Feed additive authorisation

A feed additive is unusual among regulated substances in that the animal consuming it is not the party being protected. Four distinct subjects of safety have to be addressed separately, and each one is a different study: the target animal, the human eating its meat, milk or eggs, the worker handling the premix, and the environment receiving what passes through the animal. Under Regulation (EC) No 1831/2003 the EU additionally requires that the additive work — efficacy is part of authorisation, not a marketing matter — and that requirement is what makes the dossier a scientific object rather than a filing.

The tolerance study and its margin

Target-animal safety is established by deliberate overdose. The standard design fixes a margin: the additive is fed at a multiple of the maximum intended dose — conventionally around tenfold — to the species and category it is intended for, over a period relevant to that animal’s production cycle, with a control group and endpoints covering performance, haematology, clinical chemistry and mortality. The logic is that on-farm dosing errors and premix inhomogeneity are certainties, not hypotheticals, so a substance whose safe dose sits just above its effective dose is not usable in practice however good its efficacy data.

Consumer safety is a different question, answered by residues and metabolism: what fraction of the additive or its metabolites reaches edible tissue, and does that exposure sit inside a toxicologically derived acceptable intake. This is why a substance can be entirely safe for the animal and still fail. User safety turns on inhalation and dermal exposure to a dusty powder — respiratory sensitisation is a common reason for a restriction on handling. Environmental assessment asks what excretion does to soil and water, and for trace elements such as copper and zinc it has driven the authorised maxima down independently of animal need.

Efficacy is harder than safety

Zootechnical effects are small. A few percent improvement in feed conversion ratio is commercially decisive and statistically awkward, because between-animal variation is large and the pen or the house, not the individual animal, is usually the experimental unit. The EFSA FEEDAP Panel therefore expects several independent trials, in the relevant species and category, with enough replication to detect the claimed effect — and a demonstrated effect in broilers says nothing about laying hens or piglets, so authorisations are granted species by species.

Microbial and enzymatic additives add identity requirements: the production strain is characterised by whole-genome sequencing, screened for antimicrobial resistance determinants that could transfer, and checked for toxigenic potential. The strain, not the genus, is the authorised entity.

Two regimes, not one process

An EU authorisation is a legal act naming a holder, a species, a minimum and maximum dose and conditions of use, granted for a defined period and requiring renewal. US GRAS is a different construction: the standard is general recognition of safety among qualified experts, based on publicly available evidence, and it can be self-affirmed by an expert panel without any submission. Notification to FDA is voluntary and closes with a letter stating no questions. The practical consequence is that a US GRAS conclusion attaches to a described use and does not expire, and that neither route’s conclusion transfers to the other jurisdiction.