Search

Some of the chemicals in materials used for packaging food may leak into the food, resulting in human exposure. These include so-called Non-intentionally Added Substances (NIAS), many of them being unidentified and toxicologically uncharacterized. This raises the question of how to address their safety. An approach consisting of identification and toxicologically testing all of them appears neither feasible nor necessary. Instead, it has been proposed to use the threshold of toxicological concern (TTC) Cramer class III to prioritise unknown NIAS on which further safety investigations should focus. Use of the Cramer class III TTC for this purpose would be appropriate if amongst others sufficient evidence were available that the unknown chemicals were not acetylcholinesterase inhibitors or direct DNA-reactive mutagens. While knowledge of the material and analytical chemistry may efficiently address the first concern, the second could not be addressed in this way. An alternative would be use of a bioassay capable of detecting DNA-reactive mutagens at very low levels. No fully satisfactory bioassay was identified. The Ames test appeared the most suitable since it specifically detects DNA-reactive mutagens and the limit of biological detection of highly potent genotoxic carcinogens is low. It is proposed that for a specific migrate, the evidence for absence of mutagenicity based on the Ames test, together with analytical chemistry and information on packaging manufacture could allow application of the Cramer class III TTC to prioritise unknown NIAS. Recommendations, as well as research proposals, have been developed on sample preparation and bioassay improvement with the ultimate aim of improving limits of biological detection of mutagens. Although research is still necessary, the proposed approach should bring significant benefits over the current practices used for safety evaluation of food contact materials.

This study investigated how the properties of a polymeric can coating film, such as thickness and crosslink density as well as the type of migrant, influence the migration kinetics of model migrants in an attempt to better understand, model and control the migration process. Four model migrants were used BADGE (bisphenol A diglycidyl ether), BADGE·H2O, cyclo-diBADGE and Uvitex OB, that differ in size and polarity. Fatty and aqueous food simulants were used at high temperatures (70-130°C). The apparent diffusion coefficients were found to decrease with increasing crosslink density, while they increased with increasing film thickness. The apparent activation energy of BADGE and BADGE-related compounds was calculated from the diffusion data and were high, in the range of 250-264kJmol-1. The polarity of the simulant and the polarity of the migrant were found to influence migration. The results can be used to improve existing migration models, and thereby help to reduce migration from packaging into food by using safety-by-design approaches in new product development.

The polymeric coating used in metal packaging such as cans for foods and beverages may contain residual amounts of monomers used in the production of the coating, as well as unreacted linear and cyclic oligomers. Traditionally, although designed for use with plastic food contact materials, food simulants have been used to determine the migration of monomers from coatings into foodstuffs. More recently, food simulants have also been used to determine oligomeric species migrating from can coatings. In the work reported here, the migration of both monomers and oligomers from polyester-based can coatings into food simulants and foodstuffs, some of which were towards the end of their shelf-life, is compared. The concentrations of monomers and selected oligomers in canned foods at the end of their shelf life were found to be significantly lower than those in food simulants, which in turn was lower than those in the extraction solvent acetonitrile.

Aspects of melamine migration from epoxy-based coatings into a food simulant were studied. Four commercial amino-based cross linkers were incorporated into an epoxy anhydride coating system and into an epoxy phenolic coating system. The epoxy-based coatings were formulated, applied, cured and tested for migration by retorting in contact with the food simulant, 10% ethanol, at 131°C. The commercial melamino-based cross linkers used and the model coatings that were prepared using these cross linkers contained very low or non-detectable levels of free melamine. However, during retorting, the migration of melamine from the coatings increased as the retorting time was increased. This migration process is not the more classical diffusional process but rather the result of chemical attack (hydrolysis) of the coating. For these model can coatings, a substantial fraction of the melamino cross linker was hydrolysed although, curiously, the essential functional properties of the coating are retained. In all cases, for these model systems the migration of melamine was rather low because the cross linkers are used commercially in only small amounts - typically 1-2% of the dry film weight of the coatings. For the standard retorting conditions of 1 h, migration of melamine was up to 0.4 mg kg(-1), depending on the cross linker used. The cross linker that contained the methylol functionality (-CH₂OH group) gave rise to less melamine than did the alkylated cross linkers (methylated and butylated, -OCHv and -O(CH₂)₃CH₃, respectively). This observation could prove useful in formulating coatings with even lower melamine release characteristics.

In this study, the suitability of four chemically different melamino cross-linkers for use in formulating epoxy coatings was investigated on the basis of the composition and of the tendency of cured coatings to hydrolyse to melamine and to formaldehyde when they were retorted in aqueous food simulants. The four cross-linkers were characterised for their composition identity, flow behaviour, thermal stability and for the presence of residual species. The different cross-linkers were used individually to cross-link selected epoxy coatings. The effects of the cross-linker chemistry, the curing conditions and the kinetics of the hydrolysis and subsequent migration processes, leading to melamine and formaldehyde were investigated following thermal treatments that were designed to represent the conditions of food sterilisation. The results show that each cross-linker type is different in its rheological characteristics, its solids content, its thermal behaviour and its physical properties. The chemistry of each cross-linker plays a major role in the manner in which the epoxy coatings undergo hydrolysis to release melamine and formaldehyde. The greatest migration of melamine (from an unpigmented epoxy anhydride coating, cured with the hexamethoxymethyl melamine cross-linker) into the 10% (v/v) aqueous ethanol food stimulant, after retorting at 131 °C, for 1 h was 525 μg/6 dm 2 . The greatest migration of formaldehyde into the simulant was also from this coating at 11 μg/6 dm 2 , when retorted at 131 °C for 1 h. The curing conditions affected the extent of the cross-linker hydrolysis. The influence of varying the curing time and the curing temperature was used to control the hydrolysis of the cross-linked, epoxy-based coatings. A decrease in the extent of cross-linker hydrolysis by 50–80% was achieved in all cases as the temperature of the curing was increased, in stages, from 160 °C to 200 °C.

Pigmented (TiO2), amino resin crosslinked coatings, designed for applications in can coatings’ internal lacquers, were formulated, characterized, applied, and cured. Three grades of a pigmentary form of TiO2 were characterized in terms of their particle size, their particle morphology, their zeta potential, and their moisture retention behavior. Epoxy coatings that were crosslinked using one of several, different amino crosslinkers were prepared. The effect of the presence of the TiO2 pigments on the hydrolysis of the cured coatings was monitored via the controlled retorting of the coatings. The different grades of TiO2 pigment were selected, to establish whether or not they could be used interchangeably with respect to hydrolysis and to melamine release. Also, the effects of the aging of the fluid coatings on the amount of melamine released from the coatings (after curing and retorting) were monitored. Storage under laboratory conditions for 2, 20, and 40 weeks was used for this purpose. The TiO2 pigment contributed significantly to the hydrolysis behavior of the epoxy coatings in that their presence substantially reduced the amount of melamine released and the extent of crosslinker hydrolysis. Typical results show that excluding the TiO2 pigment particles from the formulation results in there being 50% more hydrolysis of the crosslinker to melamine. With respect to the melamine release and crosslinker hydrolysis, the different grades of the pigment gave similar results.

The safety of a polyester-polyurethane can coating has been assessed using a suite of complementary analytical methods to identify and estimate the concentrations of potential chemical migrants. The polyester was based on phthalic acids and aliphatic diols. The polyisocyanate cross-linking agent was 1-isocyanato-3-isocyanatomethyl-3,5,5-trimethyl cyclohexane homopolymer (IPDI) blocked with methylethylketone oxime (MEKO) to make a one-part formulation. The overall migrate, obtained using solvent extraction of cured films, comprised almost completely of 12 cyclic and one linear polyester oligomer up to molecular weight 800 and containing up to six monomer units. These 13 oligomers covered a total of 28 isomeric forms. Other minor components detected were plasticisers and surfactants as well as impurities present in the starting materials. There was no detectable residue of either the blocked isocyanate (0.01 microg/dm(2)) used as the starting substance or the unblocked isocyanate (0.02 microg/dm(2)). The level of extractable IPDI was used as an indicator of the completeness of cure in experimental coatings. These studies revealed that there was an influence of time, temperature and catalyst content. Polymerisation was also influenced by the additives used and by the ageing of the wet coating formulation over several months. These studies allow parameters to be specified to ensure that commercial production coatings receive a full cure giving low migration characteristics.

Information is given concerning the curing characteristics of polyester–polyurethane (PEPU)-based food can coatings and the analysis of the free isocyanate by-products. The behavior of a methyl ethyl ketoxime (MEKO)-blocked isocyanate on heating was studied using thermal analytical techniques. Following this, the curing behavior of a PEPU-based can coating (PEPU01 that contained an aluminum flake pigment) and the MEKO-blocked isocyanate was investigated. A study of the unblocking of the MEKO-blocked isocyanate in the PEPU01 can coatings was undertaken with respect to the migration potential of any residual-free isocyanate compounds that could migrate from the cured PEPU can coatings to the surrounding medium (e.g., foodstuffs).

A methyl ethyl ketone oxime-blocked isocyanate was used as a crosslinking agent in the curing of polyester–polyurethane (PEPU) can coatings. The completeness of cure was assessed from the amount of residual (un-reacted) isophorone diisocyanate (IPDI) trimer that was presented in the coating after curing. This amount of residual trimer depended strongly on the temperature at which the curing was undertaken and time given for curing. The degree of cure was also assessed, correlating with the reduced IPDI trimer migration potential. The amount of residual IPDI trimer that was extractable from a typical PEPU coating rose markedly (15- to 20-fold) if the wet coating was aged up to 6 months before use. The presence of a TiO2 pigment increased the aging effect. An aluminum flake pigment had an even greater effect. The TiO2 pigment increased aging by different degrees according to the type of surface modification that had been applied to the pigment. The grade having the greatest alumina content, as a result of surface modification, gave the greatest aging. The undesirable effects of aging were counteracted by the addition of more catalyst to the aged coating just before the coating was used to coat panels. Deactivation of the acidic polymerization catalyst system (based on organotin compounds) is thought to be the cause of this aging. These findings have allowed improvements to be made to the specifications of the pigments and the catalysts used, with respect to the consistent industrial production of low migration coatings.

This paper describes development work undertaken to expand the capabilities of an existing two-dimensional probabilistic modelling approach for assessing dietary exposure to chemicals migrating out of food contact materials. A new three-level hub-coding system has been devised for coding different food groups with regards to their consumption by individuals. The hub codes can be used at three different levels representing a high, medium and low level of aggregation of individual food items. The hub codes were developed because they have a greater relevance to packaging migration than coding used (largely and historically) for nutritional purposes. Also, the hub codes will assist pan-europeanization of the exposure model in the future, when up to 27 or more different food coding systems from 27 European Union Member States will have to be assimilated into the modelling approach. The applicability of the model with the new coding system has been tested by incorporating newly released 2001 UK consumption data. The example used was exposure to a hypothetical migrant from coated metal packaging for foodstuffs. When working at the three hierarchical levels, it was found that the tiered approach gave conservative estimates at the cruder level of refinement and a more realistic assessment was obtained as the refinement progressed. The work overall revealed that changes in eating habits over time had a relatively small impact on estimates of exposure. More important impacts are changes over time in packaging usage, packaging composition and migration levels. For countries like the UK, which has sophisticated food consumption data, it is uncertainties in these other areas that need to be addressed by new data collection.

The migration issues associated with the transfer of chemicals from epoxy-based can coatings into foodstuffs have been studied in recent years. Slip agents are usually used in can coatings for ease of manufacture and to provide required physical resistance properties. Slip agents in epoxy-based can coatings are suspected of not only performing as lubricants but also having effects on the other properties of coatings, including their migration resistance. This article contains information relating to the evaluation of the influence of selected slip agents on the characteristics and properties of specific epoxy-phenolic can coatings. These characteristics and properties include those thermal properties of the coatings that relate to curing characteristics, the surface appearance of cured coating films, the wetting properties that may be associated with migration issues, and the abrasion properties of the cured epoxy-phenolic systems. It was found that three selected slip agents each affected the properties of the specific epoxy-phenolic can coating differently. One of three selected slip agents in particular reduced the interaction (“wettability”) of the specific coatings by test fluids and so may help to reduce chemical migration.

The mass transport process (migration) of a model substance, benzophenone (BZP), from LDPE into selected foodstuffs at three temperatures was studied. A mathematical model based on Fick's Second Law of Diffusion was used to simulate the migration process and a good correlation between experimental and predicted values was found. The acquired results contribute to a better understanding of this phenomenon and the parameters so-derived were incorporated into the migration module of the recently launched FACET tool (Flavourings, Additives and Food Contact Materials Exposure Tool). The migration tests were carried out at different time-temperature conditions, and BZP was extracted from LDPE and analysed by HPLC-DAD. With all data, the parameters for migration modelling (diffusion and partition coefficients) were calculated. Results showed that the diffusion coefficients (within both the polymer and the foodstuff) are greatly affected by the temperature and food's physical state, whereas the partition coefficient was affected significantly only by food characteristics, particularly fat content.

We have used a two-dimensional probabilistic model to estimate the short-term dietary exposure of UK consumers to bisphenol A diglycidyl ether (BADGE) migrating from light metal food packaging. Using three UK National Dietary and Nutrition Surveys comprising 4–7-day dietary surveys for different ages and genders, actual body weights and survey years, a sample representative of the UK population was obtained, comprising around 4200 food items. The packaging type of each food item was assigned by utilizing known packaging type from the database or by sampling from a distribution based upon market share information or expert judgement. For concentration data, we have used published data for foods or food simulants or a combination of both. The probabilistic approach allowed sensitivity analysis to evaluate the relative importance of the input parameters and placed confidence bounds on the outputs to show the effect of the uncertainties. The refined estimates gave an exposure for UK consumers, at the 97.5th percentile level, of 0.41–0.83 μ/kg body weight (bw)/day for the different age ranges and scenarios run. All estimates are well below the new tolerable daily intake value of 150 μ/kg bw/day for BADGE and its two hydrolysed forms, and are also well below the restriction value of 17 μ/kg bw/day for the other regulated BADGE derivatives. The main contributors to exposure are beverages, along with aqueous and acidic foods. This is because of the high consumption of these classes of foodstuffs, even though levels of migration into these foodstuffs and into their appropriate simulants is normally non-detectable. Reducing the non-detectable level six-fold reduced the estimate of exposure by 40–60%. Copyright © 2006 John Wiley & Sons, Ltd.

A two-dimensional probabilistic model was constructed to estimate the short-term dietary exposure of UK consumers to any generalized migrant from coated light metal food packaging. Using three UK National Dietary and Nutrition Surveys (NDNS) comprising 4-7-day dietary surveys for different age and gender groups, actual body weights and survey years, a sample representative of the dietary consumption of the UK population was obtained comprising around 4,200 food items. Interrogation of the raw data showed that the per capita consumption of food and beverage for an adult was 2.9 kg per person day(-1), which is comparable with the US FDA value of 3.0 kg. The packaging type of each food item was assigned from the survey descriptions or by sampling from distributions based upon market share information and expert judgement. Each food item was assigned to the relevant food simulant: A (aqueous), B (acidic) or D (fatty), so that simulant migration data could be used. The exposure model was used to evaluate exposure for a given level of migration and, conversely, the level of migration that could be tolerated whilst keeping within a target threshold exposure level. As examples, migration at 10 microg dm(-2) into fatty foods only resulted in an exposure ranging from 0.06 to 0.22 microg kg(-1) body (actual) weight day(-1) depending on the scenario. The model revealed that if migration from metal coatings was only into fatty foods, migration in the range 1.83-4.95 microg dm(2) (97.5th percentile, depending on the scenario) would give an exposure of less than 1.5 microg per person day(-1). This is a toxicological threshold limit used in the USA. If migration into simulants A and B is also considered to be at the same level as that for simulant D, then the level of migration for the threshold to be reached is, not surprisingly, lower (0.64-0.87 microg dm(-2)) than that if migration were only into fatty foods. In this case, clearly the main contributors to the exposure were foodstuffs represented by simulants A and B because of their importance in the diet. These estimates are based on 4- or 7-day food diaries and chronic exposure over the long-term would be expected to be lower.

A two-dimensional probabilistic model has been developed to estimate the short-term dietary exposure of UK consumers to migrants from food packaging materials. The current EU approach uses a default scenario of assuming that all individuals are 60 kg weight and consume 1 kg of food packaged in the material of interest per day. Using four UK National Dietary and Nutrition Surveys comprising 4-7 day dietary records for different age groups and survey years, a sample representative of the UK population has been obtained consuming around 4200 different food items. Each survey provides records for around 2000 individuals and supplies detailed information on the consumption of food and data on sex, height and socio-economic status which may be used to analyse the exposure of selected groups within the community. As a result we are able to address the variation in consumption of food amongst individuals, and account for actual body weights providing a more accurate representation of the 'true' exposure. The migrants bisphenol A diglycidyl ether (BADGE), di-2-ethylhexyl adipate (DEHA) and styrene were considered as specimen compounds although the methodology employed has the flexibility to adapt to other migrants and packaging types and indeed other food contaminants. Exposure for each individual is estimated by calculating and summing the individual exposure from each item in their diet, and is repeated for all individuals in each survey to produce a distribution of exposures for the population. The packaging type of each food item is assigned by utilizing known packaging types from the database or, by sampling from a distribution based upon market share information. The parameters contributing towards the exposure from a packaged dietary item are migrant concentration and item weight. Distributions are used to represent the inherent variation and uncertainty affecting these parameters. Where data on concentrations for a particular type of food are lacking, expert judgement is used to extrapolate from available data for other food types. The model can also be run using only migration data for food simulants. In this case, concentrations expected for each of the food items are assigned based on the data for the relevant food simulant. The primary outputs of the model are distributions of estimated daily intakes for the selected population. Each distribution gives the variation across the population subject to the uncertain parameters sampled in that iteration of the model. Analysing the ensemble of distributions allows us to obtain the confidence limits around estimates for percentiles due to the uncertainties. The probabilistic approach allows sensitivity analysis to evaluate the relative importance of the input parameters and places confidence bounds on the outputs to show the effect of the uncertainties and the contribution of each food type toward the overall exposure.