Uncontrolled migration from lid gaskets into oily foods

In 2004, the systematic non-compliance of lids for glass jars packed with oily foods was detected and compelled authorities to factually lift existing rules. This shows that not all industry performs satisfactory self control. It was a regulatory disaster and might, 4 years later, end in another regulatory disaster, if the major lid producers continue to refuse producing compliant lids.

Development of a regulatory disaster

Infant food in glass jars

More than 10 years ago, a Swedish group detected that the migration of epoxidised soy bean oil (ESBO) into infant food, packed in glass jars, sometimes caused the tolerable daily intake (TDI) to be exceeded (Hammarling et al. 1998). In the eyes of enforcement authorities, this is alarming and calls for a rapid correction. However, later surveys in the UK (MAFF, 1999) and 16 European countries (Fantoni and Simoneau, 2003) showed that industry failed in reducing the ESBO migration. With Directive 2005/79, the EU Commission introduced a legal limit for infant foods of 30 mg/kg, but some months ago, in Switzerland, a product exceeding 60 mg/kg ESBO was detected. This pointed out an astonishing reluctance of this industry to solve the problem. Figure 1 (see page 50) shows the internal face of a lid for infant foods. The gasket consists of a ring into which the jar rim (position shown in blue) forms an indentation to result in a safe seal. In the case of the infant food, “push on twist off” (PT) closures are used in which the gasket also forms the thread on the side wall12,13,15.

ESBO was an issue not only as such, but also since it serves as an HCl scavenger: it reacts with HCl released during curing of the PVC gasket, forming chloroderivatives (Biedermann-Brem et al., 2001). Although this reaction was intended, its products were neither identified nor their safety assured. The internal surface was commonly coated by organosols, stabilised with novolac glycidyl ether (NOGE). NOGE has neither been toxicologically evaluated (despite the presence of glycidyl functions), nor listed by the Resolution AP (96) 5 on coatings of the Council of Europe. Its use as an additive to coatings was factually banned by EU-Directive 2002/16 and then Regulation 1895/2005. Industry indeed changed to the approved bisphenol A glycidyl ether (BADGE)2.

In 2004 it was detected that the blowing agent in the gasket, azodicarbonamide, released potentially carcinogenic semicarbazide (EFSA, 2004, Stadler et al., 2004). Azodicarbonamide was replaced by a concerted action of industry. Furthermore, Elss et al. discovered that zinc and calcium soaps of 2-ethyl hexanoic acid were used as stabilisers. The use of this acid is not authorised because of some teratogenic activity – nor was there any need for it, since normal fatty acids can be used8,16.

General foods in glass jars

If migration of ESBO into infant foods containing barely 1-2% fat or oil reached 135 mg/kg, what would then be the migration into products containing more oil, such as sauces, pestos and products laid in oil (such as vegetables, mushrooms, fish, and cheese)? Indeed, the first survey, carried out in summer 2004, revealed an average migration of around 170 mg/kg and that merely two out of 84 products analysed respected the limit of 60 mg/kg (Fankhauser-Noti et al. 2005). Even for a case-hardened enforcement authority, this was an unusual neglect of compliance. However, it became even worse. Surprisingly, some oily foods contained hardly any ESBO. The analysis of the gasket revealed that there was no ESBO, but phthalates – and their migration was even higher. In June 2005, 158 products with some free oil were collected from the major distributors of the local market, almost exclusively as European products. 147 of the 158 products exceeded the limits (Fankhauser-Noti et al., 2006):

• Of 103 products with ESBO as the major plasticizer in the gasket, 91 exceeded 60 mg/kg; the mean concentration was 216 mg/kg, the maximum 1170 mg/kg.
• 17 products contained diethylhexyl phthalate (DEHP) exceeding 1.5 mg/kg (SML in Directive 2007/19). In 15 of these, the concentrations were above 180 mg/kg; the mean value was 387 mg/kg, the maximum, 825 mg/kg.
• In 25 products, diisononyl phthalate (DINP) and diisodecyl phthalate (DIDP) exceeded the SML of 9 mg/kg. 14 out of the 25, 100 mg/kg had exceeded; maximum, 740 mg/kg.
• In 5 products, the SML of 18 mg/kg for diethylhexyl adipate (DEHA) was exceeded; maximum 180 mg/kg.
• 9 products contained non-authorised plasticisers without a toxicological safety evaluation, approved by the EU authorities.

Late in 2005, products from Asia were analysed. The gaskets almost exclusively contained phthalates or non-approved plasticisers, and migration was often extremely high. The highest DEHP concentration in the packed food was 1130 mg/kg. A child of 20 kg in body weight reaches the tolerable daily intake (TDI, 0.05 mg/kg body weight) with merely 1g of such a product, which means that the content of a single jar was sufficient to bring exposure to the TDI, for almost one year. Another product (with a gasket completely covering the internal lid surface) contained 2030 mg/kg (0.2%!) diisononyl phthalate, which meant half a gram of phthalate in the jar9,10.

Reaction by industry and enforcement

These results showed a complete failure of control for the most part of 20 years. Several industry experts frankly admitted that they knew about the non-compliance of the lids for oily food, but “nobody mentioned anything”. One expert even suspected a secret agreement with the Italian government. Again, another argued that it would have been unreasonable to invest substantially and make his customers unsecure, while all others continued on. Just measuring the migration could have started an avalanche.

If “nobody mentioned anything” meant that the enforcement had failed. In fact, none of the many authorities around the world investigated the migration from lids. This is not surprising, since migration from food contact materials was grossly neglected in general. Still today, in many European countries, there is no relevant capacity in this sector (less than 5 persons engaged in this extremely broad field, even in some large European countries). In terms of amounts, migration from food contact materials exceeds the pesticide residues by 2-3 orders of magnitude and is probably the largest source of food contamination. The majority of the many 10,000 migrants have not been identified, not to speak about toxicological evaluation (Grob et al., 2006). To get these under control, a capacity far beyond that of the (successful) pesticide control would be needed14.

Even after this severe non-compliance was acknowledged, only few enforcement laboratories reacted. One reason was that most were busy with their yearly control plans and therefore unable to give this new issue a higher priority. Others were unable to start analysis in reasonable time, since the full analytical battery includes four fairly demanding methods and it takes more than a year to get them running – often even far more because of time-consuming validation and accreditation procedures. As a consequence, only a minority of the European countries took some measures, and this mostly on the basis of measurements carried out in Zürich. It may be argued that after exposure to such high amounts of plasticisers over many years, it makes little difference whether this would continue for 2-3 more years (until the problem was solved by the activity of other countries), but the event also demonstrated that most enforcement authorities are not really able to take measures within a few months, in order to protect their consumers.

The first legal nightmare

In 2004, it rapidly became apparent that no compliant closures were available for oily foods. This prompted the Swiss authorities to factually lift the legal restrictions for ESBO, in order to provide a way out for the food industry and prevent the use of plasticizers more toxic than ESBO, such as phthalates (BAG, 2004). It was planned for 2 years, but then needed to be prolonged1.

At first, the EU Commission insisted on compliance with the rules, but then had to admit that the damage in enforcing the rules would be in the order of billions of Euros. It reacted by no longer considering gaskets being within the scope of the Plastic Directive 2002/72 (although azodicarbonamide and ESBO for infant foods were regulated by its amendments 2004/1 and 2005/79). This enabled introducing a transition period. After an attempt for a new regulation on lids failed, the lids were re-introduced into the plastic regulation by Directive 2007/19 (to be implemented in July 2008). An additional regulation (“Transitional Regulation” 372/2007) temporarily increased the tolerated migration of some plasticisers to 300 mg/kg or 50 mg/dm2, overruling stricter national legislation. Owing to these regulatory problems, the lid industry was granted 4 years to solve its migration problems, which is unparalleled in the history of food contact materials.

Lid technology

Gasket material in food contact

At this point, lid technology must be introduced to explain the background. Lids are made of coated metal sheet, stamped with a ring-shaped indentation into which the gasket is placed to seal against the jar rim (Figure 2 – see page 52). To form the gasket, plastisol, a viscous liquid, is deposited into the turning lid by a pistol and cured by short heating to some 200°C. Plastisols consist of some 60% PVC powder and 40% additives, largely plasticiser, but also some slip agent (oleamide and erucamide) to facilitate opening of the lid and possibly a PVC stabiliser (Zn/Ca soaps of fatty acids).

Lids typically contain around 1g of plasticised PVC (depending on the lid diameter), of which some 250-400 mg are on the internal side of the seal against the jar rim, i.e. in direct food contact (Fankhauser-Noti and Grob 2006). This means that some 100-160 mg plasticiser is in direct food contact. Complete transfer into a jar content of, e.g., 200 g results in a concentration of 500-800 mg/kg. For relatively small molecular weight plasticisers, such as phthalates, even material from within and outside the seal may migrate, resulting in the extremely high migration sometimes observed10.

Options for improvements

The first attempts to solve the migration problem were aimed at replacing the PVC compound by another material, such as polyethylene. Soon, gaskets in ring shape laid into the lid became available, but they were suitable only for products which required no pasteurisation or sterilisation. Some Asian lids contain a polyethylene disk closing the jar, but limited thermostability again seems to restrict their use. Deposition of a PVC gasket is an approved technology of low costs, but migration must be reduced by a factor of at least 10. The present gaskets are rather large (almost 1 mm thick), but cannot be substantially reduced, owing to the tolerance in the glass rim. Hence plasticisers of lower migration are needed. Among the authorised plasticisers there is only one type with the potential to fulfill this requirement: the polyadipates. Polyadipates are oligomeric, formed by adipic acid interlinked with propane or butane diol and with various end cappings (Biedermann and Grob 2006a). They vary in molecular mass distribution, with some 5-35% material of a molecular mass below 1000 Da (Biedermann and Grob, 2006b). There is an SML of 30 mg/kg (5 mg/dm2), which only applies to the components below 1000 Da; the migration of the higher molecular mass material is only limited by the OML. The higher the molecular mass of a polyadipate, the lower is the migration of the <1000 Da components. Additionally, that of higher mass components is lower, since migration rapidly falls with masses beyond 1000 Da. However, viscosity of the resulting plastisol also increases and renders the deposition of the plastisol a severe problem – see below3,4.

The problem of migration testing

There is a further obstacle in developing compliant lids: the testing procedure to predict (in a short time), the migration, to be expected after several years. EU legislation provides some rules, such as that testing for 10 d at 40°C would reflect long term migration (Directive 82/711/EEC and its two amendments). Other points are not laid down in legislation, such as that, migration testing for lids is performed by turning the jar on its lid, with the stimulant (olive oil for oily foods) laying on the lid. EU legislation also provides that migration limits refer to the food (Regulation 1935/2004), i.e. if migration testing produces results deviating from those in real packs, these are not valid.

It was soon detected that migration of ESBO into oil, at standard testing conditions, may be up to approximately 7 times lower than into real food (Fankhauser-Noti and Grob, 2006). A later investigation (Biedermann et al., 2008) showed that the increase of temperature to 40°C only weakly increases migration and that 10 d/40°C merely simulated 24 and 39 d at 20°C for ESBO and polyadipates, respectively. Even worse: upon initial heating to simulate pasteurisation or sterilisation, ESBO is rapidly extracted, resulting in a surface layer of deplasticised PVC, hindering further migration for several weeks, i.e. testing at 10 d at 40°C could just as well be omitted6,11.

Finally, the large amount of oil in contact with the gasket neglects a relevant effect of migration: saturation of the oil at equilibrium. In reality, migration occurs stepwise. With the jar in the upright position, merely some 30 mg oil is in contact with the about 300 mg gasket, containing some 120 mg plasticiser. After a few days, migration virtually stops, as a partitioning equilibrium between the PVC and the oil is reached (which strongly differs between various plasticisers). Upon moving the jar, the oil is likely to be exchanged by fresh oil from the jar content, which initiates another migration step. This tells us that the migration depends on the history of the jar: the age may be less important than the number of times the jar was shaken.

The solution

The first polyadipate lids

In 2004, a small lid producer (IN.CAM, Campegine, Italy) started investigating polyadipates. In Autumn 2004, it was concluded that only the use of polyadipates had a chance to meet the limits and no other plasticisers could be added to reduce its high viscosity, as the OML would be exceeded otherwise. Even worse: this polyadipate had to be of a high molecular mass with a proportion of components <1000 Da below 10%. Such plastisols could not be applied with standard machinery: they cannot be sucked through the system and they are not properly separated from the pistol. Hence the machinery was rebuilt. In early summer 2005 (and after solving some further problems), the production started. In the absence of a reliable test, numerous real products were analysed, 11 of which were kept in the laboratory and regularly shaken for more than 2 years. No exceedance of the legal limits was detected (Biedermann et al, 2006), but sometimes the migration was close to the limit, confirming that only rather high molecular mass polyadipates assure compliance5.

Other producers

The other producers, among which the large ones, kept their customers suspended by negative rumors about the polyadipate lids of the competitor, claiming that no complying lids were available (although more than 100 Mio have been in use in the mean time) and continuous promises that their compliant lid would be more or less ready.

As far as visibility in the market place, several manufacturers use plastisols of a viscosity adjusted to their machinery, mixing rather low molecular mass polyadipates with low viscosity plasticisers, such as dibutyl sebacate, ESBO or acetylated monoglycerides. It is difficult to predict whether these lids will comply with the legal limits over up to several years, but the experience made by the IN.CAM lids casts doubts, and not even their producers claim compliance. It seems possible that their efforts limited to what is achievable without adjustment of the machinery might fail. IN.CAM is too small a producer to supply the lids required for the European production of oily foods (at least 1000 Mio/y). Given the short time left up until July 2008, when the conventional limits must be respected again, it seems an obvious approach that this technology is also used by other lid manufacturers.

Second regulatory nightmare

Towards the end of 2007, the lid manufacturers wanted the food producers to intervene for a prolongation of the transition regulation by several more years or even being exempted from the OML. The Commission is in a difficult position: on the one hand, the resistance of the main lid producers to take over a technology approved over about 2.5 years, seems unacceptable; on the other, with the distress of the (mostly small) food producers, the lid producers keep a strong pledge in their hands. Lifting rules due to broad non-compliance was a regulatory nightmare. To buckle to an industry, not really willing to take its responsibility and prolong the transition regulation, would be another one.

At the time this text was written (December 2007), it was open to how the story would end. It seemed clear, however, that the Commission would have little room not to yield when the lid producers succeeded in having another 2 months to pass without progress, as it will soon be too late then. They may say that in a free market nobody can force them to produce compliant lids.

What we should learn

The quite systematic non-compliance of lids for oily products for more than 15 years is unparalleled. As long as the lids were used for jam, pickles or other foods free of fat, the migration from the gaskets was negligible, but industry should not have accepted the use of these lids for oily products without investigating the predictably high migration.

As the issue of the lids was not the only one of failing compliance in the sector of the food contact materials, it seems necessary to draw more general conclusions on what should be learned to prevent such incidences in the future.

Failing control

The first conclusion is that control failed. Industry experts knew about the problem for a long time, but did not react as “nobody mentioned anything”. Even after the high migration of ESBO into infant foods became known, little was done to solve it. The use of polyadipates to reduce migration from plasticised PVC (e.g. cling films) was known for a long time (Castle et al. 1988). Enforcement authorities failed in that they did not detect the problem. This suggests that the control activities are too heavily driven by control plans, but also that public resources in this field are insufficient. According to the ‘common opinion’, the high migration of phthalates was unknown to the authorities, to the extent that the Commission classified the introduction of SMLs for phthalates as low priority. Furthermore, no relevant amounts of phthalates would be used anymore in food contact7.

Lifting rules: precedence

Lifting legal rules because of broad non-compliance is a regulatory disaster. Enforcement authorities tend to be strict with the small producers: marketing of non-compliant products is stopped and the producer may even be fined. Now, as non-compliance of lids was systematic, the culprits were not punished, received four years to solve the problem and in the end, they ask for even more time or hope to permanently adjust legislation to their favour. How should enforcement perform its job under such circumstances?

It was not an option to enforce rules in 2004, as the consequences for the food industry and farmers would have been severe. Therefore, the main conclusion is that such a case must be avoided, in order to begin with by tighter control. It is the job of the industry to check compliance. There might be a broad agreement on this, but at the bottom of the problem lies the lack of an effective control system.

Improved definition of responsibilities and control through documentation

The production of packed food typically involves a long chain of manufacturing stages and companies. Control should primarily occur within industry, but the confidentiality required by the producers hinders this process. Lid producers commonly refuse to inform the food industry about the plasticisers used. Many lids were accompanied by a certificate, but practically all these compliance declarations were unwarranted, showing the need for controlled SD. Often, they circumvented clear statements, e.g. by using negative statement, like “ESBO-free”, which covered the fact that the main plasticiser was, e.g., a phthalate. Other certificates declared “free of DEHP” when the plasticiser used was another phthalate. There were also lid producers using a plastisol of which they did not know the composition, i.e. they did not know the plasticisers they used.

New legislation (“GMP Regulation” 2023/2006 and the 4th amendment of the Plastic Directive 2007/19 specifying compliance declarations) should improve on this: it still protects the confidentiality of the producers, but clarifies responsibilities. At some point in the future, lid producers can declare compliance without naming the substances they used, but this means that they would accept responsibility and liability. They can only avoid this by delegating compliance work to the next stage, e.g. the food producer, instructing the latter about the work to be performed.

Enforcement authorities received a key role in this system, as no one else has access to supporting documentation and GMP documentation, justifying compliance declarations. They should build up control to the extent that industry can be confident in the compliance declarations they receive and that the products they buy in were manufactured, according to GMP defined as the process to assure compliance with legislation. This is a great challenge – perhaps an illusion: present enforcement is unable to meet this goal.

References

1. BAG (2004). Bundesamt für Gesundheit, Weisung 3, ESBO in Lebensmitteln. Bern, Nov. 5, 2004, http://www.bag.admin.ch/ themen/ernaehrung/00171/00460/02376/index.html?lang=de
2. Biedermann-Brem, S., Grob, K. and Biedermann, M. (2001). Analysis of reaction products (e.g. chlorohydrins) of ESBO in poly(vinyl chloride) type polymers and coatings. Mitteilungen aus Lebensmitteluntersuchung und Hygiene, 92, 515-534.
3. Biedermann, M. and Grob, K. (2006a). GC-MS characterization of oligomers in polyadipates used as plasticizers for PVC in food contact. Packaging Technology and Science, 19, 159-178.
4. Biedermann, M. and Grob, K. (2006b). Polyadipates used as plasticizers in food contact: Fraction below 1000 Da determined by size exclusion chromatography with evaporative light scattering detection and segmental response linearization or UV detection. Journal of Separation Science, 29, 114-122.
5. Biedermann, M., Fiselier, K., Marmiroli, G., Avanzini, G., Rutschmann, E., Pfenninger, S. and Grob, K. (2006). Migration from the gaskets of lids into oily foods: first results on polyadipates. European Food Research and Technology Technol (on line)
6. Biedermann, M., Fiselier, K. and Grob, K. Testing migration from the PVC gaskets in metal closures into oily foods. Trends in Food Science and Technology (on line), 2008.
7. Castle, L., Mercer, A.J. and Gilbert, J. (1988). Migration from plasticized films into food. Use of polymeric plasticizers and lower levels of di-(2-ethylhexyl) adipate plasticizer in PVC films to reduce migration into foods. Food Additives and Contaminants, 5, 277–282.
8. Elss, S., Grünewald, E., Richling, E., and Schreier, P. (2004). Occurrence of 2-ethylhexanoic acid in foods packed in glass jars. Food Additives and Contaminants, 21, 811-814. EFSA, European Food Safety Agency, (2004) http://www.efsa.eu.int/ press_room/questions_and_answers/ 44_en.html
9. Fankhauser-Noti, A., Fiselier, K., Biedermann, S., Biedermann, M., Grob, K., Armellini, F., Rieger, K., and Skjevrak, I. (2005). Epoxidized Soy Bean Oil (ESBO) migrating from the gaskets of lids into food packed in glass jars. European Food Research and Technology, 221, 416-422.
10. Fankhauser-Noti, A., Biedermann-Brem, S., and Grob, K. (2006). PVC plasticizers/additives migrating from the gaskets of metal closures into oily food: Swiss market survey June 2005. European Food Research and Technology, 223, 447-453.
11. Fankhauser-Noti, A. and Grob, K. (2006). Migration of plasticizers from plastisol gaskets of lids for glass jars into oily foods: Amount of gasket material in food contact, proportion of plasticizer migrating into food and compliance testing by simulation. Trends in Food Science and Technology, 17, 105-112.
12. Fantoni, L., and Simoneau, C. (2003). European survey of contamination of homogenized baby food by epoxidized soybean oil migration from plasticized PVC gaskets. Food Additives and Contaminants, 20, 1087-1096.
13. Hammarling, L., Gustavsson, H., Svensson, K., Karlsson, S., and Oskarson, A. (1998). Migration of epoxidized soya bean oil from plasticized PVC gaskets into baby food. Food Additives and Contaminants, 15, 203-208.
14. Grob, K., Biedermann, M., Scherbaum, E., Roth, M. and Rieger, K. (2006). Food contamination with organic materials in perspective: packaging materials as the largest and least controlled source? A view focusing on the European situation. Critical Reviews in Food Science and Nutrition, 46, 529-536.
15. MAFF (1999) Food surveillance Information Sheet No. 196, http://www.foodstandards.gov.uk/maff/archive/food/infsheet/1999/no186/186esbo.htm
16. Stadler, R.H., Mottier, P., Guy, P., Gremaud, E., Varga, N., Lalljie, S., Whitaker, R., Kintscher, J., Dudler, V., Read, W.A. and Castle, L. (2004). Semicarbazide is a minor thermal decomposition product of azodicarbonamide used in the gaskets of certain food jars. Analyst, 129, 276-281.