Transparency data

Draft Protocol for the Risk-Benefit Assessment of Fish Consumption

Published 6 July 2026

This is a paper for discussion. It does not reflect the views of the Committee and should not be cited.

Introduction

1. This document sets out the draft protocol developed by the European Food Safety Authority (EFSA) for a comprehensive risk–benefit assessment (RBA) of fish consumption. The EFSA draft protocol for the risk-benefit assessment of fish has been prepared in response to a mandate from the European Commission and establishes the scientific framework, assessment questions, methodological approach and evidence sources that will be used to evaluate the health implications of consuming fish across different population groups.

2. The assessment concerns the combined presence in fish of several chemical contaminants, namely polychlorinated dibenzo-p-dioxins and dibenzofurans (PCDD/Fs), dioxin-like polychlorinated biphenyls (DL-PCBs), per- and polyfluoroalkyl substances (PFAS), polybrominated diphenyl ethers (PBDEs), inorganic arsenic, dimethylarsinic acid (DMA), inorganic mercury and methylmercury. At the same time, the assessment will consider the recognised nutritional and health benefits associated with fish consumption.

3. The protocol was developed in accordance with EFSA’s guidance on protocol development for scientific assessments and follows the more recent EFSA guidance on risk-benefit assessment of foods. It represents a planning document describing how the assessment will be conducted rather than the assessment itself. The protocol explains how the European Commission’s terms of reference have been interpreted, how the assessment questions have been formulated and what methods will be used to answer them.

4. EFSA emphasises that the work is not intended to constitute a new risk assessment of the contaminants concerned. Instead, it will build upon existing scientific opinions and toxicological evaluations already undertaken by EFSA and other authoritative bodies. Likewise, the assessment of benefits will rely primarily on previous risk–benefit assessments and evaluations conducted by organisations such as the Norwegian Scientific Committee for Food and Environment (VKM) and the Food and Agriculture Organization/World Health Organization (FAO/WHO).

5. The protocol recognises that fish is both a source of beneficial nutrients and a potential source of exposure to environmental contaminants. The central objective is therefore to evaluate these competing effects together and estimate the overall health consequences associated with different levels and types of fish consumption.

6. The work originates from recommendations made in earlier EFSA opinions, particularly the 2018 opinion on dioxins and dioxin-like PCBs, which highlighted the need for an updated assessment of fish consumption. Since then, additional scientific developments relating to PFAS, PBDEs, arsenic compounds and mercury have further strengthened the need for a comprehensive integrated evaluation.

7. The European Commission requested that the assessment consider fish species and fish categories, including distinctions such as fatty versus lean fish, predatory versus non-predatory fish, and wild-caught versus farmed fish. The purpose is to generate evidence that can support national authorities when developing food-based dietary guidelines and consumer advice.

The draft protocol

8. EFSA clarifies that the target population for the assessment is the general population of the European Union, encompassing all age groups and life stages. Particular attention will be paid to differences between population subgroups where risks and benefits may vary.

9. The assessment is intended to be conducted as a refined Tier 3 risk–benefit assessment. This is the most advanced level of assessment within EFSA’s framework and aims to quantitatively integrate both beneficial and adverse health outcomes into common metrics that allow direct comparison.

10. Where data permit, EFSA intends to estimate probabilities of beneficial and adverse effects at different levels of fish consumption and within different population groups. The assessment may also express outcomes using integrated health measures such as disability-adjusted life years (DALYs).

11. The protocol explains that the assessment will consider evidence relating both to contaminants present in fish and to nutrients for which fish is an important dietary source. In addition to direct evidence linking fish consumption to health outcomes, evidence relating nutrient intake to health outcomes will also be considered where relevant.

12. Several activities are explicitly outside the scope of the work. The assessment will not compare fish with alternative foods, dietary patterns or supplements. It will not estimate the overall contribution of fish to nutrient intakes or contaminant exposures in the diet. Nor will it develop dietary recommendations itself. Instead, it is intended to provide scientific evidence that may subsequently be used by risk managers and national authorities.

13. The scope is limited to finfish and products derived from finfish. Other seafood categories, such as shellfish and crustaceans, are excluded. Fish oils and algal oils are also outside the scope of the mandate.

14. Health effects considered in the assessment include beneficial effects associated with fish consumption and fish-derived nutrients, together with adverse effects linked to the contaminants specified in the mandate. Effects such as food allergy and adverse effects associated with contaminants not listed in the mandate are excluded.

15. To structure the assessment, EFSA translated the mandate into one overarching assessment question and seven specific assessment questions. The overarching question asks what the overall health impacts are when risks and benefits associated with fish consumption are integrated across different fish consumption scenarios and fish types.

16. The first assessment question focuses on identifying the health effects and fish components that are relevant to the assessment. This stage forms the evidence-gathering foundation for all subsequent work. Key details of this question include the following:

• EFSA plans to identify beneficial health effects associated with fish consumption primarily through existing VKM and FAO/WHO assessments. Particular attention will be given to effects for which evidence has previously been judged to be probable. More recent literature published since late 2021 will also be examined where appropriate.
• Human evidence will receive priority. Prospective cohort studies, nested case-control studies, follow-up studies of trials and randomised controlled trials will be considered particularly valuable. Systematic reviews and meta-analyses will also play an important role.
• Adverse effects associated with contaminants will be identified primarily from EFSA contaminant opinions and the US Agency for Toxic Substances and Disease Registry (ATSDR) toxicological profile on mercury and methylmercury.
• EFSA intends to consider not only the critical effects used previously to derive health-based guidance values but also other adverse effects that may be relevant for a broader risk–benefit assessment. This reflects the principle that risk–benefit assessment considers the full spectrum of health outcomes rather than focusing solely on the most sensitive endpoint.
• Nutrients for which fish is an important dietary source will be identified through exposure and intake analyses. Associated beneficial effects will then be drawn from relevant nutritional assessments produced by EFSA’s nutrition experts.
• An additional element of the first assessment question involves examining whether adverse effects linked to contaminants have actually been evaluated in studies that used fish consumption itself as the exposure measure. EFSA notes that previous reviews suggested important gaps in this evidence base.

17. The second assessment question concerns the selection and characterisation of adverse health effects. A prioritisation process will be applied to determine which adverse outcomes should be included in the quantitative assessment. This question will consist of the following points:

• The prioritisation will rely on weight-of-evidence conclusions from previous scientific opinions. Where necessary, targeted literature reviews will be undertaken to identify newer studies that could influence prioritisation decisions or provide data suitable for dose-response modelling.
• EFSA plans to establish dose-response relationships between contaminant exposure and adverse health outcomes. Existing dose-response models from EFSA opinions will be used whenever possible. Where such models are unavailable, new analyses may be performed using available evidence.
• Several modelling approaches may be used, including benchmark dose modelling, regression-based techniques and Bayesian modelling frameworks. Human data will be preferred, although animal studies may be considered where appropriate.
• The protocol also envisages examining dose-response relationships between fish consumption itself and adverse health outcomes where sufficient evidence exists.

18. The third assessment question mirrors the second but focuses on beneficial health effects. Again, a prioritisation exercise will identify the most relevant outcomes for quantitative modelling. The following aspects will be addressed:

• Priority beneficial effects will be selected using evidence from the VKM and FAO/WHO assessments and, where necessary, supplemented by updated literature reviews.
• Dose-response relationships between fish consumption and beneficial health outcomes will be derived from existing meta-analyses or newly identified evidence. Similar analyses may be conducted for nutrients that are important components of fish.

19. The fourth assessment question concerns exposure assessment. This stage aims to determine contaminant occurrence, nutrient composition, fish consumption patterns and resulting exposures and intakes. Additional details regarding this question are as follows:

• Occurrence data for contaminants will be extracted from EFSA databases covering the period 2013–2023. These data will undergo extensive quality control procedures, including checks for duplicates, analytical quality and coding consistency.
• Particular attention will be given to ensuring that occurrence data are compatible with the datasets used in ongoing EFSA assessments, especially the forthcoming opinion on dioxins and related PCBs.
• Nutrient composition data will be obtained from EFSA’s developing European food composition database. This resource contains information from numerous national food composition databases and covers a large number of fish products.
• Where necessary, additional nutrient information may be sourced from publicly available national databases to ensure completeness.
• Consumption information will be obtained from EFSA’s Comprehensive European Food Consumption Database. This database contains individual-level dietary information collected across Member States and includes data covering a wide range of age groups.
• Surveys considered unsuitable for estimating long-term exposure, such as those based on a single reporting day, will not be used for chronic exposure assessment.
• Exposure to contaminants from fish consumption will be estimated by combining occurrence data with consumption data using EFSA’s FoodEx2 food classification system.
• The analysis will produce distributions of exposure for different age groups and population categories. Mean, low and high exposure estimates will be generated for each survey and subgroup.
• Alternative consumption scenarios will be developed and compared with a reference scenario. These scenarios will reflect different patterns of fish consumption and will form the basis for later risk–benefit comparisons.
• Nutrient intakes will be estimated using analogous methods, combining food composition information with fish consumption data.

20. The fifth assessment question concerns the characterisation of risks and benefits under both reference and alternative scenarios. Exposure estimates will be combined with dose-response relationships to estimate probabilities or magnitudes of health effects. Additional details on this question are the following points:

• EFSA may use either bottom-up or top-down approaches depending on data availability. Bottom-up approaches estimate the probability of health outcomes directly from exposure and dose-response relationships. Top-down approaches estimate the proportion of disease burden attributable to a risk factor using relative risks and population disease statistics.
• Risks and benefits may be expressed using common metrics such as disease incidence or mortality, or through broader composite measures.
• The protocol recognises that health outcomes may differ considerably in nature and severity. Consequently, additional methods may be required to transform or scale outcomes so that they can be compared within a common analytical framework.
• Although the assessment is intended to operate at Tier 3 level, Tier 2 methods may also be used to provide supporting information. Such analyses could include comparisons with health-based guidance values or dietary reference values and estimates of the proportion of individuals exceeding them.

21. The sixth assessment question focuses on integrating risks and benefits into an overall health impact estimate. Notably:

• EFSA highlights DALYs as a particularly useful composite metric because they combine information on mortality and morbidity and permit direct comparison across different diseases and risk factors.
• DALY calculations require information on disease incidence, duration and severity, together with disability weights that reflect the relative burden associated with particular health states.
• Where necessary, intermediate health effects may be transformed into disease outcomes that can be expressed using DALYs or other common metrics.
• The integrated assessment will compare disease burdens between reference and alternative consumption scenarios and calculate the overall net balance of risks and benefits.
• Results may be presented for the total population as well as for specific population groups, thereby allowing examination of population variability and potentially vulnerable groups.

22. The seventh assessment question concerns communication of the assessment outcomes. EFSA recognises that risk–benefit findings can be difficult to communicate effectively and that consumer perceptions may influence how advice is interpreted and acted upon. To address this challenge, the following points are addressed:

• The protocol proposes using EFSA’s Concern Assessment framework. This framework considers public perceptions, values, concerns, behavioural factors, social reactions and the influence of institutions and media.
• Evidence will be gathered from existing social science research, systematic reviews, media monitoring and social media analyses. Additional literature published between 2022 and 2026 will be reviewed where necessary.
• Particular emphasis will be placed on identifying individual factors that influence consumer responses, including attitudes, cognitive processes, information-seeking behaviours and trade-offs between perceived risks and benefits.
• Contextual influences such as location, seasonality, social settings and consumption occasions will also be examined.
• These findings will be used to identify audience segments with distinct information needs, perceptions and behavioural characteristics.
• Audience segmentation will support the development of communication strategies tailored to different consumer groups and exposure scenarios.
• The protocol also notes that EFSA is undertaking additional research and pilot studies aimed at testing and refining communication approaches related to fish consumption advice within selected Member States.

23. Throughout the entire assessment process, uncertainties will be systematically identified, documented and evaluated. EFSA intends to consider uncertainty at each stage of the assessment process and assess its influence on overall conclusions.

24. Overall, the protocol covers an ambitious programme of work intended to deliver one of the most comprehensive risk–benefit assessments of fish consumption undertaken to date. By integrating evidence on contaminants, nutrients, health outcomes, population exposure and consumer communication, the assessment aims to provide a scientifically robust basis for future policy development and national dietary guidance relating to fish consumption across Europe.

25. This Item is presented for discussion of the COT as a whole. The consultation period closes on 03/08/2026 and therefore the Secretariat welcomes comments until COP 17th of July, after which the Committee comments will be compiled and submitted on behalf of the COT.  A summary document will be placed in the collaboration folder for comments, or these can be sent to the Secretariat directly. Please provide page and line numbers in your comment submissions.

Questions on which views of the Committee are sought:

a) Does the COT agree with EFSA’s overall approach and methodology proposed in their draft protocol for the risk-benefit assessment of fish consumption?

b) Does the COT have any comments with regards to the defined assessment questions, and the intended use of a Tier 3 integration approach with common metrics like DALYs?

c) Does the COT have any comments on the structure, clarity, and comprehensiveness of EFSA’s draft protocol?

d) Does the COT have any other comments?

Secretariat

June 2026

Annex A to TOX/2026/27

Committee on the Toxicity of Chemicals in Food, Consumer Products and the Environment

Draft protocol for the Risk-Benefit Assessment of fish consumption in relation to the combined presence of 10 PCDD/Fs and DL-PCBs, PFAS, PBDEs, inorganic arsenic, DMA, inorganic mercury and methyl mercury in fish.

Link to the consultation: Public Consultation

Secretariat

June 2026