A spoonful of kefir grains looks faintly like wet cauliflower. Add it to milk, leave it alone, and by the next day the milk has become tart and often lightly sparkling. Strain out the grains and the same small clusters can start the process again. There is something quietly remarkable about a food culture that can be passed from one kitchen to another without anyone needing to know which species are doing the work.
Kefir has travelled from the North Caucasus to supermarket shelves across the world, gathering a rather extravagant medical reputation on the way. It is said to restore the gut microbiome, control blood sugar and strengthen immunity. Some of these claims begin with real experiments. But reading the studies raises a problem more fundamental than their sample sizes: two drinks sold as kefir may be microbiologically different foods.
A culture older than its label
Traditional milk kefir begins not with a powdered starter but with living grains: dense associations of bacteria and yeasts embedded in a matrix of microbial polysaccharides and proteins. The grains are not cereal grains. They are communities that have developed a way of persisting together. Lactic-acid bacteria ferment milk sugars, producing lactic acid and much of the drink’s tartness. Yeasts, alongside some bacteria, can also generate carbon dioxide (CO₂) and small amounts of ethanol (ethyl alcohol); carbon dioxide helps explain the occasional sparkle. The quantities are not constant: the microbes present, fermentation temperature, duration and storage conditions all matter. A review of kefir’s flavour chemistry and production identifies lactic acid, ethanol and carbon dioxide among the characteristic products. Other organisms take part in exchanges that researchers have yet to fully reconstruct.
The grains contain a characteristic polysaccharide called kefiran, and their membership changes from one culture to another. Nevertheless, a working grain can remain recognisably stable across repeated fermentations. A 2025 review by scientists based in the North Caucasus describes the paradox: the microorganisms in the surrounding milk can shift substantially, while the grain community itself retains a degree of continuity. How this particular partnership arose is still uncertain. Katchieva and colleagues’ historical and microbiological review [1] is valuable partly because it treats that uncertainty seriously.
The strongest historical association is with the North Caucasus, particularly communities in what are now Karachay-Cherkessia and Kabardino-Balkaria. Stories about kefir’s first discovery differ, and the more picturesque details resist verification. Russian physicians took an interest in the drink in the late nineteenth century; production on a larger scale followed in the early twentieth. Thus “Russian kefir” is understandable as a description of later culinary history, but not a complete account of its origin. It is not, as another circulating question sometimes suggests, a food with a distinctively Jewish origin.
The bottle on the shelf is not necessarily the same ecosystem
Keeping traditional grains alive is more complicated than using a small, defined mixture of starter organisms. Manufacturers often favour repeatability: the same taste, acidity and shelf life from batch to batch. That can produce a good fermented dairy drink, but it may not reproduce the original microbial consortium.
In 2022, microbiologists compared grain-fermented kefirs with commercial versions using DNA-based methods and chemical analysis. The bacteria dominating the traditional grains were largely Lactobacillus kefiranofaciens; in the commercial drinks studied, Lactococcus lactis predominated. The aromas clustered differently, too. The commercial samples in that experiment contained more residual lactose and galactose than the traditional ones. The differences were measurable, not merely nostalgic claims that homemade food is more authentic. The full comparison [2] is particularly revealing because it looked at both organisms and their chemical products.
These were selected samples, not an audit of every brand on earth. Nor is a more complex microbial mixture automatically healthier. But the experiment exposes a problem with reading medical studies by product name. A clinical result from a drink prepared with one microbial community might not transfer to another that happens to share its label.
Fifteen people, five dairy drinks
The most useful human test of kefir may be a modest experiment published in 2003. Researchers recruited 15 adults who had difficulty digesting lactose. After an overnight fast, each volunteer consumed test meals containing the same 20 g of lactose, supplied as milk, plain or flavoured kefir, or plain or flavoured yoghurt. For eight hours, the researchers recorded symptoms and measured hydrogen in exhaled breath. The gas comes in part from microbes fermenting lactose that was not absorbed in the small intestine.
The result was quite specific. After milk, the mean integrated breath-hydrogen response was 224 ppm·h. After plain kefir it was 87, a reduction of about 61%. Plain yoghurt gave 76, remarkably close to kefir. The flavoured kefir was less impressive at 156. Participants also reported less flatulence with the fermented drinks. In other words, plain kefir performed better than milk in this small experiment, but it did not clearly outperform plain yoghurt.
| Test drink | Breath-hydrogen response |
|---|---|
| Milk | 224 ppm·h |
| Plain kefir | 87 ppm·h |
| Plain yoghurt | 76 ppm·h |
| Flavoured kefir | 156 ppm·h |
That comparison is more informative than a slogan about “good bacteria”. Equal lactose portions allowed the researchers to compare how people handled the test drinks, rather than simply observing that a product made from fermented milk might contain less lactose. Yet the trial was small, tested particular formulations and was funded by Lifeway Foods, the kefir manufacturer that supplied the product. The funding is worth knowing; it does not, on its own, invalidate the measurements. Independent replication with other products would be more persuasive. Hertzler and Clancy’s paper [3] remains a useful demonstration of what a carefully framed claim can look like.
It is also a practical one. For some people with lactose maldigestion, kefir may be easier to tolerate than milk. That is not the same as saying that all kefir is lactose-free, or that someone with milk-protein allergy can safely drink it. Fermentation changes milk; it does not turn it into an entirely different allergen-free food.
When a promising glucose result has the wrong name
A 2026 meta-analysis pooled intervention studies of kefir and reported an average reduction of 8.46 mg/dL (around 0.47 mmol/L) in fasting blood glucose, along with a reduction in HOMA-IR, an indirect measure of insulin resistance. Other outcomes, including lipids, body measurements and inflammatory markers, did not show consistent significant improvements. The analysis [4] offers a reason to keep studying fermented milk in metabolic health, though a pooled estimate is not a promise about the next bottle someone buys.
One influential trial shows why the identity of the drink matters so much. In Iran, 60 adults with type 2 diabetes were randomly assigned to drink 600 mL a day of either a product described as probiotic kefir or a conventional fermented-milk control for eight weeks. Some glucose measures improved more in the probiotic group. But the intervention was described in terms of a defined set of organisms, including Lactobacillus casei, L. acidophilus and bifidobacteria. It was not clearly established as milk fermented by an intact traditional kefir-grain community. The trial itself [5] is worth reading with the product description in view.
This is not a reason to dismiss the trial. It is evidence about the fermented milk that was actually tested. It is a reason, however, to resist transferring its result uncritically to every grain-fermented or supermarket kefir.
A separate 2026 overview counted 28 human clinical studies of kefir. Crucially, its eligibility rules excluded fermented drinks with additional defined probiotic strains, prebiotics or synbiotics. The older Iranian trial therefore should not be treated as if it automatically qualified for that review. The apparent inconsistency between the labels used in different studies is part of the methodological story, not an error that disappears by adding all the study counts together. The 28-study overview [6] is broader in design than a collection of 28 rigorous randomised trials.
The spectacular result that still needs company
Kefir has also been examined alongside treatment for Helicobacter pylori, the stomach bacterium linked to ulcers and some gastric cancers. In one randomised study of 82 patients, 36 of 46 people assigned kefir plus an antibiotic-based regimen tested negative for the infection afterwards, compared with 18 of 36 who received the same drugs with a milk placebo. That is 78.2% versus 50.0%, a difference too large to ignore casually. Participants given kefir also reported fewer treatment-related side effects.
It would be equally careless to call kefir a cure for H. pylori. Everyone in that experiment received prescription treatment; kefir was an addition, not a substitute. The trial was small, used an older antibiotic regimen and cannot be assumed to predict outcomes in settings with different resistance patterns. The published experiment [7] provides an intriguing signal, not a contemporary treatment protocol.
The distinction between an interesting experiment and a reliable therapeutic recommendation runs through the clinical literature. A systematic review published in 2023 found 16 randomised studies across 18 publications, but judged 12 studies to be at high overall risk of bias. Positive findings appeared in several areas, yet the trials were too varied and often too weak to justify condition-specific advice. The review [8] does not turn kefir into a useless food. It shows how much stronger the evidence needs to be before the food becomes a treatment.
Do kefir microbes actually settle in the gut?
A final popular claim concerns the gut microbiome. The story usually goes like this: drink kefir, introduce good organisms, and they establish a permanent home in the intestine. The first step is plausible. The last step is not established by finding an organism in a stool sample after somebody has been drinking it. That microbe may be passing through. Even a shift in the proportions of bacterial groups does not tell us automatically whether digestion, immunity or health improved.
A 2026 systematic review identified eight human intervention studies examining the microbiome after kefir consumption. The overall changes were modest and inconsistent, although Bifidobacterium increased in three of the four studies that measured it. Some kefir-associated organisms appeared in participants’ stool samples. What remains hard to demonstrate is durable colonisation, meaningful changes in microbial function and clinical benefits attributable to those changes. The microbiome review [9] makes the gap between microbial detection and medical benefit unusually clear.
There is a pleasing irony here. Kefir grains are a relatively persistent community outside the body; the evidence for permanent residence by their members inside the human gut is much weaker. The history of the drink and the biology of the consumer are not the same experiment.
What, then, is worth drinking?
For someone who likes fermented milk, kefir needs no extravagant claim to earn its place at breakfast. It can be a pleasant source of dairy nutrients, with a taste and texture unlike ordinary milk. If lactose tolerance is the goal, the human evidence gives a limited but reasonable argument for trying a suitable product and observing the response. People with milk-protein allergy should not confuse that with tolerance to lactose.
The label deserves more attention than the marketing. Is the drink made with actual kefir grains or a defined starter culture? How much sugar is added? Is the milk pasteurised? What is its protein content, and does the manufacturer provide credible information about the culture? These questions can help a reader choose a food without pretending that a microbiological name guarantees an outcome.
Homemade kefir brings the original living community closer, though not necessarily a proven health advantage. It also brings responsibility: use pasteurised milk, hygienic equipment and a reliable culture, then refrigerate the finished drink. Fermenting raw milk does not remove the infection risks associated with unpasteurised dairy.
The stronger the claim, the more precisely the drink has to be identified. We have some good human measurements, several provocative clinical findings and an unusually interesting microbial culture. But we do not yet have one standard intervention called “kefir” whose benefits can be predicted across people and products. The original grains have survived by maintaining relationships between organisms. Kefir research will become more useful when it describes those relationships with as much care as it counts the participants.
Editorial history: Originally published in February 2024 as Kefir – A Divine Beverage. Substantially revised in October 2026.
References
[1] Katchieva PK et al. Kefir revisited: Insights from the North Caucasus. World Journal of Experimental Medicine. 2025.
[2] Nejati F et al. Traditional Grain-Based vs. Commercial Milk Kefirs, How Different Are They? Applied Sciences. 2022.
[3] Hertzler SR, Clancy SM. Kefir improves lactose digestion and tolerance in adults with lactose maldigestion. Journal of the American Dietetic Association. 2003.
[4] Hamsho M et al. Effect of different kefir dosages on inflammation status, metabolic profile, and anthropometric measurements in adults. Nutrition, Metabolism and Cardiovascular Diseases. 2026; published online 2025.
[5] Ostadrahimi A et al. Effect of probiotic fermented milk (kefir) on glycemic control and lipid profile in type 2 diabetic patients. Iranian Journal of Public Health. 2015.
[6] Fijan S et al. Kefir Consumption and Health Effects Based on Human Clinical Trials: An Overview of Literature. Healthcare. 2026.
[7] Bekar O et al. Kefir improves the efficacy and tolerability of triple therapy in eradicating Helicobacter pylori. Journal of Medicinal Food. 2011.
[8] Kairey L et al. The effects of kefir consumption on human health: a systematic review of randomized controlled trials. Nutrition Reviews. 2023.
[9] Kefir: A Potential Gut Microbiota Modulator: A Systematic Review of Human Interventional Studies. 2026.