Food Standards Agency UK - Chemical Migration from can coatings into food-terephthalic and isophthalic acid (Number 07/00)

• Terephthalic and isophthalic acids migrated into food from the coatings of 12 out of 28 samples of cans. Terephthalic acid was found at up to 0.8 mg/kg of food in 10 samples. Isophthalic acid was detected in 4 samples, at less than 0.7 mg/kg of food.
• The Committee on Toxicity advises that these levels of terephthalic and isophthalic acids are not of concern for public health based on available information, and recommends further toxicological work on these substances to see if they have endocrine disruptor activity.
• The Food Standards Agency is taking action to ensure that consumers are protected by directing industry to carry out this further toxicological work and to keep migration of these substances from can coatings into food to a minimum. The Food Standards Agency is also encouraging the European Commission to make proposals for a specific Directive on chemical migration from can coatings into food.

The Food Standards Agency carries out surveillance of chemical migration from packaging into food to ensure that consumers are adequately protected from ingesting unsafe amounts of chemicals that contaminate food in this way. Recent work by UK Government found evidence for the migration from can coatings into food of one chemical (BADGE) but not of two others, epichlorohydrin and formaldehyde.^1-3 The survey reported here was on terephthalic acid (TA) and isophthalic acid (IA).

Industry has carried out work on the effective use of these substances in can coatings. TA and IA may be present in can coatings either because they are in resins that are used to make coatings, or because of cross-contamination of coating materials. Liquid coatings that may contain TA and IA are applied to the internal surface of some metal cans, usually by a roller. The solvent in the coating is evaporated at a moderately low temperature, before the coating is cured at a temperature of approximately 200°C. The coated sheet or coil of metal is then used to make the can. The use of polyethylene terephthalate - PET - in stripes to cover the welded joint in the side of cans, could lead to migration of TA (IA is not used to make PET). These side stripes are made by applying powdered PET, by a roller or an electrostatic process, to the welded joint on the side of the can. The can is then heated to form the side stripe. After the coating or coating and side stripe have been applied the can is then filled with food or beverage, sealed and the contents sterilised, typically at 121°C for up to an hour with the product either static or rotated.

Chemical migration from can coatings into food is controlled in Great Britain under the general provisions of the Materials and Articles in Contact with Food Regulations 1987 and of the Food Safety Act 1990. Similar controls apply in Northern Ireland. Work continues at the Council of Europe to develop numerical standards of chemical migration from can coatings.

Sampling
Sampling was carried out in two phases. The first phase was to identify which products were sold in cans with coatings made from polyester resin. In the second phase samples of products packaged in cans with such coatings were purchased and the can contents analysed for the presence of TA and IA. Analysing the contents of these cans would allow contamination of the contents with TA or IA to be detected.

For the first phase, 54 samples of canned foods were purchased in November and December 1998 from supermarkets and other retail outlets in the Leatherhead and Reigate area, so that their coatings could be identified. The use of infrared (IR) spectroscopy allowed identification of polyester coatings, in particular PET, in some cans.

For the remaining cans it was necessary to establish the presence of polyester resin by hydrolysing the coating to TA or IA, followed by derivatization of the acids to their respective dimethyl esters and identification of the esters by gas chromatography-mass spectrometry (GC-MS). By using these two methods of identification 28 cans were found to contain coatings made from polyester resins.

In the second phase, samples of the products in the 28 cans shown to have been coated with polyester resin were purchased in triplicate, for the analysis of the contents. These samples were purchased from similar outlets to those in phase one, in south-east England between December 1998 and February 1999. The samples are listed in Table 1.

All samples in the first and second phases were stored unopened and away from the laboratory at ambient temperature until required for analysis. In the second phase the entire contents of each of the canned samples was homogenised and a 10 g portion analysed immediately.

Analysis
Food samples were analysed by GC-MS using deuterated TA as internal standard and derivatization to form the dimethyl esters of TA and IA prior to analysis.⁴ The contents of each can were homogenised and 0.1 ml hydrochloric acid, as well as 100 micrograms of the deuterated internal standard, was added to 10 g portions of the homogenised food sample. After extraction with methanol and drying over anhydrous sodium sulphate, the extract was reduced to 4 ml and derivatized with boron trifluoride diethyl etherate. For foods containing more than 5 per cent fat, the samples were additionally extracted with heptane (in which TA and IA are insoluble), to remove the fat, prior to derivatization. The derivatized extracts were then shaken with sodium chloride and further extracted with heptane. The heptane extracts were analysed by GC-MS. For analysis, chromatographic peak areas were compared for TA and IA at m/z 163 and m/z 194 and for the deuterated terephthalic acid at m/z 167 and m/z 198. Calibration graphs were constructed based on the response ratios for TA and IA respectively at m/z 163 to that at m/z 167 for the deuterated internal standard. All calibration graphs were rectilinear with correlation coefficients of 0.997 or better.

For both TA and IA, the limit of detection was estimated to be 0.2 mg/kg from calibration graph confidence intervals, corresponding to three times the standard deviation of the readings for blank samples, and by visual inspection of GC-MS traces. The limit of quantification was calculated to be 0.7 mg/kg corresponding to ten times the standard deviation for blank samples.

Analytical Quality Assurance
Before samples were analysed, PIRA International were required to provide evidence that results obtained by them would be of acceptable accuracy. Samples of two canned foods, a soup and a fruit cocktail in syrup, were 'spiked' with TA and IA at two undisclosed levels in each case by the Central Science Laboratory (CSL), Norwich. Additionally an undisclosed blank sample of each food had to be tested. PIRA successfully identified the blank samples and also quantified satisfactorily the 'spiked' samples. The levels, found by PIRA, were 93 - 98 per cent of those added by CSL. This performance was considered to be within an acceptable margin of analytical variation.

To help ensure analytical quality assurance during the survey, samples were analysed in batches. An analytical batch typically consisted of five duplicate analytical samples, a reagent blank and a control sample run in duplicate. Control samples were prepared by fortifying the entire contents of food cans, known not to contain TA and IA, with 8 mg/kg of TA and IA using a set of standards prepared by a different analyst. The recoveries of TA and IA obtained from the control samples analysed in each run were in the range 94 -106 per cent. These data were used to derive recovery correction factors for TA and IA that were applied to the results for the samples in the same run. Reagent blanks did not contain detectable amounts of either TA or IA, indicating that there was no extraneous source of TA or IA.

The precision of the method was plus or minus 0.11 mg/kg from the combined standard deviation of the samples analysed in duplicate at a mean level of 0.43 mg/kg. The analytical uncertainty, based on two standard deviations, was therefore plus or minus 0.22 mg/kg. In two cases (samples 6 and 12 in Table 1), where TA was found in quantifiable amounts, a second can was opened and the contents analysed in duplicate by a second analyst.

Reporting
Brand names are reported as this survey was carried out in accordance with guidelines for reporting survey results published in the Food Safety Information Bulletin in September 1997. The absence of a particular brand from Table 1 means only that the brand was not included in the survey.

The results of analysis of the can coatings and contents are given in Table 1. TA was found in 3 out of 28 samples, at or just above the limit of quantification and in 7 samples at levels between the limit of detection and limit of quantification. IA was detected in 4 out of 28 samples at levels between the limit of detection and limit of quantification.

Where TA and IA were found in canned food this was consistent with the type of resin used in the can coating (Table 1). TA is a starting material for PET. TA and IA are used to make other polyester coatings. Seven of the ten food samples containing TA were from cans coated with PET. The other three samples were from cans which contained a PET 'side stripe'. All the samples containing IA were from cans with internal surfaces coated with a polyester resin other than PET.

This survey provided evidence of migration of both TA and IA. The views of the Committee on Toxicity of Chemicals in Food, Consumer Products and the Environment (COT) were sought on the health implications of this survey. The Committee advises that the levels of TA and IA found in foods in this survey were not of concern for public health based on available information, and recommends further toxicological work on these substances (Annex 1).

The Food Standards Agency accepts the views of the COT on this survey. In the light of their recommendation that studies on endocrine disruption be carried out for TA and IA (Annex 1) the Food Standards Agency considers that it would be prudent to keep the migration of these substances into food to a minimum. The Food Standards Agency is therefore taking action to ensure that consumers are protected. The Agency is directing industry to carry out this further toxicological work recommended by the COT and to keep migration of these substances from can coatings into food to a minimum. The Food Standards Agency is also encouraging the European Commission to make proposals for a specific Directive on chemical migration from can coatings into food. There is no specific EU legislation on this at present.

1. Survey of BADGE epoxy monomer in canned foods. Food Surveillance Information Sheet No. 125, 1997.
2. Survey of chemical migration from can coatings into food and beverages 1. Formaldehyde. Food Surveillance Information Sheet No. 157, 1998.
3. Survey of chemical migration from can coatings into food and beverages 2. Epichlorohydrin. Food Surveillance Information Sheet No. 170, 1999.
4. Castle, L., Mayo, A., Crews, C. and Gilbert, J. Migration of PET oligomers from PET plastics into foods during microwave and conventional cooking and into bottled beverages. Journal of Food Protection, 1989, 52, 337-342.

Mr Patrice Mongelard
Food Standards Agency
Chemical Safety and Toxicology Division

Email: patrice.mongelard@foodstandards.gsi.gov.uk

A copy of the full report of this survey has been placed in the Food Standards Agency library. If you wish to consult a copy please contact the library for an appointment giving at least 24 hours notice or, alternatively, copies can be obtained from the library: a charge will be made to cover photocopying and postage.