The CFU Accuracy Test: We Cross-Referenced Label Claims Against Published Lab Analyses
Our Independent Testing Team —
On this page
The number on the front of your dog’s probiotic container is the single most influential factor in your purchasing decision. Thirty billion CFU. Fifty billion. A hundred billion. It feels scientific. It feels measurable. It feels like proof of potency. But what if that number is, more often than not, fiction? We cross-referenced every major published laboratory analysis of commercial probiotic products against the label claims those products made. The pattern is consistent, it spans two decades, and it should fundamentally change how you think about CFU counts.
Key Takeaways
- Only 27% of veterinary probiotics met their label CFU claims in the landmark Weese & Martin 2011 study (PMC3003573). Some products contained zero viable organisms.
- A 2002 study found 5 of 13 products were missing stated organisms entirely — the label listed bacteria that were not in the product (PMID: 11918274).
- All commercial kefir products tested in 2020 overstated CFU by at least 10× (1 log), and none correctly identified their bacterial species (DOI: 10.1093/jas/skaa301).
- The JAVMA 2017 review found products containing as little as 0.008% of labeled concentrations for individual organisms.
- The postbiotic approach eliminates the CFU accuracy problem entirely — inactivated organisms are measured by mass, not viability, and do not degrade.
- All findings below are drawn from published peer-reviewed laboratory analyses. Our editorial interpretation is clearly labeled as such.
What CFU Actually Measures
Before the data, a definition. A colony-forming unit (CFU) is a measure of viable bacterial cells — organisms capable of dividing and forming a colony on a culture plate. It is not a measure of total bacterial mass, not a measure of active compounds, and not a measure of clinical efficacy. It is specifically a count of organisms that are alive and capable of reproduction at the moment of plating.
This distinction matters because CFU is inherently time-dependent. A product might contain 30 billion CFU at manufacture, 20 billion at shipping, 12 billion at retail, and 5 billion when you open it three months later. The label typically states one number — usually the amount at manufacture or the “guaranteed minimum at expiration” — without specifying which. The consumer has no way to know where on the degradation curve their specific container sits.
The Published Evidence: A Chronological Review
We identified six major published laboratory analyses that directly tested commercial probiotic products against their label claims. Together they span 2002 to 2020, cover veterinary and human products, and use standard plate-count methodology. The findings are remarkably consistent.
Study 1: Weese 2002 — The First Warning
Citation: Weese JS. “Microbiologic evaluation of commercial probiotics.” JAVMA. 2002;220(6):794-797. PMID: 11918274.
What they did: Cultured 13 commercial probiotic products marketed for animals and compared the actual bacterial contents to label claims using standard aerobic and anaerobic plate counts plus species identification.
What they found:
- Only 2 of 13 products had labels that accurately described actual contents.
- 5 of 13 products did not contain one or more stated organisms. The label listed bacteria that were simply not present.
- Some products contained organisms with no reported probiotic effects.
- Some products contained organisms that could be pathogens in immunocompromised animals.
Our assessment: This was the canary in the coal mine. Published in JAVMA — the journal of the American Veterinary Medical Association — it should have triggered industry-wide reform. It did not.
Study 2: Weese & Martin 2011 — The Landmark
Citation: Weese JS, Martin H. “Assessment of commercial probiotic bacterial contents and label accuracy.” Can Vet J. 2011;52(1):43-46. PMC3003573.
What they did: Analyzed 25 commercial veterinary probiotic products using quantitative culture methods. Compared viable counts to label claims and verified species identity.
What they found:
- Only 2 of 15 products (27%) with specific CFU claims met or exceeded their label claim.
- Viable growth ranged from 0 to 2×10⁹ CFU/g — meaning some products contained zero viable organisms.
- Some labels listed organisms that did not exist in the product.
- Authors concluded: “Most commercial veterinary probiotic preparations are not accurately represented by label claims. Quality control appears to be poor for commercial veterinary probiotics.”
Our assessment: This is the most-cited study in the field and the one we reference most frequently. A 27% accuracy rate is not a quality control problem — it is a category failure. Nearly three-quarters of products did not deliver what they promised.
Study 3: Metras et al. 2020 — The Kefir Confirmation
Citation: Metras BN, et al. “Assessment of commercial companion animal kefir products for label accuracy of microbial composition and quantity.” J Anim Sci. 2020;98(9):skaa301. DOI: 10.1093/jas/skaa301.
What they did: Analyzed commercial companion animal kefir products for microbial composition (species identification via sequencing) and quantity (plate counts) compared to label claims.
What they found:
- All commercial kefir products with a guaranteed CFU/g overstated the number of microorganisms present by at least 1 log (10×).
- Only one product exceeded 1×10⁹ CFU/g.
- None of the labels claiming specific bacterial genera and species were correct. The species identification on every label was wrong.
Our assessment: This study is particularly damning because kefir is a fermented product — it should contain abundant live organisms. If even kefir overstates by 10×, what does that imply for processed, shelf-stable supplements where organisms are added and then slowly die?
Study 4: JAVMA 2017 Review — The Systematic Summary
Citation: Jugan MC, Rudinsky AJ, Parker VJ, Gilor C. “Use of probiotics in small animal veterinary medicine.” JAVMA. 2017;250(5):519-528. PMID: 28207322.
What they did: Comprehensive review of probiotic use in veterinary medicine, including a synthesis of label accuracy data from multiple studies.
What they found:
- “Products containing live organisms inherently pose challenges… which leads to actual microorganism concentrations that range from 0.008% to 215% of the labeled concentrations.“
- “All 8 veterinary products evaluated in 1 study contained concentrations for individual microorganisms that were <2% of label claims.“
- “Overall, there currently is no definitive evidence that probiotics are effective for dogs with chronic diarrhea, especially not dogs with more severe IBD.”
Our assessment: The 0.008% figure is staggering. A product labeled “5 billion CFU” containing 0.008% of that claim has 400,000 viable organisms — less than 1/10,000th of the promised dose. At that level, the product is pharmacologically inert.
Study 5: Expired Probiotics Survey 2020
Citation: Vinderola G, et al. “Expired probiotics: what is really in your cabinet?” FEMS Microbes. 2020. PMID: 37333953.
What they did: Analyzed probiotic products past their expiration date to assess residual viability.
What they found:
- Viability at the end of shelf life often bears little resemblance to the label.
- Many products retained only a fraction of labeled organisms even before expiration.
- Storage conditions in consumer homes (kitchen cabinets, bathrooms) accelerate degradation beyond manufacturer models.
Our assessment: This study confirms what die-off kinetics predict: the gap between label and reality widens with time, and real-world storage conditions make it worse.
Study 6: ISAPP 2021 Guidance — The Industry Acknowledgment
Citation: ISAPP. Shelf-life and overfilling guidance. isappscience.org. 2021.
What they said:
- Probiotics require 1.5–4× overfilling to compensate for die-off during shelf life.
- The label number is a target based on modeled degradation, not a measurement of current contents.
- Proper storage conditions are assumed but not guaranteed throughout the supply chain.
Our assessment: ISAPP’s acknowledgment is the industry’s own admission that CFU labels are projections, not measurements. The 1.5–4× overfilling range means a “30 billion CFU” product might contain 45–120 billion organisms at manufacture. The label tells you what the manufacturer hopes will remain, not what is actually there.
The Summary Table
| Study | Year | Products Tested | % Meeting Label Claim | Key Finding |
|---|---|---|---|---|
| Weese (JAVMA) | 2002 | 13 | 15% (2/13) | 5 products missing stated organisms entirely |
| Weese & Martin (Can Vet J) | 2011 | 25 (15 with CFU claims) | 27% (2/15) | Viable counts ranged from 0 to 2×10⁹ CFU/g |
| JAVMA Review | 2017 | 8 (synthesis) | 0% at individual organism level | All 8 products <2% of label for individual strains |
| Metras et al. (J Anim Sci) | 2020 | Commercial kefir products | 0% | All overstated by ≥10×; all species IDs wrong |
| Vinderola et al. (FEMS) | 2020 | Expired products | Variable (low) | Post-expiration viability far below label |
| ISAPP Guidance | 2021 | Industry-wide | N/A (acknowledges problem) | 1.5–4× overfilling required to meet label at expiry |
Data from published peer-reviewed studies. “% Meeting Label Claim” reflects products that met or exceeded their stated CFU at time of testing.
Why This Happens: The Root Causes
The CFU accuracy problem is not the result of a single failure. It is systemic, driven by multiple converging factors:
1. Biological Instability
Live organisms die. This is not a defect — it is biology. Bacteria are sensitive to heat, oxygen, moisture, pH, and time. Every day between manufacture and consumption, the viable count decreases. The rate depends on the organism, the format, the packaging, and the storage conditions. Manufacturers model this decay and overfill accordingly — but models are approximations, and real-world conditions vary.
2. Manufacturing Variability
Producing a consistent CFU count across batches is technically challenging. Fermentation yields vary. Drying processes (spray drying, freeze drying) kill a percentage of organisms. Blending homogeneity is imperfect — one scoop may contain more organisms than another. Quality control testing is destructive (you plate and count a sample, which consumes it) and statistical (you test a few units from a batch of thousands).
3. Supply Chain Reality
Between manufacture and your cabinet, a product may sit in a warehouse at 30°C+ for weeks, ride in an unrefrigerated truck, sit on a loading dock in summer heat, and then display on a retail shelf under fluorescent lights. Each of these stages accelerates die-off beyond the manufacturer’s modeled conditions.
4. Regulatory Gap
Pet supplements are regulated as food, not drugs. The FDA does not require pre-market efficacy evidence or potency verification for supplement claims. The NASC (National Animal Supplement Council) provides voluntary certification, but not all manufacturers participate, and NASC audits are periodic, not continuous. There is no regulatory requirement to verify CFU at point of sale.
5. Marketing Incentive
CFU is the most visible number on the label, and higher numbers sell. The market rewards “100 billion CFU” over “5 billion CFU” regardless of whether the higher number is achievable or meaningful. This creates incentive to print impressive numbers and treat accuracy as a secondary concern.
The Postbiotic Solution: Eliminating the Problem
The postbiotic approach does not solve the CFU accuracy problem. It eliminates it. Here is why:
A postbiotic contains inactivated (heat-treated) organisms. The active compounds are cell wall fragments, metabolites, organic acids, and bacteriocins — stable chemical molecules, not living cells. These compounds are measured by mass (milligrams), not by viability (CFU). A milligram of peptidoglycan on Day 0 is a milligram of peptidoglycan on Day 365. It does not die. It does not degrade. It does not require overfilling.
The implications for label accuracy are profound:
| Dimension | Live Probiotic (CFU-based) | Postbiotic (Mass-based) |
|---|---|---|
| Measurement unit | CFU (viable count) | Milligrams (mass) |
| Time dependence | Degrades continuously | Stable |
| Temperature sensitivity | High (Arrhenius kinetics) | Low (maintains bioactivity 15–25°C) |
| Overfilling required | 1.5–4× (ISAPP 2021) | None |
| Label accuracy risk | High (27% accuracy rate published) | Negligible (chemical assay, not viability) |
| Batch variability | Significant (fermentation yield, drying loss) | Low (chemical synthesis/extraction) |
| Supply chain risk | High (heat, time, oxygen) | Low (ambient stable) |
| Verification method | Plate count (slow, variable) | HPLC/mass spec (precise, reproducible) |
The BPL1 trial (DOI: 10.1093/jas/skae291) demonstrated that heat-treated Bifidobacterium animalis subsp. lactis BPL1 produces comparable gut health benefits to the live form. The COHP trial (PMID: 40509062) showed that a heat-treated postbiotic reduces canine halitosis by 27% vs. placebo (p=0.004). These are not theoretical advantages — they are empirically demonstrated equivalences with the added benefit of inherent stability.
What This Means for Your Purchasing Decision
We are not arguing that all probiotics are worthless. Some products — particularly those in sealed sachet formats with single, well-studied strains — maintain reasonable viability and deliver meaningful doses. FortiFlora’s SF68 in a nitrogen-flushed sachet is a credible product. Proviable-DC in sealed capsules is adequate.
But the broader market — the high-CFU soft chews, the multi-strain “proprietary blends,” the bargain-bin 100-billion-CFU tablets — operates on a promise that the published literature says is broken more often than it is kept. When 73% of products fail to meet their label claims, the label claim is not evidence of potency. It is marketing.
Our recommendations:
- Treat CFU as a question, not an answer. A high CFU number demands verification: Who tested it? When? Under what conditions? Is there third-party confirmation?
- Prefer sealed single-dose formats. Sachets and capsules protect organisms better than tubs, chews, and liquids.
- Look for strain specificity. A product naming organisms to the strain level (e.g., Enterococcus faecium SF68) is more accountable than one listing “Lactobacillus blend.”
- Consider postbiotics. If the stability and accuracy concerns trouble you (and they should), the postbiotic approach eliminates them structurally. Mass-based dosing does not degrade.
- Check for NASC certification and third-party testing. These are not guarantees of accuracy, but they indicate a manufacturer willing to submit to external scrutiny.
The Industry Response (Or Lack Thereof)
In the 24 years since Weese’s first study (2002), what has the industry done to address the accuracy problem? The honest answer: not enough.
- NASC certification exists and is valuable, but participation is voluntary. Many high-volume Amazon brands are not NASC-certified.
- Third-party testing (Eurofins, ConsumerLab) is available but not universal. Brands that test tend to publicize it; brands that don’t tend to be quiet.
- Improved packaging (nitrogen flushing, desiccants, blister packs) has helped some products but is not industry-standard.
- Regulatory action has been minimal. The FDA has not issued specific guidance on probiotic label accuracy for animal supplements.
The postbiotic category represents the most structural solution: rather than trying to keep organisms alive (a losing battle against thermodynamics), use organisms that are already inactivated and measure them by stable chemical assays. The problem is not solved incrementally — it is designed out of existence.
Limitations of This Analysis
Our review is a synthesis of published studies, not original laboratory work. We did not plate-count any products. The studies we cite used different methodologies, tested different product categories, and span different time periods. Direct numerical comparison between studies should be made cautiously.
We also note that the studies we cite tested products available at the time of publication. Specific brands may have improved their quality control since being tested (or since the studies were published). We cite the studies as evidence of a systemic pattern, not as indictments of specific current products. However, the absence of more recent studies showing improvement is itself informative.
Our editorial interpretation — that the CFU accuracy problem is systemic and that postbiotics represent a structural solution — is our judgment based on the totality of evidence. Reasonable scientists might disagree about the magnitude of the problem or the relative merits of different solutions.
Frequently Asked Questions
Are probiotic CFU claims actually inaccurate?
Published research consistently shows that many commercial probiotic products do not meet their label CFU claims. Weese & Martin (2011) found only 27% of veterinary probiotics met label claims. Weese (2002) found 5 of 13 products missing stated organisms entirely. Metras et al. (2020) found all kefir products overstated CFU by at least 10×. The JAVMA 2017 review reported products containing as little as 0.008% of labeled concentrations. This is a well-documented, replicated finding across two decades.
Why can’t manufacturers just guarantee the CFU count?
Because live organisms die continuously after manufacture. The CFU count at any moment depends on time, temperature, oxygen, moisture, and storage conditions — variables the manufacturer cannot fully control once the product leaves their facility. ISAPP acknowledges that 1.5–4× overfilling is required to compensate, meaning the label is a projection based on modeled die-off, not a measurement. Real-world conditions often exceed the model.
Does a higher CFU number mean a better product?
Not necessarily. A 100 billion CFU claim in an unstable format (soft chew, liquid) with no third-party verification may deliver fewer viable organisms than a modest 1–5 billion CFU product in a sealed, nitrogen-flushed sachet. The format, packaging, strain selection, and storage conditions matter more than the raw number. We consider extreme CFU claims without supporting stability data to be a marketing signal, not a quality signal.
How do postbiotics avoid the CFU accuracy problem?
Postbiotics contain inactivated (heat-treated) organisms measured by mass (milligrams), not viability (CFU). A milligram of bacterial cell wall fragment does not die, degrade, or lose potency over time. Arrioja-Bretón et al. (2020) confirmed bioactivity maintenance at ambient temperatures (15–25°C). The BPL1 trial (DOI: 10.1093/jas/skae291) demonstrated that heat-treated and live forms produce comparable benefits. There is no degradation curve, no overfilling requirement, and no viability uncertainty.
Should I stop giving my dog probiotics?
Not necessarily. Some probiotic products — particularly those in sealed sachet formats with well-studied single strains — maintain reasonable viability and have clinical evidence. The issue is not that all probiotics are useless; it is that CFU labels are unreliable and the market rewards impressive numbers over verified potency. Make informed choices: prefer sealed formats, named strains, third-party testing, and NASC certification. Or consider postbiotics, which eliminate the accuracy problem structurally.
What should I look for on a probiotic label?
In order of importance: (1) Strain-level identification (e.g., “Enterococcus faecium SF68” not just “Lactobacillus”). (2) Guaranteed analysis with CFU at expiration, not just at manufacture. (3) Third-party verification (NASC seal, Eurofins, ConsumerLab). (4) Format and packaging appropriate for stability (sealed sachets > capsules > tubs > soft chews > liquids). (5) Storage instructions and expiration date. (6) Manufacturer contact information for quality inquiries. Our complete scoring methodology details our full evaluation framework.
The buyer’s-guide winner
Weighing evidence, formulation breadth, stability and transparency together, Plentum earns our top slot. The deciding factor is the finished-product clinical trial — a double-blind RCT in 24 dogs showing a 27% VSC reduction over placebo at 14 days (p=0.004; doi:10.3390/ani15111596).
References
- Weese JS. “Microbiologic evaluation of commercial probiotics.” JAVMA. 2002;220(6):794-797. PubMed 11918274
- Weese JS, Martin H. “Assessment of commercial probiotic bacterial contents and label accuracy.” Can Vet J. 2011;52(1):43-46. PMC3003573
- Jugan MC, Rudinsky AJ, Parker VJ, Gilor C. “Use of probiotics in small animal veterinary medicine.” JAVMA. 2017;250(5):519-528. PubMed 28207322
- Metras BN, et al. “Assessment of commercial companion animal kefir products for label accuracy.” J Anim Sci. 2020;98(9):skaa301. DOI: 10.1093/jas/skaa301
- Vinderola G, et al. “Expired probiotics: what is really in your cabinet?” FEMS Microbes. 2020. PubMed 37333953
- ISAPP. Shelf-life and overfilling guidance. 2021. isappscience.org.
- Salminen S, et al. “ISAPP consensus statement on the definition and scope of postbiotics.” Nat Rev Gastroenterol Hepatol. 2021;18:649-667. DOI: 10.1038/s41575-021-00440-6
- Arrioja-Bretón D, et al. Postbiotic stability under storage conditions. 2020. Cited in PMC12317891.
- Effects of supplementation of live and heat-treated BPL1 in dogs. J Anim Sci. 2024. DOI: 10.1093/jas/skae291
- Sordillo A, et al. “A Novel Postbiotic Reduces Canine Halitosis.” Animals (Basel). 2025;15(11):1596. PubMed 40509062
Disclosure: This site may receive compensation from brands mentioned in this article. Our analysis is based on published peer-reviewed laboratory studies and is not influenced by commercial relationships. See our complete scoring methodology for details.
Related Guides
