimenopausal and postmenopausal women may experience changes in gut microbiota composition alongside hormonal transitions, though whether probiotic supplementation modifies relevant outcomes in this group is not well established.
Older adults tend to have lower gut microbiota diversity and may be more susceptible to gut dysbiosis following illness or antibiotic use, but evidence that probiotic supplementation meaningfully improves health outcomes specifically in this age group is limited. Infants and young children represent a distinct context: certain strains have meaningful evidence in paediatric diarrhoea and colic, and Bifidobacterium-dominant early microbiota appears important for immune development, though the evidence for supplementation to support this is still accumulating.
Immunocompromised individuals, including those receiving chemotherapy, those with HIV, and transplant patients, face a different risk profile. The general safety record of probiotics does not straightforwardly apply to these populations, where bacteraemia and fungaemia have been documented. Probiotic use in immunocompromised individuals should be supervised rather than self-directed.
Safety
In healthy adults, probiotics are generally well tolerated. The most commonly reported adverse effects are minor and transient, including bloating, flatulence, and altered stool consistency in the first few days of use. These typically resolve without intervention.
Serious adverse events are rare in healthy populations but have been documented in specific at-risk groups. Bacteraemia has been reported in immunocompromised patients receiving Lactobacillus-containing probiotics, and fungaemia has been reported with Saccharomyces boulardii in patients with central venous catheters or severe immune compromise. These are not common events, but they establish that probiotic use is not without risk in all contexts.
There is limited data on very long-term probiotic supplementation in healthy adults. Most trials run for weeks to months, and the effects, including any unintended effects on microbiota composition, of continuous supplementation over years are not well characterised. Given the evidence that standard probiotic doses do not substantially alter microbiota alpha-diversity in healthy individuals, concerns about permanent disruption are probably overstated, but long-term data is genuinely limited.
Interactions with antibiotics are clinically relevant. Bacterial probiotic strains are susceptible to antibiotics and should be separated by at least two hours from the antibiotic dose to maximise survival. Saccharomyces boulardii, being a yeast, is not affected by antibiotics and does not require this separation. Antifungal medications can suppress or eliminate S. boulardii and should not be taken concurrently if S. boulardii is being used therapeutically.
Some probiotic strains have been studied specifically in pregnant women without identified safety concerns, and certain strains are generally considered acceptable in pregnancy in clinical guidance. However, the safety record is strain-specific and the evidence base is not comprehensive enough to support blanket recommendations across probiotic products. Clinical guidance is appropriate for pregnant women considering supplementation.
What can reasonably be concluded
Probiotics are a biologically diverse category of interventions with real clinical evidence in specific contexts. The evidence is most replicated for prevention of antibiotic-associated diarrhoea and treatment of acute infectious diarrhoea, though effect sizes vary and guideline bodies differ in how they weigh the heterogeneity. Evidence for IBS is present but low-certainty. Evidence for broader wellness applications in healthy individuals is considerably weaker and frequently overstated in consumer-facing contexts.
Strain specificity is the central organising principle for evaluating probiotic evidence. Products that do not identify strains to the level used in clinical trials cannot be assessed against that evidence, which describes most consumer probiotic products. CFU count, species count, and marketing claims are not reliable proxies for clinical efficacy.
Safety in healthy adults is generally good, with the important caveat that immunocompromised individuals and critically ill patients require a more cautious approach. The evidence base continues to develop, and several areas, including the gut-brain axis, women's health, and long-term microbiome modulation, are areas of active research where conclusions should be held provisionally.
Where evidence is limited or outcomes are uncertain, conclusions should be treated as provisional and subject to revision as the evidence base develops.