Assessment of microwave cooking on the bioaccessibility of cadmium from various food matrices using an in vitro digestion model.

Bioaccessibility represents the maximum amount of pollutant ingested with food that is available for intestinal absorption. The measurement of bioaccessibility can achieve a more accurate risk assessment. Thus, in this study, the bioaccessibility of raw/microwave-cooked store-bought food including carrot, potato, white radish, lotus root, sweet corn, long grain rice, soybean, fleshy prawn, eastern oyster, kelp, and common carp were investigated by applying an in vitro digestion method. A validated microwave digestion/ICP-MS method was applied for determining the concentration of Cd. In this study, the concentration of Cd ranged 3.7-215.8 μg/kg fw in which carrot contained the lowest Cd while the fleshy prawn contained the highest Cd. There are no statistical differences of Cd content in microwave-cooked food and raw food except potato, lotus root, and eastern oyster. Cd in most of the cooked food materials was less bioaccessible than in raw food except sweet corn, potato, and kelp. The bioaccessibility of Cd was around 100 % in either raw or cooked potatoes. Microwave cooking caused the decreasing of bioaccessibility around 0-68 %, depending on different food matrix. Maximal decreasing of Cd bioaccessibility occurred in common carp. Thus, microwave cooking could be a feasible strategy for decreasing Cd bioaccessibility. In addition, the Cd dissolution in oral, gastric, and small intestine phase was different in different food matrix. For most of the investigated food items, Cd was largely migrated either into the oral phase (carrot, potato, white radish, lotus root, raw soybean, kelp, and common carp) or into the gastric phase (sweet corn, cooked soybean, rice, fleshy prawn, and eastern oyster). Our findings will have significant implications for food processing aiming to decrease the absorption of Cd and risk assessment analysis improvements. Further study is needed to use the animal model to validate these results.