The EU has several laws governing chemical safety, but the central pillars consist of the Registration, Evaluation, Authorisation and Restriction of Chemicals (REACH) regulation, and the Classification, Labelling and Packaging (CLP) regulation.  

In her ATLA paper, Legal incoherence in EU chemical safety law: A proportionality-based analysis of REACH and CLP, Katerina Stoykova neatly summarised the aims and scope of the two laws:  

What is REACH?

“The REACH regulation aims to ensure a high level of protection for human health and the environment from chemicals, while promoting non-animal approaches to reduce reliance on animal testing in chemical safety assessment and requiring that testing on vertebrate animals be undertaken only as a ‘last resort’. REACH governs nearly all chemical substances that are manufactured, imported or used within the EU — whether on their own, in mixtures or in other items…  

Under REACH, companies manufacturing or importing one tonne or more of a substance per year must collect data on its properties and uses, and evaluate and manage the associated risks. To this end, REACH sets out standard information requirements, which remain largely based on traditional animal tests for assessing specific human health and environmental effects.” 

What is the CLP regulation?

“Its purpose is to ensure that chemical hazards are classified consistently and communicated clearly, and it applies to all chemical substances and mixtures placed on the EU market. For a substance or mixture to be classified, its hazard information must be assessed against the classification criteria laid down in the CLP regulation.” 

What is the purpose of REACH and CLP?

Both REACH and CLP aim to keep people and the environment safe. While REACH focuses on testing and gathering data on chemicals to evaluate and restrict their uses if necessary, CLP focuses on labelling packaging to inform people of the hazards posed by chemicals. 

The EU has consistently signalled its intention to transition away from animal testing in chemical safety assessment. In 2023, in response to the European Citizens’ Initiative Save Cruelty-Free Cosmetics — Commit to a Europe Without Animal Testing, the European Commission pledged to develop, and recently published, a roadmap for phasing out animal testing in safety assessment.  

As scientific and technical knowledge advance, new non-animal technologies are being developed and the foundations of the traditional animal-based paradigm are continuously being challenged. Thus, the concept of ‘best available science’ has to adapt to incorporate scientifically appropriate non-animal methods and more flexible approaches to risk assessment. 

The problem

Since announcing the Roadmap, the EU Commission’s consultations with the public and other stakeholders have indicated broad support for:  

• Updating the regulatory context, such as streamlining the development and validation of non-animal methods for complex endpoints (e.g. neurotoxicity, carcinotoxicity, reproductive toxicity). 

• Reforming validation systems to accelerate the approval of both new and existing non-animal approaches, i.e. non-animal methods have to undergo years-long validation processes before they are accepted by regulators. 

• Strengthening collaboration and information-sharing.  

• Facilitating acceptance of non-animal methods by regulators.  

• Promoting harmonised implementation across different sectors. 

However, while the EU Commission has acknowledged that legislative revisions will become necessary once a new framework for integrating non-animal methods has been developed, little attention has been given to the current legislative design as a barrier to moving away from animal testing.  

While, in theory, both REACH and CLP regulations support non-animal approaches, their application is biased toward animal-based studies. This bias persists despite mounting scientific concerns over the limited predictive reliability of animal data, the inefficiency of animal testing in assessing chemical safety, and the increasing availability of scientifically robust non-animal methods.  

Restrictive legal drafting is a significant barrier to phasing out animal testing, even where scientific and institutional obstacles have been addressed. As testing obligations and data requirements are determined by chemical safety legislation (i.e. REACH and CLP), the way in which these regulations are written and applied can either help or hinder the adoption of non-animal methods.  

In addition, from a legal standpoint, it is essential that these regulations comply with higher-ranking EU law. 

The point made in the paper

In the paper, Stoykova highlights a major legal flaw within REACH and CLP. Higher-ranking EU law relies on the “principle of proportionality,” a constitutional rule requiring legislation to strike a fair balance between competing interests, ensuring no single priority gets absolute precedence.  

Because primary Union law protects animals as sentient beings, chemical safety must be balanced against animal welfare. However, the messy, inconsistent wording inside REACH and CLP disrupts this balance. By defaulting to animal testing and data mandates, the application of these specific regulations results in a disproportionate encroachment on animal welfare, violating the overarching EU principle of proportionality.

Scientific Limitations and Regulatory Failure

This legal imbalance forces reliance on flawed methodology. Despite being conducted under strict, standardised conditions, animal tests frequently yield inconsistent results when repeated and are notoriously poor at mimicking human biology and chemical exposure. 

This framework is also wildly inefficient: 

• The backlog: Testing is so slow and resource-heavy that out of roughly 100,000 chemicals in circulation, only 0.5% have undergone a comprehensive safety assessment. 

• The data gap: Nearly 70% of substances currently on the EU market lack any safety data whatsoever. 

• Meanwhile, non-animal methods (NAMs) are rapidly advancing, oftentimes offering a much more accurate reproduction of human responses. 

The Institutional Habit Loop

The standard defence is that non-animal methods are not yet advanced enough to replace complex safety checks, such as reproductive toxicity. Stoykova argues the actual barrier is institutional inertia. Because regulators, policymakers and academics have operated within an animal-based framework for decades, they reflexively treat animal data as the “gold standard.”  

This systemic bias manifests in how data is handled: 

• When scientists submit legally accepted, scientifically valid non-animal data to regulators like the European Chemicals Agency (ECHA), it is often viewed with skepticism. 

• Non-animal data is routinely accepted only when it can be directly benchmarked against existing animal studies. 

This cautious and risk-averse regulatory environment actively discourages the submission of modern data and entrenches an animal-centric mindset. 

An important point raised by the author “…irrespective of whether NAMs [non-animal methods] are ready to replace animal testing entirely, animal-centred legal requirements at the statutory level are already highly questionable. They entrench a paradigm whose scientific status is increasingly contested and undergoing change, and they constrain the flexibility needed to apply the best available science in pace with continuous methodological and technological advances.” 

What the implications are

The REACH regulation permits animal testing only as a last resort, and regulators are legally obliged to accept scientifically valid non-animal data, a position supported by the Court of Justice of the European Union. The current reluctance to accept non-animal data seems to also be driven by familiarity with a long-standing system and ‘path dependence’.  

When animal tests are performed for regulatory purposes, the aim is to protect human health and the environment. The job of regulators will always involve caution and conservatism. They are bound by the ‘precautionary’ principle to ensure a high level of human protection, understandably seeking to avoid error and uncertainty. As such, familiar forms of data (including from animal tests) will be more trusted. 

However, this situation could be improved with targeted training and better resourcing, to gain knowledge and confidence in the evidence provided by new non-animal methods. 

Legislative design as a major barrier

The thought put forward by Stoykova in her ATLA paper is that the legal design of the relevant regulations constitutes a major barrier to shifting away from animal testing. She states “Alongside cultural and institutional barriers, such as regulatory conservatism and a rigid validation system, legal obstacles embedded in the legislative design of REACH and CLP also impede a paradigm shift away from animal testing.” 

Furthermore, these legal requirements influence the types of non-animal methods that are prioritised for development or validation, further compounding the bias toward animal-centric data.  

To make her case, she applies the principle of “proportionality” to defend that the animal test-centered implementation of the regulations is not suitable, or necessary, to ensure adequate protection for humans, and that it also disproportionately impacts animal welfare in relation to potential benefits for humans.

Katerina hopes that “By providing this legal analysis, the paper lays the groundwork for increased discussion and research on how to future-proof legal acts, and design them in ways that allow for greater flexibility in regulatory decision-making — thereby better reflecting evolving scientific progress and socio-political values — without compromising legal certainty.” 

Thus, the take-home message is that, unless the wording of the regulations is addressed, the EU Roadmap toward the phasing out of animal testing in chemical safety assessment might have limited practical impact.