Probiotics: strain specificity, what's evidenced, and what isn't

Probiotics are a top-five UK supplement category. The trial evidence is far more nuanced than the marketing implies. "Probiotic for gut health" is roughly as specific a claim as "drug for illness"; the question is which strain, for which condition.

What probiotics actually are#

Probiotics are live microorganisms intended to confer a health benefit. The WHO definition specifies "adequate amounts"; the dose and viability matter.

The main species used commercially:

• Lactobacillus (acidophilus, rhamnosus, casei, plantarum, GG, etc.)
• Bifidobacterium (bifidum, longum, breve, infantis, etc.)
• Saccharomyces boulardii (a yeast, not bacterial)
• Streptococcus thermophilus
• Bacillus species (more recent in supplements)

Strain specificity#

This is the most-misunderstood point. Lactobacillus rhamnosus GG is a different organism from Lactobacillus rhamnosus generally. Effects vary by strain even within the same species.

The trial evidence for specific indications is built on specific strains. Generalising from "this strain works" to "this species works" to "all probiotics work" is the marketing trap.

What the trial evidence supports#

Antibiotic-associated diarrhoea (AAD). Meta-analyses consistently support certain probiotics (particularly Saccharomyces boulardii and Lactobacillus rhamnosus GG) for reducing AAD risk. Effect sizes meaningful (NNT around 10–13).

Acute infectious diarrhoea in children. Some evidence for reducing duration with S. boulardii and L. rhamnosus GG. Modest effects.

IBS-D and IBS-M. A 2018 NICE review supported a trial of probiotics for IBS. But recommendation is to try a specific product for at least 4 weeks and stop if no benefit. No strong evidence for a specific best product.

C. difficile prevention. Some evidence for S. boulardii in high-risk patients; not a primary strategy.

Colic in infants. Specific evidence for L. reuteri DSM 17938 in breast-fed infants with colic.

Lactose intolerance. Specific lactose-fermenting strains can ease symptoms.

What the trial evidence does NOT support#

• "Boosts immunity" in non-deficient adults — small surrogate-marker effects; no good outcome data
• "Weight loss": popular claim with thin evidence; small effects in some trials, mostly null
• "Mood / mental health": "psychobiotics" marketing is way ahead of the evidence
• "Skin health": mostly marketing
• "Fixes leaky gut" — "leaky gut" itself isn't a clinical diagnosis; probiotic claims around it are unsupported
• "Reverses gut dysbiosis": supplementation doesn't reliably colonise; effects are mostly transient

The colonisation question#

A key technical point: probiotic supplementation does not reliably colonise the gut. Stool studies show ingested probiotic strains are mostly transient; they appear in stool while you're taking them, then disappear once you stop.

This means:

• The "rebuild your microbiome" framing is mostly wrong
• Effects, where they exist, are functional during use, not long-term colonisation
• Stopping probiotics returns the gut to its prior state within days to weeks

This isn't a criticism. It's a clarification of what's happening biologically.

Quality and dose#

Commercial probiotic quality varies dramatically. Issues:

• CFU counts (colony-forming units): products that promise 50 billion CFU may have far fewer by the time you swallow them
• Strain authentication: independent testing often finds different strains than the label claims
• Viability: many strains die during shelf storage; refrigerated products often have higher real CFU

For specific indications with evidence (AAD, IBS), use products with clinical-trial-grade strains: Florastor (S. boulardii), Culturelle (L. rhamnosus GG), VSL#3 (multi-strain for IBS).

Fermented foods#

Yoghurt, kefir, sauerkraut, kimchi, miso, and similar foods contain live microorganisms. Are they "probiotic"?

• They contain live cultures that can be detected in stool transiently
• The strains may not match the strains studied in probiotic trials
• The dietary patterns associated with regular fermented food consumption are independently healthful
• The PREDIMED-style trials don't separate fermented foods from broader Mediterranean patterns

Fermented foods are reasonable inclusions in most diets. Whether they're "as good as" probiotics depends on the specific outcome.

What we'd actually say#

For specific evidence-supported indications (particularly antibiotic-associated diarrhoea, IBS trial, or specific paediatric uses): use trial-grade products at trial doses. Effects are real and meaningful.

For "general gut health" in healthy adults, the trial evidence for routine probiotic supplementation is thin. A diet rich in fibre and naturally fermented foods is better-evidenced for microbiome health than supplementation.

If you're going to try probiotics for a specific concern, give it 4 weeks at the trial-evidenced dose. If you don't notice anything, stop.

When to talk to a clinician#

• Persistent diarrhoea or IBS symptoms that aren't being managed
• Considering probiotics during immunosuppression (some risk of bacteraemia)
• Pregnancy and probiotic decisions
• Recurrent C. difficile infection: specialist territory
• After bowel surgery
• Severe central venous catheter dependence (Saccharomyces fungaemia risk)