A prioritization strategy for functional alternatives to bisphenol A in food contact materials

Abstract

The use of bisphenol A (BPA), a substance of very high concern, is proposed to be banned in food contact materials (FCMs) in the European Union. To prevent regrettable substitution of BPA by alternatives with similar or unknown hazardous properties, it is of importance to gain the relevant toxicological information on potential BPA alternative substances and monitor them adequately. We created an inventory of over 300 substances mentioned as potential BPA alternatives in regulatory reports and scientific literature. This study presents a prioritization strategy to identify substances that may be used as an alternative to BPA in FCMs. We prioritized 20 potential BPA alternatives of which 10 are less familiar. We subsequently reviewed the available information on the 10 prioritized less familiar substances regarding hazard profiles and migration potential obtained from scientific literature and in silico screening tools to identify a possible risk of the substances. Major data gaps regarding the hazard profiles of the prioritized substances exist, although the scarce available data give some indications on the possible hazard for some of the substances (like bisphenol TMC, 4,4-dihydroxybenzophenone, and tetrachlorobisphenol A). In addition, very little is known about the actual use and exposure to these substances. More toxicological research and monitoring of these substances in FCMs are, therefore, required to avoid regrettable substitution of BPA in FCM.

Keywords:

• Bisphenol A
• bisphenol A substitutes
• bisphenol A alternatives
• hazard assessment
• migration
• in silico
• regrettable substitution

Introduction

4,4′-(Propane-2,2-diyl) diphenol, commonly known as bisphenol A (BPA), is a high production volume chemical that is primarily used as a key building block of polycarbonate (PC) plastic and main precursor in the synthesis of epoxy resins. Both applications have intended uses for food contact materials (FCMs). PC is used in the manufacturing of reusable beverage bottles, plastic tableware, storage containers, and other FCMs. Epoxy resins are used for food preservation and long-term storage as protective linings on the inside of metal-based food and beverage cans (EFSA 2015, 2023; KIDV 2022).

Human health concerns have been raised regarding the exposure to BPA, as this has been linked to various adverse health effects, including reproductive, developmental, and immunotoxic effects (EPA 2010; EFSA 2023). BPA is also identified as a substance of very high concern (SVHC) according to the REACH regulation (Registration, Evaluation, Authorization, and Restriction of Chemicals), for being toxic to reproduction and an endocrine disruptor for both human health and the environment (ECHA 2023a). The general public may be unintendedly exposed to BPA through diet, due to the leaching of small amounts of BPA from FCMs into foodstuffs or contact with BPA containing consumer products (Cao et al. 2011; Cunha and Fernandes 2013; Agarwal et al. 2022; Wang X et al. 2022). Leaching from PC plastic may be enhanced by various factors, such as contact duration, durability, and heating of the plastic products (De Coensel et al. 2009; Lim et al. 2009; Agarwal et al. 2022). Human exposure to BPA has been confirmed by ample human biomonitoring studies that measured BPA (in the form of glucuronidated BPA) in urine or (in the form of unconjugated BPA) in blood of both children and adults (Vandenberg et al. 2010; Frederiksen et al. 2020; IPCHEM 2023).

The use of BPA is restricted for many consumer products in Europe. Its use as a monomer in plastic FCMs is authorized, with its use subject to a specific migration limit of 0.05 mg/kg in the EU (Regulation (EC) 10/2011 (EU 2020)). In addition, the use of PC infant feeding bottles is forbidden (Regulation (EC) 10/2011 (EU 2020)). The use of thermal paper containing BPA in concentrations of 0.02% by weight or more has been banned in the EU (Regulation (EC) 1907/2006 (EU 2023)). The use of BPA is also restricted in the United States of America (USA), yet not as extensively as in Europe. The U.S. Food and Drug Administration allows the use of BPA-based PC in baby bottles and sippy cups and BPA-based epoxy resins as coatings in packaging for infant formulas (FDA 2012, 2013). A proposal for further restrictions on the use of BPA and other bisphenols in the EU was recently submitted to ECHA by German authorities, but is now withdrawn and being revised following the public consultation.

In addition, the European Food Safety Authority (EFSA) recently reevaluated the health-based guidance value (HBGV) for BPA in a Scientific Opinion as new scientific literature had become available (EFSA 2023). Rather than the temporary tolerable daily intake (t-TDI) level of 4 µg/kg bw/d that was previously derived based on relative kidney weight changes in mice (EFSA 2015), a new tolerable daily intake (TDI) of 0.2 ng/kg bw/d was derived based on effects on the immune system (EFSA 2023). The derivation of this new TDI led to a myriad of comments during the public consultation of the draft opinion (EFSA 2023), which have not been adequately addressed in final opinion in the view of many (BfR 2023; EMA 2023; Zagorski and Kaminski 2023; Prueitt and Goodman 2024). Among other issues, an unclear connection between the selected endpoint and a potential adverse health outcome (apical toxicity) and the lack of consistent low-dose effects of BPA in other studies were mentioned as main drawbacks of the new TDI (BfR 2023; EMA 2023; Zagorski and Kaminski 2023; Prueitt and Goodman 2024). The implementation of EFSA’s newly derived TDI would result in further restrictions on the use of BPA, like its proposed ban in FCM in the EU, currently in preparation by the European Commission (European Commission 2023). The measure will also address the use of other bisphenols in FCMs to avoid replacing BPA with other harmful substances. Which “other bisphenols” are considered by the European Commission has not yet been disclosed.

Although there is considerable global debate about the risks associated with the exposure to BPA (Schug et al. 2013; BfR 2023; EMA 2023; Zagorski and Kaminski 2023; Prueitt and Goodman 2024), the public awareness and current restrictions of BPA resulted in the increased use of and search for BPA alternatives (ECHA 2020b; Catenza et al. 2021; Harnett et al. 2021). The use of BPA alternatives is expected to increase even more following the significant reduction of BPA’s TDI by EFSA. Manufacturers consider the use of BPA alternatives as a drop-in substitute, or the use of alternative materials to PC plastic and epoxy resin coatings, or the use of nonchemical solutions (alternative technologies) (Bakker et al. 2016).

A great number of compounds are named as a drop-in substitute to BPA. These substances are typically highly structurally related to BPA, thereby raising concerns for similar health effects (Harnett et al. 2021). Indeed, significant reservations have been made considering the drop-in replacement of BPA alternatives with regard to regrettable substitution, where a substance is replaced by another substance with similar or unknown hazardous properties (ChemAgenda 2020). It is, however, unclear whether these chemicals are actually used as drop-in substitutes of BPA, and to what extent. The constituents of epoxy resin coatings for example are typically proprietary knowledge and are not disclosed by manufacturers. This makes it difficult to identify the alternatives that pose the highest risk for public health. In 2020, we have reviewed the potential hazard characteristics of BPA drop-in replacements in consumer products (den Braver-Sewradj et al. 2020). However, as food is the main source of exposure to BPA, this study aims to make an inventory of BPA alternatives that can be used in FCM. We further developed a prioritization strategy to identify substances that can possibly be used as a drop-in alternative to BPA, as this information is scarcely available. This is of great importance to avoid regrettable substitution of BPA in FCM by alternatives with similar or unknown hazardous properties. It can be expected that this substitution and concurrent development of BPA alternatives will be accelerated with the ongoing restrictions on the use of BPA and other bisphenols in FCM in the EU (European Commission 2023). This work can inform food safety authorities on targeted monitoring of substances that can potentially replace BPA in FCM.

Material and methods

Inventory of substances that are listed in literature as BPA alternatives

In previous work, we have inventoried potential alternative substances for BPA that are used in consumer products (den Braver-Sewradj et al. 2020). That inventory was built using publications and databases from a variety of official bodies and governmental agencies, published before 2018. The current work is a follow-up of that inventory; therefore, only reports and scientific publications published in or after 2018 are included. Following completion, it was confirmed that all potential alternative substances with reported uses in FCM that were included in the previous inventory (den Braver-Sewradj et al. 2020), were included in the current inventory.

An inventory of substances that are considered alternatives to BPA was built from the following information sources (Figure 1):

1. ECHA’s assessment of regulatory needs for the group of bisphenols (ECHA 2021a)

   ECHA has performed an assessment of regulatory needs with the aim to help authorities conclude on the most appropriate way to address identified concerns for bisphenol-type substances. For this purpose, a group of similar substances was created by chemical structure searches from the substance identity information that is provided to ECHA in the registration dossier. Substances with the presence of a “bisphenol” moiety were grouped.

2. The 2017–2018 review cycle of the Identification of Risk Assessment Priorities (IRAP), performed by Environment and Climate Change Canada (ECCC) and Health Canada (HC) (Health Canada 2018)

   This review is performed under Canada’s Chemicals Management Plan (CMP) and focuses on identifying sources of new information that would constitute indicators of hazard and/or exposure.

3. Technical consultation: Proposed subgrouping of BPA structural analogues and functional alternatives – environment and climate change Canada Health Canada (Health Canada 2020)

   Following the 2017–2018 IRAP review mentioned above, BPA structural analogues and functional alternatives were identified for further consideration in a problem formulation process in the context of prioritization. This technical consultation was published to obtain input on the group of BPA analogues that was identified through the scoping and subgrouping approach taken under Canada’s CMP.

4. The Notice with respect to BPA and BPA structural analogues and functional alternatives (published in the Canada Gazette, Part I on 13 November 2021) (Health Canada 2021)

   The Government of Canada published a notice in order to gather information on the commercial status, industrial processes, and downstream use of BPA and BPA structural analogues and functional alternatives. The information was requested in order to inform prioritization decisions, risk assessment actions and risk management measures, if needed.

5. UBA report on substitution candidates for BPA and BPA analogues (2019) (UBA 2019)

   UBA performed a literature search in scientific literature and in the DEPATISnet patent database to identify environmentally relevant BPA substitution candidates.

6. The dedicated website of the French National Institute for Industrial Environment and Risks (INERIS) that promotes safer alternatives to bisphenols (INERIS 2022)

   INERIS provides tabulated information on potential alternatives to bisphenols.

7. Scientific publications

   A literature search was performed in Embase.com in the time period of 2018 to March 2022 to identify scientific publications on drop-in alternatives to BPA. The search strategy can be found in Appendix A. For substances for which CAS numbers were not provided by the authors, these were looked up in online databases ChemSpider and PubChem.

Based on the inventory, a list of BPA alternatives was generated by applying the following steps:

• Duplicates were identified by CAS number and removed from the inventory.

• General mentions of BPA substitution that could not be related to a public or meaningful name, for example, “fully biodegradable resins without BPA”, were removed from the list.

• (Building blocks for) alternative materials were removed from the list, since these were not within the scope of this study. Alternative materials in the inventory were identified from the description of applications in the publications that were used to build the inventory.

Figure 1. Overview of the various sources used to identify potential alternatives to bisphenol A.

Identification of potential functional alternatives in FCM from the inventory

Chemicals that can be used as drop-in substitutes in FCMs are required to have the same chemical functionality as BPA. Therefore, the inventory of substances was evaluated by expert judgment from an experienced material expert and an expert in chemistry. Substances included in the list as defined above (Inventory of substances that are listed in literature as BPA alternatives) were examined for functional groups that are required for polymerization. For convenience, the chemical structure of most substances has been inserted in the inventory. The following three categories were made:

• Y (yes): Substances were marked as “a potential replacement of BPA” when two or more hydroxyl (OH) groups were present, similarly to BPA. Hydroxyl groups are important for the polymerization of the substance in the polymer manufacturing process (Sadoway 2024). This does, however, not mean that the substances in this category necessarily are or can be used in BPA free PC or epoxy resins. The final use of an alternative monomer depends mainly on the final properties of a polymer or epoxy resin coating, but this has not been included in this evaluation.

• N (no): Substance lacking at least two functional hydroxyl groups were marked as “not a BPA replacement”. Also some polymers as well as oligomers and prepolymers that did not meet the requirement of two or more hydroxyl groups were considered not suitable BPA replacements based on their properties (e.g. BADGE (CAS 1675-54-3), with no free hydroxyl groups and high molecular weight).

• Q (questionable): This category comprises substances for which the use as potential BPA replacement is presently considered unlikely, but cannot be excluded. Substances are marked as “questionable” when:

  • Two or more OH groups are present; and

  • The availability of the functional groups for the polymerization reaction was considered to be affected due to steric hindrance (e.g. 4,4′-isopropylidenebis[2,6-di-tert-butylphenol] (CAS 13676-82-9)); and/or

  • The substance is known to be used as additive, e.g. antioxidant, dye, pH indicator, or biocide. Based on expert judgment by the material expert and an expert in chemistry, these substances were considered less attractive for the industry to use as BPA replacement due to factors such as high costs and properties of the substance (e.g. colors of dyes and indicators, antioxidants which are often hindered phenols; Lou et al. 2024). Future polymerization innovation or availability and costs of substances may change and thus influence the potential use of the substances.

Complementing the inventory with information on structure and hazard

More information was collected on the pool of substances that was considered as potential functional replacement for BPA (“Yes” group) (Figure 2) for the purpose of prioritization. We obtained information on additional structural characteristics, authorization for use in FCM, known hazardous substance, presence in the FCCmigex database, presence on ECHA’s list with substances proposed for further risk management measures (ECHA 2021a) (Supplementary material).

• Additional structural characteristics

  A material expert and an expert in chemistry reviewed the remaining substances and indicated the number of hydroxyl groups on the substances. Also, ionized or otherwise considered very complex substances were highlighted.

• Identification of substances with FCM authorization at EU or NL national level

  The substances identified as potential functional replacement for BPA were examined for their presence, function (when available) and migration restriction in the regulations on FCM. For authorization on EU level, the EU Regulation No. 10/2011 on plastic materials and articles intended to come into contact with food, as last amended by the 15th (Regulation (EC) 10/2011 (EU 2020)) was examined by a search on CAS numbers. This regulation contains the Union List, which is a positive list of substances that are allowed to be used in plastics. The Union List also contains some substances typically used in coatings, but for coatings the Union List is not a restrictive list; other substances can be used as well, depending on national legislation. With regard to the Dutch National Authorization, the Commodities Act Regulation on Packaging and Consumer Articles ((Staatscourant 2014), as last amended in 2022 (Staatscourant 2022)) provides detailed lists of substances that are allowed in FCMs in the Netherlands. All chapters were screened for the presence of alternative substances listed in the inventory. Chapter X – coatings is of particular interest for BPA replacements and was examined extensively for potential replacements of BPA. It allows the use of additional substances (e.g. aromatic diols) which are not included in the EU regulation (10/2011). For monomers and additives used in plastics, the Dutch regulation (Chapter I) refers to the provisions of the EU regulation on plastics.

• Identification of known substances of concern or otherwise hazardous substances

  The candidate list of Substances of Very High Concern (SVHC) (ECHA 2023a), the authorization list (Annex XIV) (ECHA 2023a) and the restriction list (Annex XVII) (ECHA 2023a) were screened for the presence of the substances in the inventory. SVHCs are substances that may have serious effects on human health and/or the environment. In the European Chemicals legislation REACH, SVHCs are classified as being carcinogenic, mutagenic, and/or reprotoxic (CMR), as being persistent, bioaccumulative, and toxic (PBT) or very persistent and very bioaccumulative (vPvB), or as substances of equivalent concern, like endocrine disruption (ED) (Wassenaar et al. 2019; ECHA 2024a, 2024b).

• Presence in FCCmigex database

  To further link the identified potential BPA alternatives to their relevance in FCM, we cross-checked their presence in the FCCmigex database. This database, created by the Food Packaging Forum, contains scientific data on measurements of migrates and extracts of FCMs and articles (Geueke et al. 2022). Information on migration or presence in FMC (i.e. extraction data) was added to the inventory.

• Substances for which regulatory risk management is proposed

  We also identified the substances in the inventory that were categorized by ECHA in their assessment of regulatory needs of a group of substances. In particular, we identified the substances with a need for restriction, based on the potential for widespread use and the information available on the potential endocrine disruptive, reproductive and PBT/vPvB properties (ECHA 2021a).

Figure 2. Overview of the selection criteria to prioritize the potential BPA alternatives. Numbers in italics are the total number of substances from the inventory applicable to the criteria.

Prioritization of substances

The pool of substances collected that was considered as potential replacement for BPA (“Yes” group) was further prioritized based on the collected information on the substances as described in the section above (Complementing the inventory with information on structure and hazard). The following criteria were used for further prioritization, which is summarized in Figure 2:

1. We prioritized the substances with two hydroxyl groups attached. Substances with three or more hydroxyl groups contain multiple reactive sites, and will likely form a cross-linked polymer, thereby losing the flexibility required for the product (with some typical examples being (2-hydroxyphenyl)bis(4-hydroxy-2,3,5-trimethylphenyl)methane (CAS 184355-68-8) or 2,6-bis(p-cresol-2-ylmethyl)-p-cresol (CAS 1620-68-4) (Patil et al. 2021). We also did not select the substances that are ionized or otherwise considered very complex, since these substances are also not expected to be readily used as alternatives to BPA in FCM (with some typical examples being tris[4,4′-thiobis(3-methyl-6-tert-butylphenol)] phosphite (CAS 36339-47-6) or thymol blue sodium salt (CAS 62625-21-2).

2. We prioritized the substances that were not authorized for use in FCM following the European or Dutch regulations. Considering a regulatory point of view, authorized substances are not the most relevant substances to select for further investigation regarding regrettable substitution. These substances are namely already evaluated, to some extent, and they were not readily considered a health risk at the time of their evaluation. We rely on this expert judgment, therefore, the authorized substances (16 substances) were excluded from the list and only the substances that are not authorized for use are prioritized in this criterion. We must note that these assessments may have been made a long time ago.

3. We prioritized the substances that are not known to be identified as hazardous substances by regulatory bodies, i.e. substances that were not identified as SVHC nor present on the authorization list or restriction list. For those substances, additional risk management measures (will) apply or phasing out can be expected. It is, therefore, less likely that these substances will be used by industry to substitute BPA.

4. We prioritized the substances that were present in the FCCmigex database. For authorities, in particular for enforcement purposes, analytical detection is imperative to monitor the use of these substances in FCM. For substances in the FCCmigex database, methods for analytical detection and/or migration of substances are published and it can therefore be assumed that monitoring of these substances is feasible.

5. Finally, we also prioritized the substances that were not identified as substances with a need for restriction by ECHA (2021). This prioritization criterion is of relevance as it is expected that industry is not likely to use substances that are marked by ECHA as BPA alternatives that may become restricted. Although ECHA’s report does not have a legal status, phasing out or restriction of these substances in the near future may reasonably be expected and, therefore, these substances were excluded from further investigation.

Data on hazard and migration of prioritized potential BPA alternatives

To identify a possible health risk of the prioritized substances, more information was obtained focusing on hazard identification and migration potential. Specific searches for information were performed as described below.

REACH registration dossiers

The REACH registration status was noted for each substance, i.e. registered, pre-registered, or not registered (ECHA 2023b). Substances that are not registered are either exempted from registration, not manufactured or imported in the EU, or do not exceed the threshold for import/manufacture of one ton per year. The registered tonnage volume was added to the inventory for the substances that were registered under REACH. It should be noted that also substances that do not reach the threshold of one ton per year may be used as alternatives to BPA and may especially lead to relevant exposure levels if the main use is in FCMs.

In addition, if available, the oral and dermal derived no effect levels (DNELs) for the general population and related most sensitive endpoint were included in the inventory for the prioritized substances as identified from the disseminated registration dossiers in the REACH database. The inhalation route was not considered since consumers (considered as the general population) are mainly exposed to potential BPA alternatives processed in materials. It was not used to prioritize the most relevant BPA alternatives for further assessment of the possible health risks.

Search on hazard data

Together with a data information specialist, a search string was developed to identify scientific literature on each prioritized substance in the Embase.com database (Embase 2023). Appendix B reports the search string that was used. In total, 134 references were identified. This search was performed in June 2023. Titles and abstracts were screened for exclusion according to the following criteria: solely in vitro studies; solely in silico studies; in vivo studies using yeast; bacteria or algae; human biomonitoring studies; occurrence or migration studies; induced disease model studies; studies not addressing the prioritized compounds and studies not written in English. Ex vivo studies, in vivo studies other than those that were excluded and case reports were included, resulting 26 full text articles. In addition, the National Toxicology Program study reports were screened for long-term or short-term toxicology/carcinogenicity studies on the prioritized substances (NTP 2023). Moreover, the literature compiled in the PubChem database on substance information was screened for additional relevant titles for each of the prioritized substances (Pubchem 2023).

Chemical similarity screening and structure–activity relationships

To obtain more insight to hazardous properties of the less familiar potential BPA alternatives, a chemical similarity screening was performed using the ZZS (Zeer Zorgwekkende Stoffen – SVHC) similarity tool as described in Wassenaar et al. (2019, 2021, 2022). This tool can be used to identify chemicals of potential concern by screening the structural similarity of a substance to that of known SVHCs. A cutoff level of >50% confidence in the model prediction was used to presume that it is likely that a substance will have a similar “concern” as the most similar SVHC substance it is compared to. This results in a positive predicted value (or precision) of ∼90% for the CMR substances, 96% for the PBT substances, and 100% for the ED substances (Wassenaar et al. 2019). Predicting negatives, i.e. if a substances is “safe”, is much less reliable using the similarity tool as it yields relatively high percentage of false negatives (Wassenaar et al. 2022). This is due to the fact that the similarity is only “measured” compared to the known SVHC substances, there is no “knowledge” in the model about chemical features that are common to “safe” substances.

In addition, the Derek Nexus software (v6.2.1, from Lhasa Ltd., Lhasa Ltd. 2023) (Judson et al. 2003) was used to assess the presence of chemical (sub)structures (so called “structural alerts”) of the potential BPA alternatives that are related to the human health endpoints estrogenicity, carcinogenicity, mutagenicity, chromosome damage in vitro/in vivo, skin irritation/corrosion, skin sensitization, and photo allergenicity. Qualitative nuance in the results is given by the terms equivocal < plausible < probable < certain (Judson et al. 2003). When the result is “less” than equivocal, it does not mean that a substance is predicted as not toxic, but no prediction for the absence of toxicological effects is made, except for the endpoints Ames mutagenicity and skin sensitization, where the alert models are considered good enough to predict that a substance will be non-mutagenic, or a non-sensitizer if none of the structural alerts for these endpoints are identified.

Search on migration data

The novel FCCmigex database was consulted to collect scientific information on the migration of the prioritized substances (Geueke et al. 2022; FCCmigex 2023). This database was last updated in April 2023. The respective CAS registry numbers of the prioritized substances were used to filter and extract data on migration from e.g. plastic and coatings into food. All database entries on the prioritized substances were recovered and summarized.

Stakeholder consultation

We also reached out to national (Dutch) trade and sector organizations to compare the theoretical identification of potential BPA alternatives based on the criteria described above and the actual substances that are used by industry. Further information on the methods used and a summary of the results can be found in Appendix C.

Results

Inventory of potential BPA alternatives and prioritization of alternative substances

An inventory of substances that are mentioned in reports and literature as possible alternatives to BPA was built from several information sources from official bodies, governmental agencies (Health Canada 2018, 2020, 2021; UBA 2019; ECHA 2021a; INERIS 2022) and scientific literature (Figure 1). The final inventory of potential BPA alternatives comprised 376 substances.

These substances were evaluated by a material expert and an expert in chemistry for their chemical functionality, i.e. the presence of at least two hydroxyl groups, and corresponding potential to act as drop-in substitute for BPA in PC plastic or epoxy resins. Based on this first screening, 152 substances were identified as potential functional BPA alternatives. One hundred and fifty-five substances were not deemed as possible drop-in substitutes and for 69 substances it was concluded that the use as drop-in substitute was not likely but could not be excluded.

A further prioritization of the substances was performed with respect to the need for monitoring of their presence in FCMs. Five inclusion criteria were, therefore, applied (Figure 2). The first criterion was based on chemical functionality of the hydroxyl groups for effective cross-linking resulting in prioritization of 95 of 152 substances. A second criterion was selected to exclude substances for which current hazard data do not suggest a risk: we identified 135 of the 152 substances that are not specifically authorized (in the EU or the Netherlands) for their use in FCMs. For the third criterion, we excluded substances that are already identified by regulatory bodies as hazardous substances, resulting in 148 of the 152 substances that were not identified as SVHC and are not included on the authorization list (Annex XIV) or the restriction list (Annex XVII). For the fourth criterion, we identified 31 of the 152 potential BPA alternatives for which analytical detection does not pose a problem as they are present in the FCCmigex database. The fifth criterion included substances that are not expected to be targeted by regulatory bodies for authorization or restriction. This concerned substances that were not categorized as substances with a need for restriction, “based on the potential for widespread use and the information available on the potential endocrine disruptive, reproductive and PBT/vPvB properties” (ECHA 2021a), a criterion applicable to 140 of the 152 substances.

Combining these prioritization criteria, 20 of the 152 substances fulfilled all five (Figure 2). Out of these substances, 10 substances were already reviewed in our previous work (den Braver-Sewradj et al. 2020) or are currently under investigation in the Partnership on the Assessment of Risks from Chemicals (PARC) Work package 5.1 for their adversity (PARC 2023). Therefore, these 10 substances were excluded from further investigation: bisphenol G (BPG; CAS 127-54-8), bisphenol C2 (BPC2; CAS 14868-03-2), bisphenol BP (BPBP; CAS 1844-01-5), bisphenol E (BPE; CAS 2081-08-5), bisphenol P (BPP; CAS 2167-51-3), bisphenol PH (BPPH; CAS 24038-68-4), bisphenol FL (BPFL; CAS 3236-71-3), bisphenol AP (BPAP; CAS 1571-75-1), bisphenol C (BPC; CAS 79-97-0), and bisphenol Z (BPZ; CAS 843-55-0). For the remaining 10 substances identified as potential BPA alternatives in FCM, we obtained additional information on hazard and migration from scientific literature. These 10 substances are: bisphenol TMC (BPTMC; CAS 129188-99-4), 4,4′-dihydroxydiphenyl ether (DHPE or p,p-oxybisphenol, CAS 1965-09-9), 2,2′‐bisphenol F (2,2′-BPF, CAS 2467-02-9), 2,4′-bisphenol F (2,4′-BPF, CAS 2467-03-0), 4,4-dihydroxybenzophenone (DHBP, CAS 611-99-4), 2,2′-bisphenol A (2,2′-BPA, CAS 7559-72-0), tetrachlorobisphenol A (TCBPA, CAS 79-95-8), 3,3′-dichlorobisphenol A (CAS 79-98-1), BPA 2EO (CAS 901-44-0), and benzophenone-6 (BP-6; CAS 131-54-4) (Table 1).

Table 1. Overview of the selected potential BPA alternatives in the current study and their CAS numbers.

Substance	CAS
Bisphenol TMC	129188-99-4
4,4′-Dihydroxydiphenyl ether	1965-09-9
2,2′‐Bisphenol F	2467-02-9
2,4′-Bisphenol F	2467-03-0
4,4-Dihydroxybenzophenone	611-99-4
2,2′-Bisphenol A	7559-72-0
Tetrachlorobisphenol A	79-95-8
3,3′-Dichlorobisphenol A	79-98-1
BPA 2EO	901-44-0
Benzophenone-6	131-54-4

Hazard profile and migration potential of prioritized potential BPA alternatives

For the prioritized potential BPA alternatives, a summary of the REACH registration status, available hazard data in the disseminated REACH registration dossier and hazard and migration data published in scientific literature is given below. Chemical similarity screening or structure–activity relationship assessments can be informative for substances when little is known regarding their hazard. As few data are available for the potential BPA alternative substances, we performed a chemical similarity screening using the ZZS-tool (SVHC-tool) and a structural alerts assessment using the Derek Nexus tool to provide information on their similarity to SVHCs and activity alerts (Table 2).

Table 2. Chemical similarity and structural alerts of the prioritized substances.

Substance	Similarity tool			Derek Nexus
CAS, common name Chemical structure	SVHC category ≥50%	Confidence	Most similar SVHC substance	Category*	Confidence	Alert
80-05-7 BPA	R substance	100	4,4′-Isopropylidene diphenol (BPA)	Estrogenicity	Plausible	Bisphenol or precursor
	ED substance	100	4,4′-Isopropylidene diphenol (BPA)	Carcinogenicity	Plausible	Bisphenol or precursor, from estrogenicity
				Skin irritation/corrosion	Plausible	Phenol
129188-99-4 BPTMC	R substance	85	4,4-Isobutylethylidene diphenol	Estrogenicity	Plausible	Alkyl phenol or precursor bisphenol or precursor
	ED substance	68	Representing 4-nonylphenol, branched^†	Carcinogenicity	Plausible	Alkyl phenol or precursor, from estrogenicity Bisphenol or precursor, from estrogenicity
				Skin irritation/corrosion	Plausible	Phenol
1965-09-9 DHPE	CM substance	54	4,4′-Oxydianiline	Chromosome damage in vitro	Plausible	4-Alkoxyphenol or 4-aminophenol
				Chromosome damage in vivo	Equivocal	4-Alkoxyphenol or 4-aminophenol
				Skin irritation/corrosion	Plausible	Phenol
2467-02-9 2,2′-BPF	R substance	50	4,4′-Isopropylidene diphenol (BPA)	Chromosome damage in vitro	Plausible	4-Alkoxyphenol or 4-aminophenol
				Skin irritation/corrosion	Plausible	Phenol
2467-03-0 2,4′-BPF	R substance	58	4,4′-isopropylidene diphenol (BPA)	Chromosome damage in vitro	Plausible	4-Alkoxyphenol or 4-aminophenol
				Hepatotoxicity	Plausible	p-Alkylphenol or derivative
				Skin irritation/corrosion	Plausible	Phenol
611-99-4 DHBP	R substance	77	4,4′-Isopropylidene diphenol (BPA)	Carcinogenicity	Plausible	Diaryl ketone
	ED substance	63	4,4′-Isopropylidene diphenol (BPA)	Skin irritation/corrosion	Plausible	Phenol
				Photo allergenicity	Plausible	Diaryl ketone or precursor
7559-72-0 2,2′-BPA	R substance	66	4,4′-Isopropylidene diphenol (BPA)	Skin irritation/corrosion	Plausible	Phenol
	ED substance	52	4,4′-Isopropylidene diphenol (BPA)
79-95-8 TCBPA	R substance	64	4,4′-Isopropylidene diphenol (BPA)	Carcinogenicity	Plausible	Polyhalogenated aromatic
				Skin irritation/corrosion	Plausible	Phenol
79-98-1 3,3′-dichlorobisphenol A	R substance	77	4,4′-Isopropylidene diphenol (BPA)	Skin irritation/corrosion	Plausible	Phenol
	ED substance	67	4,4′-Isopropylidene diphenol (BPA)
901-44-0 BPA 2EO	ED substance	59	Representing ethoxylated 4-(1,1,3, 3-tetramethylbutyl)phenol^‡	No alerts identified
	Substances with other ZZS properties	70	Representing nonylphenolethoxylate with 1-2 ethoxymonomers carboxylated (NPEO1 + 2C)
131-54-4 BP-6	No sufficiently similar SVHC identified			Carcinogenicity	Plausible	Diaryl ketone
				Photo allergenicity	Plausible	Diaryl ketone or precursor

SVHC: substance of very high concern; R: reproductive; ED: endocrine disrupting; ZZS: substances of very high concern.

* All substances were inactive for the mutagenicity alert and almost all (except DHPE, BFDGE·2H2O, and BPA2EO; non-sensitizer and BP-6; plausible (resorcinol or precursor alert) gave an equivocal skin sensitization alert. Because hardly any distinction can be made in these categories, these were not included in the table.

^† Similarity based on a selection of the isomers to aid model-optimalization.

^‡ Also covering well-defined substances and UVCB substances, polymers, and homologues.

Bisphenol TMC

Hazard

BPTMC is pre-registered under REACH and has a reported self-classification for skin irritation 2 (H315), eye irritation 2 (H319), and STOT SE 3 (H335) (ECHA 2021b). However, BPTMC has also been added to ECHA’s database under a different EC number, without a CAS identifier. Under this EC number, BPTMC has been registered under REACH with a confidential registered tonnage band (ECHA 2020a). No harmonized classification or self-classification was provided, but the oral DNEL for the general population was reported as 0.175 mg/kg bw/d based on repeated dose toxicity. This DNEL was derived from a no observed adverse effect level (NOAEL) of 35 mg/kg bw/d. No hazard was identified for acute or short-term exposure. The NOAEL from the oral repeated dose toxicity study and same assessment factors were used to derive the dermal DNEL of 0.175 mg/kg bw/d (ECHA 2020a).

Literature review revealed two additional studies that reported specific effects of BPTMC. Li XP et al. (2023) studied the ecotoxicological effects of BPTMC and reported an LC50 of 4.6 mg BPTMC/L in marine medaka embryos after four days of exposure. Developmental effects were studied from 1.5 h post fertilization (hpf) up to 14 days post fertilization (dpf) and the effects observed after daily exposure to the lowest concentrations of BPTMC (0.5–100 μg/L) manifested in an increased hatching rate, heart rate, malformation rate, mortality, and swimming velocity. At the highest exposed concentration (2 mg/L), no embryos hatched and a decrease in heart rate and swimming velocity was observed.

BPTMC was also studied in ex vivo cultures of fetal Sprague-Dawley testes at GD15 (Tardif et al. 2023). After a 48 and 72-h exposure of 10 µM BPTMC to the testes cultures, basal and luteinizing hormone (LH)-stimulated testosterone levels increased. No effect on germ cell density or Sertoli cell density was observed. However, an increase in Leydig cell density and area was observed after BPTMC treatment. The authors conclude that BPTMC induced an impact on steroidogenesis and its associated fetal Leydig cells ex vivo.

Structural similarity and structure–activity relationship

BPTMC shows the highest structural similarity to the reproductive SVHC 4,4-isobutylethylidene diphenol (85% confidence that is will be of similar reproductive concern) and the ED SVHC group of nonylphenols (68% confidence of similar concern). With 79% confidence, BPTMC is structurally similar to BPA (data not shown). Plausible activity was predicted with the Derek model for estrogenicity (due to the alert “alkyl phenol or precursor” and “bisphenol or precursor”), carcinogenicity (due to the predicted estrogenicity activity) and skin irritation/corrosion (due to the presence of the “phenol” alert), given the shared structural characteristics of the substances (Table 2).

Migration

The presence of BPTMC was analyzed in polystyrene-made food containers that were collected in China (n = 126, detection frequency 65%), Canada (n = 18, detection frequency 0%), and Poland (n = 6, detection frequency 0%) (Zhao et al. 2023). The median concentration BPTMC that was found in the containers was 3.4 ng/g. The mean partitioning coefficients into a food simulant observed after 30 min was 0.39% in tap water, 0.26% in ethanol (EtOH) 10%, 5.5% in EtOH 50% and 6.5% in corn oil). No migration was observed into steamed rice (Zhao et al. 2023). BPTMC was also analyzed in non-PC baby bottle leachates (with a detection of frequency 55% (n = 20), median = 0.36 ng/L), but not in sippy cups (n = 13) that were sold in Canada (Siddique et al. 2021). This migration was only observed up to 2 h of incubation time and when water was used as a simulant. No migration was observed when (10% and 50%) EtOH was used as a simulant to mimic acidic and fatty foods, in contrast to the previous study. It cannot be excluded, however, that BPTMC decomposed or has reacted with the simulant and has formed another chemical that was not included in the analyses. No migration over time (up to 10 days) was observed (Siddique et al. 2021). No BPTMC was detected in water and 3% acetic acid leachates of drinking bottles, baby bottles, a wine glass (made of PC plastic) or samples of canned foods or drinks obtained from stores in the Netherlands (van Leeuwen et al. 2019). This was the only study identified that analyzed BPTMC in PC plastic or epoxy resin.

Conclusion/summary

Two literature studies were available that identified in vivo developmental effects and hormonal changes in organ culture after exposure to BPTMC and no target organ of BPTMC was identified in the study summary in its REACH dossier. BPTMC shows high structural similarity to BPA and is structurally similar to other SVHCs with reproductive and endocrine disrupting properties. It induces the same estrogenicity and carcinogenicity activity alerts as BPA, although these are also prompted by its shared structure with alkyl phenol. The limited data available on the migration of BPTMC show that this substance has migration properties in material other than PC plastic albeit limited. To conclude, the in silico alerts of BPTMC in combination with the available experimental data indicate potential endocrine disrupting properties and developmental effects but more studies on the in vivo toxicity and migration of BPTMC are needed to assess its potential risk.

4,4′-Dihydroxydiphenyl ether

Hazard

DHPE (or p,p-oxybisphenol) is pre-registered under REACH and has a reported self-classification for acute toxicity 4 (H302), skin irritation 2 (H315), skin sensitization 1 (H317), eye damage 1 (H318), and STOT SE 3 (H335) (ECHA 2023f). No REACH dossier is available.

Two studies were identified that assessed the effects of DHPE in vivo. Qin et al. (2014) studied the acute toxicity of the substance in adult zebrafish (by reference of the OECD standard protocol 203). The authors report an LC50 of 9.7 mg DHPE/L in adult fish after four days of exposure, suggesting that DHPE exhibits moderate toxicity to adult zebrafish. DHPE treatment up to 6 mg/L to zebrafish embryos for four days showed hatching retardation (reduced effect on embryo incubation), reduced body length and developmental malformations such as increased rates of yolk sac edema, pericardial edema, and spine deformation.

DHPE was also studied in male ICR mice that were orally exposed to 1.2 mg/kg bw/d for 28 days (Zhang X et al. 2014). The growth rate of the mice was not affected by DHPE. Relative liver weight and kidney weight were statistically significantly increased and decreased, with 5.7% and 1.4%, respectively. Pathological changes were observed in the liver and DHPE treatment disturbed the hepatic redox status and metal homeostasis was perturbed. This indicated the occurrence of hepatic oxidative damage according to the authors.

Structural similarity and structure–activity relationship

DHPE shows highest structural resemblance to 4,4′-oxydianiline, a CM SVHC. The confidence in the similarity leading to similar (CM) concern of 54% is close to the threshold of 50%. Derek Nexus predicted possible chromosome damage in vitro (plausible) and in vivo (equivocal) because of the presence of 4-alkoxyphenol or 4-aminophenol, and skin irritation/corrosion was predicted as “plausible” due to the phenol structure present in DHPE (Table 2).

Migration

No studies were found on the migration of DHPE from PC plastic or epoxy resins. Migration of DHPE in tap water and artificial saliva from plastic drinking bottles made of the alternative material Tritan™ from five different brands for sale in Austria was analyzed but not detected (Banaderakhshan et al. 2022).

Conclusions/summary

The limited identified literature indicates signs of toxicity in both adult zebrafish and zebrafish embryo upon exposure to DBPE, and pathological changes in the liver were observed in rat exposed to DHPE. The highest structural similarity to an SVHC was shared with a substance other than BPA and activity alerts were identified based on shared structural characteristics other than that shared with BPA, indicating that this substance is less likely to exert the same effects as BPA, as compared to the other prioritized substances. No migration of DHPE was observed in the one study identified that analyzed DHPE in non-PC plastic drinking bottles. More research is required to assess the possible risk of DBPE.

2,2′-BPF

Hazard

2,2′-BPF is pre-registered under REACH and thus, no dossier is available. A self-classification for skin irritation 2 (H315), eye irritation 2 (H319), and STOT SE 3 (H335) was provided (ECHA 2023c).

One study was identified that studied the effects of 2,2′-BPF. Kubota et al. (2023) studied the effect of 24-h 2,2′-BPF exposure to wildtype zebrafish embryos at 72 hpf. The highest exposure concentration, 100 µM, was lethal to all embryos. CYP19A1b mRNA expression, that is induced by estradiol, did not change after exposure to 2,2′-BPF as compared to the control treatment, in contrast to the same exposure concentration of BPA. This study shows no observed in vivo estrogenic potency of 2,2′-BPF in zebrafish embryos, as measured by the expression of CYP19A1b mRNA.

Structural similarity and structure–activity relationship

2,2′-BPF showed just sufficient structural similarity to BPA (50% confidence) indicating that 2,2′-BPF would likely have similar reproductive toxicological concern as BPA. For the ED endpoint, the similarity to BPA is just below the 50% cutoff (49%) for a positive identification as a potential SVHC. The Derek model predicts chromosome damage in vitro and skin irritation/corrosion as “plausible” due to the 4-alkoxyphenol or 4-aminophenol and phenol alerts, respectively (Table 2).

Migration

2,2′-BPF was analyzed in leachates of aluminum beverage cans with coating (with detection frequency 100%, n = 6) and reusable sports bottles (with detection frequency 9%, n = 11) that were collected in Slovenia (Kovačič et al. 2020). The concentrations that were found in the leachates of the cans were 8.2–31 ng/L using two different simulants (3% acetic acid and 20% ethanol) and a very low concentration of 1.3 ng/L was analyzed in a used reusable sports bottle made of steel (yet, higher concentrations of BPF's dominant isomer (4-4′-BPF) were observed in the cans) (Kovačič et al. 2020). No 2,2′-BPF was detected in leachates of drinking bottles, baby bottles or a wine glass (made of PC) obtained from stores in the Netherlands (n = 13). 2,2′-BPF was detected at 0.007 ng/mL in a carbonated soft drink, but not in the other food products and drinks (n = 8) from the Dutch market (van Leeuwen et al. 2019).

Conclusions/summary

In the one identified study on toxicological effects, no clear estrogenic potential of 2,2′-BPF was shown. Although the substance is structurally similar to BPA and would likely exert similar toxicological concern, it is with lower confidence than many of the other prioritized compounds. An alert for estrogenicity (and other alerts) was also not identified based on the structure of the substance. 2,2′-BPF can migrate into foodstuffs, based on the limited available data. More research is required to assess the possible risk of the substance.

2,4′-BPF

Hazard

2,4′-BPF is pre-registered under REACH. A self-classification for Skin sensitization 1 (H317), Eye damage 1 (H318), and STOT SE 3 (H335) was provided (ECHA 2023e).

Kubota et al. (2023) also studied the effect of 24-h 2,4′-BPF exposure to wildtype zebrafish embryos at 72 hpf. CYP19 mRNA expression was induced after the exposure of the embryos with an EC50 of 35 µM, which was approximately 4.5-fold that of BPA (7.6 µM) and threefold that of BPF ((4,4′-BPF) 12 µM). The authors conclude that exhibiting one hydroxy-group at the para-position of the phenyl ring caused a slightly lower potency of the compound, as compared to exhibiting two hydroxy-groups at the para-position.

Structural similarity and structure–activity relationship

2,4′-BPF showed slightly more structural similarity to BPA than 2,2′-BPF giving 58% confidence for the prediction that 2,4′-BPF will have the same reproductive toxicity concern as BPA. Activity for chromosome damage in vitro, hepatotoxicity and skin irritation/corrosion were predicted by the Derek model as “plausible” based on the alerts 4-alkoxyphenol or 4-aminophenol, p-alkylphenol or derivative and phenol alerts, respectively (Table 2).

Migration

2,4′-BPF was analyzed in leachates of aluminum beverage cans with coating (with detection frequency 100% n = 6) and reusable sports bottles (with detection frequency 0% n = 11) that were collected in Slovenia. The concentrations that were found in the cans were 21–391 ng/L using two different simulants (3% acetic acid and 20% ethanol). (Yet, higher concentrations of BPF's dominant isomer (4-4′-BPF) were observed in the cans) (Kovačič et al. 2020).

Conclusions/summary

In the one identified study on its toxicological effects, estrogenic potential of 2,4′-BPF was shown in zebrafish embryos, albeit with a lower potency than BPA and BPF’s dominant isomer (4,4′-BPF) and no direct adversity was observed. Although the substance is structurally similar to BPA and would likely exert similar toxicological concern, no structural alerts were identified for estrogenicity based on its structural properties, in contrast to the alerts for chromosome damage in vitro, hepatotoxicity and skin irritation/corrosion based shared structural characteristics with other compounds than BPA. Similarly to 2,2′-BPF, 2,4′-BPF can migrate from coatings into foodstuffs in the one study identified. More research on the toxicological potential of this substance is needed to assess its possible risk.

4,4-Dihydroxybenzophenone

Hazard

DHBP has been registered under REACH with a tonnage band of 10–100 ton/year. A self-classification for eye irritation 2 (H319) was provided. In the dossier, the conclusion from the hazard assessment for the general population (dermal and oral route) was low hazard, but no specific threshold was derived. The reasoning was that, in the absence of specific data for this substance, data on benzophenone were used as a surrogate molecule for the evaluation (EFSA TDI 0.03 mg/kg b/d). As exposure of the substance would be in the manufacturing process under carefully controlled conditions, exposure will be minimal and additional animal testing could not be justified (ECHA 2018b).

Several studies reported on the effects of DHBP in (eco)toxicological experiments. Han et al. (2021) studied the 48-h exposure of DHBP to juvenile Daphnia magna between 6 and 24 h of age. The authors report a LC50 of 12.5 mg/L, which was considered medium-level toxicity according to classification criteria. The predicted no-effect concentration (PNEC) to evaluate ecological risk of DHBP was 0.0125 mg/L. The EC50 for acute immobility of DHBP in juvenile D. magna was reportedly 69 mg/L after 48-hours of exposure in another study (Liu H et al. 2015). This value, however, greatly contradicts the reported LC50 and no reference to this (earlier published) EC50 was mentioned in the article reporting the LC50.

In addition, fathead minnows aged between two and three months were exposed to DHBP for 14 days with concentration up to 10, 100, 500, 1000, and 5000 µg/L (Kunz et al. 2006). Although estrogenic effects were observed in vitro, no induction of vitellogenin (VTG), as a proxy for estrogenicity (as its induction is observed after estradiol exposure in the minnows) could be measured after exposure to any of the DHBP concentrations. No toxic side effects (i.e. lethargy, uncoordinated swimming, loss of equilibrium, or hyperventilation) were observed in the fish for any of the exposure concentrations.

DHBP was tested in a uterotrophic assay using juvenile 19 day old Crj:CD (SD) rats without prior ovariectomy that were exposed for three days via subcutaneous injection and a Hershberger assay using eight-week old castrated Brl Han: WIST Jcl (GALAS) rats that were exposed for 10 days via oral gavage (Yamasaki et al. 2003). An increase in uterine weight was only observed after treatment with the highest exposure concentration of 200 mg/kg bw/day. A reduction in uterine weight was observed after co-treatment with synthetic estradiol (that itself increases uterine weight), but again only at the highest exposure concentration. No changes in body weight or male sex organs were observed in the Hershberger assay. After co-treatment with testosterone propionate, a decrease in body weight gain and an increase in glans penis weight were observed after treatment with 600 mg/kg and a decrease in ventral prostate weight and bulbocavernosus/levator ani muscle (BC/LA) weight was observed after treatment with 200 mg/kg. No general toxicity signs were observed in the Hershberger assay after treatment. From this study, the authors concluded that DHBP has partial estrogen agonistic properties and that it also reduces the agonistic effect of synthetic estradiol.

Structural similarity and structure–activity relationship

DHBP shows structural similarity to BPA, both for the reproductive as well as the ED SVHC categories, with 77% and 63% confidence in the prediction that this substance will have similar concerns as BPA. Carcinogenic activity and photo allergenicity were predicted by the Derek model as plausible based on shared structural characteristics with diaryl ketones. Skin irritation/corrosion was predicted to be “plausible” based on similar structural characteristics with phenols (Table 2).

Migration

DHBP was not detected when water was used to extract it from plastic food packaging materials (plastic packaging or plastic bottles, n = 5, not further specified) obtained from local stores in China after an incubation time of 2 h at 70 °C (Li Y et al. 2019). DHBP was also not detected in any of the plastic films for food packaging from China (n = 8) (Hu et al. 2019). Concurrently, DHBP was not detected in leachates from 31 typical commercial plastic food contact samples, in any of the five different food simulants used (10% ethanol, 3% acetic acid, 20% ethanol, 50% ethanol and olive oil), for any of the different temperature conditions that were used in the migration test, according to the different intended uses of the food packaging (Wang J et al. 2016).

Conclusions/summary

No direct adversity was observed as a result of DHBP exposure but partial estrogen agonistic properties and antagonistic properties were observed in a rat uterothropic assay (only when exposed to the highest dose of DHBP), while no estrogenicity could be shown in vivo in Japanese medaka. The structural characteristics of DHBP also did not identify an alert for estrogenicity, although the substance is structurally similar to BPA and would likely exert similar reproductive and endocrine disrupting effects (with higher confidence than most other prioritized compounds). In the limited identified studies, DHBP could not be detected in leachates of plastic food contact samples. More research on the endocrine disrupting potential of this substance is needed to assess its toxicological profile and possible risk.

2,2′-BPA

Hazard

2,2′-BPA has not been registered under REACH. No additional hazard data were identified in our literature search.

Structural similarity and structure–activity relationship

Structural similarity between 2,2-BPA and BPA gives a confidence in the prediction that 2,2′-BPA will have similar reproductive toxicity concern of 66% and similar ED concern of 52%. In addition, the same structural alert as found in BPA for only the skin irritation/corrosion activity was identified (phenol) by the Derek model (Table 2).

Migration

No 2,2′-BPA was detected in leachates of drinking bottles, baby bottles or a wine glass (made of PC) obtained from stores in the Netherlands (n = 13), using water or 3% acetic acid as a simulant (van Leeuwen et al. 2019). Also, 2,2′-BPA was not detected in any of the canned foods and drink from the Dutch market (n = 8) (van Leeuwen et al. 2019).

Conclusions/summary

No information on the hazard of 2,2′-BPA could be identified in literature. It shows structural similarity to BPA and only a structural alert for plausible skin irritation/corrosion was identified similarly to BPA. It was not detected in the one study on the migration of this substance that was identified and no conclusion as to the potential risk of this compound can be drawn, as more research on the toxicological properties and migration is needed.

Tetrachlorobisphenol A

Hazard

TCBPA is pre-registered under REACH and has a reported self-classification for skin irritation 2 (H315), eye irritation 2 (H319), and STOT SE 3 (H335) (ECHA 2023d).

Various literature studies reported on the effects of TCBPA in several species. A transgenic liver tumor zebrafish model was used by Chen et al. (2022) to identify the tumor-promoting potential of TCBPA. Fluorescence, as proxy for the expression of the oncogene, was increased at 500 µg/L and 1000 µg/L exposure from 3 dpf to 7 dpf. Mortality was observed after exposure levels higher than 5000 µg/L. All transgenic zebrafish exposed to TCBPA showed progression to the hepatocellular carcinoma stage after histological examination, as compared to the hyperplasia and hepatic adenoma stages that were observed in the control. TCBPA thus exhibits liver tumor-promoting effects in zebrafish and is more potent in the induction of the oncogene as compared to BPA according to the authors (Chen et al. 2022).

Other transgenic zebrafish embryos of 4 hpf were exposed to TCBPA for five days to concentrations up to 4 mg/L to assess its developmental toxicity (Liu W et al. 2023). The 120 h-LC50 was 0.628 mg/L and at 0.5 mg/L exposure a decrease in body length was observed. The authors report abnormal swimming behavior after exposure to TCBPA and concluded that the nervous system of zebrafish might be damaged by TCBPA during the developmental period of early life-stage. Also, morphological observations revealed that TCBPA could cause the loss of head blood vessels during the developmental period (Liu W et al. 2023). Following a 24-h exposure to 1 µM TCBPA, transgenic zebrafish embryos expressed GFP as a proxy for PPARy activation. Daily exposure from 3 dpf to 11 dpf resulted in zebrafish with markedly increased levels of lipid accumulation at 11 dpf as compared to the control. In addition, the body mass index of fish exposed to TCBPA was significantly higher than that of vehicle-treated fish at 30 dpf (Liu W et al. 2023). The exposure to 1.5 mg/L TCBPA delayed the hatching rate of zebrafish embryos at 72 hpf, indicating potential developmental toxicity according to Song et al. (2014). Also TCBPA exposure was more acutely toxic than BPA, with 100% mortality of 1 mg/L exposed embryo/larvae by 120 hpf. Developmental lesions (yolk sac, pericardial edema, and hemorrhage) were indeed observed in zebrafish embryos/larvae exposed to TCBPA, already at 1 mg/L after 72 hpf. No effect of BPA treatment up to 1.5 mg/L was observed. In adult male zebrafish (2 month old), increased mortality was observed after exposure to 1.5 mg/L TCBPA for 21 days. TCBPA did not induce VTG (Song et al. 2014).

Yolk utilization and malabsorption was analyzed in zebrafish embryos exposed to TCBPA, from 2 dpf to 5 dpf. TCBPA induced a faster uptake of the yolk, with treated 5 dpf-larvae possessing smaller yolk areas than the vehicle-controls. The LC50 and lowest effect level for TCBPA exposure in this experiment is 3.6 µM and 1 µM, respectively (Kalasekar et al. 2015).

The effect of TCBPA on CYP191A1b mRNA expression was examined in the 72–96 hpf development stage of wildtype zebrafish embryos to assess its estrogen-like potency (Kubota et al. 2023). Yet, no induction of the mRNA was measured after exposure to TCBPA, whereas BPA induced CYP191A1b mRNA expression.

Embryonic exposure to TCBPA significantly inhibited the synthesis of nucleosides, amino acids, and lipids, as well as the supply of energy from the Krebs cycle, thus interfering with normal organ development and suppressing developmental processes in Oryzias melastigma, marine medaka, embryos (Ye et al. 2016). Also, lactate accumulation and dopamine pathway activation and a decrease in inhibitory neurotransmitters were observed, together leading to neural activation in response to TCBPA exposure. Notably, many of the effects observed were heritable according to the authors (Ye et al. 2016).

The effect of TCBPA on the reproduction and development of O. melastigma was assessed during various developmental stages by Huang et al. (2017). TCBPA accelerated the embryonic heartbeat (as did BPA) and led to delayed hatching and a decreased hatching rate (but no effect of BPA was observed at the same exposure concentrations) already at environmentally relevant doses (0.05 and 0.2 mg/L). TCBPA did not affect estradiol or testosterone level in female and male fish, respectively, after a four month exposure period (in contrast to BPA exposure, that resulted in declined estradiol and testosterone levels). The impact of TCBPA on the expression of ERα differed in various developmental stages (Huang et al. 2017).

Also, adult Rana nigromaculata frogs were exposed to 1 mg/L TCBPA for 14 days and the effect on reactive oxygen species (ROS)-dependent mitochondria-mediated apoptosis in the liver was investigated by Jia et al. (2022). The authors conclude that TCBPA induced hepatotoxicity is mediated by the ROS-dependent mitochondrial pathway, as the alanine transaminase level in serum and the content of ROS increased, DNA fragments were observed and Bax/Bcl-2 ratio (mRNA expression) increased dose-dependently. Adult R. nigromaculata were also exposed to concentrations up to 1 mg/L for 14 days by Zhang H et al. (2018). Sperm numbers and sperm mobility were significantly decreased and sperm deformity was significantly increased in a concentration dependent manner following exposure to TCBPA. Serum testosterone increased after exposure to several concentrations of TCBPA, while serum estradiol increased and LH and FSH decreased only at the highest exposure level. Androgen receptor mRNA expression in the testes was markedly decreased. The authors conclude that TCBPA induced reproductive toxicity in R. nigromaculata.

The effect of TCBPA on the reproduction of Caenorhabditis elegans was assessed by exposing C. elegans to environmentally relevant doses for 24 h (Yu et al. 2022). Brood size of the nematode was decreased indicating that TCBPA impaired C. elegans fertility. This effect was postulated to be mediated via genotoxic response in the germline apoptosis pathway.

BALB/c mice, 6 weeks of age, were exposed via oral gavage to 5 and 50 mg/kg/day TCBPA for 14 days to study its immunotoxic effect (Wang Y et al. 2021). The percentage CD3+ T lymphocytes decreased, whereas the percentage Treg cells increased after exposure to TCBPA. The body weights and organ coefficients of spleen and thymus in de TCBPA-treated animals did not differ from the controls. Also, pro-inflammatory and anti-inflammatory immune responses were activated by TCBPA, thereby disrupting the immune system. Moreover, uterine edema was observed in over 80% of the TCBPA-treated mice. Thickened endometrium and infiltration of inflammatory cells were observed in the tissue. The authors conclude that TCBPA can suppress the immune response in BALB/c mice, indicating that TCBPA may enhance susceptibility to infectious diseases or cancer by suppressing the immune system (Wang Y et al. 2021).

One study reported on an uterotrophic assay using ovariectomized mice to assess the estrogenic activity of TCBPA in vivo (Kitamura et al. 2005). A 118% increase in uterine weight was observed after the lowest TCBPA treatment of 20 mg/kg (as compared to 147% following the same dose of BPA), up to 164% after treatment of 500 mg/kg. The authors conclude that TCBPA exerts estrogenic activity in vivo.

Structural similarity and structure–activity relationship

TCBPA is structurally sufficiently similar to BPA to give a 64% confidence in the prediction that it will have identical reproductive toxicity concern. The same alert for skin irritation/corrosion as found in BPA is identified (phenol). Carcinogenicity is also predicted as plausible for TCBPA based on shared structural characteristics with polyhalogenated aromatics, which is not a structural feature shared with BPA (Table 2).

Migration

TCBPA was analyzed in polystyrene-made food containers that were collected in China, Canada, and Poland (n = 150), but not found in any of the samples (in contrast to other bisphenols) (Zhao et al. 2023). TCBPA was detected in commercial food contact paper from China (n = 32), or Japan/Europe (n = 42) that was ordered online with a detection frequency of 11% (Zhou et al. 2015). The average detected concentration in food contact paper (coffee filter) was 0.0097 ng/g. The detection frequency of TCBPA was much higher in bleached food contacting papers. No migration from coffee paper into coffee solution was found for TCBPA (Zhou et al. 2015). TCBPA was not detected in the analyzed virgin or recycled paper or paper board food contact products (e.g. wrapping paper, tea bag, coffee filter, and noodle cup) from Japan (n = 28) (Ozaki et al. 2004).

Conclusions/summary

Several studies showed that TCBPA induces developmental toxicity and interferes with normal (organ) development in zebrafish, possibly with higher potency than BPA. TCBPA did not show any signs of estrogenicity in several studies in zebrafish, in contrast to an indication for minor estrogenicity in an ovariectomized mouse uterotrophic assay. Some studies also indicated that TCBPA can interfere with fertility in R. nigromaculata and C. elegans. Although structurally similar to BPA, and therefore likely to have similar reproductive toxicological concern as BPA, no estrogenic activity was predicted based on its structural characteristics. An alert was predicted for carcinogenicity, given the shared structural characteristics of TCBPA with polyhalogenated aromatics. TCBPA was detected in some food contact paper material, yet its migration has not been demonstrated. More studies focusing on the hazard as well as the migration or presence of TCBPA in various types of PC plastic and epoxy resin FCM are needed to identify a possible risk.

3,3′-Dichlorobisphenol A

Hazard

3,3′-Dichlorobisphenol A is pre-registered under REACH, but no self-classification was provided (ECHA). No additional hazard data were identified in our literature search.

Structural similarity and structure–activity relationship

The structural similarity between 3,3′-dichlorobisphenol A and BPA gives a confidence in the prediction that it will have similar reproductive toxicity concern of 77% and similar ED concern of 67%. The same structural alert as for BPA was identified by the Derek model for skin irritation/corrosion (Table 2).

Migration

3,3′-Dichlorobisphenol A was not detected in the analyzed virgin or recycled paper or paper board food contact products (e.g. wrapping paper, tea bag, coffee filter, and noodle cup) from Japan (n = 28) (Ozaki et al. 2004).

Conclusions/summary

No toxicological information on 3,3′-dichlorobisphenol A could be identified. Although the substance is similar to BPA, indicating that it might have similar reproductive and ED concerns as BPA, the same estrogenicity and carcinogenicity alerts as BPA were not predicted based on its structure. The substance was not detected in the single study that analyzed it in paper food contact products and no information could be identified on the migration potential in PC plastic or epoxy resin. More research is required to assess the possible risk of this substance.

BPA 2EO

Hazard

BPA 2EO, or ethoxylated BPA, has been registered under REACH with a tonnage band of 10–100 ton/year and no self-classification was provided (ECHA 2022). An oral DNEL for the general population was reported as 0.75 mg/kg bw/d based on repeated dose toxicity. This DNEL was derived following a 28-day oral + reproduction/developmental toxicity screening (OECD 422) in rats, where the NOAEL was determined 300 mg/kg bw/d. No additional hazard data were identified in our literature search.

Structural similarity and structure–activity relationship

BPA 2EO is structurally most similar (with 59% confidence) to the ED SVHC representing ethoxylated 4-(1,1,3,3-tetramethylbutyl)phenol and the substance with other SVHC properties representing nonylphenol-ethoxylate with 1–2 ethoxymonomers carboxylated (NPEO1 + 2C) (with 70% confidence), but not to BPA. Based on this similarity model prediction, BPA 2EO might raise a concern for ED-properties which have led to the identification of nonylphenol ethoxylates as SVHCs. The Derek models do not identify any structural alert associated with toxicological effects in BPA 2EO (Table 2).

Migration

A non-targeted screening showed presence of BPA 2EO in paper-based food packaging material from the USA (n = 27). Yet, its identity and quantity were not confirmed using reference analytical standards (Sapozhnikova and Nuñez 2022).

Conclusions/summary

A NOAEL of 300 mg BPA 2EO/kg bg/d was derived from a 28-day oral + reproduction/developmental toxicity screening in rats and no additional toxicological information was identified in literature. BPA 2EO is not structurally similar to BPA itself and no effects could be predicted based on its structural characteristics. More research is needed to identify its toxicological profile. BPA 2EO was shown to be present in paper-based food packaging material in the single study that was identified, yet no information on its migration from PC plastic or epoxy resin FCMs (into food) could be obtained. More research is required to assess the possible risk of BPA 2EO.

Benzophenone-6

Hazard

BP-6 has been registered under REACH with a tonnage band of 1–10 ton/year and no self-classification was provided (ECHA 2018a).

One study on BP-6 was identified in scientific literature. Ovariectomized mice, 8 weeks of age, were exposed to BP-6 subcutaneously or via oral gavage for seven consecutive days with a highest dose of 1000 mg/kg (Ohta et al. 2012). BP-6 showed weak estrogen agonistic effects at highest dose and only by the subcutaneous route of exposure. No agonistic effect was detected by the p.o. route. Antagonistic effects were lacking by either route of exposure in this study.

Structural similarity and structure–activity relationship

BP-6 is not structurally similar to any of the registered SVHCs, above threshold of 50% confidence in the prediction. The shared structural characteristics with diaryl ketone make the Derek models predict carcinogenicity and photo-allergenicity to be “plausible” (Table 2).

Migration

BP-6 was analyzed in PC plastic containers intended to be used in microwave ovens, but not positively detected (Nerín et al. 2003). BP-6 was also not detected in any of the plastic films for food packaging from China (n = 8) (Hu et al. 2019). The compound was also not detected in the PC tableware with different colors (n = 4) (Bignardi et al. 2014).

Conclusions/summary

Weak estrogen agonistic effects after BP-6 exposure were observed in ovariectomized mice, but only after subcutaneous exposure in a single study. The substance is not structurally similar to BPA and estrogenicity was not predicted based on its structural characteristics. However, plausible carcinogenicity was predicted based on its shared structure with a compound other than BPA. BP-6 could not be detected in the few studies that analyzed the substance in FCMs. Yet, no conclusion can be drawn on the possible risk of this compound as a result of insufficient data.

Discussion

The use of BPA, a SVHC, is under heavy international debate and within the EU the substance is proposed to be banned for the use in FCM (European Commission 2023). A great number of substances have been named as alternatives to BPA. To prevent regrettable substitution of BPA by alternatives with similar or unknown hazardous properties, it is of importance to gain the relevant toxicological information on these substances and monitor them properly. Therefore, we developed an inventory and strategy to prioritize the potential BPA alternatives, in order to guide regulatory bodies in their focus on potential BPA alternatives. The focus of this work was on the substitution of BPA in FCM, i.e. in PC plastic and epoxy resins, since food and beverages are considered the main source of exposure to BPA for the general population (Peivasteh-Roudsari et al. 2023).

The inventory was made by first collecting the substances that are mentioned as BPA alternatives from reports published by ECHA (2021a), Health Canada (2018, 2020, 2021), UBA (2019), INERIS (2022), and additionally from scientific literature. For all these substances in the inventory, we collected additional information that was subsequently used for prioritization, including information on structural functionality, legislation, and REACH status.

The inventory of substances that are mentioned in reports or scientific literature contains over 300 chemical substances that have been named as potential BPA alternative. It is not feasible to test and monitor all of them. However, it is also not likely that all of these substances are in fact candidates to substitute BPA in FCM. Factors such as technical feasibility (based on chemical functionality) and economic feasibility need to be taken into account. The criteria that were used to prioritize the substances were additionally chosen from a regulatory perspective, as the foreseen ban or restriction on BPA and other bisphenols make that these substances are less likely to be used as drop-in substitute for BPA. Using different sets of criteria and information will, however, result in a different set of prioritized compounds. Selection criteria based on different fields of expertise create an added value in this research. The first prioritization step and criterion were selected by two individual experts. Although these criteria may be considered basic knowledge for chemistry, this is usually not taken into consideration in toxicology. Depending on the nature of the research question, one can select or add relevant criteria and information to focus the prioritization on a larger set of substances from the inventory. This is especially recommended for the substances that are marked as potential or questionable functional alternatives (221 substances), as these may theoretically be considered as technically feasible alternatives. Such strategy can not only be considered in the case of BPA, but for all substances for which (regrettable, or not) substitution is foreseen.

A list of 20 prioritized potential BPA alternatives was identified. The 10 more familiar substances were not further studied, since a lot of data and concerns of these substances are already collected by other initiatives (e.g. den Braver-Sewradj et al. 2020 and PARC). For the 10 less familiar of these substances, more data on hazard and migration were obtained from scientific literature and screening tools.

Hazard data on prioritized substances

Four out of the 10 prioritized substances were registered under REACH, five were pre-registered and one was not registered at all. No harmonized classification applied to any of the substances registered under REACH. Also, no specific use of the substances in FCMs was documented in the REACH dossiers. For none of the prioritized substances, a rodent bioassay (short-term or long-term) was identified.

Only few hazard data are available in literature for the prioritized potential alternatives to BPA and the amount of data varies highly between the substances. The data that are available typically focus on the effects of the compounds in the ecotoxicological field or “level 3” tests (uterotrophic assay or Hershberger assay) where adversity cannot be measured, which hampers the application for human risk assessment (den Braver-Sewradj et al. 2020; WHO 2021). Various studies, however, did report on developmental toxicity or effects on hormonal disruption which may give grounds for follow-up studies. Nonetheless, important information can be obtained from these studies, e.g. that exposure of zebrafish embryos to TCBPA indicated effects on their development in several studies, albeit exerting a seemingly potential different mode of action than BPA as there were no clear signs of estrogenicity (Kitamura et al. 2005; Song et al. 2014; Liu W et al. 2023). Nevertheless, for none of the evaluated potential BPA alternatives, a definitive conclusion as to the hazard of the substance can be drawn.

Chemical similarity screening and structure–activity relationships

As few data are available for the potential BPA alternative substances, we performed a chemical similarity screening using the ZZS-similarity tool (SVHC-similarity tool) developed by RIVM (Wassenaar et al. 2022), and we used a structural alerts assessment using the Derek Nexus software (Lhasa Ltd. 2023). There is other software available to evaluate the likelihood that BPA alternatives will give (similar) toxicological concern as BPA itself, for example, the OECD QSAR Toolbox has several modules (profiles) to assess carcinogenicity, mutagenicity, reproductive toxicity, ED-like behavior and skin irritation and sensitization (OECD 2023). Also the open-access VEGA-suite of models has multiple QSAR models available to evaluate these endpoints (VEGA 2023). As the ZZS-similarity tool and the Derek Nexus tool complement each other in the providing information, we selected these tools for the current project. The Derek software provides information on the various alert categories, where the ZZS-tool can also provide information on the discriminatory power within an alert category. The use of other or additional tools (e.g. OECD Toolbox, VEGA) can be considered in future steps. Combining the outcomes of multiple in silico tools in the prioritization of chemical substances, however, is difficult to interpret and algorithms need to be developed for such analysis.

Of all substances in the chemical similarity screening, BPTMC, DHBP, and 3,3′-dichlorobisphenol A are structurally the most similar to BPA when considering both its Reproductive (88%, 77%, and 77% confidence, respectively) as well as endocrine disrupting SVHC properties (with 68%, 67%, and 63% confidence, respectively) (Table 2). Yet, these substances do not all give rise to the same estrogenicity and carcinogenicity activity alerts as BPA. For DHBP, surprisingly, plausible carcinogenicity was predicted as a result of its shared structural characteristics with diaryl ketones rather than bisphenols (or precursor) structure. This was the same for BP-6, for which no structural similarity to BPA was observed above the set threshold. In addition, for TCBPA, plausible carcinogenicity was predicted as a result of its shared characteristics with polyhalogenated aromatics, rather than BPA. Noticeably, DHPE shows the highest structural similarity to the SVHC 4,4′-oxydianiline, a carcinogenic/mutagenic SVHC. Yet, the structural alert for mutagenicity/carcinogenicity which is present in 4,4′-oxydianiline (the aromatic amine group) is not present in DHPE (Neumann 2010; Wang S et al. 2019). The Derek models do indicate the presence of a chromosome aberration in vitro alert (the 4-alkylphenol group) in DHPE, but this alert acts via a very different potential carcinogenic mechanism than aromatic amines. An alert obtained from such an in silico screening should, therefore, not be considered absolute but should rather give rise to additional (mechanistic) research.

Migration data on prioritized substances

Few studies were identified that analyzed the presence of the prioritized substances in either FCM or leachates thereof, regardless of their REACH registration status. Even less studies were identified that detected the presence of these substances in actual foodstuffs from the market. The lack of quantifiable data can, to some extent, be a consequence of detection and quantification limits (LOD and LOQ) that are not sensitive enough. Indeed, the LOD/Q of the methods used (UHPLC, LC–HRMS, LC-ESI-MS/MS, and GC–MS/MS) between the studies differed considerably, which can partly be explained by differing quantification methods and material analyzed (see e.g. Sanchis et al. 2015; Noureddine El Moussawi et al. 2019; Kovačič et al. 2020). In addition, the reason few studies included the prioritized compounds may be related to the absence of suitable detection methods within laboratories. This highlights the need for sensitive and standardized detection (multi-)methods, especially when an exposure assessment is warranted based on the potential hazard of the identified substances. The lack of literature on migration, especially from PC plastic or epoxy resins, may also indicate a low awareness of the potential use of these substances as drop-in alternatives for BPA. However, it should also be acknowledged that substitution of BPA has proven to be highly challenging. Polyester-based or acrylic-based coatings are being developed as promising alternatives to BPA-based coatings, yet it is challenging to meet the performance of BPA and not be more costly to produce (Banna et al. 2011; Geueke 2016; Beniah et al. 2020). An epoxy resin coating should be flexible and adhesive enough, it must be resistant to high temperatures during the sterilization processes and withstand corrosion resulting from contact with a large variety of food types to ensure long-term shelf life (Beniah et al. 2020). It is therefore not unlikely that only few substances will be used as drop-in substitute combined with the increased use of alternative materials for food contact purposes. Alternatively, it is also a realistic possibility that the use of BPA will be replaced by numerous alternative substances, each for a specific application. Especially the latter scenario is of interest because (1) these substances are likely not produced in high volumes like BPA and, therefore, the hazard information requirements according to the REACH regulation are low, (2) targeted monitoring will become more challenging because exposure will be highly source-dependent, and (3) mixture toxicity will become of higher concern (Docea et al. 2021).

Note that the prioritized substances cannot be legally used in the European Member States for use in plastic FCMs as long as they are not listed in the Union List of Regulation 10/2011. The same applies to substances used in coatings, which cannot be legally used in the Netherlands because these they are not included in the positive list for coating substances in the Dutch Packaging and Consumer Articles Regulation.

Stakeholder consultation and recommendations

Based on an inquiry with stakeholders (for a full summary see Appendix C), we learned that BPA is still used to a very large extent for the production of epoxy coatings, although BPA based coatings are slowly being replaced (yet it was not disclosed by which drop-in substitutes) and activity toward the transition to non BPA-based coatings is ongoing.

With the foreseen phasing out of BPA and other bisphenols for use in FCM within the EU, the development/innovation of BPA alternatives or alternative materials/technologies (e.g. polymer coated steel and polyesters and acrylics based on authorized monomers) will be accelerated. It is, therefore, recommended to monitor the use of PC plastic and epoxy resins. Monitoring of these substances will inform on the emergence of BPA alternatives and alternative materials which is especially relevant if that will be the primary route of BPA substitution in FCM and will make drop-in substitutes less relevant for regulatory purposes.

Conclusions

This study presents a prioritization strategy to identify substances that can possibly be used as an alternative to BPA. This is of great importance to avoid regrettable substitution of BPA in FCM by alternatives with similar or unknown hazardous properties, as currently very limited information is disclosed on the actual use of BPA alternatives. To identify potential regrettable substitution, information on hazard profiles and migrating potential is required. After reviewing the available data on hazard and migration potential of the 10 prioritized alternative substances, however, we conclude that there is not enough data for each of the substances to be able to conclude that the substance does not carry a possible health risk when used in high tonnage such as BPA. Major data gaps regarding the hazard profiles of the prioritized substances exist, although the scarce available data give some indications on the possible hazard for some of the substances (like BPTMC, DHBP, and TCBPA). In addition, very little is known about the actual use and exposure to these substances, although the few studies on migration suggest that these substances are currently not extensively used in FCMs. This work can inform food safety authorities on targeted monitoring of possible BPA replacements. It is recommended to monitor the use of PC plastic and epoxy resins and emerging alternative materials, which is especially relevant if the latter will be the primary route of BPA substitution in FCM and will make drop-in substitutes less relevant for regulatory purposes.

Abbreviations
2,2′-BPA	=	2,2′-bisphenol A
2,2′-BPF	=	2,2′‐bisphenol F
2,4′-BPF	=	2,4′-bisphenol F
BC/LA	=	bulbocavernosus/levator ani muscle
BP-6	=	benzophenone-6
BPA	=	bisphenol A
BPTMC	=	bisphenol TMC
CMR	=	carcinogenic, mutagenic, and/or reprotoxic
CMP	=	Chemicals Management Plan
DHBP	=	4,4-dihydroxybenzophenone
DHPE	=	4,4′-dihydroxydiphenyl ether
DNEL	=	derived no effect level
Dpf	=	days post fertilization
ECHA	=	European Chemicals Agency
ED	=	endocrine disruption
EFSA	=	European Food Safety Authority
EtOH	=	ethanol
FCM	=	food contact material
HBGV	=	health-based guidance value
Hpf	=	hours post fertilization
LH	=	luteinizing hormone
LOD	=	limit of detection
LOQ	=	limit of quantification
NOAEL	=	no observed adverse effect level
PARC	=	Partnership on the Assessment of Risks from Chemicals
PBT	=	persistent, bioaccumulative, and toxic
PC	=	polycarbonate
REACH	=	Registration, Evaluation, Authorization, and Restriction of Chemicals
ROS	=	reactive oxygen species
SVHC	=	substance of very high concern
USA	=	United States of America
TCBPA	=	tetrachlorobisphenol A
TDI	=	tolerable daily intake
t-TDI	=	temporary tolerable daily intake
vPvB	=	very persistent and very bioaccumulative
VTG	=	vitellogenin

Acknowledgements

The authors gratefully acknowledge the preparatory work of Joantine van Esterik (RIVM), the help of data information specialist Rob van Spronsen (RIVM), the input from the participants in the stakeholder consultation and the valuable review of this manuscript by prof. Dr. Aldert Piersma (RIVM) and the anonymous external reviewers that were selected by the Editor of this journal.

Declaration of interest

No potential conflict of interest was reported by the author(s). This work was supported by the Netherlands Food and Consumer Product Safety Authority (NVWA BuRO - RBT 090469/23).

Supplemental material

Supplemental material for this article is available online here.

Appendix A

Below we provide the search string that was used to obtain the scientific literature from the Embase database. The goal of this search strategy was to identify additional drop-in alternatives to BPA to collate the inventory of substances that are considered BPA alternatives.

Search string Embase

Table

No.	Query	Results
#43	#41 OR #42	309
#42	(#30 OR #32 OR #33 OR #34 OR #39) AND [01-07-2018]/sd	296
#41	(#30 OR #32 OR #33 OR #34 OR #39) AND [2018-2022]/py	279
#40	#30 OR #32 OR #33 OR #34 OR #39	476
#39	#36 AND #37 AND #38	43
#38	'bisphenol a':ti OR bpa:ti OR 'bisphenol-a':ti	7505
#37	analog*:ti OR alternative*:ti OR substitu*:ti OR derivat*:ti	432,478
#36	#20 NOT #35	541
#35	#30 OR #32 OR #33 OR #34	433
#34	(#11 OR #12 OR #13 OR #14 OR #15 OR #16 OR #17 OR #18 OR #19) AND ([cochrane review]/lim OR [systematic review]/lim OR [meta analysis]/lim)
#33	#20 AND 'review'/it	64
#32	#1 AND #31	47
#31	'bisphenol derivative'/exp/mj	79
#30	#20 AND #29	367
#29	#21 OR #22 OR #23 OR #24 OR #25 OR #26 OR #27 OR #28	417
#28	(bisphenol* NEAR/4 derivat*):ti	80
#27	(bpa NEAR/4 derivat*):ti	5
#26	(bisphenol* NEAR/4 substitu*):ti	28
#25	(bpa NEAR/4 substitu*):ti	8
#24	(bisphenol* NEAR/4 alternative*):ti	45
#23	(bpa NEAR/4 alternative*):ti	8
#22	(bisphenol* NEAR/4 analog*):ti	234
#21	(bpa NEAR/4 analog*):ti	24
#20	#11 OR #12 OR #13 OR #14 OR #15 OR #16 OR #17 OR #18 OR #19	974
#19	#1 AND #10	139
#18	#1 AND #9	123
#17	#1 AND #8	103
#16	#1 AND #7	151
#15	#1 AND #6	120
#14	#1 AND #5	156
#13	#1 AND #4	391
#12	#1 AND #3	322
#11	#1 AND #2	50
#10	(bisphenol* NEAR/4 derivat*):ti,ab	200
#9	(bpa NEAR/4 derivat*):ti,ab	149
#8	(bisphenol* NEAR/4 substitu*):ti,ab	167
#7	(bpa NEAR/4 substitu*):ti,ab	218
#6	(bisphenol* NEAR/4 alternative*):ti,ab	164
#5	(bpa NEAR/4 alternative*):ti,ab	222
#4	(bisphenol* NEAR/4 analog*):ti,ab	472
#3	(bpa NEAR/4 analog*):ti,ab	411
#2	'bisphenol derivative'/exp	102
#1	'4,4′ isopropylidenediphenol'/exp OR 'bisphenol a':ti OR bpa:ti OR 'bisphenol-a':ti	14,106

Appendix B

Below we provide the search string that was used to obtain the scientific literature from the Embase database. The goal of this search strategy was to identify hazard information on the prioritized alternatives to BPA published in scientific literature.

Table

No.	Query	Results
#74	#62 OR #63 OR #64 OR #65 OR #66 OR #67 OR #68 OR #70 OR #71 OR #72 OR #73	136
#73	#61 AND (#3 OR #4)	17
#72	#56 AND (#3 OR #4)	0
#71	#52 AND (#3 OR #4)	4
#70	#48 AND (#1 OR #2)	55
#69	#48 AND (#3 OR #4)	70
#68	#43 AND (#3 OR #4)	1
#67	#38 AND (#3 OR #4)	23
#66	#33 AND (#3 OR #4)	23
#65	#30 AND (#3 OR #4)	4
#64	#22 AND (#3 OR #4)	1
#63	#15 AND (#3 OR #4)	5
#62	#9 AND (#3 OR #4)	7
#61	#57 OR #58 OR #59 OR #60	44
#60	'2,2 hydroxy 4 methoxybenzoyl 5*'	0
#59	'bis 2 hydroxy 4 methoxyphenyl methan*' OR '2,2-dihydroxy-4,4-dimethoxybenzophenone*'	21
#58	'benzophenone 6'	23
#57	'131 54 4':rn	9
#56	#53 OR #54 OR #55	0
#55	'2,2 isopropylidenebis p*' OR 'bisphenol a bis(2-hydroxyethyl)ether*'	0
#54	'bpa 2 eo'	0
#53	'901 44 0':rn	0
#52	#49 OR #50 OR #51	13
#51	'4,4 1methylethylidene*'	4
#50	'3,3 dichlorobisphenol*' OR '3,3 dichloro bisphenol*'	9
#49	'79 98 1':rn	0
#48	#44 OR #45 OR #46 OR #47	139
#47	'2,6 dichloro 4 2 3,5 dichloro*'	1
#46	'2,6 dichloro 4 2 3 5 dichloro*'	1
#45	'tetrachlorobisphenol a‘/exp OR 'tetrachlorobisphenol a‘ OR 'tcbpa'	139
#44	'79 95 8':rn	0
#43	#39 OR #40 OR #41 OR #42	1
#42	'2,2 bis 2 hydroxyphenylpropane*'	0
#41	'2,2 bis 2 hydroxyphenyl propane*'	1
#40	'2,2 bisphenol a*' OR '2,2-bis(2-hydroxyphenyl)propane*'	1
#39	'7559 72 0':rn	0
#38	#34 OR #35 OR #36 OR #37	85
#37	'bis 4,2,3 dihydroxypropoxy*'	8
#36	'bfdge*'	64
#35	'bisphenol f diglycidyl ether'/exp OR 'bisphenol f diglycidyl ether*' OR 'bis[4-(2,3-dihydroxypropoxy)phenyl]methane*'	67
#34	'72406 26 9':rn	0
#33	#31 OR #32	70
#32	'4,4 dihydroxybenzophenone*'	70
#31	'611 99 4':rn	0
#30	#23 OR #24 OR #25 OR #26 OR #27 OR #28 OR #29	11
#29	'4 hydroxyphenyl methyl phenol*'	5
#28	'4 hydroxyphenylmethyl phenol*'	0
#27	'2,4 hydroxyphenylmethyl phenol*'	0
#26	'2,4 hydroxyphenyl methyl phenol*' OR '2,4-dihydroxydiphenylmethane*'	1
#25	'2,4 bpf'	3
#24	'2,4 bisphenol f‘	2
#23	'2467 03 0':rn	0
#22	#16 OR #17 OR #18 OR #19 OR #20 OR #21	10
#21	'2,2 bpf'	3
#20	'2,2 bisphenol f‘	2
#19	'bisphenol f ortho'	0
#18	'2,2 methylenebisphenol*'	3
#17	'2,2 methylene bisphenol*' OR '2,2-methylenediphenol*'	3
#16	'2467 02 9':rn	0
#15	#10 OR #11 OR #12 OR #13 OR #14	18
#14	'4,4 oxydiphenol*'	5
#13	'p,p oxybisphenol*'	2
#12	'dhdpe'	1
#11	'4,4 dihydroxydiphenyl ether*'	10
#10	'1965 09 9':rn	0
#9	#5 OR #6 OR #7 OR #8	12
#8	'trimethyl-cyclohexyliden*' NEAR/2 'bisphenol*'	0
#7	'trimethylcyclohexyliden*' NEAR/2 'bisphenol*'	0
#6	'bisphenol tmc' OR 'bp tmc' OR 'phenol 4,4-(3,3,5-trimethylcyclohexylidene)bis-'	12
#5	'129188 99 4':rn	0
#4	'carcinogenicity'/exp OR 'carcinogen'/exp OR 'carcinogenesis'/exp OR 'carcinogenic activity'/exp OR carcinogen*:ti,ab	538021
#3	'reproductive toxicity'/exp OR reprotox*:ti,ab OR 'reproduct* tox*':ti,ab OR 'endocrine disruptor'/exp OR 'endocrin* disrup*':ti,ab OR 'toxicity'/exp OR 'toxicology'/exp OR 'drug toxicity'/lnk	1319342
#2	'carcinogenicity'/exp/mj OR 'carcinogen'/exp/mj OR 'carcinogenesis'/exp/mj OR 'carcinogenic activity'/exp/mj OR carcinogen*:ti	169436
#1	'reproductive toxicity'/exp/mj OR reprotox*:ti OR 'reproduct* tox*':ti OR 'endocrine disruptor'/exp/mj OR 'endocrin* disrup*':ti OR 'toxicity'/exp/mj OR 'toxicology'/exp/mj OR 'drug toxicity'/lnk	880981

Appendix C

Stakeholder consultation methods

We reached out to national (Dutch) trade and sector organizations to compare the theoretical identification of potential BPA alternatives described in this work and the actual substances that are used by industry. Trade and sector organizations were identified by conducting a web search and by personal referrals. Questions on the current use of BPA and alternatives were asked per e-mail and questionnaire and oral clarifications were requested. Three responses on the questionnaire were received, four additional respondents replied by e-mail only and from two additional respondents oral information was obtained. The following questions were asked in the questionnaire:

1. Is, according to your knowledge, BPA still used for FCM and to what extent?

2. Are there already alternative substances being used to replace BPA in FCM?

3. Which foodstuffs are mainly packed in polycarbonate or epoxy coated cans?

4. Would a possible full ban of BPA in FCM pose a challenge to pack these foodstuffs in the future?

Stakeholder consultation results

To compare the current theoretical identification of potential BPA alternatives based on our criteria to the actual substances that are used by industry, we reached out to national (Dutch) trade and sector organizations. Based on this inquiry with stakeholders, we learned that BPA is still used to a very large extent for the production of epoxy coatings, although BPA based coatings are slowly being replaced (also by alternative technologies) and activity toward the transition to non BPA-based coatings is ongoing. It appears that there are alternative substances for BPA available for use in can coatings for most foodstuffs and beverages (e.g. canned soda, soup, and fish), although higher costs, compromises on performance/shelf life or ongoing qualification/validation processes hamper them from being vastly used in industry. Yet, challenges remain for BPA-free heavy duty coatings as technologies cannot easily be transferred to another substance.

It was not disclosed which drop-in substitutes are used, as coating specifics are intellectual property of industry. The recent approval of tetramethyl bisphenol F-diglyceryl ether (TMBPF-DGE) by the Dutch Committee for safety assessment of food contact materials (CBVV) and authorization in the Netherlands (Dutch Ministry of Health Welfare and Sport 2022) can be expected to lead to a commercialization of this alternative and consequently its increased use. Polyesters and acrylics based on authorized monomers are also being used as alternative to BPA and polymer coated steel has been suggested as an adequate replacement. It is, however, not possible to predict which BPA alternatives that are possibly being used may lead to regrettable substitution.

Although BPA (and other bisphenols) are proposed to be banned in FCMs, it is expected that traces of BPA will continue to be found in the (production of) coatings for a period of time due to the ubiquitous use of BPA. An absence of threshold for enforcement, or a too low detection limit, may result in delays for substituting BPA-based coatings. It is important that analytical methods are in place that ensure a sufficient level of detection/quantification from an enforcement point of view but also for research and development purposes. Also, it is important that new EU regulations will also apply to imported goods and is sufficiently enforced, as it is clear that the use of BPA is not regulated the same way globally.