
Every few years ECHA publishes a report on how far alternatives to animal testing are actually used in REACH dossiers. This year the numbers say one encouraging thing and one uncomfortable one, and they are worth reading together.
The sixth report, published in June 2026, examined almost 13,000 registration datasets up to July 2025, with particular attention to 2022-2025.
Five points apart: more endpoints are now covered by argument than by new testing. For complex endpoints — repeated-dose toxicity, carcinogenicity — the dependence on the animal model remains.
Put in the dry prose of an agency: adaptations are often insufficiently justified, and regulators then request further data, sometimes animal data.
It means there are companies that chose the non-animal route, built a read-across or a weight of evidence, and were asked for the in vivo study anyway. Not because the alternative method was wrong: because the file supporting it did not hold up.
A read-across is not a shortcut: it is a scientific claim that has to be demonstrated. Saying that substance A behaves like substance B requires showing they are similar in the way that matters for that endpoint, and structural similarity alone is almost never enough. It takes analytical characterisation of both, the impurity profile, and increasingly mechanistic data showing the same behaviour in a biological system.
The same holds for a weight of evidence, which by definition assembles sources that would not suffice alone: if they do not converge, or the reasoning holding them together is missing, what is left is a list.
REACH requires toxicological and ecotoxicological tests to be conducted under good laboratory practice, and that applies to key studies. It does not apply to adaptations, which are not tests: a read-across, a weight of evidence, a QSAR do not produce a study to certify, they produce a justification to defend. The distinction changes where it makes sense to invest, and it is overlooked more often than you would think.
In cosmetics the European ban has been fully in force since 2013: the non-animal route was never a choice. What ECHA now measures for industrial chemicals is the question a safety assessor has always put to your ingredient file, and it is not which method you used: it is what it is built on.
On 1 June 2026 the Commission adopted the Roadmap towards phasing out animal testing, with over thirty recommendations spanning industrial chemicals, pharmaceuticals and food additives. The direction is clear, the methods are there, and what is missing sits in between: the data that has to carry the argument. That is the part we work on — analytical characterisation and mechanistic evidence, before someone else decides on the file.
Technical content and news from our laboratories, sent when there is something worth saying.