Read-Across: A Practical New Approach Methodology for Food Safety Assessments
23 Jul 2025
New approach methodologies are transforming chemical safety assessments, providing innovative, non-animal strategies that increase efficiency, lower testing costs, and improve relevance to humans.
In food safety, new approach methodologies (NAMs) are increasingly applied to complement or replace traditional toxicological methods. Among these, read-across stands out as a practical and scientifically sound approach that enables regulators to predict the safety of a substance using data from structurally or biologically similar compounds. Read-across is not a novel concept in food safety assessments. In fact, regulatory bodies have relied on read-across for decades, particularly in the evaluation of flavouring substances, where data from structurally similar compounds have been used to fill safety data gaps. However, as scientific understanding and regulatory expectations have evolved, so too have the challenges associated with applying read-across in a robust and transparent manner. Concerns around uncertainty, structural differences, and the need for mechanistic evidence underscores the necessity to improve how read-across is performed and justified. In response, regulators are developing structured guidance to help companies apply read-across more effectively in regulatory submissions.
Principles and Practical Challenges
Read-across is a data-bridging or data gap–filling strategy where information from one substance (the “source”) is used to predict the properties of another similar substance (the “target”). The fundamental principle of the read-across approach is that substances that share structural similarity can be expected to elicit similar effects. The application of data from one or only a few source substances is often called an “analogue approach”, whereas a “grouping approach” refers to the application of data from a larger subset of related substances, where the known toxicological properties follow a specific trend that can be used to infer the properties of other comparable substances.
The current best practice for selecting suitable analogues for read-across involves searching databases for structurally similar compounds; comparing molecular structures and physical-chemical properties; reviewing metabolism, kinetics, and degradation pathways; and assessing toxicity or bioaccumulation data, considering exposure routes. After thorough evaluation, one or more analogues might be deemed suitable for read-across and can then be applied qualitatively or quantitatively for risk assessment, while other substances might be rejected for being insufficiently representative. Many resources are available to support read-across approaches, including published literature and established tools—the OECD QSAR Toolbox, eChemPortal, CEFIC AMBIT tool, EPA Analog Identification Methodology (AIM) Tool, CompTox Chemical Dashboard, etc.—for accessing in vivo data. For in vitro data, platforms like Tox21 and ToxCast provide valuable information and toxicogenomics evaluations at the molecular level to strengthen read-across justification.
At a conceptual level, the read-across approach appears simple: if two substances are considered “similar”, they are expected to exhibit comparable properties. However, translating this concept into practice is considerably more complex. The primary challenge, especially in regulatory contexts, is adequately demonstrating that the source and target substances are sufficiently similar for the property or endpoint being assessed, as small structural differences can significantly impact behaviour. Thus, the analogue chosen must be relevant to the specific property or endpoint being assessed since an analogue appropriate for one endpoint (e.g., acute toxicity) may not be suitable for another (e.g., carcinogenicity).
Data availability is another major challenge in the read-across approach, as high-quality, reliable data on potential analogues are often limited, outdated, or inconsistent. Regulators expect read-across to be supported not just by structural similarity but also by mechanistic information, such as mode of action or kinetic data; however, these data are often lacking or difficult to compare across substances. Thus, integration with complementary NAMs, such as in vitro assays, in silico models, and physiologically based kinetic models, can strengthen read‑across justifications and boost regulatory confidence but requires significant technical expertise and careful interpretation. Read-across is intended to support, not replace, other toxicological data and is most effective when combined with in silico, in vitro, or in vivo evidence. Finally, proper documentation of data sources, assumptions, uncertainties, and scientific rationale is essential for transparency but can be resource intensive. These challenges underscore the need for clear regulatory frameworks and structured to support the robust use of read-across in food safety assessments.
New Guidance on Read-Across in Food Safety Assessments
The European Food Safety Authority (EFSA) recently held a public consultation on its draft guidance for applying read-across in food and feed safety assessments. The guidance outlines a structured approach on how to use data from similar substances to reduce reliance on animal testing. Key steps include problem formulation, substance characterization, source identification, data gap filling, uncertainty assessment, and reporting. Tools provided include in vitro method lists, an uncertainty template, and case studies, all of which are aimed at promoting transparent and scientifically justified read-across within a weight-of-evidence (WoE) framework.
In recent EFSA Working Group meetings, experts discussed the application of read-across to complex mixtures like food enzymes, noting challenges such as variability in production strains and structures. Nonetheless, the guidance may still be applied in these cases, with strong justification and case-by-case evaluation. Adoption of the EFSA guidance on read-across is expected by the end of 2025.
Other jurisdictions and agencies have lagged behind EFSA in establishing read-across guidance; however, the Joint FAO/WHO Expert Committee on Food Additives increasingly incorporates read-across–like approaches within its WoE evaluations for food additives and contaminants to pragmatically address data gaps, while the United States and Canada apply read-across on a case-by-case basis but have not yet established fully formalized regulatory frameworks for read-across in food safety comparable to those in the European Union.
With growing global regulatory openness and improved data integration, read-across is poised to play a central role in the future of food and chemical safety assessments. This shift is further underscored by recent policy developments - most notably the July 10th, 2025, announcement by the U.S. National Institutes of Health that it will no longer issue Notices of Funding Opportunities (NOFOs) exclusively supporting animal models. This decision reinforces the global momentum to reduce reliance on animal testing and accelerate the adoption of NAMs, such as read-across.
Intertek Is Here to Help
As regulatory expectations evolve, EFSA’s upcoming guidance on read-across offers a valuable roadmap for the food and feed industry to navigate the complex process. At Intertek, we help clients apply read-across approach effectively to fill data gaps, meet regulatory expectations and streamline submissions. If you are exploring read-across for your food or feed product, our team can work with you to ensure your read-across strategy is well-justified, documented, and regulator-ready.
