An adaptation to a standard information requirement means that instead of providing the specific information required, you provide a justification to use other available information. Your justification must either be based on general rules as explained in Annex XI to REACH, or on specific rules for each information requirement as detailed in Column 2 of Annexes VII to X.
- Follow and apply the general and specific rules for adaptations under REACH.
- Refer to the appropriate rules and state your justification clearly.
- For example, if you are providing a Qualitative or Quantitative Structure-Activity Relationship (QSAR) prediction, refer to and follow the rules of Annex XI, Section 1.3 of REACH. Similarly, if you are providing a read-across approach, refer to and follow the rules of Annex XI, Section 1.5 of REACH. These considerations apply for all QSAR or read-across predictions including those prepared with the support of the OECD QSAR Toolbox.
- Information provided using an adaptation must be as reliable as if it was generated using the test required to fulfil the information requirement.
- Make sure that the results of your adaptation are adequate for classification and labelling and for risk assessment. If they are not, ECHA will reject your adaptation.
- Some adaptations can never be accepted. For example, ECHA is currently not aware of any in vitro methods or QSAR models that reliably predict higher-tier endpoints, including repeated-dose toxicity, carcinogenicity, developmental or reproductive toxicity studies.
- ECHA can only assess information provided in your registration dossier. This means that for each source of information you need an endpoint study record containing a study summary or a robust study summary. This also applies to calculated or predicted values.
- If ECHA does not accept your adaptation, you will receive a decision. The decision explains why the adaptation was rejected and requests a standard test to be conducted and submitted by a given deadline.
- Consider if you can improve the rejected adaptation or provide a new valid adaptation addressing the deficiencies listed in ECHA’s decision. If you cannot do this by the given deadline, you must conduct the standard test as required in the decision.
- You can only rely on data from computer models if you can ensure that the model is scientifically valid, your substance falls within the applicability domain of the model and the prediction is adequate for the regulatory endpoint in question. You need to provide the related documentation in your dossier for the data to be independently assessed. If these conditions are not met, ECHA will reject your adaptation.
- You need to show that your registered substance (the target substance) is likely to have similar (eco)toxicological properties to the substances from which you want to read across (the source substances). Then you must predict properties of the target substance from study results with the source substances.
Establish structural similarity:
- To apply the read-across approach, you must demonstrate that the target and source substances are structurally similar. Report the composition of your substances and the available test materials.
- Provide detailed substance identity and analytical information for both target and source substances.
- To establish structural similarity for multi-constituent substances and those of unknown or variable composition, complex reaction products or of biological materials (UVCBs), you also need to explain the differences and similarities of the constituents between the target and source substances. This means the identity and the concentration of these constituents, as well as the variability in the concentration of these constituents. The outcome needs to be compared between the target and source substances.
- This includes details on the composition of actual test materials used in studies with source substances. The impact of these differences on the prediction of hazardous properties needs to be explained as well.
- If it is technically impossible or impractical to identify and quantify all constituents of UVCB substances, you need to consider other techniques to estimate the quantitative and qualitative comparison of the compositions between the target and the source substances. Your read-across adaptation is more likely to be accepted if you can demonstrate that you have made efforts to achieve this.
Define the group or category:
- Identify the group of substances clearly – define inclusion and exclusion criteria if you set up a category of substances.
Provide a hypothesis for the prediction of properties:
- A read-across adaptation can only be accepted if you provide a credible read-across hypothesis with a proper justification and reliable data for each endpoint.
- If your hypothesis is based on similarity through (bio)transformation, you need to provide data e.g. on toxicokinetics or metabolism.
- If your hypothesis is based on structural similarity which leads to similar properties, you need to have reliable and relevant data on the lower-tier endpoints for both source and target substances to confirm your hypothesis and possibility to predict. This could be done by bridging data, for example, from Annex VII or VIII information requirements.
Predict the properties of the target substance:
- Describe structural similarities and differences between the target and source substances. Explain how structural differences may (or may not) impact the predicted properties of the target substance. Support claims with evidence.
- Provide a data matrix with all available physico-chemical and (eco)toxicological information to enhance the comparison of properties and the prediction of unknown ones.
- Be realistic in your analysis of your data gaps. The studies that you use as source data for the read-across must be reliable and adequate for the information requirement.
- Support your hypothesis with information that allows the target substance to be directly compared with the source substance. Cases in which there is no experimental hazard information for the target substance are usually rejected.
- You need to provide comparable information for each endpoint: for instance, data on repeated dose toxicity does not necessarily support toxicity to reproduction or development. Read-across is usually endpoint-specific.
- You need to explain why the properties of the target substances may be predicted from the other substances in the group. Explain the trends you use to support your prediction – including all inconsistencies and their impact on your prediction.
- Provide robust study summaries for each source study used to support the read-across.
- Use ECHA’s Read-Across Assessment Framework (RAAF) to cross-check that your read-across adaptation is robust and complete.
- You can only rely on data from structurally similar/analogue substances if you have lawful access to the study reports and other relevant data generated with those substances.
- Apply the prediction of hazards to the target substance: classify and implement risk management measures based on the prediction.
- You need documented justification on why the sources of information provide a conclusion on the information requirement under consideration:
- In principle, any information can be part of a weight of evidence. The weight (contribution) of an individual source of information depends on its relevance to the endpoint in question and on its reliability.
- Relevance is the extent to which the data and tests are appropriate, i.e. if the information is fully reliable, what contribution does it make to the overall conclusion on the information requirement being considered.
- Reliability is the extent to which the information is correct, i.e. the inherent quality of the information. It is closely linked to the test method used to generate the data.
- Read-across approaches can be used as part of a weight-of-evidence adaptation. You must demonstrate a reliable prediction from the source substance to show that the information is relevant. For this, a read-across justification is required and it must follow the principles of the RAAF.
- (Q)SAR information can be used as part of a weight-of-evidence adaptation if it is adequately documented and the substance falls within the applicability of the used model.
- A weight-of-evidence adaptation must consist of at least two relevant and reliable sources of information.
- All weight-of-evidence adaptations are assessed and benchmarked against the information that would normally be obtained from a study performed to meet this information requirement.
- Completeness is the extent to which the available sources of information cover the data that would be obtained from the study that is normally performed for this information requirement, i.e. a comparison between key parameters covered by the sources of information and the key parameters covered by the benchmark test guideline.
- The weight-of-evidence justification must address the relevance and reliability of each individual information source as well as completeness of the information.
- ECHA can only assess information provided in your registration dossier. This means that for each source of information, you need an endpoint study record containing a study summary or robust study summary. This also applies to calculated or predicted values.
- Testing in accordance with Sections 8.6 and 8.7 of Annex VIII and in accordance with Annexes IX and X may be omitted, based on the exposure scenario(s) developed in the chemical safety report. To support an exposure-based adaptation, you have to provide exposure scenarios in the CSR. This applies even if the exposure scenarios are not required according to the current classification of the substance.
- Adequate justification and documentation must be provided. The justification has to be based on a thorough and rigorous exposure assessment in accordance with Section 5 of Annex I.
- Annex XI 3.2.(a) requires a quantitative exposure assessment. For the adaptation to be acceptable, available test data must enable a DNEL to be derived for the substance concerned that is appropriate for risk assessment. The DNEL derivation should take full account of the increased uncertainty resulting from the omission of the information requirement (i.e. applying additional uncertainty factors). It is not possible to use the DNEL derived from a 28-day sub-chronic toxicity study to waive the 90-day sub-chronic toxicity study. It is also not possible to use the DNEL from a reproductive screening study to waive the extended one generation reproductive toxicity study or the pre-natal developmental toxicity studies.