Minimisation of animal testing

3. General rules for adaptation given in UK REACH Annex 11

Annexes 7 to 10 of UK REACH establish the standard information requirements at the different registration tonnage bands. Article 13(3) states that where tests on substances are required to generate information on intrinsic properties of substances, they shall be conducted in accordance with the test methods laid down in the UK Test Methods Regulation or in accordance with other international test methods recognised by the Agency as being appropriate.

Additionally, ecotoxicological and toxicological tests and analyses must be performed in compliance with the principles of good laboratory practice (GLP) or other equivalent international standards (Article 13(4)). Therefore, new tests conducted for the purposes of UK REACH are expected to meet these requirements.

However, Annex 11 provides general rules that registrants can use to adapt the standard tests set out in Annexes 7 to 10.

Registrants must provide a well-documented and valid justification for each adaptation that they apply. Adaptations must be specific to the information requirement being addressed.

Testing does not appear scientifically necessary

There are several circumstances under which it might not be scientifically necessary to conduct the tests specified in Annexes 7 to 10.

These adaptations can be applied individually or in a weight-of-evidence approach instead of conducting new animal tests.

Alternative approaches used in accordance with Annex 11 should:

  • be adequate for classification and labelling and risk assessment
  • provide equivalent information, that is, adequate and reliable coverage of the key parameters, to the methods referred to in Article 13(3)
  • be supported with adequate and reliable documentation

Use of existing data

In the context of Annex 11, ‘existing data’ means data that were generated for purposes other than REACH, that is, before EU REACH entered into force. Data that were generated for EU REACH registration should meet the requirements of Article 13(3) and 13(4).

Sources of existing data on a substance could include ‘old’ studies, historical human data, published literature and databases. To assess if existing data are sufficient and suitable to meet the information requirement, the relevance, reliability and adequacy of the data are considered.

  • Relevance: the extent to which data and tests are appropriate for a particular hazard identification or risk characterisation. Factors to be considered are the relevance of the endpoint, test material and the test methods and conditions.
  • Reliability: the inherent quality of a test report or publication and how the experimental procedure and results are described so that the clarity and plausibility of the findings can be assessed. Reliability is therefore measured by the quality of the study, the method used, the reporting of the results and the conclusion. In registration dossiers, registrants assign a Klimisch score to communicate the reliability of each study.
  • Adequacy: defines the usefulness of data for hazard and/or risk assessment purposes. Where there is more than one study for an endpoint, the greatest weight is attached to the studies that are the most relevant and reliable.

To judge quality, it is useful to look at a relevant international standard, such as the UK Test Methods Regulation  or OECD test guidelines (oecd.org) and consider how test designs compare and if reliability criteria are achieved. ECHA has published a standard operating procedure (europa.eu) and guidance on the reporting of robust study summaries (europa.eu).

It might be possible to use data or studies that are judged to be inadequate by themselves to meet a standard information requirement, for example because they measured fewer or different parameters from the current test guidelines, within a weight-of-evidence approach.

Whether used as a stand-alone study or in a weight-of-evidence approach, as much documentation as possible of the existing data should be provided, including a detailed description of the method, test material and deviations from current test guidelines.

Weight-of-evidence assessment

This is a common scientific practice that is applied when a conclusion on the presence/absence of a hazardous property cannot be made on information from a single source alone. For example, there could be information from several studies, none of which were conducted to recognised international standards. However, when taken together, these could be used to meet the UK REACH information requirements and be sufficient for risk assessment.

Information gathered for weight-of-evidence assessment should be evaluated for its completeness and quality for the purpose of UK REACH. The evaluation of the quality of data includes assessment of adequacy of the information for classification and labelling and/or risk assessment, and the assessment of relevance and reliability.

A weight-of-evidence approach is a very efficient way of maximising the potential of older studies and non-standard approaches and avoiding new animal testing for a particular hazardous property. The OECD has published guiding principles and key elements for undertaking a weight-of-evidence assessment.

Developments in weight-of-evidence assessments include the use of Adverse Outcome Pathways (AOPs) and Integrated Approaches to Testing and Assessment (IATAs). AOPs provide mechanistic frameworks to allow for the combination of information from key mechanistic events to assess end-points without the induction of apical adverse outcomes in animals. IATAs combine multiple sources of information to conclude on the toxicity of chemicals through a weight-of-evidence analysis. The OECD is building a bank of case studies (oecd.org) to illustrate regulatory applications of IATAs.

OECD defined approaches also combine multiple sources of information, but provide fixed data interpretation protocols; therefore, they do not rely on expert judgment or a weight-of-evidence assessment. Defined approaches link to hazard classification categories for, for example, skin sensitisation or eye damage and eye irritation. Where new data or assessment are needed, defined approaches should be used to meet an information requirement where an OECD guideline exists (for example, guidelines on defined approaches for skin sensitisation and eye damage and irritation).

It is important to document how the evidence-based approach was used in a reliable, robust and transparent manner. Registrants should provide a rationale to explain how the weight-of-evidence approach adequately meets the information requirement.

To support this, in registration dossiers it is recommended that registrants:

  • prepare an end-point summary, compiled from the endpoint study records, in which they summarise the findings for the endpoint and provide a rationale for the conclusion
  • provide sufficient data for each piece of evidence that allows evaluation of the overall evidence and that demonstrates that the combined information allows for an appropriate conclusion on the substance’s properties
  • document and report the scientific considerations of the evidence and overall conclusions;
  • flag all endpoint study records that form part of the weight-of-evidence approach as such in the ‘adequacy of study’ field
  • provide robust study summaries for each study that forms part of the weight of evidence

(Quantitative) structure-activity relationships

(Q)SARs are computer-based models (termed in silico methods) that are designed to predict the physicochemical properties, the human-health effects or the environmental effects and fate of a substance from knowledge of its chemical structure. Structure-activity relationship (SAR) models (like structural alerts, expert knowledge-based systems) rely on structural features of substances to make qualitative predictions about the presence or absence of a property.

Quantitative structure-activity relationships (QSAR) models (such as statistical tools) quantitatively relate characteristics of a chemical to a measure of a particular activity.

Many models are available, either freely or via commercial software. A useful repository of acceptable (Q)SARs for use by potential registrants is the OECD QSAR Toolbox (oecd.org), which has been jointly developed by the OECD and ECHA. In the context of (Q)SARs, registrants can use the Toolbox to obtain relevant information on the intrinsic hazards of their substances, to generate chemical categories and to develop (Q)SAR models.

In principle, valid (Q)SARs could be used to fulfil one or more of the UK REACH information requirements or help build chemical categories without the need to use experimental tests for all substances in the category.

Annex 11 states that the results of (Q)SARs may be used instead of testing when all the following conditions are met:

  • the substance falls within the model’s applicability domain
  • the scientific validity of the model used for predictions has been established
  • the predictions are adequate for the purposes of classification and labelling and risk assessment
  • the information is well documented

(Q)SARs can be used to reliably predict some simple properties and end-points if the substances fall within the model’s applicability domain. Although (Q)SARs are generally not suitable as a stand-alone approach to meet (eco)toxicological information requirements, they could be combined with other information in a weight-of-evidence approach or in an integrated testing strategy. The OECD defined approaches for skin sensitisation introduced the first regulatory-accepted guideline to incorporate in silico predictions (OECD Guideline 497). This guideline outlines how in silico predictions can be used in combination with in chemico and in vitro tests to predict skin sensitisation potency in place of animal-based methods.

The acceptability for regulatory purposes of (Q)SARs depends upon consideration of:

  • the model(s) used and
  • the predictions generated

OECD member countries have agreed a set of principles for the validation, for regulatory purposes, of (Q)SAR models outlined in this guidance (oecd.org).

These principles relate to:

  • the clarity of the end-point to be predicted
  • transparency in the model algorithm
  • the definition of an applicability domain
  • appropriate measures of goodness-of-fit
  • robustness and predictivity
  • a mechanistic interpretation, if possible

The OECD (Q)SAR Assessment Framework (oecd.org) establishes 4 principles for the regulatory assessment of (Q)SAR predictions and results from single and multiple models:

  • correct input(s) to the model(s)
  • substance within the applicability domain
  • reliability of the predictions
  • the outcome’s fitness for the regulatory purpose

The (Q)SAR assessment framework provides workflows for compiling and assessing information on the model(s) used, predictions generated, results from single or multiple predictions and conclusions on the property for a given regulatory purpose.

The documentation submitted in support of a (Q)SAR should reliably demonstrate that the (Q)SAR model and its predictions adhere to these 2 sets of principles. The OECD-recommended documentation comprises:

  • (Q)SAR model reporting format (QMRF) to document the model (usually provided by the model developer)
  • (Q)SAR prediction reporting format (QPRF) to document the prediction (prediction-specific, provided by the (Q)SAR user)

The latest versions of the QMRF and QPRF are included in the (Q)SAR assessment framework.

Additional guidance (europa.eu) on the use and reporting of (Q)SARs is also available.

Grouping and read-across

The grouping of chemicals is an approach whereby two or more chemicals that are likely to be similar or follow a regular pattern as a result of similarity are considered at the same time. In terms of meeting the UK REACH information requirements, grouping is done with the aim of filling data gaps for some or all of the chemicals in a group without performing additional tests on each individual substance. This approach can therefore reduce or eliminate the need for new animal tests.

Data-gaps (physico-chemical, toxicological, ecotoxicological or environmental fate) are filled through the use of read-across. Read-across is a technique for applying measured endpoint data from one or more ‘source’ substances to predict effects for one or more ‘target’ substances that don’t have measured data for the same endpoint. Read-across can be qualitative, in which the presence (or absence) of a particular hazard is predicted. It can also be quantitative, in which the magnitude of a particular property is predicted; this can be, for example, a benchmark dose or no-observed adverse effect level (concentration) (NOAEL(C)), which is a dose or concentration at which no adverse changes are observed.

Grouping can include the identification of chemical analogues, or the formation of a chemical category. The analogue approach focuses on filling data gaps for a specific target substance, typically through the read-across of data from a single source chemical. The category approach differs in that the properties of individual chemicals within the category are assessed by evaluation of all the available information for all the category members for a given endpoint, rather than on the basis of measured data for any single chemical. A category approach can help to identify trends in the physico-chemical, (eco)toxicological or environmental fate properties of its members.

Registrants relying on grouping to meet information requirements must provide a scientific rationale for the applied analogue or category approach. Under Annex 11 of UK REACH, the establishment of structural similarity is a pre-requisite for the use of grouping and read-across. Grouping might also be based on one or more other similarities, including physico-chemical properties, chemical reactivity profile, bioactivity, conventional toxicological profile and toxicokinetics, including metabolism. Registrants must provide supporting data to justify the similarity of their source substance(s) and their suitability for read-across for predicting the relevant physico-chemical properties, human-health effects and/or environmental effects or environmental fate of the substance to be registered (target substance). The OECD QSAR toolbox can be used to generate categories and identify additional toxicological data for read-across.

Changes to Section 1.5 of Annex XI of EU REACH have provided clarity on what constitutes ‘adequate and reliable documentation’ in the context of grouping and read-across. These changes have not (yet) been reflected in Annex 11 of UK REACH. However, they represent good practice and therefore UK REACH registrants are recommended to follow them.

These state that adequate and reliable documentation shall include:

  • a robust study summary for each source study used in the adaptation
  • an explanation why the properties of the registered substance may be predicted from other substances in the group
  • supporting information to scientifically justify such explanation for prediction of properties

Registrants should be in legitimate possession of or have permission to refer to the relevant full study reports for the source substances, for example through a letter of access.

Sources of general guidance on the grouping of chemicals to help facilitate the use of read-across for regulatory purposes include the OECD guidance on the grouping of chemicals(oecd.org) and the Read-across Assessment Framework (RAAF)(europa.eu).

There is specific guidance such as the ECHA read-across assessment framework (RAAF) consideration on multi-constituent substance and UVCBs and advice on using read-across for UVCB substances for grouping UVCBs (substances of unknown or variable composition), complex reaction products and biological materials.

There is also guidance for grouping nanomaterials such as the Guidance on Grouping of Chemicals, the Gracious project and ECHA’s Recommendations for nanomaterials applicable to the Guidance on QSARs and Grouping of Chemicals. The Guidance on Grouping of Chemicals can also be applied when grouping metals and inorganic metal compounds. As grouping approaches are continuously under development, the most recent version of guidance should be used.

In vitro methods

An in vitro test is performed outside a living animal.

In vitro methods and data can be assigned to one of three groups, which determines how they can be used for the registration of substances under UK REACH.

Validated in vitro methods have been scientifically agreed according to internationally-agreed validation principles and adopted into the UK Test Methods Regulation and OECD test guidelines. They can be used as stand-alone methods to fully replace tests in animals, or combined with other information in an integrated testing strategy, as dictated by their respective OECD guidelines, IATAs and defined approaches.

Some of the UK REACH standard information requirements can be met by validated in vitro tests that have OECD test guidelines, see Table 1 under further considerations for meeting the standard information requirements.

Pre-validated in vitro methods are those that meet internationally-agreed pre-validation criteria. For example, the European Centre for the Validation of Alternative Methods (ECVAM) has set criteria for the acceptance of methods into the pre-validation process. Pre-validated methods can be used as an adaptation when the criteria in Section 1.4 of Annex 11 are met. Information to demonstrate that the method meets internationally agreed test development criteria for entry into the pre-validation phase must be included in the registration dossier.

The results from pre-validated in vitro methods can be used under UK REACH when they indicate a dangerous property; in other words, when they give a positive result.

However, if the results of such methods do not indicate a dangerous property (they give a negative result), Annex 11 states that the relevant test stated in the information requirement shall be conducted.

Non-pre-validated in vitro methods (and pre-validated methods) can provide additional information for evaluation and interpretation in combination with other sources of data, including from animal tests or other in vitro methods. They might inform on the mode-of-action of a substance and/or support adaptations allowed in Annex 11, such as the use of existing data, grouping and read-across and/or a weight-of-evidence assessment.

The purpose of such methods should be clearly and scientifically justified, with submission of the criteria for suitability where applicable (for example, where pre-validated methods provide supportive evidence).

Testing is not technically possible

It is possible to adapt some tests if it is not technically possible to conduct the test, or the substance is very hazardous, for example, owing to the generation of an explosive atmosphere. The UK Test Methods Regulation should also be consulted, as some test methods, such as the acute dermal and acute inhalation tests, identify specific circumstances where testing is inappropriate.

Exposure-based waiving

Exposure-based waiving of certain tests might be permitted in cases where it can be shown that exposure is insignificant or absent for the substance concerned. This will depend on the conditions of use and is not considered further here.

View a printable version of the whole guide

Back to top

Is this page useful?