Safety of Plastic Food Packaging: The Challenges about Non-Intentionally Added Substances (NIAS) Discovery, Identification and Risk Assessment

Several food contact materials (FCMs) contain non-intentionally added substances (NIAS), and most of the substances that migrate from plastic food packaging are unknown. This review aimed to situate the main challenges involving unknown NIAS in plastic food packaging in terms of identification, migration tests, prediction, sample preparation, determination methods and risk assessment trials. Most studies have identified NIAS in plastic materials as polyurethane adhesives (PU), polyethylene terephthalate (PET), polyester coatings, polypropylene materials (PP), multilayers materials, plastic films, polyvinyl chloride (PVC), recycled materials, high-density polyethylene (HDPE) and low-density polyethylene (LDPE). Degradation products are almost the primary source of NIAS in plastic FCMs, most from antioxidants as Irganox 1010 and Irgafos 168, following by oligomers and side reaction products. The NIAS assessment in plastics FCMs is usually made by migration tests under worst-case conditions using food simulants. For predicted NIAS, targeted analytical methods are applied using GC-MS based methods for volatile NIAS and GC-MS and LC-MS based methods for semi- and non-volatile NIAS; non-targeted methods to analyze unknown NIAS in plastic FCMs are applied using GC and LC techniques combined with QTOF mass spectrometry (HRMS). In terms of NIAS risk assessment and prioritization, the threshold of toxicological concern (TTC) concept is the most applied tool for risk assessment. Bioassays with sensitive analytical techniques seem to be an efficient method to identify NIAS and their hazard to human exposure; the combination of genotoxicity testing with analytical chemistry could allow the Cramer class III TTC application to prioritize unknown NIAS. The scientific justification for implementing a molecular weight-based cut-off (<1000 Da) in the risk assessment of FCMs should be reevaluated. Although official guides and opinions are being issued on the subject, the whole chain’s alignment is needed, and more specific legislation on the steps to follow to get along with NIAS.


Introduction
Food packaging can contain non-intentionally added substances (NIAS) as a result of the interactions between different substances in the packaging materials, between food • the contact surface size, since the more extensive the contact surface between packaging and food, the higher the migration rate [24]; • the nature of the migrant, since the more volatile and the lower the molecular weight, the greater the migration rate (also the vapor pressure, water solubility, octanol solubility and polarity) [44]; • FCM material type (e.g., impermeable, permeable, porous materials) [45]; • kinetics and thermodynamics of the migration process (how fast will a substance transfer from the FCM into food) [46,47]; • the nature of the food, since high-fat foods are reported with a high level of migration, for example [42,45]; • temperature, high temperatures demonstrate high migration rates [42,44]; • contact during serving and duration of the time of contact [42,45]. In the specific case of plastics, non-inert materials, the migration can occur from the inner side of the packaging and internal layers due to diffusion processes [18,48,49]. According to the deterministic model, the migration of substances from packaging to food  In the specific case of plastics, non-inert materials, the migration can occur from the inner side of the packaging and internal layers due to diffusion processes [18,48,49]. According to the deterministic model, the migration of substances from packaging to food occurs through diffusion, where there is a movement of molecular structures from the high concentration to the low concentration region until equilibrium is reached. Fick's second law can describe the diffusion rate [10,[50][51][52][53]. Some studies describe precisely how to use this mathematical model [42,47,53]. In addition to diffusion, there is also desorption of diffuse polymer surface molecules, sorption of compounds at the plastic-food interface and desorption of compounds in the food that describe the migration stages [24].
Migration modelling can complement or substitute actual laboratory migration experiments that are expensive and time-consuming. Thus, it has become a popular tool for researchers, industry and regulators to predict migration. Efforts have been made to modify the standard "worst-case" deterministic models to reflect more realistic migration scenarios better. These approaches comprehend probabilistic and stochastic modelling that considers variability and uncertainty in the mass transfer parameters, mechanistic or empirical migration models that can be combined with food consumption data to estimate dietary exposure to the chemicals [17,[54][55][56]. However, trained specialists should only use migration models with in-depth knowledge of chemical migration to make reliable predictions and correctly interpret the results. Gavriil et al. [55] published an extensive review to evaluate the migration models' efficiencies, concluding that these models cannot predict the total accurate migration from a different FCM to the food volume since the material might contain several completely unknown compounds. Thus, the migration studies are still the traditional method for evaluating the migration of substances from packaging to food, despite analytical challenges and cost.
Two terms of high significance that should not be confused are the overall migration (OM) and the specific migration (SM). The OM refers to the sum of all mobile substances of packaging released per unit area of the package under the influence of specific predetermined conditions. On the other hand, SM is only related to a specific known substance [57].
In most cases, and as provided by legislation ((EU) n • 10/2011), the migration study is performed with food simulants in place of the food itself, due to the high complexity that food has typically as a matrix than foods [29,58,59]. They should represent the significant physicochemical properties exhibited by each food type. When using food simulants, temperature and standard testing time should replicate the migration from the FCM into the food [42,60]. Moreover, current legislation provides for the simulation of temperature and contact period situations in which migration tests performed according to the actual situation's food storage conditions.
In the US, according to The Guidance for Industry: Preparation of Premarket Submissions for Food Contact Substances (Chemistry Recommendations) published by FDA in 2007, sponsors should provide information sufficient to permit estimation of the daily dietary consumption of the food contact substances (FCS), e.g., consumer exposure. The FDA will calculate the FCS consumption or other components consumption that might migrate to food expected daily based on analyzed or estimated levels in food or food simulants. Still following the guidance, "Although the FDA always has accepted reliable analyses of FCS in real foods, many analytes are difficult to measure in food in practice.
As an alternative, sponsors may submit migration data acquired with food simulants to reproduce the FCS's nature and migration into food. Because an FCS can contact many foods with different processing conditions and shelf lives, the submitted migration data should reflect the most severe temperature/time conditions to which the food contact article (FCA) containing the FCS will be exposed. The document brings a table with the respective recommended simulant according to the type of food. The recommendation for aqueous and acidic foods is 10% ethanol; low-and high-alcoholic foods, 10-50% ethanol; fatty food, food oil (e.g., corn oil), HB307, Miglyol 812 or others.
Two steps are involved in migration tests of substances from food packaging into food simulants: to expose the polymer packaging to the food simulant; to quantify the migrants Polymers 2021, 13, 2077 6 of 43 transferred to a food simulant in terms of OM or SM. Determining OM is a regulatory requirement in EU countries with established migration limits for FCMs substances [24].
Most of the substances that migrate from packaging to food are unknown. A recent Scientific Opinion issued by the European Food Standard Authorization (EFSA) in 2016 comments that regarding identifying and evaluating all migrating substances, experience has shown that it is necessary to focus more on finished materials and articles used during the manufacturing process. Substances used in the manufacture of FCM or objects may contain impurities originating in their manufacture. Besides, during manufacture and use, reaction and degradation products may be formed, of which oligomers may be of the dominant class. These substances have become known as NIAS and referred to as such in the Commission Regulations. Regardless of whether they are intentional or unintended, all migrant substances' safety, not just initiating substances-e.g., monomers or additives only-needs to be assessed, and the guidelines should be updated to account for this broader approach fully. This change to the finished FCM and its use requires an adjustment of the current substance listing system. The present work will discuss the definition, different aspects and types of NIAS, analytical methods of determination and quantification and the next section's risk assessment task.

NIAS in Plastic Food Packaging
Inventory lists of plastic FCMs contain several thousands of chemicals, including starting substances like monomers, production aids and additives [19,20,61]. However, these chemicals often suffer transformations, and final FCAs and FCMs may contain novel compounds that can migrate into food [19,61,62]. Figure 2 introduces the scientific terms for food contact articles (FCA), which are the combination of diverse food contact materials (FCMs) and food contact chemicals (FCCs) defined as substances used or present in the manufacture of FCMs or existing in the final FCMs or FCAs. Some FCCs are starting substances that no longer exist in the final FCM/FCA. Some FCCs are generated during the manufacture of an FCM/FCA or during the application of high temperatures, irradiation or the consumer's food packaging. FCCs comprehend intentionally added substances (IAS), such as monomers, additives, catalysts and production aids; and impurities and reaction products (oligomers, polymers, by-products and degradation products), be referred to as non-intentionally added substances (NIAS) [7].
Polymers 2021, 13, x 7 of 47 reaction products (oligomers, polymers, by-products and degradation products), be referred to as non-intentionally added substances (NIAS) [7]. The FDA classifies Food Contact Substance (FCS) as "any substance that is intended for use as a component of materials used in manufacturing, packing, packaging, transporting or holding food if such use of the substance is not intended to have any technical effect in such food". FCS could be "a single substance, such as a polymer or an antioxidant in a polymer. As a substance, it is reasonably pure". Although a polymer may be composed of diverse monomers, it has a well-defined composition. FCM is "made with the FCS and (usually) other substances. It is often (but not necessarily) a mixture, such as an antioxidant in a polymer. The composition may be variable". It is impossible to know all substances that may be formed by the degradation of additives or plastics' polymerization [1,19]. Their presence is often not known by the consumer or by the manufacturer [5,19]. If, according to Regulation EC 1935/2004 on materials and articles intended to come into contact with food, the manufacturer must ensure the safety of food contact packaging, it is clear that it is necessary to know about the NIAS formed and also to assess the risk assessment (RA) of these new substances. Still, follow- The FDA classifies Food Contact Substance (FCS) as "any substance that is intended for use as a component of materials used in manufacturing, packing, packaging, transporting Polymers 2021, 13, 2077 7 of 43 or holding food if such use of the substance is not intended to have any technical effect in such food". FCS could be "a single substance, such as a polymer or an antioxidant in a polymer. As a substance, it is reasonably pure". Although a polymer may be composed of diverse monomers, it has a well-defined composition. FCM is "made with the FCS and (usually) other substances. It is often (but not necessarily) a mixture, such as an antioxidant in a polymer. The composition may be variable".
It is impossible to know all substances that may be formed by the degradation of additives or plastics' polymerization [1,19]. Their presence is often not known by the consumer or by the manufacturer [5,19]. If, according to Regulation EC 1935/2004 on materials and articles intended to come into contact with food, the manufacturer must ensure the safety of food contact packaging, it is clear that it is necessary to know about the NIAS formed and also to assess the risk assessment (RA) of these new substances. Still, following EC (EU) n • 10/2011, unauthorized substances may be used in FCMs plastics behind a functional barrier, providing they do not migrate to the food at levels above 10 µg/kg. Thus, in Europe, NIAS could be within this approach [5].
In the United States, according to the FDA Code Federal Regulation Title 21 Sec. 170.39 on the threshold of regulation for substances used in food-contact articles, substances that may enter the diet at levels below 50 µg/kg and demonstrate that they are non-genotoxic, are exempt from need authorization indirect food additive [11]. The FDA uses the term "Indirect Food Additive", a food additive coming into contact with food as part of the packaging or processing but is not intended to be added directly to, become a constituent or have a technical outcome in/on the food. Indirect food additives cited in Title 21 of the US Code of Federal Regulations (21CFR) used in food-contact articles comprise adhesives and components of coatings (Part 175), paper and paperboard (Part 176), polymers (Part 177) and adjuvants and production aids (Part 178).
In this sense, the scientific community has been striving to define NIAS identification and determination strategies in FCMs and FCAs in general and strategies to predict the migration rate of these compounds from packaging to food and evaluate the health risks of each identified NIAS. A literature search was made using Medical Subject Headings (MeSH) terms on the Pubmed, Web of Science and Scopus databases. The initial screening process was performed from August to September 2020. Further directed searches were carried out by checking the reference lists of relevant articles. In this context, we defined a strategy composed of four phases to build the search strings: (i) identification of keywords considering the research question; (ii) synonyms based on relevant studies about nonintentionally added substances and plastic food packaging; and (iii) use of "AND" and "OR" Booleans operators.

Oligomers
Monomers' reactions may cause oligomer formation during the polymerization processes due to incomplete polymerization, side reaction products like cyclic oligomers and subsequent thermal or hydrolytic degradation [105]. "Oligomers" are defined here as substances consisting of a small number (<20) of repeating units. Oligomers can also inten-

Oligomers
Monomers' reactions may cause oligomer formation during the polymerization processes due to incomplete polymerization, side reaction products like cyclic oligomers and subsequent thermal or hydrolytic degradation [105]. "Oligomers" are defined here as substances consisting of a small number (<20) of repeating units. Oligomers can also intentionally be used as "prepolymers". These are reactive species that will be utilized as reacting blocks to manufacture polymers. When oligomers are not intentionally added, they are well within the scope of the NIAS definition, according to [7]. Oligomers do not come just from conventional polymers but also biodegradable polymers, and biopolymers also have a series of oligomers that can migrate to the food [5]. Despite this, according to other authors [105], the classification of oligomers as NIAS is still controversial due to similar properties of a homolog series of oligomers. Because the oligomer chemistry is linked to a polymer manufacturing process, it makes much sense to classify oligomers as polymer specific substances and not as NIAS. They are not explicitly regulated in the Regulation (EU) 10/2011. The extension of the EU positive list with new co-monomers as a consequence of industrial developments gives rise to an exponentially growing number of new oligomers present in food contact polymers as potential migrants.
However, according to a Scientific Opinion of EFSA (2016) [8], oligomers have become NIAS. Whether their presence is intended or not, it is needed to evaluate all migrating substances' safety and not just the starting substances-e.g., the monomers or additives alone-and the guidelines should be updated. Importantly, oligomers are considered part of polymers in the US, with no impurities. Regulation (EU) No. 10/2011 does not explicitly regulate oligomers' levels in general. However, two EFSA opinions concerning new co-monomers for polyester food-contact materials specified 50 µg kg −1 for total oligomer migration with less than 1000 Da [106,107]. The European Food Safety Authority (EFSA) states that the PA oligomers should not exceed 5 ppm (mg kg −1 food). Table 2 listed scientific publications that determined oligomers in different types of plastic food packaging. Most of the FCMs studied are polyester coatings and resins [29,63,64,108] and polyethylene terephthalate materials [28,65,66]. Polyethylene terephthalate (PET) Cyclic oligomers LC-MS [28] Virgin and recycled Polyethylene terephthalate-PET pellets

By-Products Compounds or Side Products
Side reactions occur during manufacturing the starting substances, materials, additives and the consumer's food packaging use, forming various novel products. Further, the close contact between the single materials can also lead to unwanted reaction products, often unknown [67,68]. An example of neo-formed NIAS is a primary aromatic amine (PAAs) in polyurethane (PU) adhesives formed by polyols and diisocyanate monomers' polymerization [109]. If the adhesive has not adequately cured or the ingredients have not adequately mixed, the polymerization reaction is not efficient enough, and the remaining nonpolymerized aromatic isocyanates can produce PAAs in contact with water [69]. PU can often generate other by-products compounds, the cyclic adipate 1,4,7-trioxacyclotridecane-8,13-dione, likely to derive from a chemical reaction from the interaction between two common ingredients in the adhesive formula [1]. Table 3 listed scientific publications that determined by-products in different types of plastic food packaging. Degradation of active compounds; impurities from the raw materials; additives used in the manufacture of the active polymer (citral thermal reaction products; oxidation product of citral; decomposition product of adipates used as plasticizers; impurity/reaction product/breakdown product for the additives used in the manufacture of PE materials; xanthenone derivates) UPLC-QTOF-MS [86] Polyurethane adhesives in multilayer packaging materials By-product of the polyester-based urethane 91,6-Dioxacyclododecane-7,12-dione); cyclic adipate; unknown nitrogen-compounds; unknown phenolic compounds; 1,4,7-Trioxacyclotridecane-8,13-dione HS-SPME-GC-MS [67] Polyethylene terephthalate (PET),

Breakdown Products or Degradation Products
During the polymers' manufacturer, processes like high temperatures or irradiation can cause the material's degradation, such as UV irradiation, gamma irradiation and microwave treatment [70]. Other factors like light exposure, oxidative damage or chemicals, in general, contribute to polymer degradation, causing changes in their physicochemical properties [110]. The same way is applied for the additives, e.g., antioxidants and light stabilizers since the purpose of these substances is to act as a scavenger when oxidizing conditions such as high temperatures or microwave exposure occur, and also exposure to UV light, and their mode of action is often to be oxidized earlier than the material [70,71]. Breakdown products are almost the primary source of NIAS in the FCMs. The degradation of additives and polymers leads to novel low molecular weight substances that can migrate to food [1]. Degradation products can be further divided into the degradation of polymers and additives' degradation [9].

Impurities in the Raw Materials
Raw materials and the additives may contain impurities, some of them known by the manufacturers, but minor impurities often are unknown and can persist in the final FCMs, consequently migrating to the foodstuffs [69,111]. Ink raw materials can be unique chemical substances or mixtures of many chemical substances. They comprise one or several main substances with a specific function in ink (IAS); besides these main components, the raw materials may contain other substances that will not have any specific function in ink, e.g., monomers or residues of catalysts, solvents or defoamers. Such substances are necessary to produce the raw material, but they are not needed in the printing inks. Nevertheless, such impurities are usually known due to the raw materials production process and should be specified in the supply chain to allow a risk assessment, as they should be considered "known NIAS" in the inks. However, it is not always possible to list and consider all impurities during the authorization.
An example is primary aromatic amines (PAAs) and β-naphthol in azo-pigments made for printing inks. Both substances can be present as impurities in the pigment and the final ink formulation. The azo-pigment itself is an IAS used to formulate the ink, but PAAs and β-naphthol or β-naphthol-derivates are NIAS. Another example is impurities from acrylic adhesive additives in migration tests of multilayer materials [112], Table 5 listed scientific publications that determined impurities in different types of plastic food packaging. PP/EVOH/PP; PP/EVOH + oregano Degradation of active compounds; impurities from the raw materials; additives used in the manufacture of the active polymer (citral thermal reaction products; oxidation product of citral; decomposition product of adipates used as plasticizers; impurity/reaction product/breakdown product for the additives used in the manufacture of PE materials; xanthenone derivates) UPLC-QTOF-MS [86] Polyethylene terephthalate-PET film with an acrylic resin Ethyl lauroyl arginate (LAE) impurities: N2-Dodecanoyl-l-arginine (LAS) UPLC-MS(QTOF) [87] Polypropylene (PP) Degradation products derived from phenolic antioxidants, impurity/reaction product/breakdown product of the additives, Family 1: Family formed with the reference structure; was formed by the compounds that had a similar structure constituted by a group 3,5-di-tert-butyl-4-hydroxyphenyl; Family 2: With glycerol molecule (glyceryl monostearate, glyceryl palmitate and glyceryl dihexadecanoate, an ester of an acid chain bonded to a glycerol molecule); Family 3: Dihydroxy alquilamines (amine bonded to two ethanol molecules and also an alkyl hydrocarbon chain); Family 4: ceramide and dihydroceramide (a family of waxy lipid molecules which are composed of sphingosine (an 18 carbon amino alcohol with an unsaturated hydrocarbon chain) and a fatty acid); Family 5: amides bonded by ethylene (degradation products from a lubricant losing C 2 H 4 ); Other compounds (amides come from the impurities or degradation products from erucamide and oleamide widely used as slip agents)

Contaminants
There are multiple sources of contaminant residues in the supply chains, including presses (fountain solutions, lubricants and additives), substrates (plasticizers, surfactants, stabilizers, antioxidants, resins), inks and coatings (resins, polymers, adhesives, pigments, solvents, monomers, additives) and the environment (pesticides, cleaners, fumes) [12]. Toxic elements such as As, Cd, Hg, Pb are trace elements and environmental contaminants present in the raw materials and considered NIAS in the final FCA. When recycled materials are utilized for food packaging, often undefined mixtures of chemicals present during recycling can react and form additional substances that increase the list of potential NIAS. Furthermore, the accumulation of chemicals might occur when materials are recycled many times. Thus, the prediction, identification and control of NIAS in recycled materials are challenging because of their difficulty tracing their origin [1,31]. Packaging materials produced from recycled polyethylene terephthalate are applied for direct food contact in recycled rigid containers and films. Most recycled polyethylene terephthalate packaging materials contain metal catalysts, the most common antimony [31]. In other cases, contaminants have been associated mainly with printing inks, paperboards, adhesives or substances added to color the final material [77,79] reported 134 substances, including volatile and semi-volatile compounds, in recycled LDPE and HDPE from domestic waste. The main groups were contamination from external sources, additives and additives breakdown products. Table 6 listed a few scientific publications that determined contaminants in different types of plastic food packaging, most of them recycled materials. Recycled pellets obtained from post-consumer low-density polyethylene (PC-LDPE) and high-density polyethylene (PC-HPDE) Contaminants from external sources: methyl lactate, hexyl acetate, and dimethyl butanedioate, α-methylionone, 3-(4-Isopropylphenyl)-2-methylpropionaldehyde, α-amylcinnamaldehyde, (phenylmethylene)octanal and dipropylene glycol amog others (cosmetic ingredients); alkylbenzenes (breakdown products produced by the degradation of alkylbenzene sulfonates); contamination related to food: the lactones, 5-methylfurfural, furfural and methyl hexanoate (can derive from food flavors as well as from cosmetics ingredients); furfuryl alcohol, methyl pyruvate and 2-acetyl pyridine (food flavors), methyl-2-ethylhexanoate, acetic acid, propanoic acid, pyridine and dimethyl trisulfide (rotten food products), 2,6-diisopropylnaphthalene (paper labels).

The Challenges of the NIAS Assessment
Substances with a molecular weight of more than 1000 Da are not usually considered in NIAS determination, while non-listed compounds <1000 Da are considered NIAS since substances with a higher molecular weight are not absorbed in the metabolism. An exception is applied for fluorochemicals, for which a cut-off of 1500 Da has been suggested [8], since these substances tend to have smaller molecular volumes at the same molecular mass, enabling resorption and render them potentially toxicologically relevant. However, substances migrating from FCMs may be of considerably high Mw. They may include intentionally added polymeric additives, such as polyadipates [113], and NIAS, such as oligomers, which can be formed due to an incomplete polymerization result in a later degradation of the base polymer [105]. A molecular size cut-off is typically applied in chemical risk assessment of FCMs in Europe and the US: chemicals larger than 1000 Da is not assessed. [114] published a review that pointed to the existence of FCCs with an increased intestinal permeability; this may lead to the uptake of the compound of not only low (<1000 Da) but also high (>1000 Da) molecular weight. The authors discuss the toxicological relevance of high molecular weight compounds in the gut, suggesting that the scientific justification for implementing a molecular weight-based cut-off in the risk assessment of FCMs should be reevaluated. Figure 4 reports the steps detailed by Koster et al. [7] about NIAS information collection and identification in FCMs. Figure 5 shows the usual steps involved in the NIAS assessment. According to [7], since the polymer can be a mixture of different substances, detailed description and knowledge of the starting substances and processes enormously facilitates NIAS analysis in the final product. An adequate characterization of the FCM is of utmost importance for defining what type of NIAS may be expected and which analytical techniques should be used to determine migration. This includes a complete characterization of the IAS present in the FCM, the manufacturing processes, the temperature of exposure during the production, etc. sessment of FCMs should be reevaluated. Figure 4 reports the steps detailed by Koster et al. [7] about NIAS information collection and identification in FCMs. Figure 5 shows the usual steps involved in the NIAS assessment. According to [7], since the polymer can be a mixture of different substances, detailed description and knowledge of the starting substances and processes enormously facilitates NIAS analysis in the final product. An adequate characterization of the FCM is of utmost importance for defining what type of NIAS may be expected and which analytical techniques should be used to determine migration. This includes a complete characterization of the IAS present in the FCM, the manufacturing processes, the temperature of exposure during the production, etc. After that, the next step is to identify the predicted NIAS from the IAS knowledge present in the FCM, e.g., reaction/breakdown products formed from the IAS and base materials. Moreover, the guidance strongly recommends the literature survey to predict After that, the next step is to identify the predicted NIAS from the IAS knowledge present in the FCM, e.g., reaction/breakdown products formed from the IAS and base materials. Moreover, the guidance strongly recommends the literature survey to predict NIAS's formation in each FCM and communicate with subject matter laboratories. For predicted NIAS (known NIAS), targeted analytical methods are applied, detailed in the next section.

NIAS Identification: How to Start?
Finally, the identification of unpredicted NIAS formed in the final FCM with potential migration to the foodstuffs was not predicted based on the system's chemistry considerations. For example, these NIAS can be contaminants and reaction products present in products made by complex chemistry such as coatings, rubbers and adhesives. Nontargeted screening analytical approaches should be used to assess these unpredicted NIAS, detailed in the next section [7].
EFSA published in 2016 a public scientific opinion about recent developments in the risk assessment of chemicals in food and their prospective impact on the safety assessment of substances used in FCMs [8]. As the EFSA document is the most recent European official publication on the subject, emphasis will be given to this approach. The main points addressed by the document are listed below: • although information about the substance's identity as used and its impurities is necessary, more focus on the migration potential from finished materials and articles is needed. Substances used to manufacture FCM may disappear, and it may be mainly reaction products that turn up in the migrates; • it is necessary to describe the chemical and the physical properties of a substance that are the determinants for its potential to persist or react in the final FCM and food and migration. The needed information includes: the volatility and thermal/chemical stability of the substances used as well as their impurities; the octanol-water partition coefficient (log Po/w) and the solubility of the migrating substances in solvents of different polarity and food and food simulants; their stability in food simulants and food hydrolysis in the gastrointestinal tract; possible chemical interactions with the food, leading to the generation of reaction products with or from the food; • information about the level of use, the function of the substance and the manufacturing process conditions are needed for assessing the quantities, types, and nature of potentially migrating substances. Depending on the details of information accessible, such as the nature of the plastics produced using the substance, the food characteristic the plastic materials intended to contact, and whether the FCM is intended for single or repeat use applications, a more or less refined exposure estimate may be derived. NIAS's formation in each FCM and communicate with subject matter laboratories. For predicted NIAS (known NIAS), targeted analytical methods are applied, detailed in the next section. Finally, the identification of unpredicted NIAS formed in the final FCM with potential migration to the foodstuffs was not predicted based on the system's chemistry considerations. For example, these NIAS can be contaminants and reaction products present in products made by complex chemistry such as coatings, rubbers and adhesives. Non-targeted screening analytical approaches should be used to assess these unpredicted NIAS, detailed in the next section [7]. EFSA published in 2016 a public scientific opinion about recent developments in the risk assessment of chemicals in food and their prospective impact on the safety assessment of substances used in FCMs [8]. As the EFSA document is the most recent European official publication on the subject, emphasis will be given to this approach. The main points addressed by the document are listed below: • although information about the substance's identity as used and its impurities is necessary, more focus on the migration potential from finished materials and articles is needed. Substances used to manufacture FCM may disappear, and it may be mainly reaction products that turn up in the migrates; • it is necessary to describe the chemical and the physical properties of a substance that are the determinants for its potential to persist or react in the final FCM and food and migration. The needed information includes: the volatility and thermal/chemical stability of the substances used as well as their impurities; the octanol-water partition coefficient (log Po/w) and the solubility of the migrating substances in solvents of different polarity and food and food simulants; their stability in food simulants and food hydrolysis in the gastrointestinal tract; possible chemical interactions with the

NIAS Extraction
Currently, there is no standard protocol about how to proceed with the NIAS identification/quantification. Firstly, it is necessary to decide if the analysis will be made on the packaging material itself, on the food simulant during migration tests or direct in the food in contact with the packaging [1,115]. The most recent scientific opinion published by EFSA (2016) [8] brings three main approaches: modelling, migration and direct food measurements. In the first case, the FCM could be analyzed to identify/quantify the NIAS, and mathematical models can estimate the total mass transferred from the FCM to the foodstuff, which should be validated. The document EUR 24,514 EN 2010 (applicability of generally recognized diffusion models for estimating specific migration in support of EU Directive 2002/72/E.C.) must be consulted in this kind of approach.
In the simulation of migration, according to the recommendation of EU n • 10/2011, some food simulants (see Table 7) could be placed in contact with the FCM and posteri-orly analyzed instead of the real food, with temperature conditions and contact period predetermined in the current legislation. According to the scientific opinion by EFSA (2016) [8], the EC guidelines on migration testing are under development. Finally, direct measurements in foods are needed when the simulation is impossible or not reliable. This type of approach has recently been the most used in NIAS identification/migration studies by the scientific community in different plastics FCMs [6,80,81,116]. Food simulants such as Tenax [32,68,72,73,76,83,117], water or organic solvents [25,28,31,63,69,74,[83][84][85][86][87][88]108] have been used to simulate the migration of NIAS from plastic food packaging. Table 1 identifies the simulants or foodstuffs used in several studies on NIAS assessment. Some authors reported migration studies direct in food matrices, e.g., pizza [89], sliced fresh chicken breasts [87], baby food: a mixture of fruit purées (apple, banana, pear), fruit jelly, chocolate custard [66]. It is also essential to define temperature and exposure time to simulate real conditions [69]. Another challenge is sample preparation. As chromatographic techniques are the most commonly used in NIAS identification studies, any chromatographic analysis of solid samples needs an extraction or migration step that transfers as many compounds as possible into the liquid or gaseous phase or representative for what may migrate into food [84]. Organic molecules are extracted from polymeric materials by applying a solvent in which they are soluble. Different extraction techniques for organic compounds in polymers, including NIAS, have been reported in the literature, such as liquid-liquid extraction (LLE) [81,118], Solid Phase Extraction (SPE) [30,69], dispersive liquid-liquid microextraction (DLLME) [119], ultrasonic extraction [10,90,120], headspace solid-phase microextraction (HS-SPME) [73,79,91,121], solid-phase microextraction (SPME) [67,91] fabric phase sorptive extraction (FPSE) [92], accelerated solvent extraction (ASE) [74] and QuECh-ERS [115]. When the determination is performed in a solution, most analytical procedures involve concentration steps before the instrumental analysis [1]. Sanchis et al. [115] recently published an extensive review of relevant conventional and novelty analytical procedures for the three matrices (simulants, FCM or food) proposed in recent literature. Nerin et al. [1] cite the possibility of direct analysis in polymers by using ASAP (atmospheric solid analy-sis probe), DART (direct analysis in real-time) [93,108] or DESI (desorption electrospray ionization) [117], which can be done directly using high-resolution mass spectrometry (HRMS) instruments [64,68,116,122]. These methods do not require extraction steps and do not separate the analytes. Therefore, it is a quick technique but should only analyze well-known substances due to the complicated fragmentation patterns usually obtained.
For the analysis of predicted or unpredicted NIAS, two strategies must be applied: targeted analytical methods for predicting NIAS or non-targeted or screening methods to analyze substances with a wide range of physical/chemical properties [9,94]. All analysis strategies should detect and quantify the amount of NIAS present in the FCM. This is possible for predicted and known NIAS but difficult for unpredicted NIAS since reference standards may not be available. As a practical standard, the migration level of 10 µg/kg food for NIAS is applied, as this is the level from which each migrated substance must be identified [9]. This level has been specified in the Plastics Regulation (EU) No 10/2011 for migration through a functional barrier: unauthorized but intentionally added substances may be used in FCM plastics behind a functional barrier provided they do not migrate at levels above 10 µg/kg food; substances that are carcinogenic, mutagenic or toxic for reproduction (CMR) may not be used. The threshold of 10 µg/kg is a practical limit and not based on current toxicological understanding.
Usually, NIAS's identification requires susceptible advanced analytical techniques and databases and software tools [5,9]. The general workflow for the analysis of FCMs involves the data acquisition after the selection of one or a combination of analytical techniques, such as gas chromatography (GC) for volatile and semi-volatile compounds with low polarity [78,95,[123][124][125] and liquid chromatography (LC) for thermally unstable and nonvolatile compounds; both coupled with mass spectrometry detectors (MS) [28,96,126].
The GC-MS enables mass spectral libraries for the identification (target methods) for volatile and semi-volatile compounds. For a qualitative analysis of unknown samples, GC-MS is operated in a scan mode to obtain the entire mass spectrum within the predefined peak identification mass range [4,32]. When the identification is not possible, high-resolution tandem quadrupole and time-of-flight mass analyzers (Q-TOF-MS) help identify sum formulas of unknowns (non-target methods) [33,66,69,77,82,85,86,93,127]. For non-volatile compounds, because of the absence of extensive libraries for mass spectra, the use of a Q-TOF-MS will allow accurate mass determinations of the precursor ion and the productions, providing information about fragmentation patterns. This represents excellent structural information and assures the correct identification of unknown compounds [1]. Figure 6 illustrates a decision tree for each analytical procedure to choose from NIAS identification/quantification published by [1]. Table 1 listed the most techniques utilized to determine unknown compounds in plastic food packaging, including NIAS.

Target Analysis for Predict NIAS
An internal standard should be used for this evaluation: the same or structurally very similar (isotope-labelled) compared to the NIAS under investigation. This ensures that the detector response for the internal standard and NIAS is the same or very similar. Migrates may be obtained using migration conditions simulating the intended use of the FCM. One or more internal standards should be added at a level in the range of the NIAS's regular migration [7].
GC-MS is the most frequently used technique for volatile, and semi-volatile knew organic compounds since the commercially available mass spectra library (e.g., NIST library) could help identify the compounds [32,78,80,123]. The libraries contain the MS spectra obtained by electronic impact and a quadrupole mass analyzer. An ion trap allows further fragmentation to a select mass fragment, which could better identify the compound. If the identification is not possible, the use of high-resolution mass spectrometry (HRMS) techniques such as time of flight (TOF) or Orbitrap instruments provides accurate mass measurements and full-scan spectra that will help [4,116]. The HRMS detectors can be accoupled to both LC and GC systems [1]. the use of a Q-TOF-MS will allow accurate mass determinations of the precursor ion and the productions, providing information about fragmentation patterns. This represents ex cellent structural information and assures the correct identification of unknown com pounds [1]. Figure 6 illustrates a decision tree for each analytical procedure to choose from NIAS identification/quantification published by [1]. Table 1 listed the most techniques uti lized to determine unknown compounds in plastic food packaging, including NIAS.

Target Analysis for Predict NIAS
An internal standard should be used for this evaluation: the same or structurally very similar (isotope-labelled) compared to the NIAS under investigation. This ensures that the detector response for the internal standard and NIAS is the same or very similar. Migrates may be obtained using migration conditions simulating the intended use of the FCM. One or more internal standards should be added at a level in the range of the NIAS's regular migration [7].
GC-MS is the most frequently used technique for volatile, and semi-volatile knew organic compounds since the commercially available mass spectra library (e.g., NIST li brary) could help identify the compounds [32,78,80,123]. The libraries contain the MS spectra obtained by electronic impact and a quadrupole mass analyzer. An ion trap allows For non-volatile compounds, the previous separation in liquid chromatography of the compounds present in the sample is essential for proper identification. Even though many detectors can be coupled to LC, most of them, such as UV, DAD [29], fluorescence or IR, do not provide enough information for NIAS identification, and they are more frequently used for confirmation purposes. Liquid chromatography (LC) coupled to mass spectrometry (MS) is probably the most powerful technique for analyzing non-volatile NIAS [60,63,64]. Because of the absence of extensive libraries for mass spectra acquired by LC-MS techniques, the identification process is more complicated and time-consuming than for mass spectra obtained by GC-MS [128].
As in GC, different mass analyzers can be used in LC-MS. These include quadrupole (Q) and ion trap (IT), or high resolution (HR) mass analyzers such as time of flight (TOF) and Orbitrap [65,66,97,98]. The quadrupole is frequently used for quantitative purposes due to its high sensitivity and selectivity. However, its identification capabilities are deficient due to its reduced sensitivity in full scan mode and mass accuracy. The main benefit of HRMS techniques is the possibility of collecting full scan spectra with very accurate mass measurements, which allows the analyst to perform structural elucidations of unknown or suspected compounds [1].

Non-Target Analysis for Unpredicted NIAS
Screening analysis can be conducted for unpredicted NIAS and detects predicted NIAS and IAS [9]. In this approach, an FCM or a starting substance(s) is extracted with one or more simulant/extraction solvents followed by analysis using several analytical methods to provide maximum coverage for all substances possible, e.g., headspace/solid-phase microextraction (SPME) [34], gas chromatography flame ionization detection (GC-FID) or gas chromatography-mass spectrometry (GC-MS) to detect volatile substances [77,91]; GC-FID or GC-MS for semi-volatile substances [77,129]; liquid chromatography ultraviolet detection (LC-UV) or LC high-resolution MS for non-volatile and polar compounds [25,28,60,78]; inductively coupled plasma (ICP-MS) for trace elements [99]; and nuclear magnetic resonance (NMR) for general screening [63,118]. High-resolution GC-MS may also be useful for the analysis of NIAS [7]. NIAS with a molecular weight exceeding 1000 Da is generally not considered relevant for risk assessment. This cut-off does not cover fluorinated substances (the cut-off for these substances is 1500 Da due to the smaller molecular diameter [8]. The emerging scientific consensus points towards the potential significance of substances above 1000 kDa for long-term chronic health effects [114]. Recently, a review published by Martínez-Bueno et al. [4] presents an overview of current analytical approaches based on HRMS analysis to identify unknown migrants (volatile and non-volatile compounds) plastic food packaging materials. The conclusions are that despite the complexity of identifying unknown migrants from plastic FCMs, recent advances in mass spectrometry techniques based on accurate mass measurements (HRMS) combined with improvements in hardware and software performance have contributed to facilitating this analytical process. According to the study, regarding separation techniques, a clear trend towards applying HRMS in LC for non-targeted analysis has been observed during the last years, being ESI (in positive mode) the most widely used ionization source. In the case of GC analysis, E. and APCI have been equally applied in non-targeted approaches. Two types of high-resolution mass analyzers have been used to identify non-targeted compounds released from plastic FCMs: quadrupole time-offlight (Q-TOF) [25,33,63,66,69,71,72,82,[85][86][87]93,96,98,100,127] and quadrupole-Orbitrap (Q-Orbitrap) [84,95,96,101]. To date, the Q-TOF detector has been used more often than Q-Orbitrap in the plastic FCMs field since the use of hybrid instruments provides more information about fragmentation patterns than their simple versions (TOF and Orbitrap). This represents additional structural information and ensures the correct identification of unknown compounds. Moreover, Orbitrap technology is more expensive than Q-TOF, leading to the increased application of the last one [4].
The main challenge of a non-targeted approach is the high number of signals generated by the HRMS instrument, making it necessary to group the number of relevant peaks. The authors suggested different data reduction techniques in non-target analysis, as a casecontrol comparison, intensity threshold signal to noise threshold, mass range restrictions, and homemade databases. The most used analytical strategy has been the case-control comparison, or comparing the chromatograms from control or blanks and sample material solutions as pre-processing data treatment. However, combining more than one approach was often necessary to reduce the search space's size to manageable levels. Identification of unknown migrants by GC(EI) HRMS is usually easier than LC-HRMS due to using large mass spectral databases that support the identification [4].
Wang et al. [10] studied the identification of chemicals in a commercial PVC/PE film using a UPLC-QTOF/MS method. With this robust technique, many chemicals were identified, and some recognized NIAS from diverse sources. Six additives were selected, and their migration behaviors were simultaneously assessed. The authors concluded that investigation of migration behaviors of chemicals based on the identification results obtained by the HRMS technique could be a more practical approach for the safety evaluation of commercial packaging materials. They also mentioned the necessity of quantifying chemicals (including NIAS), whose standard substances are not available. Solving this problem requires a proper selection of reference standards that can represent the target chemicals. Another aspect is identifying chemicals in some consumable materials (syringe, syringe filter, plastic tube, etc.) used for sample pretreatment. These materials may contribute to additives also present in packaging materials, resulting in blank interference.

The NIAS Risk Assessment Challenge
The traditional approach to verify the safety of chemicals in food is to perform a specific risk assessment (RA) for each chemical. There are often data available from prior research or accredited methods for the commonly investigated IAS, but there is rarely relevant data for the more elusive NIAS. Most NIAS do not have assigned chemical structures, concentration data or characterization of hazards, and few methods can obtain these data for a large group of chemicals [130]. In the EU, FCCs that are subject to authorization include IAS (e.g., starting substances and additives used in the production of food contact plastics), and in the US, these comprehend "indirect food additives", which are substances that come into contact with food and they are transferred into food but are not intended to be added to food [11]. In both the EU and the US, the specific requirements for toxicological testing of FCCs requiring authorization depend on estimated consumer exposures, which FCM manufacturers determine before marketing [11,21].
Generally, risk assessment is considered a science-based process consisting of four steps: hazard identification, dose-response assessment, exposure assessment, and risk assessment. Figure 7 shows the traditional approach to perform a specific risk assessment of food chemicals [130]. Hazard typically refers to the intrinsic properties, such as toxicity, while exposure addresses the likelihood to which a human or environmental receptor will be exposed to the intrinsic hazards. Risk is the likelihood of harm occurring. Captured into a formula, this would be: hazard x exposure potential = risk.

Strategies to Hazard Characterisation of NIAS
Hazard identification aims to recognize the potential adverse health effects mans associated with exposure to a chemical. Hazard identification requires an ade and documented review of relevant scientific information obtained from appropria tabases, peer-reviewed literature or study reports, if available. This approach emph studies in the following order: human epidemiological and safety studies, animal t logical studies, in-vitro bioassays and read-across and (quantitative) structure-activ lationships ((Q)SAR) [7].
According to the classical approach, any NIAS should go through a toxicol evaluation requiring the same toxicity data as IAS. Toxicity data of substances m collected from existing scientific information and complemented by further in vi vitro or in silico tests. However, this concept is only applicable to identified NIAS. with a known chemical structure and no toxicological data, in silico tools, may pr qualitative and/or quantitative hazard information. For example, quantitative stru activity relationships (QSAR) allow the quantitative prediction of toxicological end [80].

Bioassays
To complement the classical approach of detecting, identifying and assessing a

Strategies to Hazard Characterisation of NIAS
Hazard identification aims to recognize the potential adverse health effects in humans associated with exposure to a chemical. Hazard identification requires an adequate and documented review of relevant scientific information obtained from appropriate databases, peer-reviewed literature or study reports, if available. This approach emphasizes studies in the following order: human epidemiological and safety studies, animal toxicological studies, in-vitro bioassays and read-across and (quantitative) structure-activity relationships ((Q)SAR) [7].
According to the classical approach, any NIAS should go through a toxicological evaluation requiring the same toxicity data as IAS. Toxicity data of substances may be collected from existing scientific information and complemented by further in vivo, in vitro or in silico tests. However, this concept is only applicable to identified NIAS. NIAS with a known chemical structure and no toxicological data, in silico tools, may provide qualitative and/or quantitative hazard information. For example, quantitative structure-activity relationships (QSAR) allow the quantitative prediction of toxicological endpoints [80].

Bioassays
To complement the classical approach of detecting, identifying and assessing a single NIAS, the overall migration or extract of an FCM can be tested through in vitro bioassays [25]. This approach's interest has increased in recent years since the European regulation on chemicals (REACH) recommends that the RA of chemicals priorities in silico and in vitro methods [131]. Numerous bioassays allowing the detection of a range of bioactivities are available. Most are based on different cell types and use various biological detection principles such as transcriptional activation and cell proliferation [132]. Such tests may help detect the cumulative effects of chemical mixtures for toxicological endpoints sensitive to mixture toxicity. Extracts or migrates generating positive bioassays' responses may subsequently be fractionated and re-analyzed to identify the active substances. In vitro bioassays offer a robust and economical solution to screen the toxicity of FCMs and FCAs and assess the overall migrates' hazards. The results of in vitro tests could be used to highlight those FCMs or components of FCAs that are critical in terms of toxicological hazard, thus initiating a process for substituting chemicals of concern with more benign compounds. Alternatively, in vitro tests could be used to priorities samples for further testing to evaluate their potential to cause toxicity in vivo and provide input data for a whole risk assessment process [133]. However, the variability of available assays and sample preparation protocols request further optimization and standardization before bioassays can be used routinely [131] beyond the method's complete validation. The approach results in many peaks that require identification that can be time-consuming and expensive to perform. Also, it is not comprehensible what the predictability in terms of the safety of individual peaks determined by the screening is [134].
Veyrand et al. [135] describe an initial attempt to investigate bioassays' potential role in detecting toxicologically relevant molecules in migration simulation studies. For that purpose, the authors used as a case study a plastic cup containing the antioxidant TNPP well documented to degrade into isomers of 4NP. This chemical is known to interfere with the endocrine system, acting through activation and inhibition of the estrogen and the androgen receptors. The authors conclude that, together with analytics, bioassays contribute to identify toxicologically relevant molecules leaching from FCM and enable improved risk assessment. Ubeda et al. [25] investigated two cyclic esters' migration from multilayer packaging material based on PU adhesive and evaluated their bioaccessibility to the body using digestion assays (in vitro). The potential formation of new compounds during gastrointestinal digestion was also evaluated. The authors observed that most samples' oligomers' migration values exceeded 10 ng g−1; however, bioaccessibility studies showed decreased oligomers after digestions.

Threshold of Toxicological Concern (TTC)
The Threshold of Toxicological Concern (TTC) concept ascribes human exposure thresholds to substances with unknown toxicity and known structure [38,100]. That may be applied to evaluate materials for their potential toxicity when exposure is shallow. In the absence of toxicological data, the TTC concept is a practical risk assessment tool that sets up human exposure levels to substances considered to have no appreciable risk to human health. The TTC is currently accepted for use in the EU and the USA [4]. An Expert Group of the International Life Sciences Institute (ILSI Europe) has examined the TTC principle for its broader applicability in food safety evaluation, and they have concluded that this approach can be applied for low concentrations of chemicals in the food that lack toxicity data, provided that there is an intake estimate [7]. The authors used the TTC approach to estimate NIAS Risk Assessment in recent literature [116,[136][137][138].
According to its molecular structure, the TTC principle is based on the Cramer rules, which estimate compounds' theoretical toxicity. There are three classes of toxicity according to these rules: low (class I), moderate (class II) and high (class III). Cramer recommends a maximum intake value for each toxicity group; values for class I, II and III are 30 µg/kg bw per day, 9 µg/kg bw per day and 1.5 µg/kg bw per day, respectively [38]. The software Toxtree developed by the JRC commissioned is an open-source application that can estimate toxic hazards by applying a decision tree approach following Cramer rules (Ref http: //toxtree.sourceforge.net/. Accessed on 25 May 2021). This software can be used to estimate the theoretical toxicity of the identified compounds [4].
In 2011, Koster et al. [139] proposed a TTC approach to regulating unknown substances found in food samples, including NIAS, but NIAS's confident identification and, particularly, genotoxic substances remain an unresolved issue [139]. Genotoxicity is a term referring to the capability of chemicals to damage genetic material. Many molecular mechanisms resulting in genotoxicity have been described, such as direct covalent binding to DNA, DNA cross-linking, DNA breakage DNA intercalation, oxidative stress and inhibition topoisomerases and interference with DNA repair [132]. Figure 8 shows the stepwise approach suggested for NIAS's evaluation [139]. In 2012 the European Food Safety Authority (EFSA) published a scientific opinion on investigating options for providing advice about potential human health risks based on the concept of Threshold of Toxicological Concern (TTC). In 2019, EFSA published Guidance on using the TTC approach in food safety assessment. This Guidance provides clear In 2012 the European Food Safety Authority (EFSA) published a scientific opinion on investigating options for providing advice about potential human health risks based on the concept of Threshold of Toxicological Concern (TTC). In 2019, EFSA published Guidance on using the TTC approach in food safety assessment. This Guidance provides clear step-by-step instructions for using the TTC approach, which can be used when the substance's chemical structure is known, there are limited chemical-specific toxicity data, and the exposure can be estimated. According to the document, the TTC approach should not be applied for substances for which EU food/feed legislation claims the submission of toxicity data or when enough data are accessible for a RA or if the substance under consideration consists into one of the exclusion categories. For substances that can potentially be DNA-reactive mutagens or carcinogens based on the weight of evidence, the TTC value is 0.0025 µg/kg body weight (bw) per day. For organophosphates or carbamates, the TTC value is 0.3 µg/kg bw per day. All other substances are classified according to the Cramer classification.
In 2014, Koster et al. [134] proposed the Complex Mixture Safety Assessment Strategy (CoMSAS) to assess unknown NIAS's safety in cartoon food contact material. The basis of CoMSAS is the decision tree-based approach of the application of the TTC concept [134]. This strategy is unnecessary to identify the unknown substances present in the migration extract below 90 µg/person/day, corresponding to the Cramer class III TTC threshold; substances with a structural alert for genotoxicity should be excluded a TTC threshold of 0.15 µg/day applies. The strategy respects all decisions introduced in the original TTC concept [134,140]. Several exclusion steps are introduced to exclude highly toxic substances and other substances that are excluded from applying the TTC concept [4]. In 2018, Pieke et al. [130] suggested applying quantitative structure-activity relationships (QSAR) models for carcinogenicity, mutagenicity, and reproductive toxicity as a strategy for risk prioritization based on tentative data. Four risk assessors assess a selection of 60 chemical compounds from two FCMs to classify compounds based on potential risk. For almost 60% of cases, the assessors classified compounds as either high or low priority. Unclassified compounds are due to the lack of agreement between experts or a perceived lack of data. In the high priority group, substances are high-concentration compounds, benzophenone derivatives and dyes. The low priority compounds contained, e.g., oligomers from plasticizers and linear alkane amides. The classification provides valuable information based on tentative data and prioritizes discovered chemical compounds for pending risk assessment.
In 2019, Schilter et al. [132] published an extensive review about applying the TTC to prioritize unidentified chemicals in food contact materials. The authors concluded that for a specific migrate, the evidence for the absence of mutagenicity based on the Ames test (genotoxicity testing) and analytical chemistry and information on packaging manufacture could allow the application of the Cramer class III TTC to prioritize unknown NIAS.
The Food Contact Additives guidelines (2016) on Risk Assessment of NLS and NIAS mentions the toxicological assessment, with the proposition to identify the adverse toxicological effects that a substance could cause and to define the critical dose of a substance in the daily diet, below which the substance is not expected to pose a risk to human health (dose-response assessment or hazard characterization). The critical dietary exposure level is referred to as the Tolerable Daily Intake (TDI), generally applied for substances detected in food but not intentionally added, or the Acceptable Daily Intake (ADI) for substances intentionally added to food, expressed in mg/person/day or mg/kg body weight/day. The TDI concept assumes that a clear dose-response relationship with a threshold exists, whereas the Threshold determines the point of exposure below which no adverse effect is observable. A TDI or ADI presupposes a complete in vivo study. In Europe, The Joint FAO/WHO Expert Committee on Food Additives-JECFA uses PTWI, or provisional tolerable daily intake, for contaminants accumulated in the body. The weekly designation stresses the importance of limiting intake over a while for such substances. When using this approach, no-observed-effect levels (NOELs) or no-observed-adverse-effect levels (NOAELs) are identified in the critical studies, to which appropriate safety or uncertainty factors are applied [141].
It is essential to know that two options can be considered to set a migration level above which a safety assessment of NIAS is needed: detection above the detection limit of the analytical method or detection above a level corresponding to a safe exposure threshold, the so-called exposure-based approach. In the first option, 10 µg/kg food is the conventional European limit of detection. This value has no relation to a health-derived threshold but was introduced in the Regulation (EU) n • 10/2011 as a typical detection limit of the analytical techniques used (EU, 10/2011). This 10 µg/kg threshold, although widely used for the NIAS assessment, could represent a challenge with some FCM as they may release many NIAS exceeding this limit that is sometimes difficult or not possible to identify; in the second option, a level of migration corresponding to a safe exposure threshold (based on substance-specific data or in-silico tools) could be derived. This level is also called the Level of Interest (LOI) [7]. In the US, industry and law regulatory use a similar approach, working as consultants and performing substance evaluations for food contact applications. Due to this approach, an analytical sensitivity level of 50 µg/kg, a finding of "non-detected" is found to be reasonable. In conclusion, if it can be shown that the migration of a substance, which is not CMR classified (CMR Substance: a substance listed as carcinogenic, mutagenic or toxic to reproduction category 1A, 1B or 2 in CLP Regulation 1272/2008 Annex VI Table 3.1) and shows no genotoxic structural alerts, is less than 0.01-0.05 mg/kg food (simulant), then no further assessment is needed. However, these conservative approaches have minimal applicability due to the low values (EU: 10 µg/kg, US: 50 µg/kg).

Alternatives to Plastics: Biopolymers and Bioplastics
There are two main types of biopolymers, also called renewable polymers: those that come from living organisms and those from renewable resources that need to be polymerized [142]. Biopolymers are not new. In the 1850s, a British chemist created plastics from bio-cellulose [143]. Nevertheless, in the last two decades, the gradual replacement of synthetic polymers with biopolymer counterparts has gained significant interest, considering the ever-increasing public interest in the eco-friendly use of sustainable and safe materials [144]. With a production volume of 2.11 million tons in 2018, their market share is meagre (1% of all plastics) but expected to increase [145].
A bioplastic is obtained from a biopolymer formed in a biological system such as polyhydroxyalkanoates (PHA's), starch, cellulose and lignin. The biopolymers are, under most general conditions, also biodegradable. Bioplastics should be distinguished from synthetic (manufactured) bio-based plastics made from monomers originating from biological (once-living) systems, and bio-based polymers are not always biodegradable. Bioplastics are produced from a range of natural resources, among which agricultural products such as corn, cassava, flax fibers and agricultural by-products such as rice straw [146].
Very little is known about the chemical safety of bioplastics/biopolymers. These gaps in our knowledge are problematic because human exposure to chemicals from bioplastics and plant-based materials will increase with their increasing application [145]. Zimmermann et al. [145] extracted 43 bio-based and biodegradable products as well as their precursors, covering predominantly food contact materials made of nine material types, and characterized these extracts using in vitro bioassays and non-target high-resolution mass spectrometry. A total of 67% of the materials induced baseline toxicity, 42% induced oxidative stress, 23% induced antiandrogenic and one material induced estrogenicity. The authors tentatively identified 343 priority compounds, including monomers, oligomers, plastic additives, lubricants and non-intentionally added substances (NIAS). A comparison with conventional plastics indicates that bioplastics and plant-based materials are similarly toxic.
Specific migration of volatile compounds has been studied in two types of dishes (wheat pulp and wood) [149]. Those identified compounds considered of interest, according to existing legislation, have been quantified. The results showed that the quantified compounds are well below the specific migration limits (SML) set by the legislation, thereby showing the safety of using this biodegradable dish. Third-seven different compounds were detected in pellets and films of PLA-polyester blend samples by [150]. The results showed that new compounds from the reaction of PLA components with food simulants were present in migration solutions.
A biodegradable antioxidant active food packaging based on antioxidants from medicinal and aromatic plants incorporated into a polylactic acid matrix was designed and developed by Gavril et al. [151]. Moreover, an extensive investigation of the influence of sage and lemon balm leaves on material safety and the type of migrants was performed using migration assays. It was shown that the addition of sage and lemon balm leaves into a polylactic acid structure decreased the migration of both linear and cyclic polylactic acid oligomers, currently not legislated by European Union. Besides, total absence or decrease of migration of volatile compounds were observed when using the active films.
New bamboo-based biopolymers were evaluated by Osorio et al. [152] to ensure consumers safety. Twelve non-volatile compounds were detected in migration solutions, mainly melamine and its derivatives, coming from polymer resins from the biopolymer. Even though some of the compounds found in migration came from bamboo, such as phytosterols, most migrants came from the melamine to improve the biopolymer properties. Not only melamine but several melamine derivatives were found in migration above the limits established in European legislation. Consequently, this material does not comply with the EU legislation. The material cannot be identified as bamboo but as melamine with bamboo filler. As melamine is neither a biopolymer nor biodegradable material, promoting these kitchenware materials as bamboo can be considered fraud to consumers.
Two different biopolymer samples based on Polylactic acid (PLA) and compounds' migration to food simulants were studied by Ubeda et al. [153]. Thirty-nine different PLA oligomers made of repeated monomer units of [LA] (C 3 H 4 O 2 ) and different structures were identified. They corresponded to cyclic oligomers with [LA]n structure and two groups of linear oligomers, one with a hydroxyl group and the other one with an ethoxy group. Cyclic oligomers only appeared in the material and were not present in migration solutions. Linear oligomers HO-[LA]n-H was already present in the pellets/film, and they migrated in a higher extension to aqueous food simulants. However, linear oligomers CH3-CH2-O-[LA]n-H was not present initially in the pellets/film but were detected in migration to simulants with ethanol content. Furthermore, 5 cyclic polyester oligomers were identified in migration.

Conclusions
Plastic packaging used as food items is becoming increasingly complex, and the growing use of polymers is a fact in the coming years. Concerning the non-intentionally added substances (NIAS) in polymers, there are evident challenges. This review evidenced that further studies and strategies are needed in many different contexts, such as the definition of NIAS, identification and determination, the prediction, the sample preparation and validation, determination methods and methods validation, migration studies and validation and especially risk assessment trials and validation, consequently, the setting of maximum limits for the risk substances. This is a multidisciplinary field of study. Moreover, in many countries, there is no current legislation regarding NIAS. Despite the efforts and work published, NIAS will still be a much-explored theme over the years and remains a significant challenge for the scientific community and industry since the lack of information, guides and definitions still need resolutions.
Furthermore, with the high complexity of plastic packaging, it is a great challenge to predict all chemical reactions and by-products formed in the packaging and eventually migrate to the food. Some high points identified in this work are: • it is estimated that most of the food contact material contain NIAS and most of the substances that migrate from plastic food packaging are unknown; • several studies determining NIAS were carried out in polyurethane adhesives (PU), polyethylene terephthalate (PET), polyester coatings, polypropylene materials (PP), multilayers materials, plastic films, polyvinyl chloride (PVC), recycled materials, high-density polyethylene (HDPE) and low-density polyethylene (LDPE); • breakdown products are almost the primary source of NIAS in plastic FCMs, most of the degradation products from antioxidants; following by side reaction products and oligomers; • substances used to manufacture FCM may disappear, and it may be mainly reaction products that turn up in the migrates; • the NIAS assessment in plastics FCMs is usually made by migration tests under worstcase conditions using food simulants (Tenax, organic solvents, water) and simulating the temperature and time of exposure; • targeted analytical methods for the analysis of predicted NIAS are applied using GC-MS based methods for volatile NIAS and GC-MS and LC-MS based methods for semi-and non-volatile NIAS; • non-targeted or screening methods to analyze unknown NIAS in plastic FCMs is mainly done using GC and LC techniques combined with QTOF mass spectrometry; • for all concepts, better information transfer through the whole value chain would mainly facilitate the identification of unknown compounds since the entire supply chain has responsibility; • NIAS could be present in the materials in deficient concentrations and still be a risk; • in terms of risk assessment and prioritization for NIAS, the threshold of toxicological concern (TTC) concept is the most applied tool, by comparing the semi-quantitative concentration of the chemical compound with the estimated exposure limit; • the combination of bioassays with sensitive analytical techniques seems to be an efficient way of identifying NIAS and their hazard to human exposure; mutagenicity based on the Ames test (genotoxicity testing), together with analytical chemistry and information on packaging manufacture, could allow the application of the Cramer class III TTC to prioritize unknown NIAS; • oligomers up to a molar weight of 1000 Da seems to be relevant under the assumption of human gastrointestinal absorption, and this cut-off should be re-evaluated by legislation; • currently, there is an absence of industry-wide harmonized methodology on dealing with NIAS; • for commonly investigated IAS, there is often risk assessment data available for prior research or via accredited methods, but for the more elusive NIAS, there is rarely relevant data; • there are only a few references outside Europe and the United States regarding NIAS in food contact materials, although other territories closely follow Europe and the United States legislation; • the guidelines need to be urgently updated; • biopolymers/bioplastics are currently being developed as an attempt to replace conventional plastics; however, few studies on the safety of these materials and migration of compounds to food have been carried out and published in the literature; • the migration of NIAS from biopolymers/bioplastics reported in some studies emphasize the need to evaluate these alternative materials for hazardous compounds and NIAS and establish legislation for these specific new materials.
NIAS is still a challenge for the regulatory system, food industry, and for the scientific community, being necessary upgrading of the current legislation that meets the requirements of consumers and food producers, with the efforts of science with the publication of reliable data and commitment of the entire food production chain.