Strain, not species, and certainly not category.
The single most important fact about probiotics is the one the packaging obscures. Probiotic effects are typically strain specific and dose dependent, which means a result obtained with one strain does not carry over to another organism sold under the same species name. As one 2024 review puts it, the strain specificity of probiotics is critical, as not all strains exert the same effects, with variability even seen within the same bacterial species, and these sources of variation remain largely unexplored in dogs and cats, making it difficult to generalise the benefits of even very similar products.
Host origin appears to matter too. The probiotic characteristics of microorganisms are linked to host specificity, and for successful use a strain should ideally be of host intestinal origin. Unfortunately, most probiotics marketed for companion animals are not originally derived from dogs or cats.
This is why a label reading Lactobacillus acidophilus tells you very little. It names a species. The evidence, where it exists, attaches to a numbered strain.
What the trials actually found, including when they found nothing.
The literature is not empty, and it is not uniformly positive. Both matter.
The same product, opposite results in two species. Enterococcus faecium SF68 at 2.1 × 109 CFU/g was tested in 182 shelter dogs and 217 shelter cats. Cats fed the probiotic had fewer episodes of diarrhoea than placebo. The probiotic had no effect on the incidence of diarrhoea in dogs. One product, one dose, one setting, two species, two answers.
The same synbiotic, opposite results in two populations. In healthy shelter dogs given E. faecium at 2 × 109 CFU per capsule with fructooligosaccharides and gum arabic, the incidence of diarrhoea was lower than in untreated controls. The same combination had little effect on faecal microbial diversity, clinical scores or immune response in dogs with chronic enteropathy. Healthy animals and sick animals are not interchangeable evidence.
A clear positive in a defined disease. Twenty dogs with inflammatory bowel disease received either combination therapy with prednisone and metronidazole or the multistrain probiotic VSL#3 at 112 to 225 × 109 CFU per 10 kg for 60 days. The probiotic significantly decreased clinical and histological scores, decreased CD3+ T-cell infiltration, and produced significant increases in regulatory T-cell markers. This is among the strongest canine results in the field, and note what it took: a defined disease, a specific multistrain product, a stated dose and 60 days.
A shorter illness, shortened. A cocktail of thermostabilised Lactobacillus farciminis, Pediococcus acidilactici, Bacillus subtilis and Bacillus licheniformis with L. acidophilus reduced convalescence time in acute self-limiting gastroenteritis in 36 dogs, from 2.2 days on placebo to 1.3 days.
A benefit without a microbiome change. In healthy cats, a combination of Saccharomyces boulardii at 2.0 × 1010 CFU/g and Pediococcus acidilactici at 2.5 × 1010 CFU/g increased faecal butyrate and total short-chain fatty acids, reduced the inflammatory markers calprotectin and myeloperoxidase, and raised faecal antioxidant capacity. There were no significant changes to faecal microbiome composition. A product can do something measurable without moving the community at all, which is worth remembering before treating a microbiome report as the scoreboard for a supplement.
And a direct failure. A mixture of probiotics and prebiotics was unable to prevent or ameliorate dysbiosis in healthy research cats given clindamycin at 75 mg orally once daily for 21 days. The most commonly assumed use case, protecting the gut through a course of antibiotics, was tested and did not hold up.
Do they stay? Sometimes, briefly, and it depends on the dose.
Persistence is measured less often than it should be. Where it has been, the answers vary widely. A canine-origin strain, E. faecium EE3, given for one week to 11 healthy dogs at 109 CFU/mL, survived gastrointestinal transit and was still detectable in faeces three months after administration stopped, at 6.83 ± 0.95 log CFU/g. Lactobacillus rhamnosus GG, by contrast, showed dose-dependent recovery and at 5 × 1011 CFU per day was still recovered only four days after cessation.
Three months against four days, both in dogs, both live organisms given by mouth. Whatever else that comparison shows, it rules out any general claim about how long a probiotic lasts.
The regulatory gap, which is the real problem.
Pre-, pro- and synbiotic products are not regulated like medicines and are often sold over the counter as nutritional supplements. Because they do not fall under the same governance and regulations as medicines, from both a manufacturing and a labelling point of view, this hinders transparent assessment of the source and quality of ingredients and increases concerns about the validity of any associated health claims.
The labelling problem is specific and easy to check on any product you own. Total probiotic counts are usually given, as CFU per capsule, but in multistrain products the count of each individual organism is typically unknown. The doses, sources and proportions of the prebiotics and fibres in synbiotic formulations are rarely disclosed and may differ from the versions tested when those components were studied on their own. One product described in the literature is listed simply as containing lactobacilli, bifidobacteria and bacilli species, with strains not specified.
If effects are strain specific and dose dependent, and the label gives you neither the strain nor the per-organism dose, then the label cannot tell you whether the product matches any study.
There is also a survival question. Few studies have been published on the quality of probiotic products for pets or on probiotic survival through the gastrointestinal tract, and formulations must ensure that live organisms reach the intestine in adequate numbers to have any effect at all.
Reading the newer evidence carefully.
Two recent studies illustrate why the type of evidence matters as much as the result.
A 2025 study of a prebiotic and postbiotic blend reported a favourable shift, with increases in Bifidobacterium and reductions in proteolytic Proteobacteria. It was conducted entirely in an in vitro simulator of the canine intestine, with no animals involved, and the authors note that effects varied by donor and may not be generalisable to all dogs. It is a reason to run a trial, not a result in a live animal.
A 2025 trial of an E. faecium and Lactobacillus plantarum combination reported reduced intestinal inflammation and improved gut homeostasis. The colitis it treated was chemically induced, which is a model of inflammatory bowel disease rather than the naturally occurring canine disease an owner would recognise.
Neither study is weak. Both are answering narrower questions than a product claim built on them would suggest.
A note on who funds this work.
Most reviews in this area are written by people with commercial relationships in it. One of the 2024 reviews quoted above discloses that its senior author has received research funding from companies producing or selling biotic products, naming several, and that the first author's doctoral research was sponsored by three of them. The review is careful and its cautions are the sharpest in the set, which is rather the point: disclosure lets you read a paper properly instead of guessing. Published work on probiotics in companion animals remains greatly limited compared with the human literature, and more controlled trials are needed.
We sell microbiome testing, so the same standard applies here. Nothing in this article recommends a product, and BAARK does not sell supplements.
Where BAARK fits.
A profile describes the community. It does not validate a product.
Where the published evidence supports a specific strain for a specific purpose, a BAARK report can name it and cite the basis. Where it does not, the report says so rather than filling the gap. A cat study above improved several measurable outcomes with no change to community composition at all, which is a direct warning against using a microbiome profile as a pass or fail test for a supplement.
BAARK is wellness profiling, not a diagnostic test. It does not detect, diagnose, stage or predict the progression of any disease. Decisions about probiotics, prebiotics or any other supplement for an individual animal belong with your veterinarian.
The bottom line.
The honest summary of the canine and feline probiotic literature is that a small number of specific strains have produced real, measurable benefits in specific situations, that several well-designed tests have produced nothing, that persistence after dosing ranges from four days to three months depending on the strain and dose, and that the products on the shelf are not regulated in a way that lets you match what you are buying to what was studied. Ask which strain and how much of it. If the label cannot answer, the evidence cannot either.