№ 156 · biology

What the gut's microbes actually do, and what the ads claim

Your gut holds a large community of bacteria that competes with invaders for nutrients, and it trains the immune system. That much is well described. What a yoghurt pot can promise about it is a separate question, and the answer is narrower than the label suggests.

What the microbes do

Shealy, Yoo and Byndloss describe the gut microbiota as a microbial organ. It educates the immune system, helps extract calories from food, and blocks invading bacteria from settling in — a defence called colonization resistance. The mechanism is partly chemical. Microbes ferment dietary fibre into short-chain fatty acids (SCFAs), chiefly butyrate, acetate and propionate. Butyrate feeds the cells lining the colon, which then consume more oxygen and keep that stretch of gut anaerobic. That suits the resident oxygen-shy bacteria and disadvantages invaders such as Salmonella, which can respire using oxygen or other electron acceptors that inflammation makes available.

The immune side

Wiertsema and colleagues note that 70–80% of immune cells sit in the gut. SCFAs encourage regulatory T cells that dampen local inflammation, and they support the cells that release antimicrobial molecules. Tryptophan metabolites and bile acid derivatives act on receptors that help keep the gut in a stable state. The authors also caution that in gut–lung research it is hard to tell whether microbial changes cause a disease or follow from it.

Predict first

If a probiotic trial reports a significant rise in a “beneficial” bacterial group, what has it shown about the volunteers' health?

Check your answer

Only the composition shift. De Jesus Cano and García Menéndez argue such measures describe community structure, not function, so they cannot carry a health claim on their own.

Why trials disappoint

De Jesus Cano and García Menéndez, writing in 2026, list recurring design faults. Trials lean on diversity scores and taxonomic shifts as stand-ins for benefit. Claims get written after the data, around whatever reached significance. Symptom scores get too much weight in conditions where placebo responses are strong. Diet, antibiotics and other medications are poorly controlled, and many studies are too small. They give examples: a fermented-food trial reported a 20% rise in a diversity index and changes in inflammatory markers, with no mechanism linking the two; a probiotic trial preserved diversity during antibiotics without measuring diarrhoea. Their fix is to treat the microbiome as a mediator and judge the host outcome.

What the label may say

The Food Safety Authority of Ireland states, as of 2025, that no health claims for probiotics are approved in the EU. Applications went to EFSA and none received a positive opinion. Calling a food “probiotic” is itself treated as a health claim, because it implies benefit. In December 2024 the European Ombudsman accepted the Commission's view that evidence is insufficient to support such claims. A separate authorised claim covers yoghurt: its live cultures can be said to improve digestion of the yoghurt's lactose in people who have trouble digesting lactose, provided each gram holds at least 100 million colony forming units of the two starter organisms named in the rule.

In short

Two questions sit side by side here. What the microbes do is a body of mechanism, partly mapped. Whether a product helps a person needs, de Jesus Cano and García Menéndez argue, trials that judge a host outcome. As of 2025, no health claim for probiotics has been authorised in the EU.

Interactive Drag the slider to set a reported change in a diversity score. The host-outcome panel does not move, because the score describes the microbial community, not the person.

Schematic. Left: the change in a diversity score that a trial might report, from −30% to +30%. The marks at +20% and −25% are the two trial figures de Jesus Cano and García Menéndez cite; other slider positions are illustrative settings, not results. Right: the host outcome, which a diversity score does not measure.

Where this comes from

  1. Colonization resistance: metabolic warfare as a strategy against pathogenic Enterobacteriaceae (Current Opinion in Microbiology 64:82-90) linked only, not reproduced
    Shealy, Yoo, Byndloss · 2021
    pmc.ncbi.nlm.nih.gov/articles/PMC8612973/
  2. The Interplay between the Gut Microbiome and the Immune System in the Context of Infectious Diseases throughout Life and the Role of Nutrition in Optimizing Treatment Strategies (Nutrients 13:886) linked only, not reproduced
    Wiertsema, van Bergenhenegouwen, Garssen, Knippels · 2021
    pmc.ncbi.nlm.nih.gov/articles/PMC8001875/
  3. Why Clinical Trials of Microbiome-Targeted Interventions Often Fail to Support Health Claims: A Commentary on Probiotics and Translational Design (Microorganisms 14:470) linked only, not reproduced
    de Jesus Cano, García Menéndez · 2026
    pmc.ncbi.nlm.nih.gov/articles/PMC12942760/
  4. Probiotic Health Claims (questions and answers) linked only, not reproduced
    Food Safety Authority of Ireland · 2025
    www.fsai.ie/business-advice/nutrition/probiotic-health-claims