Every supplement label carries a claim. Few of them were checked by anyone. The European answer to “checked by anyone” has a name, a procedure and a public paper trail, and reading one of those papers is a learnable skill.
What the register actually is
Regulation (EC) No 1924/2006 is the frame. Flynn’s account of it, in the Proceedings of the Nutrition Society, is plain: a health claim may be authorised for use in the community only after a scientific assessment of the highest possible standard has been carried out by the European Food Safety Authority. The assessment sits with the EFSA Scientific Panel on Dietetic Products, Nutrition and Allergies, known as NDA, and the Panel has published extensive guidance telling applicants how to build an application that survives it.
Flynn also records what the Panel has actually accepted over the years: many well-established functions of nutrients, beneficial effects of foods and food constituents on a range of body functions, roles of nutrients in growth and development of children, and effects of nutrients on risk factors for disease. That is not a small list. The bar is high, not impossible.
What the bar is made of
Navas-Carretero and Martinez, in the International Journal of Food Science and Nutrition, spell out the design features that decide whether a cause-and-effect relationship can be established: the outcome studied, the length of the trial, sample size and type, and the transparency of reporting. Their guidance is built on the CONSORT statement together with the EFSA NDA Panel’s own guidelines.
Read that list as the answer to a marketing question. If a page says a product “works”, what was the outcome, how long was the trial, how many people were in it, and can you see the numbers? No page on the internet survives all four questions.
Three opinions, three reasons for the same outcome
The two EFSA opinions below are worth reading together, because they failed for completely different reasons and neither failure was about the ingredient being useless.
Spermidine and the hair cycle (EFSA Journal, 2011). The applicant proposed prolonging the growing phase of the hair cycle, with a target population of healthy people affected by chronic telogen effluvium. Asked why that population would be appropriate, the applicant argued that telogen effluvium has multifactorial triggers, occurs in both sexes at any age, and could be considered to affect the general population. The Panel disagreed, and not on the science: it considered the claimed effect to be related to the treatment of a pathological condition leading to shortening of the anagen phase, and therefore outside the criteria of the Regulation.
That is a classification decision. The same molecule, the same study, a different claimed effect, and the answer flips.
‘Nutrimune’ and immune defence (EFSA Journal, 2017). Here the applicant had a human intervention study from which conclusions could be drawn, plus an animal study that could support defence in the gastrointestinal tract. The Panel accepted that immune defence against pathogens in the gastrointestinal and upper respiratory tracts is a beneficial physiological effect. It still concluded that the evidence was insufficient to establish a cause and effect relationship, listing three specific problems: inconsistencies in the reporting of the process and criteria used to diagnose upper respiratory tract infection in the human study, the results of that study had not been replicated, and no evidence was provided for a plausible mechanism by which the product could act in vivo in humans.
One human study, no replication, a fuzzy endpoint, no mechanism. Any one of those would have been uncomfortable. Together they are a refusal.
The part almost nobody implements
Allaert and colleagues, in 2017, looked at the same system from outside and asked a fair question: the Regulation’s own first recital says the process is meant to let consumers “choose in full knowledge of the facts”, and after ten years the number of accepted claims is marginal. They argue that grading the weight of evidence inside the EFSA assessment — rather than a binary pass or fail — would let the public benefit from the very high standard the Panel already applies. Their point is that standards of proof range from formal proof in high-powered double-blind placebo-controlled studies down to consensus views of experts, and that grading the evidence, rather than grading the expertise, is what every recommendation and consensus meeting already does.
Whether the EU adopted that suggestion is a regulatory question, not a scientific one. It is worth knowing that the argument exists, because it tells you what an “authorised” claim does and does not imply: it means the evidence cleared a high bar, not that it was the best evidence available.
How to use this when you read a label
Three questions, in order. First: is this a nutrient function claim or a disease claim? The spermidine opinion shows where that distinction bites. Second: what was the human endpoint, and was it reported consistently? The Nutrimune opinion turned on a diagnostic procedure, which is the least glamorous possible failure point and the most common one. Third: has anybody replicated it? An unreplicated single study is not a weak version of a claim, it is the only version of the claim.
What this evidence does not show
It does not show that the two rejected products do nothing. A rejected claim is a statement about a submitted application, not about the biology of the ingredient. It does not show that every claim outside the EU register is false: the FDA and EFSA both set scientific bases for claims, but the register is a European mechanism and a product can carry a legally authorised claim in the United States without having been through it. And it does not show that consistently-substantiated claims are worth taking for any particular person — the Panel evaluates whether a cause-and-effect relationship is established, not whether the effect is large enough to matter to you.
This article describes a regulatory procedure and published scientific opinions. It is not medical advice and it is not a recommendation for any individual.
