Research dossier
Clinical research on Probiotics
10 trials reviewed across 3 indications.
Strongest evidence
Antibiotic-associated diarrhea, C. difficile, and IBS
Mechanism
Probiotics compete with pathogens for gut wall binding sites, produce short-chain fatty acids and bacteriocins, and modulate gut-immune signaling. The mechanism is non-specific; whether it translates into a clinical effect is entirely strain-dependent.
Strongest direct trial evidence is in three places: AAD prevention during antibiotics (L. rhamnosus GG and S. boulardii at 5–40 billion CFU/day), C. difficile prevention in hospitalized adults on antibiotics, and IBS symptom relief (B. infantis 35624 has the cleanest single-strain signal). Acute pediatric diarrhea benefit from older Cochrane data has been undermined by recent large North American RCTs.
Match the strain to the indication. L. rhamnosus GG and S. boulardii for antibiotic-associated diarrhea. B. infantis 35624 (Align) for IBS. Generic '30 billion CFU multi-strain' products are CFU-counting marketing — they may contain none of the strains that drove the trial signals.
Trials cited
Ford — probiotics for irritable bowel syndrome (meta-analysis)
positive · Meta-analysis
Ford et al., 2018, Alimentary Pharmacology & Therapeuticsn=5545Pooled across 53 randomized trials in 5,545 IBS patients, probiotics outperformed placebo on global symptom response and persistent symptom reduction. The honest read: the meta-analysis is positive overall, but the authors flagged that no single strain or combination has emerged as clearly superior — the field still cannot confidently match a strain to a patient.
High between-trial heterogeneity. The aggregate effect is real but the strain-to-strain variability means averaging masks responder versus non-responder products.
Whorwell — Bifidobacterium infantis 35624 for IBS
positive · RCT
Whorwell et al., 2006, American Journal of Gastroenterologyn=362Industry-fundedThe 10⁸ CFU dose of B. infantis 35624 — and only that dose — beat placebo and the other two doses on abdominal pain, bloating, and bowel-movement difficulty. The non-monotonic dose response is the cautionary tale: more CFU is not always better, and the strain plus the dose is what matters.
Industry-funded by Procter & Gamble (Align). Replicated separately in smaller trials, but the dose-response oddity has not been fully explained mechanistically.
Goldenberg — Cochrane review of probiotics for C. difficile prevention
positive · Systematic review
Goldenberg et al., 2017, Cochrane Database of Systematic Reviewsn=8672Pooled across 39 RCTs in 8,672 patients, probiotics during antibiotics cut C. difficile-associated diarrhea risk by roughly 60% versus placebo (moderate-certainty evidence). The benefit was concentrated in patients whose baseline CDAD risk was above 5%; in low-risk populations, the absolute benefit was small. Number-needed-to-treat around 12 in the at-risk subgroup.
The protective effect was strongest in studies where baseline CDAD risk exceeded 5% — typical hospitalized adults on broad-spectrum antibiotics. In low-risk outpatients, the absolute benefit shrinks.
Guo & Goldenberg — Cochrane review of probiotics for pediatric AAD
positive · Systematic review
Guo, Goldenberg et al., 2019, Cochrane Database of Systematic Reviewsn=8014Pooled across 33 trials in over 6,000 children, probiotics during antibiotics roughly halved AAD risk versus placebo, with a number-needed-to-treat of 9. High-dose preparations (≥ 5 billion CFU/day) of L. rhamnosus GG or S. boulardii drove most of the signal. Low-dose multi-strain blends did not reliably help.
Effect size depends on baseline AAD rate and on which strain was used. Generic 'multi-strain' products in the trials often did not reach the high-dose threshold and did not produce the same effect.
Allen — Cochrane review of probiotics for acute infectious diarrhea
positive · Systematic review
Allen et al., 2010, Cochrane Database of Systematic Reviewsn=8014Pooled across 63 trials in 8,014 participants, adjunctive probiotics shortened diarrhea duration by roughly 25 hours and reduced day-three stool frequency. Effect held alongside standard oral rehydration. The authors emphasized that this was add-on benefit, not a replacement for rehydration.
Updates since 2010 (Schnadower 2018 and Freedman 2018, large North American RCTs) found no benefit in outpatient pediatric gastroenteritis — undermining the strength of the original Cochrane signal in better-resourced settings.
Atopic eczema prevention and immune modulation
Mechanism
Early-life gut colonization shapes immune programming via Treg cell development and tolerance pathways. Probiotic exposure during pregnancy and infancy is hypothesized to redirect immune development away from atopic responses.
L. rhamnosus GG given perinatally halved atopic eczema incidence at age 2 in high-risk infants in Kalliomäki 2001. The 5-year follow-up showed the effect did not persist in the overall cohort, only in cesarean-delivered children. A real but narrow window — not an immune-boost claim, and not relevant to adults.
The atopic-prevention signal is for high-risk infants of atopic mothers, given perinatally. It does not extrapolate to 'probiotics boost the immune system' for healthy adults — that broad claim is not supported by clean RCT data.
Kalliomäki — Lactobacillus GG for atopic eczema prevention
positive · RCT
Kalliomäki et al., 2001, The Lancetn=132Atopic eczema incidence at age 2 was halved in infants whose mothers and themselves received L. rhamnosus GG perinatally — 23% vs 46% on placebo. The first solid signal that a specific probiotic strain delivered at the right developmental window moved a real allergic-disease endpoint.
Follow-up to age 5 (Kuitunen 2009) showed the prevention effect did not extend in the overall cohort, with persistence only in the cesarean-delivered subgroup. Treat as a perinatal-window intervention, not a permanent reset.
Kuitunen — 5-year follow-up of perinatal probiotics for IgE-allergic disease
mixed · RCT
Kuitunen et al., 2009, Journal of Allergy and Clinical Immunologyn=1018In the full cohort of 1,018 infants, the perinatal probiotic mixture did not reduce IgE-associated allergic disease at age 5. In the cesarean-delivered subgroup, the protective effect persisted (24% vs 41% on placebo). The right-strain-right-window-right-population caveat — not a universal prevention.
Subgroup-positive, overall-null. The cesarean-only finding is consistent with the gut-microbiota-seeding hypothesis but should be replicated before recommending broadly.
Vaginal flora, traveler's diarrhea, and gut-brain
Mechanism
Multiple strain-specific pathways: GR-1 + RC-14 reach the vaginal niche through gut and perineal translocation; S. boulardii survives gastric acid and competes with travel-pathogen colonization; L. helveticus R0052 + B. longum R0175 affect HPA-axis tone in early human stress trials.
Orally dosed L. rhamnosus GR-1 + L. fermentum RC-14 shifts vaginal flora and is widely used adjunctively in BV management. S. boulardii and specific Lactobacillus mixes modestly reduce traveler's diarrhea risk. Gut-brain probiotic effects on mood and stress are real but preliminary — small samples, strain-specific, not yet ready for label claims.
These are the niche use cases. Each requires the specific strain in published trials — generic 'women's probiotic' or 'mood probiotic' marketing rarely matches the actual strains tested.
Reid — oral L. rhamnosus GR-1 + L. fermentum RC-14 for vaginal flora
positive · RCT
Reid et al., 2003, FEMS Immunology and Medical Microbiologyn=64Daily oral GR-1 + RC-14 capsules shifted vaginal flora toward a Lactobacillus-dominant profile and reduced colonization by yeast and potential pathogens vs placebo. The pivotal proof-of-concept that orally dosed strains can reach and modify the vaginal niche.
Small sample. Subsequent BV-treatment trials with the same strains have shown more variable results, and the strongest case remains as adjunct to standard BV therapy rather than monotherapy.
McFarland — meta-analysis of probiotics for traveler's diarrhea
positive · Meta-analysis
McFarland, 2007, Travel Medicine and Infectious DiseasePooled across 12 trials, probiotics reduced traveler's diarrhea incidence vs placebo. S. boulardii and a Lactobacillus acidophilus + Bifidobacterium bifidum combination produced the cleanest signal. The effect size is modest — a single-digit absolute risk reduction depending on destination.
Heterogeneous destinations and pathogen exposures across the included trials. A probiotic that works in Mexico may not work in South Asia, where the pathogen mix is different.
Sarkar — psychobiotics review of bacteria-gut-brain signaling
mixed · Systematic review
Sarkar et al., 2016, Trends in NeurosciencesReviewed the small-but-growing literature on probiotic strains tested for mood, anxiety, and stress endpoints in humans (notably L. helveticus R0052 + B. longum R0175 in healthy adults; B. longum 1714 for stress). The signal exists but is preliminary — small samples, strain-specific effects, and limited replication.
Treat the gut-brain probiotic literature as preliminary. Mood-and-stress claims on probiotic labels run far ahead of the human RCT base.
10 forms of Probiotics compared
Culturelle
Lactobacillus rhamnosus GG (LGG)
Survives gastric acid; documented colonization in dosed range
Best forAntibiotic-associated diarrhea, acute pediatric diarrhea, atopic eczema preventionThe most-studied single strain in the entire probiotic literature. Specifically the GG strain — generic 'L. rhamnosus' on a label is not the same product. Culturelle is the consumer-facing branded form.
digestion5000000000–40000000000 cfuimmune5000000000–20000000000 cfuFlorastor
Saccharomyces boulardii
Yeast, not bacteria — naturally acid-resistant and antibiotic-resistant
Best forAntibiotic-associated diarrhea, C. difficile, traveler's diarrheaA non-pathogenic yeast rather than a bacterium. Because it is a fungus, common antibacterial antibiotics do not kill it — making it the natural pairing during an antibiotic course.
digestion5000000000–20000000000 cfuAlign
Bifidobacterium infantis 35624
Encapsulated to survive gastric transit; colonization documented
Best forIrritable bowel syndromeThe 35624 strain is the IBS workhorse. Whorwell 2006 showed only the 10⁸ CFU dose worked — higher CFU did not produce a larger effect. Generic B. infantis without the strain number is not the same product.
digestion100000000–1000000000 cfuLactobacillus reuteri (DSM 17938 and others)
Survives gastric transit
Best forPediatric infant colic and acute diarrhea, with strain-specific evidenceDSM 17938 has the cleanest infant-colic signal and is the strain in BioGaia products. Without the strain number, 'L. reuteri' on a label is functionally undosed for the indications that have evidence.
Lactobacillus acidophilus
Variable by strain
Best forGeneric gut support — strain-specific evidence is sparseThe most common probiotic on the market and the least strain-specific. Most products list 'L. acidophilus' with no strain number, which is the label equivalent of saying 'a dog' instead of 'a Border Collie.'
Lactobacillus plantarum (299v and others)
Acid-tolerant; documented colonization
Best forIBS bloating and abdominal pain (299v specifically)299v is the strain with the published IBS evidence. Generic 'L. plantarum' without 299v on the label is not the same product.
Lactobacillus fermentum (RC-14)
Reaches vaginal niche through gut-perineal translocation
Best forVaginal flora support, paired with L. rhamnosus GR-1The RC-14 strain is the vaginal-flora workhorse, almost always paired with L. rhamnosus GR-1. Used in the Reid 2003 trial and most subsequent vaginal-health probiotic products.
Bifidobacterium longum (BB536, 1714, and others)
Acid-tolerant; documented colonization
Best forStrain-dependent — BB536 for general gut, 1714 for stressOne of the more diverse genus-level entries. BB536 has general digestive trial data; 1714 has small early-stage stress and cognition signals.
Bifidobacterium lactis (HN019 and BB-12)
Acid-tolerant; survives shelf storage well
Best forGeneral digestive transit and constipation supportBB-12 and HN019 are the strains with most data — bowel transit and modest immune-marker effects. Common in yogurts and shelf-stable probiotic products.
'Probiotic' (no strain identification)
Unknown without strain identification
Best forMarketing — not a clinically defined productIf the label says 'probiotic blend' or '30 billion CFU multi-strain' without listing strain numbers, you are paying for CFU count, not a tested strain. The trial evidence on this page does not transfer to unlabeled blends.
Are you deficient? Symptoms, risk groups, lab tests
There is no clinical 'probiotic deficiency.' Native gut microbiota varies enormously between healthy individuals and is not measured against a deficiency cutoff. The relevant questions are dysbiosis (post-antibiotic, post-infection, IBS-associated) and indication-specific intervention.
Common symptoms
- Persistent diarrhea or constipation after antibiotic use
- Recurrent C. difficile infection
- Bloating, abdominal pain, or altered bowel habits consistent with IBS
- Recurrent bacterial vaginosis or vulvovaginal candidiasis
- Atopic eczema with strong family atopy history (in infants)
Who is at risk
e.g. amoxicillin-clavulanate, clindamycin, fluoroquinolones, broad-spectrum cephalosporins
Adults on antibiotic courses
Broad-spectrum antibiotics deplete commensal gut flora and create the niche for C. difficile overgrowth. L. rhamnosus GG and S. boulardii during the course substantially reduce AAD and CDAD risk.
Hospitalized older adults on antibiotics
Highest baseline CDAD risk and the population where probiotic prevention has the largest absolute benefit per Goldenberg 2017.
Adults with diagnosed IBS
Altered gut microbiota and gut-brain signaling. B. infantis 35624 has the cleanest single-strain signal; broader meta-analyses support multi-strain benefit but cannot identify a single best formulation.
Travelers to high-TD-risk destinations
Pathogen exposure in food and water. S. boulardii or specific Lactobacillus mixes started before travel modestly reduce traveler's diarrhea incidence.
Women with recurrent BV or vulvovaginal candidiasis
Disrupted vaginal Lactobacillus dominance. Oral GR-1 + RC-14 shifts vaginal flora; strongest case is as adjunct to standard antimicrobial therapy.
High-atopy-risk pregnancies (perinatal use)
Early-life immune programming. L. rhamnosus GG perinatally reduces 2-year atopic eczema incidence in infants of atopic mothers.
Side effects and drug interactions
Side effects
Mild gas, bloating, or transient GI discomfort
Common
Most common in the first 1–2 weeks of starting a new probiotic. Usually self-limiting as the gut adjusts.
Bacteremia or fungemia
Severe
Rare but documented in severely immunocompromised patients, neonates with central lines, and those with severe gut barrier compromise. Avoid probiotics in these populations without medical supervision.
D-lactic acidosis
Rare
Rare metabolic complication seen in short-bowel-syndrome patients on certain Lactobacillus strains that produce D-lactate. Indication-specific consideration, not a general-population concern.
Histamine sensitivity reactions
Uncommon
Some Lactobacillus strains produce histamine in the gut. People with histamine intolerance may experience flushing, headache, or GI upset on certain multi-strain products.
Drug interactions
Other
broad-spectrum antibioticsAntibacterial antibiotics kill bacterial probiotic strains. S. boulardii (a yeast) is unaffected and is the natural pairing during antibiotic courses. For Lactobacillus and Bifidobacterium products, separate dosing from the antibiotic by at least 2 hours.
Use S. boulardii or take bacterial probiotics 2 hours apart from antibiotic doses to maintain CFU through the gut.
Other
immunosuppressantsSeverely immunocompromised patients face elevated risk of probiotic-derived bacteremia or fungemia.
Do not initiate probiotics in patients on high-dose immunosuppression, post-transplant, or with central venous catheters without their care team's input.
Other critical caveats
- The strain number is the entire conversation. L. rhamnosus GG is not the same as random L. rhamnosus. B. infantis 35624 is not the same as generic B. infantis. If a product lists species without strain numbers, it is functionally unbranded — the trial evidence on this page does not transfer to it.
- '30 billion CFU multi-strain' products are CFU-counting marketing. CFU is the easy number to print on the front of a bottle; the strain mix and the per-strain dose are what would actually predict an effect. A single-strain B. infantis 35624 product at 1 billion CFU has cleaner IBS evidence than a 50-billion-CFU random multi-strain.
- Refrigeration and shelf life matter. Many bacterial probiotics lose CFU rapidly at room temperature. 'Guaranteed at expiration date' on the label is the only honest disclosure — 'guaranteed at manufacture' tells you nothing about what you're actually swallowing six months later.
- Severely immunocompromised, post-transplant, neonates with central lines, and short-bowel-syndrome patients should not start probiotics without their care team. Sepsis from probiotic-derived organisms is rare but documented.
- 'Boost your immune system' is not a clinically meaningful claim and does not have RCT support in healthy adults. The atopic-eczema prevention signal in Kalliomäki 2001 is specific to high-risk infants given a specific strain perinatally — it does not generalize to immune health in adults.
- Yogurt and fermented foods are not a substitute for clinically dosed strain-specific products. They contain live cultures, but the strains and CFU are not standardized to trial protocols. They are a fine general dietary practice; they are not the intervention if you are trying to replicate an AAD or IBS trial.
Frequently asked
What's the best probiotic to take?
There is no single best probiotic — match the strain to the indication. L. rhamnosus GG (Culturelle) or S. boulardii (Florastor) for antibiotic-associated diarrhea. B. infantis 35624 (Align) for IBS. L. rhamnosus GR-1 + L. fermentum RC-14 for vaginal flora. Generic '30 billion CFU multi-strain' is marketing, not a tested intervention.How much CFU do I need?
Strain-dependent. AAD prevention works best at ≥ 5 billion CFU/day of L. rhamnosus GG or S. boulardii. B. infantis 35624 worked in IBS at 1 billion CFU but not at higher doses — more CFU is not always better. The strain plus the dose used in the published trial is the only reliable target.Should I take a probiotic during antibiotics?
If you are at meaningful baseline risk for AAD or CDAD (hospitalized, older, broad-spectrum antibiotic course), yes — Goldenberg 2017 and the pediatric Cochrane review both support this. Use S. boulardii (which is a yeast, not killed by antibacterial antibiotics) or take bacterial probiotics 2 hours apart from the antibiotic dose. Lower-risk outpatients see smaller absolute benefit.Do probiotics boost immunity in healthy adults?
The 'immune boost' marketing claim is not well supported. The strongest immune-related probiotic signal is for atopic-eczema prevention in high-risk infants given L. rhamnosus GG perinatally — that does not extrapolate to general immune health in adults. Healthy adults taking a daily probiotic 'for immunity' have minimal RCT evidence backing the practice.Are probiotics safe?
For the vast majority of healthy adults and children, yes. Mild gas and bloating in the first 1–2 weeks are common and self-limiting. The exceptions are severely immunocompromised patients, neonates with central lines, post-transplant patients, and people with severe gut barrier compromise — sepsis from probiotic-derived organisms is rare but documented in those populations and they should not start probiotics without their care team.Refrigerated or shelf-stable — does it matter?
Yes, often. Many bacterial probiotic strains lose viability at room temperature, so a shelf-stable product needs strong encapsulation or strain selection (S. boulardii is naturally heat-stable; some Bifidobacterium lactis variants are too). The label that matters is 'CFU guaranteed at expiration date' — products that only guarantee CFU at manufacture can degrade to a fraction by the time you take them.
References
- 01Examine.com — Probiotics summary
- 02PubMed — Goldenberg 2017 (Cochrane review of probiotics for C. difficile prevention)
- 03PubMed — Ford 2018 (probiotics meta-analysis for IBS)
Last reviewed2026-05-07