Nutrition claims need human evidence: the EFSA-aligned path
Small effect sizes, biological variability and unsupervised real-world consumption make nutrition claims genuinely hard to substantiate. What an EFSA-aligned programme looks like, and where most fail.
Insight · · 3 min
Food, nutrition and FSMP products occupy an awkward position. They are held to real scientific and regulatory expectations for efficacy and tolerability, but their effects are modulated by background diet, lifestyle, microbiome composition and genetics, and the effect sizes involved are usually small. That combination makes claim substantiation genuinely difficult, and the failure rate under the European regime reflects it.
The legal frame is Regulation (EC) No 1924/2006 on nutrition and health claims made on foods. It distinguishes claims based on generally accepted scientific evidence under Article 13(1), claims based on newly developed scientific evidence or proprietary data under Article 13(5), and reduction-of-disease-risk and children's development claims under Article 14. Articles 13(5) and 14 require an application, assessed by the EFSA Panel on Nutrition, Novel Foods and Food Allergens, and authorisation before the claim may be used. Foods for special medical purposes fall under Regulation (EU) 2016/128, with its own compositional and evidentiary expectations.
EFSA's assessment turns on a single question, and it is narrower than most applicants expect: has a cause-and-effect relationship been established between consumption of the food or constituent and the claimed effect, in the population for whom the claim is intended, under the intended conditions of use? Every part of that sentence has been decisive in a negative opinion at some point.
The recurring reasons for failure are well documented in EFSA's published opinions, and they are structural rather than accidental.
The food or constituent is insufficiently characterised. A botanical extract without a defined production process, standardised marker compounds and batch consistency cannot be the subject of a cause-and-effect assessment, because the thing being claimed about is not stably defined.
The claimed effect is not accepted as a beneficial physiological effect. Applicants frequently propose an outcome that is measurable but not established as beneficial, or a biomarker whose relationship to a health outcome has not been demonstrated. Choosing an endpoint EFSA has already accepted in a comparable context is worth more than choosing a novel one that is easier to move.
The pivotal evidence is not human intervention data in the target group. Mechanistic, animal and in vitro work supports plausibility; it does not establish the relationship. Studies in a diseased population rarely support a claim aimed at the general population, and vice versa.
The studies are underpowered for the effect actually present. This is the quiet killer. Nutrition effects are small, inter-individual variability is large, and background diet is uncontrolled. A trial powered on an optimistic assumption produces a null result that then sits in the dossier permanently.
A defensible programme is therefore built backwards from the claim wording, and built early. The claim is drafted before the study is designed, with regulatory, medical and marketing in the same room, because the endpoint has to support the exact wording that will appear on the pack rather than something adjacent to it. The population is defined to match the claim's intended users. Endpoints are selected from validated, clinically relevant measures, with a documented rationale for why the marker matters. Duration reflects the biology rather than the budget. Background intake is measured, because it determines whether an effect can be detected at all. And power calculations use realistic effect sizes drawn from the literature, not from the best result in the most favourable prior study.
The second half of the programme is the part most sponsors omit. A controlled human intervention study establishes the effect under specified conditions; it does not describe what happens when the product is bought in a pharmacy or supermarket and used without supervision, alongside an ordinary diet and ordinary adherence. Real-world data covering actual conditions of use, actual consumption patterns and actual tolerability closes that gap, supports the plausibility of the claim in practice, and serves post-market obligations that arrive regardless. The bridge between the controlled trial and real consumption is where nutrition evidence programmes usually succeed or fail.
The classic mistake is treating nutrition as clinical-lite: a smaller trial, a looser protocol, a softer endpoint, on the assumption that the standard is lower because the product is a food. The standard is not lower. It is differently shaped, and in some respects harder, because the effect is smaller and the confounding is greater. Treating nutrition research with full scientific rigour, adapted to its specific biology and its specific regulatory test, is what separates a claim that survives assessment from one that does not.
Sources. Regulation (EC) No 1924/2006 on nutrition and health claims made on foods, Articles 13 and 14. Regulation (EU) 2016/128 on food for special medical purposes. EFSA Panel on Nutrition, Novel Foods and Food Allergens, general scientific guidance for stakeholders on health claim applications.
Written by the Evidilya scientific team. For interviews, references or a full publication list, use the contact page.
Talk to our team.
Bring us the evidence question you're wrestling with, we'll tell you the shortest defensible path.