Migration of melamine from can coatings cross-linked with melamine-based resins, into food simulants and foods.

Resins based on melamine-formaldehyde and related analogues such as methylolated melamine are used to cross-link coatings used inside food cans and on the metal closures of glass jars. Thirteen commercially coated cans and closures representing 80% of the European market were tested using simulants under realistic industrial heat-processing conditions for canned and jarred foods. The food simulants and the retort conditions used were 3% acetic acid for 1 h at 100 °C and 10% ethanol for 1 h at 130 °C. The highest migration level seen for melamine into simulant was 332 µg kg⁻¹. There was no detectable migration of the melamine analogues cyanuric acid (<1 µg kg⁻¹) or ammelide (<5 µg kg⁻¹) from any sample. Twelve of the thirteen samples released no detectable ammeline (<5 µg kg⁻¹) but the coating giving the highest release of melamine did also release ammeline at 8 µg kg⁻¹ with the higher of the two process temperatures used. Migration experiments into food simulant and foods themselves were then conducted using two experimental coatings made using amino-based cross-linking resins. Coated metal panels were exposed to the food simulant 10% (v/v) aqueous ethanol and to three foodstuffs under a range of time and temperature conditions both in the laboratory and in a commercial food canning facility using proprietary time and temperature conditions. The highest migration into a food was 152 µg kg⁻¹ from the first coating processed for a long time at a moderate sterilisation temperature. The highest migration into simulant was also from this coating at 220 µg kg⁻¹ when processed at 134 °C for 60 min, dropping to 190 µg k⁻¹ when processed at 123 °C for 70 min. Migration from the second coating was quite uniformly two to three times lower under all tests. These migration results were significantly higher than the levels of melamine extractable using 95% ethanol at room temperature. The experiments show that commercial canning and retorting can be mimicked in an acceptable way using laboratory tests with an autoclave or a simple pressure cooker. The results overall show there is hydrolytic degradation of the melamine cross-linked resins to release additional melamine. There is a strong influence of the temperature of heat treatment applied with foods or simulants but only a minor influence of time of heating and only a minor influence, if any, of food/simulant acidity.