Quick Facts
- TypeLive microorganisms (bacteria and yeast)
- Key GeneraLactobacillus, Bifidobacterium, Saccharomyces
- Dosage Range1-100 billion CFU/day (strain-dependent)
- Strongest EvidenceIBS symptom relief (NNT 4-7); AAD prevention
- Key PointStrain specificity matters — effects cannot be generalized across species
What Are Probiotics
Probiotics are live microorganisms that, when administered in adequate amounts, confer a health benefit on the host. That is the WHO/FAO definition — and every word matters. "Live" means the organism must survive transit to the gut. "Adequate amounts" means the dose used in clinical trials, not whatever a manufacturer puts on a label. And "health benefit" means a benefit demonstrated for that specific strain, not for probiotics as a generic category.
This distinction is critical. Lactobacillus rhamnosus GG has strong evidence for preventing antibiotic-associated diarrhea. Lactobacillus rhamnosus HN001 does not have the same evidence base. Same species, different strain, different clinical profile. Treating all probiotics as interchangeable is the single biggest mistake consumers make — and the single biggest source of misleading marketing claims.
What the Evidence Shows
| Condition | Evidence | Key Finding | Dose/Strain |
|---|---|---|---|
| IBS | Moderate | Umbrella review of 27 SRs: symptom RR 0.78-0.79, NNT 4-7[1] | 1-10 billion CFU/day |
| Antibiotic diarrhea | Moderate | Multiple meta-analyses support prevention; S. boulardii and L. rhamnosus GG strongest[2] | S. boulardii 250-500 mg/day |
| Atopic dermatitis | Moderate | 13 RCTs: SCORAD SMD -0.13 (p=0.04); requires >2 months[3] | Lactobacillus + Bifidobacterium combos |
| Weight/body composition | Moderate | Waist circumference and visceral fat reduction in obese; inconsistent for body weight | 10⁹-10⁹⁰ CFU/day |
| Depression | Moderate | 7 RCTs: significant improvement (p<0.00001); no effect on IL-6/TNF-alpha[4] | Lacto + Bifido strains |
| Immune function | Emerging | Reduced URIs in athletes and children; strain-specific[5] | Varies by strain |
| Vaginal health/BV | Emerging | 35 RCTs for gynecological infections; 10 RCTs (n=1,234) for BV recurrence[6] | L. rhamnosus, L. crispatus |
| Cognitive function | Preliminary | Some meta-analyses show support but substantially more data needed[7] | Varies |
Deep Dives
Strain Specificity: The Most Important Thing
If you take one thing from this page, let it be this: probiotic effects are strain-specific. A product that contains "Lactobacillus acidophilus" without specifying the strain designation tells you almost nothing about what it will do. Clinical trials test specific strains at specific doses for specific conditions — and those results do not transfer to other strains of the same species, let alone other species entirely.
The table below maps well-studied strains to their evidence-backed indications. If a product does not list the strain designation (the alphanumeric code after the species name), you cannot verify whether it matches any studied strain.
| Condition | Well-Studied Strains | Notes |
|---|---|---|
| IBS | B. infantis 35624, L. plantarum 299v | Symptom-specific; may not help all IBS subtypes equally |
| Antibiotic-associated diarrhea | S. boulardii CNCM I-745, L. rhamnosus GG | Start with antibiotic course; take 2 hrs apart from antibiotic dose |
| Atopic dermatitis | Lactobacillus + Bifidobacterium combinations | Multi-strain combos outperform single strains; >2 months needed |
| Depression | Lactobacillus + Bifidobacterium combinations | Gut-brain axis mechanism; no inflammatory marker changes observed |
| Bacterial vaginosis | L. rhamnosus, L. crispatus | Vaginal and oral routes studied; adjunct to standard treatment |
A "50 billion CFU, 12-strain blend" sounds impressive on a label but is scientifically meaningless without strain-level identification and evidence for that specific combination. More strains and higher CFU counts do not equal better outcomes. Match the strain to the condition — or you are guessing.
The Quality Caveat
Even where probiotic evidence exists, its quality is often poor. A sobering finding from the IBS umbrella review: 66.7% of the included systematic reviews were rated "critically low" quality by AMSTAR-2.[1] This does not mean probiotics do not work for IBS — the NNT of 4-7 is clinically meaningful. It means the certainty of that estimate is lower than the volume of published research might suggest.
This pattern recurs across probiotic research: many studies, but often small sample sizes, short durations, heterogeneous strains lumped together, and inadequate blinding. When a meta-analysis pools 15 different strains across 30 trials and reports a positive effect, that finding may be driven by 2-3 effective strains diluted by 12 ineffective ones. Strain-level subgroup analyses are essential but rarely performed.
Safety
Probiotics are generally safe in healthy individuals. The most common side effects are temporary bloating and gas when first starting supplementation, which typically resolve within 1-2 weeks as the gut microbiome adjusts.
However, probiotics carry meaningful risks in specific populations:
- Immunocompromised patients: Risk of bacteremia (bacterial probiotics) or fungemia (Saccharomyces). Cases of Lactobacillus sepsis have been documented in ICU patients and organ transplant recipients. Probiotics should be avoided or used only under medical supervision in this group.
- Critically ill patients: The PROPATRIA trial demonstrated increased mortality with probiotic use in severe acute pancreatitis. Probiotics are not appropriate for ICU settings without specific clinical guidance.
- Short bowel syndrome: Risk of D-lactic acidosis from bacterial overgrowth in the shortened gut.
- Antibiotic timing: Take probiotics at least 2 hours apart from antibiotic doses to avoid immediate killing of the probiotic organisms.
References
- Umbrella review of 27 systematic reviews: probiotics for IBS. Symptom RR 0.78-0.79, NNT 4-7. 66.7% of SRs rated critically low by AMSTAR-2. 2026. PubMed
- Meta-analysis of probiotics for antibiotic-associated diarrhea prevention. S. boulardii and L. rhamnosus GG with strongest evidence. 2025. PubMed
- Probiotics for atopic dermatitis: 13 RCTs, SCORAD SMD -0.13 (p=0.04). Benefits require >2 months of supplementation. 2026. PubMed
- Probiotics for depression: 7 RCTs, significant improvement (p<0.00001). No effect on inflammatory markers IL-6 or TNF-alpha. 2026. PubMed
- Probiotics and immune function: reduced upper respiratory infections in athletes and children. Strain-specific effects. 2026. PubMed
- Probiotics for gynecological infections: 35 RCTs overall; 10 RCTs (n=1,234) specifically for BV recurrence prevention. 2024. PubMed
- Probiotics and cognitive function: meta-analytic support but insufficient data for firm conclusions. 2025. PubMed
- Probiotics for weight management in obese populations: waist circumference and visceral fat reduction; inconsistent effects on total body weight. 2025.
- AMSTAR-2 quality assessment methodology for systematic reviews. Shea BJ, et al. BMJ. 2017.
- World Health Organization/Food and Agriculture Organization. Guidelines for the evaluation of probiotics in food. 2002.
- PROPATRIA trial: probiotics in severe acute pancreatitis. Increased mortality in treatment group. Besselink MG, et al. Lancet. 2008.