Household Ingredients Compared › Guides
Fermented foods have gone from a niche health-shop category to a mainstream supermarket shelf — kimchi, kombucha, kefir, probiotic yogurt, tempeh, miso — and the word doing most of the selling is "probiotic". The problem is that "fermented" and "probiotic" are not the same word, and most fermented foods are not probiotics. A probiotic is a specific live microbe, tested at a specific dose, for a specific benefit. Most fermented foods contain a mix of microbes that has never been tested that way. That does not make them worthless — one small, well-run trial found real changes from eating more fermented food — but it does mean the label has to be read carefully, and the strongest claims on the shelf are usually the weakest ones in the evidence.
Fermentation is food made through the wanted growth of microbes. Bacteria or yeast convert sugars and other compounds in the food into acids, alcohol, gases and flavour compounds. Lactic acid bacteria turn milk into yogurt and cabbage into kimchi; yeast and acetic acid bacteria turn sweetened tea into kombucha; moulds and bacteria turn soybeans into tempeh and miso. The process preserves the food, changes its flavour and texture, and in some cases creates vitamins — but the microbes themselves are not the point of every fermented food, and in many of them they are gone by the time you eat it.
An expert panel of the International Scientific Association for Probiotics and Prebiotics (ISAPP) defined fermented foods in 2021 and sorted common ones by whether they usually still contain live microbes when eaten:
| Usually still contains live microbes | Usually does not |
|---|---|
| Yogurt and kefir, unheated fermented vegetables such as kimchi and sauerkraut, tempeh, miso, natto, most kombuchas, most cheeses, sour cream | Bread (baked), most beers and wines (microbes removed), heat-treated or pasteurised sauerkraut, soy sauce, vinegar, some kombuchas (pasteurised) |
Two common mistakes follow from that table. First, vegetables pickled in vinegar are not fermented at all — they are preserved in acid, and no microbes were involved. Second, pasteurisation kills the microbes — a shelf-stable, pasteurised kimchi or kombucha may have been fermented once, but it contains no live cultures now, however much the label talks about fermentation.
The FAO and WHO define a probiotic as a live microorganism that, when administered in adequate amounts, confers a health benefit on the host. The ISAPP panel's version is the same idea in plainer terms: a specific strain, tested at a specific dose, for a specific benefit. That definition has three parts, and most fermented foods fail all three:
The panel's recommendation is blunt: fermented foods whose microbes have not been identified down to the strain, in variable amounts, and whose health benefits have generally not been shown, should not be labelled as containing probiotics. A product that says "probiotic" without naming a strain and a dose is using the word as marketing, not as a description.
The fermented-food trial. The best human evidence for fermented foods as a category comes from one small trial at Stanford, published in Cell in 2021. Eighteen healthy adults ate a diet rich in fermented foods for 10 weeks — going from under half a serving a day to about 6 servings a day, choosing among yogurt, kefir, fermented cottage cheese, kombucha, vegetable brine drinks and fermented vegetables such as kimchi. Gut microbe diversity rose steadily, and 19 of 93 inflammation-related proteins measured in the blood went down. The trial was small, short, limited to healthy volunteers, and its primary outcome — a measure of how immune cells respond when stimulated — did not change in either group. The results are promising, and they have not yet been repeated at a larger scale.
Probiotics in specific situations. The clearest evidence for probiotics comes from supplements, not foods, and from narrow uses. A 2019 Cochrane review of 33 trials in 6,352 children taking antibiotics found diarrhoea affected 8% of children given probiotics against 23% in control groups, with higher doses (at least 5 billion CFU a day) working better. A 2021 meta-analysis of 42 trials and 11,305 participants reported a relative risk of 0.63 for antibiotic-associated diarrhoea. For Clostridioides difficile-associated diarrhoea, a Cochrane review found incidence of 1.5% in probiotic groups against 4.0% in controls. These are concentrated supplements of specific strains, at specific doses, in specific situations — not a yogurt with a probiotic claim.
Probiotics for healthy people. The case is much weaker. A systematic review of seven randomised trials in healthy adults found no effect on gut bacteria diversity, richness or evenness compared with placebo. The US National Institutes of Health states there are currently no formal recommendations for or against probiotics in healthy people. The American College of Gastroenterology's 2021 guideline recommended against probiotics for global IBS symptoms, citing very low-quality evidence, and the American Gastroenterological Association's 2020 guideline made conditional recommendations only for preventing necrotising enterocolitis in preterm infants and C. difficile-associated diarrhoea in patients on antibiotics — and none for Crohn's disease, ulcerative colitis, pouchitis or IBS.
Singapore has no specific legal definition of "probiotic" in its food law, and no published list of microorganisms permitted as probiotics in food — unlike Malaysia, the Philippines and Thailand, which have both. What Singapore does have is the general framework: probiotic strains used in food and supplements must be assessed for safety, and health claims are subject to substantiation requirements under the Singapore Food Agency's framework, with the Health Sciences Authority covering therapeutic goods and certain health supplements. In practice, that means a product making a probiotic or functional claim in Singapore must be able to substantiate it with scientific evidence — and a claim that names no strain and no dose is difficult to substantiate, however often it appears on shelves.
Two practical consequences follow. First, "contains live cultures" is not the same as "probiotic": live cultures is a fact about the product, probiotic is a claim about a benefit, and only the second needs a strain, a dose and a trial. Second, imported products carry their home country's rules: a kombucha or yogurt imported from the US or Europe was labelled under that country's framework, and the claim on its front may mean something different from what the same words would mean under Singapore's rules. The check is the same either way — strain named, dose stated, benefit specified.
| Food | Live microbes? | What to check on the label |
|---|---|---|
| Yogurt | Usually yes | Strain names and CFU if a probiotic claim is made; total sugar per 100 g (flavoured yogurts can carry 10–15 g of sugar); whether it is plain or sweetened. Yogurt is the most studied fermented food, with more than 20 randomised trials of yogurt and cultured milk. |
| Kefir | Usually yes | Similar to yogurt, with a wider mix of bacteria and yeasts. At least one randomised trial exists. Often lower in lactose than milk, because the cultures digest some of it — see the dairy, lactose and calcium guide. |
| Kimchi | Usually yes, if refrigerated and unpasteurised | Sodium first: a typical cabbage kimchi has about 498 mg of sodium per 100 g (USDA), so a 50 g side gives about 249 mg — around a tenth of the 2,000 mg daily limit. Check whether it was pasteurised, and whether fish sauce or shrimp paste is listed (an allergen and a dietary issue for some). At least one randomised trial exists. |
| Kombucha | Usually yes, if unpasteurised | Sugar per 100 ml (residual sugar varies widely, often 3–4 g per 100 ml); alcohol — commercial kombuchas usually aim to stay at or below 0.5% ABV, and in Singapore a drink above 0.5% ABV is treated as alcoholic; whether it is pasteurised. The human evidence for kombucha is mostly laboratory and animal work, not trials. |
| Tempeh | Usually yes | Protein (about 19–20 g per 100 g) and fibre (about 9 g per 100 g) — tempeh is worth eating for the food itself, and the fermentation is a bonus. See the protein guide. |
| Miso | Usually yes, if unpasteurised | Sodium is very high — miso paste is about 3,700 mg of sodium per 100 g — but a serving is a tablespoon in soup, not 100 g. Pasteurised miso has no live microbes. |
| Sauerkraut | Only if refrigerated and unpasteurised | Sodium about 661 mg per 100 g. Shelf-stable jars are usually pasteurised and contain no live cultures. Vegetables pickled in vinegar are not fermented at all. |
| Soy sauce | No | Fermented in production, but pasteurised and filtered before bottling. The microbes are gone; what remains is salt — see the sugar and salt guide. |
Kombucha deserves a section of its own, because it is the fermented food most likely to confuse shoppers in Singapore. Fermentation naturally produces small amounts of alcohol alongside acids and flavour compounds, and the final alcohol level depends on time, temperature, sugar and storage. Most commercial kombuchas aim to stay at or below 0.5% alcohol by volume, and in Singapore a drink above 0.5% ABV is treated as an alcoholic beverage — which also exempts it from Nutri-Grade labelling. Home-brewed kombucha, especially in Singapore's tropical climate, can ferment quickly and exceed 0.5% without any lab testing to show it.
For Muslim shoppers, the question is whether trace alcohol from fermentation makes the drink impermissible. Many scholars and halal authorities allow beverages with naturally occurring, non-intoxicating trace alcohol when the product is not produced as an intoxicant — Indonesia's MUI fatwa No. 10 of 2018 treats fermented drinks under 0.5% alcohol as halal, provided no haram ingredients are used and the product is medically harmless. Singapore's MUIS has not issued a ruling on kombucha specifically; its fatwa on ethanol addresses ethanol used as a solvent in food flavouring, and is not a kombucha ruling. The practical checks are: look for an ABV statement ("≤0.5% ABV" or "non-alcoholic"), look for a MUIS halal mark if one is present, and check the HalalSG app to verify a brand's current certification. "Hard kombucha" is deliberately alcoholic and is a different product entirely.
Three checks, in order:
Pages where these checks are most useful: yogurt, kefir, kimchi, kombucha, probiotic drinks, tofu and tempeh, flavoured yogurt drinks and fresh milk, plus the fibre guide (for the fermentable fibre that feeds the gut microbiome) and the dairy, lactose and calcium guide.