Update of the risk assessment on emerging and novel brominated flame retardants in food

Abstract

The European Commission asked EFSA to update its 2012 risk assessment on Emerging and Novel brominated flame retardants (BFRs) in food. Since the previous Opinion, further information was collected for a total of 27 BFRs. The CONTAM Panel concluded that TDBPP, DBPNG and TBNPA are considered genotoxic and carcinogenic. Based on positive in vitro mutagenicity in mammalian cells, DBE‐DBCH is considered genotoxic. EBTEBPI, DBDPE, PBB‐Acr and TTBP‐TAZ are not genotoxic. Insufficient or absent data precluded the Panel to conclude on the genotoxic potential of the remaining BFRs. Reproductive/developmental toxicity and/or neurotoxicity/neurodevelopmental effects were reported for DBNPG, BEH‐TEBP, EH‐TBB, DBDPE and TDBP‐TAZTO. Reference Points were identified for TDBPP, DBNPG, BEH‐TEBP, EH‐TBB, DBDPE and TDBP‐TAZTO. No health‐based guidance values could be derived for any of the BFRs. Occurrence data were submitted to EFSA for eight BFRs (BEH‐TEBP, EH‐TBB, BTBPE, DBDPE, HBB, PBEB, PBT, TBX). Dietary exposure could be estimated only for BTBPE and HBB, due to the lack of quantitative data for the other BFRs. However, no risk characterisation could be performed for these BFRs due to the lack of a Reference Point. Using the limited available information on dietary exposure estimates reported in the scientific literature, risk characterisation could be performed for BEH‐TEBP, EH‐TBB and DBDPE, and the CONTAM Panel concluded that dietary exposure does not raise a health concern. The Panel considered that the level of certainty was 90% or higher for most of the reported conclusions. Very limited or no information on hazard or exposure were available for HBCYD, TBCO, DBHCTD, HCTBPH, BDBP‐TAZTO, DBP‐TAZTO, OBTMPI, DBS, HEEHP‐TEBP, 4’‐PeBPO‐BDE208 or TTBNPP. Recommendations were formulated to address data gaps, including the need for access to original data, and for the collection of biomonitoring data, occurrence data in food, genotoxicity information and toxicological studies to support the identification of a Reference Point.

SUMMARY

Brominated flame retardants (BFRs) are chemicals used in a wide variety of consumer/commercial products to improve their resistance to fire. Concern has been raised because of the occurrence of several chemical compounds from the group of BFRs in the environment, food and in humans. This has led to bans on the production and use of certain formulations.

The European Commission asked the European Food Safety Authority (EFSA) to update its 2010–2012 risk assessments on the different families of BFRs, i.e. hexabromocyclododecanes (HBCDDs), polybrominated diphenyl ethers (PBDEs), tetrabromobisphenol A (TBBPA) and its derivatives, brominated phenols and their derivatives, and novel and emerging BFRs. The CONTAM Panel is updating the risk assessments of different classes of BFRs in a series of separate Opinions.

The similarities in chemical properties and effects seen in the previous EFSA assessments for the different BFR families warrant the consideration of a mixture approach. The Panel on Contaminants in the Food Chain (CONTAM Panel) will evaluate the appropriateness of applying a mixture approach in an additional Opinion once the risk assessment for each BFR family has been updated. It will be based on the EFSA Guidance on harmonised methodologies for human health, animal health and ecological risk assessment of combined exposure to multiple chemicals published in 2019 and 2021.

The first four Opinions in the current series updated the risk assessments of HBCDDs in food, PBDEs in food, TBBPA and its derivatives in food and brominated phenols and their derivatives in food. This fifth Opinion updates the risk assessment of emerging and novel BFRs in food previously performed by EFSA and published in 2012.

The draft Scientific Opinion underwent a targeted consultation from 28 April to 18 May 2026. The comments received were taken into account when finalising the Scientific Opinion and are presented and addressed in Annex F.

As requested in the Terms of Reference by the EC, the current assessment considered the six BFRs assessed in the 2012 Opinion and included in Commission Recommendation 2014/118/EU:

• tris(2,3‐dibromopropyl) phosphate (TDBPP),

• bis(2‐ethylhexyl) tetrabromophthalate (BEH‐TEBP),

• 2‐Ethylhexyl 2,3,4,5‐tetrabromobenzoate (EH‐TBB),

• N,N′‐Ethylenebis(tetrabromophthalimide) (EBTEBPI),

• Hexabromocyclodecane (HBCYD),

• Dibromoneopentyl glycol (DBNPG).

It also considered the additional 21 BFRs not included in this Commission Recommendation but assessed in the previous EFSA Opinion:

• 1,2‐bis(2,4,6‐tribromophenoxy)ethane (BTBPE),

• Decabromodiphenyl ethane (DBDPE),

• 4‐(1,2‐Dibromoethyl)‐1,2‐dibromocyclohexane (DBE‐DBCH),

• 5,6‐dibromo‐1,10,11,12,13,13‐hexachloro‐11‐tricyclo[8.2.1.0^2,9]tridecene (DBHCTD),

• Hexabromobenzene (HBB),

• 1,2,3,4,7,7‐Hexachloro‐5‐(2,3,4,5‐tetrabromophenyl)‐bicyclo[2.2.1]hept‐2‐ene (HCTBPH),

• Octabromotrimethylphenyl indane (OBTMPI),

• Pentabromobenzyl acrylate (PBB‐Acr),

• Pentabromoethylbenzene (PBEB),

• Pentabromotoluene (PBT),

• Tribromoneopentyl alcohol (TBNPA),

• 1,3,5‐Tris(2,3‐dibromopropyl)‐1,3,5‐triazine‐2,4,6‐trione (TDBP‐TAZTO),

• 1,2,5,6‐Tetrabromocyclooctane (TBCO),

• 2,3,5,6‐Tetrabromo‐p‐xylene (TBX),

• 1,3‐Bis(2,3‐dibromopropyl)‐5‐(2‐propen‐1‐yl)‐1,3,5‐triazine‐2,4,6(1H,3H,5H)‐trione (BDBP‐TAZTO),

• 1‐(2,3‐Dibromopropyl)‐3,5‐diallyl‐1,3,5‐triazine‐2,4,6‐(1H,3H,5H)trione (DBP‐TAZTO),

• dibromoestyrene (DBS),

• 2‐(2‐Hydroxyethoxy)ethyl 2‐hydroxypropyl 3,4,5,6‐tetrabromophthalate (HEEHP‐TEBP),

• Tetradecabromo‐1,4‐diphenoxybenzene (4’‐PeBPO‐BDE208),

• Tris(tribromoneopentyl) phosphate (TTBNPP),

• 2,4,6‐tris(2,4,6‐tribromophenoxy)‐1,3,5‐triazine (TTBP‐TAZ).

The previous Opinion made a distinction between ‘emerging’ and ‘novel’ BFRs, but for the current update, the CONTAM Panel decided not to make this distinction.

The present assessment takes into account the occurrence data in food and biological samples submitted to EFSA after the publication of its previous Opinion on ‘Emerging and Novel’ BFRs in food, as well as the newly available scientific information of relevance to hazard identification and characterisation.

Eight of the BFRs assessed (EBTEBPI, BEH‐TEBP, DBNPG, DBDPE, TDBP‐TAZTO, PBB‐Acr, TTBNPP, TTBP‐TAZ) hold a registered status according to Regulation (EC) No 1907/2006 of the European Parliament and of the Council of 18 December 2006 concerning the Registration, Evaluation, Authorisation and Restriction of Chemicals (REACH Regulation). The studies and information reported in the REACH dossier, as submitted by the registrant(s), was considered by the CONTAM Panel to inform the hazard identification and characterisation of the BFRs under assessment, when relevant, noting that the Panel did not have access to the original data.

The CONTAM Panel also used the OECD QSAR Toolbox and VEGA QSAR models to generate predictions on Ames test results to inform on the mutagenic potential, to provide insights on the toxicological mode of action as regards genotoxicity and to identify potential for read‐across by generating evidence for analogue identification and structural and mechanistic similarity. The OECD QSAR Toolbox was additionally used to capture experimental data present in databases to complement the literature search performed.

Overall, the information available on the toxicity of the BFRs included in the Terms of Reference varied greatly across compounds, but was in general scarce. Evidence of toxicity for reproduction or development in rodents was identified for DBNPG, BEH‐TEBP, EH‐TBB and DBDPE. Evidence of neurotoxicity or developmental neurotoxicity in rodents was identified for BEH‐TEBP, EH‐TBB, DBDPE and TDBP‐TAZTO. Studies in zebrafish larvae indicated potential for developmental neurotoxicity for DBE‐DBCH.

Regarding genotoxicity, the CONTAM Panel concluded that:

• −Based on the information provided in the REACH dossier, EBTEBPI, DBDPE, PBB‐Acr and TTBP‐TAZ are not genotoxic.
• − TDBPP is genotoxic and carcinogenic in rats (in kidney) and mice (in kidney, liver, forestomach, lung).
• − DBNPG is genotoxic and carcinogenic in rats (e.g. in skin, subcutaneous tissue, mammary gland, Zymbal's gland, oral cavity, oesophagus, forestomach, small and large intestines, mesothelium, urinary bladder, lung and thyroid gland) and mice (e.g. in Harderian gland, lung, subcutaneous tissue, forestomach).
• −In the absence of conclusive negative results for in vivo genotoxicity data, TBNPA is considered genotoxic based on the positive in vitro mutagenicity and clastogenicity. TBNPA is predicted to be carcinogenic based on the positive in vitro mutagenicity and clastogenicity supported by structural similarity with the genotoxic and carcinogenic DBNPG.
• −In the absence of conclusive negative results for in vivo mutagenicity, DBE‐DBCH is considered genotoxic based on positive in vitro mutagenicity in mammalian cells.
• − BEH‐TEBP, BTBPE, HBB, PBEB and PBT are not mutagenic, but the lack of data on clastogenicity and aneugenicity did not allow the Panel to conclude on their genotoxic potential.
• −QSAR VEGA models predicted negative Ames test results for DBHCTD and TBX, but the lack of data on clastogenicity and aneugenicity did not allow the Panel to conclude on their genotoxic potential.
• −Insufficient (TTBNPP) or no data (HBCYD, EH‐TBB, TDBP‐TAZTO, TBCO, HCTBPH, BDBP‐TAZTO, DBP‐TAZTO, DBS, HEEHP‐TEBP, 4’‐PeBPO‐BDE208, OBTMPI) were identified to allow an assessment of the genotoxic potential of the remaining BFRs.

The CONTAM Panel identified a Reference Point for TDBPP, DBNPG, BEH‐TEBP, EH‐TBB, DBDPE and TDBP‐TAZTO:

• −For TDBPP, a Reference Point corresponding to a BMDL₁₀ of 0.073 mg/kg bw per day was identified, based on combined incidence of kidney tubular‐cell adenoma and carcinoma in male rats.
• −For DBNPG, a Reference Point corresponding to a BMDL₁₀ of 1.14 mg/kg bw per day for the technical product FR‐1138 (purity 78.6%), corresponding to 0.9 mg/kg bw per day for DBNPG was identified, based on increased incidence of mammary gland fibroadenoma, adenoma or carcinoma in female rats.
• −For BEH‐TEBP, a Reference Point corresponding to lowest‐observed‐adverse‐effect level (LOAEL) of 1 mg/kg bw per day was identified, based on neurobehavioural and skeletal effects in a rat study with binary mixture BEH‐TEBP and EH‐TBB.
• −For EH‐TBB, a Reference Point corresponding to LOAEL of 2 mg/kg bw per day was identified, based on neurobehavioural and skeletal effects in a rat study with binary mixture BEH‐TEBP and EH‐TBB.
• −For DBDPE, a Reference Point corresponding to a BMDL₁₀ of 1.07 μg/kg bw per day was identified based on a neurobehavioural endpoint, an increase in total wheel running activity.
• −For TDBP‐TAZTO, a Reference Point corresponding to a LOAEL of 5 mg/kg bw per day was identified, based on depressive related behaviour in the forced swim test and the tail suspension test in rodents.

Due to limited or complete lack of data on hazard identification and characterisation, the CONTAM Panel could not identify a Reference Point for the remaining BFRs included in the Terms of Reference (HBCYD, EBTEBPI, BTBPE, DBE‐DBCH, DBHCTD, HBB, HCTBPH, OBTMPI, PBB‐Acr, PBEB, PBT, TBNPA, TBCO, TBX, BDBP‐TAZTO, DBP‐TAZTO, DBS, HEEHP‐TEBP, 4’‐PeBPO‐BDE208, TTBNPP, TTBP‐TAZ).

No health‐based guidance values (HBGV) could be derived for any of the BFRs included in the Terms of Reference.

Occurrence data in human milk and blood/serum in European countries were identified in the scientific literature for BEH‐TEBP, EH‐TBB, DBDPE, BTBPE, HBB, DBE‐DBCH, PBEB, PBT, TBCO, DBHCTD, TBX and OBTMPI. The number of studies was limited and characterised by low detection frequencies, and relied on samples collected around a decade ago or more. No data on human samples in European countries was identified for the other BFRs included in the Terms of Reference.

In the previous 2012 Opinion, occurrence data in food were submitted to EFSA for three of the BFRs included in the Terms of Reference (BTBPE, DBDPE and HBB) by two countries only (Ireland and the UK) covering the years 2006–2008. For the current update, occurrence data in food were submitted to EFSA for two out of the six BFRs covered in Commission Recommendation 2014/118/EU (BEH‐TEBP and EH‐TBB) and for six out of the remaining 21 BFRs covered in the 2012 Opinion (BTBPE, DBDPE, HBB, PBEB, PBT, TBX). These data were provided by six European Countries (France, Ireland, Norway, Austria, Germany, Portugal), from samples collected between 2010 and 2025. Most results for HBB were provided by Portugal (68%) and France (28%) between 2010 and 2017. France was the main data provider for BTBPE (90%) between 2014 and 2017.

A total of 2833 analytical results were submitted to EFSA as follows: HBB (n = 2004, 82% left‐censored), BTBPE (n = 630, 78% left‐censored), DBDPE (n = 76, 96% left‐censored), BEH‐TEBP (n = 21, 100% left‐censored), EH‐TBB (n = 21, 100% left‐censored), PBEB (n = 34, 100% left‐censored), PBT (n = 34), TBX (n = 13, 100% left‐censored).

Only the data availability for HBB and BTBPE was considered appropriate to perform a dietary exposure assessment. For DBDPE, 66 samples on ‘Fish and seafood’ were available but the only three positive samples were from 2010, thus not representative of the recent situation. In addition, insufficient data were available on other food categories. Thus, the CONTAM Panel deemed it not appropriate to perform an exposure assessment for DBDPE. The limited and left‐censored data available for BEH‐TEBP, EH‐TBB, PBEB, PBT and TBX were not further considered.

For BTBPE, mean LB dietary exposure ranged from 0.001 ng/kg bw per day in ‘Adults’, ‘Elderly’ and ‘Very Elderly’ to 0.006 ng/kg bw per day in ‘Infants’ and ‘Toddlers’, while mean UB dietary exposure ranged from 0.006 ng/kg bw per day in ‘Elderly’ and ‘Very Elderly’ to 0.104 ng/kg bw per day in ‘Elderly’ and ‘Very Elderly’, across surveys. The P95 LB dietary exposure ranged from 0.002 ng/kg bw per day in ‘Elderly’ and ‘Very Elderly’ to 0.014 ng/kg bw per day in ‘Other children', while the P95 UB dietary exposure ranged from 0.011 ng/kg bw per day in ‘Verly Elderly’ to 0.328 ng/kg bw per day in ‘Elderly’, across surveys. The raw primary commodities that contributed most to the mean lower bound dietary exposure belong to the ‘Milk and dairy products’ (47% to 88%) and ‘Meat and meat products’ (9%–40%) food categories.

For HBB, mean LB dietary exposure ranged from 0.04 ng/kg bw per day in ‘Adults’, ‘Elderly’ and ‘Very Elderly’ to 0.25 ng/kg bw per day in ‘Toddlers’, while mean UB dietary exposure ranged from 0.05 ng/kg bw per day in ‘Elderly’ and ‘Very Elderly’ to 0.3 ng/kg bw per day in ‘Toddlers’, across surveys. The P95 LB dietary exposure ranged from 0.06 ng/kg bw per day in ‘Elderly’ to 0.61ng/kg bw per day in ‘Other children’, while P95 UB dietary exposure ranged from 0.09 ng/kg bw per day in ‘Elderly’ to 0.67 ng/kg bw per day in ‘Other children', across surveys. The food category that mostly contributed to the dietary exposure was ‘Milk’ (87%–98%) across all age groups.

In the scientific literature limited data on dietary exposure estimates in European countries were identified for BEH‐TEBP, EH‐TBB, DBDPE, BTBPE, DBE‐DBCH, PBEB, PBT and TBX, covering one or two European countries. No information on potential dietary exposure in European countries was identified for the remaining BFRs (TDBPP, EBTEBPI, HBCYD, DBNPG, TDBP‐TAZTO, TBCO, DBHCTD, HCTBPH, PBB‐Acr, TBNPA, DBDP‐TAZTO, DBP‐TAZTO, DBS, HEEHP‐TEBP, 4’‐PeBPO‐BDE208, TTBNPP and TTBP‐TAZ).

For breastfed infants, there were insufficient data in human milk from European countries reported in the scientific literature to make robust exposure estimates, but an indication of mean and/or high level (P95) potential exposure was possible for BEH‐TEBP, EH‐TBB, DBDPE, BTBPE, HBB, DBE‐DBCH, PBT and TBX.

Risk characterisation could be performed for two out of the six BFRs covered by Commission Recommendation 2014/118/EU (BEH‐TEBP and EH‐TBB) and for one out of the remaining 21 BFRs assessed in the 2012 Opinion (DBDPE), making use of the limited information on dietary exposure estimates from one (EH‐TBB, DBDPE) or two (BEH‐TEBP) European countries reported in the scientific literature. The margin of exposure (MOE) approach was applied to assess possible health concerns.

For BEH‐TEBP, the CONTAM Panel considered that MOEs ≥ 1500 would be sufficient to conclude that no health concern is raised for non‐neoplastic effects. This consideration was based on the default uncertainty factor of 100 (i.e. factor 10 for the extrapolation from experimental animals to humans and a factor 10 for inter‐individual human variability), an additional factor of 3 to allow for the absence of a NOAEL and the necessity of using a LOAEL, and an additional factor of 5 to account for deficiencies in the database, in particular in relation to inconsistencies between studies of non‐standard design, most studies being on mixtures rather than on the pure substance, and some studies only being available in summary form, and considering the worst‐case approach of attributing the effects solely to BEH‐TEBP. Based on the limited dietary exposure estimates from the literature and the LOAEL of 1 mg/kg bw per day based on neurobehavioural and skeletal effects in a rat study with binary mixture of BEH‐TEBP and EH‐TBB, the CONTAM Panel concluded that dietary exposure to BEH‐TEBP does not raise a health concern for non‐neoplastic effects. The Panel could not set a Reference Point for neoplastic effects, since no carcinogenicity data are available. In addition, it was not possible to conclude on the genotoxic potential of BEH‐TEBP.

For EH‐TBB, the CONTAM Panel considered that MOEs ≥ 1500 would be sufficient to conclude that no health concern is raised for non‐neoplastic effects. This consideration was based on the default uncertainty factor of 100 (i.e. factor 10 for the extrapolation from experimental animals to humans and a factor 10 for inter‐individual human variability), an additional factor of 3 to allow for the absence of a NOAEL and the necessity of using a LOAEL, and an additional factor of 5 to account for deficiencies in the database, in particular in relation to inconsistencies between studies of non‐standard design, all but one of the studies being on mixtures rather than on the pure substance, no experimental data on genotoxicity, and some studies only being available in summary form, and considering the approach of attributing the effects solely to EH‐TBB. Based on the limited dietary exposure estimates from the literature and the LOAEL of 2 mg/kg bw per day based on neurobehavioural and skeletal effects in a rat study with binary mixture of BEH‐TEBP and EH‐TBB, the CONTAM Panel concluded that dietary exposure to EH‐TBB does not raise a health concern for non‐neoplastic effects. The CONTAM Panel could not set a Reference Point for neoplastic effects, since no carcinogenicity data are available. In addition, it was not possible to conclude on the genotoxic potential of EH‐TBB.

For DBDPE, the CONTAM Panel considered that MOEs ≥ 300 would be sufficient to conclude that no health concern is raised for the critical effect. This consideration was based on the default uncertainty factor of 100 (i.e. factor 10 for the extrapolation from experimental animals to humans and a factor 10 for inter‐individual human variability), and an additional factor of 3 to account for deficiencies in the database, in particular in relation to the lack of long‐term toxicity studies, and the fact that the Reference Point comes from a study showing mild effects on a sensitive endpoint, at much lower doses than those at which effects were seen in other studies. Considering the limited dietary exposure estimates from the literature and the BMDL₁₀ of 1.07 μg/kg bw per day based on the increase of total wheel running activity, the CONTAM Panel concluded that dietary exposure to DBDPE does not raise a health concern.

No risk characterisation could be performed for any of the remaining BFRs included in the Terms of Reference. The CONTAM Panel noted that:

• −For TDBPP, DBNPG and TDBP‐TAZTO, although a Reference Point could be identified, risk characterisation was not possible due to the lack of information on potential dietary exposure.
• −For BTBPE and HBB, although an exposure assessment could be performed based on the occurrence data in food submitted to EFSA, risk characterisation was not possible due to the lack of sufficient toxicological data to identify a Reference Point.
• −Due to the lack of a Reference Point and limited information on potential dietary exposure, risk characterisation was not possible for DBE‐DBCH, PBEB, PBT and TBX.
• −Due to the lack of a Reference Point and lack of information on potential dietary exposure, risk characterisation was not possible for EBTEBPI, HBCYD, TDBP‐TAZTO, TBCO, DBHCTD, HCTBPH, PBB‐Acr, TBNPA, BDBP‐TAZTO, DBP‐TAZTO, DBS, HEEHP‐TEBP, 4’‐PeBPO‐BDE208, TTBNPP, TTBP‐TAZ and OBTMPI.

Risk characterisation was not performed for breastfed infants for any of the BFRs for which a Reference Point could be identified, because of the highly uncertain exposure estimates.

Uncertainties affecting the assessment were identified and their overall impact was quantified in those cases where the available data were sufficient to conclude on genotoxic potential, hazard characterisation or risk characterisation. The CONTAM Panel assessed the level of certainty (probability) to be 90% or more for all the reported conclusions, except for the risk characterisation for DBDPE, where the certainty of no health concern was considered to fall within the range of 66%–90%.

The CONTAM Panel made the following recommendations:

• −Occurrence data in food, obtained using appropriate analytical methods, are needed for those BFRs for which insufficient or no occurrence data were submitted to EFSA. Priority should be given to: (i) BFRs for which a Reference Point is available and/or for those that the Panel has concluded to be genotoxic: TDBPP, BEH‐TEBP, EH‐TBB, DBNPG, DBDPE, DBE‐DBCH, TDBP‐TAZTO, TBNPA, (ii) BFRs that have been detected in human samples but lack a Reference Point or conclusion on genotoxicity: BTBPE, HBB, PBEB, PBT, TBX, OBTMPI.
• −Biomonitoring data, obtained using appropriate analytical methods should be generated. Priority should be given to TDBPP, BEH‐TEBP, EH‐TBB, DBNPG, DBDPE, HBB, TDBP‐TAZTO, DBE‐DBCH and TBNPA, as these show potential for bioaccumulation, or have an identified Reference Point, or are concluded to be genotoxic.
• −Genotoxicity data are needed, most importantly for the BFRs with inconclusive (BEH‐TEBP, BTBPE, HBB, PBEB, PBT, TBX) or absent (EH‐TBB) genotoxicity information and that are known to occur in food.
• −Toxicological studies that would enable the identification of a Reference Point are needed for those BFRs for which insufficient or no studies were identified. Priority should be given to the BFRs for which evidence of potential dietary exposure was identified and/or they have been detected in human samples (BTBPE, HBB, DBE‐DBCH, PBEB, PBT, TBCO, DBHCTD, TBX, OBTMPI).
• −In the context of the One Substance One Assessment, and to ensure efficient and harmonised risk assessments, there is a need to further enable access to the original data underpinning regulatory submissions across relevant authorities.

1. INTRODUCTION

1.1. Background and Terms of Reference as provided by the requestor

BACKGROUND

Brominated flame retardants (BFRs) are anthropogenic chemicals, which are added to a wide variety of consumer/commercial products in order to improve their fire resistance. The major classes of BFRs are brominated bisphenols, diphenyl ethers, cyclododecanes, phenols, biphenyl derivatives and the emerging and novel BFRs.

Concern has been raised because of the occurrence of several chemical compounds from the group of BFRs in the environment, including feed and food, and in humans. This has led to bans on the production and use of certain formulations of polybrominated diphenyl ethers (PBDEs).

Between September 2010 and September 2012, the Scientific Panel on Contaminants in the Food of EFSA adopted six scientific Opinions on different classes of brominated flame retardants. ¹ Because in its Opinion EFSA highlighted several data gaps, hampering the consumer risk assessment for these substances, by means of Commission Recommendation 2014/118/EU on the monitoring of traces of brominated flame retardants in food, Member States were recommended to collect in 2014 and 2015 occurrence data for specific substances in specific foodstuffs.

The newly available occurrence data would enable an updated consumer exposure assessment. Furthermore, since the publication of the EFSA scientific Opinions between 2010 and 2012, new scientific information has become available, therefore it would be necessary to verify whether an update of these scientific Opinions would be appropriate, including an update of the consumer risk assessment.

TERMS OF REFERENCE

In accordance with Art. 29 (1) of Regulation (EC) No 178/2002, the European Commission asks the European Food Safety Authority

• for an updated exposure assessment for the brominated flame retardants (BFRs), covered by Recommendation 2014/118/EU, taking into account the occurrence data in food, submitted after the publication of the 2010–2012 EFSA scientific Opinions, and an updated consumer risk assessment, taking into account newly available scientific information.

• for 21 novel and emerging brominated flame retardants (BFRs), which are not covered by Recommendation 2014/118:

  • ○
    to identify newly available scientific information in the public domain that could inform the risk assessment of any of these 21 BFRs, and categorise the studies according to the field of knowledge potentially relevant or informative for the assessment, e.g. toxicity in experimental animals, epidemiological studies, toxicokinetics, mode of action, levels in human tissues. This includes the occurrence data in food submitted to EFSA or reported in the literature.
  • ○
    to assess the new scientific information retrieved for indications/signals of BFRs of potential concern among these 21 BFRs. For those for which sufficient and relevant scientific information is available, conduct an updated consumer risk assessment, including an exposure assessment, taking into account the newly available evidence.

1.2. Interpretation of the Terms of Reference

Following the request from the European Commission (EC), the CONTAM Panel will update its 2010–2012 risk assessments on the different families of BFRs: hexabromocyclododecanes (HBCDDs), polybrominated diphenyl ethers (PBDEs), tetrabromobisphenol A (TBBPA) and its derivatives, brominated phenols and their derivatives and Emerging and Novel brominated flame retardants (BFRs) (EFSA CONTAM Panel, 2011a, 2011b, 2011c, 2012a, 2012b).

The first four Opinions in the series updated the risk assessments of HBCDDs in food (EFSA CONTAM Panel, 2021), PBDEs in food (EFSA CONTAM Panel, 2024a), TBBPA and its derivatives in food (EFSA CONTAM Panel, 2024b), and brominated phenols and their derivatives in food (EFSA CONTAM Panel, 2024c). This fifth Opinion updates the risk assessment of Emerging and Novel BFRs in food previously performed by EFSA (EFSA CONTAM Panel, 2012b).

The previous Opinion defined ‘emerging’ and ‘novel’ BFRs as follows:

• − ‘Emerging BFRs’ were defined as chemicals which are applied as flame retardants and that have been identified in any environmental compartment, in wildlife, in food or in humans. It should be noted that the definition for the emerging BFRs, used in that Opinion, did not imply that there was evidence for an increasing trend in the concentration of these BFRs in the environment, in food or in human samples.
• − ‘Novel BFRs’ were defined as chemicals applied as flame retardants, with confirmed presence in materials and/or goods in concentrations above 0.1%, but not identified in environmental samples, wildlife, food or humans.

In Commission Recommendation 2014/118/EU the following Emerging and Novel BFRs were listed and will be considered in the current update to answer the first part of the Terms of Reference:

• Tris(2,3‐dibromopropyl) phosphate (TDBPP, CAS No 126‐72‐7)

• Bis(2‐ethylhexyl) tetrabromophthalate (BEH‐TEBP, CAS No 26040‐51‐7)

• 2‐Ethylhexyl 2,3,4,5‐tetrabromobenzoate (EH‐TBB, CAS No 183658‐27‐7)

• N,N′‐Ethylenebis(tetrabromophthalimide) (EBTEBPI, CAS No 32588‐76‐4)

• Hexabromocyclodecane (HBCYD, CAS No 25495‐98‐1)

• Dibromoneopentyl glycol (DBNPG, CAS No 3296‐90‐0)

The remaining Emerging and Novel BFRs considered in the 2012 EFSA Opinion and not covered by Recommendation 2014/118/EU that will be addressed in the second part of the Terms of Reference are the following:

• 1,2‐bis(2,4,6‐tribromophenoxy)ethane (BTBPE, CAS No 37853‐59‐1)

• Decabromodiphenyl ethane (DBDPE, CAS No 84852‐53‐9)

• 4‐(1,2‐Dibromoethyl)‐1,2‐dibromocyclohexane (DBE‐DBCH, CAS No 3322‐93‐8)

• 5,6‐Dibromo‐1,10,11,12,13,13‐hexachloro‐11‐tricyclo[8.2.1.0^2,9]tridecene (DBHCTD, CAS No 51936–55‐1)

• Hexabromobenzene (HBB, CAS No 87‐82‐1)

• 1,2,3,4,7,7‐Hexachloro‐5‐(2,3,4,5‐tetrabromophenyl)‐bicyclo[2.2.1]hept‐2‐ene (HCTBPH, CAS No 34571‐16‐9)

• Octabromotrimethylphenyl indane (OBTMPI, CAS No 1084889‐51‐9, 1025956–65‐3, 893843–07‐7, 155613‐93‐7)

• Pentabromobenzyl acrylate (PBB‐Acr, CAS No 59447‐55‐1)

• Pentabromoethylbenzene (PBEB, CAS No 85‐22‐3)

• Pentabromotoluene (PBT, CAS No 87‐83‐2)

• Tribromoneopentyl alcohol (TBNPA, CAS No 1522‐92‐5)

• 1,3,5‐Tris(2,3‐dibromopropyl)‐1,3,5‐triazine‐2,4,6‐trione (TDBP‐TAZTO, CAS No 52434‐90‐9)

• 1,2,5,6‐Tetrabromocyclooctane (TBCO, CAS No 3194‐57‐8)

• 2,3,5,6‐Tetrabromo‐p‐xylene (TBX, CAS No 23488‐38‐2)

• 1,3‐Bis(2,3‐dibromopropyl)‐5‐(2‐propen‐1‐yl)‐1,3,5‐triazine‐2,4,6(1H,3H,5H)‐trione (BDBP‐TAZTO, CAS No 75795‐16‐3)

• 1‐(2,3‐Dibromopropyl)‐3,5‐(2‐propen‐1‐yl)‐1,3,5‐triazine‐2,4,6(1H,3H,5H)‐trione (DBP‐TAZTO, CAS No 57829‐89‐7)

• Dibromoestyrene (DBS, CAS No 31780‐26‐4)

• 2‐(2‐Hydroxyethoxy)ethyl 2‐hydroxypropyl 3,4,5,6‐tetrabromophthalate (HEEHP‐TEBP, CAS No 20566‐35‐2)

• Tetradecabromo‐1,4‐diphenoxybenzene (4’‐PeBPO‐BDE208, CAS No 58965‐66‐5)

• Tris(tribromoneopentyl) phosphate (TTBNPP, CAS No 19186‐97‐1)

• 2,4,6‐Tris(2,4,6‐tribromophenoxy)‐1,3,5‐triazine (TTBP‐TAZ, CAS No 25713‐60‐4)

The CONTAM Panel will identify if the BFRs listed in the Terms of Reference are of potential concern through a weight of evidence assessment (EFSA Scientific Committee, 2017) of the following lines of evidence (in no particular order), noting that information on production, uses and environmental fate will not be subject to detailed review for the purpose of this exercise:

• −occurrence data in food and/or human milk,
• −studies on the genotoxicity of the substance,
• −in vivo toxicity data in experimental animals,
• −QSAR, read‐across and in vitro studies, related to genotoxicity, effects on reproduction/development, neurotoxicity and/or endocrine effects,
• −epidemiological data,
• −presence in human tissues,
• −toxicokinetic data.

The use of terms such as ‘emerging’, ‘novel’ and ‘legacy’, and inconsistent use of abbreviations for BFRs, even within the same published article, can lead to confusion. For this reason, Bergman et al. published a paper on ‘A novel abbreviation standard for organobromine, organochlorine and organophosphorus flame retardants and some characteristics of the chemicals’ (Bergman et al., 2012). This paper described a systematic way to abbreviate names of BFRs and also listed abbreviations that had been used historically. One particularly confusing example is for hexabromocyclododecanes (HBCDDs) and hexabromocyclodecane (HBCYD) where both are sometimes referred to as HBCD. In this Opinion, the abbreviations based on Bergman et al. (2012) are used.

The similarities in chemical properties and effects seen in the previous EFSA assessments for the different BFR families warrant the consideration of a mixture approach. The CONTAM Panel will evaluate the appropriateness of applying a mixture approach in an additional Opinion once the risk assessment for each BFR family has been updated. It will be based on the EFSA Guidance on harmonised methodologies for human health, animal health and ecological risk assessment of combined exposure to multiple chemicals (EFSA Scientific Committee, 2019).

1.3. Supporting information for the assessment

1.3.1. Physicochemical properties

The physicochemical properties of the BFRs under consideration were described in the previous EFSA Opinion (EFSA CONTAM Panel, 2012b). These are summarised in Table 1. Abbreviations are according to the publication of Bergman et al. (2012), who presented a harmonised set of abbreviations.

TABLE 1.

Physicochemical characteristics of the BFRs considered in this Opinion (based on Bergman et al., 2012).

Compound (common name)	Abbreviation used in this opinion	Other abbreviations	CAS	Molecular weight (g/mol)	Log Kow	pKa	Vapour pressure (Torr)
6 BFRs included in the 2012 Opinion and covered by Recommendation 2014/118/EU
Tris(2,3‐dibromopropyl) phosphate	TDBPP	tDBPrP, TBPP, TBP	126‐72‐7	697.61	3.71	NA	2.4 × 10−11
Bis(2‐ethylhexyl) tetrabromophthalate	BEH‐TEBP	bEtH‐TeBPht, TeBrDEHP, TBPH, BEHTBP	26040‐51‐7	706.14	9.34 10.2 a	NA	1.2 × 10−13
2‐Ethylhexyl 2,3,4,5‐tetrabromobenzoate	EH‐TBB	EtH‐TeBBzo, EHTeBB, EHTBB, TBB	183658‐27‐7	549.92	7.73	NA	2.8 × 10−9
N,N′‐Ethylenebis (tetrabromophthalimide)	EBTEBPI	N,N′‐EtbTeBPhtIm, BrPhtimi	32588‐76‐4	951.47	6.63	NA	1.5 × 10−27
Hexabromocyclodecane	HBCYD	HxBcDe, HBCD	25495‐98‐1	613.64	7.13 b	NA	4.2 × 10−6 b
Dibromoneopentyl glycol	DBNPG	bBMe‐PrDiOH, DBPT, BBMP, BMP	3296‐90‐0	261.94	0.41	13.6	5.2 × 10−7
Other 21 BFRs included in the 2012 Opinion
Decabromodiphenyl ethane	DBDPE	DBDiPhEt; DBDE; EBPE; DeBrPylE	84852‐53‐9	971.22	11.1	NA	6.0 × 10−15
4‐(1,2‐Dibromoethyl)‐1,2‐dibromocyclohexane	DBE‐DBCH	DiBEt‐DiBcH; TBEC; TBECH; BrCyHx	3322‐93‐8	427.8	4.82	NA	2.23 × 10−5
5,6‐Dibromo‐1,10,11,12,13,13‐hexachloro‐11‐tricyclo[8.2.1.02,9]tridecene	DBHCTD	DiBHxC‐TrcTrDen; HCDBCO	51936‐55‐1	540.76	7.62	NA	6.20 × 10−9
1,2‐Bis(2,4,6‐tribromophenoxy)ethane	BTBPE	EtH‐TeBBzo; TBB; EHTeBB	37853‐59‐1	549.92	7.73	NA	2.78 × 10−9
Hexabromobenzene	HBB	HxBBz; HxBrBz	87‐82‐1	551.49	6.11	NA	8.52 × 10−7
1,2,3,4,7,7‐Hexachloro‐5‐(2,3,4,5‐tetrabromophenyl)‐ bicyclo[2.2.1]hept‐2‐ene	HCTBPH	HxCTeBPh‐bcHen; Dec 604	34571‐16‐9	692.50	10.24	NA	1.21 × 10−10
Pentabromobenzyl acrylate	PBB‐Acr	PeBBn‐Acr; PeBrAcr; PBBA	59447‐55‐1	556.67	5.60	NA	2.7 × 10−9
Pentabromoethylbenzene	PBEB	PeBEtBz; PBrEtBz	85‐22‐3	500.65	6.76	NA	1.17 × 10−6
Pentabromotoluene	PBT	PeBT	87‐83‐2	486.62	6.25	NA	4.50 × 10−6
Tribromoneopentyl alcohol	TBNPA	tBMe‐EtOH	1522‐92‐5	324.84	2.06	14.2	1.20 × 10−5
1,3,5‐Tris(2,3‐dibromopropyl)‐1,3,5‐triazine‐2,4,6‐trione	TDBP‐TAZTO	tDiBPr‐Tazto; TBC; BrTriaz	52434‐90‐9	728.67	4.45	NA	3.07 × 10−15
1,2,5,6‐Tetrabromocyclooctane	TBCO	α−/β‐TeBcO	3194‐57‐8	427.8	5.28	NA	3.59 × 10−5
2,3,5,6‐Tetrabromo‐p‐xylene	TBX	TeBDiMeBz; p‐TBX	23488‐38‐2	421.75	6.20	NA	4.35 × 10−5
1,3‐Bis(2,3‐dibromopropyl)‐5‐(2‐propen‐1‐yl)‐1,3,5‐triazine‐2,4,6(1H,3H,5H)‐trione	BDBP‐TAZTO	bDiBPr‐A‐Tazto	75795‐16‐3	568.88	3.55	NA	8.90 × 10−12
1‐(2,3‐Dibromopropyl)‐3,5‐diallyl‐1,3,5‐triazine‐2,4,6(1H,3H,5H)‐trione	DBP‐TAZTO	DiBPr‐DiA‐Tazto	57829‐89‐7	409.07	2.66	NA	1.62 × 10−8
Dibromostyrene e (2,2‐Dibromovinyl)benzene; 3,5‐dibromostyrene	DBS	DiBSty; DBrsty	31780‐26‐4 120359‐56‐0	261.94	4.28 c 4.18 d	NA	7.94 × 10–2b 6.65 × 10–2c
2‐(2‐Hydroxyethoxy)ethyl 2‐hydroxypropyl 3,4,5,6‐tetrabromophthalate	HEEHP‐TEBP	OHEtOEt‐OHPr‐TeBPht; TeBrPht	20566‐35‐2	627.90	1.04	NA	3.59 × 10−15
Tetradecabromo‐1,4‐diphenoxybenzene	4’‐PeBPO‐BDE208	TeDB‐DiPhOBz; DPeTeBrBz	58965‐66‐5	1366.85	12.67	NA	6.88 × 10−19
Tris(tribromoneopentyl) phosphate	TTBNPP	tBbBMePrP; TrisPhos; TTBNP	19186‐97‐1	1018.46	7.55	NA	1.06 × 10−19
2,4,6‐ Tris(2,4,6‐tribromophenoxy)‐1,3,5‐triazine	TTBP‐TAZ	TrBPhO‐Taz; BrPhTriaz	25713‐60‐4	1067.43	12.97	NA	2.02 × 10−25
Octabromotrimethylphenyl indane (f)	OBTMPI	OBTrMePhIn; OBIND; Br‐Indan; OctaInd	1084889‐51‐9 1025956‐65‐3 893843‐07‐7 155613‐93‐7	867.52	15.11	NA	1.31 × 10−14

Abbreviations: CAS, Chemical Abstract Service; log K_ow, n‐octanol–water partition coefficient; pK_a, logarithm of the acid dissociation constant; NA, not applicable.

^a At 25°C taken from Reference: Bis(2‐ethylhexyl) tetrabromophthalate 100.043.099 | 60930694‐3eb6‐400e‐b8ab‐6fc23de9ddda ‐ ECHA CHEM.

^b Predicted median values taken from https://comptox.epa.gov/dashboard/chemical/properties/DTXSID70180181.

^c Predicted median values taken from https://comptox.epa.gov/dashboard/chemical/properties/DTXSID20178871.

^d Predicted median values taken from https://comptox.epa.gov/dashboard/chemical/properties/DTXSID90595612.

^e DBS includes all possible combinations of bromine atoms directly connected to the ring or the side chain, only two CAS numbers are available.

^f OBTMPI includes all possible combinations of bromine atoms directly connected to the aromatic or aliphatic rings, only 4 CAS numbers are available.

Polybrominated dibenzo‐p‐dioxins and furans (PBDD/Fs) are sometimes formed as byproducts in the manufacture of organo‐bromine compounds such as BFRs, or can be formed when they degrade. No direct evidence of their formation was found in the literature that was identified for the BFRs assessed in this Opinion.

1.3.2. Previous assessments

In 2012, the EFSA CONTAM Panel published its first risk assessment on Emerging and Novel BFRs in food (EFSA CONTAM Panel, 2012b) in which information on 17 Emerging and 10 Novel BFRs was reported. Due to the limited information identified at that time for occurrence, exposure and toxicity of the BFRs considered, the Panel could not perform a risk characterisation for any of these BFRs. Instead, the Panel made an attempt to identify those BFRs that could be a potential health concern and should be prioritised for future investigations, based on the available experimental data on behaviour in the environment, occurrence in food and toxicity.

The CONTAM Panel identified convincing evidence for genotoxicity and carcinogenicity of TDBPP and DBNPG and recommended further surveillance of their occurrence in the environment and in food. In addition, based on the limited experimental data on environmental behaviour, the Panel concluded that BTBPE and HBB could raise concerns for bioaccumulation (EFSA CONTAM Panel, 2012b).

The Panel additionally performed a modelling exercise selecting two environmental characteristics, overall persistence (Pov) and the potential for bioaccumulation, to predict the potential presence of the Emerging and Novel BFRs in food. Among the 27 BFRs, the modelling exercise predicted the following BFRs to have both a high overall persistence (Pov > 500 days) and a high potential for bioaccumulation: BTBPE, DBHCTD, EBTEBPI, HBB, HBCYD, HCTBPH, PBT, PBB‐Acr, PBEB and TBX (EFSA CONTAM Panel, 2012b).

Several bodies have performed hazard assessments for some of the BFRs considered in this Opinion, and these are briefly described below.

For TDBPP, and as noted in the previous Opinion (EFSA CONTAM Panel, 2012b), the International Agency for Research on Cancer (IARC) in its monograph on the Re‐evaluation of Some Organic Chemicals, Hydrazine and Hydrogen Peroxide classified TDBPP as probably carcinogenic to humans (Group 2A) (IARC, 1999). In December 2021, the National Toxicology Program (NTP) released its 15th Report on Carcinogens, ² where TDBPP is mentioned as reasonably anticipated to be a human carcinogen based on sufficient evidence of carcinogenicity from studies in experimental animals. The NTP based their conclusions on observations from studies reported in the IARC monograph (IARC, 1999).

Regarding DBNPG, the Danish Environmental Protection Agency, Centre for Integrated Environment and Toxicology (CETOX), published a report in 2000 based on data submitted to the United States‐Environment Protection Agency (US‐EPA), concluding that DBNPG should be considered potential carcinogenic (Simonsen et al., 2000). As noted in the previous Opinion (EFSA CONTAM Panel, 2012b), IARC published a monograph on Evaluation of Carcinogenic Risks to Humans (IARC, 2000). IARC classified DBNPG as possibly carcinogenic to humans (Group 2B). In 2018, ECHA's Committee for Risk Assessment (RAC) adopted its opinion ³ on DBNPG based on the proposal submitted in December 2016, by the Norwegian Competent authority ⁴ for harmonised classification and labelling (CLH) as mutagenic and carcinogenic. In brief, ECHA's RAC considered that DBNPG should be classified as a germ cell mutagen, Cat. 1B, based on positive results for somatic cell mutagenicity from in vitro/in vivo studies and evidence from the reproductive toxicity studies that DBNPG reaches the (female) germ cells. In addition, ECHA's RAC considered DBNPG as ‘carcinogenic, Cat. 1B’ based on the evidence that DBNPG‐induced dose‐dependent multi‐site tumours in two species, rats and mice, with tumours of human relevance, in an OECD TG 453‐study conducted by the NTP under GLP conditions (NTP, 1996).

In 2020, ECHA's RAC adopted its opinion ⁵ on TBNPA based on the proposal submitted in June 2019, by the Norwegian Competent authority for CLH as mutagenic and carcinogenic. ECHA's RAC classified TBNPA as a ‘germ cell mutagen Category 2, H341: Suspected of causing genetic defects’ based on read‐across from DBNPG, a known germ cell mutagen. Additionally, ECHA's RAC classified TBNPA as ‘carcinogen category 1B, Carc 1B, H350: May cause cancer’ based on a detailed read‐across analysis from DBNPG, which was recognised as a multi‐site and multi‐species carcinogen (ECHA's RAC opinion as Carc. 1B), despite the absence of epidemiological and animal data for TBNPA (more details available in Section 3.3.12).

For the rest of BFRs included in the Terms of Reference of the current Opinion, no other risk or hazard assessments by any international body were identified.

1.3.3. Legislation

In this Opinion, where reference is made to European legislation (Regulations, Directives, Recommendations, Decisions), the reference should be understood as relating to the most recent amendment at the time of publication of this Opinion, unless otherwise stated.

In order to protect public health, Article 2 of Council Regulation (EEC) No 315/93 ⁶ of 8 February 1993 laying down Community procedures for contaminants in food stipulates that, where necessary, maximum tolerances for specific contaminants shall be established. The BFRs considered in this Opinion are not regulated under any specific legislation within the EU for food or feed, e.g. Commission Regulation (EU) 2023/915 ⁷ of 25 April 2023 and Council Directive 2002/32/EC. ⁸

Regarding Regulation (EC) No 1907/2006 ⁹ of the European Parliament and of the Council of 18 December 2006 concerning the Registration, Evaluation, Authorisation and Restriction of Chemicals (REACH Regulation), the 27 BFRs under the Terms of Reference have different statuses under REACH, which reflects the current use of the compounds in the European Economic Area (EEA) (Table 2). Registered status refers to substances/mixtures that hold a valid registration by one or more companies and can be placed on the EEA market. Pre‐registered status refers to substances/mixtures indicated, in 2009, as being intended to be registered by at least one company in the EEA. Notified classification and labelling (C&L) refers to substances either manufactured or imported in small amounts as substances/mixtures or in articles to the EU. Under REACH, a ‘persistent, bioaccumulative and toxic’ (PBT) or ‘very persistent and very bioaccumulative’ (vPvB) assessment is required for all substances for which a chemical safety assessment is carried out. A chemical safety assessment is required for substances manufactured or imported in amounts of 10 tonnes or more per year, unless exemptions apply. Substances that persist for long periods of time in the environment and have a high potential to accumulate in biota are characterised as PBT/vPvB and are considered of specific concern because their long‐term effects are rarely predictable. Substances of very high concern (SVHC) are identified substances, proposed by a Member State or ECHA, at the request of the European Commission, with the following hazard properties: a. Substances meeting the criteria for classification as carcinogenic, mutagenic or toxic for reproduction (CMR) category 1A or 1B in accordance with Regulation (EC) No 1272/2008 on classification, labelling and packaging of chemical substances; b. Substances which are PBT/vPvB according to REACH Annex XIII. c. Substances on a case‐by‐case basis, that cause an equivalent level of concern as CMR or PBT/vPvB substances.

TABLE 2.

Identification characteristics of the BFRs considered in this Opinion and ECHA registration status and classification of substances according to REACH regulation (chemical structures provided in Section 1.3.1).

Full name	Abbreviation	CAS number	ECHA registration status a , b	Additional ECHA information a
BFRs included in the 2012 Opinion and covered by Recommendation 2014/118/EU
Tris(2,3‐dibromopropyl) phosphate	TDBPP	126‐72‐7	Pre‐registered/Notified C&L	Restricted under Annex XVII of REACH, Commission Regulation (EC) No 552/2009, shall not be used in textile articles, such as garments, undergarments and linen, intended to come into contact with the skin 12
N,N′‐Ethylenebis(tetrabromophthalimide)	EBTEBPI	32588‐76‐4	Registered (100–1000 tonnes per annum)	–
Hexabromocyclodecane	HBCYD	25495‐98‐1	Non‐registered	–
Bis(2‐ethylhexyl) tetrabromophthalate	BEH‐TEBP	26040‐51‐7	Registered (100–1000 tonnes per annum)	November 2022: SVHC (very persistent and very bioaccumulative (vPvB) meeting the criteria set out in Article 57(e) of the Regulation (EC) No 1907/2006 (REACH)) 13 European Union proposal to list BEH‐TEBP in the Stockholm Convention on POPs. 14
2‐Ethylhexyl 2,3,4,5‐tetrabromobenzoate	EH‐TBB	183658‐27‐7	Non‐registered	–
Dibromoneopentyl glycol	DBNPG	3296‐900	Registered (1–10 tonnes per annum)	Harmonised classification as mutagenic (Cat. 1B) and carcinogenic (Cat. 1B) under Regulation (EC) No 1272/2008 15 June 2021: SVHC (carcinogen meeting the criteria set out in Article 57(a) of the Regulation (EC) No 1907/2006 (REACH)) 16
Other BFRs included in the 2012 Opinion
Decabromodiphenyl ethane	DBDPE	84852‐53‐9	Registered (10,000–100,000 tonnes per annum)	October 2025: SVHC (very persistent and very bioaccumulative (vPvB) meeting the criteria set out in Article 57(e) of the Regulation (EC) No 1907/2006 (REACH)) 17 Government of Norway proposal to list DBDPE in the Stockholm Convention on POPs. 18
1,2‐Bis(2,4,6‐tribromophenoxy)ethane	BTBPE	37853‐59‐1	Pre‐registered/Notified C&L	January 2023: SVHC (very persistent and very bioaccumulative (vPvB) meeting the criteria set out in Article 57(e) of the Regulation (EC) No 1907/2006 (REACH)) 19
Hexabromobenzene	HBB	87‐82‐1	Pre‐registered/Notified C&L	–
4‐(1,2‐Dibromoethyl)‐1,2‐dibromocyclohexane	DBE‐DBCH	3322‐93‐8	Pre‐registered/Notified C&L	–
Pentabromoethylbenzene	PBEB	85‐22‐3	Pre‐registered/Notified C&L	–
Pentabromotoluene	PBT	87‐83‐2	Pre‐registered/Notified C&L	–
1,3,5‐Tris(2,3‐dibromopropyl)‐1,3,5‐triazine‐2,4,6‐trione	TDBP‐TAZTO	52434‐90‐9	Registered (100–1000 tonnes per annum)	–
1,2,5,6‐Tetrabromocyclooctane	TBCO	3194‐57‐8	Non‐registered	–
5,6‐Dibromo‐1,10,11,12,13,13‐hexachloro‐11‐tricyclo[8.2.1.02,9]tridecene	DBHCTD	51936‐55‐1	Pre‐registered	–
1,2,3,4,7,7‐Hexachloro‐5‐(2,3,4,5‐tetrabromophenyl)‐bicyclo[2.2.1]hept‐2‐ene	HCTBPH	34571‐16‐9	Pre‐registered	–
Pentabromobenzyl acrylate	PBB‐Acr	59447‐55‐1	Registered (100–1000 tonnes per annum)	–
Tribromoneopentyl alcohol	TBNPA	1522‐92‐5	Non‐registered	Harmonised classification as carcinogen (Cat. 1B) and as suspected of causing genetic defects (muta. 2) under Regulation (EC) No 1272/2008.5 July 2021: SVHC (carcinogen and suspected of causing genetic defects meeting the criteria set out in Article 57(a) of the Regulation (EC) No 1907/2006 (REACH)) 20
2,3,5,6‐Tetrabromo‐p‐xylene	TBX	23488‐38‐2	Pre‐registered/Notified C&L	–
1,3‐Bis(2,3‐dibromopropyl)‐5‐(2‐propen‐1‐yl)‐1,3,5‐triazine‐2,4,6(1H,3H,5H)‐trione	BDBP‐TAZTO	75795‐16‐3	Non‐registered	–
1‐(2,3‐Dibromopropyl)‐3,5‐diallyl‐1,3,5‐triazine‐2,4,6(1H,3H,5H)‐trione	DBP‐TAZTO	57829‐89‐7	Non‐registered	–
Dibromostyrene	DBS	31780‐26‐4	Pre‐registered/Notified C&L	–
2‐(2‐Hydroxyethoxy)ethyl 2‐hydroxypropyl 3,4,5,6‐tetrabromophthalate	HEEHP‐TEBP	20566‐35‐2	Non‐registered (previously registered)	–
Tetradecabromo‐1,4‐diphenoxybenzene	4’‐PeBPO‐BDE208	58965‐66‐5	Pre‐registered/Notified C&L	–
Tris(tribromoneopentyl) phosphate	TTBNPP	19186‐97‐1 (134376‐16‐2)	Registered (10–100 tonnes per annum)	–
2,4,6‐Tris(2,4,6‐tribromophenoxy)‐ 1,3,5‐triazine	TTBP‐TAZ	25713‐60‐4	Registered (1000–10,000 tonnes per annum)	–
Octabromotrimethylphenyl indane	OBTMPI	1084889‐51‐91025956‐65‐3893843‐07‐7 155613‐93‐7	Non‐registered	–

Abbreviations: C&L, classification and labelling; PBT, persistent, bioaccumulative and toxic; vPv, very persistent and very bioaccumulative.

^a Information extracted from the European Chemicals Agency (ECHA) website (www.echa.europa.eu, last accessed: 11/3/2026).

^b Registered: substances/mixtures that hold a valid registration by one or more companies and can be placed on the EEA market; Pre‐registered: Substances indicated, in 2009, as being intended to be registered by at least one company in the EEA; Notified classification and labelling (C&L): substances either manufactured or imported in small amounts as substances/mixtures or in articles to the EU.

None of the BFRs considered in this Opinion is listed under the restricted substances included in Annex II of the Directive 2011/65/EU ¹⁰ of the European Parliament and of the Council of 8 June 2011 on the restriction of the use of certain hazardous substances in electrical and electronic equipment (recast).

Six out of the 27 BFRs considered in this Opinion are listed in Commission Recommendation 2014/118/EU, ¹¹ on the monitoring of traces of BFRs in food. According to this regulation, the six BFRs listed should be monitored in fish and other seafood, meat and meat products (including edible offal), animal and vegetable fats and oils, milk and dairy products, eggs and egg products and food for infants and small children. Moreover, their determination should be carried out with analytical methods with a limit of quantification of 1 ng/g wet weight (ww) or lower.

2. DATA AND METHODOLOGIES

The recent updates of the EFSA risk assessments on BFRs, including the current one on Emerging and Novel BFRs, were developed applying a structured methodological approach, which involved developing a priori the protocol or strategy of the full risk assessments and performing each step of the risk assessment in line with the protocol or strategy and documenting the process. The protocol in Annex A of this Opinion contains the method that was used for all the steps of the risk assessment process, including any subsequent refinements/changes made.

The CONTAM Panel used its previous risk assessment on Emerging and Novel BFRs in food (EFSA CONTAM Panel, 2012b) as a starting point for drafting the current Opinion.

The draft Scientific Opinion underwent a targeted consultation from 28 April to 18 May 2026. The comments received were taken into account when finalising the Scientific Opinion and are presented and addressed in Annex F.

2.1. Supporting information for the assessment

Information on physicochemical properties, production and industrial use, environmental fate and levels, analytical methods, previous assessments and legislation was gathered from the previous EFSA Opinion on Emerging and Novel BFRs in Food (EFSA CONTAM Panel, 2012b), assessments by international bodies (by checking the original websites of the relevant organisations) and from current EU legislation. Literature searches were also conducted to identify new information in reviews and other peer‐reviewed publications, also in the fields of previously reported occurrence data and exposure assessments, food processing and non‐dietary exposure. The information was summarised in a narrative way based on expert knowledge and judgement, and does not aim to be comprehensive.

Details of the literature search performed are shown in Appendix A. Search strings were run in the databases indicated. The literature searches were performed in April 2024. The outcome of the searches was saved in separate EndNote files and an automatic duplicate detection run. The selection for relevance based on title and abstract and full text was done by EFSA staff and EFSA CONTAM Working Group members.

2.2. Hazard identification and characterisation

Information relevant for the sections under hazard identification and characterisation was identified by the literature searches shown in Appendix A and assessments by international bodies. The literature searches were performed in April 2024, September 2024, May 2025 and December 2025. The outcome of the searches was saved in separate EndNote files and an automatic duplicate detection run.

The selection of the scientific papers for inclusion or exclusion was based on consideration of the extent to which the study was relevant to the assessment or on general study quality considerations (e.g. sufficient details on the methodology, performance and outcome of the study, on dosing, substance studied and route of administration and on statistical description of the results). Limitations in the information used are documented in this Scientific Opinion. Studies on levels in human samples were limited to European countries as studies from other countries might not reflect the same pattern of usage of BFRs.

The CONTAM Panel also used the OECD QSAR Toolbox v4.8 and VEGA v1.2.4 for all the BFR substances subject to this Opinion. General information about VEGA ²¹ and OECD QSAR toolbox ²² can be found at their respective websites (see footnotes). The objectives were the following:

1. to capture experimental data present in databases of the OECD QSAR Toolbox to complement the literature search performed.

2. to generate predictions on Ames test results to inform on the mutagenic potential. The CONTAM Panel noted that this was useful also for substances for which experimental data are available because the comparison between VEGA QSAR predictions and experimental data provided additional reassurance on the validity of the QSAR predictions as well as an additional line of evidence. Predictions for other more complex endpoints (both genotoxic and non‐genotoxic) were not considered in this Opinion since QSAR models did not predict these with sufficient reliability for the substances considered.

3. to provide insights on the toxicological mode of action as regards genotoxicity (OECD QSAR toolbox profiler). The results from these profilers were considered further in this Opinion to provide insight on the mode of action when there were positive experimental results or, in absence of experimental results, when there were valid positive QSAR results. They were not considered further to conclude on the endpoint in the absence of other valid lines of evidence.

4. to identify potential for read‐across by generating evidence for analogue identification and structural and mechanistic similarity, also using the QSAR toolbox profilers.

The analysis included the production of QSAR predictions using VEGA software. VEGA automatically assigns a reliability score to each prediction (see Annex B). These reliability scores were then reviewed. The CONTAM Panel considered that the predictions for negative Ames test should have at least ‘GOOD reliability’ to be further evaluated, whereas predictions for positive Ames test should have at least ‘MODERATE reliability’. A higher reliability requirement was applied to negative predictions since ruling out every possible mechanism of mutagenicity requires high confidence in the QSAR result, which can be achieved when the training or test set of the QSAR model contains substances highly similar to the target, resulting in predictions with ‘GOOD reliability’. On the other hand, for positive predictions, it may be sufficient for the model to identify substances with moderate overall similarity to the target, as long as they share the specific fragment linked to mutagenicity. This can provide the assessor with sufficient confidence to use the QSAR result for further consideration. Annex B include the full reports as produced by VEGA and QSAR Toolbox used in this Opinion.

Benchmark dose (BMD) analysis was carried out according to the EFSA Scientific Committee guidance on BMD modelling (EFSA Scientific Committee, 2022) using the Bayesian BMD Modelling web‐app (https://zenodo.org/record/7334435#.Y5osYXbMLD4) available at the EFSA R4EU platform (https://efsa.openanalytics.eu/).

As indicated in the protocol for the risk assessment (Annex A), an evaluation of the inherent uncertainties in the risk assessment was performed, and it was based on the Guidance on uncertainty analysis in scientific assessments of the EFSA Scientific Committee (EFSA Scientific Committee, 2018) and the guidance on communication of uncertainty in scientific assessments (EFSA, 2019a). The aim of the uncertainty analysis was to identify uncertainties affecting the assessment and quantify their combined impact on the main conclusions. Probabilities assessed by the uncertainty analysis were reported as % certainty for the more probable outcome, following EFSA's guidance on communication of uncertainty (EFSA, 2019a). An uncertainty analysis was only performed when a conclusion could be drawn, e.g. on hazard characterisation. Details can be found in each specific sub‐section of this assessment.

2.3. Occurrence data submitted to EFSA

2.3.1. Data collection

Following a mandate from the European Commission, a call for annual collection of chemical contaminant occurrence data in food was issued by EFSA in December 2010. Since then, data have been submitted every year by a deadline agreed with the EFSA Scientific Network on Chemical Monitoring Data collection. ²³

The data submission to EFSA follows the requirements of the EFSA Guidance on Standard Sample Description 2 (EFSA, 2013). Occurrence data are managed following the EFSA standard operational procedures (SOPs) on ‘Data collection and validation' and on ‘Data analysis of food consumption and occurrence data’.

2.3.2. Data validation and analysis

Following EFSA's Technical Report on handling of occurrence data for dietary exposure assessment (EFSA, 2021) to guarantee an appropriate quality of the data used in the exposure assessment, the initial dataset was carefully evaluated by applying several data cleaning and validation steps. Special attention was paid to the identification of duplicates and to the accuracy of different parameters, such as ‘Sampling strategy’, ‘Sampling year’, ‘Sampling country’, ‘Analytical method’, Expression of result (e.g. fat weight), ‘Reporting unit’, ‘Limit of detection/quantification' and the codification of the analysed food under FoodEx2 classification (EFSA, 2011a, 2011b, 2015).

Left‐censored data were treated using the substitution method as recommended in the ‘Principles and Methods for the Risk Assessment of Chemicals in Food’ (WHO/IPCS, 2009, updated in 2020). This is the same method as indicated in the EFSA scientific report ‘Management of left‐censored data in dietary exposure assessment of chemical substances’ (EFSA, 2010). The guidance suggests that the lower bound (LB) and upper bound (UB) approach should be used for chemicals likely to be present in the food (e.g. naturally occurring contaminants, nutrients and mycotoxins). The LB is obtained by assigning a value of zero (minimum possible value) to all samples reported as lower than the LOD (< LOD) or LOQ (< LOQ). The UB is obtained by assigning the numerical value of LOD to values reported as < LOD and LOQ to values reported as < LOQ (maximum possible value), depending on whether LOD or LOQ is reported by the laboratory.

A mean LB and UB occurrence value was then calculated at each level of the FoodEx2 classification.

Mean occurrence values for specific food categories calculated on less than five analytical results were not used in the dietary exposure assessment. However, these analytical results were included in the calculation of averages for categories at higher levels of the FoodEx2 classification if at least five samples were available at the higher level. If at FoodEx2 Level 1 there were not at least 5 samples, the food category was not included in the assessment and this exclusion was discussed in the uncertainty section.

Specific food subcategories with 100% left‐censored results were included with mean LB and UB concentration of the closest parent FoodEx2 categories for which quantified results were available if there was no reason to exclude the presence of the compound in the subcategory.

Similarly, specific food subcategories for which there were no occurrence data available were attributed with the concentration of the parent FoodEx2 category where contamination could not be excluded, e.g. the mean concentration of BTBPE for ‘Mammals meat’ was attributed to ‘Pig fresh meat’, subcategory for which there were no specific occurrence data.

2.3.2.1. Analytical methods

The analytical methods for some of the BFRs may be associated with uncertainties due to the fact that analytical standards in some cases may not be readily available. As a consequence, some studies have relied on surrogate response factors and specific recovery of the analyte was not measured (Papachlimitzou et al., 2012). This introduces some uncertainty in the reported concentrations and hinders comparison across datasets.

2.4. Food consumption data

Food consumption data from the EFSA Comprehensive European Food Consumption Database (Comprehensive Database) were used to perform the dietary exposure assessment. This database contains national data on food consumption at the individual level, which are the most complete and detailed data currently available in the EU.

The food consumption data gathered in the Comprehensive Database were collected using repeated 24 or 48‐h dietary recalls or dietary records covering 1 to 9 days per individual. Owing to the differences in the methods used for data collection, direct country‐to‐country comparisons of the exposure estimates should be avoided. Details of how the Comprehensive Database is used to assess the dietary exposure to food chemicals are published in a 2011 EFSA Guidance (EFSA, 2011b).

In addition, in 2019, EFSA developed the Raw Primary Commodity (RPC) model (EFSA, 2019b). The RPC model disaggregates the food consumption data for composite foods (foods consisting of multiple components) and RPC derivatives (single‐component foods which have been physically changed by processing) reported in the Comprehensive Database into their corresponding quantities of RPCs. As the occurrence data submitted to EFSA on the novel and emerging BFRs assessed in this Opinion covered mainly raw primary commodities the RPC database was deemed the most appropriate consumption database to be used. The RPC Consumption Database is based on the Comprehensive Database version published on the 31 March 2018, and contains 39 dietary surveys from 22 different countries, covering a total of 73,029 subjects. A chronic dietary exposure assessment is relevant in the context of the Terms of Reference. For such an assessment, surveys in which food consumption data were collected over only 1 day or subjects involved only for 1 day, are not considered appropriate. Exclusion of these surveys and subjects resulted in a total of 33 dietary surveys carried out in 18 Member States covering 61,336 individuals available for the chronic dietary exposure assessment.

Table 3 provides an overview of the population groups and countries included in the dietary exposure assessment. Surveys for special population groups, i.e. lactating women and pregnant women, one survey each, are not included in the table as they are not the focus of the assessment and of limited representativeness; however the dietary exposure estimates of these two surveys will be compared to those of the general adult population.

TABLE 3.

Population groups and countries included in the chronic dietary exposure assessment.

Population group	Age range	Countries with food consumption surveys covering more than 1 day
Infants	> 12 weeks to < 12 months	Bulgaria, Denmark, Finland, Germany, Italy
Toddlers	≥ 12 months to < 36 months	Belgium, Bulgaria, Denmark, Finland, Germany, Italy, the Netherlands, Poland, Spain
Other children	≥ 36 months to < 10 years	Austria, Belgium, Bulgaria, Czechia, Denmark, Finland, France, Germany, Greece, Italy, Latvia, the Netherlands, Spain, Sweden
Adolescents	≥ 10 years to < 18 years	Austria, Belgium, Cyprus, Czechia, Denmark, Finland, France, Germany, Italy, Latvia, the Netherlands, Spain, Sweden
Adults	≥ 18 years to < 65 years	Austria, Belgium, Czechia, Denmark, Finland, France, Germany, Hungary, Ireland, Italy, Latvia, the Netherlands, Romania, Spain, Spain, Sweden
Elderly	≥ 65 years to < 75 years	Austria, Belgium, Denmark, Finland, France, Germany, Hungary, Ireland, Italy, Netherlands, Romania, Sweden
Very elderly	≥ 75 years	Austria, Belgium, Denmark, France, Germany, Hungary, Ireland, Italy, the Netherlands, Romania, Sweden

According to the EFSA Scientific Committee Guidance on the risk assessment of substances present in food intended for infants under 16 weeks of age, the exposure assessment for these infants should be carried out separately from that for older infants, following the procedure described in the guidance (EFSA Scientific Committee, 2017). Based on this guidance, infants under 16 weeks of age should be excluded from the dietary exposure estimation of the infants age group. However, due to uncertainty in the reported individual ages of infants in the Comprehensive Database, the cut‐off age was based on a validated existing age group in this database corresponding to 12 weeks of age. Thus, food consumption data of infants between 12 and 16 weeks of age were also included in the exposure assessment. As the number of children within this age range in the database is limited, it is not expected that this will have affected the exposure estimate for infants of 16 weeks up to 12 months of age. Annex C (Table C.1) provides details on the dietary surveys included in the dietary exposure assessment.

To assess the dietary exposure of breastfed infants, a scenario considering a median age of 3 months was selected, equivalent to a body weight of about 6.1 kg, with an estimated average daily human milk consumption of 800 mL and a high consumption of 1200 mL. A mean fat content in human milk of 3.5% was considered for the estimations.

2.5. Food classification

Consumption and occurrence data were codified according to the FoodEx2 classification system (EFSA, 2011a, 2011b). Since 2018, all consumption records in the Comprehensive Database as well as all occurrence data submitted to EFSA have been codified according to the FoodEx2 classification system (EFSA, 2015). The FoodEx2 classification system consists of a large number of standardised basic food items aggregated into broader food categories in a hierarchical parent–child relationship. Additional descriptors, called facets, are used to provide additional information about the codified foods (e.g. information on food processing and packaging material).

2.6. Exposure assessment

Due to the limited occurrence data on BEH‐TEBP, EH‐TBB, PBEB, PBT, TBX (all left‐censored) and DBDPE (see Section 3.1.1), the CONTAM Panel deemed it not possible to perform an exposure assessment for these BFRs. An exposure assessment was only performed for BTBPE and HBB.

For calculating the chronic dietary exposure to the three emerging BFRs, food consumption and body weight data at the individual level were retrieved from the Comprehensive Database. Occurrence data and consumption data were linked at the relevant FoodEx2 level.

Chronic dietary exposures were calculated by combining mean occurrence values of the chemicals for food samples collected in different countries (pooled European occurrence data) with the average daily consumption for each food at the individual level in each dietary survey and age class. Consequently, individual average exposures per day and body weight were obtained for all individuals. The following formula describes the calculation made:

$${\overline{e}}_i=\frac{{\sum }_{d\in D_i}{\sum }_{f\in F}{\overline{x}}_f·c_{f,d,i}}{\left|D_i\right|·b{w}_i}$$

where:

${\overline{\mathrm{e}}}_{\mathrm{i}}$ is the average exposure of individual i

${\overline{x}}_f$ is the mean chemical concentration in each food or food group f (belonging to set of foods F _i for individual i)

$c_{f,d,i}$ is the consumed amount of food f by individual i on day d

bw _i is individual body weight of individual i

d is the survey day (belonging to the set of survey days D _i for individual i)

|D _i| represents the number of survey days of individual i

The distributions of individual exposures were then used to calculate the mean and high (95th percentile) exposure per survey and per age class. These exposure estimates were obtained using the LB and UB mean concentration of BTBPE and HBB.

All analyses were run using the SAS Statistical Software (SAS enterprise guide 8.5).

2.7. Risk characterisation

The general principles of the risk characterisation for chemicals in food as described by WHO/IPCS (2009), updated in 2020) were applied as well as the different EFSA guidance documents relevant to this step of the risk assessment (see Annex A).

2.8. Uncertainty analysis

The risk assessments in this Opinion followed the normal approach of the CONTAM Panel, including elements to address standard sources of uncertainty. Uncertainty analysis was therefore performed using the approach for a standardised assessment, described in Section 3 of EFSA's Guidance on uncertainty analysis (EFSA Scientific Committee, 2018). Sources of uncertainty affecting each assessment were tabulated and their individual impacts on the assessment were evaluated qualitatively, except where the available data were clearly insufficient to perform an assessment. The overall (combined) impacts of the non‐standard uncertainties on the conclusions of each assessment regarding genotoxicity, hazard characterisation and risk ²⁴ were then evaluated quantitatively, except where no conclusion was reached. The quantitative evaluations were performed by expert group discussion rather than formal or semi‐formal expert knowledge elicitation (section 12.6 of the Guidance) since this was found to be sufficient in each case and was practical, considering the number of substances to be assessed. The resulting probabilities are reported as % certainty for the more probable outcome for each conclusion, following EFSA's guidance on communication of uncertainty (EFSA, 2019a).

3. ASSESSMENT

3.1. Occurrence data in food

3.1.1. Occurrence data in food submitted to EFSA

The occurrence data in food that were used for the human dietary exposure assessment in the current Opinion are described in this section.

In the previous Opinion, two countries, Ireland and the UK, submitted data on three emerging BFRs, i.e. BTBPE, DBDPE and HBB, covering the years 2006–2008 (EFSA CONTAM Panel, 2012b). For BTBPE, only 11% of the analytical results were reported as > LOQ. For DBDPE and HBB, all reported analytical values were < LOQ. The CONTAM Panel concluded at that time, that based on these data a meaningful exposure assessment was not possible.

Since then, and by the 8th of September 2025, a total of 2833 analytical results were submitted to EFSA (see Table 4) for HBB (n = 2004), BTBPE (n = 630), DBDPE (n = 76), BEH‐TEBP (n = 21), EH‐TBB (n = 21), PBEB (n = 34), PBT (n = 34), TBX (n = 13) sampled between year 2010 and 2025 (Annex E).

TABLE 4.

Number of analytical results and percentage of left‐censored (LC) data submitted to EFSA by compound.

Compound	n	% LC
Covered by Commission Recommendation 2014/118/EU
BEH‐TEBP	21	100
EH‐TBB	21	100
Not covered by Commission Recommendation 2014/118/EU
BTBPE	630	78
DBDPE	76	96
HBB	2004	82
PBEB	34	100
PBT	34	100
TBX	13	100
Total	2833	82

Only the data availability for HBB and BTBPE was considered appropriate to perform a dietary exposure assessment. The limited, and left‐censored, data available for the other six BFRs were not further considered.

66 samples on ‘Fish and seafood’ were available for DBDPE. However, the only three positive samples were from 2010, thus not representative of the recent situation. In addition, insufficient data were available on other food categories. Thus, the CONTAM Panel deemed it not appropriate to perform an exposure assessment for DBDPE.

A list of validation steps was applied to the available data for HBB and BTBPE before using them to estimate dietary exposure. Data providers were contacted to clarify inconsistencies identified during the data validation. The following paragraphs describe modifications that were made to the initial data set based on the feedback received and/or expert judgement.

A number of analytical results were reported as not corrected for recovery (n = 370) and no recovery factor was provided. In this case the recovery was assumed to be 100%. When no indication was provided on the application of a correction factor (n = 1794), it was assumed that the result was reported as corrected.

Analytical results reported to be related to ‘suspect sampling’ (n = 6, all for HBB) were excluded based on the fact that sampling strategy cannot be considered randomised.

For the majority of the results (n = 1408), the reported analytical method was gas chromatography–tandem mass spectrometry (GC–MS/MS), followed by gas chromatography– high resolution mass spectrometry (GC–HRMS) (n = 187), gas chromatography (GC) coupled to triple‐quadrupole–tandem massv spectrometry (GC‐QqQ‐MS–MS) (n = 185), GC–MS (n = 90). For 758 analytical results (30%), no information on the analytical method was available. All analytical results were retained.

After the removal of the above‐mentioned six ‘suspect’ samples, 2628 analytical results were included in the final dataset for HBB (n = 1998) and BTBPE (n = 630).

The number of analytical results per year and country in this final dataset is presented in Table 5. Most results for HBB were provided by Portugal (68%) and France (28%) between 2010 and 2017. France was the main data provider for BTBPE (90%) between 2014 and 2017.

TABLE 5.

Number of analytical results per year and country in the final dataset for HBB and BTBPE.

	Country/year	2010	2011	2012	2013	2014	2015	2016	2017	2020	2024	2025	Total (n)
BTBPE	France					190	183	190	1				564
	Ireland	45											45
	Norway										20	1	21
HBB	Austria									13			13
	France					190	183	190	1				564
	Germany								1		1		2
	Ireland	45											45
	Norway										20	1	21
	Portugal	276	214	257	192	204	210						1353
Total (n)		366	214	257	192	584	576	380	3	13	41	2	2628

Table 6 shows data availability and summary statistics for each Level 1 of the FoodEx2 classification of the occurrence data submitted to EFSA on HBB and BTBPE.

TABLE 6.

Number of analytical results, % of left‐censored analytical results, number of quantified samples, LB and UB mean concentrations, maximum quantified concentrations and min LOQ (all expressed as μg/kg wet weight (ww)) for each Level 1 of the FoodEx2 classification calculated from occurrence data submitted to EFSA on HBB and BTBPE.

Compound	Foodex2_L1	N results	%LC	N quant.	Mean LB	Mean UB	Max quant.	Min ResLOQ
BTBPE	Animal and vegetable fats and oils	5	100	0	0	0.01	0	–
	Eggs and egg products	56	84	9	0.0004	0.001	0.01	0.0004
	Fish and seafood	334	69	103	0.0006	0.03	0.02	0.0003
	Meat and meat products	163	88	19	0.0004	0.001	0.01	0.0003
	Milk and dairy products	72	86	10	0.00005	0.0003	0.001	0.0002
HBB	Alcoholic beverages	14	100	0	0	10	0	10
	Animal and vegetable fats and oils	16	100	0	0	6.3	0	0.6
	Eggs and egg products	59	32	40	0.001	0.002	0.016	0.0001
	Fish and seafood	334	52	161	0.001	0.003	0.041	0.0001
	Food products for young population	34	100	0	0	10	0	10
	Fruit and fruit products	688	100	0	0	10	0	10
	Grains and grain‐based products	42	100	0	0	10	0	10
	Legumes, nuts, oilseeds and spices	24	100	0	0	10	0	10
	Meat and meat products	168	33	113	0.004	0.006	0.1	0.0001
	Milk and dairy products	77	32	52	0.002	0.004	0.1	0.0001
	Starchy roots or tubers and products thereof	98	100	0	0	10	0	10
	Sugar and similar, confectionery and water‐based sweet desserts	1	100	0			0	0.0015
	Vegetables and vegetable products	443	100	0	0	10	0	10

Quantified results were only reported in food commodities of animal origin. More specifically:

• −For HBB quantified results were found in various food of animal origin: for the ‘Fish and seafood’ category (n = 161), the quantified samples referred to various types of ‘Fish meat’ (marine, diadromous, freshwater and unspecified), and to ‘Crustaceans’ and ‘Molluscs’ samples submitted by Norway and France. For these samples, concentrations ranged from 0.00005 μg/kg in ‘Fish (meat)’ to 0.041 μg/kg in mussels, with an overall 52% left censorship. For the category of ‘Meat and meat products’, 113 samples had positive results, with HBB levels quantified in fat tissue, meat, liver and liver and marinated meat, and concentrations ranging from 0.00008 μg/kg in rabbit meat to 0.147 μg/kg in sheep fat tissue, and 33% of left‐censored data. For the ‘Milk and dairy products’ and ‘Eggs and egg products’ categories, the number of positive results was 52 and 40, respectively, with values between 0.00003 and 0.101 μg/kg in milk and from 0.0002 to 0.016 μg/kg in eggs. For these latter categories, the proportion of left‐censored data was 32%.
• −For BTBPE quantified results were found for 103 samples of ‘Fish and seafood’ (0.0001–0.020 μg/kg, with highest value measured in oysters), 19 samples of ‘Meat and meat products’ (0.00005–0.015 μg/kg in sheep fat tissue), 9 samples of ‘Eggs and egg products’ (0.0004–0.013 μg/kg in whole eggs), 10 samples of ‘Milk and dairy products’ (0.0001–0.0008 μg/kg, with highest value in goat milk). The amount of left‐censored results ranged between 69% and 88%, with the highest percentage in ‘Meat and meat products’.

For all the other food categories reported in Table 6, analytical results were 100% left‐censored.

To be noted that most samples in food of plant of origin (n = 1319) for HBB (‘Grains and grain‐based products’, ‘Vegetables and vegetable products’, ‘Starchy roots or tubers and products thereof, sugar plants’, ‘Legumes, nuts, oilseeds and spices’, ‘Fruit and fruit products’, ‘Animal and vegetable fats and oils and primary derivatives thereof’ – Olive oils, ‘Alcoholic beverages’) and the samples for ‘Food products for young population' (n = 34) were analysed with an LOQ of 10 μg/kg. Analytical results for these samples were all left‐censored and were not included in the dietary exposure assessment Uncertainty introduced by the high LOQs used for the analysis of these food categories and their subsequent exclusion from the dietary exposure assessment might have caused an underestimation of the exposure.

Annex C (Table C.2) contains the mean LB and UB occurrence values used in the dietary exposure assessment for each of the FoodEx2 food category for which a mean concentration could be estimated.

3.1.2. Occurrence data in food and dietary exposure assessments in the scientific literature since the previous Opinion

This section provides an overview of the occurrence in food samples of the BFRs covered in the current Opinion from studies reported in the scientific literature for European countries since the publication of the previous EFSA Opinion (EFSA CONTAM Panel, 2012b). The overview is not intended to be exhaustive.

Details of the studies are provided in Appendix B, including reported levels of the BFRs analysed (expressed as means or ranges), type of commodity, detection frequency, country and year of sampling.

No studies on occurrence in food from European countries were found for EBTEBPI, HBCYD, DBNPG, TDBP‐TAZTO, TBCO, TBNPA, BDBP‐TAZTO, DBP‐TAZTO, DBS, HEEHP‐TEBP, 4’‐PeBPO‐BDE208, TTBNPP or TTBP‐TAZ.

For TDBPP, DBHCTD, HCTBPH and OBTMPI two studies were identified reporting levels below the LOD (Fernandes et al., 2016; Castro et al., 2020, see Appendix B).

The studies identified concerned BEH‐TEBP, EH‐TBB, DBDPE, BTBPE, HBB, DBE‐DBCH, PBEB, PBT, PBB‐Acr and TBX. Many of the studies failed to measure the BFRs since they were present at concentrations below the method LOD. The following paragraphs summarise those studies where compounds were measured above LOD and reported; more details including the LODs for the individual studies can be found in Appendix B.

Data from the scientific literature from European countries for BTBPE and HBB was broadly in line with the data submitted to EFSA (see Section 3.1.1).

BEH‐TEBP

For BEH‐TEBP, no studies in the scientific literature were reported in the previous Opinion. Since then, a few studies published in peer‐reviewed journals have been identified:

• Xu et al. (2017) reported analysis of flame retardants, including BEH‐TEBP, in 24 h duplicate diet samples collected 2013–2014 from a Norwegian cohort (n = 61), samples. BEH‐TEBP was detected in 44% of the samples with a mean concentration of 0.012 ± 0.023 ng/g ww (range from < 0.03 (LOQ) to 0.14 ng/g ww). Several studies reported measurements of BEH‐TEBP in individual food samples:

• In Germany, BEH‐TEBP was measured in blue mussels and eelpout filet composite samples, collected in 2015 from three coastal sites in the North and Baltic Seas (Fliedner et al., 2020). BEH‐TEBP ranged between < 0.0384–0.0559 ng/g ww in blue mussels and < 0.0404–0.0728 ng/g ww in eelpout filet.

In Belgium, Poma et al. (2018) reported concentrations of BFRs in individual food items (1289 samples) collected during 2015–2016. BEH‐TEBP was only found in two samples from the vegetable oil category (at 0.295 ng/g ww in mixed oils and 0.350 ng/g ww in olive oil), and in one sample from the Meat and meat products category (duck foie gras; 0.202 ng/g ww, LOD: 0.200 ng/g ww), however, the authors report possible analytical false positive results.

In the UK, Tao et al. (2017) studied the levels of BEH‐TEBP in samples of 14 different food groups collected in 2015. BEH‐TEBP was detected in 63% of the samples with mean concentrations ranging between 0.20 and 0.57 ng/g lipid in meat samples (beef, lamb, pork and chicken), < 0.10–1.1 ng/g lipid in fish samples (salmon, mackerel, tuna and trout), while mean concentration was reported as 1.8 ng/g lipid in eggs and 0.22 ng/g lipid in cheese samples.

Ma et al. (2023) also studied the levels in 108 individual food samples (covering 15 food categories) collected in the UK between 2020 and 2021. BEH‐TEBP was reported as one of the most abundant ‘emerging’ BFRs with 61% detection frequency and a mean concentration of 0.360 ng/g ww (or 11 ng/g lipid) (range: < 0.037–4.700 ng/g ww or < 0.33–65 ng/g lipid). The authors reported an increase in the levels of BEH‐TEBP between 2015 (Tao et al., 2017) and 2020–2021 in the different food categories, which was calculated at 1100%, 3000%, 140% and 11% in meat, fish, cheese and chicken eggs, respectively.

BEH‐TEBP concentrations reported in the Ma et al. (2023) study were found comparable to the concentrations reported in foodstuffs collected as part of the 6th total diet study (TDS) in China (from 24 provinces) < 0.015–3.009 ng/g ww (Niu et al., 2021) and from several e‐waste recycling sites in China (Wenling and Luqiao districts located within Taizhou City, one of the largest e‐waste recycling areas in Eastern China), i.e. 0.81–16.3 ng/g lw as reported by Labunska et al. (2015) and < 0.05–0.96 ng/g ww in home‐produced eggs reported by Zheng et al. (2016). Moreover, the concentration of BEH‐TEBP in this study were generally one order of magnitude higher than the concentrations reported in foodstuffs collected from Belgium (Poma et al., 2018), Tanzania (range: < LOD–9.8 ng/g lipid in eggs of free‐ranging chicken; Polder et al., 2016) and China, non‐e‐waste recycling areas (range: 1.11–15 ng/g lipid; Shi et al., 2016).

EH‐TBB

For EH‐TBB, no studies in the scientific literature were reported in the previous Opinion. Since then, a few studies published in peer‐reviewed journals have been identified:

• In France, Vénisseau et al. (2018) reported concentrations of BFRs in food samples in 577 food samples comprising six different food categories (fish, crustacea and molluscs, milk, sheep livers, meat and eggs). EH‐TBB was measured in the samples and low detection frequencies and concentrations were found (in 10% of fish samples: range < LOD–10.0 pg/g ww; in 18% of crustacea and molluscs: < LOD–8.0 pg/g ww, in 4% of milk < LOD–2.92 pg/g ww, in 0% of sheep livers, in 9% of meat samples: < LOD–11.79 pg/g ww and in 5% of eggs: < LOD–1.26 pg/g ww).

• Two studies retrieved from the UK reported high concentrations of EH‐TBB. Tao et al. (2017) reported concentrations of EH‐TBB in samples of 14 different food groups collected in 2015. EH‐TBB was detected in 77% of the samples analysed and mean concentrations reported ranged between 0.19 and 1.4 ng/g lipid in meat samples (beef, lamb, pork and chicken), 0.22–0.43 ng/g lipid in fish samples (salmon, mackerel, tuna and trout), while mean concentration was reported as 0.10 ng/g lipid in eggs and 0.11 ng/g lipid in cheese samples. Ma et al. (2023) reported concentrations of BFRs in UK food items collected between December 2020 and October 2021. A total of 108 individual food samples were analysed (covering 15 food categories). EH‐TBB was detected in 11% of the samples analysed with mean concentration reported at 21 pg/g ww (1.5 ng/g lipid) and range of < 4.4–290 pg/g ww (0.051–35 ng/g lipid). Overall, concentrations of EH‐TBB were found similar between 2015 (Tao et al., 2017) and 2020–2021 in the different food categories in the UK, however looking at the food categories separately, a 26% increase was observed in meat samples and a 93%, 88% and 22% decrease in cheese, eggs and fish, respectively.

• Concentrations of EH‐TBB measured in UK foodstuffs were higher in comparison to concentrations reported in France (Vénisseau et al., 2018). In contrast to BEH‐TEBP, concentrations of EH‐TBB measured in UK foodstuffs were comparable to those reported in Chinese food samples (< 30–222 pg/g ww; Niu et al., 2021), and 1–2 orders of magnitude lower than the concentrations reported (range of 3.09–62.2 ng/g lipid) in foodstuffs collected from an e‐waste recycling site in China (Labunska et al., 2015).

DBDPE

For DBDPE, one study in the scientific literature was reported in the previous Opinion (Tlustos et al., 2010). Since then, a few studies published in peer‐reviewed journals have been identified:

• Santín et al. (2013) reported concentrations of up to 130 ng/kg lipid in various species of river fish from Spain.

• Ma et al. (2023) found mean concentrations of DBDPE of 18 ng/kg ww in meat, and 45 ng/kg ww in fish and prawns from the UK. An average dietary exposure estimate of 0.066–0.24 ng/kg bw per day, and a high‐end estimate of 0.39–1.5 ng/kg bw per day was made by the authors (see Section 3.3.1.2).

• Munschy et al. (2015) found DBDPE in samples of mussels and oysters at concentrations up to 20.1 ng/kg ww in France.

• García‐Bermejo et al. (2016) reported mean concentrations of DBDPE of 0.0036 ng/kg ww in cow's milk, 24 ng/kg ww in eggs, 0.041 ng/kg ww in biscuits, 8.2 ng/kg ww in fish and 13 ng/kg ww in vegetable oil dietary supplements from Spain.

• Mohr et al. (2014) found concentrations of up to 0.87 ng/kg ww DBDPE in Spanish honey.

BTBPE

For BTBPE, only one study in the scientific literature was reported in the previous Opinion (Tlustos et al., 2010). Since then, a few studies published in peer‐reviewed journals have been identified:

• In Sweden, Sahlström et al. (2015) measured BFRs, including BTBPE, in market basket samples comprising five different food categories (fish, meat, vegetable oils, dairy products and eggs, n = 4 homogenates for each food category) collected in 2010. BTBPE was detected at low concentrations in the range 1.1–3.6 pg/g ww in fish and dairy products, and in eggs with an average of 3.9 pg/g ww.

• Ma et al. (2023) found BTBPE in samples from the UK at mean concentrations of 190 ng/kg ww in meat, 200 ng/kg ww in fish and prawns, 97 ng/kg ww in cheese and 2300 ng/kg ww in chicken eggs. An average dietary exposure estimate of 1.2–4.4 ng/kg bw per day, and a high‐end estimate of 7.1–27 ng/kg bw per day was made by the authors (see Section 3.3.2.2).

• In France, Vénisseau et al. (2018) found a mean concentration of 0.050 ng/kg ww BTBPE in crustaceous/molluscs where it was detected in 50% of samples. Tao et al. (2017) found concentrations of up to 2.4 ng/g (lipid weight) in meat, up to 0.35 ng/g lipid weight in liver, 0.78 ng/g lipid in fish, 0.18 ng/g in eggs and up to 0.20 ng/g lipid in cheese.

• Poma et al. (2014) reported concentrations of up to 0.005 ng/kg ww BTBPE in shad and whitefish (lake fish species) collected in Italy.

García‐Bermejo et al. (2016) reported mean concentrations of BTBPE of 0.35 ng/kg ww in cow milk, 0.039 ng/kg ww in soy products, 0.43 ng/kg ww eggs, 0.60 ng/kg ww in biscuits and 17 ng/kg ww in fish oil dietary supplements from Spain.

Mohr et al. (2014) found BTBPE in honey samples at mean concentrations up to 0.06 ng/kg ww for Spanish honey.

In a study on lake fish species, Babut et al. (2021) found BTBPE at concentrations up to 10 ng/kg ww in roach and burbot from Switzerland. In a study on mussels and oysters, Munschy et al. (2015) found BTBPE at concentrations up to 37.3 ng/kg ww.

HBB

For HBB, one study in the scientific literature was reported in the previous Opinion (Tlustos et al., 2010). Since then, a few studies published in peer‐reviewed journals have been identified:

• HBB was found by Zacs et al. (2021) where concentrations up to 9.82 ng/kg ww were found in meat, up to 5.71 ng/kg ww in eggs, up to 21.8 ng/kg ww in milk and dairy products, up to 17.8 ng/kg ww in fish and fish products and up to 19.5 ng/kg ww in products of plant origin from Latvia.

• Vénisseau et al. (2018) reported mean concentrations of 0.363 ng/kg ww in fish (detected in 65% of samples), 0.257 ng/kg ww in Crustaceous/molluscs (57% samples), 0.321 ng/kg ww in milk (72% samples), 0.617 ng/kg ww in eggs (68% samples), 0.633 ng/kg ww in sheep liver (57% samples) and 0.701 ng/kg ww in meat (85% samples) from France.

• Sahlström et al. (2015) reported mean concentrations of 23 ng/kg ww in fish (detected in 25% samples) and 0.95–3.1 pg/g ww in meat (detected in 25% samples) from Sweden.

• Aznar‐Alemany et al. (2017) detected HBB in 43% of fish samples available in European countries with a mean concentration of 0.04 ng/g ww in cod ( Gadus morhua ), 0.13 ng/g ww in mackerel (Scomber scombrus), 0.04 ng/g ww in monkfish (Lophius piscatorius), 0.04 ng/g ww in mussels ( Mytilus edulis ), 0.08 ng/g ww in plaice ( Pleuronectes platessa ), 1.6 ng/g ww in seabream ( Sparus aurata ) and 0.07 ng/g ww in tuna (Thunnus species).

• Trabalón et al. (2017) found up to 0.2 ng/g ww in a variety of 10 different fish species.

DBE‐DBCH

For DBE‐DBCH no information on occurrence data in the scientific literature was reported in the previous Opinion. Since then, several studies published in peer‐reviewed journals have been identified:

• DBE‐DBCH was included in a study by Zacs et al. (2021) where it was found in samples of meat at concentrations up to 34.6 ng/kg ww, 25.0 ng/kg ww in eggs, 60.9 ng/kg ww in milk and dairy products, 132 ng/kg ww in fish and fish products and up to 28.6 ng/kg ww in food of plant origin from Latvia.

• Tao et al. (2017) measured α‐DBE‐DBCH and β‐DBE‐DBCH separately in a variety of food products sampled in the UK. α‐DBE‐DBCH was found at concentrations of up to 0.95 ng/g lipid in meat, 1.6 ng/g lipid in liver samples of beef, lamb, chicken and pork, 1.3 ng/g lipid in fish, 0.42 ng/g lipid in eggs and 0.44 ng/g lipid in cheese. β‐DBE‐DBCH was found at concentrations of up to 7.6 ng/g lipid in meat, 85 ng/g lipid in liver, 39 ng/g lipid in fish, 3.0 ng/g lipid in eggs and 0.99 ng/g lipid in cheese.

• Sahlström et al. (2015) found 97 ng/kg ww α‐DBE‐DBCH and 17 ng/kg ww β‐DBE‐DBCH in fish samples in Sweden and made daily intake estimates for α‐DBE‐DBCH of 3.5 ng/day for mothers and 1.7 ng/day for toddlers based on fish intake only. For β‐DBE‐DBCH the intake estimate was 0.6 ng/day for mothers and 0.3 ng/day for toddlers again based on fish consumption only.

PBEB

For PBEB no information on occurrence data in the scientific literature was reported in the previous Opinion. Since then, several studies published in peer‐reviewed journals have been identified:

• Zacs et al. (2021) measured PBEB in samples from Latvia at concentrations up to 5.66 ng/kg ww in meat, 1.85 ng/kg ww in eggs, 2.9 ng/kg ww in milk and dairy products, 3.13 ng/kg ww in fish and fish products and 19.0 ng/kg ww in foods of plant origin.

• Sahlström et al. (2015) found PBEB in samples from Sweden at mean concentration in fish of 8.7 ng/kg ww and in meat of 0.60–2.0 ng/kg ww.

• Sühring et al. (2013) found PBEB in eels collected in Germany at concentrations of up to 100 ng/kg ww.

PBT

For PBT no information on occurrence data in the scientific literature was reported in the previous Opinion. Since then, several studies published in peer‐reviewed journals have been identified:

• Vénisseau et al. (2018) reported in samples from France mean concentrations of PBT in fish of 0.174 ng/kg ww, in milk of 0.139 ng/kg ww, in eggs of 0.236 ng/kg ww, in sheep liver of 0.214 ng/kg ww and in meat of 0.165 ng/kg ww.

• Sahlström et al. (2015) found PBT in Swedish fish samples with a mean concentration of 1.8 ng/kg ww.

• Fliedner et al. (2020) reported up to 67 ng/kg ww PBT in samples of blue mussels and eelpouts fillets from Germany.

DBHCTD

For DBHCTD no information on occurrence data in the scientific literature was reported in the previous Opinion, and no new data where this compound was found above LOD has been identified: DBHCTD was included in the study by Fernandes et al. (2016) but nothing was measured above the LOD.

HCTBPH

For HCTBPH no information on occurrence data in the scientific literature was reported in the previous Opinion, and no new data where this compound was found above LOD has been identified: HCTBPH was included in the study by Fernandes et al. (2016) but nothing was measured above the LOD.

PBB‐Acr

For PBB‐Acr no information on occurrence data in the scientific literature was reported in the previous Opinion. This compound was included in the study by Fernandes et al. (2016) but nothing was measured above the LOD.

TBX

For TBX no information on occurrence data in the scientific literature was reported in the previous Opinion. Since then, TBX was reported to be found in Crustaceous/molluscs from France (Vénisseau et al., 2018) at mean concentrations of 0.123 ng/kg ww (73% of samples) and in fish samples in Sweden (Sahlström et al., 2015) with a mean concentration of 20 ng/kg ww (100% samples).

OBTMPI

For OBTMPI no information on occurrence data in the scientific literature was reported in the previous Opinion. This compound was included in the study by Fernandes et al. (2016) but nothing was measured above the LOD.

3.1.3. Food processing and food contact materials

No data were identified in the scientific literature about the effects of cooking and processing on emerging and novel BFRs in food. However, the CONTAM Panel considers it reasonable to assume that lipophilic BFRs will behave in a similar way to other lipophilic POPs, i.e. that (i) they are stable to heat at normal cooking temperatures and (ii) will associate with lipids, which means that amounts in food may be reduced if fats released from the food during cooking or processing are discarded. On the other hand, if moisture is lost during the heating of foods but no fat is discarded, then the total amount will remain unchanged but the concentrations may increase due to weight loss.

Food contact materials such as black plastic kitchen utensils made using recycled plastics have been shown to be a source of BFRs. For example, Liu et al. (2024) investigated consumer products purchased in the USA, including food serving items and kitchen utensils from recycled plastics. Approximately 15% of the tested food contact materials contained BFRs. The authors hypothesised that contamination of the recycled plastics with plastics from electronic devices etc. may be responsible for their findings, and therefore migration from such items may be a source of additional exposure to BFRs. There are only limited specific references on the presence of the BFRs assessed in this Opinion in food contact materials including kitchen utensils, such as for BEH‐TEBP (see Section 3.2.4.3). The potential additional exposure from kitchen utensils might have caused an underestimation of the dietary exposure. However, this contribution is likely to be minor and within the range of the large uncertainty due to the small amount of available occurrence data in food.

3.2. BFRs included in the 2012 Opinion and covered in Commission Recommendation 2014/118/EU

3.2.1. Tris(2,3‐dibromopropyl) phosphate (TDBPP) [CAS number 126‐72‐7]

3.2.1.1. Production, use, analysis and environmental fate

3.2.1.1.1. Production and use

WHO/IPCS (1995) reported that TDBPP was first produced in 1950. In 1975, the commercial production of TDBPP in the USA was reported to be 4100–5400 tonnes (US‐EPA, 1976). Prior to 1977, 4500 tonnes were produced annually in the US, but by 1986, production had ceased in the US. In Japan, production in 1976 and 1977 was estimated to have been 100 and 300 tonnes per year, respectively, but had ceased by 1995 (EFSA CONTAM Panel, 2012b).

In addition to its use as a flame retardant for cellulose and triacetate and polyester fabrics, TDBPP has also been used in urethane foam and acrylic carpets and sheets, polyvinyl‐ and phenolic resins, polystyrene foam, paints, lacquers, paper, latexes and cured unsaturated polyesters products. Rigid foams containing TDBPP were used in insulation, furniture, automobile interior parts and water flotation devices. Around two thirds of the TDBPP produced in the USA in the 1970s was used in fabrics used for children's clothing which could contain about 5–10% by weight of the flame retardant (US‐EPA, 1976; Kirk‐Othmer, 1978–1984, as cited in EFSA CONTAM Panel, 2012b).

Use and exposure has greatly decreased after a ruling by the US Consumer Product Safety Commission in April 1977 when TDBPP was removed from the market due to concerns about its toxicity, but there is evidence that TDBPP was used well after that decade. ²⁵

3.2.1.1.2. Analysis

Analytical methods for TDBPP usually consist of some kind of extraction followed by clean‐up using techniques such as gel permeation chromatography (GPC) or sold phase extraction (SPE), followed by instrumental analysis using typically either GC–MS or liquid chromatography–mass spectrometry (LC–MS).

Most of the scientific literature on TDBPP is focused on textiles or environmental samples (water, sediments, etc), but examples of the methods that were reported over the last decade that were used for food or wildlife are summarised below.

Castro et al. (2020) used matrix solid‐phase dispersion (MSPD), which integrated sample extraction and clean‐up in a single step, followed by liquid chromatography–tandem mass spectrometry (LC–MS/MS) for the simultaneous determination of 18 organophosphorus flame retardants and/or plasticisers including TDBPP in marine mussel (Mytilus edulis and Mytilus galloprovincialis) samples. For TDBPP the instrumental LOD (iLOD) was 16 ng/mL which was not as good as most of the other compounds because ionisation was less efficient. This equated to a LOD of 4 ng/g dry weight in the mussel samples.

In a study on glaucous gull from the Canadian Arctic, Verreault et al. (2018) measured a range of BFRs including TDBPP. Samples were homogenised with diatomaceous earth, extracted with dichloromethane:n‐hexane (50:50) using an accelerated solvent extractor followed by SPE using cartridges containing Isolute NH₂ aminopropyl silica gel. Instrumental analysis was by LC–MS/MS, and the method had a LOD of 0.06 ng/g dry weight.

3.2.1.1.3. Environmental fate and occurrence in the environment

TDBPP has a reported log Koc value of 3.29 which accounts for its strong potential for adsorption to soil. This, along with a measured water solubility of around 8.0 mg/L, means that TDBPP is expected to leach only slowly into groundwater (EFSA CONTAM Panel, 2012b). No information on the environmental fate of TDBPP could be found, but there were several reports in the literature on concentrations in soil, sediment, water and dust.

There is very little reported on concentrations of TDBPP in environmental samples, especially from European Countries. Where studies have included TDBPP, it is often reported at very low levels, typically < LOD.

Occurrence in the environment

Soil

Zhang et al. (2021) reported on the spatial occurrence and composition profile of organophosphate esters in farmland soils from different regions of China. Whilst concentrations in soil were measured for 11 out of 12 compounds measured, TDBPP, with a LOD of 0.020 ng/g, was not found. This was consistent with earlier studies, in which TDBPP had a low detection frequency (< 20%) in the dust from rural areas of Nanjing, China.

Sediment

TDBPP was not detected (LOD = 0.03 ng/g) in a study on the optimisation and development of analytical methods for the determination of new BFRs in sediments and suspended particulate matter that included analysis of samples from the Western Scheldt estuary in the Netherlands (Brandsma et al., 2015).

TDBPP was detected in about one third of sediment samples tested from the Jiulong River Estuary and adjacent Western Taiwan Strait, at concentrations up to 0.926 ng/g in a study by Cai et al. (2022).

Dust

There are no studies reporting concentrations of TDBPP in samples of dust from Europe.

In a study on flame retardants in house dust samples from California, TDBPP was found (LOD = 20 ng/g) in 62% of samples taken in 2006 ranging from less than 20 ng/g to 8900 ng/g with a median of 35 ng/g. In samples dating from 2011, TDBPP was found in 38% ranging from less than 20 to 310 ng/g with the median being < LOD (Dodson et al., 2012).

Concentrations were much lower in a study by Li et al. (2020) reporting concentrations in dust from urban and rural domestic houses and college dormitories in Nanjing, China, where the highest concentration was just over 2 ng/g.

Water

There are a few reports of TDBPP in water, including a global systematic review by Vasseghian et al. (2022). It reported concentrations (μg/L) of different types of organophosphates in surface waters of China including TDBPP (4 samples) from 1.7 × 10⁻⁶ to 0.0016 with an average of 0.000477 and standard deviation of 0.000484.

Wildlife

Verreault et al. (2018) found TDBPP in glaucous gull from the Canadian Arctic in 14% of samples at concentrations up to 0.57 ng/g (w/w).

Xie et al. (2024) found TDBPP in liver samples of Indo‐Pacific finless porpoises from the northern South China Sea, China, taken between 2007 and 2020 at a mean concentration of 9.9 ng/g ww ranging from 1.6 to 25.8 ng/g with an LOD of 0.5 ng/g and a detection frequence of 70%.

3.2.1.2. Hazard identification and characterisation

3.2.1.2.1. Toxicokinetics

In the previous Opinion (EFSA CONTAM Panel, 2012b), there were no data on the absorption of TDBPP.

Regarding distribution, and based on the study by Nomeir and Matthews (1983), the CONTAM Panel noted that after oral administration of 1.38 mg/kg bw (¹⁴C‐TDBPP) to male Sprague–Dawley rats, radioactivity was first detected in lung, liver and kidney. Radioactivity was also measured in other tissues (blood, muscle, skin and fat). The in vitro and in vivo data on metabolism showed transformation of TDBPP into BBPP (bis(2,3‐dibromopropyl) phosphate, a major metabolite). Five other metabolites were identified in urine and bile. The clearance half‐life of radioactivity ranged from 2.5 days (blood, muscle, skin, lung and fat) to 3.8 days (liver and kidney). 24 h after exposure, the radioactivity was excreted in the urine (24% of the radioactivity) and faeces (11.5%).

In a study by Minegishi et al. (1988) ²⁶ , the authors performed studies on the absorption, distribution and excretion of flame retardants, including TDBPP, with a single administration by gavage of ¹⁴C‐labelled (50 μmol/kg) in male Wistar rats. The authors found that for TDBPP, the recovered radioactivity on day 7 was 50% in urine, 33% in faeces, 13% in expired air (as ¹⁴CO₂) and 3.7% in the body. The measured biological half‐life in blood and different tissues was between 36 and 72 h.

Since the previous Opinion, no new studies have been identified.

In summary, these data show that TDBPP is highly absorbed (> 50%), highly metabolised, distributed and accumulated in different tissues, and is eliminated mainly via urine.

3.2.1.2.2. Levels in humans

No studies were identified in the previous Opinion, nor in the current update.

3.2.1.2.3. Toxicity in experimental animals

Studies considered in the previous EFSA assessment

The previous Opinion (EFSA CONTAM Panel, 2012b) reported that nephrotoxicity (pyknosis of the renal tubular epithelial cells on day 1, necrosis on day 2, regeneration from day 3 and large nuclei formation from day 4) was observed in male Wistar rats exposed to a single oral dose of TDBPP (reported purity 98%) at 200 mg/kg bw ²⁷ and observed till day 10 (Fukuoka et al., 1988).

In a 4‐week dietary study, male weanling rats received 0, 100 or 1000 mg/kg TDBPP in the diet, equivalent to 0, 10 and 100 mg/kg bw per day, respectively (Kerst, 1974, as cited in NICNAS, 2005).

The CONTAM Panel identified a NOAEL of 10 mg/kg bw per day, based on significant decrease in body weight gain and food consumption and decreases in the weight of heart, liver, spleen, kidney and testes observed at 100 mg/kg bw per day. However, no histopathological changes were noted.

In a developmental toxicity study, pregnant Sprague–Dawley rats were exposed by gavage to TDBPP (reported purity 99.7%) at 0, 5, 25 or 125 mg/kg bw per day on GD6–15 and sacrificed on GD20 (Seabaugh et al., 1981, as cited in NICNAS, 2005). The CONTAM Panel identified a NOAEL for maternal effects of 25 mg/kg bw per day based on decreased body weight gain at 125 mg/kg bw per day and the NOAEL for developmental effects was 125 mg/kg bw per day, the highest dose tested.

Regarding genotoxicity, the previous Opinion reported that TDBPP is a bacterial mutagen both in the presence and absence of metabolic activation (with a higher activity in the presence of metabolic activation), induced gene mutations and chromosomal aberrations in mammalian test systems in vitro and micronuclei (after ip exposure) and chromosomal aberrations (after iv exposure) in animals in vivo. It induced also DNA damage in the liver, kidneys and small intestine and to a lesser extent in the brain, lungs, testes, spleen and large intestine of rats exposed to a single ip injection of TDBPP. In transgenic rodent mutation assays, exposure by gavage of TDBPP was shown to induce mutations in the kidney and liver, but not in the stomach in lacI transgenic male B6C3F1 mice (Big Blue®) and increases in lacI mutations was also observed in the cortex and outer and inner medullas of the kidney in lacI (Big Blue) transgenic F344 rats (see EFSA CONTAM Panel, 2012b).

Carcinogenic and toxic effects of TDBPP (administered as the technical product Firemaster LV‐T23P, purity: not reported (NR)) to the kidney have been investigated in rats (Reznik et al., 1979, as cited in NICNAS, 2005). F344 rats were fed diet containing 50 or 100 mg/kg TDBPP (corresponding to 2 and 4 mg/kg bw per day in males and 2.5 and 5 mg/kg bw per day in females) for 103 weeks followed by 1–2 weeks observation period prior to sacrifice. Tubular‐cell adenoma of the kidney was noted in TDBPP treated males and females. Tubular‐cell carcinoma was observed at the top dose in males. The only significant non‐neoplastic findings in kidneys were small foci of dilated and hyperplastic or dysplastic proximal convoluted tubules. These changes were seen in both males and females at the lowest dose.

The carcinogenic potential of TDBPP (administered as the technical product Firemaster LV‐T23P, purity: NR) has been tested in mice (Reznik et al., 1979, as cited in NICNAS, 2005; US‐NCI, 1978, as cited in WHO/IPCS, 1995). B6C3F1 mice were fed diet containing 500 or 1000 mg/kg TDBPP (corresponding to 60 and 120 mg/kg bw per day in males and 65 and 130 mg/kg bw per day in females) for 103 weeks followed by 1–2 weeks observation period prior to sacrifice. Tubular‐cell adenomas of the kidney were seen in TDBPP treated males and females. Tubular‐cell carcinomas were observed at the top dose in males. The only significant non‐neoplastic findings in kidneys were small foci of dilated and hyperplastic or dysplastic proximal convoluted tubules. These changes were seen in both males and females at the lowest dose. TDBPP increased the incidence of squamous‐cell papilloma and carcinoma of the forestomach and of adenoma and carcinoma of the lungs in both males and females. Moreover, in treated females there was an increased incidence of liver cell adenoma and carcinoma.

Studies published since the previous EFSA assessment

No new studies were identified since the previous Opinion. However, more details on the carcinogenicity studies of TDBPP were provided in the NTP Report (2014) (see Appendix C, Table C.1). Previously the results were only available from secondary sources.

In rats, there was no significant effect on survival. Increased incidence of tubular‐cell adenoma was observed in males and females (see Table 7). For males there was a significant positive association between the incidence of kidney tubular‐cell adenocarcinoma and TDBPP at the highest dose. Non‐neoplastic lesions were seen in renal tubules in high‐dose males and females, a few tubular cells were slightly enlarged and showed nuclear dysplasia consisting of nuclear enlargement, chromatin clumping and parachromatin clearing. A clear effect was observed at the lowest dose tested.

TABLE 7.

Summary of tumour and non‐neoplastic lesions in rats in the NTP (2014) study.

Firemaster LV‐T23P	0 mg/kg diet	50 mg/kg diet	100 mg/kg diet
Males	0 mg/kg bw per day	2 mg/kg bw per day	4 mg/kg bw per day
Kidney
Tubular‐cell adenoma	0/53 (0%)	26/54 (48%)	26/54 (48%)
Tubular‐cell adenocarcinoma	0/53 (0%)	0/54 (0%)	3/54 (6%)
Tubular‐cell adenoma or adenocarcinoma	0/53 (0%)	26/54 (48%)	29/54 (54%)
Non‐neoplastic lesions in tubules	0/53 (0%)	0/54 (0%)	6/54 (11%)
Females	0 mg/kg bw per day	2.5 mg/kg bw per day	5 mg/kg bw per day
Kidney
Tubular‐cell adenoma	0/52 (0%)	4/54 (7%)	10/54 (19%)
Non‐neoplastic lesions in tubules	0/52 (0%)	0/54 (0%)	35/54 (65%)

In mice, there was no significant effect on survival. Increased incidence of tubular‐cell adenoma in males and females was observed as well as renal tubular‐cell adenocarcinoma in males but not in females (see Table 8). Variable amounts of haemorrhage, necrosis and calcification were associated with renal tubular‐cell carcinoma. Renal tubular dysplasia was also observed in males and in females. Increased incidence of squamous‐cell papilloma of the forestomach was observed in males and females. In addition, carcinoma of the forestomach was observed in high‐dose males and both doses in females. Chronic inflammation was frequently found in the submucosa below the papillary lesion. The squamous‐cell carcinoma of the forestomach also often showed superficial hyperkeratinisation. A dose‐related increase in hepatocellular adenoma or carcinoma was noted in females. In males, no such dose‐related increase in hepatocellular neoplasms was apparent. Finally, bronchiolar/alveolar adenoma and carcinoma were seen in males and in females. A clear effect was observed at the lowest dose tested.

TABLE 8.

Summary of tumour and non‐neoplastic lesions in mice in the NTP (2014) study.

Firemaster LV‐T23P	0 mg/kg diet	500 mg/kg diet	1000 mg/kg diet
Males	0 mg/kg bw per day	60 mg/kg bw per day	120 mg/kg bw per day
Kidney
Tubular‐cell adenoma	0/54 (0%)	3/50 (6%)	9/49 (19%)
Tubular‐cell adenocarcinoma	0/54 (0%)	1/50 (2%)	5/49 (10%)
Tubular‐cell adenoma or adenocarcinoma	0/54 (0%)	4/50 (8%)	14/49 (29%)
Tubular dysplasia	0/54 (0%)	37/50 (74%)	30/49 (61%)
Forestomach
Squamous‐cell papilloma	0/51 (0%)	10/47 (21%)	11/48 (23%)
Squamous‐cell carcinoma	0/51 (0%)	0/47 (0%)	2/48 (4%)
Squamous‐cell papilloma or carcinoma	0/51 (0%)	10/47 (21%)	13/48 (27%)
Lung
Bronchiolar/alveolar adenoma	6/54 (11%)	11/44 (23%)	12/50 (24%)
Bronchiolar/alveolar carcinoma	6/54 (11%)	8/44 (18%)	13/50 (26%)
Bronchiolar/alveolar adenoma or carcinoma	12/54 (22%)	18/44 (41%)	25/50 (50%)
Females	0 mg/kg bw per day	65 mg/kg bw per day	130 mg/kg bw per day
Kidney
Tubular‐cell adenoma	0/55 (0%)	2/50 (4%)	2/46 (4%)
Tubular dysplasia	0/55 (0%)	1/50 (2%)	12/46 (26%)
Forestomach
Squamous‐cell papilloma	2/53 (4%)	10/48 (21%)	18/44 (41%)
Squamous‐cell carcinoma	0/53 (0%)	4/48 (8%)	4/44 (9%)
Squamous‐cell papilloma or carcinoma	2/53 (4%)	14/48 (29%)	22/44 (50%)
Liver
Hepatocellular adenoma	4/54 (7%)	11/50 (22%)	15/49 (31%)
Hepatocellular carcinoma	7/54 (13%)	12/50 (24%)	20/49 (41%)
Hepatocellular adenoma or carcinoma	11/54 (20%)	23/50 (46%)	35/49 (71%)
Lung
Bronchiolar/alveolar adenoma	3/55 (5%)	8/50 (16%)	14/50 (28%)
Bronchiolar/alveolar carcinoma	1/55 (2%)	1/50 (2%)	3/50 (6%)
Bronchiolar/alveolar adenoma or carcinoma	4/55 (7%)	9/50 (18%)	17/50 (34%)

Summary of toxicity in experimental animals

TDBPP is nephrotoxic. It is genotoxic, and carcinogenic in rats (in kidney) and mice (in kidney, forestomach, liver and lung).

3.2.1.2.4. Computational analysis

The QSAR Toolbox identified experimental data for the Ames test on TDBPP as retrieved by the previous Opinion (see above and Annex B).

VEGA models predicted positive Ames test results for TDBPP with ‘GOOD reliability’ (three models). The Panel critically reviewed the predictions and considered them valid mainly due to the presence of relevant analogues known by the model with experimental and predicted values concordant with the prediction for the target compound (see Annex B). The CONTAM Panel noted that this positive prediction is consistent with the available experimental data.

The QSAR Toolbox profilers provided the following information on the structural alerts and predicted mode of toxicity: according to the ‘DNA binding by OECD’ and ‘DNA alerts for Ames, CA and MNT by OASIS’ profilers, the alert ‘Vicinal Dihaloalkanes’ is triggered in TDBPP because of the two Br atoms in the vicinal position (see Figure 1). In the QSAR Toolbox, this alert is associated with the production of DNA adducts. In addition, the ‘in vitro mutagenicity (Ames test) by ISS’ profiler flags the presence of an alert associated to an alkyl ester of phosphonic acid, as of concern for genotoxicity due to the presence of electrophilic sites. The QSAR Toolbox profiling results are provided in Annex B, while additional information on the alerts triggered can be consulted within the software.

FIGURE 1.

Chemical structures of the six BFRs included in the 2012 Opinion and covered by Recommendation 2014/118/EU: Tris(2,3‐dibromopropyl) phosphate (TDBPP), N,N′‐ethylenebis(tetrabromophthalimide) (EBTEBPI), Dibromoneopentyl glycol (DBNPG), Bis(2‐ethylhexyl) tetrabromophthalate (BEH‐TEBP), 2‐Ethylhexyl 2,3,4,5‐tetrabromobenzoate (EH‐TBB), Hexabromocyclodecane (HBCYD).

The CONTAM Panel noted that the computational analysis supports the available experimental data.

3.2.1.2.5. Observations in humans

No studies were identified in the previous Opinion, nor in the current update.

3.2.1.2.6. Studies in other model systems and mode of action

Some in vitro studies on the mode of action of TDBPP were identified on metabolic disturbances, mitochondrial toxicity and mineralocorticoid effects (Gu et al., 2019; Ma et al., 2021; Zhang et al., 2017). However, none of them were related to the observed effects in experimental animals.

No in vivo studies in mammalian species were available on neurobehaviour and reproductive effects, which have been identified as critical effects for other BFR families. Therefore, the CONTAM Panel evaluated one study in a neuronal cell line and a study in zebrafish identified as possible indicators of such hazards.

In vitro, the effects of 50 μM TDBPP on cell number, cell growth and differentiation into dopaminergic or cholinergic phenotypes were studied in differentiated PC12 cells after 6 days of exposure (Dishaw et al., 2011). Results showed a significant reduction in cell number without any impairment of cell growth and a significant increase in the emergence of both dopaminergic and cholinergic neurophenotypes.

In zebrafish (Dishaw et al., 2014), larvae exposure to TDBPP from 0 to 5 days postfertilisation showed the greatest toxicity (expressed as the percentage of mortality) at a concentration of 1 μM and higher. Neurotoxicity has been studied through the measurement of the larvae swimming activity on the 6th day post‐fertilisation. TDBPP (0.56 and 1 μM) significantly altered the locomotor activity level whereas lower levels of exposure (0.1, 0.18 or 0.31 μM) did not. In addition, hypoactivity was observed with TDBPP only during the light phase of the cycle. In zebrafish larvae, TDBPP was reported to be more potent than the known neurotoxicant chlorpyrifos, which was included in the same study for comparison.

Taken together, these two studies provide some evidence of neurotoxicity for TDBPP.

3.2.1.2.7. Consideration of critical effects and dose–response analysis

The CONTAM Panel concluded that TDBPP is nephrotoxic. It is genotoxic and carcinogenic in rats and mice. The critical studies are the carcinogenicity studies in rats and mice (see Section 3.2.2.3). No studies in humans were identified.

The Panel performed benchmark dose (BMD) modelling according to the 2022 EFSA Guidance on the use of the BMD approach in risk assessment (EFSA Scientific Committee, 2022). The results of the BMD modelling for the critical studies in rats and mice (NTP, 2014) are summarised in Table 9. Details of the BMD analyses, including the individual reports of the modelling are shown in Annex D. For quantal data the default benchmark response (BMR) of 10% was applied as recommended in the EFSA BMD guidance (EFSA Scientific Committee, 2022).

TABLE 9.

Benchmark dose (BMD) modelling for the critical studies of TDBPP (NTP, 2014; for details of the BMD analyses see Annex D).

Study	Observed effect, (sex)	BMDL10	BMD10	BMDU10
		(mg/kg bw per day)
Rats	Kidney tubular‐cell adenoma, F	2.26	3.73	6.66
	Kidney tubular‐cell adenoma or adenocarcinoma, M	0.073	0.411	0.859
Mice	Kidney adenoma or adenocarcinoma, M	42.4	69.9	97.7
	Kidney adenoma, F	142	255	375
	Stomach papilloma or carcinoma, M	13.5	40.8	66.3
	Stomach papilloma or carcinoma, F	14.8	33.3	55.5
	Lung adenoma or carcinoma, M	18.8	57.5	256
	Lung adenoma or carcinoma, F	35.9	73.1	137
	Liver adenoma or carcinoma, F	13.4	31.7	60.5

The lowest BMDL₁₀ was 0.073 mg/kg bw per day, based on combined incidence of kidney tubular‐cell adenoma or carcinoma in male rats. Combining the incidence of these tumours was considered appropriate because renal adenomas and adenocarcinomas are part of the same neoplastic progression pathway. The CONTAM Panel noted that the dose‐spacing in this study was not optimal since there was only a two‐fold difference between the two dose levels tested, and both showed a similar tumour incidence. However, modelling of the data on the sum of kidney tubular‐cell adenoma and carcinoma in male rats gave an acceptable fit. Therefore, the CONTAM Panel selected this lowest BMDL₁₀ as the Reference Point.

No Reference Point could be established for the non‐neoplastic effects that were identified as critical endpoints for other BFRs evaluated, e.g. neurotoxicity and reproductive toxicity.

Since TDBPP is both genotoxic and carcinogenic, the CONTAM Panel considered that no health‐based guidance value (HBGV) can be established as any level of exposure could theoretically lead to cancer (EFSA Scientific Committee, 2005). Instead, the margin of exposure (MOE) approach would normally be used for risk characterisation. For genotoxic and carcinogenic compounds, and according to EFSA Scientific Committee (2005); EFSA Scientific Committee (2025), a MOE of 10,000 or higher, if it is based on the BMDL₁₀ for carcinogenicity in an animal study, would be of low concern from a public health point of view and might be considered as a low priority for risk management actions.

3.2.1.3. Occurrence data and dietary exposure assessment

No occurrence data on TDBPP in food were submitted to EFSA, and only one study in the scientific literature was identified reporting levels below the LOD (Castro et al., 2020, see Section 3.1.2). No estimation of the dietary exposure was possible.

3.2.1.4. Risk characterisation

No risk characterisation was possible due to the lack of information on potential dietary exposure.

3.2.1.5. Uncertainty analysis

The uncertainty analysis was performed as described in Section 2.8. Due to the lack of data on occurrence and dietary exposure for TDBPP, uncertainty analysis was performed only for the hazard assessment. All the identified sources of uncertainty affecting the hazard assessment were assessed as having negligible or low impact on the conclusions (Table 10).

TABLE 10.

Uncertainties identified and their impact on the outcome of the hazard identification/characterisation of TDBPP.

Description of the uncertainty		Impact of the uncertainty on the hazard identification and characterisation a
Chemical composition and analytical methods
Dosing and chemical composition, including impurities	Uncertainty associated with the dose in the critical studies used in the risk assessment	0 – Negligible impact. Studies conducted used gavage dosing (developmental toxicity study) or via feed (carcinogenicity studies).
Hazard identification and characterisation
ADME	ADME	1 – Low impact. Only two studies identified. TDBPP is highly absorbed (> 50%), highly metabolised, distributed and accumulated in different tissues, and is eliminated mainly via urine.
	Information on transfer rate to animal products	0 – Negligible impact. No studies identified, but no impact on risk assessment.
	Transfer via mother's milk	0 – Negligible impact. No studies are available.
Toxicity studies in experimental animals: endpoints and study design	Limitations in the study design of the critical study that can result in uncertainties	1 – Low impact on hazard identification regarding the carcinogenicity effects.
	Relevance for humans of the adverse effect	0 – Negligible impact. TDBPP is both genotoxic and carcinogenic, thus relevant for humans.
Genotoxicity	Uncertainty in the assessment of genotoxicity	0 – Negligible impact. Strong evidence that TDBPP is genotoxic both in vitro and in vivo (see Section 3.2.1.2.3).
MOA	Uncertainties associated with the data on mode of action	1 – Low impact. Despite the strong evidence of genotoxicity in vitro and in vivo, no information on the mode of action has been identified.
Selection of Reference Point	Selection of BMDL and BMR	2 – Medium impact. In the carcinogenicity study, only two doses, and dose‐spacing closer than recommended by current test guidelines. BMR of 10% for quantal data as recommended in the EFSA 2022 Guidance. BMD criteria satisfied. BMD criteria satisfied. Both tested doses resulted in similar responses, so the BMD estimates were strongly influenced by the prior distributions used.
	Weaknesses in non‐critical studies, or absence of studies, and uncertainty about whether the endpoints they tested might have been critical if the weaknesses were not present	1 – Low impact. No studies were available on reproductive toxicity and neurotoxicity/neurobehaviour, which have been identified as critical effects for other BFRs evaluated, which are not genotoxic and carcinogenic.

Abbreviations: ADME, absorption, distribution, metabolism, excretion; BMD, benchmark dose; BMDL, benchmark dose lower confidence limit; BMR, benchmark response; MOA, mode of action; TK, toxicokinetic.

^a 0 – Uncertainty with negligible impact; 1 – Uncertainty with low impact; 2 – Uncertainty with medium impact; 3 – Uncertainty with high impact. This ordinal scale provides a relative assessment of the impact of the individual uncertainties as an aid to quantifying their combined impact by expert judgement, as described in Section 11 of EFSA Scientific Committee (2018).

Uncertainty analysis: Genotoxicity and carcinogenicity

Uncertainties affecting the evidence on the genotoxic potential of TDBPP (see Section 3.2.1.2.3.) were assessed to have negligible impact on the assessment (Table 10). TDBPP is genotoxic in vitro and in vivo, and carcinogenic in both mice and rats. Based on the overall evidence, and following an expert group discussion, the CONTAM Panel judged it was almost certain (99%–100% certainty) that TDBPP is genotoxic and carcinogenic.

Uncertainty analysis: Reference Point for carcinogenicity

The Reference Point for carcinogenicity was affected by uncertainties arising from the number and spacing of doses in the critical study (NTP, 2014), which were assessed as having medium impact on the hazard characterisation (Table 10). Only two dose levels were used, at 2 and 4 mg/kg bw per day, resulting in similar responses (48% and 54% incidence of tumours) far above the BMR.

EFSA's Guidance on BMD analysis states that, in such cases, the available information may not be sufficient for estimating the BMD (EFSA Scientific Committee, 2022) and, with only two active doses and a control, informative prior distributions would be needed for some of the parameters in the model to make the model identifiable (EFSA Scientific Committee, 2022).

Nevertheless, results of the BMD analysis for the present case met the acceptability criteria contained in the EFSA BMD Guidance. In the advanced output available from the software (see Annex D, section D.2.1 and Figures therein) comparison between prior and posterior distributions for each parameter in the model showed peaked posteriors for all parameters, and less difference between prior and posterior for the technical parameter d, which influences the shape and slope of the fitted dose–response models. This indicates that the resulting model fits were driven mainly by the prior for this technical parameter d, which is therefore informative for this dataset. Recent simulation studies that include similar datasets have demonstrated that the default prior used for the technical parameter d produces BMD estimates which meet the acceptability criteria and BMDLs with appropriate coverage, i.e. close to the nominal 95% probability that BMDL < true BMD (Aerts et al., 2022).

The construction of the default prior used for the technical parameter d was inspired by the priors used in the US‐EPA Bayesian implementation of BMD analysis. It is biologically appropriate because it closely resembles the distribution of median d estimates obtained for data from more than 3000 NTP datasets analysed by Kremer et al. (2026). Furthermore, the mass of the prior for values below 1 is restricted to around 20%, which avoids estimation of the BMDL being implausibly low or on the boundary set by the optimisation algorithm as occurs when using a frequentist analysis, where prior distributions are not involved (Kremer et al., 2026).

Based on these considerations, and following an expert group discussion, the CONTAM Panel judged with at least 90% certainty (i.e. probability) that the Reference Point for TDBPP of 0.073 mg/kg bw per day would not be lower if the non‐standard uncertainties were resolved.

3.2.2. N,N'‐Ethylenebis(tetrabromophthalimide) (EBTEBPI) [CAS number 32588‐76‐4]

3.2.2.1. Production, use, analysis and environmental fate

3.2.2.1.1. Production and use

The previous Opinion (EFSA CONTAM Panel, 2012b) reported that EBTEBPI was used as an additive in high‐impact polystyrene (HIPS) (Weil & Levchik, 2009) and that it was stated to be stable against light and heat. It may be used in polyethylene, polypropylene, thermoplastic polyesters, polyamide, ethylene propylene‐diene terpolymers (EPDM rubbers) and other synthetic rubbers, polycarbonate, ethylene copolymers, ionomer resins, epoxies and textile treatments (WHO/IPCS, 1997).

ECHA, in its investigation report on aromatic BFRs (ECHA, 2024), stated that EBTEBPI is notified in 7% of notifications received for aromatic BFRs following a call for evidence to gather information on the uses. Most commonly reported sectors for the use of EBTEBPI were automotive followed by electrical and electronic equipment (EEE), while aerospace and defence, buildings and construction and textile sectors were reported sporadically.

3.2.2.1.2. Analysis

At the time of the previous Opinion (EFSA CONTAM Panel, 2012b), no information could be identified on methods of analysis of EBTEBPI in environmental or food samples. No new information has been identified.

3.2.2.1.3. Environmental fate and occurrence in the environment

It was stated in the previous Opinion (EFSA CONTAM Panel, 2012b) that Hardy (2004) performed experimental studies on the bioconcentration of a number of BFRs in carp (Cyprinus carpio). The results on EBTEBPI were not fully conclusive, but a bioconcentration factor of < 0.3 or < 33 was indicated depending on the concentration tested. The experimental set‐up did not include exposure via feed. No new information on environmental fate or on occurrence in the environment was identified.

ECHA, in its investigation report on aromatic BFRs (ECHA, 2024), performed release estimation to identify the key substances, sectors and lifecycle steps for the overall release of aromatic BFRs to the environment. Among the aromatic BFRs investigated, EBTEBPI was found to contribute to 1.1% to the releases.

3.2.2.2. Hazard identification and characterisation

3.2.2.2.1. Toxicokinetics

In the previous Opinion (EFSA CONTAM Panel, 2012b), the Panel identified a study where female rats were treated with EBTEBPI (¹⁴C‐labelled) by gavage at 0.67 mg/kg bw, daily for 14 days. The Panel noted that the main route of excretion was via the faeces (65% of the dose), whereas 15% of the dose was recovered in the urine. No accumulation was found in tissues.

No new studies were identified since the previous Opinion.

3.2.2.2.2. Levels in humans

No studies were identified in the previous Opinion, nor in the current update.

3.2.2.2.3. Toxicity in experimental animals

Studies considered in the previous EFSA assessment

The previous Opinion reported that in 28‐day and 90‐day dietary studies with EBTEBPI in Sprague–Dawley rats, the NOAELs identified by the CONTAM Panel were the highest dose tested, which were equivalent to about 1000 mg/kg bw per day (US‐EPA HPVIS, 2012; HSBD, 2005, as cited in EFSA CONTAM Panel, 2012b).

In a developmental toxicity study, pregnant female Sprague–Dawley rats were administered EBTEBPI by gavage at doses of up to 1000 mg/kg bw daily from GD6 to GD15 (Rodwell, 1988a, as cited in US‐EPA HPVIS, 2012; HSBD, 2005). In another study, pregnant New Zealand White rabbits were administered EBTEBPI by gavage at a dose of 1000 mg/kg bw daily from GD7 to GD19 (Rodwell, 1988b, as cited in US‐EPA HPVIS, 2012; HSBD, 2005). In both studies, the NOAEL for maternal and developmental toxicity identified by the CONTAM Panel were the highest dose tested, 1000 mg/kg bw per day.

Regarding genotoxicity, the previous Opinion reported that the mutagenicity of EBTEBPI had been investigated in a number of in vitro systems. It was not mutagenic in several tests in S. Typhimurium and E. coli or in Saccharomyces cerevisiae (D4) (see EFSA CONTAM Panel, 2012b).

EBTEBPI had no porphyrinogenic activity in vitro in cultured chick embryo liver cells.

Studies published since the previous EFSA assessment

No new studies have been identified except information on genotoxicity in the REACH dossier ²⁸ as submitted by the registrant(s). The CONTAM Panel did not have access to the original data. The REACH dossier reported that:

• − EBTEBPI (purity NR) was not mutagenic to five strains of S. Typhimurium (TA98, TA100, TA1535, TA1537 and TA1538) and E. coli, WP2 uvrA when tested with or without S9 metabolic activation (Aroclor‐induced rat liver S9 (1982).
• − EBTEBPI (purity NR) was reported to be negative in an in vitro chromosomal aberration assay in CHO cells in the presence or absence of metabolic activation (Aroclor‐induced rat hepatic S9) (2008).
• − EBTEBPI (purity NR) was reported to be negative in an in vitro gene mutation test in Chinese hamster V79 cells using the hprt gene in the presence or in the absence of metabolic activation (phenobarbitone/beta‐naphthoflavone rat liver S9‐mix) (2016).

Summary of toxicity in experimental animals

Based on the limited data available, EBTEBPI did not show effects in short‐term (28‐ and 90‐day) toxicity studies in rats or in developmental toxicity studies in rats and rabbits, up to 1000 mg/kg bw per day, the highest dose tested. No chronic toxicity/carcinogenicity, reproductive toxicity or developmental neurotoxicity studies were identified.

According to the data reported in the REACH dossier, EBTEBPI was not mutagenic in S. Typhimurium TA98, TA100, TA1535 and TA1537 and E. coli WP2uvrA with or without metabolic activation. It was also negative in a gene mutation assay at the hprt locus in V79 cells and in a chromosomal aberration test in CHO cells in presence or absence of metabolic activation. The aneugenic potential of EBTEBPI has not been investigated. No in vivo genotoxicity data were identified, but would not be required considering the negative in vitro data (EFSA Scientific Committee, 2011). Based on this information, the CONTAM Panel concluded that EBTEBPI is not genotoxic.

3.2.2.2.4. Computational analysis

The QSAR Toolbox identified experimental data for the Ames test on EBTEBPI as retrieved by the previous Opinion (see above and Annex B).

VEGA models predicted negative Ames test results for EBTEBPI with ‘GOOD reliability’ (three models) and with ‘MODERATE reliability’ (one model). The Panel critically reviewed the predictions and considered them valid mainly due to the presence of relevant analogues known by the model with experimental and predicted values concordant with the prediction for the target compound (see Annex B). The CONTAM Panel noted that this negative prediction is consistent with the available experimental data.

3.2.2.2.5. Observations in humans

No studies were identified in the previous Opinion, nor in the current update.

3.2.2.2.6. Mode of action

In the previous Opinion, two studies were reported, neither of them providing useful information for the hazard assessment. No new studies were identified.

3.2.2.2.7. Consideration of critical effects and dose–response analysis

Based on the information reported in the REACH dossier, the CONTAM Panel concluded that EBTEBPI is not genotoxic.

Because of the limited data available, the CONTAM Panel concluded that it was not possible to identify a Reference Point.

3.2.2.3. Occurrence data and dietary exposure assessment

No occurrence data on EBTEBPI in food were submitted to EFSA, and no occurrence or exposure data were identified in the literature, either in the previous Opinion or in the current update (see Section 3.1.2). Therefore, no estimation of the dietary exposure was possible.

3.2.2.4. Risk characterisation

No risk characterisation was possible due to the lack of a Reference Point and lack of information on potential dietary exposure.

3.2.2.5. Uncertainty analysis

Due to the lack of data on occurrence and dietary exposure for EBTEBPI, and the insufficiency of the available toxicological data for establishing a Reference Point, uncertainty analysis was performed only for the assessment of genotoxic potential. In view of the negative in vitro data in bacteria and mammalian cells (gene mutation and chromosomal damage), with no identified sources of uncertainty, the CONTAM Panel judged with > 95% certainty (i.e. probability) that EBTEBPI is not genotoxic.

3.2.3. Hexabromocyclodecane (HBCYD) [CAS number 25495‐98‐1]

3.2.3.1. Production, use, analysis and environmental fate

3.2.3.1.1. Production and use

There is very little information in the scientific literature about HBCYD. It is unfortunate that there is great similarity in both name and abbreviations often used for hexabromocyclododecane (HBCDDs; HBCD) and hexabromocyclodecane (HBCYD, HBCD). This leads to confusion sometimes even within the same published article (see Section 1.2).

The previous EFSA Opinion (EFSA CONTAM Panel, 2012b) stated that HBCYD together with tetrakis(hydroxymethyl)phosphonium salts may be used as an alternative to cyclic methylphosphonates in cotton‐polyester blends (Weil & Levchik, 2009).

3.2.3.1.2. Analysis

No new literature has been identified that could be unambiguously associated with HBCYD (and not confused with HBCDDs).

3.2.3.1.3. Environmental fate and occurrence in the environment

No new information has been identified in the literature since the previous Opinion that could be unambiguously associated with HBCYD (and not confused with HBCDDs).

3.2.3.2. Hazard identification and characterisation

The CONTAM Panel could not identify any toxicokinetic, toxicity or epidemiological studies on HBCYD. Regarding the mode of action, in the previous Opinion one study was reported, but it did not provide useful information for the hazard assessment.

It was therefore not possible to identify a Reference Point for HBCYD, and it was not possible to conclude on its genotoxic potential.

Computational analysis

The QSAR Toolbox did not identify any experimental data for the Ames test on HBCYD (see Annex B). VEGA models predicted negative Ames test results for HBCYD with ‘LOW reliability’ (four models). These predictions were not further considered due to their limited reliability (see Section 2.2).

3.2.3.3. Occurrence data and dietary exposure assessment

No occurrence data on HBCYD in food were submitted to EFSA, and no occurrence or exposure data were identified in the literature, either in the previous Opinion or in the current update. Therefore, no estimation of the dietary exposure was possible.

3.2.3.4. Risk characterisation

No risk characterisation was possible due to the lack of a Reference Point and lack of information on potential dietary exposure.

3.2.4. Bis(2‐ethylhexyl) tetrabromophthalate (BEH‐TEBP) [CAS number 26040‐51‐7]

3.2.4.1. Production, use, analysis and environmental fate

3.2.4.1.1. Production and use

The previous EFSA Opinion (EFSA CONTAM Panel, 2012b) reported that BEH‐TEBP is used as an additive flame retardant replacing PentaBDE in many applications, e.g. in polyurethane foam, PVC and neoprene rubber and in wire and cable insulation. It is also used as a plasticiser (Andersson et al., 2006; de Wit et al., 2011). BEH‐TEBP may also be used in combination with EH‐TBB (see Section 3.1.5), in differing proportions resulting in commercial mixtures with different trade names.

Annual production in the United States of America (US) was 450–4500 tonnes during 1990–2006 (US‐EPA, 2006, 2010, as cited by de Wit et al., 2011).

3.2.4.1.2. Analysis

Little has changed since the previous EFSA Opinion (EFSA CONTAM Panel, 2012b) where it was reported that BEH‐TEBP can be analysed using multi‐analyte methods (i.e. methods that analyse BEH‐TEBP together with other BFRs) (e.g. Covaci et al., 2011; Klif, 2010). Extraction of BEH‐TEBP is usually done with solvent mixtures of medium polarity such as cyclohexane:acetone or dichloromethane:n‐hexane. Clean‐up can be performed using gel permeation chromatography (GPC) for lipid removal, florisil (Klif, 2010) or alternatively it can be isolated using pressurised liquid extraction (PLE) followed by clean‐up using acid silica and subsequent carbon column clean‐up (Zhou et al., 2010). Instrumental analysis is usually performed using either GC–MS or LC–MS based methods.

Since the previous Opinion was published, there have been many publications on BEH‐TEBP in dust and environmental samples but very few of food. Bichon et al. (2018) reported on ‘simultaneous analysis of historical, emerging and novel BFRs in food and feed’ (including BEH‐TEBP) using a common extraction and purification method. The method used falls into the description above and was based on clean‐up using acidified silica, florisil and carbon columns, followed by analysis using gas and liquid chromatography coupled to high resolution and tandem mass spectrometry.

3.2.4.1.3. Environmental fate and occurrence in the environment

There is very little information about the environmental fate of BEH‐TEBP but given the similar chemical characteristics it can be assumed to behave in a broadly similar way to the other BFRs. There is one study reported by de Jourdan et al. (2013) on the environmental fate and persistence of three novel brominated flame retardants in aquatic mesocosms, where Firemaster BZ‐54, a commercial mixture containing BEH‐TEBP and EH‐TBB in a ratio of 1:4 was one of the compounds investigated. It was found that BEH‐TEBP (and other compounds) was environmentally persistent in both the particulate and the sediment compartments. The median dissipation times (DT50) differed in each compartment, with more rapid disappearance in the particulate (9–30 days) compared with the sediment compartment (> 100 days). The degradation products were more concentrated in the particulate compartment and corresponded to known photodegradation products; the proposed degradation pathway being through the hydrolysis of the ester groups to form an anhydride. The ratio of EH‐TBB to BEH‐TEBP differed in the mesocosm compartments compared with the technical product used for treatment, indicating increased degradation of EH‐TBB relative to BEH‐TEBP (de Jourdan et al., 2013).

Occurrence in the environment

Air

Drage et al. (2016) reported on concentrations of flame retardants (including BEH‐TEBP) in air and soil on a transect in the West Midlands (UK). BEH‐TEBP was found in a third of the air samples tested, but could not be quantified due to the low recovery of the relevant internal standards.

In studies by Salamova et al. (2014) and Hao et al. (2020) where BFRs were measured in air from the arctic region, BEH‐TEBP was included but not detected.

BEH‐TEBP has also been investigated in air in studies that were conducted outside Europe and concentrations were always very low (e.g. an average of 3.61 pg/m³ was reported for BEH‐TEBP in air samples collected in a test home in Harbin, China, between January 2017 and June 2018).

ECHA, in its investigation report on aromatic BFRs (ECHA, 2024) conducted a review and found several reports of BEH‐TEBP in air. Concentrations were generally between < LOD and 60 pg/m³ except where there were known or suspected sources of contamination, e.g. near e‐waste dismantling sites. Higher concentrations were also found in air samples taken from Arctic locations as is common for persistent organic pollutants in general (Carlsson et al., 2018).

Soil

Drage et al. (2016) reported on concentrations of flame retardants (including BEH‐TEBP) in air and soil on a transect in the West Midlands (UK). BEH‐TEBP was not found in soil. The ECHA (2024) review identified a study on BFRs in indoor and outdoor media in Stockholm, Sweden but BEH‐TEBP was not detected in soil (Newton et al., 2015). No studies with quantified data for BEH‐TEBP in soil could be found for European countries or elsewhere globally.

Sediment

No data were identified for BEH‐TEBP in sediment from European countries, but there were some studies from outside Europe (Chokwe et al., 2019; Olukunle & Okonkwo, 2015; Xu et al., 2012). Concentrations were < 60 ng/g dry weight, where detected.

Dust

For many BFRs dust can be an important source of exposure. This is especially the case for toddlers and infants due to low body weight, high crawling activity and high frequency of hand‐to‐mouth contact. A summary of reported occurrence of BEH‐TEBP in dust samples from European countries, and estimates of exposure via dust when reported by the authors is given in Appendix D (Table D.1). Concentrations reported are typically < 2 ng/g to 5000 ng/g but values as high as 25,000 were reported for dust from UK offices (Tao et al., 2016). Exposure estimates are typically a few 10s of ng/kg bw per day with a high value reported of 47 ng/kg bw per day for children in a study conducted in Spain (Esplugas et al., 2022). These values are comparable to the exposure estimates reported in the literature for BEH‐TEBP from the diet, e.g. 0.91 to 3.3 ng/kg bw per day across population groups (Ma et al., 2023) and 1.2 ng/kg bw per day for the 95th percentile (Xu et al., 2017) (see Section 3.2.4.3).

Water

No reports on BEH‐TEBP in inland water were identified for Europe. Guo et al. (2017) measured the concentrations of 65 flame retardants in water samples from five Lake Michigan tributaries. The mean concentrations of BEH‐TEBP from five different sites investigated ranged from 83 to 690 pg/L.

The ECHA investigation report on aromatic BFRs (ECHA, 2024) identified a study (Carlsson et al., 2018) that reported concentrations of up to 7100 ng/L (7,100,000 pg/L) in marine water from the Arctic. Other studies identified in their search reported concentration that were < 3 ng/L.

Wildlife

Most of the studies on BEH‐TEBP in wildlife that are reported are on either fish or birds. There is one study from Europe (Sühring et al., 2013) on BFRs in eels from German rivers. Where detected, concentrations were < 7.4 ng/g. Other studies identified were from outside Europe and concentrations, when detected in fish, birds or bird eggs, were of the same order of magnitude.

The ECHA investigation report on aromatic BFRs (ECHA, 2024) identified reports of BEH‐TEBP in whale, seal, bear and gull of up to 6090 pg/g lipid weight taken from the Norwegian Arctic, and concentrations of around 400 pg/g dry weight in tree leaves and shoots (Fagus sylvatica or Populus nigra italica) from Germany. Marine specimens from the North Sea had concentrations of < 75 pg/g.

3.2.4.2. Hazard identification and characterisation

3.2.4.2.1. Toxicokinetics

No information relating to toxicokinetics in mammals was described in the previous Opinion (EFSA CONTAM Panel, 2012b). Since then, several studies have been identified.

Because of the similar structure between DEHP and BEH‐TEBP, Springer et al. (2012) hypothesised that the cleavage by spontaneous‐ or esterase‐mediated hydrolysis of BEH‐TEBP could lead to a monoester metabolite mono‐(2‐ethyhexyl) tetrabromophthalate (TBMEHP), as was observed for DEHP with the formation of MEHP. The authors studied the in vitro metabolism of BEH‐TEBP using commercial hepatic porcine esterase, and confirmed the formation of TBMEHP.

Roberts et al. (2012) investigated the metabolism of BEH‐TEBP in human and rat liver microsomes. The authors did not detect metabolites of BEH‐TEBP in human or rat subcellular fractions.

Bearr et al. (2012) studied the in vitro metabolism of the flame‐retardant mixture Firemaster BZ‐54, containing BEH‐TEBP in mice (and other non‐mammalian species) hepatic subcellular fractions. The authors found that BEH‐TEBP was poorly metabolised (metabolites had not been characterised).

Phillips et al. (2016) studied the maternal transfer of Firemaster 550 components containing BEH‐TEBP in Wistar rat. Rats were orally exposed to 300 or 1000 mg of Firemaster 550 for 10 consecutive days during gestation (GD9–18) or lactation (PND3–12). The authors concluded that low levels of BEH‐TEBP were transferred to the fetus during both lactation and gestation.

Silva et al. (2016) orally administered a single dose of 500 mg/kg bw of the technical product Uniplex FRP‐45, reported to contain > 95% BEH‐TEBP plus < 5% of unknown impurities (the authors reported the identification of EH‐TBB < 5%) to female Sprague–Dawley rats. The authors identified tetrabromophthalic acid (TBPA) and tetrabromobenzoic acid (TBBA) in the urine and serum of rat (faeces were not collected).

Knudsen, Sanders, and Birnbaum (2017) investigated the toxicokinetics of BEH‐TEBP in female Sprague–Dawley rats and male B6C3F1 mice by administering a single dose and repeated exposure of [¹⁴C]‐BEH‐TEBP by gavage and iv route at 0.07 mg/kg or 7 mg/kg. The authors showed that in rats, after oral administration, less than 0.3% of BEH‐TEBP dose was excreted in urine, whereas 98% of the radioactivity was eliminated in faeces by 72 h. The authors noted that less than 0.1% of the oral dose (0.07 mg/kg) was eliminated in the bile over the first 24 h. After iv injection (0.07 mg/kg), BEH‐TEBP was metabolised to TBMEHP (% conversion not reported); both the parent compound and metabolite were eliminated via biliary excretion. The authors observed that male B6C3F1 mice appeared to excrete a similar proportion of the dose, but eliminated BEH‐TEBP more rapidly via the faecal route than female rats.

After repeated administration, the same authors assessed accumulation potential of BEH‐TEBP in female Sprague–Dawley rats by measuring radioactivity recoveries in excreta and tissues collected 24 h after 10 daily oral administrations of BEH‐TEBP at 0.07 mg/kg (Knudsen, Sanders, & Birnbaum, 2017). The authors stated that BEH‐TEBP could accumulate in some tissues (mainly in liver and adrenal tissues); specifically, the study results showed that concentrations in adrenal and liver tissue increased after 10 exposures compared to a single administration but the % of radioactivity in adrenal and liver tissue was 0.01% and 0.4%, respectively.

Chen et al. (2020) administered BEH‐TEBP to female Balb/c mice by gavage (8 μg/mL, approximately 200 μg/kg bw) and analysed for metabolites after 4 and 8 h post exposure. The authors did not detect metabolites in blood.

In summary, after oral administration in rodents BEH‐TEBP was poorly absorbed and eliminated mainly by the faecal route as a parent compound. After repeated exposure to BEH‐TEBP accumulation in some tissues has been reported (mainly in liver and adrenal tissues).

3.2.4.2.2. Levels in humans

No information on levels of BEH‐TEBP in human samples was reported in the previous Opinion. Since then, several studies have been identified reporting concentrations of BEH‐TEBP as well as other BFRs covered in this assessment in samples from European countries. Details of these studies are reported in Appendix E (Table E.1), and a brief summary of the results related to BEH‐TEBP are reported below.

Regarding human milk samples,

• In the study by Čechová et al. (2017), BEH‐TEBP was analysed in human milk samples from Norway (2003–2006, n = 305), the Netherlands (2011–2014, n = 116) and Slovakia (2011–2012, n = 37) along with 10 PBDEs and 17 other alternative halogenated flame retardants. BEH‐TEBP was reported to have a detection frequency in the human milk samples analysed of < 29% (LOD = 0.023 ng/g lipid). Mean and P95 concentrations across countries ranged from 0.613 to 1.34 ng/g lipid and from 1.432 to 2.750 ng/g lipid, respectively, and BEH‐TEBP concentrations were reported to be the highest among the analysed halogenated flame retardants across all countries.

• Tao et al. (2017) analysed human milk samples collected in the UK in 2010 (n = 25) and 2014–2015 (n = 10) for several BFRs. In 2010, the levels of BEH‐TEBP ranged from < 0.1 to 4.6 ng/g lipid (detection frequency 36%), while in 2014–2015 a higher detection frequency (50%) was reported and concentrations ranged from < 0.1 to 0.73 ng/g lipid, with a mean value of 0.25 ng/g lipid.

• Sahlström et al. (2015) did not detect BEH‐TEBP above the LOD (< 26 pg/g ww) in two pooled human milk samples from Sweden collected in 2009–2010.

Regarding levels in blood/serum

• Sahlström et al. (2014) did not detect BEH‐TEBP above the LOD (30 pg/25 μL) in any of the 24 matched serum samples from mothers and their toddlers collected in Sweden between 2009 and 2010.

• BEH‐TEBP was also reported below the LOD in 61 serum samples collected in 2013 from a Norwegian cohort (LOD = 7.0 pg/serum sample, volume not reported, Tay et al., 2019).

• Van der Schyff et al. (2023) reported on flame retardants in children from nine European countries from the HBM4EU aligned studies. BEH‐TEBP was investigated in blood from France only collected between 2014 and 2016, and was reported to be below the LOQ (0.7 μg/L) in all samples analysed.

• Sales et al. (2017) did not detect BEH‐TEBP above the LOQ (5 pg/g serum) in one pool of human cord blood from Belgium.

3.2.4.2.3. Toxicity in experimental animals

Studies considered in the previous EFSA assessment

No studies in experimental animals were identified in the previous Opinion.

Studies published since the previous EFSA assessment

Since the previous Opinion, several new studies have been identified.

• BEH‐TEBP (purity NR) was administered to groups of 8 male Balb/c mice by gavage at 0 or 200 mg/kg bw per day for 28 days and then the livers were examined for histopathological findings and for biochemical parameters (Yin & Zhang, 2021). The authors reported that BEH‐TEBP resulted in hepatocyte oedema, narrowing of the sinusoids and dilation of the central vein.

• Three groups of Balb/c male mice (5 weeks of age, n = 8/group) were daily exposed by gavage to saline or BEH‐TEBP at 0, 20 or 200 mg/kg bw per day for 28 days and examined for their learning and memory performances in a Morris water maze (Bao & Jing, 2021). Behavioural results did not show learning impairments during the 5‐day training period of the test. The memory retention of the location of the platform was tested at the 7th day of testing with results showing non‐significant dose‐related reductions in the time spent in the platform quadrant and the number of crossings of the platform area.

• Two groups of Balb/c male mice (8–9 weeks of age, n = 10/group) were daily exposed to saline added with DMSO or BEH‐TEBP at a dose of 20 mg/kg bw per day for 60 days and explored for their learning and memory performances in a step‐down passive avoidance test and a Morris water maze (Xiong et al., 2024). The results showed learning and memory disturbances in both tests reflected by a significant increase in the number of errors in the 1 and 24 h‐ step‐down test, and a significant reduction in the latency and number of the platform crossed in the Morris water maze. The CONTAM Panel noted that this study provided limited information due to the single dose level and unclear information on the route of administration of BEH‐TEBP, and did not consider it further.

• A 90‐day repeated dose oral toxicity study has been conducted in groups of 10 male and 10 female Wistar Han rats with BEH‐TEBP (purity NR) administered by gavage doses of 0, 100, 300 and 1000 mg/kg bw per day (Fulcher, 2018; REACH dossier, ²⁹ as submitted by the registrant(s)). The study was performed in compliance with GLP and OECD TG408. There were no treatment‐related findings in the wide range of observations, including mortality, body and organ weights, functional performance tests of motor activity and grip strength, sensory reactivity, ophthalmology, haematology, blood chemistry and histopathology. The NOAEL was 1000 mg/kg bw per day, the highest dose tested.

• In a comparative study on effects of a number of BFRs on liver and thyroid toxicity, conducted by the NTP, BEH‐TEBP was administered for 5 days via oral gavage to groups of 6 male Harlan Sprague–Dawley rats at doses of 0, 0.07, 0.71, 7.1, 71 and 706 mg/kg bw per day (NTP, 2019a; Shockley et al., 2020a). After the 5‐day exposure, there were no dose‐related effects on survival, body weight, gross pathology, organ weights or histopathology of the liver or thyroid gland or levels of thyroid hormones (NTP, 2019a; Shockley et al., 2020a). Shockley et al., (2020a) in an investigation of metabolic oxidative stress and liver disease pathways found no changes in the transcriptome related to these parameters, indicating that liver toxicity would not be anticipated also after longer exposure.

• Yuan et al. (2024a) dosed 6‐weeks old Sprague–Dawley male rats with BEH‐TEBP (purity 95%) at 0, 5, 50 and 500 mg/kg bw per day for 28 days by gavage. BEH‐TEBP exposure resulted in significant dose‐related increases in serum TT4 and FT4 levels at 50 and 500 mg/kg bw per day, with corresponding decreases in TSH and relative thyroid weight, which were significant at the highest dose. There was no effect on TT3 or FT3. The authors reported that histopathological analysis of the thyroid showed that the thyroid follicular epithelium in the high‐dose group was flatter and thinner, and the follicular cavity was markedly dilated, compared with the control group.

• Fetal testosterone production was unaltered on GD18 in rats where the pregnant dams were exposed to BEH‐TEBP at a single dose of 750 mg/kg bw per day from GD14 to GD18, which is a critical period for sexual differentiation of the reproductive tract (Furr et al., 2014).

• The same daily dose of 750 mg/kg per day given by daily oral gavage to pregnant rats from GD17 to GD21 did not induce formation of multinucleated germ cells in the testes of fetuses at GD21 (Spade et al., 2018).

Information on repeated dose toxicity studies was available in the REACH dossier on BEH‐TEBP, as submitted by the registrant(s). The CONTAM Panel did not have access to the original data. The REACH dossier reported that:

• −In a 28‐day feeding study (OECD TG 407, OECD, 2008), CD rats were exposed to 0, 22, 223 or 2331 mg/kg bw per day to BEH‐TEBP (purity > 95%). The authors stated that only minor changes were observed and only at the highest dose (slightly lower bodyweight gain in females, marginally low alanine aminotransferase (ALT) serum levels compared to concurrent controls in females, and marginally low phosphorus concentrations in animals of both sexes).
• −In a developmental toxicity study in Sprague–Dawley rats (OECD TG414, OECD, 2018), pregnant female Sprague–Dawley rats were exposed by gavage from GD5 to GD19 to 0, 250, 500 or 1000 mg/kg bw per day BEH‐TEBP (purity 90.6%). No treatment‐related effects in dams were observed. Some changes in skeletal ossification were recorded in fetuses/offspring at the highest dose. However, the authors considered these to be normally occurring variations. The authors reported a NOAEL for both maternal and developmental toxicity of 1000 mg/kg bw per day, the highest dose tested.

Some studies considered binary mixtures of BEH‐TEBP and EH‐TBB (Schkoda et al., 2023; Witchey et al., 2020, 2023) or the technical product Firemaster BZ‐54 which contains only BEH‐TEBP and EH‐TBB (Hays & Kirman, 2017). Considering the limited knowledge of effects with experiments with single chemicals, the CONTAM Panel found it useful to consider evidence from these mixture studies with known composition and limited impurities. These studies are briefly described below.

Witchey et al. (2020) exposed pregnant Wistar rats orally to a number of flame retardant mixtures including a binary mixture of BEH‐TEBP and EH‐TBB in the relative proportions present in the technical product Firemaster 550 (68.1% EH‐TBB and 31.9% BEH‐TEBP). The binary mixture was administered on a wafer cookie starting at 72 h after mating and continued until PND21 at a single dose level of approximately 3.3 mg/kg bw per day (equivalent to 1.05 mg/kg bw per day BEH‐TEBP and 2.25 mg/kg bw per day EH‐TBB), based on the initial weight of the dams. There was a decrease in body weight gain in the offspring throughout the observation period. From PND65, the pups were randomly selected for behavioural testing. Changes in socioemotional behaviour were reported with deficits greater in female offspring than in males. PND1 pups from this study were subjected to transcriptomic and lipidomic analyses (Witchey et al., 2023) (see Section 3.2.4.2.5 on mode of action).

Subsequently, Schkoda et al. (2023) analysed tissues collected from the animals reported in Witchey et al. (2020) for skeletal effects. Bone health and parameters such as cortical and trabecular density were evaluated in offspring at 8 months using X‐ray micro‐computed tomography. Serum was collected from the trunk blood at the time of killing and used for analysis of 5‐hydroxytryptophan and osteocalcin. BFR exposure of dams resulted in reduced bone mineral density of the femur in female offspring, with no effects observed in males. The average marrow area of the femur was also decreased in female offspring but not in males. There were no changes in 5‐hydroxytryptophan or osteocalcin concentrations in the serum of BFR exposed offspring.

Hays and Kirman (2017) summarised data from three unpublished studies carried out on effects of Firemaster BZ‐54 in rodents and used these for risk assessment of this technical product. Firemaster BZ‐54 contained 70% EH‐TBB and 30% BEH‐TEBP. For two of these studies, Hays and Kirman (2017) did not report any data other than NO(A)ELs/LO(A)ELs. For the third study, dose–response data were reported for changes in body weight and spleen weight. The three studies were conducted in the same strain of Sprague–Dawley rats [Crl:CD(SD)]. The studies are briefly summarised below:

• −Study 1 (WIL, 1997, as reported in Hays & Kirman, 2017): In a 28‐day study, groups of six rats of each sex were given oral (gavage) doses of 0, 160, 400 or 1000 mg/kg bw per day of Firemaster BZ‐54 dissolved in corn oil. Males in the highest dose group exhibited reduced food consumption and body weight, while females showed these effects across all dose groups. Clinical observations revealed relaxed vaginal opening in all treated female groups and salivation in both sexes at the highest dose. Renal tubule epithelial regeneration was observed in a dose‐dependent manner in exposed males from all treatment groups and in 100% of the exposed female animals. The authors conducting the study identified a lowest observed effect level (LOEL) of 160 mg/kg bw per day for BZ‐54. In the absence of access to the data to support this conclusion, the CONTAM Panel was not able to use this study for hazard characterisation.
• −Study 2 (MPI, 2008a, as reported in Hays & Kirman, 2017): In a 2008 developmental toxicity study, groups of 25 female rats were administered oral (gavage) doses of 0, 50, 100 or 300 mg/kg per day of Firemaster BZ‐54 in peanut oil from gestation days 6 through 19. Maternal toxicity, evidenced by decreased food consumption, body weight gains and sparse abdominal hair, was noted in the mid and high‐dose groups. Fetal body weights were reduced in these dose groups, and skeletal malformations (fused cervical neural arches) and variations were observed in a small number of fetuses at the highest dose. The authors conducting the study identified a NOAEL of 50 mg/kg bw per day for Firemaster BZ‐54. In the absence of access to the data to support this conclusion, the CONTAM Panel was not able to use this study for hazard characterisation.
• −Study 3 (MPI, 2008b, as reported in Hays & Kirman, 2017): A two‐generation reproductive toxicity study was conducted with Firemaster BZ‐54. Groups of 19–23 female rats were administered oral (gavage) doses of 15, 50 or 165 mg/kg bw per day of Firemaster BZ‐54 in peanut oil for at least 10 weeks prior to mating, continuing through mating, gestation and up to lactation day 21. Parental and F1 females showed lower body weights and body weight gains, with F1 males being slightly affected. Pups from both generations exhibited reduced body and spleen weights. No other treatment‐related effects were observed. Based on the observed effects, the authors conducting the study identified a NOAEL of 50 mg/kg bw per day for parental and neonatal toxicity, while the NOAEL for reproductive performance and fertility was identified to be 165 mg/kg bw per day.

New data were also available for a technical product containing BEH‐TEBP, i.e. Firemaster 550 (Baldwin et al., 2017; Gillera et al., 2020; Krentzel et al., 2021; Marinello et al., 2023; Patisaul et al., 2013; Phillips et al., 2016; Rock et al., 2018; Schkoda et al., 2023; Witchey et al., 2020, 2023). The CONTAM Panel concluded that data on this technical product added little to the body of evidence for the effects of BEH‐TEBP (and EH‐TBB) and were not suitable for identifying a Reference Point for use in risk characterisation. This is because the studies provide limited information on composition and impurities, and Firemaster 550 included organophosphate flame retardants.

Genotoxicity

Since the previous Opinion, genotoxicity studies were available in the REACH dossier, as submitted by the registrant(s), and the registrant concluded that BEH‐TEBP is not genotoxic. ECHA did not make any conclusions on mutagenicity of BEH‐TEBP as it does not have a harmonised classification for mutagenicity under the Classification, Labelling and Packaging (CLP) Regulation (EC) No 1272/2008. The CONTAM Panel did not have access to the original data. The REACH dossier reported that:

• BEH‐TEBP (purity NR) was reported to be negative in a reverse mutation assay (both plate incorporation and pre‐incubation methods) in S. Typhimurium TA98, TA100, TA1535 and TA1537 and E. coli WP2uvrA in the presence and absence of S9 mix from Aroclor‐induced rat liver. It was also reported that BEH‐TEBP (RC9927) to be negative in pour‐plate assays in S. Typhimurium TA98, TA100, TA1535, TA1537 and TA1538. BEH‐TEBP (purity 90.6%) was also negative in a V79/hprt gene mutation assay with and without metabolic activation (tested up to 5000 μg/mL; 4 and 24 h treatment).

• In an in vitro chromosome aberration test in human lymphocytes with and without metabolic activation, BEH‐TEBP (RC9927) showed a weak clastogenic activity only at the highest tested concentration (1000 μg/mL) in the absence of any cytotoxicity.

• Two in vivo micronucleus tests in bone marrow of CD‐1 mice after single ip or sub‐acute (5 days) dermal administration of BEH‐TEBP (purity > 95%) at doses up to 2000 mg/kg bw were reported to be negative (Substance Evaluation Conclusion document, KEMI, 2020).

Summary of toxicity in experimental animals

Toxicity studies on BEH‐TEBP have indicated effects on the liver (hepatocyte oedema, narrowing of the sinusoids and dilation of the central vein) in mice following oral administration for 28 days at 200 mg/kg bw per day (only dose tested). No significant effects on learning or memory retention were reported in mice exposed to BEH‐TEBP for 28 days (0, 20 or 200 mg/kg bw per day), whereas learning and memory performance was impaired in mice exposed daily to 20 mg/kg bw per day for 60 days. In rats, BEH‐TEBP exposure for 28 days by oral gavage resulted in significant increases in serum TT4 and FT4 levels at 50 and 500 mg/kg bw per day, with a corresponding decrease in TSH and thyroid weight, which were significant at the highest dose, but no effect on TT3 or FT3. No effects were reported in a 90‐day study in rats at doses up to 1000 mg/kg bw per day, or in a 5‐day study at BEH‐TEBP doses up to 706 mg/kg bw per day, including on the thyroid.

According to information provided in the REACH dossier, as submitted by the registrant(s), BEH‐TEBP was not mutagenic in bacteria or in mammalian cells in vitro, but was a weak clastogen in human lymphocytes in vitro. In vivo, it was negative in micronucleus tests in mice after ip or dermal exposure.

In rats, exposure of a binary mixture of BEH‐TEBP and EH‐TBB (BEH‐TEBP: 1.05 mg/kg bw per day; EH‐TBB: 2.25 mg/kg bw per day) via dams during gestation and lactation resulted in changes in socioemotional behaviour with deficits greater in female offspring than in males (Witchey et al., 2020). Effects of exposure of pregnant rats to this binary mixture resulted in reduced bone mineral density and average marrow area of the femur in female offspring, with no effects observed in males (Schkoda et al., 2023).

Three unpublished studies on Firemaster BZ‐54 containing 70% EH‐TBB and 30% BEH‐TEBP were summarised in the publication of (Hays & Kirman, 2017), only one of which (MPI, 2008b) was described with sufficient detail to allow evaluation by the CONTAM Panel. In this two‐generation reproduction toxicity study, decreased bodyweights and spleen weights were reported, particularly in the F1 pups, with a NOAEL of 50 mg/kg bw per day of the technical product, corresponding to 15 mg BEH‐TEBP and 35 mg EH‐TBB /kg bw per day.

From the available studies, the lowest LOAEL for the binary mixture of BEH‐TEBP and EH‐TBB is 3.3 mg/kg bw per day (equivalent to 2.25 mg EH‐TBB and 1.05 mg BEH‐TEBP/kg bw per day), identified based on changes in socioemotional behaviour deficits and skeletal effects in female offspring of rat dams exposed to the mixture during gestation and lactation.

3.2.4.2.4. Computational analysis

The QSAR Toolbox identified experimental data for the Ames test on BEH‐TEBP (see above and Annex B).

VEGA models predicted negative Ames test results for BEH‐TEBP with ‘GOOD reliability’ (three models) and with ‘MODERATE reliability’ (one model). The Panel critically reviewed the predictions and considered them valid mainly due to the presence of relevant analogues known by the model with experimental and predicted values concordant with the prediction for the target compound (see Annex B). The CONTAM Panel noted that this negative prediction is consistent with the available experimental data.

3.2.4.2.5. Observations in humans

In the previous Opinion, no human studies on BEH‐TEBP were reported. Since then, eight publications representing six epidemiological studies were identified. In Appendix F an overview of the epidemiological studies identified reporting on the BFRs considered in this Opinion is shown, including those on BEH‐TEBP.

The studies on BEH‐TEBP pertained to populations in China (four studies) or the USA (two studies). Various endpoints were assessed including thyroid hormone profile (three studies), thyroid cancer (one study), type 2 diabetes (one study), BMI (one study) and respiratory disease (including lung function assessment) (one study). One study assessed DNA methylation. Exposure was assessed in ambient air and wristbands (one study), in dust (one study) or in serum (four studies).

Guo et al. (2018) in a cross‐sectional study collected blood samples from 174 school students (age 10 years) living near a petrochemical complex in South China. PBDE congeners and eight other flame retardants were measured, of which six are assessed in the current Opinion (see Appendix F). Associations with thyroid hormone levels were evaluated for the single chemicals and for sums of the PBDEs and NFRs. BEH‐TEBP was associated with lower FT4 levels (β −0.21, 95% CI –0.42, −0.009), but no associations were found for T3, T4, FT3 and TSH.

Wang, Romanak, et al. (2019) assessed the association between 10 BFRs (of which six are assessed in the current Opinion, see Appendix F) and thyroid hormone profile (FT3, FT4 and TSH) in rural Central Appalachia (USA) (n = 101). Exposure was assessed in ambient air and in wristbands. In wristbands, BEH‐TEBP and EH‐TBB were the most abundant BFRs, with median concentrations of 31 and 87 ng/g wristband, respectively. The combined contribution of these two chemicals to the sum of 10 BFRs was 94%. Among the multiple analyses performed, no statistically significant associations with the thyroid hormone profiles were reported for BEH‐TEBP.

Liu, Li, et al. (2022) in a case–control study with a cross‐sectional exposure assessment collected 481 serum samples from patients with thyroid cancer (n = 242) and healthy controls (n = 239) in Shandong Province, eastern China. The levels of 9 BFRs (of which seven are assessed in the current Opinion, see Appendix F) and 9 organophosphate esters, thyroid hormones and serum lipid parameters, were measured in all the participants at the time of diagnosis and the authors assessed the association between the levels of BFRs in serum and thyroid cancer risk. BEH‐TEBP was detected in less than 20% of samples and was not further analysed.

Mendy et al. (2023, 2024) assessed the association between the levels of BFRs in dust and respiratory disease (wheezing, respiratory infections, hay fever/allergies, lung function) in children at age 5 in the USA using a birth cohort (HOME study). Exposure was assessed prenatally (Mendy et al., 2024; n = 342) or at age 1 (Mendy et al., 2023; n = 234) via house dust concentrations (μg/g dust) and house dust loadings (μg/m²). As regards the prenatal exposure adjusted analyses, both prenatal BEH‐TEBP log₁₀ dust concentration (RR, 1.57; 95% CI, 1.28, 1.94) and prenatal BEH‐TEBP log₁₀ dust loading (RR, 1.28; 95% CI, 1.08, 1.52) were statistically significantly associated with wheeze. In the adjusted analyses, BEH‐TEBP log₁₀ loading at age 1 was statistically marginally significantly associated with hay fever/allergies (RR, 1.12; 95% CI, 1.00, 1.24). BEH‐TEBP log₁₀ loading at age 1 was also statistically significantly associated with peak expiratory flow (mL/min, −9.05; 95% CI, −17.67, −0.43) but not with forced expiratory volume. The same study was also used to evaluate associations between prenatal levels of BFRs and cord blood DNA methylation in an epigenome‐wide association study (Mendy et al., 2025; n = 252). This study found house dust concentrations of BEH‐TEBP to be associated with higher ATXN10 (cg07527826) methylation and with lower FAM159B (cg12662072) methylation. In the discussion of these results, the authors stated that the FAM159B gene is involved in cell signalling and regulation and was associated with psoriasis, body size and body impedance. They also stated that ATXN10 is a gene involved in neuron function and bone mineral density and height.

Zheng, Chen, et al. (2025) assessed the association between mixed serum exposure to dechlorane plus and novel brominated flame retardants and type 2 diabetes risk in a cross‐sectional case–control study in China, including 60 type 2 diabetes patients and 57 healthy controls. Serum levels of seven BFRs were measured (of which five are assessed in the current Opinion, see Appendix F). After adjusting for age, sex and high‐density lipoprotein, all mixture models consistently showed a significant positive association between mixed dechlorane plus/BFRs exposure and diabetes risk. Concentrations of BEH‐TEBP were not statistically significantly different between cases and controls in unadjusted comparisons (adjusted associations for single exposures were not reported) and BEH‐TEBP was not a main contributor to the mixture effects.

Zheng, Sun, et al. (2025) studied the association between BFRs and BMI in a cross‐sectional study set in Inner Mongolia, China (n = 77). Serum levels of seven BFRs were measured (of which five are assessed in the current Opinion, see Appendix F). Unadjusted correlations and linear regressions did not show an association between BEH‐TEBP and BMI. Adjusted models are not presented.

In summary, the available epidemiological evidence on BEH‐TEBP pertained to populations in China or the USA and it was very limited in terms of the total number of available studies (n = 7) and the number of available studies per endpoint (n = 1–2). Moreover, the exposure assessed in three out the seven studies was done in environmental samples without additional dietary/food data and studies were mostly cross‐sectional in design (apart from Mendy et al., 2023, 2024, 2025). No statistically significant associations were observed for thyroid function, apart from a lower FT4 in one cross‐sectional study. A statistically significant signal was observed for the association between BEH‐TEBP levels measured in dust and wheeze, hay fever/allergies and peak expiratory flow. One study reported alterations in DNA methylation.

3.2.4.2.6. Studies in other model systems and mode of action

In the previous Opinion, the only identified study with BEH‐TEBP was performed at concentrations up to 30 μM in chicken embryonic hepatocytes in vitro (Egloff et al., 2011). No effects on RNA expression levels or cell viability were observed. The previous Opinion also reported a study in which groups of fathead minnows (Pimephales promelas) were exposed via the diet to either of the technical products Firemaster 550 (BEH‐TEBP: 745 μg/g feed, EH‐TBB: 1658 μg/g feed) or Firemaster BZ‐54 (BEH‐TEBP: 907 μg/g feed, EH‐TBB: 2087 μg/g feed). The fish were fed 6% of their body weight per day resulting in daily intake of 45 mg/kg bw for BEH‐TEBP and 99 mg/kg bw for EH‐TBB in the FM 550 feed, and 54 mg/kg bw for BEH‐TEBP and 125 mg/kg bw for EH‐TBB in the BZ‐54 feed. The fish were dosed for 56 days followed by a 22‐day recovery period (Bearr et al., 2010). There were no effects on fish weight or length. Blood and liver samples were investigated for DNA strand breaks using the comet assay. A significant increase in percent tail DNA was reported in liver cells but not in blood cells (Bearr et al., 2010).

Since then, several studies have been identified (see Appendix G, Table G.1), and a brief summary is provided below.

Dose‐related increased serum TT4 and FT4 levels in rats following exposure to 50 and 500 mg/kg bw per day for 28 days (Yuan et al., 2024a) suggest interference with thyroid hormone regulation. The consensus of several in vitro studies examining effects on the thyroid hormone system is that BEH‐TEBP has low anti‐thyroid hormone and thyroid agonistic activities. However, BEH‐TEBP at a concentration of 3.75 μM inhibited DIO 3,3′‐T2 and T3 formation (Smythe et al., 2017). BEH‐TBPH treatment of human thyroid follicular epithelial Nthy ori3–1 cells resulted in increased mRNA and protein levels of TG, TPO and IL‐10 (at 100 nM) (Yuan et al., 2024a). Knock down of IL‐10 diminished the effects of BEH‐TBPH on these thyroid proteins, suggesting that IL‐10 is a target through which BEH‐TBPH influences thyroid function.

BEH‐TEBP and its carboxylic acid metabolites exhibited anti‐glucocorticoid activity by direct competition with the glucocorticoid receptor in mammalian cell models (Klopčič et al., 2016). No glucocorticoid agonistic activities were observed. BEH‐TEBP Activated PXR with EC50 values of 2.0 μM and showed PXR antagonist activity (IC50 = 13.9 μM). BEH‐TEBP also upregulated CYP3A4 mRNA expression which is PXR dependent. Zebrafish exposed to BEH‐TEBP via the water (0.2–200 nM) for up to 2 months showed dose‐dependent increase of visceral fat accumulation, which was significant at 2 nM exposure (Fu, Zhu, et al., 2024). Increased blood glucose, triglycerides and liver fat accumulation were observed and signs of NAFLD in males in which the liver transcriptome indicated upregulation of glycolysis and insulin‐signalling pathways. The CONTAM Panel noted that PPARγ dysregulation is a plausible mechanism for the metabolic effects of BEH‐TEBP. The phenotypes observed (visceral fat, hyperglycaemia, triglyceride elevation and NAFLD‐like liver) mirror classic PPARγ overactivation, with compensatory glycolysis and insulin‐signalling gene upregulation reinforcing metabolic stress (Gavrilova et al., 2003; Panasyuk et al., 2012).

Several neurotoxicity studies have been identified showing the potency of BEH‐TEBP to impair some mechanistical endpoints. In mice, BEH‐TEBP (20 or 200 mg/kg per day, 28 days, p.o.) impaired the hippocampal expression of proteins related to synaptic plasticity (BDNF, p‐CREB, PSD‐95) and cell signalling (GRP78, PERK and C/EBP homologous protein CHOP), and altered levels of oxidative stress markers (MDA, SOD and GSH) (Bao & Jing, 2021). The daily exposure of mice to BEH‐TEBP (20 mg/kg per day, 60 days, route of administration unclear) activated the extracellular signal‐regulated kinase 1/2 (ERK1/2) signalling pathway in hippocampus associated with oxidative stress responses and apoptosis (Xiong et al., 2024). Alternative in vivo models were also used to assess the neurotoxicity of BEH‐TEBP. In planarians (flatworms), BEH‐TEBP (0.01–100 μM) was tested in adult and regenerating animals and showed no effects in this model (Zhang et al., 2019). A study performed in wild‐type zebrafish larvae (Liu, Song, et al., 2022) showed behavioural alterations limited to a reduction of phototactic response and concomitant decreases in several opsin genes expression. The same study reported impairments in blood–brain barrier permeability in vivo using a transgenic zebrafish model exposed to BEH‐TEBP (0–20 μM per day, 5 days post‐fertilisation), and in vitro in hCMEC/D3 cells. BEH‐TEBP or its metabolite mono‐(2‐ethyhexyl) tetrabromophthalate were tested for their toxicity in human umbilical vein endothelial cells (Chen et al., 2020; Xiang et al., 2017). Results related to the cell cycle, apoptosis, cell viability and specific gene expression (eNOS, VEGF1, PECAM1) showed the ability of the metabolite to cause higher toxicity and to affect gene expression to a greater extent than the parent compound BEH‐TEBP. In SH‐SY5Y neuroblastoma cells, no effects on cell viability, mitochondrial function, neural plasticity and development, promotion of oxidative stress and inflammation of BEH‐TEBP have been reported after 2–24 h of exposure at doses ranging 2.5–20 μM, suggesting absence of toxicity in this neural cell line at the low concentrations tested (Esplugas et al., 2023).

Experiments in fish indicate an effect of BEH‐TEBP on reproduction. Feeding medaka a diet containing Firemaster 550 (containing BEH‐TEBP and EH‐TBB) resulted in downregulation of the vast majority of genes along the hypothalamus‐pituitary‐gonadal‐liver axis. There was an associated reduction in cumulative egg production. Exposure of zebrafish with BEH‐TEBP (4.00, 15.95 and 63.80 μg/g of food) for 21 days induced a dose‐related reduction of fertility in F0 females (Ma et al., 2018). Mating behaviour was reduced in male zebrafish exposed to low nanomolar concentrations of BEH‐TEBP in the water (Fu, Zhu, et al., 2024). Experiments on zebrafish sperm showed reduced cell viability, total mobility and swimming speed measurements at 0.01 μM BEH‐TEBP and higher concentrations. In silico docking of BEH‐TEBP to the human androgen receptor (AR) indicated strong binding (−119 kcal/mol) similar to that of the known AR agonist, flutamide (−122 kcal/mol). Whilst BEH‐TEBP showed no induction of AR‐mediated reporter gene expression in the hormone‐induced two‐hybrid androgen receptor yeast AR‐GRIP1 assay, it inhibited dehydroxytestosterone‐induced expression by 9%, 7%, 11% and 22% in the 0.125, 0.25, 0.5 and 1 μM exposure groups, respectively. BEH‐TEBP reduced mouse TM3 Leydig cell testosterone production at an exposure concentration of 200 μM with no effects at lower (0.2–20 μM) concentrations. Furthermore, BEH‐TEBP inhibited the proliferation of mouse Leydig cells by 1%, 3%, 7% and 22% at concentrations of 0.2, 2, 20 and 200 μM, respectively. These results suggest that BEH‐TEBP interferes with reproduction in zebrafish. Possible relevance to mammalian reproduction is indicated by in vitro experiments in mouse Leydig cells, which showed effects on androgen receptor (AR) signalling, transactivation experiments with the human AR in yeast and by in silico docking experiments with the human AR (Fu, Hua, et al., 2024, Appendix G).

3.2.4.2.7. Consideration of critical effects and dose–response analysis

The toxicological properties of BEH‐TEBP are not well‐characterised. No long‐term studies were identified.

In the absence of access to the original data reported in the REACH dossier, the CONTAM Panel could not make an independent evaluation of the possible genotoxicity of BEH‐TEBP, but noted that this compound does not have a harmonised classification for mutagenicity under the CLP Regulation (EC) No 1272/2008 based on the available data. The CONTAM Panel concluded that BEH‐TEBP does not induce gene mutations in bacteria and mammalian cells. There was evidence of clastogenicity in vitro at a concentration of 1000 μg/mL, although this was not confirmed in non‐oral in vivo in the erythrocyte micronucleus tests. However, it was not possible to verify whether the compound reached the bone marrow in the in vivo tests. Following the EFSA Opinion on genotoxicity (EFSA Scientific Committee, 2011), the Panel could not conclude on the genotoxic potential of BEH‐TEBP.

Studies conducted with BEH‐TEBP alone indicated possible effects on the liver in mice. Neurotoxicity studies pointed to induction of oxidative stress in the hippocampus, but produced contradictory neurobehavioural results possibly related to methodological differences (e.g. study design, route of administration, doses used and behavioural endpoints assessed) (Bao & Jing, 2021; Fulcher, 2018; Xiong et al., 2024).

Yuan et al. (2024a) reported dose‐related increases in serum TT4 and FT4 at 50 and 500 mg/kg bw per day, together with decreased TSH and decreased relative thyroid weight, and changes in thyroid histopathology in rats exposed to BEH‐TEBP for 28 days. However, the CONTAM Panel noted that no changes in histopathology were seen in a 90‐day OECD guideline study at doses up to 1000 mg/kg bw per day (Fulcher, 2018).

Administration of a binary mixture of BEH‐TEBP and EH‐TBB, at a combined dose level of approximately 3.3 mg/kg bw per day, to rats during pregnancy and lactation resulted in effects on socioemotional behaviour and bone. In view of the differences between the design and endpoints assessed in the studies with BEH‐TEBP alone and those with binary mixtures of BEH‐TEBP and EH‐TBB, it is not possible to attribute the effects to one or both of the components.

Three unpublished studies on a technical product containing 30% BEH‐TEBP and 70% EH‐TBB were summarised in a publication by Hays and Kirman (2017). According to Hays and Kirman (2017), two of these studies reported developmental and/or renal effects, but the data were not available to the CONTAM Panel and therefore could not be assessed. In the third study (MPI, 2008b), reduced body and spleen weight were reported in a two‐generational reproductive study, and the data were presented in the publication of Hays and Kirman (2017).

The available epidemiological studies did not provide sufficient evidence to characterise the hazards.

The only data suitable for dose–response modelling were the continuous data sets on spleen and body weights from the MPI (2008b), as cited by Hays & Kirman, 2017). The CONTAM Panel applied benchmark dose (BMD) modelling on these data, following the EFSA guidance on BMD modelling (EFSA SC, 2022), with a Benchmark Response (BMR) of 10% for body weight and a BMR of 15% for spleen weight based on the consideration in the JMPR guidance document that increases of less than 10% or 15% of body weight or spleen weight might be non‐adverse (JMPR, 2015). The outcome is shown in Table 11). The lowest calculated BMDLs calculated on the combined dose of BEH‐TEBP and EH‐TBB were for the female F1 generation, at 31 mg/kg bw per day for bodyweight, corresponding to 10 mg/kg bw per day BEH‐TEBP.

TABLE 11.

Benchmark dose (BMD) modelling for the study of Firemaster BZ‐54 (70% EH‐TBB and 30% BEH‐TEBP) in rats (for details of the BMD analyses see Annex D).

Study	Observed effect	BMDL10	BMD10	BMDU10
		mg/kg bw per day
MPI (2008b), as cited by Hays & Kirman, 2017)	Body weight offspring a
	F1 male	37	95	215
	F1 female	31	110	390
	F2 male	120	147	177
	F2 female	49	92	166
		BMDL 15	BMD 15	BMDU 15
	Spleen weight offspring a
	F1 male	78	124	184
	F1 female	56	98	174
	F2 male	94	133	174
	F2 female	59	101	177

^a Sex used as covariate.

The CONTAM Panel noted the lack of consistency in the findings of different studies, possibly related to different study designs and endpoints examined.

The findings reported at the lowest dose levels were on neurobehavioural and skeletal effects in the study of a binary mixture of BEH‐TEBP and EH‐TBB in rats (Schkoda et al., 2023; Witchey et al., 2020, 2023), which provided a BEH‐TEBP dose of 1.05 mg/kg bw per day. The CONTAM Panel noted that the greater solubility and oral absorption of EH‐TBB compared to BEH‐TEBP (see Section 3.2.4.2.1 and 3.2.5.2.1) indicates that the effects are more likely to be attributable to EH‐TBB. However, in the absence of information on relative potency, the CONTAM Panel does not know what the relative contributions of the two substances are, and therefore to be conservative made a worst‐case assumption for each substance. Accordingly, in order to identify a Reference Point specifically for BEH‐TEBP, the Panel attributed all of the findings to BEH‐TEBP and adopted the LOAEL value of 1 mg/kg bw per day (rounded) as the Reference Point, noting that this approach is highly conservative.

The Panel concluded that, due to the limitations and uncertainties in the current database, the establishment of a health‐based guidance value for BEH‐TEBP was not appropriate. Instead, the MOE approach was used for risk characterisation (EFSA Scientific Committee, 2025).

Usually, an MOE of 100, covering variability with respect to kinetic and dynamic differences between animal species and humans (factor 4 × 2.5 = 10) and within the human population (factor 3.2 × 3.2 = 10), is considered sufficient to conclude that exposure does not raise a health concern.

To allow for the absence of a NOAEL and the necessity of using a LOAEL, the EFSA Guidance on selected default values indicates that an additional uncertainty factor will be needed, the size of which should be determined on a case‐by‐case basis (EFSA Scientific Committee, 2012). ECHA, in its Guidance on information requirements and chemical safety assessment (ECHA, 2012) proposed the application of ‘an assessment factor ranging from 3 (as minimum/majority of cases) to 10 (as maximum/exceptional cases)’ with the following considerations: ‘An AF of 3 may be more appropriate for instance in situations, where the effects at the LOAEL are mild, or the LOAEL represents the lower boundary of the exposure range in which the effect is observed. Higher numerical values should be considered in situations where the effects at the LOAEL are severe and irreversible, or the shape of the dose–response curve is shallow or the quality of the study (e.g. group sizes, statistical methods, study design, exposure data) gives rise to uncertainties about the reliability of the identified LOAEL’ (ECHA, 2012). In this instance, since the effects at the LOAEL for BEH‐TEBP are mild, the CONTAM Panel considered that a factor of 3 is sufficient to extrapolate from a LOAEL to NOAEL.

According to the EFSA Scientific Committee Guidance on selected default values, an additional factor can be considered in case of deficiencies in the database on a case‐by‐case basis. A default value has not been proposed, as it will be directly dependent on the data set available (EFSA Scientific Committee, 2012). The WHO/IPCS (1994, 1999) has recommended a factor of 3 or 5 if there are minor deficiencies in the database and a factor of 10 if there are major deficiencies in the database. There are limitations in the database for BEH‐TEBP, such as inconsistencies between studies of non‐standard design, most studies being on mixtures rather than on the pure substance, and some studies only being available in summary form, so that they could not be fully assessed by the Panel. Taking into account also the sensitivity of the endpoints examined in the critical study, and the worst‐case approach of attributing the effects solely to BEH‐TEBP, the CONTAM Panel considered an additional factor of 5 to be appropriate for deficiencies in the database.

Overall, the CONTAM Panel considered that MOEs ≥ 1500 (100 for inter and intra‐species variability, 3 for LOAEL to NOAEL extrapolation and 5 for the gaps in the toxicological database) do not raise a health concern for non‐neoplastic effects.

The CONTAM Panel could not set a Reference Point for neoplastic effects, since no carcinogenicity data are available. In addition, it was not possible to conclude on the genotoxic potential of BEH‐TEBP.

3.2.4.3. Occurrence data and dietary exposure assessment

Limited occurrence data in food were submitted to EFSA (see Section 3.1.1) and no estimation of the dietary exposure based on these data was possible.

In the scientific literature, several studies were identified reporting on the occurrence of BEH‐TEBP and an overview is given in Section 3.1.2 and Appendix B.

Two studies were identified reporting estimates of dietary exposure to BEH‐TEBP and these are summarised below:

• Exposure estimates for BEH‐TEBP in the UK population were calculated based on daily consumption data for UK citizens from the National Diet and Nutrition Survey (2008–2019) and occurrence data from samples collected in 2020–2021 (Ma et al., 2023). Average dietary exposures to BEH‐TEBP ranged from 0.91 to 3.3 ng/kg bw per day across population groups (from 0 up to > 75 years old), while high‐end exposure estimates, assuming food contaminated at the average concentrations consumed at the mean rate + 2 standard deviations, ranged from 5.3 to 20 ng/kg bw per day (Ma et al., 2023).

• Exposure estimates were reported in the Norwegian adult population for BEH‐TEBP based on 24‐h duplicate diet method from samples collected in 2013–2014. Dietary intake of BEH‐TEBP was 1.2 ng/kg bw per day, in the 95^th percentile. Intake of red meat, poultry products and fruits was weakly correlated to dietary exposure to BEH‐TEBP. The authors suggested that dietary exposure to BEH‐TEBP from fruit intake might be related to the usage of this compound as a plasticiser in wrapping film and sheeting used as food packaging (Xu et al., 2017).

For breastfed infants, there were insufficient data in human milk from European countries to make robust exposure estimates. To provide an indication of the potential exposure, the CONTAM Panel made an estimation for average and high consumption of human milk (see Section 2.4) considering the available data reported in the literature (Čechová et al., 2017; Tao et al., 2017, see Section 3.2.4.2.2).

The exposure scenario based on average human milk consumption and the highest reported mean concentration in human milk across three European countries (Norway, the Netherlands and Slovakia) of 1.34 ng/g lipid (Čechová et al., 2017), resulted in an estimate of 6.15 ng/kg bw per day. The study by Tao et al. (2017) reported a lower mean concentration in samples from the UK of 0.25 ng/g lipid that would result in an average exposure of 1.15 ng/kg bw per day.

The exposure scenario based on high human milk consumption and the highest reported P95 concentration of 2.75 ng/g lipid (Čechová et al., 2017), resulted in an estimate of 18.9 ng/kg bw per day. The study by Tao et al. (2017) reported only maximum concentrations that were not used for these estimations.

It should be noted that these estimates are based on human milk samples from four European countries that were collected one or two decades ago. It is unknown whether this is representative of current exposure to this BFR.

Non‐dietary exposure

Several studies were identified on the non‐dietary exposure to BEH‐TEBP. Whilst exposure estimates were not made, child car seats were identified as a source of exposure to BEH‐TEBP with concentrations up to 39,000 ng/g (Harrad et al., 2023).

A range of BFRs including BEH‐TEBP were measured in a variety of black plastic kitchen utensils by Kuang et al. (2018). Concentrations up to 30,000 ng/g were found and it was estimated that exposure via cooking in median and high exposure scenarios could be as high as 1651 ng per day.

Estimated daily dermal exposure to BEH‐TEBP for adults from hand wipe and settled dust data (direct and indirect methods used to assess dermal exposure) were 61 pg/kg bw per day and 23 pg/kg bw per day at the 95th percentile high exposure scenario (Tay et al., 2018).

Silicone (polydimethylsiloxane or PDMS) wristbands and cotton T‐shirts were used to assess the exposure of e‐waste recyclers in Dhaka, Bangladesh to a variety of BFRs by Wang, Peris, et al. (2020). The median daily exposure doses via dermal transfer from the front of the T‐shirt to the front body trunk was in the range 0.005–0.12 ng/kg bw per day with a median value of 0.02 ng/kg bw per day for BEH‐TEBP.

Some estimates of exposure to BEH‐TEBP from dust have been made and these are collated in Appendix D. They are typically < 1 ng/kg bw per day. These non‐dietary sources of exposure to BEH‐TEBP are several orders of magnitude smaller than estimates of dietary exposure (Section 3.2.4.4 above). Exposure to BEH‐TEBP from dust is also similarly lower than exposure from the diet (Appendix D).

In a study by Sjöström et al. (2024), to study children's exposure to halogenated flame retardants and organophosphate esters through dermal absorption and hand‐to‐mouth ingestion in Swedish preschools, it was found that BEH‐TEBP was one of the highest contributors to exposure in terms of mass. Exposure was estimated using handwipes and 2.8 ng BEH‐TEBP per handwipe was found.

In summary, non‐dietary sources of exposure to BEH‐TEBP were lower than estimates made for dietary exposure reported in the literature (ranging from 0.91 to 3.3 ng/kg bw per day).

3.2.4.4. Risk characterisation

Insufficient occurrence data in food on BEH‐TEBP were submitted to EFSA to allow robust estimates of dietary exposure for the EU population to be made. Limited data obtained from the scientific literature indicated that dietary exposure from food in two European countries (UK and Norway, see Section 3.2.4.3) was up to 20 ng/kg bw per day at P95 exposure. The Panel noted this would result in MOEs in the region of 10⁶, which does not raise a health concern for non‐neoplastic effects. The indicative exposure estimates obtained for breastfed infants (see Section 3.2.4.3) would result in similar MOEs.

Regarding potential neoplastic effects, the EFSA guidance on the threshold of toxicological concern (TTC) specifies that, for substances raising a concern about potential DNA‐reactive mutagenicity and/or carcinogenicity, if exposure is below 2.5 ng/kg bw per day, the substance is considered of low probability of adverse health effects (EFSA Scientific Committee, 2019). Although the CONTAM Panel could not conclude on the genotoxic potential of BEH‐TEBP, the Panel noted that the dietary exposure from two studies from Europe obtained from the literature are above the TTC for genotoxic compounds, and therefore further genotoxicity information is needed.

3.2.4.5. Uncertainty analysis

The uncertainty analysis was performed as described in Section 2.8. Uncertainties affecting the hazard assessment for BEH‐TEBP are listed in Table 12.

TABLE 12.

Uncertainties identified and their impact on the outcome of the hazard identification/characterisation and risk characterisation of BEH‐TEBP.

Description of the uncertainty		Impact of the uncertainty on the hazard identification and characterisation a
Chemical composition and analytical methods
Dosing and chemical composition, including impurities	Uncertainty associated with the dose in the critical studies used in the risk assessment	2 – Medium impact. The critical study is a study in which a binary mixture was tested. A conservative (worst‐case) assumption was made that the effects were solely due to BEH‐TEBP.
Hazard identification and characterisation
ADME	ADME in relation to the critical studies	1 – Low impact: Several studies showed that BEH‐TEBP is poorly absorbed and eliminated mainly by the faecal route as a parent compound.
	Information on transfer rate to animal products	0 – Negligible impact. No studies identified, but no impact on risk assessment.
	Transfer via mother's milk	1 ‐ Low impact. BEH‐TEBP can be transferred during lactation and gestation.
Toxicity studies in experimental animals: endpoints and study design	Limitations in the design of the studies that can result in uncertainties	2 – Medium impact. The number of studies is limited, most have a non‐standard design and/or tested mixtures of BFRs, and findings are inconsistent. However, the effects seen at lowest doses represent sensitive endpoints.
	Relevance for humans of the adverse effect.	0 – Negligible impact: The effects seen at the lowest doses are on behaviour and bone. These are considered relevant for humans even though there are no supporting epidemiological data.
Genotoxicity	Uncertainty in the assessment of genotoxicity	3 – High impact. The CONTAM Panel did not have access to the original data and could not conclude on the genotoxic potential of BEH‐TEBP (see Section 3.2.4.2).
MOA	Uncertainties associated with the data on mode of action	1 – Low impact. There is limited information on the mode of action of BEH‐TEBP for the critical effects.
Selection of Reference Point	Weaknesses in the critical study	2 – Medium impact. Single dose level study. Study with binary mixture.
	Weaknesses in non‐critical studies, or absence of studies, and uncertainty about whether the endpoints they tested might have been critical if the weaknesses were not present	2 – Medium impact. Inconsistencies between studies of non‐standard design, most studies being on mixtures rather than on the pure substance, and some studies only being available in summary form, so that they could not be fully assessed by the Panel. No long‐term or carcinogenicity studies. These deficiencies in the database were addressed by including an additional uncertainty factor of 5 in the MOE that would not raise a health concern (see below).
Description of the uncertainty		Impact on the risk characterisation a
Margin of exposure (MOE)	–	1 – Low impact. The usual factor of 100 (to cover for variability with respect to kinetic and dynamic differences between animal species and humans, and within the human population) is considered sufficient. The additional factors of 3 (to account for the absence of a NOAEL) and 5 (for deficiencies in the database, as mentioned above), are considered sufficient.
Risk characterisation (non‐neoplastic effects)	–	1 – Low impact. Not possible for EFSA to perform an exposure assessment since insufficient occurrence data were submitted to EFSA. Based on dietary exposure estimates reported in the literature for two European countries (UK and Norway), the resulting MOEs did not raise a health concern. It is uncertain whether this is representative of the situation in other European countries. The uncertainty affecting the exposure estimates from the literature is of large magnitude, but its impact on the risk characterisation is low due to the very large size of the calculated MOE.

Abbreviations: ADME, absorption, distribution, metabolism, excretion; BMD, benchmark dose; BMDL, benchmark dose lower credible limit; BMR, benchmark response; MOA, mode of action; TK, toxicokinetic.

^a 0 – Uncertainty with negligible impact; 1 – Uncertainty with low impact; 2 – Uncertainty with medium impact; 3 – Uncertainty with high impact. This ordinal scale provides a relative assessment of the impact of the individual uncertainties as an aid to quantifying their combined impact by expert judgement, as described in section 11 of EFSA Scientific Committee (2018).

Uncertainty analysis: Genotoxicity

The CONTAM Panel did not have access to the original data and could not conclude on the genotoxic potential of BEH‐TEBP (see Section 3.2.4.2), so no further analysis of the uncertainties was performed.

Uncertainty analysis: Hazard characterisation

The CONTAM Panel identified substantial uncertainties affecting the hazard assessment (Table 12) that did not allow the establishment of a HBGV. Instead, the MOE approach was used, addressing the hazard uncertainties by applying UFs of 100 (to cover for variability with respect to kinetic and dynamic differences between animal species and humans, and within the human population), 3 (to account for the absence of a NOAEL) and 5 (for deficiencies in the database).

Considering the combined impact of the uncertainties and the extent to which they are addressed by the applied UFs, the CONTAM Panel judged with 90%–99% certainty (i.e. probability) that, when using the Reference Point of 1 mg/kg bw per day, MOEs ≥ 1500 would not raise a health concern for non‐neoplastic effects.

Uncertainty analysis: Risk characterisation

Insufficient occurrence data in food on BEH‐TEBP were submitted to EFSA to allow robust estimates of dietary exposure for the European population to be made. Based on the dietary exposure data from food from the scientific literature from two European countries (UK and Norway, see Section 3.2.4.3), the CONTAM Panel noted this would result in MOEs in the region of 10⁶.

Considering these very large MOEs, and the uncertainties affecting the exposure estimates reported from two European countries, the CONTAM Panel judged with 95–99% certainty (i.e. probability) that current exposure of the European population to BEH‐TEBP does not raise a health concern for non‐neoplastic effects.

3.2.5. 2‐Ethylhexyl 2,3,4,5‐tetrabromobenzoate (EH‐TBB) [CAS number 183658‐27‐7]

3.2.5.1. Production, use, analysis and environmental fate

3.2.5.1.1. Production and use

The previous EFSA Opinion (EFSA CONTAM Panel, 2012b) reported that EH‐TBB was found in several additive flame‐retardant products used as replacements for PBDE mixtures. It is one of the two major components (35%) of Firemaster 550, which is used as a replacement for PentaBDE in polyurethane foam applications (de Wit et al., 2011) and can also be found in Firemaster BZ‐54 (70%) (Ma et al., 2011). In both commercial products BEH‐TEBP (see Section 3.2.4) is also present (de Wit et al., 2011; Ma et al., 2011). Stapleton et al. (2008) reported that the EH‐TBB/BEH‐TEBP ratio observed in Firemaster 550 commercial mixture is generally 4:1 by mass, and studied the combined occurrence of both compounds in foam samples taken from different baby products, e.g. car seats, changing table pads sleep positioners, mattresses and pillows (Stapleton et al., 2011). It was also found that all samples containing EH‐TBB and BEH‐TEBP also contained triphenylpolyphosphate, which is another ingredient in Firemaster 550.

3.2.5.1.2. Analysis

It was reported in the previous Opinion that EH‐TBB can be analysed using the same methods as for PBDEs, BEH‐TEBP and DBDPE (Covaci et al., 2011; Papachlimitzou et al., 2012). It can be extracted using toluene, dichloromethane or mixtures of solvents (e.g. hexane:acetone or MTBE:acetone). Clean‐up is usually done by GPC, sometimes combined with silica column fractionation, silica gel or florisil where a more polar solvent is required for elution. Carbon columns can also be used to separate EH‐TBB. Instrumental determination is usually done by some type of GC–MS or LC–MS.

3.2.5.1.3. Environmental fate and occurrence in the environment

There is evidence that like many other BFRs, EH‐TBB is persistent and bioaccumulates in wildlife and at higher trophic levels (e.g. Lee et al., 2019; Lee & Kwon, 2016). In a study reported by de Jourdan et al. (2013) described earlier under BEH‐TEBH, on the environmental fate and persistence of three novel BFRs in aquatic mesocosms, EH‐TBB was the other main component of the commercial mixture Firemaster BZ‐54 (see Section 3.2.4).

Occurrence in the environment

Soil

No data were identified from the EFSA literature search for EH‐TBB in European soils. A review of soil contamination by PBDEs and other BFRs by McGrath et al. (2017) showed that EH‐TBB was generally not found, but had been reported in soil close to a waste disposal site in Australia at concentrations ranging from < LOD (0.02 ng/g dry weight) to 1.79 ng/g (mean 0.3 ng/g) (McGrath et al., 2017). It had also been reported in soils from a forested area of China by Zheng et al. (2015) in concentrations ranging from < LOD (0.0024 ng/g) to 1.6 ng/g with a mean of 0.184 ng/g, and in soil from around an e‐waste processing workshop in Viet Nam (Someya et al., 2016) at concentrations between < LOD and 0.32 ng/g. EH‐TBB was not found in soils at the other sites investigated.

EH‐TBB was measured in soil from Stockholm, Sweden, in a study by Newton et al. (2015) identified by a literature search conducted by ECHA (2024). EH‐TBB was below the LOD in all samples; the LOD was stated as 3.2 pg per sample and samples of 1.5 g were initially taken although some variation can occur due to the amount of moisture and organic content.

Sediment

As for soils, data on EH‐TBB could only be found in the search conducted by EFSA for sediment samples taken from outside of Europe. Chokwe et al. (2019) reported levels of 0.5–17 ng/g with a mean of 3.25 ng/g for sediment from the Vaal River in South Africa.

La Guardia et al. (2013) investigated inland and coastal sediments in the Thekwini Metropolitan Municipality, South Africa and found EH‐TBB at concentrations up to 13,900 ng/g with a mean of 545 ng/g.

Olukunle and Okonkwo (2015) measured BFRs in leachates and sediments from selected municipal solid waste landfill sites in Gauteng Province, South Africa. In sediment, concentrations of EH‐TBB ranged from < LOD to 56 ng/g with a mean of 10 ng/g; in leachate concentrations were 8.7 pg/L to 142 pg/L with a mean of 60 pg/L.

The ECHA literature search for the report on aromatic BFRs (ECHA, 2024) identified reports of EH‐TBB in sediment from the North Sea and River Elbe in Germany at concentrations up to 0.258 ng/g dry weight.

Dust

As mentioned earlier (Section 3.2.4.1.3), for many BFRs dust can be an important source of exposure, and this is especially the case for toddlers and infants due to low body weight, high crawling activity and high frequency of hand‐to‐mouth contact. A summary of reported occurrence of EH‐TBB in dust samples from European countries and estimates of exposure via dust when reported by the authors, is given in Appendix D (Table D.2). Concentrations reported are typically a few tens of ng/g but a value as high as 690 ng/g was reported by Larsson et al. (2018) for dust from pre‐school environments. Fromme et al. (2014) gave the highest exposure estimate for EH‐TBB from dust that could be identified which was 63 ng/kg bw per day for toddlers from Germany.

Given that reported estimates of exposure from diet are all less than 2 ng/kg bw per day (see Section 3.2.5.4 below), dust may be an important source of exposure for EH‐TBB at least for some individuals.

Water

No studies reporting values for concentrations of EH‐TBB in European water were identified by the EFSA literature search. Outside of Europe, Shanmuganathan et al. (2018) reported levels of up to around 500 ng/L in influent water at Canadian water plants.

Wang et al. (2017) reported on novel brominated flame retardants in seawater near a coastal mariculture area of the Bohai Sea, China. The concentration of EH‐TBB was found to be in the range of 0.81–6.69 ng/L in seawater. This was compared to a value of 1.3–7.9 pg/L in water from the Great Lakes (Venier et al., 2014).

The ECHA investigation report on aromatic BFRs identified three studies from Europe (ECHA, 2024). Carlsson et al. (2018) measured new brominated flame retardants including EH‐TBB in water from the Arctic but concentrations were below the LOD which was not stated (LOQ for water was stated to be 0.2 pg/L). Valls‐Cantenys et al. (2013) reported on the application of polydimethylsiloxane rod extraction to the determination of 16 halogenated flame retardants in water, including river water and Ria and Sea water from coastal areas. EH‐TBB was not detected in any sample using the method which had a LOD of 1.3 ng/L.

Wildlife

There are only limited data for EH‐TBB in wildlife worldwide. Berger et al. (2023) reported concentrations of 0.34–37 ng/g lipid in the blubber of harbour seal pups and 1.8–41 ng/kg lipid in liver of harbour seal pups from the NorthWest Atlantic region.

Lewis et al. (2020) measured BFRs including EH‐TBB in Adelie penguin colonies in East Antarctica. Most samples were < LOD but concentrations of up to 0.08 ng/g dry weight were found.

Gentes et al. (2012) found EH‐TBB in liver of ring‐billed gulls nesting in the St Lawrence River, downstream from Montreal, Canada at concentrations ranging from < LOD–1.55 ng/g ww. All samples of plasma were < LOD.

EH‐TBB was found in fewer than half of glaucous gulls from Cape Dorset in NU (Eastern Canadian Arctic) at concentrations up to 0.92 ng/g ww.

Vorkamp et al. (2018) measured halogenated flame retardants in peregrine falcon eggs from Greenland. EH‐TBB was found in 54% samples at concentrations between < 0.620 and 54.3 ng/g lipid with a mean concentration of 4.61 ng/g lipid.

The ECHA literature search for the report on aromatic BFRs also identified reports of EH‐TBB in whale, seal, bear, gull from northern Norway at concentrations up to 1050 pg/g lipid and in eelpout and blue mussels from the North Sea, Germany at concentrations of up to 14.3 pg/g (ECHA, 2024).

3.2.5.2. Hazard identification and characterisation

3.2.5.2.1. Toxicokinetics

No information relative to toxicokinetic in mammals had been described in the previous Opinion (EFSA CONTAM Panel, 2012b). Since then, several studies have been identified.

Roberts et al. (2012) investigated the metabolism of EH‐TBB in human and rat liver microsomes. The authors found that EH‐TBB was metabolised to 2,3,4,5‐tetrabromobenzoic acid (TBBA).

Hoffman et al. (2014) collected in adult volunteers (n = 64), urine, hand wipe samples and dust samples from their homes where FM550 was detected in the home environment. The authors indicated that urinary TBBA could be a useful biomarker of Firemaster 550 exposure.

Hoffman et al. (2024) analysed the urine from 180 children (mean age: 4.5 years, ranging from 38 to 73 months) for TBBA and the serum for 11 BFRs. The authors collected dust, handwipe, wristband and foam in the home environment and analysed them for the same BFRs. The authors found a significant association between concentrations of EH‐TBB in dust, handwipe and wristband with the detection of TBBA in urine.

Bearr et al. (2012) studied the in vitro metabolism of the flame retardant mixture, Firemaster BZ‐54 containing EH‐TBB (as well BEH‐TEBP), with hepatic subcellular fractions from mice and fathead minnow, common carp and snapping turtle. The authors found that BEH‐TEBP was metabolised to greater extent in mice compared to other species (fathead minnow and carp) (metabolites have not been characterised).

Phillips et al. (2016) studied the maternal transfer of Firemaster 550 components containing EH‐TBB (as well as BEH‐TEBP) in Wistar rat. Rats were orally exposed to approximately 1 or 3.3 mg/kg bw per day of Firemaster 550 for 10 consecutive days during gestation (GD9–18) or lactation (PND3–12). The authors concluded that EH‐TBB accumulates in tissue largely via lactational transfer than during gestational transfer.

Knudsen et al. (2016) investigated the toxicokinetic of EH‐TBB in female Sprague–Dawley rats and male B6C3F1 mice by administering a single dose and repeated exposure of [¹⁴C]‐EH‐TBB by gavage and iv route at 0.1, 1, 10 or 100 mmol/kg (0.05–55 mg/kg). In female rats following oral administration, the authors showed that EH‐TBB was well absorbed (85%) in comparison to iv‐injected rats. No accumulation of EH‐TBB or its metabolites was observed in tissues after single administration of up to 100 mmol/kg or repeated administration of 0.1 mmol/kg to rats. EH‐TBB was metabolised into TBBA, TBBA‐glycine and TBBA‐sulfate that were eliminated primarily in urine. In faeces, only EH‐TBB and TBBA were found, and TBBA‐sulfate was detected. More than 90% of administered [¹⁴C]‐radioactivity at all dose levels was eliminated in urine and faeces within 48 h, and less than 1% remained in the tissues after 72 h. At 0.1 mmol/kg, based on radioactivity recovery in excreta, the excretion in urine and faeces were 51% and 47% of the dose after 72 h, respectively. Similar excretion pathways were observed in male B6C3F1/Tac mice at 0.1 mmol/kg. From the data collected, the authors performed on a one‐compartment model, and calculated an absorption half‐life of 4 h, a distribution/elimination half‐life of 5 h with a clearance rate of 35 ± 10 mL/h.

Chen et al. (2020) administered to female Balb/c mice by gavage (8 μg/mL) and searched the metabolites after 4 and 8 h post exposure. Compared to BEH‐TEBP, the authors indicated that EH‐TBB was metabolised into TBBA.

In summary, following oral administration in rodents, EH‐TBB is well absorbed, readily metabolised and eliminated by both urinary and faecal routes.

3.2.5.2.2. Levels in humans

No information on levels of EH‐TBB in human samples was reported in the previous Opinion. Since then, several studies have been identified reporting concentrations of EH‐TBB as well as other BFRs covered in this assessment in samples from European countries. Details of these studies are reported in Appendix E (Table E.1), and a summary of the results related to EH‐TBB are reported below.

Regarding human milk samples,

• In the study by Čechová et al. (2017) described above, EH‐TBB was analysed in human milk samples from Norway (2003–2006, n = 305), the Netherlands (2011–2014, n = 116) and Slovakia (2011–2012, n = 37) with a detection frequency between 47% and 69%. Mean and P95 concentrations across countries ranged from 0.037 to 0.240 ng/g lipid and from 0.091 to 0.683 ng/g lipid, respectively.

• Tao et al. (2017) analysed human milk samples collected in the UK in 2010 (n = 25) and 2014–2015 (n = 10) for several BFRs. In 2010, the levels of EH‐TBB ranged from < 0.01 to 2.1 ng/g lipid (detection frequency 44%), while in 2014–2015 a higher detection frequency (90%) was reported and concentrations ranged from < 0.01 to 0.48 ng/g lipid, with a mean value of 0.21 ng/g lipid.

• Sahlström et al. (2015) did not detect EH‐TBB above the LOD (< 13 pg/g ww) in two pooled human milk samples from Sweden collected in 2009–2010.

Regarding levels in blood/serum,

• Sahlström et al. (2014) did not detect EH‐TBB above the LOD (10 pg/25 μL) in any of the 24 matched serum samples from mothers and their toddlers collected in Sweden between 2009 and 2010.

• In the study by Tay et al. (2019) on 61 serum samples collected in 2013 from a Norwegian cohort, EH‐TBB was also reported below the LOD in all samples (LOD = 36 pg/serum sample, volume not reported).

• Sales et al. (2017) detected EH‐TBB at 50 pg/g serum in one pool of human cord blood from Belgium.

3.2.5.2.3. Toxicity in experimental animals

Studies considered in the previous EFSA assessment

In the previous Opinion (EFSA CONTAM Panel, 2012b), no experimental data were identified on EH‐TBB tested as a single chemical.

Studies published since the previous EFSA assessment

Since the previous Opinion, the only study to consider EH‐TBB alone, was the NTP comparative study reported above for BEH‐TEBP (see Section 3.2.4.2) (NTP ³⁰ ; Shockley et al., 2020a). In this case, male Harlan Sprague–Dawley rats were administered EH‐TBB for 5 days via oral gavage in corn oil at doses of 0, 0.05, 0.55, 5.5, 55 and 550 mg/kg bw per day (n = 6). There were no dose‐related effects on survival, body weight, gross pathology or histopathology. The relative heart weight was decreased at doses of 0.55 mg/kg bw per day and higher, with a very flat dose–response relationship (less than 10% decrease at all doses). In view of the very small change over a wide range of doses (three orders of magnitude), the CONTAM Panel considered this effect not to be adverse and possibly incidental. The relative kidney weight was also reduced, but at the highest dose only. A few significant changes in haematology (reticulocytes, monocytes, large unstained cells) and clinical chemistry (bile salts, sodium, chloride, albumin) were found but these seem to be unrelated to the treatment. The liver and thyroid gland of the control and highest dose group were examined for neoplastic and non‐neoplastic lesions and none were found. Both T3 and FT4 concentrations in serum were reduced at the highest EH‐TBB dose of 550 mg/kg bw per day. However, the observed numerical decrease in TT4 was not statistically significant.

As mentioned in Section 3.2.4.2.3, new data were also available for studies that considered binary mixtures of EH‐TBB and BEH‐TEBP (Schkoda et al., 2023; Witchey et al., 2020, 2023), or the technical product Firemaster BZ‐54 which contains EH‐TBB and BEH‐TEBP (Hays & Kirman, 2017) and these are described in Section 3.2.4.2.3.

Studies on a technical product containing EH‐TBB, i.e. Firemaster 550, were also identified (see Section 3.2.4.2.3). As for BEH‐TEBP, the CONTAM Panel concluded that data on this technical product added little to the body of evidence for the effects of EH‐TBB and were not suitable for identifying a Reference Point for use in risk characterisation.

Summary of toxicity in experimental animals

From the single study with exposure to EH‐TBB alone, few effects were reported, i.e. reduced T3 and free T4 at 550 mg/kg bw per day, with a NOAEL of 55 mg/kg bw per day (Shockley et al., 2020a).

A LOAEL of 3.3 mg/kg bw per day has been reported for developmental behavioural and skeletal effects of the binary mixture of EH‐TBB and BEH‐TEBP, in which EH‐TBB contributed 2.25 mg/kg bw per day (Schkoda et al., 2023; Witchey et al., 2020, 2023). From a study on a technical product containing 70% EH‐TBB and 30% BEH‐TEBP, the CONTAM Panel calculated a BMDL₁₀ of approximately 30 mg/kg bw per day for decreased bodyweight and spleen weight in a two‐generation study (MPI, 2008b, as reported in Hays & Kirman, 2017), corresponding to 20 mg/kg bw per day EH‐TBB.

No long‐term studies, reproductive toxicity studies or experimental data on the genotoxic potential of EH‐TBB were identified.

3.2.5.2.4. Computational analysis

The QSAR Toolbox did not identify any experimental data for the Ames test on EH‐TBB (see Annex B). VEGA models predicted negative Ames test results for EH‐TBB with ‘MODERATE reliability’ (four models). These predictions were not further considered due to their limited reliability (see Section 2.2).

3.2.5.2.5. Read‐across

In the absence of experimental data on the genotoxic potential of EH‐TBB, the CONTAM Panel explored the possibility to perform read‐across from BEH‐TEBP, PBB‐Acr and HEEHP‐TEBP (for details on these compounds see Section 3.2.4, 3.3.11 and 3.3.17, respectively) regarding their structural similarity obtained from the Similarity Matrix reported by ‘istSimilarity’ (tool available in VEGA HUB platform;¹⁹ Floris et al., 2014): profiling results from the OECD QSAR Toolbox indicated endpoint specific similarity for the four substances, which do not trigger any DNA binding alerts. (Q)SAR predictions from VEGA software indicated that all four of these substances would be non‐mutagenic in the Ames test, noting the ‘MODERATE reliability’ of the predictions for EH‐TBB. See Annex B (Table B.4) for details.

However, the CONTAM Panel noted that the uncertainty in the read‐across conclusion would be too high considering: (i) the doubts on in vitro clastogenicity for BEH‐TEBP (see Section 3.2.4) and PBB‐Acr (see Section 3.3.11) (contrasting results) and (ii) the major differences in kinetics (absorption) between BEH‐TEBP and EH‐TBB, which do not enable to extrapolate the results of the in vivo clastogenicity/aneugenicity studies of BEH‐TEBP to the target compound EH‐TBB. Moreover, the Panel was unable to conclude on the genotoxic potential of BEH‐TEBP.

3.2.5.2.6. Observations in humans

In the previous Opinion, no human studies were identified. Since then, six publications representing four epidemiological studies were identified. In Appendix F, an overview of the epidemiological studies identified reporting on the BFRs considered in this Opinion, including those on EH‐TBB, is shown.

Various endpoints were assessed including thyroid hormone profile (three studies), thyroid cancer (one study) and respiratory disease (one study). Exposure was assessed either in ambient air and wristbands (one study), in dust (one study) or in serum (two studies).

In the study by Guo et al. (2018) on the association with thyroid hormone levels in 10‐year‐old students living near a petrochemical complex in South China (for details see description in Section 3.2.4.2), EH‐ TBB levels were associated with lower FT4 levels (β −0.48, 95%CI ‐0.80, −0.15), but no associations were found for T3, T4, FT3 and TSH.

In the study by Wang, Romanak, et al. (2019) that assessed the association between several BFRs and thyroid hormone profile (FT3, FT4 and TSH) in rural Central Appalachia (USA) (see details in Section 3.2.4.2), EH‐TBB and BEH‐TEBP were the most abundant novel BFRs in wristbands, with median concentrations of 87 and 31 ng/g wristband, respectively. The combined contribution of these two chemicals to the sum of 10 BFR concentrations was 94%. Among the multiple analyses performed, no statistically significant associations with the thyroid hormone profiles were reported for EH‐TBB.

In the study by Liu, Li, et al. (2022) that assessed the association between the levels of several BFRs in serum and thyroid cancer risk in Shandong Province (China) (for details see description in Section 3.2.4.2), EH‐TBB was detected in 21% of the control population (geometric mean, 0.1 ± 3.0 ng/g lipid weight). Among the multiple analyses performed, no statistically significant associations were reported for EH‐TBB.

In the studies by Mendy et al. (2023, 2024) that assessed the association between the levels of BFRs in dust and respiratory disease in the USA HOME study (for details see description in Section 3.2.4.2), as regards the prenatal exposure adjusted analyses, both EH‐TBB log₁₀ dust concentration (RR, 1.51; 95% CI, 1.23, 1.85) and EH‐TBB log₁₀ dust loading (RR, 1.26; 95% CI, 1.08, 1.47) were associated with wheeze. In the adjusted analyses, EH‐TBB log₁₀ loading at age 1 was statistically significantly associated with respiratory infections (RR, 1.34; 95% CI, 1.01, 1.79). The same study was also used to evaluate associations between prenatal levels of BFRs and cord blood DNA methylation in an epigenome‐wide association study (Mendy et al., 2025). EH‐TBB dust concentrations were not statistically significantly associated with single CpG sites in the study, but when differentially methylated regions (DMRs) were considered, EH‐TBB concentrations were associated with DMRs overlapping with genes involved in cell division and DNA repair, inflammation and immunity, and gene regulation, as stated by the authors.

In summary, the available epidemiological evidence on EH‐TBB pertained to populations in China or the USA and it was very limited in terms of the total number of available studies (six studies) and the number of available studies per endpoint (1–2 endpoints). Moreover, the exposure assessed in three out of the six studies was done in environmental samples without additional dietary/food data. No statistically significant associations were observed for thyroid function or thyroid cancer, apart from a decrease in FT4 level associated with EH‐TBB levels in one cross‐sectional study. A statistically significant signal was observed for the association between EH‐TBB levels measured in dust and wheeze and respiratory infections.

3.2.5.2.7. Other model systems and mode of action

The previous Opinion reported the study in fathead minnows describe above for BEH‐TEBP, in which fish were exposed via the diet to either of the technical products Firemaster 550 (BEH‐TEBP: 745 μg/g feed, EH‐TBB: 1658 μg/g feed) or Firemaster BZ‐54 (BEH‐TEBP: 907 μg/g feed, EH‐TBB: 2087 μg/g feed) (Bearr et al., 2010).

Since then, several studies have been published and an overview of the studies identified since the previous Opinion is given in Appendix G (Table G.1) and a brief summary is provided below.

Reduced serum T3 and FT4 levels in rats suggest interference with thyroid hormone regulation. The consensus of several in vitro studies examining effects on the thyroid hormone system is that EH‐TBB has low anti‐thyroid hormone and thyroid agonistic activities.

Experiments in fish indicate a possible effect of EH‐TBB on reproduction. Exposure of zebrafish with EH‐TBB (3.00, 13.63 and 54.53 μg/g of food) for 21 days induced a dose‐related reduction of fertility in F0 females (Ma et al., 2018). Feeding medaka a diet containing Firemaster 550 (containing BEH‐TEBP and EH‐TBB) for 21 days resulted in downregulation of the vast majority of genes along the hypothalamus‐pituitary‐gonadal‐liver axis. This was associated with a reduction in cumulative egg production. There is only limited evidence from a number of in vitro studies that EH‐TBB (or BEH‐TEBP) directly interfere with sex steroid hormones. A maximal anti‐oestrogenic effect of 62% was measured at 0.5 mg L⁻¹ of EH‐TBB in the YES assay (Saunders et al., 2013).

EH‐TBB and its carboxylic acid metabolites exhibited competitive inhibition of glucocorticoids for the glucocorticoid receptor in mammalian cell models (Klopčič et al., 2016). No glucocorticoid agonistic activities were observed. EH‐TBB (10 μM) and its metabolite, 2,3,4,5‐tetrabromobenzoic acid (10 μM), activated PXR and showed PXR antagonist activity. EH‐TBB also upregulated CYP3A4 mRNA expression which is PXR dependent.

Other in vitro effects of EH‐TBB included increased hepatocyte lipid accumulation and reduced expression of genes involved in lipid metabolism, such as PPARγ and PGC (Maia et al., 2022). However, EH‐TBB did not affect PPARα and PPARγ activities in transactivation assays in Chinese hamster ovary (CHO) cells or human liver (HepG2) cells (Belcher et al., 2014; Skledar et al., 2016).

There is some in vitro evidence that EH‐TBB has impact on vitamin D signalling pathway. Namely, EH‐TBB 250 nM significantly reduced expression of proteins involved in vitamin D metabolism and pathways (CYP2R1, CYP24A1, RxRa) in hepatic cells but not in MCF‐7 cells, and impaired cell cycle response to vitamin D (Sousa et al., 2022).

EH‐TBB did not show AHR activity in reporter gene assays (Saunders et al., 2013). Whereas dietary exposure of rats to EH‐TBB resulted in upregulation of AHR target genes, with Cypa1a showing up to 162‐fold induction (at 550 mg/kg bw per day), there was no corresponding increase in CYP1A1 activity (Shockley et al., 2020a). Activity of 7‐ethoxyresorufin deethylase (EROD) was increased in double‐breasted cormorant hepatocytes exposed to EH‐TBB, indicative of AHR activity (Pagé‐Larivière et al., 2018). Considering the absence of activity in an AHR reporter gene assay, any direct effect of EH‐TBB on AHR activation requires verification.

EH‐TBB and its metabolite 2,3,4,5 tetrabromobenzoic acid decreased cell viability and increased LDH release from the cells in human umbilical vein endothelial cells exposed for 24h. The metabolite of EH‐TBB increased specific gene expression (VEGFR1, eNOS, PECAM1) and reduced NO release in cells, and increased platelet number of mouse serum whereas the parent compound EH‐TBB did not. All these results suggest a higher level of toxicity for the metabolite compared to the parent compound (Chen et al., 2020).

Exposure of neuroblastoma cells SH‐SY5Y with EH‐TBB (2.5–20 μM) for 24 h showed no effects on cell viability, mitochondrial function, neural plasticity and development, promotion of oxidative stress and inflammation, suggesting the absence of toxicity in this neural cell line at the low μM concentrations tested (Esplugas et al., 2023).

3.2.5.2.8. Consideration of critical effects and dose–response analysis

The CONTAM Panel noted the lack of consistency in the findings of different studies, possibly related to different study designs and endpoints examined. No data on the genotoxic potential of EH‐TBB were identified, and read‐across was considered inappropriate (see above).

The findings reported at the lowest dose levels were neurobehavioural and skeletal effects in the study of a binary mixture of EH‐TBB and BEH‐TEBP (Schkoda et al., 2023; Witchey et al., 2020, 2023), which provided an EH‐TBB dose of 2.25 mg/kg bw per day. However, in the absence of information on relative potency, the CONTAM Panel does not know what the relative contributions of the two substances are, and therefore to be conservative made a worst‐case assumption for each substance. Accordingly, in order to identify a Reference Point specifically for EH‐TBB the CONTAM Panel attributed all findings to EH‐TBB, and adopted the LOAEL value of 2 mg/kg bw per day (rounded) as the Reference Point.

The Panel concluded that, due to the limitations and uncertainties in the current database, the establishment of a health‐based guidance value for EH‐TBB was not appropriate. Instead, the MOE approach should be used for risk characterisation (EFSA Scientific Committee, 2025).

Usually, an MOE of 100, covering variability with respect to kinetic and dynamic differences between animal species and humans (factor 4 × 2.5 = 10) and within the human population (factor 3.2 × 3.2 = 10), is considered sufficient to conclude that exposure does not raise a health concern.

To allow for the absence of a NOAEL and the necessity of using a LOAEL, the EFSA Guidance on default values indicates that an additional uncertainty factor will be needed, the size of which should be determined on a case‐by‐case basis (EFSA Scientific Committee, 2012). ECHA, in its Guidance on information requirements and chemical safety assessment (ECHA, 2012) proposed the application of ‘an assessment factor ranging from 3 (as minimum/majority of cases) to 10 (as maximum/exceptional cases)’ with the following considerations: ‘An AF of 3 may be more appropriate for instance in situations, where the effects at the LOAEL are mild, or the LOAEL represents the lower boundary of the exposure range in which the effect is observed. Higher numerical values should be considered in situations where the effects at the LOAEL are severe and irreversible, or the shape of the dose–response curve is shallow or the quality of the study (e.g. group sizes, statistical methods, study design, exposure data) gives rise to uncertainties about the reliability of the identified LOAEL’ (ECHA, 2012). In this instance, the CONTAM Panel considered that since the effects at the LOAEL for EH‐TBB are mild, the factor of 3 was sufficient to extrapolate from a LOAEL to a NOAEL.

According to the EFSA Scientific Committee Guidance on selected default values, an additional factor can be considered in case of deficiencies in the database on a case‐by‐case basis. A default value has not been proposed, as it will be directly dependent on the data set available (EFSA Scientific Committee, 2012). The WHO/IPCS (1994, 1999) recommended a factor of 3 or 5 if there are minor deficiencies in the database and a factor of 10 if there are major deficiencies in the database. There are limitations in the database for EH‐TBB, such as inconsistencies between studies of non‐standard design, all but one of the studies being on mixtures rather than on the pure substance, no experimental data on genotoxicity and some studies only being available in summary form, so that they could not be fully assessed by the Panel. Taking into account also the sensitivity of the endpoints examined in the critical study, and the approach of attributing the effects solely to EH‐TBB, the CONTAM Panel considered an additional factor of 5 to be appropriate for deficiencies in the database.

As a result, the CONTAM Panel considered that MOEs ≥ 1500 (100 for inter and intra‐species variability, 3 for LOAEL to NOAEL extrapolation and 5 for the gaps in the toxicological database) do not raise a health concern for non‐neoplastic effects.

The CONTAM Panel could not set a Reference Point for neoplastic effects, since no carcinogenicity data are available. In addition, it was not possible to conclude on the genotoxic potential of EH‐TBB.

3.2.5.3. Occurrence data and dietary exposure assessment

Limited occurrence data in food (all left‐censored) were submitted to EFSA (see Section 3.1.1) and no estimation of the dietary exposure was possible.

In the scientific literature, no information was identified for EH‐TBB in the previous EFSA Opinion (EFSA CONTAM Panel, 2012b), but since then, a few studies published in peer‐reviewed journals have been identified reporting on occurrence data on EH‐TBB; an overview is given in Section 3.1.2.

Two studies were identified reporting estimates of dietary exposure to EH‐TBB, and these are summarised below:

• Exposure estimates for EH‐TBB in the UK population were calculated based on daily consumption data for UK citizens from the national diet and nutrition survey (2008–2019) and occurrence data from samples collected in 2020–2021. Average dietary intakes of EH‐TBB ranged from 0.053 to 0.19 ng/kg bw per day, while high‐end estimates, assuming food contaminated at the average concentrations consumed at the mean rate + 2 standard deviations, ranged from 0.31 to 1.2 ng/kg bw per day, across population groups (from 0 up to > 75 years old) (Ma et al., 2023).

• The average daily intake of EH‐TBB was reported for the Chinese population at 1.33 and 0.97 ng/kg bw per day for the 5th (2009−2012) and 6th (2015−2018) TDSs, respectively. Interestingly, contribution of vegetables to the EH‐TBB exposure in the Chinese population was the greatest in both the 5th and 6th TDSs (48.5 and 39.2%, respectively), followed by meat (18.2 and 14.9%, respectively) and cereals (13.7 and 28.6%, respectively).

For breastfed infants, there were insufficient data in human milk from European countries reported in the literature to make robust exposure estimates. As for BEH‐TEBP, to provide an indication of the potential exposure, the CONTAM Panel made an estimation for average and high consumption of human milk (see Section 2.4) considering the available data reported in the literature (Čechová et al., 2017; Tao et al., 2017, see Section 3.2.5.2.2).

The exposure scenario based on average human milk consumption and the highest reported mean concentration in human milk across three European countries (Norway, the Netherlands and Slovakia) of 0.24 ng/g lipid (Čechová et al., 2017), resulted in an estimate of 1.10 ng/kg bw per day. The study by Tao et al. (2017) reported a lower mean concentration in samples from the UK of 0.21 ng/g lipid that would result in an average exposure of 0.96 ng/kg bw per day.

The exposure scenario based on high human milk consumption and the highest reported P95 concentration of 0.683 ng/g lipid (Čechová et al., 2017), resulted in an estimate of 4.7 ng/kg bw per day. The study by Tao et al. (2017) reported only maximum concentrations that were not used for these estimations.

It should be noted that these estimates are based on human milk samples from four European countries that were collected one or two decades ago. It is unknown whether this is representative of current exposure to this BFR.

3.2.5.4. Risk characterisation

Insufficient occurrence data in food on EH‐TBB were submitted to EFSA to allow robust estimates of dietary exposure for the EU population to be made. Limited data obtained from the scientific literature indicated that average dietary exposure from food in the UK was < 0.2 ng/kg bw per day and high‐end estimates up to 1.2 ng/kg bw per day (see Section 3.2.5.3). The Panel noted this exposure in the UK population would result in MOEs in the region of 10^6–7, which does not raise a health concern for non‐neoplastic effects. The indicative exposure estimates obtained for breastfed infants (see Section 3.2.5.3) would result in similar MOEs.

Regarding potential neoplastic effects, the EFSA guidance on the threshold of toxicological concern (TTC) specified that for substances raising a concern about potential DNA‐reactive mutagenicity and/or carcinogenicity, if exposure is below 2.5 ng/kg bw per day, the substance is considered of low probability of adverse health effects (EFSA Scientific Committee, 2019). Although the CONTAM Panel could not conclude on the genotoxic potential of EH‐TBB, the Panel noted that dietary exposure from one study in Europe obtained from the literature is below the TTC for genotoxic compounds, indicating a low probability of adverse health effects.

3.2.5.5. Uncertainty analysis

The uncertainty analysis was performed as described in Section 2.8. The identified sources of uncertainty affecting the hazard assessment are listed in Table 13.

TABLE 13.

Uncertainties identified and their impact on the outcome of the hazard identification/characterisation and risk characterisation of EH‐TBB.

Description of the uncertainty		Impact of the uncertainty on the hazard identification and characterisation a
Chemical composition and analytical methods
Dosing and chemical composition, including impurities	Uncertainty associated with the dose in the critical studies used in the risk assessment	2 – Medium impact. The critical study is a study in which a binary mixture was tested. A conservative (worst‐case) assumption was made that the effects were solely due to EH‐TBB.
Hazard identification and characterisation
ADME	ADME in relation to the critical studies	1 – Low impact. Some studies have been identified. EH‐TBB is well absorbed, highly metabolised and eliminated mainly via urine and faeces.
	Information on transfer rate to animal products	0 – Negligible impact. No studies identified, but no impact on risk assessment
	Transfer via mother's milk	1 – Low impact. Rodent data showed a lactational transfer. EH‐TBB has been detected in human milk
Toxicity studies in experimental animals: endpoints and study design	Limitations in the design of the studies that can result in uncertainties	2 – Medium impact. The number of studies is limited, most have a non‐standard design and/or tested mixtures of BFRs, and findings are inconsistent. However, the effects seen at lowest doses represent sensitive endpoints.
	Relevance for humans of the adverse effect.	0 – Negligible impact: The effects seen at the lowest doses are on behaviour and bone. These are considered relevant for humans even though there are no supporting epidemiological data.
Genotoxicity	Uncertainty in the assessment of genotoxicity	3 – High impact. No experimental data identified. No opportunity for read‐across.
MOA	Uncertainties associated with the data on mode of action	1 – Low impact. There is limited information on the mode of action of EH‐TBB for the critical effects.
Selection of Reference Point	Weaknesses in the critical study	2 – Medium impact. Single dose level study. Study with binary mixture.
	Weaknesses in non‐critical studies, or absence of studies, and uncertainty about whether the endpoints they tested might have been critical if the weaknesses were not present	2 – Medium impact. Inconsistencies between studies of non‐standard design, most studies being on mixtures rather than on the pure substance. No long‐term or carcinogenicity studies. These deficiencies in the database were addressed by including an additional uncertainty factor of 5 in the MOE that would not raise a health concern (see below).
Description of the uncertainty		Impact on the risk characterisation a
Margin of exposure (MOE)	–	1 – Low impact. The usual factor of 100 (to cover for variability with respect to kinetic and dynamic differences between animal species and humans, and within the human population) is considered sufficient. The additional factors of 3 (to account for the absence of a NOAEL) and 5 (for deficiencies in the database, as mentioned above), are considered sufficient. However, it was noted that the assumption that the effects are solely attributed to EH‐TBB is less conservative than the corresponding assumption for BEH‐TEBP, due to differences in solubility/absorption.
Risk characterisation (non‐neoplastics effects)	–	1 – Low impact. Not possible for EFSA to perform an exposure assessment since insufficient occurrence data were submitted to EFSA. Based on dietary exposure estimates reported in the literature for one European country (UK), the resulting MOEs did not raise a health concern. It is uncertain whether this is representative of the situation in other European countries. The uncertainty affecting the exposure estimates from the literature is of large magnitude, but its impact on the risk characterisation is low due to the very large size of the calculated MOE.

Abberviations: ADME, absorption, distribution, metabolism, excretion; BMD, benchmark dose; BMDL, benchmark dose lower credible limit; BMR, benchmark response; MOA, mode of action; TK, toxicokinetic.

^a 0 – Uncertainty with negligible impact; 1 – Uncertainty with low impact; 2 – Uncertainty with medium impact; 3 – Uncertainty with high impact. This ordinal scale provides a relative assessment of the impact of the individual uncertainties as an aid to quantifying their combined impact by expert judgement, as described in section 11 of EFSA Scientific Committee (2018).

Uncertainty analysis: Genotoxicity

No experimental data were identified for EH‐TBB and the CONTAM Panel did not identify any suitable source substances with adequate data to support assessment by read‐across. Consequently, the CONTAM Panel could not conclude on the genotoxic potential of EH‐TBB (see Section 3.2.5.2), and no further analysis of the uncertainties was performed.

Uncertainty analysis: Hazard characterisation

The CONTAM Panel identified substantial uncertainties affecting the hazard assessment (Table 13) that did not allow the establishment of a HBGV. Instead the MOE approach was used, addressing the hazard uncertainties by applying UFs of 100 (to cover for variability with respect to kinetic and dynamic differences between animal species and humans, and within the human population), 3 (to account for the absence of a NOAEL) and 5 (for deficiencies in the database).

Considering the combined impact of the uncertainties and the extent to which they are addressed by the applied UFs, the CONTAM Panel judged with 90–99% certainty (i.e. probability) that, when using the Reference Point for EH‐TBB of 2 mg/kg bw per day, MOEs ≥ 1500 would not raise a health concern for non‐neoplastic effects.

Uncertainty analysis: Risk characterisation

Insufficient occurrence data in food on EH‐TBB were submitted to EFSA to allow robust estimates of dietary exposure for the EU population to be made. Based on dietary exposure data from food from the scientific literature from one European country (UK, see Section 3.2.5.3), the Panel noted this would result in MOEs in the region of 10^6–7.

Considering these very large MOEs, and the uncertainties affecting the exposure estimates, the CONTAM Panel judged with 90%–99% certainty (i.e. probability) that current exposure of the European population to EH‐TBB does not raise a health concern for non‐neoplastic effects. Compared to the situation with BEH‐TEBP, it was noted that the certainty is reduced due to exposure data (one European country only), and less conservativism in the hazard characterisation (regarding the assumption that effects are solely due to EH‐TBB).

3.2.6. Dibromoneopentyl glycol (DBNPG) [CAS number 3296‐900]

3.2.6.1. Production, use and environmental fate

3.2.6.1.1. Production and use

The previous Opinion reported that there were more than 50 suppliers of DBNPG in the world. It is supplied as FR‐522 to make flame retardant unsaturated polyester resins (Weil & Levchik, 2009). NTP (1996), described DBNPG as a reactive flame retardant that is used primarily in unsaturated polyester resins for moulded products and in rigid polyurethane foams. No new information about the production and use of DBNPG was identified.

3.2.6.1.2. Analysis

No information was identified on methods for environmental or food samples for DBNPG.

3.2.6.1.3. Environmental fate and occurrence in the environment

No information could be found about the environmental fate of DBNPG or on levels in wildlife or the environment.

3.2.6.2. Hazard identification and characterisation

3.2.6.2.1. Toxicokinetics

For the current update the CONTAM Panel identified two studies published before the previous Opinion that are reported here for completeness. These two studies referred to 2,2‐Bis(bromomethyl)‐1,3‐propanediol (called BBMP or BMP), which corresponds to DBNPG.

Hoehle et al. (2009) investigated the ADME of DBNPG in rats. Male F344 rats (n = 4 animals per dose) were administered by gavage with a single dose of 10, 100, 150, 300 and 600 mg/kg of ¹⁴C‐labelled DBNPG, and 10 and 15 mg/kg by iv. In parallel, the authors performed repeated dose studies with 100 mg/kg administered daily by oral gavage for 1, 5 or 10 days. DBNPG was rapidly absorbed (C_max reached at 40 min) and efficiently metabolised to a glucuronide metabolite. The authors found that less than 1% of the dose (100 mg/kg bw) was retained in tissues at 72 h: adipose tissues, liver, kidneys, muscle and skin contained 0.2%, 0.7%, 0.1%, 0.3% and 0.3% of the initial dose, respectively. Within 24 h, 89% of the administered radioactivity had been recovered in the urine. The elimination in faeces accounted for less than 10% of the dose at 72 h.

In the repeated daily oral dose study, this route and rate of elimination was similar to the single administration. After oral administration, the majority of the radioactivity detected in plasma extracts was a glucuronide metabolite. DBNPG glucuronide was excreted in the bile, and eliminated predominantly in the urine. Because of enterohepatic recirculation, a low percentage of the dose was eliminated in the faeces.

Rad et al. (2010) investigated the in vitro glucuronidation of DBNPG using both hepatic microsomes from human and male Fisher rat, human intestinal microsomes and also human hepatic recombinant hepatics UDP‐glucuronosyltransferase (UGT) enzymes. The authors reported that DPNPG was converted to a monoglucuronide metabolite in human intestinal microsomes, and to a lesser extent in human hepatic microsomes. The rate of DBNPG glucuronidation in rat hepatocyte was 90‐fold higher than in human (hepatocyte). From also human hepatic recombinant UGT enzymes, the authors reported that only UGT2B7 was involved in the glucuronidation of DBNPG with a similar rate to that of human microsomes.

In a satellite experiment with hepatic microsomes from different species, the authors compared the glucuronidation activity between rats (male and female F344), mice (male B6C3F1), male Syrian hamster, male rhesus monkey and humans. The order of hepatic microsomal glucuronidation activity of DBNPG was rats, mice > hamsters > monkeys > humans.

In summary, DBNPG is rapidly absorbed in rats after oral exposure and metabolised in the liver to a monoglucuronide conjugate. The elimination of DBNPG glucuronide in bile and enterohepatic recycling results in a low systemic bioavailability of DBNPG. The vast majority (89%) of radiolabelled DBNPG was recovered in the urine 24 h following administration by oral gavage. Similar metabolism has been described in humans in vitro.

3.2.6.2.2. Levels in humans

No studies were identified in the previous Opinion, nor in the current update.

3.2.6.2.3. Toxicity in experimental animals

Studies considered in the previous EFSA assessment

The previous Opinion reported data on four oral studies in rodents administered the technical product FR‐1138, containing 78.6% DBNPG, 6.9% 2,2‐bis(bromomethyl)‐1‐bromo‐3‐hydroxypropane (TBNPA, see Section 3.3.12), 6.6% 2,2‐bis(hydroxymethyl)‐1‐bromo‐3‐hydroxypropane, 0.2% pentaerythritol and 7.7% dimers and structural isomers. FR‐1138 was administered for 13 weeks at doses up to 1700 or 1600 mg/kg bw per day in male and female rats, respectively and 3000 or 2900 mg/kg bw per day in male and female mice, respectively. The most sensitive effects were papillary degeneration and necrosis of the kidney and hyperplasia of the transitional‐cell epithelium of the urinary bladder (Elwell et al., 1989; NTP, 1996). The lowest NOAEL reported in these studies was 100 mg/kg bw per day for males of both rats and mice.

Decrease of male reproductive organ weights was only observed with concurrent reduction of body weight in 13 weeks feeding studies with the technical product FR‐1138 (76.8% DBNPG) in mice at 1300 and 3000 mg/kg bw per day, and in rats at 1700 mg/kg bw per day. Oestrous cyclicity and spermatozoal parameters were not affected (NTP, 1996).

The previous Opinion (EFSA CONTAM Panel, 2012b) reported also a two‐generation reproduction toxicity study, CD‐1 mice were fed a diet containing 0, 141, 274 and 589 mg/kg bw per day of DBNPG (purity 87%) (the following impurities were reported: 2,2‐bis(hydroxymethyl)‐1‐bromo‐3‐hydroxypropane (6.7%), 1‐bromo‐3‐hydroxypropane (5.5%) and 3,3‐bis(bromomethyl)‐oxetane (0.5%), 0.3% water) (Treinen et al., 1989). Reduction of postpartum body weights and weight gain in males and females was noted at ≥ 274 mg/kg bw per day as well as a dose‐dependent decrease in live pup weight. The number of litters per pair and the number of live pups born per litter were significantly reduced at 589 mg/kg bw per day. In the F1 generation, weight gain was significantly reduced at the two highest dose levels in males and at the highest dose level in females. The number of live pups born per litter and adjusted live pup weight was decreased at 589 mg/kg bw per day. Cross‐breeding of F0 exposed males and females of the high‐dose group with F0 controls after 11 weeks of dosing resulted in a reduction of the number of live pups per litter and of pup weight only when exposed females were mated with control males. Histological examination revealed kidney lesions in treated male and female rats. The number of follicles was reduced in the ovaries of cross‐bred F0 females in the high‐dose group. The same effect was found in F1 females of the two high‐dose groups. There was no effect on reproductive organ weights and oestrual cyclicity.

Fischer 344/N rats were fed a diet containing the technical product FR‐1138 (containing 78.6% DBNPG), at concentrations of 0, 2500, 5000 or 10,000 mg/kg feed for 2 years. Average daily doses of FR‐1138 were 0, 100, 200 or 430 mg/kg bw for males and 0, 115, 230 or 460 mg/kg bw for females (NTP, 1996). Survival at 2 years of male and female rats in the two highest dose groups was significantly lower than controls. In both males and females, significantly increased incidences of neoplasms in various organs were observed, with males exhibiting a wider range of affected organs than females. Details of the tumour incidence reported by NTP (1996) were not included in the previous Opinion, but are now provided below (see Table 14).

TABLE 14.

Summary of % tumours in rats in the NTP (1996) study with the technical product FR‐1138 (78.6% DBNPG).

FR‐1138 doses a	0 mg/kg diet	2500 mg/kg diet	5000 mg/kg diet	10,000 mg/kg diet
Males	0 mg/kg bw per day	100 mg/kg bw per day	200 mg/kg bw per day	430 mg/kg bw per day	Historical control data
Mammary gland (Fibroadenoma or adenoma)	0/51 (0%)	4/53 (8%)*	7/51 (14%)**	7/55 (13%)**	0–12
Oral cavity (Squamous cell papilloma or carcinoma)	0/51 (0%)	4/53 (8%)*	9/51 (18%)**	10/55 (18%)**	0–4
Oesophagus (Squamous cell papilloma or carcinoma)	0/51 (0%)	0/53 (0%)	1/51 (2%)	6/55 (11%)**	N.A.
Forestomach (Papilloma)	0/51 (0%)	0/53 (0%)	0/51 (0%)	1/55 (2%)	0–2
Small intestine (Adenoma or carcinoma)	0/51 (0%)	0/53 (0%)	0/51 (0%)	2/53 (4%)	0–4
Large intestine (Adenomatous polyp)	0/51 (0%)	0/53 (0%)	3/51 (6%)	4/55 (7%)	0–2
Skin neoplasms (Squamous cell papilloma, keratoacanthoma, squamous cell carcinoma, trichoepithelioma, basal cell adenoma and basal cell carcinoma)	4/51 (8%)	6/53 (11%)	14/51 (27%)**	24/55 (44%)**	2–16
Subcutaneous tissues (Fibroma, fibrosarcoma or sarcoma)	2/51 (4%)	9/53 (17%)**	13/51 (25%)**	16/55 (29%)**	0–16
All organs malignant mesothelioma	0/51 (0%)	3/53 (6%)	8/51 (16%)**	9/55 (16%)**	0–8
Kidney (Tubular adenoma)	0/51 (0%)	0/53 (0%)	1/51 (2%)	3/55 (5%)**	0–6
Urinary bladder (Papilloma or carcinoma)	0/51 (0%)	0/53 (0%)	1/51 (2%)	3/55 (5%)	0–2
Lung (Alveolar/bronchiolar adenoma or carcinoma)	1/51 (2%)	1/53 (2%)	3/51 (6%)	4/55 (7%)*	0–8
Thyroid gland (Adenoma or carcinoma)	0/51 (0%)	2/53 (4%)	6/51 (12%)*	3/55 (5%)	0–6
Pancreas (Adenoma)	1/51 (2%)	2/53 (4%)	4/51 (8%)*	3/53 (6%)	0–10
Haematopoietic system (Mononuclear cell leukaemia)	27/51 (53%)	29/53 (55%)	40/51 (78%)**	34/55 (62%)**	32–62
Zymbal's gland (Adenoma or carcinoma)	2/51 (4%)	1/53 (2%)	4/51 (8%)	5/55 (9%)	0–4
Females	0 mg/kg bw per day	115 mg/kg bw per day	230 mg/kg bw per day	460 mg/kg bw per day	Historical control data
Mammary gland (Fibroadenoma, adenoma or carcinoma)	27/50 (54%)	47/51 (92%)**	47/53 (89%)**	47/52 (90%)**	8–64
Oral cavity (Squamous cell papilloma or carcinoma)	2/50 (4%)	3/51 (6%)	5/53 (9%)	6/52 (12%)	0–6
Oesophagus (Squamous cell papilloma)	0/50 (0%)	0/51 (0%)	1/53 (2%)	10/52 (19%)**	N.A.
Thyroid gland (Adenoma or carcinoma)	0/50 (0%)	0/51 (0%)	2/53 (4%)	4/52 (8%)**	0–6

^a * Means significant different from the control by the logistic regression test p < 0.05, ** means significant different from the control by the logistic regression test p < 0.01.

In addition, non‐neoplastic effects were observed: in males at several doses, in the kidney (focal atrophy, papillary degeneration, transitional epithelial hyperplasia (pelvis) and papillary epithelial hyperplasia), lung (alveolar/bronchiolar hyperplasia), thyroid gland (follicular cell hyperplasia), seminal vesicle (hyperplasia), pancreas (hyperplasia), urinary bladder (hyperplasia) and forestomach (mucosal hyperplasia) and in females, in the kidney (focal atrophy, papillary degeneration and papillary epithelial hyperplasia) at the highest dose only.

B6C3F1 mice were fed a diet containing 0, 312, 625 or 1250 mg/kg of the technical product FR‐1138 (containing 78.6% DBNPG) for 2 years. Average daily doses of FR‐1138 were 0, 35, 70 or 149 mg/kg bw for males and 0, 40, 80 or 170 mg/kg bw for females (NTP, 1996). Survival in both males and females of the highest dose group was significantly lower than that of controls. Details of the tumour incidence reported by NTP (1996) were not included in the previous Opinion, but are now provided below (see Table 15).

TABLE 15.

Summary of % tumours in mice in the NTP (1996) study with the technical product FR‐1138 (78.6% DBNPG).

FR‐1138 doses a	0 mg/kg diet	312 mg/kg diet	625 mg/kg diet	1250 mg/kg diet
Males	0 mg/kg bw per day	35 mg/kg bw per day	70 mg/kg bw per day	149 mg/kg bw per day	Historical control data
Harderian gland (Adenoma or carcinoma)	4/50 (8%)	7/51 (14%)	16/50 (32%)**	22/49 (45%)**	0–20
Lung (Alveolar/bronchiolar adenoma or carcinoma)	15/50 (30%)	11/51 (22%)	16/50 (32%)	25/49 (51%)**	4–32
Forestomach (Papilloma or carcinoma)	0/50 (0%)	3/51 (6%)	3/50 (6%)	4/49 (8%)*	0–6
Kidney (Adenoma)	0/49 (0%)	0/51 (0%)	3/50 (6%)	2/49 (4%)	0–2
Females	0 mg/kg bw per day	40 mg/kg bw per day	80 mg/kg bw per day	170 mg/kg bw per day	Historical control data
Harderian gland (Adenoms or carcinoma)	3/52 (6%)	12/50 (24%)**	13/51 (25%)**	19/50 (38%)**	0–10
Lung (Alveolar/bronchiolar adenoma or carcinoma)	5/52 (10%)	5/50 (10%)	15/51 (29%)**	19/50 (38%)**	2–26
Forestomach (Squamous cell papilloma)	0/52 (0%)	1/50 (2%)	5/51 (10%)*	3/50 (6%)*	0–14
Mammary gland (Carcinoma)	0/52 (0%)	0/50 (0%)	1/50 (2%)	3/49 (6%)	0–10
Circulatory system (Haemangioma or haemangiosarcoma)	1/52 (2%)	2/50 (4%)	0/51 (0%)	5/50 (10%)*	0–8
Subcutaneous tissue (Fibrosarcoma or sarcoma)	0/52 (0%)	1/50 (2%)	4/51 (8%)	12/50 (24%)**	0–8

^a * Means significant different from the control by the logistic regression test p < 0.05, ** means significant different from the control by the logistic regression test p < 0.01.

Regarding genotoxicity, the previous Opinion reported that DBNPG (purity 84% or 96.3%) was mutagenic in S. Typhimurium strain TA100 in the presence of liver S9 fraction from Aroclor‐induced Syrian hamsters. A dose‐related increase in chromosomal aberrations was found in the presence of metabolic activation (rat liver) in Chinese hamster ovary cells (NTP, 1996; purity NR). In vivo, the technical product FR‐1138 (containing 78.6% DBNPG) induced significant increases in the frequencies of micronucleated normochromatic erythrocytes in peripheral blood of male and female mice exposed for 13 weeks via feed. Inconsistent results (one positive and one negative) were noted in bone marrow micronucleus tests in male mice, where DBNPG (purity NR) was administered by gavage three times at 24 h intervals. A bone marrow micronucleus test in mice, where DBNPG (purity NR) was administered ip, was positive only in females (NTP, 1996). Additional data from reproductive reports led to the conclusion that DBNPG reaches the germ cells (see EFSA CONTAM Panel, 2012b).

Studies published since the previous EFSA assessment

Three genotoxicity studies were identified.

The effects of time (1–24 h) and concentration (5–100 μM) on induction of DNA strand breaks by DBNPG (purity of 98%, vehicle 0.5% EtOH) were assessed via the alkaline comet assay in SV‐40 immortalised human urothelial cells (UROtsa). DNA strand breaks (measured by an increase in % DNA in tail) were observed at 1 and 3 h following incubation of cells with non‐cytotoxic concentrations of DBNPG (25 μM). Results indicated also that 1 h exposure (5–100 μM) induced a concentration‐dependent increase in the % DNA in the comet tails. Strand breaks were not present after 6 h of incubation. There was evidence for DBNPG oxidative stress: an elevation of intracellular ROS formation as well as induction of Nrf2 and HSP70 protein levels. Moreover, DNA strand breaks were attenuated when cells were pre‐treated with N‐acetyl‐l‐cysteine (NAC). In addition, oxidative base modifications were revealed when a lesion specific endonuclease, human 8‐hydroxyguanine DNA glycosylase 1 (hOGG1) was introduced into the comet assay (Kong et al., 2011).

An in vitro study investigated the susceptibility of target (UROtsa cells) and non‐target cells (primary rat hepatocytes) to DBNPG‐induced genotoxicity. No increase in DNA strand breaks was detected after 1 h exposure of rat hepatocytes to 25–100 μM DBNPG. In contrast at the same concentrations and exposure time a concentration‐dependent increase in DNA strand breaks was observed in UROtsa cells. To assess the ability of DBNPG to covalently bind to DNA, UROtsa cells and rat hepatocytes were incubated with [14C]‐DBNPG (10–250 μM) for 1 or 24 h (Kong et al., 2013). Following 1 h incubation, the amount of [14C]‐DBNPG equivalents bound to DNA was similar for hepatocytes and UROtsa cells. However, after 24 h incubation with [14C]‐DBNPG, the extent of binding measured in rat hepatocytes was significantly lower than that measured in UROtsa cells (13–111 and 48–1041 pmol/mg DNA, respectively) (Kong et al., 2013). In addition, HPLC analysis indicated that DBNPG was not metabolised and/or eliminated in UROtsa cells at any of the concentrations tested but was extensively converted to a monoglucuronide in hepatocytes. These results demonstrate that a target cell line (UROtsa cells) is more susceptible to DBNPG‐induced DNA damage in comparison to non‐target cells (hepatocytes). This increased susceptibility may relate to the deficiency of metabolic and/or antioxidant capacity in target cells.

The increased susceptibility of target cells to DBNPG‐induced genotoxicity was confirmed in an in vivo mammalian alkaline Comet assay performed on urinary bladder and liver cells (not validated) from Sprague–Dawley male rats exposed by gavage to 0, 300, 600 mg DBNPG/kg bw per day (purity 98%), twice at 21 h apart. Animals were sacrificed 3 h after the second administration. Increased DNA damage was observed in the urinary bladder but not in the liver. In these experimental conditions, no decrease in body weight and no abnormal necropsy or histopathological findings were observed (Wada et al., 2014; REACH dossier, ³¹ as submitted by the registrant(s)).

Summary of toxicity in experimental animals

In sub‐chronic toxicity studies, the technical product FR‐1138 (78.6% DBNPG) induced papillary degeneration (from 400 mg/kg bw per day in male rats and at 1600 mg/kg bw per day in female rats) and necrosis (from 500 mg/kg bw per day in male mice and at 2900 mg/kg bw per day in female mice) of the kidney and hyperplasia of the transitional‐cell epithelium of the urinary bladder in rats and mice (at 1600 mg/kg bw per day in male rats, and 1300 and 1200 mg/kg bw per day in male and female mice, respectively). In a two‐generation reproductive toxicity in mice with DBNPG (87% purity), fetotoxicity was observed in presence of maternal toxicity as well as a decrease in fertility. Crossover mating indicated a specific effect on the female reproductive capacity.

DBNPG (purity 84% or 96.3%, depending on the study) was genotoxic in vitro. The technical product FR‐1138 (78.6% DBNPG) is genotoxic and carcinogenic in a variety of organs in rats (e.g. skin, subcutaneous tissue, mammary gland, Zymbal's gland, oral cavity, oesophagus, forestomach, small and large intestines, mesothelium, urinary bladder, lung and thyroid gland,) and mice (e.g. Harderian gland, lung, forestomach, subcutaneous tissue).

No in vivo studies in mammalian species were available on neurobehaviour effects, which has been identified as critical effect for other BFR families.

3.2.6.2.4. Computational analysis

The QSAR Toolbox identified experimental data for the Ames test on DBNPG as retrieved by the previous Opinion (see above and Annex B).

VEGA models predicted positive Ames test results for DBNPG with ‘GOOD reliability’ (four models). The Panel critically reviewed the predictions and considered them valid mainly due to the presence of relevant analogues known by the model with experimental and predicted values concordant with the predictions for the target compound (see Annex B). The CONTAM Panel noted that this positive prediction is consistent with the available experimental data.

The QSAR Toolbox profilers provided the following information on the structural alerts and predicted mode of toxicity: DBNPG triggers the ‘Aliphatic halides’ alert according to the ‘DNA binding by OECD’ profiler, and the ‘Aliphatic halogens’ alert according to the ‘in vitro mutagenicity (Ames test) by ISS’ profiler because of the aliphatic Br atom (see Figure 1). In the QSAR Toolbox, this alert is associated with DNA alkylation by nucleophilic substitution of second order (SN2) mechanism. The QSAR Toolbox profiling results are provided in Annex B, while additional information on the alerts triggered can be consulted within the software.

The CONTAM Panel noted that the computational analysis supports the available experimental data.

3.2.6.2.5. Observations in humans

No studies were identified in the previous Opinion, nor the current update.

3.2.6.2.6. Studies in other model systems and mode of action

Two new studies on mode of action have been identified.

Induction of DNA strand breaks by DBNPG was shown in alkaline comet assays in immortalised urothelial (UROtsa) cells after 1 and 3 h of exposure to non‐cytotoxic concentrations, but not at later time points up to 24 h (Kong et al., 2011, 2013). There was evidence for DBNPG related oxidative stress: an elevation of intracellular ROS formation, which was statistically significant after 20 min of exposure, induction of oxidised DNA bases (analysed after 1 h exposure only), as well as induction of Nrf2 (significant after 3, 6 and 24 h of exposure) and HSP70 protein levels (significant after 6 and 18 h of exposure). Even if the induced damage seems rapidly repaired (see above), these early genotoxic events may, in part, contribute to the DBNPG‐induced carcinogenesis observed in rodents.

The susceptibility of target (UROtsa cells) and non‐target cells (primary rat hepatocytes) to DBNPG‐induced genotoxicity was investigated (see above). In contrast to hepatocytes, DBNPG exhibited genotoxic potential in UROtsa cells as evidenced by the concentration‐dependent increase in DNA strand breaks and DNA binding. Total content of intracellular GSH quantified in UROtsa cells was 4‐fold lower than that in hepatocytes. HPLC analysis indicated DBNPG was not metabolised in UROtsa cells at any of the concentrations tested (10–250 μM, up to 24 h incubation) in the absence or presence of metabolic co‐factors (UDPGA and NADPH). In hepatocytes, it was extensively converted to a monoglucuronide. This is a less DNA‐reactive metabolite which produced less DNA strand breaks than DBNPG, with no evidence of other metabolites being formed. The extent of glucuronidation by rat hepatocytes was dependent on the substrate ([¹⁴C]‐DBNPG) concentration and duration of exposure. There was extensive metabolism at low substrate concentrations (10–50 μM) and incomplete conjugation of DBNPG and persistence of free parent DBNPG at high concentrations. This is most likely due to the concentration‐dependent saturation of UDP‐glucuronosyltransferases. The degree of covalent binding of [¹⁴C]‐DBNPG equivalents to cellular DNA was similar in both cell types at 1 h but clearly different at 24 h. The authors of the study proposed that increased susceptibility of DBNPG to induce DNA damage in UROtsa cells may be related to the deficiency of antioxidant and/or metabolic capabilities of these cells (Kong et al., 2013).

3.2.6.2.7. Consideration of critical effects and dose–response analysis

The CONTAM Panel concluded that DBNPG (purity 84% or 96.3%, depending on the study) is genotoxic in vitro, and that the technical product FR‐1138 (containing DBNPG with a purity of 78.6%) is genotoxic and carcinogenic in rats and mice.

The Panel noted that one of the impurities of the technical product FR‐1138 is TBNPA, covered in this Opinion in Section 3.3.12. TBNPA was shown to be mutagenic and clastogenic in vitro. In vivo, it was negative in a micronucleus test in mice exposed by the oral route; however, exposure of bone marrow was not demonstrated.

TBNPA has been classified by the ECHA's RAC as a ‘Germ cell mutagen Category 2, H341: Suspected of causing genetic defects’ (ECHA RAC, 2020). No carcinogenicity study on TBNPA has been identified but based on read‐across from DBNPG, ECHA's RAC supports the classification of TBNPA as ‘Carc 1B, H350: May cause cancer’.

No information was identified on the other main impurity of the technical product FR‐1138. In the absence of knowledge of the relative potency of these compounds, and for the purpose of hazard characterisation of DBNPG, the carcinogenic effects observed in animal studies with FR‐1138 were attributed to DBNPG.

Papillary degeneration and necrosis of the kidney and hyperplasia of the transitional‐cell epithelium of the urinary bladder were observed in rats and mice exposed by gavage or by feed to the technical product FR‐1138 (containing 78.6% DBNPG) for 13 weeks. Depression of reproductive organ weight was only observed with concurrent reduction of body weight in male and female mice, and in male and female rats (in mice at 1300 and 3000 mg/kg bw per day, and in rats at 1700 mg/kg bw per day).

In a two‐generation reproductive toxicity, mice were fed with a test substance containing 87% DBNPG, fetotoxicity was observed in presence of maternal toxicity at doses ≥ 274 mg/kg bw per day as well as decrease in fertility was observed at 589 mg/kg bw per day. Crossover mating between F0 mice exposed to 589 mg/kg bw per day and control F0 mice indicated a specific effect on the female reproductive capacity. No in vivo studies in mammalian species were available on neurobehaviour effects, which has been identified as critical effect for other BFR families.

The critical studies are the carcinogenicity studies in rats and mice since other effects were observed only at higher dose levels (see Section 3.2.6.2). No studies in humans were identified.

The CONTAM Panel performed benchmark dose (BMD) modelling according to the 2022 EFSA Guidance on the use of the BMD approach in risk assessment (EFSA Scientific Committee, 2022). The results of the BMD modelling for the critical studies in rats and mice (NTP, 1996) are summarised in Tables 16 and 17, respectively. Details of the BMD analyses, including the individual reports of the modelling are shown in Annex D.

TABLE 16.

Benchmark dose (BMD) modelling for the critical study with the technical product FR‐1138 (containing 78.6% DBNPG) in rats (NTP, 1996) (for details of the BMD analyses see Annex D).

Observed effect a	BMDL10	BMD10	BMDU10
	mg/kg bw per day
Mammary gland
Fibroadenoma or adenoma in M	142	333	1000
Fibroadenoma, adenoma or carcinoma in F	1.16	11	45
Oral cavity
Squamous cell papilloma or carcinoma in M	88	201	451
Squamous cell papilloma or carcinoma in F	397	777	1310
Lung
Alveolar/bronchiolar adenoma or carcinoma in M	441	787	1230
Oesophagus
Squamous cell papilloma in F	322	405	493
Squamous cell papilloma or carcinoma in M b	–	–	–
Forestomach
Squamous cell papilloma in M c	–	–	–
Thyroid gland
Adenoma or carcinoma in M	327	718	1221
Adenoma or carcinoma in F b	–	–	–
Skin neoplasms
Squamous cell papilloma, keratoacanthoma, squamous cell carcinoma, trichoepithelioma, basal cell adenoma and basal cell carcinoma in M	86	151	258
Subcutaneous tissues
Fibroma, fibrosarcoma or sarcoma in M	39	133	338
Large intestine
Adenoma or carcinoma in M	397	674	1198
Small intestine
Adenoma or carcinoma in M c	–	–	–
Haematopoietic system
Mononuclear cell leukaemia in M	148	700	1233
Urinary bladder
Papilloma or carcinoma in M b	–	–	–
Kidney
Tubular adenoma in M b	–	–	–
Pancreas
Adenoma in M	476	877	1247
All organs malignant mesothelioma in M	127	251	617
Zymbal's gland adenoma or carcinoma in M	431	791	1240

^a Based on number of rats bearing disease.

^b There is insufficient evidence that there is a substantial dose‐effect.

^c Tumours occurred only at the highest dose.

TABLE 17.

Benchmark dose (BMD) modelling for the critical study with the technical product FR‐1138 (containing 78.6% DBNPG) in mice (NTP, 1996) (for details of the BMD analyses see Annex D).

Observed effect a	BMDL10	BMD10	BMDU10
	mg/kg bw per day
Harderian gland
Adenoma or carcinoma in M	23	44	79
Adenoma or carcinoma in F	12	43	101
Lung
Alveolar/bronchiolar adenoma or carcinoma in M	52	109	338
Alveolar/bronchiolar adenoma or carcinoma in F	37	71	136
Mammary gland
Carcinoma in F	190	333	491
Forestomach
Squamous cell papilloma in F	138	272	481
Squamous cell papilloma or carcinoma in M	116	238	421
Circulatory system
Haemangioma or haemangiosarcoma in F b	–	–	–
Kidney
Adenoma in M b	–	–	–
Subcutaneous tissue
Fibrosarcoma or sarcoma in F	75	109	156

^a Based on number of mice bearing disease.

^b There is insufficient evidence that there is a substantial dose‐effect.

For quantal data the default benchmark response (BMR) of 10% was applied as recommended in the EFSA BMD guidance (EFSA Scientific Committee, 2022).

The lowest BMDL₁₀ for the technical product FR‐1138 (containing 78.6% DBNPG) was 1.16 mg/kg bw per day, based on increased incidence of mammary gland fibroadenoma, adenoma or carcinoma in female rats. The CONTAM Panel noted that the dose‐spacing in this study was not optimal since the dose levels only differed by two‐fold and the incidence was in the region of 90% at all tested doses. However, DBNPG is clearly a multi‐site carcinogen, and modelling of the data for mammary gland fibroadenoma, adenoma or carcinoma in female rats gave an acceptable fit. Therefore, the CONTAM Panel selected this lowest BMDL₁₀ as the Reference Point.

Although the tumour incidences in the treated groups do not increase with dose, the statistically significant increase in fibroadenoma incidence across all three DBNPG treated groups compared to controls is a concern, as the effect is consistent and biologically plausible. This plausibility is partially supported by mechanistic evidence, including genotoxicity and potential hormonal effects, given that fibroadenomas in rats are often associated with hormonal imbalance. Therefore, a Reference Point could be established based on the observed increase in rat mammary fibroadenoma incidence. Assuming the observed effects were due entirely to DBNPG, and correcting for the purity of the test material, the CONTAM Panel established a Reference Point of 0.9 mg/kg bw per day for DBNPG.

Since the technical product FR‐1138 (containing 78.6% DBNPG) is both carcinogenic and genotoxic, the CONTAM Panel considered that no HBGV can be established as any level of exposure could theoretically lead to cancer (EFSA Scientific Committee, 2005). Instead, the MOE approach would normally be used for risk characterisation. For genotoxic and carcinogenic compounds, and according to EFSA Scientific Committee (2005); EFSA Scientific Committee (2025), an MOE of 10,000 or higher, if it is based on the BMDL₁₀ from an animal carcinogenicity study, would be of low concern from a public health point of view and might be considered as a low priority for risk management actions.

3.2.6.3. Occurrence data and dietary exposure assessment

No occurrence data on DBNPG in food were submitted to EFSA, and no occurrence or exposure data were identified in the literature, either in the previous Opinion or in the current update. Therefore, no estimation of the dietary exposure was possible.

3.2.6.4. Risk characterisation

No risk characterisation was possible due to the lack of information on potential dietary exposure.

3.2.6.5. Uncertainty analysis

Due to the lack of data on occurrence and exposure for DBNPG, uncertainty analysis was performed only for the hazard assessment, using the methods described in Section 2.8. All but one of the identified sources of uncertainty affecting the hazard assessment were assessed as having negligible or low impact on the conclusions (Table 18).

TABLE 18.

Uncertainties identified and their impact on the outcome of the hazard identification/characterisation of DBNPG.

Description of the uncertainty		Impact of the uncertainties on the hazard identification and characterisation a
Chemical composition and analytical methods
Dosing and chemical composition, including impurities	Uncertainty associated with the dose and composition of the test substance in the critical studies used in the risk assessment	2 – Medium impact. Most of the studies are conducted with the technical product FR‐1138 (containing 78.6% DBNPG, 6.6% 2,2‐bis(hydroxymethyl)‐1‐bromo‐3‐hydroxypropane, 6.9% 2,2‐bis(bromomethyl)‐1‐bromo‐3‐hydroxypropane, 0.2% pentaerythritol, and 7.7% dimers and structural isomers) using gavage dosing or via feed (13 weeks repeated toxicity studies, carcinogenicity studies). There is uncertainty about the impact of the impurities on the toxicity.
Hazard identification and characterisation
ADME	ADME in relation to the critical studies	1 – Low impact. Only two studies have been identified. DBNPG is highly absorbed (> 50%), highly metabolised, distributed and accumulated in different tissues and is eliminated mainly via urine.
	Information on transfer rate to animal products	0 – Negligible impact. No studies identified, but no impact on risk assessment.
	Transfer via mother's milk	0 – Negligible impact. No studies identified, but no impact on the hazard characterisation.
Toxicity studies in experimental animals: endpoints and study design	Limitations in the study design of the studies that can result in uncertainties	1 – Low impact. The repeated dose/carcinogenicity and reproductive toxicity studies were performed on the technical product FR‐1138. No studies were available on neurotoxicity/neurobehaviour, which has been identified as critical effect for other BFRs evaluated.
	Relevance for humans of the adverse effect.	0 – Negligible impact: It is both genotoxic and carcinogenic, thus relevant for humans.
Genotoxicity	Uncertainty in the assessment of genotoxicity	1 – Low impact. DBNPG (purity 84% or 96.3%, depending on the study) is genotoxic in vitro. The technical product FR‐1138 is genotoxic in vivo: positive micronucleus test in mice after 13 weeks feeding exposure. Some inconsistent results in in vivo bone marrow micronucleus test (See Section 3.2.6.2.3).
MOA	Uncertainties associated with the data on mode of action	0 – Negligible impact. It is both genotoxic and carcinogenic thus relevant for humans.
Selection of Reference Point	Assumption regarding impurities	1 – Low impact. The Reference Point of 0.9 mg/kg bw per day for DBNPG based on increased incidence of mammary gland fibroadenoma, adenoma or carcinoma in female rats, assumes that the observed effects in studies with the technical product FR‐1138 are caused entirely by DBNPG. One of the impurities of FR‐1138 is TBNPA (see Section 3.3.12). TBNPA was shown to be mutagenic and clastogenic in vitro. In vivo, it was negative in a micronucleus test in mice exposed by the oral route; however, exposure of bone marrow was not demonstrated. TBNPA has been classified by the ECHA's RAC as a germ cell mutagen and carcinogenic 1B based on read‐across from DBNPG (ECHA RAC Opinion, 2020). No information was identified on the other main impurity.
	Selection of BMDL and BMR	0 – Negligible impact. BMR of 10% for quantal data as recommended in the EFSA 2022 Guidance. BMD criteria satisfied. 1 – Low impact. Dose levels in this study only differed by two‐fold and the incidence was in the region of 90% at all tested doses. However, modelling of the data provided an acceptable fit, and selection of the lowest BMDL10 as the Reference Point is considered to be conservative.
	Weaknesses in non‐critical studies, or absence of studies, and uncertainty about whether the endpoints they tested might have been critical if the weaknesses were not present	1 – Low impact. No studies were available on neurotoxicity/neurobehaviour, which has been identified as critical effect for other BFRs evaluated.

Abbreviations: ADME, absorption, distribution, metabolism, excretion; BMD, benchmark dose; BMDL, benchmark dose lower credible limit; BMR, benchmark response; MOA, mode of action; TK, toxicokinetic.

^a 0 – Uncertainty with negligible impact; 1 – Uncertainty with low impact; 2 – Uncertainty with medium impact; 3 – Uncertainty with high impact. This ordinal scale provides a relative assessment of the impact of the individual uncertainties as an aid to quantifying their combined impact by expert judgement, as described in section 11 of EFSA Scientific Committee (2018).

Uncertainty analysis: Genotoxicity and carcinogenicity

Uncertainties affecting the evidence on the genotoxic potential of DBNPG (see Section 3.2.6.2.3) were assessed to have low impact on the assessment (Table 18). DBNPG (purity 84% or 96.3%, depending on the study) is genotoxic in vitro, and the technical product FR‐1138 (containing DBNPG with a purity of 78.6%) is genotoxic and carcinogenic in rats and mice. There is evidence of DNA strand breaks and covalent binding to DNA with DBNPG (purity 98%).

Based on the overall evidence, and taking into account of the associated uncertainties (Table 18), and following an expert group discussion, the CONTAM Panel judged with 95%–99% certainty (i.e. probability) that DBNPG is genotoxic and carcinogenic.

Uncertainty analysis: Reference Point

The Reference Point for carcinogenicity was affected by uncertainties arising from the number and spacing of doses in the critical study (0, 115, 230, 460 mg/kg bw per day), which were assessed as having low impact on the hazard characterisation (Table 18). The critical study was carried out with a technical product in which DBNPG represents 78.6%, and included impurities that could contribute to the toxicity/carcinogenicity. Since the remaining non‐standard uncertainties were all considered to have a negligible or low impact, the Reference Point of 0.9 mg/kg bw per day, which assumed that the effects were entirely attributed to DBNPG, should be conservative.

Considering this, and the general considerations around the prior distributions used in the Bayesian BMD analysis (see Section  3.2.1.5 ), the CONTAM Panel judged with at least 95% certainty (i.e. probability) that the Reference Point of DBNPG of 0.9 mg/kg bw per day would not be lower if the non‐standard uncertainties were resolved.

3.3. Other BFRs included in the 2012 Opinion

According to the Terms of Reference, the CONTAM Panel assessed the 21 BFRs that were also included in the 2012 Opinion, but not covered in Commission recommendation 2014/118/EU, as per the following steps outlined in Sections 1.1 and 1.2:

1. Identification via a literature search since the previous Opinion, of scientific information in the public domain relative to any of the 21 BFRs, and categorisation of the studies according to field of knowledge, e.g. toxicity in experimental animals, epidemiological studies, toxicokinetics, mode of action, levels in human tissues in European countries, occurrence data in food submitted to EFSA and reported in the literature in European countries.

2. Assessment of the evidence for indications/signals of BFRs of potential concern for human health among the 21 BFRs. This was done through a weight of evidence assessment (EFSA Scientific Committee, 2017) as outlined in Section 1.2 on the Interpretation of the Terms of Reference and under Data and Methodologies. Details of this assessment per individual compound are given in Sections 3.3.1 to 3.3.21.

3. For those BFRs for which sufficient and relevant scientific information is available, perform hazard characterisation and exposure assessment, and risk characterisation.

Occurrence data in food was submitted to EFSA for six out of the 21 BFRs considered, i.e. DBDPE, BTBPE, HBB, PBEB, PBT and TBX. An overview of the data submitted is provided in Section 3.1.1. The number of samples was sufficient to perform a dietary exposure assessment only for BTBPE and HBB. For details see Sections 3.3.2 and 3.3.3, respectively.

Literature searches were performed as described in Section 2.2. The studies retrieved were categorised according to the field of knowledge (see above) and the sections below provide an overview of the data identified.

The CONTAM Panel assessed the evidence available for these 21 BFRs, when possible established Reference Points, and highlighted those BFRs of potential concern, e.g. regarding genotoxic potential. An overview is provided in Table 19, which includes also the 6 BFRs covered in Section 3.2.

TABLE 19.

Overview of the data availability for the 27 BFRs considered in the current Opinion. Where no data is identified it is indicated with a dash.

	Occurrence in food in European countries a	Dietary exposure assessment in European countries b	Human biomonitoring data from European countries c	Toxicokinetics	Toxicity studies in mammalian experimental animals d , e	Genotoxicity f	Carcinogenicity d	Observations in humans (epidemiological studies)	Reference point risk characterisation
TDBPP [CAS no 126–72‐7] Section 3.2.1	Literature: < LOD (ES)	–	–	Two oral studies: highly absorbed (> 50%), highly metabolised, distributed and accumulated in different tissues, eliminated mainly via urine	10‐day single dose study in rats, 4‐week dietary study in rats, 20‐day developmental study in rats	Genotoxic QSAR predictions: positive Ames test	Carcinogenic in rats (in kidney) * and mice (in kidney, liver, forestomach, lung)	–	BMDL10 = 0.073 mg/kg bw per day (combined incidence of kidney tubular‐cell adenoma or adenocarcinoma in M rats) MOEs ≥ 10,000 would be of low concern (genotoxic and carcinogenic compound) No risk characterisation possible (lack of data on potential dietary exposure)
EBTEBPI [CAS no 32588–76‐4] Section3.2.2	Literature: one study (failed to display any response in GC‐ or LC–MS)	–	–	One oral study: no accumulation in tissues, eliminated in faeces (65%) and 15% of the dose recovered in urine.	28‐day dietary study in rats, 90‐day dietary study in rats, 20‐day developmental study in rats and rabbits	Not genotoxic QSAR predictions: negative Ames test	–	–	–
HBCYD [CAS no 25495–98‐1] Section 3.2.3	–	–	–	–	–	–	–	–	–
BEH‐TEBP [CAS no 26040–51‐7] Section3.2.4	Data submitted to EFSA: 21 samples (100% left‐censored) Literature: > LOD (NO, DE, BE, UK)	– Literature: two studies (NO, UK)	Human milk: > LOD (NO, NL, SK, UK) < LOD (SE) Blood/serum: < LOD (SE, NO, FR Cord blood: < LOD (BE)	Eight studies: in rodents, poorly absorbed, eliminated mainly via faeces as a parent compound, potential accumulation in liver and adrenal tissues	5‐day study in rats, 28‐day study in rats, 28‐day dietary study in rats, two 28‐day study in mice (2), 90‐day study in rats, single‐dose study in rats (2), developmental study in rats (OECD TG414), 21‐day studies in rats with binary mixture BEH‐TEBP/EH‐TBB*	Inconclusive (negative Ames test, evidence of clastogenicity in vitro not confirmed in vivo by a negative micronucleus tests with no evidence of bone marrow exposure) QSAR predictions: negative Ames test	‐	Six studies. No association for thyroid function/thyroid cancer, type 2 diabetes, BMI. Association between prenatal and infant dust levels and child wheeze, hay fever/allergies and peak expiratory flow (dust exposure). Association with DNA methylation alterations.	LOAEL = 1 mg/kg bw per day (neurobehavioural and skeletal effects in a rat study with binary mixture BEH‐TEBP/EH‐TBB) MOEs ≥ 1500 do not raise a health concern for non‐neoplastic effects Risk characterisation based on dietary exposure estimates from the literature: no health concern for non‐neoplastic effects Dietary exposure estimates in the open literature are above the TTC for substances raising a concern about potential DNA‐reactive mutagenicity and/or carcinogenicity
EH‐TBB [CAS no 183658–27‐7] Section3.2.5	Data submitted to EFSA: 21 samples (100% left‐censored) Literature: > LOD (NO, BE (< LOQ), FR, UK)	– Literature: one study (UK)	Human milk: > LOD (NO, NL, SK, UK) < LOD (SE) Blood/serum: < LOD (SE, NO) Cord blood: > LOD (BE)	Seven studies: in rodents, well absorbed, readily metabolised, eliminated both in urine and faeces	5‐day study in rats, 21‐day studies in rats with binary mixture BEH‐TEBP/EH‐TBB*	–	–	Four studies. No associations for thyroid function/thyroid cancer. Association between prenatal and infant dust levels and child wheeze and respiratory infections. Association with DNA methylation alterations.	LOAEL = 2 mg/kg bw per day (neurobehavioural and skeletal effects in a rat study with binary mixture BEH‐TEBP/EH‐TBB) MOEs ≥ 1500 do not raise a health concern for non‐neoplastic effects Risk characterisation based on dietary exposure estimates from the literature: no health concern for non‐neoplastic effects Dietary exposure estimates in the open literature are below the TTC value for substances raising a concern about potential DNA‐reactive mutagenicity and/or carcinogenicity (the lowest TTC value for any type of effect)
DBNPG [CAS no 3296–90–0] Section3.2.6	–	–	–	Two studies: in rats, rapidly absorbed, metabolised in the liver to a monoglucuronide conjugate. Elimination in bile and enterohepatic recycling (low systemic bioavailability). Similar metabolism in humans in vitro	13‐week studies in rats, two‐generation reproduction toxicity study in mice	Genotoxic QSAR predictions: positive Ames test	Carcinogenic in rats (e.g. skin, sc tissue, mammary gland, Zymbal's gland, oral cavity, oesophagus, forestomach, small and large intestines, mesothelium, urinary bladder, lung and thyroid gland) * and mice (e.g. Harderian gland, lung, sc tissue, forestomach)	–	BMDL10 = 1.14 mg/kg bw per day for FR‐1138 (technical product, purity 78.6%) corresponding to 0.9 mg/kg bw per day for DBNPG (increased incidence of mammary gland fibroadenoma, adenoma or carcinoma in female rats) MOEs ≥ 10,000 would be of low concern (genotoxic and carcinogenic compound) No risk characterisation possible (lack of data on potential dietary exposure)
DBDPE [CAS no 84852–53‐9] Section3.3.1	Data submitted to EFSA: 76 samples (96% left‐censored) Literature: > LOD (ES, UK, FR, PT)	– Literature: One study (UK)	Human milk: > LOD (IRE, UK) < LOD (IRE, SE) Blood/serum: > LOD (NO, FR, SI, EL) < LOD (SE, NO) Cord blood: < LOD (BE)	Three studies: in rodents, poorly absorbed, excreted mainly by faecal route (unchanged form)	5‐day study in rats, 28‐day studies in rats (2), 90‐day studies in rats (3), 30‐day study in mice, 6‐week study in mice, 7‐week study in mice, developmental studies in rats (2), in rabbits (1) and in mice (2), EOGRTS in rats + neurobehavioural cohort, 30‐day and 14‐day studies in pregnant mice development + neurobehavior*, neurodevelopmental study in mice	Not genotoxic	–	Five cross‐sectional studies. Inconclusive results for associations between serum DBDPE levels and TT3, TT4. One study reported association with higher odds of thyroid cancer. No association for lipids, liver and other thyroid biomarkers.	BMDL10 = 1.07 μg/kg bw per day (increased total wheel running activity in mice) MOEs ≥ 300 do not raise a health concern Risk characterisation based on dietary exposure estimates from the literature: no health concern
BTBPE [CAS no 37853–59‐1] Section 3.3.2	Data submitted to EFSA: 630 samples (78% left‐censored) Literature: > LOD (SE, UK, FR, IT, CH, ES)	Dietary exposure estimated by EFSA Literature: One study (UK)	Human milk: > LOD (SE, NO, SK, NL, UK) < LOD (IRE, SE) Blood/serum: > LOD (NO) < LOD (NO, SE) Cord blood: < LOQ (BE)	Two studies: in rats, poor oral absorption, more than 80% radioactivity excreted in faeces	Acute study in rats and dogs, 5‐day study in rats, 28‐day study in rats, 30‐day study in rats (2)	Inconclusive (negative Ames test, no data on clastogenicity or aneugenicity) QSAR predictions: negative Ames test	–	Three cross‐sectional studies. No association between serum BTBPE levels and thyroid hormone levels or thyroid cancer. Increase in total, direct and indirect bilirubin levels in one study.	– Dietary exposure estimates in the open literature are below the TTC value for substances raising a concern about potential DNA‐reactive mutagenicity and/or carcinogenicity (the lowest TTC value for any type of effect)
HBB [CAS no 81–82‐1] Section 3.3.3	Data submitted to EFSA: 2004 samples (82% left‐censored) Literature: > LOD (LV, FR, SE, ES)	Dietary exposure estimated by EFSA –	Human milk: > LOD (NO, SK, NL) < LOD (IRE, SE, UK) Blood/serum: > LOD (NO) < LOD (SE) Cord blood: < LOD (BE)	Three studies: accumulated in adipose tissue of rats. Half‐life of 48 days based on a body burden approach. Rapidly metabolised into debrominated metabolites in rats	Studies in rats with 3 days, 12 weeks and ‘chronic’ oral exposure. Developmental studies with limited design: in pregnant rats exposed GD5 to 15 and in rats exposed from 2 weeks before mating to PND21	Inconclusive (negative Ames test, no data on clastogenicity or aneugenicity) QSAR predictions: negative Ames test	–	Five cross‐sectional studies. No association between HBB serum levels and thyroid hormone profiles, liver injury biomarkers, BMI or thyroid cancer risk. HBB was contributor to association between mixed NBFRs exposure and type 2 diabetes risk	–
DBE‐DBCH [CAS no 3322‐93‐8] Section 3.3.4	Literature: > LOD (LV, UK, SE)	– Literature: One study (SE)	Human milk: > LOD (UK, SE, NO, NL, SK) Blood/serum: > LOD (SE) < LOD (NO)	Three studies: metabolism via CYP P450 into OH‐ and debrominated metabolites. β‐DBE‐DBCH binding affinities described	28‐day study in rats	Genotoxic (mutagenic in vitro only in mammalian cells; no evidence of clastogenicity; no in vivo data)	–	One cross‐sectional study: no association with serum thyroid hormone levels in children	–
PBEB [CAS no 85–22–3] Section 3.3.5	Data submitted to EFSA: 34 samples (100% left‐censored) Literature: > LOD (LV, SE, DE)	– Literature: One study (SE)	Human milk: > LOD (NO, SK) < LOD (SE, NL, UK) Blood/serum: < LOD (SE, NO)	–	28‐day study in rats, perinatal exposure in mice	Inconclusive (negative Ames test, no data on clastogenicity or aneugenicity) QSAR predictions: Negative Ames test	–	Five cross‐sectional studies: association between serum PBEB and increased TT3 levels, not with other thyroid hormones. No association with type 2 diabetes, unadjusted correlation with BMI.	–
PBT [CAS no 87–83‐2] Section 3.3.6	Data submitted to EFSA: 34 samples (100% left‐censored) Literature: > LOD (FR, SE, DE)	– Literature: One study (SE)	Human milk: > LOD (NO, NL, SK) < LOD (SE, UK) Blood/serum: < LOD (SE, NO)	–	91‐day study in rats, developmental toxicity study in rats	Inconclusive (negative Ames test, no data on clastogenicity or aneugenicity) QSAR predictions: Negative Ames test	–	Four cross‐sectional studies: one reported association between serum PBT and decreased AST/ALT liver biomarkers. One reported association with decreased TT3 levels. No association for other thyroid hormones, thyroid cancer, type 2 diabetes or BMI	–
TDBP‐TAZTO [CAS no 52434–90‐9] Section 3.3.7	–	–	–	Two studies: in mice, accumulated in kidney, liver, heart and brain in mice. CYP2E1, CYP3A4 and CYP2C9 involved in the in vitro metabolism (human liver microsome). Four metabolites detected	14‐day study in mice, 28‐day study in rats and mice, 4‐week study in rats, 30‐day study in mice (neuro)*, 6‐month study in rats (neuro), combined repeated dose/reproductive/developmental study in rats (OECD TG422)	–	–	–	LOAEL = 5 mg/kg bw per day (depressive‐related behaviour in the forced swim test and the tail suspension test in rodents) MOEs ≥ 900 do not raise a health concern No risk characterisation possible (lack of data on potential dietary exposure)
TBCO [CAS no 3194‐57‐8] Section 3.3.8	–	–	Human milk: < LOD (NO, NL. SK) Blood/serum: < LOD (NO)	–	–	–	–	–	–
DBHCTD [CAS no 51936–55‐1] Section 3.3.9	Literature: < LOD (UK)	–	Human milk: < LOD (NL, NO, SK) Blood/serum: < LOD (NO, SE)	–	5‐day study in rats	Inconclusive (negative Ames test by QSAR prediction, no clastogenicity and aneugenicity) QSAR predictions: negative Ames test	–	One cross‐sectional study: serum DBHCTD levels associated with decreased levels of serum T3, T4, FT4 and TSH in children	–
HCTBPH [CAS no 34571–16‐9] Section 3.3.10	Literature: < LOD (UK)	–	–	–	–	–	–	–	–
PBB‐Acr [CAS no 59447–55‐1] Section 3.3.11	Literature: < LOD (UK)	–	–	–	28‐day study in rats	Not genotoxic QSAR predictions: negative Ames test	–	–	–
TBNPA [CAS no 1522‐92‐5] Section 3.3.12	–	–	–	–	14‐day study in rats, 28‐day study in rats, 30‐day dietary study in rats, 90‐days study in rats, developmental study in rats	Genotoxic (mutagenic in bacteria and mammalian cells in vitro; positive clastogenicity in vitro; inconclusive data in vivo with no evidence of bone marrow exposure) QSAR predictions: positive Ames test	Predicted carcinogenic (positive in vitro mutagenicity and clastogenicity supported by structural similarity with the genotoxic and carcinogenic DBNPG)	–	–
TBX [CAS no 23488–38‐2] Section 3.3.13	Data submitted to EFSA: 13 samples (100% left‐censored) Literature: > LOD (FR, SE)	‐ Literature: One study (SE)	Human milk: > LOD (NO, NL) < LOD (SE, UK, SK) Blood/serum: < LOD (SE, NO)	–	–	Inconclusive (negative Ames test by QSAR prediction, no clastogenicity and aneugenicity) QSAR predictions: negative Ames test	–	Three cross‐sectional study: TBX concentrations in wristbands not statistically significant associated with serum thyroid hormone profiles. Serum TBX levels not associated with type 2 diabetes or BMI.	–
BDBP‐TAZTO [CAS no 75795–16‐3] Section 3.3.14	–	–	–	–	–	–	–	–	–
DBP‐TAZTO [CAS no 57829–89‐7] Section 3.3.15	–	–	–	–	–	–	–	–	–
DBS [CAS no 31780–26‐4, 120,359–56‐0] Section 3.3.16	–	–	–	–	–	–	–	–	–
HEEHP‐TEBP [CAS no 20566–35‐2] Section 3.3.17	–	–	–	–	Range‐finding study in rats	–	–	–	–
4’‐PeBPO‐BDE208 [CAS no 58965–66‐5] Section 3.3.18	–	–	–	–	–	–	–	–	–
TTBNPP [CAS no 19186–97‐1] Section 3.3.19	–	–	–	One toxicokinetic study with the technical product in rats: rapid excretion following oral administration. Within 24 h, 81–85% of the dose was recovered in faeces and 2% in urine, indicating a low oral bioavailability of approximately 2%.	28‐day study in rats, 90‐day study in rats, reproductive and developmental studies in rats	Inconclusive (insufficient reporting accessible to conclude)	–	–	–
TTBP‐TAZ [CAS no 25713–60–4] Section 3.3.20	–	–	–	One toxicokinetic study with a technical product in rats: very low oral bioavailability (estimated at less than 0.2%).	Acute oral study in rats, 28‐day study in rats, 91‐day study in rats, developmental study in rats (OECD TG414)	Not genotoxic	–	–	–
OBTMPI [CAS no 1084889–51‐9/1,025,956–65‐3/893,843–07‐7/155,613–93‐7] Section 3.3.21	Literature: < LOD (UK)	–	Blood/serum: < LOD (SE) > LOD (NO)	–	–	–	–	–	–

^a Occurrence data in food submitted to EFSA or availability of studies reporting on occurrence data in food from European countries identified in the open literature.

^b Dietary exposure assessment performed by EFSA or availability of studies reporting on dietary exposure from European countries identified in the open literature.

^c Studies reporting on occurrence in human samples identified in the open literature.

^d The critical study selected for the identification of the Reference Point is indicated with an asterisk (*).

^e Refers to the design of the toxicity studies in experimental animals identified.

^f The conclusion on genotoxicity is based on the available experimental data. The outcome of the computational analysis is reported only when the VEGA QSAR predictions on Ames test were considered valid.

3.3.1. Decabromodiphenyl ethane (DBDPE) [CAS number 84852‐53‐9]

DBDPE is a non‐polymeric additive flame retardant, considered by many stakeholders as an alternative to the PBDE technical mixture DecaBDE. It is widely used in the automotive industry, for electrical and electronic equipment. For textiles it can be used at concentrations making up to 40% of the material produced.

The physicochemical properties of DBDPE are reported in Table 1. DBDPE is registered in REACH (10,000–100,000 tonnes per annum) and it has been found to meet the vPvB criteria (see Table 2).

3.3.1.1. Hazard identification and characterisation

3.3.1.1.1. Toxicokinetics

In the previous Opinion (EFSA CONTAM Panel, 2012a, 2012b), it was reported that after a single oral dose of 1000 mg/kg bw in Sprague–Dawley rats, DBDPE was poorly absorbed by the gastrointestinal tract (Environment Agency, England and Wales, 2007).

Since the previous Opinion, two studies have been identified (Knudsen, Sanders, Hughes, et al., 2017; Zheng et al., 2022):

• Knudsen, Sanders, Hughes, et al. (2017) investigated the disposition of DBDPE in female Sprague–Dawley rats and male B6C3F1/Tac mice. A single dose of 0.02 mg/kg bw of ¹⁴C‐DBDPE was administered via oral and iv route in rats and mice. A second group of rats was administered for 10 days with 0.02 mg/kg bw per day of ¹⁴C‐DBDPE via oral route.

• After a single oral administration to Sprague–Dawley female rats, 75% of the dose was retrieved in faeces after 24h. In a pilot study with bile‐duct cannulated rat, the authors found that 0.1% of the dose was recovered in bile which suggests that radioactivity recovered in faeces represents unabsorbed dose. Similar results were obtained in male B6C3F1 mice. The authors conducted an i.v. study in rats, but did not calculate oral bioavailability. Less than 1% of the radioactivity was retained in tissues collected at 72 h in rat.

• The results showed that DBDPE was poorly absorbed, and mainly excreted (98%) by the faecal route (unchanged form) within 72 h after a single oral administration to rats and mice.

• After 10 daily oral administrations to Sprague–Dawley female rats, the authors measured 2% ± 0.4% and 94% ± 8% of the radioactivity in urine and faeces, respectively. The concentrations in adrenal and liver tissue increased after 10 days exposure. However, the radioactivity retained in these tissues collected 24 h after the final dose was low (0.06% and 0.008% of the total dose in liver and adrenal tissues, respectively).

• Zheng et al. (2022) studied the biotransformation of several BFRs including DBDPE, to evaluate whether this was responsible for the presence of 2,4,6‐TBP found in humans using human and rat liver microsomes. No information was reported by the authors on the biotransformation of DBDPE except that it did not result in the formation of 2,4,6‐TBP.

Furthermore, data reported in the REACH dossier, ³² as submitted by the registrant(s), support a very low oral bioavailability of DBDPE in rodents following oral administration of radiolabelled DBDPE. It should be noted that the CONTAM Panel could not independently assess these data as it did not have access to the original study reports.

3.3.1.1.2. Levels in humans

The previous Opinion reported a study on five serum samples from Sweden in which DBDPE could not be identified above the LOD (1.03 ng/g lipid) (Karlsson et al., 2007).

Since then, several studies have been identified reporting on the levels of DBDPE in human samples from European countries, including human milk and blood/serum. These studies are briefly described below, and details can be found in Appendix E (Table E.1).

Regarding human milk samples,

• DBDPE was reported to be below the LOD (2.2 ng/g lipid) in 11 pooled human milk samples collected in 2010 from 109 first‐time mothers in Ireland (Pratt et al., 2013). In a later study on 16 pools of human milk collected between 2016 and 2018 still in Ireland, DBDPE was detected in 3 out of the 16 pools analysed, with concentrations ranging from < LOD to 4.6 ng/g lipid (LOD 2.5 ng/g lipid) and a mean value of 0.74 ng/g lipid (Wemken et al., 2020).

• In Sweden, Sahlström et al. (2015) analysed two pooled human milk samples collected in 2009–2010. DBDPE was not detected in any of the two pooled samples (< 11 pg/g ww).

• Tao et al. (2017) analysed DBDPE in milk samples from UK mothers collected in 2010 (n = 25) and 2014–2015 (n = 10). In 2010, the levels of DBDPE ranged from < 0.78 to 250 ng/g lipid (detection frequency 4%), while in 2014–2015 a slightly higher detection frequency (10%) was reported and concentrations ranged from < 0.78 to 58 ng/g lipid.

Regarding blood/serum samples,

• Cequier et al. (2013) developed an analytical method for the determination of PBDEs and other BFRs in serum samples. The method was applied to the analysis of 10 serum samples from Norway and DBDPE was not detected in any of the samples (LOD = 20 pg/mL). In a later study, Cequier et al. (2015) also reported levels of DBDPE below the LOD in serum samples from 46 women from a Norwegian mother–child cohort established in 2012.

• DBDPE was also measured in 61 serum samples from a Norwegian cohort collected in 2013 with a reported geometric mean of 3.4 ng/g lipid (detection frequency 2%, LOD 42 pg/serum sample (Tay et al., 2019).

• In Sweden, Sahlström et al. (2014) did not detect DBDPE above the LOD (15 pg/25 μL) in any of the 24 matched serum samples from mothers and their toddlers collected 2009–2010.

• Van der Schyff et al. (2023) reported on flame retardants in children from nine European countries from the HBM4EU aligned studies. DBDPE was investigated in blood from France, Greece and Slovenia, with a detection frequency of 13–18% (LOD = 0.013 μg/L). No further details were provided as compounds with detection frequencies < 40% were excluded by the authors for statistical analysis.

• Sales et al. (2017) did not detect DBDPE above the LOD of 13 pg/g serum in one pool of human cord blood from Belgium.

3.3.1.1.3. Toxicity in experimental animals

Studies considered in the previous EFSA assessment

In the previous Opinion several studies were reported. No clinical signs, changes in clinical chemistry or haematology were observed in male and female Sprague–Dawley rats exposed by gavage to DBDPE (purity 96.3%) for 90 days up to 1000 mg/kg bw per day (Hardy et al., 2002). Statistically significant differences were only found between control and female rats exposed to 1000 mg/kg bw per day in absolute and relative liver weights.

No significant changes in body, liver and kidney weights (absolute or relative) were observed in male rats orally exposed for 90 days to 100 mg DBDPE/kg bw per day (Wang et al., 2010). A significant increase in the T3 level was found in the treated group suggesting that DBDPE can alter thyroid hormone homeostasis. However, there was no effect on T4 level, and TSH was not measured.

No evidence of maternal toxicity, developmental toxicity or teratogenicity was noted in rats and rabbits exposed by gavage up to 1250 mg DBDPE/kg bw per day from GD6 to GD15 for rats and GD6 to GD18 for rabbits (as cited in Hardy et al., 2010).

Regarding genotoxicity, the previous Opinion reported that DBDPE was not genotoxic in reverse mutation assays in S. Typhimurium and E. coli WP2 and did not induce chromosomal aberrations in V79 Chinese hamster lung cells (San and Wagner, 1991; Stankowski, 1988, as cited in Hardy et al., 2010) (see EFSA CONTAM Panel, 2012b).

Studies published since the previous EFSA assessment

A number of studies were identified since the previous Opinion. Some studies were identified in the scientific literature while others were found in the REACH dossier³², as submitted by the registrant(s). For the latter, the CONTAM Panel did not have access to the original study reports.

An overview of these studies can be found in Appendix C (Table C.1).

DBDPE was also included in the same comparative study of a number of BFRs conducted by the NTP and described in Section 3.2.4.2.3 (Shockley et al., 2020a). Male Harlan Sprague–Dawley rats exposed to DBDPE (purity 94%) by gavage for 5 days to 0 (corn oil), 0.1, 0.97, 9.71, 97.1 and 970 mg/kg bw per day showed no effects on serum TT4, TT3 or TSH nor on absolute liver weight. The clinical chemistry endpoints and the haematology endpoints measured were unchanged compared to concurrent controls. There was no effect on UDP GT1a1 liver enzyme level. No histological changes were present in the liver or thyroid gland (Shockley et al., 2020a). The NOAEL was 970 mg/kg bw per day.

A number of publications reported on a 28‐day study in male Sprague–Dawley rats exposed by gavage to DBDPE (purity > 98%) at 0 (corn oil), 5, 50, 500 mg/kg bw per day conducted by the same research group (Gao et al., 2022; Jing et al., 2019; Li et al., 2021; Sun et al., 2020; Wang, Chen, et al., 2019; Xue et al., 2022; Zheng et al., 2021). The results are described below:

• −At the highest dose group, DBDPE was reported to cause morphological changes in the heart and abdominal aorta (Jing et al., 2019). It also induced ultrastructural damage in the heart: slight mitochondrial swelling and cristae disorder at 50 mg/kg bw per day, and additionally, mitochondrial swelling, membrane breach and more cristae disorder, or vacuolation at 500 mg/kg bw per day. Ultrastructural damage (endothelial cell nucleus swelling, subendothelial space expansion, endothelial cell shedding and internal elastic lamina thinning) in the aorta was also observed at 50 and 500 mg/kg bw per day, and ultrastructural changes of smooth muscle cells at 500 mg/kg bw per day. There was no alteration in serum creatine kinase (CK) activity, whereas lactate dehydrogenase (LDH) activity showed a significant increase at 500 mg/kg bw per day. Additionally, DBDPE led to endothelial dysfunction, as evidenced by the Endothelin‐1 (ET‐1) and intercellular adhesion molecule‐1 (ICAM‐1) elevation at the highest dose group. DBDPE‐induced inflammation was characterised by the upregulation of key inflammatory mediators detected in the serum, including interleukin‐1β (IL‐1β), IL‐6 and tumour necrosis factor alpha (TNFα) at 50 and 500 mg/kg bw per day. In the heart, superoxide dismutase (SOD) and glutathione peroxidase (GSH‐Px) activities were significantly increased at all doses and malondialdehyde (MDA) content was significantly increased at the two highest doses. The authors concluded that DBDPE could induce oxidative stress, inflammation and eventually lead to endothelial dysfunction and cardiovascular injury (Jing et al., 2019). The NOAEL was 5 mg/kg bw per day. Zheng et al. (2021) reported also morphological damage in the abdominal aortas and ultrastructural changes at the two DBDPE highest doses.
• −The two highest DBDPE doses induced cardiomyocyte injury and fibrosis. Cellular apoptosis increased also in these groups. Ultrastructural observation of cardiomyocytes revealed that DBDPE could induce mitochondria damage and dissolution (Gao et al., 2022).
• −Decreased serum FT3 at ≥ 50 mg/kg bw per day and increased TSH and thyrotropin‐releasing hormone (TRH) were observed at 500 mg/kg bw per day; however, there was no effect on TT4, TT3 and FT4 (Wang, Chen, et al., 2019). DBDPE exposure caused significant changes in histological structure of the thyroid gland: smaller size of follicular cavities and disorder of the array of thyroid follicular epithelial cells. Other dose‐related changes were also observed: increased height of the thyroid follicular epithelial cells, swollen and vacuolated follicular cells, oedema in follicular epithelial cells, infiltration of mast cells in the follicular stroma, exfoliated epithelial cells and decreased average colloid area. In addition, there were changes in ultrastructure of the thyroid gland (e.g. atrophy of the rough endoplasmic reticulum and vacuolation of the mitochondria). No morphological changes were observed in the pituitary gland and hypothalamus. Oxidative damages of the thyroid gland (dose‐related decreased SOD and increased MDA content) were also observed. The authors concluded that DBDPE exposure could disrupt thyroid function and the underlying mechanism was oxidative stress and perturbations of the hypothalamic–pituitary–thyroid (HPT) axis (Wang, Chen, et al., 2019). The LOAEL was 5 mg/kg bw per day based on dilatation of the rough endoplasmic reticulum, swelling of mitochondria and disappearance of mitochondrial crests.
• − DBDPE had no effect on liver (absolute or relative) weight but caused liver morphological changes (feathery necrosis at ≥ 50 mg/kg bw per day and irregular arrangement of hepatic cords, and inflammatory cell infiltration at the two highest doses) (Sun et al., 2020). It induced oxidative stress (increases in MDA content and decreases in SOD activity at high dose), increased γ‐glutamyl transferase (50 and 500 mg/kg bw per day) and glucose levels (500 mg/kg bw per day) in serum (Sun et al., 2020). The LOAEL was 50 mg/kg bw per day.
• −Exposure to DBDPE resulted in histopathological changes in the testis (significant seminiferous epithelium deletion) and decreased sperm number at 50 and 500 mg/kg bw per day, decreased motility at 5, 50 and 500 mg/kg bw per day, and increased sperm malformation rates at 500 mg/kg bw per day (Li et al., 2021). Oxidative stress was observed in the testis with significant increases in MDA content and significant decreases in T‐SOD activity at 500 mg/kg bw per day. DBDPE exposure led to significant increase of TUNEL‐positive cells in the 50 and 500 mg/kg bw per day groups indicating testicular apoptosis (Li et al., 2021).
• − DBDPE caused histopathological changes in the testis (morphological changes and expanded and vacuolised mitochondria and fractured or vanished mitochondrial cristae at 50 and 500 mg/kg bw per day) (Xue et al., 2022). It reduced also sperm quantity (at 50 and 500 mg/kg bw per day) and motility (at all doses), and raised the malformation rate (at 50 and 500 mg/kg bw per day). DBDPE had an effect on the clinical chemistry parameters in rats: increased glucose, LDL‐C and TC levels. Glucose levels were significantly negatively correlated with sperm motility and sperm concentration, and were significantly positively correlated with sperm malformation rate. Furthermore, LDL‐C levels showed a significant negative correlation with sperm motility and sperm concentration and a significant positive correlation with sperm malformation rate. Immunofluorescence analysis showed that the DNA damage marker γ‐H2AX increased significantly in rat testicular cells after exposure for 28 days to 500 mg DBDPE/kg bw per day (Xue et al., 2022). The NOAEL was 5 mg/kg bw per day.

Male and female Sprague–Dawley rats were exposed to DBDPE (purity 96.3% DBDPE, 3.6% dodecabromodiphenyl ethane) by gavage for 28 days (OECD TG 407) to 0 (corn oil), 125, 400 and 1250 mg/kg bw per day. Exposure was followed by a 14‐day withdrawal in subsets of the control and high‐dose groups only. The NOAEL was 1250 mg/kg bw per day, the highest dose tested (REACH dossier, as submitted by the registrant(s)).

Male and female Sprague–Dawley rats were exposed to DBDPE (purity 96.3% DBDPE, 3.6% Dodecabromodiphenyl ethane) by gavage for 90 days (OECD TG 408) to 0 (corn oil), 100, 320 and 1000 mg/kg bw per day. At the end of the 90‐day treatment period all animals in the low and mid dose groups and 10 randomly selected rats/sex in the control and high‐dose groups were sacrificed. The remaining rats in the control and high‐dose groups were continued untreated for a 28‐day recovery period prior to sacrifice. Mean liver weights were increased at 1000 mg/kg bw per day. Low‐grade microscopic liver changes were observed in male rats. The liver changes were characterised by minimal to slight hepatocellular vacuolation (high‐dose males) and minimal to slight centrilobular hepatocytomegaly (high, and possibly mid dose, males), and were considered an adaptive response. No changes were found in the livers of female rats. The liver changes resolved after a 28‐day recovery period. No effects were observed on the reproductive organs. The NOAEL was 1000 mg/kg bw per day (REACH dossier, as submitted by the registrant(s)).

Male C57BL/6J mice were exposed by gavage to 0, 5 or 50 mg DBDPE (purity NR)/kg bw per day for 7 weeks (Qi et al., 2025). The bodyweights of the mice in the 50 mg/kg bw per day group were slightly lower. At both doses, there were statically significant increases in relative testicular weight. DBDPE decreased the sex hormone levels (serum and intratesticular testosterone et both doses, LH and FSH at 50 mg/kg bw per day) and sperm count. The steroidogenic acute regulatory protein (StAR) decreased in a dose‐dependent manner. DBDPE impaired testicular morphology (testicular damage characterised most notably by a reduction in cell layers, disorganisation of the germinal epithelium and pronounced loss of germ cells, especially at 50 mg/kg bw per day). DBDPE dose‐dependently increased the sperm malformation rate. Oxidative stress (increase of MDA content and decrease of SOD levels in testicular homogenates) and apoptosis maybe potential target pathways of DBDPE (Qi et al., 2025).

Male ApoE−/− mice were exposed by gavage to 0 (corn oil) or 100 mg DBDPE (98% purity)/kg bw per day for 6 weeks. DBDPE promoted atherosclerosis in ApoE−/− mice. The number of lesions in the aortic roots increased; the plaques formed in the lumen increased, and the plaque area increased significantly. Exposure significantly increased the lipid content in the aortic root. No significant change was observed in the area of aortic fibrosis in the DBDPE group. ROS levels in plaques in the aortic roots of ApoE−/− mice exposed to DBDPE significantly increased, suggesting that DBDPE caused oxidative stress in the aorta (Li et al., 2025).

Sprague–Dawley rats were exposed to DBDPE (purity NR) by gavage at 0, 5, 50 or 500 mg/kg bw per day for 8 weeks (Cui et al., 2025). It was shown that DBDPE exposure led to a dose‐related statistically significant increase in relative liver weight from the lowest dose. Five mg/kg bw per day caused mild tissue injury (mild loosening of hepatocyte arrangement was observed in certain regions) and metabolic disturbances. At 50 mg/kg bw per day, a substantial accumulation of lipid vacuoles was detected in the liver, indicative of steatosis. Exposure at 500 mg/kg bw per day resulted in prominent lipid vacuole formation within hepatocytes, also consistent with steatosis, accompanied by marked alterations in cellular morphology. In addition, inflammatory infiltration and ballooning degeneration were observed in some areas of the liver, suggesting significant metabolic injury. DBDPE exposure resulted in significant increases in total cholesterol (TC), triglycerides (TG) and LDL levels but significant decreases in HDL levels only at low dose levels. The LOAEL was 5 mg/kg bw per day.

Male and female Balb/C mice were exposed to DBDPE (purity 98.5%), by gavage for 30 days to 0 (corn oil), 5, 20, 100 and 200 mg/kg bw per day (Sun et al., 2018). There were no signs of toxicity and no significant treatment effect on body weight, or liver‐to‐body weight ratios. The levels of ALT and aspartate aminotransferase (AST) increased markedly in a dose‐dependent manner. Blood glucose levels were increased according to the dose and time of exposure. There was also an increase in TSH, T3 and fT3 at 200 mg/kg bw per day. UDPGT, PROD and EROD activities were increased significantly at ≥ 100 mg/kg bw per day. Histopathologic liver changes were characterised by hepatocyte hypertrophy and cytoplasmic vacuolisation (Sun et al., 2018). The LOAEL was 5 mg/kg bw per day based on the increased ALT serum level.

Reproductive/developmental toxicity

During the perinatal period (GD6‐PND21), pregnant ICR mice were exposed to DBDPE (purity 96%) by gavage to 0 and 100 μg/kg bw (Yan et al., 2018; REACH dossier, as submitted by the registrant(s)). After weaning, male offspring were fed on either a low‐fat diet or a high‐fat diet for 12 weeks. DBDPE increased the risk of obesity in mouse offspring and affected triglyceride synthesis, bile secretion, purine synthesis, mitochondrial function and glucose metabolism. Furthermore, the use of high‐fat diet feeding may further exacerbate these effects. The authors claimed that their results show that early‐life exposure to low doses of DBDPE may promote the development of metabolic dysfunction, which in turn can induce obesity (Yan et al., 2018; REACH dossier).

Pregnant ICR mice were exposed to DBDPE (purity NR) by gavage at doses of 0, 0.05, 0.5 and 5 mg/kg bw per day from GD1 to GD17 (Xie et al., 2025). Fetal size was measured on GD18. Exposure caused placental damage, resulting in fetal growth restriction (FGR). Fetal weight was decreased in the 0.5 and 5 mg/kg bw groups. Crown‐rump length decreased and the rate of FGR was markedly increased at the highest dose but not significantly at lower doses. A dose‐dependent reduction in placental weight following DBDPE exposure was observed (at the two highest doses). No significant difference was found in placental diameter, however, there was a notable decrease in the fetal‐to‐placental weight ratio (at 0.5 and 5 mg/kg bw per day). A significant reduction in the total placental area, as well as a decrease in both the placental spongiotrophoblast and labyrinth areas were observed at all doses. In humans, a reduced fetus‐to‐placental weight is a marker for a variety of adverse outcomes in the infants as well as of prognostic value predictive of later life health outcomes (Beaumont et al., 2023; Kinjo et al., 2024; Rivera‐Núñez et al., 2025). Additionally, the spongiotrophoblast‐to‐labyrinth area ratio was increased in exposed groups (statistically significant at 0.5 mg/kg bw per day). There was also a decrease in placental microvascular density in exposed mice (Xie et al., 2025). The LOAEL was 0.05 mg/kg bw per day.

Female Kunming mice were orally exposed to DBDPE (purity NR) for 30 days at 50 μg/kg bw per day (Shi, Xu, et al., 2022). DBDPE‐exposed female mice were mated with untreated male mice, and pre‐ and post‐implantation embryos were isolated. A loss of embryos (statistically significant decrease in the rate of 2‐cell and 8‐cell embryos and blastocysts formation) and a statistically significant lower litter size were noted in DBDPE‐exposed mice. Moreover, from GD14.5 onwards, the embryo length of DBDPE‐exposed mice was significantly reduced, indicating delayed fetal development.

In another study in the same mouse strain by the same authors, female Kunming mice were exposed to DBDPE (purity NR) by gavage for 14 days to 0, 5, 50 and 500 μg/kg bw per day. After treatment, female mice were intraperitoneally injected with PMSG and 48 h later with HCG and then mated with sexually mature males (Shi, Qiu, et al., 2022). DBDPE inhibited preimplantation embryo development in a dose‐dependent manner by affecting the circadian rhythm (from 50 μg/kg bw per day) and endogenous melatonin levels (decrease serum melatonin in serum from 5 μg/kg bw per day in 2‐cell embryo and 8‐cell embryo). The highest dose significantly reduced the rate of 2‐cell embryos. The ratio of the 4‐cell embryo, 8‐cell embryo and morula dramatically decreased at 50 and 500 μg/kg bw per day and even the lowest dose resulted in a significant decrease in the rate of blastocyst formation. However, DBDPE exposure did not affect the size of the blastocyst. It was shown that exogenous melatonin restored DBDPE‐induced developmental inhibition and apoptosis of preimplantation embryos. It mitigated DBDPE‐induced mitochondrial dysfunction and mechanical changes in blastocysts and consequently ameliorated the adverse reproductive outcomes of DBDPE‐exposed blastocysts. The LOAEL was 5 μg/kg bw per day. The results related to developmental neurotoxicity outcomes in these two studies are reported below.

DBDPE (purity NR) was administered by gavage at doses of 0, 100, 320 and 1000 mg/kg bw per day in corn oil to female Crl:CD (SD) rats (F0 dams) from GD6 through lactation day 21. There were no test substance‐related effects observed in this study consistent with systemic toxicity. There was no evidence of maternal toxicity at any dose level. There were no effects on reproductive performance or on litter or offspring developmental parameters. There were no effects on mean male and female pup body weight gains, on postweaning body weight or the onset of sexual maturation as measured by the age of attainment of vaginal opening or preputial separation. The NOAEL for systemic toxicity was 1000 mg/kg bw per day (the highest dose tested) (REACH dossier, as submitted by the registrant(s)).

DBDPE (purity 96.3% DBDPE, 3.6% dodecabromodiphenyl ethane) was administered by gavage at doses of 0, 125, 400 and 1250 mg/kg bw per day in corn oil to female Sprague–Dawley rats from GD6 through lactation day 21 (OECD TG 414). No evidence of maternal or developmental toxicity or teratogenicity was observed in treated rats. The NOAEL was 1250 mg/kg bw per day (the highest dose tested) (REACH dossier, as submitted by the registrant(s)).

DBDPE (purity 96.3% DBDPE, 3.6% dodecabromodiphenyl ethane) was administered by gavage at doses of 0, 125, 400 and 1250 mg/kg bw per day in methylcellulose to female rabbits from GD6 through GD18 (OECD TG414). No evidence of maternal toxicity, developmental toxicity or teratogenicity was observed in treated rabbits. The NOAEL was 1250 mg/kg bw per day (the highest dose tested) (REACH dossier, as submitted by the registrant(s)).

DBDPE (purity NR) was tested in an extended one‐generation reproductive toxicity study with developmental neurotoxicity cohort (OECD TG443) (REACH dossier, as submitted by the registrant(s)). Male and female Sprague–Dawley rats were exposed by gavage to 0, 100, 300 and 1000 mg/kg bw per day. Parental males were exposed for 10 weeks before mating up to necropsy after litters were weaned. Parental females were exposed for 10 weeks before mating, up to 2 weeks during mating, during resulting pregnancies, up to necropsy on day 23 of lactation. F1 males and females were exposed from weaning until termination of respective cohort.

• −Cohort 1A: Reproductive/developmental toxicity testing ‐ treated from weaning to approximately 13 weeks of age.
• −Cohort 1B: Spare cohort ‐ treated from weaning to approximately 14 weeks of age.
• −Cohort 1C: Additional cohort for assessment of sexual maturation ‐ treated from weaning to completion of sexual maturation (approximately seven to 8 weeks of age).

Two additional cohorts were included to assess the developmental neurotoxicity of DBDPE due to the uncertainties of the first neurotoxicity study reported in the developmental toxicity/teratogenicity section of the REACH dossier: Cohort 2A: Developmental neurotoxicity testing (neurobehavioural testing followed by neuro‐histopathology assessment as adults) – treated from weaning up to approximately day 75 of age (see below). Cohort 2B: Developmental neurotoxicity testing – no direct treatment, assigned to neuro‐histopathology assessment at weaning (see below). Administration of DBDPE at dose levels up to and including 1000 mg/kg bw per day was well tolerated with no treatment‐related mortality, no adverse effects on general condition, body weight, food consumption, thyroid hormones, sperm parameters or macropathology in either the F0 or F1 generation. Oestrous cycle, mating performance, fertility, gestation length and gestation index were unaffected by treatment. For F1 litters, the general condition of offspring, litter size, offspring survival, sex ratio and development were unaffected by parental treatment at dose levels up to and including 1000 mg/kg bw per day. In addition, sexual maturation (selected F1), spleen immunophenotyping (F1 Cohort 1A) and ovarian follicle counts (F1 Cohort 1A) were unaffected at all dose levels. There was no adverse effect on anogenital distance. Minor non‐adverse effects were restricted to F0 and F1 male animals that showed a higher incidence of minimal or slight macrovesicular/microvesicular fatty change in the liver at all dose levels; this pathology correlated with the increased body weight relative liver weight for F0 males that received 1000 mg/kg bw per day. The authors concluded that the NOAEL for both systemic toxicity and reproductive/developmental toxicity is 1000 mg/kg bw per day, the highest dose tested.

Developmental neurotoxicity

Several studies were identified and details are reported in Appendix C (Table C.2).

In the same study by Shi, Xu, et al. (2022) reported above in which female Kunming mice were orally exposed to DBDPE (purity NR) at 50 μg/kg per day (gavage) for 30 days, a set of females were given hormones (Pregnant mare's serum gonadotropin (PMSG) and Human chorionic gonadotropin (HCG)) before being mated with untreated males and 1‐, 2‐ and 8‐cell embryos collected not later than 56 h after HCG injection. Blastocytes were collected 94 h after HCG and fetuses taken on GD11.5, GD14.5 and GD18.5. The other set of females was maintained to have pups that were tested for their behavioural performances at 6 weeks of age. At 6 weeks of age, DBDPE‐exposed mice exhibited a higher level of activity in the open‐field whereas anxiety‐related behaviours measured in the open‐field and the light–dark box were unaffected. Results obtained from the two objects recognition test showed that DBDPE exposure resulted in memory and cognitive impairments in both male and female mice. Social deficits were observed only in DBDPE‐exposed females assessed for their social abilities (social memory and novelty) in the three‐chamber test. All these results suggest the developmental toxicity and the neurotoxicity of a preconception exposure to DBDPE.

In the same study of Shi, Qiu, et al. (2022), female Kunming mice were exposed by gavage to DBDPE (purity NR) at 0, 5, 50 and 500 μg/kg bw per day for 14 days and were then tested for their circadian rhythm using a voluntary wheel running test. This test consisted in keeping mice individually in cages within light‐tight chambers. The animals were first exposed to 12:12‐h light/dark cycle for 4 days and exposed to a complete 24 h dark period for the 4 following days to measure their own endogenous circadian rhythm after 14 days of exposure (Shi, Qiu, et al., 2022). Clock gene expression was studied in the suprachiasmatic nucleus of the same animals before and during the test. Results suggested that DBDPE exposure led to impaired circadian rhythm disorders at doses of 50 and 500 μg/kg bw per day.

In the study reported in the developmental toxicity/teratogenicity section of the REACH dossier,³⁹ DBDPE (purity NR) was administered by gavage at doses of 0, 100, 320 and 1000 mg/kg bw per day to female Crl:CD (SD) rats from GD6 to PND21. The animals were divided into four subsets with two of them used for the behavioural testing of the offspring. Detailed clinical observations, open‐field observations and hindlimb and forelimb grip strength were conducted on subset A at PND21, 35, 45 and 60. Motor activity of one pup/sex/litter of the same subset was evaluated on PND13, 17 21 and 61 using a SmartFrame Cage Rack system with two‐dimensional infrared photobeam grid during a session of 60 min.

One pup/sex/litter of the same subset was tested on PND24 and PND63 for their auditory startle response (ASR). Prepulse inhibition and ASR habituation were tested at both timepoints. This test is a conditioning paradigm that allows to test the reflexive contraction of skeletal muscles in response to an abrupt acoustic stimulus, named ASR. The startle stimulus used for each trial was 115 ± 5 dB mixed‐frequency noise burst stimulus approximately 40 milliseconds (ms) in duration. ASR magnitude is reduced with the repetition of the same stimulus, or the presentation of a weaker acoustic or non‐acoustic stimulus before the ASR‐inducing stimulus, a phenomenon known as prepulse inhibition. Trials including prepulse stimuli 4, 8 and 12 dB above the background noise (70 dB) were applied in a sequence of 50 trials ending with the 115 dB stimulus alone.

Learning and memory performances were assessed using a passive avoidance paradigm. One pup/sex/litter from subset B was tested at PND23 and 30, and from subset A at PND64 and 71. The passive avoidance test is designed to evaluate the conditioning learning and memory abilities of the animals. It consists of a two‐chamber apparatus with one being a dark compartment in which they may receive an electric shock. The test assesses the ability of the animal to learn to avoid a painful stimulus with the amount of time that it stays in the lighted chamber being indicative of its learning performances.

No effects of DBDPE on detailed clinical observations (DCOs), motor activity, auditory startle response, and learning and memory performances were reported in both sexes at any dose level at any age.

Subsets A and C were used for necropsy, including brain morphometric measurements. The results showed some test substance‐related morphometric changes in brains of male rats on PND22 and PND72 at 100, 320 and 1000 mg/kg bw per day. There were lower group mean morphometric brain measurements in the cortex (Level 1), hippocampus (Level 3) and cerebellum (Level 5) on PND22 and PND72. The morphometric changes were not associated with statistically significant changes in brain weight or macroscopic brain measurements, although there was a slight non‐significant decrement of group mean brain weight in the highest dose group of males at PND22. There were no microscopic changes in brain, spinal cord, nerve roots or ganglia. There were no test substance‐related changes in female rats.

Peer review by two independent pathologists suggested that these were considered ambiguous findings and an artefact of the brain section and measuring method could not be excluded. Thus, the authors of the study concluded that a NOAEL for developmental neurotoxicity could not be determined with sufficient certainty for males (REACH dossier).

Two additional cohorts were added in the extended one‐generation reproductive toxicity study described above (REACH dossier³⁹) and studied for developmental neurotoxicity of DBDPE. Behaviour of the animals of the cohort 2A was assessed using a functional observational battery (FOB) testing at PND70. The FOB is a non‐invasive procedure investigating various behavioural and activity‐related parameters with the objective to detect gross functional deficits in rodents. Auditory startle response habituation, reactivity investigations and motor activity (ambulatory and rearing) were unaffected by treatment. There were no in cage or in hand observations that could be related to treatment, and arena observations were limited to a slight increase in the incidence of females with elevated/slightly abnormal gait in the treated groups but with no evidence for a dose response. All differences in brain morphometry parameters were considered by the study authors to be incidental as they were inconsistent between sexes, inconsistent between cohorts, the magnitude was considered small, had no histopathological correlates and/or had no significant mean brain weight differences.

Genotoxicity

Since the previous Opinion, three studies have been reported in the REACH dossier³², as submitted by the registrant(s). The CONTAM Panel had no access to the original data. The REACH dossier reported that:

• DBDPE (96.3% DBDPE, 3.6% dodecabromodiphenyl ethane) was negative in a reverse mutation assay in S. Typhimurium TA 1538, TA 1535, TA 1537, TA 98, TA 100 and E. coli WP2 uvrA (OECD TG471) in presence and absence of metabolic activation (Aroclor‐induced rat liver microsomes).

• DBDPE (96.3% DBDPE, 3.6% dodecabromodiphenyl ethane) was negative in an in vitro chromosomal aberration test in Chinese hamster lung fibroblasts (V79) (OECD TG473) in presence and absence of metabolic activation (Aroclor‐induced rat liver microsomes).

• DBDPE (purity 94%–99%; contains nonabromodiphenyl oxide and octabromodiphenyl oxide as impurities) was negative in an in vitro gene mutation assay in L5178Y/TK+/‐ mouse lymphoma cells with and without metabolic activation (S9‐mix of Aroclor 1254‐induced male F344/N rats).

Summary of toxicity in experimental animals

Morphological and ultrastructural changes were observed in the heart and abdominal aorta of rats exposed by gavage for 28 days to doses ≥ 50 mg DBDPE/kg bw per day (Gao et al., 2022; Jing et al., 2019; Zheng et al., 2021). Cardiomyocyte injury and fibrosis were also observed as well as apoptosis and mitochondria damage.

DBDPE promoted atherosclerosis in ApoE−/− mice exposed by gavage to 100 mg/kg bw per day for 6 weeks. It was shown that DBDPE caused oxidative stress in the aorta (Li et al., 2025).

DBDPE induced hepatic lipid metabolism disorders in rats exposed by gavage. Steatosis, accompanied by marked alterations in cellular morphology were observed. In addition, there were inflammatory infiltration and ballooning degeneration in some areas of the liver, suggesting significant metabolic injury. DBDPE exposure resulted in significant increases in TC, TG and LDL levels (Cui et al., 2025).

Liver toxicity (histopathological and hepatic enzymes changes) was observed in mice exposed for 30 days to ≥ 5 mg DBDPE/kg bw per day (Sun et al., 2018) or in rats exposed for 28 days to 50 mg DBDPE/kg bw per day (Sun et al., 2020). In mice, DBDPE may induce drug‐metabolising enzymes, interfere with thyroid hormone levels and could also affect the glucose metabolism homeostasis.

Changes in histological structure and ultrastructure of the thyroid gland and in thyroid hormone levels were noted in rats exposed to ≥ 50 mg/kg bw per day DBDPE by gavage for 28 days (Wang, Chen, et al., 2019). DBDPE exposure could disrupt thyroid function through oxidative stress and perturbations of HPT axis.

Twenty‐eight‐day gavage exposure of male rats to DBDPE result in testicular damage in physiological structure (≥ 50 mg/kg bw per day), decreased sperm number (≥ 50 mg/kg bw per day) and motility (≥ 5 mg/kg bw per day) and increased sperm malformation rates (500 mg/kg bw per day) (Li et al., 2021; Xue et al., 2022).

Exposure of male mice to DBDPE by gavage at 5 or 50 mg/kg bw per day for 7 weeks increased the relative testicular weight. DBDPE decreased the sex hormone levels and sperm count. DBDPE impaired testicular morphology (testicular damage characterised most notably by a reduction in cell layers, disorganisation of the germinal epithelium and pronounced loss of germ cells, especially at 50 mg/kg bw per day) and dose‐dependently increased the sperm malformation rate. Oxidative stress and apoptosis are potential target pathways of DBDPE (Qi et al., 2025).

Exposure of ICR mice to DBDPE (from 0.05 mg/kg bw per day) during gestation caused placental damage, resulting in fetal growth restriction (FGR from 0.5 mg/kg bw per day) (Xie et al., 2025).

DBDPE oral exposure affected oocyte development in mice (Shi & Feng, 2021; Shi, Qiu, et al., 2021; Yan et al., 2022). Asymmetric division of oocytes was markedly impaired (Shi & Feng, 2021). DBDPE exposure in utero caused also low birth weight in cubs and then the cubs grow rapidly during lactation. A loss of embryos and a significant lower litter size were noted in female Kunming mice exposed to 50 μg DBDPE/kg bw per day for 30 days before mating. The embryo length of DBDPE‐exposed mice was also significantly reduced, indicating delayed fetal development (Shi, Xu, et al., 2022). In addition, DBDPE (5–500 μg/kg bw per day) inhibited preimplantation embryo development in a dose‐dependent manner by affecting the circadian rhythm and endogenous melatonin levels (Shi, Qiu, et al., 2022) (see Section  3.3.1.1.6 on mode of action).

Brain and behavioural toxicity of DBDPE related to learning and memory, spontaneous activity and social behaviour were reported in Kunming mice gestationally exposed at 50 μg/kg bw per day for 30 days (Shi, Xu, et al., 2022) whereas impaired circadian rhythm disorders were observed at doses of 50 and 500 μg/kg bw per day for 14 days (Shi, Qiu, et al., 2022). Two developmental neurotoxicity studies reported in the REACH dossier did not show any effects of DBDPE on learning memory, motor activity, auditory startle and general behaviour assessed through a functional observational battery in rats early exposed to DBDPE (GD6‐PND21) at doses of 100, 320 and 1000 mg/kg bw per day.

These contradictory results obtained at very different doses and with different rodent species from one study to another raise the question of the neurotoxicity of DBDPE, which has to be investigated in further studies.

No evidence of maternal toxicity, developmental toxicity or teratogenicity was noted in rats and rabbits exposed during gestation by gavage up to 1250 mg DBDPE/kg bw per day (Hardy et al., 2010; REACH dossier).

In an extended one‐generation reproductive toxicity in rats reported in the REACH dossier, administration of DBDPE at dose levels up to and including 1000 mg/kg bw per day did not induce adverse effects in the parental generation nor in the offspring. The NOAEL for both systemic toxicity and reproductive/developmental toxicity is 1000 mg/kg bw per day, the highest dose tested.

No long‐term or carcinogenicity study was identified.

According to the data reported in the REACH dossier, DBDPE was not mutagenic in bacteria or in mammalian cells and did not induce chromosomal aberrations in Chinese hamster lung cells. Based on this information, the CONTAM Panel concluded that DBDPE is not genotoxic.

3.3.1.1.4. Computational analysis

The QSAR Toolbox identified experimental data for the Ames test on DBDPE as retrieved by the previous Opinion (see above and Annex B).

VEGA models predicted negative Ames test results for DBDPE with ‘MODERATE reliability’ (four models). These predictions were not further considered due to their limited reliability (see Section 2.2).

3.3.1.1.5. Observations in humans

In the previous Opinion, no human studies on DBDPE were reported. Since then, six publications were identified. In Appendix F an overview of the epidemiological studies identified reporting on the BFRs considered in this Opinion, including those on DBDPE, is shown.

Five of the included publications pertained to populations in China (two were from the same study population) and one was located in the USA. Exposure was assessed in serum in four studies; the fifth study assessed exposure from indoor air and wristbands. Various endpoints were assessed including thyroid hormone profile (five studies), thyroid cancer (one study), serum lipids (one study), liver injury markers (one study) and changes in mitochondrial DNA copy number (one study).

In a study by Zhao et al. (2021), five of the BFRs assessed in the current Opinion (see Appendix F) were determined in 172 serum samples (118 males and 54 females) collected from non‐occupational residents of a major BFR‐producing region in North China. The age of the participants varied from 21 to 58 with an average of 36 years. All the five BFRs presented high detection frequencies (> 90%), and DBDPE was the most abundant one. Using a cross‐sectional study design, the relationship between the serum levels of the BFRs and a series of thyroid hormones, serum lipids and liver injury biomarkers, was evaluated. Results showed that a 10‐fold increment in the serum DBDPE level was associated with lower TT3 level after adjustment for confounders (−0.037 nmol/L) [95% CI: −0.070, −0.003]. DBDPE was not associated with any of the other biomarkers.

In another cross‐sectional study from the same region by Chen et al. (2019), a total of 302 blood samples were collected from two populations in the largest DBDPE manufacturing area located in North China: 133 occupationally exposed workers from a DBDPE manufacturing plant and 169 non‐ occupationally exposed residents from a nearby food processing plant. The levels of DBDPE and thyroid function parameters (TT4, FT4, TT3, FT3, TSH, TG‐Ab and TPO‐Ab) were measured in serum samples. Serum concentrations of DBDPE ranged from 3.148 to 54,360 ng/g lipid, with a geometric mean of 332.6 ng/g lipid. A 10‐fold higher DBDPE concentration was associated with 4.73 nmol/L higher TT4 [95% confidence interval (CI): 2.75, 6.71] and 0.046 nmol/L [95% CI: 0.012, 0.081] higher TT3, corresponding to approximately 4.73% [95% CI: 2.75%–6.71%] and 2.38% [95% CI: 0.62%–4.20%] higher concentrations, respectively. DBDPE in serum was also significantly and positively associated with the concentrations of TG‐Ab and TPO‐Ab.

In the same study population, Chen et al. (2022) also assessed the association between DBDPE and changes in mitochondrial DNA copy number in whole blood (in 131 workers and 161 non‐occupationally exposed residents). A 10‐fold increase in the serum BDBPE concentration was associated with a reduced DNA copy number [β = −0.063, 95% CI: −0.088 to −0.038], corresponding to a decrease of 6.91% [95% CI: −9.65% to −4.17%].

In the study by Guo et al. (2018) on the association with thyroid hormone levels in 10‐year‐old students living near a petrochemical complex in South China (for details see description in Section 3.2.4.2), DBDPE levels were not associated with the serum thyroid hormone levels. The sum of the 13 PBDE congeners and the sum of all eight new flame retardants were associated with higher TT3 levels. The authors concluded that this positive association was mainly due to the dominant contribution of BDE‐209.

In the study by Liu, Li, et al. (2022) that assessed the association between the levels of several BFRs in serum and thyroid cancer risk in Shandong Province (China) (for details see description in Section 3.2.4.2), overall, and in sex‐stratified logistic regression models, DBDPE exhibited significantly higher ORs in the third and fourth quartiles in both male and female participants but confidence intervals were wide indicating imprecise effect estimates (OR males: 3rd quartile: 33.35 [95%CI: 4.45, 250.02], 4th quartile: 13.56 [95%CI: 2.30, 79.85]; OR females: 3rd quartile: 1.93 [95%CI 0.92, 4.03], 4th quartile: 4.13 [95%CI: 2.04, 8.34]). DBDPE was not associated with thyroid hormone levels in the control participants.

In the USA study by Wang, Romanak, et al. (2019) (for details see Section 3.2.4.2), DBDPE concentrations, measured in wristbands of 101 adult study participants, were not statistically significantly associated with thyroid hormone profiles.

In summary, the available epidemiological evidence on DBDPE pertained to a limited number of cross‐sectional studies in populations in China (n = 5) and the USA (n = 1), five of which were studies in adults and one was in children. Results for thyroid hormone levels are inconsistent across studies, with positive, inverse and null associations reported for TT3 in relation to serum DBDPE levels, as well as one study reporting increased TT4 levels. Further, one study reported association with higher odds of thyroid cancer, and one reported changes in mitochondrial DNA copy number. No statistically significant associations were observed for lipid levels, liver injury markers or other thyroid function markers.

3.3.1.1.6. Studies in other model systems and mode of action

A number of studies to inform the MOA of DBDPE were identified in the scientific literature since the previous Opinion. An overview of those studies is given in Appendix G (Table G.2).

In summary, morphological and ultrastructural changes were observed in the heart and abdominal aorta after 28‐day gavage exposure of rats to DBDPE (Jing et al., 2019; Gao et al., 2022; Zheng et al., 2021, see above). These effects can be induced by oxidative stress and/or inflammation that can lead to endothelial dysfunction and cardiovascular injury. Cardiomyocyte injury and fibrosis were also observed as well as apoptosis and mitochondria damage.

Li et al. (2025) showed that exposure to DBDPE promotes the formation of atherosclerotic plaque in aortic roots of ApoE−/− mice by increasing macrophage recruitment and foam cell formation.

Liver toxicity (histopathological and hepatic enzyme changes) was observed in mice (Sun et al., 2018) or rats exposed for 30 or 28 days to DBDPE (Sun et al., 2020). DBDPE may induce drug‐metabolising enzymes including CYP450s and UDPGT, interfere with thyroid hormone levels mediated by AHR and CAR signalling pathways and could also affect the glucose metabolism homeostasis.

The impact of DBDPE on the liver in rats and L02 cells (immortalised hepatocyte line) was investigated. DBDPE exposure could induce lipid metabolism disorders, up‐regulate the expressions of Notch pathway genes and promote Arrestin domain containing protein 1‐ mediated microvesicles (ARMMs)‐derivation (Cui et al., 2025).

Changes in histological structure and ultrastructure of the thyroid gland and in thyroid hormone levels were noted in rats exposed by gavage to DBDPE for 28 days (Wang, Chen, et al., 2019). The authors noted that DBDPE exposure disrupted thyroid function associated with oxidative stress and perturbations of HPT axis.

Twenty‐eight‐day gavage exposure of male rats to DBDPE resulted in testicular damage in physiological structure, decreased sperm number and motility and increased sperm malformation rates (Li et al., 2021; Xue et al., 2022). By shortening telomere length and reducing telomerase activity DBDPE caused also cell senescence and apoptosis in testis of rat (Li et al., 2021). There was a dose‐related significant increase in apoptotic index in testis at 50 and 500 mg/kg bw per day. Consistently, immunofluorescence analysis showed that the DNA damage marker γ‐H2AX increased significantly in rat testicular cells after exposure for 28 days to 500 mg DBDPE/kg bw per day (Xue et al., 2022). Oxidative stress was demonstrated in the testis (Li et al., 2021). DBDPE‐induced glycolipid metabolism disorder and promoted apoptosis of spermatogenic cells through the mitochondrial‐mediated apoptosis pathway, leading to reproductive toxicity by decreasing sperm quantity and quality (Xue et al., 2022). No 1‐ or 2‐generation reproductive toxicity studies are available.

DBDPE oral exposure affected oocyte development in mice (Shi & Feng, 2021; Shi, Qiu, et al., 2021; Yan et al., 2022). Asymmetric division of oocytes was markedly impaired (Shi & Feng, 2021). DBDPE exposure in utero caused also low birth weight in pups and then the pups grow rapidly during lactation. In utero exposure to DBDPE affects the whole‐body metabolic profile of newborns; the pups exhibited accelerated energy utilisation and lipid accumulation for rapid growth and ultimately obesity (Yan et al., 2022). DBDPE exposure impaired mitochondrial function, causing oxidative damage, autophagy and apoptosis in oocytes (Shi, Qiu, et al., 2021).

DBDPE did not affect spindle rotation in oocytes after fertilisation, but led to a decrease of pronuclei (PN) in zygotes from Kunming mice (see above for description of the study). DBDPE was found to interfere with the self‐assembly of F‐actin in pronuclei of embryo from Kunming mice resulting in their size reduction in zygotes. In addition, DBDPE reduced the expression of the Zscan4c transcription factor involved in maintaining genome stability during embryogenesis. This early developmental instability may lead to metabolic and structural defects observed in later life. The authors noted that these data indicated that DBDPE exposure might affect embryonic development and result later in behavioural impairments (learning and memory, spontaneous activity, social behaviour) (Shi, Xu, et al., 2022).

DBDPE (5, 50, 500 μg/kg bw per day) exposure‐induced circadian rhythm disorder in adult female Kunming mice exposed by gavage for 14 days. It decreased also endogenous melatonin levels during pregnancy, which impaired preimplantation embryo development. DBDPE exposure impaired embryos by inducing oxidative stress, abnormal DNA methylation, mitochondrial dysfunction and mechanical property changes. Finally, DBDPE‐induced embryo damage led to adverse reproductive outcomes (Shi, Qiu, et al., 2022).

To investigate the mechanisms of reproductive toxicity of DBDPE, experiments using the mouse premeiotic germ cell line GC‐2 were performed. Results of transcriptomics from GC‐2 cells indicated that DBDPE lead to changes in oxidative stress, apoptosis, inflammation, proliferation and differentiation. Furthermore, DBDPE exposure inhibited the Kelch‐like ECH‐associated protein1‐NF‐E2‐related factor2 (Keap1‐Nrf2) signalling pathway (Qi et al., 2025).

Exposure of ICR mice to DBDPE during gestation caused placental damage, resulting in fetal growth restriction (FGR) (Xie et al., 2025). Mechanistically, the mitochondrial dysfunction induced by gestational DBDPE exposure initiates oxeiptosis (a cell death mechanism, which involves reactive oxygen species (ROS)‐dependent activation of Keap1‐Phosphoglycerate mutase family member 5 (PGAM5)‐Apoptosis‐inducing factor mitochondria associated 1 (AIFM1), referred to as the (Keap1‐PGAM5‐AIFM1)) in placental trophoblasts, exacerbating placental injury and ultimately leading to FGR.

Neurotoxicity of DBDPE has been suggested in vitro through the exposure of human neuroblastoma SK‐N‐SH cells (Dong, Wang, Wang, et al., 2023; Dong, Wang, Zhang, et al., 2023). Results showed the ability of DBDPE to induce cell apoptosis through the initiation of key cellular events such as mitochondrial dysfunction, ROS production, DNA damage through caspase‐dependent and ‐independent pathways. Neurotoxicity of DBDPE was assessed also in zebrafish showing a dose‐related increase in the locomotor activity of larvae exposed to DBDPE (0, 3, 30, 300 μg/L) during the first 5 days after fertilisation (Li et al., 2024). The exposure of zebrafish embryos over the 120 h postfertilisation showed significant increases in the locomotor activity during the dark–light transition period, and the whole‐body contents of various neurotransmitters like glutamate, epinephrine, norepinephrine and different metabolic changes suggesting a mitochondrial dysfunction as a potential mechanism for DBDPE neurotoxicity (Yang, Zhu, et al., 2023). The hyperactivity observed has been related also to the disruption of the cholinergic neurotransmission system with ChAT as a potential target for neurotoxicity of DBDPE (Hua et al., 2022). A transgenerational study was conducted to assess the brain (protein and neurotransmitter contents, global DNA methylation and specific gene expression) and behaviour (sociability and recognition) toxicity of DBDPE in three successive generations (Sun et al., 2023). The results showed some specific brain and behaviour alterations in the female offspring whereas the males remained quite unaffected, suggesting the ability of DBDPE exposure at environmentally relevant doses to induce transgenerational neurotoxicity in zebrafish.

3.3.1.1.7. Consideration of critical effects and dose–response analysis

Based on the information reported in the REACH dossier, the CONTAM Panel concluded that DBDPE is not genotoxic. No long‐term or carcinogenicity studies were identified.

Toxicological effects (histopathological, enzymatic or hormonal effects) were observed in the heart, abdominal aorta, liver or thyroid in some sub‐acute toxicity studies in rats at doses ≥ 50 mg/kg bw per day. In most of the sub‐acute and sub‐chronic studies performed according to the OECD test guidelines and reported in the REACH dossier, no effect was observed up to 1000 mg/kg bw per day (the highest dose tested).

No evidence of maternal toxicity, developmental toxicity or teratogenicity was noted in rats and rabbits exposed during gestation by gavage up to 1250 mg DBDPE/kg bw per day and DBDPE up to and including 1000 mg/kg bw per day did not induce adverse effects in the parental generation nor in the offspring in an extended one‐generation reproductive toxicity in rats. However, testicular damage was observed in rats exposed for 28 days to ≥ 50 mg/kg bw per day and exposure of ICR mice to DBDPE (from 0.05 mg/kg bw per day) during gestation caused placental damage, resulting in fetal growth restriction (FGR from 0.5 mg/kg bw per day).

Two studies were performed in female Kunming mice exposed to low doses of DBDPE (5–500 μg/kg bw per day and showed effects on embryos (loss of embryos, decrease in litter size and reduction in embryo length) indicating delayed fetal development (Shi, Qiu, et al., 2022; Shi, Xu, et al., 2022). DBDPE inhibited also preimplantation embryo development in a dose‐dependent manner by affecting the circadian rhythm and endogenous melatonin levels (see Appendix C, Table C.1).

Brain and behavioural toxicity of DBDPE related to learning and memory, spontaneous activity and social behaviour were reported in Kunming mice gestationally exposed at 50 μg/kg bw per day for 30 days (Shi, Xu, et al., 2022).

It is to be noted that two developmental neurotoxicity studies and an EOGRTS (REACH dossier) did not show any effect of DBDPE on learning memory, motor activity, auditory startle and general behaviour assessed through a functional observational battery in rats early exposed to DBDPE at doses up to 1000 mg/kg bw per day.

The CONTAM Panel noted the different effect levels reported among these studies, that could be related to different rodent species, different study designs, and using different tests which explored different aspects of neurobehaviour (see details in Appendix C, Table C.2). Therefore, it can be concluded that the results of the studies are not contradictory, and all evidence was included for hazard characterisation.

The CONTAM Panel selected the study by Shi, Qiu, et al. (2022) as the critical study. Critical effects were the decrease in the rate of blastocyst formation and the increase in total wheel running activity in Kunming mice exposed during gestation to DBDPE. The decrease in the rate of blastocyst formation was the most sensitive effect. It was dose‐related and observed from the lowest dose; while the effects on the other embryo stages appeared only at the two highest doses. Likewise, the effect in wheel running activity was only significant at the two highest doses.

The CONTAM Panel performed BMD modelling according to the 2022 EFSA Guidance on the use of the BMD approach in risk assessment (EFSA Scientific Committee, 2022). The Panel selected a benchmark response (BMR) of 10% for the neurodevelopmental effects, considering the high variability typically observed in neurobehavioural endpoints. For the rate of blastocyst formation, the Panel selected a BMR of 10%. This approach is consistent with that applied in the EFSA CONTAM Panel PBDEs Opinion (EFSA CONTAM Panel, 2024a). The results of the BMD modelling for the critical study in Kunming mice (Shi, Qiu, et al., 2022) are summarised in Table 20. Details of the BMD analyses, including the individual reports of the modelling are shown in Annex D.

TABLE 20.

Benchmark dose (BMD) modelling for the study with DBDPE in female mice (Shi, Qiu, et al., 2022).

Observed effect	BMDL10	BMD10	BMDU10
	μg/kg bw per day
Neurobehavioural effects
Total wheel running activity (individual raw data) LD a	1.07	2.64	6.07
Total wheel running activity (individual raw data) DD a	2.03	3.76	5.86
Reproductive effects
Rate of blastocyst formation (individual raw data) b	0.63	4.10	288
Rate of blastocyst formation (individual raw data) no gamma models b	0.62	4.00	303

^a DD (24‐h dark cycle) and LD (12:12‐h light/dark cycle) used as covariate.

^b Criteria not satisfied (none of the models provide an adequate fit to the data, BMDU/L > 50).

The lowest BMDL₁₀ for DBDPE was 1.07 μg/kg bw per day, based on the increase of total wheel running activity (LD). The CONTAM Panel selected this lowest BMDL₁₀ as the Reference Point. Although the decrease in the rate of blastocyst formation yielded a lower BMDL, the BMD models did not provide an adequate fit to the data, resulting in a credible interval too wide to be considered reliable (see Table 20 and BMD reports in Annex D).

The CONTAM Panel concluded that, due to the limitations and uncertainties in the current database, the establishment of a health‐based guidance value for DBDPE was not appropriate. Instead, the MOE approach was used for risk characterisation (EFSA Scientific Committee, 2025).

Usually, an MOE of 100, covering variability with respect to kinetic and dynamic differences between animal species and humans (factor 4 × 2.5 = 10) and within the human population (factor 3.2 × 3.2 = 10), is considered sufficient to conclude that exposure does not raise a health concern.

According to the EFSA Scientific Committee Guidance on selected default values, an additional factor can be considered in case of deficiencies in the database on a case‐by‐case basis. A default value has not been proposed, as it will be directly dependent on the data set available (EFSA Scientific Committee, 2012). The WHO/IPCS (1994, 1999) recommended a factor of 3 or 5 if there are minor deficiencies in the database and a factor of 10 if there are major deficiencies in the database. The CONTAM Panel noted that the limitations in the database for DBDPE are the lack of long‐term toxicity studies, while there are several studies on repeated dose toxicity, developmental and reproductive toxicity, and neurodevelopmental toxicity. It was noted that DBDPE is not genotoxic, and the Reference Point comes from a study showing mild effects on a sensitive endpoint, at much lower doses than those at which effects were seen in other studies. Therefore, the CONTAM Panel considered an additional factor of 3 to be sufficient for deficiencies in the database.

Overall, the CONTAM Panel considered that MOEs ≥ 300 (100 for inter and intra‐species variability, and 3 for the gaps in the toxicological database) do not raise a health concern.

3.3.1.2. Occurrence data and dietary exposure assessment

Occurrence data of DBDPE in food were submitted to EFSA (see Section 3.1.1). Due to the limited number of quantified results (n = 3), the CONTAM Panel considered these data insufficient to perform any exposure estimation.

In the scientific literature, several studies were identified reporting on the occurrence of DBDPE and an overview is given in Section 3.1.2.

One study was identified reporting estimates of dietary exposure to DBDPE. In UK, Ma et al. (2023) estimated the average and high‐end daily dietary intake for a number of BFRs, including DBDPE, for all population groups. Concentrations of BFRs in UK food items were obtained from a total of 108 samples collected between 2020 and 2021 from three supermarkets. Consumption data for UK citizens of all age groups were obtained from the National Diet and Nutrition Survey (2008–2019). Average dietary exposure to DBDPE was calculated in the range of 0.066–0.24 ng/kg bw per day, while the high‐end estimations, assuming food contaminated at the average concentrations consumed at the mean rate +2 standard deviations, were estimated between 0.39 and 1.5 ng/kg bw per day across population groups (from 0 up to > 75 years old).

For breastfed infants, there were insufficient data in human milk from European countries reported in the literature to make robust exposure estimates. To provide an indication of the potential exposure, the CONTAM Panel made an estimation for average and high consumption of human milk (see Section 2.4) considering the available data reported in the literature (Tao et al., 2017; Wemken et al., 2020, see Section 3.3.1.1.2).

The exposure scenario based on average human milk consumption and the only reported mean concentration in human milk samples from Ireland of 0.74 ng/g lipid (Wemken et al., 2020), resulted in an estimate of 3.4 ng/kg bw per day. The study by Tao et al. (2017) did not report on mean levels.

For the exposure scenario based on high human milk consumption none of the studies identified reported P95 concentrations, only maximum concentrations. Thus no estimation of exposure was performed.

It should be noted that these estimates are based on human milk samples from one European country collected about a decade ago, and with low detection frequencies. It is unknown whether this is representative of current exposure to this BFR.

3.3.1.3. Risk characterisation

Insufficient occurrence data on DBDPE in food were submitted to EFSA to allow robust estimates of dietary exposure for the EU population to be made. Limited data obtained from the scientific literature indicated that average dietary exposure from food in the UK was up to 0.24 ng/kg bw per day and high‐end estimates were up to 1.5 ng/kg bw per day. The selected Reference Point for DBDPE, based on the increase of total wheel running activity, is 1.07 μg/kg bw per day, i.e. 1070 ng/kg bw per day. This MOE of about 700 does not raise a health concern, recognising the limitations in the exposure data, but also the low Reference Point derived from a study in which effects were observed at much lower doses than in other studies (μg/kg bw per day compared to mg/kg bw per day).

3.3.1.4. Uncertainty analysis

The uncertainty analysis was performed using the methods described in Section 2.8. Identified sources of uncertainty affecting the assessment are documented in Table 21.

TABLE 21.

Uncertainties identified and their impact on the outcome of the hazard identification/characterisation and risk characterisation of DBDPE.

Description of the uncertainty		Impact of the uncertainty on the hazard identification and characterisation a
Chemical composition and analytical methods
Dosing and chemical composition, including impurities	Uncertainty associated with the dose in the critical studies used in the risk assessment	2 – Medium impact. Purity not reported in most studies, including the critical study.
Hazard identification and characterisation
ADME	ADME in relation to the critical studies	1 – Low impact. Studies have been identified. DBDPE is poorly absorbed, and eliminated via faeces (which mainly represent the unabsorbed dose). An increase of concentration in liver and adrenal tissues has been observed after 10 days of exposure, but the radioactivity retained in theses tissues was very low (0.06% and 0.008% of the total dose).
	Information on transfer rate to animal products	0 – Negligible impact. No studies identified, but no impact on risk assessment.
	Transfer via mother's milk	1 – Low impact. No studies have been identified in rodent. DBDPE was rarely detected or was reported to be below the LOD in human milk.
Toxicity studies in experimental animals: endpoints and study design	Limitations in the design of the studies that can result in uncertainties	1 – Low impact. The CONTAM Panel noted the lack of consistency in the findings of the studies, but considered that these were possibly related to different rodent species, different study designs and using different tests which explored different aspects of neurobehaviour (see Appendix C, Table C.2). It was therefore concluded that the results of the studies are not contradictory, and all evidence was included for hazard characterisation.
	Relevance for humans of the adverse effect.	1 – Low impact. The critical effects selected (total wheel running activity and decrease in the rate of blastocyst formation) to establish the Reference Point are relevant for humans.
Genotoxicity	Uncertainty in the assessment of genotoxicity	1 – Low impact. DBDPE is not genotoxic. No access to the original data (REACH dossier), but all the in vitro tests which cover every genetic endpoint are negative.
MOA	Uncertainties associated with the data on mode of action	0 – Negligible impact. There are suitable mechanistic data which support the relevance of the critical effects.
Selection of Reference Point	Weaknesses in the critical study	1 – Low impact. BMD modelling criteria not met for rate of blastocyst formation. Potential pseudo‐replication due to lack of information on which blastocysts derived from which dam. However, this was not selected as the Reference Point.
	Weaknesses in non‐critical studies, or absence of studies, and uncertainty about whether the endpoints they tested might have been critical if the weaknesses were not present	1 – Low impact. No long‐term or carcinogenicity study. The compound is not genotoxic and a factor of 3 was applied to account for deficiencies in the database.
Description of the uncertainty		Impact on the risk characterisation a
Margin of exposure (MOE)	–	1 – Low impact. The usual factor of 100 (to cover for variability with respect to kinetic and dynamic differences between animal species and humans, and within the human population) is considered sufficient. The additional factor of 3 for deficiencies in the database is considered sufficient.
Risk characterisation	–	2 – Medium impact. Not possible for EFSA to perform an exposure assessment since insufficient occurrence data were submitted to EFSA. Based on dietary exposure estimates reported in the literature for one European country (UK), the resulting MOEs did not raise a health concern. It is uncertain whether this is representative of the situation in other European countries. However, the calculated MOE is more than twice that identified as being of low concern.

Abbreviations: ADME, absorption, distribution, metabolism, excretion; BMD, benchmark dose; BMDL, benchmark dose lower credible limit; BMR, benchmark response; MOA, mode of action; TK, toxicokinetic.

^a 0 – Uncertainty with negligible impact; 1 – Uncertainty with low impact; 2 – Uncertainty with medium impact; 3 – Uncertainty with high impact. This ordinal scale provides a relative assessment of the impact of the individual uncertainties as an aid to quantifying their combined impact by expert judgement, as described in section 11 of EFSA Scientific Committee (2018).

Uncertainty analysis: Genotoxicity

Uncertainties affecting the evidence on the genotoxic potential of DBDPE (see Section 3.3.1.1) were assessed to have low impact on the assessment (Table 21). The CONTAM Panel did not have access to the original data (REACH dossier), but the in vitro tests were all negative and covered every genetic endpoint.

Based on the overall evidence, and taking into account the associated uncertainties (Table 21), and following an expert group discussion, the CONTAM Panel judged with > 95% certainty (i.e. probability) that DBDPE is not genotoxic.

Uncertainty analysis: Reference Point

The CONTAM Panel identified substantial uncertainties affecting the hazard assessment for DBDPE (Table 21) that did not allow the establishment of a HBGV. Instead, the MOE approach was used, addressing the hazard uncertainties by applying UFs of 100 and 3.

The CONTAM Panel noted that the limitations in the database for DBDPE are the lack of long‐term toxicity studies, while there are several studies on repeated dose toxicity, developmental and reproductive toxicity, and neurodevelopmental toxicity. It was noted that DBDPE is not genotoxic, and the Reference Point comes from a study showing mild effects on a sensitive endpoint, at much lower doses than those at which effects were seen in other studies. Therefore, the CONTAM Panel considered an additional factor of 3 to be sufficient for deficiencies in the database.

Considering the combined impact of the uncertainties and the extent to which they are addressed by the applied UFs, the CONTAM Panel judged with 90%–95% certainty (i.e. probability) that, when using the Reference Point of 1.07 μg/kg bw per day, MOEs ≥ 300 would not raise a health concern for non‐neoplastic effects.

Uncertainty analysis: Risk characterisation

Insufficient occurrence data in food on DBDPE were submitted to EFSA to allow robust estimates of dietary exposure for the EU population to be made. Based on data from the scientific literature for one European country (UK, see Section 3.3.1.2), the CONTAM Panel noted this would result in an MOE of 700.

The CONTAM Panel noted that the reported high exposure estimates were based on mean concentrations plus 2 SDs and referred only to the UK.

Considering the MOE of 700 for the UK exposure estimates, the uncertainties affecting extrapolation to exposure in other European countries and uncertainties affecting the Reference Point (see above), the CONTAM Panel judged with 66–90% certainty (i.e. probability) that current P95 exposure of the European population to DBDPE does not raise a health concern.

3.3.2. 1,2‐Bis(2,4,6‐tribromophenoxy)ethane (BTBPE) [CAS number 37853‐59‐1]

BTBPE is a non‐polymeric additive flame retardant that has been used as a replacement for the PBDE technical mixture OctaBDE in polystyrene, thermoplastics and resins (EFSA CONTAM Panel, 2012b).

The physicochemical properties of BTBPE are reported in Table 1. It has been identified as a substance of very high concern (SVHC) due to its vPvB properties (see Table 2). BTBEP is currently not registered under REACH, and thus its production volume and use within the EU should be lower than one tonne per year. Nevertheless, the measured concentrations during service life (indoor dust) and end‐of‐life (indoor air at WEEE recycling facility) are comparable to those of DBDPE (see Section 3.3.1), which is the aromatic BFR with the highest registered aggregated volume at ~26,700 tonnes/year. As there are currently no active registration dossiers under REACH for this substance, the quantities present in imported articles into the EU do not give rise to registration obligations under REACH.

3.3.2.1. Hazard identification and characterisation

3.3.2.1.1. Toxicokinetics

In the previous Opinion (EFSA CONTAM Panel, 2012a, 2012b), two studies in rats were identified (Hakk et al., 2004; Nomeir et al., 1993). The results of these studies showed that BTBPE was poorly absorbed; after oral administration of ¹⁴C‐BTBPE, less than 1% of radioactivity was retrieved in urine whereas more than 80% was excreted in faeces (Hakk et al., 2004; Nomeir et al., 1993).

Since the previous Opinion, no relevant studies were identified.

3.3.2.1.2. Levels in humans

The previous Opinion reported a study on five serum samples from Sweden in which BTBPE could not be identified above the LOD (1.31 ng/g lipid) (Karlsson et al., 2007).

Since then, several studies have been identified in the scientific literature reporting on the levels of BTBPE in human samples from European countries, including human milk and blood/serum. These studies are briefly described below and details can be found in Appendix E (Table E.1).

Regarding human milk samples,

• BTBPE was reported to be below the LOD (0.2–0.3 ng/g lipid) in 11 pooled human milk samples collected in 2010 from 109 first‐time mothers in Ireland (Pratt et al., 2013).

• In Sweden, BTBPE was detected in one out of two pooled human milk samples collected in 2009–2010 in Sweden at the LOD (1.0 pg/g ww) (Sahlström et al., 2015).

• Čechová et al. (2017) analysed BTBPE in human milk samples from Norway (2003–2006, n = 305), the Netherlands (2011–2014, n = 116) and Slovakia (2011–2012, n = 37) with a detection frequency between 22 and 26%. Mean and P95 concentrations across countries ranged from 0.023 to 1.103 ng/g lipid and from 0.078 to 2.380 ng/g lipid, respectively.

• In the study by Tao et al. (2017) that analysed several BFRs in human milk samples collected in the UK in 2010 (n = 25) and 2014–2015 (n = 10), the levels of BTBPE in the samples from 2010 ranged from < 0.1 to 56 ng/g lipid (detection frequency 28%), while in 2014–2015 a higher detection frequency (40%) was reported and concentrations ranged from < 0.1 to 0.71 ng/g lipid.

Regarding blood/serum samples,

• Cequier et al. (2013) detected BTBPE in 2 out of 10 serum samples from Norway (sampling year not reported) at concentrations of 0.35 and 0.48 ng/g lipid (LOD = 2.3 pg/mL). In a later study, Cequier et al. (2015) also analysed BTBPE in serum samples from 46 women from a Norwegian mother–child cohort established in 2012 with median (mean) levels of 0.19 (0.21) ng/g lipid (detection frequency 9%).

• BTBPE was reported to be below the LOD in 61 serum samples collected in 2013 from a Norwegian cohort (LOD = 17 pg/serum sample, volume not reported, Tay et al., 2019) as well as in 24 matched serum samples from mothers and their toddlers collected 2009–2010 in Sweden (LOD = 10 pg/25 μL, Sahlström et al., 2014).

• Sales et al. (2017) detected BTBPE just below the LOQ of 0.1 pg/g serum in one pool of human cord blood from Belgium.

3.3.2.1.3. Toxicity in experimental animals

Studies considered in the previous EFSA assessment

The previous EFSA Opinion cited three studies on toxicity of BTBPE in experimental animals (EFSA CONTAM Panel, 2012b). Only one of these concerned toxicity in mammalian species. Acute toxicity studies of BTBPE indicated an oral LD50 > 10,000 mg/kg bw in rats and dogs, with no obvious effects observed in rats exposed to 35 mg/kg bw per day via the diet for 14 days (Matthews, 1984, as cited by Nomeir et al., 1993). Additionally, BTBPE impacted hepatic gene expression related to xenobiotic biotransformation, lipid metabolism and thyroid hormone regulation, increasing cytochrome Cyp1a4 and Cyp1A5 mRNA levels and altering deiodinase expression with decrease in type III iodothyronine 5′‐deiodinase and increase in type I iodothyronine deiodinase mRNA levels (Egloff et al., 2011).

Regarding genotoxicity, the previous Opinion reported that mutagenicity assays with yeast (Saccharomyces cerevisiae) and in Salmonella strains TA98, TA100, TA1535, TA1537 and TA1538, both in the presence and absence of rat liver S9 fraction, were negative (WHO/IPCS, 2005) (see EFSA CONTAM Panel, 2012b).

Studies published since the previous EFSA assessment

Since the previous Opinion, a few additional experimental animal studies informing on toxicity of BTBPE have been published.

Cui, Xiao, Wang, et al. (2024) investigated the effects of BTBPE on the intestinal mucosa. Sprague–Dawley male rats (6–7 weeks old) were treated with BTBPE at 0, 2.5, 25, 250 mg/kg bw by oral gavage once daily for 30 days. Histological changes of the rat jejunum and ileum were observed through H&E staining and qualitatively reported. A small amount of inflammatory cell infiltration was observed in the 2.5 mg/kg bw per day group and this was further increased in the 25 and 250 mg/kg bw per day group. Goblet cells were slightly reduced, and epithelial cells were slightly damaged in both the 2.5 and 25 mg/kg bw per day groups. In the group treated with 250 mg/kg bw per day, the number of goblet cells was much decreased and epithelial cells were severely damaged. In a complementary study in Sprague–Dawley male rats subjected to an identical treatment regime there was a dose‐dependent increase in the relative proportion of necrotic cells in the liver, which was statistically significant at 25 and 250 mg/kg bw per day (Cui, Xiao, Zhang, et al., 2024). This was associated with dose‐related changes in expression of genes related to antioxidant defence and inflammation. Cui et al. (2026) reported further data on liver effects from the same or an identical experiment. Dose‐dependent effects on liver fibrosis markers were noted, with most markers being significantly elevated at 25 mg/kg bw per day and one (Masson stain) at 2.5 mg/kg bw per day.

Male Sprague–Dawley rats (180–220 g) were treated with 2.5, 25 or 250 mg/kg bw BTBPE daily by gavage for 30 days (Zhang et al., 2023). There was a small but statistically significant increase in body weight of the animals in the 250 mg/kg bw group. There was no change in relative or absolute thymus or liver weights. Serum FT4 concentrations were significantly lower in all treatment groups, compared with the control, but no dose–response relationship was observed. TT4 concentration was reduced in the 25 mg/kg bw group only. FT3 and TT3 concentrations were unaltered.

In a later study by the same laboratory, male Sprague–Dawley rats (6–7‐week‐old, 180–220 g) were given BTBPE (0 or 250 mg/kg bw per day) by daily oral gavage for 28 days followed by 12 h fasting before they were killed (Yuan et al., 2024b). TT4 and FT4 concentrations were numerically reduced by average in the treated group, but this was not statistically significant.

Shockley et al. (2020a) exposed 7‐week‐old male Harlan Sprague–Dawley rats to BTBPE given by daily oral gavage at doses of 0, 0.07, 0.69, 6.88, 68, 680 mg/kg bw per day for 5 days. No treatment‐related liver histopathologic alterations were reported, and there were no changes in expression of genes in the liver.

3.3.2.1.4. Computational analysis

The QSAR Toolbox identified experimental data on BTBPE for the Ames test (see above and Annex B).

VEGA models predicted negative Ames test results for BTBPE with ‘GOOD reliability’ (three models) and with ‘MODERATE reliability’ (one model). The Panel critically reviewed the predictions and considered them valid mainly due to the presence of relevant analogues known by the model with experimental and predicted values concordant with the prediction for the target compound (see Annex B). The CONTAM Panel noted that this negative prediction is consistent with the available experimental data.

3.3.2.1.5. Observations in humans

No studies were identified in the previous Opinion. Since then, three new studies were identified all pertaining to populations in China. The endpoints evaluated were thyroid hormone profile (three studies), liver injury biomarkers (one study) and thyroid cancer (one study). In Appendix F an overview of the epidemiological studies identified reporting on the BFRs considered in this Opinion, including those on BTBPE, is shown.

In the study by Guo et al. (2018) on the association with thyroid hormone levels in 10‐year‐old students living near a petrochemical complex in South China (for details see description in Section 3.2.4.2), the authors did not find an association between BTBPE levels and serum thyroid hormone levels in the samples from the 174 school students.

In the study by Zhao et al. (2021) that assessed the cross‐sectional relationship between the serum levels of five BFRs and a series of thyroid/liver injury biomarkers in residents of a major BFR‐producing region in China (for details see description in Section 3.3.1.1), a 10‐fold increment in the serum level of BTBPE was associated with higher bilirubin levels: total bilirubin level (0.869 μmol/L) [95% CI: 0.175, 1.564], direct bilirubin (DBIL) level (0.231 μmol/L) [95% CI: 0.075, 0.388] and indirect bilirubin level (0.638 μmol/L) [95% CI: 0.091, 1.185]. There were no statistically significant associations with thyroid hormone levels.

In the study by Liu, Li, et al. (2022) that assessed the association between the levels of several BFRs in serum and thyroid cancer risk in Shandong Province (China) (for details see description in Section 3.2.4.2), BTBPE was not associated with thyroid cancer risk and was not associated with thyroid hormone levels in the control participants.

In summary, the available epidemiological evidence on BTBPE is limited to three cross‐sectional studies in populations in China, two of which were studies in adults and one was in children. None of these three studies observed associations between BTBPE and thyroid hormone levels. One study observed an increase in total, direct and indirect bilirubin levels. The one study that studied thyroid cancer risk, reported no association.

3.3.2.1.6. Studies in other model systems and mode of action

Within the literature retrieved on BTBPE, several modes of action have been proposed to explain BTBPE toxicity. An overview of the studies identified since the previous Opinion is given in Appendix G (Table G.3) and a brief summary is provided below.

Mitochondrial dysfunction may be a predominant mechanism, with evidence of impaired mitochondrial functions in zebrafish testes, including ultrastructural damage, increased apoptosis‐inducing factor expression and disordered respiratory chain function (Zheng et al., 2024). Studies in human endothelial cells (HUVECs) have shown reduced mitochondrial membrane potential and decreased ATP production at low micromolar concentrations (Shi, Wang, et al., 2021). Mitochondrial dysfunction might be associated with oxidative stress which has been reported as another pathway, with studies demonstrating increased ROS production and MDA content alongside decreased antioxidant defences such as reduced GSH content and SOD activity (Shi, Wang, et al., 2021). Both apoptotic and necroptotic cell death pathways are activated by BTBPE, as evidenced by increased expression of apoptotic markers including CASP3, CYTC, BAX and BCL2 in endothelial cells (Shi, Wang, et al., 2021), and necroptosis via RIP1, RIP3 and MLKL activation in intestinal tissue (Cui, Xiao, Wang, et al., 2024). In rat liver cells, BTBPE necrosis was reported to be mediated by GAS5 and NUAK1, through binding to miR‐743a‐5p (Cui, Xiao, Zhang, et al., 2024). Liver fibrosis following BTBPE exposure was found to involve activation of hepatic stellate cells by macrophage extracellular traps (Cui et al., 2026). Hepatic stellate cells that came in contact with macrophage extracellular showed PPAR‐mediated excessive collagen and αSMA production, leading to hepatic fibrosis. Endocrine disruption mechanisms include anti‐oestrogenic activity at concentrations of 12.5 μM and above (Ezechiáš et al., 2012), predicted AR antagonistic activity from QSAR models (Bajard et al., 2021), and in vivo disruption of sex hormone regulation (Choi et al., 2021). Inflammatory signalling through the IRX3/NOS2 pathway has been identified as promoting inflammation and necroptosis in intestinal tissue (Cui, Xiao, Zhang, et al., 2024), whilst widespread effects on cellular protein phosphorylation levels across multiple compartments have also been documented in zebrafish (Zheng et al., 2024).

Notably, some studies have reported negative findings that help define the toxicological profile of BTBPE. No effects on hepatic gene expression were observed in rats at doses up to 680 mg/kg body weight per day for 5 days (Shockley et al., 2020a, 2020b), and no inhibition of thyroid hormone metabolism through DIO and SULT activity was detected in human liver preparations, possibly attributable to the absence of hydroxyl groups in the BTBPE chemical structure (Smythe et al., 2017).

Overall, the data indicate that reproductive toxicity mediated through mitochondrial dysfunction and endocrine disruption are of concern for BTBPE, with effects occurring at relatively low concentrations in zebrafish. Male reproductive toxicity has been shown in zebrafish at concentrations ranging from 0.01 to 10 μg/L in water, with observed impairments in mating behaviour, reduced sperm quality and decreased reproductive success (Zheng et al., 2024). These reproductive effects are accompanied by disrupted sex hormone regulation, including elevated E2 levels, increased E2/11‐KT ratio and altered oestrogen receptor expression (Choi et al., 2021). Vascular endothelial toxicity has also been observed, with cytotoxic effects appearing at 5 μM and above in human umbilical vein endothelial cells, and inhibition of cell migration and tube formation at 10 μM (Shi, Wang, et al., 2021). Additionally, intestinal damage characterised by compromised gut barrier integrity, loss of tight junction proteins, inflammatory gene activation and necroptosis has been documented in Sprague–Dawley (SD) rats following oral exposure to BTBPE or in vitro in rat intestinal ICE6 cells (Cui, Xiao, Wang, et al., 2024, see above).

3.3.2.1.7. Consideration of critical effects and dose–response analysis

The CONTAM Panel noted that the database for BTBPE is limited and characterised by significant gaps. A small number of experimental animal studies identified intestinal effects, and thyroid hormone effects without effects on the thyroid gland. No sub‐chronic or long‐term studies, or studies addressing possible developmental and reproductive toxicity or developmental neurotoxicity, were identified.

BTBPE was negative in Ames test, but no data on clastogenicity or aneugenicity were identified. Therefore, the CONTAM Panel could not conclude on the genotoxic potential of BTBPE.

The available evidence includes limited observations in humans (three cross‐sectional studies in Chinese populations).

Due to major limitations in the database, the CONTAM Panel could not identify a Reference Point.

3.3.2.2. Occurrence data and dietary exposure assessment

Occurrence data of BTBPE in food were submitted to EFSA (see Section 3.1.1) and the CONTAM Panel considered these data sufficient to perform an exposure estimation (see below). In the scientific literature, several studies were identified reporting on the occurrence of BTBPE and an overview is given in Section 3.1.2.

Regarding the data submitted to EFSA, Table 22 shows the LB and UB summary statistics of the estimated chronic dietary exposure to BTBPE, across surveys for each age group.

TABLE 22.

Mean and P95 LB and UB summary statistics of the estimated chronic dietary exposure to BTBPE, across surveys for each age group.

Age group	Mean dietary exposure (ng/kg bw per day)					P95 dietary exposure (ng/kg bw per day) a
	N	LB		UB		N	LB		UB
		Min	Max	Min	Max		Min	Max	Min	Max
Infants b	5	0.002	0.004	0.010	0.050	4	0.006	0.011	0.035	0.179
Toddlers b	8	0.005	0.006	0.029	0.058	5	0.008	0.011	0.053	0.253
Other children	17	0.003	0.006	0.018	0.061	17	0.006	0.014	0.046	0.259
Adolescents	16	0.002	0.003	0.012	0.040	16	0.003	0.006	0.017	0.209
Adults	16	0.001	0.002	0.008	0.074	16	0.003	0.004	0.016	0.303
Elderly	13	0.001	0.002	0.006	0.104	13	0.002	0.003	0.012	0.328
Very elderly	11	0.001	0.002	0.006	0.104	8	0.002	0.003	0.011	0.308

Abbreviations: bw, body weight; LB, lower bound; UB, upper bound.

^a The 95th percentile estimates obtained on dietary surveys with fewer than 59 observations may not be statistically robust (Meeker et al., 2017) and are therefore not included in this table.

^b Exposure estimates do not include potential exposure to BTBPE from human milk and infant and follow‐on formula due to lack of occurrence data. This might have caused an underestimation of the estimates for ‘Infants’ and ‘Toddlers’.

The mean LB dietary exposure ranged from 0.001 ng/kg bw per day in ‘Adults’, ‘Elderly’ and ‘Very Elderly’ to 0.006 ng/kg bw per day in ‘Infants’ and ‘Toddlers’, across surveys. The mean UB dietary exposure ranged from 0.006 ng/kg bw per day in ‘Elderly’ and ‘Very Elderly’ to 0.104 ng/kg bw per day in ‘Elderly’ and ‘Very Elderly’, across surveys.

The P95 LB dietary exposure ranged from 0.002 ng/kg bw per day in the ‘Elderly’ and ‘Very Elderly’ to 0.014 ng/kg bw per day in ‘Other children', across surveys. The P95 UB dietary exposure ranged from 0.011 ng/kg bw per day in ‘Very Elderly’ to 0.328 ng/kg bw per day in ‘Elderly’, across surveys.

Dietary exposure estimates to BTBPE in the surveys of pregnant women and lactating women (one survey each), were within the range of exposure estimates in the adult population and will not be further discussed.

Detailed summary statistics on the exposure estimates to BTBPE, for each dietary survey are presented in Annex C (Table 3).

Table 23 shows the number of surveys in which food categories at the Level 1 of the FoodEx2 classification have contributed more than 10% to the total mean dietary exposure to BTBPE for each age group. The range of percentage contribution to the mean exposure across surveys is shown in parenthesis. The raw primary commodities that contributed more than 10% to the total mean LB dietary exposure in the highest number of surveys belong to the ‘Milk and dairy products’ (47%–88%) and ‘Meat and meat products’ (9%–40%) FoodEx2 Level 1 food categories. This includes their contribution through the use of the raw primary commodity in derivatives and composite foods such as composite dishes (meat, egg, fish‐based meals, soups), fine bakery wares, desserts, cheese and butter.

TABLE 23.

Number of surveys (and in parenthesis the range of percentage contribution to the mean exposure across surveys) in which raw primary commodities under the indicated food categories at the Level 1 of the FoodEx2 classification have contributed more than 10% to the total mean LB dietary exposure to BTBPE for each age group (in parenthesis in the header the total number of surveys).

Food	Infants (5)	Toddlers (8)	Other children (17)	Adolescents (16)	Adults (16)	Elderly (13)	Very elderly (11)	N surveys
Milk and dairy products	5 (80.7–87.8)	8 (60.4–82.8)	17 (58.7–80.3)	16 (50.9–74.8)	16 (46.4–71.9)	13 (47.3–71.3)	11 (52.6–72.5)	86
Meat and meat products	4 (9.3–18.3)	8 (11.5–22.3)	17 (12.9–25.3)	16 (18.2–30.6)	16 (20.9–40.3)	13 (18.7–37.3)	11 (17.6–34.3)	85
Fish and seafood	(0–0)	(0–0)	1 (0.5–10.4)	2 (1–15.6)	3 (1.2–15.2)	1 (0.8–13.4)	(0–0)	7
Eggs and egg products	(0–0)	1 (1.3–10.4)	1 (3.1–10.3)	1 (2.8–10.8)	1 (4.3–11)	1 (3.6–12.1)	1 (3.9–10)	6

The contribution to the total mean LB dietary exposure to BTBPE of the raw primary commodities under each food category at Level 1 and Level 3 of the FoodEx2 classification in each survey is available in Annex C (Table C.4 and Table C.5, respectively).

In the literature, Ma et al. (2023) found in samples from the UK collected in 2020–2021, BTBPE at mean concentrations of 190 ng/kg in meat, 200 ng/kg in fish and prawns, 97 ng/kg in cheese and 2300 ng/kg in chicken eggs. Exposure estimates for BTBPE in the UK population were calculated based on daily consumption data for UK citizens from the national diet and nutrition survey (2008–2019). Average dietary intakes of BTBPE ranged from 1.2 to 4.4 ng/kg bw per day, while high‐end estimates, assuming food contaminated at the average concentrations consumed at the mean rate +2 standard deviations, ranged from 7.1 to 27 ng/kg bw per day, across population groups (from 0 up to > 75 years old).

For breastfed infants, there were insufficient data in human milk from European countries reported in the literature to make robust exposure estimates. To provide an indication of the potential exposure, the CONTAM Panel made an estimation for average and high consumption of human milk (see Section 2.4) considering the available data reported in the literature (Čechová et al., 2017; Tao et al., 2017, see Section 3.3.2.1.2).

The exposure scenario based on average human milk consumption and the highest reported mean concentration in human milk across three European countries (Norway, the Netherlands and Slovakia) of 1.103 ng/g lipid (Čechová et al., 2017), resulted in an estimate of 5.06 ng/kg bw per day. The study by Tao et al. (2017) did not report mean concentrations.

The exposure scenario based on high human milk consumption and the highest reported P95 concentration of 2.38 ng/g lipid (Čechová et al., 2017), resulted in an estimate of 16.4 ng/kg bw per day. The study by Tao et al. (2017) reported only maximum concentrations in samples from the UK, that were not used for these estimations.

It should be noted that these estimates are based on human milk samples from three European countries that were collected one or two decades ago. It is unknown whether this is representative of current exposure to this BFR.

3.3.2.3. Risk characterisation

In the absence of a Reference Point for BTBPE, it is not possible to perform an MOE approach to risk characterisation. However, the CONTAM Panel concluded that, because occurrence data for BTBPE in food were submitted to EFSA and exposure could be estimated, an indication should be given as to whether dietary exposure indicates a health concern.

The EFSA guidance on the TTC specifies that, for substances raising a concern about potential DNA‐reactive mutagenicity and/or carcinogenicity, if exposure is below 2.5 ng/kg bw per day, the substance is considered of low probability of adverse health effects (EFSA Scientific Committee, 2019). The next lowest TTC value is 300 ng/kg bw per day. Although the CONTAM Panel could not conclude on the genotoxic potential of BTBPE, the Panel noted that the highest dietary exposure estimates obtained from the occurrence data in food submitted to EFSA are below all of the TTC values, indicating a low probability of adverse health effects. However, this is based on a dietary exposure assessment with some significant sources of uncertainty (see Section 3.3.2.4).

3.3.2.4. Uncertainty analysis

The assessment of genotoxicity was inconclusive, and it was not possible to establish a Reference Point or conclude on risk characterisation for BTBPE. Identified sources of uncertainty affecting the exposure assessment are documented in Table 24.

TABLE 24.

Uncertainties identified and their impact on the outcome of the assessment of BTBPE.

Description of the uncertainty		Impact of the uncertainty on the assessment a
Occurrence data
Analytical measurements	Performance (e.g. specificity for the target compounds) of the analytical method (GC–ECD, GC–MS, etc)	1 – Low impact. Most of the data were analysed by GC–MS using isotope labelled standards. MS methods are considered to be reliable. Also, where the information about the analytical method was missing it was assumed samples were analysed with reliable methods.
	Proportion of left‐censored data and magnitude of difference between risk estimates for LB and UB exposures	2 – Medium impact. This uncertainty is handled through the substitution methods and with the estimation of LB and UB exposure estimates. Due to the inclusion in the occurrence dataset, of left‐censored analytical results with relatively high LOQs, maximum LB estimates are 88% to 99% lower than the maximum UB estimates, contributing to the uncertainty of the estimations.
	Consideration of recovery (e.g. correction carried out or not)	0 – Negligible impact. No reported results that needed to be corrected.
Data reporting	Potential errors in reporting the occurrence data (e.g. in the classification of the food category, unit of measurement, parameter, fat vs. whole weight, etc.) – unidentified errors (not apparent from the data provided)	1 – Low impact. Data cleaning procedures that identify outliers and clarifications received by data providers aim to solve most of the possible reporting errors. Although some errors might remain unidentified, their overall impact on the results should be low.
	Missing information in reporting the occurrence data (e.g. analytical method)	1 – Low impact. All information considered relevant was available or if missing was considered to have a low impact
	Missing or unclear information about the treatment/processing applied prior to the analysis of the sample that is submitted to EFSA	0 – Negligible impact. No information available on the impact of food processing on the concentration of BTBPE.
	Use of food categories at high (often not enough specified) FoodEx/FoodEx2 level	1 – Low impact. 92% of the analytical results were reported at Foodex2 Level 3 or more detailed.
	Uncertainty in the reporting of fat content by data providers and need for conversion to whole weight	1 – Low impact. 46% of results were reported on a fat weight basis and were converted to whole weight using the provided fat percentages. This conversion was considered to have a low impact on the exposure estimates.
Representativeness of the data	Limited number of analytical results per food categories expected to contribute to the exposure	1 – Low impact. Enough analytical results were available for each of the main raw primary commodities; however, no data were available for foods of plant origin and foods for the young population.
	Extrapolation of data from one food category to others, and other assumptions	1 – Low impact. Occurrence data have been extended to all similar foods when there was no reason to assume BTBPE was not present in the concerned similar foods. This allows to include in the exposure assessment also foods for which no specific occurrence data are available. This could lead to an under or over estimation of the true exposure but the impact is considered to be low on the risk assessment.
	Sampling strategy not fully random	0 – Negligible impact. All results were reported as being sampled based on randomised design (no ‘Suspect sampling’ reported).
	Uneven distribution of the data per year (e.g. recent years not sufficiently represented)	2 – Medium impact. 89% of analytical results were available for samples sampled between year 2014 and 2016. Ma et al. (2023) reported an increase of BTBPE concentration in food between 2015 and 2020 in the UK. The use of data sampled in less recent years might have caused an underestimation of exposure to BTBPE.
	Uneven distribution of the data per country (e.g. large number of MSs not sufficiently represented)	1 – Low impact. Data on BTBPE were submitted by 2 Member States plus Norway. Considering a common market, these data were considered sufficiently representative for exposure assessment.
	Limited number of analytical results per variables that could explain higher/lower levels, such as production method (e.g. wild vs. farmed), processing (e.g. peeled vs. raw), etc.	0 – Negligible impact. The additional information provided is limited but no specific aspects have been identified to have a major impact.
Consumption data
Data reporting	Unidentified errors in reporting consumption data, e.g. in the classification of the food, portion size, body weight estimation, memory errors, capacity to report details in dietary surveys. Different dietary survey methodologies (e.g. dietary record vs. 24‐h re‐call), dietary software, interview options, use of portion‐size measurement. Use of national standard recipes and ingredients factors for composite dishes (potentially leading to, e.g. underestimation of minor ingredients, overestimation of standard ingredients). Different sample sizes and response rates of the dietary surveys. Long‐term (chronic) exposure assessed based on few days of consumption per individual. Information about processing/cooking method not consistently reported in consumption records.	1 – Low impact. Uncertainties and limitations related to the use of the EFSA Comprehensive Food Consumption Database have been described by EFSA (EFSA, 2011a). These uncertainties are common to dietary exposure assessments performed using the Comprehensive Database, and have the potential to cause either an over‐ or underestimation of the exposure. Additional uncertainties introduced by the RPC model (e.g. recipes for disaggregation of composite foods and reverse yield factors for derivatives) shall also be considered as standard uncertainties. No specific additional uncertainties affecting the food consumption data used for the BTBPE exposure assessment were identified.
Representativeness of the data	Availability of food consumption data for special population groups, including consumers only of specific foods of special interest, or following special diets.	1 – Low impact. Consumers of mostly contaminated foods are covered by 95th percentile exposure estimates. Pregnant and lactating women surveys provided similar exposure results as adults from the general population. No other specific population groups that might have a particular exposure to BTBPE were identified.
Exposure estimates
Non‐dietary exposure	Sources of exposure other than dietary – how important is dietary exposure to the total	1 – Low impact. Exposure from dust can vary greatly between individuals and age classes, although not as great as exposure from diet. It can be particularly relevant for young children. Limited data on dermal and inhalation and other potential non‐dietary routes of exposure.
Exposure assessment scenario	Consumers loyalty to specific brands or from specific local areas not considered.	0 – Negligible impact. Exposure assessment was made for the general population. Brand loyalty and consumer's only scenarios were considered not relevant for the BTBPE exposure assessment. High exposures due to variability across individuals are covered by the 95th percentile of the exposure estimates.
Genotoxicity
Genotoxicity	Uncertainty in the assessment of genotoxicity	3 – High impact. Limited data identified. Negative data on induction of reverse mutations in bacteria, and no data on clastogenicity or aneugenicity.

Abbreviations: ADME, absorption, distribution, metabolism, excretion; BMD, benchmark dose; BMDL, benchmark dose lower credible limit; BMR, benchmark response; MOA, mode of action; TK, toxicokinetic.

^a 0 – Uncertainty with negligible impact; 1 – Uncertainty with low impact; 2 – Uncertainty with medium impact; 3 – Uncertainty with high impact. This ordinal scale provides a relative assessment of the impact of the individual uncertainties as an aid to quantifying their combined impact by expert judgement, as described in section 11 of EFSA Scientific Committee (2018).

3.3.3. Hexabromobenzene (HBB) [CAS number 87‐82‐1]

HBB is a non‐polymeric additive BFR. The physicochemical properties of HBB are reported in Table 1. HBB is currently not registered under REACH (see Table 2), and thus its production volume and use within the EU should be lower than one tonne per year. However, it has been identified in the environment at concentrations comparable to DBDPE (ECHA, 2024).

3.3.3.1. Hazard identification and characterisation

3.3.3.1.1. Toxicokinetics

In the previous Opinion two studies were cited. Yamaguchi et al. (1986) investigated the distribution and elimination of HBB in male Wistar rats after a single oral dose of 0.2 mg HBB per animal. The authors measured HBB in brain, liver, kidney, adipose tissue and blood until 40 days after exposure. The maximum concentrations were observed in adipose tissue (after 1 day). Compared to the other tissues, the concentration in liver decreased rapidly, indicating a rapid metabolism of HBB. The authors found a biphasic elimination, with half‐lives of 0.7 (initial phase or alpha) and 48 days (terminal phase or beta) for HBB. Koss et al. (1982) administered HBB at 16.6 mg/kg bw every other day by gavage in female Wistar rats during 2 weeks. The authors identified the formation of pentabromobenzene in faeces, probably due to a reductive debromination of HBB.

For the current update, the CONTAM Panel identified three additional studies; a second study by Yamaguchi et al. (1988), Sapota et al. (1997) ³³ and Shin et al. (2023).

Yamaguchi et al. (1988) administered a single oral dose of 0.2 mg HBB to male Wistar rats. Rats were sacrificed at 1, 2, 4 and 8 h, and 1, 3, 10, 20 and 40 days after exposure, respectively. The authors measured the highest HBB concentration in adipose tissues, liver, brain and kidneys. HBB was rapidly metabolised into debrominated metabolites, pentabromobenzene (PeBB), tetrabromobenzene (TeBB) and tribromobenzene (TrBB). Among the three measured metabolites, TeBB accumulated in adipose tissues, liver, brain and kidneys. The half‐life of total HBB and its metabolite was 16 days.

The authors reported that 10% of the administered dose was excreted as HBB and its debrominated metabolites in faeces. After 10 days, the remaining HBB and its debrominated metabolites were 20%. The authors suggested that 70% of the HBB could be metabolised to other compounds such as hydroxylated compounds (Yamaguchi et al., 1988).

Sapota et al. (1997) administered a single oral dose of 600 mg/kg and 4500 mg/kg (¹⁴C‐HBB) in female Wistar rats. The highest measured radioactivity was: fat tissue > adrenals > sciatic nerve > liver > lung > brain > kidney > muscle. The maximum radioactivity concentration was observed between 2 and 8 h in the different tissues, except for fat tissues where the maximum concentration was observed at 48 h after the administration.

Within 72 h after administration, 1.2% and 75% of the administered dose were excreted in urine and faeces, respectively. The authors reported a terminal half‐life of 440 h (18 days). The percentage of radioactivity remaining in rats 48 h after exposure was between 24 and 30%.

Shin et al. (2023) performed a metabolomics study using hepatic cells treated by various concentrations of HBB during 24 or 48 h and with a period of observation of 21 days. The authors found 35 metabolites with a decreasing concentration within 15–17 days.

In summary, the available data showed that HBB and its metabolites accumulated in lipid‐rich tissues.

3.3.3.1.2. Levels in humans

The previous Opinion did not identify any study on levels of HBB in human samples from European countries, but reported on studies in samples from Japan and China (EFSA CONTAM Panel, 2012b).

Since then, studies have been identified reporting on the levels of HBB in human samples from European countries, including human milk and blood/serum. These studies are briefly described below and details can be found in Appendix E (Table E.1).

Regarding human milk samples,

• HBB was reported to be below the LOD (0.2–0.3 ng/g lipid) in 11 pooled human milk samples collected in 2010 from 109 first‐time mothers in Ireland (Pratt et al., 2013).

• Similarly, HBB was not detected (< 0.9 pg/g ww) in either of two pooled human milk samples collected in 2009–2010 in Sweden (Sahlström et al., 2015) nor in human milk samples collected in the UK in 2010 (n = 25) and 2014–2015 (n = 10) (LOD not reported) (Tao et al., 2017).

• Čechová et al. (2017) analysed HBB in human milk samples from Norway (2003–2006, n = 116), the Netherlands (2011–2014, n = 116) and Slovakia (2011–2012, n = 37) with a detection frequency between 44 and 88%. Mean and P95 concentrations across countries ranged from 0.036 to 0.048 ng/g lipid and from 0.099 to 0.170 ng/g lipid, respectively.

Regarding blood/serum,

• Cequier et al. (2013) detected HBB in 3 out of the 10 serum samples from Norway analysed at concentrations of 0.04, 0.07 and 0.08 ng/g lipid (LOD = 0.30 pg/mL). In a later study, Cequier et al. (2015) also analysed HBB in serum samples from 46 women from a Norwegian mother–child cohort established in 2012 with median (mean) levels of 0.03 (0.07) ng/g lipid (detection frequency 48%).

• HBB was also measured in 61 serum samples from a Norwegian cohort collected in 2013 with a detection frequency of 5% and a reported geometric mean of 5.9 ng/g lipid (Tay et al., 2019).

• In Sweden, Sahlström et al. (2014) did not detect HBB above the LOD (1 pg/25 μL) in any of the 24 matched serum samples from mothers and their toddlers collected between 2009 and 2010.

• Sales et al. (2017) did not detect HBB above the LOD of 5 pg/g serum in one pool of human cord blood from Belgium.

3.3.3.1.3. Toxicity in experimental animals

Studies considered in the previous EFSA assessment

Several toxicological studies on HBB were described in the previous Opinion (EFSA CONTAM Panel, 2012b). Only those with oral administration and in vitro studies are considered here. A number of studies demonstrated effects on liver microsomal enzymes at various doses from 15 to 3024 mg/kg bw per day. Excretion of porphyrins in urine was reported following ‘chronic’ intragastric exposure to HBB at 15, 75 or 375 mg/kg bw per day, providing further evidence of hepatotoxicity. HBB was also reported to increase liver weight of pups of dams fed with 80 mg/kg feed from before mating until PND21 but was not teratogenic. It did not activate the AHR in a CALUX assay or the human nuclear androgen receptor in HepG2 cells at the concentrations tested.

Regarding genotoxicity, the previous Opinion reported that HBB was negative in a Salmonella mutation assay with and without metabolic activation (EFSA CONTAM Panel, 2012b).

Studies published since the previous EFSA assessment

No new studies were identified.

3.3.3.1.4. Computational analysis

The QSAR Toolbox identified experimental data for the Ames test on HBB (see above and Annex B).

VEGA models predicted negative Ames test results for HBB with ‘GOOD reliability’ (three models) and with ‘MODERATE reliability’ (one model). The Panel critically reviewed the predictions and considered them valid mainly due to the presence of relevant analogues known by the model with experimental and predicted values concordant with the prediction for the target compound (see Annex B). The CONTAM Panel noted that this negative prediction is consistent with the available experimental data.

3.3.3.1.5. Observations in humans

No studies were identified in the previous Opinion. Since then, five publications were identified, four pertaining to populations in China and one in the USA. Exposure was assessed in serum in all studies in China, and in indoor air and wristbands in the USA study. The endpoints assessed were thyroid hormone profile (three studies), thyroid cancer (one study), type 2 diabetes (one study), BMI (one study) and liver injury markers (one study). In Appendix F an overview of the epidemiological studies identified reporting on the BFRs considered in this Opinion, including those on HBB, is shown.

In the study by Zhao et al. (2021) that assessed the cross‐sectional relationship between the serum levels of five BFRs and a series of thyroid/liver injury biomarkers in residents of a major BFR‐producing region in China (for details se description in Section 3.3.1.1), no associations were observed between HBB serum levels and any of the thyroid or liver biomarkers.

The study by Liu, Li, et al. (2022) that assessed the association between the levels of several BFRs in serum and thyroid cancer risk in Shandong Province (China) (for details see description in Section 3.2.4.2), did not observe an association between HBB serum levels and thyroid cancer risk in either males or females. Further, HBB was not associated with thyroid hormone levels in the control participants.

In the USA study by Wang, Romanak, et al. (2019) (for details see Section 3.2.4.2), HBB concentrations, measured in wristbands of 101 adult study participants, were not statistically significantly associated with thyroid hormone profiles.

In the study by Zheng, Chen, et al. (2025) in China (for details see Section 3.2.4.2), serum HBB concentrations were higher in type 2 diabetes cases (n = 60) than in healthy controls (n = 57) in unadjusted comparisons. Mixture models, including seven BFRs and two dechlorane plus, consistently showed a significant positive association between mixed dechlorane plus/BFRs exposure and diabetes risk, and HBB was identified as a major contributor to the mixture effect.

In the study by Zheng, Sun, et al. (2025) in China (for details see Section 3.2.4.2), serum HBB concentrations did not show statistically significant associations with BMI. Adjusted models are not presented.

In summary, the five epidemiological studies that assessed health outcomes related to HBB are cross‐sectional studies, four in Chinese populations and one in a USA population. In none of these studies statistically significant associations were observed between HBB serum levels and thyroid hormone profiles, liver injury biomarkers or thyroid cancer risk. In a small case–control study, HBB appeared to be a major contributor to the association between mixture exposure and type 2 diabetes risk.

3.3.3.1.6. Studies in other model systems and mode of action

Few studies are available, and an overview of the studies identified since the previous Opinion is given in Appendix G (Table G.4). There is some evidence of effects on vitamin D metabolism (Sousa et al., 2022) and oxidative stress (Usenko et al., 2016). Bajard et al. (2019, 2021) have reported agonistic activity towards the progesterone and androgen receptors, but the evidence supporting this is unclear.

3.3.3.1.7. Consideration of critical effects and dose–response analysis

Very few toxicological data are available for HBB. There is some evidence for hepatotoxicity. Three epidemiological studies did not show statistically significant associations between HBB serum levels and thyroid hormone profiles, liver injury biomarkers or thyroid cancer risk.

HBB did not induce mutations in S. Typhimurium with and without metabolic activation. However, there is no information on clastogenicity or aneugenicity therefore an overall conclusion on genotoxicity cannot be reached.

Overall, the data were not sufficient to identify a Reference Point.

3.3.3.2. Occurrence data and dietary exposure assessment

Occurrence data of HBB in food were submitted to EFSA (see Section 3.1.1) and the CONTAM Panel considered these data sufficient to perform an exposure estimation (see below). In the scientific literature, several studies were identified reporting on the occurrence of HBB and an overview is given in Section 3.1.2.

Table 25 shows the LB and UB summary statistics of the estimated chronic dietary exposure to HBB, across surveys for each age group.

TABLE 25.

Mean and P95 LB and UB summary statistics of the estimated chronic dietary exposure to HBB, across surveys for each age group.

Age group	Mean dietary exposure (ng/kg bw per day)					P95 dietary exposure (ng/kg bw per day) a
	N	LB		UB		N	LB		UB
		Min	Max	Min	Max		Min	Max	Min	Max
Infants b	5	0.07	0.17	0.08	0.19	4	0.25	0.44	0.29	0.49
Toddlers b	8	0.19	0.25	0.23	0.30	5	0.34	0.46	0.39	0.53
Other children	17	0.11	0.21	0.14	0.25	17	0.22	0.61	0.26	0.67
Adolescents	16	0.05	0.11	0.07	0.14	16	0.10	0.23	0.13	0.26
Adults	16	0.04	0.06	0.06	0.08	16	0.08	0.12	0.11	0.15
Elderly	13	0.04	0.05	0.05	0.07	13	0.06	0.11	0.09	0.14
Very elderly	11	0.04	0.06	0.05	0.08	8	0.08	0.12	0.10	0.15

Abbreviations: bw, body weight; LB, lower bound; UB, upper bound.

^a The 95th percentile estimates obtained on dietary surveys with fewer than 59 observations may not be statistically robust (Meeker et al., 2017) and are therefore not included in this table.

^b Exposure estimates do not include potential exposure to HBB from human milk and infant and follow‐on formula due to lack of occurrence data. This might have caused an underestimation of the estimates for infants and toddlers.

The mean LB dietary exposure ranged from 0.04 ng/kg bw per day in ‘Adults’, ‘Elderly’ and ‘Very Elderly’ to 0.25 ng/kg bw per day in ‘Toddlers’, across surveys. The mean UB dietary exposure ranged from 0.05 ng/kg bw per day in ‘Elderly’ and ‘Very Elderly’ to 0.3 ng/kg bw per day in ‘Toddlers’, across surveys.

The P95 LB dietary exposure ranged from 0.06 ng/kg bw per day in ‘Elderly’ to 0.61ng/kg bw per day in ‘Other children', across surveys. The P95 UB dietary exposure ranged from 0.09 ng/kg bw per day in ‘Elderly’ to 0.67 ng/kg bw per day in ‘Other children', across surveys.

Dietary exposure estimates to HBB in the surveys of pregnant women and lactating women (one survey each), were within the range of exposure estimates in the adult population and will not be further discussed.

The food category that mostly contributes to the dietary exposure was ‘Milk’ across all age groups with contribution percentages between 87% and 98% including its contribution from its use in derivatives and composite foods such as cheese, butter, soups and fine bakery wares.

Detailed summary statistics on the exposure estimates to HBB, for each dietary survey are presented in Annex C (Table C.3). The contribution to the total mean LB dietary exposure to HBB of raw primary commodities under each food category at Level 1 and Level 3 of the FoodEx2 classification in each survey is available in Annex C (Table C.4 and Table C.5, respectively).

For breastfed infants, there were insufficient data in human milk from European countries reported in the literature to make robust exposure estimates. To provide an indication of the potential exposure, the CONTAM Panel made an estimation for average and high consumption of human milk (see Section 2.4) considering the available data reported in the literature. Only one study reported levels of HBB above the LOD (Čechová et al., 2017, see Section 3.3.3.1.2).

The exposure scenario based on average human milk consumption and the highest reported mean concentration in human milk across three European countries (Norway, the Netherlands and Slovakia) of 0.048 ng/g lipid (Čechová et al., 2017), resulted in an estimate of 0.22 ng/kg bw per day.

The exposure scenario based on high human milk consumption and the highest reported P95 concentration of 0.17 ng/g lipid (Čechová et al., 2017), resulted in an estimate of 1.17 ng/kg bw per day.

It should be noted that these estimates are based on human milk samples from three European countries that were collected one or two decades ago. It is unknown whether this is representative of current exposure to this BFR.

3.3.3.3. Risk characterisation

In the absence of a Reference Point for HBB, it is not possible to perform an MOE approach to risk characterisation. Given the bioaccumulation potential of HBB, the CONTAM Panel did not consider it appropriate to apply a TTC approach (EFSA Scientific Committee, 2019).

3.3.3.4. Uncertainty analysis

The assessment of genotoxicity was inconclusive and it was not possible to establish a Reference Point or conclude on risk characterisation for HBB. Identified sources of uncertainty affecting the exposure assessment and genotoxicity are documented in Table 26.

TABLE 26.

Uncertainties identified and their impact on the outcome of the assessment of HBB.

Description of the uncertainty		Impact of the uncertainty on the assessment a
Occurrence data
Analytical measurements	Performance (e.g. specificity for the target compounds) of the analytical method (GC‐ECD, GC–MS, etc)	1 – Low impact. Most of the data were analysed by GC–MS using isotope labelled standards. MS methods are considered to be reliable. Also, where the information about the analytical method was missing it was assumed samples were analysed with reliable methods.
	Proportion of left‐censored data and magnitude of difference between risk estimates for LB and UB exposures	1 – Low impact. This uncertainty is handled through the substitution methods and with the estimation of LB and UB exposure estimates. LB estimates are up to 24% lower than the maximum UB estimates, contributing to the uncertainty of the estimations.
	Consideration of recovery (e.g. correction carried out or not)	0 – Negligible impact. No reported results that needed to be corrected.
Data reporting	Potential errors in reporting the occurrence data (e.g. in the classification of the food category, unit of measurement, parameter, fat vs. whole weight, etc.) – unidentified errors (not apparent from the data provided)	1 – Low impact. Data cleaning procedures that identify outliers and clarifications received by data providers aim to solve most of the possible reporting errors. Although some errors might remain unidentified, their overall impact on the results should be low. To be noted that an outlier occurrence value in milk was identified but the data providers confirmed its correctness. Removing the outlier would have lowered the highest LB mean exposure by 79%.
	Missing information in reporting the occurrence data (e.g. analytical method)	1 – Low impact. All information considered relevant was available or if missing was considered to have a low impact
	Missing or unclear information about the treatment/processing applied prior to the analysis of the sample that is submitted to EFSA	0 – Negligible impact. No information available on the impact of food processing on the concentration of HBB.
	Use of food categories at high (often not enough specified) FoodEx/FoodEx2 level	1 – Low impact. 96% of the analytical results were reported at FoodEx2 Level 3 or more detailed.
	Uncertainty in the reporting of fat content by data providers and need for conversion to whole weight	1 – Low impact. 15% of results were reported on a fat weight basis and were converted to whole weight using the provided fat percentages. This conversion was considered to have a low impact on the exposure estimates.
Representativeness of the data	Limited number of analytical results per food categories expected to contribute to the exposure	2 – Medium impact. Enough analytical results were available for each of the main raw primary commodities of animal origin. However, data for food of plant origin, honey (only one sample) and Food products for the young population were all left‐censored and analysed with a very high LOQ and excluded from the assessment. This might have caused an underestimation of exposure.
	Extrapolation of data from one food category to others, and other assumptions	1 – Low impact. Occurrence data have been extended to all similar foods when there was no reason to assume HBB was not present in the concerned similar foods. This allows to include in the exposure assessment also foods for which no specific occurrence data are available. This could lead to an under or over estimation of the true exposure, but the impact is considered to be low on the risk assessment.
	Sampling strategy not fully random	0 – Negligible impact. Analytical results reported as ‘Suspect sampling’ were excluded from the assessment.
	Uneven distribution of the data per year (e.g. recent years not sufficiently represented)	1 – Low impact. Most analytical results were available for samples sampled between the years 2014 and 2016 thus might not reflect the current occurrence of HBB in food.
	Uneven distribution of the data per country (e.g. large number of MSs not sufficiently represented)	1 – Low impact. Data on HBB were submitted by 2 Member States plus Norway. Considering a common market, these data were considered sufficiently representative for exposure assessment.
	Limited number of analytical results per variables that could explain higher/lower levels, such as production method (e.g. wild vs. farmed), processing (e.g. peeled vs. raw), etc.	0 – Negligible impact. The additional information provided is limited but no specific aspects have been identified to have a major impact.
Consumption data
Data reporting	Unidentified errors in reporting consumption data, e.g. in the classification of the food, portion size, body weight estimation, memory errors, capacity to report details in dietary surveys. Different dietary survey methodologies (e.g. dietary record vs. 24‐h re‐call), dietary software, interview options, use of portion‐size measurement. Use of national standard recipes and ingredients factors for composite dishes (potentially leading to, e.g. underestimation of minor ingredients, overestimation of standard ingredients). Different sample sizes and response rates of the dietary surveys. Long‐term (chronic) exposure assessed based on few days of consumption per individual. Information about processing/cooking method not consistently reported in consumption records.	1 – Low impact. Uncertainties and limitations related to the use of the EFSA Comprehensive Food Consumption Database have been described by EFSA (EFSA, 2011a). These uncertainties are common to dietary exposure assessments performed using the Comprehensive Database, and have the potential to cause either an over‐ or underestimation of the exposure. Additional uncertainties introduced by the RPC model (e.g. recipes for disaggregation of composite foods and reverse yield factors for derivatives) shall also be considered as standard uncertainties. No specific additional uncertainties affecting the food consumption data used for the HBB exposure assessment were identified.
Representativeness of the data	Availability of food consumption data for special population groups, including consumers only of specific foods of special interest or following special diets.	1 – Low impact. Consumers of mostly contaminated foods are covered by 95th percentile exposure estimates. Pregnant and lactating women surveys provided similar exposure results as adults from the general population. No other specific population groups that might have a particular exposure to HBB were identified.
Exposure estimates
Non‐dietary exposure	Sources of exposure other than dietary – how important is dietary exposure to the total	1 – Low impact. Exposure from dust can vary greatly between individuals and age classes, although not as great as exposure from diet. It can be particularly relevant for young children. Limited data on dermal and inhalation and other potential non‐dietary routes of exposure.
Exposure assessment scenario	Consumers loyalty to specific brands or from specific local areas not considered.	0 – Negligible impact. Exposure assessment was made for the general population. Brand loyalty and consumer's only scenarios were considered not relevant for the HBB exposure assessment. High exposures due to variability across individuals are covered by the 95th percentile of the exposure estimates.
Genotoxicity
Genotoxicity	Uncertainty in the assessment of genotoxicity	3 – High impact. Limited data identified. Negative data on induction of reverse mutations in bacteria, and no data on clastogenicity or aneugenicity.

Abbreviations: ADME, absorption, distribution, metabolism, excretion; BMD, benchmark dose; BMDL, benchmark dose lower credible limit; BMR, benchmark response; MOA, mode of action; TK, toxicokinetic.

^a 0 – Uncertainty with negligible impact; 1 – Uncertainty with low impact; 2 – Uncertainty with medium impact; 3 – Uncertainty with high impact. This ordinal scale provides a relative assessment of the impact of the individual uncertainties as an aid to quantifying their combined impact by expert judgement, as described in section 11 of EFSA Scientific Committee (2018).

3.3.4. 4‐(1,2‐Dibromoethyl)‐1,2‐dibromocyclohexane (DBE‐DBCH) [CAS number 3322‐93‐8]

DBE‐DBCH is used as an additive flame retardant mainly in polystyrene and polyurethane products and in electrical cable coating and some construction materials (EFSA CONTAM Panel, 2012b). There are four DBE‐DBCH stereoisomers (α‐, β‐, γ‐ and δ‐DBE‐DBCH) consisting of two enantiomeric pairs. Technical grade DBE‐DBCH consists of near equimolar amounts of α‐ and β‐DBE‐DBCH, whereas γ‐ and δ‐DBE‐DBCH appear in the technical mixture when heated at temperatures above 120°C.

The physicochemical properties of DBE‐DBCH are reported in Table 1. DBE‐DBCH is currently not registered under REACH, and thus its production volume and use within the EU should be lower than one tonne per year (see Table 2).

3.3.4.1. Hazard identification and characterisation

3.3.4.1.1. Toxicokinetics

No toxicokinetic studies were identified in the previous Opinion. Since then, three studies have been identified.

• Chu et al. (2012) investigated the metabolism of DBE‐DBCH (α‐ and β‐isomers) in male Wistar Han rat liver microsomes. The authors found that both α‐ and β‐DBE‐DBCH can be debrominated and metabolised into quantifiable metabolites (OH‐DBE‐DBCH and (OH)2‐DBE‐DBCH metabolites).

• Nguyen et al. (2017) studied the metabolism of DBE‐DBCH (β‐DBE‐DBCH isomer) in human Liver Microsomes. The authors showed that DBE‐DBCH was metabolised via cytochrome P450 enzyme into hydroxylated and debrominated metabolites (OH‐DBE‐DBCH, (OH)2‐DBE‐DBCH, dihydroxy tribromoethyl cyclohexane or (OH)2‐TriBECH, and bromo‐(1,2‐dibromocyclohexyl) acetic acid or DBCBA).

• In a computational toxicokinetic study, Ma et al. (2020) measured the affinities of β‐DBE‐DBCH in human CYPs 1A2, 2A6, 2B6, 2C9, 2C19, 2D6, 2E1 and 3A4 by using molecular docking, MD simulations and density functional theory. The authors found that β‐DBE‐DBCH binding affinities may follow the order of 2A6 > 2C9 > 2B6 > 2E1 > 3A4 ≈ 2C19 ≈ 1A2 > 2D6. Among all CYPs, 2A6, 2C19, 2E1 and 3A4 isoforms have been reported as the specific CYP isoforms responsible for the metabolism of β‐DBE‐DBCH.

3.3.4.1.2. Levels in humans

The previous Opinion did not report any study on levels of DBE‐DBCH in human samples. Since then, several studies have been identified reporting on the levels of DBE‐DBCH in human milk and blood/serum from European countries. These studies are briefly described below, and details can be found in Appendix E (Table E.1).

Regarding human milk samples,

• In Sweden, Sahlström et al. (2015) analysed 2 pooled human milk samples collected in 2009–2010: α‐DBE‐DBCH was detected in both pooled samples at a mean concentration of 4 pg/g ww, while β‐DBE‐DBCH was detected in one of the pooled samples at the LOD of 0.5 pg/g ww.

• In the study by Čechová et al. (2017), DBE‐DBCH isomers were analysed in human milk samples from Norway (2003–2006, n = 305), the Netherlands (2011–2014, n = 116) and Slovakia (2011–2012, n = 37) with a detection frequency between 25% and 32%. Mean and P95 concentrations across countries ranged from 0.017 to 0.038 ng/g lipid and from 0.035 to 0.083 ng/g lipid, respectively.

• Tao et al. (2017) analysed human milk samples collected in the UK in 2010 (n = 25) and 2014–2015 (n = 10) for several BFRs and noted that the most abundant compound was DBE‐DBCH (α‐ and β‐DBE‐DBCH). Regarding α‐DBE‐DBCH, in 2010 the levels of ranged from < 0.13 to 1.7 ng/g lipid (detection frequency 20%), while in 2014–2015 a detection frequency of 100% was reported and concentrations ranged from < 0.3 to 1.1 ng/g lipid, with a mean value of 0.67 ng/g lipid. Regarding β‐DBE‐DBCH, in 2010 the levels of ranged from < 0.13 to 38 ng/g lipid with a mean value of 6.8 ng/g lipid (detection frequency 76%), while in 2014–2015 a detection frequency of 100% was reported and concentrations ranged from 0.43 to 10 ng/g lipid, with a mean value of 2.5 ng/g lipid. The authors reported that this mean concentration was comparable to that of BDE‐47 or α‐HBCDD (2.8 or 2.1 ng/g lipid, respectively).

Regarding levels in blood/serum,

• In Sweden, Sahlström et al. (2014) analysed 24 matched serum samples form mothers and their toddlers collected between 2009 and 2010, and detected α‐DBE‐DBCH above the LOQ in two toddler serum samples (at 0.7 and 1.8 ng/g lipid), and β‐DBE‐DBCH in one of these toddler's samples at 1.3 ng/g lipid.

• DBE‐DBCH was also measured in 61 serum samples collected in 2013 from a Norwegian cohort, and it was reported below the LOD for α‐ and β‐DBE‐DBCH (LOD = 3.9 and 4.8 pg/serum sample, respectively, volume not reported, Tay et al., 2019).

3.3.4.1.3. Toxicity in experimental animals

Studies considered in the previous EFSA assessment

No data on toxicity in experimental mammalian species were identified in the previous Opinion (EFSA CONTAM Panel, 2012b).

Studies published since the previous EFSA assessment

Since the previous Opinion, one study in rodents has been published. Rats were fed technical grade DBE‐DBCH in the diet at concentrations of 0, 10, 50, 250, 1250 or 5000 mg/kg (corresponding to 0, 0.9, 4.2, 21, 98 and 329 mg/kg bw per day in males and 0, 0.8, 3.9, 19 and 321 mg/kg bw per day in females for 28 days (Curran et al., 2017). Top dose animals were euthanised due to rapid weight loss. There were dose‐dependent clinical and histopathological changes consistent with mild hepatic and renal inflammation and increased serum testosterone in males. The authors identified NOAELs of about 4 mg/kg bw per day and also performed BMD modelling.

Genotoxicity

Three studies were identified informing on the genotoxicity of DBE‐DBCH ³⁴ :

• DBE‐DBCH (supplied by the NTP chemical repository [Radian Corporation, Austin, TX]) was negative in S. Typhimurium TA98, TA100, TA1535, TA1537 and TA97 strains using a pre‐incubation assay, both in the absence and presence of metabolic activation provided by liver S‐9 from Aroclor‐induced male Sprague–Dawley rats and Syrian hamsters (Zeiger et al., 1987).

• DBE‐DBCH was tested for its ability to induce sister chromatid exchanges (SCEs) and chromosomal aberrations in CHO cells with and without a metabolic activation system (rat liver S9 fraction). It was negative for induction of chromosomal aberrations under both experimental conditions, while increased SCE frequency was reported at cytotoxic concentrations (independent of metabolic activation) (Loveday et al., 1989).

• DBE‐DBCH also tested positive for the induction of mutations in the L5178Y tk⁺/tk⁻ mouse lymphoma cell forward mutation assay in the absence of metabolic activation (McGregor et al., 1991). These study results were reported in NTP (2007). ³⁵

In summary, in vitro, DBE‐DBCH is not mutagenic in bacteria and did not induce chromosomal aberrations in CHO cells (with and without metabolic activation). Positive results were reported; however, for induction of SCEs in the same cell line (with S9 from rat liver) and of gene mutations in L5178Y mouse lymphoma cells (in the absence of metabolic activation). Based on the available data the CONTAM Panel concluded that DBE‐DBCH is mutagenic in mammalian cells in vitro. No in vivo data are available. The aneugenic potential was not investigated.

3.3.4.1.4. Computational analysis

The QSAR Toolbox identified experimental data for the Ames test on DBE‐DBCH (see above and Annex B).

VEGA models predicted positive Ames test results for DBE‐DBCH with ‘LOW reliability’ (three models) and negative Ames test results with ‘LOW reliability’ (one model). These predictions were not further considered due to their limited reliability.

3.3.4.1.5. Observations in humans

No studies were identified in the previous Opinion. Since then, one publication was identified, pertaining to a population in China. In Appendix F, an overview of the epidemiological studies identified reporting on the BFRs considered in this Opinion is shown, including the one on DBE‐DBCH.

In the study by Guo et al. (2018) on the association with thyroid hormone levels in 10‐year‐old students living near a pretrochemical complex in South China (for details see description in Section 3.2.4.2), DBE‐DBCH levels (no distinction between the diastereomers) were not associated with the serum thyroid hormone levels.

3.3.4.1.6. Studies in other model systems and mode of action

An overview of the studies identified since the previous Opinion is given in Appendix G (Table G.5) and a brief summary is provided below.

The 2012 Opinion reported two toxicity studies on juvenile brown trout exposed to β‐DBE‐DBCH in the diet at concentrations of 2.02, 14.7 and 118.4 pmol/g fat for 56 days. There were no effects on liver and gonad somatic indices or serum levels of steroid hormones (Gemmill et al., 2011). A subsequent study at the same doses reported reduced plasma tT4 at the top dose and a significant increase in mean thyroid epithelial cell height in all dose groups (Park et al., 2011). No new studies in fish were identified.

With respect to mode of action, previously reported studies identified activation of the androgen receptor (AR) in HepG2 cells with the γ‐ and δ‐ isomers more potent than α‐ and β‐DBE‐DBCH (Khalaf et al., 2009; Larsson et al., 2006). More recent studies with varying model systems have provided further evidence for this association (Bereketoglu et al., 2021; Kharlyngdoh et al., 2016; Kharlyngdoh et al., 2018; Wong et al., 2016).

Caspase‐3 and measures of oxidative stress were elevated only at cytotoxic concentrations in a macrophage cell line (Wang, Wei, et al., 2020) and therefore the relevance is unclear.

In zebrafish larvae, there are indications of developmental and locomotor behavioural toxicity, immunotoxicity, oxidative stress and proapoptotic effects (Wang, Wei, et al., 2019) and of delayed hatching and toxicity (Pradhan et al., 2013). Studies in cultured Purkinje neurons suggested potential neurotoxicity of β‐DBE‐DBCH (Stojak et al., 2019).

3.3.4.1.7. Consideration of critical effects and dose–response analysis

DBE‐DBCH has shown evidence of effects on liver and kidneys in a short term (28‐day) study in rats. Studies in model systems indicate potential for neurotoxicity, effects on development and the thyroid, and activation of the AR.

DBE‐DBCH is not mutagenic in bacteria, but it induced gene mutations in mammalian cells in vitro. It did not induce chromosomal aberrations in CHO cells. No in vivo data are available. The CONTAM Panel considered that DBE‐DBCH is mutagenic in vitro in mammalian cells.

Taking into account the in vitro mutagenicity and lack of long‐term studies, the CONTAM Panel was not able to identify a Reference Point for DBE‐DBCH.

The CONTAM Panel noted that there are four stereoisomers of DBE‐DBCH: α‐ and β‐DBE‐DBCH which are present in the technical product, and γ‐ and δ‐DBE‐DBCH, which are formed when the technical product is heated at temperatures above 120°C. MOA studies indicate that γ‐ and δ‐DBE‐DBCH are more potent than α‐ and β‐DBE‐DBCH in activation of the AR. In principle, it is possible that γ‐ and δ‐DBE‐DBCH could enter the environment as a result of combustion of materials containing DBE‐DBCH as a flame retardant. However, most studies do not distinguish between the isomers. Therefore, the relative potencies could not be addressed.

3.3.4.2. Occurrence data and dietary exposure assessment

No occurrence data of DBE‐DBCH in food were submitted to EFSA, but some studies in the scientific literature were identified (see Section 3.1.2 for an overview).

One study was identified reporting estimates of dietary exposure to DBE‐DBCH. Sahlström et al. (2015) reported occurrence levels and exposure of α‐ and β‐DBE‐DBCH, among other BFRs from analysis in Swedish market basket food samples collected in 2010, food baskets comprising of fish, meat, vegetable oils, dairy products and eggs (n = 4 food baskets). Both α‐ and β‐DBE‐DBCH were detected only in fish samples, at 97 and 17 pg/g ww respectively, while concentration in other food categories were below the LOD (See Section 3.1.2 and Appendix B). The dietary exposure was based on measured concentrations and intake assumptions from questionnaire data about dietary habits of the participants. The estimated median dietary daily intake for both mothers and toddlers (n = 20) based on fish intake only, was 3.5 and 1.7 ng per day for α‐DBE‐DBCH, respectively and 0.6 and 0.3 ng per day for β‐DBE‐DBCH, respectively.

For breastfed infants, there were insufficient data in human milk from European countries reported in the literature to make robust exposure estimates. To provide an indication of the potential exposure, the CONTAM Panel made an estimation for average and high consumption of human milk (see Section 2.4) considering the available data reported in the literature (Čechová et al., 2017; Tao et al., 2017, see Section 3.3.4.1.2).

The exposure scenario based on average human milk consumption and the highest reported mean concentration in human milk for DBE‐DBCH isomers across three European countries (Norway, the Netherlands and Slovakia) of 0.038 ng/g lipid (Čechová et al., 2017), resulted in an estimate of 0.17 ng/kg bw per day. The study by Tao et al. (2017) analysed α‐ and β‐DBE‐DBCH in samples from the UK: for α‐DBE‐DBCH the reported mean concentration of 0.67 ng/g lipid would result in an average exposure of 3.1 ng/kg bw per day, while for β‐DBE‐DBCH, the reported mean concentrations of 6.8 and 2.5 ng/g lipid for samples collected in 2010 and 2014–2015, respectively, would result in average exposures of 31.2 and 11.5 ng/kg bw per day, respectively.

The exposure scenario based on high human milk consumption and the highest reported P95 concentration for DBE‐DBCH isomers of 0.083 ng/g lipid (Čechová et al., 2017), resulted in an estimate of 0.57 ng/kg bw per day. The study by Tao et al. (2017) reported only maximum concentrations that were not used for these estimations.

It should be noted that these estimates are based on human milk samples from four European countries that were collected one or two decades ago. It is unknown whether this is representative of current exposure to this BFR.

3.3.4.3. Risk characterisation

No risk characterisation was possible, due to the lack of a Reference Point and the limited data on dietary exposure.

3.3.4.4. Uncertainty analysis

Due to the lack of data on occurrence and exposure for DBE‐DBCH, uncertainty analysis was performed only for the conclusion on genotoxicity of DBE‐DBCH using the methods described in Section 2.8.

Based on the positive in vitro mutagenicity in mammalian cells of DBE‐DBCH in the absence of conclusive negative results for in vivo mutagenicity data, the CONTAM Panel concluded that DBE‐DBCH is genotoxic. Based on the overall evidence, and taking into account the associated uncertainties, the CONTAM Panel judged with > 95% certainty (i.e. probability) that DBE‐DBCH is genotoxic.

3.3.5. Pentabromoethylbenzene (PBEB) [CAS number 85‐22‐3]

PBEB is a non‐polymeric additive flame retardant. PBEB is a compound of high persistency in the environment, although it may undergo reductive and photolytical debromination. It has been reported to be used as an additive flame retardant in unsaturated polyesters, styrene butadiene copolymers and in other textiles and also in circuit boards, adhesives, wire and cable coatings, and in polyurethane foam (EFSA CONTAM Panel, 2012b).

The physicochemical properties of PBEB are reported in Table 1. PBEB is currently not registered under REACH (see Table 2), and thus its production volume and use within the EU should be lower than one tonne per year. It has been predicted by QSAR to be ‘(very)persistent’, ‘very bioaccumulative’ and ‘toxic’ (ECHA, 2024).

3.3.5.1. Hazard identification and characterisation

3.3.5.1.1. Toxicokinetics

No information on toxicokinetics was reported in the previous Opinion, nor in the current update.

3.3.5.1.2. Levels in humans

The previous Opinion did not report on any study in European countries, but identified one study from China in which PBEB was not detected in serum samples from office cleaners, university students or policemen (Zhu et al., 2009).

Since then, some studies in human samples from European countries, including human milk and blood/serum, have been identified. These studies are briefly described below and details can be found in Appendix E (Table E.1).

Regarding human milk samples,

In Sweden, Sahlström et al. (2015) did not detect PBEB in 2 pooled human milk samples collected in 2009–2010 (< 0.6 pg/g ww).

In the study by Čechová et al. (2017), PBEB was analysed in human milk samples from Norway (2003–2006, n = 305), the Netherlands (2011–2014, n = 116) and Slovakia (2011–2012, n = 37). A low detection frequency of 2% and 3% was reported in samples from Norway and Slovakia, while it was not detected in samples from the Netherlands.

PBEB was also not detected in human milk samples collected in the UK in 2010 (n = 25) and 2014–2015 (n = 10) (LOD not reported) (Tao et al., 2017).

Regarding blood/serum,

PBEB was neither detected above the LOD in 24 matched serum samples from mothers and their toddlers collected in Sweden between 2009 and 2010 (LOD = 0.9 pg/25 μL, Sahlström et al. (2014), nor in 61 serum samples collected in 2013 from a Norwegian cohort (LOD = 14 pg/serum sample, volume not reported, Tay et al., 2019).

3.3.5.1.3. Toxicity in experimental animals

Studies considered in the previous EFSA assessment

In the previous Opinion (EFSA CONTAM Panel, 2012b), the information given on hazard identification of PBEB was that the LD50 in rabbits following dermal exposure exceeded 8000 mg/kg bw.

Regarding genotoxicity, the previous Opinion reported that PBEB was negative in the Ames test in S. Typhimurium strains TA98, TA100, TA1535 and TA1537 with and without metabolic activation.

Studies published since the previous EFSA assessment

Since then, some studies on the effects of PBEB on rodents have been published.

• Pregnant CD‐1 mice were orally administered PBEB at 0, 0.2, 2 and 20 mg/kg bw per day in olive oil from gestation until the day before delivery, with dosing resumed on PND3 until weaning (Wang et al., 2024). Offspring were then fed a high‐fat ‘Western diet’. The perinatally exposed male (but not female) offspring exhibited liver injury, which was significant at the highest dose only (20 mg/kg bw per day), but without clear dose–response. Inflammation with elevated plasma ALT and AST markers and histological liver damage was observed, which is consistent with the observed hepatotoxicity. PBEB exposure led to overexpression of TWEAK in liver macrophages, promoting M1 macrophage polarisation and activating NF‐κB and NLRP3 inflammasome pathways, triggering pyroptosis. These cellular and molecular effects were predominantly observed at 20 mg/kg bw per day and in male mice. The NOAEL was 2 mg/kg bw per day based on liver inflammation scores in males.

• In a study designed to study mechanisms of obesogenic effects of PBEB, pregnant CD‐1 mice were orally administered PBEB at 0, 0.2, 2 and 20 mg/kg bw per day in olive oil from gestation until the day before delivery, with dosing resumed on PND3 until weaning with offspring maintained until 12 weeks of age (Xu et al., 2023). Perinatal exposure to 20 mg/kg bw per day (but not lower doses) caused increased absolute weight and hypertrophy of epididymal white adipose tissue in male offspring, without a clear dose–response. The NOAEL was 2 mg/kg bw per day based on increased weight gain and epididymal white adipose tissue hypertrophy in male offspring.

• Five‐week‐old Sprague–Dawley rats were administered PBEB by gavage at 0, 0.2, 2 and 20 mg/kg bw per day for 28 days to investigate its effect on thyroid function (Lu et al., 2020). In females, serum TT3 and relative thyroid weight decreased at all doses. Serum TT4 and TSH increased in both sexes. Thyroid follicles became depleted and epithelial cell layers thickened at ≥ 2 mg/kg bw per day but it was unclear if this refers to males or females or both. PBEB altered thyroid transcriptome profiles with 17, 42 and 119 differentially expressed genes at the three doses, including downregulation of Tg, Tpo, Slc5a5, Ttr and Dio1, and upregulation of Tshb. The LOAEL was 0.2 mg/kg bw per day based on increased serum TSH and reduced thyroid weight.

3.3.5.1.4. Computational analysis

The QSAR Toolbox identified experimental data for the Ames test on PBEB (see above and Annex B).

VEGA models predicted negative Ames test results for PBEB with ‘GOOD reliability’ (three models) and with ‘LOW reliability’ (one model). The Panel critically reviewed the predictions and considered them valid mainly due to the presence of relevant analogues known by the model with experimental and predicted values concordant with the prediction for the target compound (see Annex B). The CONTAM Panel noted that this negative prediction is consistent with the available experimental data.

3.3.5.1.5. Observations in humans

No studies were identified in the previous Opinion. Since then, five publications were identified, four studies pertaining to populations in China, and one study in the USA. Exposure was assessed in serum in all Chinese studies, and in indoor air and wristbands in the USA study. Various endpoints were assessed including thyroid hormone profile (three studies), thyroid cancer (one study), type 2 diabetes (one study), BMI (one study) and liver injury markers (one study). In Appendix F an overview of the epidemiological studies identified reporting on the BFRs considered in this Opinion, including those on PBEB, is shown.

In the study by Zhao et al. (2021) that assessed the cross‐sectional relationship between the serum levels of five BFRs and a series of thyroid/liver injury biomarkers in residents of a major BFR‐producing region in China (for details se description in Section 3.3.1.1), serum PBEB levels were associated with a higher serum TT3 levels, but not with the other thyroid hormones and liver biomarkers measured.

In the study by Liu, Li, et al. (2022) that assessed the association between the levels of several BFRs in serum and thyroid cancer risk in Shandong Province (China) (for details see description in Section 3.2.4.2), PBEB was detected in less than 20% of samples and was not further analysed.

In the USA study by Wang, Romanak, et al. (2019) (for details see Section 3.2.4.2), PBEB concentrations, measured in wristbands of 101 adult study participants, were not statistically significant associated with thyroid hormone profiles.

In the type 2 diabetes case–control study in China by Zheng, Chen, et al. (2025) (for details see Section 3.2.4.2), serum PBEB concentrations were not statistically significantly different between cases and controls in unadjusted comparisons (adjusted associations for single exposures were not reported) and PBEB was not a main contributor to the mixture effects.

In the study by Zheng, Sun, et al. (2025) in China (for details see Section 3.2.4.2), serum PBEB concentrations showed a statistically significant positive correlation with BMI. Adjusted models were not presented.

In summary, there are five cross‐sectional studies that assessed PBEB and four of these assessed associated health outcomes; in the fifth study PBEB was not further analysed due to low detection levels. Statistically significant associations were observed between PBEB serum levels and TT3 levels in one study, and PBEB and BMI in another (unadjusted correlation) but no other associations were reported.

3.3.5.1.6. Mode of action

An overview of the studies identified since the previous Opinion is given in Appendix G (Table G.6).

In mouse 3T3‐L1 cells, PBEB selectively activated retinoid X receptor α (RXRα) but not the peroxisome proliferator‐activated receptor γ (PPARγ), disrupting AMPK and PI3K/AKT signalling pathways involved in lipid metabolism (Xu et al., 2023). Inhibiting TWEAK/Fn14 signalling reduced inflammatory markers and inflammasome activation in vitro (Wang et al., 2024).

In cell‐based dual‐luciferase assays using CV‐1 cells, PBEB acted as a thyroid receptor β (TRβ) antagonist with an IC₅₀ of 9.82 × 10⁻⁷ M, showing no agonist activity between 10⁻¹⁰ and 10⁻⁶ M (Lu et al., 2020). Molecular modelling indicated that PBEB binds within the TRβ ligand‐binding pocket, inducing conformational changes in helices H11 and H12.

Yeast reporter gene assays showed PBEB exhibited no oestrogenic or androgenic or anti‐oestrogenic and anti‐androgenic activities up to 4.1 μM (Bajard et al., 2021).

3.3.5.1.7. Consideration of critical effects and dose–response analysis

PBEB was negative in an Ames test, but no data were found on clastogenicity or aneugenicity making it impossible to conclude on possible genotoxicity.

Two developmental studies in mice were available. Both of these studies focus on specific effects (liver inflammation or obesogenic effects) and were mechanistic in nature. One involved administration of a high‐fat diet and reported liver injury in male offspring only. The second reported increase in absolute weight and hypertrophy of epididymal white adipose tissue in male offspring. The NOAEL in both studies was 2 mg/kg bw per day, but a clear dose–response was lacking in either study. A third animal study focused on effects of PBEB on the thyroid system. They found a dose‐dependent increase in TSH and reduced thyroid weight in a 28‐day study in rats, with a LOAEL of 0.2 mg/kg bw per day. The mode of action for this effect could be that PBEB acts as an antagonist of TRβ. Observations in humans indicate an association between serum PBEB levels and an increase in serum TT3 levels (Zhao et al., 2021).

Due to the limited scope of the three available animal studies – particularly the narrow range of endpoints investigated and the use of a high‐fat diet in one study – as well as the absence of sub‐chronic or long‐term studies and the lack of studies addressing developmental toxicity, reproductive toxicity or developmental neurotoxicity, the CONTAM Panel was unable to identify a suitable Reference Point for PBEB.

3.3.5.2. Occurrence data and dietary exposure assessment

Limited occurrence data in food were submitted to EFSA (see Section 3.1.1) and no estimation of the dietary exposure based on these data was possible.

In the scientific literature, several studies were identified reporting on the occurrence of PBEB and an overview is given in Section 3.1.2 and Appendix B.

One study was identified reporting estimates of dietary exposure to PBEB. Sahlström et al. (2015) reported occurrence levels and exposure of PBEB, among other BFRs from analysis in Swedish market basket food samples collected in 2010, food baskets comprising of fish, meat, vegetable oils, dairy products and eggs (n = 4 food baskets). PBEB was detected in fish (100% detection frequency) and meat (50% detection frequency) samples, at 8.7 and 0.60–2.0 pg/g ww respectively, while the concentration in other food categories was below the LOD. The dietary exposure was based on measured concentrations and intake assumptions from questionnaire data about dietary habits of the participants. The estimated median dietary daily intake for both mothers and toddlers (n = 20) based on fish intake only, was 0.3 and 0.2 ng per day, respectively.

No estimation of the exposure of breastfed infants was possible as the studies from European countries identified in the literature reported values < LOD or with low detection frequency (< 3%).

3.3.5.3. Risk characterisation

No risk characterisation was possible, due to the lack of a Reference Point and the limited data on dietary exposure.

3.3.6. Pentabromotoluene (PBT) [CAS number 87‐83‐2]

PBT is a reactive flame retardant primarily used in textiles, and can be released from oligomeric BFRs, such as BC‐58 (a TBBPA carbonate oligomer) and FR‐1025 (a pentabromobenzyl acrylate oligomer) (EFSA CONTAM Panel, 2012b).

The physicochemical properties of PBT are reported in Table 1. PBT is currently not registered under REACH (see Table 2), and thus its production volume and use within the EU should be lower than one tonne per year.

3.3.6.1. Hazard identification and characterisation

3.3.6.1.1. Toxicokinetics

No information on toxicokinetics was identified in the previous Opinion, nor in the current update.

3.3.6.1.2. Levels in humans

The previous Opinion did not report any study on levels of PBT in human samples. Since then, several studies have been identified in human milk and blood/serum from European countries. These studies are briefly described below, and details can be found in Appendix E (Table E.1).

Regarding human milk samples,

• Čechová et al. (2017) analysed PBT in human milk samples from Norway (2003–2006), the Netherlands (2011–2014) and Slovakia (2011–2012) with a detection frequency between 85% and 100%. Mean and P95 concentrations across countries ranged from 0.028 to 0.049 ng/g lipid and from 0.063 to 0.089 ng/g lipid, respectively.

• In Sweden, Sahlström et al. (2015) analysed two pooled human milk samples collected in 2009–2010. PBT was not detected in any of the two pooled samples (< 0.4 pg/g ww).

• PBT was below the LOD in human milk samples collected in the UK in 2010 (n = 25) and 2014–2015 (n = 10) (LOD not reported) (Tao et al., 2017).

Regarding levels in blood/serum,

• Sahlström et al. (2014) did not detect PBT above the LOD (0.9 pg/25 μL) in any of the 24 matched serum samples from mothers and their toddlers collected in Sweden between 2009 and 2010.

• PBT was also reported below the LOD in 61 serum samples collected in 2013 from a Norwegian cohort (LOD = 14 pg/serum sample, volume not reported, Tay et al., 2019).

3.3.6.1.3. Toxicity in experimental animals

Studies considered in the previous EFSA assessment

The previous Opinion reported on some studies on PBT. In a sub‐chronic study during 91 days in male and female Sprague–Dawley rats fed a diet containing PBT (purity 98%) at 0, 0.05, 0.5, 5.0, 50 or 500 mg/kg (about 0, 0.003–0.004, 0.03–0.04, 0.35–0.4, 3.5–4.0 or 34–40 mg/kg bw per day, covering both males and females), clinical observations revealed no adverse effects (Chu et al., 1987; Danish EPA, 2000). Mild histopathological changes were observed in the liver (e.g. anisokaryosis, accentuated zonation), kidney (e.g. eosinophilic inclusions in the proximal tubules) and thyroid gland (e.g. reduced follicular size, increased epithelial heights). A NOAEL of 5 mg PBT/kg feed (equivalent to 0.35 mg/kg bw per day) was identified by the authors (Chu et al., 1987; Danish EPA, 2000). In another study no adverse fetal effects were observed when doses up to 600 mg/kg bw were given orally to rat dams during organogenesis (Danish EPA, 2000, as cited in SFT, 2009).

Regarding genotoxicity, the previous Opinion also reported that according to Zeiger et al. (1987) (as reported by Danish EPA, 2000), PBT was not mutagenic in an Ames test (pre‐incubation assay) with S. Typhimurium TA98, TA100, TA1535 and TA1537 with and without metabolic activation (hepatic S9 fraction from Aroclor 1254‐treated rats and hamsters). No data are available on chromosomal aberrations or other genotoxic effects.

Studies considered in the previous EFSA assessment

Since the previous Opinion, no new studies have been identified.

3.3.6.1.4. Computational analysis

The QSAR Toolbox identified experimental data for the Ames test on PBT (see above and Annex B).

VEGA models predicted negative Ames test results for PBT with ‘GOOD reliability’ (three models) and with ‘LOW reliability’ (one model). The Panel critically reviewed the predictions and considered them valid mainly due to the presence of relevant analogues known by the model with experimental and predicted values concordant with the prediction for the target compound (see Annex B). The CONTAM Panel noted that this negative prediction is consistent with the available experimental data.

3.3.6.1.5. Observations in humans

No studies were identified in the previous Opinion. Since then, four publications were identified, pertaining to populations in China. Exposure was assessed in serum in all studies, and various endpoints were assessed including thyroid hormone profile (two studies), thyroid cancer (one study), liver injury markers (one study), type 2 diabetes (one study) and BMI (one study). In Appendix F, an overview of the epidemiological studies identified reporting on the BFRs considered in this Opinion is shown, including those on PBT.

In the study by Zhao et al. (2021) that assessed the cross‐sectional relationship between the serum levels of five BFRs and a series of thyroid/liver injury biomarkers in residents of a major BFR‐producing region in China (for details see description in Section 3.3.1.1), PBT serum levels were associated with a lower ratio of liver biomarkers AST/ALT. No associations were observed for any of the other thyroid hormones or liver injury biomarkers.

In the study by Liu, Li, et al. (2022) that assessed the association between the levels of several BFRs in serum and thyroid cancer risk in Shandong Province (China) (for details see description in Section 3.2.4.2), serum PBT levels were not associated with thyroid cancer risk. In control participants, PBT was associated with lower total T3 levels (−3.14%; 95% CI: −6.42, −0.30).

In the study by Zheng, Chen, et al. (2025) in China (for details see Section 3.2.4.2), serum PBT concentrations were not statistically significantly different between type 2 diabetes cases and controls in unadjusted comparisons (adjusted associations for single exposures were not reported) and PBT was not a main contributor to the mixture effects.

In the study by Zheng, Sun, et al. (2025) in China (for details see Section 3.2.4.2), serum PBT concentrations did not show statistically significant associations with BMI. Adjusted models are not presented.

3.3.6.1.6. Mode of action

The previous Opinion reported that PBT was found to be a very weak agonist (more than six orders of magnitude less potent than 2,3,7,8‐TCDD) of the AHR in the CALUX bioassay (Brown et al., 2004).

An overview of the studies identified since the previous Opinion is given in Appendix G (Table G.7).

Preadipocyte (3T3‐L1) cell line and a hepatocyte (HepG2) cell line were used to investigate the possible lipid metabolism disruption caused by PBT. PBT did not significantly affect preadipocytes proliferation, but increased lipid accumulation in hepatocytes and also display a significant increase of PPARγ gene expression (Maia et al., 2022).

MCF‐7 cells were treated with PBT using short and long‐term (4 weeks) exposure protocols. Cell viability was not altered; however, cell proliferation was reduced after 24 h or 4 weeks exposure (Sousa et al., 2022).

Bajard et al. (2019) collected in vivo toxicity data on 62 FRs to establish prioritisation categories based on data availability and toxicological concern; the authors classified PBT as having ‘scarce toxicological data indicating lower toxicological concern'.

3.3.6.1.7. Consideration of critical effects and dose–response analysis

Mild histopathological changes were observed in the liver, kidney and thyroid gland in rats exposed to PBT via diet for 91 days. The NOAEL was ≈ 0.35 mg/kg bw per day. No adverse fetal effects were observed in rats exposed by gavage up to 600 mg/kg bw per day during organogenesis.

PBT was negative in an Ames test in the presence or absence of metabolic activation. No data on clastogenicity or aneugenicity were identified. Therefore, the CONTAM Panel cannot conclude on the genotoxic potential of PBT.

No long‐term studies, reproductive toxicity studies or neurodevelopmental toxicity studies have been identified.

The CONTAM Panel concluded that the database on PBT is too weak to identify a Reference Point.

3.3.6.2. Occurrence data and dietary exposure assessment

Limited occurrence data in food were submitted to EFSA (see Section 3.1.1) and no estimation of the dietary exposure based on these data was possible.

In the scientific literature, several studies were identified reporting on the occurrence of PBT and an overview is given in Section 3.1.2 and Appendix B.

One study was identified reporting estimates of dietary exposure to PBT. Sahlström et al. (2015) reported occurrence levels and exposure of PBT, among other BFRs from analysis in Swedish market basket food samples collected in 2010, food baskets comprising of fish, meat, vegetable oils, dairy products and eggs (n = 4 food baskets). PBT was detected only in fish (100% detection frequency) samples, at 1.8 pg/g ww, while concentration in other food categories were below the LOD. The dietary exposure was based on measured concentrations and intake assumptions from questionnaire data about dietary habits of the participants. The estimated median dietary daily intake for both mothers and toddlers (n = 20) based on fish intake only, was 0.07 and 0.03 ng per day, respectively.

For breastfed infants, there were insufficient data in human milk from European countries reported in the literature to make robust exposure estimates. To provide an indication of the potential exposure, the CONTAM Panel made an estimation for average and high consumption of human milk (see Section 2.4) considering the available data reported in the literature. Only one study reported levels above the LOD (Čechová et al., 2017, see Section 3.3.6.1.2).

The exposure scenario based on average human milk consumption and the highest reported mean concentration in human milk across three European countries (Norway, the Netherlands and Slovakia) of 0.049 ng/g lipid (Čechová et al., 2017), resulted in an estimate of 0.22 ng/kg bw per day.

The exposure scenario based on high human milk consumption and the highest reported P95 concentration of 0.089 ng/g lipid (Čechová et al., 2017), resulted in an estimate of 0.61 ng/kg bw per day.

It should be noted that these estimates are based on human milk samples from three European countries that were collected one or two decades ago. It is unknown whether this is representative of current exposure to this BFR.

3.3.6.3. Risk characterisation

No risk characterisation was possible, due to the lack of a Reference Point and the limited data on dietary exposure.

3.3.7. 1,3,5‐Tris(2,3‐dibromopropyl)‐1,3,5‐triazine‐2,4,6‐trione (TDBP‐TAZTO) [CAS number 52434‐90‐9]

TDBP‐TAZTO is used as an additive flame retardant as a mixture of two congeners, one substituted with one and the other with two 2,3‐dibromopropyl groups (DBP‐TAZTO and BDBP‐TAZTO), respectively (EFSA CONTAM Panel, 2012b). This compound is the subject of a review by Bar and Szychowski (2022).

The physicochemical properties of TDBP‐TAZTO are reported in Table 1. TDBP‐TAZTO is registered in REACH (100–1000 tonnes per annum) (see Table 2).

3.3.7.1. Hazard identification and characterisation

3.3.7.1.1. Toxicokinetics

No information on toxicokinetics was reported in the previous Opinion. Since then, two studies have been identified.

Tong et al. (2021) developed an innovative method (matrix‐assisted laser desorption/ionisation imaging mass spectrometry (MALDI–IMS) using 1,5‐diaminonaphthalene hydrochloride and silver trifluoromethanesulfonate (NDA/AgOTf) as the matrix for the in‐situ localisation) for TDBP‐TAZTO quantification and distribution analyses in mice organs. ICR mice (18 males and 18 females) were treated by gavage to 80 or 160 mg/kg bw and sacrificed at day 15 or 30. The authors found that TDBP‐TAZTO accumulates in mice organs in the following order: kidney > liver > heart > brain.

Ma, Ma, et al. (2024) studied the in vitro metabolism of TDBP‐TAZTO using human liver microsome. The major CYP450 isoforms involved in the metabolism of TDBP‐TAZTO were the CYP2E1 and 3A4 followed by the CYP2C9. The authors detected four metabolites: a hydroxylated debrominated metabolite (C₁₂H₁₆Br₅N₃O₄), a dihydroxylated metabolite (C₁₂H₁₅Br₆N₃O₅), a desaturated debrominated metabolite (C₁₂H₁₄Br₅N₃O₃) and an N‐dealkylated metabolite (C₉H₁₁Br₄N₃O₃).

3.3.7.1.2. Levels in humans

No studies were identified in the previous Opinion, nor in the current update.

3.3.7.1.3. Toxicity in experimental animals

Studies considered in the previous EFSA assessment

No data on toxicity in experimental mammalian species were reported in the previous Opinion (EFSA CONTAM Panel, 2012b).

Studies published since the previous EFSA assessment

Since the previous Opinion several studies have been identified in the scientific literature addressing the toxicity of TDBP‐TAZTO (Han et al., 2022; Li et al., 2015; Zhang et al., 2024; Zhou et al., 2018) and two studies focused on the neurotoxic effects (Dong et al., 2015; Ye et al., 2015).

Li et al. (2015) studied the effects of TDBP‐TAZTO on growth, blood biochemical parameters organs and related gene expression in both male and female Balb/c mice daily administered by gavage for 28 days at 0, 2, 10 or 50 mg/kg bw per day. The results indicated that TDBP‐TAZTO exposure had no effect on body weight gain and food intake. The serum ALT was increased in male mice at all 3 dose levels, although not in a dose‐dependent manner, while the relative liver weight was significantly increased only in males exposed to 50 mg/kg bw per day. A higher relative kidney weight was observed in males exposed to 10 mg/kg bw per day whereas the index was decreased in females treated with 50 mg/kg bw per day. Histopathological analyses showed both liver and lung lesions (liver: cell necrosis and ballooning degeneration; lung: epithelium hyperplasia, mitochondrial swelling and inflammatory cell infiltration) with increasing severity according to the level of dose.

Male Wistar rats (4‐week‐old) were administered daily with TDBP‐TAZTO for 4 weeks by gavage at doses of 0, 2 or 50 mg/kg bw per day (Zhou et al., 2018). No variations of body weight nor relative organ weights were observed at the end of the period of exposure between the three groups, except for a significant increase of the relative brain weight in the two TDBP‐TAZTO‐exposed groups compared to controls. Free T4 and estradiol blood levels were decreased in the 2 and the 50 mg/kg bw per day‐treated groups, respectively. AST and ALT activities were significantly reduced in 50 mg/kg per day‐exposed animals compared to control rats. Slight histopathological lesions were reported in lung (diffuse thickened epithelial walls, partial fibrosis and vasodilatation), spleen (enlargement in the marginal zone) and kidney (tissular congestion and inflammation) according to the dose of exposure and the organ.

TDBP‐TAZTO liver toxicity was assessed in vivo (rats) and in vitro (HepG2 cells) (Han et al., 2022). The in vivo part of the study consisted of four groups of 20 Wistar rats each (4‐week‐old, half male and half female in each group) daily administered with TDBP‐TAZTO by gavage for 28 days. The dose levels were 0, 313, 625 and 1250 mg/kg bw per day. Body and organ weights, blood chemistry and histopathological parameters were analysed at the end of the exposure period. The body weight and the relative organ weights revealed insignificant differences between the four groups except for the liver that showed a significant increase in the relative organ/body weight with the dose of TDBP‐TAZTO. The blood chemistry results including liver AST and ALT activities, total protein and albumin contents, lipid (CT and TG) concentrations and haematological parameters showed higher values in TDBP‐TAZTO‐exposed rats compared to controls which confirmed the liver to be a target organ for this BFR.

Zhang et al. (2024) studied the respiratory effects of TDBP‐TAZTO. Male C57BL/6J mice (5–6 weeks old) were daily administered by gavage for 14 days at a dose of 20 mg/kg bw per day. Parameters including lung function, body weight, organ indices, pathological alterations and the regulatory expression of lung inflammatory signalling were assessed in TDBP‐TAZTO‐treated male mice. Mice were also co‐treated with estradiol (E2) to assess its potential protective role against TDBP‐TAZTO‐induced damages. Results showed the ability of TDBP‐TAZTO to induce significant damages in the lung tissue, including haemorrhage, alveolar wall thickening, infiltration of inflammatory cells into the alveolar space and interstitium, and collagen deposition. The lung injury scores, as indicator of the severity of lung injury, were significantly higher in TDBP‐TAZTO‐exposed mice than in controls. Lung function was impaired as reflected by the increase in physiological variables as the accumulated volume and the airway pressure–time index in TDBP‐TAZTO‐treated animals compared to controls. Immunohistochemical analysis revealed that TDBP‐TAZTO exposure reduced ERα expression and increased nuclear factor kappa‐light‐chain‐enhancer of activated B cells (NF‐κB) expression. Co‐exposure of estradiol with TDBP‐TAZTO significantly mitigated the COPD‐like phenotypes. The authors concluded that TDBP‐TAZTO induced chronic obstructive pulmonary disease (COPD)‐like phenotypes in lungs of mice orally exposed, with a role for the ERα‐mediated inflammatory pathway in the physiopathology of COPD induced by this compound.

Dong et al. (2015) administered TDBP‐TAZTO (purity > 97%) at 0, 5 or 50 mg/kg bw per day to male Balb/c mice daily by gavage for 30 days and studied for brain and behaviour toxicity. Depressive‐related behaviour was measured in the forced swim test and the tail suspension test, locomotor activity in the open‐field and apoptosis‐related protein expression (Bax and Bcl‐2) in the hippocampus of the same animals. TDBP‐TAZTO induced a significant dose‐dependent increase in depressive‐related behaviour in both tests whereas locomotor activity remained unchanged.

Ye et al. (2015) administered male Sprague–Dawley rats daily by gavage with TDBP‐TAZTO (purity > 97%) at 0, 5 or 50 mg/kg bw per day for 6 months and studied for brain and behaviour toxicity during the last 2 weeks of the 6th month. Depressive‐related behaviour was measured in the forced swim test, locomotor activity in the open‐field, and spatial learning and memory performances in the Morris water maze. Brains and adrenal glands were removed and studied for histopathological and immunohistochemical examination and measurements of oxidative stress markers, pro‐inflammatory cytokines and apoptosis‐related proteins. TDBP‐TAZTO dose‐dependently induced significant increases in depressive‐related behaviour and in plasma ACTH and serum corticosterone levels. TDBP‐TAZTO also impaired the learning and memory performances of the animal at both doses whereas locomotor activity remained unchanged.

Additional studies were available in the REACH dossier, ³⁶ as submitted by the registrant(s). The CONTAM Panel did not have access to the original data. The REACH dossier reported that:

• The technical product AP 729 (purity NR) was tested in a Combined Repeated Dose Toxicity Study with the Reproduction/Developmental Toxicity Screening Test (OECD TG 422) in Wistar rats by oral gavage at doses of 0, 100, 300 and 1000 mg/kg bw per day. AP 729 was administered for 2 weeks prior to mating, during mating and post‐mating periods for males, during pregnancy and up to Lactation Day 13 for females. There were no treatment‐related effects on systemic, reproduction and fertility parameters and on F1 generation pups up to and including the highest dose tested of 1000 mg/kg bw per day. Moreover, no treatment‐related neurological abnormalities were observed at any of the doses tested. The NOAEL was 1000 mg/kg bw per day, the highest dose tested (REACH dossier).

Genotoxicity

Since the previous Opinion, two studies have been cited in the REACH dossier, as submitted by the registrant(s): an in vitro mammalian cell micronucleus test (OECD Guideline 487), and an in vivo mammalian alkaline comet assay (OECD Guideline 489). However, no information on the protocol or results was available in the publicly available REACH dossier due to confidentiality restrictions.

3.3.7.1.4. Computational analysis

The QSAR Toolbox identified experimental data for the Ames test on TDBP‐TAZTO, that referred to the REACH dossier (see above and Annex B). In the absence of further details, the CONTAM Panel could not interpret this information.

VEGA models predicted positive Ames test results for TDBP‐TAZTO with ‘MODERATE reliability’ (three models) and negative Ames test results with ‘MODERATE reliability’ (one model). The CONTAM Panel assessed the validity of the positive predictions with moderate reliability according to VEGA, including the relevance of the analogues with experimental data identified by the software and the accuracy and consistency of their predictions. The Panel noted that the analogues were not sufficiently similar to the target compound to be confident in the reliability of the predictions, and thus these predictions were not further considered.

3.3.7.1.5. Observations in humans

No studies were reported in the previous Opinion, nor in the current update.

3.3.7.1.6. Studies in other model systems and on mode of action

The previous Opinion reported two studies in zebrafish and one study in cultured cells (EFSA CONTAM Panel, 2012a, 2012b). Exposure of adult zebrafish to 0, 0.25, 1 or 4 mg/L TDBP‐TAZTO in water for 28 days had no effect on survival or growth. Significant damage to the liver and gill was observed at 4 mg/L. Thyroid hormones, testosterone, estradiol, liver superoxide dismutase activity, malondialdehyde content and brain cholinesterase activity were not affected (Zhang et al., 2011). Zebrafish eggs and embryos exposed to TDBP‐TAZTO at 0, 0.5, 1, 2.5, 5 or 10 μg/mL of water showed higher mortality at longer exposure periods in the three highest dose groups, and hatching rate was reduced in the four highest dose groups. Dose‐dependent defects in inflation of the gas bladder were observed in the three highest dose groups. Electron microscopy examination of the gas bladder revealed changes in the electron density of the cytoplasmic vesicles and disruption of the mitochondrial cristae (Li et al., 2011). The one in vitro study reported, showed that TDBP‐TAZTO was cytotoxic for primary cultured cerebellar granule neurons at concentrations in the range of 5–10 μM (Qu et al., 2011).

New publications related to brain effects and mode of action of TDBP‐TAZTO have been published since the previous Opinion, most of them related to impacts on steroidogenesis and oestrogenic and androgenic activities. Several in vitro studies also explored the neurotoxicity of TDBP‐TAZTO. An overview of these studies is given in Appendix G (Table G.8) and a brief summary is provided below.

In vitro, TDBP‐TAZTO impacted cell viability and triggered apoptosis and oxidative stress in SH‐SY5Y neuronal cell line (Dong et al., 2015; Szychowski et al., 2021) whereas neuronal differentiation remains unaffected (Szychowski et al., 2021). Dose‐dependent increases in apoptosis‐related and protein expressions and pro‐inflammatory cytokine levels were observed in hippocampus of adult rats exposed to TDBP‐TAZTO for 6 months (Ye et al., 2015). However, such effects on apoptosis and neuroinflammation were not observed in primary mouse astrocytes exposed for 48 h to TDBP‐TAZTO (Szychowski & Skóra, 2023a).

In the same model of primary mouse astrocyte cell culture, TDBP‐TAZTO disturbed the production of progesterone, testosterone and oestradiol, and the expression of key‐related enzymes involved in the hormonal production (Szychowski & Skóra, 2023b), suggesting the ability of this compound to dysregulate sex steroid hormone production. The effects of TDBP‐TAZTO on steroidogenesis, production of steroid hormones, enzyme‐related expression and interactions with hormonal receptors were studied in vitro in various cell lines, including H295R adrenocortical cells (Li et al., 2016), MCF‐7 human breast adenocarcinoma cells (Cao et al., 2018; Krivoshiev et al., 2016), MDA kb‐2 epithelial cells (Bajard et al., 2023) and mouse‐derived spermatogonic GC‐1 spg cells (Tabecka‐Lonczynska et al., 2023,b, 2024).

TDBP‐TAZTO showed a significant anti‐oestrogenic activity in MCF‐7 cells co‐exposed to 17β‐estradiol (Krivoshiev et al., 2016) by affecting the ERα‐mediated signalling pathway using an in vitro ER‐dependent luciferase reporter gene assay (Cao et al., 2018). TDBP‐TAZTO exhibited anti‐androgenic properties in MDA kb‐2 cells alone or in combination with other anti‐androgenic FRs using the same type of luciferase reporter gene assay (Bajard et al., 2023); this finding confirmed the predicted androgen receptor antagonist activity of this compound from two QSAR models (Bajard et al., 2021).

In mouse‐derived spermatogonic GC‐1 spg cells, TDBP‐TAZTO impaired the process of steroidogenesis by reducing the expression of proteins and enzymes involved in the synthesis of gonadal hormones, including progesterone and slightly, estradiol and testosterone (Tabecka‐Lonczynska, Kaczka, & Kaleniuk, 2023). The ability of TDBP‐TAZTO to inhibit sex hormone production was also demonstrated in H295R adrenocortical cells (Li et al., 2016) due to the interference of this compound with steroidogenic gene expression when activated with the co‐exposure to the cAMP inducer forskolin. The impact of TDBP‐TAZTO on steroidogenesis and hormone production was related to its ability to disrupt androgenic‐signalling pathways, to modulate autophagy processes, and to affect cell proliferation and survival in GC‐1 spg cells (Tabecka‐Lonczynska et al., 2024). The ability of TDBP‐TAZTO to affect AhR, PPARγ and ER‐related gene expression and then to dysregulate the steroid‐related signalling pathway in the same model of mouse‐derived spermatogonic GC‐1 spg cells (Tabecka‐Lonczynska, Skóra, et al., 2023) suggests that TDBP‐TAZTO may affect sex steroid hormone production.

Li et al. (2015) reported the modifications in the expression of apoptosis‐related genes (p53 and bcl‐2) in liver and lung of both male and female Balb/c mice daily administered by gavage for 28 days at 0, 2, 10 or 50 mg/kg bw per day, suggesting that mitochondria is one of the major targets for TDBP‐TAZTO and one of the pathways that led to toxic effects of TDBP‐TAZTO in such organs.

ERα gene expression was significantly decreased in male Wistar rats (4‐week‐old) daily administered daily with TDBP‐TAZTO for 4 weeks by gavage at doses of 0, 2 or 50 mg/kg bw per day whereas CYP19A1 mRNA expression increased with the dose, but it was not significant. Concomitant histopathological lesions were reported in lung, spleen and kidney according to the dose of exposure and the organ but it was not significant (Zhou et al., 2018).

TDBP‐TAZTO exhibited cell damages and apoptosis in A549 adenocarcinoma human alveolar basal epithelial cells and triggered oxidative stress at concentrations lower than the ones inducing cytotoxicity and apoptosis (Szychowski & Skóra, 2023c). TDBP‐TAZTO also interfered with mTOR‐PPARγ signalling pathway and p62 autophagy pathway, disrupting the cell functioning given the major role of both signalling pathways in cell metabolism and differentiation.

In HepG2 cells, TDBP‐TAZTO (0, 12.5, 25 and 50 μM, cell exposure for 72 h) impaired the cell growth and induced apoptosis via mitochondrial pathway and death receptor signalling pathway (Han et al., 2022).

Bajard et al. (2019) collected in vivo toxicity data on 62 FRs to establish prioritisation categories based on data availability and toxicological concern; the authors classified TDBP‐TAZTO as having ‘scarce toxicological data indicating toxicological concern'.

3.3.7.1.7. Consideration of critical effects and dose–response analysis

Although two studies on genotoxicity were cited in the REACH dossier, there was no information on the protocol or results reported. Thus, the CONTAM Panel was unable to conclude on the genotoxic potential of TDBP‐TAZTO.

Experimental studies of TDBP‐TAZTO in animals include 28‐day studies and a combined repeated dose toxicity study with the reproduction/developmental toxicity screening test, plus several neurotoxicity studies from the literature. Based on the available data, neurotoxicity appears to be the critical effect, with effects on depressive‐related behaviour in the forced swim test and the tail suspension test at the lowest tested dose of 5 mg/kg bw per day (Dong et al., 2015; Ye et al., 2015).

Estimated BMDLs were obtained (see Annex D) but did not meet the acceptability criteria specified in the Guidance on BMD analysis (EFSA Scientific Committee, 2022).

There were no long‐term/carcinogenicity studies identified.

The CONTAM Panel identified the LOAEL of 5 mg/kg bw per day as the Reference Point for TDBP‐TAZTO.

The CONTAM Panel concluded that due to the limitations and uncertainties in the current database, the establishment of a health‐based guidance value for TDBP‐TAZTO was not appropriate. Instead, the MOE approach was used for risk characterisation (EFSA Scientific Committee, 2025).

Usually, an MOE of 100, covering variability with respect to kinetic and dynamic differences between animal species and humans (factor 4 × 2.5 = 10) and within the human population (factor 3.2 × 3.2 = 10), is considered sufficient to conclude that exposure does not raise a health concern.

To allow for the absence of a NOAEL and the necessity of using a LOAEL, the EFSA Guidance on selected default values indicates that an additional uncertainty factor will be needed, the size of which should be determined on a case‐by‐case basis (EFSA Scientific Committee, 2012). ECHA, in its Guidance on information requirements and chemical safety assessment (ECHA, 2012) proposed the application of ‘an assessment factor ranging from 3 (as minimum/majority of cases) to 10 (as maximum/exceptional cases)' with the following considerations: ‘An AF of 3 may be more appropriate for instance in situations, where the effects at the LOAEL are mild, or the LOAEL represents the lower boundary of the exposure range in which the effect is observed. Higher numerical values should be considered in situations where the effects at the LOAEL are severe and irreversible, or the shape of the dose–response curve is shallow or the quality of the study (e.g. group sizes, statistical methods, study design, exposure data) gives rise to uncertainties about the reliability of the identified LOAEL’ (ECHA, 2012). In this instance, since the effects at the LOAEL for TDBP‐TAZTO are mild, the CONTAM Panel considered that a factor of 3 is sufficient to extrapolate from a LOAEL to NOAEL.

According to the EFSA Scientific Committee Guidance on selected default values, an additional factor can be considered in case of deficiencies in the database on a case‐by‐case basis. A default value has not been proposed, as it will be directly dependent on the data set available (EFSA Scientific Committee, 2012). The WHO/IPCS (1994, 1999) recommended a factor of 3 or 5 if there are minor deficiencies in the database and a factor of 10 if there are major deficiencies in the database. There are limitations in the database for TDBP‐TAZTO, such as absence of long‐term/carcinogenicity studies. Taking into account also the sensitivity of the endpoints examined in the critical study, the CONTAM Panel considered an additional factor of 3 to be appropriate for deficiencies in the database.

Overall, the CONTAM Panel considered that MOEs ≥ 900 (100 for inter and intra‐species variability, 3 for LOAEL to NOAEL extrapolation and 3 for the gaps in the toxicological database) do not raise a health concern for non‐neoplastic effects.

The CONTAM Panel could not set a Reference Point for neoplastic effects, since no carcinogenicity data are available. In addition, two studies on genotoxicity were cited in the REACH dossier but there was no information on the protocol or results reported. Thus, it was not possible to conclude on the genotoxic potential of TDBP‐TAZTO.

3.3.7.2. Occurrence data and dietary exposure assessment

No occurrence data on TDBP‐TAZTO in food were submitted to EFSA, and no occurrence or exposure data were identified in the literature, either in the previous Opinion or in the current update. Therefore, no estimation of dietary exposure was possible.

3.3.7.3. Risk characterisation

No risk characterisation was possible, due to the lack of information on potential dietary exposure.

3.3.7.4. Uncertainty analysis

The uncertainty analysis was performed as described in Section  2.8 . Uncertainties affecting the assessment for TDBP‐TAZTO are listed in Table 27.

TABLE 27.

Uncertainties identified and their impact on the outcome of the hazard identification/characterisation and risk characterisation of TDBP‐TAZTO.

Description of the uncertainty		Impact of the uncertainty on the hazard identification and characterisation a
Chemical composition and analytical methods
Dosing and chemical composition, including impurities	Uncertainty associated with the dose in the critical studies used in the risk assessment	0 – Negligible impact. Purity > 97% in both studies used for risk assessment
Hazard identification and characterisation
ADME	ADME in relation to the critical studies	2 – Medium impact. Only two studies identified, with no information on absorption, nor elimination
	Information on transfer rate to animal products	0 – Negligible impact. No studies identified, but no impact on risk assessment.
	Transfer via mother's milk	3 – High impact. No studies identified.
Toxicity studies in experimental animals: endpoints and study design	Limitations in the design of the studies that can result in uncertainties	1 – Medium impact. Six studies identified in the literature (one 14‐day toxicity study in mice, three 28‐day toxicity studies in rats and mice. Two neurotoxicity studies: one in mice (30 days) and one in rats (6 months) + a Combined Repeated Dose Toxicity Study in the REACH dossier with the Reproduction/Developmental Toxicity Screening Test in rats.
	Relevance for humans of the adverse effect	0 – Negligible impact. Neurotoxicity appears to be the critical effect, with effects on depressive‐related behaviour in the forced swim test and the tail suspension test at the lowest tested dose of 5 mg/kg bw per day. These effects are relevant for humans.
Genotoxicity	Uncertainty in the assessment of genotoxicity	3 – High impact. Two studies on genotoxicity were cited in the REACH dossier, but there was no information on the protocol or results reported. Thus, the CONTAM Panel was unable to conclude on the genotoxic potential of TDBP‐TAZTO.
MOA	Uncertainties associated with the data on mode of action	1 – Low impact. A significant set of new studies (13 studies) published since the previous Opinion, using various in vitro models. Four studies explored the toxicity on neuronal and glial cells. Seven recent studies showed the anti‐oestrogenic properties of TDBP‐TAZTO.
Selection of Reference Point	Weaknesses in the critical study	2 – Medium impact. Only two doses and control. Estimated BMDLs did not meet the acceptability criteria specified in the Guidance on BMD analysis
	Weaknesses in non‐critical studies, or absence of studies, and uncertainty about whether the endpoints they tested might have been critical if the weaknesses were not present	2 – Medium impact. No sub‐chronic or chronic/carcinogenicity studies identified.
Description of the uncertainty		Impact on the risk characterisation a
Margin of exposure (MOE)	–	1 – Low impact. MOEs ≥ 900 (100 for inter and intra‐species variability, 3 for LOAEL to NOAEL extrapolation and 3 for the gaps in the toxicological database.

Abbreviations: ADME, absorption, distribution, metabolism, excretion; BMD, benchmark dose; BMDL, benchmark dose lower credible limit; BMR, benchmark response; MOA, mode of action; TK, toxicokinetic.

^a 0 – Uncertainty with negligible impact; 1 – Uncertainty with low impact; 2 – Uncertainty with medium impact; 3 – Uncertainty with high impact. This ordinal scale provides a relative assessment of the impact of the individual uncertainties as an aid to quantifying their combined impact by expert judgement, as described in section 11 of EFSA Scientific Committee (2018).

Uncertainty analysis: Hazard characterisation

The CONTAM Panel identified substantial uncertainties affecting the hazard assessment (Table 27) that did not allow the establishment of a HBGV. Instead, the MOE approach was used, addressing the hazard uncertainties by applying UFs of 100 (to cover for variability with respect to kinetic and dynamic differences between animal species and humans, and within the human population), 3 (to account for the absence of a NOAEL) and 3 (for deficiencies in the database).

Considering the combined impact of the uncertainties and the extent to which they are addressed by the applied UFs, the CONTAM Panel judged with at least 90% certainty (i.e. probability) that, when using the Reference Point of 5 mg/kg bw per day, MOEs ≥ 900 would not raise a health concern for non‐neoplastic effects.

3.3.8. 1,2,5,6‐Tetrabromocyclooctane (TBCO) [CAS number 3194‐57‐8]

TBCO is an additive flame retardant. Technical TBCO consists of 34% α‐HBCO and 66% β‐TBCO, and interconversion of these isomers has been proposed as a possibility (EFSA CONTAM Panel, 2012b).

The physicochemical properties of TBCO are reported in Table 1. TBCO is currently not registered under REACH (see Table 2), and thus its production volume and use within the EU should be lower than one tonne per year.

3.3.8.1. Hazard identification and characterisation

3.3.8.1.1. Toxicokinetics

No information on toxicokinetics was identified in the previous Opinion, nor in the current update.

3.3.8.1.2. Levels in humans

The previous Opinion did not report any study on levels of TBCO in human samples. Since then, one study in human milk and one study in blood/serum from European countries have been identified, both reporting TBCO below the LOD (Appendix E, Table E.1):

• In the study by Čechová et al. (2017) in which human milk samples from Norway (2003–2006, n = 305), the Netherlands (2011–2014, n = 116) and Slovakia (2011–2012, n = 37) were analysed, TBCO was not detected in any sample (LOD = 0.002 ng/g lipid).

• TBCO was measured in 61 serum samples collected in 2013 from a Norwegian cohort and it was reported below the LOD in all cases (LOD = 7.3 pg/serum sample, volume not reported, Tay et al., 2019).

3.3.8.1.3. Toxicity in experimental animals

The previous Opinion contained no information on hazard identification for TCBO but it highlighted TBCO had potential for bioaccumulation (EFSA CONTAM Panel, 2012b).

Since then, the CONTAM Panel has not identified any experimental animal studies in mammalians on effects of TBCO. All available animal research on TBCO was conducted using fish models, specifically Japanese medaka and zebrafish exposed through waterborne or oral exposure, or exposure via maternal transfer to the oocytes (see Section 3.3.8.1.6).

3.3.8.1.4. Computational analysis

The QSAR Toolbox did not identify any experimental data for the Ames test on TBCO (see Annex B). VEGA models predicted negative Ames test results for TBCO with ‘MODERATE reliability’ (three models) and positive Ames test results with ‘LOW reliability’ (one model). These predictions were not further considered due to their limited reliability.

3.3.8.1.5. Observations in humans

No studies were reported in the previous Opinion, nor in the current update.

3.3.8.1.6. Studies in other model systems and mode of action

Studies on TBCO have revealed reproductive, developmental, and cytotoxic effects across multiple biological systems using various exposure pathways. An overview of those studies is given in Appendix G (Table G.9) and a brief summary is provided below.

In fish models, reproductive and endocrine effects are those most studied, with reduced fecundity ranging from 18% to 56.8% observed following oral exposure with dietary levels between 58 and 1034 μg/g feed (Devoy, Raza, Jones, et al., 2023; Devoy, Raza, Kleiner, et al., 2023; Saunders et al., 2015). Impaired oocyte maturation represents another effect, with significant reductions in maturation stimulated by 17α,20β‐dihydroxy‐4‐pregnen‐3‐one, used as maturation inducing hormone (MIH), occurring through both in vitro waterborne exposure at concentrations of 20–200 μg/L and in vivo oral exposure at 100–1000 μg/g food for 21 days (Raza et al., 2023). The reproductive effects extend across generations through maternal transfer, with F1 fecundity reduced by 33–34% at both low (40.6 μg/g) and high (1034 μg/g) oral doses in the F0 generation (Devoy, Raza, Kleiner, et al., 2023). These effects persisted in the F2 generation with a 29.8% reduction, although F3 fish showed complete recovery (Devoy, Raza, Kleiner, et al., 2023).

Developmental effects have been studied in fish. Exposure of Japanese medaka embryos to TBCO via water at 100–1000 μg/L from 2 h post‐fertilisation until 1 day post‐hatch significantly impaired hatching, causing both delayed hatching time and reduced success rates (p < 0.001) (Sun et al., 2016). A background‐colour preference assay was devised to assess vision of the fish larvae. The visual performance by larvae was compromised at relatively low waterborne concentrations, with significant impairment observed in larvae exposed from 2 h post‐fertilisation until 14 days post‐hatch at 10 and 100 μg/L (Sun et al., 2016). Cardiac function was also affected through waterborne exposure from 2 to 5 h post‐fertilisation, with decreased heart rates documented at 100 μg/L and 1000 μg/L (Sun et al., 2016). In zebrafish, waterborne exposure from 1 to 120 h post‐fertilisation at concentrations of 2, 20 or 200 μg/L produced dose‐dependent mortality, spine curvature and uninflated swim bladders, with significant effects at 20 μg/L (Van Essen, Devoy, et al., 2021). Maternal transfer is an exposure pathway in fish. F0 Japanese medaka fed oral diets containing low (74.7 μg/g feed) or high (663 μg/g feed) TBCO doses produced F1 females which showed 26.0% and 56.8% reduced fecundity, respectively, associated with dose‐dependent transfer of TBCO to oocytes (Devoy, Raza, Jones, et al., 2023). Both doses also impaired MIH‐stimulated oocyte maturation with 23.4% and 20.0% reductions (Devoy, Raza, Jones, et al., 2023). This demonstrates that in fish TBCO can be transferred from parent to offspring and that this could cause reproductive impairment in the next generation without direct oral exposure.

Cytotoxicity studies in human neuroblastoma cells using direct in vitro exposure revealed that TBCO reduced cell viability at concentrations of 0.5 μM and above (Shi et al., 2020). This cytotoxicity was accompanied by increased membrane permeability, apoptosis, elevated intracellular calcium concentrations and increased reactive oxygen species production (Shi et al., 2020). In vitro endocrine activity assays indicated multiple modes of action. In porcine testicular cells, TBCO elevated both testosterone and oestradiol at concentrations between 0.03 and 3.0 mg/L (Mankidy et al., 2014). Anti‐androgenic activity reached a maximum of 59% at 1500 mg/L in the yeast androgen screen (YAS) assay, while exposure to 15 mg/L in H295R cell assays produced a 3.3‐fold increase in oestradiol (Saunders et al., 2013). Additionally, a review of the literature and the ToxCast database identified presence of anti‐oestrogenic activity of TBCO (Bajard et al., 2021).

The mechanisms underlying TBCO toxicity appear to be multifaceted and operate at various levels of biological organisation. Disruption of steroidogenesis represents a primary mechanism, evidenced by altered expression of key steroidogenic enzymes including CYP11a, CYP19a, CYP17a and CYP21a (Devoy, Raza, Kleiner, et al., 2023; Mankidy et al., 2014; Saunders et al., 2015). Studies using oral exposure documented downregulation of cholesterol metabolism genes and changes in luteinizing hormone receptor expression (Devoy, Raza, Kleiner, et al., 2023; Saunders et al., 2015). Interestingly, some studies found no consistent changes in plasma oestradiol or vitellogenin levels despite significant transcript alterations (Raza et al., 2023), suggesting post‐transcriptional regulation or compensatory mechanisms.

Impaired oocyte maturation in fish may have involved disruption of insulin‐like growth factor signalling pathways and altered microRNA expression that regulates these pathways, as observed in zebrafish following oral exposure for 14 days (Van Essen, Alcaraz, et al., 2021). Notably, the effects on MIH‐stimulated maturation occurred independently of plasma hormone levels in both in vitro and in vivo exposure scenarios (Raza et al., 2023; Van Essen, Alcaraz, et al., 2021), suggesting direct effects on the oocyte maturation machinery independent of systemic hormonal changes.

Developmental toxicity mechanisms, observed through waterborne embryonic exposure of fish, involve downregulation of ribosomal proteins, which affects protein biosynthesis and impairs embryo development and hatching pathways (Sun et al., 2016). Specific developmental processes appear to be targeted through altered expression of genes involved in eye development and phototransduction, including atoh7, col15a1b, recoverin and various opsin genes (Sun et al., 2016). Cardiac development and function are compromised through changes in myosin isoforms and cytochrome c oxidase expression, while calcium signalling pathways are disrupted through altered expression of genes such as cacng6a (Sun et al., 2016). These transcriptomics changes were associated with apical effects which are summarised above.

The hypothalamus‐pituitary‐gonadal‐liver axis in fish was affected in an organ‐specific and dose‐dependent manner following oral dietary exposure (Saunders et al., 2015). Lower doses particularly induced upregulation of oestrogen‐responsive genes including vitellogenin II, choriogenin H and ERα, while higher doses affected regulatory feedback mechanisms throughout the axis (Saunders et al., 2015).

In summary, there are few studies on mode of action of TBCO in mammalian systems. In fish, visual performance impairment represents the most sensitive endpoint investigated, occurring at just 10 μg/L through waterborne exposure during early development. Developmental abnormalities in zebrafish (mortality, spine curvature, uninflated swim bladders) showed significant effects at 20 μg/L waterborne exposure. Reduced fecundity in the same test system appeared at relatively low dietary doses of 40.6–58 μg/g feed, with reductions ranging from 18 to 56.8%. In terms of cellular effects, cytotoxicity in human neuroblastoma cells occurred at 0.5 μM, accompanied by membrane damage, apoptosis and oxidative stress. TBCO demonstrates multifaceted toxicity through disruption of steroidogenesis, affecting key enzyme expression (CYP11a, CYP19a, CYP17a, CYP21a) and showing both anti‐androgenic and anti‐oestrogenic activities. Developmental effects involve downregulation of ribosomal proteins and genes controlling eye development, cardiac function and calcium signalling. In fish, maternal transfer occurs, with reproductive impairment persisting through F2 generations before F3 recovery.

3.3.8.1.7. Consideration of critical effects and dose–response analysis

The CONTAM Panel could not identify any toxicokinetic, toxicity in experimental mammalian species or epidemiological studies on TBCO.

It was therefore not possible to identify a Reference Point for TBCO, and it was not possible to conclude on its genotoxic potential.

3.3.8.2. Occurrence data and dietary exposure assessment

No occurrence data on TBCO in food were submitted to EFSA, and no occurrence or exposure data were identified in the literature, either in the previous Opinion or in the current update. Therefore, no estimation of the dietary exposure was possible.

No estimation of the exposure of breastfed infants was possible as the studies from European countries identified in the literature reported all values < LOD.

3.3.8.3. Risk characterisation

No risk characterisation was possible due to the lack of a Reference Point and lack of information on potential dietary exposure.

3.3.9. 5,6‐Dibromo‐1,10,11,12,13,13‐hexachloro‐11‐tricyclo[8.2.1.0^2,9]tridecane (DBHCTD) [CAS number 51936‐55‐1]

DBHCTD is a tricycloaliphatic compound with both bromine and chlorine substituents. It can form 10 stereoisomers among which six form pairs of enantiomers (EFSA CONTAM Panel, 2012b).

The physicochemical properties of DBHCTD are reported in Table 1. DBHCTD is currently not registered under REACH (see Table 2), and thus its production volume and use within the EU should be lower than one tonne per year.

3.3.9.1. Hazard identification and characterisation

3.3.9.1.1. Toxicokinetics

No information on toxicokinetics was identified in the previous Opinion, nor in the current update.

3.3.9.1.2. Levels in humans

The previous Opinion did not report any study on levels of DBHCTD in human samples. Since then, some studies have been identified in human milk and blood/serum from European countries, all reporting levels of DBHCTD below the LOD. These studies are briefly described below and details can be found in Appendix E (Table E.1).

Regarding human milk,

• In the study by Čechová et al. (2017) in which human milk samples from Norway (2003–2006, n = 305), the Netherlands (2011–2014, n = 116) and Slovakia (2011–2012, n = 37) were analysed, DBHCTD was not detected in any sample (LOD = 0.004 ng/g lipid).

Regarding blood/serum samples,

• Cequier et al. (2013) did not detect DBHCTD above the LOD in any of the 10 serum samples from Norway analysed (LOD = 5.4 pg/mL), or in any of the 46 serum samples from women from a Norwegian mother–child cohort established in 2012 (Cequier et al., 2015).

• DBHCTD was also reported below the LOD in 61 serum samples collected in 2013 from a Norwegian cohort (LOD = 63 pg/serum sample, volume not reported, Tay et al., 2019), and in 24 matched serum samples from mothers and their toddlers collected in Sweden between 2009 and 2010 (LOD = 7.1 pg/25 μL, Sahlström et al., 2014).

3.3.9.1.3. Toxicity in experimental animals

Studies considered in the previous EFSA assessment

No data on toxicity in experimental mammalian species were identified in the previous Opinion.

Studies published since the previous EFSA assessment

Since then, the comparative study of a number of BFRs conducted by the NTP has been published. DBHCTD was administered for 5 days via oral gavage to groups of 6 male Harlan Sprague–Dawley rats at doses of 0, 0.05, 0.54, 5.4, 54 and 541 mg/kg bw per day. After the 5‐day exposure, the dose levels of 54 and 541 mg/kg bw resulted in centrilobular hypertrophy and increased absolute liver weight (NTP, 2019b; Shockley et al., 2020a). Transcriptome changes associated with the Nrf2 pathway, metabolic and/or liver disease were reported at the top dose (Shockley et al., 2020a, 2020b).

3.3.9.1.4. Computational analysis

The QSAR Toolbox did not identify any experimental data for the Ames test on DBHCTD (see Annex B).

VEGA models predicted negative Ames test results for DBHCTD with ‘GOOD reliability’ (two models) and with ‘MODERATE reliability’ (one model). The Panel critically reviewed the predictions and considered them valid mainly due to the presence of relevant analogues known by the model with experimental and predicted values concordant with the prediction for the target compound (see Annex B).

3.3.9.1.5. Observations in humans

No studies were identified in the previous Opinion. Since then, one publication was identified, pertaining to a population in China. In Appendix F an overview of the epidemiological studies identified reporting on the BFRs considered in this Opinion is shown, including the one on DBHCTD.

Guo et al. (2018) studied the association of serum DBHCTD levels with thyroid hormone levels in 10‐year‐old students living near a petrochemical complex in South China (for details see description in Section 3.2.4.2). Serum DBHCTD levels were statistically significantly associated with lower levels of serum T3, T4, FT4 and TSH.

3.3.9.1.6. Studies in other model systems and mode of action

Bajard et al. (2019) collected in vivo toxicity data on 62 FRs to establish prioritisation categories based on data availability and toxicological concern; the authors classified DBHCTD as having ‘insufficient toxicological data to evaluate the toxicological concern'.

3.3.9.1.7. Consideration of critical effects and dose–response analysis

No experimental data on genotoxicity were identified. Computational analysis indicates that DBHCTD does not induce mutations in S. Typhimurium, but no information is available on clastogenicity or aneugenicity. It is therefore not possible to conclude on genotoxicity overall.

One 5‐day‐study in rats indicates possible effects on the liver at doses of 54 mg/kg bw per day and higher. No other toxicity data were available.

In view of the absence of toxicological studies except for the single 5‐day study, no Reference Point could be identified.

3.3.9.2. Occurrence data and dietary exposure assessment

No occurrence data on DBHCTD in food were submitted to EFSA. In the literature, one study included DBHCTD, but it was not detected above the LOD (Fernandes et al., 2016, see Section 3.1.2).

No exposure data were identified in the literature, either in the previous Opinion or in the current update. Therefore, no estimation of the dietary exposure was possible.

No estimation of the exposure of breastfed infants was possible as the studies from European countries identified in the literature reported all values < LOD.

3.3.9.3. Risk characterisation

No risk characterisation was possible due to the lack of a Reference Point and lack of information on potential dietary exposure.

3.3.10. 1,2,3,4,7,7‐Hexachloro‐5‐(2,3,4,5‐tetrabromophenyl)‐bicyclo[2.2.1]hept‐2‐ene (HCTBPH) [CAS number 34571‐16‐9]

HCTBPH belongs to the group of ‘dechlorane’ chemicals. It is a tricyclic compound with three chiral carbons, leading to the formation of three stereoisomers each one forming enantiomeric pairs. Dechloranes were widely used as substitutes for Mirex and used primarily as a pesticide but also as an additive flame retardant in plastics, rubber, paint, paper and electrical equipment (EFSA CONTAM Panel, 2012b).

The physicochemical properties of HCTBPH are reported in Table 1. HCTBPH is currently not registered under REACH (see Table 2), and thus its production volume and use within the EU should be lower than one tonne per year.

3.3.10.1. Hazard identification and characterisation

3.3.10.1.1. Toxicokinetics

No information on toxicokinetics was identified in the previous Opinion, nor in the current update.

3.3.10.1.2. Levels in humans

No studies were identified in the previous Opinion, nor in the current update.

3.3.10.1.3. Toxicity in experimental animals

No data on toxicity in experimental mammalian species were identified in the previous Opinion, nor in the current update.

3.3.10.1.4. Computational analysis

The QSAR Toolbox did not identify any experimental data for the Ames test on HCTBPH (see Annex B). VEGA models predicted negative Ames test results for HCTBPH with ‘MODERATE reliability’ (three models) and positive Ames test results with ‘LOW reliability’ (one model). These predictions were not further considered due to their limited reliability.

3.3.10.1.5. Observations in humans

No studies were reported in the previous Opinion, nor in the current update.

3.3.10.1.6. Studies in other model systems and mode of action

Bajard et al. (2021) reviewed the in vitro activity of 52 BFRs towards five nuclear receptors associated with reproductive outcomes, retrieving data from published articles and ToxCast/Tox21 programmes; the authors reported that no data were retrieved for HCTBPH, and that the opensource platform Endocrine Disruptome model predicted medium probability of binding to the antagonist conformation of the AR.

3.3.10.1.7. Consideration of critical effects and dose–response analysis

The CONTAM Panel could not identify any toxicokinetic, toxicity or epidemiological studies on HCTBPH.

It was therefore not possible to identify a Reference Point for HCTBPH, and it was not possible to conclude on its genotoxic potential.

3.3.10.2. Occurrence data and dietary exposure assessment

No occurrence data on HCTBPH in food were submitted to EFSA. In the literature, one study included HCTBPH, but it was not detected above the LOD (Fernandes et al., 2016, see Section  3.1.2 ).

No exposure data were identified in the literature, either in the previous Opinion or in the current update. Therefore, no estimation of the dietary exposure was possible.

3.3.10.3. Risk characterisation

No risk characterisation was possible due to the lack of a Reference Point and lack of information on potential dietary exposure.

3.3.11. Pentabromobenzyl acrylate (PBB‐Acr) [CAS number 59447‐55‐1]

PBB‐Acr is reactive flame retardant and as with other reactive BFRs is found at relatively lower levels in the environment. PBB‐Acr is a polar pentabromobenzyl alcohol derivative that is reactive, and suitable for polymerisation reactions, via the acrylate group (EFSA CONTAM Panel, 2012b).

The physicochemical properties of PBB‐Acr are reported in Table 1. PBB‐Acr is currently registered under REACH (100–1000 tonnes per annum) (see Table 2).

3.3.11.1. Hazard identification and characterisation

3.3.11.1.1. Toxicokinetics

No information on toxicokinetics was identified in the previous Opinion, nor in the current update.

3.3.11.1.2. Levels in humans

No studies were identified in the previous Opinion, nor in the current update.

3.3.11.1.3. Toxicity in experimental animals

Studies considered in the previous EFSA assessment

No data on toxicity in experimental mammalian species were identified in the previous Opinion (EFSA CONTAM Panel, 2012b).

Studies published since the previous EFSA assessment

Since then, one repeated dose toxicity study was described in the REACH dossier ³⁷ , as submitted by the registrant(s). The CONTAM Panel did not have access to the original study report. The REACH dossier reported that:

• Sprague–Dawley rats were exposed by gavage for 28 days to 0, 15, 150 or 1000 mg/kg bw per day of the technical product FR‐1025M (containing 95% PBB‐Acr and 5% PBB‐OH). Some animals from the control and highest dose groups were retained for a minimum two‐week recovery period. Dose‐related clinical signs were observed, e.g. increased salivation, alopecia and greasy fur. Animals had a lower bodyweight gain during the treatment period at 1000 mg/kg bw per day and recovered after. This reflects the lower food consumption. Increased water consumption was also seen in both sexes receiving 1000 mg/kg bw per day. Small disturbances in haematological parameters were recorded following treatment for male and female rats at 1000 mg/kg bw per day. For rats receiving 150 mg/kg bw per day, slightly lower haemoglobin content, lower total red blood cell counts and slightly shorter thrombotest times were seen for males only. Small electrolyte disturbances were seen at the highest dose in both sexes and triglyceride levels were notably increased and urea nitrogen level were decreased in females. At 150 mg/kg bw per day, most of these disturbances were also noted but to a lesser degree. Urinary volume was increased and the specific gravity was lower for rats of both sexes treated at 150 and 1000 mg/kg bw per day. Increased absolute kidney weights was noted for female rats at 150 and 1000 mg/kg bw per day. Macroscopic changes at 1000 mg/kg bw per day include enlarged kidneys for five females, irregular cortical scarring of the kidneys for one male and two females and alopecia for three females; two females treated at 150 mg/kg bw per day had enlarged kidneys. Renal tubular lesions consisting of cortical wedge and ray‐shaped tubular epithelial basophilia were noted at 1000 mg/kg bw per day in four males and five females. The CONTAM Panel identified a NOAEL of 15 mg/kg bw per day.

Genotoxicity

Since the previous Opinion, information on genotoxicity studies was available in the REACH dossier, as submitted by the registrant(s). The CONTAM Panel did not have access to the original data. The REACH dossier reported that:

• FR‐1025M (containing 95% PBB‐Acr and 5% pentabromobenzyl alcohol) was reported to be negative in reverse mutation assays (plate incorporation assays according to OECD TG471) in S. Typhimurium TA98, TA100, TA1535, TA1537, TA1538 and E. coli WP2 and WP2uvrA in the presence and absence of S9 mix from Aroclor‐induced rat liver.

• In an in vitro chromosome aberration test in CHO cells (OECD TG473), FR‐1025M (containing 95% PBB‐Acr and 5% pentabromobenzyl alcohol) was reported to be negative in the presence and absence of S9 mix.

3.3.11.1.4. Computational analysis

The QSAR Toolbox identified experimental data for the Ames test on PBB‐Acr (see above and Annex B).

VEGA models predicted negative Ames test results for PBB‐Acr with ‘GOOD reliability’ (three models) and with ‘LOW reliability’ (one model). The Panel critically reviewed the predictions and considered them valid mainly due to the presence of relevant analogues known by the model with experimental and predicted values concordant with the prediction for the target compound (see Annex B). The CONTAM Panel noted that this negative prediction was consistent with the available experimental data.

3.3.11.1.5. Observations in humans

No studies were reported in the previous Opinion, nor in the current update.

3.3.11.1.6. Mode of action

Ezechiáš et al. (2012) used yeast reporter‐gene assays to determine the potential to interfere with oestrogenic and androgenic pathways, and reported that PBB‐Acr exhibited slight anti‐androgenic activity, as it was able to inhibit testosterone activation of AR with a nominal IC50 of 3.0 μM.

Bajard et al. (2021) reviewed the in vitro activity of 52 BFRs towards five nuclear receptors associated with reproductive outcomes, retrieving data from published articles and ToxCast/Tox21 programmes; the authors reported that no data were retrieved for PBB‐Acr, and that the opensource platform Endocrine Disruptome model predicted medium probability of binding to the antagonist conformation of the AR. Bajard et al. (2019) collected in vivo toxicity data on 62 FRs to establish prioritisation categories based on data availability and toxicological concern; the authors classified PBB‐Acr as having ‘insufficient toxicological data to evaluate the toxicological concern'.

3.3.11.1.7. Consideration of critical effects and dose–response

Changes in clinical signs, body weight gain, haematology, biochemical parameters and macroscopic and histopathological changes in the kidney were observed in rats exposed for 28 days to FR‐1025M (containing 95% PBB‐Acr and 5% pentabromobenzyl alcohol) at 150 mg/kg bw per day. This was the only toxicological study available.

According to the data reported in the REACH dossier, FR‐1025M (containing 95% PBB‐Acr and 5% pentabromobenzyl alcohol) was not mutagenic in bacteria and not clastogenic in CHO cells in the presence and absence of metabolic activation. The aneugenic potential was not investigated. Based on this in vitro information, the CONTAM Panel concluded that PBB‐Acr is not genotoxic.

In the absence of long‐term toxicity data, reproductive/developmental toxicity or neurodevelopmental toxicity studies, a Reference Point could not be identified.

3.3.11.2. Occurrence data and dietary exposure assessment

No occurrence data of PBB‐Acr in food were submitted to EFSA, but some studies in the scientific literature were identified (see Section 3.1.2 for an overview).

No exposure data were identified in the literature, either in the previous Opinion or in the current update. Therefore, no estimation of the dietary exposure was possible.

3.3.11.3. Risk characterisation

No risk characterisation was possible due to the lack of a Reference Point and lack of information on potential dietary exposure.

3.3.11.4. Uncertainty analysis

Due to the limited data on hazard, and lack of occurrence and dietary exposure data for PBB‐Acr, uncertainty analysis was performed only for the conclusion on genotoxicity of PBB‐Acr using the methods described in Section 2.8.

Information on genotoxicity was retrieved from the REACH dossier, reporting negative results for all the necessary studies. No sources of uncertainty were indicated, but the CONTAM Panel did not have access to the original data. Taking this into account, the CONTAM Panel judged with > 95% certainty (i.e. probability) that PBB‐Acr is not genotoxic.

3.3.12. Tribromoneopentyl alcohol (TBNPA) [CAS number 1522‐92‐5]

The previous Opinion (EFSA CONTAM Panel, 2012b) reported that TBNPA is used together with tris(chloroalkyl) phosphate to form a base for the technical product SaFRon 6600 (Weil & Levchik, 2009). As many as 79 commercial sources of TBNPA have been indicated. It has a relatively strong resistance to biodegradation (0%–7% degradation after 28 days).

The physicochemical properties of TBNPA are reported in Table 1. TBNPA is currently not registered under REACH (see Table 2), and thus its production volume and use within the EU should be lower than one tonne per year.

3.3.12.1. Hazard identification and characterisation

3.3.12.1.1. Toxicokinetics

No information on toxicokinetics was identified in the previous Opinion, nor in the current update.

3.3.12.1.2. Levels in humans

No studies were identified in the previous Opinion, nor in the current update.

3.3.12.1.3. Toxicity in experimental animals

Studies considered in the previous EFSA assessment

No data on toxicity in experimental mammalian species were identified in the previous Opinion.

Regarding genotoxicity, the previous Opinion reported that TBNPA was tested in the Ames test using S. Typhimurium TA98, TA100, TA1535 and TA1537 in presence and absence of metabolic activation (10% Aroclor 1254 induced rat or hamster liver S9). TBNPA was not mutagenic in the absence and in the presence of rat liver S9; whereas it was positive in TA100 and TA1535 ³⁸ in the presence of hamster liver S9 (EFSA CONTAM Panel, 2012b; Mortelmans et al., 1986).

Studies published since the previous EFSA assessment

Since then, no studies were identified in the literature search, but the ECHA CLH Report (2019) and the ECHA RAC Opinion (2020) proposing harmonised classification and labelling at EU level of TBNPA, did report several studies. The CONTAM Panel did not have access to the original data. The ECHA CLH Report and the ECHA RAC Opinion reported that:

• −In a 30‐day oral study in Sprague–Dawley rats with TBNPA (purity 98%) at 0, 10, 30, 100 or 300 mg/kg bw per day (Anonymous, 1973), kidney damage and urinary bladder hyperplasia were observed in males from 100 mg/kg bw per day. Additional treatment‐related effects were: increase in serum urea nitrogen content in males receiving 300 mg/kg bw per day and decrease in serum glutamic‐pyruvic transaminase (SGPT) in males at the two top doses. No changes were noted in any of the females. The reported NOAELs were 30 mg/kg bw per day for males based on the kidney and urinary bladder effects and 300 mg/kg bw per day for females.
• −In a 14‐day study, Crl:CD(SD) rats were administered TBNPA (purity 98.4%) by gavage at doses of 0, 100, 300 or 1000 mg/kg bw per day (Anonymous, 2011). Enlargement of the liver with associated dark areas was seen in one female at 1000 mg/kg bw per day. Urine staining occurred in males and females in the top dose group. Males receiving the top dose were killed early on day 4 of treatment for animal welfare reasons. The reported NOAEL was 300 mg/kg bw per day.
• −In a 28‐day study, Sprague–Dawley rats were exposed by gavage to TBNPA (purity 97%) at 0, 30, 150 or 500 mg/kg bw per day and the animals were allowed to recover for 14 days (Anonymous, 2015). Increased liver weight (predominantly at ≥ 150 mg/kg bw per day), and slight minimal centrilobular hypertrophy were observed. Slightly higher kidney weights were observed in females in the low dose group and in males in the medium dose group. All findings showed full recovery, with the exception of kidney weights which remained slightly elevated at the end of the recovery period for males in the top dose group. A dose‐related increase in urinary volume and slightly high total protein and glucose output in males was seen. All of these findings showed full recovery. At the top dose group, transient slightly low sodium concentration and slightly high potassium concentration were noted. No treatment‐related changes in sperm count and motility were observed. Vaginal lavages which were taken early morning during the 3‐week period from all females, prior to termination of the animals showed no treatment‐related changes in the oestrus cycle. In addition, there were no dose‐related changes in weight of ovaries, seminal vesicles, testis, ureter, uterus or vagina in comparison to control animals. The reported NOAEL was 500 mg/kg bw per day.
• −In a 90‐day study, TBNPA (97% pure) was administered by gavage to Sprague–Dawley rats at doses of 0, 50, 150 or 450 mg/kg bw per day and the animals were allowed to recover for 28 days (Anonymous, 2020). There was an increase in blood urea nitrogen at 450 mg/kg bw per day and in creatinine at ≥ 150 mg/kg bw per day in males. This increase was associated with microscopic findings of increased eosinophilic droplets in the tubular epithelium consisting of tubular casts in kidneys. Microscopic changes were noted in kidneys and urinary bladder in males. In kidneys, increased eosinophilic droplets were noted in the tubular epithelium in the cortex of six males treated at 150 mg/kg bw per day and all males treated at 450 mg/kg bw per day but not in females. A single incidence of papillary necrosis was also observed at 450 mg/kg bw per day males. Diffuse epithelial hyperplasia was noted in urinary bladder of six males at 150 mg/kg bw per day and all males and one female at 450 mg/kg bw per day. All the changes reversed at the end of recovery period. The reported NOAEL was 50 mg/kg bw per day.

The ECHA CLH Report (2019) also reported on a developmental toxicity study in Sprague–Dawley rats:

Pregnant females were exposed by gavage from GD6‐19 to TBNPA (FR‐513, purity 97.6%) at doses of 0, 100, 300, 500/1000 mg/kg bw per day. The highest dose was reduced from 1000 to 500 mg/kg bw per day due to post dosing toxicity. Two animals were killed due to animal welfare reasons. Gravid uterine weight in the dosed animals was lower. No clear findings of developmental toxicity were observed. Minor effects on ossification in the medium and high‐dose groups were within the range of the historical control data (Anonymous, 2016).

No multi‐generation reproductive toxicity study has been retrieved. However, in the 28‐day oral study in rats with a 14‐day recovery period described above, some relevant reproduction parameters were investigated but no indications of effects on sexual function and fertility have been identified (Anonymous, 2015; as cited in ECHA CLH Report, 2019).

No long‐term or carcinogenicity studies on TBPNA alone have been identified.

The CONTAM Panel noted that TBNPA was one of the impurities (about 7%) of the technical products containing DBNPG (see Section 3.2.6). The ECHA CLH Report (2019) included further information as follows: TBNPA and DBNPG are classified within the Small Brominated Linear and Branched Alkyl Alcohols group, comprising 61 compounds. These substances were identified by several (Q)SAR models, including the OECD QSAR Toolbox and exhibit highly similar chemical structures. In terms of chemical reactivity, TBNPA and DBNPG share several characteristics, such as comparable electrophilic properties, the presence of bromine and hydroxyl groups attached to primary carbon atoms and the potential for radical activation. Additionally, the aliphatic halogen moiety serves as a structural alert for both carcinogenicity and mutagenicity. Consequently, TBNPA and DBNPG were found to cluster within the same (Q)SAR‐based groupings associated with genotoxic and carcinogenic potential.

Genotoxicity

Since the previous Opinion, the ECHA CLH Report (2019) and the ECHA RAC Opinion (2020) reported on several studies that are summarised below. The CONTAM Panel did not have access to the original data. The ECHA CLH Report and the ECHA RAC Opinion (2020) reported that:

• − TBNPA (FR‐513, purity 98%) was negative in a reverse mutation assay in S. Typhimurium TA1535, TA1537, TA 98 and TA 100 in the absence of S9 mix, but was positive in the presence of S9 mix (Aroclor‐1254 induced hamster liver S9 mix) in TA1535 and TA100 (Anonymous, 1996).
• − TBNPA (up to 500 μg/mL, FR‐513, purity 97%) was negative in an in vitro gene mutation test in mouse lymphoma L5178Y cells (TK locus) in the absence of S9 mix, whereas it was positive in the presence of S9 mix. Cytotoxicity was seen at 333 μg/mL and above with and without metabolic activation (Anonymous, 2004).
• −The ability of TBNPA (100–2000 μg/mL, FR‐513, purity 97%) to induce chromosomal aberrations was tested in vitro in human peripheral lymphocytes in the absence and the presence of metabolic activation (Aroclor‐1254 induced rat liver S9 mix). TBNPA was positive in the presence of S9 and in the absence of S9 at 1000 μg/mL. Cytotoxicity was observed as low as at 100 μg/mL, with S9 (Anonymous, 2004).
• −In vivo, TBNPA (FR‐513, purity 98.1%) did not induce micronuclei in bone marrow cells of NMRI mice exposed by the oral route up to 300 mg/kg bw, however, it was not demonstrated that the bone marrow had been exposed (Anonymous, 2007).
• − TBNPA (FR‐513, purity NR) was negative in vivo in a UDS test in rat hepatocytes exposed (2 or 16h) by gavage up to 2000 mg/kg bw (Anonymous, 2007).

TBNPA has been classified by the ECHA's RAC as a ‘germ cell mutagen Category 2, H341: Suspected of causing genetic defects’. Based on read‐across from DBNPG, ECHA's RAC supported the classification of TBNPA as ‘Carc 1B, H350: May cause cancer’.

3.3.12.1.4. Computational analysis

The QSAR Toolbox identified experimental data for the Ames test on TBNPA (see above and Annex B).

VEGA models predicted positive Ames test results for TBNPA with ‘GOOD reliability’ (four models). The Panel critically reviewed the predictions and considered them valid mainly due to the presence of relevant analogues known by the model with experimental and predicted values concordant with the prediction for the target compound (see Annex B). The CONTAM Panel noted that this positive prediction was consistent with the available experimental data.

The QSAR Toolbox profilers provided the following information on the structural alerts and predicted mode of toxicity: TBNPA triggers the ‘Aliphatic halides’ alert according to the ‘DNA binding by OECD’ profiler, and the ‘Aliphatic halogens’ alert according to the ‘in vitro mutagenicity (Ames test) by ISS’ profiler because of the aliphatic Br atom (see Figure 2). In the QSAR Toolbox, this alert is associated with DNA alkylation by nucleophilic substitution. The QSAR Toolbox profiling results are provided in Annex B, while additional information on the alerts triggered can be consulted within the software.

FIGURE 2.

Chemical structures of the other 21 BFRs included in the 2012 Opinion: Decabromodiphenyl ethane (DBDPE); 4‐(1,2‐Dibromoethyl)‐1,2‐dibromocyclohexane (DBE‐DBCH); 5,6‐Dibromo‐1,10,11,12,13,13‐hexachloro‐11‐tricyclo[8.2.1.0^2,9]tridecene (DBHCTD); 1,2‐Bis(2,4,6‐tribromophenoxy)ethane (BTBPE); Hexabromobenzene (HBB); 1,2,3,4,7,7‐Hexachloro‐5‐(2,3,4,5‐tetrabromophenyl)‐bicyclo[2.2.1]hept‐2‐ene (HCTBPH); Pentabromobenzyl acrylate (PBB‐Acr); Pentabromoethylbenzene (PBEB); Pentabromotoluene (PBT); Tribromoneopentyl alcohol (TBNPA); 1,3,5‐Tris(2,3‐dibromopropyl)‐1,3,5‐triazine‐2,4,6‐trione (TDBP‐TAZTO); 1,2,5,6‐Tetrabromocyclooctane (TBCO); 2,3,5,6‐Tetrabromo‐p‐xylene (TBX); 1,3‐Bis(2,3‐dibromopropyl)‐5‐‐(2‐propen‐1‐yl)‐1,3,5‐triazine‐2,4,6(1H,3H,5H)‐trione (BDBP‐TAZTO); 1‐(2,3‐Dibromopropyl)‐3,5‐diallyl‐1,3,5‐triazine‐2,4,6(1H,3H,5H)‐trione (DBP‐TAZTO); Dibromostyrene (DBS); 2‐(2‐Hydroxyethoxy)ethyl 2‐hydroxypropyl 3,4,5,6‐tetrabromophthalate (HEEHP‐TEBP); Tetradecabromo‐1,4‐diphenoxybenzene (4’‐PeBPO‐BDE208); Tris(tribromoneopentyl) phosphate (TTBNPP); 2,4,6‐Tris(2,4,6‐tribromophenoxy)‐1,3,5‐triazine (TTBP‐TAZ); Octabromotrimethylphenyl indane (OBTMPI).

The CONTAM Panel noted that the computational analysis supports the available experimental data.

3.3.12.1.5. Observations in humans

No studies were reported in the previous Opinion, nor in the current update.

3.3.12.1.6. Mode of action

Bajard et al. (2019) collected in vivo toxicity data on 62 FRs to establish prioritisation categories based on data availability and toxicological concern; the authors classified TBNPA as having ‘insufficient toxicological data to evaluate the toxicological concern'. The same authors reviewed the in vitro activity of 52 BFRs towards five nuclear receptors associated with reproductive outcomes, retrieving data from published articles and ToxCast/Tox21 programmes; the authors reported that there are studies showing that TBNPA has some antagonistic activity towards the AR and PR, and agonistic activity towards AHR (Bajard et al., 2021).

3.3.12.1.7. Consideration of critical effects and dose–response analysis

No toxicokinetic data were identified. A limited number of short‐term toxicity studies were available in the ECHA CLH Report (2019) and ECHA RAC Opinion (2020). Kidney damage and urinary bladder hyperplasia were observed in a 30‐day oral study in rats with TPNPA. In a 28‐day study in rats exposed by gavage to TBNPA increased liver weight and slight minimal centrilobular hypertrophy were observed as well as slightly higher kidneys. A dose‐related increase in urinary volume and slightly high total protein and glucose output. Administration of TBNPA to rats by gavage for 90 days induced an increase in blood urea nitrogen and in creatinine. This increase was associated with microscopic findings of increased eosinophilic droplets in the tubular epithelium consisting of tubular casts in kidneys and diffuse epithelial hyperplasia in the urinary bladder in males. The NOAEL was 50 mg/kg bw per day.

No clear findings of developmental toxicity were observed in a study in rats where pregnant females were exposed by gavage from GD6‐19 to TBNPA up to doses of 500/1000 mg/kg bw per day.

No multi‐generation reproductive toxicity study has been retrieved. However, in a 28‐day oral study in rats with a 14‐day recovery period, where some relevant reproduction parameters were investigated, no indications of effects on sexual function and fertility have been identified.

No long‐term or carcinogenicity studies on TBPNA have been identified.

According to the data reported in the ECHA CLH Report (2019) and the ECHA RAC Opinion (2020), TBNPA is mutagenic in vitro and clastogenic. Based on this information, and in the absence of conclusive negative results for in vivo genotoxicity data, the CONTAM Panel concluded that it is likely that TBNPA is genotoxic. The carcinogenic potential of TBNPA is inferred from the positive in vitro mutagenicity and clastogenicity, together with predictions based on structural similarity/read‐across considerations with DBNPG, which is genotoxic and carcinogenic (see Section 3.2.6).

The CONTAM Panel noted that the limited available data provide evidence of liver, kidney and urinary bladder toxicity but, particularly in the absence of carcinogenicity studies, the data are insufficient to identify a Reference Point.

3.3.12.2. Occurrence data and dietary exposure assessment

No occurrence data on TBNPA in food were submitted to EFSA, and no occurrence or exposure data were identified in the literature, either in the previous Opinion or in the current update. Therefore, no estimation of the dietary exposure was possible.

3.3.12.3. Risk characterisation

No risk characterisation was possible due to the lack of a Reference Point and lack of information on potential dietary exposure.

3.3.12.4. Uncertainty analysis

Due to the lack of data on occurrence and exposure for TBNPA, uncertainty analysis was performed only for the conclusion on genotoxicity and carcinogenicity of TBNPA using the methods described in Section 2.8.

Based on the positive in vitro mutagenicity and clastogenicity of TBNPA in the absence of conclusive negative results for in vivo genotoxicity data, the CONTAM Panel concluded that it is likely that TBNPA is genotoxic and carcinogenic. Supporting this, the Panel noted the structural similarity with DBNPG which is genotoxic and carcinogenic.

Based on the overall evidence, and taking into account the associated uncertainties, the CONTAM Panel judged with 95%–99% certainty (i.e. probability) that TBNPA is genotoxic and carcinogenic.

3.3.13. 2,3,5,6‐Tetrabromo‐p‐xylene (TBX) [CAS number 23488‐38‐2]

TBX is an environmentally stable aromatic compound with the potential to undergo reductive debromination (EFSA CONTAM Panel, 2012b). TBX is a non‐polymeric additive with a QSAR prediction to have the properties ‘(very)persistent’, ‘very bioaccumulative’ and ‘toxic’.²⁸

The physicochemical properties of TBX are reported in Table 1. TBX is currently not registered under REACH (see Table 2), and thus its production volume and use within the EU should be lower than one tonne per year.

3.3.13.1. Hazard identification and characterisation

3.3.13.1.1. Toxicokinetics

No information on toxicokinetics was identified in the previous Opinion, nor in the current update.

3.3.13.1.2. Levels in humans

The previous Opinion did not report any study on levels of TBX in human samples. Since then, five studies have been identified in human milk and serum samples:

Regarding human milk,

• In Sweden, Sahlström et al. (2015) analysed 2 pooled human milk samples collected in 2009–2010. TBX was not detected in any of the two pooled samples (LOD < 0.5 pg/g ww).

• Tao et al. (2017) analysed human milk samples collected in the UK in 2010 (n = 25) and 2014–2015 (n = 10) and TBX was reported to be below the LOD (not reported).

• In the study by Čechová et al. (2017), TBX was analysed in human milk samples from Norway (2003–2006, n = 305), the Netherlands (2011–2014, n = 116) and Slovakia (2011–2012, n = 37). It was not detected in samples from Slovakia (2011–2012, n = 37, LOD = 0.001 ng/g lipid), while detection frequencies of 15% and 69% were reported for samples from Norway and the Netherlands, respectively, with mean and P95 concentrations across the two countries ranging from 0.019 to 0.035 ng/g lipid and from 0.038 to 0.128 ng/g lipid, respectively.

Regarding levels in blood/serum,

• TBX was not detected above the LOD (0.9 pg/25 μL) in any of the 24 matched serum samples from mothers and their toddlers collected in Sweden between 2009 and 2010 (Sahlström et al., 2014).

• Another study reported TBX below the LOD in 61 serum samples collected in 2013 from a Norwegian cohort (LOD = 14 pg/serum sample, volume not reported, Tay et al., 2019).

3.3.13.1.3. Toxicity in experimental animals

No data on toxicity in experimental mammalian species were identified in the previous Opinion, nor in the current update.

3.3.13.1.4. Computational analysis

The QSAR Toolbox did not identify any experimental data for the Ames test on TBX (see Annex B).

VEGA models predicted negative Ames test results for TBX with ‘GOOD reliability’ (three models). The Panel critically reviewed the predictions and considered them valid mainly due to the presence of relevant analogues known by the model with experimental and predicted values concordant with the prediction for the target compound (see Annex B).

3.3.13.1.5. Observations in humans

In the previous Opinion, no human studies on TBX were reported. Since then, three publications were identified, two pertaining to Chinese populations and one pertaining to a USA study population. Outcomes included thyroid hormone profiles (one study), type 2 diabetes (one study) and BMI (one study). In Appendix F, an overview of the epidemiological studies identified reporting on the BFRs considered in this Opinion, including those on TBX, is shown.

In the cross‐sectional study by Wang, Romanak, et al. (2019) in the USA (for details see Section 3.2.4.2), TBX concentrations were measured in wristbands in 101 adult study participants. TBX concentrations in the wristbands were not statistically significant associated with serum thyroid hormone profiles.

In the study by Zheng, Chen, et al. (2025) in China (for details see Section 3.2.4.2), serum TBX concentrations were not statistically significantly different between type 2 diabetes cases and controls in unadjusted comparisons (adjusted associations for single exposures were not reported) and TBX was not a main contributor to the mixture effects.

In the study by Zheng, Sun, et al. (2025) in China (for details see Section 3.2.4.2), serum TBX concentrations did not show statistically significant associations with BMI. Adjusted models are not presented.

3.3.13.1.6. Mode of action

Bajard et al. (2019) collected in vivo toxicity data on 62 FRs to establish prioritisation categories based on data availability and toxicological concern; the authors classified TBX as having ‘insufficient toxicological data to evaluate the toxicological concern'. The same authors reviewed the in vitro activity of 52 BFRs towards five nuclear receptors associated with reproductive outcomes, retrieving data from published articles and ToxCast/Tox21 programmes; the authors reported that no data were retrieved for TBX, and that the opensource platform Endocrine Disruptome model predicted medium probability of binding to the antagonist conformation of the AR (Bajard et al., 2021).

3.3.13.1.7. Consideration of critical effects and dose–response analysis

No experimental data on genotoxicity were identified. Computational analysis predicts that TBX does not induce mutations in S. Typhimurium, but no information is available on clastogenicity or aneugenicity. It is therefore not possible to conclude on genotoxicity overall.

The CONTAM Panel could not identify any toxicokinetic, toxicity or epidemiological studies on TBX.

It was therefore not possible to identify a Reference Point for TBX.

3.3.13.2. Occurrence data and dietary exposure assessment

Limited occurrence data in food were submitted to EFSA (see Section 3.1.1) and no estimation of the dietary exposure based on these data was possible.

In the scientific literature, several studies were identified reporting on the occurrence of TBX and an overview is given in Section 3.1.2 and Appendix B.

One study was identified reporting estimates of dietary exposure to TBX. Sahlström et al. (2015) reported occurrence levels and exposure of TBX, among other BFRs from analysis in Swedish market basket food samples collected in 2010, food baskets comprising of fish, meat, vegetable oils, dairy products and eggs (n = 4 food baskets). TBX was detected only in fish (100% detection frequency) samples, at 20 pg/g ww, while concentration in other food categories were < LOD. The dietary exposure was based on measured concentrations and intake assumptions from questionnaire data about dietary habits of the participants. The estimated median dietary daily intake for both mothers and toddlers (n = 20) based on fish intake only, was 0.7 and 0.3 ng per day, respectively.

For breastfed infants, there were insufficient data in human milk from European countries reported in the literature to make robust exposure estimates. To provide an indication of the potential exposure, the CONTAM Panel made an estimation for average and high consumption of human milk (see Section 2.4) considering the available data reported in the literature. Only one study reported levels above the LOD (Čechová et al., 2017, see Section 3.3.13.1.2).

The exposure scenario based on average human milk consumption and the highest reported mean concentration in human milk across three European countries (Norway, the Netherlands, Slovakia) of 0.035 ng/g lipid (Čechová et al., 2017), resulted in an estimate of 0.16 ng/kg bw per day.

The exposure scenario based on high human milk consumption and the highest reported P95 concentration of 0.128 ng/g lipid (Čechová et al., 2017), resulted in an estimate of 0.88 ng/kg bw per day.

It should be noted that these estimates are based on human milk samples from three European countries that were collected one or two decades ago. It is unknown whether this is representative of current exposure to this BFR.

3.3.13.3. Risk characterisation

No risk characterisation was possible, due to the lack of a Reference Point and the limited data on dietary exposure.

3.3.14. 1,3‐Bis(2,3‐dibromopropyl)‐5‐(2‐propen‐1‐yl)‐1,3,5‐triazine‐2,4,6(1H,3H,5H)‐trione (BDBP‐TAZTO) [CAS number 75795‐16‐3]

BDBP‐TAZTO is generally associated with TDBP‐TAZTO (see Section 3.3.7) which is used as an additive flame retardant. TDBP‐TAZTO is used as a mixture of the congeners substituted with one and two 2,3‐dibromopropyl groups (DBP‐TAZTO and BDBP‐TAZTO), respectively (EFSA CONTAM Panel, 2012b).

The physicochemical properties of BDBP‐TAZTO are reported in Table 1. BDBP‐TAZTO is currently not registered under REACH (see Table 2), and thus its production volume and use within the EU should be lower than one tonne per year.

3.3.14.1. Hazard identification and characterisation

3.3.14.1.1. Toxicokinetics

No information on toxicokinetics was identified in the previous Opinion, nor in the current update.

3.3.14.1.2. Levels in humans

No studies were identified in the previous Opinion, nor in the current update.

3.3.14.1.3. Toxicity in experimental animals

No data on toxicity in experimental mammalian species were identified in the previous Opinion, nor in the current update.

3.3.14.1.4. Computational analysis

The QSAR Toolbox did not identify any experimental data for the Ames test on BDBP‐TAZTO (see Annex B).

VEGA models predicted positive Ames test results with ‘LOW/MODERATE reliability’ (three models) and negative Ames test results with ‘MODERATE reliability’ (one model). The Panel assessed the validity of the positive predictions with moderate reliability according to VEGA, including the relevance of the analogues with experimental data identified by the software and the accuracy and consistency of their predictions. The Panel noted that the analogues were not sufficiently similar to the target compound to be confident in the reliability of the predictions, and thus these predictions were not further considered.

3.3.14.1.5. Observations in humans

No studies were reported in the previous Opinion, nor in the current update.

3.3.14.1.6. Mode of action

Bajard et al. (2019) collected in vivo toxicological data on 62 FRs to establish prioritisation categories based on data availability and toxicological concern; the authors classified BDBP‐TAZTO as having ‘insufficient toxicological data to evaluate the toxicological concern'. The same authors reviewed the in vitro activity of 52 BFRs towards five nuclear receptors associated with reproductive outcomes, retrieving data from published articles and ToxCast/Tox21 programmes; the authors reported that no data were retrieved for BDBP‐TAZTO, and that the opensource platform Endocrine Disruptome model predicted medium probability of binding to the antagonist conformation of the AR (Bajard et al., 2021).

3.3.14.1.7. Consideration of critical effects and dose–response analysis

The CONTAM Panel could not identify any toxicokinetic, toxicity or epidemiological studies on BDBP‐TAZTO.

It was therefore not possible to identify a Reference Point for BDBP‐TAZTO, and it was not possible to conclude on its genotoxic potential.

3.3.14.2. Occurrence data and dietary exposure assessment

No occurrence data on BDBP‐TAZTO in food were submitted to EFSA, and no occurrence or exposure data were identified in the literature, either in the previous Opinion or in the current update. Therefore, no estimation of the dietary exposure was possible.

3.3.14.3. Risk characterisation

No risk characterisation was possible due to the lack of a Reference Point and the lack of information on potential dietary exposure.

3.3.15. 1‐(2,3‐Dibromopropyl)‐3,5‐diallyl‐1,3,5‐triazine‐2,4,6(1H,3H,5H)‐trione (DBP‐TAZTO) [CAS number 57829‐89‐7]

DBP‐TAZTO is generally associated with TDBP‐TAZTO (Section 3.3.7) which is used as an additive flame retardant. TDBP‐TAZTO is used as a mixture of two congeners, one substituted with one and the other with two 2,3‐dibromopropyl groups (DBP‐TAZTO and BDBP‐TAZTO), respectively (EFSA CONTAM Panel, 2012b).

The physicochemical properties of DBP‐TAZTO are reported in Table 1. DBP‐TAZTO is currently not registered under REACH (see Table 2), and thus its production volume and use within the EU should be lower than one tonne per year.

3.3.15.1. Hazard identification and characterisation

3.3.15.1.1. Toxicokinetics

No information on toxicokinetics was identified in the previous Opinion, nor in the current update.

3.3.15.1.2. Levels in humans

No studies were identified in the previous Opinion, nor in the current update.

3.3.15.1.3. Toxicity in experimental animals

No data on toxicity in experimental mammalian species were identified in the previous Opinion, nor in the current update.

3.3.15.1.4. Computational analysis

The QSAR Toolbox did not identify any experimental data for the Ames test on DBP‐TAZTO (see Annex B).

VEGA models predicted positive Ames test results with ‘LOW/MODERATE reliability’ (three models) and negative Ames test results with ‘MODERATE reliability’ (one model). The CONTAM Panel assessed the validity of the positive predictions with moderate reliability according to VEGA, including the relevance of the analogues with experimental data identified by the software and the accuracy and consistency of their predictions. The Panel noted that the analogues were not sufficiently similar to the target compound to be confident in the reliability of the predictions, and thus these predictions were not further considered.

3.3.15.1.5. Observations in humans

No studies were reported in the previous Opinion, nor in the current update.

3.3.15.1.6. Mode of action

Bajard et al. (2019) collected in vivo toxicological data on 62 FRs to establish prioritisation categories based on data availability and toxicological concern; the authors classified DBP‐TAZTO as having ‘insufficient toxicological data to evaluate the toxicological concern'. The same authors reviewed the in vitro activity of 52 BFRs towards five nuclear receptors associated with reproductive outcomes, retrieving data from published articles and ToxCast/Tox21 programmes; the authors reported that no data were retrieved for DBP‐TAZTO, and that the opensource platform Endocrine Disruptome model predicted medium probability of binding to the antagonist conformation of the AR (Bajard et al., 2021).

3.3.15.1.7. Consideration of critical effects and dose–response analysis

The CONTAM Panel could not identify any toxicokinetic, toxicity or epidemiological studies on DBP‐TAZTO.

It was therefore not possible to identify a Reference Point for DBP‐TAZTO, and it was not possible to conclude on its genotoxic potential.

3.3.15.2. Occurrence data and dietary exposure assessment

No occurrence data on DBP‐TAZTO in food were submitted to EFSA, and no occurrence or exposure data were identified in the literature, either in the previous Opinion or in the current update. Therefore, no estimation of the dietary exposure was possible.

3.3.15.3. Risk characterisation

No risk characterisation was possible due to the lack of a Reference Point and the lack of information on potential dietary exposure.

3.3.16. Dibromostyrene (DBS) [CAS number 31780‐26‐4, 120359‐56‐0]

DBS is likely to occur as a mixture of isomers and is used for production of poly(dibromostyrene) (EFSA CONTAM Panel, 2012b). DBS is a reactive BFR with a QSAR prediction to be borderline ‘Persistent’ and borderline ‘Bioaccumulative’.²⁸

The physicochemical properties of DBS are reported in Table 1. DBS is currently not registered under REACH (see Table 2), and thus its production volume and use within the EU should be lower than one tonne per year.

3.3.16.1. Hazard identification and characterisation

3.3.16.1.1. Toxicokinetics

No information on toxicokinetics was identified in the previous Opinion, nor in the current update.

3.3.16.1.2. Levels in humans

No studies were identified in the previous Opinion, nor in the current update.

3.3.16.1.3. Toxicity in experimental animals

No data on toxicity in experimental mammalian species were identified in the previous Opinion, nor in the current update.

3.3.16.1.4. Computational analysis

The QSAR Toolbox did not identify any experimental data for the Ames test on DBS (see Annex B). VEGA models predicted negative Ames test results with ‘LOW reliability’ (four models). These predictions were not further considered due to their limited reliability.

3.3.16.1.5. Observations in humans

No studies were reported in the previous Opinion, nor in the current update.

3.3.16.1.6. Mode of action

Bajard et al. (2021) reviewed the in vitro activity of 52 BFRs towards five nuclear receptors associated with reproductive outcomes, retrieving data from published articles and ToxCast/Tox21 programmes; the authors reported that no data were retrieved for DBS, and reported that the opensource platform Endocrine Disruptome model predicted medium probability of binding to the antagonist conformation of the AR.

3.3.16.1.7. Consideration of critical effects and dose–response analysis

The CONTAM Panel could not identify any toxicokinetic, toxicity or epidemiological studies on DBS.

It was therefore not possible to identify a Reference Point for DBS, and it was not possible to conclude on its genotoxic potential.

3.3.16.2. Occurrence data and dietary exposure assessment

No occurrence data on DBS in food were submitted to EFSA, and no occurrence or exposure data were identified in the literature, either in the previous Opinion or in the current update. Therefore, no estimation of the dietary exposure was possible.

3.3.16.3. Risk characterisation

No risk characterisation was possible due to the lack of a Reference Point and due to the lack of information on potential dietary exposure.

3.3.17. 2‐(2‐Hydroxyethoxy)ethyl 2‐hydroxypropyl 3,4,5,6‐tetrabromophthalate (HEEHP‐TEBP) [CAS number 20566‐35‐2]

HEEHP‐TEBP is a highly polar tetrabromophthalic acid ester with one chiral carbon leading to formation of an enantiomeric pair. It is a phthalate ester with expected low environmental reactivity. Proposed uses are with production of foam for insulation (EFSA CONTAM Panel, 2012b). HEEHP‐TEBP is a reactive BFR with a QSAR prediction to be ‘(very)persistent’ and ‘very mobille’.²⁸

The physicochemical properties of HEEHP‐TEBP are reported in Table 1. HEEHP‐TEBP is currently not registered under REACH (see Table 2), and thus its production volume and use within the EU should be lower than one tonne per year.

3.3.17.1. Hazard identification and characterisation

3.3.17.1.1. Toxicokinetics

No information on toxicokinetics was identified in the previous Opinion, nor in the current update.

3.3.17.1.2. Levels in humans

No studies were identified in the previous Opinion, nor in the current update.

3.3.17.1.3. Toxicity in experimental animals

Studies considered in the previous EFSA assessment

No data on toxicity of HEEHP‐TEBP in experimental mammalian species were identified in the previous Opinion. The previous Opinion reported on one study with the technical product Saytex RB‐79 but no information on composition was reported. In that (range‐finding) study, the Saytex RB‐79 was given to female rats at a single gavage dose at 2000 mg/kg bw, and no effects were observed.

Regarding genotoxicity, the previous Opinion reported that the technical product Saytex RB‐79 (purity NR) tested in an Ames test with and without metabolic activation was not mutagenic. The technical product PHT4‐Diol (purity NR) was examined for mutagenic activity in S. Typhimurium and in Saccharomyces cerevisiae strain D4 with and without metabolic activation. No indication for mutagenicity was found (EFSA CONTAM Panel, 2012b).

Studies published since the previous EFSA assessment

No new studies were identified since the previous Opinion.

3.3.17.1.4. Computational analysis

The QSAR Toolbox did not identify any experimental data for the Ames test on HEEHP‐TEBP (see Annex B). VEGA models predicted negative Ames test results with ‘MODERATE reliability’ (four models). These predictions were not further considered due to their limited reliability.

3.3.17.1.5. Observations in humans

No studies were reported in the previous Opinion, nor in the current update.

3.3.17.1.6. Mode of action

No studies were identified in the previous Opinion, nor in the current update.

3.3.17.1.7. Consideration of critical effects and dose–response analysis

The CONTAM Panel could not identify any toxicokinetic, toxicity study in experimental animals, epidemiological studies or in vitro studies on HEEHP‐TEBP. One acute toxicity study and mutagenicity studies in bacteria and yeast conducted with the technical product Saytex RB‐79 and PHT4‐Diol did not show evidence of effects, but no information on the composition of these technical products was available.

It was therefore not possible to identify a Reference Point for HEEHP‐TEBP, and it was not possible to conclude on its genotoxic potential.

3.3.17.2. Occurrence data and dietary exposure assessment

No occurrence data on HEEHP‐TEBP in food were submitted to EFSA, and no occurrence or exposure data were identified in the literature, either in the previous Opinion or in the current update. Therefore, no estimation of the dietary exposure was possible.

3.3.17.3. Risk characterisation

No risk characterisation was possible due to the lack of a Reference Point and the lack of information on potential dietary exposure.

3.3.18. Tetradecabromo‐1,4‐diphenoxybenzene (4’‐PeBPO‐BDE208) [CAS number CAS 58965‐66‐5]

4’‐PeBPO‐BDE208 is a perbrominated compound that is expected to undergo photolysis in the environment. The high molecular mass may limit its bioavailability. It is used in high performance polyamide and linear polyester engineering resins and alloys (EFSA CONTAM Panel, 2012b).

The physicochemical properties of 4’‐PeBPO‐BDE208 are reported in Table 1. 4’‐PeBPO‐BDE208 is currently not registered under REACH (see Table 2), and thus its production volume and use within the EU should be lower than one tonne per year.

3.3.18.1. Hazard identification and characterisation

3.3.18.1.1. Toxicokinetics

No information on toxicokinetics was identified in the previous Opinion, nor in the current update.

3.3.18.1.2. Levels in humans

No studies were identified in the previous Opinion, nor in the current update.

3.3.18.1.3. Toxicity in experimental animals

No data on toxicity in experimental mammalian species were identified in the previous Opinion, nor in the current update.

3.3.18.1.4. Computational analysis

The QSAR Toolbox did not identify any experimental data for the Ames test on 4’‐PeBPO‐BDE208 (see Annex B). VEGA models predicted negative Ames test results with ‘LOW/MODERATE reliability’ (four models). These predictions were not further considered due to their limited reliability.

3.3.18.1.5. Observations in humans

No studies were reported in the previous Opinion, nor in the current update.

3.3.18.1.6. Mode of action

Hill et al. (2018) investigated the potential thyroidogenicity of 4’‐PeBPO‐BDE208 using an in vitro competitive protein binding assay with human thyroid hormone, transthyretin (hTTR) and albumin (hALB). The authors reported that 4’‐PeBPO‐BDE208 competed with T4 in binding to hTTR and hALB.

Bajard et al. (2019) collected in vivo toxicological data on 62 FRs to establish prioritisation categories based on data availability and toxicological concern; the authors classified 4’‐PeBPO‐BDE208 as having ‘insufficient toxicological data to evaluate the toxicological concern'. The same authors reviewed the in vitro activity of 52 BFRs towards five nuclear receptors associated with reproductive outcomes, retrieving data from published articles and ToxCast/Tox21 programmes; the authors reported that no data were retrieved for 4’‐PeBPO‐BDE208, and that the opensource platform Endocrine Disruptome model failed to predict due to the high molecular weight of this BFR (Bajard et al., 2021).

3.3.18.1.7. Consideration of critical effects and dose–response analysis

The CONTAM Panel could not identify any toxicokinetic, toxicity or epidemiological studies on 4’‐PeBPO‐BDE208.

It was therefore not possible to identify a Reference Point for 4’‐PeBPO‐BDE208, and it was not possible to conclude on its genotoxic potential.

3.3.18.2. Occurrence data and dietary exposure assessment

No occurrence data on 4’‐PeBPO‐BDE208 in food were submitted to EFSA, and no occurrence or exposure data were identified in the literature, either in the previous Opinion or in the current update. Therefore, no estimation of the dietary exposure was possible.

3.3.18.3. Risk characterisation

No risk characterisation was possible due to the lack of a Reference Point and lack of information on potential dietary exposure.

3.3.19. Tris(tribromoneopentyl) phosphate (TTBNPP) [CAS number 19186‐97‐1]

TTBNPP has been used in polypropylene products for uses in, e.g. carpets and stadium seats, and has potential for use in polyesters, polyamides, polystyrenes and polyolefins (EFSA CONTAM Panel, 2012b).

The physicochemical properties of TTBNPP are reported in Table 1. TTBNPP is currently not registered under REACH (see Table 2), and thus its production volume and use within the EU should be lower than one tonne per year.

3.3.19.1. Hazard identification and characterisation

3.3.19.1.1. Toxicokinetics

No information on toxicokinetics was identified in the previous Opinion.

Since then, toxicokinetics data have been added by the registrant(s) to the REACH dossier³⁸. According to the REACH dossier, the technical TTBNPP product FR‐370 (purity 99.8%) was rapidly excreted in rats, primarily via faeces (81%–85%) and urine (2%) within 24 h of oral administration. This indicates very low oral bioavailability (2%) and no accumulation potential. The CONTAM Panel could not assess the data because it did not have access to the original study.

3.3.19.1.2. Levels in human samples

No studies were identified in the previous Opinion, nor in the current update.

3.3.19.1.3. Toxicity in experimental animals

Studies considered in the previous EFSA assessment

No data on toxicity in experimental mammalian species were identified in the previous Opinion.

Studies published since the previous EFSA assessment

Since the previous Opinion, three studies were reported in the REACH dossier, ³⁹ as submitted by the registrant(s), for the technical product FR‐370 or PB‐370. The CONTAM Panel did not have access to the original studies. The REACH dossier reported that:

• In a 28‐day oral feeding study (OECD TG 407), male and female Sprague–Dawley CD rats were exposed to PB‐370 (purity NR) at 0, 4000, 8000 or 20,000 mg/kg feed. No effects were observed up to 20,000 mg/kg feed (the highest dose tested, corresponding to 1361–1959 mg/kg bw per day for males and 1691–2081 mg/kg bw per day for females).

• In a 90‐day oral feeding study (OECD TG 408), male and female Sprague–Dawley CD rats were given PB‐370 (purity NR) at 0, 2000, 10,000 or 20,000 mg/kg feed (corresponding to 0, 158, 679 or 1358 mg/kg bw per day for males, and 0, 169, 843 or 1685 mg/kg bw per day for females). No effects were observed up to 20,000 mg/kg feed (the highest dose tested, corresponding to 1358 mg/kg bw per day in males and 1685 mg/kg bw per day in females).

• In a developmental toxicity study (OECD TG 414), pregnant female Sprague–Dawley rats (CD (Crl:CDR (SD)IGS BR) were exposed by gavage from GD1–19 to PB‐370 (reported purity ‘99.1%’) at 0, 100, 300 or 1000 mg/kg bw per day. There were no adverse effects on the dams or on embryo‐fetal survival. Slightly reduced fetal weight and marginally increased placental weight were observed at 1000 mg/kg bw per day. No information on statistical significance was reported. The NOAEL for maternal toxicity was 1000 mg/kg bw per day. In the absence of the original data the CONTAM Panel could not conclude on a NOAEL on developmental toxicity.

• In a reproduction/developmental toxicity screening test (OECD TG 421), male and female Sprague–Dawley rats were exposed orally via feed to FR‐370 (reported purity ‘99.8%’) at 0, 2000, 10,000 or 20,000 mg/kg feed (corresponding to 0, 160, 760 or 1590 mg/kg bw per day for males and 0, 400, 1780 or 4110 mg/kg bw per day for females). F0 animals were treated for 28 days before mating until termination either after all litters have been born for the males or after weaning of F1 generation for the females. F1 animals were treated from weaning until approximately 6 weeks of age. No adverse effects were observed in parental animals nor in F1 generation. The NOAEL for parental toxicity and reproduction was 20,000 mg/kg feed (the highest dose tested, corresponding to 1590 mg/kg bw per day for males and 4110 mg/kg bw per day for females).

The first three studies were also reported in an US‐EPA Report on alternatives assessment for decabromodiphenyl ether (US‐EPA, 2014). The CONTAM Panel had no access to the original data.

Genotoxicity

Since the previous Opinion, information on genotoxicity was retrieved from the REACH dossier, as submitted by the registrant(s) for the technical product FR‐370. The CONTAM Panel did not have access to the original data. The REACH dossier reported that:

• FR‐370 (purity NR) was not mutagenic with and without metabolic activation in reverse mutation assays in S. Typhimurium (TA1535, TA1537, TA1538, TA98, TA100) or in the TK gene mutation assay in L5178Y mouse lymphoma cells (reported purity ‘99.6%’), and did not induce chromosomal aberrations in CHO Chinese hamster ovary cells with and without metabolic activation. Equivocal results for induction of numerical aberrations were also reported in CHO cells.

The CONTAM Panel noted that the available information in the REACH dossier is insufficient to conclude on the genotoxic potential of the substance (cytotoxicity levels are often unclear, positive/equivocal results are scattered and not associated with a dose–response relationship, and precipitation of the test substance may have affected the reliability of the results). The same assays were reported in the US‐EPA Report on alternatives assessment for decabromodiphenyl ether (US‐EPA, 2014).

3.3.19.1.4. Computational analysis

The QSAR Toolbox identified experimental data for Ames test on TTBNPP (see above and Annex B).

VEGA models predicted positive Ames test results with ‘MODERATE reliability’ (one model) and negative Ames test results with ‘LOW/MODERATE reliability’ (three models). The CONTAM Panel assessed the validity of the positive predictions with moderate reliability according to VEGA, including the relevance of the analogues with experimental data identified by the software and the accuracy and consistency of their predictions. The Panel noted inconsistent data for different similar analogues, and thus these predictions were not further considered.

3.3.19.1.5. Observations in humans

No studies were reported in the previous Opinion, nor in the current update.

3.3.19.1.6. Mode of action

Bajard et al. (2019) collected in vivo toxicological data on 62 FRs to establish prioritisation categories based on data availability and toxicological concern; the authors classified TTBNPP as having ‘scarce toxicological data indicating lower toxicological concern'. In a later study, Bajard et al. (2023) reported that TTBNPP did not exhibit anti‐androgenic activity in an in vitro reporter gene assay.

3.3.19.1.7. Consideration of critical effects and dose–response analysis

Several studies have been performed on PB‐370 or FR‐370. In these studies, no adverse effects were reported in 28‐day or 90‐day feeding toxicity studies in rats up to the highest dose tested 20,000 mg/kg feed (corresponding to 1358 mg/kg bw per day in males and 1685 mg/kg bw per day in females for the 90‐day study) (purity NR). In a developmental toxicity study in rats exposed by gavage from GD1 to 19, there were no adverse effects on the dams or on embryo‐fetal survival up to the highest dose 1000 mg/kg bw per day. Slightly reduced fetal weight and marginally increased placental weight were reported at 1000 mg/kg bw per day (reported purity ‘99.1%’). No adverse effects were reported in the parental or F1 generation in a reproduction/developmental toxicity screening test in rats up to the highest dose tested 20,000 mg/kg feed (reported purity ‘99.8%’).

The CONTAM Panel could not identify any toxicokinetic, long‐term, neurodevelopmental or epidemiological studies on TTBNPP.

Based on the in vitro data reported in the REACH dossier, the CONTAM Panel could not conclude on the genotoxicity of TTBNPP.

Based on the limited data available, and in the absence of access to the original study reports, the CONTAM Panel could not identify a Reference Point.

3.3.19.2. Occurrence data and dietary exposure assessment

No occurrence data on TTBNPP in food were submitted to EFSA, and no occurrence or exposure data were identified in the literature, either in the previous Opinion or in the current update. Therefore, no estimation of the dietary exposure was possible.

3.3.19.3. Risk characterisation

No risk characterisation was possible due to the lack of a Reference Point and lack of information on potential dietary exposure.

3.3.20. 2,4,6‐Tris(2,4,6‐tribromophenoxy)‐1,3,5‐triazine (TTBP‐TAZ) [CAS number 25713‐60‐4]

TTBP‐TAZ is a non‐polymeric additive BFR. It has been assessed to be not PBT (‘persistent, bioaccumulative and toxic’) and not vPvB (‘very persistent and very bioaccumulative’). ⁴⁰ TTBP‐TAZ may biotransform to the substance 2,4,6‐TBP, for which there is a concern for PMT (‘persistent, mobile, toxic’) or vPvM (‘very persistent, very mobile’) (Zheng et al., 2022). It is also possible that the hydrolysis of TTBP‐TAZ can form 2,4,6‐TBP.

TTBP‐TAZ occurs as technical products with different tradenames, such as SR 245 or FR‐245, that have been reported to be mono‐constituent substance (NICNAS, 2006).

The physicochemical properties of TTBP‐TAZ are reported in Table 1. TTBP‐TAZ is registered under REACH (1000–10,000 tonnes per annum) (see Table 2).

3.3.20.1. Hazard identification and characterisation

3.3.20.1.1. Toxicokinetics

No information on toxicokinetics was identified in the previous Opinion.

Since then, and based on the information in the REACH dossier,⁴⁰ as submitted by the registrant(s), the technical TTBP‐TAZ product, FR‐245 (purity 99.9%), was reported to have very low oral bioavailability in rats (< 0.2%), and no accumulation potential has been found. The CONTAM Panel could not assess these data as it did not have access to the original study.

3.3.20.1.2. Levels in humans

No studies were identified in the previous Opinion, nor in the current update.

3.3.20.1.3. Toxicity in experimental animals

Studies considered in the previous EFSA assessment

No data on toxicity in experimental mammalian species were identified in the previous Opinion.

Studies published since the previous EFSA assessment

Since the previous Opinion, studies with rats were summarised in the REACH dossier, ⁴¹ as submitted by the registrant(s) and in NICNAS (2006) for the technical products FR‐245 and SR‐245. The CONTAM Panel did not have access to the original study reports. These sources reported that:

• −An acute oral (gavage) toxicity study in rats with a single dose of the technical product SR‐245 (reported as 100% purity) of 2000 mg/kg bw was conducted and no mortalities were observed (LD50 > 2000 mg/kg bw).
• −In a 28‐day oral (gavage) toxicity study (OECD TG 407) in Crj:CD(SD) rats, animals were given 0, 10, 50, 250 or 1000 mg/kg bw per day of the technical product FR‐245 (reported as 100% purity). Two recovery groups were treated with the high dose (1000 mg/kg bw per day) or the vehicle for 28 days and then maintained without treatment for further 14 days. Clinical signs, body weight development, food and water consumption were monitored during the study. Haematology, blood chemistry and urinalysis were evaluated for all non‐recovery group animals at the end of the treatment period and for recovery group animals at the end of the recovery period. All animals were subjected to gross necropsy examination and histopathological evaluation of selected tissues from non‐recovery high dose and control animals. In females, decreased reticulocytes count in the 10 mg/kg bw per day and higher groups, decreased mean corpuscular haemoglobin in the 50 and 250 mg/kg per day groups were noted. However, these were deemed not to be dose related and were not present in the animals that were allowed to recover before sacrifice. The NOEL was reported to be > 1000 mg/kg bw per day.
• −A 91‐day oral gavage toxicity study (OECD TG 408) in Crj:CD(SD) rats with a 28‐day recovery period was conducted using FR‐245 (reported as 99.7% purity). Three treatment groups (100, 350 and 1000 mg/kg bw per day) comprising 10 male and 10 female rats each received daily doses for 91 consecutive days, alongside a vehicle control group. Two recovery groups (five males and five females) received either the high dose (1000 mg/kg bw per day) or vehicle for 91 days, followed by a 28‐day treatment‐free period. The study assessed reproductive and neurotoxicity endpoints in addition to standard parameters. Daily oral gavage administration of FR‐245 to Sprague–Dawley rats for 91 consecutive days produced no adverse effects on body weight, food consumption, haematology, blood biochemistry, urinalysis, organ weight, or gross and histopathological findings throughout the in‐life phase, necropsy or histological assessment. No treatment‐related changes were observed in sperm count or motility. Vaginal lavages showed no treatment‐related alterations in the oestrous cycle. Furthermore, no dose‐related changes in ovarian, seminal vesicle, testicular, ureteral, uterine or vaginal organ weights were detected compared with controls. The NOEL was reported to be > 1000 mg/kg bw per day.
• −A developmental toxicity study in Crl:CD(SD) rats was conducted in accordance with OECD TG 414 to evaluate the potential maternal and developmental toxicity of the technical product BTAC‐245 (purity NR). Doses of 250, 500 and 1000 mg/kg bw per day test substance were administered by gavage during GD5–19. No effects of test substance treatment were observed in maternal rats across any of the parameters examined. At caesarean section, no treatment‐related abnormalities were noted in mean gravid uterine weights, numbers of corpora lutea and implantations, percentage implantation index, or pre‐ and post‐implantation loss in any treated groups. Regarding fetuses, the number of live fetuses, percentage fetal deaths, fetal body weight, placental weights and sex ratio in treated groups were comparable to the control group. No treated groups demonstrated adverse effects of test substance treatment in external, visceral or skeletal examinations of live fetuses. Based on the absence of toxicity (maternal and fetal) and absence of teratogenic findings, the authors concluded that the dose level of 1000 mg/kg per day of TTBP‐TAZ represented the NOAEL for both maternal rats and fetuses.

In summary, from the scientific data identified it may be concluded that sub‐chronic exposure to the technical products SR 245, FR‐245 and BTAC‐245 containing TTBP‐TAZ did not have adverse effects on adult rats up to 1000 mg/kg bw per day. This dose did not cause systemic toxicity or developmental toxicity, and did not affect reproductive organs.

Genotoxicity

Since the previous Opinion, information on the same genotoxicity studies was available in the REACH dossier, as submitted by the registrant(s) and in NICNAS (2006). The CONTAM Panel did not have access to the original data. These sources reported that:

The technical products FR‐245 and SR‐245 (reported purity 100%) were not mutagenic with and without metabolic activation in reverse mutation assays in S. Typhimurium (TA1535, TA1537, TA98, TA100), and SR‐245 (purity NR) did not induce chromosomal aberrations in human lymphocytes with and without metabolic activation. The technical product BTAC‐245 (reported purity 99.8%) and FR‐245 were not mutagenic in the TK gene mutation assay in L5178Y mouse lymphoma cells with and without metabolic activation.

Based on the information reported in the REACH dossier, the CONTAM Panel concluded that TTBP‐TAZ is not genotoxic.

3.3.20.1.4. Computational analysis

The QSAR Toolbox identified experimental data for the Ames test on TTBP‐TAZ (see above and Annex B). VEGA models predicted negative Ames test results with ‘LOW/MODERATE reliability’ (four models). These predictions were not further considered due to their limited reliability.

3.3.20.1.5. Observations in humans

No studies were reported in the previous Opinion, nor in the current update.

3.3.20.1.6. Mode of action

No studies were identified in the previous Opinion, nor in the current update.

3.3.20.1.7. Consideration of critical effects and dose–response analysis

The CONTAM Panel could not identify any toxicokinetic, toxicity or epidemiological studies on TTBP‐TAZ.

Based on the limited data available identified, SR/FR‐245 did not show short‐term (28‐ and 90‐day) or developmental toxicity in rats at doses up to 1000 mg/kg bw per day, the highest dose tested.

No chronic/carcinogenicity, reproductive or neurodevelopmental toxicity studies were identified.

The technical product SR/FR‐245 was not mutagenic in bacteria or in mammalian cells, and did not induce chromosomal damage. Based on the information reported in the REACH dossier, the CONTAM Panel concluded that these in vitro data indicated that TTBP‐TAZ was not genotoxic.

The CONTAM Panel did not have information on the content of TTBP‐TAZ in the technical products tested.

Because of the limited data available, the CONTAM Panel concluded that it was not possible to identify a Reference Point.

3.3.20.2. Occurrence data and dietary exposure assessment

No occurrence data on TTBP‐TAZ in food were submitted to EFSA, and no occurrence or exposure data were identified in the literature, either in the previous Opinion or in the current update. Therefore, no estimation of the dietary exposure was possible.

3.3.20.3. Risk characterisation

No risk characterisation was possible due to the lack of a Reference Point and lack of information on potential dietary exposure.

3.3.20.4. Uncertainty analysis

Due to the limited data on hazard, occurrence and exposure for TTBP‐TAZ, uncertainty analysis was performed only for the conclusion on genotoxicity of TTBP‐TAZ using the methods described in Section  2.8 .

Information on genotoxicity was retrieved from the REACH dossier and NICNAS (2006), reporting negative results for all the necessary studies. No sources of uncertainty were indicated. Taking this into account, the CONTAM Panel judged with > 95% certainty (i.e. probability) that TTBP‐TAZ is not genotoxic.

3.3.21. Octabromotrimethylphenyl indane (OBTMPI) [CAS number 1084889‐51‐9, 1025956‐65‐3, 893843‐07‐7, 155613‐93‐7]

OBTMPI is a non‐polymeric additive BFR that has been assessed by QSAR to be borderline ‘Persistent’, ‘Bioaccumulative’, ‘Toxic’ (ECHA, 2024). OBTMPI contains eight bromine atoms on the two aromatic rings. It has one chiral carbon, leading to the formation of one enantiomeric pair (EFSA CONTAM Panel, 2012b).

The physicochemical properties of OBTMPI are reported in Table 1. OBTMPI is currently not registered under REACH (see Table 2), and thus its production volume and use within the EU should be lower than one tonne per year.

3.3.21.1. Hazard identification and characterisation

3.3.21.1.1. Toxicokinetics

No information on toxicokinetics was identified in the previous Opinion, nor in the current update.

3.3.21.1.2. Levels in humans

No studies were reported in the previous Opinion. Since then, two studies have been identified reporting levels in blood in European countries (see Appendix E, Table E.1):

• Sahlström et al. (2014) did not detect OBTMPI above the LOD (18 pg/25 μL) in any of the 24 matched serum samples from mothers and their toddlers collected in Sweden between 2009 and 2010.

• OBTMPI was also measured in 61 serum samples collected in 2013 from a Norwegian cohort with a low detection frequency of 2%, and a reported geometric mean of 1.3 ng/g lipid (Tay et al., 2019).

3.3.21.1.3. Toxicity in experimental animals

No data on toxicity in experimental mammalian species were identified in the previous Opinion, nor in the current update.

3.3.21.1.4. Computational analysis

The QSAR Toolbox did not identify any experimental data for the Ames test on OBTMPI (see Annex B). VEGA models predicted negative Ames test results with ‘LOW/MODERATE reliability’ (four models). These predictions were not further considered due to their limited reliability.

3.3.21.1.5. Observations in humans

No studies were reported in the previous Opinion, nor in the current update.

3.3.21.1.6. Mode of action

Bajard et al. (2019) collected in vivo toxicological data on 62 FRs to establish prioritisation categories based on data availability and toxicological concern; the authors classified OBTMPI as having ‘insufficient toxicological data to evaluate the toxicological concern'. The same authors reviewed the in vitro activity of 52 BFRs towards five nuclear receptors associated with reproductive outcomes, retrieving data from published articles and ToxCast/Tox21 programmes; the authors reported that no data were retrieved for OBTMPI, and that the opensource platform Endocrine Disruptome model failed to predict due to the high molecular weight of this BFR (Bajard et al., 2021).

Das et al. (2021) performed in silico toxicity predictions and reported that OBTMPI exhibited some oestrogen antagonist, androgen antagonist and oestrogen binding capability.

3.3.21.1.7. Consideration of critical effects and dose–response

The CONTAM Panel could not identify any toxicokinetic, toxicity or epidemiological studies on OBTMPI.

It was therefore not possible to identify a Reference Point for OBTMPI, and it was not possible conclude on its genotoxic potential.

3.3.21.2. Occurrence data and dietary exposure assessment

No occurrence data on OBTMPI in food were submitted to EFSA. In the literature, one study included OBTMPI but it was not detected above the LOD (Fernandes et al., 2016, see Section 3.1.2).

No exposure data were identified in the literature, either in the previous Opinion or in the current update. Therefore, no estimation of the dietary exposure was possible.

3.3.21.3. Risk characterisation

No risk characterisation was possible due to the lack of a Reference Point and lack of information on potential dietary exposure.

4. CONCLUSIONS

This assessment is an update of the EFSA CONTAM Panel Scientific Opinion on Emerging and Novel Brominated Flame Retardants (BFRs) in Food published by EFSA in 2012. The update focuses on the six BFRs considered in the 2012 Opinion and covered in Commission Recommendation 2014/118/EU (TDBPP, EBTEBPI, HBCYD, BEH‐TEBP, EH‐TBB and DBNPG) and the remaining 21 BFRs considered in the 2012 Opinion (BTBPE, DBDPE, DBE‐DBCH, DBHCTD, HBB, HCTBPH, OBTMPI, PBB‐Acr, PBEB, PBT, TBNPA, TDBP‐TAZTO, TBCO, TBX, BDBP‐TAZTO, DBP‐TAZTO, DBS, HEEHP‐TEBP, 4’‐PeBPO‐BDE208, TTBNPP, TTBP‐TAZ). It takes into account the occurrence data in food and biological samples submitted to EFSA after the publication of the previous Opinion, as well as the newly available scientific information of relevance to hazard identification and characterisation.

4.1. Hazard identification and characterisation

• The information available on the toxicity of the BFRs included in the Terms of Reference varied greatly across compounds, but was in general scarce.

• Regarding genotoxicity, the CONTAM Panel concluded that:

  • −
    Based on the information provided in the REACH dossier, EBTEBPI, DBDPE, PBB‐Acr and TTBP‐TAZ are not genotoxic.
  • −
    TDBPP is genotoxic and carcinogenic in rats (in kidney) and mice (in kidney, liver, forestomach, lung).
  • −
    DBNPG is genotoxic and carcinogenic in rats (e.g. in skin, subcutaneous tissue, mammary gland, Zymbal's gland, oral cavity, oesophagus, forestomach, small and large intestines, mesothelium, urinary bladder, lung and thyroid gland) and mice (e.g. in Harderian gland, lung, subcutaneous tissue, forestomach).
  • −
    In the absence of conclusive negative results for in vivo genotoxicity data, TBNPA is considered genotoxic based on the positive in vitro mutagenicity and clastogenicity. TBNPA is predicted to be carcinogenic based on the positive in vitro mutagenicity and clastogenicity supported by structural similarity with the genotoxic and carcinogenic DBNPG.
  • −
    In the absence of conclusive negative results for in vivo mutagenicity, DBE‐DBCH is considered genotoxic based on positive in vitro mutagenicity in mammalian cells.
  • −
    BEH‐TEBP, BTBPE, HBB, PBEB and PBT are not mutagenic, but the lack of data on clastogenicity and aneugenicity did not allow the Panel to conclude on their genotoxic potential.
  • −
    QSAR VEGA models predicted negative Ames test results for DBHCTD and TBX, but the lack of data on clastogenicity and aneugenicity did not allow the Panel to conclude on their genotoxic potential.
  • −
    Insufficient (TTBNPP), or no data (HBCYD, EH‐TBB, TDBP‐TAZTO, TBCO, HCTBPH, BDBP‐TAZTO, DBP‐TAZTO, DBS, HEEHP‐TEBP, 4’‐PeBPO‐BDE208, OBTMPI) were identified to allow an assessment of the genotoxic potential of the remaining BFRs.

• Evidence of toxicity for reproduction or development in rodents was identified for DBNPG, BEH‐TEBP, EH‐TBB and DBDPE.

• Evidence of neurotoxicity or developmental neurotoxicity in rodents was identified for BEH‐TEBP, EH‐TBB, DBDPE and TDBP‐TAZTO. Studies in zebrafish larvae indicated potential for developmental neurotoxicity for DBE‐DBCH.

• The CONTAM Panel identified a Reference Point for TDBPP, DBNPG, BEH‐TEBP, EH‐TBB, DBDPE and TDBP‐TAZTO:

  • −
    For TDBPP, a Reference Point corresponding to a BMDL₁₀ of 0.073 mg/kg bw per day was identified, based on combined incidence of kidney tubular‐cell adenoma and carcinoma in male rats.
  • −
    For DBNPG, a Reference Point corresponding to a BMDL₁₀ of 1.14 mg/kg bw per day for the technical product FR‐1138 (purity 78.6%), corresponding to 0.9 mg/kg bw per day for DBNPG was identified, based on increased incidence of mammary gland fibroadenoma, adenoma or carcinoma in female rats.
  • −
    For BEH‐TEBP, a Reference Point corresponding to LOAEL of 1 mg/kg bw per day was identified, based on neurobehavioural and skeletal effects in a rat study with binary mixture BEH‐TEBP and EH‐TBB.
  • −
    For EH‐TBB, a Reference Point corresponding to LOAEL of 2 mg/kg bw per day was identified, based on neurobehavioural and skeletal effects in a rat study with binary mixture BEH‐TEBP and EH‐TBB.
  • −
    For DBDPE, a Reference Point corresponding to a BMDL₁₀ of 1.07 μg/kg bw per day was identified based on a neurobehavioural endpoint, an increase in total wheel running activity.
  • −
    For TDBP‐TAZTO, a Reference Point corresponding to a LOAEL of 5 mg/kg bw per day was identified, based on depressive‐related behaviour in the forced swim test and the tail suspension test in rodents.

• Due to limited or complete lack of data on hazard identification and characterisation, the CONTAM Panel could not identify a Reference Point for the remaining BFRs included in the Terms of Reference (HBCYD, EBTEBPI, BTBPE, DBE‐DBCH, DBHCTD, HBB, HCTBPH, OBTMPI, PBB‐Acr, PBEB, PBT, TBNPA, TBCO, TBX, BDBP‐TAZTO, DBP‐TAZTO, DBS, HEEHP‐TEBP, 4’‐PeBPO‐BDE208, TTBNPP, TTBP‐TAZ).

• No health‐based guidance values (HBGV) could be derived for any of the BFRs included in the Terms of Reference.

4.2. Occurrence in human samples

• Occurrence data in human milk and/or blood/serum in European countries were identified for BEH‐TEBP, EH‐TBB, DBDPE, BTBPE, HBB, DBE‐DBCH, PBEB, PBT, TBCO, DBHCTD, TBX and OBTMPI.

• The number of studies was limited and characterised by low detection frequencies, and relied on samples collected around a decade ago or more.

• No data on human samples in European countries was identified for the other BFRs included in the Terms of Reference.

4.3. Occurrence in food and dietary exposure for the European population

• Occurrence data in food were submitted to EFSA for two out of the six BFRs covered in Commission Recommendation 2014/118/EU (BEH‐TEBP and EH‐TBB) and for six out of the remaining 21 BFRs covered in the 2012 Opinion (BTBPE, DBDPE, HBB, PBEB, PBT, TBX).

• The occurrence data in food submitted to EFSA were provided by six European Countries and were derived from samples collected between 2010 and 2025.

• Dietary exposure could be estimated for BTBPE and HBB, while the limited and predominantly left‐censored data submitted to EFSA for the other BFRs were not sufficient to perform exposure estimations.

  • −
    For BTBPE, mean LB dietary exposure ranged from 0.001 ng/kg bw per day in ‘Adults’, ‘Elderly’ and ‘Very Elderly’ to 0.006 ng/kg bw per day in ‘Infants’ and ‘Toddlers’, while mean UB dietary exposure ranged from 0.006 ng/kg bw per day in ‘Elderly’ and ‘Very Elderly’ to 0.104 ng/kg bw per day in ‘Elderly’ and ‘Very Elderly’, across surveys. The P95 LB dietary exposure ranged from 0.002 ng/kg bw per day in ‘Elderly’ and ‘Very Elderly’ to 0.014 ng/kg bw per day in ‘Other children', while the P95 UB dietary exposure ranged from 0.011 ng/kg bw per day in ‘Very Elderly’ to 0.328 ng/kg bw per day in ‘Elderly’, across surveys
  • −
    For HBB, mean LB dietary exposure ranged from 0.04 ng/kg bw per day in ‘Adults’, ‘Elderly’ and ‘Very Elderly’ to 0.25 ng/kg bw per day in ‘Toddlers’, while mean UB dietary exposure ranged from 0.05 ng/kg bw per day in ‘Elderly’ and ‘Very Elderly’ to 0.3 ng/kg bw per day in ‘Toddlers’, across surveys. The P95 LB dietary exposure ranged from 0.06 ng/kg bw per day in ‘Elderly’ to 0.61ng/kg bw per day in ‘Other children', while P95 UB dietary exposure ranged from 0.09 ng/kg bw per day in ‘Elderly’ to 0.67 ng/kg bw per day in ‘Other children', across surveys

• Limited data on dietary exposure estimates in European countries were identified in the scientific literature for BEH‐TEBP, EH‐TBB, DBDPE, BTBPE, DBE‐DBCH, PBEB, PBT and TBX.

• No information on potential dietary exposure in European countries was identified for the remaining BFRs (TDBPP, EBTEBPI, HBCYD, DBNPG, TDBP‐TAZTO, TBCO, DBHCTD, HCTBPH, PBB‐Acr, TBNPA, DBDP‐TAZTO, DBP‐TAZTO, DBS, HEEHP‐TEBP, 4’‐PeBPO‐BDE208, TTBNPP and TTBP‐TAZ).

• There were insufficient data to make robust exposure estimates for breastfed infants, but an indication of mean and/or high level (P95) potential exposure was possible for BEH‐TEBP, EH‐TBB, DBDPE, BTBPE, HBB, DBE‐DBCH, PBT and TBX.

4.4. Risk characterisation

• Risk characterisation could be performed for two out of the six BFRs covered by Commission Recommendation 2014/118/EU (BEH‐TEBP and EH‐TBB) and for one out of the remaining 21 BFRs assessed in the 2012 Opinion (DBDPE), making use of the limited information on dietary exposure estimates from one (EH‐TBB, DBDPE) or two (BEH‐TEBP) European countries reported in the scientific literature. The margin of exposure (MOE) approach was applied to assess possible health concerns.

  • −
    For BEH‐TEBP, based on the limited dietary exposure estimates from the literature and the LOAEL of 1 mg/kg bw per day based on neurobehavioural and skeletal effects in a rat study with binary mixture of BEH‐TEBP and EH‐TBB, the CONTAM Panel concluded that dietary exposure to BEH‐TEBP does not raise a health concern for non‐neoplastic effects.
  • −
    For EH‐TBB, based on the limited dietary exposure estimates from the literature and the LOAEL of 2 mg/kg bw per day based on neurobehavioural and skeletal effects in a rat study with binary mixture of BEH‐TEBP and EH‐TBB, the CONTAM Panel concluded that dietary exposure to EH‐TBB does not raise a health concern for non‐neoplastic effects.
  • −
    For DBDPE, based on the limited dietary exposure estimates from the literature and the BMDL₁₀ of 1.07 μg/kg bw per day based on the increase of total wheel running activity, the CONTAM Panel concluded that dietary exposure to DBDPE does not raise a health concern.

• No risk characterisation could be performed for any of the other BFRs included in the Terms of Reference, due to insufficient or lack of data on hazard characterisation and/or dietary exposure.

4.5. Uncertainty analysis

• Uncertainties affecting the assessment were identified and their overall impact was quantified in those cases where the available data were sufficient to conclude on genotoxic potential, hazard characterisation or risk characterisation. The CONTAM Panel assessed the level of certainty (probability) to be 90% or more for all the reported conclusions, except for the risk characterisation for DBDPE, where the certainty of no health concern was lower and considered to fall within the range of 66%–90%.

5. RECOMMENDATIONS

The CONTAM Panel made the following recommendations:

• Occurrence data in food, obtained using appropriate analytical methods, are needed for those BFRs for which insufficient or no occurrence data were submitted to EFSA. Priority should be given to: (i) BFRs for which a Reference Point is available and/or for those that the Panel has concluded to be genotoxic (TDBPP, BEH‐TEBP, EH‐TBB, DBNPG, DBDPE, DBE‐DBCH, TDBP‐TAZTO, TBNPA), (ii) BFRs that have been detected in human samples but lack a Reference Point or conclusion on genotoxicity (BTBPE, HBB, PBEB, PBT, TBX, OBTMPI).

• Biomonitoring data, obtained using appropriate analytical methods, should be generated. Priority should be given to TDBPP, BEH‐TEBP, EH‐TBB, DBNPG, DBDPE, HBB, TDBP‐TAZTO, DBE‐DBCH and TBNPA, as these show potential for bioaccumulation, or have an identified Reference Point, or are concluded to be genotoxic.

• Genotoxicity data are needed, most importantly for the BFRs with inconclusive (BEH‐TEBP, BTBPE, HBB, PBEB, PBT, TBX) or absent (EH‐TBB) genotoxicity information, and that are known to occur in food.

• Toxicological studies that would enable the identification of a Reference Point are needed for those BFRs for which insufficient or no studies were identified. Priority should be given to the BFRs for which evidence of potential dietary exposure was identified and/or they have been detected in human samples (BTBPE, HBB, DBE‐DBCH, PBEB, PBT, TBCO, DBHCTD, TBX, OBTMPI).

• In the context of the One Substance One Assessment, and to ensure efficient and harmonised risk assessments, there is a need to further enable access to the original data underpinning regulatory submissions across relevant authorities.

ABBREVIATIONS

4’‐PeBPO‐BDE208

tetradecabromo‐1,4‐diphenoxybenzene

ABS

acrylonitrile butadiene styrene

AhR

arylhydrocarbon receptor

ALD

approximate lethal doses

APCI

atmospheric pressure chemical ionization

AR

androgen receptor

b.w.

body weight

BAF

bioaccumulation factor

BBPP

bis(2,3‐dibromopropyl) phosphate

BCF

bioconcentration factor

BDBP‐TAZTO

1,3‐Bis(2,3‐dibromopropyl)‐5‐(2‐propen‐1‐yl)‐1,3,5‐triazine‐2,4,6(1H,3H,5H)‐trione

BEH‐TEBP

bis(2‐ethylhexyl) tetrabromophthalate

BFR

brominated flame retardant

BMD

benchmark dose

BMDL

Benchmark dose lower confidence limit

BMDU

Benchmark dose upper confidence limit

BMF

Biomagnification factor

BMR

Benchmark response

BTBPE

1,2‐Bis(2,4,6‐tribromophenoxy)ethane

CALUX

chemically activated luciferase expression

CAR

constitutive androgen receptor

CAS

Chemical Abstract Service

CONTAM Panel

Panel on Contaminants in the Food Chain

CRM

certified reference material

CTD

characteristic travel distance

CYP

cytochrome P‐450

DBDBE

decabromodibenzyl ether

DBDPE

decabromodiphenyl ethane

DBE‐DBCH

4‐(1,2‐dibromoethyl)‐1,2‐dibromocyclohexane

DBHCTD

5,6‐dibromo‐1,10,11,12,13,13‐hexachloro‐11‐tricyclo[8.2.1.0^2,9]tridecene

DBNPG

dibromoneopentyl glycol

DBP‐TAZTO

1‐(2,3‐dibromopropyl)‐3,5‐diallyl‐1,3,5‐triazine‐2,4,6(1H,3H,5H)‐trione

DBS

dibromostyrene

DHT

dihydrotestosterone

DMSO

dimethyl sulfoxide

DOH

2,6‐dioxaspiro[3.3]‐heptane

E2

17β‐estradiol

EBTBPI

N,N′‐Ethylenebis(tetrabromophthalimide)

ECHA

European Chemicals Agency

ECNI

electron capture negative ionisation

EHD

empirical half distance

EH‐TBB

2‐Ethylhexyl 2,3,4,5‐tetrabromobenzoate

EI‐HRMS

electron impact‐high resolution mass spectrometry

EPDM

ethylene propylene‐diene terpolymers

EROD

ethoxyresorufin‐Q‐deethylase

ESI

electrospray ionisation

ESIS

European Chemical Substances Information System

GC × GC–TOF‐MS

two‐dimensional gas chromatography coupled to time‐of‐flight mass spectrometry

GC–ECNI‐MS

gas chromatography–mass spectrometry using electron capture negative ionisation

GD

gestation day

GPC

gel permeation chromatography

GSH

glutathione

HBB

hexabromobenzene

HBCDDs

hexabromocyclododecanes

HBCYD

Hexabromocyclodecane

HCTBPH

1,2,3,4,7,7‐hexachloro‐5‐(2,3,4,5‐tetrabromophenyl)‐bicyclo[2.2.1]hept‐2‐ene

HEEHP‐TEBP

2‐(2‐hydroxyethoxy)ethyl 2‐hydroxypropyl 3,4,5,6‐tetrabromophthalate

HepG2

hepatocellular carcinoma cell line

HIPS

high‐impact polystyrene

i.p.

intraperitoneal

IUPAC

International Union of Pure and Applied Chemistry

Kaw

air–water partition coefficient (Henry's Law constant)

Koa

octanol–air partition coefficient

Koc

organic carbon–water partition coefficient

Kow

octanol–water partition coefficient

LB

lower bound

LC

liquid chromatography

LC–APCI‐MS

liquid chromatography–tandem mass spectrometry using atmospheric pressure chemical ionisation

LC–APPI‐MS

liquid chromatography–mass spectrometry using atmospheric pressure photoionisation

LD50

lethal Dose, 50%

LOAEL

lowest‐observed‐adverse‐effect level

LOD

limit of detection

LOQ

limit of quantification

LRAT

long‐range atmospheric transport

MN

micronucleus

mRNA

messenger RNA

MW

molecular weight

NOAEL

no‐observed‐adverse‐effect level

OBTMPI

octabromotrimethylphenyl indane

PBB‐Acr

pentabromobenzyl acrylate

PBBs

polybrominated biphenyls

PBDD/Fs

polybrominated dibenzo‐p‐dioxins and furans (PBDD/Fs)

PBDEs

polybrominated diphenyl ethers

PBEB

pentabromoethylbenzene

PBT

‘persistent, bioaccumulative and toxic’

PBT

pentabromotoluene

PCBs

polychlorinated biphenyls

PLE

pressurised liquid extraction

PND

post natal day

QSAR

quantitative structure–activity relationship

REACH

Registration, Evaluation and Authorisation of Chemicals

RfD

Reference Dose

ROS

reactive oxygen species

SN2

substitution nucleophilic bimolecular

SPE

solid‐phase extraction

T3

triiodothyronine

TBBPA

tetrabromobisphenol A

TBCO

1,2,5,6‐tetrabromocyclooctane

TBNPA

tribromoneopentyl alcohol

TBX

2,3,5,6‐tetrabromo‐p‐xylene

TDBPP

tris(2,3‐dibromopropyl) phosphate

TDBP‐TAZTO

1,3,5‐tris(2,3‐dibromopropyl)‐1,3,5‐triazine‐2,4,6‐trione

TSH

thyroid‐stimulating hormone

TT3

total triiodothyronine

TTBNPP

tris(tribromoneopentyl) phosphate

TTBP‐TAZ

2,4,6‐ tris(2,4,6‐tribromophenoxy )‐1,3,5‐triazine

TTR

transthyretin

UB

upper bound

UF

uncertainty factor

UV

ultraviolet

vPvB

‘very persistent and very bioaccumulative’

ww

whole weight

WWTP

waste water treatment plant

Τ4

thyroxine

REQUESTOR

European Commission

QUESTION NUMBER

EFSA‐Q‐2018‐00436

COPYRIGHT FOR NON‐EFSA CONTENT

EFSA may include images or other content for which it does not hold copyright. In such cases, EFSA indicates the copyright holder and users should seek permission to reproduce the content from the original source.

PANEL MEMBERS

Helle Katrine Knutsen, Agneta Åkesson, Vasileios Bampidis, Margherita Bignami, Laurent Bodin, James Kevin Chipman, Gisela Degen, Antonio Hernández‐Jerez, Tim Hofer, Christer Hogstrand, Stefano Landi, Jean‐Charles Leblanc, Kyriaki Machera, Evangelia Ntzani, Isabelle P. Oswald, Guido Rychen, Salomon Sand, Katharina Vejdovszky and Barbara Viviani.

Supporting information

ANNEX A Protocol for the human risk assessments related to the presence of brominated flame retardants (BFRs) in food

ANNEX D Benchmark dose (BMD) modelling

ANNEX F Outcome of the targeted consultation on the draft scientific Opinion on the update of the risk assessment on emerging and novel BFRs in food

APPENDIX A. Literature searches

TABLE A.1 Details of the literature search

“tris(2,3‐dibromopropyl) phosphate” OR TDBPP OR “N,N′‐ethylenebis(tetrabromophthalimide)” OR EBTEBPI OR Hexabromocyclodecane OR HBCYD OR “bis(2‐ethylhexyl) tetrabromophthalate” OR BEH‐TEBP OR “2‐ethylhexyl 2,3,4,5‐tetrabromobenzoate” OR EH‐TBB OR “dibromoneopentyl glycol” OR DBNPG OR 126‐72‐7 OR 32588‐76‐4 OR 25495‐98‐1 OR 26040‐51‐7 OR 183658‐27‐7 OR 3296‐90‐0 Sources: Web of Science (all databases) and PUBMED Period: 01.01.2011–23.04.2024	23.4.2024	546 results Removal of duplicates: 333
“tris(2,3‐dibromopropyl) phosphate” OR TDBPP OR “N,N′‐ethylenebis(tetrabromophthalimide)” OR EBTEBPI OR Hexabromocyclodecane OR HBCYD OR “bis(2‐ethylhexyl) tetrabromophthalate” OR BEH‐TEBP OR “2‐ethylhexyl 2,3,4,5‐tetrabromobenzoate” OR EH‐TBB OR “dibromoneopentyl glycol” OR DBNPG OR 126‐72‐7 OR 32588‐76‐4 OR 25495‐98‐1 OR 26040‐51‐7 OR 183658‐27‐7 OR 3296‐90‐0 Sources: Web of Science (all databases) and PUBMED Period: 23.04.2024‐03.09.2024	3.9.2024	12 results Removal of duplicates: 7
(following the update of the Terms of Reference) 37853‐59‐1 OR BTBPE OR HxBrPoxE OR bTBPhOEt OR 1,2‐Bis(2,4,6‐tribromophenoxy)ethane OR “1,2‐Bis(2,4,6‐tribromophenoxy) ethane” OR “1,1'‐[1,2‐Ethanediyl‐bis(oxy)] bis[2,4,6‐tribromo]benzene” OR (TBE AND “flame retardant”) 84852‐53‐9 OR DBDPE OR DBDE OR DeBrPylE OR DBDiPhEt OR “Decabromodiphenyl ethane” OR “1,1'‐(1,2‐Ethanediyl) bis[2,3,4,5,6‐pentabromo‐benzene]” OR EBPE 3322‐93‐8 OR DBE‐DBCH OR DiBEt‐DiBcH OR TBECH OR BrCyHx OR 1,2‐Dibromo‐4‐(1,2‐dibromoethyl)‐cyclohexane OR 1‐(1,2‐Dibromoethyl)‐3,4‐dibromocyclohexane OR (TBEC AND “flame retardant”) 51936‐55‐1 OR DBHCTD OR HCDBCO OR Hexachlorocyclopentadienyl‐dibromocyclooctane OR 7,8‐Dibromo‐1,2,3,4,11,11‐hexachloro‐1,4,4a,5,6,7,8,9,10,10a‐decahydro‐1,4‐methanobenzocyclooctane OR 5,6‐Dibromo‐1,10,11,12,13,13‐hexachloro‐11‐tricyclo[8.2.1.02,9]tridecene OR 7,8‐Dibromo‐1,2,3,4,11,11‐hexachloro‐1,4,4a,5,6,7,8,9,10,10a‐decahydro‐1,4‐methanobenzo[8]annulene 87‐82‐1 OR HxBrBz OR Hexabromobenzene OR Perbromobenzene OR (HBB AND “flame retardant”) 34571‐16‐9 OR HCTBPH OR “Dec 604” OR 1,2,3,4,7,7‐Hexachloro‐5‐(2,3,4,5‐tetrabromophenyl)‐bicyclo[2.2.1]hept‐2‐ene OR Hexachlorocyclopentadiene‐tetrabromostyrene OR “Hexachlorocyclopentadiene tetrabromostyrene” OR 5‐(Tetrabromophenyl)‐1,2,3,4,7,7‐hexachloro‐2‐norbornene OR “2‐Norbornene, 1,2,3,4,7,7‐hexachloro‐5‐(tetrabromophenyl)‐(8CI)” 1084889‐51‐9 OR 1025956‐65‐3 OR 893843‐07‐7 OR 155613‐93‐7 OR OBTMPI OR OBIND OR Br‐Indan OR OctaInd OR “Octabromotrimethylphenyl indane” OR 4,5,6,7‐Tetrabromo‐1,1,3‐trimethyl‐3‐(2,3,4,5‐tetrabromophenyl)‐2,3‐dihydro‐1H‐indene OR 2,3‐Dihydro‐1,1,3‐trimethyl‐3‐phenyl‐octabromo‐1H‐indane 59447‐55‐1 OR PBB‐Acr OR PeBrAcr OR “2,3,4,5,6‐Pentabromophenyl‐2‐Propenoic acid methyl ester” OR “Pentabromobenzyl acrylate” OR “2,3,4,5,6‐Pentabromobenzyl acrylate” OR (PBBA AND “flame retardant”) 85‐22‐3 OR PbrEtBz OR Pentabromoethylbenzene OR 1,2,3,4,5‐Pentabromo‐6‐ethylbenzene OR (PBEB AND “flame retardant”) 87‐83‐2 OR PeBT OR 1,2,3,4,5‐Pentabromo‐6‐methylbenzene OR Pentabromotoluene OR Pentabromomethylbenzene OR (PBT AND “flame retardant”) NOT (“proton beam therapy”) 1522‐92‐5 OR TBNPA OR tBMe‐EtOH OR “Tribromoneopentyl alcohol” OR Tribromoneopentanol OR “Pentaerythritol tribromide” OR “Pentaerythritol tribromohydrin” OR 2,2,2‐Tris(bromomethyl)ethanol OR 3‐Bromo‐2,2‐bis(bromomethyl)propanol 52434‐90‐9 OR TDBP‐TAZTO OR BrTriaz OR 1,3,5‐Tris(2,3‐dibromopropyl)‐1,3,5‐triazine‐2,4,6(1H,3H,5H)‐trione OR “Tris(2,3‐dibromopropyl) isocyanurate” OR “Tri(2,3‐dibromopropyl) isocyanurate” OR 1,3,5‐Tris(2,3‐dibromopropyl)‐2,4,6‐trioxohexahydrotriazine OR (TBC AND “flame retardant”) NOT (tuberculosis OR COVID) 3194‐57‐8 OR Tetrabromocyclooctane OR (TBCO AND “flame retardant”) 23488‐38‐2 OR p‐TBX OR TeBDiMeBz OR 2,3,5,6‐Tetrabromo‐p‐xylene OR 2,3,5,6‐Tetrabromo‐1,4‐dimethylbenzene OR 1,4‐Dimethyltetrabromobenzene OR 1,2,4,5‐Tetrabromo‐3,6‐dimethylbenzene OR (TBX AND “flame retardant”) 75795‐16‐3 OR BDBP‐TAZTO OR bDiBPr‐A‐Tazto OR 1,3‐Bis(2,3‐dibromopropyl)‐5‐(2‐propen‐1‐yl)‐1,3,5‐triazine‐2,4,6(1H,3H,5H)‐trione OR 1‐allyl‐3,5‐bis(2,3‐dibromopropyl)‐1,3,5‐triazinane‐2,4,6‐trione OR 1,3‐Bis(2,3‐dibromopropyl)‐5‐allyl‐1,3,5‐triazine‐2,4,6(1H,3H,5H)‐trione 57829‐89‐7 OR DBP‐TAZTO OR DiBPr‐DiA‐Tazto OR 1‐(2,3‐Dibromopropyl)‐3,5‐diallyl‐1,3,5‐triazine‐2,4,6(1H,3H,5H)‐trione OR 1‐(2,3‐Dibromopropyl)‐3,5‐di‐2‐propenyl‐1,3,5‐triazine‐2,4,6(1H,3H,5H)‐trione	3.9.2024	3304 results in total Removal of duplicates: 1506
31780‐26‐4 OR DiBSty OR Dbrsty OR Dibromoethenylbenzene OR Dibromostyrene OR (DBS and “flame retardant”) 20566‐35‐2 OR HEEHP‐TEBP OR OHEtOEt‐OHPr‐TeBPht OR TeBrPht OR “Phthalic acid, tetrabromo‐, 2‐(2‐hydroxyethoxy)ethyl 2‐hydroxypropyl ester” OR “2‐(2‐Hydroxyethoxy)ethyl 2‐hydroxypropyl 3,4,5,6‐tetrabromophthalate” OR “3,4,5,6‐Tetrabromo‐1,2‐benzenedicarboxylic acid 1‐[2‐(2‐hydroxyethoxy)ethyl] 2‐(2‐hydroxypropyl) ester” 58965‐66‐5 OR 4'‐PeBPO‐BDE208 OR TeDB‐DiPhOBz OR DpeTeBrBz OR Tetradecabromo‐1,4‐diphenoxybenzene OR “Pentabromophenoxy‐nonabromodiphenyl ether” OR 1,4‐Bis(pentabromophenoxy)tetrabromobenzene OR 1,2,4,5‐Tetrabromo‐3,6‐bis(2,3,4,5,6‐pentabromophenoxy)‐benzene OR 1,2,4,5‐Tetrabromo‐3,6‐bis(pentabromophenoxy)‐benzene 19186‐97‐1 OR tBbBMePrP OR TrisPhos OR TTBNP OR “Tris(tribromoneopentyl) phosphate” OR 3‐Bromo‐2,2‐bis(bromomethyl)‐1‐Propanol 1,1',1”‐phosphate OR “2,2‐Bis‐(bromomethyl)‐3‐bromo‐1‐propanol phosphate” 25713‐60‐4 OR TTBP‐TAZ OR tTrBPhO‐Taz OR BrPhTriaz OR 2,4,6‐Tris(2,4,6‐tribromophenoxy)‐1,3,5‐triazine OR Tris(tribromophenoxy)‐s‐triazine OR “Tri(tribromophenyl) cyanurate” OR Tris(2,4,6‐tribromophenoxy)‐s‐triazine Sources: Web of Science (all databases) and PUBMED Exclusion criteria: meetings and only abstract, PHD thesis Language: English only Period: 1.1.2011–3.9.2024
Same search strings as above all substances Sources: Web of Science (all databases) and PUBMED Language: English only Period: 3.9.2024‐2.5.2025	2.5.2025	107 results in total Removal of duplicates: ‐‐‐
Same search strings as above all substances Sources: Web of Science (all databases) and PUBMED Language: English only Period: 2.5.2025‐18.12.2025	18.12.2025	115 results in total Removal of duplicates: 87

APPENDIX B. Occurrence of the BFRs assessed in this Opinion in food samples from European countries as reported in the literature

TABLE B.1 Literature data on occurrence in food samples from European countries of the BFRs assessed in this Opinion

Country	Year	Food matrices	Food matrices BFRs detected (detection frequency)	Mean (range)	Exposure estimates	Reference
General food categories
Portugal, Spain, Slovenia	2010–2012	Honey samples: Spain (8), Portugal (4) and Slovenia (4) n = 16	BTBPE: Spain Portugal Slovenia DBDPE: Spain Portugal Slovenia	BTBPE: 0.06 pg/g ww < LOD pg/g ww < LOD pg/g ww DBDPE: 0.32 pg/g ww 0.87 pg/g ww < LOD pg/g ww	NR	Mohr et al. ( 2014 ). Levels of brominated flame retardants (BFRs) in honey samples from different geographic regions.
Sweden	2010	Food baskets comprising: fish, meat, vegetable oils, dairy products and eggs n = 4 food baskets	DBDPE: All samples (0%) BTBPE: Fish (25%) Dairy pr. (25%) Eggs (50%) EH‐TBB: All samples (0%) BEH‐TEBP: All samples (0%) α‐DBE‐DBCH: Fish (100%) β‐DBE‐DBCH: Fish (100%) TBX: Fish (100%) PBT: Fish (100%) PBEB: Fish (100%) Meat (50%) HBB: Fish (25%) Meat (25%)	DBDPE: < LOD (pg/g ww) BTBPE: 1.1–3.6 (pg/g ww) 1.1–3.6 (pg/g ww) 3.9 (pg/g ww) EH‐TBB: < LOD (pg/g ww) BEH‐TEBP: < LOD (pg/g ww) α‐DBE‐DBCH: 97 (pg/g ww) β‐DBE‐DBCH: 17 (pg/g ww) TBX: 20 (pg/g ww) PBT: 1.8 (pg/g ww) PBEB: 8.7 (pg/g ww) 0.60–2.0 (pg/g ww) HBB: 23 (pg/g ww) 0.95–3.1 (pg/g ww)	Estimated median dietary daily intake (ng per day): Mothers/toddlers (n = 20) based on fish intake only: α‐DBE‐DBCH: 3.5/1.7 β‐DBE‐DBCH: 0.6/0.3 TBX: 0.7/0.3 PBT: 0.07/0.03 PBEB: 0.3/0.2	Sahlström et al. ( 2015 ). Estimated intakes of brominated flame retardants via diet and dust compared to internal concentrations in a Swedish mother–toddler cohort.
UK	2013	Eggs, processed meats, poultry and carcase meats, milk and dairy products, offals and shellfish, processed foods (of vegetable origin) and others including fruit, vegetable and bread Total n = 297	PBT TBX PBEB DBHCTD HCTBPH PBB‐Acr OBTMPI DBDPE BTBPE HBB	< LOD (0.01 ng/g ww)	NR	Fernandes et al. ( 2016 ). Bromine content and brominated flame retardants in food and animal feed from the UK
Spain	2010‐2012	Cow's milk (n = 3) Soy products (n = 3) Eggs (n = 3) Biscuits (n = 2) Fish oil dietary supplement (n = 6) Vegetable oil diet supplement (n = 1) Total n = 18	BTBPE: Cow's milk Soy products Eggs Biscuits Fish oil diet suppl Veg oil diet suppl DBDPE: Cow's milk Soy products Eggs Biscuits Fish oil diet suppl Veg oil diet suppl	BTBPE: 0.35 pg/g ww 0.039 pg/g ww 0.43 pg/g ww 0.60 pg/g ww 17 pg/g ww < LOD pg/g ww DBDPE: 0.0036 pg/g ww < LOD pg/g ww 24 pg/g ww 0.041 pg/g ww 8.2 pg/g ww 13 pg/g ww	NR for the compounds in this Opinion (only the sum of BFRs is estimated based on fish oil consumption only, where BTBPE and DBDPE had negligible contribution)	García‐Bermejo et al. ( 2016 ). Applicability of gas chromatography (GC) coupled to triple‐quadrupole (QqQ) tandem mass spectrometry (MS/MS) for polybrominated diphenyl ether (PBDE) and emerging brominated flame retardant (BFR) determinations in functional foods enriched in Omega‐3.
UK	2015	Meat (n = 12 samples of beef, lamb, chicken, pork) Liver (n = 6 samples of beef, lamb, chicken, pork liver) Fish (n = 9 samples of salmon, tuna, trout, mackerel) Eggs (n = 1) Cheese (n = 2) Ttal n = 30	α‐DBE‐DBCH: Meat Liver Fish Eggs Cheese β‐DBE‐DBCH: Meat Liver Fish Eggs Cheese EH‐TBB: Meat Liver Fish Eggs Cheese	α‐DBE‐DBCH: 0.58–0.95 ng/g lw < 0.26–1.6 ng/g lw 0.48–1.3 ng/g lw 0.42 ng/g lw 0.44 ng/g lw β‐DBE‐DBCH: 1.9–7.6 ng/g lw 34–85 ng/g lw 4.4–39 ng/g lw 3.0 ng/g lw 0.99 ng/g lw EH‐TBB: 0.19–1.4 ng/g lw 0.19–1.6 ng/g lw 0.22–0.43 ng/g lw 0.10 ng/g lw 0.11 ng/g lw	Consumption data: National diet and nutrition survey report published by Public Health England and the Food Standards Agency (2014) Exposure estimations NR for the compounds in this Opinion (only the sum all BFRs: Average: 26–89 ng per day High‐end*: 162–467 ng per day * High‐end estimations were made assuming high‐end food intakes of mean + 2SD	Tao et al. ( 2017 ). Emerging and legacy flame retardants in UK human milk and food suggest slow response to restrictions on use of PBDEs and HBCDD.
			BTBPE: Meat Liver Fish Eggs Cheese BEH‐TEBP: Meat Liver Fish Eggs Cheese DBDPE: Meat Liver Fish Eggs Cheese	BTBPE: < 0.040–2.4 ng/g lw < 0.006–0.35 ng/g lw < 0.040–0.78 ng/g lw 0.18 ng/g lw 0.20 ng/g lw BEH‐TEBP: 0.20–0.57 ng/g lw 0.69–5.8 ng/g lw < 0.10–1.1 ng/g lw 1.8 ng/g lw 0.22 ng/g lw DBDPE: < 1.1–4.6 ng/g lw < 1.6–7.6 ng/g lw < 0.63–21 ng/g lw < 1.2 ng/g lw < 0.74 ng/g lw
Norway	2013–2014	Whole‐day food samples n = 61	BEH‐TEBP All samples (44%)	BEH‐TEBP 0.012 ng/g ww	95th daily dietary intake estimate for adult population: BEH‐TEBP: 1.2 ng/kg bw/day	Xu et al. ( 2017 ). Assessment of dietary exposure to organohalogen contaminants, legacy and emerging flame retardants in a Norwegian cohort.
Belgium	2015–2016	Fish and fish products (n = 61) Meat and meat products (n = 35) Milk and dairy products (n = 38) Food for infants and small children (n = 18) Animal and vegetable fat (n = 9) Grains and grain‐based products (n = 7) Eggs and egg products (n = 4) Potatoes and derived products (n = 4) Other food (n = 4) Food supplements (n = 1) Vegetables (n = 2) Total n = 183	BEH‐TEBP: Meat and meat products Animal and vegetable fat Other categories EH‐TBB: All categories HBB: All categories BTBPE: All categories	LOQs: BEH‐TEBP and EH‐TBB: 200 pg/g ww, HBB: 20 pg/g ww, BTBPE: 10 pg/g ww BEH‐TEBP: 6 pg/g ww 72 pg/g ww < LOQ EH‐TBB: < LOQ HBB: < LOQ BTBPE: < LOQ	NR for the BFRs assessed in this Opinion	Poma et al. ( 2018 ). Occurrence of selected halogenated flame retardants in Belgian foodstuff.
France	2014‐2016	Fish (n = 114) Crustaceous/molluscs (n = 154) Milk (n = 72) Eggs (n = 57) Sheep liver (n = 28) Meat (n = 152) Total n = 577	PBEB All categories (< 14%) PBT: Fish (71%) Crustaceous/molluscs (45%) Milk (75%) Eggs (75%) Sheep liver (61%) Meat (72%) HBB: Fish (65%) Crustaceous/molluscs (57%) Milk (72%) Eggs (68%) Sheep liver (57%) Meat (75%) TBX: Crustaceous/molluscs (73%) Other categories (< 43%) EH‐TBB: All categories (< 18%) BTBPE: Crustaceous/molluscs (50%) Other categories (< 17%)	PBEB < LOD pg/g ww PBT: 0.174 pg/g ww < LOD pg/g ww 0.139 pg/g ww 0.236 pg/g ww 0.214 pg/g ww 0.165 pg/g ww HBB: 0.363 pg/g ww 0.257 pg/g ww 0.321 pg/g ww 0.617 pg/g ww 0.633 pg/g ww 0.701 pg/g ww TBX : 0.123 pg/g ww < LOD pg/g ww EH‐TBB: < LOD pg/g ww BTBPE: 0.050 pg/g ww < LOD pg/g ww	NR	Vénisseau et al. ( 2018 ). Occurence of legacy and novel brominated flame retardants in food and feed in France for the period 2014 to 2016.
Latvia	2016–2019	Meat (n = 8) Eggs (n = 8) Milk and dairy products (n = 8) Fish and fish products (n = 8) Samples of plant origin (n = 8) Total n= 40	DBE‐DBCH: Meat Eggs Milk and dairy Fish and fish Plant origin PBEB: Meat Eggs Milk and dairy Fish and fish Plant origin HBB: Meat Eggs Milk and dairy Fish and fish Plant origin	DBE‐DBCH: < 3.70–34.6 ng/kg ww 6.46–25.0 ng/kg ww < 2.60–60.9 ng/kg ww < 3.90–132 ng/kg ww < 6.30–28.6 ng/kg ww PBEB: 0.34–5.66 ng/kg ww < 0.43–1.85 ng/kg ww 0.11–2.9 ng/kg ww 2.20–3.13 ng/kg ww 0.52–19.0 ng/kg ww HBB: 1.38–9.82 ng/kg ww 0.93–5.71 ng/kg ww 1.15–21.8 ng/kg ww 2.85–17.8 ng/kg ww 3.7–19.5 ng/kg ww	NR for the compounds in this Opinion (only the Sum BFRs: Average: 0.47 ng/kg bw per day)	Zacs et al. ( 2021 ). Polybrominated diphenyl ethers (PBDEs), hexabromocyclododecanes (HBCDD), dechlorane‐related compounds (DRCs) and emerging brominated flame retardants (EBFRs) in foods: The levels, profiles and dietary intake in Latvia.
UK	2020–2021	Meat (36 samples of beef; lamb; chicken; pork; turkey) Fish/Prawns (42 samples of salmon; prawn; cod; tuna; basa; trout; tilapia; mackerel) Cheese (15 samples) Chicken eggs (15 samples) n = 108	PBT: Meat Fish/Prawns Cheese Chicken eggs PBEB: Meat Fish/Prawns Cheese Chicken eggs HBB: Meat Fish/Prawns Cheese Chicken eggs EH‐TBB: Meat Fish/Prawns Cheese Chicken eggs BTBPE: Meat Fish/Prawns Cheese Chicken eggs BEH‐TEBP: Meat Fish/Prawns Cheese Chicken eggs DBDPE: Meat Fish/Prawns Cheese Chicken eggs	PBT: < 0.091 ng/g lw (< 3.1 pg/g ww) < 0.26 ng/g lw (< 2.9 pg/g ww) < 0.024 ng/g lw (< 5.9 pg/g ww) < 0.034 ng/g lw (< 2.3 pg/g ww) PBEB: < 0.037 ng/g lw (< 1.3 pg/g ww) < 0.10 ng/g lw (< 1.2 pg/g ww) < 0.0099 ng/g lw (< 2.4 pg/g ww) < 0.014 ng/g lw (< 0.94 pg/g ww) HBB: < 0.041 ng/g lw (< 1.4 pg/g ww) < 0.12 ng/g lw (< 1.30 pg/g ww) < 0.011 ng/g lw (< 2.7 pg/g ww) < 0.015 ng/g lw (< 1.0 pg/g ww) EH‐TBB: 0.63 ng/g lw (43 pg/g ww) 3.3 ng/g lw (17 pg/g ww) < 0.073 ng/g lw (< 18 pg/g ww) < 0.10 ng/g lw (< 6.9 pg/g ww) BTBPE: 4.3 ng/g lw (190 pg/g ww) 22 ng/g lw (200 pg/g ww) 0.39 ng/g lw (97 pg/g ww) 36 ng/g lw (2300 pg/g ww) BEH‐TEBP: 8.5 ng/g lw (510 pg/g ww) 20 ng/g lw (390 pg/g ww) 0.52 ng/g lw (140 pg/g ww) 2.0 ng/g lw (130 pg/g ww) DBDPE: 0.38 ng/g lw (18 pg/g ww) 1.2 ng/g lw (45 pg/g ww) < 0.31 ng/g lw (< 51 pg/g ww) < 0.31 ng/g lw (< 21 pg/g ww)	Consumption data: Obtained from the National Diet and Nutrition Survey (2008–2019) Exposure estimates: EH‐TBB: Average: 0.053–0.19 ng/kg bw per day High‐end*: 0.31–1.2 ng/kg bw per day BTBPE: Average: 1.2–4.4 ng/kg bw per day High‐end*: 7.1–27 ng/kg bw per day BEH‐TEBP: Average: 0.91–3.3 ng/kg bw per day High‐end*: 5.3–20 ng/kg bw per day DBDPE: Average: 0.066–0.24 ng/kg bw per day High‐end*: 0.39–1.5 ng/kg bw per day * High‐end estimations made assuming high‐end food intakes of mean+ 2 SD	Ma et al. ( 2023 ). Temporal trends in concentrations of brominated flame retardants in UK foodstuffs suggest active impacts of global phase‐out of PBDEs and HBCDD.
Fish and seafood
Switzerland	2018	Lake fish species: Roach (n = 20), Burbot (n = 11) Total  n= 31	LOQ = 0.005 ng/g ww BTBPE Roach (25%) Burbot (9%) DBDPE Roach (100%) Burbot (91%) HBB Roach (25%) Burbot (0%) PBB Roach (5%) Burbot (0%)	BTBPE < LOQ–0.01 ng/g ww < LOQ–0.01 ng/g ww DBDPE < LOQ–1.54 ng/g ww < LOQ–2.89 ng/g ww HBB < LOQ–0.02 ng/g ww < LOQ ng/g ww PBB < LOQ–0.005 ng/g ww < LOQ ng/g ww	NR	Babut et al. ( 2021 ). Legacy and alternative halogenated flame retardants in Lake Geneva fish.
Baltic Sea – North Sea	2015	Blue mussels soft body (BM), Eelpouts filet (EF) n = NR	EH‐TBB: BM EF BEH‐TEBP: BM EF BTBPE: BM EF DBDPE: BM EF HBB: BM EF PBEB: BM EF PBT: BM EF	EH‐TBB: < 0.0125–< 0.0143 (ug/kg ww) < 0.0117–< 0.0126 (ug/kg ww) BEH‐TEBP: < 0.0384–0.0559 (ug/kg ww) < 0.0404–0.0728 (ug/kg ww) BTBPE: < 0.0083–< 0.0106 (ug/kg ww) < 0.0087–< 0.0094 (ug/kg ww) DBDPE: < 0.0751–0.109 (ug/kg ww) < 0.0789–< 0.0855 (ug/kg ww) HBB: < 0.0008–< 0.0011 (ug/kg ww) < 0.0009–< 0.001 (ug/kg ww) PBEB: < 0.0003 (ug/kg ww) < 0.0003 (ug/kg ww) PBT: 0.004–0.0067 (ug/kg ww) 0.0072–0.013 (ug/kg ww)	NR	Fliedner et al. ( 2020 ). Chemicals of emerging concern in marine specimens of the German Environmental Specimen Bank
Spain	NR	Mussels n = 7	TDBPP: (0%)	TDBPP: < LOD (9 ng/g fat)	NR	Castro et al. ( 2020 ). Determination of 18 organophosphorus flame retardants/plasticizers in mussel samples by matrix solid‐phase dispersion combined to liquid chromatography–tandem mass spectrometry.
Albania, Greece, Italy, Norway, Portugal, Spain and UK	2016	Mussels n = 17	LOD range (0.19–19.3 ng/g lw)	DBDPE: < LOD–4.35 (ng/g lw) HBB: < LOD (ng/g lw) PBEB: < LOD (ng/g lw)	NR	Aznar‐Alemany et al. ( 2018 ). Halogenated and organophosphorus flame retardants in European aquaculture samples
EU countries	2014–2015	Various fish species: cod (3), mackerel (11), monkfish (4), mussel (10), plaice (1), salmon (3), seabream (2) and tuna (4) n = 38	DBDPE : All samples (0%) LOD/LOQ: 9.66/32.2 (ng/g lw) HBB (43%): LOD/LOQ: 0.20/0.67 (ng/g lw) Cod Mackerel Monkfish Mussel Plaice Salmon Seabream Tuna PBEB (50%): All samples LOD/LOQ: 0.18/0.61 (ng/g lw)	DBDPE: < LOD HBB: 10.7 (ng/g lw) 0.04 (ng/g ww) 0.68 (ng/g lw) 0.13 (ng/g ww) 17.6 (ng/g lw) 0.04 (ng/g ww) 2.76 (ng/g lw) 0.04 (ng/g ww) 8.61 (ng/g lw) 0.08 (ng/g ww) < LOD 65.8 (ng/g lw) 1.6 (ng/g ww) 1.23 (ng/g lw) 0.07 (ng/g ww) PBEB: < LOQ	NR for the BFRs assessed in this Opinion Lipid content % average (0.4% for Cod, 1.5% for mussels, 19% for mackerel, 0.2% for monkfish, 0.9% for plaice, 2.5% for seabream, 6% for tuna) Formula Cww = (Clw × lw%)/100	Aznar‐Alemany et al. ( 2017 ). Occurrence of halogenated flame retardants in commercial seafood species available in European markets.
Spain	NR	n = 10 fish species (sole, hake, sardine, tuna, codfish, shrimp, salmon, mackerel, squid and mussel)	PBEB: All species HBB: All species DBDPE: All species	PBEB: < 0.01–< 0.02 ng/g ww HBB: < 0.01–0.2 ng/g ww DBDPE: < 0.004–< 0.01 ng/g ww	Estimated dietary intake (ng/kg bw/day) for the general population are shown in a graph, including the sum of the PBDEs and MeO‐PBDEs in addition to the 3 congeners here considered.	Trabalón et al. ( 2017 ). Human exposure to brominated flame retardants through the consumption of fish and shellfish in Tarragona County (Catalonia, Spain).
Italy, Greece, Slovenia, Croatia, Bosnia and Herzegovina and Serbia	2014–2015	Various fish species from 3 rivers (Adige, Evrotas, Sava) Total n = 27 samples	LOD: 0.002–19.3 ng/g lw HBB, PBEB, DBDPE: Adige (0%) Evrotas (0%) Sava (10%)	Sum of HBB, PBEB, DBDPE: ND ND 2.94 ng/g lw	NR	Giulivo et al. ( 2017 ). Occurrence of halogenated and organophosphate flame retardants in sediment and fish samples from three European river basins.
France	2011–2012	Mussels and oysters n total = NR	LOQ: 0.02–0.84 pg/g ww HBB: All samples BTBPE: All samples DBDPE: All samples	HBB: 0.1–2.3 pg/g ww BTBPE: 0.7–37.3 pg/g ww DBDPE: < 0.8–20.1 pg/g ww	NR	Munschy et al. ( 2015 ). Occurrence of legacy and emerging halogenated organic contaminants in marine shellfish along French coasts.
Italy	2011–2012	Shad and whitefish (lake fish) n total = 123	LOD = 0.1 ng/g dry weight Lipid content average (20%)	PBEB: < LOD HBB: < LOD–2.4 ng/g lw BTBPE: 0.2–25.3 ng/g lw DBDPE: < LOD	NR	Poma et al. ( 2014 ). Concentrations and trophic interactions of novel brominated flame retardants, HBCD and PBDEs in zooplankton and fish from Lake Maggiore (Northern Italy).
Spain	2010	Various fish species from rivers n total = 48	PBEB: All samples HBB: All samples DBDPE: All samples	PBEB: ND (LOD/LOQ: 179/596 pg/g lw) HBB: ND (LOD/LOQ: 133/444 pg/g lw) DBDPE: < LOD–130 pg/g lw (LOD/LOQ: 2050/6834 pg/g lw)	NR	Santín et al. ( 2013 ). Emerging and historical halogenated flame retardants in fish samples from Iberian rivers.
Germany	NR	Eel (yellow, silver) n total = 80	PBEB: All samples BEH‐TEBP: All samples	PBEB: < LOD–0.1 ng/g ww (LOD: 0.062–0.12 ng/g ww) BEH‐TEBP: < LOD–0.027 ng/g ww (LOD: 0.008–0.016 ng/g ww)	NR	Sühring et al. ( 2013 ). Brominated flame retardants and dechloranes in eels from German Rivers.

Abbreviations: LOD/LOQ, limit of detection/quantification; lw, lipid weight; ND, not detected; NR, not reported; ww, wet weight.

APPENDIX C. Toxicity studies in experimental animals

Table C.1.

Studies in experimental animals on the six BFRs covered in Commission recommendation 2014/118/EU (TDBPP, EBTEBPI, BEH‐TEBP, EH‐TBB, DBNPG) and on DBDPE

Compound Purity	Species Number of animals Age and sex	Route of administration Exposure doses Study duration	Outcome	NOAEL/LOAEL	Reference
TDBPP (Section 3.2.1 )
TDBPP Purity: NR Michigan Chemical Corporation, under the trade name Firemaster LV‐T23	Rats (F344) 55 M and 55 F per dose group 6 weeks old	Oral exposure (feed) 0, 50, 100 mg/kg feed (corresponding to 0, 2 and 4 mg/kg bw per day in M and 0, 2.5 and 5 mg/kg bw per day in F) Repeated exposure: 103 weeks + additional observation period of 1 or 2 weeks.	No significant association between increased dosage and mortality. Kidney: ‐ Increased incidence of tubular‐cell adenoma (M: 0%, 48%, 48%, respectively; F: 0, 7%, 19%, respectively). For M, significant positive association between incidence of kidney tubular‐cell adenocarcinoma and TDBPP (6% in high‐dose M). ‐ Non‐neoplastic lesions in renal tubules: In 11% of high‐dose M and 65% of high‐dose F, a few tubular cells were slightly enlarged and showed nuclear dysplasia consisting of nuclear enlargement, chromatin clumping and parachromatin clearing.	LOAEL = 2 mg/kg bw per day	NTP ( 2014 ), Reznik et al. ( 1979 )
TDBPP Purity: NR Firemaster LV‐T23 (Michigan Chemical Corporation)	Mice (F344) 50 M and 50 F per dose group 6 weeks old	Oral exposure (feed) 0, 500, 1000 mg/kg feed (corresponding to 0, 60, 120 mg/kg bw per day in M, and to 0, 65, 130 mg/kg bw per day in F) Repeated exposure: 103 weeks + additional observation period of 1 or 2 weeks.	No significant association between increased dosage and accelerated mortality. Kidney: – Increased incidence of tubular‐cell adenoma (M: 0%, 6% and 18%, respectively; F: 0%, 4% and 4%, respectively). – Renal tubular‐cell adenocarcinoma observed in M (0%, 2% and 10%, respectively) but not in F. Variable amounts of haemorrhage, necrosis and calcification were associated with renal tubular‐cell carcinoma. – Renal tubular dysplasia observed in M (74% of low dose and 61% of high dose) and in F (2% of low dose and 26% of high dose). Forestomach: – Increased incidence of squamous‐cell papilloma (M: 0%, 21% and 23%, respectively; F: 2%, 21%, 41%, respectively). Carcinoma of the forestomach observed in 4% of the high‐dose M and 8% and 9% of the treated F. Chronic inflammation frequently found in the submucosa below the papillary lesion. The squamous‐cell carcinoma of the forestomach also often showed superficial hyperkeratinisation. – A basal cell carcinoma of the forestomach seen in a highest dose M and a leiomyosarcoma of the stomach wall in a highest dose F. There was one squamous‐cell carcinoma of the oesophagus in a high‐dose F. Liver: – Dose‐related increase in hepatocellular adenoma or carcinoma in F mice (11%, 46% and 71%, respectively). In M mice, no such dose‐related increase in hepatocellular neoplasms was apparent. Lung: – Bronchiolar/alveolar adenoma and carcinoma seen in M (22%, 41% and 50%, respectively) and in F (7%, 18% and 34%, respectively). Cystadenoma of the Harderian gland in F (0%, 8% and 4%, respectively) and in M (2%, 2% and 4%, respectively).	LOAEL = 60 mg/kg bw per day	NTP ( 2014 ), Reznik et al. ( 1979 )
BEH‐TEBP (Section 3.2.4 )
BEH‐TEBP Purity: NR	Rats (Harlan Sprague–Dawley) 3‐4 rats per dose group Adult pregnant F	Repeated exposure Oral (gavage) Vehicle: corn oil 0, 750 mg/kg per day From GD14 to GD18	Fetal testosterone production unaltered at GD18.	‐	Furr et al. ( 2014 )
BEH‐TEBP Purity: NR	Rats (Harlan Sprague‐ Dawley) 3 rats per dose group Adult pregnant F Dams mated on PND90	Repeated exposure Oral (gavage) Vehicle: corn oil 0, 750 mg/kg per day From GD17 to GD21	Exposure did not induce formation of multinucleated germ cells in the testes of fetuses at GD21.	–	Spade et al. ( 2018 )
BEH‐TEBP Purity: 95% OECD TG408 GLP	Rats (Wistar Han) 10 rats per dose group 6‐8 weeks M, F	Repeated exposure Oral (gavage) 0, 100, 300, 1000 mg/kg bw per day Vehicle: Arachis oil 90‐day exposure	No treatment‐related effects in the wide range of observations, including mortality, body and organ weights, behavioural assessments, functional performance tests of motor activity and grip strength, sensory reactivity, ophthalmology, haematology, blood chemistry or histopathology.	NOAEL = 1000 mg/kg bw per day	Fulcher ( 2018 )
BEH‐TEBP Purity: 99%	Rats (Harlan Sprague–Dawley) 6 rats per dose group 7‐week‐old M	Repeated exposure Oral (gavage) 0, 0.1, 1, 10, 100, 1000 μmol/kg bw per day (equal to 0, 0.07, 0.71, 7.1, 71, 706 mg/kg bw per day) Vehicle: corn oil 5‐day exposure	No dose‐related effects on survival, body weight, gross pathology, organ weights or histopathology or the liver or thyroid gland or levels of thyroid hormones. No transcriptome changes associated with oxidative stress, metabolic pathway transcripts or liver disease.	NOAEL = 706 mg/kg bw per day	NTP ( 2019b ), Shockley et al. ( 2020a )
BEH‐TEBP Purity: NR	Mice (Balb/c) 8 mice per dose group 5‐week‐old M	Repeated exposure Oral (gavage) 0, 200 mg/kg bw per day Vehicle: corn oil 28 days exposure	BEH‐TEBP resulted in hepatocyte oedema, narrowing of the sinusoids and dilation of the central vein. Levels of ROS, MDA, 8‐OH‐dG and DNA‐protein cross‐links increased, and levels of GSH decreased. Western blotting analysis showed upregulation of the endoplasmic reticulum stress related proteins GRP78, PERK and CHOP. The effects of BEH‐TEBP were attenuated by co‐administration of 25 mg/kg bw per day curcumin, which was attributed to its antioxidant properties	LOAEL = 200 mg/kg bw per day (only dose tested)	Yin and Zhang et al. (2021)
BEH‐TEBP Purity: NR	Mice (Balb/c) 8 mice per dose group 5‐week‐old M	Repeated exposure Oral 0, 20, 200 mg/kg bw per day Vehicle: water saline solution 28 days exposure	No changes in spatial learning and memory performances (Morris water maze). The behavioural effects of BEH‐TEBP remained the same in animals co‐administered with 25 mg/kg bw per day curcumin, which was attributed to its antioxidant properties.	NOAEL = 20 mg/kg bw per day (two doses tested)	Zhang and Yin ( 2021 )
BEH‐TEBP Purity: 95%	Rats (Harlan Sprague–Dawley) 12 rats per dose group 6‐week‐old M	Repeated exposure Oral (gavage) 0, 5, 50, 500 mg/kg bw per day Vehicle: corn oil 28 days exposure	No changes in body weight, blood biochemical parameters and lipid metabolism. Thyroids and relative thyroid weights in top dose animals were smaller. Also in the top dose rats, thyroid follicular epithelium was flatter and thinner and the follicular cavity was dilated. TT4 and FT4 were significantly increased in the mid and top dose animals in a dose‐related manner. TSH was decreased at the top dose. No changes in TT3 or FT3.	NOAEL = 5 mg/kg bw per day	Yuan et al. ( 2024 )
EH‐TBB (Section 3.2.5 )
EH‐TBB Purity: 97.9%	Rats (Harlan Sprague–Dawley) 6 rats per dose group 7‐week‐old M	Repeated exposure Oral (gavage) 0, 0.1, 1, 10, 100, 1000 μmol/kg bw per day, equal to 0, 0.05, 0.55, 5.5, 55 and 550 mg/kg bw per day. Vehicle: corn oil 5‐day exposure	No statistically significant treatment‐related effects on mortality, body weight or liver lesions after 5 days of exposure. Liver transcriptomics showed elevated CYP1A1 and CYP1A2 at the top two doses, and activation of liver disease transcripts at the top dose.	NOAEL = 550 mg/kg bw per day	Shockley et al. ( 2020a )
Binary mixtures BEH‐TEBP and EH‐TBB (Section 3.2.4. and 3.2.5 )
Binary mixture BEH‐TEBP and EH‐TBB (68.1% EH‐TBB and 31.9% BEH‐TEBP)	Rats (Wistar) 8 per dose group c. 300g Pregnant F	Repeated exposure Oral (on a wafer cookie) 0, 1000 ug/day Estimated mass fraction in FM‐550 was EB‐TBB 29.7% and BEH‐TEBP 13.9%, corresponding to EH‐TBB and BEH‐TEBP doses in the binary mixture experiment of 2.25 mg/kg bw per day EH‐TBB and 1.05 mg/kg bw per day BEH‐TEBP GD0–PND21 Vehicle: sesame oil 1M and 1F pup per litter randomly selected for behavioural tasks from PND65 to PND247‐261	Dams had more female pups that controls. Decreased bw of M and F pups at most time points. Increase relative liver and pituitary weight in M pups but not F pups. F offspring showed decreased latency to enter the centre of an open field and increased numbers of marbles buried. No significant effects in M pups. No significant effects of exposure on sociability with a stranger rat. Running wheel activity in the light period was elevated in M offspring. Authors concluded that developmental exposure to the binary mixture BEH‐TEBP+EH‐TBB affected socioemotional behaviour. Effects were exposure, sex‐ and task‐specific. F offspring showed more deficits than M.	–	Witchey et al. ( 2020 )
Binary mixture BEH‐TEBP and EH‐TBB (68.1% EH‐TBB and 31.9% BEH‐TEBP)	Rats (Adult Wistar, 8 months) Dams/pups Controls: N = 8 M, 8 F OPFR: N = 8 M, 8 F BFR: N = 8 M, 8 F FM 550: N = 8 M, 10 F	Oral 0, 1000 μg/day (~3.3 mg/kg bw per day) Estimated mass fraction in FM‐550 was EH‐TBB 29.7%, BEH‐TEBP 13.9%, corresponding to EH‐TBB and BEH‐TEBP doses in the binary mixture experiment of 2.25 mg/kg bw per day EH‐TBB and 1.05 mg/kg bw per day BEH‐TEBP, respectively. Vehicle: sesame oil Repeated exposure Dams dosed from 72 h after mate pairing until PND21	In F: reduction in Bone Mineral Density (BMD). In M: Bone Volume Fraction was decreased in distal femur and an increase in ↑ Serum calcium.	–	Schkoda et al. ( 2023 ).
DBNPG (Section 3.2.6 )
DBNPG Purity: technical product FR‐1138 (containing 78.6% DBNPG, 6.6% 2,2‐bis(hydroxymethyl)‐l‐bromo‐3‐hydroxypropane, 6.9% 2,2‐bis(bromomethyl)‐l‐bromo‐3‐hydroxypropane, 0.2% pentaerythritol and 7.7% dimers and structural isomers)	Rats (F344/N) 60 M and 60 F per dose group 6 weeks old	Oral exposure (feed) 0, 2500, 5000 or 10,000 mg/kg feed (Average daily doses of FR‐1138 were 0, 100, 200 or 430 mg/kg bw for M and 0, 115, 230 or 460 mg/kg bw for F) Repeated exposure 2 years	Survival at 2 years of M and F rats in the two highest dose groups was significantly lower than controls. In both M and F, significantly increased incidences of neoplasms in various organs were observed, with M exhibiting a wider range of affected organs than F. In M: mammary gland, oral cavity, oesophagus, forestomach, small intestine, large intestine, skin neoplasms, subcutaneous tissues, all organs malignant mesothelioma, kidney, urinary bladder, lung, thyroid gland, pancreas, haematopoietic system and Zymbals’ gland. In F: mammary gland, oral cavity, oesophagus and thyroid gland. For more details See Table 14. Non‐neoplastic effects were observed: in M at several doses, in the kidney (focal atrophy, papillary degeneration, transitional epithelial hyperplasia (pelvis) and papillary epithelial hyperplasia), lung (alveolar/bronchiolar hyperplasia), thyroid gland (follicular cell hyperplasia), seminal vesicle (hyperplasia), pancreas (hyperplasia), urinary bladder (hyperplasia) and forestomach (mucosal hyperplasia) and in F, in the kidney (focal atrophy, papillary degeneration and papillary epithelial hyperplasia) at the highest dose only.	LOAEL = 100 mg FR‐1138/kg bw per day	NTP ( 1996 )
DBNPG Purity: technical product FR‐1138 (containing 78.6% DBNPG, 6.6% 2,2‐bis(hydroxymethyl)‐l‐bromo‐3‐hydroxypropane, 6.9% 2,2‐bis(bromomethyl)‐l‐bromo‐3‐hydroxypropane, 0.2% pentaerythritol and 7.7% dimers and structural isomers)	Mice (B6C3F1) 60 M and 60 F per dose group 6 weeks old	Oral exposure (feed) 0, 312, 625 or 1250 mg/kg feed (Average daily doses of FR‐1138 were 0, 35, 70 or 149 mg/kg bw for M and 0, 40, 80 or 170 mg/kg bw for F) Repeated exposure 2 years	Survival in both M and F of the highest dose group was significantly lower than that of controls. In both males and females, significantly increased incidences of neoplasms in various organs were observed. In M: Harderian gland, lung, forestomach and kidney. In F: Harderian gland, lung, forestomach, mammary gland, circulatory system and subcutaneous tissue. See Table 15.	LOAEL = 35 mg FR‐1138/kg bw per day	NTP ( 1996 )
DBNPG Purity 87% (impurities: 2,2‐bis(hydroxymethyl)‐1‐bromo‐3‐hydroxypropane (6.7%), 1‐bromo‐3‐hydroxypropane (5.5%) and 3,3‐bis(bromomethyl)‐oxetane (0.5%), 0.3% water)	Mice (CD‐1)	Oral exposure (feed) 2G reproduction toxicity study 0, 141, 274 and 589 mg/kg bw per day of DBNPG	Reduction postpartum BW and weight gain in M and F was at ≥ 274 mg/kg bw per day Dose‐dependent decrease in live pup weight. Significant reduction of the number of litters per pair and the number of live pups born per litter at 589 mg/kg bw per day. In the F1 generation, significant reduction of weight gain at the two highest dose levels in M and at the highest dose level in F. The number of live pups born per litter and adjusted live pup weight was decreased at 589 mg/kg bw per day. Cross‐breeding of F0 exposed M and F of the high‐dose group with F0 controls after 11 weeks of dosing resulted in a reduction of the number of live pups per litter and of pup weight only when exposed F were mated with control M. Histological examination revealed kidney lesions in treated M and F. Reduction of the number of follicles in the ovaries of cross‐bred F0 F in the high‐dose group. The same effect was found in F1 F of the two high‐dose groups. No effect on reproductive organ weights and oestrual cyclicity.	LOAEL = 141 mg/kg bw per day	Treinen et al. ( 1989 )
DBDPE (Section 3.3.1 )
DBDPE Purity: 96.3%	Rats (Sprague–Dawley)	0, 100, 300 and 1000 mg/kg bw per day 90‐day gavage	Effects on liver weight in F at 1000 mg/kg bw per day.	NOAEL = 300 mg/kg bw per day	Hardy et al. ( 2002 )
DBDPE Purity: NR	Rats (Sprague–Dawley) M	0, 100 mg/kg bw per day 90‐day gavage	Increase in T3 level.	–	Wang et al. (2010)
DBDPE Purity: 94%	Rats (Harlan Sprague–Dawley) M	0, 0.1, 0.97, 9.71, 97.1, 970 mg/kg bw per day 5‐day gavage	No effects on serum TT4, TT3 or TSH. No effects on absolute liver weight. Clinical chemistry endpoints and haematology endpoints unchanged compared to controls. No liver or thyroid histological changes.	NOAEL = 970 mg/kg bw per day	Shockley et al. ( 2020a )
DBDPE Purity: NR	Rats (Sprague–Dawley) M, F OECD TG407	0, 125, 400, 1250 mg/kg bw per day 28‐day gavage + 14‐day recovery period	–	NOAEL = 1250 mg/kg bw per day	REACH dossiera
DBDPE Purity: NR	Rats (Sprague–Dawley) M, F OECD TG408	0, 100, 320, 1000 mg/kg bw per day 90‐day gavage + 28‐day recovery period (C and H doses)	–	NOAEL = 1000 mg/kg bw per day	REACH dossiera
DBDPE Purity: > 98%	Rats (Sprague–Dawley) M	0, 5, 50, 500 mg/kg bw per day 28‐day gavage	Heart and abdominal aorta: morphological changes, ultrastructural damage, Oxydative stress, Endothelial dysfunction. Inflammation. Cardiomyocyte injury (mitochondria damage and dissolution) and fibrosis, cellular apoptosis. Thyroid: decrease FT3 and increase TSH serum levels, changes in histological structure and in ultrastructure of the thyroid gland, oxidative stress, perturbations of HPT axis. Liver: morphological changes, oxidative stress, increase γ‐GT and glucose level. Testis: histopathological changes, decrease sperm number and motility, increase sperm malformation rates, oxidative stress and apoptosis.	LOAEL = 5 mg/kg bw per day	Jing et al. ( 2019 ), Wang, Romanak, et al. ( 2019 ), Sun et al. ( 2020 ), Zheng et al. ( 2021 ), Li et al. ( 2021 ), Gao et al. ( 2022 ), Xue et al. ( 2022 )
DBDPE Purity: NR	Mice (C57BL/6J) M	0, 5, 50 mg/kg bw per day 7‐week gavage	Testis: decrease weight, decrease sex hormones levels, testicular damage, increase sperm malformations.	LOAEL = 5 mg/kg bw per day	Qi et al. ( 2025 )
DBDPE Purity: 98%	Mice (ApoE−/−) M	0, 100 mg/kg bw per day 6‐week gavage	Aortic roots: lesions, atherosclerosis, increase lipid content. Oxidative stress	LOAEL = 100 mg/kg bw per day	Li et al. ( 2025 )
DBDPE Purity: 98.5%	Mice (Balb/C) M, F	0, 5, 20, 100, 200 mg/kg bw per day 30‐day gavage	Increase serum ALT and AST levels, increase glucose levels. Increase TSH and T3 and fT3 at 200 mg/kg bw per day. Hepatocyte hypertrophy.	LOAEL = 5 mg/kg bw per day	Sun et al. ( 2018 )
DBDPE Purity: NR	Rats (Sprague–Dawley) Pregnant F	0, 100, 320, 1000 mg/kg bw per day GD6‐LD21 Development gavage OECD TG414	– See Table C.2 for developmental neurotoxicity outcomes.	NOAEL maternal and developmental toxicity = 1000 mg/kg bw per day	REACH dossiera
DBDPE Purity: NR	Rats (Sprague–Dawley) Pregnant F	0, 125, 400, 1250 mg/kg bw per day Development gavage GD6‐LD21 OECD TG414	–	NOAEL maternal and developmental toxicity = 1250 mg/kg bw per day	REACH dossiera
DBDPE Purity: NR	Rabbits (NZW) Pregnant F	0, 125, 400, 1250 mg/kg bw per day Development gavage GD6‐GD18 OECD TG414	–	NOAEL maternal and developmental toxicity = 1250 mg/kg bw per day	REACH dossiera
DBDPE Purity: NR	Rats (Sprague–Dawley) Extended one‐generation reproductive toxicity with developmental neurotoxicity cohort M, F	0, 100, 300, 1000 mg/kg bw per day Cohort 1A: Reproductive/developmental toxicity testing – treated from weaning to approximately 13 weeks of age. Cohort 1B: Spare Cohort – treated from weaning to approximately 14 weeks of age. Cohort 1C: Additional cohort for assessment of sexual maturation – treated from weaning to completion of sexual maturation (approximately seven to 8 weeks of age). OECD TG443	– See Table C.2 for developmental neurotoxicity outcomes.	NOAEL for both systemic toxicity and reproductive/developmental toxicity = 1000 mg/kg bw per day	REACH dossier (a)
DBDPE Purity: 98%	Mice (ICR) Pregnant F	0, 100 μg/kg bw per day Gavage GD6‐PND21 After weaning, male offspring were fed on either a low‐fat diet or a high‐fat diet for 12 weeks	Increased risk of obesity in offspring and effects on triglyceride synthesis, bile secretion, purine synthesis, mitochondrial function and glucose metabolism	LOAEL = 100 μg/kg bw per day	Yan et al. ( 2018 ), REACH dossiera
DBDPE Purity: NR	Mice (ICR) Pregnant F	0, 0.05, 0.5, 5 mg/kg bw per day Gavage GD1‐GD17 Fetal size measured on GD18	Increase fetal growth restriction, decrease crown‐rump length, decrease fetal weight, decrease in the fetal‐to‐placental weight ratio, reduction in the total placental area, decrease in both the placental spongiotrophoblast and labyrinth areas, placental damage, decrease in placental microvascular density	LOAEL = 0.05 mg/kg bw per day	Xie et al. ( 2025 )
DBDPE Purity: NR	Mice (Kunming) Pregnant F	0, 50 μg/kg bw per day 30‐day exposure A set of exposed F were given hormones (PMSG and HCG) before being mated with untreated M. Pre‐ and post‐implantation embryos isolated.	Loss of embryos (statistically significant decrease in the rate of 2‐cell and 8‐cell embryos and blastocysts), decrease litter size. From GD14.5: decrease embryo length indicating delayed fetal development. See Table C.2 for developmental neurotoxicity outcomes.	LOAEL = 50 μg/kg bw per day	Shi, Xu, et al. ( 2022 )
DBDPE Purity: NR	Mice (Kunming) 14‐day exposure Pregnant F	0, 5, 50, 500 μg/kg bw per day After treatment, F mice were i.p. injected with PMSG and 48 h later with HCG and then mated with sexually mature M.	DBDPE inhibited preimplantation embryo development in a dose‐dependent manner by affecting the circadian rhythm (measured using a voluntary wheel running test) (from 50 μg/kg bw per day) and endogenous melatonin levels (decrease serum melatonin from 5 μg/kg bw per day in 2‐cell embryo and 8‐cell embryo). The highest dose significantly reduced the rate of 2‐cell embryos. The ratio of the 4‐cell embryo, 8‐cell embryo and morula dramatically decreased at 50 and 500 μg/kg bw per day and even the lowest dose resulted in a significant decrease in the blastocyst rate. See Table C.2 for developmental neurotoxicity outcomes.	LOAEL = 5 μg/kg bw per day	Shi, Qiu, et al. ( 2022 )

Abbreviations: F, females; GD, gestational day; HCG, human chorionic gonadotropin; LOAEL, lowest observed adverse effect level; M, males; NOAEL, no observed adverse effect level; NR, not reported; PMSG, Pregnant mare's serum gonadotropin; PND, postnatal day.

^a REACH dossier: 1,1'‐(ethane‐1,2‐diyl)bis[pentabromobenzene] 100.076.669|578c0ca4‐2d04‐4f31‐99bb‐149c3ae9578a ‐ ECHA CHEM.

Table C.2.

Developmental neurotoxicity studies on DBDPE in experimental animals

Species compound purity	Species route of administration exposure doses study duration	Test used to assess behavioural endpoints	Outcome	Reference
DBDPE Purity: NR	Mice (Kunming) F 0, 50 μg/kg per day Gavage For 30 days Exposed F mated with untreated M	Pups tested on week 6 of age for behavioural performances. Open‐field test: (5–10 mice/group from 3 litters, free exploration for 5 min). Variables related to locomotor activity: total distance, average speed, running time, static time, walking time; variables related to anxiety‐related behaviour: corner zone duration, central zone duration. Dark/light transition test: (7 mice/group from 3 litters, 5 min). Divided into two boxes, a dark one connected with an illuminating one (50 ± 0.5 Lux). Number of transitions between the two boxes, time spent in each box (the longer the mouse stays in the light box, the lower the anxiety level of the animal is, contrary regarding the dark side). Novel object recognition test: to evaluate the preference for novelty and working memory (7 mice/group from 3 litters). Three sessions: habituation (5 min, any object inside the cage), training (10 min, two similar objects, exploring time of each object) and testing (10 min, one object similar to those used during the training session, the other is a new one, exploring time of each object). Three‐chamber test: (7 mice/group from 3 litters). 3 chambers with a cylinder cage in two of them, the right and the left ones whereas the central chamber remains empty. 3 phases: (i) habituation (5 min, free exploration of the complete maze), (ii) social preference test (5 min, a mouse is placed in one cylinder cage in one chamber, the other remains empty), exploring time of the mouse vs. the empty cage, (iii) social novelty test (5 min, one cage with a mouse that is the same as used for the social preference test, one new unknown mouse in the other cage), exploring time of the unknow mouse vs. the familiar one.	Open‐field test: significant increases in the total distance travelled and the velocity only in DBDPE exposed females vs. controls. No effects on anxiety. Dark–light transition test: No significant variations of variables related to anxiety (time spent in each box) in both sexes. As observed in the OF, activity level of females significantly increased vs. controls (total distance travelled, velocity, walking time) Novel object recognition test: Significant increase in the time spent to explore the novel object in M and F control mice compared to the time spent to explore the ‘old’ one. No significant variations in M and F‐exposed mice. Significant increase in the activity level in F exposed to DBDPE vs. controls, not in M. Three‐chamber test: Social preference test: significant increases in the time spent to explore the mouse vs. the empty cage in control and DBDPE‐exposed M and control F. No difference in DBDPE‐exposed F. Social novelty test: same significant variations regarding the exploration of the unknow mouse vs. the familiar one in control and DBDPE‐exposed M and control F. No differences in DBDPE‐exposed F. Overall, DBDPE induced a significant increase in the activity level in F and not in M, and impaired the learning and memory performances of both DBDPE‐exposed M and F. F exposed to DBDPE were impaired for their social recognition (social memory) performances.	Shi, Xu, et al. (2022)
DBDPE Purity: NR	Kunming mice (F) 0, 5, 50, 500 μg/kg per day Gavage For 14 days Exposed F mated with untreated M	Voluntary wheel running test: To assess the locomotor activity according to the circadian rhythm Mice kept individually in cages within light‐tight chambers after 14 days of DBDPE exposure. First, the mice were exposed to a 12:12‐h light/dark (LD) cycle (lights on at 07:00, off at 19:00) for 4 days to adapt to the environment. Then, the mice were exposed to a 24‐h dark (DD) cycle for 4 days to demonstrate their own endogenous circadian rhythm. Clock gene expression was studied in the suprachiasmatic nucleus of the same animals before and during the test.	In the LD cycle, mice in the four groups were active during the dark phase of the cycle and almost inactive during the light phase, showing an obvious circadian rhythm. After entering the DD cycle, the control group and 5 μg/kg bw/day DBDPE‐treated mice showed normal circadian rhythm behaviour. However, mice from the 50 and 500 μg/kg bw/day DBDPE groups showed active movement during the 12h‐period that corresponds to the light phase of the cycle, which indicated an endogenous circadian rhythm disorder. The daily activity of the mice was also analysed, and the results showed that the total wheel running activity and the wheel running activity during the dark phase of the cycle in the 50 and 500 μg/kg bw per day DBDPE‐exposed groups were significantly increased in LD and DD, and the total rest bouts were significantly reduced, demonstrating the disorder of activity rhythm.	Shi, Qiu, et al. (2022)
DBDPE Purity: NR	Rats (Sprague–Dawley) 0, 100, 300 and 1000 mg/kg bw per day Gavage (P) M: 10 weeks before mating up to necropsy after litters were weaned. (P) F: 10 weeks before mating, up to 2 weeks during mating, during resulting pregnancies, up to necropsy on Day 23 of lactation. (F1) M and F: From weaning until termination of respective cohort. Cohort 2A: Developmental neurotoxicity testing (neurobehavioural testing followed by neuro‐histopathology assessment as adults) – treated from weaning up to approximately Day 75 of age. Cohort 2B: Developmental neurotoxicity testing – no direct treatment, assigned to neurohistopathology assessment at weaning.	Neurobehavioural Screening – Cohort 2A Functional observational battery on all Cohort 2A animals on day 70 ± 1 of age In cage observations: Abnormal motor movements (e.g. fasciculation, convulsions, stereotypy, excessive sniffing, licking, grooming and twitches), Palpebral closure, Posture, Tremor In hand observations: Ease of removal from cage, Exophthalmos, Fur condition, Piloerection, Reactivity to handling, Salivation/lacrimation, Vocalisation on handling, Observation in the arena – 2‐min recording period:, Activity count, Arousal, Abnormal motor movements (e.g. fasciculation, convulsions, stereotypy, excessive sniffing, licking, grooming and twitches), Faecal count, Gait, Palpebral closure, Posture, Rearing count, Urination, Tremor Reactivity investigations: Approach response, Body temperature, Body weight, Grip strength: fore and hindlimb, Landing foot splay, Pupil closure reflex, Righting reflex, Auditory startle reflex, Tail pinch response, Touch response (pinna reflex), Proprioception F1 Cohort 2A: Auditory Startle Response Habituation: Habituation and latency to peak values. Motor Activity: Motor activity including assessment for both ambulatory activity (low beam) and rearing activity (high beam) In Cage Observations In Hand Observations Arena Observations Brain morphometry	Auditory startle response habituation, reactivity investigations and motor activity (ambulatory and rearing) were unaffected by treatment. There were no in cage or in hand observations that could be related to treatment and arena observations were limited to a slight increase in the incidence of females with elevated/slightly abnormal gait in the treated groups but with no evidence for a dose response. All differences in brain morphometry parameters were considered incidental as they were inconsistent between sexes, inconsistent between cohorts, the magnitude was considered small, had no histopathological correlates, and/or had no significant mean brain weight differences.	REACH dossier ‐ Extended 1 generation reproductive study (2 added cohorts) REACH dossier: 1,1'‐(ethane‐1,2‐diyl)bis[pentabromobenzene] 100.076.669 | 578c0ca4‐2d04‐4f31‐99bb‐149c3ae9578a ‐ ECHA CHEM
DBDPE Purity: NR	Rats (Sprague–Dawley) Gavage 0, 100, 320, 1000 mg/kg bw per day F GD6‐PND21 F1 Pups: exposure via the dam during gestation and lactation only. Test till PND30	Detailed clinical observations: for weaning and postweaning rat offspring consisted of handling observations and open‐field observations (details not available). F1 In addition, hindlimb and forelimb grip strength were evaluated. F1 Offspring Behavioural Observations: − Motor Activity on PND13, 17, 21 and 61. Testing was performed following the detailed clinical observations on PND21 and prior to weaning. − Auditory Startle on PND24 and 63. − Prepulse inhibition (PPI) and habituation of the startle response evaluated at both time points. − Passive Avoidance on PND23 (Subset B) and 64 (Subset A) − Using a passive avoidance, trials‐to‐criterion approach, learning and memory were assessed on PND23 and PND30, respectively (Subset B), and on PND64 and 71, respectively (Subset A),	No effects of DBDPE on detailed clinical observations, motor activity, auditory startle, learning and memory were reported in both sexes at any dose level at any age. Test substance‐related morphometric changes in brains of M rats on PND22 and PND72 at 100, 320 and 1000 mg/kg bw per day. Lower group mean morphometric brain measurements in the cortex (Level 1), hippocampus (Level 3), and cerebellum (Level 5) on PND22 and PND72. Morphometric changes not associated with statistically significant changes in brain weight or macroscopic measurement of the brain, although there was a slight decrement of group mean brain weight in the highest dose group of M at PND22. No microscopic changes in brain, spinal cord, nerve roots or ganglia. No changes in F rats for the macroscopic examinations and measurements, or neuropathology at any dose level at any age. NOAEL for systemic toxicity = 1000 mg/kg bw per day (highest dose tested). DBDPE‐related findings in the offspring were limited to morphometric changes in brains of M rats on PND22 and PND72 at 100, 320 and 1000 mg/kg bw per day. The NOAEL for developmental neurotoxicity cannot be determined with sufficient certainty for M.	REACH dossier ‐ Neurotoxicity study in rats REACH dossier: 1,1'‐(ethane‐1,2‐diyl)bis[pentabromobenzene] 100.076.669 | 578c0ca4‐2d04‐4f31‐99bb‐149c3ae9578a ‐ ECHA CHEM

APPENDIX D. Occurrence of BEH‐TEBP and EH‐TBB in dust samples from European countries

Table D.1.

Occurrence of BEH‐TEBP in dust samples from European countries, and estimate of exposure via dust when reported by the authors

Country sampling year	Number of samples	Concentration	Exposure estimate	Reference
Belgium 2008	Homes: n = 39 Offices: n = 6	Mean, median, range (ng/g dust): Homes: 212 ± 867, 13, < 2–5004 Offices: 95 ± 89, 64, 16–265	Mean, 95th percentile (ng/kg bw/day): Worst‐case scenario Adults: 0.02, 0.32 Toddlers: 0.40, 8.61	Ali et al. ( 2011 ). Novel brominated flame retardants in Belgian and UK indoor dust: Implications for human exposure.
UK 2007‐2008	Classrooms: n = 36	Classrooms: 381 ± 1151, 96, < 2–6175
Sweden 2006	Apartments: n = 6	Range (ng/g): 260–950	NR	Sahlström et al. ( 2012 ). Clean‐up method for determination of established and emerging brominated flame retardants in dust.
Sweden 2016‐2017	Offices: n = 10	Median (ng/g): 265	Range (ng/kg bw/day): Adult – Dust ingestion 0.0017–1.3 (sum of some emerging and novel BFRs)	Tao et al. ( 2019 ). Organohalogenated Flame Retardants and Organophosphate Esters in Office Air and Dust from Sweden
Sweden 2016	Home, n = 19 Preschool, n = 11	Mean (ng/g): Home: 150 Preschool: 83	Range (ng/kg bw/day): Children – Dust ingestion 0.037–1.1 (sum of some emerging and novel BFRs)	Tao et al. ( 2023 ). Organohalogenated flame retardants and organophosphate esters from home and preschool dust in Sweden: Pollution characteristics, indoor sources and intake assessment
Germany NR	n = 20	Mean, median, range (ng/g): 436, 343, 25–2274	Average intake, high intake (pg/kg bw): Adults: 147, 348 Toddlers: 1715, 4055	Fromme et al. (2014). Polybrominated diphenyl ethers (PBDEs), hexabromocyclododecane (HBCD) and "novel" brominated flame retardants in house dust in Germany.
Sweden 2012	n = 27	Geometric Mean, median (range) (ng/g): 160, 140 (< 33–1500)	NR	Newton et al. ( 2015 ). Emerging Flame Retardants, PBDEs, and HBCDDs in Indoor and Outdoor Media in Stockholm, Sweden.
Sweden 2009–2010	n = 27 For intake estimation: n = 20	Geometric Mean, median (range) (ng/g): 62, 61 (10–340)	Median, range (ng per day): Mothers: 1.6 (< 0.22–7.6) Toddlers: 3.3 (< 0.44–15)	Sahlström et al. ( 2015 ). Estimated intakes of brominated flame retardants via diet and dust compared to internal concentrations in a Swedish mother–toddler cohort.
Spain	Homes: n = 5 Schools: n = 4 Theatres: n = 3 University: n = 1 Research institute: n = 1	Homes: 17.3–231 ng/g Schools:54–290 ng/g Theatres: 159–508 ng/g University: 89.4 ng/g Research institute: 289 ng/g	NR	Cristale et al. ( 2016 ). Occurrence and sources of brominated and organophosphorus flame retardants in dust from different indoor environments in Barcelona, Spain.
Norway 2013–2014	n = 60	Geometric Mean, median, range (ng/g): 490, 300, 26–3100	Median, 5th–95th percentile (pg/kg bw per day): Adults: 34, 5.8‐130	Tay et al. ( 2017 ). Human Exposure to Legacy and Emerging Halogenated Flame Retardants via Inhalation and Dust Ingestion in a Norwegian Cohort.
Greece NR	n = 30 cars (1–19 years old)	Mean, median, range (ng/g): 233, 88.7, 7.75‐1553	Median, 5th–95th percentile (ng/kg bw per day): Adults: 2.53E‐03, 3.71E‐04 – 2.15E‐02 Toddlers: 1.77E‐02, 2.60E‐03 – 1.50E‐01	Besis et al. ( 2017 ). Legacy and novel brominated flame retardants in interior car dust ‐ Implications for human exposure.
Greece 2018	n = 20 Classroom	Mean, median, range (ng/g): 7.3, 2.9, 0.70–46	NR	Besis et al. ( 2023 ). Hazardous organic pollutants in indoor dust from elementary schools and kindergartens in Greece: Implications for children's health
UK 2015	Kitchen: n = 30 Living room/bedroom: n = 30	Median, range (ng/g): Kitchen: 36, 2.7–420 Living Room: 75, 7.8–630	NR	Kuang et al. ( 2016 ). Concentrations of “legacy” and novel brominated flame retardants in matched samples of UK kitchen and living room/bedroom dust.
UK 2013–2014	n = 5 Particle size 1 (P1): 125–250 μm Particle size 2 (P2): 63–125 μm Particle size 3 (P3): 25–63 μm Non‐fractionated bulk dust samples (BD)	Mean, median, range (ng/g): P1: 671, 220, 73–2659 P2: 972, 333, 79–4359 P3: 1265, 355, 94–5519 BD: 746, 320, 80‐3187	NR	Al‐Omran and Harrad ( 2016 ). Distribution pattern of legacy and "novel" brominated flame retardants in different particle size fractions of indoor dust in Birmingham, United Kingdom.
UK 2014–2015	n = 12	Mean, median, range (ng/g): Living room: 306 ± 348, 175, 64–1299 Bedroom: 339 ± 380, 131, 43–1139 Household vacuum: 233 ± 256, 121, 33–890	NR	Al‐Omran et al. ( 2017 ). Influence of sampling approach on concentrations of legacy and "novel" brominated flame retardants in indoor dust.
UK 2013–2014	n = 238 samples from 3 homes	Average range (ng/g): Home 1: 66–323 Home 2: 105–445 Home 3: 926–7049	NR	Al‐Omran et al. ( 2018 ). Within‐room and within‐home spatial and temporal variability in concentrations of legacy and "novel" brominated flame retardants in indoor dust.
UK 2021	n = 30 Living rooms, n = 6 Kitchens, n = 6 Bedrooms, n = 18	Mean, median (ng/g): 1000, 190	Adult Mean, median, high (ng/kg bw/day) Dust ingestion: 0.19, 0.036, 3.1	Ma, Y., Stubbings, W. A., et al.. (2024). Impact of Legislation on Brominated Flame Retardant Concentrations in UK Indoor and Outdoor Environments: Evidence for Declining Indoor Emissions of Some Legacy BFRs
UK 2013 Norway 2013‐2014	Houses, n = 10 Offices, n = 12 Houses, n = 10	Mean, median, range (ng/g): Houses: 116, 106, 18‐234 Offices: 678, 248, 25.3‐2541 Houses: 89.6, 27.1, 7.9‐426	Adult median, max (ng/kg bw/day) Dust ingestion worst‐case scenario UK offices: 0.11, 0.72 Dust ingestion worst‐case scenario UK houses: 0.076, 0.167 Dust ingestion worst‐case scenario Norway houses: 0.019, 0.304	Kademoglou et al. ( 2017 ). Legacy and alternative flame retardants in Norwegian and UK indoor environment: Implications of human exposure via dust ingestion
Finland NR	n = 40	Mean, median, max (ng/g): 243, 106, 1927	3‐year‐old children (pg/kg bw/day): Dust ingestion: 460 Dermal: 6.5	Rantakokko et al. ( 2019 ). Concentrations of brominated and phosphorous flame retardants in Finnish house dust and insights into children's exposure.
Norway 2012	Living room, n = 48 Classroom, n = 6	Mean, median, max (ng/g): Living room: 132, 78.5, 809 Classroom: 99.9, 103, 151	Dermal contact, Ingestion (pg/kg bw/day): Children‐Household/school: 37, 2.6 × 102 Children‐Household: 22, 152 Adult women‐Household: 17, 68	Cequier et al. ( 2014 ). Occurrence of a Broad Range of Legacy and Emerging Flame Retardants in Indoor Environments in Norway.
Germany 2015	n = 83	Geometric mean, median, 95th percentile, range (ng/g dust): 80, 78, 340, < 0.18−2600	4‐year‐old children Geometric Mean, 95th percentile (ng/kg bw/day): Oral: 0.14, 0.58 Dermal: 0.005, 0.02	Larsson et al. ( 2018 ). Brominated Flame Retardants and Organophosphate Esters in Preschool Dust and Children's Hand Wipes.
Germany NR	n = 5	Mean, range (ng/g): 377.5, 24.8–535.2	NR	Kopp et al. ( 2012 ). Analysis of common and emerging brominated flame retardants in house dust using ultrasonic assisted solvent extraction and on‐line sample preparation via column switching with liquid chromatography–mass spectrometry.
UK 2013–2015	Offices, n = 42 Houses, n = 30	Mean, median, range (ng/g): Houses: 240, 65, 16–3500 Offices: 1000, 160, 54–25,000	Toddlers: 4.7 ng/day Adults: 2.6 ng/day	Tao et al. ( 2016 ). Emerging and legacy flame retardants in UK indoor air and dust: evidence for replacement of PBDEs by emerging flame retardants?
Czech Republic 2013	n = 28	Mean, median, range (ng/g): 60±13, 42, ND–373	NR	Venier et al. ( 2016 ). Brominated flame retardants in the indoor environment — Comparative study of indoor contamination from three countries.
Spain 2019–2020	n = 9 (2 offices, 3 schools, 4 homes)	Mean, median, range (ng/g): 753±554, 861, 137–1540	Mean, range (ng/kg bw/day): Adults: 0.24 ± 0.32, 0.00–11.8 Children: 1.01 ± 1.18, 0.02–47.4	Esplugas et al. ( 2022 ). Emerging and legacy flame retardants in indoor air and dust samples of Tarragona Province (Catalonia, Spain).
Italy 2019	n = 5 (2 houses, 3 work places)	Mean (ng/g): Houses: 323.7 Work places: 64.5	Range (ng/kg bw/day): Adults – Ingestion + dermal: 1.1–22	Simonetti et al. ( 2020 ). Occurrence of Halogenated Pollutants in Domestic and Occupational Indoor Dust

Abbreviation: NR, not reported.

Table D.2.

Occurrence of EH‐TBB in dust samples from European countries, and estimate of exposure via dust when reported by the authors

Country sampling year	Number of samples	Concentration	Exposure estimate	Reference
Belgium 2008	Homes: n = 39 Offices: n = 6	Mean, median, range (ng/g dust): Homes: 20 ± 73, 1, < 2–436 Offices: 12 ± 13, 7, < 2–31	Mean, 95th percentile (ng/kg bw/day): Worst‐case scenario Adults: 0.0, 0.05 Toddlers: 0.08, 1.14	Ali et al. ( 2011 ). Novel" brominated flame retardants in Belgian and UK indoor dust: Implications for human exposure.
UK 2007–2008	Classrooms: n = 36	45 ± 55, 25, < 2–289
Sweden 2006	Apartments: n = 6	Range (ng/g): 25–440	NR	Sahlström et al. ( 2012 ). Clean‐up method for determination of established and emerging brominated flame retardants in dust.
Sweden 2009–2010	n = 27	Geometric Mean, median, range (ng/g): 6.9, 2.6, < 0.29–280	Median, range (ng/day): Mothers: 0.075, < 0.0062–8.2 Toddlers: 0.15, < 0.012–16	Sahlström et al. ( 2015 ). Estimated intakes of brominated flame retardants via diet and dust compared to internal concentrations in a Swedish mother–toddler cohort.
Sweden 2016–2017	Offices: n = 10	Median (ng/g): 73	Range (ng/kg bw/day): Adult – Dust ingestion 0.0017–1.3 (sum of some emerging and novel BFRs)	Tao et al. ( 2019 ). Organohalogenated Flame Retardants and Organophosphate Esters in Office Air and Dust from Sweden
Sweden 2016	Home, n = 19 Preschool, n = 11	Mean (ng/g): Home: 7.5 Preschool: 6.3	Range (ng/kg bw per day): Children ‐ Dust ingestion 0.037–1.1 (sum of some emerging and novel BFRs)	Tao et al. ( 2023 ). Organohalogenated flame retardants and organophosphate esters from home and preschool dust in Sweden: Pollution characteristics, indoor sources and intake assessment
Germany	n = 20	Mean, median, range (ng/g): 4.2, < 3.0, < 3.0–13.6	Average intake, high intake (pg/kg bw): Adults: 1, 5 Toddlers: 8, 63	Fromme et al. (2014). Polybrominated diphenyl ethers (PBDEs), hexabromocyclododecane (HBCD) and "novel" brominated flame retardants in house dust in Germany.
Germany 2015	n = 100	95th percentile, range (ng/g dust): 540, < 0.98−690	95th percentile (ng/kg bw per day): 4‐year‐old children Oral: 0.92 Dermal: 0.03	Larsson et al. ( 2018 ). Brominated Flame Retardants and Organophosphate Esters in Preschool Dust and Children's Hand Wipes.
Sweden 2012	n = 27	Geometric Mean, median, range (ng/g): 11, 9.1, < 2.5–65	NR	Newton et al. ( 2015 ). Emerging Flame Retardants, PBDEs and HBCDDs in Indoor and Outdoor Media in Stockholm, Sweden.
Spain NR	Homes: n = 5 (detected in 2) Schools: n = 4 (no detected) Theatres: n = 3 (detected in 1) University: n = 1 (no detected) Research institute: n = 1	Homes: 8.1–62.1 ng/g Theatre: 14.3 ng/g Research Institute: 12.9 ng/g	NR	Cristale et al. ( 2016 ). Occurrence and sources of brominated and organophosphorus flame retardants in dust from different indoor environments in Barcelona, Spain.
Norway 2013–2014	n = 60	Geometric Mean, median, range (ng/g): 42, 25, < 0.31–280	Median, 5th–95th percentile (pg/kg bw/day): Adults: 4.6, 0.094–26	Tay et al. ( 2017 ). Human Exposure to Legacy and Emerging Halogenated Flame Retardants via Inhalation and Dust Ingestion in a Norwegian Cohort.
Greece NR	n = 30 cars (1–19 years old)	Mean, median, range (ng/g): 38.3, 15.2, 0.511–244	Median, 5th–95th percentile (ng/kg/day): Adults: 4.34E‐04, 2.67E‐05–5.10E‐03 Toddlers: 3.04E‐03, 1.87E‐04–3.57E‐02	Besis et al. ( 2017 ). Legacy and novel brominated flame retardants in interior car dust ‐ Implications for human exposure.
Greece 2018	n = 20 classroom	Mean, median, range (ng/g): 0.91, 0.59, 0.10‐2.6	NR	Besis et al. ( 2023 ). Hazardous organic pollutants in indoor dust from elementary schools and kindergartens in Greece: Implications for children's health
UK 2015	Kitchen: n = 30 Living room or bedroom: n = 30	Median, range (ng/g): Kitchen: 4.1, < 0.2–290 Living Room: 12, < 0.2–450 ng/g	NR	Kuang et al. ( 2016 ). Concentrations of "legacy" and novel brominated flame retardants in matched samples of UK kitchen and living room/bedroom dust.
UK 2013–2014	n = 5 Particle size 1 (P1): 125–250 μm Particle size 2 (P2): 63‐125μm Particle size 3 (P3): 25–63 μm Non‐fractionated bulk dust samples (BD)	Mean, median, range (ng/g): P1: 9.6, 6.4, 1.8‐ 25.2 P2: 12.2, 8, 2.5‐44 P3: 11.9, 7.8, < 0.5‐36.1 BD: 9.2, 5.8, 4‐22.9	NR	Al‐Omran and Harrad ( 2016 ). Distribution pattern of legacy and "novel" brominated flame retardants in different particle size fractions of indoor dust in Birmingham, United Kingdom.
UK 2014–2015	n = 12	Mean, median, range (ng/g): Living room: 6.9 ± 6, 6.4, 0.5–21.2 Bedroom: 6.4 ± 7.9, 3.1, 0.5– 24.2 Household vacuum: 4.9 ± 4.9, 3.9, 0.5–13.5	NR	Al‐Omran et al. ( 2017 ). Influence of sampling approach on concentrations of legacy and "novel" brominated flame retardants in indoor dust.
UK 2021	N = 30 Living rooms, n = 6 Kitchens, n = 6 Bedrooms, n = 18	Mean, median (ng/g): 20, 4.5	Adult Mean, median, high (pg/kg bw/day) Dust ingestion: 3.8, 0.85, 57	Ma, Stubbings, et al. ( 2024 ). Impact of Legislation on Brominated Flame Retardant Concentrations in UK Indoor and Outdoor Environments: Evidence for Declining Indoor Emissions of Some Legacy BFRs
UK 2013 Norway 2013–2014	Houses, n = 10 Offices, n = 12 Houses, n = 10	Mean, median, range (ng/g): Houses: 10.7, 5.0, 2.5–32.0 Offices: 29.4, 18.1, 1.6–143 Houses: 5.1, 5.4, 1.3–9.2	Adult Median, max (ng/kg bw per day) Dust ingestion worst‐case scenario UK offices: 0.007–0.049 Dust ingestion worst‐case scenario UK houses: 0.004–0.023 Dust ingestion worst‐case scenario Norway houses: 0.004–0.007	Kademoglou et al. ( 2017 ). Legacy and alternative flame retardants in Norwegian and UK indoor environment: Implications of human exposure via dust ingestion
Finland		Mean, median, maximum (ng/g): 8.91, 5.46, 33.4	3‐year‐old children Ingestion, dermal (pg/kg per day): 23, 0.74	Rantakokko et al. ( 2019 ). Concentrations of brominated and phosphorous flame retardants in Finnish house dust and insights into children's exposure.
Germany NR	n = 5	Mean, range (ng/g): < LOQ (3 ng/g) , < LOQ‐8	NR	Kopp et al. ( 2012 ). Analysis of common and emerging brominated flame retardants in house dust using ultrasonic assisted solvent extraction and on‐line sample preparation via column switching with liquid chromatography–mass spectrometry.
Norway 2012	Living room, n = 48 Classroom, n = 6	Mean, median, maximum (ng/g): Living room: 16.1, 2.54, 245 Classroom: 2.67, 3.32, 5.72	Dermal contact, ingestion (pg/kg bw/day): Children‐Household/school: 1.2, 8.2 Children‐Household: 0.70, 4.9 Adult women‐Household: 0.54, 2.2	Cequier et al. ( 2014 ). Occurrence of a Broad Range of Legacy and Emerging Flame Retardants in Indoor Environments in Norway.
UK 2013–2015	Offices, n = 42 Houses, n = 30	Mean, median, range (ng/g): Houses: 21, 10, < 0.010–85 Offices: 120, 31, < 0.010–2000	Toddlers: 0.35 ng/day Adults: 0.19 ng/day	Tao et al. ( 2016 ). Emerging and legacy flame retardants in UK indoor air and dust: evidence for replacement of PBDEs by emerging flame retardants?
Spain 2019–2020	n = 9 (2 offices, 3 schools, 4 homes)	Mean, median, range (ng/g): 22.7 ± 16.5, 20.4, 3.30–49.4	Mean, range (ng/kg/day): Adults: 0.12 ± 0.11, 0.01–2.94 Children: 0.25 ± 0.24, 0.01–9.94	Esplugas et al. ( 2022 ). Emerging and legacy flame retardants in indoor air and dust samples of Tarragona Province (Catalonia, Spain).
Czech Republic 2013	n = 28	Mean, median, range (ng/g): 17 ± 5.4, 7.8, ND‐150	NR	Venier et al. ( 2016 ). Brominated flame retardants in the indoor environment — Comparative study of indoor contamination from three countries.
Italy 2019	n = 5 (2 houses, 3 work places)	Mean (ng/g): Houses: 11.9 Work places: 9.7	Range (ng/kg bw per day): Adults – Ingestion + dermal: 0.16–0.58	Simonetti et al. ( 2020 ). Occurrence of Halogenated Pollutants in Domestic and Occupational Indoor Dust

Abbreviation: NR, not reported.

APPENDIX E. Levels in human samples of the BFRs assessed in this Opinion from European countries and published since the previous Opinion

Table E.1.

Levels of the BFRs considered in this Opinion in human milk and serum from European countries as reported in studies published since the previous Opinion (EFSA CONTAM Panel, 2012b)

Country Year	Number of samples	Concentration	Detection frequency (LOD/LOQ)	Reference
Human milk
Ireland 2010	n = 11 pooled human milk samples from 109 first‐time mothers	DBDPE: ND BTBPE: ND HBB: ND	DBDPE: 0% (LOD = 2.2 ng/g lipid) BTBPE: 0% (LOD = 0.2–0.3 ng/g lipid) DBDPE: 0% (LOD = 0.2–0.3 ng/g lipid)	Pratt et al. (2013). Brominated and fluorinated organic pollutants in the breast milk of first‐time Irish mothers: is there a relationship to levels in food?
Sweden 2009–2010	n = 2 pooled human milk samples	Mean (pg/g ww): BEH‐TEBP: ND EH‐TBB: ND DBDPE: ND BTBPE: < LOQ α‐DBE‐DBCH: 4.0 β‐DBE‐DBCH: < LOQ TBX: ND PBT: ND PBEB: ND HBB: ND	BEH‐TEBP: 0 pooled samples (LOD = 26 pg/g ww) EH‐TBB: 0 pooled samples (LOD = 13 pg/g ww) DBDPE: 0 pooled samples (LOD = 11 pg/g ww) BTBPE: 1 pooled sample (LOD‐LOQ = 1.0–3.4 pg/g ww) α‐DBE‐DBCH: 2 pooled samples (LOD‐LOQ = 0.50–1.7 pg/g ww) β‐DBE‐DBCH: 1 pooled sample (LOD‐LOQ = 0.50–1.7 pg/g ww) TBX: 0 pooled samples (LOD = 0.5 pg/g ww) PBT: 0 pooled samples (LOD = 0.4 pg/g ww) PBEB: 0 pooled samples (LOD = 0.6 pg/g ww) HBB: 0 pooled samples (LOD = 0.9 pg/g ww)	Sahlström et al. (2015). Estimated intakes of brominated flame retardants via diet and dust compared to internal concentrations in a Swedish mother‐toddler cohort.
Norway (NO) 2003–2006 The Netherlands (NL) 2011–2014 Slovakia (SK) 2011–2012	n = 305 n = 116 n = 37	Mean, P95 (ng/g lipid): EH‐TBB: 0.086, 0.100 (NO) 0.240, 0.683 (NL) 0.037, 0.091 (SK) BEH‐TEBP: 0.613, 1.432 (NO) 1.34, 2.75 (NL) 0.622, 1.887 (SK) BTBPE: 0.586, 0.140 (NO) 1.103, 2.380 (NL) 0.023, 0.078 (SK) HBB: 0.036, 0.102 (NO) 0.048, 0.099 (NL) 0.041, 0.170 (SK) DBE‐DBCH: 0.038, 0.083 (NO) 0.019, 0.035 (NL) 0.017, 0.042 (SK) PBEB 0.005, NR (NO) ND (NL) 0.002, NR (SK) PBT: 0.049, 0.063 (NO) 0.028, 0.065 (NL) 0.039, 0.089 (SK) DBHCTD: ND (NO) ND (NL) ND (SK) TBCO: ND (NO) ND (NL) ND (SK) TBX 0.019, 0.038 (NO) 0.035, 0.128 (NL) ND (SK)	EH‐TBB: 47% (NO), 69% (NL), 59% (SK) (LOD = 0.001 ng/g lipid) BEH‐TEBP: 29% (NO), 15% (NL), 27% (SK) (LOD = 0.023 ng/g lipid) BTBPE: 24% (NO), 26% (NL), 22% (SK) (LOD = 0.001 ng/g lipid) HBB: 44% (NO), 88% (NL), 62% (SK) (LOD = 0.002 ng/g lipid) DBE‐DBCH: 25% (NO), 26% (NL), 32% (SK) (LOD = 0.001 ng/g lipid) PBEB: 2% (NO), 0% (NL), 3% (SK) (LOD = 0.001 ng/g lipid) PBT: 87% (NO), 85% (NL), 100% (SK) (LOD = 0.002 ng/g lipid) DBHCTD: 0% (NO), 0% (NL), 0% (SK) (LOD = 0.004 ng/g lipid) TBCO: 0% (NO), 0% (NL), 0% (SK) (LOD = 0.002 ng/g lipid) TBX: 15% (NO), 69% (NL), 0% (SK) (LOD = 0.001 ng/g lipid) For concentrations < LOD, the value of LOD/√2 was used.	Čechová et al. (2017). Legacy and alternative halogenated flame retardants in human milk in Europe: Implications for children's health.
UK 2010 UK 2014–2015	n = 25 n = 10	Mean, median (min–max) (ng/g lipid): BEH‐TEBP: NR, < 0.1 (< 0.1–4.6) EH‐TBB: NR, < 0.1 (< 0.01–2.1) DBDPE: NR, < 0.78 (< 0.78–250) BTBPE: NR, < 0.1 (< 0.1–56) α‐DBE‐DBCH: NR, < 0.13 (< 0.13–1.7) β‐DBE‐DBCH: 6.8, 3.1 (< 0.13–38) PBT, HBB, TBX, PBEB: ND BEH‐TEBP: 0.25, < 0.1 (< 0.1–0.73) EH‐TBB: 0.21, 0.16 (< 0.01–0.48) DBDPE: NR, < 0.78 (< 0.78–58) BTBPE: NR, < 0.1 (< 0.1–0.71) α‐DBE‐DBCH: 0.67, 0.60 (< 0.3–1.1) β‐DBE‐DBCH: 2.5, 1.2 (0.43–10) PBT, HBB, TBX, PBEB: ND	BEH‐TEBP: 36% (LOD = 0.004 pg/uL) EH‐TBB: 44% (LOD = 0.004 pg/uL) DBDPE: 4% (LOD = 0.44 pg/uL) BTBPE: 28% (LOD = 0.046 pg/uL) α‐DBE‐DBCH: 20% (LOD = 0.07 pg/uL) β‐DBE‐DBCH: 76% (LOD = 0.07 pg/uL) PBT, HBB, TBX, PBEB: 0% BEH‐TEBP: 50% (LOD = 0.004 pg/uL) EH‐TBB: 90% (LOD = 0.004 pg/uL) DBDPE: 10% (LOD = 0.44 pg/uL) BTBPE: 40% (LOD = 0.046 pg/uL) α‐DBE‐DBCH: 100% (LOD = 0.07 pg/uL) β‐DBE‐DBCH: 100% (LOD = 0.07 pg/uL) PBT, HBB, TBX, PBEB: 0% For concentrations < LOQ, the value of LOQ/2 was used.	Tao et al. (2017). Emerging and legacy flame retardants in UK human milk and food suggest slow response to restrictions on use of PBDEs and HBCDD.
Ireland (IRE)	n = 16 pooled human milk samples	Average, median (min‐max) (ng/g lipid): DBDPE: 0.74, < 2.5 (< 2.5–4.6)	DBDPE: 3/16 pooled samples (LOD = 2.5 ng/g lipid)	Wemken et al. (2020). Emerging and legacy brominated flame retardants in the breast milk of first time Irish mothers suggest positive response to restrictions on use of HBCDD and Penta‐ and Octa‐BDE formulations.
Serum
Norway NR	n = 10	DBDPE: ND BTBPE: 0.35, 0.48 HBB: 0.04, 0.07, 0.08 DBHCTD: ND Single values reported for each of the 10 samples (ng/g lipid)	DBDPE (LOD = 20 pg/mL): 0% BTBPE (LOD = 2.3 pg/mL): 20% HBB (LOD = 0.30 pg/mL): 30% DBHCTD (LOD = 5.4 pg/mL): 0%	Cequier et al. (2013). Determination of emerging halogenated flame retardants and polybrominated diphenyl ethers in serum by gas chromatography mass spectrometry.
Sweden 2009–‐‐2010	n = 24 (5 samples for α‐ and β DBE‐DBCH) n = 5 standard reference maternal serum (for the other congeners)	Range (ng/g lipid) α‐ and β‐DBE‐DBCH: Toddler's serum: LOD–1.8 Maternal serum: ND‐LOD Mean (pg/g wet weight) DBDPE: 418 BTBPE: 430 α‐DBE‐DBCH: 97 TBX: 7.1 PBEB: 19 HBB: 331 PBT: ND DBHCTD: ND OBTMPI: ND	α‐DBE‐DBCH (LOD = 0.3 ng/g lipid) β‐DBE‐DBCH (LOD = 0.4 ng/g lipid) DBDPE (LOD = 15 pg/serum sample) BTBPE (LOD = 10 pg/serum sample) α‐DBE‐DBCH (LOD = 6.4 pg/serum sample) TBX (LOD = 0.9 pg/serum sample) PBEB (LOD = 0.9 pg/serum sample) HBB (LOD = 1.0 pg/serum sample) PBT (LOD = 0.9 pg/serum sample) DBHCTD (LOD = 7.1 pg/serum sample) OBTMPI (LOD = 18 pg/serum sample)	Sahlström et al. (2014). Brominated flame retardants in matched serum samples from Swedish first‐time mothers and their toddlers.
Norway 2012	n = 46 mothers	Mean (ng/g lipid) DBDPE: ND BTBPE: 0.21 HBB: 0.07 DBHCTD: ND	DBDPE (NR): 0% BTBPE (NR): 9% HBB (NR): 48% DBHCTD (NR): 0%	Cequier et al. (2015). Comparing human exposure to emerging and legacy flame retardants from the indoor environment and diet with concentrations measured in serum.
Belgium NR	n = 1 pooled human cord blood	Mean (pg/g) BEH‐TEBP: ND EH‐TBB: 50 DBDPE: ND BTBPE: < LOQ	BEH‐TEBP (LOD, LOQ = 5, 15 pg/g) EH‐TBB (LOD, LOQ = 10, 40 pg/g) DBDPE (LOD, LOQ = 13, 30 pg/g) BTBPE (LOD, LOQ = 0.24, 1 pg/g)	Sales et al. (2017). Simultaneous determination of dechloranes, polybrominated diphenyl ethers and novel brominated flame retardants in food and serum.
Norway 2013–2014	n = 61 (16 males and 45 females)	Range (ng/g lipid) EH‐TBB: < LOD BEH‐TEBP: < LOD DBDPE: < 1.4‐27 BTBPE: < LOD DBE‐DBCH: < LOD OBTMPI: < 0.52–10 PBEB: < LOD PBT: < LOD TBCO: < LOD TBX: < LOD DBHCTD: < LOD	EH‐TBB (LOD = 36 pg/serum sample) BEH‐TEBP (LOD = 7.0 pg/serum sample) DBDPE (LOD = 42 pg/serum sample): 2% BTBPE (LOD = 17 pg/serum sample) DBE‐DBCH (LOD =3.9 pg/serum sample) OBTMPI (LOD =15 pg/serum sample): 2% PBEB (LOD =14 pg/serum sample) PBT (LOD =14 pg/serum sample) TBCO (LOD =7.3 pg/serum sample) TBX (LOD =14 pg/serum sample) DBHCTD (LOD = 63 pg/serum sample)	Tay et al. (2019). Serum concentrations of legacy and emerging halogenated flame retardants in a Norwegian cohort Relationship to external exposure.
European countries from the HBM4EU aligned studies (France, Greece, Slovenia) 2014–2021	n = 413, 55, 149 children blood	BEH‐TEBP: NR DBDPE: NR	BEH‐TEBP (LOD = 0.7 ug/L): 0%, NA, NA DBDPE (LOD = 0.013 ug/L): 13%, 16%, 18%.	Van der Schyff et al. (2023). Exposure to flame retardants in European children ‐ Results from the HBM4EU aligned studies.

Abbreviations: LOD, limit of detection; LOQ, limit of quantification; n, number of samples; ww, wet weight.

APPENDIX F. Overview of epidemiological studies identified for the BFRs assessed in this Opinion and published since the previous Opinion

TABLE F.1 Overview of studies in humans that have investigated the BFRs assessed in the current Opinion

Reference	Guo et al. ( 2018 ). Association between serum polybrominated diphenyl ethers, new flame retardants and thyroid hormone levels for school students near a petrochemical complex, South China.	Chen et al. ( 2019 ). Thyroid function and decabromodiphenyl ethane (DBDPE) exposure in Chinese adults from a DBDPE manufacturing area.	Wang, Chen, et al. (2019). Association between thyroid function and exposures to brominated and organophosphate flame retardants in rural Central Appalachia.	Zhao et al. ( 2021 ). Serum levels of novel brominated flame retardants (NBFRs) in residents of a major BFR‐producing region: Occurrence, impact factors and the relationship to thyroid and liver function.	Chen et al. ( 2022 ). Reduced mitochondrial DNA copy number in occupational workers from brominated flame retardants manufacturing plants.	Liu, Li, et al. (2022). Exposure to novel brominated flame retardants and organophosphate esters and associations with thyroid cancer risk: A case–control study in eastern China.	Mendy et al. ( 2023 ). Exposure to dust organophosphate and replacement brominated flame retardants during infancy and risk of subsequent adverse respiratory outcomes.	Mendy et al. ( 2024 ). Prenatal exposure to replacement flame retardants and organophosphate esters and childhood adverse respiratory outcomes.	Mendy et al. ( 2025 ). Epigenome‐wide association study of gestational exposure to organophosphate esters and replacement brominated flame retardants with newborn DNA methylation.	Zheng, Chen, et al. (2025). Overlooked environmental risks of type 2 diabetes: Evidence from a case–‐control study on the roles of dechlorane plus (DPs) and novel brominated flame retardants (NBFRs).	Zheng, Sun, et al. (2025). Exposure to novel brominated flame retardants (NBFRs) drives obesity via PPARγ‐mediated adipogenesis: A cross‐ethnic study in aging agro‐pastoral communities.
Country	China	China	USA	China	China	China	USA	USA	USA	China	China
Effects studied	Thyroid hormone levels (T3, T4, FT3, FT4, TSH),	Thyroid hormone levels in blood (TT4, FT4, TT3, FT3, TSH, TG‐Ab, TPO‐Ab9 in serum)	Thyroid hormone levels in blood (FT3, FT4, TSH)	Thyroid hormones (TT4, FT4, TT3, FT3, TSH, TG‐Ab, TPO‐Ab), serum lipids, liver injury biomarkers	Mitochondrial DNA copy number	Thyroid cancer risk, thyroid hormone levels (TT4, FT4, TT3, FT3, TSH)	Wheezing, respiratory infections, hay fever/allergies, lung function	Wheezing, respiratory infections, hay fever/allergies, lung function	DNA methylation	Type 2 diabetes	BMI
BFRs included in the 2012 Opinion and covered by Recommendation 2014/118/EU
TDBPP	–	–	–	–	–	–	–	–	–	–	–
EBTEBPI	–	–	–	–	–	–	–	–	–	–	–
HBCYD	–	–	–	–	–	–	–	–	–	–	–
BEH‐TEBP	Yes	–	Yes	–	–	Yes	Yes	Yes	Yes	Yes	Yes
EH‐TBB	Yes	–	Yes	–	–	Yes	Yes	Yes	Yes	–	–
DBNPG	–	–	–	–	–	–	–	–	–	–	–
Other BFRs included in the 2012 Opinion
DBDPE	Yes	Yes	Yes	Yes	Yes	Yes	–	–	–	–	–
BTBPE	Yes	–	–	Yes	–	Yes	–	–	–	–	–
HBB	–	–	Yes	Yes	–	Yes	–	–	–	Yes	Yes
DBE‐DBCH	Yes	–	–	–	–	–	–	–	–	–	–
PBEB	–	–	Yes	Yes	–	Yes	–	–	–	Yes	Yes
PBT	–	–	–	Yes	–	Yes	–	–	–	Yes	Yes
TDBP‐TAZTO	–	–	–	–	–	‐	–	–	–	–	–
TBCO	–	–	–	–	–	–	–	–	–	–	–
DBHCTD	Yes	–	–	–	–	–	–	–	–	–	–
HCTBPH	–	–			–	‐	–	–	–	–	–
PBB‐Acr	–	–	–	–	–	–	–	–	–	–	–
TBNPA	–	–	–	–	–	–	–	–	–	–	–
TBX	–	–	Yes	–	–	–	–	–	–	Yes	Yes
BDBP‐TAZTO	–	–	–	–	–	–	–	–	–	–	–
DBP‐TAZTO	–	–	–	–	–	–	–	–	–	–	–
DBS	–	–	–	–	–	–	–	–	–	–	–
HEEHP‐TEBP	–	–	–	–	–	–	–	–	–	–	–
4'‐PeBPO‐BDE208	–	–	–	–	–	–	–	–	–	–	–
TTBNPP	–	–	–	–	–	–	–	–	–	–	–
TTBP‐TAZ	–	–	–	–	–	–	–	–	–	–	–
OBTMPI	–	–	–	–	–	–	–	–	–	–	–

APPENDIX G. Studies informing the mode of action and studies on other model systems for the BFRs assessed in this Opinion

Table G.1.

Studies informing the mode of action and studies on other model systems of BEH‐TEBP and EH‐TBB

Reference	Comments
Ezechiáš et al. ( 2012 ). Hormonal activities of new brominated flame retardants	Reporter gene assays were used to assess the ability of BEH‐TEBP to activate or inhibit the transcriptional activity of human ER and AR. At a concentration of 14.1 μM BEH‐TEBP there was no oestrogenic, anti‐oestrogenic or androgenic activity.
Saunders et al. ( 2013 ). In vitro endocrine disruption and TCDD‐like effects of three novel brominated flame retardants: TBPH, TBB, & TBCO.	EH‐TBB, BEH‐TEBP and TBCO did not show AHR agonistic effects in two reporter gene assays. Maximal anti‐estrogenic effect of 62% was observed from EH‐TBB at 0.5 mg L−1 in the YES assay. BEH‐TEBP and TBCO had maximal anti‐androgenic effects in the YAS assay of 74% and 59% at 300 mg L−1 and 1500 mg L−1, respectively. Using the H295R assay, 0.05 mg L−1, 15 mg L−1 and 15 mg L−1 EH‐TBB, BEH‐TEBP and TBCO, respectively resulted in a 2.8‐fold, 5.4‐fold and 3.3‐fold increase in concentrations of E2.
Mankidy et al. ( 2014 ). Effects of novel brominated flame retardants on steroidogenesis in primary porcine testicular cells.	The effects of EH‐TBB, TBCO and BEH‐TEBP on steroidogenesis was studied in a porcine primary testicular cell model. TBCO and TBPH stimulated production of sex hormones testosterone and oestradiol was observed. Effects may have been mediated by changes in expression of genes encoding enzymes involved in the steroidogenesis pathway.
Fic et al. ( 2014 ). Estrogenic and androgenic activities of TBBA and TBMEPH, metabolites of novel brominated flame retardants, and selected bisphenols, using the XenoScreen XL YES/YAS assay.	In the XenoScreen XL YES/YAS assay, EH‐TBB and BEH‐TEBP, had no or very weak (anti)estrogenic or (anti)androgenic activities. Metabolites of TBBA and TBMEPH, showed anti‐oestrogenic (IC50 for TBBA = 31.75 uM and IC50 for TBMEPH = 0.265 uM) and anti‐androgenic (IC50 for TBBA = 73.95 uM and IC50 for TBMEPH = 2.92 uM) activities.
Saunders et al. ( 2015 ). A mixture of the novel brominated flame retardants TBPH and TBB affects fecundity and transcript profiles of the HPGL‐axis in Japanese medaka.	BEH‐TEBP and EH‐TBB, the two main components of Firemaster 550, were investigated in a 21‐day fecundity study in Japanese medaka for their potential effects on reproductive endocrinology‐related genes and fecundity following dietary exposure. This study followed OECD test 229 for short‐term reproductivity testing in fish. Medaka were fed a diet containing either 138 μg BEH‐TEBP and 144 μg EH‐TBB/g feed (wet weight) or 1422 μg BEH‐TEBP and 1474 μg EH‐TBB/g feed (wet weight) in addition to an unexposed control. The fish were fed at a ration of 6% of the bw per day. There were no treatment‐related mortalities or changes in liver and gonad weights. Cumulative production of eggs was reduced by 32% in females exposed to the higher concentration of the mixture. No effects on fecundity was observed at the lower concentration. Both male and female medaka showed downregulation of the vast majority of genes along the hypothalamus‐pituitary‐gonadal‐liver axis. In males, key genes involved in steroidogenesis, such as CYP17, 3β‐HSD and HMGR, were significantly downregulated, suggesting a disruption in testosterone synthesis. In females, expression of erα was significantly lower in the liver, while genes associated with production of egg yolk protein (vitellogenin 2 and choriogenin h) were upregulated.
Krivoshiev et al. ( 2016 ). Assessing in‐vitro estrogenic effects of currently‐used flame retardants.	A number of flame retardants were tested for oestrogenic activity in the MCF‐7 proliferation assay. BEH‐TEBP did not exhibit significant activity in this assay.
Klopcic et al. ( 2016 ). Comparison of in vitro hormone activities of novel flame retardants TBB, TBPH and their metabolites TBBA and TBMEPH using reporter gene assays.	Anti‐androgenic, anti‐thyroid hormonal and anti‐glucocorticoid activities were tested in luciferase reporter gene assays in mammalian cell models. Both BEH‐TEBP and EH‐TBB exhibited anti‐glucocorticoid activity by direct competition with the glucocorticoid receptor. BEH‐TEBP and EH‐TBB, and their carboxylic acid metabolites also exhibited anti‐androgenic and anti‐thyroid hormonal activities, without agonistic activities on the respective receptors. BEH‐TEBP was the most active.
Skledar et al. ( 2016 ). New brominated flame retardants and their metabolites as activators of the pregnane X receptor.	Transactivation assays were used to investigate the effects of BEH‐TEBP and EH‐TBB on nuclear receptor activities. Neither of the chemicals exhibited marked activity towards PPARa and PPARg. BEH‐TEBP and EH‐TBB both activated PXR in this system. BEH‐TEBP and its metabolite, mono(2‐ethylhexyl) tetrabromophthalate, showed PXR agonist activity, with EC50 values of 2.0 and 5.5 μM, respectively. BEH‐TEBP also showed PXR antagonist activity (IC50 = 13.9 μM). Molecular docking experiments indicated that similar binding of BEH‐TEBP and EH‐TBB to PXR. Additionally, exposure to BEH‐TEBP, EH‐TBB and their metabolites upregulated CYP3A4 mRNA expression, a gene regulated by PXR.
Usenko et al. ( 2016 ). Evaluation of Common Use Brominated Flame Retardant (BFR) Toxicity Using a Zebrafish Embryo Model	Zebrafish embryos were exposed to a number of flame retardants at concentrations of 1.25–20 mg/L from 6 to 168 post‐fertilisation and assessed for toxicity, the rate of spontaneous movement and acetylcholinesterase and glutathione‐S‐transferase activities. EH‐TBB was moderately toxic with an LC50 value of 7 mg/L but did not exhibit effects on the other parameters assessed. BEH‐TEBP had no effect at any of the tested concentrations.
Smythe et al. ( 2017 ). Impacts of Unregulated Novel Brominated Flame Retardants on Human Liver Thyroid Deiodination and Sulfotransferation.	Human liver microsomal and cytosolic fractions were used to explore the potential inhibitory effects of BEH‐TEBP and EH‐TBB on deiodinases (DIO) and sulfotransferases (SULT) following activation with T3 or T4. BEH‐TEBP inhibited DIO 3,3′‐T2 and T3 formation at the highest concentration tested (3.75 μM). There were no other effects on DIO and SULT activities.
Xiang et al. ( 2017 ). Effects of novel brominated flame retardant TBPH and its metabolite TBMEHP on human vascular endothelial cells: Implication for human health risks.	BEH‐TEBP and its metabolite mono‐(2‐ethyhexyl) tetrabromophthalate were tested separately for their effects and associated molecular mechanisms on human umbilical vein endothelial cells (0.02, 1, 2, 5, 10, 20 or 50 μg/mL). BEH‐TEBP induced early apoptosis after G2/M cell cycle arrest at 10 μg/mL. Its metabolite showed a higher toxicity (marked G0/G1 cell cycle arrest, high level of apoptosis and over expression of p53, cyclin‐dependent kinase inhibitors p21 and p27, underexpression cyclin D1, CDK2, CDK6, Bcl‐2).
Tung et al. ( 2017 ). Firemaster 550 and its components isopropylated triphenyl phosphate and triphenyl phosphate enhance adipogenesis and transcriptional activity of peroxisome proliferator‐activated receptor (Ppargamma) on the adipocyte protein 2 (aP2) promoter.	The authors investigated the adipogenic effects of Firemaster 550 and its components, focusing on the activation of the PPARγ receptor and the subsequent effects on adipogenesis. Firemaster 550 and some of its components including BEH‐TEBP induced lipid accumulation in murine 3T3‐L1 preadipocytes, but EH‐TBB did not. The adipogenic effect was dose‐dependent, showing significant lipid accumulation at 10 μM. Exposure to Firemaster 550 (100 uM), enhanced the transcriptional activity of the aP2 promoter via PPARγ. This was confirmed through reporter assays and chromatin immunoprecipitation (ChIP) experiments, showing PPARγ recruitment to the aP2 enhancer region. BEH‐TEBP (20 uM) did not cause PPARγ activation. EH‐TBB was not tested for this endpoint. In conclusion, Firemaster 550 promotes adipogenesis through PPARγ activation, but the main components causing this effect did not include BEH‐TEBP and EH‐TBB.
Kollitz et al. ( 2018 ). Chemical Mixtures Isolated from House Dust Disrupt Thyroid Receptor β Signaling.	The study tested 137 house dust samples and found that 41% of the extracts significantly antagonised TRβ signalling. Extracts contained 0.045–1266 μg EH‐TBB/mL and 0.566–564.4 μg BEH‐TEBP/mL. The concentrations of 12 flame retardants in the dust extracts were found to be significantly correlated with the observed TRβ antagonism with PBDE congeners showing the strongest correlations. However, when these flame retardants were tested individually up to concentrations of 10‐7 pg/mL (10‐16 g/L), they did not show any significant antagonistic activity.
Ma et al. ( 2018 ). Multigenerational Effects and Demographic Responses of Zebrafish (Danio rerio) Exposed to Organo‐Bromine Compounds	Exposure of zebrafish via feed at 3 levels of contamination of BEH‐TEBP (4.00, 15.95 and 63.80 μg/g) and EH‐TBB (3.00, 13.63 and 54.53 μg/g) for 21 days resulted in a dose‐related reduction of fertility in F0 females. The authors predicted extinction of the zebrafish population over multiple generations in 20 years.
Pagé‐Larivière et al. ( 2018 ). Prioritization of 10 organic flame retardants using an avian hepatocyte toxicogenomic assay.	BEH‐TEBP was exposed to hepatocytes isolated from chicken and double‐breasted cormorant hepatocytes for 24 h at concentrations of 0.1–300 uM. Cell viability, activity of 7‐ethoxyresorufin deethylase (EROD) and transcriptomics were assessed. EROD activity was induced indicating AHR activation.
Zhang et al. ( 2019 ). Screening for neurotoxic potential of 15 flame retardants using freshwater planarians.	BEH‐TEBP and EH‐TBB were tested for neurotoxicity using a planarian automated platform able to assess morphological and behavioural endpoints in adult or regenerating animals. BEH‐TEBP and EH‐TBB were tested at 5 different doses (0.01, 0.1, 1.0, 10, 100 μM). No effects of these two compounds were observed in regenerating and adult planarians.
Shockley et al. (2020a). Comparative toxicity and liver transcriptomics of legacy and emerging brominated flame retardants following 5‐day exposure in the rat.	Male Harlan SD rats were orally exposed to EH‐TBB or BEH‐TEBP for 5 days at doses ranging from 0.1 to 1000 μmole/kg bw per day. No treatment‐related liver histopathologic alterations were present with either of the test substances. EH‐TBB exposure for 5 days caused transcriptome changes in the rat liver with 2, 2 and 66 significantly regulated transcripts identified at doses of 0.55, 55 and 550 mg/kg bw per day, respectively. Several genes that are targets of the AHR (Cyp1a1, Cyp1a2, Gsta3, Nqo1 and Slc39a4) were upregulated. Of these, Cyp1a1 showed the strongest induction (162‐fold at 550 mg/kg bw per day). However, there was no corresponding increase in CYP1A1 activity, or indeed activities of several other xenobiotic metabolism enzymes analysed. The genes, Gsta3 and Nqo1, are also regulated by NRF2, which would be indicative of oxidative stress. The authors calculated a transcriptome BMDL for EH‐TBB, based on Gene Ontology Process categories, of 110.77 mg/kg per day. Other groups of rats were exposed to BEH‐TEBP at doses of 0.07, 0.71, 7.06, 70.6 or 706 mg/kg bw per day for 5 days. There were no significant liver disease‐related transcript changes following exposure to BEH‐TEBP up to the highest dose.
Chen et al. ( 2020 ). Effects of novel brominated flame retardants and metabolites on cytotoxicity in human umbilical vein endothelial cells.	Human umbilical vein endothelial cells were exposed to EH‐TBB and BEH‐TEBP, and their respective metabolites 2,3,4,5‐tetrabromo benzoic acid (TBBA) and mono‐(2‐ethyhexyl) tetrabromophthalate for 24h. Concentrations of 0.1, 1, 10, 50, 100 or 200 μg/mL were tested for cell viability, and of 0.8 and 8 μg/mL for LDH activity. Significant reduction of cell viability was observed with the two metabolites (50 or 10μM, respectively) whereas the parent compounds did not. Significant increases in LDH activity were measured with 8 μg/mL of BEH‐TEBP and the two metabolites. Significant gene overexpression (eNOS, VEGFR1) was reported with TBBA. Significant reductions in VEGFR1 expression and NO release were noted with BEH‐TEBP. Significant increases in adhesion molecule gene expression PECAM1 were observed with both metabolites 0.8 μg/mL whereas no changes were noted with the two parent compounds. In vivo, an acute oral administration of the two metabolites (0.8 μg/mL) induced a significant increase in platelet number. Metabolites caused higher toxicity and affected gene expression to a greater extent than their parent compounds EB‐TBB and BEH‐TEBP.
Yin and Zhang et al. (2021). Effects of bis(2‐ethylhexyl)‐2,3,4,5‐tetrabromophthalate on liver injury in Balb/c mice.	Levels of ROS, MDA, 8‐OH‐dG and DNA‐protein cross‐links were increased, and levels of GSH were decreased by BEH‐TEBP treatment. Western blotting analysis showed upregulation of the endoplasmic reticulum stress related proteins GRP78, PERK and CHOP by BEH‐TEBP. The effects of BEH‐TEBP were attenuated by co‐administration of 25 mg/kg bw per day curcumin, which was attributed to its antioxidant properties.
Liu, Song, et al. (2022). Bis (2‐ethylhexyl)‐2, 3, 4, 5‐tetrabromophthalate showed poor penetrability but increased the permeability of blood brain barrier: Evidences from in vitro and in vivo studies	BEH‐TEBP was tested for its toxicity on the blood–brain barrier (BBB) using hCMEC/D3 cells and in zebrafish models. In addition, behavioural toxicity of BEH‐TEBP in zebrafish was assessed. In vitro, BEH‐TEBP did not impair the cell viability and had a low level of penetrability through the BBB. In vivo measurement of BBB permeability using transgenic zebrafish exposed to BEH‐TEBP (0, 0.02, 0.2, 2 or 20 μM, 0 to 5 days post‐fertilisation) showed a significant increase in BBB permeability in fishes exposed to 2 and 20 μM BEH‐TEBP through modifications of the gene expression of tight junction proteins. Concomitant increases in the head content of BEH‐TEBP were measured at the two same doses in wild‐type zebrafish larvae. The assessment of neurotoxicity in wild‐type zebrafish larvae did not show any effects of the exposure to BEH‐TEBP (0, 0.02, 0.2, 2 or 20 μM, 0 to 5 days post‐fertilisation) on the neurodevelopment, neurotransmission and locomotor activity of the animals. Within the same experiment, the 2 and 20 μM exposed groups showed a significant decrease in the phototactic behaviour compared to the control group and a reduction in the expression of all the opsin genes measured in the same two groups. This study suggests a limited neurotoxicity of BEH‐TEBP regarding the neurodevelopment and neurotransmission of the brain, and the general activity of the animals. However, the alteration of phototactic response may be indicative of another form of neurotoxicity more related to the predator avoidance response.
Maia et al. ( 2022 ). Impact of brominated flame retardants on lipid metabolism: An in vitro approach.	EH‐TBB was studied for its ability to disturb lipid metabolism. Preadipocyte (3T3‐L1) cell line and hepatocyte (HepG2) cell line were exposed to EH‐TBB acutely (50, 100 or 250 nM, 24h) or chronically (250 nM, 4 weeks). Preadipocyte proliferation and differentiation were not affected by EH‐TBB (50‐250 nM, 24h) whereas the fatty acid composition has been changed with a short chain fatty acid decrease between day 3 and day 11 after exposure. In hepatocytes, 4‐week exposure to EH‐TBB 250 nM induced significant increases in lipid accumulation and long chain fatty acide concentration in culture medium. EH‐TBB reduced the expression of lipid metabolism‐related genes like PPARγ and PGC‐1α.
Sousa et al. ( 2022 ). Brominated flame retardants effect in MCF‐7 cells Impact on vitamin D pathway.	EH‐TBB was assessed for its toxicity on MCF‐7 breast cancer cells and its impact on vitamin D metabolism. Cells were incubated with EH‐TBB for 24 h in a range 10‐500 nM resulted only in a significant reduction of cell proliferation at a dose of 10 nM. The exposure of cells to 250 nM for 4 weeks induced early disturbances in cell migration after injury with values returning to control ones later whereas cell proliferation and MMP‐9 prtein expression remained unchanged. EH‐TBB 250 nM 4 weeks significantly reduced vitamin D gene expression in human hepatic cells (CYP2R1) and MCF‐7 cells (CYP24A1, RxRα), and impaired cell cycle response to vitamin D.
Esplugas et al. ( 2023 ). In vitro neurotoxic potential of emerging flame retardants on neuroblastoma cells in an acute exposure scenario	The acute neurotoxicity of BEH‐TEBP and EH‐TBB on neuroblastoma cells SH‐SY5Y was assessed at concentrations ranging 2.5–20 μM during 2–24 h. Cell viability, mitochondrial function, neural plasticity and development, promotion of oxidative stress and inflammation were not affected by BEH‐TEBP or EH‐TBB suggesting the absence of toxicity in this neural cell line.
Witchey et al. ( 2023 ). Impacts of Gestational FireMaster 550 (FM 550) Exposure on the neonatal cortex are Sex Specific and largely Attributable to the Organophosphate Esters.	PND1 siblings of the animals used for behavioural studies (Witchey et al., 2020) were assessed for transcriptomic and lipidomic differences in the cortex. Differentially expressed genes indicated multiple mechanisms included dysregulation of mitochondrial function and disruption of cholinergic and glutamatergic systems. Affected lipid classes included ceramides, sphingomyelins and triacyl glycerides. Comparison with organophosphate ester flame retardants (OPFR) suggested that OPFRs contribute more than the BFRs to the developmental neurotoxicity of Firemaster 550.
Yuan et al. ( 2024a ). IL‐10‐TGTPO‐T4 axis, the target of bis (2‐ethylhexyl) tetrabromophthalate on thyroid function imbalance.	Six‐weeks old Sprague–Dawley male rats were dosed with BEH‐TEBP (95% pure) control (corn oil), 5, 50 and 500 mg/kg bw per day for 28 days by gavage. The treatment altered the expression of thyroid‐specific proteins, including thyroglobulin (TG), thyroid peroxidase (TPO) and sodium iodide symporter (NIS). The mRNA and protein levels of TG and TPO were significantly elevated, indicating that BEH‐TBPH might stimulate thyroid hormone synthesis in rats. BEH‐TBPH exposure also affected the expression of cytokines, particularly IL‐6 and IL‐10. Both IL‐6 and IL‐10 were upregulated in thyroid tissues of rats treated with BEH‐TBPH. The elevated levels of IL‐10 were notably linked to increased expression of TG and TPO. The effects of BEH‐TBPH were further investigated in human thyroid follicular epithelial Nthy ori3–1 cells, where BEH‐TBPH treatment also increased the mRNA and protein levels of TG, TPO and IL‐10 at 100 nM, which was the highest concentration tested. Further experiments using RNA interference to knock down IL‐6 and IL‐10 showed that IL‐10 was involved in the upregulation of TG and TPO. Silencing IL‐10 diminished the effects of BEH‐TBPH on these thyroid proteins, suggesting that IL‐10 is a target through which BEH‐TBPH influences thyroid function.
Fu, Zhu, et al. (2024). Bis(2‐ethylhexyl)‐tetrabromophthalate Poses a Higher Exposure Risk and Induces Gender‐Specific Metabolic Disruptions in Zebrafish Liver.	Waterborne exposure to 2 uM BEH‐TEBP for 28 days resulted in predominant accumulation in the liver. The liver of males accumulated more BEH‐TEBP than that in females. Effects of BEH‐TEBP in the liver were studies after 2‐month exposure to 0, 0.2, 2, 20, 200 nM. Female zebrafish showed increased glycogen levels and reduced cholesterol, while male zebrafish exhibited altered lipid and glucose metabolism, with increased blood glucose, triglycerides and liver fat accumulation. There was dose‐dependent increase in visceral fat accumulation which was significant at 2 nM exposure in both sexes. Male zebrafish had higher visceral fat accumulation and more severe liver damage, including signs of non‐alcoholic fatty liver disease (NAFLD), compared to females. Male liver transcriptome analysis indicated a significant upregulation of glycolysis and insulin‐signalling pathways, contributing to these metabolic changes. Histopathological analysis revealed liver damage in both sexes, with male zebrafish showing more severe inflammation and oxidative stress.
Yuan et al. ( 2024b ). Effects of 1, 2‐bis (2,4, 6‐tribromophenoxy) ethane and bis (2‐ethylhexyl) tetrabromophthalate on serum metabolic and lipid profiles in male rats.	Adult rats were orally exposed (gavage) to BEH‐TEBP at 500 mg/kg bw per day for 28 days and serum samples collected for metabolomics and lipidomics analysis. BEH‐TEHBP induced specific metabolic and lipidomic changes (19 metabolites and 70 lipids differentially affected). Reliable differential substances possibly related with changes in thyroid function (dehydroascorbic acid, acylcarnitine and diglyceride) were identified to be impacted by BEH‐TEBP. Cyclooxygenase‐2, an enzyme related to oxidative stress, was selected as a potent target for BEH‐TEBP. BEH‐TEBP was showed to impact the rat metabolism possibly related to changes in thyroid function.
Fu, Hua, et al. (2024). Integrated Studies on Male Reproductive Toxicity of Bis(2‐ethylhexyl)‐tetrabromophthalate: in Silico, in Vitro, ex Vivo, and in Vivo.	BEH‐TEBP was found to bind stably to the androgen receptor (AR) in in silico molecular docking studies. It exhibited antagonistic, but not agonistic, activity on AR signalling using an AR‐GRIP1 yeast assay system. Mouse Leydig cells were exposed to BEH‐TEBP at 0, 0.2, 2, 20 and 200 μM (equal to 0, 141, 1412, 14,122 and 141 228 μg/L). BEH‐TEBP inhibited the proliferation of Leydig cells by 0.8%, 3.1%, 7.3% and 22.5% at concentrations of 0.2, 2, 20 and 200 μM, respectively. Leydig cells showed inhibited testosterone production at 200 μM. In zebrafish, BEH‐TEBP exposure impaired sperm motility and movement patterns at all concentrations tested. Male zebrafish (F0) exposed to BEH‐TEBP displayed reduced mating behaviour when paired with unexposed females with significant changes observed at all concentrations starting at 0.002 μM. AR expression was specifically suppressed in the brain, providing possible mechanistic evidence for endocrine‐disrupting potential via central nervous system pathways. The offspring of exposed males showed dose‐dependent changes in behavioural locomotor patters with significance observed at concentrations of 0.02 μM and higher, indicating transgenerational effects.

Abbreviation: NR, not reported.

Table G.2.

Studies informing the mode of action of DBDPE

Reference	Summary of main conclusions
Wang et al. ( 2010 ). Comparative tissue distribution, biotransformation and associated biological effects by decabromodiphenyl ethane and decabrominated diphenyl ether in male rats after a 90‐day oral exposure study	In male rats orally exposed for 90 days to 100 mg DBDPE/kg bw per day. The CYP3A2 mRNA was slightly increased in liver tissue, but other Phase 1 biotransformation enzymes (CYP1A1, CYP2B1, CYP2B2) were not affected at the mRNA level. The authors suggested that DBDPE may induce PXR‐dependent gene expression (induction of CYP3A biotransformation enzymes), but no AHR and CAR‐dependent gene expression and consequent possible adverse effects.
Sun et al. ( 2012 ). Cytotoxicity and apoptosis induction in human HepG2 hepatoma cells by decabromodiphenyl ethane	HepG2 cells were cultured in the presence of DBDPE at 0 (DMSO), 3.125, 6.25, 12.5, 25.0, 50.0, 100.0 μg/mL for 24, 48 and 72 h, respectively. DBDPE inhibited HepG2 viability (MTT test) in a time and dose‐dependent manner within a range of 12.5 to 100 mg/L and for 48 and 72 h. Induction of apoptosis was detected at 12.5–100 μg/mL at 48 and 72 h accompanied with overproduction of ROS. Moreover, N‐acetyl‐L‐cysteine (ROS scavenger), significantly reduced DBDPE‐induced ROS levels and increased HepG2 cells viability.
Sun et al. ( 2014 ). Sub‐acute effect of decabromodiphenyl ethane on hepatotoxicity and hepatic enzyme activity in rats	Male and female Wistar rats were exposed by gavage for 28 days to DBDPE at 0, 50, 100, 250, 500, 1000 mg/kg bw per day. No significant changes observed in liver weight (absolute or relative) following exposure. The Glu, TBA, ALT and AST levels following exposure to various doses of DBDPE were markedly higher. The Glu level was higher after exposure to ≥ 250 mg/kg bw per day. DBDPE (≥ 250 mg/kg bw per day) increased UDPGT, 7‐pentoxyresorufin O‐depentylase (PROD) and luciferin benzylether debenzylase activities, and induced CYP2B2, CYP3A1 and CYP3A2 mRNA, CYP3A2 and CYP2B1/2 proteins. Increasing UDPGT activity with increasing dose suggest that oral DBDPE exposure induces drug‐metabolising enzymes in rats via the CAR/PXR signalling pathway. According to the study authors, the induction of CYPs and co‐regulated enzymes of phase II biotransformation may affect the homeostasis of endogenous substrates, including thyroid hormones, which may, in turn, alter glucose metabolism.
Smythe et al. ( 2017 ). Impacts of unregulated novel brominated flame retardants on human liver thyroid deiodination and sulfotransferation	The inhibitory effects of DBDPE on DIO and SULT activity were investigated using human in vitro liver microsomal and cytosolic bioassays (concentrations: 0.13 to 130 nM or 260 nM). The potential inhibition of hepatic ORD and IRD activity in the presence of DBDPE was monitored by evaluating the production of T3 and rT3 from T4, respectively. There was almost complete inhibition of T3 and 3,3’‐T2 formation from T4 and 3,3’‐T2 formation from rT3 by DBDPE at the highest dose tested (260 nM). A concentration−response relationship was observed for formation of T3 from T4 IRD and for formation of 3,3’‐T2 from rT3 ORD. Significant inhibition of T2 formation from T4 ORD/IRD was only observed at the highest dose level. Inhibition of 3,3’‐T2 sulfation was investigated by monitoring sulfotransferase activity in the presence of DBDPE. No statistically significant inhibition of SULT activity was observed (up to 130 nM).
Wang, Chen, et al. (2019). A comparison of the thyroid disruption induced by decabrominated diphenyl ethers (BDE‐209) and decabromodiphenyl ethane (DBDPE) in rats	Male Sprague–Dawley rats were exposed to DBDPE by gavage for 28 days to 0, 5, 50 or 500 mg/kg bw per day. Doses 50 and 500 mg/kg bw per day could downregulate the expression of TSHR, NIS, TPO and TG genes in thyroid gland. Slight decreased TT4 and TT3 levels in serum after DBDPE exposure might be caused by impair of key components of TH biosynthesis, such as TSHR, NIS, TG and TPO. High levels of DBDPE exposure decreased the expression of Pax8 genes which might explain the decrease of TG, TPO, NIS and TSH receptor (TSHR) expression, eventually leading to reduced synthesis of T4 and T3. DBDPE exposure could up‐regulate the expression of TRHR gene in the pituitary compared to control. According to the authors, reduced thyroid hormone levels in serum had a feedback effect to the pituitary, and as a consequence, TSH synthesis was increased.
Sun et al. ( 2020 ). Hepatotoxicity of decabromodiphenyl ethane (DBDPE) and decabromodiphenyl ether (BDE‐209) in 28‐day exposed Sprague‐Dawley rats	Sprague–Dawley male rats were exposed by gavage to DBDPE (purity > 98%) for 28 days to 0, 50, 500 mg/kg bw per day. DBDPE downregulated PXR and CAR at the two highest doses, and CYP3A1 and CYP3A2 expression. No significant change was found on TNF‐a and IL‐6 in the rat liver in the exposed groups.
Choi et al. ( 2021 ). Investigation on Sex Hormone‐Disruption Effects of Two Novel Brominated Flame Retardants (DBDPE and BTBPE) in Male Zebrafish (Danio rerio) and Two Human Cell Lines (H295R and MVLN).	In male zebrafish, 14‐day exposure to DBDPE significantly increased 17β‐estradiol (E2) concentrations. In H295R cells, an E2/T ratio showed an increasing trend by DBDPE exposure.
Li et al. ( 2021 ). BDE‐209 and DBDPE induce male reproductive toxicity through telomere‐related cell senescence and apoptosis in SD rat	28‐day exposure of male Sprague–Dawley rats to DBDPE could damage the telomeric function by shortening telomere length and reducing telomerase activity at 500 mg/kg bw per day, which consequently caused cell senescence and apoptosis in testis of rat. The numbers of senescent cells per seminiferous tubule were increased significantly in the 5, 50 and 500 mg/kg bw per day and the expression of p53, p21 and p16 were increased in the 50 and 500 mg/kg bw per day. According to the authors, these effects could contribute to the decline of sperm quality and quantity.
Shi and Feng ( 2021 ). Decabromodiphenyl ethane exposure damaged the asymmetric division of mouse oocytes by inhibiting the inactivation of cyclin‐dependent kinase 1	Three‐week‐old female mice were orally exposed to DBDPE at 0, 0.05, 0.5, 5, 50 μg/kg bw per day for 30 days. Exposure did not affect the first polar body extrusion of oocytes, but asymmetric division of oocytes was markedly impaired at 5 and 50 μg/kg bw per day, which resulted in oocytes with larger polar bodies. DBDPE exposure inhibited the spindle migration and membrane protrusion in oocytes during anaphase of meiosis I (anaphase I), thereby impairing asymmetric division. DBDPE exposure suppressed the inactivation of cyclin‐dependent kinase 1 (Cdk1), resulting in the decrease of cytoplasmic formin2 (FMN2)‐mediated F‐actin polymerisation in oocytes at the onset of anaphase I. Simultaneously, DBDPE exposure damaged the structural integrity of the spindle and the perpendicular relationship between spindle and cortex. These together led to the failure of spindle migration and membrane protrusion required for oocytes asymmetric division. Finally, DBDPE exposure injured the development of blastocysts, leading to blastocyst apoptosis.
Shi, Qiu, et al. (2021). The toxic effects and possible mechanisms of decabromodiphenyl ethane on mouse oocyte	DBDPE inhibited the rate of germinal vesicle breakdown (GVBD), first polar body extrusion (at 20–100 μM) and fertilisation of mouse oocytes (at 50 μM). After 14 h exposure, the hardness of zona pellucida metaphase II oocytes markedly increased and the spindle morphology was abnormal. DBDPE exposure‐induced abnormal mitochondrial distribution, mitochondrial dysfunction and ATP deficiency. It also downregulated the expression of antioxidant‐related genes (Sod2, Gpx1) and increased the level of ROS in oocytes. Autophagy occurred in treated oocytes with high expression of autophagy‐related protein (LC3) and genes (Lc3, Beclin1). In addition, DBDPE significantly upregulated the protein (Bax) and mRNA (Bax, Caspase3) levels of pro‐apoptosis genes. However, the protein and mRNA expression of anti‐apoptosis genes Bcl‐2 was downregulated in exposed oocytes. The authors concluded that DBDPE exposure impaired mitochondrial function, causing oxidative damage, autophagy and apoptosis in oocytes.
Zheng et al. ( 2021 ). NLRP3 inflammasome‐mediated endothelial cells pyroptosis is involved in decabromodiphenyl ethane‐induced vascular endothelial injury	Sprague–Dawley rats were exposed to DBDPE (purity ≥ 98%) at 0, 5, 50, 500 mg/kg bw per day by gavage for 28 days. At the two highest doses there were increases in the protein levels of IL‐1b and IL‐18 in the abdominal aorta, demonstrating inflammatory reaction. Doses of 50 and 500 mg/kg bw per day induced the activation of NLRP3 inflammasome, induced pyroptosis through activated caspase‐1 in abdominal aorta endothelium of rats and thus determined the important roles of NLRP3 inflammasome and caspase‐1 in cardiovascular diseases. In vitro, human vascular endothelial cells (HAECs) were treated for 24 h with 0 (DMSO), 6.25, 12.5, 25, 50, 100 mM DBDPE. DBDPE decreased viability, induced cytotoxicity (increase LDH activity) at concentrations ≥ 12.5 mM. It induced ROS generation and caused HAECs pyroptosis at 25 and 100 mM, which was evidenced by the elevated expression of Nod‐like receptor protein‐3 (NLRP3), ASC and caspase‐1 in DBDPE‐treated group. The authors constructed NLRP3 knockdown HAECs by lentivirus‐mediated short hairpin RNA (shRNA). It was shown that NLRP3 knockdown downregulated DBDPE‐induced increases of caspase‐1 activity and caspase‐1, ASC and NLRP3 mRNA and protein expression levels. The authors indicated that DBDPE caused HAECs pyroptosis by activating NLRP3 inflammasome and caspase‐1.
Gao et al. ( 2022 ). Fat mass and obesity‐associated gene (FTO) hypermethylation induced by decabromodiphenyl ethane causing cardiac dysfunction via glucolipid metabolism disorder	In the same male rats exposed for 28 days as in Sun et al. (2020), DBDPE induced significant disruption of myocardial systolic and diastolic function as shown by the decreased level of MYH6 and increased level of SERCA2 (the two key proteins involved in the maintenance of homeostasis during myocardial contractile and diastolic processes). Mitochondrial respiration‐related protein was measured, and the expressions of UCP2 decreased in all the treated groups in a dose‐dependent manner. The ATP level in heart tissue of rats was significantly reduced at 500 mg/kg bw per day. DBDPE also increased the serum levels of glucose (at 500 mg/kg bw per day) and low‐density lipoprotein but decreased the content of high‐density lipoprotein (at the two high doses). The expressions of FTO in the 50 and 500 mg/kg bw per day groups were reduced. Methylation level of the promoter region of the FTO gene were increased by DBDPE. The authors concluded that DBDPE exposure‐induced glycolipid metabolism disorder contributes to myocardial dysfunction through cardiomyocytes fibrosis and apoptosis via the mitochondrial‐mediated apoptotic pathway resulting from mitochondrial damage and dysfunction. FTO hypermethylation mediated this pathogenetic process via the PI3K/AKT/GLUT2, PPARγ/RXRα and mitochondrial apoptosis pathways. AC16 cells (Ventricular Cardiomyocyte Cell Line) were exposed to 0, 25, 50, 100 μM DBDPE, 0 μM DBDPE+12 μM 5‐aza and 100 μM DBDPE+12 μM 5‐aza for 24 h. The methylation rates were elevated critically in 100 μM DBDPE. FTO level significantly decreased in 50 and 100 μM groups, while 5‐aza, the inhibitor of DNA methylation, could significantly reverse the level induced by DBDPE. Cells viability gradually decreased with the increase of DBDPE concentration. DBDPE at 100 μM downregulate the level of MYH6. Additionally, the levels of SERCA2 increased in the 50 and 100 μM groups. In addition, the ROS level in 100 μM group was significantly increased when compared to that of control. DBDPE significantly decreased the level of ATP. 5‐aza could partially reverse the changed expressions of MYH6 and SERCA2 induced by DBDPE. Also, 5‐ aza could reduce the increase of ROS and antagonise the decrease of ATP level caused by DBDPE. The apoptotic rate of the cells increased at 100 μM. The UCP2 level was significantly decreased in all DBDPE exposure groups indicating the effect of DBDPE on mitochondria function.
Hua et al. ( 2022 ). Decabromodiphenyl ethane induced hyperactivity in developing zebrafish at environmentally relevant concentrations.	Zebrafish embryos were exposed to DBDPE (0, 2.91, 9.71, 29.14, 97.12 μg/L) from 2 to 120 h post‐fertilisation (hpf). Behavioural assessment of embryonic motor reflexes and larvae locomotor activity was performed at various times during the post‐fertilisation period. Morphological organisation of muscle fibres and motor neurons was assessed at 36 and 120 hpf. Whole‐body neurotransmitter content, cholinergic‐related enzyme activities (choline acyltransferase, ChAT and acetylcholine esterase, AchE) and specific genes expression related to brain development and nicotinic acetylcholine receptors were measured at 120 hpf. Embryonic exposure to DBDPE caused hyperactivity without affecting the development of secondary motoneuron axons and muscle fibres. However, further results implicated that DBDPE may affect the locomotor regulatory network via different mechanisms at lower and higher concentrations. Whole‐body neurotransmitter concentrations were significantly reduced at the lower dose of DBDPE whereas specific increases in acetylcholine content and ChAT activity, and concomitant reduction in AchE, were observed only at the two highest levels of dose. Molecular docking results showed direct interaction of DBDPE with ChAT. Results suggested that hyperactivity observed at higher concentrations of DBDPE was probably related to disruption of the cholinergic system with ChAT as a potential target.
Jing et al. ( 2022 ). Oxidative stress and endoplasmic reticulum stress contributed to hepatotoxicity of decabromodiphenyl ethane (DBDPE) in L‐02 cells	Human hepatocytes (L‐02) were exposed to 0 (DMSO), 6.25, 12.5, 25, 50 and 100 μM DBDPE. It caused L‐02 cell viability decrease (25–100 μM), increased lactate LDH (at 25 and 50 μM) and AST (concentration‐dependent) and ALT (at 25 and 50 μM) levels, ultrastructural damage and apoptosis. Moreover, DBDPE induced oxidative stress (i.e. increased ROS generation and MDA levels and decreased GSH content, SOD activity and mitochondrial membrane potential at 50 μM) and ER stress as evidenced by the elevated PERK and IRE‐1a expression levels. The CYP3A expression level was decreased in cells exposed o 50 μM. Pretreatment of L‐02 cells with antioxidant NAC and endoplasmic reticulum stress inhibitor 4‐PBA inhibited DBDPE‐induced oxidative stress, endoplasmic reticulum stress, CYP3A expression decrease and apoptosis.
Shi, Xu, et al. (2022). Decabromodiphenyl ethane affects embryonic development by interfering with nuclear F‐Actin in zygotes and leads to cognitive and social disorders in offspring mice.	Female Kunming mice were exposed for 30 days to 50 μg DBDPE/kg bw per day. After treatment, female mice were i.p. injected with PMSG and 48 h later, they were injected with HCG. Then, the females were mated with sexually matured males. DBDPE exposure did not affect spindle rotation in oocytes after fertilisation, but led to a decrease of pronuclei (PN) in zygotes. DBDPE interfered with the self‐assembly of F‐actin in pronuclei of embryo, resulting in their size reduction in zygotes. In addition, DBDPE reduced the expression of the Zscan4c transcription factor involved in maintaining genome stability during embryogenesis. The authors concluded that this early developmental instability may lead to metabolic and structural defects observed in later life, and that DBDPE exposure might affect embryonic development and result in behavioural defects in F1 mice related to learning and memory, spontaneous activity and social behaviour. The authors reported that DBDPE significantly increased γ‐H2AX expression in both male and female pronuclei (PN), suggesting that DBDPE exposure elevated DNA damage and that the accumulation of DNA damage was associated with reduced F‐actin levels. However, the Panel noted that increased γ‐H2AX expression in zygotes is not necessarily indicative of DNA damage. Instead, γ‐H2AX can also be involved in normal developmental processes, such as chromatin remodelling after fertilisation and genome reprogramming.
Shi, Qiu, et al. (2022). Exogenous melatonin protects preimplantation embryo development from decabromodiphenyl ethane‐induced circadian rhythm disorder and endogenous melatonin reduction.	DBDPE (5, 50, 500 μg/kg bw per day) exposure‐induced circadian rhythm disorder in adult female Kunming mice exposed by gavage for 14 days. It decreased also endogenous melatonin levels during pregnancy, which impaired preimplantation embryo development. DBDPE exposure impaired embryos by inducing oxidative stress, abnormal DNA methylation, mitochondrial dysfunction and mechanical property changes. DBDPE‐induced embryo damage led to adverse reproductive outcomes. DBDPE exposure led to endogenous rhythm disorder in mice because it reduced the expression of BMAL1. After pregnancy, DBDPE‐exposed mice with circadian rhythm disorders exhibited an inhibition of preimplantation embryo development. Therefore, DBDPE exposure‐induced circadian rhythm disorder resulting in embryo damage. The concentration of melatonin in the serum and oviduct of DBDPE‐exposed mice decreased significantly. The authors suggested that the circadian rhythm disorder and decrease in endogenous melatonin in DBDPE‐exposed mice inhibited preimplantation embryo development. It was also shown that DBDPE might affect epigenetic modifications in early embryos. Melatonin rescued the abnormal DNA methylation induced by DBDPE to improve preimplantation embryo development. The number of mitochondria dramatically increase to match the energy requirements of blastocyst formation, and mitochondrial dysfunction or deficiency induces developmental arrest of early embryos. It was shown that ATP levels in the blastocysts in DBDPE‐treated mice were reduced. This was related to the decrease in MMP and the number of mitochondria induced by DBDPE. The decline in the number of mitochondria may have been caused by the decreased expression of Drp1, which regulates mitochondrial division The authors reported that the number and functional disorder of mitochondria caused by DBDPE directly led to a reduction in melatonin synthesis, and that since mitochondria are also the location of melatonin functions, supplementation with melatonin rescued the mitochondrial dysfunction induced by DBDPE.
Yan et al. ( 2022 ). In utero exposure to decabromodiphenyl ethane causes rapid growth in mice cubs by activating glycogenolysis and lipid synthesis	Pregnant mice were exposed to DBDPE at 0, 0.1, 10 mg/kg bw by oral gavage daily until the cubs were born. At PND1, the female to male ratio was counted. Then 8 cubs were randomly selected in each dam and continued to be fed (8 cubs each dam, 10 dam each group). Exposure to DBDPE during pregnancy had no significant effect on the body weight of pregnant mice. The litter size and the sex ratio were not significantly changed. However, DBDPE exposure in utero caused low birth weight in cubs and then the cubs grow rapidly during lactation. It was shown that in utero exposure to DBDPE affects the whole‐body metabolic profile of newborns; the cubs absorbed the energy quickly and underwent active glucose metabolism and lipid accumulation for rapid growth. After 25 days of rapid energy accumulation, cubs exposed to DBDPE were sufficient to alter metabolic profiles and body composition such as increased lipid content, triglyceride and cholesterol levels, and ultimately lead to early symptoms of obesity. The authors concluded that exposure to DBDPE may cause cubs to undergo active glycolysis pathway during which large amounts of pyruvate were produced, thereby promoting lipid accumulation and ultimately obesity.
Xue et al. ( 2022 ). Decabromodiphenyl ethane induces male reproductive toxicity by glycolipid metabolism imbalance and meiotic failure	Male Sprague–Dawley rats were exposed by gavage to 0, 5, 50 and 500 mg DBDPE/kg bw per day. The meiosis regulator MILI (piwi‐like RNA‐mediated gene silencing 2) expression was downregulated in the cytoplasm of spermatocytes at 50 and 500 mg/kg bw per day. DBDPE downregulated the expressions of spermatogenesis‐ and oogenesis‐specific helix‐loop‐helix transcription factor 1 (Sohlh1), and synaptonemal complex proteins 1 and 3 (SYCP1 and SYCP3) at these doses. The Cyclin A expression was reduced in all treated groups and the expression of cyclin‐dependent kinase 2 (CDK2) was downregulated only in the highest dose group. DBDPE activated the mitochondrial apoptotic pathway in the testes. The Bcl‐2/Bax ratio was found to be downregulated, while the expressions of Cytochrome C, Caspase‐9 and Caspase‐3 were upregulated in the 500 mg/kg bw per day group only. The expression of Cleaved Caspase‐ 9 and Cleaved Caspase‐3 increased in the 50 and 500 mg/kg bw per day groups. So, DBDPE exposure result in meiotic failure leading to chromosomal association disorder in meiosis and spermatocyte cycle arrest. DBDPE produced mitochondrial damage and dysfunction and significantly increased serum glucose and low‐density lipoprotein cholesterol levels. Also, DBDPE caused meiotic failure through testicular DNA damage, reduced the expressions of meiotic regulatory factors (Sohlh1, MILI, CDK2 and CyclinA) and synaptonemal complex proteins SYCP1 and SYCP3. The authors concluded that DBDPE induced glycolipid metabolism disorder and promoted apoptosis of spermatogenic cells through the mitochondrial‐mediated apoptosis pathway, leading to reproductive toxicity by decreasing sperm quantity and quality.
Dong, Wang, Zhang, et al. (2023). DBDPE and ZnO NPs synergistically induce neurotoxicity of SK‐N‐SH cells and activate mitochondrial apoptosis signaling pathway and Nrf2‐mediated antioxidant pathway.	The present study assessed the neurotoxicity of DBDPE alone or in combination with Zinc oxide particles on human neuroblastoma SK‐N‐SH cells. Cells were exposed to DBDPE (0, 0.1, 0.5, 1.0, 2.5, 5.0, 12.5 and 25 mg/L) for 12, 24, 36 and 48 h. DBDPE induced a significant reduction in cell viability after 24 h of exposure at doses of 5.0, 12.5 and 25 mg/L. Concomitant LDH activity increases were observed at the same doses. DBDPE 25 mg/L induced a significant increase in ROS production measured through the oxidation of the fluorescent probe DCFH‐DA, the MDA production, the SOD activity and the intracellular GSH content. DBDPE 25 mg/L induced mitochondrial damages and cell apoptosis reflected by significant reductions in mitochondrial membrane potential, and concomitant increases in caspase‐3 and ‐9 activities, and in the expression of apoptosis‐related genes and proteins (Bax, Bcl‐2, caspase‐3, Cyt C). Proportions of apoptotic cells were significantly increased after exposure to the same dose of DBDPE. The Nrf2/ARE signalling pathway was activated with DBDPE 12.5 and 25 mg/L to restore the redox stat in response to stimulation by DBDPE and induction of oxidative stress. The results suggest the neurotoxicity of DBDPE with the induction of oxidative stress and apoptosis as potential mechanisms of toxicity.
Dong, Wang, Wang, et al. (2023). Investigating the adverse outcome pathways (AOP) of neurotoxicity induced by DBDPE with a combination of in vitro and in silico approaches.	The neurotoxicity of DBDPE was investigated through a combination of in silico and in vitro approaches in human neuroblastoma SK‐N‐SH cells. Molecular docking simulation showed the ability of DBDPE to interact with PARP1 (poly (ADP‐ribose) polymerase‐1), an enzyme that participates to a caspase‐independent apoptosis pathway and to change the secondary structure of the protein. In SK‐n‐SH cells, DBDPE 12.5 and 25 mg/L induced significant increases in ROS production and MDA content, and reductions in SOD activity and GSH cell content. The same doses of DBDPE increased significantly the apoptosis rate measured through the staining of the Hoechst 33342 fluorescent probe. DBDPE 12.5 and 25 mg/L induced mitochondrial damages reflected by significant reductions in mitochondrial membrane potential and ATP production. The same doses of DBDPE increased PARP1 activity and AIF (apoptosis‐induced factor) translocation from mitochondria to the nucleus, leading to caspase‐independent apoptosis. Caspase‐3 and ‐9 activities and apoptosis‐related gene and protein (Bax, Bcl‐2, caspase‐3, Cyt C) expression were also increased after exposure to DBDPE 12.5 and 25 mg/L. Such results suggest the neurotoxicity of DBDPE and its ability to induce cell apoptosis by initiating key cellular events such as mitochondrial dysfunction, ROS production and DNA damage through caspase‐dependent and ‐independent pathways.
Sun et al. ( 2023 ). Evaluation and Mechanistic Study of Transgenerational Neurotoxicity in Zebrafish upon Life Cycle Exposure to Decabromodiphenyl Ethane (DBDPE).	F0 zebrafish embryos were exposed to DBDPE (0, 0.1, 1 and 10 nM) from fertilisation (2 h hpf) until 8 months old. Then, sexually matured F0 adults were paired to generate F1 embryos that were cultured until 8 months old. F2 embryos were obtained by pairing F1 adults and cultured until they were 8 months old. Contrary to F0, F1 and F2 generations were reared in clean water. Growth and behavioural performances (shoaling behaviour and social preference test) were studied in male and female adults of each generation (n = 8/sex/treatment/generation). Protein and neurotransmitter in eggs and brains from adult males and females in all treatment groups and generations were also measured. DNA damage (8‐OHdG), global DNA methylation (5‐mC) and transcriptomic sequencing were determined in male and female adult brains in all treatment groups and generations. Cell cycle and apoptosis were assessed in the brain of female zebrafish in all treatment groups and generations. The results showed that the effects on growth and shoaling behaviour of life cycle exposure to DBDPE were impaired in the female offspring whereas they remained quite unaffected in the male offspring. Sociability and recognition with the conspecifics were also impaired in DBDPE exposed F0 generation and the unexposed offspring (F1 and F2). The protein contents in the brain from both male and female adult brains were decreased in F0 generation, female F1 eggs and female F1 and F2 brains whereas the male eggs and adult brains (F1 and F2) were not affected. Specific neurotransmitter contents (Ache, dopamine, epinephrine, glutamate) were changed in F0 male and female brains, remained unaffected in F1 animals, and increased in F2 male and female fish. In addition, the disruption of neurotransmitters in lower dose groups in F2 females was stronger than that in the higher dose groups. In addition, the tested neurotransmitters were also altered in F1 and F2 eggs, suggesting a specific disturbance of the parental transfer of neurotransmitters to zygotes. A lack of changes in the F1 generation and opposite changing trends in the F0 and F2 generations were observed regarding the results of 8‐OHdG and 5‐mC brain levels and specific gene expression. In females, exposure of DBDPE led to cell cycle regulation disorders in the brain of all 3 generations, whereas apoptosis was observed in only F2 offspring from F0 animals exposed to DBDPE 10 nM. Taken together, such results indicated the ability of DBDPE exposure at environmentally relevant doses to induce transgenerational neurotoxicity in zebrafish.
Yang, Zhu, et al. (2023). Mitochondrial Dysfunction Was Involved in Decabromodiphenyl Ethane‐Induced Glucolipid Metabolism Disorders and Neurotoxicity in Zebrafish Larvae.	Zebrafish embryos were exposed to DBDPE (0, 50, 100, 200, 400 μg/L) from 2 h post‐fertilisation (hpf) to 120 hpf. Motor activity, whole‐body neurotransmitters and gene transcription measurements were done after 120 h of exposure. Mitochondrial respiration and oxidative stress measurements were also performed at the end of the period of exposure. In addition, a nontargeted metabolomic analysis was performed in control and 400 μg/L‐exposed larvae after 120 hpf. DBDPE 400 μg/L significantly increased in zebrafish larvae the malformation and heart beat rates, the locomotor activity during the dark–light transition period, and the whole‐body contents of various neurotransmitters like glutamate, epinephrine and norepinephrine, and acetylcholine whereas the content of GABA and the serotonin metabolite 5‐HIAA was significantly reduced. The untargeted metabolomic results showed different changes of metabolites upon DBDPE exposure, involved in glycolipid metabolism, oxidative phosphorylation and oxidative stress, and suggest a potential mechanism related with mitochondrial dysfunction for DBDPE toxicity. Exploration of mitochondria upon DBDPE exposure revealed inhibition in mitochondrial oxidative respiration, and reductions in respiratory chain complex activities, ATP content and mitochondrial membrane potential. All these results suggest the mitochondrial dysfunction is involved in DBDPE‐induced brain toxicity in zebrafish larvae.
Li et al. ( 2024 ). Transcriptome‐based approach to identify mechanisms underlying locomotor abnormality induced by decabromodiphenyl ethane in zebrafish larvae.	Zebrafish larvae were exposed to DBDPE (0, 3, 30, 300 μg/L) over the 5 days post‐fertilisation. Larvae were monitored for their locomotor activity at 120 hpf and transcriptomic sequencing performed at the same stage. Specific assessment of heart rate was performed in fishes of the 0, 30 and 300 μg/L ‐ exposed groups with adding 0.5 mg/L salbutamol. Results showed dose‐related increases in locomotor activity which was significant at the highest dose of DBDPE. Transcriptomic analyses indicated that pathways associated with cardiac muscle contraction and retinol metabolism. Locomotor hyperactivity was shown to be related to increasing heart rate linked to inhibition of Na+/K+ – ATPase activity whereas no histological alterations were identified. Larvae eye histology and retinoid contents remained unchanged in DBDPE‐exposed animals. Such results suggested the developmental toxicity of DBDPE in a zebrafish model of exposure.
Xie et al. ( 2025 ). Decabromodiphenyl ethane exposure‐mediated mitochondrial dysfunction drives oxeiptosis in placental trophoblasts and induces fetal growth restriction	Exposure of ICR mice to DBDPE during gestation caused placental damage, resulting in fetal growth restriction (FGR). DBDPE exposure induces oxeiptosis in placental trophoblasts via activating KEAP1/PGAM5/AIFM1 signalling. A significant reduction in the phosphorylation level of AIFM1 Ser116 in placental trophoblasts was observed, specifically correlating with the activation of oxeiptosis. Further investigations were performed in vitro in HTR8/SVneo cells. Metabolomic and transcriptomic analyses suggested that DBDPE exposure disrupts the oxidative phosphorylation (OXPHOS) pathway, thereby impairing mitochondrial function. Treatment with MitoQ, a mitochondria‐targeted antioxidant, effectively reversed DBDPE‐induced oxeiptosis in placental trophoblasts, alleviating the negative effects of DBDPE on placental damage and FGR. Mechanistically, the mitochondrial dysfunction induced by gestational DBDPE exposure initiates oxeiptosis in placental trophoblasts, exacerbating placental injury and ultimately leading to FGR.
Li et al. ( 2025 ). Brominated Flame Retardants Exposure Promotes the Formation of Atherosclerotic Plaque in Aortic Roots of ApoE–/–Mice by Increasing Macrophage Recruitment and Foam Cell Formation.	The authors investigate the effects of DBDPE on the development of atherosclerosis and explored the underlying mechanisms using an in vitro cell model and ApoE−/− mice. DBDPE (10 μmol/L, 24 h) significantly inhibited the viability of human aortic endothelial cells (HAECs) and induced the generation of ROS. Macrophages are the main immune cells involved in the formation of atherosclerosis, and the phagocytosis of Ox‐LDL by macrophages to form foam cells plays a very important role in the occurrence and development of atherosclerosis. DBDPE significantly enhanced the content of Ox‐LDL in THP‐1 macrophages, which promoted the formation of foam cells. In the in vivo study, DBDPE exposure significantly increased the plaque area and lipid content in the aortic root of mice. It also significantly increased the ROS level in plaques and promoted the expression level of adhesion molecule ICAM‐1, which enhanced the recruitment of macrophages. Transcriptome analysis showed that differentially expressed genes (DEGs) were significantly enriched in signalling pathways related to oxidative stress and lipid metabolism.
Qi et al. ( 2025 ). The single and combined effects of decabromodiphenyl ethane and mixed microplastics on male mice reproductive toxicity.	To investigate the mechanisms of reproductive toxicity of DBDPE, experiments using the mouse premeiotic germ cell line GC‐2 were performed. DBDPE exposure can lead to changes in thousands of differentially expressed genes (DGEs), in which the cell cycle, p53 signalling pathway, apoptosis, FoxO signalling pathway and TNF signalling pathway change significantly. The results also suggest that oxidative stress (increase of ROS level, increase of cellular MDA content and decrease of SOD level after 24 h exposure to 16 mg/L DBDPE) and apoptosis maybe potential target pathways of DBDPE induce reproductive toxicity. Protein expression levels of Nrf2 and Keap1 in mice testis were also examined by western blot analysis. Male mice were exposed by gavage to 0, 5 or 50 mg DBDPE/kg bw per day for 7 weeks. The protein expression levels of Nrf2 decreased with increasing doses of DBDPE. Conversely, Keap1 protein levels were increased dose‐dependently with DBDPE. The genes related to the p53 apoptosis pathway in mouse testes were investigated. The mRNA levels of p53 and Bax were both increased significantly following DBDPE exposure. The trend of Bcl‐2 was opposite to that of p53. The results showed that the toxicity of DBDPE is associated with oxidative stress and apoptosis. Results of transcriptomics from GC‐2 cells indicated that DBDPE lead to changes in oxidative stress, apoptosis, inflammation, proliferation and differentiation. Furthermore, DBDPE exposure inhibited the Kelch‐like ECH‐associated protein1‐NF‐E2‐related factor2 (Keap1‐Nrf2) signalling pathway.
Cui et al. ( 2025 ). ARMMs modulate Notch signaling pathway in DBDPE‐induced liver lipid metabolism disorders: An in vivo and in vitro study.	The impact of DBDPE on the liver in rats and L02 cells (immortalised hepatocyte line) was investigated. Sprague–Dawley rats were exposed by gavage to 0, 5, 50 or 500 mg DBDPE/kg bw per day for 8 weeks. L02 cells were exposed for 24 h to 0, 6.25, 12.5, 25 and 50 μM DBDPE. Low‐dose DBDPE exposure slightly promoted L02 cell proliferation. However, high‐dose DBDPE exposure significantly inhibited L02 cell proliferation, suggesting that DBDPE is also cytotoxic. It was shown that DBDPE exposure increased the number of lipid droplets in L02 cells through Nile Red staining and colorimetric assays, with significant elevations in TC, TG, HDL and LDL levels. The ultrastructural changes in L02 cells also supported this finding. TEM results showed that, compared with the control group, L02 cells exposed to DBDPE exhibited numerous lipid vacuoles. DBDPE exposure could induce lipid metabolism disorders, up‐regulate the expressions of Notch pathway genes and promote Arrestin domain containing protein 1‐ mediated microvesicles (ARMMs)‐derivation. Subsequently, siRNAs were used to inhibit the Notch pathway and the ARMMs‐derivation in L02 cells respectively. Inhibition of Notch or ARMMs‐derivation could reverse the lipid accumulation induced by DBDPE, and the inhibition of ARMMs‐derivation could suppress the activation of the Notch pathway. Extracellular vesicles from L02 cells exposed to DBDPE co‐culture with L02 cells confirmed that ARMMs can directly activate the Notch signalling pathway and induce lipid accumulation.

Abbreviations: AchE, acetylcholine esterase; ALT, alanine aminotransferase; AST, aspartate aminotransferase; CAR, constitutive androstane receptor; ChAT, choline acetyltransferase; DIO, thyroid hormone deiodinase; FGR, fetal growth restriction; GSH, glutathione; LDH, lactate dehydrogenase; MDA, malondialdehyde; NAC, N‐acetyl‐l‐cysteine; PARP1, poly(ADP‐ribose) polymerase‐1; PXR, pregnane X receptor; ROS, reactive oxygen species; SOD, superoxide dismutase; SULT, thyroid hormone sulfotransferase.

Table G.3.

Mode of action studies on BTBPE

Reference	Summary of main conclusions
Ezechiáš et al. ( 2012 ). Hormonal activities of new brominated flame retardants.	Reporter gene assays were used to assess the ability of BTBPE at concentrations up to 14.5 μM to activate or inhibit the transcriptional activity of human ER and AR. At a concentration of 14.5 μM BTBPE, there was no oestrogenic or androgenic activity. BTBPE (12.5 μM and higher) showed anti‐oestrogenic activity.
Smythe et al. ( 2017 ). Impacts of Unregulated Novel Brominated Flame Retardants on Human Liver Thyroid Deiodination and Sulfotransferation.	The inhibitory effects of BTBPE on DIO and SULT activity were investigated using human in vitro liver microsomal and cytosolic bioassays (concentrations from 3.64 to 364 nM). No significant inhibitory effects were observed for BTBPE at any of the concentrations tested. The authors suggest that the lack of activity is likely due to BTBPE's chemical structure. Unlike hydroxylated polybrominated diphenyl ethers (OH‐PBDEs), BTBPE lacks a hydroxyl group, which previous research has shown to be critical for binding to DIO and SULT active sites.
Bajard et al. ( 2019 ). Prioritization of hazards of novel flame retardants using the mechanistic toxicology information from ToxCast and Adverse Outcome Pathways.	In vivo toxicological data were collected on 62 flame retardants to establish five prioritisation categories based on data availability and toxicological concern. The authors classified BTBPE as having ‘scarce toxicological data indicating lower toxicological concern’.
Shockley et al. (2020b). Transcriptomic data from the rat liver after 5 days of exposure to legacy or emerging brominated flame retardants.	Harlan Sprague–Dawley rats (7 weeks old): 0, 0.07, 0.69, 6.88, 68, 680 mg/kg bw per day for 5 days. There were no dose‐related changes in gene expression in the liver.
Bajard et al. ( 2021 ). Endocrine disrupting potential of replacement flame retardants‐Review of current knowledge for nuclear receptors associated with reproductive outcomes.	The authors reviewed the in vitro activity of 52 BFRs towards five nuclear receptors (oestrogen, androgen, glucocorticoid, progesterone and Ah receptors) associated with reproductive outcomes, retrieving data from open literature and ToxCast/Tox21 programmes. No data was retrieved for BTBPE. The opensource platform Endocrine Disruptome model predicted BTBPE to have AR antagonistic activity.
Choi et al. ( 2021 ). Investigation on Sex Hormone‐Disruption Effects of Two Novel Brominated Flame Retardants (DBDPE and BTBPE) in Male Zebrafish (Danio rerio) and Two Human Cell Lines (H295R and MVLN).	In male zebrafish, disruption of sex hormone regulation was suggested after exposure to BTBPE, i.e. the increasing trend of E2 levels, E2/11‐ketotestosterone (11‐KT) ratio and oestrogen receptor alpha (erα) and erβ gene expression levels.
Shi, Wang, et al. (2021). HBCD, TBECH, and BTBPE exhibit cytotoxic effects in human vascular endothelial cells by regulating mitochondria function and ROS production.	Human umbilical vein endothelial cells (HUVECs) were used to investigate cellular effects of BTBPE. Cytotoxicity was observed at concentrations of 5 μM and higher. This was accompanied by increased expression of CASP3, CYTC, BAX and BCL2 as well as ROS release and MDA content. These were dose dependent reduced mitochondrial membrane potential and ATP production as well as decreased GSH content and SOD activity. Cell migration and tube formation were inhibited at concentrations of 10 μM and above.
Cui, Xiao, Zhang, et al. (2024). 1,2‐bis(2,4,6‐tribromophenoxy) ethane, a novel brominated flame retardant, disrupts intestinal barrier function via the IRX3/NOS2 axis in rat small intestine.	BTBPE exposure significantly increased IRX3 and NOS2 expression in the jejunum and ileum, promoting inflammatory gene activation and necroptosis through RIP1, RIP3 and MLKL, including their phosphorylated forms. Immunohistochemistry confirmed increased NOS2 signals. BTBPE also reduced key tight junction proteins (ZO1, Occludin, Claudin1), weakening intestinal barrier integrity and causing mucus layer loss. These effects suggest compromised gut protection against harmful agents. Inhibition of IRX3 or NOS2 using siRNA or pharmacological inhibitors effectively reduced necroptosis and restored tight junctions, indicating IRX3/NOS2 activation as a major pathway in BTBPE‐induced intestinal damage.
Zheng et al. ( 2024 ). Mechanistic Insights into 1,2‐bis(2,4,6‐tribromophenoxy)ethane‐Induced Male Reproductive Toxicity in Zebrafish.	Adult male zebrafish were exposed to BTBPE via the water (0, 0.01, 0.1, 1 and 10 μg/L) for 28 days. BTBPE affected mating behaviour and reduced sperm quality and reproductive success. Proteomic and phosphoproteomic analysis revealed a predominant impact on mitochondrial functions and a widespread effect across cellular compartments on protein phosphorylation level. Ultrastructural damage, increased expression of apoptosis‐inducing factor and disordered respiratory chain confirmed the involvement of mitochondrial impairment in zebrafish testes.

Table G.4.

Studies on the mode of action of HBB

Reference	Main conclusions
Ezechiáš et al. ( 2012 ). Hormonal activities of new brominated flame retardants.	HBB did not affect the transcriptional activity of human oestrogen and androgen receptors in 2 yeast reported‐gene assays.
Usenko et al. ( 2016 ). Evaluation of Common Use Brominated Flame Retardant (BFR) Toxicity Using a Zebrafish Embryo Model.	Exposure of zebrafish embryos to HBB affected markers of oxidative stress but had no effect on spontaneous movement.
Bajard et al. ( 2019 ). Prioritization of hazards of novel flame retardants using the mechanistic toxicology information from ToxCast and Adverse Outcome Pathways.	In vivo toxicological data were collected on 62 flame retardants to establish five prioritisation categories based on data availability and toxicological concern. The authors classified HBB as having ‘scarce toxicological data indicating toxicological concern’.
Bajard et al. ( 2021 ). Endocrine disrupting potential of replacement flame retardants‐Review of current knowledge for nuclear receptors associated with reproductive outcomes.	The authors reviewed the in vitro activity of 52 BFRs towards five nuclear receptors (oestrogen, androgen, glucocorticoid, progesterone and Ah receptors) associated with reproductive outcomes, retrieving data from open literature and ToxCast/Tox21 programmes. The authors claim that there are studies showing HBB has agonistic activity towards the PR and AR, but this is not supported by inspection of the cited references.
Chen et al. ( 2021 ). Bioconcentration and developmental neurotoxicity of novel brominated flame retardants, hexabromobenzene and pentabromobenzene in zebrafish.	Zebrafish embryos were exposed to 0, 30, 100 and 300 μg/L from 2 until 144 h post‐fertilisation. HBB did not cause developmental toxicity (assessed as hatching rates, malformations, grown and survival). Free swimming activity and locomotor behaviour in response to dark‐to‐light stimulation were decreased at the highest concentration. Mode of action studies indicated that HBB binds to and inhibits AChE. Altered expression of genes associated with central nervous system development was also reported.
Sousa et al. ( 2022 ). Brominated flame retardants effect in MCF‐7 cells: Impact on vitamin D pathway.	Exposure to 250 nM HBB for 4 weeks resulted in increased cell proliferation, expression of MMP‐9 and altered the vitamin D pathway in MCF‐7 cells. Exposure for 24 h had no effects.

Abbreviations: AChE, acetylcholinesterase; Ah, aryl hydrocarbon; AR, androgen receptor; BFRs, brominated flame retardants; HBB, hexabromobenzene; MCF‐7, Michigan Cancer Foundation‐7; MMP‐9, matrix metalloproteinase‐9; PR, progesterone receptor.

Table G.5.

Studies on the mode of action of DBE‐DBCH

Reference	Main conclusions
Pradhan et al. ( 2013 ). The brominated flame retardant TBECH activates the zebrafish (Danio rerio) androgen receptor, alters gene transcription and causes developmental disturbances.	Zebrafish eggs were exposed to 1 and 10 uM αβ‐ or γδ‐ DBE‐DBCH. In silico modelling demonstrated binding to the zebrafish androgen receptor, supported by in vitro and in vivo data. DBE‐DBCH was found to delay hatching while morphological abnormalities and juvenile mortality were observed at higher concentrations. qRT‐PCR analysis showed alterations of multiple genes involved in chondrogenesis, metabolism and stress response.
Kharlyngdoh et al. ( 2016 ). TBECH, 1,2‐dibromo‐4‐(1,2 dibromoethyl) cyclohexane, alters androgen receptor regulation in response to mutations associated with prostate cancer.	Effect of AR mutations (AR(W741C) and AR(T877A)) on the transcriptional activation following exposure of cells to the 4 diastereoisomers of DBE‐DBCH individually and as the αβ and γδ. All showed evidence of androgenic activity with no marked differences.
Wong et al. ( 2016 ). The Effects of the Organic Flame‐Retardant 1,2‐Dibromo‐4‐(1,2‐dibromoethyl) Cyclohexane (TBECH) on Androgen Signaling in Human Prostate Cancer Cell Lines.	Androgen receptor target gene expression in the human LNCaP prostate cancer cell line. Both αβ‐ and γδ‐ DBE‐DBCH showed evidence of being partial agonists of the AR.
Curran et al. ( 2017 ). Toxicologic effects of 28‐day dietary exposure to the flame retardant 1,2‐dibromo‐4‐(1,2‐dibromoethyl)‐cyclohexane (TBECH) in F344 rats.	Rats were exposed to 0, 10, 50, 250, 1250 or 5000 mg/kg (0, 0.9, 4.2, 21, 98 and 329 mg/kg bw per day in males and 0, 0.8, 3.9, 19 and 321 mg/kg bw per day in females) technical DBE‐DBCH in the diet for 28 days. Top dose animals were euthanised due to rapid weight loss. There were dose‐dependent clinical and histopathological changes consistent with mild hepatic and renal inflammation and increased serum testosterone in M. The authors identified NOAELs of about 4 mg/kg bw per day and also performed dose–response modelling.
Kharlyngdoh et al. ( 2018 ). Androgen receptor modulation following combination exposure to brominated flame‐retardants.	Various BFRs tested for androgen‐receptor modulation, compared to activity of DBE‐DBCH‐γδ. Not DBE‐DBCH‐αβ. Not clear if relevant.
Bajard et al. ( 2019 ). Prioritization of hazards of novel flame retardants using the mechanistic toxicology information from ToxCast and Adverse Outcome Pathways.	In vivo toxicological data were collected on 62 flame retardants to establish five prioritisation categories based on data availability and toxicological concern. The authors classified DBE‐DBCH as having ‘scarce toxicological data indicating toxicological concern’.
Stojak et al. ( 2019 ). Acute b‐tetrabromoethylcyclohexane (b‐TBECH) treatment inhibits the electrical activity of rat Purkinje neurons.	The effects of acute exposure to β‐DBE‐DBCH on the electrical activity of cultured Purkinje neurons were studied using cell‐attached patch clamp methodology. Cells were obtained from Sprague–Dawley rat embryos removed at the gestational day 20 and were maintained for 14‐21 days in vitro before electrophysiology assays. DBE‐DBCH was tested at 4 levels of concentration (0.01, 0.1, 1 and 10 μM). DBE‐DBCH decreased action potential activity in a dose‐dependent way with an apparent IC50 of 396 nM, suggesting the potential neurotoxicity of this compound.
Wang, Wei, et al. (2019). Evaluation of development, locomotor behavior, oxidative stress, immune responses and apoptosis in developing zebrafish (Danio rerio) exposed to TBECH (tetrabromoethylcyclohexane).	DBE‐DBCH at concentrations of 0.01–10 uM reduced hatchability, diminished heart rate and increased malformations in zebrafish larvae. DBE‐DBCH also decreased free swimming distance and affected measures of oxidative stress and pro‐inflammatory genes. The authors concluded that DBE‐DBCH caused developmental and locomotor behavioral toxicity, immunotoxicity, oxidative stress and proapoptotic effects in early life zebrafish.
Wang, Wei, et al. (2020). Tetrabromoethylcyclohexane (TBECH) exhibits immunotoxicity in murine macrophages.	RAW 264.7 cells were incubated with concentrations of DBE‐DBCH ranging from 0.01 to 100 uM. Cell viability (MTT assay) was significantly decreased at 50–100 uM. Caspase‐3 activity and some measures of oxidative stress were elevated only at the highest concentration, and therefore the relevance cannot be assessed. The mRNA expression of some pro‐inflammatory cytokines, whilst the LPS‐stimulated mRNA expression was downregulated. The authors concluded that DBE‐DBCH is immunotoxic to macrophages.
Bereketoglu et al. ( 2021 ). The brominated flame retardants TBECH and DPTE alter prostate growth, histology and gene expression patterns in the mouse.	Castrated mice were used to test for the androgenic effect of a 1:1 mixture of the diastereoisomers of DBE‐DBCH dosed s.c. at 0 or 1 mg/kg bw per day for 9 days. DBE‐DBCH‐αβ had no effect on body or prostate weights, whereas a mixture of DBE‐DBCH diastereoisomers formed following a fire (DBE‐DBCH‐γδ) increased both of these parameters. DBE‐DBCH‐γδ also resulted in epithelial and glandular alterations in the prostate and changes in androgen‐responsive genes, comparable to the effects of testosterone. The effects of DBE‐DBCH‐αβ were less pronounced.
Bajard et al. ( 2021 ). Endocrine disrupting potential of replacement flame retardants‐Review of current knowledge for nuclear receptors associated with reproductive outcomes.	The authors reviewed the in vitro activity of 52 BFRs towards five nuclear receptors (oestrogen, androgen, glucocorticoid, progesterone and Ah receptors) associated with reproductive outcomes, retrieving data from open literature and ToxCast/Tox21 programmes. For DBE‐DBCH, the authors reported that 5 assays were identified indicating that it is an AR agonist. The authors reported that no assays were identified relating to ER, GR, PR or AHR.

Table G.6.

Studies on the mode of action of PBEB

Reference	Main conclusions
Ezechiáš et al. ( 2012 ). Hormonal activities of new brominated flame retardants.	Yeast reporter gene assays were deployed to determine the potential of PBEB and 10 other compounds to interfere with oestrogenic and androgenic pathways. The assays indicated that none of the test compounds showed oestrogenic or androgenic activities up to the highest concentrations tested, which was 4.1 μM for PBEB.
Lu et al. ( 2020 ). Pentabromoethylbenzene Exposure Induces Transcriptome Aberration and Thyroid Dysfunction: In Vitro, in Silico, and in Vivo Investigations.	5‐week‐old Sprague–Dawley rats (N = 10) were exposed daily by gavage to 0, 0.2, 2 and 20 mg/kg bw per day for 28 days. In F, serum T3 and relative thyroid weight were decreased at all doses. Serum levels of T4 and TSH were increased in both sexes. Morphologically, thyroid follicles became depleted and epithelial cell layers thickened at ≥ 2 mg/kg. PBEB exposure altered the transcriptome profile of thyroid with 17, 42 and 119 differentially expressed genes at the three doses. This included genes of relevance for TH homeostasis with downregulation of Tg, Tpo, Slc5a5, Ttr and Dio1, and upregulation of Tshb. In cell‐based dual‐luciferase assays using African green monkey kidney (CV‐1) cells, PBEB acted as a TRβ antagonist with an IC₅₀ of 9.82 × 10‐7 M, showing no agonist activity between 10‐10 and 10‐6 M. Molecular simulations based on human TRß indicated that PBEB binds within the TRβ ligand‐binding pocket, inducing conformational changes in helices H11 and H12.
Xu et al. ( 2023 ). The mechanism underlying pentabromoethylbenzene‐induced adipogenesis and the obesogenic outcome in both cell and mouse model.	Pregnant CD‐1 mice (N = 6) were orally administered with olive oil (control), 0.2, 2 and 20 mg/kg bw per day PBEB (dissolved in olive oil) until the day prior to delivery. Oral gavage was resumed on PND3 and continued until weaning. The offspring was maintained until 12 weeks of age. Perinatal PBEB exposure to 20 mg/kg bw per day (but not lower doses) caused increased weight gain and epididymal white adipose tissue hypertrophy in male offspring (but not in females). Experiments in mouse 3T3‐L1 cells showed that PBEB selectively activated RXRα rather than PPARγ. Transcriptomic analyses of 3T3‐L1 cells revealed that PBEB disrupted AMPK and PI3K/AKT signalling pathways, which are involved in lipid metabolism and adipocyte maintenance.
Wang et al. ( 2024 ). Perinatal exposure to PBEB aggravates liver injury via macrophage‐derived TWEAK in male adult offspring mice under western diet.	Pregnant CD‐1 mice (N = 6) were orally administered with olive oil (control), 0.2, 2 and 20 mg/kg bw per day PBEB (dissolved in olive oil) until the day prior to delivery. Oral gavage was resumed on PND3 and continued until weaning. The offspring were then fed on a ‘Western diet’. Male, but not female, offspring exposed to PBEB exhibited significant liver injury and inflammation at 20 mg/kg bw per day. Elevated plasma markers such as ALT and AST, alongside histological liver damage, confirmed hepatotoxicity. Mechanistically, PBEB exposure led to the overexpression of tumour necrosis factor‐like weak inducer of apoptosis (TWEAK) in liver macrophages. This cytokine promoted M1 macrophage polarisation and activated the NF‐κB and NLRP3 inflammasome pathways, triggering pyroptosis, an inflammatory form of cell death. Inhibiting TWEAK/Fn14 signalling reduced inflammatory markers and inflammasome activation in vitro.

Abbreviations: ALT, alanine aminotransferase; AST, aspartate aminotransferase; PPARγ, peroxisome proliferator‐activated receptor γ; RXRα, retinoid X receptor α; T3, triiodothyronine; T4, thyroxine; TRβ, thyroid receptor β; TSH, thyroid‐stimulating hormone; TWEAK, tumour necrosis factor‐like weak inducer of apoptosis.

Table G.7.

Studies on the mode of action of PBT

Reference	Main conclusions
Bajard et al. ( 2019 ). Prioritization of hazards of novel flame retardants using the mechanistic toxicology information from ToxCast and Adverse Outcome Pathways.	In vivo toxicological data were collected on 62 flame retardants to establish five prioritisation categories based on data availability and toxicological concern. The authors classified PBT as having ‘scarce toxicological data indicating toxicological concern’.
Bajard et al. ( 2021 ). Endocrine disrupting potential of replacement flame retardants‐Review of current knowledge for nuclear receptors associated with reproductive outcomes.	The authors reviewed the in vitro activity of 52 BFRs towards five nuclear receptors (oestrogen, androgen, glucocorticoid, progesterone and Ah receptors) associated with reproductive outcomes, retrieving data from open literature and ToxCast/Tox21 programmes. No data was retrieved for PBT. The opensource platform Endocrine Disruptome model predicted medium probability of binding to the antagonist conformation of the AR.
Sousa et al. ( 2022 ). Brominated flame retardants effect in MCF‐7 cells: Impact on vitamin D pathway.	MCF‐7 cells were treated with PBT using short and long‐term (4 weeks) exposure protocols. Cell viability was not altered, however cell proliferation was reduced after 24 h or 4 weeks exposure. Changes in cell protein content can be indicative of cellular proliferation but also of other cellular events such as ribosomal impairment, nucleotide availability or DNA damage, a consequence of the prolonged exposure to BFRs. The authors reported that these findings were corroborated with the gene expression of MMP‐9 and the injury assay results. The authors reported that MMP‐9 is involved in the inflammatory response, angiogenesis, wound healing and cellular differentiation. An elevated expression of MMP‐9 has been associated with breast cancer development, angiogenesis, metastasis and invasion. PBT significantly reduced the expression of MMP‐9. Cell migration was diminished especially in the first 6 h after injury. When MCF‐7 cells were exposed to BFRs, cell cycle was arrested in phase G2/M, possibly due to DNA damage.
Maia et al. ( 2022 ). Impact of brominated flame retardants on lipid metabolism: An in vitro approach.	Pre‐dipocyte (3T3‐L1) cell line and a hepatocyte (HepG2) cell line were used to investigate the possible lipid metabolism disruption caused by PBT. PBT do not significantly affect preadipocytes proliferation, but increased lipid accumulation in hepatocytes and also display a significant increase of PPARγ gene expression. Lipid accumulation in the cells can occur by diverse mechanisms. Cells treated with PBT showed an increase in IL‐1β and a decrease in PGC‐1α gene expression, alongside with a tendency to a decrease in adiponectin gene expression. Moreover, PPARγ gene expression was increased after PBT treatment along with an increase in lipid accumulation.

Abbreviation: PPARγ, peroxisome proliferator‐activated receptor γ.

Table G.8.

Studies on the mode of action of TDBP‐TAZTO

References	Main conclusions
Dong et al. ( 2015 ). Tris‐(2,3‐dibromopropyl) isocyanurate induces depression‐like behaviors and neurotoxicity by oxidative damage and cell apoptosis in vitro and in vivo.	Adult mice daily exposed to TDBP‐TAZTO for 30 days, 2 doses (5 and 50 mg/kg/day, gavage), and studied for activity (OF) and depression (tailed suspension test, forced swim test). Neural mechanism were studied in mice brain and in vitro (Sh‐SY5Y cells) for apoptosis and oxidative stress. Overexpression of proapoptotic proteins and upregulation of oxidative stress markers were observed in hippocampus at the highest dose of 50 mg/kg bw per day. In vitro, TDBP‐TAZTO (0, 12.5, 25 and 50 μM, 48 h of exposure) damaged SH‐SY5Y cells and induced cell apoptosis and oxidative stress in a dose‐dependent manner
Ye et al. ( 2015 ). Tris‐(2,3‐Dibromopropyl) Isocyanurate, a New Emerging Pollutant, Impairs Cognition and Provokes Depression‐Like Behaviors in Adult Rats.	Adult male rats daily exposed to TDBP‐TAZTO for 6 months, 2 doses (5 and 50 mg/kg per day, gavage). Animals were tested for activity (OF), depression (forced swim test) and spatial learning and memory (Morris water maze). Oxidative stress, apoptosis, synaptic plasticity and inflammation were explored in rat hippocampus. ACTH and corticosterone were measured to assess stress level of the animals. Dose‐dependent increases in hippocampal pro‐apoptosis proteins and pro‐inflammatory cytokine levels and reductions in synaptic plasticity markers (BDNF and synaptophysine) and GSH and SOD antioxidant actors were observed. Histological and immunohistochemical results indicated that TDBP‐TAZTO caused significant increases in the number of apoptotic cells and GFAP‐positive cells, and lead to marked neuronal damage in hippocampal area. These results demonstrated that 6‐month TDBP‐TAZTO exposure impaired hippocampal neurons relatively to the hippocampal‐related behavioural disturbances reported in the same study.
Krivoshiev et al. ( 2016 ). Assessing in‐vitro estrogenic effects of currently‐used flame retardants. Toxicology in Vitro 33, 153‐162	TDBP‐TAZTO was investigated for its potential endocrine‐disruptive activity using the MCF‐7 flow cytometric proliferative assay. MCF‐7 human breast adenocarcinoma cells were exposed to TDBP‐TAZTO 100 μM for 24 h and assessed for cytotoxicity and proliferation. The results showed the significant anti‐estrogenic activity of TDBP‐TAZTO when co‐exposed with 17β‐estradiol that remained limited compared to control and to the effects induced by other BFRs tested.
Li et al. ( 2016 ). Effects of tris(2,3‐dibromopropyl) isocyanurate on steroidogenesis in H295R cells.	The effects of TDBP‐TAZTO (0.1‐100 μM) on steroidogenesis and adrenocortical function were studied in vitro using the H295R cell line. The effects of co‐exposure with the cAMP inducer forskolin 10 μM were also studied. Estradiol and testosterone production and steroidogenesis‐related gene expression were measured after 72 h of exposure. The results showed the absence of cytotoxicity and the significant reduction in the hormone production level only at the dose of 100 μM. Adding forskolin with TDBP‐TAZTO strongly increased the production of both hormones compared to control. TDBP‐TAZTO alone did not change the level of expression of all genes studied whereas it blocked forskolin‐induced mRNA expression of specific adrenal steroidogenic enzymes (StAR, 3bHSD2, CYP17, CYP21, CYP19). The present results suggest the ability of TDBP‐TAZTO to be reprotoxic.
Cao et al. ( 2018 ). Anti‐estrogenic activity of tris(2,3‐dibromopropyl) isocyanurate through disruption of co‐activator recruitment: experimental and computational studies. Archives of Toxicology 92, 1471‐1482	The potential anti‐estrogenic activity of TDBP‐TAZTO has been studied using the MCF‐7/BUS cell proliferation assay (E‐Screen bioassay) and the MVLN cells/ERα luciferase reporter gene assay (MVLN assay). TDBP‐TAZTO significantly inhibited 17β‐estradiol‐induced cell proliferation in the E‐Screen assay in a dose range between 10‐9 and 10‐7 M. The luciferase activity induced by 17β‐estradiol measured in the MVLN assay was also significantly inhibited by TDBP‐TAZTO in a dose‐dependent manner. Moreover, neither TDBP‐TAZTO nor 17β‐estradiol affected proliferation of the ERα‐negative breast cancer cell line MDA‐MB‐231. These experimental results confirmed that TDBP‐TAZTO has anti‐estrogenic effects by affecting the ERα‐mediated signalling pathway.
Bajard et al. ( 2019 ). Prioritization of hazards of novel flame retardants using the mechanistic toxicology information from ToxCast and Adverse Outcome Pathways.	In vivo toxicological data were collected on 62 flame retardants to establish five prioritisation categories based on data availability and toxicological concern. The authors classified TDBP‐TAZTO as having ‘scarce toxicological data indicating toxicological concern’.
Bajard et al. ( 2021 ). Endocrine disrupting potential of replacement flame retardants‐Review of current knowledge for nuclear receptors associated with reproductive outcomes.	The authors reviewed the in vitro activity of 52 BFRs including TDBP‐TAZTO towards five nuclear receptors (oestrogen, androgen, glucocorticoid, progesterone and Ah receptors) associated with reproductive outcomes, retrieving data from open literature and ToxCast/Tox21 programmes. The authors reported that TDBP‐TAZTO was predicted as having an androgenic receptor antagonist activity from 2 QSAR models.
Szychowski et al. ( 2021 ). Tris (2,3‐Dibromopropyl) Isocyanurate (TDBP‐TAZTO or TBC) Shows Different Toxicity Depending on the Degree of Differentiation of the Human Neuroblastoma (SH‐SY5Y) Cell Line.	The effects of TDBP‐TAZTO on neuronal differentiation were tested in vitro using neuroblastoma SH‐SY5Y cell line as a model of neuronal differentiation. Cells were differentiated for 7 and 14 days using retinoic acid. TDBP‐TAZTO was applicated at doses ranging 1–100 nM and 1‐100 μM and the expression and/or activity of CYP1A1, CYP1A2 and CYP2B6 enzymes, and AhR and TUBB3 genes were assessed 24 h after the application. Cells were also studied for cytotoxicity (resaruzin reduction and LDH activity assays), apoptosis (caspase‐3 activity) and reactive oxygen production (H2DCFDA fluorimetric assay). The results showed a decrease in the ability of TDBP‐TAZTO to induce EROD activity during the differentiation process, which could be an effect of cell differentiation into neurons. Changes of the EROD activity after stimulation of AhR antagonist (CAY10464) and agonist (βNF) correlated with variations in AhR and CYP1A1 expression during differentiation whereas no mRNA expression of CYP1A2 and CYP2B6 was detected in all types of SH‐SY5Y cells (undifferentiated, differentiated for 7 or 14 days). TUBB3 gene expression highly increased in cells differentiated for 7 and 14 days, confirming the effective differentiation of SH‐SY5Y cells into neurons and the absence of effects of TDBP‐TAZTO on cell differentiation.
Han et al. ( 2022 ). In vivo and in vitro study on hepatotoxicity of Tris‐(2, 3‐dibromopropyl) isocyanurate exposure via mitochondrial and death receptor pathway.	TDBP‐TAZTO liver toxicity was assessed in vitro using HepG2 cells (Han et al., 2022). Cells were exposed for 72 h to TDBP‐TAZTO (0, 12.5, 25 or 50 μM). The results showed the ability of this compound to induce apoptosis via mitochondrial pathway and death receptor signalling pathway.
Bajard et al. ( 2023 ). Anti‐androgenic activity of novel flame retardants in mixtures: Newly identified contribution from tris(2,3‐dibromopropyl) isocyanurate (TDBP‐TAZTO).	TDBP‐TAZTO has been tested for its androgenic and anti‐androgenic activities alone or in mixture. The interaction of TDBP‐TAZTO with androgenic receptor (AR) was studied employing an in vitro AR‐dependent luciferase reporter gene assay in MDA‐kb2 cells. TDBP‐TAZTO showed a certain degree of cytotoxicity with an IC50 of 62.9 μM. This compound did not activate the Ah receptor in a significant way at concentrations below the IC50 value whereas it was the only one compound tested in the study that exhibited inhibition of dihydrotestosterone‐induced AR receptor activity with an IC50 of 10.8 μM. In addition, TDBP‐TAZTO contributed to anti‐androgenic activity when combined with six other anti‐androgenic FRs and showed a contribution to the overall mixture effect at a lower concentration than its individual effective dose. The present study provided experimental evidence for the anti‐androgenic properties of TDBP‐TAZTO and its endocrine‐disrupting potency.
Szychowski and Skóra ( 2023c ). Involvement of peroxisome proliferator‐activated receptor gamma (PPARa) and autophagic pathways in the mechanism of action of the tris(2,3‐dibromopropyl) isocyanurate (TDBP‐TAZTO or TBC) flame retardant in the lung adenocarcinoma (A549) cells in vitro.	The present study aimed to assess the involvement of the PPARγ receptor and some autophagic proteins (mTOR and p62) in the toxicity of TDBP‐TAZTO in adenocarcinomic human alveolar basal epithelial cell cultures (A549). Cells were exposed to TDBP‐TAZTO (0.001‐100 μM) and assessed for the impact of this xenobiotic on basic cell metabolism parameters (metabolic activity, ROS production, LDH release and caspase‐3 activity) and on the gene and protein expression of CAT, SOD1, PPARγ, mTOR and p62, which are involved in oxidative stress and autophagy processes. The results showed that TDBP‐TAZTO induced cell damages and apoptosis only at the highest micromolar concentrations (10, 50 and 100 μM) whereas TDBP‐TAZTO showed the ability to trigger oxidative stress and affected the mRNA expression of antioxidant enzymes (SOD1 and CAT) at the lower concentrations (1 and 10 μM) than in the case of apoptosis, suggesting that the apoptosis was ROS independent. The pharmacological modulation of PPARγ receptor agonist indicated that TDBP‐TAZTO acts in the A549 cell line probably through activation of the mTOR‐PPARγ pathway and could interfere with the p62 autophagy pathway.
Szychowski and Skóra ( 2023b ). Disruption of neurosteroid synthesis and release by tris(2,3‐dibromopropyl)isocyanurate in primary mouse cortical astrocytes in vitro.	The study aimed to determine the impact of TDBP‐TAZTO on the neurosteroid (progesterone, testosterone and estradiol) production and secretion as well as the mRNA expression of key enzymes involved in its production in a model of primary mouse astrocyte cell culture. Effects of TDBP‐TAZTO (1 and 10 μM) on hormone production, and gene expression of various enzymes involved in the production and metabolism of such hormones (Actb, Gadph, p53, Mki67, Cyp19a1, Cyp17a1 and Hsd17b4) were measured after 48 h of exposure. Results showed a significant dose‐dependent decrease of the expression of the proliferative marker Ki67. In the same time; TDBP‐TAZTO increased progesterone production with a significant decrease in testosterone production at both concentrations, which is accompanied by a decrease in Cyp17a1 mRNA expression, that is, the main enzyme metabolising progesterone to testosterone. Moreover, TDBP‐TZATO increased progesterone secretion in the culture medium at both concentrations. TDBP‐TAZTO also induced an increase in the production of estradiol at the concentration of 1 μM correlated with the higher level of expression of Cyp19a1 mRNA, the enzyme metabolising testosterone to estradiol. The present in vitro study suggests that TDBP‐TAZTO have endocrine‐disrupting properties with potential consequences for brain development and functioning.
Szychowski and Skóra ( 2023a ). Tris(2,3‐dibromopropyl)Isocyanurate has an Effect on Inflammation Markers in Mouse Primary Astrocytes In vitro.	The impact of TDBP‐TAZTO on neuroinflammation and apoptosis was tested in a model of primary mouse astrocyte cell culture. Dose‐related effects of TDBP‐TAZTO (1–100 nM and 1–100 μM) on caspase‐1 and ‐3 activities, DNA cell staining, oxidative stress and inflammation‐related proteins were measured after 6, 24 or 48 h of exposure. TDBP‐TAZTO increased caspase‐1 and caspase‐3 activities at the two highest concentrations of 50 and 100 μM. TDBP‐TAZTO also increased the level of such inflammation markers as Cat, IL‐1β and IL‐1βR1 proteins and decreased the level of the proliferation marker (Ki67). No changes in astrocyte morphology and formation of apoptotic bodies were observed. The results suggest the lack of toxicity of TDBP‐TATZO on astrocytes in vitro at the low levels of concentration used.
Tabecka‐Lonczynska, Kaczka, and Kaleniuk (2023). Involvement of estrogen receptor alpha (ERα) and impairment of steroidogenesis after exposure to tris(2,3‐dibromopropyl) isocyanurate (TBC) in mouse spermatogenic (GC‐1 spg) cells in vitro.	The effects of TDBP‐TAZTO 10 μM (48 h of exposure) on the expression of nuclear receptors involved in the formation of steroid hormones, proteins and enzymes responsible for steroidogenesis and the levels of steroid hormones (estradiol, progesterone and testosterone) were tested in the GC‐1 spg cell line as a mouse model of spermatogenic cells in vitro. Effects of co‐exposure to TDBP‐TAZTO with BHPI 10 μM (Erα antagonist) were also studied. The results indicated the ability of TDBP‐TAZTO to increase the Erα mRNA and protein expression whereas Pparγ, AhR and IGF‐1R gene expression were reduced compared to control cells. In addition, a decrease in the levels of most of the analysed proteins and enzymes involved in steroid conversion was observed. A significant decrease was observed in the production and secretion of progesterone to the culture medium compared to the control. Slight non‐significant decreases in estradiol and testosterone were also observed. Co‐exposure with BHPI modulated the effects of TDBP‐TAZTO alone, suggesting the involvement of Era in the mechanism of TDBP‐TAZTO toxicity. These results demonstrated that TDBP‐TAZTO affects the process of steroidogenesis by reducing the expression of proteins and enzymes of major contribution to the synthesis of gonadal hormones, suggesting the endocrine disrupting potency of this compound.
Tabecka‐Lonczynska, Skóra, et al. (2023). Reprotoxic Effect of Tris(2,3‐Dibromopropyl) Isocyanurate (TBC) on Spermatogenic Cells In Vitro. Molecules 28, xxx.	The effects of TDBP‐TAZTO on metabolic activity (resazurin reduction assay), apoptosis (caspase‐3 activity and protein expression) and signalling key genes (Ki67, p53, Pparγ, Ahr, Esr1 mRNA expression) were evaluated in mouse spermatogonic GC‐1 spg cells. Cells were exposed to increasing concentrations of TDBP‐TAZTO (1, 10, 50, 100 nM and 1, 10, 50, 100 μM) for 48 h. Effects of co‐exposure to TDBP‐TAZTO 10 μM with BHPI 10 μM (Erα antagonist), estradiol 100 nM or letrozole 1 μM (aromatase inhibitor) were also tested. TDBP‐TAZTO (10, 50 or 100 μM) induced significant decreases in resaruzin reduction, and concomitant increases in caspase‐3 activity and expression at levels of 10 and 100 μM. Significant increases in Ki67, p53, Pparγ and Esr1 mRNA expression were observed with TDBP‐TAZTO 10 μM whereas the Ahr expression was decreased compared to controls. An increase in Pparγ mRNA levels and a decrease in Ahr and Esr gene expression were observed in cells co‐exposed to TDBP‐TAZTO and estradiol. These results suggest the significant impact of TDBP‐TAZTO on the dysregulation of the steroid‐related signalling pathway in male reproductive cell line, suggesting the toxicity of this compound on male fertility.
Yang, Li, et al. (2023). Effects of fibrous microplastics on the accumulation of tris(2,3‐dibromopropyl) isocyanurate and behavior of zebrafish via water‐ and foodborne exposure routes.	The present study investigated the uptake and depuration and the behavioural toxicity of TDBP‐TAZTO (5 μg/L) in zebrafish. The effects of co‐exposure with fibrous microplastics were also investigated. Adult female wild‐type zebrafish were exposed to TDBP‐TAZTO (5 μg/L added in artificial freshwater) for 7 days. TDBP‐TAZTO depuration was evaluated after 4 additional days in clean non‐contaminated freshwater. Foodborne exposure to TDBP‐TAZTO was performed through the feeding of zebrafish with clean or contaminated daphnids (10 per fish and day). The accumulation of TDBP‐TAZTO in fish muscle via waterborne exposure increased over time to reach a maximum and comparable concentration at 3 and 7 days of exposure. TDBP‐TAZTO content in muscle was significantly reduced by 60% after the period of depuration. The TDBP‐TAZTO accumulation in zebrafish muscle after 7 days of foodborne exposure was quite low, two magnitude orders lower than waterborne exposure. TDBP‐TAZTO did induce significant behavioural changes related to swimming trajectories, distance and cumulative duration of manic, active and static states of zebrafish whatever the way of exposure (waterborne and foodborne exposure).
Tabecka‐Lonczynska et al. ( 2024 ). Unraveling the anti‐androgenic mechanism of tris(2,3‐dibromopropyl) isocyanurate (TBC) via the non‐classical testosterone pathway and steroidogenesis: Potential human reproductive health implications.	The effects of TDBP‐TAZTO 10 μM alone or combined with testosterone 10 μM or flutamide 1 μM (anti‐androgen compound) were evaluated in mouse‐derived spermatogonic GC‐1 spg cells after 48 h of exposure. Cells were assessed for their metabolic activity (resaruzin reduction and LDH activity assays), apoptosis (caspase‐3 activity) and a large set of protein expression, involved for a part in non‐classical testosterone pathway and steroidogenesis. TDBP‐TAZTO induced a reduction in metabolic activity and an increase in LDH release, indicating cell damages, while caspase‐3 activity was slightly reduced. TDBP‐TAZTO exposure led to a reduction in progesterone synthesis alone or combined with testosterone or flutamide whereas estradiol synthesis was decreased with TDBP‐TAZTO alone only. Testosterone production remained unaffected in all conditions of exposure. TDBP‐TAZTO exposure downregulated SRC‐EGFR‐AKT‐p33MAPK protein production affecting non‐classical androgen signalling. In addition, TDBP‐TAZTO reduced StaR and HSD3b1 levels suggesting impaired steroidogenesis. TDBP‐TAZTO was able to initiate autophagy through the significant increase in the Beclin1 production level. These results indicate the potential of TDBP‐TAZTO to disrupt steroid hormone synthesis and androgen signalling pathways, to modulate autophagy processes and to affect cell proliferation and survival. Taken together, these results confirm the potential reproductive toxicity of TDBP‐TAZTO.

Table G.9.

Studies on the mode of action of TBCO

References	Main conclusions
Saunders et al. ( 2013 ). In vitro endocrine disruption and TCDD‐like effects of three novel brominated flame retardants: TBPH, TBB, & TBCO.	TBCO produced: (i) No TCDD‐like effects in the H4IIE assay. (ii) No agonistic effects in the YES/YAS assays. (iii) A maximal anti‐androgenic effect of 59% at 1500 mg L‐1 in the YAS assay. (iv) A 3.3‐fold increase in concentrations of E2 15 mg L‐1 exposure in the H295R assay.
Mankidy et al. ( 2014 ). Effects of novel brominated flame retardants on steroidogenesis in primary porcine testicular cells.	The effect of TBCO on sex steroid production were investigated in porcine primary testicular cells. The concentration of T and E2 in the conditioned media was elevated at exposures to 0.03 mg and 3.0 mg TBCO/L. Expression of Cyp17a and 21a were upregulated at the higher dose.
Saunders et al. ( 2015 ). Effects of the brominated flame retardant TBCO on fecundity and profiles of transcripts of the HPGL‐axis in Japanese medaka.	Medaka were fed diets containing 58 or 607 μg TBCO/g food (wet mass). Cumulative fecundity decreased 18% in fish exposed to 58 μg/g, while 607 μg/g showed no significant difference from controls. Organ‐specific, dose‐dependent transcript alterations of genes involved in the hypothalamus‐pituitary‐gonadal‐liver axis occurred in both sexes. At 607 μg/g, downregulation of genes involved in steroidogenesis, cholesterol metabolism and regulatory feedback was observed in gonads of both sexes, though these changes did not affect fertilisation or fecundity. At 58 μg/g, livers from both sexes showed increased expression of oestrogen‐responsive genes including vitellogenin II, choriogenin H and ERα.
Sun et al. ( 2016 ). Combined Transcriptomic and Proteomic Approach to Identify Toxicity Pathways in Early Life Stages of Japanese Medaka (Oryzias latipes) Exposed to 1,2,5,6‐Tetrabromocyclooctane (TBCO).	Japanese medaka (Oryzias latipes) embryos were exposed to 10, 100 or 1000 μg/L TBCO from 2 h post‐fertilisation until 1 day post‐hatch. TBCO accumulated 0.43‐1.3 × 104‐fold in embryos with accumulation rates of 1.7–1.8 per day. Exposure to 100 and 1000 μg/L significantly impaired hatching time and success (p < 0.001). Transcriptomic and proteomic analyses of embryos exposed to 100 μg/L revealed enrichment in pathways related to embryo development, hatching, visual perception and cardiac function. Targeted bioassays confirmed these predictions: visual performance was significantly reduced at 10 μg/L (p = 0.013) and 100 μg/L (p = 0.028), while heart rates decreased significantly at 100 μg/L (p = 0.014) and 1000 μg/L (p < 0.001).
Shi et al. ( 2020 ). Cytotoxicity of hexabromocyclododecane, 1,2‐dibromo‐4‐(1,2‐dibromoethyl) cyclohexane and 1,2,5,6‐tetrabromocyclooctane in human SH‐SY5Y neuroblastoma cells.	Cytotoxicity of TBCO was studied in human SH‐SY5Y neuroblastoma cells. TBCO reduced cell viability at concentrations of 0.5 μM and above. This was associated with increased cell membrane permeability and apoptosis. TBCO caused increased [Ca2+]i and at high concentrations elevated ROS levels.
Bajard et al. ( 2021 ). Endocrine disrupting potential of replacement flame retardants‐Review of current knowledge for nuclear receptors associated with reproductive outcomes.	The authors reviewed the in vitro activity of 52 BFRs towards five nuclear receptors (oestrogen, androgen, glucocorticoid, progesterone and Ah receptors) associated with reproductive outcomes, retrieving data from open literature and ToxCast/Tox21 programmes. The authors reported that there is evidence from literature and ToxCast database on anti‐oestrogenic activity of TBCO.
Van Essen, Devoy, et al. (2021). Effects of the brominated flame retardant, TBCO, on development of zebrafish (Danio rerio) embryos.	Zebrafish larvae were exposed to TBCO at 2, 20 or 200 μg/L from 1 to 120 hpf. Mortality was dose‐dependently increased, the incidences of spine curvature and uninflated swimbladder increased with significance at 20 μg/L while heart rate was dose‐dependently decreased at all concentrations. Effects of TBCO on swimbladder formation was investigated at 10 and 100 μg/L. There was no effect on cyp1a1 mRNA expression.
Van Essen, Alcaraz, et al. (2021). The brominated flame retardant, TBCO, impairs oocyte maturation in zebrafish (Danio rerio).	Sexually mature female zebrafish were given a diet of 35.3 or 628.8 μg TBCO/g food for 14 days after which, stage IV oocytes were isolated to assess maturation in response to 17α,20β‐dihydroxy‐4‐pregnen‐3‐one, aka maturation inducing hormone (MIH). Both doses reduced oocyte maturation. Transcriptomics implicated changes in expression of insulin‐like growth factor signalling and expression of microRNAs regulating this pathway in the reduction of oocyte maturation.
Devoy, Raza, Kleiner, et al. (2023). The brominated flame retardant, 1,2,5,6‐tetrabromocyclooctane (TBCO), causes multigenerational effects on reproductive capacity of Japanese medaka (Oryzias latipes).	TBCO at low (40.6 μg/g) and high (1034.4 μg/g) doses reduced F1 fecundity by 33.9% and 33.3%, respectively. High‐dose effects persisted in F2 (29.8% reduction), but F3 recovered completely. TBCO caused multigenerational, not transgenerational, reproductive effects. Abundances of mRNA for cholesterol side chain cleavage enzyme (cyp11a), aromatase (cyp19a) and luteinizing hormone receptor (lhr) were differentially expressed in gonads from F1 females.
Devoy, Raza, Jones, et al. (2023). Japanese medaka (Oryzias latipes) exposed via maternal transfer to the brominated flame retardant, 1,2,5,6‐tetrabromocyclooctane (TBCO), experience decreased fecundity and impaired oocyte maturation.	F0 medaka fed low (74.7 μg/g) or high (663 μg/g) TBCO diets produced F1 females with 26.0% and 56.8% reduced fecundity, respectively. Both doses impaired 17α,20β‐dihydroxy‐4‐pregnen‐3‐one, aka maturation inducing hormone (MIH) stimulated oocyte maturation (23.4% and 20.0% reduction). Vitellogenin transcript abundance decreased dose‐dependently, while E2 concentrations remained unchanged.
Raza et al. ( 2023 ). Impairment of oocyte maturation as a mechanism of decreased fecundity in Japanese medaka (Oryzias latipes) exposed to the brominated flame retardant, 1,2,5,6‐tetrabromocyclooctane (TBCO).	Stage IX oocytes exposed in vitro to 20 and 200 μg/L TBCO showed significantly reduced MIH‐stimulated maturation, while 0 and 2 μg/L had no effect. For in vivo assessment, sexually mature fish were fed control, 100 or 1000 μg/g TBCO‐spiked food for 21 days. Both TBCO concentrations significantly decreased fecundity and oocyte maturation. Plasma 17β‐estradiol (E2) levels and hepatic vitellogenin (vtgI and vtgII) transcript abundances showed no significant differences between treatments.

ANNEX A. Protocol for the human risk assessments related to the presence of brominated flame retardants (BFRs) in food

The Annex is provided as a separate pdf file containing the risk assessment protocol applied by the CONTAM Panel to update the previous risk assessments of brominated flame retardants (BFRs) in food.

ANNEX B. Computational analysis for the assessment of emerging and novel BFRs

The Annex is provided as a separate pdf file and xls file containing the details and outcome of the similarity matrix, profilers, information sources, QSAR predictions and read across exercise. The files are available on EFSA's Knowledge Junction Community on Zenodo at: https://doi.org/10.5281/zenodo.21026968.

ANNEX C. Occurrence data on emerging and novel BFRs in food submitted to EFSA, dietary surveys per country and age group, and detailed results of the human chronic dietary exposure assessment

The Annex is provided as a separate excel file and contains the summary statistics of the occurrence data on emerging and novel BFRs in food submitted to EFSA, and data about the dietary surveys per country and age group available in the EFSA Comprehensive Database used in the dietary exposure assessment, as well as the detailed results of the human chronic dietary exposure assessment to emerging and novel BFRs. The file is available on EFSA's Knowledge Junction Community on Zenodo at: https://doi.org/10.5281/zenodo.21027469.

ANNEX D. Benchmark dose (BMD) modelling

The Annex is provided as a separate pdf file containing the results of the benchmark dose (BMD) analysis.

ANNEX E. Raw occurrence data on emerging and novel bfrs in food

The raw occurrence data on the emerging and novel BFRs in food extracted from EFSA Data Warehouse by 8 September 2025 and sampled between years 2010 and 2025 are available on EFSA's Knowledge Junction Community on Zenodo at: https://doi.org/10.5281/zenodo.21027554.

ANNEX F. Outcome of the targeted consultation on the draft scientific Opinion on the update of the risk assessment on emerging and novel BFRs in food

The Annex is provided as a separate pdf file containing the comments received during the targeted consultation and the replies by the CONTAM Panel, and this is available under the Supporting Information section on the online version of the Scientific output.

EFSA CONTAM Panel (EFSA Panel on Contaminants in the Food Chain) , Knutsen, H. K. , Åkesson, A. , Bampidis, V. , Bignami, M. , Bodin, L. , Chipman, J. K. , Degen, G. , Hernández‐Jerez, A. , Hofer, T. , Hogstrand, C. , Landi, S. , Leblanc, J.‐C. , Machera, K. , Ntzani, E. , Oswald, I. P. , Rychen, G. , Sand, S. , Vejdovszky, K. , … Hogstrand, C. (2026). Update of the risk assessment on emerging and novel brominated flame retardants in food. EFSA Journal, 24(7), e10187. 10.2903/j.efsa.2026.10187