Probiotics: Strain Specificity Matters More Than the CFU Number on the Front

The probiotic aisle is a marketing war fought over one number: colony-forming units per serving. Fifty billion. One hundred billion. Two hundred and fifty billion. The number gets bigger every year and it is almost entirely the wrong thing to look at.

Probiotic strains are as different from each other as dogs are from wolves. They share ancestors, they look similar under a microscope, and they occupy overlapping niches — and yet a Lactobacillus rhamnosus GG is not interchangeable with a Lactobacillus rhamnosus GR-1, any more than a border collie is interchangeable with a gray wolf. The strain designation — the letters and numbers after the species name — is doing all the pharmacological work.

This piece walks through what actually matters when you read a probiotic label, which strains have real human evidence for specific conditions, which claims are marketing filler, and how to think about the newer bacteria (Akkermansia, Faecalibacterium) coming out of the microbiome-research pipeline.

The taxonomy that matters

A probiotic organism is fully identified by three levels of naming.

• Genus. The broadest grouping. Lactobacillus, Bifidobacterium, Saccharomyces, Akkermansia. Genera share general biology but can differ enormously in behavior.
• Species. The next level down. Lactobacillus rhamnosus, Lactobacillus acidophilus, Lactobacillus plantarum. Species within a genus can occupy very different niches and produce very different metabolites.
• Strain. The specific isolate. Lactobacillus rhamnosus GG. Saccharomyces boulardii CNCM I-745. Bifidobacterium infantis 35624. This is the level at which clinical trials are conducted and at which effects can be attributed.

A label that says "contains Lactobacillus acidophilus, 10 billion CFU" tells you almost nothing. It tells you the genus and species. It does not tell you which strain. And because strains within the same species can have opposite effects on some outcomes, that is a big omission.

A label that says "contains Lactobacillus rhamnosus GG (LGG), 10 billion CFU" is a real product. The strain has been isolated, characterized, deposited in a public culture collection, and studied. You can look up its trials. You can predict what it does.

The strains with actual evidence, by condition

Probiotic trials are messy. Effect sizes are usually modest. Publication bias is a real concern. But for a small number of strain-condition pairs, the evidence is consistent enough to call it real.

Antibiotic-associated diarrhea

Lactobacillus rhamnosus GG (LGG) and Saccharomyces boulardii both have Cochrane review support for preventing antibiotic-associated diarrhea. Number needed to treat around 10 to 15. Start within 48 hours of the antibiotic and continue a few days after. S. boulardii is a yeast, so it survives the concurrent antibiotic.

Clostridioides difficile prophylaxis

Saccharomyces boulardii CNCM I-745 has the strongest evidence for reducing recurrence of C. diff infection when combined with standard antibiotic therapy. This is a niche use case with real clinical stakes and should be run through a clinician rather than self-directed.

Traveler's diarrhea

Saccharomyces boulardii again, plus LGG, have modest evidence for reducing incidence of traveler's diarrhea in high-risk destinations. Effect size is small; food and water hygiene is a bigger lever.

Irritable bowel syndrome

Bifidobacterium infantis 35624 has the cleanest RCT signal in IBS, particularly the diarrhea-predominant subtype. Whorwell 2006 showed meaningful improvement in composite IBS symptom scores at 10 to the 8 CFU per day — not billions. Modest but reproducible, and picked up in gastroenterology society guidelines.

VSL#3 or Visbiome — a specific eight-strain formulation — has evidence for IBS symptoms in some trials, but the evidence base is stronger for pouchitis and ulcerative colitis maintenance (see below).

Ulcerative colitis and pouchitis

VSL#3 / Visbiome has the strongest probiotic evidence in inflammatory bowel disease. Multiple RCTs support its use in maintaining remission in ulcerative colitis and in preventing pouchitis after ileoanal anastomosis. This is a therapeutic use case that requires physician supervision. The formulation and dose are specific.

Vaginal and urinary tract health

Lactobacillus rhamnosus GR-1 and Lactobacillus reuteri RC-14 have evidence for restoring vaginal flora after antibiotic disruption and for reducing recurrence of bacterial vaginosis. Note that these are different strains from LGG despite the shared species.

Infant colic

Lactobacillus reuteri DSM 17938 has moderate evidence for reducing crying time in breastfed infants with colic, though effect sizes vary across trials. Formula-fed infants may not benefit similarly.

What the evidence does not support

Read this list carefully. Most of the retail probiotic market is built on it.

• Generic "immune support." The phrase is meaningless as regulated language. The FDA does not let real pharmaceuticals claim to boost immunity. Probiotics do interact with mucosal immunity, but that does not translate into fewer colds or reduced illness in any consistent way across trials.
• Generic "gut health." A bottle that promises to "restore your gut" without naming the strain or the condition is selling vibes. Some strains help with specific conditions. There is no probiotic that broadly "heals the gut."
• Weight loss. Some interesting mouse data, mixed and mostly negative human data. Do not buy a probiotic for weight loss.
• Mood and "psychobiotics." The gut-brain axis is real. The clinical evidence for probiotic supplementation improving depression or anxiety in humans is still preliminary and inconsistent. Some early signals with Bifidobacterium longum 1714 and Lactobacillus helveticus R0052 plus B. longum R0175. Not ready for prime time.
• Skin. Some signal for atopic dermatitis in infants with specific Lactobacillus strains taken by mothers during pregnancy. Not established for adult acne, eczema, or rosacea.
• Probiotic multivitamins and probiotic protein powders. Almost always nonsense. The strain identity is usually not disclosed, the CFU count at the time of consumption is rarely guaranteed, and the delivery matrix is not designed for probiotic survival.

Why CFU count is the wrong axis

CFU stands for colony-forming units — a count of live organisms. The number matters, but not the way the marketing implies.

The B. infantis 35624 trial that got the strain into IBS guidelines used 10 to the 8 CFU. That is 100 million. The label on a competing IBS-marketed probiotic might advertise 50 billion CFU. The 50 billion product may or may not contain the strain that was actually studied. Fifty billion of a strain with no evidence for IBS is not better than one hundred million of a strain with evidence.

Higher CFU is not automatically better. Trial doses are matched to what worked in trials. Once you exceed the therapeutic dose for a given strain, more colonies do not translate into more effect. The extra bacteria simply pass through.

CFU count matters in two specific ways:

• It must exceed the studied dose for that strain. If the trial used 10 billion CFU of LGG and the bottle you buy has 100 million CFU of LGG, you are underdosing.
• It must be guaranteed at expiration, not at manufacture. Probiotic organisms die on the shelf. Reputable brands state the CFU count at the end of shelf life. Marketing brands state the count at manufacture, then let the number degrade over months of storage. Look for the phrase "guaranteed potency at expiration."

Delivery: the stomach acid problem

Most probiotic bacteria are killed by stomach acid. A capsule that does not protect them delivers a small fraction of the label CFU count to the intestine, where the bacteria are supposed to act.

Serious probiotic products handle this in one of three ways:

• Enteric coating. The capsule is coated to dissolve in the alkaline environment of the small intestine rather than the acidic stomach. Effective and inexpensive.
• Delayed-release capsule technology. The capsule shell is engineered to open only after passing through the stomach. Effective.
• Naturally acid-resistant organisms. Saccharomyces boulardii is a yeast that survives stomach acid without protective encapsulation. That is one reason it is popular for antibiotic-related indications.

A probiotic in an unprotected capsule taken with food is largely wasted. If the label does not mention delivery protection and the strain is a typical acid-sensitive Lactobacillus or Bifidobacterium, most of the dose is dying in the stomach.

Refrigeration. Some strains require refrigeration to maintain viability. Others are shelf-stable when properly lyophilized. If a product requires refrigeration and you bought it from a shelf, the CFU count at your consumption is unknown.

Prebiotics, probiotics, postbiotics — the terminology

The market has adopted three overlapping terms. Understanding them cuts through a lot of noise.

• Probiotics. Live organisms intended to confer a health benefit. This is what most of this article is about.
• Prebiotics. Fibers that feed beneficial bacteria already living in your gut. Inulin, fructooligosaccharides (FOS), galactooligosaccharides (GOS), partially hydrolyzed guar gum. Prebiotics can be genuinely useful and are cheaper than probiotics. They also sometimes cause gas and bloating during the first two weeks.
• Postbiotics. The metabolites bacteria produce (short-chain fatty acids like butyrate, for example) or heat-killed bacterial fragments. A newer category with less evidence but genuine interest, particularly for people who cannot tolerate live probiotics.

A synbiotic is a combination of a specific prebiotic and a specific probiotic engineered to work together. In practice, most bottles labeled synbiotic are a probiotic strain and some added chicory root fiber — real but modest.

The newer bacteria: Akkermansia and friends

The next generation of probiotics comes out of microbiome-sequencing research rather than dairy fermentation history. Two organisms are receiving the most attention.

Akkermansia muciniphila. A mucin-degrading anaerobe that lives in the gut lining. Depommier 2019 published a small human trial showing improvements in insulin sensitivity, waist circumference, and inflammatory markers after three months of pasteurized (not live) Akkermansia supplementation. Interesting signal, small trial, requires replication. Live Akkermansia is difficult to formulate because it is a strict anaerobe.

Faecalibacterium prausnitzii. A butyrate producer with observational associations to gut health. Not yet commercially available in most jurisdictions.

The Akkermansia in the current retail market is either pasteurized (killed with the intent of postbiotic-style effect) or in strain formulations that combine Akkermansia with other organisms and prebiotic fibers. Treat these as emerging rather than established. The mechanistic story is promising. The clinical trial base is still small.

How to read a probiotic label

1. Genus, species, and strain designation for every organism. "Lactobacillus rhamnosus GG" is a real label entry. "Lactobacillus acidophilus" alone is not.
2. CFU count per strain at expiration. Not at manufacture. If the product does not say "at expiration" or "guaranteed potency," the count is degrading in storage.
3. Delivery protection. Enteric coating or delayed-release capsule. Otherwise you are feeding stomach acid.
4. Storage requirements. Refrigerated or shelf-stable. If it needs refrigeration and it was on an ambient shelf, walk away.
5. Match strain to condition. Do not buy LGG for IBS or B. infantis 35624 for C. diff. Match the strain to what the strain has been studied for.
6. Third-party testing. Genuine potency verification is rarer here than in other supplement categories, but a few brands publish independent COAs.

Why supplement scores in this category cluster Provisional

Vyvata's supplement scoring is transparency-weighted. Probiotics score at the low end of the Provisional band more often than not, because the industry has particularly weak norms around strain disclosure at label level, guaranteed potency at expiration, and independent verification. Brands with strong strain-level science — Culturelle for LGG, Florastor for S. boulardii, Align for B. infantis 35624, Visbiome for the eight-strain UC formulation — would score higher on a probiotic-specific transparency rubric.

The honest summary

Probiotics work for a small number of specific strain-condition pairs. LGG and S. boulardii for antibiotic-associated diarrhea. S. boulardii for C. diff prophylaxis. B. infantis 35624 for IBS. VSL#3/Visbiome for pouchitis and UC maintenance. A few others in narrower indications.

Outside those pairings, most probiotic marketing is vibes. The bigger-CFU-count arms race is chasing the wrong axis. What matters is the strain designation, the delivery, the storage, and whether the strain has actually been tested for the condition you are trying to influence.

If you are taking a probiotic and cannot say the strain name, you are guessing. Look up the strain. Match it to the condition. Take it for at least four weeks before evaluating. And do not spend $80 a month on a general "gut health" product when the specific strain-condition pairs are cheaper and better-studied.