Risk assessmentindependent cross-checkfull text checked I-FSH-18 in-depth review

EFSA: PFAS in food (2020)

Risk to human health related to the presence of perfluoroalkyl substances in food.

Overview

EFSA Panel on Contaminants in the Food Chain (EFSA CONTAM Panel) et al.. EFSA J 2020 · PMID 32994824 · DOI 10.2903/j.efsa.2020.6223

Fish meat, fruit and eggs were the main dietary PFAS sources; EFSA set a weekly limit of 4.4 ng/kg body weight, which parts of the population exceed. Freshwater fish such as eel had the highest levels; salmon/trout were low.

Does the study support the specific claim? Unclear · Methodological quality: high

Original publication

Plain-language summary

Study explained simply

What did the researchers want to find out?: Set safe intake.
How was the study done?: EFSA risk assessment (BMD and PBPK modelling; EU occurrence data).
How many people, animals or samples were studied?: 69,433 food analytical results from 16 European countries.
What was actually found?: TWI 4.4 ng/kg bw/week; adults' average exposure up to 5 times and high exposure up to 16 times the TWI (lower-bound estimates); fish meat a main source.
What do the results mean in everyday life?: Fish is a PFAS source regardless of farming.
What are the study's limitations?: 92% of results below detection; most data from three countries; based on one critical human study.
What does it say about Elawady's specific claim?: Context.
Do the results apply to Germany and Europe?: High.
What can explicitly not be concluded from this study?: Whether farmed or wild fish contain more PFAS; individual health risk.

Scientific in-depth review

1. Full reference

Authors: EFSA Panel on Contaminants in the Food Chain (EFSA CONTAM Panel) et al.
Journal: EFSA J
Year: 2020
Identifier: PMID 32994824 DOI 10.2903/j.efsa.2020.6223
Verification: verified – PubMed efetch 2026-10-25
Retracted: no
Corrections: none
Extent checked: full text
Conflicts of interest and funding: EFSA scientific opinion requested by the European Commission; prepared by the CONTAM Panel and working group under EFSA's independence and declarations-of-interest policy; no specific conflicts reported in the full text

2. Research question

Risks of PFASs in food

3. Study design and population

Study design: Risk assessment – EFSA CONTAM Panel scientific opinion (mandate from the European Commission; public consultation Feb-Apr 2020): hazard identification from animal and human studies, BMD modelling on human immune data, PBPK modelling, and dietary exposure assessment from EU occurrence data
Population: EU population
Country: EU
Period: Occurrence data from samples collected 2000-2016 (74% after 2007); opinion 2020
Sample size: 97,434 food analytical results from 16 European countries; 69,433 after cleaning (26 PFASs; 92% left-censored); 67,839 used for exposure; fish meat n=23,572 results (~33% of data); ~84% of samples from Germany, Norway and France

4. Methods and endpoints

Exposure / intervention: Sum of PFOA, PFNA, PFHxS and PFOS (equal potency assumed) from food
Comparator: TWI 4.4 ng/kg bw per week for the sum of 4 PFASs
Endpoints: Critical effect: reduced vaccine antibody response in children (immune system); also serum cholesterol, liver, birth weight considered
Statistical methods: Benchmark dose modelling (BMDL10 17.5 ng/mL serum in 1-year-olds), PBPK modelling to derive maternal intake (0.63 ng/kg bw per day), lower- and upper-bound exposure estimates; structured uncertainty analysis

5. Actual results

Results: TWI 4.4 ng/kg bw per week (sum of PFOA, PFNA, PFHxS, PFOS). Mean LB exposure: adolescents/adults 3-22, P95 9-70 ng/kg bw per week (highest mean exceeds TWI 5-fold, highest P95 16-fold); toddlers/other children mean 6-46, P95 19-96; UB estimates 4- to 49-fold higher but less reliable. Main contributors to combined exposure: 'Fish meat', 'Fruit and fruit products' and 'Eggs and egg products'. Highest mean PFOS among fish in carp, eel (mean LB 9.23 ug/kg), roach, perch, bream, barbel and sardine; salmon and trout mean LB PFOS 0.31 ug/kg (88% left-censored). The opinion does not compare farmed and wild fish.
Absolute risks: not reported (exposure vs TWI)
Authors' conclusion: Parts of the European population exceed the TWI of 4.4 ng/kg bw per week, which is of concern; exposure assessment carries high uncertainty (left-censored data, limited countries).

6. Strengths

Comprehensive authoritative assessment based on human data for the critical endpoint
Large EU occurrence dataset
Explicit uncertainty evaluation and public consultation

7. Weaknesses

92% of results left-censored; large LB-UB gap
Occurrence data concentrated in 3 countries
TWI rests on a single critical human study (vaccine response in 1-year-olds) and PBPK assumptions
Equal potency assumed for 4 PFASs
No farmed-vs-wild fish comparison

8. Risk of bias

Orienting tool: No standard tool fits (regulatory risk assessment); structured judgement of validity threats
Type of assessment: structured assessment based on the full text by a single reviewer (no formal double review)
Full formal assessment: no
Overall risk of bias: some concerns
Rationale: Single-reviewer structured judgement based on the full text; not formal (needs two independent reviewers). Changed from not_assessable to some_concerns.

8. Risk of bias
Domain	Judgement	Rationale
Hazard characterisation	some concerns	Critical endpoint from observational human studies with possible confounding; BMD/PBPK assumptions documented.
Occurrence data representativeness	some concerns	~84% from three countries; high left-censoring.
Exposure modelling	some concerns	LB vs UB differences up to 49-fold; LB judged more reliable and corroborated by serum levels.
Transparency and process	low	Public consultation, detailed uncertainty table, full documentation.

Full-text review (25 October 2026): Added dataset size and left-censoring, BMDL/PBPK derivation, exposure ranges and exceedance factors (5-fold mean, 16-fold P95), fish-species PFOS levels (eel 9.23 vs salmon/trout 0.31 ug/kg), uncertainty sources; noted no farmed-vs-wild comparison; bias overall changed from not_assessable to some_concerns.

9. Assessment of the statistical interpretation

Exceedance conclusions rest on lower-bound estimates, which may underestimate exposure, but are supported by measured serum levels. Fish-species means are dominated by freshwater species; salmon/trout values are low and largely below LOQ.

10. Relevance to the specific claim

Context: fish (wild and farmed alike) is a main PFAS source; not addressed by the farmed/wild rule.

11. Independent contrary findings

12. Transferability to Germany and Europe

High.

13. Overall methodological judgement

Authoritative EU opinion: fish meat is one of the main dietary PFAS sources, with the highest PFOS levels in freshwater species such as carp, eel and perch and in sardine; salmon/trout levels were low. Wild fish are not exempt from PFAS, but the opinion gives no farmed-vs-wild comparison.

Multidimensional assessment
Multidimensional assessment
Authenticity of the source	verified
Methodological quality	high
Risk of bias	some concerns
Precision of results	not applicable
Directness	indirect
Reproducibility	unknown
Consistency with the overall evidence	unknown
Regional transferability	high
Adequacy of the interpretation (by the archive)	not applicable

14. Open questions

Farmed vs wild PFAS levels; current occurrence data outside the three main reporting countries.

15. Linked fact checks

Natural wild sea fish is the cleanest food you can get; wild sea fish (including large predators such as grouper/hamour, tuna, and smoked mackerel) is allowed while farmed fish is forbidden. – Studies contradicting or limiting the claim