A regression-based model to predict chemical migration from packaging to food

Area of Science:

• Food science
• Analytical chemistry
• Risk assessment

Background:

• Chemical migration from packaging materials poses a risk to human food safety.
• Current process-based migration models (PMM) have limited applicability and low predictive accuracy.
• Accurate prediction of chemical migration is crucial for prioritizing food safety concerns.

Purpose of the Study:

• To develop a more accurate model for predicting chemical migration from packaging to food.
• To statistically relate migration levels to chemical, food, packaging, and experimental properties.
• To improve the prioritization of chemicals in food safety assessments.

Main Methods:

• Developed a linear mixed-effects model (LMM) to predict chemical fraction transferred (F_C).
• Analyzed relationships between F_C and chemical molecular weight, food fat content, and packaging crystallinity.
• Validated the LMM's predictive performance against a PMM using a coefficient of efficiency (CoE).

Main Results:

• A negative correlation was observed between chemical molecular weight and F_C.
• Higher food fat content increased F_C, influenced by the migrant's octanol-water partitioning coefficient.
• Large chemicals (MW > 400 g/mol) migrated more readily from low-crystallinity packaging.
• The LMM demonstrated significantly higher predictive performance (CoE = 0.21) than the PMM (CoE = -5.24).

Conclusions:

• The developed empirical LMM effectively predicts chemical migration from food packaging.
• This model enhances the ability to prioritize chemical contaminants in the absence of direct measurements.
• The LMM offers a more reliable tool for human exposure assessments in food safety.