Dietary exposure to heterocyclic amines by the Portuguese population: comparison of different exposure assessment methods

To estimate dietary exposure to heterocyclic amines (HAs) in the Portuguese population, three methods of attributing HAs occurrence data to consumed food items were compared. Participants are part of the National Food, Nutrition and Physical Activity Survey (n = 5005, 3-84y). Food consumption was collected through two non-consecutive 24 h dietary recalls or food diaries. HAs occurrence was extracted from the literature. Margins of exposure (MOE) were estimated for different cancer types. Three methods were considered: 1.FoodEx2 Hierarchy model (FH) - random HAs values were attributed to missing data, considering the closest item in the occurrence dataset according to the FoodEx2 hierarchy; 2.FoodEx2 Hierarchy Median model (FHM) - median HAs values of the closest items in the dataset were attributed to missing data; 3.Regression Tree model (RTM) - an RTM was used to predict mean occurrence values of homogeneous food groups based on the FoodEx2 hierarchy. The FH method presents higher dietary exposure for the 95th percentile and lower MOE values, appearing to estimate better. Red and white meat were the main contributors to HAs intake. The Portuguese population's dietary exposure to HAs seems to be safe. However, MOE values for prostate cancer in children may raise concern about potential future adult risk from early-life exposures.